Vcsel with anisotropic curved mirror

Abstract

This invention provides a novel and innovative VCSEL device and system. In one example, a VCSEL device includes a cavity mirror having a curved mirror surface of a VCSEL and an anisotropic radius of curvature (ROC) of the curved mirror surface, wherein the ROC includes four directions: +x, +y, -x, and -y. The ROC in at least one direction is within a range greater than the cavity length of the VCSEL and less than a predefined ROC value for a standard beam width (ROCUL), and the ROC in at least one of the other directions is outside this range.

Description

Vertical cavity surface emission laser with anisotropic curved mirror This invention generally relates to light-emitting elements and devices comprising a vertical-cavity surface-emitting laser (VCSEL). One type of VCSEL described herein may be a vertical-cavity surface-emitting laser or a vertical external-cavity surface-emitting laser. In a VCSEL, laser light is allowed to resonate between two light-reflecting layers, typically distributed Bragg reflector (DBR) layers, which may be referred to as mirror layers. The VCSEL structure, including the DBR layers, cavity length, aperture size, and shape, affects the shape and properties of the emitted laser beam. This invention provides novel and innovative VCSEL devices and systems. In one example, a VCSEL endoscope includes a curved mirror of a VCSEL and an anisotropic radius of curvature (ROC) of the curved mirror, wherein the ROC includes four directions: +x, +y, -x, and -y. The ROC in at least one direction is within a range greater than the cavity length of the VCSEL and less than a predefined ROC value for a standard beam width (ROCUL), and the ROC in at least one of the other directions is outside this range. Additional features and advantages of the disclosed methods and apparatus are described in the following embodiments and figures and will become apparent from the following embodiments and figures. The features and advantages described herein are not exhaustive, and specifically, many additional features and advantages will become apparent to those skilled in the art from the figures and description. Furthermore, it should be noted that the language used in the specification is chosen primarily for readability and instructional purposes and does not limit the scope of the invention. VCSELs can have one or more curved mirrors or DBRs that can be configured to focus or adjust the properties of the laser beam. In VCSELs with curved mirrors, the beam properties are largely influenced by the shape of the mirrors, including the radius of curvature (ROC). A mirror can also refer to a lens. Typically, mirrors in a VCSEL are circularly symmetrical; however, anisotropy within the mirrors can have a beneficial effect on controlling the VCSEL's modes, such as achieving stable polarization. In recent years, VCSELs with curved mirrors have demonstrated excellent performance using group III nitride material systems. Furthermore, the ability to control the shape of a lateral mode by changing mirror parameters has been demonstrated using mirror ROC control of the lateral mode. However, these devices lack polarization control due to the symmetry of the device (such as a circularly symmetrical curved mirror) and the symmetry of the crystal orientation. This lack of polarization preference can lead to several problems, such as: polarization switching of devices with lifetime and current; a low polarization ratio; a polarization ratio that varies with lifetime and current; and random polarization direction requiring individual measurements for each device to be evaluated. Although there are certain limitations to working with VCSELs with curved mirrors, there are three main methods by which polarization can generally be controlled in VCSELs: using materials with polarization-dependent gain, using a grating to induce a polarization dependence on reflection, or breaking the symmetry of cavity net gain / loss, such as etching a mirror with an elliptical pattern. To achieve polarization dependence of gain, this is typically achieved through anisotropic strain in the active region. This strain can be structurally induced or inherent due to crystal orientation. Inducing structural strain can make processing very complex and often requires a complete redesign of the process. For materials such as group III nitrides, polarization can be induced by using non-polar orientation. In this polar structure, anisotropic biaxial strain in the half-polar and non-polar orientation wells causes a splitting of the valence band, leading to recombination preferentially producing a single polarization. However, this method typically requires the use of non-polar oriented substrates. These substrates are usually much more expensive and have lower quality than conventional c-planar oriented substrates. Gratings have become a typical method for controlling the polarization of arsenide-based devices. Simply etching lines on top of a DBR or depositing dielectric strips makes specular reflectivity polarization-dependent, thus promoting a higher reflectivity in polarized lasers. However, reflectivity is heavily influenced by the grating size and therefore requires fine-tuned processing. A more serious problem is that the grating size decreases as the designed wavelength decreases. This means that for VCSELs emitting in the visible spectrum, e-beam lithography is typically required to form the grating, significantly increasing fabrication complexity and time. Finally, the mirror or structure can be modified to introduce anisotropic losses into the cavity. This is achieved by etching anisotropic shapes (such as ellipses) into the mirror structure. Although this is a simple variation or addition to a standard manufacturing process, it is quite sensitive to misalignment and can increase the losses visible from the desired laser pattern, thereby degrading device performance. The present invention addresses the problem discussed above by providing the use of a VCSEL cavity with improved beam properties by using a curved mirror with different radii of curvature along different axes, the axial relationship being defined by the electrical aperture parameters along those axes. Typically, as discussed in this invention, the top of a lens or curved mirror is defined as the center of mass of the optical mode intensity of the laser mode or the "highest" point of the lens, or can be considered as a point within 2 µm of the center of mass of the electrical aperture. When referring to the axial direction as discussed in this invention, the axis is designed such that the minimum ROC of the lens is aligned with the x-axis or y-axis. In other words, the direction having the highest second derivative of the lens profile captured at the "top of the lens" is defined as parallel to the x-axis or y-axis. Therefore, the direction with the maximum curvature is set to x or y. For example, the height z(x, y) of a lens: Figure 1 shows a curved mirror 102 with a different radius of curvature (ROC) 106 and a corresponding electrical aperture 