Optical amplifier

Abstract

An optical amplifier having a front side, a back side, and an optical resonant cavity, the cavity having a length defined between the front side and the back side, the cavity including a waveguide having an optical confinement factor and a width defined transverse to the length of the cavity, the waveguide has an increasing width in a direction along the length of the cavity, and the optical confinement factor of the waveguide decreases in the direction along the length of the cavity in the portion of the cavity.

Description

[0001] This application is a divisional application. The original application has the application number 201980068369.7 and the original application date is October 31, 2019. The entire contents of the original application are incorporated herein by reference. Background Technology

[0002] This invention relates to optical amplifiers, and for example, to improving the power efficiency of such devices.

[0003] High-performance, low-cost semiconductor optical amplifier (SOA) modules are used in high-capacity, high-speed optical access networks.

[0004] Traditional optical amplifiers typically include a semiconductor block having a front or positive cavity surface, a back or back cavity surface opposite the front cavity surface, and an optical resonant cavity formed between the two. Traditionally, the cavity comprises active layers inserted between p-type or n-type semiconductor material layers and defines a waveguide structure, which can be a ridge waveguide (RW) or a buried heterostructure (BH) waveguide. One or more coatings, such as anti-reflection (AR) films, can be applied to the front and back cavity surfaces.

[0005] A standard approach to fabricating high-saturation-power SOAs is to design a device with a low optical confinement factor. The optical confinement factor of a waveguide is typically defined as the proportion of guided modes contained within the active region or waveguide core (i.e., the structural cross-section that generates optical gain in the SOA device). In such devices, only a low percentage of guided modes are confined within the active region that generates light. Examples include the rarefied SOA disclosed in US2004 / 0196540 A1 and the planar-coupled optical waveguide structure disclosed by Juodawlkis et al. in IEEE Photonics Tech. Lett., Vol. 2, No. 17, February 2005.

[0006] A simple tapered SOA design (where the active width increases along the cavity direction) is also used to increase the saturation output power (P) of the SOA. sat ). Summary of the Invention

[0007] Although these tapered designs are well-known, they are limited in the available P sat Increases are limited. This is because, for ridge waveguides, the optical confinement factor, and therefore the gain per unit length of the SOA, typically increases slightly as the ridge width of the amplifier increases. The same is true for devices with embedded heterostructure waveguides, where the confinement factor increases as the active width of the waveguide increases.

[0008] Therefore, W2Γ1 / W1Γ2 will P sat The increase in value is limited to the first approximation. W1 and W2 are the ridge widths at the beginning and end of the taper, respectively, while Γ1 and Γ2 are the modal constraint factors at the beginning and end of the taper, respectively.

[0009] It is also necessary to pay attention to the rate of increase of the ridge width, because the tapered design can support higher-order transverse dies, so the tapered part needs to be designed to maintain a near-insulated state.

[0010] There is a need to provide SOA devices with improved output saturation power characteristics and their manufacturing methods.

[0011] A first aspect provides an optical amplifier having a front side, a back side, and an optical resonant cavity, the cavity having a length defined between the front side and the back side, and including a waveguide having an optical confinement factor and a width transverse to the length defined by the cavity, wherein, in at least a portion of the cavity, the waveguide has an increasing width in a direction along the length of the cavity, and in the portion of the cavity, the optical confinement factor of the waveguide decreases in the direction along the length of the cavity.

[0012] By increasing the guide width while reducing the optical confinement factor along the SOA cavity length, the power efficiency of the device is improved, and carrier injection can be utilized more effectively.

[0013] For a given saturation level, this invention also enables a reduction in SOA length. The chip size is reduced, allowing more devices to be accommodated per wafer, resulting in lower chip costs compared to existing commercially available devices.

[0014] The direction can be towards the back, corresponding to the magnification direction along the SOA cavity.

[0015] For one portion of the cavity other than the aforementioned portion, the waveguide has a constant width. For the other portion of the cavity, the waveguide can have a constant optical confinement factor. In this way, the variable portion of the device is integrated with a cross-section of constant width and optical confinement factor at the input and output of the SOA device. A wide, low-confinement region adjacent to the output surface can provide uniform guideway dimensions on the output cavity surface.

[0016] The waveguide can be a ridge waveguide. The waveguide can also be an embedded heterostructure waveguide. Therefore, the described technique is compatible with a variety of device structures.

