Silicon Nanopore Arrays: Fabrication and Applications for DNA Sensing MIAO ZHANG Doctoral Thesis in Physics School of Engineering Sciences KTH Royal Institute of Technology Stockholm, Sweden 2018 KTH School of Engineering Sciences TRITA-SCI-FOU 2018:18 SE-100 44 Stockholm ISBN 978-91-7729-804-5 SWEDEN Akademisk avhandling som med tillstånd av Kungl Tekniska högskolan framlägges till offentlig granskning för avläggande av teknologie doktorsexamen i fysik freda- gen den 15 juni 2018 klockan 10:00 i Sal C, Electrum, Kungl Tekniska Högskolan, Kistagången 16, Kista, Stockholm. Cover photo: A dark field optical image of 4-by-4 nanopores on a Si membrane with buried oxide residuals beneath. © Miao Zhang, May 2018 Tryck: Universitetsservice US AB To my beloved parents v Abstract Nanopore biomolecule sensing and sequencing has emerged as a simple but powerful tool for single molecule studies over the past two decades. By elec- trophoretically driving single molecules through a nanometer-sized pore, often sitting in an insulating membrane that separates two buffer solutions, ionic current blockades can be detected to reveal rich information of the molecules, such as DNA length, protein size and conformation, even nucleic acid se- quence. Biological protein pores, as well as solid-state nanopores have been used, but both suffer from relatively low throughput due to the lack of abil- ity to scale up to a large array. In this thesis, we tackled the throughput issue from the fabrication aspect as well as from the detection aspect, aim- ing at a parallel optical single molecule sensing on an array of well-separated nanopores. From the fabrication aspect, several lithography-based self-regulating meth- ods were tested to obtain nanopore arrays in silicon membranes, including anisotropic KOH etching, thermal oxidation-induced pore shrinkage, metal- assisted etching and electrochemical etching. Among those, the most success- ful method was the electrochemical etching of silicon. By electron-beam or photo lithography, the positions of the pores were defined on a silicon mem- brane. Followed by anisotropic KOH etching, inverted pyramids were formed as etching pits. The nanopores were then formed by anodic etching of silicon in HF. Using this concept, the size of the pores does not depend on the lithog- raphy step; only the positions of pores were defined by lithography. In this way, an array of ∼ 900 pores with an average entrance diameter of 18 ± 4 nm was fabricated on a 120 µm × 120 µm membrane. From the detection aspect, parallel readout of fluorescence signals from the labelled DNA molecules while translocating through an array of nanopores was performed using a wide-field microscope with a relatively fast CMOS camera recording at 1 KHz frame rate. Statistics of duration and frequency of the translocation events were extracted and studied. It was found that the event duration decreases with rising excitation laser power. This can be attributed to a laser-induced heating effect. Simulation suggested that a sig- nificant thermal gradient was generated at the pore vicinity by the excitation laser due to photon absorption by the silicon membrane. Such temperature rise affects all mass transport in a solution via a viscosity change. The ther- mal effect has also been proven by that conductance of an array of nanopores scales with the laser power. The thermal effect on the translocation frequency has been studied systematically as well. Due to thermophoresis of DNA in a thermal gradient, the thermophoretic force serves as a repulsion force, op- posing the electrophoretic force at the pore vicinity, depleting molecules away from the pore. Because of the molecule-size-dependent thermal depletion, a size-dependent translocation frequency was observed. This can be potentially used for a high throughput molecule sorting by adjusting the balance between the thermophoretic force and the electrophoretic force. Keywords: nanopore, array, electrochemical etching, DNA, optics, thermophoresis vi Sammanfattning Detektion av biomolekyler med hjälp av nanoporer har under de senaste två decennierna vuxit fram som ett enkelt men kraftfullt verktyg för studi- er av enstaka molekyler. Genom att driva molekyerna elektroforetiskt (med elektriskt fält) genom en nanometer-stor por, som ofta sitter i ett isoleran- de membran som separerar två buffertlösningar, kan molekylerna detekteras genom att jonströmmen genom membranet delvis blockeras. Detta kan ge detaljerad information om de detekterade molekylerna, såsom t ex ett pro- teins storlek, längden på en DNA-sekvens, och även sekvensen på ingående nukleinsyror. Naturliga, biologiska protein-baserade porer, liksom nanoporer gjorda i fasta material har använts, men båda lider av relativt låg effektivitet på grund av svårigheten att skala upp tekniken till stora matriser av nano- porer. I denna avhandling, tacklas denna fråga från både tillverknings-sidan såväl som från detektions-sidan genom att använda en matris av nanoporer för parallell optisk detektering av enskilda molekyler vilket resulterar i en hög ge- nomströmningshastighet. Detta möjliggörs genom att nanoporerna separeras optiskt, dvs med några mikrometers mellanrum. För att tillverka matriser av nanoporer i kisel-membran har flera olika självreglerande metoder undersökts experimentellt, bland annat anisotrop KOH etsning för att minska porstorleken, termisk oxidation för att krym- pa porer, metall-inducerad etsning samt elektrokemisk etsning. Bland dessa befanns den elektrokemiska metoden ge bäst resultat. För att bestämma posi- tionen av porerna på membranet användes optisk litografi eller elektronstråle- litografi varefter KOH etsning resulterade i pyramid-formade gropar. Dessa initierade sedan etsningen av nanoporer genom membranet i den efterföljan- de elektrokemiska etsprocessen. Genom detta förfarande bestäms inte stor- leken (diametern) på porerna av litografi-processen utan endast deras lägen på membranet. På detta sätt kunde 900 porer med en medel-diameter på 18 ± 4 nm tillverkas på ett membran av storleken 120 µm × 120µm. Detektering av de fluorescens-inmärkta DNA-molekylerna utfördes paral- lellt genom avläsning av fluorescenssignaler när de passerade genom matrisen av nanoporor med hjälp av ett mikroskop med en relativt snabb CMOS- kamera med video-sekvenser på upp till 1 KHz. Statistiken över tiden för passage genom membranet av enskilda molekyler och frekvens av händelser- na extraherades och studerades. Det visade sig att tiden för passage mins- kade med ökande laser excitation. Detta kan hänföras till en laser-inducerad uppvärmnings-effekt. Uppskattning av denna effekt genom dator-simulering visade att en termisk gradient bildas i närheten av porerna på grund av upp- värmning av kisel-membranet genom absorption av laser-strålen. En sådan temperaturhöjning påverkar all mass-transport i en lösning via ändringar av viskositeten. Den termiska effekten bevisas också av att konduktansen genom matrisen av nanoporer ökar linjärt med lasereffekten. Den termiska effekten på frekvensen av molekyl-passager har också studerats systematiskt. På grund av termofores av DNA i en termisk gradient, verkar den termoforetiska kraf- ten som en repulserande kraft som motsätter sig den elektroforetiska kraften i porområdet. På grund av att utarmningen av molekyler vid porerna beror på storleken observerades en storleks-beroende passage-frekvens. Detta skulle vii potentiellt kunna användas för molekylsortering med hög genomströmnings- hastighet genom att justera balansen mellan den termoforetiska kraften och den elektroforetiska kraften. viii Acknowledgements PhD study is a life changing experience. Over the past six years, I have paddled in the ocean of knowledge, often found myself lost in the mist. In the end, I finally found a spot to dive in. The ocean had then revealed its true beauty to me, the beauty of understanding the unknown. All this cannot happen without a number of great people whom I would like to express my sincerest gratitude. First, I would like to thank Prof. Jan Linnros for giving me this oppor- tunity to work on this fascinating project and for supporting me to carry on for six years! His vision has brought to me the frontier of field. He is always generous to his students with his time, knowledge, advice and encouragement. I really enjoyed being your student, always feeling equal and respected as an individual. I would also like to thank Assoc. Prof. Ilya Sychugov for his inspiring discussions, heart-warming encouragement and contagious laughter. You are like a lighthouse, guiding me to the right methodology. I have learned from you the shortcut to truly understand something is to go back to the basic— get crystal clear about the definition and read the manual. Talking to you is always a pleasure. Not just limited to physics, I also enjoyed very much our tea-break chats about China, Japan and ice hockey. Along my studies, I have received many helps from our collaborators, former project members and lab technicians. My gratitude goes to Assoc. Prof. Niclas Roxhed, my co-supervisor, for his guidance in micro-fabrication; Prof. Joakim Lundberg, Prof. Afshin Ahmadian, Pelin Sahlén, Anders Jemt and David Redin for offering us DNA samples; Chonmanart Ngampeerapong for his hard working on electrochemical etching; Aki-Kimmo Kallio for sharing with me everything he knew about the project. I am also grateful to the Myfab-Electrum laboratory staff for maintaining an efficient public lab with an open and friendly atmosphere and especially to my cleanroom mentors, Aleksandar Radojcic and