104. In this case, the ROC... -x The minimum ROC 106 of the lens is parallel to the x-axis. When defining the sign of ROC 106, a positive value corresponds to a convex lens 102a, an infinite or near-infinite value refers to a plane mirror 102b, and a negative value refers to a concave lens 102c, as illustrated in Figure 2. It should be understood that a curved mirror or lens (e.g., a DBR) with ROC 106 can have multiple layers (not shown in Figure 2), as illustrated and discussed below. Also, as discussed herein, unless the shape of the curved mirror is explicitly referred to as convex, nearly flat, or concave, curved lens 102a, 102b, or 102c can generally refer to curved mirror 102. Figure 2 illustrates different notation conventions for ROC 106 associated with a cavity 208 configured to emit a laser beam from a VCSEL and a planar reflector or mirror 210 (e.g., DBR), as described in this invention. The curved mirror will have an ROC value {ROC 106} as illustrated with reference to Figure 1. x ROC -x ROC y ROC -y This allows the sign of the curvature of a mirror to differ in some directions. For example, the ROC 106 of a curved mirror may be convex in some directions and concave or flat in others (e.g., 102a in the x direction, 102b or 102c in the y direction). When defining the difference in ROC values, there exists a threshold value referred to herein as ROCUL. In an exemplary embodiment, ROCUL is ROC 106, which is 1 / e in a particular direction. 2 The beam width at the vertical position defined by the aperture is equal to the electrical aperture diameter of the device in that direction. For a device with an oxide-defined aperture 104, this would correspond to the beam waist in cavity 208 at the level of oxide 212. For a device with a radius {r x , r -x ,r y , r -y The aperture is 104, and there exists a corresponding ROCUL {ROCUL} in each direction. x ROCUL -x ROCUL y ROCUL -y In mathematics, in this example, for the intensity distribution of a Gaussian beam, g(x) in the direction x is given by: 1 / e 2 Strength at The following is given. Therefore, when At that time, a roculsive event occurs. For a Gaussian beam in a cavity 208 having a curved mirror 102 and a plane mirror 210, the beam width at the plane mirror 210 is... By means of the following: Where λ represents the wavelength of the optical mode, n represents the refractive index, L represents the cavity length (between 210 and 102), and R represents the radius of curvature of the curved mirror 102. Assuming the aperture 104 is defined near the plane mirror 210, these two equations can be combined to give the following definition of ROCUL in a direction x: In another exemplary embodiment, ROCUL is a ROC 106, where half the beam width in a particular direction is equal to the electrical aperture diameter of the device in that direction at a position defined by the aperture. In this exemplary embodiment, the 1 / e intensity is... The following is given, and this different relationship runs through the mathematical relationships described above. Any other suitable value of ROCUL can be used, for example, for 1 / e to 1 / e... 2 The beam width is within the range of [the specified beam width]. Therefore, it should be understood that the above relationship can vary depending on the ROCUL set for a particular implementation, as the threshold value set for ROCUL can depend on the nearby structure. Furthermore, ROCUL is set such that the ROC anisotropy is sufficiently different in different directions to produce the desired effect (e.g., polarization control) for a given structure. As discussed in the examples below, the directions of ROC 106 can be described using the following terms: the "S" direction corresponds to the direction in which loss is introduced into the mode, and the values ​​in this direction belong to the set S; and the "Q" direction corresponds to the direction in which the mirror intends to function normally, and the values ​​in this direction belong to the set Q. Those familiar with this technique will understand that aperture 104 is defined by lateral current limitation, and the size of aperture 104 is determined by the effective width of the current flow at the layer or gain region where the most losses are introduced. Typically, the layer with the most losses for a mode is the region of action; however, this is not always the case. Metal contacts, implants, or oxide layers are all other instances that can be used to define the aperture 104. For specific examples of mirror shape variations described below, unless otherwise specified, the given values ​​may generally correspond to the VCSEL cavity 208 described herein, although this description is not intended to limit the scope of the invention but rather to provide a suitable and typical exemplary embodiment. The growth substrate is composed of GaN, and the substrate orientation corresponds to a c-direction with a cutting error within ±3°. The epi structure is composed of an n-type material, followed by an active region with a peak gain of 3±1 quantum wells at 450 nm. A p-type layer exists above the active region. A current aperture 104 is formed into the p-type layer containing B or Al ions by ion implantation. To define the shape and size of the aperture 104, photolithography can be used to define a pattern of an implantation barrier layer, such as a thick photoresist or metal. This barrier material can be placed around the aperture location to prevent implantation in that area and provide electrical isolation. For embodiments where the aperture size and shape are not specified, a circular aperture 104 with a diameter of 4 μm can be used. Next, a current diffusion layer is grown or deposited using another n-type region having a tunneling contact or using a transparent conductive oxide (such as ITO). Metal contacts are deposited on the current diffuser and down to the n-type layer to achieve electrical injection through aperture 104. To form the optical cavity 208, the substrate is thinned, and a plane mirror 210 is deposited on one side while a curved mirror 102 is placed on the other side. For example, the total cavity length can be 20 μm. Mirror 102 can be made of Ta, which has a thickness of ¼ of the optical wavelength in a material corresponding to a free space wavelength of 450 nm. 2O 5 and SiO It consists of two alternating layers. These layers form a dielectric distributed Bragg reflector. In one example, the curved lens or mirror 102 is formed by spin-coating a photoactive polymer. The polymer is then photo-excited and developed through a mask to form a designed pattern depending on the desired lens shape. The top of mirror 102 is aligned to the center of aperture 104. The sample is then heated to cause the polymer to reflow and form a curved shape (e.g., 102a). This curved pattern is then transferred to the GaN surface beneath the polymer by dry etching. Finally, a dielectric layer is deposited on the curved surface to form mirror 102. The different curvatures of mirrors 102a, 102b, and 102c can be achieved by changing the shape and fluid properties (such as viscosity or surface tension) of the polymer before reflow. Therefore, selecting a suitable resin is often an important parameter in determining optimal manufacturing conditions. As described herein, the long side of a rectangular photoresist pattern will generally refer to its length, and the short side to the width of the pattern. For the case of an array described herein, the same general procedure is used, although multiple apertures are defined for each mirror 102. Figure 3 illustrates a general procedure 300 for forming a curved mirror 102 on a substrate 302 by using photoresist 304 to form a shape and etching according to a typical procedure to produce one of the desired shapes of the curved mirror 102, as discussed above. In this example illustrated in Figure 3, mirror 102 has a near-infinite ROC in the middle and an ROC of approximately 70 μm at the ends. Figure 4 illustrates a schematic cross-section of a VCSEL device 400, wherein the curved mirror 102a has a symmetrical ROC 106 defined on Q in the illustrated cross-section, indicating a direction not intended to introduce additional losses. This direction will be parallel to the y-axis for use in the following description of embodiments of the invention. An approximate threshold value of the edge of the resonant mode shape is depicted by lines 402s and 402b (e.g., half-power beam width). In the illustrated example of device 400, ROCUL is ~250 μm. In an exemplary embodiment, cavity 208 may be a complete cavity structure of a VCSEL or an external cavity structure. In an exemplary embodiment, laser cavity 208 is made of GaN, InP, GaAs, or another semiconductor material. It should be understood that similar reference element symbols may be omitted where interpretation or description of the structure is unnecessary. Figure 5A is a schematic cross-section of a VCSEL device 500 in a second direction (e.g., the x-direction). Figure 5B is a schematic cross-section of the VCSEL device of Figure 5A, presented as a plan view. In a first exemplary embodiment shown in Figures 5A and 5B, the mirror 102b has a different ROC 106 along the x-direction and a constant ROC 106 along the y-direction, wherein the outer shape 510 of the mirror shape is longer in the x-direction (e.g., 150 μm) than in the y-direction. The ROC 106 along the x-direction approaches infinity in the middle of the mirror 102b at the location of the aperture 104. In other words, the mirror 102b is substantially flat along the x-direction and curved along the y-direction, as if described in three dimensions by a cylindrical shape. This curvature can be achieved by creating a rectangular pattern of polymer before reflow so that the long direction has sufficient length relative to the short side of the lens to flatten along the top. For example, using an n-type photoresist with a specific viscosity and height after spin, a mirror 102b with a total length of 150 μm and a width of 50 μm allows the ROC 106 along the x-direction to be stabilized through one-third of the mirror length where the aperture 104 is placed. The aperture 104 is placed in the stabilization region, for example, with its center positioned at 75 μm along the length of the mirror 102b. The ROC 106 along the y-direction can be set to a constant value of 70 µm. An approximate threshold value of the edge of the resonant mode shape is depicted by lines 502a and 502b (e.g., half-power beam width). Figure 6A is a schematic cross-section of a VCSEL device 600 in a second direction (e.g., the x-direction). Figure 6B is a schematic cross-section of the VCSEL device of Figure 6A, presented as a plan view. In a second exemplary embodiment shown in Figures 6A and 6B, the ROC 106 in the x-direction may not yet be stable in the front and rear thirds (i.e., the left and right sides) of the lens 102b. In cases where a preference for a single polarization is desired, such as when a device with a known polarization direction and a low polarization ratio is desired, the position of the aperture 104 can be shifted to the end of the lens 102b. In this way, the number of directions introducing losses can be reduced. For example, if the aperture 104 is placed at the left third of the lens 102b, as shown in Figures 6A and 6B, the axial direction of introduced losses is the +x direction. However, the -x direction has a loss ROC 106 that is less than that introduced by the aperture 104 when placed in the center of the lens 102b, thereby reducing the loss introduced in that direction and reducing the polarization preference in the y direction. In other exemplary embodiments, the difference in ROC 106 between the Q and S directions can be altered to advantageously produce a polarization preference for a desired level. In which aperture 104 is positioned within the ROC... y= 70 μm and ROC x In the example at the center of the infinite mirror 102b, similarly presented in the embodiments of Figures 5A to 6B, this structure is considered a reference structure for the following discussion relating to Figures 7A to 7D. Except for adjusting the placement of the aperture 104 on the mirror 102 to change the number of axes introducing loss, the ROC 106 of the mirror 102 in the loss direction can be adjusted as further discussed below. Figures 7A to 7D are schematic cross-sections of various exemplary VCSEL devices with different ROC values ​​in a second direction having corresponding planar views. In one example, ROC... x The ROCUL value can be reduced from infinity as shown in Figure 7A to ~250 μm as shown in Figure 7B to smoothly modulate the difference in loss between x-polarization and y-polarization. The lens shapes 702b and 710b shown in Figure 7B advantageously provide a reduced polarization preference relative to the structure of Figure 7A with lens shapes 702a and 710a. This can be practically achieved by reducing the length of the polymer pattern so that the mirror forms a smooth surface with a near-constant second derivative near the top of lens 102, allowing the definition of a finite positive ROC 106. This example structure in Figure 7B can be formed by fabricating a rectangular pattern with a length of 80 μm and a width of 50 μm. Alternatively, in another exemplary embodiment, polarization preference can be further increased by using a mirror with a negative ROC value in the x-direction, as shown in FIG7C, which can be described as a saddle-point geometry of mirror lens shapes 702c and 710c. Using a concave ROC 106 in the x-direction will amplify the beam and result in additional losses in terms of absorption and diffraction. Therefore, the loss in the x-direction can be greatly increased, which can be advantageous in cases where there is low loss outside the aperture, and thus the beam needs to be significantly amplified to achieve a sufficient loss differential for stable polarization control. Alternatively, if the ROCUL value is large enough that the formation of a mirror 102 with a ROC 106 greater than ROCUL is difficult, such as in the case of a shorter cavity 208, a concave mirror 102c provides a further increase in the loss differential between directions. A saddle point can be created by carefully considering the polymer length. For certain surface tension values, the end of the polymer will tend to create a bulge in the cross-section after reflow, as shown in FIG8. By reducing the cavity length, the effect regions of the two protrusions 810a and 810b begin to interact, thereby forming a smooth curved shape 800. For example, this shape can be produced by a rectangle with a length of 100 μm and a width of 40 μm. Finally, in another exemplary embodiment, the ROC 106 in the lossy direction can be altered to introduce significant loss through a different mechanism. Instead of increasing the ROC 106 to make the beam wider and have greater overlap with the lossy region, the ROC 106 can be reduced to below the cavity length, as illustrated in Figure 7D. With an ROC 106 having a length less than the cavity length, it can be considered as the mirror "over-focusing" and resulting in an unstable resonant mode. Therefore, in which the ROC... y The standard is 70 μm and ROC x In cases where the cavity length is less than 15 μm (e.g., at 15 μm), the resonant mode component aligned only along the y-direction is then stable. This can be achieved by making the length of the polymer rectangle corresponding to the Q-direction a standard value of 40 μm, while the width corresponding to the S-direction can be 18 μm. Regarding the exemplary embodiments described in Figures 7A to 7D above, all such exemplary designs can achieve a ROC of 70 μm. y Figure 7A has an infinite ROC. x Figure 7B has a ROC that is less than infinity and equal to or greater than ROCUL, such as 250 μm; Figure 7C has a negative ROC. x Such as -500 μm; and Figure 7D has a ROC greater than zero but less than one of the cavity lengths. x For example, 15 μm. As briefly mentioned above, Figure 8 is a flowchart of a process for reducing a length to form a saddle point. Figure 8 illustrates how reducing the length of a polymer rectangle can give a saddle point, for example, by comparing a first length 802 with a shorter length 804, creating a saddle point with a shorter length 806, or even a shorter length 808. It should be understood that if the length is too long (e.g., 802) or too short (e.g., 808), a saddle point is unlikely to be formed. The acceptable length range can be controlled by changing the properties of the photoresist. It should be noted that the ROC value near aperture 104 is a relevant feature of the mirror shape used to control beam properties and mode control. Therefore, so far, mirror 102 has been depicted using an area with curved edges. However, mirror 102 may also have one or both ends truncated, as shown with reference to Figures 9A to 9C. All three structures have an ROC that does not intend to introduce additional losses. y (such as 70 μm) and an infinite ROC x(e.g., >5,000 μm). In an exemplary embodiment, a ROC 106 larger than 5,000 μm can be considered substantially infinite and provides a substantially flat surface, as the ROC 106 can be provided by at least one order of magnitude larger than the other direction, and preferably two orders of magnitude larger. As illustrated in Figures 9A to 9C, the differences in the occupied areas 910a, 910b, and 910c may be due to the photomask design affecting the reflow properties at the ends of the mirror, or may be due to etching of certain portions of the mirror due to additional process steps. Figure 9A shows a plan view of an occupied area pattern 910a and a cross-sectional view of mirror shape 902a for the previously described mirror shape. Figure 9B shows a plan view of an occupied area pattern 910b and a cross-sectional view of mirror shape 902b, with one side truncated. Figure 9C shows a plan view of an occupied area pattern 910b and a cross-sectional view of mirror shape 902b, with both sides truncated. Besides altering the mirror's radius of curvature, the aperture shape can be modified to affect the loss profile within cavity 208. Since the fundamental spatial dependence of the loss profile is typically determined by the overlap between the lateral pattern profile and the aperture shape, in some embodiments, the shape of aperture 104 relative to the pattern defined by curved mirror 102 can be quite important. Similar to changing the relative ROC values ​​in the x and y directions, the shape of aperture 104 can be modified to modulate polarization preference. Having an infinite ROC... x and greater than the cavity length and less than the ROCUL y One of the ROC y The mirror 102 will again be used as a reference device for the following discussion of another exemplary embodiment, but the aperture shape will be changed in a manner that can be used in conjunction with any of the previously mentioned variations in the mirror curvature. Figure 10 is a plan view of various exemplary apertures of a VCSEL device according to an exemplary embodiment of the present invention. One aperture shape 104a is circular, as previously described with reference to the examples above. Figure 10 also shows one aperture shape 104b as an ellipse, one aperture shape 104c as a rectangle, and one aperture shape 104d as a rhombus. The different aperture shapes 104a to 104d in the illustrated group can be implemented using the same mirror parameters (e.g., 102a and 102b). Changing the aperture x and y dimensions affects the rhodolol in that direction, and thus the loss in that direction. To increase the preference for polarization directions parallel to the y-direction, an elliptical aperture shape 104b can be used, wherein the radius of the aperture in the x-direction can be reduced relative to the radius in the y-direction. This reduces ROCULx and increases the loss seen in that direction. For example, an aperture 104b can be used with a radius of 1.5 μm in the x-direction and 2 μm in the y-direction. Alternatively, the loss in the y-direction can be reduced by increasing the aperture radius (r) in that direction. y To increase ROCUL y To reduce, as shown in Figure 10. For example, using one of the r values ​​of 2 μm. x and one of 3 μm r y The elliptical lens will make ROCUL y Increased to ~1200 μm. In addition to using smooth, curved aperture shapes, edged or polygonal shapes, such as rectangular aperture shape 104c, rhomboid aperture shape 104d, or triangular shapes, can be used. These shapes can have different side lengths to modulate the loss profile, similar to the use of an elliptical aperture shape 104b, and also have more potential to modulate the gain profile. In the previous exemplary embodiments, the VCSEL structure was configured to adjust the overlap between the mode and the net gain / loss profile by changing the shape of the mode relative to the lateral loss profile while assuming a constant gain profile. However, in the absence of perfectly uniform injection, as is the case in most practical systems, the gain profile can also be different. Typically, due to current distribution issues, the current can be more concentrated at the edges of the aperture. This configuration does not provide a polarization preference for circularly symmetrical apertures, although it may have