OSA's Digital Library

Optics Express

Optics Express

  • Editor: C. Martijn de Sterke
  • Vol. 19, Iss. 3 — Jan. 31, 2011
  • pp: 2165–2180
« Show journal navigation

Pixel-based OPC optimization based on conjugate gradients

Xu Ma and Gonzalo R. Arce  »View Author Affiliations


Optics Express, Vol. 19, Issue 3, pp. 2165-2180 (2011)
http://dx.doi.org/10.1364/OE.19.002165


View Full Text Article

Acrobat PDF (928 KB)





Browse Journals / Lookup Meetings

Browse by Journal and Year


   


Lookup Conference Papers

Close Browse Journals / Lookup Meetings

Article Tools

Share
Citations

Abstract

Optical proximity correction (OPC) methods are resolution enhancement techniques (RET) used extensively in the semiconductor industry to improve the resolution and pattern fidelity of optical lithography. In pixel-based OPC (PBOPC), the mask is divided into small pixels, each of which is modified during the optimization process. Two critical issues in PBOPC are the required computational complexity of the optimization process, and the manufacturability of the optimized mask. Most current OPC optimization methods apply the steepest descent (SD) algorithm to improve image fidelity augmented by regularization penalties to reduce the complexity of the mask. Although simple to implement, the SD algorithm converges slowly. The existing regularization penalties, however, fall short in meeting the mask rule check (MRC) requirements often used in semiconductor manufacturing. This paper focuses on developing OPC optimization algorithms based on the conjugate gradient (CG) method which exhibits much faster convergence than the SD algorithm. The imaging formation process is represented by the Fourier series expansion model which approximates the partially coherent system as a sum of coherent systems. In order to obtain more desirable manufacturability properties of the mask pattern, a MRC penalty is proposed to enlarge the linear size of the sub-resolution assistant features (SRAFs), as well as the distances between the SRAFs and the main body of the mask. Finally, a projection method is developed to further reduce the complexity of the optimized mask pattern.

© 2011 Optical Society of America

1. Introduction

Due to the resolution limits of optical lithography systems, the electronics industry has relied on resolution enhancement techniques (RET) to compensate and minimize mask distortions as they are projected onto semiconductor wafers [1

1. A. K. Wong, Resolution Enhancement Techniques (SPIE Press, 2001), Vol. 1. [CrossRef]

,2

2. X. Ma and G. R. Arce, Computational Lithography, Wiley Series in Pure and Applied Optics, 1st ed. (John Wiley and Sons, 2010). [CrossRef]

]. Resolution in optical lithography obeys the Rayleigh resolution limit R=kλNA, where λ is the wavelength, NA is the numerical aperture, and k is the process constant which can be minimized through RET methods [3

3. S. A. Campbell, The Science and Engineering of Microelectronic Fabrication, 2nd ed. (Publishing House of Electronics Industry, 2003).

6

6. L. Liebmann, S. Mansfield, A. Wong, M. Lavin, W. Leipold, and T. Dunham, “TCAD development for lithography resolution enhancement,” IBM J. Res. Dev. 45, 651–665 (2001). [CrossRef]

]. In optical proximity correction (OPC), mask amplitude patterns are modified by the addition of the sub-resolution assistant features (SRAFs) that can pre-compensate for imaging distortions [1

1. A. K. Wong, Resolution Enhancement Techniques (SPIE Press, 2001), Vol. 1. [CrossRef]

, 2

2. X. Ma and G. R. Arce, Computational Lithography, Wiley Series in Pure and Applied Optics, 1st ed. (John Wiley and Sons, 2010). [CrossRef]

]. In PBOPC, the mask is divided into small pixels, each of which is modified during the optimization process. The optimized PBOPC is then fabricated by two approaches: the raster scanning e-beam method and the vector scanning e-beam method [7

7. U. Okoroanyanwu, Chemistry and Lithography, 1st ed. (in review, SPIE Press, 2010).

]. In raster scanning, each pixel on the mask pattern is exposed by the electron-beam moving back and forth row by row. Due to the low throughput of the raster scanning method, the vector scanning approach is now preferred for mask fabrication. In the vector scanning method, the turned off electron-beam first moves on a vector path and directly reaches the region to be exposed. Then, the electron-beam is turned on to expose the nether feature on the PBOPC. After that, the electron-beam is turned off and moves again until the entire pattern is completely exposed [7

7. U. Okoroanyanwu, Chemistry and Lithography, 1st ed. (in review, SPIE Press, 2010).

]. Tiny features are either unmanufacturable or disadvantageous for the mask writing accuracy. Since each pixel on the PBOPC may be flipped independently, the two critical issues in PBOPC are the computational complexity in the design and the manufacturability of the optimized mask.

Several OPC optimization algorithms have been proposed in the literature. These range from heuristic and empirically-based design rules to optimization-based inverse algorithms. Sherif, et al. derived an iterative approach to generate binary masks in incoherent diffraction-limited imaging systems [8

8. S. Sherif, B. Saleh, and R. Leone, “Binary image synthesis using mixed integer programming,” IEEE Trans. Image Proc. 4(9), 1252–1257 (1995). [CrossRef]

]. Liu and Zakhor developed a binary and phase shifting mask (PSM) design strategy based on the branch and bound algorithm and simulated annealing [9

9. Y. Liu and A. Zakhor, “Binary and phase shifting mask design for optical lithography,” IEEE Trans. Semicond. Manuf. 5(2), 138–152 (1992). [CrossRef]

]. Erdmann proposed the optimization of the mask and illumination parameters with a genetic algorithm [10

10. A. Erdmann, R. Farkas, T. Fühner, B. Tollküuhn, and G. Kokai, “Towards automatic mask and source optimization for optical lithography,” Proc. SPIE 5377, 646–657 (2004). [CrossRef]

]. Granik described and compared solutions of inverse mask problems [11

11. Y. Granik, “Solving inverse problems of optical microlithography,” Proc. SPIE 5754, 506–526 (2004). [CrossRef]

,12

12. Y. Granik, “Fast pixel-based mask optimization for inverse lithography,” J. Microlithogr., Microfabr., Microsyst. 5, 043002 (2006). [CrossRef]

]. All of the methods mentioned above, however, are not based on gradient type optimization and thus the searching process for a suitable solution is either computationally expensive or not efficient.