[0017] The front side is a plane, and the length from the front side to the back side is measured in a direction perpendicular to the front side.

[0018] The waveguide is defined on a semiconductor substrate. Therefore, waveguide layer deposition can be achieved using conventional manufacturing methods.

[0019] The amplifier also includes a pair of electrodes, and the optical amplifier is configured to cause the cavity to emit light by applying a current between the electrodes. This amplifies the light propagating along the cavity.

[0020] The cavity includes a first semiconductor layer of a first doping type, a second semiconductor layer of a second doping type opposite to the first type, and an active region located between the first and second semiconductor layers, the first and second semiconductor layers extending in a direction extending between the front and back sides of the amplifier. This structure is compatible with common semiconductor product manufacturing methods.

[0021] The cavity includes a confinement modification layer, the thickness of which and / or its dimension transverse to the length of the cavity affects the optical confinement factor of the waveguide. This facilitates the modification of the optical confinement factor along the cavity.

[0022] The waveguide is symmetrical about an axis parallel to the length of the cavity. This waveguide structure is easy to configure and manufacture.

[0023] The front and / or back surfaces are cleaved surfaces coated with an antireflection film. The coating prevents light from reflecting back into the cavity from the cavity surface, thereby optimizing the magnification effect.

[0024] The cavity is tilted relative to a direction perpendicular to the back surface. This reduces the cavity surface reflectivity and decreases gain fluctuation.

[0025] A second aspect provides a method for fabricating a variable-thickness optical structure for use in an optical amplifier, the method comprising: defining a series of material regions on a substrate, the materials having gaps along a length between them, the average size of the gaps increasing in a first direction along the length; and heating the materials to soften them and vary their thickness along the length. This allows for the realization of a smooth, tapered-confinence modified layer suitable for fabricating the optical amplifier devices described herein.

[0026] The optical structure can be a confinement layer for the optical resonant cavity of an optical amplifier.

[0027] The present invention also provides a method for manufacturing the optical amplifier provided in the first aspect, wherein the method includes manufacturing a variable-thickness optical structure for the optical amplifier provided in the second aspect and further embodiments thereof. Attached Figure Description

[0028] The invention will now be described by way of example with reference to the accompanying drawings.

[0029] In the attached diagram:

[0030] Figure 1 A side view of an example SOA device is shown.

[0031] Figure 2 It shows Figure 1 A cross-sectional view of the AA in the SOA device.

[0032] Figure 3 It shows Figure 1 A plan view of the SOA device.

[0033] Figure 4 A plan view of another example of an SOA device is shown.

[0034] Figure 5 It shows Figure 4 The device is shown in cross-sectional diagrams along (a) BB section and (b) CC section.

[0035] Figure 6 It shows in Figure 4 The TE0 mode profile is calculated at each end of the tapered portion of section 20 in the middle.

[0036] Figure 7 Simulation results of power loss in TM0 and TM1 modes with waveguide tapered lengths are shown.

[0037] Figure 8 Another example of a waveguide used in SOA devices is shown.

[0038] Figure 9 An alternative structure for an SOA device with an etched, restricted modification layer is shown.

[0039] Figure 10 A schematic cross-sectional view of an SOA device with an embedded heterostructure waveguide is shown, equivalent to: Figure 4 (a) BB section and (b) CC section of the device.

[0040] Figure 11 The use of a variable-mark periodic grating to achieve variation of the optical confinement factor along the waveguide length is illustrated.

[0041] Figure 12 It shows Figure 11 Reflow of the grating in the middle is used to achieve a gradual change in the thickness of the modified layer: (a) before reflow, (b) after reflow.

[0042] Figure 13 This illustrates the required lateral alignment between the confined modification layer and the ridge waveguide when the width of the confined modification layer is tapered. Detailed Implementation

[0043] Figure 1 A side view of one form of SOA is shown. The device includes a semiconductor block having a front or positive cavity surface 1, a back or negative cavity surface 2 opposite to the front or positive cavity surface, and an optical resonant cavity formed between the two. The total length of the cavity is L. The cavity can be straight, as shown, or it can be tilted relative to a direction perpendicular to the positive cavity surface.

[0044] Preferably, the front cavity surface and the back cavity surface are parallel and aligned. The back cavity surface may be orthogonal to the length of the cavity. Preferably, the front cavity surface is orthogonal to the length of the cavity. The anti-reflection (AR) film 3 may be coated on one of the front cavity surface and the back cavity surface, or preferably, coated on both the front cavity surface and the back cavity surface.

[0045] The cavity includes an active layer 5, which can be inserted into, respectively, such as Figure 6 and Figure 7 The diagram shows the layers of p-type and n-type semiconductor materials. A confinement modification layer, or simply a confinement layer, is used, such as... Figure 8 As shown, its thickness can taper along at least a portion of the cavity's length. The cavity may also contain additional semiconductor outrigger (OR) layers located beneath layers 5, 6, 7, and 8. In this example, the semiconductor layers are made of InP. However, other semiconductor materials, such as GaAs, can also be used.

[0046] For example, in one embodiment, the device may include an InP base layer beneath an overhang layer forming most of the structure. Layers such as the overhang layer may be made of InGaAsP, which has a lattice-matched InP but a higher refractive index and a bandgap wavelength significantly smaller than that of the gain material to maintain low optical loss. The active layer and surrounding layers may consist of materials including various compositions and types of InGaAsP, InGaAlAs, InAlAs, or InGaAs that can transmit transversely electric (TE) and transverse magnetic (TM) polarized waves, providing gain, waveguide, and other functions within the correct wavelength range.

[0047] The cavity material can be selectively doped in the regions of the p-type and n-type layers 6 and 7. The optical resonant cavity also includes, for example, waveguide layer 4 and etched waveguide layers, or in short waveguides, which provide lateral waveguides. Figure 9 As shown, light is emitted from the waveguide end on the back of the SOA device.

[0048] Figure 2 It shows along Figure 1A cross-sectional view of AA. At each point along the length of the cavity, waveguide layer 4 has a width w transverse to the length defined by the cavity. In the vertical section, active layer 5, confinement modification layer 8, and semiconductor layers 6 and 7 all form part of the waveguide and guide light in the vertical direction. Thus, the waveguide is formed by any layer or combination of layers with a refractive index higher than that of the surrounding substrate and cladding, thereby providing light guidance through the layers.

[0049] Waveguides also possess an optical confinement factor Γ (Gamma). The optical confinement factor of a waveguide is defined as the proportion of light contained within the guided mode in the active region or waveguide core, which is the structural cross-section that generates optical gain in an SOA device. The optical confinement factor of a waveguide can vary with position along the length of the cavity.

[0050] like Figures 1 to 3 As illustrated in the example, a ridge waveguide can be formed by etching parallel trenches in the material on either side of waveguide layer 4, creating protruding isolation bands typically less than 5 μm wide and hundreds of μm long. A material with a refractive index lower than that of the waveguide material can be deposited on the sides of the ridge to guide injected current into the ridge. Alternatively, the ridge can be surrounded by air on three sides that do not contact the substrate beneath the waveguide. The ridge provides lateral waveguide characteristics because the other layers in the structure above the active layer have a higher refractive index than the air or passivation layer (typically silicon dioxide or silicon nitride) in the etched trench region.

[0051] Reference Figure 1 The thickness of the optical confinement layer 8 decreases along the length of the cavity from the front cavity surface 1 to the back cavity surface 2. The thickness of the confinement modification layer 8 affects the optical confinement factor of the waveguide. As the thickness of the confinement modification layer decreases, the optical confinement factor of the waveguide also decreases, and a lower proportion of guided mode light is contained in the active region or waveguide core.

[0052] like Figure 3 As shown, the waveguide also has an increasing width w along the length of the cavity, the direction being from the front to the back of the cavity. Therefore, as... Figure 1 As shown in the side view, as the waveguide width increases, the optical confinement factor of the waveguide decreases along the cavity length in the stated direction due to the gradual decrease in the thickness of the confinement modification layer 8. In this example, the ridge width narrows at the device input at the positive cavity surface 1 and widens in the direction toward the output cavity surface 2.

[0053] Figure 4A plan view of another example of an SOA with a ridged waveguide is shown. In this embodiment, waveguide 4 has a cross-section 20 of variable width in a direction perpendicular to the cavity length and other cross-sections of uniform width. In this example, the ridge width at the device input at the positive cavity face 1 is narrow and constant, and widens from the center of the cavity across region 20 toward the output cavity face. A wide, low-restriction region adjacent to the output face 2 is used to provide uniform rail dimensions on the split output cavity face. The relative lengths of the constant width and tapered cross-sections can be set according to the desired output power of the SOA, as tapering needs to begin once the signal power in the rails approaches the safe power limit to avoid saturation. The minimum tapering length can also be set by ensuring that the waveguide tapering remains nearly adiabatic to minimize scattering losses along its length.

[0054] Figure 5 (a) and (b) respectively show Figure 4 A schematic diagram of the SOA's cross-sectional structure along sections BB and CC. Figure 5 The ridge width of the cross-section in (a) is 2.4 μm and the modified layer is confined by 1.2Q. Figure 5 In (b), the ridge width is 5.0 micrometers and the modified layer has been etched away, thus providing a large, low-confined mode. In this embodiment, there is an additional OR layer below the active layer. These layers can further reduce the waveguide confinement factor by pulling modes down from the active layer.

[0055] Therefore, SOA has a narrow, high-confinence ridge at the SOA input where the amplifier-mode power is lower. On the back side of the device where the amplifier-mode power is greater than the input-mode power, the ridge width tapers upwards, while the thickness of the confinement modification layer tapers downwards.

[0056] The optical confinement variation is achieved using a confinement modification layer 8 located close to the active core of the device. When the confinement modification layer 8 is thicker and has a narrower ridge width, close to the input cavity surface of the device, a single-mode waveguide can be provided, exhibiting a high optical confinement factor and high gain per unit length, while the signal power is lower. As the ridge or mesa width increases, the thickness of the confinement modification layer decreases towards the rear (output) cavity surface of the device, reducing the optical confinement factor and gain per unit length, while maintaining higher optical power.

[0057] This device can achieve high gain per unit length with low drive current under low signal strength conditions, and reduce the gain per unit length on the output cavity surface, thereby providing more current to prevent saturation under high signal power conditions.

[0058] Figure 5 The structures of (a) and (b) were modeled in FimmWave. Tables 1 and 2 show the FimmWave simulation frame and the calculated pattern size. Figure 6 (a) and Figure 6 (b) shows that in Figure 4 The TE0 mode profile is calculated at each end of the tapering section 20. (See also:) Figure 7 As shown, the simulation of tapering loss also indicates that for tapering lengths greater than 200 μm, the losses in both TE0 and TM0 modes are very small, indicating that tapering is adiabatic.

[0059] Table 1

[0060]

[0061] Table 2

[0062]

[0063] For semiconductor optical amplifier devices, the saturated output power (output with gain compressed to 3dB) at a fixed current density is set as a first approximation using the optical confinement factor Γ (Gamma). To avoid using signal mode in SOA (for batch and multiple quantum well (MQW) active devices), the signal output power typically needs to be kept about 5dB below the saturated output power limit.

[0064] To compare the simulation results for tapered structures, Table 3 presents the results for a conventional ridge waveguide SOA design with a fixed confinement factor (optimized for low current) and the same active core, with the Γ* length scaled to show the correlation between the confinement factor Γ, chip length L, and total current and the saturation output of a conventional SOA with a fixed confinement factor and a uniform ridge width corresponding to... Figure 4 The SOA core (BB cross-section) and low-gamma cavity surface (CC cross-section) are described. This calculation assumes identical internal losses for each design. Table 3 shows that to achieve 20 dBm output power at 45 degrees Celsius for a standard fixed-width and fixed-limiting factor SOA, the chip length needs to be approximately 6.6 mm, and the current approximately 2.1 A.

[0065] Table 3

[0066]

[0067] The simulations in Table 4 show that, by using a tapered waveguide, the same output power can be achieved with a chip length of approximately 2.9 mm and a drive current of approximately 0.56 A. This not only improves power efficiency by 5 dB but also reduces chip length by more than 3 dB, thereby lowering chip cost.

[0068] The current, gain, and drive current of the variable-limit SOA (in 100μm steps) are calculated based on a total gain of 20dB. Since the calculated saturation power for the more stringent mode is 11dBm (assuming a transmit power of 0dBm), tapering begins after 1.4mm when 11dB gain is achieved. After a tapered design of 1.7mm, the total gain reaches 20dB.

[0069] Table 4

[0070]

[0071]

[0072] By increasing the guide width while reducing the optical confinement factor along the SOA cavity length, the power efficiency of the device is improved, and carrier injection can be utilized more effectively. The above methods are limited to traveling wave devices (e.g., SOA).

