CN108614390A - A kind of source mask optimization method using compressed sensing technology - Google Patents
A kind of source mask optimization method using compressed sensing technology Download PDFInfo
- Publication number
- CN108614390A CN108614390A CN201810645227.XA CN201810645227A CN108614390A CN 108614390 A CN108614390 A CN 108614390A CN 201810645227 A CN201810645227 A CN 201810645227A CN 108614390 A CN108614390 A CN 108614390A
- Authority
- CN
- China
- Prior art keywords
- light source
- vector
- mask
- optimization
- value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 70
- 238000005457 optimization Methods 0.000 title claims abstract description 59
- 238000005516 engineering process Methods 0.000 title claims abstract description 29
- 239000011159 matrix material Substances 0.000 claims description 15
- 238000005286 illumination Methods 0.000 claims description 4
- 238000005070 sampling Methods 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims 2
- 230000005540 biological transmission Effects 0.000 claims 1
- 230000003993 interaction Effects 0.000 claims 1
- 238000003384 imaging method Methods 0.000 abstract description 36
- 230000008569 process Effects 0.000 abstract description 20
- 238000001459 lithography Methods 0.000 description 13
- 229920002120 photoresistant polymer Polymers 0.000 description 11
- 238000010586 diagram Methods 0.000 description 7
- 230000000694 effects Effects 0.000 description 4
- 230000003044 adaptive effect Effects 0.000 description 3
- 230000000903 blocking effect Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 3
- 238000000206 photolithography Methods 0.000 description 3
- 238000011478 gradient descent method Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000000671 immersion lithography Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000011112 process operation Methods 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G03—PHOTOGRAPHY; CINEMATOGRAPHY; ANALOGOUS TECHNIQUES USING WAVES OTHER THAN OPTICAL WAVES; ELECTROGRAPHY; HOLOGRAPHY
- G03F—PHOTOMECHANICAL PRODUCTION OF TEXTURED OR PATTERNED SURFACES, e.g. FOR PRINTING, FOR PROCESSING OF SEMICONDUCTOR DEVICES; MATERIALS THEREFOR; ORIGINALS THEREFOR; APPARATUS SPECIALLY ADAPTED THEREFOR
- G03F1/00—Originals for photomechanical production of textured or patterned surfaces, e.g., masks, photo-masks, reticles; Mask blanks or pellicles therefor; Containers specially adapted therefor; Preparation thereof
-
- G—PHYSICS
- G03—PHOTOGRAPHY; CINEMATOGRAPHY; ANALOGOUS TECHNIQUES USING WAVES OTHER THAN OPTICAL WAVES; ELECTROGRAPHY; HOLOGRAPHY
- G03F—PHOTOMECHANICAL PRODUCTION OF TEXTURED OR PATTERNED SURFACES, e.g. FOR PRINTING, FOR PROCESSING OF SEMICONDUCTOR DEVICES; MATERIALS THEREFOR; ORIGINALS THEREFOR; APPARATUS SPECIALLY ADAPTED THEREFOR
- G03F1/00—Originals for photomechanical production of textured or patterned surfaces, e.g., masks, photo-masks, reticles; Mask blanks or pellicles therefor; Containers specially adapted therefor; Preparation thereof
- G03F1/68—Preparation processes not covered by groups G03F1/20 - G03F1/50
- G03F1/70—Adapting basic layout or design of masks to lithographic process requirements, e.g., second iteration correction of mask patterns for imaging
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Exposure And Positioning Against Photoresist Photosensitive Materials (AREA)
Abstract
本发明公开了一种采用压缩感知技术的光源掩模优化方法，本发明将光源优化SO问题构造为在解为非负约束条件下求解2范数的图像恢复问题，即：s.t.其中约束条件线性方程组使优化后光源对应的空间像尽量接近目标成像值。另一方面，本发明将掩模优化MO问题构造为含稀疏正则项和低秩正则项的图像优化问题，即：其中约束条件非线性方程组使优化后的掩模和光源对应的空间像观测数据尽量接近目标图形上的观测数据，约束条件能够进一步降低优化过程中方程组的数目，约束条件能够保证优化过程中掩模的复杂度尽量降低。
The invention discloses a light source mask optimization method using compressed sensing technology. The invention constructs the SO problem of light source optimization as solving the image restoration problem of 2-norm under the non-negative constraint condition, namely: st where the constrained linear equations Make the aerial image corresponding to the optimized light source as close as possible to the target imaging value. On the other hand, the present invention constructs the mask optimization MO problem as an image optimization problem containing a sparse regular term and a low-rank regular term, namely: where the constrained nonlinear equations Make the aerial image observation data corresponding to the optimized mask and light source as close as possible to the observation data on the target graphic, Constraints can further reduce the number of equations in the optimization process, Constraint conditions can ensure that the complexity of the mask is minimized during the optimization process.
Description
技术领域technical field
本发明涉及光刻分辨率增强技术领域，具体涉及一种采用压缩感知技术的光源掩模优化方法。The invention relates to the technical field of lithographic resolution enhancement, in particular to a light source mask optimization method using compressed sensing technology.
背景技术Background technique
光学光刻是目前的主流光刻技术，它利用光学投影成像原理，通过步进扫描曝光的方式将掩模上的集成电路图形转移到涂有光刻胶的晶片上。随着半导体行业的快速发展，超大规模集成电路的特征尺寸也在不断缩小。而光刻系统是IC制造业中的主要核心设备，目前主流的光刻技术主要采用光源波长193nm的深紫外浸没式光刻技术。Optical lithography is the current mainstream lithography technology. It uses the principle of optical projection imaging to transfer the integrated circuit pattern on the mask to the wafer coated with photoresist through step-and-scan exposure. With the rapid development of the semiconductor industry, the feature size of VLSI is also shrinking. The lithography system is the main core equipment in the IC manufacturing industry. The current mainstream lithography technology mainly adopts the deep ultraviolet immersion lithography technology with a light source wavelength of 193nm.
随着光刻技术节点进入45-14nm，集成电路的关键尺寸已进入深亚波长范围，此时必须采用分辨率增强技术(resolution enhancement technique，简称RET)来进一步提高光刻系统的分辨率和成像保真度。光源-掩模优化(source-mask optimization，简称SMO)技术是提高光刻成像分辨率和图形保真度的重要方法之一，SMO利用光源和掩模在成像过程中的耦合关系，在对掩模进行预畸变的同时优化光源的强度分布，能够进一步降低工艺因子和提高光刻系统成像性能。以往的SMO技术主要采用梯度下降法的优化算法对像素化的光源和掩模进行优化，但是由于像素化的光源和掩模在优化的过程中需要处理庞大的数据量，传统优化算法的运算效率将显著降低，为此，针对SMO技术的运算效率问题还有待进一步解决。As the lithography technology node enters 45-14nm, the critical dimensions of integrated circuits have entered the deep sub-wavelength range. At this time, the resolution enhancement technique (RET) must be used to further improve the resolution and imaging of the lithography system. Fidelity. Source-mask optimization (SMO) technology is one of the important methods to improve the resolution and image fidelity of lithography imaging. Optimizing the intensity distribution of the light source while pre-distorting the mode can further reduce the process factor and improve the imaging performance of the lithography system. In the past, the SMO technology mainly used the optimization algorithm of the gradient descent method to optimize the pixelated light source and mask. However, due to the huge amount of data that needs to be processed during the optimization process of the pixelated light source and mask, the computational efficiency of the traditional optimization algorithm will be significantly reduced. For this reason, the problem of computing efficiency for SMO technology has yet to be further resolved.
相关文献(Journal of the Optical Society of America A，2013,30:112-123)提出了一种基于梯度下降法的SMO技术，该方法针对光源和掩模的优化次序，提出了三种不同的优化方法，分别是同步型、交替型和混合型优化方法，其中以混合型的优化效果最为理想，但是上述方法在优化过程中需要大量的运行时间，运行效率需要进一步提高。Related literature (Journal of the Optical Society of America A, 2013, 30:112-123) proposes a SMO technique based on the gradient descent method. This method proposes three different optimizations for the optimization order of the light source and the mask. The methods are synchronous, alternating and hybrid optimization methods, among which the hybrid optimization effect is the most ideal, but the above methods require a lot of running time in the optimization process, and the operating efficiency needs to be further improved.
相关文献(Optics Express,2017,25:7131-7149)提出了一种基于自适应压缩感知技术的快速SO方法，该方法假设光源图形在某组稀疏基上是稀疏的，即光源图形在该组基上的绝大部分系数值等于0或接近于0。之后，该方法根据Related literature (Optics Express, 2017, 25:7131-7149) proposes a fast SO method based on adaptive compressed sensing technology, which assumes that the light source graph is sparse on a certain set of sparse Most of the coefficient values on the basis are equal to or close to zero. Afterwards, the method is based on
蓝噪声采样方法在电路版图的关键区域内选取若干观测点，构造SO优化数学模型，并采用线性布莱格曼算法求解上述的SO问题，获得优化后的光源图形。The blue noise sampling method selects several observation points in the key area of the circuit layout, constructs the SO optimization mathematical model, and uses the linear Bregman algorithm to solve the above SO problem, and obtains the optimized light source pattern.
但是以上方法存在两点不足：第一，上述方法采用的线性布莱格曼算法需要在每次优化过程中对光源的优化变量强制置零，此操作必然会影响SO问题的最优解，从而影响光刻系统的成像性能；第二，上述方法仅仅考虑到线性压缩感知技术在光源优化问题中的应用，并未将非线性压缩感知技术应用到掩模的优化问题中，因此最终的光刻成像性能还不是最优的结果。However, there are two shortcomings in the above method: first, the linear Bregman algorithm used in the above method needs to force the optimization variable of the light source to zero in each optimization process, and this operation will inevitably affect the optimal solution of the SO problem, thus affect the imaging performance of the lithography system; secondly, the above method only considers the application of the linear compressive sensing technology in the light source optimization problem, and does not apply the nonlinear compressive sensing technology to the mask optimization problem, so the final lithography Imaging performance is not yet optimal results.
综上所述，传统SMO方法的运算效率，以及基于线性压缩感知技术的SO最优解问题均有待进一步改善和提高。In summary, the computational efficiency of the traditional SMO method and the SO optimal solution problem based on linear compressed sensing technology need to be further improved and improved.
发明内容Contents of the invention
有鉴于此，本发明提供了一种采用压缩感知技术的光源掩模优化方法，采用梯度投影的稀疏重构算法GPSR求解光源优化SO问题，优化过程中每次迭代更新所得到的都是非负解，没有强制置零相关操作，在算法运算时间相近的情况下，优化结果更加接近实际真实值，能够进一步提高光刻成像性能；且对于掩模优化MO问题，对目标图形和空间成像进行降采样，从而降低了优化方程组的数目，能够更好地提高运算效率。In view of this, the present invention provides a light source mask optimization method using compressed sensing technology, using the sparse reconstruction algorithm GPSR of gradient projection to solve the SO problem of light source optimization, and each iterative update in the optimization process obtains a non-negative solution , there is no mandatory zero-setting related operation. In the case of similar algorithm operation time, the optimization result is closer to the actual real value, which can further improve the lithography imaging performance; and for the mask optimization MO problem, the target graphics and spatial imaging are down-sampled , thereby reducing the number of optimization equations and improving the operational efficiency better.
为达到上述目的，本发明的技术方案包括如下步骤：In order to achieve the above object, the technical solution of the present invention comprises the following steps:
步骤101、将光源初始化为NS×NS的光源图形J，将掩模图形M和目标图形栅格化为N×N的图形，其中NS和N均为整数值。Step 101. Initialize the light source as a N S × N S light source pattern J, and set the mask pattern M and the target pattern Rasterized as N×N graphics, where N S and N are both integer values.
步骤102、对所述光源图形J进行逐点扫描，并将所述光源图形J转化为N2×1的光源向量所述光源向量的元素值等于所述光源图形J的对应像素值。Step 102, scan the light source pattern J point by point, and convert the light source pattern J into an N 2 ×1 light source vector The light source vector The element value of is equal to the corresponding pixel value of the light source pattern J.
对掩模图形M进行逐点扫描，并将M转化为N2×1的掩模向量所述掩模向量的元素值等于所述掩模图形M的对应像素值。Scan the mask pattern M point by point, and convert M into a mask vector of N 2 ×1 The mask vector The element value of is equal to the corresponding pixel value of the mask pattern M.
对目标图形进行逐点扫描，并将转化为N2×1的目标向量所述目标向量的元素值等于目标图形的对应像素值。to the target graphic scan point by point, and Convert to N 2 x 1 target vector The target vector The element value of is equal to the target graph The corresponding pixel value of .
步骤103、选定两组组基函数ΨJ和ΨM，使得光源向量和掩模向量分别在ΨJ和ΨM上是稀疏的；将光源向量在ΨJ上展开得到掩模向量在ΨM上展开得到其中和分别为展开后的系数。Step 103, select two groups of basis functions Ψ J and Ψ M , so that the light source vector and the mask vector are sparse on ΨJ and ΨM respectively; the light source vector Expand on ΨJ to get mask vector Expand on Ψ M to get in and are the expanded coefficients, respectively.
步骤104、采用初始的掩模图形M计算照明交叉系数ICC矩阵Icc，其大小为N2×NS 2；并对目标图形和ICC矩阵Icc降采样分别得到和 Step 104, using the initial mask pattern M to calculate the illumination cross coefficient ICC matrix I cc , the size of which is N 2 × NS 2 ; and the target pattern and ICC matrix I cc downsampling to get and
步骤105、将光源优化SO问题构造为如下的形式：Step 105, constructing the light source optimization SO problem into the following form:
其中指的优化结果；为向量的2范数；λ为权重系数；为向量的1范数；作为约束条件。in refer to The optimization results; as a vector 2 norm of ; λ is the weight coefficient; as a vector 1 norm of ; as a constraint.
步骤106、采用梯度投影的稀疏重构算法GPSR求解步骤105中的光源优化SO问题，获得对应最优光源图形的的优化结果计算优化后的光源图形为Ψ'J是ΨJ的转置。Step 106: Using the gradient projection sparse reconstruction algorithm GPSR to solve the light source optimization SO problem in step 105, and obtain the corresponding optimal light source graph The optimization result of Calculate the optimized light source graph as Ψ'J is the transpose of ΨJ .
步骤107、根据步骤106中优化的光源计算空间成像I(θM)并扫描得到所述空间成像I(θM)的空间成像向量并对和分别进行降采样得到和 Step 107, calculate the spatial imaging I(θ M ) according to the light source optimized in step 106 and scan to obtain the spatial imaging vector of the spatial imaging I(θ M ) and to and downsampling to get and
步骤108、将掩模优化MO问题构造为如下形式：Step 108, constructing the mask optimization MO problem into the following form:
其中是的优化结果；为向量的低秩正则项，为向量的稀疏正则项，α和β分别为和正则项的权重系数。in Yes The optimization results; as a vector The low-rank regularization term of , as a vector The sparse regularization term of , α and β are respectively and The weight coefficient of the regularization term.
步骤109、采用分裂布莱格曼算法Split Bregman求解步骤108中的MO问题，获得对应最优掩模图形的向量计算优化后的掩模图形为Ψ'M为ΨM的转置。Step 109, using the split Bregman algorithm Split Bregman to solve the MO problem in step 108, and obtain the vector corresponding to the optimal mask pattern Calculate the optimized mask pattern as Ψ' M is the transpose of Ψ M.
进一步地，步骤107中，所述对和分别进行降采样得到和具体为：Further, in step 107, the and downsampling to get and Specifically:
步骤201、所述目标图形为大小为N×N的矩阵；中对应目标图形中不透光部分元素值设置为1，透光部分元素值设置为0，对的行和列分别隔K个像素进行取值得到降采样后的矩阵大小用N/K×N/K表示。Step 201, the target graphic is a matrix of size N×N; Set the element value of the opaque part in the corresponding target graph to 1, and set the element value of the transparent part to 0. The row and column of the value are taken at intervals of K pixels to get Downsampled Matrix The size is represented by N/K×N/K.
步骤202、对所述空间成像I(θM)的行和列分别隔K个像素进行取值得到I(θM)降采样后的矩阵Ik(θM)，大小为N/K×N/K。Step 202, the row and column of the spatial imaging I(θ M ) are valued at intervals of K pixels to obtain the matrix I k (θ M ) after downsampling of I(θ M ), the size of which is N/K×N /K.
有益效果：Beneficial effect:
本发明的目的是提供一种采用压缩感知技术的光源掩模优化SMO方法。首先，针对SO问题采用蓝噪声采样和自适应投影矩阵相结合的方法构造约束条件线性方程组，之后，将光源优化SO问题转换为求解2范数的图像恢复问题，采用GPSR重构算法对光源图形进行优化。相比现有的线性布莱格曼算法，本发明中的GPSR重构算法不仅能够保证相似的运算效率，而且能够很好地避免优化过程产生的负数解，优化得到的结果更加接近光源的实际真实值，能够进一步提高光刻系统的成像性能，线性压缩感知技术在光源优化问题中的应用；其次，对于掩模优化MO问题，采用隔点取值的方法对目标图形和空间成像进行降采样，利用降采样后的目标图形和空间成像构造低维的损失函数，为了进一步加速优化效率和提高掩模的可制造性，本发明还在损失函数中加入稀疏和低秩两项正则项加以约束。The purpose of the present invention is to provide a light source mask optimization SMO method using compressed sensing technology. Firstly, the method of combining blue noise sampling and adaptive projection matrix is used to construct the constrained linear equations for the SO problem. After that, the SO problem of light source optimization is transformed into the image recovery problem of solving the 2-norm, and the GPSR reconstruction algorithm is used to reconstruct the light source. Graphics are optimized. Compared with the existing linear Bregman algorithm, the GPSR reconstruction algorithm in the present invention can not only ensure similar computing efficiency, but also can well avoid the negative solution generated in the optimization process, and the optimized result is closer to the actual light source The real value can further improve the imaging performance of the lithography system, and the application of linear compressed sensing technology in the light source optimization problem; secondly, for the mask optimization MO problem, the target graphics and spatial imaging are down-sampled by using the method of taking values at intervals , using the down-sampled target graph and spatial imaging to construct a low-dimensional loss function, in order to further accelerate the optimization efficiency and improve the manufacturability of the mask, the present invention also adds sparse and low-rank regular terms to the loss function to constrain .
附图说明Description of drawings
图1为本发明涉及的采用压缩感知技术的SMO方法的流程图。FIG. 1 is a flow chart of the SMO method using compressed sensing technology involved in the present invention.
图2为采用传统混合型SMO方法得到的优化光源图形、掩模图形及其在额定曝光量下最佳焦平面处的光刻胶中成像示意图。Fig. 2 is a schematic diagram of the optimized light source pattern, mask pattern and its imaging in the photoresist at the best focal plane under the rated exposure amount obtained by using the traditional hybrid SMO method.
图3为K＝2时，采用本发明中的基于压缩感知技术的SMO方法得到的优化光源图形、掩模图形及其在额定曝光量下最佳焦平面处的光刻胶中成像示意图。3 is a schematic diagram of the optimized light source pattern, mask pattern and its imaging in the photoresist at the best focal plane at the rated exposure amount obtained by using the SMO method based on compressed sensing technology in the present invention when K=2.
图4为K＝4时，采用本发明中的基于压缩感知技术的SMO方法得到的优化光源图形、掩模图形及其在额定曝光量下最佳焦平面处的光刻胶中成像示意图。FIG. 4 is a schematic diagram of the optimized light source pattern, mask pattern and its imaging in the photoresist at the best focal plane at the rated exposure amount obtained by using the SMO method based on the compressed sensing technology in the present invention when K=4.
图5为采用传统混合型SMO方法和本发明中基于压缩感知技术的SMO方法优化后得到的光刻系统工艺窗口对比示意图。FIG. 5 is a schematic diagram of the process window comparison of the lithography system obtained after the optimization of the traditional hybrid SMO method and the SMO method based on compressive sensing technology in the present invention.
具体实施方式Detailed ways
下面结合附图并举实施例，对本发明进行详细描述。The present invention will be described in detail below with reference to the accompanying drawings and examples.
本发明的原理：实际光刻系统成像性能通常采用成像误差、掩模复杂度、工艺窗口等指标进行评价。为了提高光刻成像保真度、降低掩模复杂度和扩大工艺窗口，本发明将SO问题构造为在解为非负约束条件下求解2范数的图像恢复问题，即：The principle of the present invention: the imaging performance of the actual photolithography system is usually evaluated by indicators such as imaging error, mask complexity, and process window. In order to improve the fidelity of lithographic imaging, reduce mask complexity and expand the process window, the present invention constructs the SO problem as solving the image restoration problem of 2-norm under the non-negative constraint condition, namely:
其中约束条件线性方程组使优化后光源对应的空间像尽量接近目标成像值。where the constrained linear equations Make the aerial image corresponding to the optimized light source as close as possible to the target imaging value.
另一方面，本发明将MO问题构造为含稀疏正则项和低秩正则项的图像优化问题，即：On the other hand, the present invention constructs the MO problem as an image optimization problem containing a sparse regular term and a low-rank regular term, namely:
其中约束条件非线性方程组使优化后的掩模和光源对应的空间像观测数据尽量接近目标图形上的观测数据，约束条件能够进一步降低优化过程中方程组的数目，约束条件能够保证优化过程中掩模的复杂度尽量降低。where the constrained nonlinear equations Make the aerial image observation data corresponding to the optimized mask and light source as close as possible to the observation data on the target graphic, Constraints can further reduce the number of equations in the optimization process, Constraint conditions can ensure that the complexity of the mask is minimized during the optimization process.
本发明提供了一种采用压缩感知技术的光源掩模优化方法，其流程如图1所示，包括：The present invention provides a light source mask optimization method using compressed sensing technology, the process of which is shown in Figure 1, including:
步骤101、将光源初始化为NS×NS的光源图形J，将掩模图形M和目标图形栅格化为N×N的图形，其中NS和N均为整数值。Step 101. Initialize the light source as a N S × N S light source pattern J, and set the mask pattern M and the target pattern Rasterized as N×N graphics, where N S and N are both integer values.
步骤102、对所述光源图形J进行逐点扫描，并将所述光源图形J转化为N2×1的光源向量所述光源向量的元素值等于所述光源图形J的对应像素值。Step 102, scan the light source pattern J point by point, and convert the light source pattern J into an N 2 ×1 light source vector The light source vector The element value of is equal to the corresponding pixel value of the light source pattern J.
对掩模图形M进行逐点扫描，并将M转化为N2×1的掩模向量所述掩模向量的元素值等于所述掩模图形M的对应像素值。Scan the mask pattern M point by point, and convert M into a mask vector of N 2 ×1 The mask vector The element value of is equal to the corresponding pixel value of the mask pattern M.
对目标图形进行逐点扫描，并将转化为N2×1的目标向量所述目标向量的元素值等于目标图形的对应像素值。to the target graphic scan point by point, and Convert to N 2 x 1 target vector The target vector The element value of is equal to the target graph The corresponding pixel value of .
步骤103、选定两组组基函数ΨJ和ΨM，使得光源向量和掩模向量分别在ΨJ和ΨM上是稀疏的，即光源向量和掩模向量在ΨJ和ΨM基上展开后的大部分系数为0或接近于0；将光源向量在ΨJ上展开得到掩模向量在ΨM上展开得到其中和分别为展开后的系数。Step 103, select two groups of basis functions Ψ J and Ψ M , so that the light source vector and the mask vector are sparse on ΨJ and ΨM , respectively, that is, the light source vector and the mask vector Most of the coefficients expanded on Ψ J and Ψ M bases are 0 or close to 0; the light source vector Expand on ΨJ to get mask vector Expand on Ψ M to get in and are the expanded coefficients, respectively.
步骤104、采用初始的掩模图形M计算照明交叉系数ICC(illumination crosscoefficient)矩阵Icc，其大小为N2×NS 2；并对目标图形和ICC矩阵Icc降采样分别得到和s是标识，本发明中采用蓝噪声和自适应矩阵采样方法。Step 104, using the initial mask pattern M to calculate the illumination cross coefficient ICC (illumination crosscoefficient) matrix I cc , the size of which is N 2 × NS 2 ; and the target pattern and ICC matrix I cc downsampling to get and s is a sign, and blue noise and adaptive matrix sampling methods are adopted in the present invention.
步骤105、将光源优化SO问题构造为如下的形式：Step 105, constructing the light source optimization SO problem into the following form:
其中指的优化结果；为向量的2范数；λ为权重系数；为向量的1范数；作为约束条件。in refer to The optimization results; as a vector 2 norm of ; λ is the weight coefficient; as a vector 1 norm of ; as a constraint.
步骤106、采用梯度投影的稀疏重构算法GPSR求解步骤105中的光源优化SO问题，获得对应最优光源图形的的优化结果计算优化后的光源图形为Ψ'J是ΨJ的转置。Step 106: Using the gradient projection sparse reconstruction algorithm GPSR to solve the light source optimization SO problem in step 105, and obtain the corresponding optimal light source graph The optimization result of Calculate the optimized light source graph as Ψ'J is the transpose of ΨJ .
步骤107、根据步骤106中优化的光源计算空间成像I(θM)并扫描得到所述空间成像I(θM)的空间成像向量并对和分别进行降采样得到和 Step 107, calculate the spatial imaging I(θ M ) according to the light source optimized in step 106 and scan to obtain the spatial imaging vector of the spatial imaging I(θ M ) and to and downsampling to get and
本发明实施例中，步骤107中，对和分别进行降采样得到和具体为：In the embodiment of the present invention, in step 107, the and downsampling to get and Specifically:
步骤201、所述目标图形为大小为N×N的矩阵；中对应目标图形中不透光部分元素值设置为1，透光部分元素值设置为0，对的行和列分别隔K个像素进行取值得到降采样后的矩阵大小用N/K×N/K表示；Step 201, the target graphic is a matrix of size N×N; Set the element value of the opaque part in the corresponding target graph to 1, and set the element value of the transparent part to 0. The row and column of the value are taken at intervals of K pixels to get Downsampled Matrix The size is represented by N/K×N/K;
步骤202、对所述空间成像I(θM)的行和列分别隔K个像素进行取值得到I(θM)降采样后的矩阵Ik(θM)，大小为N/K×N/K。Step 202, the row and column of the spatial imaging I(θ M ) are valued at intervals of K pixels to obtain the matrix I k (θ M ) after downsampling of I(θ M ), the size of which is N/K×N /K.
步骤108、将掩模优化MO问题构造为如下形式：Step 108, constructing the mask optimization MO problem into the following form:
其中是的优化结果；为向量的低秩正则项，为向量的稀疏正则项，α和β分别为和正则项的权重系数。in Yes The optimization results; as a vector The low-rank regularization term of , as a vector The sparse regularization term of , α and β are respectively and The weight coefficient of the regularization term.
步骤109、采用分裂布莱格曼算法Split Bregman求解步骤108中的MO问题，获得对应最优掩模图形的向量计算优化后的掩模图形为Ψ'M为ΨM的转置。Step 109, using the split Bregman algorithm Split Bregman to solve the MO problem in step 108, and obtain the vector corresponding to the optimal mask pattern Calculate the optimized mask pattern as Ψ' M is the transpose of Ψ M.
本发明的实施实例：Implementation example of the present invention:
如图2所示为采用传统混合型SMO方法得到的优化光源图形、优化掩模图形及其在额定曝光量下最佳焦平面处的光刻胶中成像示意图。201为采用传统混合型SMO方法得到的优化光源图形，白色代表发光区域，黑色代表不发光区域。202为采用传统混合型SMO方法得到的掩模图形，白色代表开口区域，黑色代表阻光区域，其关键尺寸为45nm。203为采用201作为光源、202作为掩模，不考虑曝光量变化和离焦效应时，在理想焦平面处的光刻胶中成像，成像误差为2888，其中成像误差定义为光刻胶中成像与目标图形的欧拉距离的平方，掩模复杂度为197，其中掩模复杂度的计算是采用专业的Calibre仿真软件进行的测试，整个SMO优化流程运算时间为1170991秒。Figure 2 is a schematic diagram of the optimized light source pattern, the optimized mask pattern and the imaging in the photoresist at the best focal plane under the rated exposure amount obtained by using the traditional hybrid SMO method. 201 is an optimized light source pattern obtained by using the traditional hybrid SMO method, white represents the light-emitting area, and black represents the non-light-emitting area. 202 is a mask pattern obtained by using a traditional hybrid SMO method, white represents an opening region, black represents a light blocking region, and its critical dimension is 45nm. 203 is to use 201 as the light source and 202 as the mask. When the exposure amount change and the defocus effect are not considered, the image is formed in the photoresist at the ideal focal plane, and the imaging error is 2888. The imaging error is defined as the imaging error in the photoresist The square of the Euler distance from the target pattern, the mask complexity is 197, and the mask complexity is calculated using professional Caliber simulation software. The operation time of the entire SMO optimization process is 1,170,991 seconds.
如图3所示为K＝2的情况下，采用基于CS技术的SMO方法得到的优化光源图形、优化掩模图形及其在额定曝光量下最佳焦平面处的光刻胶中成像示意图。301为采用基于CS技术的SMO方法得到的优化光源图形，白色代表发光区域，黑色代表不发光区域。302为采用基于CS技术的SMO方法得到的掩模图形，白色代表开口区域，黑色代表阻光区域，其关键尺寸为45nm。303为采用301作为光源、302作为掩模，不考虑曝光量变化和离焦效应时，在理想焦平面处的光刻胶中成像，成像误差为1376，掩模复杂度为92，整个SMO优化流程运算时间为118532秒。As shown in Figure 3, in the case of K=2, the optimized light source pattern, optimized mask pattern and its imaging in the photoresist at the best focal plane under the rated exposure amount obtained by using the SMO method based on CS technology are schematic diagrams. 301 is an optimized light source pattern obtained by adopting the SMO method based on CS technology, white represents a light-emitting area, and black represents a non-light-emitting area. 302 is a mask pattern obtained by adopting the SMO method based on CS technology, white represents an opening region, black represents a light blocking region, and its critical dimension is 45nm. 303 is to use 301 as the light source and 302 as the mask. When the exposure amount change and defocus effect are not considered, the image is formed in the photoresist at the ideal focal plane. The imaging error is 1376, the mask complexity is 92, and the entire SMO is optimized. The operation time of the process is 118532 seconds.
如图4所示为K＝4的情况下，采用基于CS技术的SMO方法得到的优化光源图形、优化掩模图形及其在额定曝光量下最佳焦平面处的光刻胶中成像示意图。401为采用基于CS技术的SMO方法得到的优化光源图形，白色代表发光区域，黑色代表不发光区域。402为采用基于CS技术的SMO方法得到的掩模图形，白色代表开口区域，黑色代表阻光区域，其关键尺寸为45nm。403为采用401作为光源、402作为掩模，不考虑曝光量变化和离焦效应时，在理想焦平面处的光刻胶中成像，成像误差为2580，掩模复杂度为151，整个SMO优化流程运算时间为96992秒。As shown in Figure 4, in the case of K=4, the optimized light source pattern, optimized mask pattern and its imaging in the photoresist at the best focal plane under the rated exposure amount obtained by using the SMO method based on CS technology are schematic diagrams. 401 is an optimized light source pattern obtained by adopting the SMO method based on CS technology, white represents a light-emitting area, and black represents a non-light-emitting area. 402 is a mask pattern obtained by using the SMO method based on CS technology, white represents the opening area, black represents the light blocking area, and its critical dimension is 45nm. 403 is to use 401 as the light source and 402 as the mask. When the exposure amount change and defocus effect are not considered, the image is formed in the photoresist at the ideal focal plane. The imaging error is 2580, and the mask complexity is 151. The entire SMO is optimized The process operation time is 96992 seconds.
如图5所示为传统混合型SMO方法和本发明中的SMO方法优化后得到的光刻系统工艺窗口对比图。501为传统混合型SMO方法得到的工艺窗口，502为K＝2时，本发明中的SMO方法得到的工艺窗口，503为K＝4时，本发明中的SMO方法得到的工艺窗口。FIG. 5 is a comparison chart of process windows of the lithography system obtained after optimization of the traditional hybrid SMO method and the SMO method of the present invention. 501 is the process window obtained by the traditional hybrid SMO method, 502 is the process window obtained by the SMO method in the present invention when K=2, and 503 is the process window obtained by the SMO method in the present invention when K=4.
对比图2、3、4、5可知，相比现有的传统混合型SMO方法，本发明中的SMO方法具有更高的运算效率，掩模复杂更低，优化之后光刻系统的成像性能更好。Comparing Figures 2, 3, 4, and 5, it can be seen that compared with the existing traditional hybrid SMO method, the SMO method in the present invention has higher computing efficiency, lower mask complexity, and better imaging performance of the photolithography system after optimization. it is good.
综上所述，以上仅为本发明的较佳实施例而已，并非用于限定本发明的保护范围。凡在本发明的精神和原则之内，所作的任何修改、等同替换、改进等，均应包含在本发明的保护范围之内。To sum up, the above are only preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.
Claims (2)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810645227.XA CN108614390B (en) | 2018-06-21 | 2018-06-21 | Light source mask optimization method adopting compressed sensing technology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810645227.XA CN108614390B (en) | 2018-06-21 | 2018-06-21 | Light source mask optimization method adopting compressed sensing technology |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108614390A true CN108614390A (en) | 2018-10-02 |
| CN108614390B CN108614390B (en) | 2020-04-03 |
Family
ID=63665588
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810645227.XA Active CN108614390B (en) | 2018-06-21 | 2018-06-21 | Light source mask optimization method adopting compressed sensing technology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108614390B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109656107A (en) * | 2019-01-30 | 2019-04-19 | 北京理工大学 | A kind of photolithography light source graph optimization method using transformation L1 norm constraint |
| CN109711376A (en) * | 2018-12-29 | 2019-05-03 | 重庆邮电大学 | A Multiscale Sparse Blue Noise Sampling Method Based on Optimal Transport Theory |
| CN110441994A (en) * | 2019-07-31 | 2019-11-12 | 北京理工大学 | A kind of quick light source optimization method using compressed sensing |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102708543A (en) * | 2012-04-19 | 2012-10-03 | 北京工商大学 | Fluorescent microscopic image restoring method based on blind deconvolution and sparse representation and device thereof |
| CN103901738A (en) * | 2014-03-18 | 2014-07-02 | 北京理工大学 | Light-source optimization method adopting compressed sensing technology |
| CN106444301A (en) * | 2016-11-21 | 2017-02-22 | 北京理工大学 | Source optimization method by adopting self-adaptive compressed sensing technology |
| US20170366763A1 (en) * | 2016-06-20 | 2017-12-21 | Joseph Hsuhuan Lin | Methods and Systems for Time-Encoded Multiplexed Imaging |
| CN107622515A (en) * | 2017-09-06 | 2018-01-23 | 郑州大学 | Physical reconstruction method of compressed light field |
-
2018
- 2018-06-21 CN CN201810645227.XA patent/CN108614390B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102708543A (en) * | 2012-04-19 | 2012-10-03 | 北京工商大学 | Fluorescent microscopic image restoring method based on blind deconvolution and sparse representation and device thereof |
| CN103901738A (en) * | 2014-03-18 | 2014-07-02 | 北京理工大学 | Light-source optimization method adopting compressed sensing technology |
| US20170366763A1 (en) * | 2016-06-20 | 2017-12-21 | Joseph Hsuhuan Lin | Methods and Systems for Time-Encoded Multiplexed Imaging |
| CN106444301A (en) * | 2016-11-21 | 2017-02-22 | 北京理工大学 | Source optimization method by adopting self-adaptive compressed sensing technology |
| CN107622515A (en) * | 2017-09-06 | 2018-01-23 | 郑州大学 | Physical reconstruction method of compressed light field |
Non-Patent Citations (2)
| Title |
|---|
| XU MA 等: "Fast optical proximity correction method based on nonlinear compressive sensing", 《OPTICS EXPRESS》 * |
| XU MA 等: "Lithographic source optimization based on adaptive projection compressive sensing", 《OPTICS EXPRESS》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109711376A (en) * | 2018-12-29 | 2019-05-03 | 重庆邮电大学 | A Multiscale Sparse Blue Noise Sampling Method Based on Optimal Transport Theory |
| CN109711376B (en) * | 2018-12-29 | 2022-10-11 | 重庆邮电大学 | Multi-scale sparse blue noise sampling method based on optimal transmission theory |
| CN109656107A (en) * | 2019-01-30 | 2019-04-19 | 北京理工大学 | A kind of photolithography light source graph optimization method using transformation L1 norm constraint |
| CN109656107B (en) * | 2019-01-30 | 2020-05-26 | 北京理工大学 | Photoetching light source graph optimization method adopting transformation L1 norm constraint |
| CN110441994A (en) * | 2019-07-31 | 2019-11-12 | 北京理工大学 | A kind of quick light source optimization method using compressed sensing |
| CN110441994B (en) * | 2019-07-31 | 2021-06-15 | 北京理工大学 | A Fast Light Source Optimization Method Using Compressed Sensing |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108614390B (en) | 2020-04-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Ma et al. | Pixel-based OPC optimization based on conjugate gradients | |
| CN102692814B (en) | Light source-mask mixed optimizing method based on Abbe vector imaging model | |
| Li et al. | Efficient source and mask optimization with augmented Lagrangian methods in optical lithography | |
| CN108228981B (en) | OPC model generation method based on neural network and prediction method of experimental pattern | |
| CN106125511A (en) | Low error suseptibility multiple target source mask optimization method based on vector imaging model | |
| CN110187609A (en) | A Deep Learning Approach to Computational Lithography | |
| CN106444301B (en) | A kind of light source optimization method using self-adapting compressing cognition technology | |
| CN108614390A (en) | A kind of source mask optimization method using compressed sensing technology | |
| CN104133348B (en) | A kind of adaptive optical etching system light source optimization method | |
| CN108693715A (en) | Promote the multiple target light source and photomask optimization method of full filed optical patterning uniformity | |
| Ma et al. | Fast lithography aerial image calculation method based on machine learning | |
| Ma et al. | Optimization of lithography source illumination arrays using diffraction subspaces | |
| Ma et al. | Hybrid source mask optimization for robust immersion lithography | |
| CN113495435A (en) | Digital mask projection lithography optimization method and system | |
| Lin et al. | Learning-based compressive sensing method for EUV lithographic source optimization | |
| CN102096336A (en) | Method for determining illumination intensity distribution of light source of photoetching process | |
| CN109543330A (en) | A kind of optical adjacent correction method pixel-based and system of Self Matching | |
| US11899374B2 (en) | Method for determining an electromagnetic field associated with a computational lithography mask model | |
| Ma et al. | Gradient-based inverse extreme ultraviolet lithography | |
| US10571799B1 (en) | Hessian-free calculation of product of Hessian matrix and vector for lithography optimization | |
| CN103901738A (en) | Light-source optimization method adopting compressed sensing technology | |
| CN109709772B (en) | A semi-implicit light source mask collaborative optimization method based on accelerated operator splitting | |
| CN117572718A (en) | Inversion lithography method capable of adaptively reducing mask complexity | |
| CN115469555B (en) | Spatial image prediction and image quality optimization method for sensor chip projection lithography machine | |
| CN102707563B (en) | Light source and mask alternate optimization method based on Abbe vector imaging model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |