https://lippmaa.issp.u-tokyo.ac.jp/contact https://lippmaa.issp.u-tokyo.ac.jp/diffusion/ https://lippmaa.issp.u-tokyo.ac.jp/diffusion/danim.gif interdiffusion Illustration of interdiffusion between thin film layers at high temprature https://lippmaa.issp.u-tokyo.ac.jp/diffusion/lfo https://lippmaa.issp.u-tokyo.ac.jp/diffusion/cai098-rheed.png RHEED oscillation RHEED specular spot intensity oscillations during LaFeO3 growth. https://lippmaa.issp.u-tokyo.ac.jp/diffusion/cai098-2.png STM after depo STM image of a LaFeO3 film surface after deposition at 600°C on a 0.05% Nb:SrTiO3 substrate. 400 × 400 nm2 https://lippmaa.issp.u-tokyo.ac.jp/diffusion/cai098-ann.png STM after anneal STM image after annealing. 400 × 400 nm2 https://lippmaa.issp.u-tokyo.ac.jp/diffusion/cai098-ffm.png AFM and FFM after anneal AFM and FFM images of the annealed surface. 1000 × 1000 nm2 https://lippmaa.issp.u-tokyo.ac.jp/diffusion/caiciss1.png CAICISS spectra CAICISS time of flight data before (A) and after (B) annealing the film https://lippmaa.issp.u-tokyo.ac.jp/dynamics/ https://lippmaa.issp.u-tokyo.ac.jp/dynamics/polished.png Polished SrTiO3 surface AFM image of a commercial polished SrTiO3 substrate https://lippmaa.issp.u-tokyo.ac.jp/dynamics/annealed.png Annealed SrTiO3 Step-and-terrace surface image of an annealed SrTiO3 substrate https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/ https://lippmaa.issp.u-tokyo.ac.jp/gallery/labicon.png Lab life Events in Lippmaa laboratory https://lippmaa.issp.u-tokyo.ac.jp/gallery/workicon.png Kashiwa life Conference posters presented by Lippmaa lab students and staff https://lippmaa.issp.u-tokyo.ac.jp/gallery/lifeicon.png Kashiwa life Various tiny creatures that live in Kashiwa https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0010.jpg Life in Lippmaa lab 2000 October, Nakagawa on the way to a WOE meeting in Switzerland. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0012.jpg Life in Lippmaa lab 2001 June, (TIT) Nakagawa, Ohkubo, Shiroki, Kato Terai, Itaka, Kawasaki, Ohnishi, Ohtani in the common space. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0014.jpg Life in Lippmaa lab 2001 June, (TIT) Tsukazaki and Nakagawa checking the fire extinguisher. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0016.jpg Life in Lippmaa lab 2001 September, Shibuya at a seminar. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0018.jpg Life in Lippmaa lab 2002 October, Kawasaki, Triscone, Lippmaa, and Hwang at the Florida WOE meeting. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0019.jpg Life in Lippmaa lab 2002 December, Lippmaa, Ohnishi, and Shibuya visiting Nakagawa at Bell Labs. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0020.jpg Life in Lippmaa lab 2004 April, Cherry blossom party in the Kashiwanoha park. Meguro, Kawamura, Ohnishi, Lippmaa, Uozumi, Jimi. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0024.jpg Life in Lippmaa lab 2004 May, Itaka in Kodomonokuni with Osawa and https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0030.jpg Life in Lippmaa lab 2004 December, Uozumi and Meguro opening Christmas presents. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0040.jpg Life in Lippmaa lab 2004 December, Jimi doing a small performance during a party. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0060.jpg Life in Lippmaa lab 2006 September, Lab picture just before Shibuya\'s departure. From the left: Sato, Meguro, Lippmaa, Nishio, Shibuya, Kondo, Ohnishi, Urata, Kawamura. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0070.jpg Life in Lippmaa lab 2007 June, Kawamura-san\'s birthday cake. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0080.jpg Life in Lippmaa lab 2007 July, Kang and the Mouse. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0090.jpg Life in Lippmaa lab 2007 July, Takeuchi-sensei won a bowling prize! https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0100.jpg Life in Lippmaa lab 2007 July, Abe having his hair cut. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0110.jpg Life in Lippmaa lab 2007 August, Hunter from Maryland at the Takeuchi farewell BBQ. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0120.jpg Life in Lippmaa lab 2007 August, Itaka at the summer BBQ. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0130.jpg Life in Lippmaa lab 2007 August, Matvejeff at the summer BBQ. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0140.jpg Life in Lippmaa lab 2007 September, Nishio in Sapporo at the JSAP fall meeting. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0144.jpg Life in Lippmaa lab 2008 November, Ion scattering spectrometer and STM. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0146.jpg Life in Lippmaa lab 2009 February, Ablation plume in a miniature deposition chamber. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0148.jpg Life in Lippmaa lab 2009 August, Thin film substrate being heated in a deposition chamber. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0149.jpg Life in Lippmaa lab 2009 September, Film thickness calibration sample. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0150.jpg Life in Lippmaa lab 2009 September, Kikuzuki, Ohtsuka, and Kozuka at the Toyama JSAP meeting. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0160.jpg Life in Lippmaa lab 2009 September, Grilling fish in front of the MegaGauss building. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0170.jpg Life in Lippmaa lab 2009 October, Tired-looking Kikuzuki and Ohtsuka preparing for the open campus. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0180.jpg Life in Lippmaa lab 2009 October, Demonstrating an excimer laser gun to visitors. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0182.jpg Life in Lippmaa lab 2009 August, Nishio introducing the lab to high-school students. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0190.jpg Life in Lippmaa lab 2009 October, Takahashi and Hikita after finishing the open campus presentations. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0200.jpg Life in Lippmaa lab 2009 October, Matvejeff and Bell discussing the finer points of cooking octopus. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0202.jpg Life in Lippmaa lab 2009 October, Directional transport anisotropy measurement in a nanowire sample. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0204.jpg Life in Lippmaa lab 2009 November, Kikuzuki learning how to use the die bonder. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0210.jpg Life in Lippmaa lab 2009 December, Santa\'s visit https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0212.jpg Life in Lippmaa lab 2010 January, A magnetostriction measurement system built by students. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0214.jpg Life in Lippmaa lab 2010 February, Kikuzuki, Ohtsuka, Takahashi, and Nishio cleaing the office. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0216.jpg Life in Lippmaa lab 2010 March, Matvejeff receiving the Marubun foundation grant. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0220.jpg Life in Lippmaa lab 2010 September, Cole and Kawasaki at the Applied Physics meeting https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0222.jpg Life in Lippmaa lab 2011 January, Preparing the internal wiring of a cryostat. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0224.jpg Life in Lippmaa lab 2011 March, Assembly of a low-temperature vacuum probing station. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0230.jpg Life in Lippmaa lab 2011 April, Harada and Nishio at a farewell party https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0240.jpg Life in Lippmaa lab 2011 May, Birthday party with Harada, Takahashi, Kawasaki, and Yoshida https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0250.jpg Life in Lippmaa lab 2011 July, 10-year anniversary party. Back: Sato, Cole, Harada, Yoshida, Tsubouchi, Peltier, Matsumoto, Ohkubo, Ogawa, Middle: Itaka, Ohtsuka, Kikuzuki, Nishio, Terai, Kozuka, Kumigashira, Nakagawa, Sasamura, Front: Kawamura, Ohnishi, Lippmaa, Koinuma, Takahashi, Kawasaki https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0252.jpg Life in Lippmaa lab 2011 November, Nice clouds over Kashiwa. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0254.jpg Life in Lippmaa lab 2012 January, Snowy day in Kashiwa. https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0260.jpg Life in Lippmaa lab 2012 April, Hanami, Yoshioka, Ogawa, Peltier, Lippmaa, Tsubouchi, Harada, Hou, Front: Kawamura, Kawasaki https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0270.jpg Life in Lippmaa lab 2013 March, Farewell party, Hou, Tsubouchi, Ohkubo, Lippmaa, Takahashi, Harada, Harada, Shimizu, Kawasaki, Nishio, Ogawa, Yoshida https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0280.jpg Life in Lippmaa lab 2013 June, Hou, Matvejeff, Lippmaa in the lab https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0290.jpg Life in Lippmaa lab 2013 July, Takahashi birthday https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0300.jpg Life in Lippmaa lab 2013 October, Farewell party, Ahvenniemi, peltier, Matvejeff https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0310.jpg Life in Lippmaa lab 2014 March, Kawasaki receiving the JSAP prize https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0320.jpg Life in Lippmaa lab 2014 October, Lippmaa demonstrates a van de Graaf generator during an open campus event https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0330.jpg Life in Lippmaa lab 2015 February, Visiting Shinkosha, Hou, Shono, Kawasaki, Takahashi, Lippmaa, Nishitani, Ohtsuki https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0340.jpg Life in Lippmaa lab 2015 April, Welcome party, Hou, Kawasaki, Osawa, Shono, Lee, Takahashi, Lippmaa, Nishitani, Ohtsuki https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0350.jpg Life in Lippmaa lab 2015 April, Sakai\'s birthday, Osawa, Hou, Lee, Takahashi, Sakai, Lippmaa, Kawamura, Shono, Kawasaki https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0360.jpg Life in Lippmaa lab 2015 May, Student Guidance event https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0370.jpg Life in Lippmaa lab 2015 May, Laboratory introduction guidance https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0380.jpg Life in Lippmaa lab 2015 May, After finishing a guidance event, Kawasaki, Osawa, Lee https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0390.jpg Life in Lippmaa lab 2015 June, Hou, Lee, Osawa, and Shono with a prize at the ISSP beer party https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0400.jpg Life in Lippmaa lab 2015 June, Kawamura birthday https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0410.jpg Life in Lippmaa lab 2015 October, After Kawasaki\'s PhD predefence: Hou, Shono, Kawamura, Ohtsuki, Osawa, Mihee Lee, Jiyeon Lee, Lippmaa, Sakai, Sakai, Kawasaki, Kawasaki https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0420.jpg Life in Lippmaa lab 2015 October, Kawasaki receives the Gold Award at the STAC-9 meeting https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0430.jpg Life in Lippmaa lab 2015 December, Sakai wedding: Shono, Hou, Sakai, Sakai, Lee https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0450.jpg Life in Lippmaa lab 2015 December, Santa Claus distributing gifts at the laboratory Christmas party https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0460.jpg Life in Lippmaa lab 2016 January, Celebrating Lippmaa\'s first half-century https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0470.jpg Life in Lippmaa lab 2016 February, Valentine\'s day chocolate party https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0480.jpg Life in Lippmaa lab 2016 March, Kawasaki and Shono graduation https://lippmaa.issp.u-tokyo.ac.jp/gallery/lab/Images/0490.jpg Life in Lippmaa lab 2016 May, Sakai leaving Kashiwa: Osawa, Takahashi, Hosokawa, Hou, Sakai, Kihara, Lippmaa, Lee https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/ https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/lifeicon.png Life https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN0242.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN0242.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN0242.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN0670.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN0670.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN0670.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN0677.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN0677.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN0677.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN0715.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN0715.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN0715.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN0798.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN0798.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN0798.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN0806.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN0806.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN0806.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN0811.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN0811.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN0811.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN0839.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN0839.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN0839.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN0855.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN0855.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN0855.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN1127.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN1127.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN1127.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN1146.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN1146.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN1146.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN1762.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN1762.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN1762.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN2158.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN2158.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN2158.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN2180.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN2180.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN2180.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN2217.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN2217.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN2217.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN2667.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN2667.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN2667.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN2698.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN2698.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN2698.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN2770.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN2770.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN2770.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN3272.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN3272.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN3272.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN3274.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN3274.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN3274.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN3282.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN3282.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN3282.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN3318.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN3318.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN3318.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN3325.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN3325.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN3325.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN3505.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN3505.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN3505.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN3508.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN3508.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN3508.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN3512.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN3512.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN3512.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN3524.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN3524.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN3524.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN3532.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN3532.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN3532.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN3542.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN3542.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN3542.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN4366.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN4366.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN4366.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN4388.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN4388.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN4388.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN4398.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN4398.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN4398.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN4407.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN4407.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN4407.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP0076.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP0076.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP0076.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP0079.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP0079.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP0079.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP0080.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP0080.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP0080.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP0084.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP0084.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP0084.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP0090.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP0090.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP0090.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP0117.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP0117.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP0117.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP0123.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP0123.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP0123.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP0126.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP0126.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP0126.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP0128.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP0128.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP0128.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP0140.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP0140.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP0140.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP0141.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP0141.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP0141.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP0144.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP0144.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP0144.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP0145.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP0145.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP0145.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP0146.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP0146.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP0146.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP0147.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP0147.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP0147.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP0148.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP0148.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP0148.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP0276.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP0276.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP0276.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP0278.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP0278.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP0278.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP0294.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP0294.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP0294.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP0334.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP0334.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP0334.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP0354.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP0354.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP0354.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP0368.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP0368.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP0368.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP0427.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP0427.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP0427.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP0430.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP0430.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP0430.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP0442.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP0442.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP0442.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP1197.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP1197.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP1197.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP1310.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP1310.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP1310.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP1316.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP1316.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP1316.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP1377.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP1377.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP1377.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP1385.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP1385.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP1385.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP1386.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP1386.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP1386.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP1404.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP1404.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP1404.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP1405.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP1405.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP1405.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP1498.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP1498.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP1498.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP1635.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP1635.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP1635.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP1637.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP1637.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP1637.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP1652.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP1652.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP1652.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP1710.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP1710.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP1710.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP1819.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP1819.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP1819.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP1840.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP1840.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP1840.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP1842.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP1842.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP1842.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP1866.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP1866.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP1866.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP1910.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP1910.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP1910.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP1917.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP1917.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP1917.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP1918.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP1918.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP1918.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP1924.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP1924.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP1924.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP1927.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP1927.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP1927.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP1929.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP1929.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP1929.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP1932.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP1932.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP1932.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP1945.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP1945.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP1945.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP3363.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP3363.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP3363.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP3365.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP3365.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP3365.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP3369.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP3369.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP3369.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP3380.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP3380.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP3380.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP3434.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP3434.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP3434.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP3448.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP3448.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP3448.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP3459.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP3459.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP3459.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP3538.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP3538.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP3538.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP3545.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP3545.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP3545.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP3553.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP3553.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP3553.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP3558.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP3558.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP3558.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP3562.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP3562.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP3562.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP3604.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP3604.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP3604.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP3623.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP3623.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP3623.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP3631.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP3631.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP3631.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP3632.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP3632.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP3632.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/PHON0624.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/PHON0624.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/PHON0624.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/PHON0722.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/PHON0722.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/PHON0722.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/PHON1017.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/PHON1017.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/PHON1017.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN0452.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN0452.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN0452.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN0502.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN0502.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN0502.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN0514.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN0514.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN0514.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN0541.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN0541.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN0541.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN2406.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN2406.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN2406.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN2413.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN2413.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN2413.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP0465.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP0465.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP0465.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP0741.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP0741.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP0741.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP0775.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP0775.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP0775.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP2036.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP2036.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP2036.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP2082.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP2082.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP2082.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP2083.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP2083.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP2083.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP2085.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP2085.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP2085.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP2299.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP2299.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP2299.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP2406.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP2406.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP2406.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP2408.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP2408.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP2408.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP2435.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP2435.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP2435.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP2437.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP2437.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP2437.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP2437.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP2440.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP2440.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP2440.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP2443.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP2443.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP2443.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP2488.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP2488.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP2488.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP2910.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP2910.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP2910.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP2931.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP2931.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP2931.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP2932.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP2932.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP2932.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP2980.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP2980.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP2980.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP2986.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP2986.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP2986.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP2992.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP2992.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP2992.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP3066.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP3066.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP3066.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/IMGP3101.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/IMGP3101.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/IMGP3101.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Large/DSCN4436.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Small/DSCN4436.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/Tiny/DSCN4436.JPG Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/nogl.jpg Insect gallery Images of various insects found in Kashiwa https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/nogl.png Insect gallery Images of various insects found in Kashiwa https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/M01.jpg Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/M02.jpg Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/M03.jpg Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/M04.jpg Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/M05.jpg Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/M06.jpg Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/M07.jpg Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/M08.jpg Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/M09.jpg Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/M10.jpg Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/M11.jpg Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/life/M12.jpg Kashiwa insect gallery https://lippmaa.issp.u-tokyo.ac.jp/gallery/ https://lippmaa.issp.u-tokyo.ac.jp/gallery/workicon.png Poster Gallery Conference posters presented by Lippmaa lab students and staff https://lippmaa.issp.u-tokyo.ac.jp/gallery/labicon.png Picture galleries https://lippmaa.issp.u-tokyo.ac.jp/gallery/workicon.png Picture galleries https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/ https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3600-05.jpg ferrite InFeCoO4 There are many well-known magnetic materials among spinel ferrites. The grain orientation and size can be controlled when spinels are grown on mismatched cubic substrates. Varying the grain size has a strong effect on the magnetic structure of the film due to the appearance of spin glass or superparamagnetic states. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3600-2.jpg ferrite InFeCoO4 There are many well-known magnetic materials among spinel ferrites. The grain orientation and size can be controlled when spinels are grown on mismatched cubic substrates. Varying the grain size has a strong effect on the magnetic structure of the film due to the appearance of spin glass or superparamagnetic states. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3600-4.jpg ferrite InFeCoO4 There are many well-known magnetic materials among spinel ferrites. The grain orientation and size can be controlled when spinels are grown on mismatched cubic substrates. Varying the grain size has a strong effect on the magnetic structure of the film due to the appearance of spin glass or superparamagnetic states. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3603-05.jpg Ruddlesden-Popper LSMRO Ruddlesden-Popper (RP) phases form natural, thermodynamically stable, and atomically sharp interfaces that are nearly lattice matched with perovskites. Growing very thin RP layers can thus be an interesting route to fabricating structurally and electronically sharp tunnel barriers and quantum wells in oxides. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3603-2.jpg Ruddlesden-Popper LSMRO Ruddlesden-Popper (RP) phases form natural, thermodynamically stable, and atomically sharp interfaces that are nearly lattice matched with perovskites. Growing very thin RP layers can thus be an interesting route to fabricating structurally and electronically sharp tunnel barriers and quantum wells in oxides. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3603-4.jpg Ruddlesden-Popper LSMRO Ruddlesden-Popper (RP) phases form natural, thermodynamically stable, and atomically sharp interfaces that are nearly lattice matched with perovskites. Growing very thin RP layers can thus be an interesting route to fabricating structurally and electronically sharp tunnel barriers and quantum wells in oxides. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3601-05.jpg BeO-ZnO BeO is the lightest stable oxide with a band gap of over 10 eV and excellent heat conductivity. BeO is isostructural with ZnO, and it is interesting to attempt to form alloy phases, despite the huge lattice mismatch. We use alternate deposition from BeO and ZnO targets to make the alloy films. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3601-2.jpg BeO-ZnO BeO is the lightest stable oxide with a band gap of over 10 eV and excellent heat conductivity. BeO is isostructural with ZnO, and it is interesting to attempt to form alloy phases, despite the huge lattice mismatch. We use alternate deposition from BeO and ZnO targets to make the alloy films. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3601-4.jpg BeO-ZnO BeO is the lightest stable oxide with a band gap of over 10 eV and excellent heat conductivity. BeO is isostructural with ZnO, and it is interesting to attempt to form alloy phases, despite the huge lattice mismatch. We use alternate deposition from BeO and ZnO targets to make the alloy films. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3602-05.jpg CaHfO3 (La,Sr)TiO3 SrTiO3 FET Carrier confinement at oxide interfaces may give rise interesting electronic states. Structurally sharp quantum wells grown by PLD can be used to study the transport behavior as a function of carrier density and distribution. In doped-channel FETs, the carrier density can be tuned by a combination of doping and field effect. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3602-2.jpg CaHfO3 (La,Sr)TiO3 SrTiO3 FET Carrier confinement at oxide interfaces may give rise interesting electronic states. Structurally sharp quantum wells grown by PLD can be used to study the transport behavior as a function of carrier density and distribution. In doped-channel FETs, the carrier density can be tuned by a combination of doping and field effect. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3602-4.jpg CaHfO3 (La,Sr)TiO3 SrTiO3 FET Carrier confinement at oxide interfaces may give rise interesting electronic states. Structurally sharp quantum wells grown by PLD can be used to study the transport behavior as a function of carrier density and distribution. In doped-channel FETs, the carrier density can be tuned by a combination of doping and field effect. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3638-05.jpg CaHfO3 (La,Sr)O SrTiO3 FET Carrier confinement at oxide interfaces may give rise interesting electronic states. Structurally sharp quantum wells grown by PLD can be used to study the transport behavior as a function of carrier density and distribution. In doped-channel FETs, the carrier density can be tuned by a combination of doping and field effect. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3638-2.jpg CaHfO3 (La,Sr)O SrTiO3 FET Carrier confinement at oxide interfaces may give rise interesting electronic states. Structurally sharp quantum wells grown by PLD can be used to study the transport behavior as a function of carrier density and distribution. In doped-channel FETs, the carrier density can be tuned by a combination of doping and field effect. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3638-4.jpg CaHfO3 (La,Sr)O SrTiO3 FET Carrier confinement at oxide interfaces may give rise interesting electronic states. Structurally sharp quantum wells grown by PLD can be used to study the transport behavior as a function of carrier density and distribution. In doped-channel FETs, the carrier density can be tuned by a combination of doping and field effect. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3613-05.jpg DyScO3 SrTiO3 FET One of the largest challenges in fabricating oxide FETs is to find a suitable gate insulator that can be grown epitaxially on the semiconductor channel, such as SrTiO3. One good candidate is DyScO3, but despite good lattice matching, there is still a strong low-temperature threshold bias shift, indicating that the interface is not clean. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3613-2.jpg DyScO3 SrTiO3 FET One of the largest challenges in fabricating oxide FETs is to find a suitable gate insulator that can be grown epitaxially on the semiconductor channel, such as SrTiO3. One good candidate is DyScO3, but despite good lattice matching, there is still a strong low-temperature threshold bias shift, indicating that the interface is not clean. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3613-4.jpg DyScO3 SrTiO3 FET One of the largest challenges in fabricating oxide FETs is to find a suitable gate insulator that can be grown epitaxially on the semiconductor channel, such as SrTiO3. One good candidate is DyScO3, but despite good lattice matching, there is still a strong low-temperature threshold bias shift, indicating that the interface is not clean. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3617-05.jpg DyScO3 SrTiO3 FET One of the largest challenges in fabricating oxide FETs is to find a suitable gate insulator that can be grown epitaxially on the semiconductor channel, such as SrTiO3. One good candidate is DyScO3, but despite good lattice matching, there is still a strong low-temperature threshold bias shift, indicating that the interface is not clean. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3617-2.jpg DyScO3 SrTiO3 FET One of the largest challenges in fabricating oxide FETs is to find a suitable gate insulator that can be grown epitaxially on the semiconductor channel, such as SrTiO3. One good candidate is DyScO3, but despite good lattice matching, there is still a strong low-temperature threshold bias shift, indicating that the interface is not clean. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3617-4.jpg DyScO3 SrTiO3 FET One of the largest challenges in fabricating oxide FETs is to find a suitable gate insulator that can be grown epitaxially on the semiconductor channel, such as SrTiO3. One good candidate is DyScO3, but despite good lattice matching, there is still a strong low-temperature threshold bias shift, indicating that the interface is not clean. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3631-05.jpg DyScO3 SrTiO3 FET One of the largest challenges in fabricating oxide FETs is to find a suitable gate insulator that can be grown epitaxially on the semiconductor channel, such as SrTiO3. One good candidate is DyScO3, but despite good lattice matching, there is still a strong low-temperature threshold bias shift, indicating that the interface is not clean. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3631-2.jpg DyScO3 SrTiO3 FET One of the largest challenges in fabricating oxide FETs is to find a suitable gate insulator that can be grown epitaxially on the semiconductor channel, such as SrTiO3. One good candidate is DyScO3, but despite good lattice matching, there is still a strong low-temperature threshold bias shift, indicating that the interface is not clean. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3631-4.jpg DyScO3 SrTiO3 FET One of the largest challenges in fabricating oxide FETs is to find a suitable gate insulator that can be grown epitaxially on the semiconductor channel, such as SrTiO3. One good candidate is DyScO3, but despite good lattice matching, there is still a strong low-temperature threshold bias shift, indicating that the interface is not clean. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3604-05.jpg DyScO3 breakdown The FET switching performance is strongly affected by trapped charge and the maximum gate field that can be applied to the device. DyScO3 films can sustain higher breakdown fields than CaHfO3 while also having cleaner interfaces when grown epitaxially on SrTiO3. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3604-2.jpg DyScO3 breakdown The FET switching performance is strongly affected by trapped charge and the maximum gate field that can be applied to the device. DyScO3 films can sustain higher breakdown fields than CaHfO3 while also having cleaner interfaces when grown epitaxially on SrTiO3. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3604-4.jpg DyScO3 breakdown The FET switching performance is strongly affected by trapped charge and the maximum gate field that can be applied to the device. DyScO3 films can sustain higher breakdown fields than CaHfO3 while also having cleaner interfaces when grown epitaxially on SrTiO3. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3609-05.jpg DyScO3 SrTiO3 FET The FET switching performance is strongly affected by trapped charge and the maximum gate field that can be applied to the device. DyScO3 films can sustain higher breakdown fields than CaHfO3 while also having cleaner interfaces when grown epitaxially on SrTiO3. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3609-2.jpg DyScO3 SrTiO3 FET The FET switching performance is strongly affected by trapped charge and the maximum gate field that can be applied to the device. DyScO3 films can sustain higher breakdown fields than CaHfO3 while also having cleaner interfaces when grown epitaxially on SrTiO3. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3609-4.jpg DyScO3 SrTiO3 FET The FET switching performance is strongly affected by trapped charge and the maximum gate field that can be applied to the device. DyScO3 films can sustain higher breakdown fields than CaHfO3 while also having cleaner interfaces when grown epitaxially on SrTiO3. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3618-05.jpg SrTiO3 single crystal FET Preparing good source and drain electrodes in epitaxial SrTiO3 FETs is complicated by the high growth temperatures needed for high-quality gate insulator growth. One of the most effective ways of fabricating electrodes is ion milling, which produces a conducting, oxygen-deficient surface layer that can survive high-temperature processing. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3618-2.jpg SrTiO3 single crystal FET Preparing good source and drain electrodes in epitaxial SrTiO3 FETs is complicated by the high growth temperatures needed for high-quality gate insulator growth. One of the most effective ways of fabricating electrodes is ion milling, which produces a conducting, oxygen-deficient surface layer that can survive high-temperature processing. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3618-4.jpg SrTiO3 single crystal FET Preparing good source and drain electrodes in epitaxial SrTiO3 FETs is complicated by the high growth temperatures needed for high-quality gate insulator growth. One of the most effective ways of fabricating electrodes is ion milling, which produces a conducting, oxygen-deficient surface layer that can survive high-temperature processing. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3619-05.jpg SrTiO3 FET process Preparing good source and drain electrodes in epitaxial SrTiO3 FETs is complicated by the high growth temperatures needed for high-quality gate insulator growth. One of the most effective ways of fabricating electrodes is ion milling, which produces a conducting, oxygen-deficient surface layer that can survive high-temperature processing. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3619-2.jpg SrTiO3 FET process Preparing good source and drain electrodes in epitaxial SrTiO3 FETs is complicated by the high growth temperatures needed for high-quality gate insulator growth. One of the most effective ways of fabricating electrodes is ion milling, which produces a conducting, oxygen-deficient surface layer that can survive high-temperature processing. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3619-4.jpg SrTiO3 FET process Preparing good source and drain electrodes in epitaxial SrTiO3 FETs is complicated by the high growth temperatures needed for high-quality gate insulator growth. One of the most effective ways of fabricating electrodes is ion milling, which produces a conducting, oxygen-deficient surface layer that can survive high-temperature processing. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3605-05.jpg SrTiO3 Argon milled electrodes Preparing good source and drain electrodes in epitaxial SrTiO3 FETs is complicated by the high growth temperatures needed for high-quality gate insulator growth. One of the most effective ways of fabricating electrodes is ion milling, which produces a conducting, oxygen-deficient surface layer that can survive high-temperature processing. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3605-2.jpg SrTiO3 Argon milled electrodes Preparing good source and drain electrodes in epitaxial SrTiO3 FETs is complicated by the high growth temperatures needed for high-quality gate insulator growth. One of the most effective ways of fabricating electrodes is ion milling, which produces a conducting, oxygen-deficient surface layer that can survive high-temperature processing. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3605-4.jpg SrTiO3 Argon milled electrodes Preparing good source and drain electrodes in epitaxial SrTiO3 FETs is complicated by the high growth temperatures needed for high-quality gate insulator growth. One of the most effective ways of fabricating electrodes is ion milling, which produces a conducting, oxygen-deficient surface layer that can survive high-temperature processing. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3615-05.jpg DyScO3 fixed charge The morphology of the insulator film has a large effect on FET performance because the typical device size is larger than the thin film grain size. Due to this, the gate insulator may hold significant fixed charge that is directly visible as a threshold shift in FETs. This problem can be mitigated by film growth optimizations and by device scaling. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3615-2.jpg DyScO3 fixed charge The morphology of the insulator film has a large effect on FET performance because the typical device size is larger than the thin film grain size. Due to this, the gate insulator may hold significant fixed charge that is directly visible as a threshold shift in FETs. This problem can be mitigated by film growth optimizations and by device scaling. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3615-4.jpg DyScO3 fixed charge The morphology of the insulator film has a large effect on FET performance because the typical device size is larger than the thin film grain size. Due to this, the gate insulator may hold significant fixed charge that is directly visible as a threshold shift in FETs. This problem can be mitigated by film growth optimizations and by device scaling. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3637-05.jpg DyScO3 fixed charge The morphology of the insulator film has a large effect on FET performance because the typical device size is larger than the thin film grain size. Due to this, the gate insulator may hold significant fixed charge that is directly visible as a threshold shift in FETs. This problem can be mitigated by film growth optimizations and by device scaling. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3637-2.jpg DyScO3 fixed charge The morphology of the insulator film has a large effect on FET performance because the typical device size is larger than the thin film grain size. Due to this, the gate insulator may hold significant fixed charge that is directly visible as a threshold shift in FETs. This problem can be mitigated by film growth optimizations and by device scaling. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3637-4.jpg DyScO3 fixed charge The morphology of the insulator film has a large effect on FET performance because the typical device size is larger than the thin film grain size. Due to this, the gate insulator may hold significant fixed charge that is directly visible as a threshold shift in FETs. This problem can be mitigated by film growth optimizations and by device scaling. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3630-05.jpg SrTiO3 FET scaling The morphology of the insulator film has a large effect on FET performance because the typical device size is larger than the thin film grain size. Due to this, the gate insulator may hold significant fixed charge that is directly visible as a threshold shift in FETs. This problem can be mitigated by film growth optimizations and by device scaling. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3630-2.jpg SrTiO3 FET scaling The morphology of the insulator film has a large effect on FET performance because the typical device size is larger than the thin film grain size. Due to this, the gate insulator may hold significant fixed charge that is directly visible as a threshold shift in FETs. This problem can be mitigated by film growth optimizations and by device scaling. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3630-4.jpg SrTiO3 FET scaling The morphology of the insulator film has a large effect on FET performance because the typical device size is larger than the thin film grain size. Due to this, the gate insulator may hold significant fixed charge that is directly visible as a threshold shift in FETs. This problem can be mitigated by film growth optimizations and by device scaling. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3606-05.jpg LaTiO3 nanodots in SrTiO3 Lateral fractional-layer structures, such as nanowire arrays, can be used to study the effects of inhomogeneity on the transport properties of delta-doped layers and to probe the effects of vertical interdiffusion. Metallic LaTiO3 nanowires grown along the step edges of a step-and-terrace SrTiO3 substrate is one particularly convenient model system for such studies. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3606-2.jpg LaTiO3 nanodots in SrTiO3 Lateral fractional-layer structures, such as nanowire arrays, can be used to study the effects of inhomogeneity on the transport properties of delta-doped layers and to probe the effects of vertical interdiffusion. Metallic LaTiO3 nanowires grown along the step edges of a step-and-terrace SrTiO3 substrate is one particularly convenient model system for such studies. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3606-4.jpg LaTiO3 nanodots in SrTiO3 Lateral fractional-layer structures, such as nanowire arrays, can be used to study the effects of inhomogeneity on the transport properties of delta-doped layers and to probe the effects of vertical interdiffusion. Metallic LaTiO3 nanowires grown along the step edges of a step-and-terrace SrTiO3 substrate is one particularly convenient model system for such studies. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3607-05.jpg LaTiO3 nanowires in SrTiO3 Lateral fractional-layer structures, such as nanowire arrays, can be used to study the effects of inhomogeneity on the transport properties of delta-doped layers and to probe the effects of vertical interdiffusion. Metallic LaTiO3 nanowires grown along the step edges of a step-and-terrace SrTiO3 substrate is one particularly convenient model system for such studies. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3607-2.jpg LaTiO3 nanowires in SrTiO3 Lateral fractional-layer structures, such as nanowire arrays, can be used to study the effects of inhomogeneity on the transport properties of delta-doped layers and to probe the effects of vertical interdiffusion. Metallic LaTiO3 nanowires grown along the step edges of a step-and-terrace SrTiO3 substrate is one particularly convenient model system for such studies. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3607-4.jpg LaTiO3 nanowires in SrTiO3 Lateral fractional-layer structures, such as nanowire arrays, can be used to study the effects of inhomogeneity on the transport properties of delta-doped layers and to probe the effects of vertical interdiffusion. Metallic LaTiO3 nanowires grown along the step edges of a step-and-terrace SrTiO3 substrate is one particularly convenient model system for such studies. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3608-05.jpg LaTiO3 nanowires in SrTiO3 Lateral fractional-layer structures, such as nanowire arrays, can be used to study the effects of inhomogeneity on the transport properties of delta-doped layers and to probe the effects of vertical interdiffusion. Metallic LaTiO3 nanowires grown along the step edges of a step-and-terrace SrTiO3 substrate is one particularly convenient model system for such studies. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3608-2.jpg LaTiO3 nanowires in SrTiO3 Lateral fractional-layer structures, such as nanowire arrays, can be used to study the effects of inhomogeneity on the transport properties of delta-doped layers and to probe the effects of vertical interdiffusion. Metallic LaTiO3 nanowires grown along the step edges of a step-and-terrace SrTiO3 substrate is one particularly convenient model system for such studies. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3608-4.jpg LaTiO3 nanowires in SrTiO3 Lateral fractional-layer structures, such as nanowire arrays, can be used to study the effects of inhomogeneity on the transport properties of delta-doped layers and to probe the effects of vertical interdiffusion. Metallic LaTiO3 nanowires grown along the step edges of a step-and-terrace SrTiO3 substrate is one particularly convenient model system for such studies. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3612-05.jpg LaTiO3 / SrTiO3 fractional-layers La doping transforms semiconducting SrTiO3 into a good metal even when the La atoms only form a single LaO delta-doping layer. Fractional lateral structures can be used to study at what fractional coverage a percolative conduction path forms in a two-dimensional layer. Such ultrathin heterostructures show a curious competition between 2-dimensional localization and 3-dimensional metallicity. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3612-2.jpg LaTiO3 / SrTiO3 fractional-layers La doping transforms semiconducting SrTiO3 into a good metal even when the La atoms only form a single LaO delta-doping layer. Fractional lateral structures can be used to study at what fractional coverage a percolative conduction path forms in a two-dimensional layer. Such ultrathin heterostructures show a curious competition between 2-dimensional localization and 3-dimensional metallicity. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3612-4.jpg LaTiO3 / SrTiO3 fractional-layers La doping transforms semiconducting SrTiO3 into a good metal even when the La atoms only form a single LaO delta-doping layer. Fractional lateral structures can be used to study at what fractional coverage a percolative conduction path forms in a two-dimensional layer. Such ultrathin heterostructures show a curious competition between 2-dimensional localization and 3-dimensional metallicity. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3614-05.jpg LaTiO3 single-layer magnetotransport La doping transforms semiconducting SrTiO3 into a good metal even when the La atoms only form a single LaO delta-doping layer. Fractional lateral structures can be used to study at what fractional coverage a percolative conduction path forms in a two-dimensional layer. Such ultrathin heterostructures show a curious competition between 2-dimensional localization and 3-dimensional metallicity. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3614-2.jpg LaTiO3 single-layer magnetotransport La doping transforms semiconducting SrTiO3 into a good metal even when the La atoms only form a single LaO delta-doping layer. Fractional lateral structures can be used to study at what fractional coverage a percolative conduction path forms in a two-dimensional layer. Such ultrathin heterostructures show a curious competition between 2-dimensional localization and 3-dimensional metallicity. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3614-4.jpg LaTiO3 single-layer magnetotransport La doping transforms semiconducting SrTiO3 into a good metal even when the La atoms only form a single LaO delta-doping layer. Fractional lateral structures can be used to study at what fractional coverage a percolative conduction path forms in a two-dimensional layer. Such ultrathin heterostructures show a curious competition between 2-dimensional localization and 3-dimensional metallicity. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3611-05.jpg SrTiO3 photoconductivity The operation of various electronic devices and the performance of functional materials is dependent on non-equilibrium carriers. The mobility of such carriers depends on the defect density in the semiconductor. Photoconductivity is one simple way of characterizing the presence and density of structural defects and impurities that affect non-equilibrium carrier transport in oxide semiconductors. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3611-2.jpg SrTiO3 photoconductivity The operation of various electronic devices and the performance of functional materials is dependent on non-equilibrium carriers. The mobility of such carriers depends on the defect density in the semiconductor. Photoconductivity is one simple way of characterizing the presence and density of structural defects and impurities that affect non-equilibrium carrier transport in oxide semiconductors. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3611-4.jpg SrTiO3 photoconductivity The operation of various electronic devices and the performance of functional materials is dependent on non-equilibrium carriers. The mobility of such carriers depends on the defect density in the semiconductor. Photoconductivity is one simple way of characterizing the presence and density of structural defects and impurities that affect non-equilibrium carrier transport in oxide semiconductors. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3636-05.jpg Rh:SrTiO3 films Rh:SrTiO3 is a p-type semiconductor that works as a hydrogen evolution catalyst in photoelectrochemical water splitting powered by sunlight. Rh:SrTiO3 thin films can be used to determine the role of the Rh4+/3+ dopant valence on the electronic structure and hydrogen evolution activity under visible light irradiation. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3636-2.jpg Rh:SrTiO3 films Rh:SrTiO3 is a p-type semiconductor that works as a hydrogen evolution catalyst in photoelectrochemical water splitting powered by sunlight. Rh:SrTiO3 thin films can be used to determine the role of the Rh4+/3+ dopant valence on the electronic structure and hydrogen evolution activity under visible light irradiation. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3636-4.jpg Rh:SrTiO3 films Rh:SrTiO3 is a p-type semiconductor that works as a hydrogen evolution catalyst in photoelectrochemical water splitting powered by sunlight. Rh:SrTiO3 thin films can be used to determine the role of the Rh4+/3+ dopant valence on the electronic structure and hydrogen evolution activity under visible light irradiation. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3641-05.jpg Rh:SrTiO3 electronic structure Rh:SrTiO3 is a p-type semiconductor that works as a hydrogen evolution catalyst in photoelectrochemical water splitting powered by sunlight. Rh:SrTiO3 thin films can be used to determine the role of the Rh4+/3+ dopant valence on the electronic structure and hydrogen evolution activity under visible light irradiation. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3641-2.jpg Rh:SrTiO3 electronic structure Rh:SrTiO3 is a p-type semiconductor that works as a hydrogen evolution catalyst in photoelectrochemical water splitting powered by sunlight. Rh:SrTiO3 thin films can be used to determine the role of the Rh4+/3+ dopant valence on the electronic structure and hydrogen evolution activity under visible light irradiation. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3641-4.jpg Rh:SrTiO3 electronic structure Rh:SrTiO3 is a p-type semiconductor that works as a hydrogen evolution catalyst in photoelectrochemical water splitting powered by sunlight. Rh:SrTiO3 thin films can be used to determine the role of the Rh4+/3+ dopant valence on the electronic structure and hydrogen evolution activity under visible light irradiation. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3628-05.jpg Rh,Ir:SrTiO3 photocatalyst Although the electronic structures of Rh- and Ir-doped SrTiO3 are very similar, Rh:SrTiO3 is a p-type hydrogen evolution photocatalyst, while Ir:SrTiO3 is an n-type oxygen evolution photocatalyst. A comparison of the electronic spectra can be used to draw a direct link between the electronic structure and the photoelectrochemical activity. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3628-2.jpg Rh,Ir:SrTiO3 photocatalyst Although the electronic structures of Rh- and Ir-doped SrTiO3 are very similar, Rh:SrTiO3 is a p-type hydrogen evolution photocatalyst, while Ir:SrTiO3 is an n-type oxygen evolution photocatalyst. A comparison of the electronic spectra can be used to draw a direct link between the electronic structure and the photoelectrochemical activity. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3628-4.jpg Rh,Ir:SrTiO3 photocatalyst Although the electronic structures of Rh- and Ir-doped SrTiO3 are very similar, Rh:SrTiO3 is a p-type hydrogen evolution photocatalyst, while Ir:SrTiO3 is an n-type oxygen evolution photocatalyst. A comparison of the electronic spectra can be used to draw a direct link between the electronic structure and the photoelectrochemical activity. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3632-05.jpg Rh,Ir:SrTiO3 photocatalyst Although the electronic structures of Rh- and Ir-doped SrTiO3 are very similar, Rh:SrTiO3 is a p-type hydrogen evolution photocatalyst, while Ir:SrTiO3 is an n-type oxygen evolution photocatalyst. A comparison of the electronic spectra can be used to draw a direct link between the electronic structure and the photoelectrochemical activity. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3632-2.jpg Rh,Ir:SrTiO3 photocatalyst Although the electronic structures of Rh- and Ir-doped SrTiO3 are very similar, Rh:SrTiO3 is a p-type hydrogen evolution photocatalyst, while Ir:SrTiO3 is an n-type oxygen evolution photocatalyst. A comparison of the electronic spectra can be used to draw a direct link between the electronic structure and the photoelectrochemical activity. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3632-4.jpg Rh,Ir:SrTiO3 photocatalyst Although the electronic structures of Rh- and Ir-doped SrTiO3 are very similar, Rh:SrTiO3 is a p-type hydrogen evolution photocatalyst, while Ir:SrTiO3 is an n-type oxygen evolution photocatalyst. A comparison of the electronic spectra can be used to draw a direct link between the electronic structure and the photoelectrochemical activity. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3639-05.jpg Rh:SrTiO3 XES and XAS The photoelectrochemical activity of oxide semiconductors is dependent on the locations of occupied and unoccupied in-gap states induced by doping. The location of the unoccupied states can be seen in X-ray absorption spectra, while the occupied states can be probed by x-ray emission spectroscopy. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3639-2.jpg Rh:SrTiO3 XES and XAS The photoelectrochemical activity of oxide semiconductors is dependent on the locations of occupied and unoccupied in-gap states induced by doping. The location of the unoccupied states can be seen in X-ray absorption spectra, while the occupied states can be probed by x-ray emission spectroscopy. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3639-4.jpg Rh:SrTiO3 XES and XAS The photoelectrochemical activity of oxide semiconductors is dependent on the locations of occupied and unoccupied in-gap states induced by doping. The location of the unoccupied states can be seen in X-ray absorption spectra, while the occupied states can be probed by x-ray emission spectroscopy. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3616-05.jpg VO2 film bending VO2 is well known for the metal-insulator transition that occurs slightly above room temperature. This is a charge ordering transition that also involves a structural change. The structural transition can be driven by dynamic mechanical strain by bending a crystal, inducing a change in conductivity. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3616-2.jpg VO2 film bending VO2 is well known for the metal-insulator transition that occurs slightly above room temperature. This is a charge ordering transition that also involves a structural change. The structural transition can be driven by dynamic mechanical strain by bending a crystal, inducing a change in conductivity. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3616-4.jpg VO2 film bending VO2 is well known for the metal-insulator transition that occurs slightly above room temperature. This is a charge ordering transition that also involves a structural change. The structural transition can be driven by dynamic mechanical strain by bending a crystal, inducing a change in conductivity. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3633-05.jpg VO2 magnetite magnetic field sensor VO2 is well known for the metal-insulator transition that occurs slightly above room temperature. This is a charge ordering transition that also involves a structural change. The structural transition can be driven by dynamic mechanical strain by bending a crystal, inducing a change in conductivity. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3633-2.jpg VO2 magnetite magnetic field sensor VO2 is well known for the metal-insulator transition that occurs slightly above room temperature. This is a charge ordering transition that also involves a structural change. The structural transition can be driven by dynamic mechanical strain by bending a crystal, inducing a change in conductivity. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3633-4.jpg VO2 magnetite magnetic field sensor VO2 is well known for the metal-insulator transition that occurs slightly above room temperature. This is a charge ordering transition that also involves a structural change. The structural transition can be driven by dynamic mechanical strain by bending a crystal, inducing a change in conductivity. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3644-05.jpg PbTiO3 ferroelectric poling Ferroelectric materials generally form complicated domain structures. An electrochemical techniques has been used to impose a high and uniform electric field on a PbTiO3 film, inducing a homogeneous domain structure over a wide area. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3644-2.jpg PbTiO3 ferroelectric poling Ferroelectric materials generally form complicated domain structures. An electrochemical techniques has been used to impose a high and uniform electric field on a PbTiO3 film, inducing a homogeneous domain structure over a wide area. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3644-4.jpg PbTiO3 ferroelectric poling Ferroelectric materials generally form complicated domain structures. An electrochemical techniques has been used to impose a high and uniform electric field on a PbTiO3 film, inducing a homogeneous domain structure over a wide area. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3647-05.jpg BaSnO3 superstructure Thin films occasionally exhibit spontaneous phase separation. A periodic composition variation has been observed in BaSnO3 films where the periodicity of the composition modulation depends on the growth rate. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3647-2.jpg BaSnO3 superstructure Thin films occasionally exhibit spontaneous phase separation. A periodic composition variation has been observed in BaSnO3 films where the periodicity of the composition modulation depends on the growth rate. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3647-4.jpg BaSnO3 superstructure Thin films occasionally exhibit spontaneous phase separation. A periodic composition variation has been observed in BaSnO3 films where the periodicity of the composition modulation depends on the growth rate. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3634-05.jpg Magnetite nanopyramids Magnetite Fe3O4 is a ferrimagnetic ferroelectric below the Verwey transition temperature. Epitaxial magnetite nanopyramids can be grown on SrTiO3 in a process where solid-phase dewetting is used to form (001)-oriented seed crystals. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3634-2.jpg Magnetite nanopyramids Magnetite Fe3O4 is a ferrimagnetic ferroelectric below the Verwey transition temperature. Epitaxial magnetite nanopyramids can be grown on SrTiO3 in a process where solid-phase dewetting is used to form (001)-oriented seed crystals. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3634-4.jpg Magnetite nanopyramids Magnetite Fe3O4 is a ferrimagnetic ferroelectric below the Verwey transition temperature. Epitaxial magnetite nanopyramids can be grown on SrTiO3 in a process where solid-phase dewetting is used to form (001)-oriented seed crystals. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3643-05.jpg Magnetite nanopyramids Magnetite Fe3O4 is a ferrimagnetic ferroelectric below the Verwey transition temperature. Epitaxial magnetite nanopyramids can be grown on SrTiO3 in a process where solid-phase dewetting is used to form (001)-oriented seed crystals. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3643-2.jpg Magnetite nanopyramids Magnetite Fe3O4 is a ferrimagnetic ferroelectric below the Verwey transition temperature. Epitaxial magnetite nanopyramids can be grown on SrTiO3 in a process where solid-phase dewetting is used to form (001)-oriented seed crystals. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3643-4.jpg Magnetite nanopyramids Magnetite Fe3O4 is a ferrimagnetic ferroelectric below the Verwey transition temperature. Epitaxial magnetite nanopyramids can be grown on SrTiO3 in a process where solid-phase dewetting is used to form (001)-oriented seed crystals. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3642-05.jpg Ferroelectric polarization in magnetite Magnetite Fe3O4 is a ferrimagnetic ferroelectric below the Verwey transition temperature. Epitaxial magnetite nanopyramids can be grown on SrTiO3 in a process where solid-phase dewetting is used to form (001)-oriented seed crystals. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3642-2.jpg Ferroelectric polarization in magnetite Magnetite Fe3O4 is a ferrimagnetic ferroelectric below the Verwey transition temperature. Epitaxial magnetite nanopyramids can be grown on SrTiO3 in a process where solid-phase dewetting is used to form (001)-oriented seed crystals. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3642-4.jpg Ferroelectric polarization in magnetite Magnetite Fe3O4 is a ferrimagnetic ferroelectric below the Verwey transition temperature. Epitaxial magnetite nanopyramids can be grown on SrTiO3 in a process where solid-phase dewetting is used to form (001)-oriented seed crystals. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3640-05.jpg Magnetic insulator manganites Magnetic insulators can be used to construct spin-filtering tunnel junctions. Pr0.8Ca0.2Mn1-yScyO3 is an insulator with a tunable magnetic field strength. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3640-2.jpg Magnetic insulator manganites Magnetic insulators can be used to construct spin-filtering tunnel junctions. Pr0.8Ca0.2Mn1-yScyO3 is an insulator with a tunable magnetic field strength. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3640-4.jpg Magnetic insulator manganites Magnetic insulators can be used to construct spin-filtering tunnel junctions. Pr0.8Ca0.2Mn1-yScyO3 is an insulator with a tunable magnetic field strength. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3645-05.jpg SrTiO3 crystal surface composition SrTiO3 is a very common substrate material for oxide thin film growth. In heterostructure studies, it is necessary to control the surface stoichiometry of the substrate. The Sr content on the substrate surface is measured by ion scattering spectroscopy and various treatment procedures can be used to optimize the substrate surface structure and composition. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3645-2.jpg SrTiO3 crystal surface composition SrTiO3 is a very common substrate material for oxide thin film growth. In heterostructure studies, it is necessary to control the surface stoichiometry of the substrate. The Sr content on the substrate surface is measured by ion scattering spectroscopy and various treatment procedures can be used to optimize the substrate surface structure and composition. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3645-4.jpg SrTiO3 crystal surface composition SrTiO3 is a very common substrate material for oxide thin film growth. In heterostructure studies, it is necessary to control the surface stoichiometry of the substrate. The Sr content on the substrate surface is measured by ion scattering spectroscopy and various treatment procedures can be used to optimize the substrate surface structure and composition. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3635-05.jpg Oxides introduction Every spring we have introduction seminars for new students. We propose research topics for students interested in doing their MSc or PhD degree in the area of oxide thin film, surface, or heterostructure studies. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3635-2.jpg Oxides introduction Every spring we have introduction seminars for new students. We propose research topics for students interested in doing their MSc or PhD degree in the area of oxide thin film, surface, or heterostructure studies. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3635-4.jpg Oxides introduction Every spring we have introduction seminars for new students. We propose research topics for students interested in doing their MSc or PhD degree in the area of oxide thin film, surface, or heterostructure studies. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3671-05.jpg Lab introduction Every spring we have introduction seminars for new students. We propose research topics for students interested in doing their MSc or PhD degree in the area of oxide thin film, surface, or heterostructure studies. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3671-2.jpg Lab introduction Every spring we have introduction seminars for new students. We propose research topics for students interested in doing their MSc or PhD degree in the area of oxide thin film, surface, or heterostructure studies. https://lippmaa.issp.u-tokyo.ac.jp/gallery/work/Images/DSCN3671-4.jpg Lab introduction Every spring we have introduction seminars for new students. We propose research topics for students interested in doing their MSc or PhD degree in the area of oxide thin film, surface, or heterostructure studies. https://lippmaa.issp.u-tokyo.ac.jp/guide https://lippmaa.issp.u-tokyo.ac.jp/guim/intro1.png Research topics Suggested research topics for new graduate students cover oxide thin film materials, surfaces, interfaces, and analysis techniques https://lippmaa.issp.u-tokyo.ac.jp/guim/intro1s.png Research topics Suggested research topics for new graduate students cover oxide thin film materials, surfaces, interfaces, and analysis techniques https://lippmaa.issp.u-tokyo.ac.jp/guim/intro-pld.png PLD plume View inside a PLD chamber showing an ablation plume https://lippmaa.issp.u-tokyo.ac.jp/guim/intro-perovskite.png Perovskite structure Perovskite unit cell and individual AO and BO2 layers https://lippmaa.issp.u-tokyo.ac.jp/guim/intro-mag.png Magnetic materials La2NiMnO6 crystal structure, polarization model, and x-ray diffraction pattern. Magnetite Fe3O4 nanopyramids formed by spontaneous solid-phase dewetting. https://lippmaa.issp.u-tokyo.ac.jp/guim/intro-cat.png Overview of photocatalyst development Principle of photocatalytic water splitting, nanostructures in Iridium-doped SrTiO3, x-ray emission spectroscopy of a Rhodium-doped SrTiO3 film and images of several Rhodium-doped thin films. https://lippmaa.issp.u-tokyo.ac.jp/guim/intro-ir.png Iridium nanopillar STEM image of an Iridium nanopillar embedded in SrTiO3 and a structural model https://lippmaa.issp.u-tokyo.ac.jp/guim/intro-nano.png Nanoscale transport Percolative transport in metallic nanodot arrays. Metallic conductivity appears at a critical nanodot density. STM images of LaTiO3 nanodots on a SrTiO3 surface. https://lippmaa.issp.u-tokyo.ac.jp/guim/intro-rp.png Ruddlesden-Popper interface Model and STEM image of a Ruddlesden-Popper type interface between two perovskite lattices. Gated magnetotransport of a LaTiO3 delta-doped layer in SrTiO3. https://lippmaa.issp.u-tokyo.ac.jp/guim/intro-surf.png FFM, RHEED, CAICISS Examples of surface analysis techniques; friction-force microscopy (FFM), reflection high-energy electron diffraction (RHEED), and coaxial impact-collision ion scattering spectroscopy (CAICISS) https://lippmaa.issp.u-tokyo.ac.jp/guim/PosterApr.png Laboratory introduction poster https://lippmaa.issp.u-tokyo.ac.jp/guim/Guide6.png Introduction cartoon Cartoon-style description of the methods and tools used in our laboratory. https://lippmaa.issp.u-tokyo.ac.jp/hardware/minipld/ https://lippmaa.issp.u-tokyo.ac.jp/hardware/minipld/CHtop.png Mini PLD chamber Top view of the chamber. https://lippmaa.issp.u-tokyo.ac.jp/hardware/minipld/CTfinger.png Mini PLD chamber Sample height and rotation manipulator. https://lippmaa.issp.u-tokyo.ac.jp/hardware/minipld/CTrheed.png Mini PLD chamber Heating laser optics. https://lippmaa.issp.u-tokyo.ac.jp/hardware/minipld/CTscreen.png Mini PLD chamber Chamber overpressure relief valves. https://lippmaa.issp.u-tokyo.ac.jp/hardware/minipld/CTview.png Mini PLD chamber High-pressure gauges. https://lippmaa.issp.u-tokyo.ac.jp/hardware/minipld/CHangle.png Mini PLD chamber Loading chamber and flow-control angle valve. https://lippmaa.issp.u-tokyo.ac.jp/hardware/minipld/CHback.png Mini PLD chamber Back of the loading chamber and the RHEED gun. https://lippmaa.issp.u-tokyo.ac.jp/hardware/minipld/CHbottom.png Mini PLD chamber Bottom of the chamber. https://lippmaa.issp.u-tokyo.ac.jp/hardware/minipld/CHgun.png Mini PLD chamber Back of the RHEED gun. https://lippmaa.issp.u-tokyo.ac.jp/hardware/minipld/CHlaser.png Mini PLD chamber Excimer laser entry viewport. https://lippmaa.issp.u-tokyo.ac.jp/hardware/minipld/CHprep.png Mini PLD chamber Gas lines of the loading chamber. https://lippmaa.issp.u-tokyo.ac.jp/hardware/minipld/CHpys.png Mini PLD chamber Photoemission chamber. https://lippmaa.issp.u-tokyo.ac.jp/hardware/minipld/CHrod.png Mini PLD chamber Excimer laser viewport and sample transfer rod. https://lippmaa.issp.u-tokyo.ac.jp/hardware/minipld/CHscreen.png Mini PLD chamber RHEED screen and cold finger. https://lippmaa.issp.u-tokyo.ac.jp/hardware/minipld/CHtrans.png Mini PLD chamber Loading chamber front view. https://lippmaa.issp.u-tokyo.ac.jp/hardware/minipld/CHturbo.png Mini PLD chamber Main chamber turbo pump. https://lippmaa.issp.u-tokyo.ac.jp/hardware/minipld/CHviewc.png Mini PLD chamber Front view of the chamber. https://lippmaa.issp.u-tokyo.ac.jp/hardware/minipld/CBgun.png Mini PLD chamber Target manipulator and RHEED differential pumping line. https://lippmaa.issp.u-tokyo.ac.jp/hardware/minipld/CBprep.png Mini PLD chamber Loading chamber transfer rod and gas lines. https://lippmaa.issp.u-tokyo.ac.jp/hardware/minipld/CBscreen.png Mini PLD chamber Target manipulator and pressure-control turbo. https://lippmaa.issp.u-tokyo.ac.jp/hardware/minipld/CBturbo.png Mini PLD chamber Target manipulator and flow-control turbo. https://lippmaa.issp.u-tokyo.ac.jp/hardware/ https://lippmaa.issp.u-tokyo.ac.jp/hardware/walkman/controls https://lippmaa.issp.u-tokyo.ac.jp/hardware/walkman/BNCS.jpg Analog signals Analog signal connections https://lippmaa.issp.u-tokyo.ac.jp/hardware/walkman/ControlboardS.jpg Digital signals Digital signals controller https://lippmaa.issp.u-tokyo.ac.jp/hardware/walkman/panel.png Control panel LabVIEW control panel of chamber functions. https://lippmaa.issp.u-tokyo.ac.jp/hardware/walkman/design https://lippmaa.issp.u-tokyo.ac.jp/hardware/walkman/section.png Chamber cross-section Cross section of the chamber https://lippmaa.issp.u-tokyo.ac.jp/hardware/walkman/chamber170101.jpg Outside view Outside view of the chamber. The RHEED gun is on the left. https://lippmaa.issp.u-tokyo.ac.jp/hardware/walkman/inside170101.jpg Inside view Inside view of the chamber, showing the sample stage. https://lippmaa.issp.u-tokyo.ac.jp/hardware/walkman/FlangeSideS.jpg Top flange The top flange of the chamber. https://lippmaa.issp.u-tokyo.ac.jp/hardware/walkman/Inside.jpg Target stage View of the chamber with the top flange removed. https://lippmaa.issp.u-tokyo.ac.jp/hardware/walkman/heating https://lippmaa.issp.u-tokyo.ac.jp/hardware/walkman/LDpower.png laser power calibration Diode laser output power vs. drive current. https://lippmaa.issp.u-tokyo.ac.jp/hardware/walkman/PDpower.png Photodiode voltage Photodiode output voltage vs. diode power. https://lippmaa.issp.u-tokyo.ac.jp/hardware/walkman/mounting.png sample mounting Sample mounting for high-temperature operation. https://lippmaa.issp.u-tokyo.ac.jp/hardware/walkman/heatings14.png Heating efficiency Measured substrate temperature during four heating runs at different heating laser beam sizes. https://lippmaa.issp.u-tokyo.ac.jp/hardware/walkman/Sample1.jpg uneven heating Focused laser spot, thin absorber https://lippmaa.issp.u-tokyo.ac.jp/hardware/walkman/W005.jpg STO at 950C SrTiO3 substrate being annealed at 950°C https://lippmaa.issp.u-tokyo.ac.jp/hardware/walkman/ld https://lippmaa.issp.u-tokyo.ac.jp/hardware/walkman/Diodelaser.jpg Diode laser The diode laser head with an attached fiber and the power supply. https://lippmaa.issp.u-tokyo.ac.jp/hardware/walkman/cooler.jpg Chiller The water cooler for the laser head. https://lippmaa.issp.u-tokyo.ac.jp/hardware/walkman/HeaterFront.jpg heater front panel Front panel of the heater unit. The optical fiber plugs into the connector on the left. https://lippmaa.issp.u-tokyo.ac.jp/hardware/walkman/LaserHead.jpg Laser head The laser head mounted in the heater unit in the instrument rack. https://lippmaa.issp.u-tokyo.ac.jp/hardware/walkman/HeatingLensS.jpg Focusing optics The top flange of the chamber https://lippmaa.issp.u-tokyo.ac.jp/hardware/walkman/ https://lippmaa.issp.u-tokyo.ac.jp/hardware/walkman/fronts.jpg Chamber view The chamber soon after delivery with most of the larger vacuum components attached. https://lippmaa.issp.u-tokyo.ac.jp/hardware/walkman/yag https://lippmaa.issp.u-tokyo.ac.jp/hardware/walkman/YAG.jpg YAG laser Quantel Brilliant YAG lasers with higher-harmonic generator units. https://lippmaa.issp.u-tokyo.ac.jp/hardware/walkman/ChamberMar2.jpg Laser mounting The YAG laser mounted at the bottom of the chamber. The black box in front of the chamber is the RHEED camera housing https://lippmaa.issp.u-tokyo.ac.jp/hardware/walkman/Target1c.jpg Target surface Target surface after 800 laser pulses. A deep track is clearly visible. https://lippmaa.issp.u-tokyo.ac.jp/ https://lippmaa.issp.u-tokyo.ac.jp/nav0.svg Home navigation https://lippmaa.issp.u-tokyo.ac.jp/lecture/advmat https://lippmaa.issp.u-tokyo.ac.jp/lecture/macros https://lippmaa.issp.u-tokyo.ac.jp/lecture/isback.svg isback function Animated trace of the isback function https://lippmaa.issp.u-tokyo.ac.jp/lecture/back2.svg back function Animated trace of the back function https://lippmaa.issp.u-tokyo.ac.jp/lecture/notes https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys01-01.svg Seminar 01 notes: Seeing the world through sensors: objective measurements and subjective understanding Institute for Solid State Physics Mikk Lippmaa Wednesdays, 10:25 - 12:10 KOMCEE K302 25.04.2018 https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys01-02.svg Seminar 01 notes: Seminar location KOMCEE K302 https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys01-03.svg Seminar 01 notes: Seminar dates https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys01-04.svg Seminar 01 notes: Seminar format We have 16 people: divided into 5 or 6 groups of 2 or 3 people each. We learn by experiment. Each group will get tasks to solve in group work. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys01-05.svg Seminar 01 notes: Get to know each other Everybody here is interested in natural sciences. Are you interested in using phones/tablets in experiments? Why are measurements important for science? https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys01-06.svg Seminar 01 notes: Software for this seminar series Data processing: octave-online.net Data exchange and communication: fys2018.slack.com Presentations: Google Slides https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys01-07.svg Seminar 01 notes: Chromebook Don't close the lid of your Chromebook while working! These Chromebooks are in kiosk mode, closing the lid will log you out. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys01-08.svg Seminar 01 notes: Using slack The address is fys2018.slack.com. Enter user name and password given on your name tag. Enter sends, shift-Enter makes a new line. Use #general channel. Send DM (Direct Message) to me or other students. You must send comments or questions to me after each seminar! https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys01-09.svg Seminar 01 notes: The scientific method What is the scientific method? What is special about scientific way of thinking? How scientific thinking differs from everyday thought? Knowledge that has a high probability of being true. How can we evaluate the reliability of knowledge? Consider observation, prior knowledge, Analysis Model, Theory. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys01-10.svg Seminar 01 notes: A simple example How long is this piece of plastic? Take one of the rulers and measure the length. Write the answer on paper. Plot the histogram on the whiteboard. It should be easy to determine the true value. We can find the most probable value and estimate the likely error margin. You can see that even a very simple measurement done by 19 people is not very reliable (the error margin is quite large) Q: How would you answer a question like Is this piece of plastic longer than 130 mm, yes or no? https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys01-11.svg Seminar 01 notes: Prior knowledge How can we evaluate the reliability of knowledge? When direct observation is not possible, we may use other sources (authorities): teachers, books, journals. How to evaluate the reliability of an authority? Who would you trust (who is most likely to be correct)? Q: How to evaluate the reliability of a source (like a book or a Web search result)? https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys01-12.svg Seminar 01 notes: Scientific thinking Three components of scientific thinking: Comparison with empirical (natural) evidence, Logical reasoning and Critical thinking. Question: Is coffee good for you? search Google... Use the source! The actual conclusions from the study Always seek the original source, verify that the data and were obtained in a reliable way and that the conclusions are appropriate. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys01-13.svg Seminar 01 notes: Reliable sources How can we evaluate the reliability of an information source? Very high on the authority scale Publications in many top-level journals Still, the results have to be evaluated on their own merits, not by simply relying on a well-known and respected source. The papers shown above contained fake and fabricated data In this one instance, 28 papers withdrawn Nature, Science, Physical Review B, Physical Review Letters, Advanced Materials, Applied Physics Letters) How can the scientific process protect against fraud? Another recent example: fabricated peer review (April 2017) https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys01-14.svg Seminar 01 notes: Scientific thinking (again) Three components of scientific thinking: Comparison with empirical (natural) evidence. Logical reasoning. Critical thinking. Empirircal data: Compare published results with independently acquired empirical data. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys01-15.svg Seminar 01 notes: Empirical evidence Empirical: based on direct observation, not theory or pure thought Empirical evidence: - can be experienced by others too - is repeatable (usually) Other forms of evidence (not useful for science): - hearsay (unknown source) - testimony (unreliable) - revelation (unrepeatable, not observable by others) https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys01-16.svg Seminar 01 notes: Critical thinking on the Web How does a prism split white light? (Known since at least Newton's experiments 300 years ago) On the Web, we have many sources and many publishers. Google image search for the phrase 'prism spectrum': Many different images. Obviously, many are wrong. But which one is correct? read the web pages: Pink Floyd (the rock band) album cover Pintrest (LGBT spectrum) Based on the description, which one do you think is correct? take a prism and a flashlight and try, which is correct. Answer: Pink Floyd got it right, all others are incorrect! Most web pages are worth nothing (in a scientific sense) https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys01-17.svg Seminar 01 notes: How to get empirical evidence? A:Direct observation through our senses. But not everything can be directly sensed by humans. We use our hearing, vision, smell, taste, touch, etc. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys01-18.svg Seminar 01 notes: Sensors This brings us to sensors: When our direct senses are not sufficient, we use technical help -- various sensors that make processes in nature accessible to our senses. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys01-19.svg Seminar 01 notes: Objective or subjective Objective - relating to an object or condition - Independent of the observer, - Observable by any observer, - Requires an agreed scale. Examples of obective measurements: What about length, weight, beauty, or human character? Which measurements can be objective? Q: Why is objective observation necessary for the scientific method? A: Required for replication, which lets us improve the probability that the measurement result is 'correct'. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys01-20.svg Seminar 01 notes: Scientific language Distinguish between fact and opinion: Statement: 'I have a headache' (subjective or objective?) 'My leg is swollen' (subjective or objective?) Subjective observation: Opinion, belief, suspicion, assumption, etc. Variable (in time and between persons) Language: We feel..., We think..., We believe..., ... - Law: courts try to avoid subjective evidence - Medicine: consider subjective evaluation from patient but use objective observation for diagnosis. Besides science, where is the objective/subjective distinction important? Objective observation: The resistance was 20 at a temperature of 4 K https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys02-01.svg Seminar 02 notes: Seeing the world through sensors: objective measurements and subjective understanding Institute for Solid State Physics Mikk Lippmaa Wednesdays, 10:25 - 12:10 KOMCEE K302 02.05.2018 https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys02-02.svg Seminar 02 notes: Nexus 9 sensors NFC antenna Camera Magnetometer Accelerometer Gyroscope Light sensor Pressure sensor Near-field communications Microphone GPS/GNSS https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys02-03.svg Seminar 02 notes: iPhone 6 sensors Gyroscope Accelerometer Bosch Sensortech BMA280 Invensense MPU-6700 Barometer Bosch Sensortech BMP280 https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys02-04.svg Seminar 02 notes: Smartphones and tablets Android: Sensor Kinetics Sensor Kinetics iOS: iTunes Sensor Kinetics. The Sensor Kinetics application shows which sensors are available in your device. -You can view numeric values -You can plot data -You can save data (Android, pro version) List up the sensors that are available in your own phone: https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys02-05.svg Seminar 02 notes: Collecting and processing data Measure and record (Sensor Kinetics) Processing and plotting (Octave) Slack Use your group channel https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys02-06.svg Seminar 02 notes: Measurements with Sensor Kinetics Record data from multiple sensors at the same time Select the sampling rate (determines bandwidth) We care about the primary sensors: - accelerometer - gyroscope - magnetometer https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys02-07.svg Seminar 02 notes: Single sensor recording Data from one sensor can be plotted in one panel Most measurements produce vector data: x, y, z components are shown in red, green, blue Units depend on sensor Start / Stop buttons to control data acquisition. Clear to erase data when stopped. Data can be saved when acquisition is stopped. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys02-08.svg Seminar 02 notes: Saving data from a single sensor Saved data files for a single sensor. Other sensors have separate data panels. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys02-09.svg Seminar 02 notes: Saving data from multiple sensors Select suitable sampling rates for each sensor and click Multi Sensor Recorder. Select sensors to record. Stop when data collection has ended. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys02-10.svg Seminar 02 notes: Sensor data types We have at least three sensors: - gyroscope - accelerometer - magnetic field. Scalar values (length, volume, time, etc.) Vector values (velocity, magnetic field, acceleration) give three components. We can measure independently the x, y, and z components. The data saved by Sensor Kinetics is in a binary format that we cannot conveniently read, but when we share the data, it is converted to CSV (comma-separated value) format. CSV files can be read by many spreadsheet, plotting, and data manipulation programs. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys02-11.svg Seminar 02 notes: Data upload to Slack Select a file from the list. Pick Share from the menu Select E-mail (we actually save to Slack) Select Slack. Check that you have your group Channel Send to Slack. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys02-12.svg Seminar 02 notes: Move data from Slack to Octave Open octave-online.net in Chrome. Close the welcome panel, open the hamburger menu Use the username and password written on your namecards. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys02-13.svg Seminar 02 notes: Octave-online user interface Type commands at the Command prompt. Octave is a scientific programming language and data processing tool. You give commands, Octave calculates and provides the result. Try to give Octave a command demo surf 1 at the command prompt and press Enter. This is a demonstration of a three-dimensional surface plot https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys02-14.svg Seminar 02 notes: How we use Octave Measure, Copy data to Slack, Process data, Copy images to slack, use Google Slides for presentations. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys02-15.svg Seminar 02 notes: Plot a csv file in Octave Find your file in slack, click the Download button and transfer a .csv file from Slack to Octave. You will see a notification when the file has been downloaded from Slack to your Chromebook. The data file is now in Downloads folder of your Chromebook and we need to upload it to Octave-online. The file contains data for x, y, and z direction acceleration, so we get three plots. The data file is now in Octave-online and we can load it with the csv function https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys02-16.svg Seminar 02 notes: Copy screenshots from Octave to Slack Click Copy to clipboard, take a screenshot on Chromebook of a part of the display. Press Ctrl-V to paste image or press Alt-space and select paste. It can be simply pasted into Slack https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys02-17.svg Seminar 02 notes: Copy image data from Octave to Slack Images can be downloaded from Octave and uploaded to Slack: Click the Plot Window button A plot window opens, click png to download the image The image is now in the Downloads folder of the Chromebook. Click the Plot Window button again to close the plot window. Set a title, group, and add a comment: upload the image to Slack https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys02-18.svg Seminar 02 notes: Upload Octave macros or functions to Slack Octave calculations should be documented. Macro files should be uploaded to slack: In Slack, press the menu Select the macro text or press Ctrl-a to select all text and copy with Ctrl-c. In octave, select a macro or text snippet Paste the text with Ctrl-v. Add name, group, comment The macro file is now in Slack. Other group members can download and use the same macro or function for their calculations. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys02-19.svg Seminar 02 notes: Preparing presentation slides We can use Google Slides in Slack to prepare presentations Add a title (with group) select group add a comment Select your group channel, select Google Docs file and Presentation Google slides will open in a new window or tab. NB! Chrome may ask permission to NB! open a pop-up window. Please allow the pop-up. You can copy and paste images from Slack to Google Slides. You can also use download/upload to transfer images from Slack to Slides without copy/paste. All changes are automatically saved. Several people can edit the slides at the same time. Select file Your presentation file is now in Slack and other group members can open it directly from Slack. Find your presentation file in the Google Drive window https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys03-01.svg Seminar 03 notes: Seeing the world through sensors: objective measurements and subjective understanding Institute for Solid State Physics Mikk Lippmaa Wednesdays, 10:25 - 12:10 KOMCEE K302 09.05.2018 3 https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys03-02.svg Seminar 03 notes: First measurement Purpose: measure the distance of movement (on the order of 1 m) Available sensor: accelerometer. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys03-03.svg Seminar 03 notes: Position, velocity, acceleration Position, velocity, and acceleration are linked: velocity is given by the time derivative of position and acceleration is the derivative of velocity. Calculate the position x(t) by integrating the velocity data v(t). We can easily check the result, because we can measure the distance with a ruler. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys03-04.svg Seminar 03 notes: Sensor Kinetics setup Open the menu, Select Preferences, Select fastest options for these settings, Select the longest chart time span https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys03-05.svg Seminar 03 notes: Measure acceleration We have 6 groups. Pick your measurement path and record the acceleration. Record background for a few seconds at the beginning and the end. For patterns 1 to 3, stop briefly at each position shown. For pattern 4, stop between the vertical and horizontal loops. Perform the circle in pattern 5 in one continuous move. For pattern 6, stop once after the vertical move. Note: keep the tablet parallel to the original position. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys03-06.svg Seminar 03 notes: Measured examples https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys03-07.svg Seminar 03 notes: Data analysis 1. Copy the path image for your group from the seminar notes and upload to your group Slack channel. 2. Measure the acceleration. 3. Try to estimate if the data is good enough for analysis. 4. Save your data, transfer to Slack. 5. Move data from Slack to Octave. 6. Try to view the data with the csv function. csv 7. Upload your data plot to Slack. You should now have your data in Octave: https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys03-08.svg Seminar 03 notes: Data types Scalar (just single numbers): Vector (a collection of numbers). All new variables are listed at the left edge of the screen. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys03-09.svg Seminar 03 notes: Loading and viewing data Prepare vectors for each axis data and for time: It is convenient to use separate vectors for each axis. Start by loading acceleration data into matrix A. The data can now be plotted or processed, for example plots the x-axis acceleration data vs time as shown. Documenting your work can be done by copying the plots to Slack. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys03-10.svg Seminar 03 notes: How to work with a subset of the measured data? We can select the range from the plo. We already know how to pick sections of a vector or a matrix. To see a range of lines from the matrix, we can select by line numbers. We can plot all data for one axis with octave. How can we find the correct line numbers if we want to plot data from 4 to 6 seconds, for example? The answer is to search the time vector for the appropriate line numbers. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys03-11.svg Seminar 03 notes: Finding a time point index Let's see how the point function works. Works for the full data matrix too (if time values are in the first column). The point function has thus returned the correct index. We can see why this function is useful by looking at a larger data file, similar to the acceleration data that we measured with the Nexus tablets. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys03-12.svg Seminar 03 notes: Creating a new function The file can now be edited in the file window panel. Delete the text on the first line. Enter the function. This function calculates the average value of x-axis acceleration for the data matrix A between time points t1 and t2. Averaging can be an effective way of reducing noise. An example of how this function can be used: The data is rather noisy, but we can estimate the signal offset by calculating an average. Looking at the function text line by line: Shows that this is a new function declaration. The point function is used to find the row number point that corresponds to the value of t1 in matrix A. You can try to make a new function avey for calculating averages of y-axis avey y acceleration data. Remember to store the function in file avey.m https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys03-13.svg Seminar 03 notes: Plotting data We can use the point function to plot only certain parts of data. Two plot functions (plotx and ploty) are provided for quickly viewing the plotx ploty x- or y-axis acceleration data. The Octave hold on command can be used to prevent plot clearing. The plot can be released with hold off. This is not a very good plot because: - lines too thin - no axis labels - no plot title - no plot legends A deco macro is provided to decorate a plot and make it easier to read. Plots can also be combined by adding several data pairs to the plot command plot This would produce the same plot as shown above. Various commands can be included in the deco function to improve the appearance of deco a plot Details on commands that can control the appearance of plots can be found in the online documentation for Octave. The decoration function can be edited to fit different plots. You can make new functions as you need them. Plot color and line or marker types can be set when a plot is generated The default color is light blue. A plot with red circles can be made with This draws a dashed line without markers and sets the line width to 3 This command changes the font to Times and the sets the font size to 21, which makes values on the axes easier to see. Axis labels and a plot title can be added with the following commands The plot frame can be made thicker https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys03-14.svg Seminar 03 notes: Vector plots Line plots cannot be used to represent more complex data. Vector plots are often used to illustrate the properties of vector fields, where each data point is given by a location, a direction, and a magnitude. One example is a weather map that shows the variations of the direction and strength of wind. A vector is defined by a magnitude and a direction. For two-dimensional space, it is often convenient to use x, y components. To plot vectors, we need to know the measurement point measurement point and the vector components vector components Using a vector plot, start by defining the measurement points 2-dimensional vector plot. The plot looks nicer if the axes are set to a 1:1 x/y aspect ratio https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys04-01.svg Seminar 04 notes: Seeing the world through sensors: objective measurements and subjective understanding Institute for Solid State Physics Mikk Lippmaa Wednesdays, 10:25 - 12:10 KOMCEE K302 16.05.2018 4 https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys04-02.svg Seminar 04 notes: Analyzing acceleration data Our purpose is to integrate acceleration to get velocity and then integrate the velocity to get the position. Assume that we start at zero velocity and zero distance. Start with the acceleration graph shown below and try to sketch the shape of the velocity and position graphs. Draw graphs for velocity and position. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys04-03.svg Seminar 04 notes: Calculating velocity We can use a simple numeric integration function to calculate the velocity the same way as we did by hand. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys04-04.svg Seminar 04 notes: Velocity example Calculate and plot position. We can now try the same example in Octave https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys04-05.svg Seminar 04 notes: Integrating measured data Plot of the measured example data. We can see the acceleration data and identify sections with up-down motion where the x-direction acceleration only shows noise and sections where a large acceleration signal is visible and with opposite sign for the right and the left moves. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys04-06.svg Seminar 04 notes: Velocity calculation Integrate the acceleration data. Velocity seems to be changing when the tablet was stopped! https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys04-07.svg Seminar 04 notes: Sensor data problems Integrate the acceleration data. Note that the acceleration is not actually zero when the tablet isn't moving! This offset is caused by a sensor offset. Can we perhaps subtract a constant? What is the cause of acceleration offsets? https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys04-08.svg Seminar 04 notes: Accelerometer offsets The accelerometer background shifts depending on the direction (and value) of previous acceleration. The accelerometer shows hysteresis in addition to offset. To handle the offset and hysteresis, we need to know how the sensor works. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys04-09.svg Seminar 04 notes: Finding technical information Two major search engines: Go to google.co.jp May change automatically to google.co.jp Use www.google.com for the U.S. version of Google https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys04-10.svg Seminar 04 notes: Search results Search results depend on language. English keyword search: device information. Japanese keyword search: mostly blogs. Google and Bing behave differently. For an English keyword search: Google and Bing are similar. For a Japanese keyword search: Bing results are mostly irrelevant or useless https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys04-11.svg Seminar 04 notes: Nexus 7 accelerometer Google suggests ifixit teardown page for Nexus 7. We can find a nice review of what is inside a Nexus 7 tablet. Invensense MPU-6050 gyroscope and accelerometer. A picture of the circuit board in the tablet. We can find several useful words: MEMS Six-Axis 3-axis gyroscope 3-axis accelerometer. There are several versions with different measurement ranges. Which version is in the Nexus 7 tablet? Detailed information is in the datasheet. At the bottom of the page we start to find technical details Where are these sensors used? What is inside the accelerometer sensor? https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys04-12.svg Seminar 04 notes: Search tips 1:images Often the fastest way to find good links is to look at images in the Google search. We can quickly find what is inside the MPU-6050 device https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys04-13.svg Seminar 04 notes: Search tips 2:patents Invensense product page mentioned patented technology, so look at related patents. We already have pictures of what is inside the sensor, so it is easy to find suitable patents by looking at the patent drawings. Navigate to patents.google.com. Patents usually have a description how the device woks. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys04-14.svg Seminar 04 notes: Search tips 3:scholar Academic literature can be found in Google Scholar. Navigate to scholar.google.com. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys04-15.svg Seminar 04 notes: Search tips 4:keywords Google searches can be made more specific to find exactly what you want with the help of special keywords. Navigate to google.com and try the following searches https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys04-16.svg Seminar 04 notes: MPU-6050 accelerometer What we have found so far: 1. The sensor model is MPU-6050 2. It is made by Invensense 4. Device contains a gyroscope and an accelerometer 5. It is a 6-axis device 6. It is a MEMS device 7. Device is on a silicon die 8. Range is +/- 2g 9. Sensitivity is 16384 levels / g 10. Used in phones, tablets, vehicles, cameras 11. Patented Nasiri fabrication process 12. MEMS wafer and CMOS electronics https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys05-01.svg Seminar 05 notes: Seeing the world through sensors: objective measurements and subjective understanding Institute for Solid State Physics Mikk Lippmaa Wednesdays, 10:25 - 12:10 KOMCEE K302 23.05.2018 5 https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys05-02.svg Seminar 05 notes: Accelerometer working principle Test mass spring spring frame acceleration Test mass (no acceleration). Measure Calculate force Calculate acceleration The test mass in MPU-6050 is a spring-mounted movable silicon sheet. It is therefore known as a MEMS (micro-electro-mechanical system) device. The shape of the MEMS layer can be seen in the microscope images x-axis gyroscope y-axis gyroscope z-axis gyroscope z-axis acc. x-axis acc. y-axis acc. xyz gyroscope z-axis acc. x-axis acc. y-axis acc. MPU-6050 MPU-6500 https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys05-03.svg Seminar 05 notes: ADXL-335 accelerometer A simpler accelerometer: ADXL-335, manufactured by Analog Devices. Device information can be found at www.analog.com. Devices mounted on a covenient circuit board are available from akizukidenshi.com. The datasheet of the device from Akizuki. Top layer support spring that supports the moving MEMS test mass. Anchor to bottom layer. Stationary fingers of the position sensing capacitor. Opening the accelermeter chip, we find the MEMS test mass and under that the structure of the support frame, shown in the photos below. You can view the same chip under a microscope during the seminar. Movement sensing capacitor fingers Spring Broken-off anchor leg 0.7 mm. A closer view of the moving top layer bottom die (fixed) anchor (between top and bottom dies) spring fixed finger (bottom) moving finger (top) test mass (top die) movement Position measurement capacitor : Capacitor fingers (fixed lower die) Capacitor fingers (moving upper die). The capacitance cannot be measured directly. The capacitance signal is usually converted to a voltage or time signal for digital measurement. The purpose of our original experiment was to determine the movement path of a tablet. The measurement is actually a long sequence of operations: Calculate acceleration Measure time Calculate sensor shift Capacitor finger sizes, distances Calculate capacitance Support spring constant Calculate force MEMS layer mass Calculate velocity Calculate position. The springs are made by cutting a meandering pattern in silicon, forming a thin but flexible support. However, these springs are not perfect and the root cause of the sensor offset and hysteresis that we saw in the measurement. After rapid movement, the MEMS sheet doesn't quite return to the original position, leading to an offset signal in the sensor output. The magnitude and direction of the offset depends on the direction and acceleration of a previous move. We can correct for this by simply subtracting the offset, but that is why we have to stop for a moment after each move segment, to determine the new offset value. The position of the moving MEMS layer is measured with the help of interdigitated capacitors consisting of a layer of moving fingers that are interleaved with fixed fingers attached to the bottom die layer. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys05-04.svg Seminar 05 notes: Accelerometer summary 1. Acceleration applies a force on the moving plate (F = ma). Force generated by acceleration and force due to gravity cannot be distinguished. That is why the accelerometer also senses gravity. Force causes the moving platelet to shift. The force is balanced by springs. Distance of platelet movement is given by Hooke law. Movement is detected with an interdigitated capacitor sensor, where capacitance is proportional to distance. Capacitance is converted to digital numbers. A CMOS computer in the accelerometer generates a digital (or analog) signal that is proportional to the measured acceleration. 5. The output signal is recorded by the Nexus 7 tablet. The tablet plots and saves the data. 6. The offset change is caused by nonideal springs in the accelerometer sensor. We have to design the experiment so that the offsets can be determined and subtracted (use short moves and stops in between). 1 https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys05-05.svg Seminar 05 notes: Optimal measurement We can now consider the optimal measurement conditions 1. Signal must be within measurement range (2g) 2. Signal must be above noise level 3. Signal should be much larger than the sensor offsets 3. Measurement should be faster than drift 4. Measurement should be much slower than the sampling rate (or the bandwidth of the sensor) (over range) Too slow: noisy (signal very noisy) Too slow sampling (not enough data points for analysis) Good data https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys05-06.svg Seminar 05 notes: Correcting acceleration offsets Before we start correcting the acceleration signal background, consider the following questions: 1. Can we change the measurement procedure to eliminate or minimize the offset problem? 2. Do we need to modify the data or is there an algorithmic way of eliminating the offsets? 3. If we change the acceleration data, can we still trust the calculated result? But, some points in this range are not part of the background. Background points are close to zero, because the accelerometer offsets are always small (between 0.2 and 0.6 in this dataset). We always stopped for about a second between moves. This means that there are always many background points in a row. We saw that the acceleration data cannot be correctly integrated due to offsets. We thus need to find way to calculate the offset profile, subtract it from the original data and then integrate the acceleration data to get the velocity. How can we automatically determine the background from the measured data? Let's look more closely at a typical measurement result know from measurement that in these regions the tablet was not moving, so these parts correspond to background. Looking at the data, which characteristics of the signal can we use to distinguish the background parts from the movement signal? Zoom some more: We can thus consider a simple algorithm: Look at each data point: If the value is higher than a maximum limit (like 0.6), it is a signal point If the value is lower than the minimum limit (like 0.2), is is a signal point Look at a few neighbors on either side of the point: If any of the neighbors are outside of the limits, it is a signal point If all neighbors are within limits, it is a background point. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys05-07.svg Seminar 05 notes: Finding the background The isback function correctly recognizes the background parts of the data. Test that the function works correctly. We can now write an Octave function that looks at one data point and gives us a 0 or 1 answer. If it is a background point, return 1. If it isn't a background point, return 0. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys05-08.svg Seminar 05 notes: Extracting the background Loop over all data points We can now pick out from the original data all those points that are identified as background points by the isback function. We can see that the background shape has been correctly extracted. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys05-09.svg Seminar 05 notes: Using the background data How to calculate the background values for the data points where we have no background points? We use interpolation. We now have the background extracted, but if we look more closely, we see that there are gaps in the data. If we wish to subtract the background from the acceleration data, we need to fill those gaps. We can thus interpolate the necessary data values to subtract the sensor offset from the acceleration data. The same method can be used to estimate the accelerometer background when the tablet is moving. This function subtracts the calculated background shape from the original data. The interpolation and subtraction is carried out for each element of the A vector. A The fixback function returns the corrected data vector. The background has now been subtracted and is therefore zero. This data can be integrated to obtain the velocity of the tablet during the moves. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys05-10.svg Seminar 05 notes: Integrating acceleration This produces incorrect velocity values due to the accelerometer offsets. Calculate acceleration data background, check that background is reasonable. The correct background points are found for all sections where the tablet was stopped. No unwanted points are detected within the move sections. Perform acceleration data correction. The corrected acceleration should be zero in the stop sections. Integrate the corrected data: Result is a much more reasonable velocity plot. However, we see that the calculated velocity is not zero in the stop sections The velocity background is not correct. Some of the last move at around 8 seconds is incorrectly identified as background. We need to make the search width larger and try again. Calculate the background for the velocity data Recalculate with larger number of neighbor points (5 instead of 2) for picking background data points Calculate the corrected velocity data. All tools are now available to process the measured acceleration data. The corrected velocity data (blue) is now zero in sections where the tablet was stopped. Integrate the corrected velocity data to get the tablet position This is a reasonable x-axis position plot x for the first movement pattern 1 Repeat the same process for the y-axis data. We have succeeded in calculating to movement path of the tablet. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys05-11.svg Seminar 05 notes: Making improvements We discussed in the seminar the question of relaibility and validity of scientific results. We can now try to evaluate the reliability of our calculation result. Use critical thinking! The calculated path looks just like the original task, the horizontal and vertical lines look perfectly straight, the angles are exactly 90 degrees. 1 But, the measurement was done by sliding a tablet on the table by hand. There were no guides, so can the movement have really been so perfect? Maybe we should look again at our calculations. The calculated velocity of the tablet showed background shifts. This is caused by imperfect integration, so the calculated velocity at the end of a move is not exactly zero. This way we could properly find background parts of the data. We can now notice that identified some parts of the signal as background. This consideration gives us an important lesson. Even though we were not trying to make our results prettier than it should be, our simple algorithm eliminated an important part of the original signal. The original background detection algorithm in the isback function relied on isback minimum and maximum limits and a search of neighboring points. How can we improve this algorithm so that small signals that fall in the detection range are not considered as background? We can assume that while the tablet is not moving, we only measure noise, which means that the data doesn't change much. When the tablet moves, much larger variations of data values would be expected. Let's check if this assumption is true: Calculate point-by-point differences in data. We can thus try to add one more test to the isback function, assigning only those points isback to background that are close in value to neighboring points. The new function is isbackdelta. This function is nearly identical to isback. The only addition is on lines 23 to 26. isback If the absolute value of the difference between the current point and the previous point is larger than delta, the function returns 0 (not background). We also make a new background calculation function that uses isbackdelta instead isbackdelta of the original isback function. The name of the new function is backdelta. We can see that the tablet actually did move sideways slightly during each move segment delta https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys05-12.svg Seminar 05 notes: Making more improvements Can we make more improvements? Take another look at the extracted background for x-axis acceleration data. The fixback function interpolates the background between point p1 and p2, which may be far from the average background value. We could calculate the average background value for each stop section of the data and use the averages instead, as illustrated by the red line. This would change slightly the integrated velocity and position graphs. This background calculation can be done with the backave function, which is a slightly modified version of the backdelta function. The background plots now consist of a very small number of points, just two point for each stop region: one at the beginning and the other at the end. We can see that the complexity of the calculation increased, but the result is very nearly the same as we had before. At some point, the effort of improving the calculation is no longer justified by an improvement of the final result. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys05-13.svg Seminar 05 notes: Documenting calculations Octave allows commands to be listed in macro files that are created the same way as functions. The commands for the tablet path analysis can be entered into a single file. For example, to process the data in the file move1.csv using the backdelta function for move1.csv background correction, we can create a macro file. Using macro files helps to reduce typing but more importantly, provides a convenient way of documenting how the calculation was done. We could see that calculation and algorithm details can have a big effect on the result. It is therefore important to provide details of calculation with the final result. This allows others to evaluate the reliability of the result. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys05-14.svg Seminar 05 notes: Lessons learned We have finished the first experiments. What did we learn? 1. Scientific approach to observing nature often requires sensors or devices to get quantitative, reproducible, and objective data. 2. The experiment must be designed to fit a particular sensor, considering parameters like sensitivity, range, noise, linearity, bandwidth, etc. 3. Even when observations are done with an instrument, there is still a subjective aspect to science, because the result depends on how a person designs the measurement procedure. 4. Measurements are often indirect. We wanted to measure the tablet movement path shape, but we had no distance sensor. We measured acceleration instead and had to calculate the position. 5. Calculations involve many subjective choices regarding the choice of data processing algorithms and parameters used in calculations. Remember the difficulties with extracting the signal backgrounds. 6. None of the sensors that we use are perfect. We have to consider the imperfect operation of a device in the calculation process. 7. There is no single correct answer. Calculation results depend on algorithm choices. If we don't have some other measurement to compare to, there is no objective reference to determine if a result is correct or not. 8. The question of correcteness of a result does not have a yes/no answer. Some results are better (more accurate) than others. Think about the different path shapes that we obtained with different background extraction algorithms. 9. Always look at measurement and calculation results with a critical eye. Look for any aspects you don't understand. We learned that results may look too good and results may be incorrect if we incorrectly ignore some aspects of the data. Think about handling the small acceleration signals that were comparable in magnitude to the sensor offsets. 10. Since the result is a combination of measurement and processing, we have to have accurate documentation of how the measurement was done, which instruments (sensors) were used, and how the data was processed. Without documentation, the result is worthless, because we cannot evaluate how reliable the result is. 11. Calculation details include many subjective decisions. It is thus especially important to document the calculation details. You can use Octave macro files to record all commands used for processing your data. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys06-01.svg Seminar 06 notes: Seeing the world through sensors: objective measurements and subjective understanding Institute for Solid State Physics Mikk Lippmaa Wednesdays, 10:25 - 12:10 KOMCEE K302 06.06.2018 6 https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys06-02.svg Seminar 06 notes: Oscilloscope demonstrations We have several kinds of sensors in the tablets. We use the accelerometer, the magnetometer, the barometer, the microphone, and maybe the gyroscope. An oscilloscope is a useful tool for looking at the signal characteristics of various sensors. An oscilloscope shows the variation of an input volatge as a function of time. This particular oscilloscope has two channels, so we can see two signals simultaneously, plotted with a blue line and a yellow line. One box corresponds to a time interval of 400 ms https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys06-03.svg Seminar 06 notes: Signal types Time Signal AC signal : time-variable but zero average Time Signal DC signal : time-variable with non-zero average DC component AC component Oscilloscope input can be in DC or AC mode. In AC mode the constant DC component is removed with a capacitor oscilloscope signal amplitude pure AC signal (almost no DC offset) We could see AC and DC signals in acceleration measurements. The z-axis reading always z included a DC offset of 9.8 m/s2. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys06-04.svg Seminar 06 notes: Sensor signals We look at the output of the ADXL335 accelerometer on an oscilloscope while waving the sensor a few times in the y-axis direction and then in the x-axis direction. The y-axis output signal is shown in yellow, the x-axis signal is shown in blue. The accelerometer output is a voltage signal with a DC offset of about 1.3 V. The waving motion causes the output to oscillate around this constant offset. The oscilloscope measurement range is 200 mV per division (dotted lines), which means that the maximum signal variation that we can see is from 0.2 V to 1.8 V. Note that the zero-level marks are about 200 mV above the bottom of the screen. The waving motion amplitude has to be quite large to generate an easily measurable signal at this scale. The sensitivity of the ADXL335 accelerometer is 300 mV/g, which means that the acceleration amplitude is about 1g. You can play the video clip in the bottom corner to see how the sensor movement corresponds to the measured output signal. The oscilloscope is in DC-coupled mode, which is useful for measuring static output values. In the example below, the sensor is slowly tilted around the y-axis to either y side by 90. The sensor senses gravity just like any other force and thus the output either increases or decreases, depending on the sensor tilt direction. The x-axis output changes by x 300 mV, whic corresponds to 1g change in acceleration. You can play the video clip to see how the output signal varies during the tilting motion. We can now switch the oscilloscope to AC mode, which means that we no longer see the 1.3 V DC offset. The oscilloscope range can now be changed to 20 mV per division and we can measure much smaller movement amplitudes (lower acceleration). The measurement now also show the output noise of the sensor. The smallest acceleration that we can measure is limited by noise. We can see that the noise amplitude is about 4 mV, which corresponds to an acceleration of 0.01g (0.1 m/s2). 2 The amplitude of the measured signal is 60 mV (0.2 g). You can see in the video that the movement amplitude is much smaller. Compare the movement range to the size of the circuit board. The next example show the output signal of a Murata ENC03-R Coriolis force gyroscope sensor. Unlike the MPU-6050 MEMS accelerometr-gyroscope, this device uses a piezoelectric crystal sensor. The output signal is proportional to the rotation rate of the device with a sensitivity of 0.65 mV per deg/s. The sensor is rotated by two full turns in about 1.5 seconds. The rotation rate is thus about 480 deg/s, which should give an output voltage of 320 mV. The measured volatge is noisy, but the average signal is about 300 mV. The oscilloscope display is set at 200 mV/division. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys06-05.svg Seminar 06 notes: Noise and drift The AC and DC components of a signal are closely related to noise and drift. Noise is generally a rapidly changing signal component that doesn't change the average value of the signal. Random noise effects can often be reduced by averaging a signal over a long period of time. We could see this technique with the use of the backave function to extract the accelerometer offsets. Drift is a slow change of the signal over time. Drift effects can be eliminated in a similar way as we used for eliminating the accelerometer offsets. This is possible if the drift is much slower than the experiment time scale. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys06-06.svg Seminar 06 notes: Digital sampling Amplitude Time Analog data is sampled (measured) at certain intervals. The interval maybe constant or variable. The output signal of a sensor device is usually an analog value, like a voltage or a change of capacitance. Our tablets work with digital signals that are produced by an analog-to-digital converter. Such converters measure the value of the analog signal at some time intervals and provide numbers that correspond to the magnitude of the original signal. For example, the digital oscilloscope that we used in the sensor demonstrations accepts voltage signals at the inputs, converts the voltage signals to arrays of numbers, and then plots these numbers on the screen as voltage-time plots. The function is similar to the SensorKinetics application on the Nexus 7 tablets. The sampling rate of the measurement should be selected to fit the maximum expected frequency of the signal. Too low sampling rates will not give an accurate digital representation of the signal while too high sampling produces large volume of data that carry no useful information. (You will see this in the oscillating weight experiment). Digital signals are usually stored and processed in a computer that uses binary logic. Numbers are therefore represented as binary numbers. Decimal number Binary number. The precision of the digital representation of an analog signal can be improved by increasing the number of bits (bit depth). Increasing bit depth increases data volume and cost. The optimal bit depth is usually set by the noise level of the original signal. Digitizing noise doesn't usually improve the final analysis result. 8-bit graphics (few colors, few pixels) 64-bit graphics (full color, many pixels) CD music 16-bit samples (65536 levels) at 44.1 kHz sampling rate Hi-res audio 24-bit samples (16777216 levels) at 192 kHz sampling rate Where can you see digital sampling in everyday life? Example 1: computer graphics (games) Example 2: digital audio https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys06-07.svg Seminar 06 notes: Accelerometer characteristics Useful Octave functions: max(a) the maximum value of the vector a. max(a) a min(a) the minimum value of the vector a. min(a) a We see from the raw data plot that the range of measured values is +/- 2g (g = 9.81 m/s2, 2g = 19.62 m/s2) The values -19.866 and +19.36 are not quite symmetric around zero. The offset is -0.253 m/s2 but this depends on movement history. Useful Octave functions: diff(a) calculate differences between elements in vector a. abs(a) calculate absolute values of elements of vector a. round(a) round elements of vector a to nearest integer. sort(a) sort the elements of vector a. unique(a) select only unique elements of vector a. We need to find the smallest changes that occur in the data. We can find the smallest difference in the data with the help of these functions. For example: Create sample data Calculate differences between elements Absolute values of differences Sorted list of differences Sorted list of unique difference values. We used the unique function, so why are there unique equal values in this list? Answer: rounding. Look at full precision with format long command format long So, the values are all different, but we don't need so many digits The smallest (non-zero) difference is 0.0012 m/s2 2 and all values are multiples of 0.0012. The maximum value was 2g = 19.62 m/s2 2 there are thus 19.62 / 0.0012 = 16350 possible values Actual number is +/- 16384 values, which means that the accelerometer digital signal is a 15-bit number. Maximum range is 215 = 32768 or -16384 to +16383 15 The actual resolution should thus be We can round the numbers to 4 fractional digits: The Nexus 7 tablet accelerometer characteristics can be determined by looking at the data collected with SensorKinetics. We saw from Invensense's home page for the MPU-6050 sensor that there are various versions of the accelerometer chip with different measurement ranges. We can find which version our tablets have by measuring rapid up-down movement that saturates the sensor (rapidly shaking a tablet). Another important sensor characteristic is the resolution (bit depth). We can determine the reslution by acquiring at a slowly varying signal and looking for the smallest recorded difference in values. We can now use the same method to find the smallest differences in the slow.csv data. For example, look at element s(2) Multiply by 104 to move the floating point four places 4 Round to integer Divide by 104 to get a number rounded to for 4 fractional digits Put all these together in one command and look at the smallest 10 values https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys06-08.svg Seminar 06 notes: A pressure sensor Interlink Electronics FSR-406 is a pressure sensor. It's pressure response is highly nonlinear and the sensor is primarily meant for sensing touch. The resistance of the device decreases when the active area is pressed. When the active area is pushed with a finger, a conducting ink layer is pressed against a sheet with two interdigitated electrodes. Resistance drops and is somewhat proportional to the applied pressure. This type of sensor might be used as a user interface element, in portable music players, for example. conducting ink The technical details of the FSR-402 device can be found in the datasheet. We are interested in the force response of the sensor. To measure the response, we start piling small weights (stainless M8 nuts) on the sensor and observe the resistance change as we add weight and then as we remove weight. Each nut weighs about 5 grams. The sensor is connected to the oscilloscope, so we can measure the volatge change on the sensor. R FSR oscilloscope FSR-406 Pressure response curve Weight (Nuts) Voltage (V) 39 mV You can click the left/right arrows to view the data for each point in the graph. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys06-09.svg Seminar 06 notes: Sensor response The force sensing resistor device showed several typical features of a nonideal sensor. (This is normal behavior for a touch sensor, which is intentionally nonlinear). Most of these features are illustrated in this plot. All sensor outputs include noise, either from the sensor itself or from the Noise environment (noisy room, vibration due to trafffic, etc.). Detection limit: The smallest excitation that the sensor can detect above noise. Detection limit For example, the smallest weight that can be measured. Saturation: The maximum excitation that can be measured. For the Nexus 7 tablet Saturation accelerometers, the range was 2 g. Hysteresis: The memory effect of a sensor, where the output depends on the previous Hysteresis values. We saw hysteresis in the background offset behavior of the accelerometers in the Nexus 7 tablets. Sensititvity: Each sensor has a certain specified sensitivity, such as mV / g for the Sensititvity accelerometer. Sensitivity is given by the slope of the sensor response plot. Nonlinearity: The sensitivity of the sensor often varies with excitation. Sensitivity Nonlinearity is zero below the detection limit and above the saturation limit. Excitation (e.g., weight) Response (output signal) Can measure down to zero signal An ideal sensor would have a linear response (constant slope) from zero to maximum measurement range. The detection limit would be zero, no saturation, and no hysteresis. In practice, non-ideal behavior muxt be considered in the measurement. The most important characteristics of a sensor device are usually listed in the datasheet. Linearity error is usually given as a percentage of the measurement range or %FS, where FS stands for 'full scale'. The sensor performance is usually affected by various outside parameters. Most common parameter is temperature, which may affect the offset, sensitivity, noise level, or other sensor characteristics. The operating temperature range is therefore given in the datasheet. Various temperature coefficients can be found in the gyroscope datasheet for sensitivity, zero offset, and the guaranteed range of operating temepratures. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys06-10.svg Seminar 06 notes: Sensor errors Sensor output is always an approximation of a true value. Accuracy and precision describe ow good this approximation is. Accuracy describes the correctness of the measurement result. It tells you Accuracy how close the measured value is to the 'true' value. Precision descibes the variation of measurement results when the same measurement Precision is repeated. There is always some variation, perhaps due to noise. Accurate but not precise Not accurate but precise Accurate and precise Systematic errors: Caused by temperature, observer mistake, incorrect calibration, etc. Systematic errors Systematic errors can be corrected or compensated for. Systematic errors can be found by comparing different measurements that use different sensors, setups, people, etc. Random errors: Caused by various forms of noise, either intrinsic or related to Random errors the measurement setup. Random errors can be reduced by repeated measurements and averaging. random error systematic error Accuracy and precision describe various random or systematic measurement errors. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys06-11.svg Seminar 06 notes: Group projects Next week we start group projects where you can plan and design your experiment, perform measurements, process your data, and prepare a short presentation. planning, measurements measure and process data process data,prepare presentation presentations, discussion https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys06-12.svg Seminar 06 notes: Group 1: Magnetic navigation The magnetic field direction is uiquely defined in the vicinity of a magnet. This can be used to determine the position of a magnet relative to a magnetometer. 1. Measure the magnetic field distribution in a single plane above a magnet. 2. Visualize the field distribution (magnitude and field direction) in 3-dimensional plots. 3. Move a small magnet above the tablet. Calculate the magnet position using the field distrbution data measured earlier. Vector field mapping 3D visualization of a vector field Data interpolation Data modeling Tasks: Techniques: https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys06-13.svg Seminar 06 notes: Group 2: Accelerometer position We know that the Nexus 7 tablet contains an accelerometer sensor. Your task is to find the location (x,y coordinates) of the sensor in the tablet. x y 1. Find the x and y coordinates relative to the tablet edge. x y 2. Pay special attention to the accuracy of the result when you plan the measurement. 3. Compare your result with Nexus 7 teardown descriptions on the Web. Collecting simultaneous data from several sensors Data averaging Linear fitting Tasks: Techniques: x y https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys06-14.svg Seminar 06 notes: Group 3: Speed of a dart Use sound recordings to determine the speed of a NERF gun dart. 1. Measure the velocity of a dart. 2. How variable is the velocity (dart-to-dart)? 3. How variable is the velocity between darts? 4. Do you need to worry about the speed of sound? Sound analysis Triggering from data Statistics, distribution of data values Tasks: Techniques: https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys06-15.svg Seminar 06 notes: Group 4: Weight of air The barometer in a Nexus 9 tablet or the Galaxy S3 phone can be used to measure air pressure. 1. Determine the resolution of the barometer. 2. Characterize the noise level, response time, and drift. 3. Measure the height of the KOMCEE building. 4. How high/low pressures can your lungs generate? How much variations is there between persons? 5. Find way to use the barometer to measure weight. Sensor characterization Indirect measurements Noisy data with few data points (lung pressure) Tasks: Techniques: low pressure high pressure https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys06-16.svg Seminar 06 notes: Group 5: Damped oscillations 1. Measure up-down oscillations of a long period. 2. Observe the amplitude damping. 3. Fit the damped oscillation data with a mathematical model function. 4. Measure the speed of longitudinal and transverse waves in a long stretched spring. Data modeling Data fitting Trigger point detection Tasks: Techniques: A long spring attached to a tablet can be used to measure simple oscillatory motion with the built-in accelerometer https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys06-17.svg Seminar 06 notes: Group 6: Secret message 1. Find the magnetic signal in the wire. 2. Read the magnetic message. 3. Plot data, determine how the message is encoded. 4. Read the message by eye. 5. Read the message with Octave. 6. Write a new message. Pattern identification Digital information encoding Pattern extraction Tasks: Techniques: A magnetic message has been encoded and recorded on a magnetic wire. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys06-18.svg Seminar 06 notes: Homework 1. Think about how you would perform the measurement. 2. Think about what kind of data you expect. 3. Think about how you would process the data. 4. Think about how you can prepare a Slides presentation where you explain: - What your measurement task was - How you planned the measurement - How you carried out your measurement (pictures?) - Plots of your original data - Explain how to processed the data - Show your result - Discuss problems, questions, suggestions for your experiment. 5. Think about how to share work: - who is in charge of experiment - who does data processing, plotting etc. - who prepares documentation and presentation https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys07-01.svg Seminar 07 notes: Seeing the world through sensors: objective measurements and subjective understanding Institute for Solid State Physics Mikk Lippmaa Wednesdays, 10:25 - 12:10 KOMCEE K302 13.06.2018 7 https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys07-02.svg Seminar 07 notes: Other sensors Our tablets have several other useful sensors that we can use: - Accelerometer and gyroscope (gravity, acceleration, rotation rate) - Barometer (pressure, available in Nexus 9 and Galaxy S3) - Microphone (sound, spectrum, time) - Magnetometer (direction, magnetic field) - Camera (color, shape, brightness, etc.) We take a quick look at a gyroscope, a magnetometer, and a MEMS microphone. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys07-03.svg Seminar 07 notes: Magnetometer The most common technique for measuring magnetic fields with miniature sensors is to use the Hall effect in a semiconductor. The Hall effect is based on the Lorentz force experienced by electrons moving in a magnetic field. electron charge velocity vector of electron magnetic flux vector The cross product of v and B means that the is perpendicular to both of these vectors. The effect of the Lorentz force can be seen in a simple demonstration with a magnet and a wire loop. Connecting a battery will force the wire loop to swing to right or left, depending on the direction of the current flow. The Hall effect is observed in semiconductors: 1. A current passes through a semiconductor along the x-axis direction. x 2. When a magnetic field is applied in the vertical z-axis direction, electrons are z pushed to left or right (along the y axis) y depending on the field direction. 3. A Hall voltage Vxy is generated V xy between the left and right edges of the semiconductor sheet. The voltage is proportional to the field strength. A Hall device can sense only one component of the magnetic field (in this case, the vertical component). All x, y, z components can be measured with three x y z Hall devices positioned at 90 to each other. This design can be seen in an Asahi Kasei AK8970 3-axis electronic compass x-axis Hall device x y-axis Hall device y z-axis Hall device z This device design is large, requires complicated assembly, and is costly. Smartphones and tablets therefore use a planar Hall device that can measure all field components with the help of a magnetically soft FeNi permalloy magnetic concentrator. The effect of a magnetically soft ferromagnet in a magnetic field can be seen in the shape of the field lines. Magnetic field lines bend near the soft magnet. Even for a x-oriented magnetic field, a x z-axis component can be generated close to z the edges of the magnetic concentrator. Readout circuit Hall effect sensors can sense only the z-axis component. The two field components can be separated becuse both sensors see Bz with the same sign, while the modified Bx field changes sign at the H1 and H2 locations. In practical electronic compass devices, the concentrator is a FeNi metal disk placed on top of the sensor die, as shown in this cross-sectional scanning electron microscope image. Si die sensor layer die coating FeNi concentrator The placement of the 8 diamond-shaped Hall sensor devices along the perimeter of the disk-shaped concentrator can be seen in the plan view of the sensor die. This example is for the Asahi Kasei AKM8973 magnetometer that was used in older iPhones. Chipworks v https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys07-04.svg Seminar 07 notes: Gyroscope Gyroscopes use the Coriolis ef- fect to detect rotation. The Coriolis effect causes storms to spin counter-clockwise in the northern hemisphere and clockwise in the southern hemisphere. If we know the velocity of an object v and the direction of the rotation vector, the direction of the Coriolis force is given by the vector product. Coriolis force gyroscopes thus have an internal moving component and a mechanism to measure the lateral Coriolis force. A test mass moving in the x-axis direction with a velocity v would sense FCoriolis in the y direction if subjected to clockwise rotational y motion around the z-axis as shown here. However, a sensor cannot continue moving in one diection and such a test mass sensor would also sense acceleration in the y-axis y direction. Practical sensors thus use oscillating motion and may use two coupled masses moving in opposite directions. The Coriolis foce would be felt in opposite directions by the two masses (since the velocities are in opposite directions). Measuring the difference of forces between m1 and m2 can be used to eliminate the sensitivity of a gyroscope to acceleration forces. We have looked at a MEMS gyroscope in the Nexus 7 and used a piezo bimorph device in the oscilloscope demonstrations. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys07-05.svg Seminar 07 notes: Piezo gyroscope A bimorph gyroscope uses a piezoelectric crystal. A piezo crystal generates a voltage when squeezed or stretched. When a volatge is applied, the crystal itself stretches, shrinks. A piezoelectric crystal can thus work as an actuator or as a sensor. The piezo bimorph is driven to oscillate at its resonance frequency of about 30 kHz. The bimorph now works as a test mass (m1, see above) that m 1 is periodically moving in the up-down direction. If the bimorph is rotated around its long axis, the Coriolis force will push the bimorph sideways. piezoelectric bimorph is a device that consists of two piezoelectric layers (bi-) attached in a way to generate a shape change (morph) when a volatge is applied to the top and bottom electrodes. To achieve bending movement, the polarization directions of the two piezoelectric layers (P1 and P2) point in the opposite directions. When a bias is applied, one layer expands while the other shrinks, causing the bimorph to bend up or down. Murata ENC-03 gyroscope Piezo bimorph Electrodes and support Electrodes and support. Bimorph movement without rotation (use the PLAY button to view the animation) Bimorph movement with rotation (use the PLAY button to view the animation) The effect of the Coriolis force is the greatest when the bimorph movement is the fastest. For an oscillating bimorph, this happens twice during each oscillation period when the up-down bending is zero. The sideways bending of the bimorph distorts the piezoelectric crystals. This distortion can be seen more clearly in a cross-section view of the bimorph structure. The surfaces highlighted are clearly unsymmetric, becoming thinner on the outer edge and thicker on the inner edge. This difference can be sensed, since a shape change generates a surface charge in a piezoelectric material. The top (red) and bottom (blue) electrodes of a bimorph can be used to make the bimorph flex and oscillate in the up-down direction. In a piezo bimorph gyroscope, the top electrode is further split into two parts to sense the bimorph distortions. If there is no rotation around the long axis of the bimorph, the voltage difference the two top electrodes is zero. If the gyroscope rotates, the Coriolis force will periodically distort the bimorph, giving a vltage difference that is larger than zero while the bimorph is moving up (U>0) and less than zero (U<0) when the U U bimorph is moving down. The rotation rate of the gyroscope can be calculated from the magnitude of the measured voltage difference between the two top electrodes. One gyroscope device can sense rotation around just one axis. For the ENC-03 sensor, the output voltage is 0.67 mV for a rotation rate of 1 deg/s. To measure rotation rates around several axes, more than one gyroscope device can be combined. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys07-06.svg Seminar 07 notes: MEMS microphone Our tablets contain tiny MEMS microphones. We look more closely at a microphone made by Knowles, model SPU0414HR5H-SB. This is a silicon MEMS device. 3 mm Bruker This is an enlarged view of the microphone device. A scanning microscope shows that there are several layers in the device. The yellow and orange layers re two electrodes that form a capacitor capacitance varies with membrane distance charge These layers correspond to two electrode sheets that form a capacitive distance measurment device, simlar to the accelerometer that we looked at before. Either voltage, charge, or current can be detected to measure the change of capacitance when the membranes moves due to changes in air pressure in a sound wave. There are many similar MEMS microphone design. We can see the micromechanical details in scanning electron microscope images from SITRI and Chipworks. 20 m Silicon wafer Silicon wafer Moving diaphragm Top electrode and support The cross-section image shows the silicon wafer that is 250 to 300 m thick and a back acoustic cavity. The actual microphone structure is at the top of the wafer and less than 10 m thick. A closer look at the membrane structure shows the different electrode layers. We can see the shape of the acoustic cavity, the polysilicon membrane and the silicon nitride support at the top. The support layer holds the polysilicon top electrode. Since the capacitance of the device is given by, the sensitivity of the microphone improves if d is small. We can appreciate just how thin the device is by compring the microphone membrane image with a typical human hair on the same scale. The holes in the top support are needed to separate the membrane from the support and to allow air to move. The thickness of the whole structure is just 6 m. The membrane thickness is about 1 m. We have looked at several MEMS devices from InvenSense and Knowles. These manufacturers don't actually make their own MEMS devices. These are known as fabless manufacturers, meaning that they don't have their own fabrication plants, The MEMS device fabrication is done in foundries. One of these is the Sony Semiconductor Kyushu corp. Some of the key technologies are the etching tools Anisotropic deep rective ion etching Isotropic volume etching We can look how these tools are used to manufacture the microphone device https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys07-07.svg Seminar 07 notes: Photolithography Microfabrication is based on photolithography. Some examples from my lab. Start with a single crystal Si wafer (the substrate) Expose a photoresist through a photomask Produce a patterned microcircuit Step 1: Deposit a thin film Step 1 Substrate (Si wafer) nm scale Step 2: Coat surface with a photoresist (light-sensitive polymer) Step 2 m scale The photoresist is originally a liquid that is spin-coated on the wafer surface. The resist is then heated, until it forms a hard layer on the surface. Step 3: Place a photomask above the wafer Step 3 quartz plate Step 4: Expose ultraviolet light Step 4 Cr metal pattern UV light A photomask is a quartz (transparent to ultraviolet light) plate with pattern that is transferred to the wafer surface. The ultraviolet light induces a reaction in the photoresist. In this example, the exposed photoresist becomes more easily soluble (positive photoresist). Step 5: Develop the resist Step 5 Step 6: Wash off the exposed resist. The mask pattern has now Step 6 been transferred to the resist layer on the film surface. Developer Dissolve the exposed (or unexposed) parts of the resist Step 8: Wash away the etched parts of the films. The original Step 8 mask pattern has now been transferred to the film layer, but the resist is still there too. Step 7: Etch the film through Step 7 the openings in the resist Acid Remover Clean away the leftover resist Step 10: Final patterned film Step 10 Step 9: Use a remover (acetone) to Step 9 dissolve the remaining resist These basic steps are repeated many times in the manufacturing process to build each layer of the device structure. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys07-08.svg Seminar 07 notes: Si (001) substrate Manufacturing process One microphone process is disclosed in US patent 5,573,679. The purpose is to manufcture the following structure: Si substrate empty cavity dielectric empty cavity flexible membrane (bottom plate Ti, TiN, Al, ...) poly-Si dielectric fixed top electrode C air SiN SiN spacer SiO2 2 SiO2 2 Si (001) substrate Protective SiN layer Bottom membrane, SiN Bottom electrode, Al (determines microphone area) Step 1: Deposit a SiN membrane, pattern the bottom electrode Step 1 Polycrystalline silicon spacer Step 2: Deposit SiO2 or TiN barriers, polysilicon spacer, and a metal top electrode Step 2 2 Al SiO2 2 Step 3: Deposit SiN top dielectric, pattern bottom cavity Step 3 SiN backside cavity KOH slow etch Step 4: Deposit SiOx mask, photoresist and reactive ion etch to open the electrode holes Step 4 x SiOx x resist ion etching backside cavity Step 5: Dicing, cavity etch, resist removal Step 5 SF6 dry etching 6 dicing dicing Many devices are manufactured in parallel on a single wafer. The final step of wafer processing is called dicing, where the wafer is cut into individual microphone devices. Step 6: Packaging Step 6 https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys07-09.svg Seminar 07 notes: Microphone response The main benefit of the MEMS microphones is the small size. For a microphone, the most important characteristic is the frequency response curve, which tells us what the sensitivity of the device is as a function of the sound frequency. For a tablet or a mobile phone, the main application is to capture human voice, which means that we look at frequencies between about 300 and 3000 Hz. Sensitivity (dB) Frequency (Hz) The frequency response curve of the SPU0414HR5H-SB microphone is the following. The data is normalized to the response at 1 kHz. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys07-10.svg Seminar 07 notes: Decibel Time-dependent signal amplitudes are often shown in dB For signal power. Microphone sensitivity is usually given as dB (SPL), where the SPL (Sound Pressure Level) reference is the human hearing limit of 20 Pa. A Most physical quantities are measured on a linear absolute scale, like temperature or voltage. When measuring sound wave amplitudes, an absolute pressure scale is inconvenient because human hearing is logarithmic, rather than linear. The most quiet sound that humans can hear corresponds to an air pressure of 20 Pa, while the loudest sounds that don't damage hearing is about 20 Pa. An increase of 10 in sound wave pressure feels about twice as loud. It is therefore convenient to use a logarithmic scale, but it is not possible to take a logarithm of a physical unit. A ratio scale is therefore used, which tells us how much larger or smaller the amplitude of a wave is compared to a reference. This ratio can be given on a logarithmic scale, since the ratio is dimensionless. The ratio unit is Bel, but a more common unit is 1/10 of a Bel, or a decibel (dB). https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys07-11.svg Seminar 07 notes: Bandwidth Every sensor has a minimum and maximum frequency that can be measured. Bandwidth is the frequency range where response of the sensor drops to -3 dB (power drops to half). The cut-off level may be set at different values, but it is common to see references to '3 dB bandwidth' or similar specs. BW ~18 kHz BW ~75 kHz Slower, but more linear. Sensitivity does not change much. Good for sound intensity measurement. Faster, but inaccurate. Response peaks strongly at 20 kHz. Good for acoustic positioning. The bndwidth can be found in the sensor data sheets. The plots below show the frequency response curves for two microphone models. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys07-12.svg Seminar 07 notes: Response time Response time (rise time) shows how quickly a sensor responds to a fast input change. The response time is directly linked to bandwidth, but it is sometimes easier to use time units instead of frequency.In general, the 3dB bandwidth. The response time (and bandwidth) often depends on the size of the sensor (mass, heat capacity, capacitance, etc.) https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys07-13.svg Seminar 07 notes: Settling time Another time-dependent sensor signal charateristic is the settling time. The settling time tells us how long it takes for the sensor output signal to become stable after a fast input change. For example, an accelerometer may show some oscillation in the output because the test mass vibrates for short time if the accelerometer is hit by very sharp acceleration (dropped on a hard surface). The settling time is related to the shape of the frequency response curve. 1 10 100 1000 10000 20 Frequency (Hz) Signal. A long settling time may require a delay after an input change before the sensor output signal reading can be taken. Resonance settling time Ringing If the frequency response curve of a sensor shows a strong peaking behavior as shown by the red plot, the sensor output signal is likely to oscillate. This oscillation is called ringing, at it may make the settling time significantly longer than the response time. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys07-14.svg Seminar 07 notes: Group projects ALL data must be uploaded to Slack. ALL macros, comments to Slack. ALL plots, images to Slack DM me for private comments, post group discussion in your group channel. If you don't know what to do, DM me. Octave problems, DM me. For Octave data processing, refer to the list of functions provided for you at https://lippmaa.issp.u-tokyo.ac.jp/lecture/macros https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys07-15.svg Seminar 07 notes: Group 1 notes The magnetic field direction is uniquely defined in the vicinity of a magnet. This can be used to determine the position of a magnet. Task 1: Measure the magnetic field distribution Task 1 in a single line above a tablet by moving a small test magnet. Use the plastic magnet holder with slots. Instead of moving a sensor (tablet) over a magnet, you can fix the position of the tablet and move the magnet instead. Use the small magnet to find the approximate position of the magnetic field sensor in the tablet. WARNING: Don't put the magnet too close to the WARNING: tablet. It will permanently magnetize parts of the tablet, creating an unwanted background. Use the plastic boxes and tape to position the magnet holder above the tablet at a height were the magnet doesn't saturate the magnetometer. It is inconvenient to make a separate measurement at each data point. You saw in our earlier tablet accelerometer analysis how several move segments could be analyzed when there is a quiet period with no motion between each movement. Software could easily separate the movemet from the quiet background parts. We can use the same idea here. - move the magnet to the next slot - collect data for about 10 seconds repeat move wait Hints: Calculate averages for each ~1-second section and look for several 'quiet' Hints sections in a row. We know that that's where the magnet didn't move. 1 second intervals change quiet Consider : Consider - How can you handle backround? - What is the optimal sampling rate? Consider : Consider - Magnetic metals affect the magnetic field distribution. Where in the classroom is the best place to set up the experiment? Look at example data measured by moving a magnet every 5 seconds. The field values can be picked out of the data by looking for sections where fthe ield doesn't change (magnet is not moving). You can use the boxave function for this purpose. It will return boxave one data point for each quiet period in the data. boxave We can try to process sample data, where the blue plot shows the x-axis magnetic x field component. Note that the measurement includes stable background measurements at the start and end of the plot. This example shows how the field points can be extracted from a single long field measurement. The boxave function has two additional boxave parameters to control the data extraction: boxave(T, M, points, limit) boxave T M points limit Time values of the measured data Field values (x, y, or z axis) of the measured data x y z Number of data points in an averaging box Maximum allowed variation of field values for a 'quiet' box. The default value for the points parameter is 50. If the sampling rate is low, there may points be less than 50 data points in each quiet region where the magnet didn't move. In this case, the points parameter should be set to a lower value. The best box width is about points one quarter of the quiet period. The default value for the limits parameter limits is 2. This value can be adjusted if the data is noisy. Problems may be caused by unintenionally touching the magnet or shaking the desk where the measurement is done. Two points found for one position magnet touched but not moved This problem can be solved by using a larger box width and a lower threshold: If a few incorrectly found data points remain in the data set and adjustment of the boxave parameters is not effective, it may be boxave easier to delete those points manually with the remove function remove remove The plot on the right shows the extracted points for the z-axis field component. The z first and last data points correspond to field values that were measured while the magnet was moved far away from the tablet. these are background points and we can see that the background is not zero. This background must be subtracted from the data. We pick the first and last data points into new vectors Gzt and Gzv and interpolate field values to get the background plot (shown in green). The background can now be subtracted from x and z components a vector plot in Octave was explained in the May 9 seminar. So far, we have the magnetic field components in Bxv and Bzv, but we also have to tell Octave where the measurement points were. Assuming a 1 cm spacing, we can plot the vector plot of the field direction and magnitude. Perform the measurement Store processing and plotting commands in an Octave script Document the measurement, processing, and the result in Slack and Slides. Task 3: Measure the magnetic field distribution Task 3 in a plane. Use the plastic sheet with a 2D array of holes for magnet positioning. Task 2: Visualize the field distribution in a vector plot Task 2 Perform the two-dimensional x/y mapping experiment using a 12 12 grid of measurement points. Try to position the magnetic field sensor close to the center of the grid. back ground back ground back ground The measurement procedure is the same as before, except that the whole grid of 12 12 points can be measured in one sequence as shown in the diagram. You saw that background correction is generally necessary. It is therefore important to measure bacground in the beginning, between each line of points, and at the end of the sequence. Together with the background points, you should have an array of 13 12 + 1 = 157 data points. You can extract the data with boxave as explained before. boxave The result should look something like this: If the 157 data points can be correctly extracted from the data, you should have three vectors with equal size holding data for the x-, y-, and z-axis field components. x- y- z- The data can be reshaped into a matrix with the Octave reshape command. reshape This command takes the first 156 elements of the Bzv vector and reshapes it into a 13 12 matrix where the first row contains the background data for each column. The background can now be subtracted and removed from the matrix the matrix transpose here ( the ' character) Just like the linear mapping, we need to tell Octave where the measurement points were (the x, y coordinates) to preparea vector plot. We don't care about the absolute x y positions in mm, so we can use just an array of numbers generated with meshgrid. meshgrid The Octave surf command can be used to view a 3D surface plot of the result. surf The same process should be repeated for the x and y components, producing plots x y similar to those: x-axis data x y-axis data y z-axis data z We can set the z-coordiantes to zero for all points with the Octave zeros command. z zeros The 3D vector plot can now be generated with default view top view set with view(0,90) Note the minus in front of By. Some component directions may need to be reversed to get the correct plot. This depends on the tablet orientation. measurement plane The vector plot shows the expected field directions. You can see the reversal of the z-axis component in the corners of the mapping area along the diagonal. Task 4: Move the magnet to a few points on the measurement grid. Task 4 Using the previous mapping data, try to determine the magnet position. This set of experiments aims to practice the following techniques: Time multiplexing: Sequencing of measurements so that data from different points Time multiplexing in space is stored in sequential time slots. Vector field: Measuring and visualizing the spatial distribution of a vector field. Vector field Patience and finger skills: Performing a long experiment that requires constant Patience and finger skills attention. One mistake in operation may degrade the whole dataset. Vectors and matrices: Working with and converting between matrix and vector Vectors and matrices representations of data. Data reuse: The final task requires the use of earlier measurements as reference data. Data reuse This cannot be done well if the original experiment is not properly documented. The field mapping experiment gave us a 12 12 array of field points. We now go back and measure the field strength at a few points (7 shown here) and our purpose is to find the coordinates of these points by comparing the field values to the earlier 12 12 mapping results. The idea would be to take the x, y, z field components at an arbitrary point and find which 12 12 mapping grid point has the same (or closest) field component values. That point would tell us the x-y coordinates x y of our measurement point. We thus need to compare vectors and determine how similar they are. One simple way is to calculate a 'distance' between two vectors. For points in 2D space, the distance between points at coordinates Let's say we place the magnet at point 1 in the image and measure field components. We can now calculate the distance from this vector F to all points in our mapping grid. mapping data and separate x, y, z components x y z Extract 13 12 grid points Reshape into matrices Subtract background Remove background rows The mapping data is now in Bx, By, Bz. We can pick one data point from this grid for testing, for example the point at row 3, column 8. Calculate the distances from this vector to all the 12 12 grid points. This creates a new distance matrix. octave:16> Note the element-by-element calculation of squares. We can plot the distances matrix Minimum distance is here The minimum value and the location of the minimum can now be calculated Octave tells us the that the closest vector is indeed at row 3 and column 8 of the mapping data. Strong maximum The distance matrix plot shows a very strong maximum peak, which is caused by the differences of field vector magnitudes. When searching for the closest mapping vector, it is advisable to normalize all vectors by their length, so that the vector similarity is decided by cmparing vector directions only. This can be done by slightly modifying the calculation (restarting from Octave step 16) result is of course the same, but the distance matrix doesn't show strong peaking, which would be good for handling noisier data. Minimum distance is here No peaking Calculate lengths for all mapping vectors and divide each vector with its own lentgh. Do the same length normalization for the search vector https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys07-16.svg Seminar 07 notes: Group 2 notes This set of experiments aims to practice the following techniques: Experiment planning: Several possible approaches can be taken to solving the problem. Experiment planning Need to consider methods and select the optimal approach. Simultaneous measurements: Collecting data from several sensors simultaneously. Simultaneous measurements Combining measurement data from several sensors in the data processing. Noisy data: The measurements are affected by mechanical noise. Noisy data Improving measurement accuracy by repeating experiments and averaging results. Vector manipulation: Learn how to perform calculations with data vectors. Vector manipulation Linear regression: Learn to use linear least squares fitting to improve reliability of Linear regression measurement results. We saw that our tablets contain various sensor devices. These are usually MEMS or semiconductor circuits somewhere in the tablet. The purpose is to find the location of the accelerometer device in the tablet without looking inside. A turntable is provided with a tablet holder. The tablet can be rotated around a fixed point and the tablet distance from the rotation center and the position or orientation of the tablet relative to the rotating arm can be changed. The distances from the rotation center to the edge of the tablet can be measured with a ruler. Task 1: Determine the accelerometer distance Task 1 from the rotation center using acceleration data and the rotation time. The location of the accelerometer can be found by considering rotational motion. We know that in rotational motion, which means that if we measure the acceleration a, we can calculate the a distance of the sensor from the center point, r. However, we need to know r the linear velocity v. v We cannot directly measure velocity, but we can measure the time it takes to make one full circle and calculate the velocity, Therefore, How can you measure the rotation time? Should you use a large or a small rotation radius? Is there an optimal rotation radius? There are several choices for measuring the rotation time (period): Choice 1: Magnetic field direction. The x and y axes show sine wave shapes. The x x y Measure period from this wave. Choice 2: Rotation rate sensor. This gives the tablet angle as a function of time, but it is a 'derived' sensor, calculated from the gyroscope data and it has a problem with background drift. Another option is to use the gyroscope sensor to measure the rotation rate in rad/s. This is known as the angular velocity. The acceleration of the rotating tablet is in this case given by Load the magnetometer data into Octave, use the limits function to convert data to 0/1 limits values, edges function to find the time values edges of each rotation, diff to calculate the time diff differences, and mean to get the average mean period. (mean and diff are Octave functions) mean diff T limits edges x-axis x data after limits The average rotation period was thus 2 s, but we can see that the rotation rate was not constant, varying between 1.79 and 2.44 s. Since the tablet rotation is done by hand, there is a significant error in the rotation time measurement. This is the largest source of positioning errors for this measurement method. Task 2: Determine the accelerometer Task 2 distance from the rotation center using the accelerometer and the gyroscope data. The acceleration data needs to be saved together with the magnetometer data. Use the "Multi-Sensor Recorder" option in SensorKinetics. Select only those sensors that you actually need in the selection list. Remember to check the recording Rate. Using the fastest rate will produce very large data files. A sampling rate of 50 Hz (Game or UI) is sufficient. The data for individual sensors can then be saved and uploaded to Slack and Octave-online. Load the accelerometer data into Octave and plot it. Find in this plot the time period where the acceleration value is more-or-less stable. In this plot, for example, an appropriate time period would be from 7 s to 14 s. We need to calculate the average acceleration in this time period and show the average value in the plot. The example is for the y-axis y acceleration. The x-axis data can be processed x in the same way. The distance r from the rotation center to the accelerometer can now be calculated as The calculated rotation radius is thus 491.76 mm for this measurement. You can either extract the average a and a values from the accelerometer and gyroscope data and calculate the radius r, r or calculate r values for each time point r and average the calculated r values. r For example, if the accelerometer data is in a30.csv and the gyroscope data is in g30.csv: Calculate r for each data point. r Note the element-by-element division (./) and square (.^2) Find a suitable time range for averaging r(t). r(t ( In this case from 2 to 10 seconds. Calculate the average r from t1=2 to t2=10 seconds Plot the time dependence of r in blue r Plot the average value of r Task 3: Determine both x and y Task 3 x y coordinates of the accelerometer sensor in the tablet. In this example, the sensor distance from the rotation center is 394.37 mm. So far, in Tasks 1 and 2 you measured the distance of the accelerometer sensor from the center of rotation. However, just knowing r doesn't give you the x, y coordinates. the accelerometer sensor is not positioned directly above the rotation arm (in the x direction), the tablet will x show both x and y components of acceleration. This effect can x y be seen in the acceleration plot for Task 2. Although the main acceleration component is in the y-axis direction (red), y there is also a small x-axis component (blue). x sensor You can shift the tablet along the x direction in small steps x and find a position where the x-axis acceleration component x is at a minimum. You can then measure the ay component and the y-direction y distance of the sensor from the tablet edge. x ax x The zero crossing gives you the desired x position x Once the y-axis distance is known, the tablet can be rotated 90 and the measurement y repeated to measure the x-direction distance from the edge. x Caution! You have to rotate the tablet at approximately the same speed every time, otherwise searching for the zero crossing of either ax or ay becomes very inaccurate. a x a y Task 4: Try to improve accuracy by measuring Task 4 the x and y positions simultaneously. x y 300 mm r = 394 mm calculating the rotation radius is easier if we combine accelerometer (a) and gyroscope () measurements. a In this case and for x, y components: We can now use radiusx and radiusy functions to calculate radiusx radiusy the components of r for a set of different distances x. For example, measuring a and for 5 distances x gives this data zero crossing The calculated rx values can now be plotted r x and a line can be fitted to the data. The point where the line crosses zero gives you the x x value from the tablet edge to the accelerometer sensor position. r x sensor position In this case, the zero crossing was at -111.3 mm, which means that the sensor distance from the left edge of the tablet holder was 111.3 mm. The ry values can be obtained from the same measurements and should be r y the same for all x values. An average can thus be taken of the measured ry values. x r y https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys07-17.svg Seminar 07 notes: Group 3 notes This set of experiments aims to practice the following techniques: Sound detection: Using microphone data for time measurement. Sound detection Sound analysis: Analysis of sound wave features to select gun and dart sounds. Sound analysis Triggering: Finding time reference points in sound data. Triggering Statistical analysis: Determining the errors in measuring dart speeds Statistical analysis (which are very variable) Error sources: Considering various possible sources of systematic errors. Error sources When a NERF dart is fired from the gun, the gun makes a sharp clicking sound. Another sharp sound can be heard when the dart hits a wall. These sounds can be recorded with the microphone in the tablet. We can analyze the sound wave to determine the dart flight time and thus we can calculate the speed of the dart . Task 1: Determine the speed of a dart. Task 1 The experiment requires careful consideration of the measurement setup. It is useful to consider the following points: - How quiet should the room be? Can the measurement be done in a noisy classroom? - Do you need to consider the speed of sound? - What would be a suitable (optimal) distance for measurement? What happens if the distance is too short? Too long? - How long recordings do you need to make? What is the data size? Sound can be recorded on the Nexus tablets with the Sound Analyser PRO application. Start/Stop Restart Settings menu Before starting to record, open the settings menu and set the following options: Displayed graph: Colorplot Autoscale: checked FFT size: 1 Sound saving format: .wav Pinch the time axis to fit about 4 sec on display Note that the wav files contains 2 44100 samples per second. Try to record very short clips of no more than 2 seconds. Otherwise the files will be large and it becomes inconvenient to work with the files in Octave. As you can see in the oscilloscope plot shown above, the zero line is in the middle of the graph and the recorded signal can be either positive or negative (AC coupled signal). Sound data can be loaded into Octave with the wav function. This places the time values in t and the sound intensity into s, but directly plotting this data may be impossible due to the plot size limits in Octave-online. Plotting of longer sound recordings can be done with wavplot. We need to cut out the section of data that includes the shot and the dart hit sounds, removing noise features. Since we only need to look at the signal amplitude, we can take the absolute value of s. The data is noisy, but we are interested in the first sound of the shot and the hit. The noise can be mostly eliminated with the envelope function that traces the envelope moving maximum value of the sound wave. The sound edges can be seen more clearly after the envelope calculation. We can now select a discrimination level for converting the analog signal to a digital signal. A reasonable level would be 20000 for this dataset. After limits, we have a binary 0/1 signal. The time values of the zero-to-one transitions can be found with the edges edges function. Time differences can be found with the octave diff function and averaged diff with mean if necessary. If the flight distance is 4 m, the velocity would be 4/0.28397 = 14 m/s. edges Task 2: Repeat measurements for a blue dart and a red dart. Task 2 Is there a systematic difference in velocities? You will need to make several recordings of dart shots. Compare the flight times of a few shots to determine a suitable number of shots to calculate average velocities for the blue and red darts. You may use the octave std function to estimate the errors. std Task 3: Consider possible error sources. Task 3 Try to think of all reasonable sources of systematic or random errors in the dart velocity measurement. For example: - How accurately did you measure the flight distance? - How big effect does an error in distance have on the velocity? - How much does the distance vary shot-to-shot? - Is there a time delay between the gun sound and the flight of the dart? ... Suggest solutions to eliminate the errors that you can identify. For error sources that cannot be eliminated, how should you report their effect. Task 4: Consider in detail the possibility that there is a delay Task 4 between the sound of the gun and the actual flight of the dart. Consider how to use the dart velocity measurements to determine if there is a delay between the gun sound and dart flight. Measure dart flight time at several distances, plot the results and fit a straight line. The straight line fit gives a slope and a zero intercept. What is the physical meanings of the slope and zero inercept? How would you decide how many measurements to make at each distance? How would the shape of this graph be affected by air resistance? Air resistance would cause the dart to slow down over time. shot hit distance https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys07-18.svg Seminar 07 notes: Group 4 notes This set of experiments aims to practice the following techniques: Sensor analysis: Determining the barometer sensor characteristics. Sensor analysis Ratio measurements: Measuring a physical value by comparing to a reference. Ratio measurements Environment: Measuring environmental characteristics that are constantly varying. Environment Physiology: Measuring human body performance characteristics Physiology Indirect measurement: Practice repurposing a sensor, using a barometer as a Indirect measurement device to measure weight. wikimedia Pressure sensors are available in the Nexus 9 tablet and in the Galaxy phone. low pressure high pressure Air pressure varies with height. The barometer in the tablet is sufficiently sensitive to detect even small height changes. The pressure measurements can be used in a variety of ways. Task 1: Determine the characteristics Task 1 of the pressure sensor: resolution, noise level, response time, stability. Air consists of 78% of N2 (M=28), 2 21% O2 (M=32), and 1% Ar (M=40), 2 which gives a molar mass of 29 g/mol. At normal pressure, 1 mol of gas has a volume of 22.4 l, which we can view as a volume with a height of 1 m and a base area of 0.0224 m2. 2 The mass of 1 mol is 29 g, which creates a downward force (weight) of 0.28 N. The pressure difference over a height of 1 m is thus .28 N / .0224 m2 = 12.7 Pa = 0.13 hPa. The plot shows the variation of pressure when the tablet was moved up and down by 1 m every 30 seconds. The pressure change is close to 0.13 hPa as expected. The pressure sensor signal shows several interesting features. resolution of the data is limited by readout, perhaps in SensorKinetics. We can see from the plot that the smallest difference in the data is 0.01 hPa, which corresponds to a height difference of about 8 cm. The height of the tablet was changed quickly, in less than 1 s. The pressure reading changes much more slowly. There appears to be an exponential relaxation, probably due to data filtering. We can try to approximate the shape with p = p0 - A (1 - exp(-(t-t0)/)), where is the p p 0 A t-t0) - 0 sensor response time. We can see that the initial pressure p0 = 1007.54 hPa and the p 0 pressure change is 0.18 hPa. The signal change starts at 61.8 s. We can plot the response function shapes and look for a value that fits the measured data best. the sensor response time is 2 s. This means that when we measure a change of pressure, we should wait about 5 s for the reading to stabilize. Another problem is the long-term stability of the pressure measurement. Outdors, pressure changes with wind and weather. Indoors, pressure is affected by opening doors or windows, ventilation, air conditioning, etc. These variations correspond to an effective height change of about 50 cm. What is the appropriate time scale for performing pressure measurements. Try to estimate the minimum and maximum time limits that should be used when planning an experiment. Task 2: Measure the height of the KOMCEE building with the Task 2 help of the pressure sensor in the Nexus 9 tablet. Here is an example measurement in a taller building, going down the stairs from the 12th floor to the second basement level and back up. Slight background drift is visible The pressure change is about 5 hPa, which corresponds to 38 m. Measure the pressure change for a known height reference, such as a desk in the classroom and calculate the pressure change for a 1 m height change. Measure the pressure difference between the bottom and top floors of KOMCEE and calculate the building height. Task 3: Measure the highest and lowest air pressure Task 3 that you can generate with your lungs. Task 4: Construct a device to measure weight with Task 4 the pressure sensor in the Nexus 9 or Galaxy. A plastic storage box is provided with a coupling for a plastic tube. The tablet can be placed inside the box and operated with the Bluetooth mouse. Attach the plastic tube to the coupling on the box, start the pressure recording and try how high or low pressures you can generate by blowing air into the box or sucking air out of the box. Note that the box will leak even for fairly low air pressures. How can you solve this problem? You have already analyzed the response time of the sensor. Considering the response time, how long should you hold your breath during this experiment? Is it easier to maintain an overpressure in the box or reduced pressure? Why? Considering the box area of 25 cm 25 cm, how large is the force acting on the cover at an overpressure of 100 hPa? The plot shows a measurement of high and low pressures. In this test, the overpressure measurement is limited by air leak from the box. Can you improve the measurement with the available supplies? The minimum pressure measurement is easier because there is less leak. Try to measure minimum pressures for different people. You can ask other students to help. How much variation is there between people? The back of the plastic box is flexible. Placing a weight (M20 stainless nuts, 58g each) in the center of the back panel of the box will compress the air inside. As long as the air doesn't leak out, the pressure in the box is proportional to the weight. Adding nuts one-by-one should produce a plot similar to the one shown here. The number of nuts, n, on the box is shown for n each pressure step. For one or two nuts, the pressure change is monotonic, but for larger number of nuts, especially for n = 5, there is a clear n drop of pressure after a few seconds. This indicates that some air escaped from the box. Due to that, the pressure steps are not equal for adding or removing nuts. Finally, when just one nut is left, air leaks back into the box (at 180 s). Perform the weight measurement experiment. Try to avoid air leaks. Compare the pressure change caused by one 58g nut to the air pressure effect of a 1 m height change. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys07-19.svg Seminar 07 notes: Group 5 notes This set of experiments aims to practice the following techniques: Experiment planning: Planning ahead how much data is needed for an experiment. Experiment planning Data modeling: Using a mathematicl model to analyze physicl behavior. Data modeling Parametrizing: Learn how to choose parameters to describe a physical process. Parametrizing how many prameters are needed. What is the optimal number. Time analysis: Measuring the speed of sound from a round-trip time. Wave analysis: Compare the speed of transverse and longitudinal waves in a Wave analysis lossy medium (coil spring). We have seen how to use the accelerometer in the Nexus tablets to analyze the motion of the tablet itself. The accelerometer can also be used to analyze processes in objects attached to a tablet. In these experiments, the object is a long spring. Task 1: Measure vertical oscillations of the spring with and Task 1 without an extra weight attached to the tablet. Determine the oscillation period. Task 2: Fit the up-down oscillation data with an exponentially damped Task 2 cosine function to determine the damping factor. Find a suitable place in the KOMCEE building to hang the spring and tablet. The spring will stretch to a length of about 4 m when the extra weight is added to the tablet. Select accelerometer recording in SensorKinetics, start the recording, and stertch the spring by about 0.5 m. Let the spring go and let it swing for a few minutes. Make sure that you pull the spring straight down. Avoid sideways or twisting motion. After a few minutes, view the accelerometer data on the tablet screen. Determine the approximate oscillation period in seconds. Consider the data volume (file size). The spring can oscillate for a very long time, you may measure oscillations for 10 minutes or more. If you record data at the highest sampling rate, the data file will be too large to process conveniently in octave-online. If you record too few data points for each oscillation, you cannot perform data analysis. You should have approximately 10 data points per oscillation. (Refer to our earlier discussion on digital data and sampling). Select the appropriate sampling rate and record spring oscillations for at least 5 minutes. Be careful to take note of how much you stretched the spring. Make another measurement with a larger starting amplitude (stretch the spring more). Be careful not to stretch the spring so much that it will completely compress at the highest point in the oscillation. Attach the additional weight to the tablet holder (steel CF70 flange). Repeat the two measurements for two different initial stretching lengths. No extra weight Extra weight no stretch slight stretch large stretch no stretch slight stretch large stretch For the heavier load, the oscillations can last for a very long time. Here's an example measured for 5000 seconds. Each oscillation can be seen by zooming in on a narrower section of the data. The measured value is the acceleration. You already know how to calculate the position of the tablet by numeric integration. Do you need to integrate the data to analyze the damping behavior, assuming the oscillations follow a cosine function? Note that the measured value is the vertical acceleration. Therefore, the background level is approximately 9.8 m/s2. 2 For further analysis, it is advisable to use a ~100 second section of data from the beginning and a 100 second section from the end. determine the oscillation period by finding the background value at the beginning of the data (between 18 and 20 seconds), subtract the background value from the data and only keep the positive values with the max max function. We can then find the peak positions for each oscillation with findpeaks, convert the findpeaks found point indices to time values, and plot the acceleration data and the peak points. The peak point time values are in pt. From here we can get the time differences of neighboring points with diff and the average value with mean. diff mean octave:9> mean(diff(pt)) octave:9> ans = 3.3878 The average period is thus 3.39 s. An exponentially damped cosine function can be written as where y is the acceleration value, y0 is the background value and should be close to 9.8, y y 0 is the damping factor that we wish to find, t is the time and t0 is the starting time of t t 0 about 21 s, T is the oscillation period, which we found to be about 3.39 s, and A is the T A oscillation amplitude, and is the initial phase of the cosine function. We can use the damp function to calculate this function and compare our calculated damp model with the actual measurement data. For the first try, we can set We can see that the model doesn't quite fit our data: - the starting point t0 should be is less than 23 t 0 - the offset y0 is less than 9.8 y 0 - the period T is close to the correct value T - the amplitude A should be less than 6 A - the damping factor should be larger We can calculate a new fit, check the plot, and repeat the process until a good fit is found between the model and the experimental data. This is called data fitting and it is usually done with a nonlinear least squares algorithm, but here we fit the data manually and simply by comparing plots. Problems with the value of each parameter are easy to see by eye: Adjust each of the parameters of the damp function to get the best fit. These parameters seem to give a good fit for the first 100 seonds of the data. We can now see how well these parameters fit a later section of the data, at around 5 minutes after start The plot shows that the amplitude of the fit (red line) is too small compared to the data (blue line). This indicates that the damping factor is larger at the beginning of the spring oscillations and means that the energy loss is larger at the beginning of the experiment. Consider possible reasons why the tablet may be losing energy faster in the beginning. Suggest ways to modify the damping function to improve the fit between the model and the experiment. One possible option is to use a sum of several exponents. However, for our simple model, the damping factor describes the loss of mechanical energy to heat. If you improve the fit by adding parameters to the model, what is the physical interpretation of these additional parameters? For example, if you have two damping factors, 1 and 2. 1 2 Hint: if no clear physical process can be associated with additional parameters, it may be more appropriate to use a simpler model and add a larger error margin to the parameters that we obtain from the fitting process. damp https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys07-20.svg Seminar 07 notes: Group 5 continued Task 3: Measure the velocity of longitudinal Task 3 waves in the spring. Task 4: Measure the velocity of transverse Task 4 waves in the spring. Attach a tablet to the spring, stretch the spring until it doesn't sag much and launch longitudinal waves into the spring. Measure the spring length. Use the accelerometer to record bounces of the wave at the tablet. can see spikes in the accelerometer signal that correspond to the wave hitting the tablet. We would like to determine the average time difference between the spikes, marked T in the plot. We can do this by converting the T accelerometer signal into a digital form with the limits function. An appropriate limit value limits for this dataset is -0.3 m/s2, shown with the red line in the plot. limits the processing is similar to the NERF dart flight time measurement. The time points of the 0-to-1 transitions in the data can be found with the edges edges function. The time differences are calculated with diff and averaged with mean. diff mean edges The average round-trip time is thus 0.495 s. We can estimate the random errors by looking at individual time differences in D. We can see that the variations are on D the order of 0.001 s. The velocity of the wave can now be calculated if the length of the spring is known. Perform longitudinal wave velocity measurements for several spring lengths and plot the calculated velocities as a function of the spring length. Use a Web search to find which physical parameters of the spring determine the velocity of a longitudinal wave. The measurement and analysis are the same as for the longitudinal waves, but depending on how the tablet is held, the signal can be much noisier. Accelerometer spikes It is difficult to distinguish from this plot where the wave hits the tablt. The data can be made more clear by looking for rapid changes that would occur only when the wave reaches the tablet. A simple way is to calculate the differences between data points with diff. We can then diff try to use the same thresholding technique with limits as we did for the longitudinal limits waves. Note that the data length changes by one if you use the Octave diff function. You can diff use the difference function to avoid this difference problem. difference This plot shows the accelerometer spikes much more clearly. We can now set a threshold limit for extracting the time points for the spikes. You can see that the negative spikes are more clearly visible, so we can set the limits threshold at -0.15. limits The accelerometer spikes can now be clearly seen in the red plot. From here on, the round-trip time calculations is the same as before. We can see that the average round-trip time is 0.49 s, but the error are clearly larger than for the longitudinal wave measurement if you compare the values before averaging. The values vary between 0.42 s and 0.57 s. A reasonable way to report the result is that the round-trip time is 0.5 s. Measure the longitudinal wave round-trip time for several spring lengths. Does the round-trip time depend on the spring length? Does the velocity of the wave depend on the spring length? Method 1: Method 1 Method 2: Method 2 Let's look at another noisy example of acceleration spikes recorded in a transverse wave velocity measurement experiment. We look at the y-axis acceleration data. y The data is very noisy. It is difficult to see the individual acceleration spikes even by eye. To process this data, we need to remove the background offset. The background level can be calculated by averaging the signal over the 107 to 108 s section. This value can then be subtracted from the data and we plot the absolute value. The data is still very noisy, but we can remove most of the noise spikes by very slight averaging with the moving average mavg function. mavg mavg We can now start to see the acceleration spikes, but the simple thresholding and edge finding as we did earlier would not work well due to the large variations in the spike shapes. We can instead use the Octave findpeaks findpeaks function returns two vectors: pv holds the Findpeaks pv calculated peak values and pp holds the peak pp indices in ayv. The times corresponding to ayv the peak positions are picked from the time vector on line 12. The data used in the analysis starts at index m2, which is whie m2 the index is calculated as m2 + pp. m2 pp The time values of the peaks are: time differences are: The average and standard deviation are The round-trip time is thus 0.50 0.04 s. the Octave findpeaks function has several parameters that you can check with indpeaks help findpeaks in Octave. The MinPeakDistance parameter is set to 25, which means help findpeaks MinPeakDistance that valid peaks cannot be closer than 25 data points. You probably don't need to change this value. The MinPeakHeight parameter is the smallest peak amplitude that is detected. This value can be set to avoid detecting noise as a peak. Usual values are between 0.1 and 1. Using findpeaks is generally more reliable findpeaks than simple limits thresholding. Use this method to measure the wave velocity for several spring lengths. You should obtain a plot similar to the example shown here for spring lengths of 1 to 6 meters. Why is the plot linear? The data suggests that the wave round-trip time is always constant and does not depend on the length of the spring. Can you explain this result? https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys07-21.svg Seminar 07 notes: Group 6 notes This set of experiments aims to practice the following techniques: Indetifying patterns: Finding structure in unknown data . Indetifying patterns Feature extraction: Using a state machine to extrat features from data. Feature extraction Encoding: Learn about data encoding techniques. Encoding Magnetic recording: Learn how data can be recording in magnetic media . Magnetic recording Task 1: Use the provided wire winding tool and Task 1 observe magnetic fields close to the wire. Record the magnetic field signal by winding the wire over the magnetometer sensor in the tablet. Ferromagnetic materials can be permanently magnetized by a strong magnetic field. Magnetic data storage is possible by magnetizing parts of a magnetic recording medium, which is usually a tape or a disk (e.g., computer hard disk), but can also be a magnetic wire. The earliest magnetic recording devices used magnetic wire. We also use a magnetic wire in a magnetic recording experiment. A convenient choice is a coaxial cable with a steel shield, which happens to be a soft ferromagnet. Note: The magnetic message Note recorded in the wire can be very easily damaged by magnetic fields. Only use the Nexus 7 (2012) tablets. Do not bring the Nexus 9 or Nexus 7 (2013) tablets close to the wire. These tablets contain strong magnets and will therefore erase the message. Attach the wire winding mechanism to your desk with C-clamps. Note: Make sure that there are no metal objects close to the Note measurement area. Check for the location of metal legs under the desk surface. Set the tablet to record magnetometer data. Move the tablet around the center of the winding mechanism and find a position where you get the strongest signal when you wind the wire. Note: Bending the wire sharply damages the message. Note Choose a winding direction so as to minimize wire bending. good bad After determining the best tablet position, rewind the wire, start a new recording in SensorKinetics and wind the wire slowly until the end. Note: Later processing is much easier Note if you keep a constant slow winding constant speed of about 1 turn in 2 seconds. Plotting the measured data, you should look at the x-axis component (parallel to the wire). x The data should look similar to the example here. Task 2: Analyze the structure of the signal. Task 2 Try to identify characters and bits. Find which signal corresponds to 0 and 1 bits. Plot the signal for single characters. We know that the signal represents text. Text consists of characters. Each character is represented by a number (encoding) and the numbers are stored in binary form, using 0 and 1 bits. Start by looking at the signal details, discard unwanted start and end parts. Let's plot the beginning and end parts. We have some unwanted signal at the beginning up to about 4.4 s. After that, we see groups of spikes. These groups (5 to 12 seconds, 14 to 22 seconds) might be characters. The message thus contain 12 or 13 characters. It isn't quite clear where the mesage ends. There is a group of spikes at 95 to 102 s, and another group from 102 to 108. We can select the time range from 4.2 to 110 s for processing. We can now look more closely at the signal for a few characters. Each character group appears to contain 7 waves. Perhaps these are bits? We can see that there are two distinct types of waves, which we could call single and double waves. these two wave shapes probably correspond to binary zero and one. But which is which? looking at the first character, we have two choices: We don't yet know in which direction we should read the message. We thus have four possible ways of reading this character. left-to-right: 011 1001 or 100 0110 right-to-left: 100 1110 or 011 0001 Converting binary numbers to decimal: For the first character, we thus have 4 possible numbers. We write the numbers for the first three characters and the corresponding letters, assuming ASCII encoding: Assume ASCII code, which is the most commonly used choice. Find the ASCII encoding table on the Web. Perform the same analysis for your message and find which way the bits should be read and which wave types correspond to the zero and one bits. Try to read your message by plotting each character and writing down all possible readings, as was done above for the first three characters. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys07-22.svg Seminar 07 notes: Group 6 continued Task 3: Use Octave to read the message automatically. Task 3 Try to develop an Octave function to read the message. Our purpose is to automatically distinguish between the waves corresponding to zero and one bits, assemble the bits into numbers, and convert these numbers into text. The simplest way to approach this problem is to distinguish between the length of the to wave types. narrow wide Start by fixing the zero line (background) of the data. We use the same functions as for the accelerometer data. Looking at the raw measured data, we see that the back- ground remains between 1 and 7. The background has now been removed. Detecting oscillatory features may be easier by looking at a derivative signal. The derivative can be calculated with the difference function. difference Original signal Derivative Absolute value of derivative The sharp spikes in the absolute value plot can be mostly removed by slight averaging with moving average. We use the mavg mavg function with an averaging width of 5 points. The zero and one bits have now been converted into peaks with different widths. We can extract the widths with the limits limits function, as used by several other groups. difference mavg limits signal has now been turned into a square wave, where the peaks correspond to the original magnetic field spikes. We are interested in the widths of these peaks, so we need to find the time positions of both rising and falling edges. We have the edges function, but it only looks for rising edges edges. We can modify our data to convert all rising and falling edges into similar spike signals. We use the same trick as on line 9. rising edge falling edge We can now use the edges edges function to find the time points of each edge and calculate the time differences between neighboring edges with diff. diff edges We now have along vector of time differences and we need to start to divide the data into characters. Let's look what is in xd for the first character. The time differences list contains 14 values for one character as marked in the above plot. The "1" bits are marked green (short waves), the "0" bits are marked purple (long waves), nd the spaces between bits are marked black. We can see that the space at the end of the character is longer (1.68 s). We thus have 14 numbers for each character and we can rearrange this list into a matrix, where each column corresponds to one character. The length of the xd vector is 197, which means that we have data for 14 characters (1414 = 196). A vector can be reshaped into a matrix with the reshape command: We can see that in this data the short waves have a length of 0.35 to 0.45 s, while the long waves have a length of 0.67 to 0.83 s for the first character. We can try to set the limit between "long" and "short" waves at 0.58 s. The data can be converted to binary data simply by comparing against a constant value, such as 0.58. All smaller values are set to "1" while larger values are set to "0". We don't need those numbers, so we keep only the bit lengths. Increment index by 2 The vector now conatins only the bit lengths. This matrix cab be converted into decimal numbers by matrix multiplication with a vector that contains the powers of 2. The numbers can be converted into text with the char function. char We see that the first 2 letters "FI" look ok, but the rest of the message is garbled. The problem is that the boundary between "short" and "long" waves depends on the absolute width of each bit. If the winding speed of the wire was not constant during the measurement, there may be some variation in the bit lengths. We can view the data in xm that was calculated on line 22 and determine a suitable xm boundary value. These values can be put in a vector xl. proper boundary values become smaller towards the end. This means that when the data was measured, the wire winding speed gradually increased towards the end. We repeat the calculation with these limits. Success! The message could be read with Octave. Consider problems here. Our reading was not automatic because our processing cannot properly handle reading speed changes. We can improve the reliability of the readout by looking not only the width but also the structure of each bit. Method 1: peak width Method 1 Method 2: counting peaks Method 2 The second method makes use of zero-crossing analysis. This is a common signal processing technique. The purpose is to find the points where the signal crosses zero. zero-crossing points The zero-crossing points are marked with red circles in the plot. We can see that it is possible to distinguish the zero and one bits according to the number of times the signal crosses zero (2 or 4 times). This is a more reliable approach than looking at the peak width, because the number of zero crossings doesn't depend on the reading speed. The signal processing start the same way as before up to the averaged signal calculation. data Remove noisy start and end parts Correct the background Calculate averaged difference signal We can now find the edges of each bit using the Octave zerocrossing function. zerocrossing The threshold is set by subtracting 2 from the data befor finding zero-crossing points. The plot shows the averaged data after subtrating 2 and the zero-crossing points found by Octave. The values in be give the start and end times be for each bit. Compare the first few be values be with the plot. The next task is to look at the individual waves inside each bit. The zero crossings are calculated for the background-corrected data in fx with a thershold of -2. We now know where each bit is in the data and we can count the number of zero- crossings within each bit. The zero-crossing points can be seen more clearly by zooming in on the first two bits and comparing the xave and fx plots. xave fx blue dots mark the bit start and end points and the red dots mark the zero crossings within each bit. We see that the bit at 4.5 to 5 s has to red zero-crossing points, while the second bit at 6 s has 4 red zero-crossing points. If we count the number of zero-crossing points within each bit, we can distinguish between the two bit shapes. We build vector c that holds the number of c zero crossings for each bit. Initialize an empty vector c c Bit index counter Loop over bit edges select start edges The loop index i takes values 1, 3, 5, 7, 9, ... i When i = 1, the bit start time is in be(i) and the bit end time is in be(i+1). i be(i) be(i+1) The values are shown above in the list on Octave line 13. The loop uses boolean comparison to find all values of zc that are within the zc time range set by be(i) and be(i+1). Look at the red dots in the last plot. be(i) be(i+1) shows that there are two zero crossing points (indices 1 and 2). Try the same thing for the second bit The second bit contains 4 zero crossings. We don't need the actual indices, we just need to know how many zero crossings there are. This can be done with length. length The second bit contains 4 zero crossings Looking at the contents of vector c, we see the zero-crossing counts for the c first character (bits 1 to 7). We can now decide that any bits with fewer than 3 zero crossings is a "1", others are "0". This shows that the bits of the first character are 1000110, which equals 70 in decimal and corresponds to the letter "F" in ASCII encoding. Check how many bits we have and reshape the bits into a matrix with one character in each column. We have 98 bits, which means we have 14 characters (7 * 14 = 98 bits) Conversion of an array of bits to decimal numbers and characters can be done the same way as we did before, by matrix multiplication The last character is not valid, so our message length is just 13 characters. Note that this method of decoding the message is not sensitive to small changes in the wire winding speed. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys07-23.svg Seminar 07 notes: Group 6 continued Task 4: Erase the old message from the wire and Task 4 write a new message. An existing magnetic field pattern in the wire can be erased by magnetizing the wire uniformly in one direction. 1. Rewind the wire to one end 2. Take the large Neodymium magnet and place close to the wire as ashown in the picture. The green dot should be just under the wire. 3. Wind the wire slowly over the magnet to the other end. 4. Remove the large magnet, and put it far away from the wire. 5. Rewind the wire to the beginning. Decide which message you wish to write. The wire is long enough to hold up to about 14 characters if encoded in 7-bit ASCII. For this experiment, just one or two characters is sufficient. Write down the characters and the ASCII codes corresponding to those characters. Convert the decimal numbers to binary codes. For most letters, it is sufficient to use the 7 least-significant bits. Write down the binary bit pattern that you wish to write. For example, if you wish to write "STENDEC", the codes are: Take the provided small Neodymium magnet in a plastic handle. Note that there are "0" and "1" written on two sides of the handle. letter decimal binary The message in binary (keeping 7 bits). Wind the wire a few turns without bringing magnets close to the wire. Hold the small magnet far from the wire. Turn the plastic handle so that either "0" or "1" points up, depending on which bit you wish to write. Move the magnet until it touches the wire as shown in the picture. Move the magnet away from the wire. Rotate the handle of the wire spool by one turn. Write the next bit. Between characters, rotate the spool handle by two turns to make it easire to distinguish later where the bits are. Rewind the wire and try to read the bits with the tablet. You should be able to see a similar signal as in Task 1. Can you read your message with Octave? https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys07-24.svg Seminar 07 notes: Octave: element-by-element operations Performing calculations with arrays (vectors) is very convenient in Octave since we don't need to write functions to process each element, for example Care must be taken when performing calculations between vectors: error: operator nonconformant arguments This error occurs because we cannot perform a vector multiplication. Often we don't need a vector multiplication, but want to operate on each element of the vector independently. This can be done with the dot operators. same kind of dot operators apply for powers. Octave gives a warning if you forget. error: for x^A, A must be a square matrix. Use .^ for elementwise power. The element operation is not needed for the + operator since there is no confusion over how the operation should be carried out. Built-in functions like sqrt can operate sqrt on vectors directly. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys07-25.svg Seminar 07 notes: Octave: linear fitting It is often necessary to fit a straight line through noisy data points. We can set up a set of (x, y) data and plot it with circles x y An easy way to fit a line through this data is to use the polynomial fitting function polyfit. We want a straight line, which is the polyfit first-order polynomial: These two numbers are the coefficients of a This linear plot can be added to the plot by holding the previous plot data https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys07-26.svg Seminar 07 notes: Experiment preparation Look over the experiment supplies for your group. Read carefully the task descriptions for your group. Ask questions if something is not clear. You will have three weeks to do the experiments and to prepare a presentation (6/20, 6/27, 7/4). Note that it is more important to understand the process of handling experimental data, rather than finishing all tasks. Our purpose is to think about how emprical observation of nature (measurement) is converted into knowledge that can be shared and trusted. Keep note of this when preparing or performing experiments. You should discuss the method, difficulties, problems, reliability, etc. in the final presentation, not just the measurement result. Homework 1. Read through the experiment task descriptions 2. Ask questions on slack to prepare for next week's experiments. Time is limited, so when unsure what or how to do, ask! https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys08-01.svg Seminar 08 notes: Seeing the world through sensors: objective measurements and subjective understanding Institute for Solid State Physics Mikk Lippmaa Wednesdays, 10:25 - 12:10 KOMCEE K302 20.06.2018 8 https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys08-02.svg Seminar 08 notes: Group 1 progress The single line and area mapping experiments have been finished, raw data appears reasonable. There are some bumps int the data (10 to 12 s) that may cause some difficulties with data processing. The area mapping experiment data looks ok. The magnetic field sensor is not in the exact center of the mapping are, but that isn't a problem for the mapping expriment. Practice using the boxave function. You will need to adjust the parameters, so check the homepage for details. Try to prepare vector plots of the linear data today. Try to process the grid data with boxave. boxave This will show if the data is clean enough or you have to repeat the measurement. If the grid data processing works well, try to prepare the x, y, and z component x y z 3D surface plots. You have several tasks that you can distribute between all three group members. Try to share the work equally. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys08-03.svg Seminar 08 notes: Group 2 progress You started with simultaneous recording of acceleration and rotation rate data. There was a problem with the Octave hold command. hold Octave usually creates a new plot every time a plot command is used. If you wish to put several plots together in a single panel, you can use hold: You thus get two separate plots. plot(t,y1) plot(t,y2) The two plots now appear together. Confusion may happen with hold because hold it will remain in effect until released. You can avoid confusion with expicit hold on / hold off commands hold on hold off Or, you can combine two plots in one command if the data is available The script for processiing the acceleration data and calculating the ditance of the sensor from the rotation center is working. You can generate the legend text automatically The final processing script from slack: Consider how many measurements you need to make at each tablet positon. Make sevarl measurements, calculate the sensor distance and look at the spread of values. You can improve precision by averaging, but you need to look how many measurements should be averaged. With too few measurements, the result will be inaccurate, but with too many measurements, you just waste time. Which tablet positions should you use to get both x and y positions? x y How accurate is the result? How can you estimate accuracy? https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys08-04.svg Seminar 08 notes: Group 3 progress You measured the sound from 50 shots for the blue and red darts. You can load the sound waves into Octave with the wav function, but plotting wav the data may be difficult due to plot size limits in Octave-online. wavplot It will reduce the plot size but tries to keep the shape of the wave unchanged. Details are on seminar home page. shot and hit sounds are at 1.5s and 1.8 s. We can cut out that section for analysis and easier plotting. Cut data from 1.4 to 2 s: Note the use of a smaller decimating parameter 8 for wavplot. The default wavplot value is 256. use the smallest value that allows Octave to produce a plot. This minimizes signal shape distortion in a plot. It doesn't affect the data processing. You have enough data. Process the flight times and calculate dart velocities. You can try to estimate the error bars when reporting the velocity. You have very many measurements, so you could calculate the standard deviation with the Octave std function. std https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys08-05.svg Seminar 08 notes: Group 4 progress First, understand your sensor performance The precision of data readout from the sensor is limited by software? You can clearly see individual steps in the data. 1. How large are these steps? Is this SensorKinetics/Android data storage precision problem or is this a bit-level sensor readout problem, i.e., limited sensor resolution? 2. How large is the corresponding effecive height change for air pressure? The smallest pressure step corresponds to a height difference of 0.089 m. The plot shows the time response behavior of the sensor. Try to change plot scales to determine if this is an exponential response. Hint: To handle exponential data, remove background and plot data on a logarithmic scale or take a log of the data and plot that. 3. What does an exponential function look like when plotted on a logarithmic scale? 4. Continue with the building height measurement. The pressure step size is 0.01 hPa. This seems to be Android or SensorKinetics limit, it is not a binary step. Subtracted background, changed sign. Take a natural logarithm of the data, plot points, fit line. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys08-06.svg Seminar 08 notes: Group 5 progress Look at the long oscillation. Try to plot beginning part and attempt to fit data Use the dampl function and find best parameters. Can you improve the calculation? Also look at the x and z components. There is soe signal there. Why? During oscillation, energy is lost. Can you guess where is the energy going? Plot traveling wave patterns. Look at the dart shooting group for analysis ideas. Found a good place for the spring oscillation measurement. Longitudinal wave round-trip measurement Accelerometer data Thresholded data Damping measurement. Fitting an exponentially damped cosine to the data First try Fit is improving Perpendicular traveling wave measurement. Measured data Thresholding https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys08-07.svg Seminar 08 notes: Group 6 progress You could get the first reading. 1 Plot single characters. 2 Try to guess how to read characters. 3 Which wave shape corresponds to '0', which wave shape is a '1'. Plot these waves shapes. 4. Cut a single character from the data 5. Try to use Octave functions to extract binary data. Plotting single characters Each bit can be clearly seen in this data. Found two bit shapes Which bit shape corresponds to "0" and which one corresponds to a "1"? Start by looking at interesting characters like this one, where almost all bits are the same. Just one different bit. Another strange character is here. There is an unusual field shape in the beginning of the data. Is this a reading problem or recorded data problem? Consider this character. If single wave = "0", double wave = "1": left-to-right: 1001101 right-to-left: 1011001 If single wave = "1", double wave = "0": left-to-right: 0110010 right-to-left: 0100110 Looks like single="0", double="1", right-to-left is the correct reading. this letter should be "Y". But there are problems too. This character has only 6 bits. This is probably a recording mistake! The character with just one different bit that was shown earlier would have a character code 0100000. What letter is this? https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys08-08.svg Seminar 08 notes: Homework 1. Make sure that all Octave scripts are uploaded to Slack 2. Make sure that all useful plots are copied to Slack 3. Start preparing the presentation slides by copying results to your group's Slides file. 4. Most groups have their measurement data. When you run into problems during processing, post questions to Slack. I will respond on slack. 5. Next week is the final chance to do measurements. Plan carefully which measurements you still need to do next week. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys09-01.svg Seminar 09 notes: Seeing the world through sensors: objective measurements and subjective understanding Institute for Solid State Physics Mikk Lippmaa Wednesdays, 10:25 - 12:10 KOMCEE K302 27.06.2018 9 https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys09-02.svg Seminar 09 notes: Octave notes Look closely at Octave messages. Errors and warnings are important! For example, when uploading a file claps.csv, which has a size of 888 kBytes claps.csv there are 9 warnings: Uploading claps.csv: If your file is large and causes you to exceed your space limit (51200 KiB), the file may be incomplete. Note that this is not an error. Octave-online is just warning you that there is a maximum file size limit of 50 MBytes. You can find your disk usage with the disk command. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys09-03.svg Seminar 09 notes: Miving average The raw data from a measurement can sometimes be quite noisy. How can we make the data easier to process? First question: can you improve the measurement? If that is not possible, we need to filter the data, but note that this changes the data and may affect our analysis result in subtle ways. One way to reduce noise is to use averaging. When the measurement samplng rate is much higher that required for analysis, simple averaging can be used. The original data is divided into block and an average value is calculated for each block. For example, averaging with a block size of 4 is illustrated here: The averaged data (red) is smoother and the number of data points is reduced. It is sometimes easier to use averaging that doesn't change the number of data points and affects sharp transitions less than simple block averaging. This technique is knonw as "moving average". As an example, we can look at the accelerometer data from Group 2. Calculate and plot the averaged data can see that the averaged data is smoother and the data length has been reduced 20 times, from 2080 data points in the original data to just 104 points in the averaged data. point p p Width: +/- w The moving average is calculated for each data point. For point p, calculate an average p over neighbors in the range +/- w. w The number of data points doesn't change. data moving average average The difference between block average and moving average can be seen in the processing of the magnetic field data from Group 6. The moving average (red) eliminates the unwanted "fast" features but still follows the data reasonably well, unlike the block average for the same averaging width (11 points). The moving average function is mavg. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys09-04.svg Seminar 09 notes: Group 1 progress The linear mapping has been finished and vector plots are ready, The 2D mapping worked ok. The magnetc field points have been found (with some manual correction of points where the data was noisy). The magnetic field y-axis component y The 3D surface plots are ready as well. x y z The plots are nice, but why is there an upturn in the y-axis 3d data? y Please copy to Slack all macros used for generating these plots. The 3d vector plot shows nicely how the magnetic field strength and direction vary in the measurement plane. The top-view projection shows that the tablet sensor was a little bit off frm the center of the mapping area The boxave function worked correctly boxave for most data points but there were some problems, probably because the magnet was touched when it should have not moved. This can confuse the boxave algorithm. boxave You still have one more task, to try magnetic navigation in a plane. Try to put the tablet back in the same place where it was during the 2D mapping. (I hope you documented the measurement properly!) 1. Move the little magnet to a few positions (red dots) and measure the field readings. 2. Extract magnetic field components for each position. 3. Can you use your earlier mapping result to determine the positions of the red dots? How would you do this? https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys09-05.svg Seminar 09 notes: Group 2 progress There are several tablet measurements for several rotation distances. Macro files for each measurement: The results seem to be reasonable. Try to finish measurements today and post the processing results on Slack (processed data plots and measured sensor distance from rotation center for each measurement). Look at the linear fitting instructions in Seminar 7 notes. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys09-06.svg Seminar 09 notes: Group 3 progress Very nice macro to process 50 measurements. Plot some typical sound plots and at least for one measurement, add plots for all calculation steps to slack. On line 12, you could do v(i) = ... to collect the velocity values in vector. You can then plot them nicely. Look more closely at possible error sources. List up as many as you can think of. Example: The gun sound is quite long. You now used the first sound edge for velocity calculation. Do you know when the dart actually started moving relative to the gun sound? How to measure the delay? https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys09-07.svg Seminar 09 notes: Group 4 progress Started with the analysis of the sensor resolution and time constant. resolution Time constant. Assume exponential behavior, extract characteristic time by log plotting and linear fitting This shows the approximate response curve Similar result with linear fitting Today, try to measure the building height. If you have time, measure lung pressure (minimum and maximum) Building height measurement. Calibrate against th height of a desk, 71 cm. Measured air pressure at the basement level and on the 5th floor. Estimated building height is 21 meters. Plot of a floor-to-floor air pressure variation. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys09-08.svg Seminar 09 notes: Group 5 progress You have divided the tasks between all group members very nicely: Damped oscillation fit, longitudinal wave, transverse wave Try to get the final number values today for all measurements. Consider what aspects of the measurement may affect the accuracy of the results. Try to plot dampled oscillation fits for the beginning and end parts of the oscillation. If you have time, you can make short measurements (about 300 seconds) of oscillations with different initial pull distance (larger amplitude). Damping factor is different! Middle part fit Much better fit with a lower damping factor: .002 not .003 The measurement shows that the damping factor (energy loss) is higher in the beginning of the experiment. Why might that happen? https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys09-09.svg Seminar 09 notes: Group 6 progress You could find the encoding details and read the message by visually analyzing the measured magnetic field profiles. One character Single bits. There were recording mistakes with some bits. Please re-read the new recording. 1. Read the full message by eye today 2. Try to follow processing steps shown in seminar 7 handout https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys10-01.svg Seminar 10 notes: Seeing the world through sensors: objective measurements and subjective understanding Institute for Solid State Physics Mikk Lippmaa Wednesdays, 10:25 - 12:10 KOMCEE K302 04.07.2018 10 https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys10-02.svg Seminar 10 notes: Presentation preparation Presentation time: 5 to 8 minutes + discussion (total 10 min/group) Consider 1 to 2 minutes / slide, prepare 4 to 8 pages. For example: 1 Title, group member names 2 Purpose, experiment description 3 Explain how you actually did the experiment Was there any difference from your original plan, why? Was there any difference from your original plan, why? 4 Show the original unprocessed data that you obtained 5 Explain data processing steps and the result Main points only, we can look at you group's Slack channel for details 6 What was difficult in the experiment or processing? How reliable is your result? What limits the accuracy of the result? You can adjust the length of different sections, depending on your work. NOTE: Our purpose is to learn how the scientific process works. NOTE: Explaining what and how you did is more important than the result. Today: - Prepare presentation Homework 1. Work together with your group members to finsih the presentation slides. 2. Decide how you share the presentation between group members. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys11-01.svg Seminar 11 notes: Seeing the world through sensors: objective measurements and subjective understanding Institute for Solid State Physics Mikk Lippmaa Wednesdays, 10:25 - 12:10 KOMCEE K302 11.07.2018 11 https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys11-02.svg Seminar 11 notes: Seminar presentations 10:45 - 10:55 Determining the accelerometer position in a Nexus 7 tablet by rotary motion. 11:00 - 11:10 Measuring the speed of a NERF dart by analyzing recorded sound. 11:30 - 11:40 Modeling the exponential decay of a damped harmonic oscillator. 10:30 - 10:40 Mapping the field distribution of a magnetic dipole. 11:15 - 11:25 Characteristics of the MEMS barometer in a Nexus 9 tablet. 11:45 - 11:55 Using a Hall magnetometer to recover a message from a wire recording. 11:55 - Closing 5-8 min presentation, 10 min total https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys11-03.svg Seminar 11 notes: Group 1 presentation https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys11-04.svg Seminar 11 notes: Group 2 presentation https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys11-05.svg Seminar 11 notes: Group 3 presentation https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys11-06.svg Seminar 11 notes: Group 4 presentation https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys11-07.svg Seminar 11 notes: Group 5 presentation https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys11-08.svg Seminar 11 notes: Group 6 presentation https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys11-09.svg Seminar 11 notes: Where we started How science works ideas concepts observations facts reliable knowledge about nature scientific method Reliable = most likely to be true L How to make reliable observations of nature? Remember the first task of measuring the length of a piece of plastic. Our results varied by almost a centimeter! Even a simple measurement can be surprisingly difficult. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys11-10.svg Seminar 11 notes: What we learned Direct observation through our senses. Very few things can be measured quantitatively by our senses. We use our hearing, vision, smell, taste, touch, etc., but not everything can be directly sensed by humans. That's why we need physical sensors - objective (not subjective like a human) - reproducible (anybody can measure the same thing and get the same result) - limitations (must know how to use a sensor and how to convert a sensor output into useful information) We need empirical (observe nature) evidence to decide what is true in the scientific sense, but where can we get empirical evidence and how? We saw that there are many steps in obtaining empirical evidence, converting observation of nature into knowledge, and sharing the results of an observation with others in an objective manner. 1. Planning a measurement What do you want to measure (observe)? How are you going to measure it? 2. 'Raw' measurement How do you control the measurement environment? How do you control the subject of the measurement? What kind of data do you collect? 3. Processing results What information can be extracted from raw data? How can you extract useful information? How do you estimate the reliability of the result? 4. Communication How can you explain the measurement and the result? How do you document the results so that other may reproduce your experiment? https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys11-11.svg Seminar 11 notes: Why these experiments? Accelerometer position: Measurement planning: - Need to understand the physics of rotational motion before you can do the measurement - Several ways to do the measurement, must select one method - Relatively noisy data, learn how to extract accurate numbers Dart velocity: Indirect measurement: - Use sound to measure velocity - Data extrapolation to extract 'hidden' parameters (gun sound delay) - Easy to repeat experiment, can do statistics for error analysis Magnetic field mapping: Data visualization: - Visualizing vector data - Visualizing three-dimensional data - Using time sequencing to collect data - importance of careful experiments - documenting all details (to go back to erlier data) The experiments that we did in this seminar were selected to demonstrate various aspects of collecting empirical evidence and processing observation data: Air pressure: Sensor characterization: - How to recognize sensor limitations - Measuring time-variable quantities - Selecting an optimal time scale depending on what we want to see (weather or height change) - Envirnomental measurements, not so clear what is signal, what is noise, and what is the sensor response Magnetic message Data recovery: - Extracting information from a complex signal - Automated pattern recognition - Signal integrity (missing letters) - Signal validity (wrong letters) Spring oscillation: Data fitting: - Extracting physical parameters by fitting data - How to select a model (subjective element) - How to interpret data (meaning of different parameters) - How to use (role of) models that are not perfect - Measuring materials properties (internal friction) https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys11-12.svg Seminar 11 notes: Scientific process What did you like? What did you hate? Many aspects of science science is a group effort Experiment Analysis Theory, Models Presentation Which is your favorite? You could pratice various parts of the scientific process of creating knowledge. https://lippmaa.issp.u-tokyo.ac.jp/lecture/Notes/fys11-13.svg Seminar 11 notes: License Creative Commons BY-NC-SA https://lippmaa.issp.u-tokyo.ac.jp/lecture/seminar https://lippmaa.issp.u-tokyo.ac.jp/lecture/Syllabus.svg Syllabus figure Illustration of the role of sensors in observing natural phenomena https://lippmaa.issp.u-tokyo.ac.jp/lecture/Moves.png Sample move data Path shapes used in the accelerometer measurements https://lippmaa.issp.u-tokyo.ac.jp/lecture/kals.jpg KALS classroom view https://lippmaa.issp.u-tokyo.ac.jp/lecture/exp2.jpg supplies Experiment supplies provided fo the seminar https://lippmaa.issp.u-tokyo.ac.jp/lecture/tablets.jpg tablets Experiments are done with the help of various Nexus tablets https://lippmaa.issp.u-tokyo.ac.jp/lecture/view2.jpg experiment 1 Reading a magnetic message https://lippmaa.issp.u-tokyo.ac.jp/lecture/view3.jpg experiment 2 Acceleration measurement https://lippmaa.issp.u-tokyo.ac.jp/lecture/view4.jpg experiment 3 Dart speed measurement https://lippmaa.issp.u-tokyo.ac.jp/lecture/seminar.jpg classroom view Seminar classroom in the KOMCEE building https://lippmaa.issp.u-tokyo.ac.jp/lecture/demo.jpg demonstrations Sensor demonstrations on an oscilloscope https://lippmaa.issp.u-tokyo.ac.jp/lecture/tools.jpg tools Experiment supplies https://lippmaa.issp.u-tokyo.ac.jp/lecture/K301-group1-poster1.png Group 1 poster Summary of sensor location search experiments https://lippmaa.issp.u-tokyo.ac.jp/lecture/K301-group2-poster1.png Group 2 poster Summary of magnetic field mapping experiments https://lippmaa.issp.u-tokyo.ac.jp/lecture/K301-group3-poster1.png Group 3 poster Summary of gait analysis https://lippmaa.issp.u-tokyo.ac.jp/lecture/ https://lippmaa.issp.u-tokyo.ac.jp/lecture/tmo https://lippmaa.issp.u-tokyo.ac.jp/links https://lippmaa.issp.u-tokyo.ac.jp/link/issp.jpg Institute for Solid State Physics https://lippmaa.issp.u-tokyo.ac.jp/link/kiban.jpg Dept. of Advanced Mat. Sci. School of Frontier Sciences https://lippmaa.issp.u-tokyo.ac.jp/link/u-tokyo.jpg University of Tokyo https://lippmaa.issp.u-tokyo.ac.jp/link/ut-mate.png UT-mate https://lippmaa.issp.u-tokyo.ac.jp/link/utask-web.png UTask-web https://lippmaa.issp.u-tokyo.ac.jp/link/ISSPonline.png ISSP online journals https://lippmaa.issp.u-tokyo.ac.jp/link/Pubs.png ISSP publications database https://lippmaa.issp.u-tokyo.ac.jp/link/L-Matsumoto.jpg Y. Matsumoto, Tohoku Univ. https://lippmaa.issp.u-tokyo.ac.jp/link/L-Kawasaki.jpg M. Kawasaki, U. Tokyo https://lippmaa.issp.u-tokyo.ac.jp/link/L-Koinuma.jpg H. Koinuma, NIMS https://lippmaa.issp.u-tokyo.ac.jp/link/L-Oshima.jpg M. Oshima, U. Tokyo https://lippmaa.issp.u-tokyo.ac.jp/link/L-Kumigashira.jpg H. Kumigashira, KEK https://lippmaa.issp.u-tokyo.ac.jp/link/L-Ohkubo.jpg I. Ohkubo, NIMS https://lippmaa.issp.u-tokyo.ac.jp/link/L-Harada.jpg T, Harada, Tohoku Univ. https://lippmaa.issp.u-tokyo.ac.jp/link/L-Hasegawa.jpg T. Hasegawa, U. Tokyo https://lippmaa.issp.u-tokyo.ac.jp/link/L-Hwang.jpg H. Y. Hwang, Stanford https://lippmaa.issp.u-tokyo.ac.jp/link/L-Takeuchi.jpg I. Takeuchi, Maryland https://lippmaa.issp.u-tokyo.ac.jp/link/L-Tybell.jpg T. Tybell, NTNU https://lippmaa.issp.u-tokyo.ac.jp/nano/ https://lippmaa.issp.u-tokyo.ac.jp/people https://lippmaa.issp.u-tokyo.ac.jp/project https://lippmaa.issp.u-tokyo.ac.jp/nav1.svg Sitemap-1 Sitemap for projects https://lippmaa.issp.u-tokyo.ac.jp/pld.jpg PLD plume Oxide thin films grown by Pulsed Laser Deposition https://lippmaa.issp.u-tokyo.ac.jp/step3d.png SrTiO3 surface AFM image of an annealed step-and-terrace SrTiO3 substrate surface https://lippmaa.issp.u-tokyo.ac.jp/rpif.png Ruddlesden-Popper interface Two dimensional electronic states at oxide heterointerfaces https://lippmaa.issp.u-tokyo.ac.jp/RingsFC.png Oxide nanorings Templated oxide nanostructures https://lippmaa.issp.u-tokyo.ac.jp/publications https://lippmaa.issp.u-tokyo.ac.jp/rheed/ https://lippmaa.issp.u-tokyo.ac.jp/rheed/oneml.png One monolayer growth Surface model and intensity variation during the growth of one SrTiO3 unit cell layer. The green arrow illustrates the electron beam. https://lippmaa.issp.u-tokyo.ac.jp/rheed/islanim.gif Island growth Island growth https://lippmaa.issp.u-tokyo.ac.jp/rheed/lblanim.gif Layer-by-layer growth Layer-by-layer growth https://lippmaa.issp.u-tokyo.ac.jp/rheed/sfanim.gif Step-flow growth Step-flow growth https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/ https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/s52s.png Etched SrTiO3 steps As-supplied wet-etched SrTiO3 AFM image https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/cai117s.png caiciss Temperature dependence of Sr segregation. CAICISS TOF peaks: Sr (green), Ti (red), O (blue). https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/rt25.png 25 nm STM SrTiO3 STM, 25 × 25 nm2 https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/rt100ag.png 100 nm STM SrTiO3 STM, 100 × 100 nm2 https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/rt200ag.png 200 nm STM SrTiO3 STM, 200 × 200 nm2 https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/60007.png 100 nm STM SrTiO3 STM at 600°C, 100 × 100 nm2 https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/635.png 110 nm STM SrTiO3 STM at 635°C, 110 × 110 nm2 https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/700.png STM at 700C SrTiO3 STM at 700°C, 110 × 110 nm2 https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/sto1.gif 635C STM animation 635°C, image size 110 nm × 30 nm, 37 sec/frame timelapse avi(562 Kb) mov(113 Kb). https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/sto1.avi https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/sto1.gif SrTiO3 635°C annealing Time lapse of STM images taken during the annealing of a SrTiO3 surface at 635°C https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/sto1.mov https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/sto1.gif SrTiO3 635°C annealing Time lapse of STM images taken during the annealing of a SrTiO3 surface at 635°C https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/sto10.gif 707C stm animation 707°C, image size 90 nm × 90 nm, 37 sec/frame timelapse avi(5.3 Mb) mov(823 Kb). https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/sto10.avi https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/sto10.gif SrTiO3 707°C annealing Time lapse of STM images taken during the annealing of a SrTiO3 surface at 707°C https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/sto10.mov https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/sto10.gif SrTiO3 707°C annealing Time lapse of STM images taken during the annealing of a SrTiO3 surface at 707°C https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/sto12.gif 721C stm animation 721°C, image size 60 nm × 75 nm, 37 sec/frame timelapse avi(1.5 Mb) mov(274 Kb). https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/sto12.avi https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/sto12.gif SrTiO3 721°C annealing Time lapse of STM images taken during the annealing of a SrTiO3 surface at 721°C https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/sto12.mov https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/sto12.gif SrTiO3 721°C annealing Time lapse of STM images taken during the annealing of a SrTiO3 surface at 721°C https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/sto21.gif 784C stm animation 784°C, image size 230 nm × 175 nm, 90 sec/frame timelapse avi(1.1 Mb) mov(154 Kb). https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/sto21.avi https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/sto21.gif SrTiO3 784°C annealing Time lapse of STM images taken during the annealing of a SrTiO3 surface at 784°C https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/sto21.mov https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/sto21.gif SrTiO3 784°C annealing Time lapse of STM images taken during the annealing of a SrTiO3 surface at 784°C https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/sto26.gif 800C stm animation 800°C, image size 28 nm × 16 nm, 10 sec/frame timelapse avi(1.7 Mb) mov(287 Kb). https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/sto26.avi https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/sto26.gif SrTiO3 800°C annealing Time lapse of STM images taken during the annealing of a SrTiO3 surface at 800°C https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/sto26.mov https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/sto26.gif SrTiO3 800°C annealing Time lapse of STM images taken during the annealing of a SrTiO3 surface at 800°C https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/800.png STM at 800C SrTiO3 STM at 800°C, 240 × 240 nm2 https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/atomic https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/s900.png SrTiO3 terraces STM after 900°C anneal, 300 × 300 nm2. https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/s1000.png Flat terrace STM after 1000°C, imaged at 600°C, 300 nm × 180 nm. https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/a800.png SrTiO3 bumps STM after 10-minute 900°C anneal, 10 × 10 nm2. https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/ao800.png 800C 10 nm × 5 nm, 800°C. https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/sec1000.png Cross section of a terrace Cross section of a terrace https://lippmaa.issp.u-tokyo.ac.jp/sto/anneal/ordered.png SrTiO3 2x2 surface Ordered 2×2 surface https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/edge https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/simpledge.png Model edge configurations Possible edge cell configurations https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/pits00.png starting image Initial state of SrTiO3 etch pit simulation for zero-angle miscut. https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/pits09.png final image Final state of SrTiO3 etch pit simulation for zero-angle miscut. https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/pits00.mov https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/pits00.png Etch pit shape simulation Etching simulation for a substrate where the miscut is along the a or b axis https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/pits30.png starting image Initial state of SrTiO3 etch pit simulation for random miscut direction. https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/pits39.png final image Final state of SrTiO3 etch pit simulation for random miscut direction. https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/pits30.mov https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/pits30.png Etch pit shape simulation Etching simulation for a substrate where the miscut is along a random direction https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/ https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/etcht.png in situ etching AFM images of SrTiO3 substrates: (a) polished and BHF etched for (b) 5 min., (c) 8 min., (d) 12 min.1000 × 1000 nm2. https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/speed.png Etching rate plot Step edge etching rate for several pH values. https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/etchingm.png Step-flow etching model Step-flow etching of SrTiO3, forming a step-and-terrace surface terminated by a perfect TiO2 layer. https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/par https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/pits https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/Pits.png Pit shape simulations Simulated etch pit shapes for various etching parameters https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/pit30.png Starting pit shape starting image for pit shape simulations https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/speed2.png Step edge speed Step edge movement speed https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/speed3.png Kink site effect Comparing the effect of kink site etching rate change https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/speed4.png Fast kink-site etching Step edge motion for very fast kink-site etching rates https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/smallpit.gif Small pit etching animation Step-flow etching of a small pit https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/bigpit.gif Large pit etching animation Step-flow advance of a long edge https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/shape https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/Pits-0.png Pit shape simulations https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/Pits-1.png Pit shape simulations https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/Pits-2.png Pit shape simulations https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/Pits-3.png Pit shape simulations https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/Pits-4.png Pit shape simulations https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/sim https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/etchmodel.png Etching model description Unit cells of the etching model. The free edge counts are shown for typical unit cell types. https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/etch000.png simulation starting image Step configuration at the start of an etching simulation https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/etchface.png Simulation example 1 Simulation result with stable edges, unstable islands. QuickTime animation (300kB) https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/etch4f.mov https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/etchface.png Etch pit simulation Simulation of step-flow etching at the corner of an etch pit https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/etchwavy.png Simulation example 2 Simulation result with metastable islands. QuickTime animation (1MB) https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/etch4w.mov https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/etchwavy.png Etch pit simulation Simulation of nearly isotropic etching at the corner of an etch pit https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/steps https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/edgestart.png Step angle simulations Initial step configurations: the step angles are 0, 15, 30, and 45 degrees. https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/edgefinal.png Step angle simulations Final step configurations. https://lippmaa.issp.u-tokyo.ac.jp/sto/etch/edgefinds.png Automatically detected step edge positions Averaged step edges marked on a simulation result image. https://lippmaa.issp.u-tokyo.ac.jp/sto/ https://lippmaa.issp.u-tokyo.ac.jp/sto/stoball200.gif SrTiO3 unit cell SrTiO3 unit cell. Ti-violet, Sr-green, O-blue. https://lippmaa.issp.u-tokyo.ac.jp/sto/stoball.gif SrTiO3 unit cell SrTiO3 unit cell. Ti-violet, Sr-green, O-blue. https://lippmaa.issp.u-tokyo.ac.jp/sto/sto100s.png SrTiO3 (001) SrTiO3 (001) https://lippmaa.issp.u-tokyo.ac.jp/sto/sto100.png SrTiO3 (001) SrTiO3 (001) https://lippmaa.issp.u-tokyo.ac.jp/sto/sto110s.png SrTiO3 (110) SrTiO3 (110) https://lippmaa.issp.u-tokyo.ac.jp/sto/sto110.png SrTiO3 (110) SrTiO3 (110) https://lippmaa.issp.u-tokyo.ac.jp/sto/sto111s.png SrTiO3 (111) SrTiO3 (111) https://lippmaa.issp.u-tokyo.ac.jp/sto/sto111.png SrTiO3 (111) SrTiO3 (111) https://lippmaa.issp.u-tokyo.ac.jp/sto/substrate https://lippmaa.issp.u-tokyo.ac.jp/sto/idealw.png Step-and-terrace SrTiO3 surface An AFM image and a structural model of a step-and-terrace SrTiO3 surface https://lippmaa.issp.u-tokyo.ac.jp/sto/polished.png Polished STO Substrate surface produced by chemical-mechanical polishing https://lippmaa.issp.u-tokyo.ac.jp/sto/etched.png Etched STO AFM image of a SrTiO3 surface during wet etching in buffered HF acid https://lippmaa.issp.u-tokyo.ac.jp/sto/annealed.png Annealed STO Etched SrTiO3 substrate after annealing in reducing conditions https://lippmaa.issp.u-tokyo.ac.jp/sto/atomic.png Atomic resolution STO The annealed surface may show various reconstructions related to segregated atoms