[0073] This invention provides an improved design that significantly enhances the power efficiency of current generation in high-saturation-power SOA devices and enables an SOA that provides high saturation-power output while significantly reducing drive current (approximately 3 times in the calculated embodiment).

[0074] In future module designs, power consumption will be of paramount importance, and energy saving will be key.

[0075] For a given SOA saturation power (approximately twice that in the calculated embodiments), the present invention is also able to reduce its length. The chip size is reduced, allowing more devices to be housed per wafer, resulting in lower chip costs compared to existing commercially available devices.

[0076] Although initial simulations have been performed on devices with very high power (20dBm), the performance advantages also apply to designs for SOAs with a variety of output saturation powers, one of which is the reduction in required drive current.

[0077] This approach is applicable not only to standalone SOA devices but also to monolithic devices that incorporate SOA as an output power booster, including standard telecom lasers, modulators, and photonic integrated circuits. It can also be applied to applications beyond communications to improve the power efficiency of III-V devices such as master oscillator poweramplifiers (MOPAs).

[0078] In another embodiment, such as Figure 8As shown, a short taper design is employed after the input cavity surface, which spans at least 200 μm across the cavity, to ensure thermal insulation and reduce the far-field angle of the device's input cavity surface. In the simulations conducted, for the same 20 dB gain and 20 dBm Psat value, an increase of approximately 100 μm in chip length resulted in an increase of approximately 30 mA in drive current.

[0079] In another embodiment, such as Figure 9 As shown in the cross-sectional view, the cavity has a ridge waveguide structure, in which the confinement modification layer is etched into part of the ridge above the active layer to reduce potential lateral current diffusion. This results in only a small change to the achieved confinement factor. This structure will also be compatible with SOA devices with buried heterostructure (BH) waveguides (and with the ridge waveguides described in previous embodiments). Figure 10 (a) and (b) of 10 respectively show the edge equivalent of the device with the BH waveguide. Figure 4 A schematic diagram of the cross-sections of sections BB and CC.

[0080] Therefore, the present invention can be applied to SOA devices with ridge and BH waveguides.

[0081] The waveguide / cavity may also be tilted relative to the direction perpendicular to the front of the cavity, causing the waveguide to tilt upwards towards the back. The waveguide can be tilted at 5°, 10°, 15°, or 20° relative to the length of the cavity.

[0082] The aim is to limit reflection and scattering losses by restricting the gradual change in the thickness of modified layer 8.

[0083] One method to limit variations in the thickness of the modified layer is to use a variable tooth distributed feedback (DFB) grating, where the percentage of grating teeth varies along the length of the tapered section, such as... Figure 11 As shown. Figure 11 A side view along the length of the ridge is shown, illustrating a schematic diagram of the variable-marking-period "grating" layer 10 in a tapered cross-section of the device. For a 1300nm SOA, assuming a basic period target of 1220nm, the grating period needs to be approximately 190nm, i.e., the unit gap width X is approximately 95nm.

[0084] In one embodiment, limiting the thickness of the modified layer to 100 nm results in a grating coupling coefficient (Kappa) of 155 cm⁻¹, which can function as a strong reflector. Therefore, as Figure 12 As shown in (a), preferably, before regeneration, the uncovered grating layer 10 is preferentially heated to reflow the layer, thereby reducing reflection by smoothing the layer thickness, as... Figure 12As shown in (b). This can limit the reflection and scattering losses relative to the unflowed grating along the grating length and achieve a gradual change in the thickness of the modified layer 8.

[0085] If the variable mark period grating is not flowing, gain clamping can be used. Since the modal index of the guided mode varies along the tapered section, the reflection at each point will differ in wavelength. It is best to use a period with reflections on the very short wavelength side of the gain peak (for the fundamental period near the 50:50 mark period cross section in the tapered section).

[0086] Alternatively, the cycle can be chirped along the tapered portion to provide gain clamping.

[0087] Other methods can also be used to achieve variable thickness-constrained modified layers. Some of these alternative techniques are described below.

[0088] Selective area growth (SAG) can be used, which is easy to implement but tends to limit the variation in layer thickness to 4 or 5 times. That is, it cannot be completely removed at the low-constraint end of the structure, which reduces the range of confinement factor modifications.

[0089] Grayscale etching can also be used, which is achieved by using a grayscale lithography mask that is etched during the etching of the confinement modification layer.

[0090] Variable etching depths can also be used through selective etching. This method, published by NTT in Elec. Lett., Vol. 12, No. 53, utilizes a variable-width mask layer to take advantage of the unpredictable variation in the etching rate of inductively coupled plasma or reactive ion etching within narrow mask channels. The etching rate increases with increasing mask gap size.

[0091] The width of the layer can be gradually reduced to approximately 0.1 μm using stepper lithography or electron beam lithography, thereby gradually thinning the lateral width of the confined modified layer so that the layer does not extend across the entire ridge width. However, as... Figure 13 As shown, precise lateral alignment is required between the modified layer and the lateral ridge. Figure 13 The diagram from top to bottom shows how to change the width of the limiting modified layer 8 and align it with the main waveguide ridge 4.

[0092] The applicant hereby individually discloses each individual feature described herein, as well as any combination of two or more such features. With ordinary knowledge of those skilled in the art, such features or combinations can be implemented as a whole based on this specification, regardless of whether such features or combinations of features solve any of the problems disclosed herein; and without limiting the scope of the claims. This application demonstrates that aspects of the invention can be constituted by any such individual features or combinations of features. In view of the foregoing description, various modifications can be apparent to those skilled in the art within the scope of this invention.

Claims

1. An optical amplifier having a front side, a back side, and an optical resonant cavity, the optical resonant cavity having a length defined between the front side and the back side, the optical resonant cavity including a waveguide having an optical confinement factor and a width transverse to the length defined by the optical resonant cavity, characterized in that, In at least a portion of the optical resonant cavity, the waveguide has an increasing width in the direction along the length of the optical resonant cavity, and in that portion of the optical resonant cavity, the optical confinement factor of the waveguide decreases in the direction along the length of the optical resonant cavity.

2. The optical amplifier according to claim 1, characterized in that, The length of the optical resonant cavity is oriented towards the back side.

3. The optical amplifier according to claim 1 or 2, characterized in that, For the portion of the optical resonant cavity other than the aforementioned portion, the waveguide has a constant width.

4. The optical amplifier according to claim 3, characterized in that, For the other portion of the optical resonant cavity, the waveguide has a constant optical confinement factor.

5. The optical amplifier according to any one of the preceding claims, characterized in that, The waveguide is a ridge waveguide.

6. The optical amplifier according to any one of claims 1 to 4, characterized in that, The waveguide is an embedded heterostructure waveguide.

7. The optical amplifier according to any one of the preceding claims, characterized in that, The front side is a plane, and the length from the front side to the back side is measured in a direction perpendicular to the front side.

8. The optical amplifier according to any one of the preceding claims, characterized in that, The waveguide is defined on a semiconductor substrate.

9. The optical amplifier according to any one of the preceding claims, characterized in that, The amplifier also includes a pair of electrodes, and the optical amplifier is configured to cause the optical resonant cavity to emit light by applying a current between the electrodes.

10. The optical amplifier according to any one of the preceding claims, characterized in that, The optical resonant cavity includes a first semiconductor layer of a first doping type, a second semiconductor layer of a second doping type opposite to the first type, and an active region located between the first semiconductor layer and the second semiconductor layer, the first and second semiconductor layers extending in a direction extending between the front and back sides of the amplifier.

11. The optical amplifier according to any one of the preceding claims, characterized in that, The optical resonant cavity includes a confinement modification layer, the thickness of which and / or the dimension transverse to the length of the optical resonant cavity affects the optical confinement factor of the waveguide.

12. The optical amplifier according to any one of the preceding claims, characterized in that, The waveguide is symmetrical about an axis parallel to the length of the optical resonant cavity.

13. The optical amplifier according to any one of the preceding claims, characterized in that, The front and / or back sides are cleaved surfaces coated with an antireflective film.

14. The optical amplifier according to any one of the preceding claims, characterized in that, The optical resonant cavity is tilted relative to a direction perpendicular to the back surface.

15. A method for manufacturing a confinement layer for a variable-thickness optical structure in an optical amplifier, particularly an optical resonant cavity, characterized in that, The method includes: A series of material regions are defined on a substrate, the materials having gaps along a length, the average size of the gaps increasing in a first direction along the length; The material is heated to soften it and to vary its thickness along its length.

Images