a limited effect on elliptical apertures. However, controlling the current injection heterogeneity relative to the mode shape generated by an anisotropic mirror can lead to a further preference for a particular polarization. For example, the electric field is typically concentrated around a corner, and using a rhomboid aperture can lead to an increase in the injected current around the corner. By changing the angle of the corner, the electric field profile can be controlled to increase the current injected along the axis to be polarized, thereby increasing the polarization preference in that direction. Alternatively, the net gain / loss can be controllably reduced in this way. For example, consider an aperture shape of 10⁴d, where 2 μm of a y and 1.5 μm of a x This will give a loss preference similar to that of an elliptical mirror of the same radius, but due to incomplete current injection, there may be an enhanced gain difference between the two directions. The use of a curved mirror with a non-circular aperture is advantageous because the beam shape is greatly influenced by the mirror shape. This means that even when using, for example, a rectangular aperture shape, the mirror shape can be chosen to maintain the output beam's relative circular symmetry. For example, for a beam with a diameter of 2 μm... x and one of 3 μm r y and ~250 μm corresponds to ROCUL x and ~1200 μm ROCUL y One of the rectangular aperture shapes 104b can be selected as a mirror 102 with a constant radius of curvature of 500 μm. This will be due to ROC x > ROCUL x This results in losses in the x-direction, but since L < ROC y < ROCUL y Laser emission parallel to the y-direction will still be allowed. In this case, although it has a rectangular aperture shape 104b, the output beam can be circularly symmetrical, thereby retaining one of the many advantages of VCSEL devices. For a rhombus-shaped aperture, different shapes of the rhombus (e.g., narrow, normal, wide) can provide different effects on the current distribution within the aperture 104d. A higher current region can be positioned in a narrow corner and can therefore have higher gain, which can enhance or reduce overlap with a preferred mode. A wide corner can be a current region lower than a narrow corner, while an interior region further away from either corner can be an even lower current region. Therefore, to provide a neutral gain alignment, the aperture shape 104d illustrated in Figure 10 can have a x = a y In this manner, the central region of aperture 104d is a region of lowest current, the center of the edge is a region of higher current, and the corners are a region of higher current than the center of the edge of 104d. To provide an enhanced gain alignment, aperture shape 104d may have a x > a y This allows for maximum current concentration at the top and bottom corners, thereby enhancing gain. To provide a reduced gain alignment, the aperture shape 104d can have a... x < a y This results in the maximum current concentration at the left and right corners, thereby reducing the gain. Therefore, with enhanced gain alignment, the top and bottom of aperture shape 104d have more current, while the middle has relatively less current; conversely, with reduced gain alignment, the right and left portions of aperture shape 104d have more current, while the middle has relatively less current. Those skilled in this art will understand that other aperture shapes are similarly affected by the effect of corners on current distribution. The method of using anisotropic mirrors and / or aperture parameter control modes of the present invention is not limited to a single emitter, but is applicable to multiple emitters or an array of devices using the same lens. Figure 11 is a diagram of various exemplary embodiments of an array of emitters on interconnected mirrors. One example of implementing a multi-emitter device is a 1-D array 111a of VCSEL devices along the same mirror, as shown in Figure 11. Incorporating all devices onto the same mirror allows for smaller spacing between devices because the area occupied at the ends of the mirror does not need to take into account the minimum device spacing. This implementation can be extended from a 1-D emitter array to a 2-D emitter array. The array symmetry can be changed depending on the desired emission characteristics of the array. Figure 11 shows an array 111b with triangular unit cells, having emitters at regular distances from each other. Similarly, Figure 11 shows an array 111c with rectangular unit cells, having emitters at regular distances from each other. A triangular array 111b can induce three distinct polarization directions, each 60 degrees apart, while a rectangular array 111c can produce two directions, each 90 degrees apart. Well-defined polarization directions and ratios are advantageous for applications seeking relatively average emission from known polarizations, such as polarization-segmented multiplexing of information transmission. Furthermore, the number of polarizations can be increased by placing emitters around a circular mirror. Figure 11 shows an array 111d in a circular shape. If the emitters are relatively uniformly spaced at a distance of ≤π / 2 radians, the total emission of the emitter group can be completely polarization-independent. This implementation reduces spotting of emitted light and is highly advantageous for display applications. However, the chips do not need to be regularly spaced because array 111e displays a plurality of emitters with different spacings. Arrays 111 of these shapes in Figure 11 can be tiled to create device arrays larger than the illustrated array, and the device spacing can vary depending on the application. For example, a tiled triangular base can be used to create a hexagonal lattice structure. Finally, the anisotropy of the mirror can be aligned with the orientation of the associated crystal in the device. Group III nitride devices, based on both semi-polarized and non-polarized configurations, exhibit a polarization preference parallel to the α direction as the tilt angle increases. By aligning the mirror so that additional losses are perpendicular to the α direction, a single polarization preference can be enhanced. This can be useful for low-tilt-angle semi-polarized orientations or in situations where the device structure causes strain in the well, thereby altering the band structure. Alternatively, if it is desired to suppress the polarization direction, the mirror losses can be aligned parallel to the α direction, although this may degrade device performance. The main effect of this invention is that it allows control of the cavity mode loss / gain profile in a VCSEL using a curved mirror, which can then be used to advantageously achieve device polarization control. The mode parameters of a VCSEL are determined by the overlap of different modes, defined by their lateral, longitudinal, and polarization properties, resulting in a net gain / loss in the cavity. By increasing the loss for a specific mode, those modes can be suppressed from laser emission. Therefore, the desired mode