Fabrication Of 0.06 /spl mu/m Poly-si Gate Using Duv Lithography With A Designed Si/sub x/O/sub y/N/sub z/ Film As An Arc And Hardmask

Wei W. Lee, Qizhi He, Hanratty, Rogers, Chatterjee, Kraft, Chapman
{"title":"Fabrication Of 0.06 /spl mu/m Poly-si Gate Using Duv Lithography With A Designed Si/sub x/O/sub y/N/sub z/ Film As An Arc And Hardmask","authors":"Wei W. Lee, Qizhi He, Hanratty, Rogers, Chatterjee, Kraft, Chapman","doi":"10.1109/VLSIT.1997.623733","DOIUrl":null,"url":null,"abstract":"We report fabrication of sub-0.1 pm poly-Si gates using conventional DUV lithography with an optimized Si,O,N, film. This film has dual functions: reducing substrate reflectivit to <1%, and serving as a hardmask for the pol Si etch. &ith an aggressive etch bias process, linewidtxs down to 0.06ym are achieved with good linewidth control ( 3 ~ 1 2 n m ) and a near erfect lineari . Excellent optical manufacturable PECVD deposition process. uniformity of the n a n t k of the AR z is obtained with a Introduction Advances in microlithography have enab!ed si nificant success in VLSI and ULSI technolog While suk0. lpm resist patterning awaits e-beam and %-ray lithography to become widely available, we show that existin DUV using an optimized Si,O,N ARC (anti-reflective coating) and hardmask layer. LPeVD SIN, as an ARC for patterning sub-0.5ym oly gates is reported in 11 and poly reported in [2]. However, these techni ues cannot be used nonuniform gas &strigution causes unacce table optical linewidths 2 0.12pm. DUV or anic ARC has several substrate reflectivity R) and &e infamous undercut resist reduction and etch of the relatively thick ARC layer leaves less resist available to mask the pol Si etch. In contrast, use of Si,O N, results in less resist 7;s~. Moreover, unlike organic ARE, Si,O,N has high selectivity during the polySi overetch. This hardmask roperty is critical in patterning marginal. SixO$J D%6 ARC on metal substrates was initially reported by Ogawa [3], but not used as an etch hardmask. Here we show that an o timized Si,OyN, for reflectance is more severe. Sup ression of substrate obtain sub-0.1 pm poly-Si gates with excelfent CD control. DUV ARC Film Design and Deposition The Si,OyN, refractive indices can be tuned over a wide range by varying its composition. Fig. 1 shows the real and imaginary parts of the refractive index (n and k) as the composition of Si,O N, is tuned by varying the deposition gas flow ratios. ?he Si,OyN, stack is deposited in a commercial P5000 PECVD system. A particle test on 700 wafers shows that there are <20 particles per wafer at size >0.16pm. The Si,O,N, film and backside of wafer is analyzed by TXRF and SIMS: S, C1, K, Ca, Ti, Cr, Fe, Ni, Cu A1 and Zn are at trace level or below detection limit. PECVD provides much better uniformity of n (0=0.19%) and k (0=1.4%) than LPCVD, which is important for ARC implementation. Our methodology to design the thickness and composition of the Si,O,N, ARC starts with Pr0 l i th i2~ simulations to determine the n and k needed for R ~ 1 % at various ARC thicknesses. Fig. 2 plots the simulated contour maps of R at 248nm at thicknesses of 20? 29, 40 nm. We superpose the experimental n vs. k (i.e., Fig.1) on these contours to determine whether at a particular thickness the n and k for R <1% coincide with what can be achieved with our Si,OyN, deposition process. Thus technology may be extended to achieve sub-0.lpm f eatures patterning with etch b a s and organic ARC E or i-line is to pattern sub-0.1 m oly gates. In t R e LPCVD process nonuniformity and biased etch with .i-line K RC achieved disadvantages compared to Si Oy. a,. In addition to higher profile, poor etch se i ectivity to resist during linewidth