Transcript
Correcting Image Placement Error Using Registration Control (RegC) Technology Erez Graitzer1 , Gunter Antesberger2, Guy Ben-Zvi , Avi Cohen1 , Dmitriev Vladimir1 , Stephanie Winkelmeier2 1
1. Carl Zeiss SMS, Pixer Technologies, Karmiel Israel 2. Advanced Mask Technology Center, Dresden, Germany Corresponding authors: Erez Graitzer (
[email protected])
1. ABSTRACT The 2009 ITRS update specifies wafer overlay control as one of the major tasks for the sub 40 nm nodes. Wafer overlay is strongly dependent on mask image placement error (registration errors or Reg errors)1 in addition to CD control and defect control. The specs for registration or mask placement accuracy are twice as difficult in some of the double patterning techniques (DPT). This puts a heavy challenge on mask manufacturers (mask shops) to comply with advanced node registration specifications. Registration test masks as well as production masks were measured on a standard registration tool and the registration error was calculated and plotted. A specially developed algorithm was used to compute a correction lateral strain field that would minimize the registration error. A laser based prototype RegCTM tool was used to generate a strain field which corrected for the pre measured registration errors. Finally the post registration error map was measured. The resulting residual registration error field with and without scale and orthogonal errors removed was calculated. In this paper we present first results of registration control experiments using the prototype RegCTM tool. Key words: Image placement, Registration, Wafer, Overlay, Photomask, Laser, RegC
2. INTRODUCTION Carl Zeiss SMS has developed new a technology named RegC™ that enables the user (mask shop) to correct registration errors and improve image placement of a manufactured mask. The process is based on the principle of the CDC tool which uses a laser to write pixels inside the mask bulk in order to attenuate the light transmission through the photomask2,3. The RegCTM process enables the mask maker to bring an out of spec mask into the specification limits and to increase the mask manufacturing yield. Figure 1 shows the Registration Control process flow in the mask shop. The conventional methods of feeding back the systematic registration error to the E-beam writer and re-writing the mask are becoming difficult and expensive and not sufficient for the advanced nodes especially for double pattering (DP) technologies. Figure 2 shows the RegC block diagram, the system has two main optical sub systems. The first one is an optical setup that includes a laser, a beam delivery path, a beam steering device and a focusing objective that is used to generate the deformation elements or pixels. The second sub system is a metrology system that is being used to measure and characterize the deformation element properties. These properties will be used later during the job computation. In this work we will present some initial results to demonstrate the RegC technology. In this paper "fused silica", "quartz (Qz)" and mask "blank substrate" are used interchangeably. Optical Microlithography XXIV, edited by Mircea V. Dusa, Proc. of SPIE Vol. 7973, 797312 · © 2011 SPIE · CCC code: 0277-786X/11/$18 · doi: 10.1117/12.879885
Proc. of SPIE Vol. 7973 797312-1
Mask E-beam writer
Registration Metrology
Reg in Spec ?
Yes Pellicle Mount
Registration Metrology through pellicle
No New Tool: Correct registration No new mask writing
No
New Tool: Correct registration Thru Pellicle
Reg in Spec ?
Yes
Ship
Figure 1: RegC process flow
Proc. of SPIE Vol. 7973 797312-2
Metrology System for deformation element characteristics determination (process calibration)
Beam Scanner
Illumination optics
Deformation element top view
Generation Deformation elements
Beam delivery
Pulsed Laser
Beam Shaper
Focusing Opt.
Photomask Sensor
Figure 2 RegC block diagram
3. THE REGISTRATION CONTROL (REGC) PRINCIPLES Intravolume laser writing at certain conditions causes a predictable deformation element in the quartz (Qz) material. This deformation can be described by a physical mathematical model that represents the deformation caused by RegC element. The deformed zone inside the Qz bulk is a 3 dimensional volume of fused silica which has a slightly different morphological organization of the atoms with a slightly less dense packing, or lower density. The zone with lower density expands and pushes away the adjacent atoms and thus deforms the whole bulk of the Qz piece. Due to the elastic amorphous property of fused silica this deformation behaves almost truly elastically without critical breakage (cracks). In other words, when discussing very small deformations on the order of ppb or even ppm, fused silica is practically rubber. A special model that was developed to describe the cumulative effect of multitude of pixels on the Qz substrate takes into account the physical properties of fused silica such as its Young Modulus, its Poisson ratio etc. The model has been verified experimentally and provides a laser material dependent parameter called the Mode Signature (MS). The MS defines the magnitude and angle of deformation induced by writing a laser pixel at given conditions. The Mode Signature can be used first to calculate and predict the deformation and hence the registration error induced by writing a given array of pixels, for example CDC type of pixels. Second and relevant for the RegC process the MS is used to calculate a set of pixels needed to compensate for a given Reg error map. The current RegC process can only induce expansion pixels. This means that the average mask dimension after the RegC process will always be larger than before the process. This means that the absolute value of Reg after RegC will typically be larger than the absolute Reg error before the process, except for rare cases the whole mask error was contracted relative to the target.
Proc. of SPIE Vol. 7973 797312-3
However this is not a limitation since the target of the RegC process is not to remove the absolute Reg errors, but rather to remove only the uncorrectable errors. As is well known to the mask and litho industry the scanners have the ability to correct for all systematic linear errors which have rotational, orthogonal and scale components (in short "Scale and Ortho"). The main issue with registration of masks is the uncorrectable residuals, those Reg errors which are left over after the scanner has done its job. These residuals are typically 6-8 nm 3S in advanced 40nm nodes and below. However the specs at these nodes are 4-8 nm and in sub 20 nm nodes especially in double patterning technologies, will be <4 nm. Therefore the task of the RegC process is to remove the uncorrectable residuals from ~8 nm to ~4 nm, or about 50% improvement in the 2X and 1X nodes. Because of the importance of the scanner ability to correct for scale and ortho, all Reg metrology tools report in addition to raw registration errors also the scale and ortho (S/O) removed residual errors. These are the values which typically interest mask makers and their fab customers who are interested eventually at mask to mask overlay in the scanner. The ability of the scanner to remove specific Reg errors is shown in Fig 3.
Reg correction ability of a scanner Reg error - Raw
Reg error – S\O removed
XMin-11.96 , XMax 2.00, X3S 7.25 YMin -0.99, YMax 1.48, Y3S 1.51
XMin -9.91, XMax 3.04, X3S 6.33 YMin -0.55, YMax 0.90, Y3S 1.01
Figure 3a (left) shows the raw registration error while Figure 3b (right) shows the results after S\O removal. Note that the large vectors on the top right have been reduced in size but not so significantly
The RegC principle is to take the registration error ("the problem”) (Figure4a) and to apply the required deformation (Figure4b) that will enable higher efficiency (Figure 4c) of the scanner S\O removal capabilities.
Proc. of SPIE Vol. 7973 797312-4
=
4a Reg error - raw
4bThe required Reg Ch
4c The superimposed (problem +Reg change) can be efficiently corrected by scaling
4d After scanner S\O removal 3S X&Y < 0.3 nm
=
Figure 4a (upper left) shows the raw registration error while Figure 4b (upper right) shows the required registration change in order to bring all the errors to a correctable field. The final Reg error shown in Figure 4c (lower left) is the sum of the original Reg error and the induced change. Figure 4d (lower right) shows very low residual error after S\O removal. The target of the RegC software is to calculate the required change shown in 4b and the RegC laser process induces this change by writing the relevant field of pixels in the bulk of the mask.
4. EXPERIMENTAL RESULTS The first step in the experiment was to test and to prove that the mathematical model is accurate and can accurately predict the registration change. A registration plate was measured by a registration tool (IPRO 4) before the RegC process was applied. The registration measurement error was estimated by 0.8 nm (long term and short term error components4. Then a pre-calculated RegC process was applied and the plate was sent to "Post" registration measurement. The induced registration change (Post Reg – Pre Reg) was compared to the predicted values. The RegC job map and the results are plotted in Figure 5. As can be seen the calculated predicted values are very well in agreement with the measured value and reflect an R2 of 0.99 in both X and Y axis and the mean deviation between them is only 0.72 nm difference.
Proc. of SPIE Vol. 7973 797312-5
Planed Pixels Density
Pixels density arbitrary units Red - high dens. Blue – low dens
Registration change - Difference (Predicted minus Measured)
Predicted and Measured Registration change
Blue – predicted Green - measured
Mean deviation of measured from predicted Reg = 0.72 nm
Figure 5a (left) shows the applied pixels density in the photomask Figure 5b (center) shows the predicted (blue) registration change and the measured registration change (green) due to RegC process. Figure 5c (right) shows the difference (Predicted minus Measured) for each location.
The second step was to prove that this RegC technology can improve registration errors on a production mask. In this step two production masks were selected to be processed. In this experiment we have selected to process only the mask periphery while the exposure field was left un-processed. Figure 6 shows the experimental results for both masks. In order to improve the process efficiency the process was split into two steps where in each step different deformation properties were applied. The results show 0.5 nm (X) and 1.5 nm (Y) registration improvement in mask #1 and 0.4 nm (X) and 0.8 nm (Y) registration improvement in mask #2. The efficiency of the periphery process is limited to about 20% since the process is being done "far" from the registration problem. Nevertheless low spatial frequency registration distortions or placement errors close to the border could be corrected sufficiently. In spite of the relatively limited improvement efficiency the masks have been brought from out of spec to in spec for registration, which is a very promising result when the mask manufacturing process is at its limits.
Proc. of SPIE Vol. 7973 797312-6
Step #1
Exposure Field
Mask #1
Exposure Field Step #2
Exposure Field
Blue = Post; Red = Pre Pre 3S [nm] Post 3S [nm] Actual improvement [%]
X 5.79 5.09 12.1
Y 8.63 7.16 17.0
Pre and post process registration error (S/O removed)
Step #1
Exposure Field
Mask #2
Pre and post process registration error (S/O removed)
Step #2
Blue = Post; Red = Pre Pre 3S [nm] Post 3S [nm] Actual improvement [%]
X 6.51 6.11 6.1
Y 7.27 6.47 11.0
Figure 6a (upper) shows mask #1 two steps process and the registration improvement, figure 6b shows the same for mask #2
Proc. of SPIE Vol. 7973 797312-7
5. CONCLUSIONS It was proven that a registration correction strain field can be computed using a special algorithm and that a laser based correction method can be used to effectively reduce the registration error in the mask by using this method without significantly affecting any other mask property.
6. FUTURE WORK Although these experiments have shown that a mask which was rejected based on its Reg could be saved and brought into spec by treating the non active area, it is recognized that a much higher value could be achieved with improvements on the order of 50% by applying the RegC process over the whole mask area. For this purpose Zeiss is developing the next stage where the whole mask area is treated. In addition to that, more and more chip manufacturers start to specify not only mask registration per se but also mask to mask overlay, which adds even more challenges to the mask maker. We propose to use in the future the RegC tool to control the mask to mask overlay. Figure 7 is a proposed process flow for mask to mask overlay control. This work is in process with several customers and will be presented in a future conference.
Correct registration
No Mask A E-beam writer
Registration Metrology
Yes
Reg in Spec ?
Yes
Overlay in Spec ? Mask B E-beam writer
Registration Metrology
Reg in Spec ?
Yes
No
No Correct registration
New Tool: Correct overlay Between Mask A or Mask B
Figure 7 shows the proposed process flow for mask to mask overlay correction
Proc. of SPIE Vol. 7973 797312-8
Ship Masks
7. AKNOWLEDGEMENTS AMTC is a joint venture of GLOBALFOUNDRIES and TOPPAN Photomasks and gratefully acknowledges the financial support by the German Federal Ministry of Education and Research (BMBF)
8. REFERENCES
1. Schultz et al SPIE Microlithography 2007 6581-13 "Meeting overlay requirements for future technology nodes with in-die overlay metrology" 2. Pforr et al BACUS 2007 6730-107 "Performance comparison of techniques for intra-field CD control improvement" 3. Ben-Zvi et al BACUS 2010 7823-11 "Process window improvement by CDC" 4. Enkrich, C., Antesberger, G., Loeffler, O., et al., "Registration measurement capability of VISTEC LMS IPRO4 with focus on small features," Proceedings of SPIE Vol. 7028, 70282Y (2008).
Proc. of SPIE Vol. 7973 797312-9