can be selected by correctly aligning the gain spectrum and increasing the loss of all other modes. In a curved mirror VCSEL cavity 208, the beam size and therefore its overlap with the lossy region outside the aperture 104 are controlled by the mirror ROC 106. However, by using an asymmetric mirror, an asymmetry in the beam shape can be introduced. This asymmetry gives a directionality to the loss profile, thereby promoting laser emission of modes with strong oscillations in the low-loss direction by suppressing modes with greater overlap in the lossy direction. Therefore, a mode with a single oscillation direction can be generated together with a polarization direction defined based on the asymmetry in the curved mirror. The method and structure disclosed in this paper employ a relatively simple modification to introduce anisotropic loss into the laser cavity, which promotes laser emission with a single polarization parallel to the low-loss direction. This minor modification to the process requires only simple adjustments to current manufacturing lines and results in high yields, while reducing implementation costs. This technology is independent of material systems and wavelength, thus allowing for a wide range of applicability. Furthermore, by maintaining a better polarization ROC, there should be no additional loss for that mode, thereby allowing for the preservation of polarization-indiscriminate device performance. Changing the mirror shape allows for fine-tuning of polarization preference, which can be combined with adjustments to the aperture shape to enhance this preference. By aligning the mirror's loss direction with a non-circular aperture, loss can be maximized, and beyond that, simple modifications to the mirror are permissible. For example, aligning the ROC=infinity direction with the smaller radius direction of an elliptical aperture will result in more additional polarization preference than for a circular aperture. Finally, the mirror manufacturing method is applicable to the fabrication of device arrays. The devices can be aligned along a single curved mirror or a lattice to provide different numbers of known polarizations. Having a predefined polarization direction can be beneficial for applications such as polarization-dependent multiplexing. If a curved mirror is used instead of a straight mirror, the polarization can be smoothly rotated along the lens. Furthermore, a device array around a toroidal mirror will have a total output that is relatively independent of polarization. This reduces speckle noise, which is highly beneficial for display quality. In one exemplary embodiment, a VCSEL laminar flowmeter with an anisotropic radius of curvature is configured to provide mode control. Considering the four directions +x, +y, -x, -y, the radius of curvature in at least two directions (+x, -x or +y, -y) is greater than the cavity length and less than ROCUL, and at least one of the other directions is outside this range. In a further exemplary embodiment, the aperture may have a circular shape, an oblong shape, a rectangular shape, or a triangular shape. In a further exemplary embodiment, the aperture has one of the shapes that combine (add or subtract) any of the previous shapes. In a further exemplary embodiment, the mirror has a radius of curvature along one axis that is nearly infinite (>5,000) or substantially flat, and is finite along another axis that is greater than the cavity length. In a further exemplary embodiment, the aperture is one-third of the lens length. In an alternative embodiment, the aperture is located in the first third of the lens length. In another exemplary embodiment, the mirror has a radius of curvature along at least one axis that is less than the cavity length, and in another (other) direction it is greater than the cavity length and is finite. In a further exemplary embodiment, the radius of curvature along at least one axis is concave and convex in another direction, the other direction corresponding to at least one direction having a positive ROC, while at least another has a negative ROC. In another exemplary embodiment, the device uses a plurality of apertures arranged using the same lens. In a further exemplary embodiment, the aperture is arranged in a 1-D array, a triangular lattice, a rectangular lattice, in a circular pattern, and / or at different intervals. In one of the exemplary embodiments provided in another perspective diagram illustrated as shown below for the mathematical relationships, "S" corresponds to the direction in which loss is introduced into the pattern, the values ​​in this direction belonging to set S, and the direction "Q" is the direction in which the mirror intends to function normally, the values ​​in this direction belonging to set Q. In this exemplary embodiment, the aperture may have any of the shapes discussed above. In a further exemplary embodiment, the ROC is expressed as follows: In one example, the aperture may be located at one-third of the lens length. In another example, the aperture may be located at one-third of the lens length. In a further exemplary embodiment, the ROC is expressed as follows: In these examples, the device may use multiple aperture arrangements of the same lens, including, for example, a 1-D array, a triangular dot array, a rectangular dot array, a circular pattern, and / or arranged at different intervals. Therefore, the present invention can be generally implemented to provide the advantageous effects discussed above by improving beam characteristics through anisotropic ROC values ​​in different directions of a curved mirror, wherein the curves in different directions are different from each other, as shown, for example, in the mathematical relations described above. Alternatively or alternatively, this advantageous effect can be provided by setting an aperture shape differently in multiple directions to provide anisotropic ROCUL in different directions. Thus, in an alternative embodiment, introducing anisotropy into one of the anisotropic curved mirrors in the aperture shape (e.g., rectangular, rhomboid, elliptical, or combinations of shapes) satisfies the mathematical relations described above and similarly provides the advantageous effects, as discussed above regarding beam properties. In this example, a VCSEL includes an anisotropic curved mirror and an anisotropic aperture, wherein the cavity length is less than a ROC, and in at least one direction (e.g., along the Q direction of the y-axis), the ROC is less than ROCUL, and in another direction (e.g., along the S direction of the x-axis), the ROC is greater than ROCUL or less than the cavity length. In this example, any combination of aperture shape, aperture placement, and / or beam shaping properties discussed above can be presented, for example, in accordance with the aforementioned mathematical relationships. Furthermore, as described above, any combination of beam shaping properties including mirror curvature and aperture shape can be combined to achieve anisotropic beam properties (e.g., focus and loss, polarization control) that are advantageous to a VCSEL device with a curved mirror. It should be understood that the relative sizes of the components within the diagram may not be proportional and are presented solely for illustrative and explanatory purposes. The following provides a list of exemplary embodiments: Example 1. A VCSEL device comprising: a curved mirror of a VCSEL; and a radius of curvature (ROC) of the curved mirror, which is anisotropic, wherein the ROC includes four directions, the four directions being +x, +y, -x, and -y, wherein the ROC in at least one direction is within a range of a predefined ROC value that is greater than a cavity length of the VCSEL and less than a standard beam width (ROCUL), and the ROC in at least one of the other directions is outside this range. Example 2. The VCSEL device as in Example 1, wherein one aperture of the VCSEL has a circular shape. Example 3. The VCSEL device as in Example 1, wherein one of the apertures of the VCSEL has an elongated oval shape. Example 4. The VCSEL device as in Example 1, wherein one aperture of the VCSEL has a rectangular shape. Example 5. A VCSEL device as in Example 1, wherein one aperture of the VCSEL has a triangular shape. Example 6. The VCSEL device as in Example 1, wherein one aperture of the VCSEL has a shape including at least two of a combination of a circular shape, an oblong shape, a rectangular shape, or a triangular shape. Example 7. The VCSEL device as in Example 1, wherein the ROC system along a first axis is nearly infinite such that the endoscope is substantially flat along the axis, and the ROC system along a different second axis is greater than the cavity length and is finite. Example 8. The VCSEL device as in Example 7, wherein one aperture is located in one-third of the first axis. Example 9. The VCSEL device as in Example 7, wherein an aperture is located in the front third of one of the first axes. Example 10. The VCSEL device as in Example 1, wherein the ROC along a first axis is less than the cavity length, and the ROC along a different second axis is greater than the cavity length and is finite. Example 11. The VCSEL device as in Example 10, wherein the ROC system along the first axis is negative and the ROC system along the second axis is positive. Example 12. The VCSEL device as in Example 1, wherein a plurality of apertures are arranged across the curved mirror surface. Example 13. The VCSEL device as in Example 12, wherein the plurality of apertures are arranged in a one-dimensional array. Example 14. The VCSEL device as in Example 12, wherein the plurality of apertures are arranged in a triangular lattice. Example 15. The VCSEL device as in Example 12, wherein the plurality of apertures are arranged in a rectangular dot matrix. Example 16. The VCSEL device as in Example 12, wherein the plurality of apertures are arranged in a circular pattern. Example 17. The VCSEL device as in Example 12, wherein the plurality of apertures are placed at different intervals. Example 18. The VCSEL device as in Example 1, wherein the ROCUL is half the power beam width. Example 19. A VCSEL device as in Example 1, wherein the ROC is configured to control one mode of the VCSEL. Example 20. A VCSEL device as in Example 1, wherein the laser cavity is made of GaN, InP, GaAs or another semiconductor material. Example 21. A VCSEL endoscope, comprising: a curved mirror of a VCSEL; and a radius of curvature (ROC) of the curved mirror, which is anisotropic, wherein the ROC includes four directions, the four directions being +x, +y, -x, and -y, wherein the ROC in at least one direction is within a range of a predefined ROC value that is greater than a cavity length of the VCSEL and less than a standard beam width (ROCUL), and the ROC in at least one of the other directions is outside this range. Example 22. A light-emitting device comprising a VCSEL cavity mirror as described in Example 21. It should be understood that those skilled in the art will appreciate the various changes and modifications that can be made to the exemplary embodiments described herein. Such changes and modifications can be made without departing from the spirit and scope of the subject matter and without diminishing its intended advantages. Therefore, it is intended that such changes and modifications be covered by the claims of the appended invention. 102: Curved mirror; 102a: Convex lens; 102b: Plane mirror; 102c: Concave lens; 104: Aperture; 104a: Aperture shape; 104b: Aperture shape; 104c: Aperture shape; 104d: Aperture shape; 106: Radius of curvature (ROC) 111a: Array 111b: Array 111c: Array 111d: Array 111e: Array 208: Cavity 210: Plane Mirror 212: Oxide 300: General Process 302: Substrate 304: Photoresist 400: VCSEL Device 402b: Line 500: VCSEL Device 502a: Line 502b: Line 510: Shape 600: VCSEL Device 702a: Shape 702b: Shape 702c: Shape 710a: Shape 710b: Shape 710c: Shape 800: Shape 802: Length 804: Length 806: Length 808: Length 810a: Protrusion 810b: Protrusion 902a: Shape 902b: Shape 910a: Occupied Area 910b: Occupied Area 910c: Occupied Area x: Direction y: Direction Figure 1 is a diagram of a curved mirror of a VCSEL according to one embodiment of the present invention. Figure 2 is a cross-sectional view of a VCSEL, one of the examples of the radius of curvature according to various embodiments of the present invention. Figure 3 is a flowchart of a procedure for forming a curved mirror according to an embodiment of the present invention. Figure 4 is a schematic diagram of a cross-section of a VCSEL device according to an embodiment of the present invention in a first direction. Figure 5A is a schematic cross-section of a VCSEL device according to a first exemplary embodiment of the present invention in a second direction. Figure 5B is a schematic cross-section of the VCSEL device of Figure 5A according to a first exemplary embodiment of the present invention in a plan view. Figure 6A is a schematic cross-section of a VCSEL device according to a second exemplary embodiment of the present invention in a second direction. Figure 6B is a schematic cross-section of the VCSEL device of Figure 6A according to a second exemplary embodiment of the present invention in a plan view. Figures 7A to 7D are schematic diagrams of cross-sections of various example VCSEL devices according to exemplary embodiments of the present invention in a second direction and in corresponding plan views. Figure 8 is a flowchart of a procedure for reducing the length of a saddle point according to a second example embodiment of the present invention. Figure 9 is a cross-sectional view of a VCSEL, illustrating one example of the mirror-occupied area according to various embodiments of the present invention. Figure 10 provides a schematic diagram of various exemplary apertures of a VCSEL device in a plan view according to an exemplary embodiment of the present invention. Figure 11 is a diagram of various arrays of transmitters on an interconnected mirror according to one of the exemplary embodiments of the present invention. 102: Curved Mirror 104: Aperture 106: Radius of curvature (ROC) x: direction y: direction

Claims

1. A vertical cavity surface-emitting laser (VCSEL) device, comprising: A cavity mirror has a curved mirror surface and a radius of curvature (ROC) of the curved mirror surface, which is anisotropic, wherein the ROC includes four directions, the four directions being +x, +y, -x, and -y, the ROC in two directions being within a range of a predefined ROC value that is greater than a cavity length of the VCSEL and less than a standard beam width (ROCUL), and the ROC in the other two directions being outside the range and being a negative value.

2. The VCSEL device of claim 1, wherein one aperture of the VCSEL has a circular shape.

3. The VCSEL device of claim 1, wherein one of the apertures of the VCSEL has an elongated oval shape.

4. The VCSEL device of claim 1, wherein one aperture of the VCSEL has a rectangular shape.

5. The VCSEL device of claim 1, wherein one aperture of the VCSEL has a triangular shape.

6. The VCSEL device of claim 1, wherein one aperture of the VCSEL has a shape including at least two of a combination of a circular shape, an oblong shape, a rectangular shape or a triangular shape.

7. The VCSEL device of claim 1, wherein one aperture is located in one-third of one of the +x and -x directions.

8. The VCSEL device of claim 1, wherein one aperture is located in the first third of one of the +x and -x directions.

9. The VCSEL device of claim 1, wherein the ROC system is negative along the +x and -x directions and positive along the +y and -y directions.

10. The VCSEL device of claim 1, wherein a plurality of apertures are arranged across the curved mirror surface.

11. The VCSEL device of claim 10, wherein the plurality of apertures are arranged in a one-dimensional array.

12. The VCSEL device of claim 10, wherein the plurality of apertures are arranged in a triangular lattice.

13. The VCSEL device of claim 10, wherein the plurality of apertures are arranged in a rectangular dot matrix.

14. The VCSEL device of claim 10, wherein the plurality of apertures are arranged in a circular pattern.

15. The VCSEL device of claim 10, wherein the plurality of apertures are placed at different intervals.

16. The VCSEL device of claim 1, wherein the ROCUL is half the power beam width.

17. The VCSEL device as claimed in claim 1, wherein the ROC is configured to control one mode of the VCSEL.

18. The VCSEL device of claim 1, wherein the laser cavity is made of GaN, InP, GaAs or another semiconductor material.

19. A VCSEL endoscope, comprising: A curved mirror of a VCSEL; and a radius of curvature (ROC) of the curved mirror, which is anisotropic, wherein the ROC includes four directions, the four directions being +x, +y, -x, and -y, the ROC in two directions being within a range of a predefined ROC value greater than a cavity length of the VCSEL and less than a standard beam width (ROCUL), and the ROC in the other two directions being outside the range and being a negative value.

20. A light-emitting device comprising a VCSEL cavity mirror as claimed in claim 19.

Images