poly-Si over to ogra h w i ere the resist thicknesses are DUV can be employed to pattern PO P y-Si gates where the reflections and the etch selectivity o P Si,O N, enables us to matching the n and k for low R with reliable deposition conditions we o timize the Si,OyN, composition as n=2.15, k=0.63 and thictness as 29nm. Fig. 3 shows R vs. ARC thickness for this com osition. It indicates a comfortably wide process window (%O*SOA) for the film deposition. Lithogra hy and Etch Figure 4 shows the egect of the DUV ARC film on 0.18pm resist patterning on pol -Si over topography. Notching is completely eliminatei by im lementing the ARC over the poly-Si ate stack. Excelrent dense and isolated profiles of D&V resist over ARC/ oly-Si are >20% and the depth-of-focus is 1.2ym. We have patterned 0.22 to 0.26pm resist lines on polySi over 350nm LOCOS top0 raph using conventional DUV lithography with our AkC d m . To achieve subO.lpm gate length we use a plasma etcher to reduce the patterned resist to sub-0.lpm. During this etch the selectivity of Si,O,N, to resist allows us to optimize the etch for minimum resist loss. Finally a multiple step ARC/ oly-Si etch anisotro ically removes the exposed ost-etch poly-Si gate len ths. Resist linewidths are 0.228.26pm in 0.01 m intervak Using etch bias of 0.16pm, a hearity of sub-0.1 m poly-Si lines with small standard deviation (30=0.01$ is obtained. Fig. 7 shows the sub0.1 pm resist profile after linewidth reduction etch. Fi 8 shows the 0.06pm poly-Si line over a LOCOS to ograpfy. The final gate size which can be achievetis mainly determined by the resist linewidth reduction etch. Remaining resist height is important since it controls whether the poly-Si line is intact after poly etch. The use of Si,O N, is crucial to fabricate 0.06pm poly-Si gate with DUV lithography. Our results show that with or anic ARC 0.08pm is the maximum etch bias possible (Fig.f). Etch of a thick (160nm) organic ARC which does not have ood selectivity removes the resist top surface resulting an 8 not enough resist is left to complete the linewidth reduction etch. The selectivity of Si,OyN, durin the poly-Si overetch further he1 s in maintaining oly Si fines intact where the resist thicLess is margin$ such as for poly-Si over topography. Fig. 9 shows a top-view SEM micro raph of poly-Si lines over LOCOS topography. Broken Enes are observed due to oor resist pattern and low etch selectivity of or anic ARC.%o such roblem is encountered when our S i x O S ARC is used. F!g. 10 shows schematicall the hnewidzth reduction etch processes with Si,OyN A R z and with organic ARC. Si,O,N, serves as both an ARC and a hardmask for the fabrication of sub-0.1 pm poly-Si gate. Conclusions The fabrication of 0.06ym poly-Si gate with conventional DUV lithograph has been successfilly demonstrated using a designed '&,OyN, as DUV ARC and hardmask for linewidth reduction etch techniques. Excellent linearity and small standard deviation for poly gate lengths at sub-0.1 pm range are achieved. Lithography and etch processes implementing the Si,O N, are com ared to that with an organic ARC. It is found That that bo& the optical properties and the etch properties of Si,O N, are critical to achieving the fine poly-Si linewidths. b s e of commercial deposition equipment makes implementation of this process in a manufacturing environment feasible. shown in Fig. 5 . The exposure latitude for 0.2 s pm lines is poly-fi. Fig. 6 summarizes t t: e resist pattern linewidths and oly-Si gates d\" own to 0.06pm are obtained. Excellent 131 4-93081 3-75-1 197 1997 Symposium on VLSl Technology Digest of Technical Papers Acknowledgments We thank Chris Bencher for technical discussions and Trace Hurd for metal contamination analysis.","PeriodicalId":414778,"journal":{"name":"1997 Symposium on VLSI Technology","volume":"1 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"1997-06-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"11","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"1997 Symposium on VLSI Technology","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/VLSIT.1997.623733","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 11

Abstract

We report fabrication of sub-0.1 pm poly-Si gates using conventional DUV lithography with an optimized Si,O,N, film. This film has dual functions: reducing substrate reflectivit to <1%, and serving as a hardmask for the pol Si etch. &ith an aggressive etch bias process, linewidtxs down to 0.06ym are achieved with good linewidth control ( 3 ~ 1 2 n m ) and a near erfect lineari . Excellent optical manufacturable PECVD deposition process. uniformity of the n a n t k of the AR z is obtained with a Introduction Advances in microlithography have enab!ed si nificant success in VLSI and ULSI technolog While suk0. lpm resist patterning awaits e-beam and %-ray lithography to become widely available, we show that existin DUV using an optimized Si,O,N ARC (anti-reflective coating) and hardmask layer. LPeVD SIN, as an ARC for patterning sub-0.5ym oly gates is reported in 11 and poly reported in [2]. However, these techni ues cannot be used nonuniform gas &strigution causes unacce table optical linewidths 2 0.12pm. DUV or anic ARC has several substrate reflectivity R) and &e infamous undercut resist reduction and etch of the relatively thick ARC layer leaves less resist available to mask the pol Si etch. In contrast, use of Si,O N, results in less resist 7;s~. Moreover, unlike organic ARE, Si,O,N has high selectivity during the polySi overetch. This hardmask roperty is critical in patterning marginal. SixO$J D%6 ARC on metal substrates was initially reported by Ogawa [3], but not used as an etch hardmask. Here we show that an o timized Si,OyN, for reflectance is more severe. Sup ression of substrate obtain sub-0.1 pm poly-Si gates with excelfent CD control. DUV ARC Film Design and Deposition The Si,OyN, refractive indices can be tuned over a wide range by varying its composition. Fig. 1 shows the real and imaginary parts of the refractive index (n and k) as the composition of Si,O N, is tuned by varying the deposition gas flow ratios. ?he Si,OyN, stack is deposited in a commercial P5000 PECVD system. A particle test on 700 wafers shows that there are <20 particles per wafer at size >0.16pm. The Si,O,N, film and backside of wafer is analyzed by TXRF and SIMS: S, C1, K, Ca, Ti, Cr, Fe, Ni, Cu A1 and Zn are at trace level or below detection limit. PECVD provides much better uniformity of n (0=0.19%) and k (0=1.4%) than LPCVD, which is important for ARC implementation. Our methodology to design the thickness and composition of the Si,O,N, ARC starts with Pr0 l i th i2~ simulations to determine the n and k needed for R ~ 1 % at various ARC thicknesses. Fig. 2 plots the simulated contour maps of R at 248nm at thicknesses of 20? 29, 40 nm. We superpose the experimental n vs. k (i.e., Fig.1) on these contours to determine whether at a particular thickness the n and k for R <1% coincide with what can be achieved with our Si,OyN, deposition process. Thus technology may be extended to achieve sub-0.lpm f eatures patterning with etch b a s and organic ARC E or i-line is to pattern sub-0.1 m oly gates. In t R e LPCVD process nonuniformity and biased etch with .i-line K RC achieved disadvantages compared to Si Oy. a,. In addition to higher profile, poor etch se i ectivity to resist during linewidth poly-Si over to ogra h w i ere the resist thicknesses are DUV can be employed to pattern PO P y-Si gates where the reflections and the etch selectivity o P Si,O N, enables us to matching the n and k for low R with reliable deposition conditions we o timize the Si,OyN, composition as n=2.15, k=0.63 and thictness as 29nm. Fig. 3 shows R vs. ARC thickness for this com osition. It indicates a comfortably wide process window (%O*SOA) for the film deposition. Lithogra hy and Etch Figure 4 shows the egect of the DUV ARC film on 0.18pm resist patterning on pol -Si over topography. Notching is completely eliminatei by im lementing the ARC over the poly-Si ate stack. Excelrent dense and isolated profiles of D&V resist over ARC/ oly-Si are >20% and the depth-of-focus is 1.2ym. We have patterned 0.22 to 0.26pm resist lines on polySi over 350nm LOCOS top0 raph using conventional DUV lithography with our AkC d m . To achieve subO.lpm gate length we use a plasma etcher to reduce the patterned resist to sub-0.lpm. During this etch the selectivity of Si,O,N, to resist allows us to optimize the etch for minimum resist loss. Finally a multiple step ARC/ oly-Si etch anisotro ically removes the exposed ost-etch poly-Si gate len ths. Resist linewidths are 0.228.26pm in 0.01 m intervak Using etch bias of 0.16pm, a hearity of sub-0.1 m poly-Si lines with small standard deviation (30=0.01$ is obtained. Fig. 7 shows the sub0.1 pm resist profile after linewidth reduction etch. Fi 8 shows the 0.06pm poly-Si line over a LOCOS to ograpfy. The final gate size which can be achievetis mainly determined by the resist linewidth reduction etch. Remaining resist height is important since it controls whether the poly-Si line is intact after poly etch. The use of Si,O N, is crucial to fabricate 0.06pm poly-Si gate with DUV lithography. Our results show that with or anic ARC 0.08pm is the maximum etch bias possible (Fig.f). Etch of a thick (160nm) organic ARC which does not have ood selectivity removes the resist top surface resulting an 8 not enough resist is left to complete the linewidth reduction etch. The selectivity of Si,OyN, durin the poly-Si overetch further he1 s in maintaining oly Si fines intact where the resist thicLess is margin$ such as for poly-Si over topography. Fig. 9 shows a top-view SEM micro raph of poly-Si lines over LOCOS topography. Broken Enes are observed due to oor resist pattern and low etch selectivity of or anic ARC.%o such roblem is encountered when our S i x O S ARC is used. F!g. 10 shows schematicall the hnewidzth reduction etch processes with Si,OyN A R z and with organic ARC. Si,O,N, serves as both an ARC and a hardmask for the fabrication of sub-0.1 pm poly-Si gate. Conclusions The fabrication of 0.06ym poly-Si gate with conventional DUV lithograph has been successfilly demonstrated using a designed '&,OyN, as DUV ARC and hardmask for linewidth reduction etch techniques. Excellent linearity and small standard deviation for poly gate lengths at sub-0.1 pm range are achieved. Lithography and etch processes implementing the Si,O N, are com ared to that with an organic ARC. It is found That that bo& the optical properties and the etch properties of Si,O N, are critical to achieving the fine poly-Si linewidths. b s e of commercial deposition equipment makes implementation of this process in a manufacturing environment feasible. shown in Fig. 5 . The exposure latitude for 0.2 s pm lines is poly-fi. Fig. 6 summarizes t t: e resist pattern linewidths and oly-Si gates d" own to 0.06pm are obtained. Excellent 131 4-93081 3-75-1 197 1997 Symposium on VLSl Technology Digest of Technical Papers Acknowledgments We thank Chris Bencher for technical discussions and Trace Hurd for metal contamination analysis.
查看原文
分享 分享
微信好友 朋友圈 QQ好友 复制链接
本刊更多论文
用设计的Si/sub x/O/sub y/N/sub z/薄膜作为电弧和硬掩膜,用Duv光刻技术制备0.06 /spl mu/m多晶硅栅极
我们报告了使用优化的Si,O,N薄膜,使用传统DUV光刻技术制造低于0.1 pm的多晶硅栅极。该膜具有双重功能:将基材反射率降低至0.16pm。利用TXRF和SIMS对硅片的Si、O、N、薄膜和背面进行了分析:S、C1、K、Ca、Ti、Cr、Fe、Ni、Cu、A1和Zn均处于痕量水平或低于检测限。PECVD提供了比LPCVD更好的n(0=0.19%)和k(0=1.4%)的均匀性,这对ARC的实现很重要。我们设计Si,O,N, ARC的厚度和成分的方法从Pr0 1开始,通过2~ 2的模拟来确定在不同的ARC厚度下R ~ 1%所需的N和k。图2为R在248nm处,厚度为20?29,40 nm。我们将实验n与k(即图1)叠加在这些轮廓上,以确定在特定厚度下,n和k的R是否为20%,焦深是否为1.2ym。我们使用传统的DUV光刻技术和我们的AkC d m在超过350nm的LOCOS top0图上绘制了0.22至0.26pm的抗蚀线。实现subO。LPM栅极长度我们使用等离子蚀刻器将图案电阻降低到0 LPM以下。在此蚀刻过程中,硅、氧、氮的选择性使我们能够优化蚀刻以实现最小的电阻损失。最后,多步ARC/多晶硅刻蚀各向异性去除暴露的成本刻蚀多晶硅栅极长度。利用0.16pm的蚀刻偏压,获得了质量小于0.1 m的多晶硅线,其标准偏差较小(30=0.01美元)。图7显示了线宽减小蚀刻后低于0.1 pm的抗蚀剂轮廓。图8显示了0.06pm多晶硅线在LOCOS上的位置。最终能达到的栅极尺寸主要取决于蚀刻抗蚀剂的线宽减小。剩余抗蚀剂高度很重要,因为它控制着多晶硅线在聚蚀刻后是否完好无损。使用Si,O N,是用DUV光刻制作0.06pm多晶硅栅极的关键。我们的研究结果表明,0.08pm是最大的蚀刻偏差可能(图f)。蚀刻厚的(160nm)有机电弧,它没有很好的选择性去除抗蚀剂的顶部表面,导致8没有足够的抗蚀剂留下,以完成线宽减小蚀刻。在多晶硅复刻过程中,硅OyN的选择性进一步提高了在抗蚀剂厚度小于0的情况下(如多晶硅复刻)保持多晶硅细粒完整的能力。图9显示了LOCOS地形上多晶硅线的SEM俯视图。由于有机电弧的抗蚀性差和低的蚀刻选择性,观察到断裂的ene。在使用我们的ssi x sarc时,不会遇到这样的问题。F, g。图10所示为Si、oyna、rz和有机电弧还原刻蚀过程示意图。Si,O,N作为ARC和硬掩膜,用于制作低于0.1 pm的多晶硅栅极。结论采用设计的'&,OyN,作为DUV电弧和硬掩膜,采用线宽减小蚀刻技术,成功地在传统DUV光刻机上制备了0.06ym多晶硅栅极。在0.1 pm以下的范围内实现了极好的线性和小的标准偏差。实现Si,O, N的光刻和蚀刻工艺与有机电弧相比较。结果表明,硅的光学性能和蚀刻性能是实现良好的多晶硅线宽的关键。商用沉积设备的B - s - e使得该工艺在制造环境中实现是可行的。如图5所示。曝光纬度为0.2 s pm线为多保真度。图6总结了得到的电阻图样线宽和0.06pm范围内的硅硅栅极。我们感谢Chris Bencher的技术讨论和Trace Hurd的金属污染分析。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 去求助
来源期刊
自引率
0.00%
发文量
0
期刊最新文献
Energy funnels - A new oxide breakdown model Fully Planarized Stacked Capacitor Cell With Deep And High Aspect Ratio Contact Hole For Gigs-bit DRAM Impact Of Trench Sidewall Interface Trap In Shallow Trench Isolation On Junction Leakage Current Characteristics For Sub-0.25 /spl mu/m CMOS Devices 0.25 /spl mu/m salicide CMOS Technology Thermally Stable Up To 1,000/spl deg/C With High TDDB Reliability Dielectric Planarization Using Mn203 Slurry
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
现在去查看 取消
×
提示
确定
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
已复制链接
已复制链接
快去分享给好友吧!
我知道了
×
扫码分享
扫码分享
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 联系我们:info@booksci.cn
Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。
Copyright © 2023 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1