An Analytical Dual-Wavelength Phase Retrieval Method

Through the dual-wavelength phase recovery method, the Fresnel reflection formula and vector analysis method are used to construct higher-order equations, which solves the data redundancy and convergence problems in the iterative phase recovery algorithm, and achieves efficient phase recovery and accurate optical thickness recovery.

CN114092588BInactive Publication Date: 2024-11-15XIAN UNIV OF POSTS & TELECOMM
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202111370374.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-11-18
Publication Date
2024-11-15
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

Existing iterative phase recovery algorithms require a large amount of intensity data when improving convergence performance, resulting in high data redundancy and affecting high-speed and real-time imaging applications.

Method used

Using the dual-wavelength phase recovery method, two lasers of different wavelengths are incident vertically along the main optical path through two lasers of different wavelengths. High-order equations are constructed using Fresnel reflection formula and vector analysis method to solve to restore the phase information of the sample, avoiding iterative optimization. The inuniqueness and convergence stagnation of the algorithm.

Benefits of technology

Efficient phase recovery is achieved, avoiding the problem of solution inequality and convergence stagnation of iterative algorithms, and can accurately restore phase information with an optical thickness exceeding one wavelength without the need for unwraping operations.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114092588B_ABST
    Figure CN114092588B_ABST
Patent Text Reader

Abstract

The present invention relates to coherent diffraction imaging, and in particular to an analytical dual-wavelength phase recovery method, which is used to solve the shortcomings of existing iterative phase recovery algorithms that require a large amount of intensity data to improve convergence performance, resulting in high data redundancy. The analytical dual-wavelength phase recovery method no longer uses an iterative optimization algorithm, but obtains an analytical solution to the phase by solving a high-order equation group, avoiding the problems of non-uniqueness of the iterative algorithm solution and convergence stagnation.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to coherent diffraction imaging, and in particular to an analytical dual-wavelength phase recovery method. Background Art

[0002] Coherent diffraction imaging (CDI) is a lensless imaging method that records intensity information and reconstructs the sample using phase retrieval technology. It has developed rapidly in recent years and is widely used in the fields of nanoscale biomaterial structure measurement, integrated circuit chip structure and surface morphology non-destructive measurement (Science, 2015, 348, 530–535; Optica, 2015, 3, 827–835).

[0003] Phase recovery technology is the core technology of CDI, including direct phase recovery algorithm and iterative phase recovery algorithm; TIE (Transport Intensity Equation) method is a typical direct phase recovery algorithm, which obtains the phase information of the object by intensity changes along the optical axis. However, under coherent illumination, the symmetry of the light intensity transmission equation of the TIE method is low, making it difficult to achieve high-precision phase reconstruction; the iterative phase recovery algorithm originated from the GS (Gerchberg-Saxton) algorithm, which requires repeatedly applying amplitude constraints between the spatial domain (sample plane) and the frequency domain (detector plane). After a certain number of iterations, phase recovery is achieved, but the GS algorithm has problems such as iteration stagnation and non-uniqueness of the solution.

[0004] In order to improve the convergence of the original GS algorithm, people have proposed various improved algorithms, such as taking the axial distance, illumination angle, wavelength, etc. as physical constraints in the iterative phase retrieval algorithm to force the iterative phase retrieval algorithm to converge to the true phase of the object (Optics Letters, 2005, 30(8): 833-835; Optics Letters, 2014, 39(2): 193-196; Optics Express, 2018, 26(17): 21418-21432). In addition, based on the iterative projection idea of ​​the GS algorithm, the hybrid input-output (HIO) algorithm, the error reduction (ER) algorithm, the Yang-Gu (YG) algorithm, etc. have also been proposed to improve the convergence performance of the iterative phase retrieval algorithm.

[0005] Compared with the GS algorithm and the HIO algorithm, the PIE (Ptychographic Iteration Engine) algorithm and the ePIE (Extended Ptychographic Iterative Engine) algorithm have greatly improved convergence due to the high data redundancy generated by the large amount of spatially overlapping intensity data collected during the iteration process, overcoming the problem of slow or even stagnant convergence in traditional CDI methods (Physical Review Letters, 2004, 93(2):023903; Ultramicroscopy, 2005, 103(2):153-16; Nature Photonics, 2013, 7(9):739-45). However, the PIE algorithm and the ePIE algorithm still have the following disadvantages: (1) a large amount of intensity data is required to ensure stable convergence; (2) multiple iterations are required to obtain a reliable solution. The above limitations affect their application in high-speed and real-time imaging. Summary of the invention

[0006] The purpose of the present invention is to solve the disadvantage that the existing iterative phase recovery algorithm requires a large amount of intensity data to improve the convergence performance, resulting in high data redundancy, and to provide an analytical dual-wavelength phase recovery method.

[0007] In order to solve the deficiencies of the above-mentioned prior art, the present invention provides the following technical solutions:

[0008] An analytical dual-wavelength phase retrieval method is characterized in that it comprises the following steps:

[0009] Step (1): respectively start the first light source and the second light source, so that lasers of different wavelengths emitted by the two light sources are vertically incident on the same position of the sample along the main light path, and the CCD image sensor collects two diffraction patterns;

[0010] Step (2): Using the two diffraction patterns obtained in step (1), the phase information of the sample is restored by solving a high-order equation system:

[0011] (2.1) According to the Fresnel reflection formula, calculate the total complex amplitude of the first light source in the two diffraction patterns and the total complex amplitude of the second light source in the two diffraction patterns

[0012] λ1 is the wavelength of the first light source. When λ=λ1, the total complex amplitude of the total reflection field of the first light source in the two diffraction patterns is for:

[0013]

[0014] Where A0 represents the amplitude of the incident field, n represents the refractive index of the sample, and d is the thickness of the sample;

[0015] λ2 is the wavelength of the second light source. When λ=λ2, the total complex amplitude of the total reflection field of the second light source in the two diffraction patterns is for:

[0016]

[0017] (2.2) By vector analysis, the and Perform complex amplitude superposition and then use the cosine theorem to construct a high-order equation only about the sample thickness d:

[0018]

[0019]

[0020] (2.3) Solve the higher-order equation obtained in step (2.2) to recover the phase information of the sample.

[0021] Furthermore, in step (2.3), the step of solving the high-order equation obtained in step (2.2) to recover the phase information of the sample is specifically:

[0022] (2.3.1) The initial value of the function m(i) (i = 1, 2, ..., I) is a zero matrix, and the thickness is expressed as d(pΔx, qΔy), where p = 1, 2, ..., P, q = 1, 2, ..., Q, I, P, Q are positive integers greater than or equal to 2, then the high-order equation obtained in step (2.2) is expressed as:

[0023]

[0024]

[0025] (2.3.2) Let d = a, i = 1, p = 1, q = 1; where a represents the initial value of the sample thickness in nm, and (p, q) represents the coordinate value of each pixel of the sample;

[0026] (2.3.3) Let d(i) = (a+i)·10 -6 , then m1(i) and m2(i) are expressed as follows:

[0027]

[0028]

[0029] Where abs{} represents the absolute value of the orientation quantity;

[0030] m(i)=(m1(i)) 2 +(m2(i)) 2

[0031] (2.3.4) If i=I, execute step (2.3.5); otherwise, let i=i+1 and return to step (2.3.3);

[0032] (2.3.5) Let [k,M] = min(m(i)), where min() represents the minimum value of the orientation quantity, k represents the minimum value of the vector, and M represents the coordinates of the minimum value of the vector;

[0033] (2.3.6) The thickness of the sample is d(pΔx,qΔy) = (a+M)·10 -6 ;

[0034] (2.3.7) If p=P, execute step (2.3.8); otherwise, set p=p+1 and return to step (2.3.3);

[0035] (2.3.8) If q=Q, execute step (2.3.9); otherwise, set q=q+1 and return to step (2.3.3);

[0036] (2.3.9) The phase of the sample is Complete the recovery of the sample's phase information.

[0037] Further, in step (2.2), the vector analysis method is used to respectively analyze the and The specific process of performing complex amplitude superposition and then using the cosine theorem to construct a high-order equation only about the sample thickness d is:

[0038] make Among them, A1 represents the amplitude information of the diffraction pattern corresponding to λ1, and A2 represents the amplitude information of the diffraction pattern corresponding to λ2. A1 and A2 are obtained through the CCD image sensor. and Respectively and Corresponding phase information;

[0039] Through vector analysis Complex amplitude superposition is performed, including the following two cases:

[0040] Case 1: When Using the law of cosines we get:

[0041]

[0042] Case 2: When Using the law of cosines we get:

[0043]

[0044] because Therefore A1 2 Expressed as:

[0045]

[0046] Similarly, through vector analysis Perform complex amplitude superposition, A2 2 Expressed as:

[0047]

[0048] In the above formula Expressed as The high-order equation only related to the sample thickness d is obtained:

[0049]

[0050]

[0051] Furthermore, in step (2.1), the total complex amplitude of the first light source in the two diffraction patterns is calculated and the total complex amplitude of the second light source in the two diffraction patterns The specific steps are as follows:

[0052] According to the Fresnel reflection formula, the complex amplitude information of each diffraction pattern in step (1) is the superposition of two reflection fields. The complex amplitude U1 of the first reflection and the complex amplitude U2 of the second reflection are expressed as:

[0053]

[0054]

[0055] in, represents the amplitude information of the first reflection, represents the amplitude information of the second reflection, exp(jπ) represents the phase change of light from the optically sparse medium to the optically dense medium during the first reflection. represents the phase change produced by the second reflection process, λ represents the incident wavelength, and d represents the thickness of the sample;

[0056] Get the total complex amplitude of the total reflection field of the first light source in the two diffraction patterns for:

[0057]

[0058] Get the total complex amplitude of the second light source in the total reflection field of the two diffraction patterns for:

[0059]

[0060] Compared with the prior art, the present invention has the following beneficial effects:

[0061] The present invention discloses an analytical dual-wavelength phase recovery method, which no longer uses an iterative optimization algorithm, but obtains an analytical solution to the phase by solving a high-order equation group, avoiding the problems of non-uniqueness of the iterative algorithm solution and convergence stagnation. Moreover, the phase distribution in this method is not obtained by inverse tangent calculation, and the phase is not limited to [-π, π]. Therefore, even if the optical thickness exceeds one wavelength, the phase can be accurately recovered, and the unwrapping operation is no longer required. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 It is a schematic diagram of the structure of a dual-wavelength imaging system;

[0063] Figure 2 It is a schematic diagram of the superposition of two reflection fields when light is vertically incident on the sample surface in the present invention;

[0064] Figure 3 For the present invention The first case of complex amplitude vector superposition;

[0065] Figure 4 For the present invention The second case of complex amplitude vector superposition;

[0066] Figure 5 The flowchart of the algorithm for solving the high-order equation in step (2.3) of the present invention;

[0067] Figure 6 The phase distribution recovered by the present invention for a phase object having a measurement range less than one wavelength;

[0068] Figure 7 for Figure 6 The original phase distribution of the medium phase object;

[0069] Figure 8 for Figure 6 Thickness distribution of mid-phase objects;

[0070] Fig. 9 The phase distribution recovered by the present invention for a phase object having a measurement range greater than one wavelength;

[0071] Fig.10 for Fig. 9 The original phase distribution of the medium phase object;

[0072] Fig.11 for Fig. 9 Thickness distribution of mid-phase objects.

[0073] The accompanying drawings are described as follows: 1-first light source, 2-second light source, 3-first polarizer, 4-second polarizer, 5-beam combiner, 6-beam expansion and wavefront shaping device, 7-beam splitter, 8-sample, 9-imaging lens, 10-CCD image sensor, 11-computer. DETAILED DESCRIPTION

[0074] The present invention will be further described below in conjunction with the accompanying drawings and exemplary embodiments.

[0075] An analytical dual-wavelength phase retrieval method based on a dual-wavelength imaging system, such as Figure 1 As shown, the dual-wavelength imaging system includes a second light source 2, a second polarizer 4, an imaging objective lens, a CCD image sensor 10, and a first light source 1, a first polarizer 3, a beam combiner 5, a beam expander and wavefront shaping device 6 and a beam splitter 7 arranged in sequence along the main light path; the laser wavelengths of the first light source 1 and the second light source 2 are 550nm and 760nm respectively, and the laser of the second light source 2 enters the beam combiner 5 after passing through the second polarizer 4 and then exits along the main light path; the exit light of the beam splitter 7 includes a first exit light that is vertically incident on the sample 8 along the main light path and a second exit light that is reflected by the sample 8 and then enters the CCD image sensor 10 through the imaging lens 9 perpendicular to the main light path; the output end of the CCD image sensor 10 is connected to a computer 11 for acquiring a diffraction pattern.

[0076] An analytical dual-wavelength phase recovery method comprises the following steps:

[0077] Step (1): respectively start the first light source 1 and the second light source 2, so that lasers of different wavelengths emitted by the two light sources are vertically incident on the same position of the sample 8 along the main light path, and the CCD image sensor 10 collects two diffraction patterns;

[0078] Step (2): Using the two diffraction patterns obtained in step (1), the phase information of sample 8 is restored by solving a high-order equation group:

[0079] (2.1) According to the Fresnel reflection formula, calculate the total complex amplitude of the first light source 1 in the two diffraction patterns: and the total complex amplitude of the two diffraction patterns of the second light source 2

[0080] According to the Fresnel reflection formula, when light is incident vertically on the surface of sample 8, the reflectivity R and transmittance T are expressed as:

[0081]

[0082] in, is the reflection coefficient, is the transmission coefficient, A0 represents the amplitude of the incident field, and A0 is usually set to 1; n represents the refractive index of the sample 8, which is 1.52 in this embodiment;

[0083] Reference Figure 2 Based on the Fresnel reflection formula, the complex amplitude information of each diffraction pattern in step (1) is the superposition of two reflection fields. The first reflection occurs on the surface of sample 8 along the A′B direction. The second reflection includes one reflection and two transmissions. First, it is transmitted into sample 8 along the AC direction. When it reaches the bottom of sample 8, it is reflected along the C′A″ direction and finally transmitted along the A″B′ direction. When the two beams of light along the A′B direction and the A″B′ direction meet, interference occurs.

[0084] The complex amplitude U1 of the first reflection and the complex amplitude U2 of the second reflection are expressed as:

[0085]

[0086]

[0087] in, represents the amplitude information of the first reflection, represents the amplitude information of the second reflection, exp(jπ) represents the phase change caused by the first reflection process, represents the phase change produced by the second reflection process, λ represents the incident wavelength, and d represents the thickness of sample 8;

[0088] λ1 is the wavelength of the first light source 1. When λ=λ1, the total complex amplitude of the total reflection field of the first light source 1 in the two diffraction patterns is for:

[0089]

[0090] λ2 is the wavelength of the second light source 2. When λ=λ2, the total complex amplitude of the total reflection field of the second light source 2 in the two diffraction patterns is for:

[0091]

[0092] (2.2) By vector analysis, the and Perform complex amplitude superposition, and then use the cosine theorem to construct a high-order equation only about the thickness d of sample 8:

[0093] make have to:

[0094]

[0095]

[0096] Wherein, A1 represents the amplitude information of the diffraction pattern corresponding to λ1, and A2 represents the amplitude information of the diffraction pattern corresponding to λ2. A1 and A2 are obtained by the CCD image sensor 10. and Respectively and Corresponding phase information;

[0097] Through vector analysis Perform complex amplitude superposition, Figure 3 and Figure 4 Two cases of complex amplitude vector superposition are shown, where the length of the vector represents the amplitude information and the angle between the vector and the x-axis represents the phase information;

[0098] Case 1: When and The angle between Using the law of cosines we get:

[0099]

[0100] Case 2: When and The angle between Using the law of cosines we get:

[0101]

[0102] because Therefore A1 2 Expressed as:

[0103]

[0104] Similarly, through vector analysis Perform complex amplitude superposition, A2 2 Expressed as:

[0105]