Poonawala and Milanfar recently introduced a PBOPC optimization framework for coherent illumination sources relying on the SD algorithm [13

13. A. Poonawala and P. Milanfar, “Fast and low-complexity mask design in optical microlithography–an inverse imaging problem,” IEEE Trans. Image Process. 16(3), 774–788 (2007). [CrossRef] [PubMed]

]. Ma and Arce generalized this algorithm to partially coherent illumination (PCI) systems [14

14. X. Ma and G. R. Arce, “Binary mask optimization for inverse lithography with partially coherent illumination,” J. Opt. Soc. Am. A 25(12), 2960–2970 (2008). [CrossRef]

]. Subsequently, Ma and Arce proposed a three-dimensional OPC optimization algorithm, and a simultaneous source and mask optimization algorithm [15

15. X. Ma and G. R. Arce, “Binary mask optimization for forward lithography based on boundary layer model in coherent systems,” J. Opt. Soc. Am. A 26(7), 1687–1695 (2009). [CrossRef]

, 16

16. X. Ma and G. R. Arce, “Pixel-based simultaneous source and mask optimization for resolution enhancement in optical lithography,” Opt. Express 17(7), 5783–5793 (2009). [CrossRef] [PubMed]

]. In each iteration of the gradient-based OPC optimization, the aerial image needs to be recalculated due to the updated pixel values on the mask [14

14. X. Ma and G. R. Arce, “Binary mask optimization for inverse lithography with partially coherent illumination,” J. Opt. Soc. Am. A 25(12), 2960–2970 (2008). [CrossRef]

]. The aerial image calculation is, in particular, the most time-consuming step in the OPC optimization process. Thus, the overall runtime of the gradient-based OPC optimization is approximately proportional to the iteration number. It has been proven that the SD solution linearly converges to the optimal solution with respect to the iteration number, while the Newton’s method has a quadratic convergence speed, and the nonlinear CG method converges in linear or superlinear speed [17

17. J. Nocedal and S. J. Wright, Numerical Optimization, 2nd ed. (Springer, 2006).

]. In practice, the CG method has much faster convergence than SD. Since the SD algorithm has very slow convergence characteristics, as compared to other efficient optimization algorithms, the OPC SD optimization framework is not efficient, and other approaches such as the CG method and Newton’s method are therefore needed [17

17. J. Nocedal and S. J. Wright, Numerical Optimization, 2nd ed. (Springer, 2006).

]. Furthermore, unlike the Newton-class of optimization algorithms, the CG method only needs the first derivative of the cost function [17

17. J. Nocedal and S. J. Wright, Numerical Optimization, 2nd ed. (Springer, 2006).

]. This property makes the CG method as simple to implement as the SD algorithm. In our future work, we plan to study and develop an in-depth analysis of the convergence properties of CG algorithms, and compose a separate manuscript devoted to this topic exclusively. To this end, this paper develops OPC optimization based on the CG method to seek optimal OPC mask patterns. Compared to the SD algorithm, the CG method often leads to greater image fidelity using much less iterations, thus effectively reducing the overall runtime. A partially coherent illumination system model is assumed the partially coherent imaging system is represented by the Fourier series expansion model [14

14. X. Ma and G. R. Arce, “Binary mask optimization for inverse lithography with partially coherent illumination,” J. Opt. Soc. Am. A 25(12), 2960–2970 (2008). [CrossRef]

, 18

18. B. E. A. Saleh and M. Rabbani, “Simulation of partially coherent imagery in the space and frequency domains and by modal expansion,” Appl. Opt. 21(15), 2770–2777 (1982). [CrossRef] [PubMed]

].

A second contribution of the paper is to introduce a mask rule check (MRC) penalty to obtain more desirable manufacturability characteristics for the optimized mask. Currently, a set of regularization and post-processing techniques have been proposed to reduce the mask complexity [13

13. A. Poonawala and P. Milanfar, “Fast and low-complexity mask design in optical microlithography–an inverse imaging problem,” IEEE Trans. Image Process. 16(3), 774–788 (2007). [CrossRef] [PubMed]

, 19

19. X. Ma and G. R. Arce, “Generalized inverse lithography methods for phase-shifting mask design,” Opt. Express 15, 15066–15079 (2007). [CrossRef] [PubMed]

, 20

20. X. Ma and G. R. Arce, “PSM design for inverse lithography with partially coherent illumination,” Opt. Express 16 (24), 20126–20141 (2008). [CrossRef] [PubMed]

]. However, they are inadequate to generate MRC-favorable optimized masks. MRC is used currently in the mask design process to constrain the mask geometries [21

21. K. Kato, K. Nishizawa, and T. Inoue, “Advanced mask rule check (MRC) tool,” Proc. SPIE 6283, 62830O (2006). [CrossRef]

, 22

22. K. Kato, Y. Taniguchi, K. Nishizawa, and M. Endo, “Mask rule check using priority information of mask patterns,” Proc. SPIE 6730, 67304F (2007). [CrossRef]

]. Usually, the minimum linear size of the SRAFs, and the distance from the SRAFs to the main body of the mask are forced to be above some prescribed thresholds. However, most of the existing MRC methods are applied to the post-optimized mask pattern, and serve as post-processing steps, which lead to sub-optimal solutions. In contrast, the proposed approach in this paper embeds the MRC penalty in the cost function resulting in MRC-favorable optimal masks.

A third contribution of this paper is the development of a projection method to further removes the small SRAFs. The projection method projects the pixel values on the mask between the subspace of the optimal solution and the subspace supported by the low frequency components in two-dimensional discrete cosine transform (2D-DCT) domain. Simulation shows that the resulting optimized OPC has much lower complexity.

The remainder of the paper is organized as follows. The partially coherent imaging model is discussed in Section 2. The OPC optimization process based on the CG method under partially coherent system is developed in Section 3. The MRC penalty is introduced in Section 4. The projection method is proposed in Section 5. Conclusions are provided in Section 6.

2. Partially Coherent Imaging Model

2.1. Partially Coherent Illumination System

Partially coherent illumination has been shown to improve the theoretical resolution limit in lithography. Thus, practical lithography systems often operate under partially coherent illumination due to non-zero width sources and off-axis illumination from spatially extended sources. A schematic of an optical lithography system with partially coherent illumination is illustrated in Fig. 1. The light source with a wavelength of λ is placed at the focal plane of the first condenser, illuminating the mask. Common illuminations sources include dipole, quadrupole and annular shapes, all introducing partial coherence. The image of the photomask is formed by the projection optics onto the wafer [1

1. A. K. Wong, Resolution Enhancement Techniques (SPIE Press, 2001), Vol. 1. [CrossRef]

]. The partial coherence factor σ=ab is defined as the ratio between the size of the source image and the pupil. Radiation of partially coherent light has been shown to be described as an expansion of coherent modes added incoherently in the image plane [18

18. B. E. A. Saleh and M. Rabbani, “Simulation of partially coherent imagery in the space and frequency domains and by modal expansion,” Appl. Opt. 21(15), 2770–2777 (1982). [CrossRef] [PubMed]

].

Fig. 1 Optical lithography system with partially coherent illuminations.

According to the Hopkins imaging model, the light intensity distribution exposed on the wafer in the PCI system is bilinear and described by [2

2. X. Ma and G. R. Arce, Computational Lithography, Wiley Series in Pure and Applied Optics, 1st ed. (John Wiley and Sons, 2010). [CrossRef]

]
I(r)=M(r1)M(r2)γ(r1r2)h*(rr1)h(rr2)dr1dr2,
(1)
where r = (x,y) is the coordinate on the image plane located on the wafer. r1 = (x1, y1) and r2 = (x2, y2) are the coordinates on the object plane located on the mask. M(r) is the mask pattern, γ(r1r2) is the complex degree of coherence, and h(r) represents the amplitude impulse response of the optical system. The complex degree of coherence γ(r1r2) is generally a complex number, whose magnitude represents the extent of optical interaction between two spatial locations r1 = (x1, y1) and r2 = (x2, y2) of the light source [1

1. A. K. Wong, Resolution Enhancement Techniques (SPIE Press, 2001), Vol. 1. [CrossRef]

]. The complex degree of coherence in the spatial domain is the inverse 2-D Fourier transform of the illumination shape. The Hopkins imaging model assumes that the light diffraction from the masks is invariant with respect to the direction of the incident light. The diffraction spectrum for oblique incidence is obtained by a simple shift of the diffraction spectrum for vertical incident light. This assumption is less accurate for modern high NA projection imaging systems using strong off-axis illuminations. In such cases, PBOPC optimization with more rigorous imaging models needs to be explored as this case is left for future research work. In general, Eqn. (1) is tedious to compute, both analytically and numerically [23

23. B. Salik, J. Rosen, and A. Yariv, “Average coherent approximation for partially cohernet optical systems,” J. Opt. Soc. Am. A 13(10, 2086–2090) (1996). [CrossRef]

]. In the following, the Fourier series expansion model is applied to formulate the imaging process of a partially coherent system in the discrete domain, and accelerate the calculation of Eqn. (1) [18

18. B. E. A. Saleh and M. Rabbani, “Simulation of partially coherent imagery in the space and frequency domains and by modal expansion,” Appl. Opt. 21(15), 2770–2777 (1982). [CrossRef] [PubMed]

].

2.2. Fourier Series Expansion Model

Assume the mask is constrained in the square area A defined by x,y[D2,D2]. Thus, for the computations involved in Eqn. (1), the only values of γ(r) needed are those inside the square area Aγ defined by x,y ∈ [−D,D]. Applying the 2-D Fourier series expansion, γ(r) can be rewritten as
γ(r)=mΓmexp(jω0mr),
(2)
and
Γm=1D2Aγγ(r)exp(jω0mr)dr,
(3)
where ω0 = π/D, m = (mx, my), mx and my are integers, and · is the inner-product operation. Substituting Eqn. (2) into Eqn. (1), the light intensity on the wafer is given by where
I(r)=mΓm|M(r)hm(r)|2,
(4)
where
hm(r)=h(r)exp(jω0mr).
(5)
It is observed from Eqn. (4) that PCI leads to the superposition of coherent systems. Since the Fourier series expansion model is based on direct discretization of the Hopkins imaging model, they have the same accuracy. For the annular illumination, the complex degree of coherence is
γ(r)=J1(2πr/2Dcu)2πr/2DcuDcu2Dcl2J1(2πr/2Dcl)2πr/2Dcl,
(6)
where r=x2+y2. The corresponding Fourier series coefficients are
Γm={4Dcu2Dcl2πD2(Dcl2Dcu2)forD/2Dcl|m|D/2Dcu0elsewhere,
(7)
where Dcl and Dcu are the coherent lengths of the inner and outer circles respectively. σinner=λ2DclNA and σouter=λ2DcuNA are the corresponding inner and outer partial coherence factors. The convolution kernel h(r) is defined as the Fourier transform of the circular lens aperture with cutoff frequency NA/λ[24

24. M. Born and E. Wolfe, Principles of Optics (Cambridge University Press, 1999).

, 25

25. R. Wilson, Fourier Series and Optical Transform Techniques in Contemporary Optics (John Wiley and Sons, 1995).

]; therefore,
h(r)=J1(2πrNA/λ)2πrNA/λ.
(8)

3. Inverse Lithography Optimization

The forward imaging process is illustrated in Fig. 2, where the Fourier series expansion model is used in the image formation stage, and a hard threshold function is used to approximate the photoresist effect. The output pattern of the optical system is binary. Further, since the derivative of the sigmoid function exists, it is used to approximate the hard threshold function. The sigmoid function is
sig(x)=11+exp[a(xtr)],
(11)
where tr is the process threshold, and a dictates the steepness of the sigmoid function.

Fig. 2 Approximated forward process model.

Following the definitions above, the following notations are used:
  1. The MN×N matrix represents the mask pattern with a N2 × 1 equivalent raster scanned vector representation, denoted as m.
  2. A convolution matrix Hm is a N2 × N2 matrix with an equivalent two-dimensional filter hm.
  3. The desired N × N binary output pattern is denoted as . It is the desired light distribution sought on the wafer. Its vector representation is denoted as .
  4. The output of the sigmoid function is the N × N image denoted as:
    Z=sig{mΓm|Hm{m_}|2}.
    (12)
    The equivalent vector is denoted as z.
  5. The hard threshold version of Z is the binary output pattern denoted as Zb. Its equivalent vector is denoted as zb, with all entries constrained to 0 or 1.
  6. The optimized N × N gray mask denoted as minimizes the distance between Z and , ie,
    M^=argmindM(sig{mΓm|Hm{m_}|2},Z˜).
    (13)
    Its equivalent vector is denoted as [0,1].
  7. The binary optimized mask b is the quantization of . Its equivalent vector is denoted as b, with all entries constrained to 0 or 1.

Given the gray level pattern z_=sig{mΓm|Hm{m_}|2}, the ith entry in this vector can be represented as
z_i=11+exp[amΓm|j=1N2hijmm_j|2+atr],i=1,N2,
(14)
where hijm is the i, jth entry of the filter. In the optimization process, is searched to minimize the square of the L2 norm of the difference between z and . Therefore,
m^_=argminm^_{F(m_)},
(15)
where the cost function F(·) is defined as:
F(m_)=z˜_z_22=i=1N2(z˜_iz_i)2,
(16)
where ‖ · ‖2 is the L2 norm and zi is represented in Eqn. (14). In order to reduce the above bound-constrained optimization problem to an unconstrained optimization problem, we adopt the parametric transformation [13

13. A. Poonawala and P. Milanfar, “Fast and low-complexity mask design in optical microlithography–an inverse imaging problem,” IEEE Trans. Image Process. 16(3), 774–788 (2007). [CrossRef] [PubMed]

]. Let
m_j=1+cos(θ_j)2,j=1,N2,
(17)
where θj ∈ (−∞, ∞) and mj ∈ [0,1]. Defining the vector θ = [θ1,...,θN2]T, the optimization problem is formulated as
θ^_=argminθ_{F(θ_)}=argminθ_{i=1N2(z˜i11+exp[amΓm|j=1N2hijm1+cosθ_j2|2+atr])2}.
(18)

The CG method is used to optimize the above problem. CG method was originally designed to solve linear systems with positive definite coefficient matrices [26

26. M. R. Hestenes and E. Stiefel, “Methods of conjugate gradients for solving linear systems,” J. Res. Natl. Bur. Stand. 49, 409–436 (1952).

]. Fletcher and Reeves generalized the linear CG method to nonlinear problems [27

27. R. Fletcher and C. M. Reeves, “Function minimization by conjugate gradients,” Comput. J. 7, 149–154 (1964). [CrossRef]

]. The CG method generates a set of gradient estimate vectors {P0, P1,...,Pl} with some properties, along which the cost function is successively minimized. Each vector Pi is updated from Pi−1, and is the linear combination of the past gradients. Thus, each new minimizing direction Pi inherits the useful information from its ancestor. In the past, a variety of CG methods have been proposed, such as Fletcher-Reeves method, Polak-Ribière method, and so on [17

17. J. Nocedal and S. J. Wright, Numerical Optimization, 2nd ed. (Springer, 2006).

]. In this paper, we adopt the Fletcher-Reeves method. The comparison among these CG algorithms and the method to select the proper algorithm are out of the scope of this paper.

Let Θ ∈ ℜN×N is the equivalent matrix of θ. The gradients ΔF(Θ) can be calculated as follows [14

14. X. Ma and G. R. Arce, “Binary mask optimization for inverse lithography with partially coherent illumination,” J. Opt. Soc. Am. A 25(12), 2960–2970 (2008). [CrossRef]

]:
F(Θ)=a×sinΘ{mΓm(hm)[(Z˜Z)Z(1N×NZ)((hm)*M)]}+a×sinΘ{mΓm(hm)*[(Z˜Z)Z(1N×NZ)(hmM)]},
(19)
where ΔF(Θ) ∈ ℜN×N, ⊙ is the element-by-element multiplication operator, (hm)° rotates hm by 180° degrees in both row and column directions, * is the conjugate operation, and 1N×N ∈ ℜN×N has all entries equal to 1. Assuming Θk is the kth iteration result, then the CG method is summarized as following:

Iteration 0: Given the desired binary mask pattern , calculate
Θ0(i,j)={4π5forZ˜(i,j)=0π5forZ˜(i,j)=1,i,j=1,2,,N,
(20)
According to Eqn. (19), assigning Θ0(i, j) = π or Θ0(i, j) = 0 would degrade the gradient to 0 and therefore the iteration would be terminated. Thus, we select the angle values 4π5 and π5, which are close to the ideal values. In addition, we initialize
P0=F(Θ0),
(21)
where Θ0,P0 ∈ ℜN×N.

Stop:

Until the termination condition is reached.

In the following, we address the advantages of the CG method, by comparing it to the performance of SD method. In this simulation, annular illumination is applied with the inner and outer partial coherence factors σinner = 0.3 and σouter = 0.4. For both methods, the convolution kernel is shown in Eqn. (8) with NA = 1.25 and λ = 193nm. We assume h(r) vanishes outside the area Ah defined by x, y ∈ [−56.25nm, 56.25nm]. In the sigmoid function, we assign the parameters a = 25 and tr = 0.19. The global threshold is tm = 0.5, the pixel size is 5.625nm × 5.625nm. Generally, smaller step lengths lead to stable but slower convergence, while larger step lengths lead to faster convergence or divergence. Based on this guideline, we set the step length equal to 1, which is moderate for the underlying optimization problem.

ILT optimization is an ill-posed problem, where numerous inputs may result in the same output. Therefore, regularization approaches are usually applied to bias the solution space to sample solutions with some favorable properties [31

31. C. Vogel, Computational Methods for Inverse Problems (SIAM Press, 2002). [CrossRef]

]. Regularization is formulated as follows:
m^_=argminm^_{F(m_)+γR(m_)},
(28)
where F(m) is the data-fidelity term and R(m) is the regularization term which is used to reduce the solution space and constrain the optimized results. γ is the user-defined parameter to reveal the weight of the regularization. In the above simulation, the quadratic penalty [13

13. A. Poonawala and P. Milanfar, “Fast and low-complexity mask design in optical microlithography–an inverse imaging problem,” IEEE Trans. Image Process. 16(3), 774–788 (2007). [CrossRef] [PubMed]

] and the wavelet penalty [19

19. X. Ma and G. R. Arce, “Generalized inverse lithography methods for phase-shifting mask design,” Opt. Express 15, 15066–15079 (2007). [CrossRef] [PubMed]

] are applied to reduce the pattern error introduced by post-processing described in Eqn. (26), and to reduce the mask complexity, respectively. The formulae of the quadratic penalty and the wavelet penalty are summarized in Appendix A. Thus, the cost function of Eqn. (16) is adjusted as
J(m_)=F(m_)+γQRQ(m_)+γWRW(m_).
(29)
However, making γQ and γW too large can result in divergence of the optimization process. In the following simulations, the regularization weights of the quadratic penalty and wavelet penalty are γQ = 0.01 and γW = 0.025, respectively.

Figure 3 shows the performance comparison between the masks optimized with the SD and CG methods, respectively. Figure 3(a) is the desired pattern with dimension of 1035nm × 1035nm, and Fig. 3(d) is its output pattern on the wafer with pattern error of 5052. Figure 3(b) is the optimized OPC pattern from the SD method, and Fig. 3(e) is its output pattern with pattern error of 1368. Figure 3(c) is the optimized OPC pattern from the CG method, and Fig. 3(f) is its output pattern with pattern error of 1288. Compared to the SD method, the CG method reduces the pattern error by 6%

Fig. 3 Performance comparison between the SD method and the CG method; (a) desired pattern ; (b) optimized OPC pattern from the SD method; (c) optimized OPC pattern from the CG method; (d) the output pattern when (a) is used as input; (e) the output pattern when (b) is used as input; (f) the output pattern when (c) is used as input.

Figure 4 illustrates the convergence comparison, where x-axis is the iteration number and y-axis is the pattern error. In Fig. 4, the convergence of SD and CG methods are shown in black solid line and blue dashed line, respectively. The SD method converges to pattern error of 1368 after 150 iterations using 996 seconds, while the CG method converges to a pattern error of 1288 after 33 iterations using 220 seconds. The performance and convergence comparison between the two approaches are summarized in the 2nd to 3rd rows of Table 1.

Fig. 4 Convergence comparison of the SD method (black solid line), the CG method (blue dashed line), the CG method with MRC penalty (green dotted line), and the projection method (red dot-dashed line).

Table 1. Comparison of performance and convergence

table-icon
View This Table

4. MRC Penalty

Although the wavelet penalty has been applied in the above simulation, numerous small opening and blocking areas are still contained in Fig. 3(b) and 3(c). In addition, the SRAFs are too close to the main body of the mask pattern. These characteristics make the mask fabrication more difficult. Figure 5 illustrates the influence of the wavelet penalty weight on the optimized OPCs and their output patterns. Figures 5(a), 5(b) and 5(c) show the optimized OPCs with γW = 0.025, 0.035 and 0.045, respectively. Figures 5(d), 5(e) and 5(f) show the output patterns corresponding to their above OPC patterns. It is shown that increasing the wavelet penalty weight γW cannot enlarge the dimensions of the SRAFs or the distances between the SRAFs and the main body of the mask. In addition, larger wavelet penalty weight will increase the pattern error. Particularly, Fig. 5(f) shows that when γW = 0.045, some bridges appear in the output pattern. Therefore, the wavelet penalty falls short to generate MRC-favorable optimized masks.

Fig. 5 Influence of the wavelet penalty weight on the optimized OPCs and their output patterns; (a) optimized OPC pattern with γW = 0.025; (b) optimized OPC pattern with γW = 0.035; (c) optimized OPC pattern with γW = 0.045; (d) the output pattern when (a) is used as input; (e) the output pattern when (b) is used as input; (f) the output pattern when (c) is used as input.

In order to make the solution have more desirable manufacturability properties, a MRC penalty is proposed in this section. MRC is used currently in the mask design process to constrain the mask geometries. Usually, the minimum linear size of the SRAFs, and the distance from the SRAFs to the main body of the mask are forced to be above some prescribed thresholds. However, most of the existing MRC methods are applied to the post-optimized mask pattern, and serve as post-processing steps, which lead to sub-optimal solutions. In contrast, the proposed MRC penalty in this paper is embedded in the optimization cost function. The MRC penalty term is formulated as
RM(M)=1N×1T[(MtmN×N)(gM)]1N×1,
(30)
where tmN×N ∈ ℜN×N has all entries equal to tm, g is referred to as the MRC filter, gM is a weighted sum of each pixel value and its neighbors, and MtmN×N judges whether the pixel value is above tm and carved on the binary mask or not. tm is a constant typically defined as 0.5. The meaning of the MRC penalty is explained as follows. Consider the value of a pixel located at (x0, y0) that is larger than tm, i.e., p(x0, y0) > tm. Then, (MtmN×N)(x0, y0) > 0. If we want to obtain a low penalty value, we need to keep (gM)(x0, y0) as large as possible, which means that we want all pixel values around p(x0, y0) to be larger than tm. The similar analysis applies for the case of p(x0, y0) < tm. In other word, it is preferable to have the value of p(x,y) −tm around the location of (x0, y0) have the same sign of p(x0, y0) − tm, thus both of the mask opening and blocking areas are concentrated and enlarged. In our simulations, tm = 0.5, and g = 13×3, which means that we equally sum each pixel value with its eight neighboring pixel values. Through this way, the linear sizes of the SRAFs, and the distances from the SRAFs to the main body of the mask will be simultaneously increased. Eqn. (30) can be reformulated as
RM(m_)=k=1N2(tmm_k)(j=1N2gk,jm_j).
(31)
As derived in Appendix B, the gradient of RM (Θ) is
RM(Θ)=(4.52gM)(0.5sinΘ),
(32)
thus, the cost function in Eqn. (29) is adjusted as
J(m_)=F(m_)+γQRQ(m_)+γWRW(m_)+γMRM(m_),
(33)
Figure 6(a) shows the resulting OPC pattern by optimizing Eqn. (33), where the desired pattern is the same as Fig. 3(a) and γM = 0.005. Figure 6(b) shows the output pattern corresponding to Fig. 6(a). The pattern error, iteration number and the runtime are summarized in the 4th row of Table 1. The convergence of this simulation is shown by the green dotted line in Fig. 4. It is observed that the MRC penalty effectively enlarges the linear sizes of SRAFs and the distances from the SRAFs to the main body of the mask. As a tradeoff, the pattern error is increased by 21%.

Fig. 6 Performance comparison between the the CG method with MRC penalty and the projection method; (a) optimized OPC pattern from the CG method with MRC penalty; (b) the output pattern when (a) is used as input; (c) optimized OPC pattern from the projection method; (d) the output pattern when (c) is used as input.

5. Projection Method

The MRC penalty proposed in Section 4 is an effective way to seek the MRC-favorable optimal solution of the ILT problem. However, it is observed from Fig. 6(a) that some small SRAFs still remain in the OPC pattern. If we extend the dimension of the MRC filter g in Eqn. (30), the pattern error will be dramatically increased, while the number of the small SRAFs cannot be further reduced. This phenomenon is illustrated in Fig. 7. Figures 7(a), 7(b) and 7(c) show the optimized OPCs with g = 13×3, 15×5 and 17×7, respectively. Figures 7(d), 7(e) and 7(f) show the output patterns corresponding to their above OPC patterns. It is shown that increasing the MRC filter dimension leads to more details on the optimized OPC. In addition, larger MRC filter dimension will increase the pattern error. One approach to ameliorate the pattern distortions as the window size of the MRC filter is increased is to replace the linear filter by a median-type edge-preserving filter [32

32. M. P. McLoughlin and G. R. Arce, “Deterministic properties of the recursive separable median filter,” IEEE Trans. Acoust., Speech, Signal Process. 35, 98–106 (1987). [CrossRef]

34

34. K. Barner and G. R. Arce, “Permutation filters - a class of nonlinear filters based on set permutations,” IEEE Trans. Signal Process 42, 782–798 (1994). [CrossRef]

]. The cost function representations of such nonlinear filters, however, are not differentiable and thus are not easily integrated into the optimization.

Fig. 7 Influence of the MRC filter dimension on the optimized OPCs and their output patterns; (a) optimized OPC pattern with g = 13×3; (b) optimized OPC pattern with g = 15×5; (c) optimized OPC pattern with g = 17×7; (d) the output pattern when (a) is used as input; (e) the output pattern when (b) is used as input; (f) the output pattern when (c) is used as input.

In order to further improve the manufacturability of the optimized OPC pattern, a projection method is proposed hereafter. Optical lithography system has a low-pass filter property that cuts off the high frequency components of the mask pattern, and constructs the print image on the wafer by the low frequency components. It has been shown that only low frequency components of the mask in the 2D-DCT domain significantly influence the image formation of the optical lithography system [20

20. X. Ma and G. R. Arce, “PSM design for inverse lithography with partially coherent illumination,” Opt. Express 16 (24), 20126–20141 (2008). [CrossRef] [PubMed]

]. Based on this principle, the proposed projection method is depicted as Fig. 8.

Fig. 8 Scheme of the projection method.

In Fig. 8, the input is the initial pattern Θ0 as described in Eqn. (20). In stage I, the CG method with the MRC penalty is applied to obtain a raw optimized gray-level mask pattern . If satisfies the prescribed termination conditions, we threshold by tm to obtain the final optimized binary mask b. Otherwise, we move to stage II. In stage II, we cut off the high frequency components of in the 2D-DCT domain, resulting in M̂′. Subsequently, in stage III we threshold M̂′ to obtain M̂′b. In addition, we update Θ according to M̂′b using method described in Eqn. (20). The updated Θ is used as the input of the stage I to continue the loop. Let SF represent the subspace of the optimal solutions of Eqn. (10). SDCT represents the subspace supported by the low frequency components in the 2D-DCT domain. The proposed method iteratively projects the pixel values on the mask between SF and SDCT, and drives the final solution into the intersection between the two subspaces.

Figure 6(c) shows the resulting OPC pattern from the projection method, and Fig. 6(d) shows the corresponding output pattern. The pattern error, iteration number and the runtime are summarized in the 5th row of Table 1. In stage I, the CG method is iterated for 18 iterations in each projection loop. In stage II, we maintain 5995 low frequency components. The projection loop is terminated when the pattern error is reduced to a tolerable level, or the overall iteration number reaches the prescribed upper limit. The convergence of this simulation is shown by the red dot-dashed line in Fig. 4. It is observed that the projection method effectively improves the manufacturability of the mask. On the other hand, the pattern error is unavoidably increased by 20.6% compared to the CG method with MRC penalty.

6. Conclusion

This paper proposes a set of approaches for inverse lithography based on the CG method. The Fourier series expansion model is used to formulate the optimization problem. Compared to the SD algorithm, the CG method results in lower pattern error and faster convergence speed. Subsequently, the MRC penalty and the projection method are proposed to further reduce the mask complexity. As a tradeoff, the pattern errors are increased caused by the absence of numerous SRAFs. Simulations illustrate that our approaches are effective and practical.

A. Appendix A

The quadratic penalty term is RQ(m) = 4mT (1m). For each pixel value, the corresponding penalty is the quadratic function r(mi)=1 − (2mi − 1)2, i = 1,...,N2. The gradient of Rq(m) is RQ(Θ)=12(8M+4)sin(Θ).

The wavelet penalty term is
RW(m_)=h112+h122+h(N2)(N2)2+v112+v122+v(N2)(N2)2+d112+d122+d(N2)(N2)2,
(34)
where
hij=m(2(i1)+1)(2(j1)+1)m(2(i1)+1)(2(j1)+2)+m(2(i1)+2)(2(j1)+1)m(2(i1)+2)(2(j1)+2),
(35)
vij=m(2(i1)+1)(2(j1)+1)+m(2(i1)+1)(2(j1)+2)m(2(i1)+2)(2(j1)+1)m(2(i1)+2)(2(j1)+2),
(36)
dij=m(2(i1)+1)(2(j1)+1)m(2(i1)+1)(2(j1)+2)m(2(i1)+2)(2(j1)+1)+m(2(i1)+2)(2(j1)+2),
(37)
for i,j=1,,N2. The gradient of the wavelet penalty is given as
RWθ_(2(i1)+p)(2(j1)+q)=12sinθ_(2(i1)+p)(2(j1)+q)×(3m_(2(i1)+p)(2(j1)+q)m_(2(i1)+p1)(2(j1)+q)m_(2(i1)+p)(2(j1)+q1)m_(2(i1)+p1)(2(j1)+q1)),
(38)
where i,j=1,,N2,p,q=1 or 2, p1 = (p + 1) mod 2 and q1 = q + 1) mod 2.

B. Appendix B

The derivation of Eqn. (32) is as following:
RM(m_)m_k=p=1N2[(tmm_p)(j=1N2gpjm_j)]/m_k=[(tmm_k)(j=1N2gkjm_j)]/m_k+p=1,pkN2[(tmm_p)(j=1N2gpjm_j)]/m_k=[(tmm_k)(j=1,jkN2gkjm_j)]/m_k+[(tmm_k)gkkm_k]/m_k+p=1,pkN2[(tmm_p)gpkm_k]/m_k.
Since g = 13×3, gkk = 1 and gpk = gkp. Therefore,
RM(m_)m_k=(j=1,jkN2gkjm_j)+(tmm_km_k2)/m_k+p=1,pkN2gpk(tmm_p)=(j=1,jkN2gkjm_j)+tm2m_k(p=1,pkN2gkpm_p)+tmp=1,pkN2hpk=2(j=1N2gkjm_j)+2gkkm_k+tm2m_k+tmp=1N2hpktmhkk=2(j=1N2gkjm_j)+tmp=1N2hpk.
Since tm = 0.5 and p=1N2hpk=9, RM(M)=(4.52gM). Therefore,
RM(Θ)=RM(M)M(Θ)=(4.52gM)(0.5sinΘ).
(39)

References and links

1.

A. K. Wong, Resolution Enhancement Techniques (SPIE Press, 2001), Vol. 1. [CrossRef]

2.

X. Ma and G. R. Arce, Computational Lithography, Wiley Series in Pure and Applied Optics, 1st ed. (John Wiley and Sons, 2010). [CrossRef]

3.

S. A. Campbell, The Science and Engineering of Microelectronic Fabrication, 2nd ed. (Publishing House of Electronics Industry, 2003).

4.

F. Schellenberg, “Resolution enhancement technology: the past, the present, and extensions for the future,” Proc. SPIE 5377, 1–20 (2004). [CrossRef]

5.

F. Schellenberg, Resolution Enhancement Techniques in Optical Lithography (SPIE Press, 2004).

6.

L. Liebmann, S. Mansfield, A. Wong, M. Lavin, W. Leipold, and T. Dunham, “TCAD development for lithography resolution enhancement,” IBM J. Res. Dev. 45, 651–665 (2001). [CrossRef]

7.

U. Okoroanyanwu, Chemistry and Lithography, 1st ed. (in review, SPIE Press, 2010).

8.

S. Sherif, B. Saleh, and R. Leone, “Binary image synthesis using mixed integer programming,” IEEE Trans. Image Proc. 4(9), 1252–1257 (1995). [CrossRef]

9.

Y. Liu and A. Zakhor, “Binary and phase shifting mask design for optical lithography,” IEEE Trans. Semicond. Manuf. 5(2), 138–152 (1992). [CrossRef]

10.

A. Erdmann, R. Farkas, T. Fühner, B. Tollküuhn, and G. Kokai, “Towards automatic mask and source optimization for optical lithography,” Proc. SPIE 5377, 646–657 (2004). [CrossRef]

11.

Y. Granik, “Solving inverse problems of optical microlithography,” Proc. SPIE 5754, 506–526 (2004). [CrossRef]

12.

Y. Granik, “Fast pixel-based mask optimization for inverse lithography,” J. Microlithogr., Microfabr., Microsyst. 5, 043002 (2006). [CrossRef]

13.

A. Poonawala and P. Milanfar, “Fast and low-complexity mask design in optical microlithography–an inverse imaging problem,” IEEE Trans. Image Process. 16(3), 774–788 (2007). [CrossRef] [PubMed]

14.

X. Ma and G. R. Arce, “Binary mask optimization for inverse lithography with partially coherent illumination,” J. Opt. Soc. Am. A 25(12), 2960–2970 (2008). [CrossRef]

15.

X. Ma and G. R. Arce, “Binary mask optimization for forward lithography based on boundary layer model in coherent systems,” J. Opt. Soc. Am. A 26(7), 1687–1695 (2009). [CrossRef]

16.

X. Ma and G. R. Arce, “Pixel-based simultaneous source and mask optimization for resolution enhancement in optical lithography,” Opt. Express 17(7), 5783–5793 (2009). [CrossRef] [PubMed]

17.

J. Nocedal and S. J. Wright, Numerical Optimization, 2nd ed. (Springer, 2006).

18.

B. E. A. Saleh and M. Rabbani, “Simulation of partially coherent imagery in the space and frequency domains and by modal expansion,” Appl. Opt. 21(15), 2770–2777 (1982). [CrossRef] [PubMed]

19.

X. Ma and G. R. Arce, “Generalized inverse lithography methods for phase-shifting mask design,” Opt. Express 15, 15066–15079 (2007). [CrossRef] [PubMed]

20.

X. Ma and G. R. Arce, “PSM design for inverse lithography with partially coherent illumination,” Opt. Express 16 (24), 20126–20141 (2008). [CrossRef] [PubMed]

21.

K. Kato, K. Nishizawa, and T. Inoue, “Advanced mask rule check (MRC) tool,” Proc. SPIE 6283, 62830O (2006). [CrossRef]

22.

K. Kato, Y. Taniguchi, K. Nishizawa, and M. Endo, “Mask rule check using priority information of mask patterns,” Proc. SPIE 6730, 67304F (2007). [CrossRef]

23.

B. Salik, J. Rosen, and A. Yariv, “Average coherent approximation for partially cohernet optical systems,” J. Opt. Soc. Am. A 13(10, 2086–2090) (1996). [CrossRef]

24.

M. Born and E. Wolfe, Principles of Optics (Cambridge University Press, 1999).

25.

R. Wilson, Fourier Series and Optical Transform Techniques in Contemporary Optics (John Wiley and Sons, 1995).

26.

M. R. Hestenes and E. Stiefel, “Methods of conjugate gradients for solving linear systems,” J. Res. Natl. Bur. Stand. 49, 409–436 (1952).

27.

R. Fletcher and C. M. Reeves, “Function minimization by conjugate gradients,” Comput. J. 7, 149–154 (1964). [CrossRef]

28.

H. P. Crowder and P. Wolfe, “Linear convergence of the conjugate gradient method,” IBM J. Res. Dev. 16, 431–433 (1972). [CrossRef]

29.

A. Cohen, “Rate of convergence of several conjugate gredient algorithms,” SIAM J. Numer. Anal. 9, 248–259 (1972). [CrossRef]

30.

K. Ritter, “On the rate of superlinear convergence of a class of variable metric methods,” Numerische Mathematik 35, 293–313 (1980). [CrossRef]

31.

C. Vogel, Computational Methods for Inverse Problems (SIAM Press, 2002). [CrossRef]

32.

M. P. McLoughlin and G. R. Arce, “Deterministic properties of the recursive separable median filter,” IEEE Trans. Acoust., Speech, Signal Process. 35, 98–106 (1987). [CrossRef]

33.

S. Kalluri and G. R. Arce, “Fast algorithms for weighted myriad computation by fixed-point search,” IEEE Trans. Signal Process 48(1), 159–171 (2000). [CrossRef]

34.

K. Barner and G. R. Arce, “Permutation filters - a class of nonlinear filters based on set permutations,” IEEE Trans. Signal Process 42, 782–798 (1994). [CrossRef]

OCIS Codes
(100.3190) Image processing : Inverse problems
(110.4980) Imaging systems : Partial coherence in imaging
(110.5220) Imaging systems : Photolithography

ToC Category:
Image Processing

History
Original Manuscript: November 10, 2010
Revised Manuscript: December 25, 2010
Manuscript Accepted: December 26, 2010
Published: January 20, 2011

Citation
Xu Ma and Gonzalo R. Arce, "Pixel-based OPC optimization based on conjugate gradients," Opt. Express 19, 2165-2180 (2011)
http://www.opticsinfobase.org/oe/abstract.cfm?URI=oe-19-3-2165


Sort:  Author  |  Year  |  Journal  |  Reset  

References

  1. A. K. Wong, Resolution Enhancement Techniques (SPIE Press, 2001), Vol. 1. [CrossRef]
  2. X. Ma and G. R. Arce, Computational Lithography, Wiley Series in Pure and Applied Optics, 1st ed. (JohnWiley and Sons, 2010). [CrossRef]
  3. S. A. Campbell, The Science and Engineering of Microelectronic Fabrication, 2nd ed. (Publishing House of Electronics Industry, 2003).
  4. F. Schellenberg, “Resolution enhancement technology: the past, the present, and extensions for the future,” Proc. SPIE 5377, 1–20 (2004). [CrossRef]
  5. F. Schellenberg, Resolution Enhancement Techniques in Optical Lithography (SPIE Press, 2004).
  6. L. Liebmann, S. Mansfield, A. Wong, M. Lavin, W. Leipold, and T. Dunham, “TCAD development for lithography resolution enhancement,” IBM J. Res. Develop. 45, 651–665 (2001). [CrossRef]
  7. U. Okoroanyanwu, Chemistry and Lithography, 1st ed. (in review, SPIE Press, 2010).
  8. S. Sherif, B. Saleh, and R. Leone, “Binary image synthesis using mixed integer programming,” IEEE Trans. Image Process. 4(9), 1252–1257 (1995). [CrossRef]
  9. Y. Liu and A. Zakhor, “Binary and phase shifting mask design for optical lithography,” IEEE Trans. Semicond. Manuf. 5(2), 138–152 (1992). [CrossRef]
  10. A. Erdmann, R. Farkas, T. Fühner, B. Tollkühn, and G. Kokai, “Towards automatic mask and source optimization for optical lithography,” Proc. SPIE 5377, 646–657 (2004). [CrossRef]
  11. Y. Granik, “Solving inverse problems of optical microlithography,” Proc. SPIE 5754, 506–526 (2004). [CrossRef]
  12. Y. Granik, “Fast pixel-based mask optimization for inverse lithography,” J. Microlithogr., Microfabr, Microsyst. 5, 043002 (2006). [CrossRef]
  13. A. Poonawala, and P. Milanfar, “Fast and low-complexity mask design in optical microlithography–an inverse imaging problem,” IEEE Trans. Image Process. 16(3), 774–788 (2007). [CrossRef] [PubMed]
  14. X. Ma and G. R. Arce, “Binary mask optimization for inverse lithography with partially coherent illumination,” J. Opt. Soc. Am. A 25(12), 2960–2970 (2008). [CrossRef]
  15. X. Ma and G. R. Arce, “Binary mask optimization for forward lithography based on boundary layer model in coherent systems,” J. Opt. Soc. Am. A 26(7), 1687–1695 (2009). [CrossRef]
  16. X. Ma and G. R. Arce, “Pixel-based simultaneous source and mask optimization for resolution enhancement in optical lithography,” Opt. Express 17(7), 5783–5793 (2009). [CrossRef] [PubMed]
  17. J. Nocedal and S. J. Wright, Numerical Optimization, 2nd ed. (Springer, 2006).
  18. B. E. A. Saleh and M. Rabbani, “Simulation of partially coherent imagery in the space and frequency domains and by modal expansion,” Appl. Opt. 21(15), 2770–2777 (1982). [CrossRef] [PubMed]
  19. X. Ma and G. R. Arce, “Generalized inverse lithography methods for phase-shifting mask design,” Opt. Express 15, 15066–15079 (2007). [CrossRef] [PubMed]
  20. X. Ma and G. R. Arce, “PSM design for inverse lithography with partially coherent illumination,” Opt. Express 16(24), 20126–20141 (2008). [CrossRef] [PubMed]
  21. K. Kato, K. Nishizawa, and T. Inoue, “Advanced mask rule check (MRC) tool,” Proc. SPIE 6283, 62830O (2006). [CrossRef]
  22. K. Kato, Y. Taniguchi, K. Nishizawa, and M. Endo, “Mask rule check using priority information of mask patterns,” Proc. SPIE 6730, 67304F (2007). [CrossRef]
  23. B. Salik, J. Rosen, and A. Yariv, “Average coherent approximation for partially cohernet optical systems,” J. Opt. Soc. Am. A 13, 2086–2090) (1996). [CrossRef]
  24. M. Born and E. Wolfe, Principles of Optics (Cambridge University Press, 1999).
  25. R. Wilson, Fourier Series and Optical Transform Techniques in Contemporary Optics (John Wiley and Sons, 1995).
  26. M. R. Hestenes and E. Stiefel, “Methods of conjugate gradients for solving linear systems,” J. Res. Natl. Bur. Stand. 49, 409–436 (1952).
  27. R. Fletcher and C. M. Reeves, “Function minimization by conjugate gradients,” Comput. J. 7, 149–154 (1964). [CrossRef]
  28. H. P. Crowder and P. Wolfe, “Linear convergence of the conjugate gradient method,” IBM J. Res. Develop. 16, 431–433 (1972). [CrossRef]
  29. A. Cohen, “Rate of convergence of several conjugate gredient algorithms,” SIAM J. Numer. Anal. 9, 248–259 (1972). [CrossRef]
  30. K. Ritter, “On the rate of superlinear convergence of a class of variable metric methods,” NumerischeMathematik 35, 293–313 (1980). [CrossRef]
  31. C. Vogel, Computational Methods for Inverse Problems (SIAM Press, 2002). [CrossRef]
  32. M. P. McLoughlin and G. R. Arce, “Deterministic properties of the recursive separable median filter,” IEEE Trans. Acoust. Speech Signal Process. 35, 98–106 (1987). [CrossRef]
  33. S. Kalluri and G. R. Arce, “Fast algorithms for weighted myriad computation by fixed-point search,” IEEE Trans. Signal Process. 48(1), 159–171 (2000). [CrossRef]
  34. K. Barner and G. R. Arce, “Permutation filters - a class of nonlinear filters based on set permutations,” IEEE Trans. Signal Process. 42, 782–798 (1994). [CrossRef]

Cited By

Alert me when this paper is cited

OSA is able to provide readers links to articles that cite this paper by participating in CrossRef's Cited-By Linking service. CrossRef includes content from more than 3000 publishers and societies. In addition to listing OSA journal articles that cite this paper, citing articles from other participating publishers will also be listed.


« Previous Article  |  Next Article »

OSA is a member of CrossRef.

CrossCheck Deposited