In situ collector cleaning and extreme ultraviolet

In Situ Collector Cleaning And Extreme Ultraviolet-Free PDF

  • Date:23 Jun 2020
  • Views:21
  • Downloads:0
  • Pages:9
  • Size:2.04 MB

Share Pdf : In Situ Collector Cleaning And Extreme Ultraviolet

Download and Preview : In Situ Collector Cleaning And Extreme Ultraviolet


Report CopyRight/DMCA Form For : In Situ Collector Cleaning And Extreme Ultraviolet


Transcription:

In situ collector cleaning and extreme ultraviolet reflectivity restoration. by hydrogen plasma for extreme ultraviolet sources. Daniel T Elg, Department of Nuclear Plasma and Radiological Engineering Center for Plasma Material Interactions. University of Illinois at Urbana Champaign Urbana Illinois 61801. John R Sporre, Department of Nuclear Plasma and Radiological Engineering Center for Plasma Material Interactions. University of Illinois at Urbana Champaign Urbana Illinois 61801 and IBM Corporation Albany. New York 12203,Gianluca A Panici, Department of Nuclear Plasma and Radiological Engineering Center for Plasma Material Interactions. University of Illinois at Urbana Champaign Urbana Illinois 61801. Shailendra N Srivastava, Applied Research Institute University of Illinois at Urbana Champaign Champaign Illinois 61820. David N Ruzica, Department of Nuclear Plasma and Radiological Engineering Center for Plasma Material Interactions.
University of Illinois at Urbana Champaign Urbana Illinois 61801. Received 14 December 2015 accepted 4 February 2016 published 23 February 2016. Laser produced Sn plasmas used to generate extreme ultraviolet EUV light for lithography cause. the release of Sn ions and neutrals in the EUV source chamber These Sn atoms condense and. deposit on the multilayer collector optic which reduces its ability to reflect EUV light This lowers. the source throughput and eventually necessitates downtime for collector cleaning In this paper an. in situ plasma based collector cleaning technique is presented and experimentally demonstrated. First the technique is shown to completely clean a 300 mm diameter stainless steel dummy collec. tor Second simulations and secondary ion mass spectroscopy depth profiles show that the tech. nique does not erode the real multilayer mirrors Finally EUV reflectivity measurements. demonstrate the ability of the technique to restore EUV reflectivity to Sn coated multilayer mirrors. This technique has the potential to be used in conjunction with source operation eliminating. cleaning related source downtime V C 2016 American Vacuum Society. http dx doi org 10 1116 1 4942456, I INTRODUCTION availability for high volume manufacturing HVM. In recent decades massive advances have been made in Additionally after EUV has gained HVM insertion source. the semiconductor industry by adherence to Moore s law power requirements will continue to rise as the feature size. which states that the number of transistors on a single inte continues to shrink 5. grated circuit chip must double every two years 1 In just 30 Both EUV power to the wafer and source availability are. years the minimum feature size on a chip has shrunk from hampered by the need for collector cleaning EUV photons. 1 lm to 14 nm 2 This progress has been enabled by consist are created by a dense Te 20 eV ne 1019 cm 3 laser. ent advances in optical lithography Among the parameters produced Sn plasma 4 6 7 Due to the poor reflectivity and. which affect the minimum resolution of a lithography system high transmissivity of all known solids the optics which. is the wavelength of the light source used to pattern photore focus these photons must employ Bragg reflection by means. sist on Si wafers Historically that wavelength was near to of 7 nm thick Mo Si bilayers which cause Bragg reflection. or smaller than the minimum feature size 3 Since the adop of 13 5 nm light 8 10 Such optics are known as multilayer. tion of the 193 nm excimer laser in 2001 however the wave mirrors MLMs The first of these mirrors the collector. length used in high volume optical lithography has not optic is directly exposed to the EUV plasma which deposits. decreased Accordingly there is a motivation to enable fur Sn on the collector and degrades EUV reflectivity EUVR. ther size reduction by reducing the wavelength used in While debris mitigation techniques such as magnetic mitiga. lithography tion to deflect ions and buffer gas to deflect neutrals. In particular research has focused on extreme ultraviolet exist 11 12 no debris mitigation technique can completely. EUV lithography which uses a 13 5 nm light source eliminate Sn deposition on the collector Thus as Sn accu. While EUV sources have shown remarkable progress in mulates EUV power at the wafer is reduced until the collec. recent years 4 they cannot yet meet the required power and tor must be either cleaned or replaced incurring costs and. The best way to minimize downtime is to clean the collec. Electronic mail druzic illinois edu tor while in the chamber in situ This can be accomplished. 021305 1 J Vac Sci Technol A 34 2 Mar Apr 2016 0734 2101 2016 34 2 021305 8 30 00 C 2016 American Vacuum Society. V 021305 1, Redistribution subject to AVS license or copyright see http scitation aip org termsconditions IP 40 139 112 162 On Tue 23 Feb 2016 15 12 05. 021305 2 Elg et al In situ collector cleaning and EUV reflectivity restoration 021305 2. with hydrogen radicals which etch Sn by forming the gas. SnH4 Hydrogen radicals have been previously shown to etch. Sn 13 14 However these experiments have been performed by. utilizing a hot filament radical source and then blowing the. radicals at an Sn coated sample While this is a possible tech. nique its application to a real EUV system could necessitate. the insertion of a delivery system in front of the collector. causing downtime and could be subject to radical diffusion. and recombination on the walls of the delivery system in the. The novel cleaning solution described in this paper is to. create the radicals directly on the collector surface by using. the collector itself to drive a capacitively coupled hydrogen. plasma This paper shows successful cleaning of a 300 mm. stainless steel dummy collector optic by means of this tech FIG 2 Color online a Collector is installed with electrically isolating. nique and the removal rates are measured Simulations and Teflon clamps b The collector driving a hydrogen plasma with the collec. tor itself acting as the antenna, secondary ion mass spectroscopy SIMS depth profiles are. undertaken to show that the plasma does not erode different. multilayer mirror surfaces Finally the technique is shown to radicals then reactively etch Sn by forming SnH4 For the. restore EUV reflectivity to Sn coated MLMs The develop experiments shown in this paper the hydrogen pressure was. ment of an in situ cleaning technique without a delivery sys 65 mTorr and the flow rate was 500 sccm The gas was. tem has the potential to run at the same time as the EUV injected through an inlet behind the center hole of the. source enabling restoration EUV reflectivity and source. A picture of XCEED is shown in Fig 1 Pictures of the. power throughput without any cleaning related downtime. dummy collector with and without a plasma are shown in. Fig 2 A circuit diagram is shown in Fig 3,II EXPERIMENTAL SETUP AND PLASMA SOURCE. Deposition was carried out in a separate chamber with a. Etching was performed in the Xtreme Commercial EUV DC magnetron operating at 30 mA of current in approxi. Exposure Diagnostic XCEED chamber XCEED originally mately 3 mTorr of Ar A quartz crystal monitor QCM was. designed as a Xe based discharge produced EUV source used to measure deposition thickness The entire collector. was repurposed to hold a stainless steel dummy collector was coated with Sn For removal rate experiments masked. optic The collector was 300 mm in diameter and was isolated Si witness plates were attached along a collector radius in. from the chamber ground with polytetrafluoroethylene clamps order to yield measurements of local removal rate as shown. The collector was attached through a matching network to a in Fig 4 For experiments involving MLM samples these. 300 W 13 56 MHz RF source and a capacitively coupled samples instead were placed on the collector area and some. hydrogen plasma was broken down on the surface of the col bare Si area had been exposed to the plasma while other. lector This plasma creates H radicals as well as ions that can parts of each area had not This allowed for measurement of. produce H radicals upon impact with the surfaces 15 the various interfaces by the profilometer In particular each sam. ple was split into four quadrants each of which had been. exposed to a different set of conditions etched Sn was coated. with Sn and exposed to the etching plasma etched Si was. never coated with Sn but was exposed to the etching plasma. FIG 1 Color online XCEED is shown with the collector installed For. etching experiments the chamber on the cart at left was attached to the. former EUV source at right The collector was driven with 300 W 13 56 FIG 3 Circuit diagram of the plasma source setup is shown The collector is. MHz RF power through an electrical feedthrough which allowed for electri isolated inside XCEED and is attached to a 300 W 13 56 MHz RF supply A. cal connection to the electrically isolated dummy collector matching network serves to minimize reflected power. J Vac Sci Technol A Vol 34 No 2 Mar Apr 2016, Redistribution subject to AVS license or copyright see http scitation aip org termsconditions IP 40 139 112 162 On Tue 23 Feb 2016 15 12 05.
021305 3 Elg et al In situ collector cleaning and EUV reflectivity restoration 021305 3. It is known that SnH4 easily decomposes and redeposits Sn. upon collision with metal surfaces 17 Despite Sn coverage of. the entire collector redeposition was not able to prohibit col. lector cleaning Complete etches were observed for 20 50. and 100 nm experiments Profilometry indicated no difference. in height between the etched Sn and etched Si quadrants. Additionally SEM images indicated that the etched Sn quad. rants were devoid of Sn and composed solely of pristine Si. A comparison of the etched Sn quadrant and the masked. Sn quadrant of a 20 nm sample is shown in Fig 6 Figure 7. shows a backscattered electron image of all four quadrants. of one of the 50 nm samples This alternative SEM technique. is sensitive not to topology but to material composition. thus the fact that the Etched Sn quadrant appears to have the. same darkness as the etched Si quadrant is indicative of a. complete etch, When coated with 200 nm of Sn the collector was not. completely cleaned after 2 h of etching Due to incomplete. etching removal rates could be calculated Witness plates. FIG 4 Diagram of the collector is shown with Si witness plates attached in analyzed on the profilometer yielded the removal rates. five different positions The entire collector was coated with Sn during depo shown in Fig 8 Two scans were taken for each sample. sitions however to measure local removal rates Si witness plates were also. placed on the collector during deposition and etching These were later ana. lyzed in a profilometer, masked Sn was coated with Sn but not exposed to the etching. plasma and masked Si was never coated with Sn or exposed. to the etching plasma A diagram is shown in Fig 5, SEM and AFM were also used for characterization of cer. tain samples A Langmuir probe was used to determine plasma. potential theory and operation are described in Ref 16 For. MLM surface damage experiments depth profiles were deter. mined with SIMS to see if etching had removed the MLM. capping layer For EUV reflectivity experiments the advanced. light source synchrotron at Lawrence Berkeley National. Laboratory was used to determine EUV reflectivity,III RESULTS AND DISCUSSION. A Removal rate experiment, Sn removal experiments were carried out for initial depo.
sitions of 20 50 100 and 200 nm Each etch was carried out. for 2 h After each experiment was completed samples were. taken to the profilometer and SEM for characterization. FIG 5 Masking was employed during etching and deposition to yield four. quadrants on each witness plate each quadrant had been exposed to differ. ent conditions etched Sn was coated with Sn and exposed to the etching FIG 6 SEM images show the difference between the plasma cleaned section. plasma etched Si was never coated with Sn but was exposed to the etch of a witness plate and the section that was coated with Sn but not exposed to. ing plasma masked Sn was coated with Sn but not exposed to the etching plasma a The masked Sn quadrant shows grains of deposited Sn indicat. plasma and masked Si was never coated with Sn or exposed to the etch ing the condition of the surface before etching b The etched Sn quadrant. ing plasma which was formerly Sn coated appears pristine after plasma cleaning. JVST A Vacuum Surfaces and Films, Redistribution subject to AVS license or copyright see http scitation aip org termsconditions IP 40 139 112 162 On Tue 23 Feb 2016 15 12 05. 021305 4 Elg et al In situ collector cleaning and EUV reflectivity restoration 021305 4. the 200 nm samples revealed that after spending 2 h exposed. to the etching plasma the etched Si quadrant had a rough. ness of only 3 2 A Such a low roughness is close to the typi. cal roughness value for a polished and very carefully. handled Si wafer 1 5 A 18 Such a small increase in rough. ness can be attributed to the fact that the sample was handled. and cut outside a cleanroom thus the measured value of. 3 2 A does not indicate plasma caused surface damage. A deeper investigation of surface damage was undertaken. through stopping and range of ions in matter SRIM model. ing and SIMS depth profiles of plasma cleaned MLM sam. ples First to give an estimate of possible ion energ. In situ collector cleaning and extreme ultraviolet reflectivity restoration by hydrogen plasma for extreme ultraviolet sources Daniel T Elg Department of Nuclear Plasma and Radiological Engineering Center for Plasma Material Interactions University of Illinois at Urbana Champaign Urbana Illinois 61801 John R Sporre

Related Books