[0106] In the above formula Expressed as The high-order equation only related to the thickness d of sample 8 is obtained:

[0107]

[0108]

[0109] (2.3)Reference Figure 5, solve the higher-order equation obtained in step (2.2) to recover the phase information of sample 8:

[0110] (2.3.1) The initial value of the function m(i) (i = 1, 2, ..., I) is a zero matrix, I is selected according to the actual situation, and the thickness is expressed as d(pΔx, qΔy), where p = 1, 2, ..., P, q = 1, 2, ..., Q, Q, I, P, Q are positive integers greater than or equal to 2, then the high-order equation obtained in step (2.2) is expressed as:

[0111]

[0112]

[0113] (2.3.2) Let d = a, i = 1, p = 1, q = 1; where a represents the initial value of the thickness of sample 8 in nm, and (p, q) represents the coordinate value of each pixel of sample 8;

[0114] (2.3.3) Let d(i) = (a+i)·10 -6 , then m1(i) and m2(i) are expressed as follows:

[0115]

[0116]

[0117] Where abs{} represents the absolute value of the orientation quantity;

[0118] m(i)=(m1(i)) 2 +(m2(i)) 2

[0119] (2.3.4) If i=I, execute step (2.3.5); otherwise, let i=i+1, that is, add 1 to the value of i, and return to step (2.3.3);

[0120] (2.3.5) Let [k, M] = min(X(i)), where min() represents the minimum value of the orientation quantity, k represents the minimum value of the vector, and M represents the coordinates of the minimum value of the vector;

[0121] (2.3.6) The thickness of sample 8 is d(qΔx,qΔy)=(a+M)·10 -6 ;

[0122] (2.3.7) If p=P, execute step (2.3.8); otherwise, set p=p+1, i.e., add 1 to the value of p, and return to step (2.3.3);

[0123] (2.3.8) If q=Q, execute step (2.3.9); otherwise, set q=q+1, i.e., add 1 to the value of q, and return to step (2.3.3);

[0124] (2.3.9) The phase of sample 8 is obtained as Complete the recovery of the sample's phase information.

[0125] Reference Figures 6 to 8 , this embodiment uses the above phase recovery method to perform phase recovery on a phase object whose measurement range is less than one wavelength; Figure 6 is the phase distribution recovered by the above phase recovery method; Figure 7 is the original phase distribution of sample 8, the number of pixels is 600×600, and the refractive index n is 1.52; Figure 8 This is the thickness distribution of sample 8, with a maximum thickness of 0.188um; Figure 6 and Figure 7 It can be seen that the phase distribution of sample 8 is successfully restored by the method proposed in the present invention, which proves that this method can have a good imaging effect on complex amplitude sample 8.

[0126] Reference Figures 9 to 11 ,Another embodiment uses the above phase recovery method to perform phase recovery on a phase object with a measurement range greater than one wavelength; Fig. 9 is the phase distribution recovered by the above phase recovery method; Fig.10 is the original phase distribution of sample 8, the number of pixels is 1026×1026, and the refractive index n is 1.52; Fig.11 This is the thickness distribution of sample 8, with a maximum thickness of 0.376um; Fig. 9 and Fig.10 It can be seen that the phase distribution of sample 8 is successfully restored by the method proposed in the present invention, which proves that this method can have a good imaging effect on complex amplitude sample 8.

[0127] The above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit them. For ordinary professional and technical personnel in the field, the specific technical solutions recorded in the aforementioned embodiments can be modified, or some of the technical features therein can be replaced by equivalents, and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions protected by the present invention.

Claims

1. An analytical dual-wavelength phase retrieval method, It is characterized in that The following steps are involved: Step (1): respectively start the first light source and the second light source, so that lasers of different wavelengths emitted by the two light sources are vertically incident on the same position of the sample along the main light path, and the CCD image sensor collects two diffraction patterns; Step (2): Using the two diffraction patterns obtained in step (1), the phase information of the sample is restored by solving a high-order equation system: (2.1) According to the Fresnel reflection formula, calculate the total complex amplitude of the first light source in the two diffraction patterns and the total complex amplitude of the second light source in the two diffraction patterns λ 1 is the wavelength of the first light source, when λ=λ 1 When the total complex amplitude of the total reflection field of the first light source in the two diffraction patterns is for: Among them, A 0 represents the amplitude of the incident field, n represents the refractive index of the sample, and d is the thickness of the sample; λ 2 is the wavelength of the second light source, when λ=λ 2 When the total complex amplitude of the second light source in the total reflection field of the two diffraction patterns is for: (2.2) By vector analysis, the and Perform complex amplitude superposition and then use the cosine theorem to construct a high-order equation only about the sample thickness d: Among them, A 1 Represents λ 1 Corresponding to the amplitude information of the diffraction pattern, A 2 Represents λ 2 Amplitude information corresponding to the diffraction pattern; (2.3) Solving the high-order equation obtained in step (2.2) to recover the phase information of the sample; (2.3.1) The initial value of the function m(i) (i = 1, 2, ..., I) is a zero matrix, and the thickness is expressed as d(pΔx, qΔy), where p = 1, 2, ..., P, q = 1, 2, ..., Q, I, P, Q are positive integers greater than or equal to 2, then the high-order equation obtained in step (2.2) is expressed as: (2.3.2) Let d = a, i = 1, p = 1, q = 1; where a represents the initial value of the sample thickness in nm, and (p, q) represents the coordinate value of each pixel of the sample; (2.3.3) Let d(i) = (a+i)·10 -6 , then m 1 (i) and m 2 (i) is expressed as follows: Where abs{} represents the absolute value of the orientation quantity; m(i)=(m 1 (i)) 2 +(m 2 (i)) 2 (2.3.4) If i=I, execute step (2.3.5); otherwise, let i=i+1 and return to step (2.3.3); (2.3.5) Let [k,M] = min(m(i)), where min() represents the minimum value of the orientation quantity, k represents the minimum value of the vector, and M represents the coordinates of the minimum value of the vector; (2.3.6) The thickness of the sample is d(pΔx,qΔy) = (a+M)·10 -6 ; (2.3.7) If p=P, execute step (2.3.8); otherwise, set p=p+1 and return to step (2.3.3); (2.3.8) If q=Q, execute step (2.3.9); otherwise, set q=q+1 and return to step (2.3.3); (2.3.9) The phase of the sample is Complete the recovery of the sample's phase information.

2. The analytical dual-wavelength phase recovery method according to claim 1, Features: In step (2.2), the vector analysis method is used to analyze the and The specific process of performing complex amplitude superposition and then using the cosine theorem to construct a high-order equation only about the sample thickness d is: make Among them, A 1 Represents λ 1 Corresponding to the amplitude information of the diffraction pattern, A 2 Represents λ 2 Corresponding to the amplitude information of the diffraction pattern, A 1 and A 2 Obtained through CCD image sensor, and Respectively and Corresponding phase information; Through vector analysis Complex amplitude superposition is performed, including the following two cases: Case 1: When Using the law of cosines we get: Case 2: When Using the law of cosines we get: because Therefore A 1 2 Expressed as: Similarly, through vector analysis Perform complex amplitude superposition, A 2 2 Expressed as: In the above formula Expressed as The high-order equation only related to the sample thickness d is obtained:

3. The analytical dual-wavelength phase recovery method according to claim 2, Features: In step (2.1), the total complex amplitude of the first light source in the two diffraction patterns is calculated and the total complex amplitude of the second light source in the two diffraction patterns The specific steps are as follows: According to the Fresnel reflection formula, the complex amplitude information of each diffraction pattern in step (1) is the superposition of two reflection fields. The complex amplitude of the first reflection is U 1 and the complex amplitude U of the second reflection 2 Expressed as: in, represents the amplitude information of the first reflection, represents the amplitude information of the second reflection, exp(jπ) represents the phase change that occurs during the first reflection process, represents the phase change produced by the second reflection process, λ represents the incident wavelength, and d represents the thickness of the sample; Get the total complex amplitude of the total reflection field of the first light source in the two diffraction patterns for: Get the total complex amplitude of the second light source in the total reflection field of the two diffraction patterns for: