Accurate calibration method for quarter-wavelength layer thickness of DBR (Distributed Bragg Reflector) reflector

Abstract

The invention relates to a method for accurately calibrating the quarter-wavelength layer thickness of a DBR (Distributed Bragg Reflector). The distributed Bragg reflector (DBR) is formed by alternately growing two materials with the optical thickness of 1 / 4 wavelength. The deviation of the thickness of the DBR enables the reflectivity of the DBR to be reduced, the bandwidth to be reduced and the reflection phase shift to be deviated. In order to accurately control the optical thickness of each layer of the DBR to be 1 / 4 wavelength, the invention provides two complementary resonant cavity comparison structures. By comparing the relative sizes of the cavity mode wavelengths of the two structures, whether the thickness of each layer of the DBR is 1 / 4 of the standard wavelength or not can be judged, and the thickness of each layer can be corrected to be 1 / 4 of the wavelength. The method makes up the defect that the conventional DBR reflection spectrum cannot accurately measure the thickness of each layer of the DBR, has the advantages of simplicity, accuracy and no damage compared with other thickness detection methods, is not only suitable for accurate calibration of the 1 / 4 layer thickness of the DBR, but also can be used for accurate characterization of the thickness of optical thin films made of other materials.

Description

Technical Field

[0001] The invention relates to a precise calibration method for a quarter-wavelength layer thickness of a DBR reflector, and belongs to the field of semiconductor optoelectronics. Background Art

[0002] Distributed Bragg reflectors (DBRs) are made of two materials with different refractive indices grown alternately, and the optical thickness of each layer is one-quarter wavelength. Multiple pairs of DBRs can form a wide-bandwidth, high-reflectivity reflector. DBRs are widely used in optoelectronic devices. For example, DBRs grown on light-emitting diode (LED) substrates can increase the light extraction efficiency of LEDs and reduce the substrate's absorption of light. Similarly, DBRs are also an important component of vertical cavity surface emitting lasers (VCSELs). With the high reflectivity characteristics of DBRs, the light field can form a stable oscillating standing wave in the resonant cavity, significantly reducing the threshold current of VCSELs.

[0003] An ideal DBR requires that the optical thickness of each layer should be one quarter of a wavelength. When the optical thickness of the DBR deviates by one quarter of a wavelength, the characteristics of the DBR will be affected, such as: (1) the reflectivity of the DBR will decrease, and the decrease in reflectivity will lead to an increase in the threshold of the VCSEL, or even the device will not be able to lasing; (2) the reflection bandwidth of the DBR will decrease, and the reduction in bandwidth will reduce the light extraction efficiency of the LED; (3) the reflection phase shift of the DBR will change, and the reflection phase shift will affect the penetration depth of the DBR, that is, the effective cavity length of the VCSEL, and thus affect the modulation rate of the VCSEL.

[0004] Currently, there are many methods to characterize and control the thickness of DBR, which can be divided into in-situ and non-in-situ testing according to whether the test is performed during growth. Common in-situ monitoring is to test the reflection spectrum during growth. Since the refractive index of the material at the growth temperature is different from that of the material at room temperature, this method is affected by the growth temperature, and the temperature fluctuation during growth will also affect the judgment of the film thickness. In addition, the quartz crystal oscillator method is also an in-situ monitoring method with the advantage of simple equipment, but the probe life of the quartz crystal oscillator is relatively short. Focused ion beam-scanning electron microscope (FIB-SEM) is the most direct non-in-situ testing method. This method cuts the grown DBR and then measures the cross section to obtain a very clear picture, but FIB-SEM requires the test sample to have good conductivity, otherwise the sample is prone to charging effect, resulting in blurred images, which in turn affects the judgment of thickness. In addition, this method is damaging to the sample, the sample preparation cost is high, and the test cycle is long. White light reflectance spectrum is the most commonly used non-in-situ testing method. After the DBR growth is completed, the peak reflectivity and peak reflection wavelength are obtained by measuring its white light reflectance spectrum, and then the growth parameters are further adjusted to achieve the target reflection wavelength. However, this method obtains the optical thickness of a pair of DBRs and cannot determine whether the thickness of each layer of the DBR is a quarter wavelength. Summary of the invention

[0005] The purpose of the present invention is to characterize the optical thickness of each layer of DBR and accurately calibrate it to a quarter wavelength. The present invention proposes a method for calibrating the thickness of DBR layer using two complementary resonant cavity structures. This method can not only determine the degree of deviation of each layer of DBR from the 1 / 4 wavelength layer thickness, but also can accurately adjust the thickness of each layer to 1 / 4 wavelength. This method has the advantages of simplicity and accuracy, and is non-destructive to the detection of DBR.

[0006] The structure of DBR can be found in Figure 1 , DBR (120) consists of N to M H (121) and M L (122) alternating growth composition, where M H (121) and M L The thickness of the (122) layer is d H and d L , the refractive index is n H and n L (n H >n L ), the entire structure is sequentially epitaxially grown on a refractive index of n S (110)M s On the substrate. For a standard DBR structure, the thickness of each layer must meet

[0007]

[0008] Among them, λ DBR is the center wavelength of the DBR reflection spectrum, and M H (121) and M L (122) The material corresponds to λ DBR In experiments, the thickness of DBR is usually characterized by the white light reflectance spectrum, and the growth parameters of DBR in VCSEL are calibrated according to the reflectivity R(λ) at different wavelengths λ. However, this calibration method can only calibrate the optical thickness of a pair of DBRs, that is, n H d H +n L d L , it is not certain whether the thickness of each layer of DBR is

[0009] The optical thickness of a pair of DBRs is n H d H +n L d L Fixed to If the optical thickness of each DBR layer is not equal to Defined as non-standard DBR. In order to characterize the degree to which the thickness of each layer of the non-standard DBR is offset by 1 / 4 wavelength, a high refractive index material M is introduced. H (121) Layer thickness offset factor f H ,Right now

[0010]

[0011] In the formula, f H The value range is 0<f H <2. When f H =1, Corresponding to the standard DBR. When f H ≠1, corresponding to non-standard DBR, and when 0<f H <1, the high refractive index (121)M H The layer thickness is less than Low refractive index (122)M L The layer thickness is greater than When 1<f H <2, the high refractive index (121)M H The layer thickness is greater than Low refractive index (122)M L The layer thickness is less than

[0012] To determine the high refractive index material M H Layer thickness offset factor f H Here, a method for calibration using two complementary resonant cavity structures is invented. The two complementary resonant cavity structures are shown in FIG2 , where the two structures are respectively called high refractive index resonant cavity and low refractive index resonant cavity. Each structure consists of three parts, namely, an upper DBR reflector, a lower DBR reflector, and a resonant cavity sandwiched between the reflectors. The material types and thicknesses of the layers of the upper and lower reflectors are Figure 1 The parameters of the materials of each layer of the DBR to be tested are the same. For the high (low (330)) refractive index resonant cavity, the cavity material (240) is the same as the high (low (122)) refractive index layer (121) of the DBR to be tested, that is, M H (M L ), the thickness of the cavity (240) is 4 times the thickness of the high (low (330)) refractive index layer of the DBR to be tested, i.e. 4d H (4d L ), forming a 1λ resonant cavity. Considering the coherent superposition of the phase, the upper DBR (250) of the high refractive index resonant cavity has a logarithm of K, and the lower DBR (230) is combined with an additional layer of low refractive index M L (220) forms K+1 / 2 pairs of DBR reflectors, while the number of pairs of the lower DBR (320) of the low refractive index resonant cavity is K pairs, and the upper DBR (340) and the high refractive index M H (350) forms K+1 / 2 pairs of DBR reflectors.

[0013] The cavity mode wavelength of the high refractive index resonant cavity is The cavity mode wavelength of the low refractive index resonant cavity is With the f H increases with the increase of Then with f H decreases with the increase of H =1, When 0<f H When <1, the high refractive index M in DBR H The thickness of (121) is smaller than that of (122) with low refractive index. L The layer thickness, When 1<f H <2, the high refractive index (121)M H The layer thickness is greater than the low refractive index M L The layer thickness of (122) For different f H , and The average is a nearly H is an irrelevant constant, and fH The further away from 1, the greater the difference in the wavelengths of the two cavity modes. Therefore, according to the difference of the cavity modes of the two complementary resonant cavity structures, With f H The relationship can be used to determine the deviation of the thickness of each layer of DBR situation.

[0014] The present invention provides a process for calibrating a DBR to a standard DBR using a complementary resonant cavity calibration method. (1) Based on the growth time t of the high refractive index layer and the low refractive index layer of the DBR to be calibrated, H and t L , respectively grow a high refractive index resonant cavity and a low refractive index resonant cavity to obtain two comparative complementary resonant cavity epitaxial wafers. (2) Test the reflection spectrum of the complementary resonant cavity epitaxial wafer and obtain the reflection spectrum of the complementary resonant cavity epitaxial wafer under the current growth time. and (3) If The growth time of the standard DBR is t H and t L (4) If The average value of the cavity modes of the two complementary resonant cavities As λ DBR , simulated λ DBR Lower Δλ c With f H (5) According to the measured And combined with Δλ c ~f H Relationship, get f H (6) According to f H Update H and t L t H / f H and t L / (2-f H ). (7) Repeat steps (1) to (6) until the condition of step (3) is met.

[0015] The method for accurately calibrating the quarter-wavelength thickness of a DBR reflector described in the present invention has some important advantages over the traditional optical film thickness test, which are reflected in the following aspects:

[0016] 1. High accuracy

[0017] Since the DBR thickness accurate calibration method of the present invention determines the deviation degree of DBR thickness through the cavity mode of the resonant cavity, it is easier to determine the position of the cavity mode than through the central wavelength of the DBR reflection spectrum. In addition, the reflection spectrum test required can be performed at room temperature without considering the interference caused by temperature.

[0018] 2. Convenience

[0019] For growth equipment without in-situ monitoring, it is necessary to use external means such as white light reflectance spectrum or FIB-SEM for testing. The present invention is based on the growth of two complementary resonant cavity structures and the testing of white light reflectance spectrum to characterize and calibrate the DBR layer thickness. Compared with FIB-SEM cutting to test the thickness, the present invention does not require additional complex sample preparation processes such as FIB-SEM. In addition, FIB-SEM requires the sample to have good conductivity, otherwise a charging effect will occur, which will affect the characterization of dielectric optical materials.

[0020] 3. Wide range of applications

[0021] The present invention only requires an additional reflectance spectrum test instrument, which is the most basic characterization tool. The present invention is applicable to growth equipment with or without online monitoring. Moreover, this method is not only for the precise calibration of the DBR quarter layer thickness, but also for the calibration of optical film thickness of other materials. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 It is a structural diagram of DBR.

[0023] FIG. 2 is a schematic diagram of two complementary resonant cavity structures.

[0024] Figure 3 It is a schematic diagram of the process of the present invention.

[0025] Figure 4 For different f H The reflection spectrum.

[0026] Figure 1 The two materials that make up DBR are M H and M L , where M H M L The thickness of the layers is d H and d L , the refractive index is n H and n L (n H >n L ), the entire structure is sequentially epitaxially grown on a refractive index of n S M s (110) on the substrate. 110: substrate material M s , whose refractive index is n s 120: DBR in the DBR structure diagram, the number of periods is N. 121: High refractive index material M in DBR H , whose refractive index is n H , thickness is d H . 122: Low refractive index material M in DBRL , whose refractive index is n L , thickness is d L

[0027] Figure 2a is a high refractive index resonant cavity in a complementary resonant cavity structure. Figure 2b For the high (low) refractive index resonant cavity, the cavity material is the same as the high (low) refractive index layer of the DBR to be tested, that is, M H (M L ), the thickness of the cavity is 4 times the thickness of the high (low) refractive index layer of the DBR to be tested, that is, 4d H (4d L ), forming a 1λ resonant cavity. Considering the coherent superposition of phases, the total number of pairs of upper DBR and lower DBR of the high refractive index resonant cavity is K and K+1 / 2 pairs, respectively, while the total number of pairs of upper DBR and lower DBR of the low refractive index resonant cavity is K+1 / 2 and K pairs, respectively.

[0028] Figure 2a 210: Substrate Materials M s , whose refractive index is n s . 220: Low refractive index material M L , whose refractive index is n L , thickness is d L 230: The lower DBR of the high refractive index resonant cavity structure, the period number is K. 231: The high refractive index material M in the lower DBR of the high refractive index resonant cavity structure H , whose refractive index is n H , thickness is d H . 232: Low refractive index material M in the lower DBR of the high refractive index resonant cavity structure L , whose refractive index is n L , thickness is d L . 240: Resonant cavity with high refractive index resonant cavity structure, material is M H , whose refractive index is n H , thickness is 4d H 250: The upper DBR of the high refractive index resonant cavity structure, the period number is K. 251: The low refractive index material M in the upper DBR of the high refractive index resonant cavity structure L , whose refractive index is n L , thickness is d L . 252: High refractive index material M in the upper DBR of the high refractive index resonant cavity structure H , whose refractive index is n H , thickness is d H .

[0029] Figure 2b 310: Substrate Materials M s , whose refractive index is ns 320: The lower DBR of the low-refractive-index resonant cavity structure, the period number is K. 321: The low-refractive-index material M in the lower DBR of the low-refractive-index resonant cavity structure L , whose refractive index is n L , thickness is d L . 322: High refractive index material M in the lower DBR of the low refractive index resonant cavity structure H , whose refractive index is n H , thickness is d H 330: The resonant cavity of the low refractive index resonant cavity structure, the material is M L , whose refractive index is n L , thickness is 4d L 340: Upper DBR of low refractive index resonant cavity structure, with period number K. 341: High refractive index material M in the upper DBR of low refractive index resonant cavity structure H , whose refractive index is n H , thickness is d H . 342: Low refractive index material M in the upper DBR of the low refractive index resonant cavity structure L , whose refractive index is n L , thickness is d L . 350: High refractive index material M H , whose refractive index is n H , thickness is d H

[0030] Note: The total lower DBR in FIG2(a) is K+1 / 2 pairs of total lower DBRs formed by combining K pairs of lower DBRs (230) and half a pair of DBRs (220).

[0031] Note: The total upper DBR in FIG2( b ) is K+1 / 2 pairs of total upper DBRs formed by combining K pairs of upper DBRs ( 340 ) and half a pair of DBRs ( 350 ). DETAILED DESCRIPTION

[0032] Take the lower DBR used in red VCSEL as an example. The (121)M H and (122)M L The layer materials are Al 0.5 Ga 0.5 As and AlAs, the corresponding growth times are t H and t L , the number of cycles N is 21.

[0033] First, EMCORE D125 MOCVD was used to grow Figure 1DBR structure, the substrate is 2 inches (100) <111> The GaAs with a 15° deviation uses TMAl, TMGa and TMIn as the III-group metal organics, high-purity AsH3 as the V-group source, and high-purity H2 as the carrier gas. The pressure in the reaction chamber is 80mbar, the growth temperature is 600℃, and the substrate speed during growth is 1000rpm. 0.5 Ga 0.5 The time taken by As and AlAs is t H =65.19s and t L =76.06s. After the growth is completed, the white light reflection spectrum of DBR is tested using the micro-area transmission and reflection test system. The test shows that the central wavelength of DBR is about 670 nm Place.

[0034] In order to determine whether the current DBR is a standard DBR, two complementary resonant cavity structures corresponding to AlGaAs were grown, where K = 10. After the growth was completed, the white light reflection spectrum was tested. The test results showed that the cavity mode of the high refractive index resonant cavity The cavity mode of the low refractive index resonant cavity is 673.11nm. is 667.32nm. The average value of the two Approximately 670nm, the difference between the two cavity modes Δλ c is 5.79nm. Therefore, t H =65.19s and t L =76.06s The DBR obtained is a non-standard DBR. c With f H The relationship is: Δλ c =5.79nm corresponding to f H is 1.035. H and t L The time was updated to 65.19s / 1.035 and 76.06s / (2-1.035) respectively, and two complementary resonant cavity test structures were regrown, and the white light reflection spectrum was tested. The results showed that the cavity modes of the two complementary resonant cavities after calibration were 669.71nm and 670.39nm respectively, which were almost equal. Therefore, the DBR obtained using the corrected growth time can be considered as the standard DBR.

[0035] 1. The logarithm K of the DBR in the complementary resonant cavity structure, we choose K as 10. The value range of K is not clearly specified. Of course, the larger the K, the higher the reflectivity of the upper and lower DBRs in the complementary resonant cavity structure, so that the cavity mode of the two complementary resonant cavity structures is easier to see when testing the white light reflection spectrum (because the larger the K, the higher the DBR reflectivity, the cavity mode is very obvious as a specific position in the reflection spectrum. Figure 4 Note: The left side of this figure is fH <1, f in the middle H =1, the right side is f H >1. HRIC stands for high refractive index cavity, LRIC stands for low refractive index cavity, and the two asterisks are the positions of the cavity modes. ), but the larger the K, the more time it takes to grow the complementary cavity structure. Therefore, considering the above, if the value of K is too small, it will affect the judgment of the cavity mode position, and if it is too large, it will increase the experimental time and cost. You can make a choice based on the actual situation.

[0036] 2. The formula for standard DBR is shown in formula (1) in the manual. The formula for non-standard DBR is shown in formula (2)(3)(4) in the manual. First of all, it should be clarified that n in these four formulas H and n L When we select two materials to form DBR, namely M H and M L After that, it is known. Therefore, for the standard DBR, that is, formula (1), as long as we know λ DBR , then its thickness d H and d L But for non-standard DBR, knowing λ DBR , we also need to know f H In order to deduce the thickness d H and d L .

[0037] 3. According to the analysis of point 2 above, if the two materials selected to constitute the DBR are known, we need to know the central wavelength λ of the DBR under the current growth conditions. DBR and f H It can be deduced that d H and d L Through simulation (transmission matrix method, a method specifically used to calculate the transmittance and reflection characteristics of multilayer films), we found that the average value of the cavity mode of the two complementary resonant cavity structures is Very close to the central wavelength λ corresponding to DBR DBR , regardless of the center wavelength λ DBR and f H For this reason, we assume that the average value of the cavity mode of the two complementary resonant cavity structures is That is, the central wavelength λ of the DBR under the current growth conditions DBR ,Right now Then, experimentally, we grow two complementary resonant cavity structures and obtain the reflection spectrum. and There is λ DBR ,Right now Then for the non-standard DBR equations (2), (3), and (4), we have only one unknown variable, namely, f H (Because f H If we know the thickness d H and d L You will know).

[0038] 4. When λ DBR Knowing this, we can simulate different f H Down, and The value of the cavity mode difference of the two complementary resonant cavity structures is obtained. With f H The relationship obtained by comparing the simulation With f H Grown from relationships and experiments and You can get f H .

[0039] 5. DBR calibration process Figure 3 As shown (Note: CRC in the flowchart is the abbreviation of the two complementary resonant cavity structure). (1) According to the growth time t of the high refractive index layer and the low refractive index layer of the DBR to be tested H and t L , and two comparative complementary resonant cavity epitaxial wafers are grown. (2) Test the reflection spectrum of the complementary resonant cavity epitaxial wafer and obtain the reflection spectrum of the complementary resonant cavity epitaxial wafer under the current growth time. and (3) If The growth time of the standard DBR is t H and t L (4) If The average value of the cavity modes of the two complementary resonant cavities As λ DBR , simulation λ DBR Lower Δλ c With f H (5) According to the measured Combined with the Δλ obtained by simulation c ~f H Relationship, get f H (6) According to f H Update H and t L t H / f H and t L / (2-f H ). (7) Repeat steps (1) to (6) until the condition of step (3) is met.

[0040] During the preparation of DBR, the growth rate of each layer is constant, and the layer thickness of DBR is proportional to the growth time. Therefore, the DBR layer thickness (d H and d L ) is the adjustment of the growth time (t H and t L ) adjustments.

[0041] The quarter-wavelength thickness mentioned in the specification actually refers to the optical thickness n H d H and n L d L , that is, the optical thickness of each layer of the quasi-DBR is one quarter of the wavelength, not the physical thickness of the quasi-DBR is one quarter of the wavelength.

Claims

1. A method for accurately calibrating the quarter-wavelength layer thickness of a DBR reflector, characterized in that: By growing two complementary resonant cavity structures and testing the white light reflection spectra of the two structures, the degree of DBR layer thickness deviation by a quarter wavelength under the current growth conditions is obtained based on the cavity mode difference of the two structures, and then the calibration is performed; from the vertical section of the two complementary resonant cavity structures, both structures are composed of three parts, namely the upper DBR and the lower DBR, and the resonant cavity sandwiched between the upper and lower DBRs; The cavity material of the high refractive index resonant cavity in the two complementary resonant cavity structures is the same as the high refractive index material in the DBR to be tested, and the thickness of the cavity is 4 times the thickness of the high refractive index layer in the DBR to be tested; the material parameters and thickness of the upper and lower DBRs constituting the high refractive index resonant cavity structure are the same as the parameters of the materials of each layer of the DBR to be tested, and the number of pairs of the upper and lower DBRs is K and K+1 / 2 pairs respectively; The cavity material of the low refractive index resonant cavity in the two complementary resonant cavity structures is the same as the low refractive index material in the DBR to be tested, and the thickness of the cavity is 4 times the thickness of the low refractive index layer in the DBR to be tested; the material parameters and thickness of the upper and lower DBRs constituting the low refractive index resonant cavity structure are the same as the parameters of the materials of each layer of the DBR to be tested, and the number of pairs of the upper and lower DBRs is K+1 / 2 and K pairs respectively; By only growing two complementary resonant cavity structures and testing the white light reflection spectrum to obtain the cavity modes of the two structures, the central wavelength of the DBR can be determined, which is the average value of the cavity modes of the two complementary resonant cavity structures. When the central wavelength of the DBR is determined, the relationship between the difference in the cavity modes of the two complementary resonant cavity structures and the degree of deviation can be calculated by simulation when the DBR is deviated by one quarter of the current central wavelength. The degree to which the DBR layer thickness is offset by a quarter wavelength under the current growth conditions can be determined by the difference in the cavity modes of the two complementary resonant cavity structures and the DBR center wavelength corresponding to the current growth conditions.

2. The method according to claim 1, characterized in that The following steps are involved: (1) According to the growth time t of the high refractive index layer and the low refractive index layer of the DBR to be tested H and t L , respectively growing a high refractive index resonant cavity and a low refractive index resonant cavity, and obtaining two relatively complementary resonant cavity epitaxial wafers; (2) Test the reflection spectrum of the complementary resonant cavity epitaxial wafer and obtain the reflection spectrum of the complementary resonant cavity epitaxial wafer under the current growth time. and (3) If The growth time of the standard DBR is t H and t L ; (4) If The average value of the cavity modes of the two complementary resonant cavities As λ DBR , simulated λ DBR Lower Δλ c With f H relationship; (5) According to the measured And combined with Δλ c ~f H Relationship, get f H ; (6) According to f H Update H and t L t H / f H and t L / (2-f H ); (7) Repeat steps (1) to (6) until the condition of step (3) is met; wherein the high refractive index material M H Layer thickness offset factor f H The value range is 0 <f H <2; the cavity mode wavelength of the high refractive index resonant cavity is The cavity mode wavelength of the low refractive index resonant cavity is

Images