UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

ZnO nanostructures for sensing and photovoltaic devices Mohseni Kiasari, Nima 2014

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata


24-ubc_2014_spring_mohsenikiasari_nima.pdf [ 32.2MB ]
JSON: 24-1.0166914.json
JSON-LD: 24-1.0166914-ld.json
RDF/XML (Pretty): 24-1.0166914-rdf.xml
RDF/JSON: 24-1.0166914-rdf.json
Turtle: 24-1.0166914-turtle.txt
N-Triples: 24-1.0166914-rdf-ntriples.txt
Original Record: 24-1.0166914-source.json
Full Text

Full Text

ZnO Nanostructures for Sensingand Photovoltaic DevicesbyNima Mohseni KiasariB.Sc., Sharif University of Technology, Tehran, Iran, 2007M.Sc., Imperial College London, London, United Kingdom, 2008A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinThe Faculty of Graduate and Postdoctoral Studies(Electrical and Computer Engineering)THE UNIVERSITY OF BRITISH COLUMBIA(Vancouver)April 2014c© Nima Mohseni Kiasari 2013AbstractIn this PhD thesis, vertical arrays of zinc oxide (ZnO) nanowires (NWs)are synthesized in a CVD system and then deposited on patterned electrodesusing dielectrophoresis (DEP). The nanowire devices illustrate 4 orders ofmagnitude increase in conductivity when exposed to ultra violet (UV) irra-diation of 1220 µW/cm2. The UV response has a fast component, due toelectron-hole generation, as well as a slower component, attributed to therelease of oxygen. Moreover, due to the increased electron density in thepresence of UV, the type of oxygen species on the surface of ZnO changesto more reactive negative ions. In addition, when the pressure is decreasedto 0.05 mBar, the conductivity of the NWs increases ∼ 2 and 3.5 times forNWs with 300-nm and 100-nm diameter, respectively. For the first time, UVirradiation is used to improve the carbon monoxide (CO) sensing propertiesof ZnO. When exposed to 250 µW/cm2 UV irradiation, not only the sensitiv-ity increases more than 75%, but also a repeatable and recoverable responseis obtained, which is due to formation of more reactive oxygen ions. Forthe same reason, when the temperature is elevated, higher sensitivity to COiiis achieved. The devices demonstrate exponential sensitivities of more than5 decades to 60% increase in relative humidity (RH) at room temperature,which is a record for ZnO NW based RH sensors.A novel, low-cost and simple technique is developed for fabrication of sen-sors based on solution processed ZnO nanoparticles (NPs) by simply sketch-ing the electrode lines and painting the NP ink. Sensors show 2000 timesincrease in conductivity when exposed to 1220 µW/cm2 UV irradiation andmore than 200% increase in current when exposed to 5-mins of CO pulse atroom temperature.Furthermore, this thesis presents efficient (3.8%) inverted organic pho-tovoltaic devices based on a P3HT:PCBM bulk heterojunction blend withimproved charge-selective layers. ZnO NP films with different thicknesses aredeposited on the transparent electrodes as a nano-porous electron-selectivecontact layer. The optimized inverted devices show exceptional short circuitcurrent, which is related to increased quantum efficiency.iiiPrefaceMy contributions during the PhD has resulted in the following publica-tions and conference presentations.Journal Publications• N. Mohseni Kiasari, S. Soltanian, B. Gholamkhass, P. Servati, ”Sketch-ing Functional, Ubiquitous ZnO Nano-Sensors on Paper” Accepted forpublication, RSC Advances.• N. Mohseni Kiasari, S. Soltanian, B. Gholamkhass, P. Servati, ”Envi-ronmental gas and light sensing using ZnO nanowires,” IEEE Transac-tions on Nanotechnology, 13 (2014) 368-374.• N. Mohseni Kiasari, S. Soltanian, B. Gholamkhass, P. Servati, ”Roomtemperature ultra-sensitive resistive humidity sensor based on singlezinc oxide nanowire,” Sensors and Actuators A: Physical, 182 (2012)101-105.• B. Gholamkhass, N. Mohseni Kiasari, P. Servati, ”An efficient invertedivorganic solar cell with improved ZnO and gold contact layers,” OrganicElectronics, 13 (2012) 945-953.• N. Mohseni Kiasari, P. Servati, ”Dielectrophoresis-assembled ZnO nanowireoxygen sensors,” Electron Device Letters, IEEE, 32 (2011) 982-984..Conference Presentations and Papers• P. Servati, B. Gholamkhass, S. Soltanian, R. Rahmanian, N. MohseniKiasari, Z. Jiang, F. Ko, J. Shen and A. Aljaafari, ”Nanostructuredelectrodes and photoactive layers for efficient, stable and flexible or-ganic photovoltaic device,” ECS 2013 Toronto. (Oral Presentation)• N. Mohseni Kiasari, S. Soltanian, B. Gholamkhass and P. Servati,”Sketching high performance ZnO nanoelectronic sensors on paper,”MRS Spring 2013 San Francisco. (Oral Presentation)• N. Mohseni Kiasari, B. Gholamkhass, P. Servati,”Hybrid organic-inorganicphotovoltaic devices with composite nanomesh transparent electrodeon clear plastic substrates,” MRS spring 2012 meeting, San Francisco,California, USA. (Poster Presentation)• B. Gholamkhass, N. Mohseni Kiasari, J. Shen, S. Soltanian, and P. Ser-vati, ”High short circuit current, stable and efficient inverted organic so-lar cell with nanostructured electron- and hole-electrodes,” MRS spring2012 meeting. (Oral Presentation)v• N. Mohseni Kiasari, J. Shen, B. Gholamkhass, S. Soltanian, P. Servati,”Well-aligned zinc oxide nanowire arrays for transparent electrode ap-plications,” 2011 IEEE Photonics Conference (PHO), pp. 561-562, 9-13Oct. 2011. (Oral Presentation)• N. Mohseni Kiasari, P. Servati, ”ZnO nanowrie for oxygen sensing,”15th Canadian Semiconductor Conference, 15-17 August 2011, Van-couver Canada. (Oral Presentation)The chapters of this thesis were prepared and written by the author withthe help and supervision of Dr. Peyman Servati and conducted in UBC’sFlexible Electronics and Energy Lab (FEEL). All of the experiments, exceptchapter 5, were designed and carried out by the author with the help andguidance of Dr. Saeid Soltanian. Dr. Jun Shen helped for the developmentof chemical vapour deposition system. All the experimental and theoreticaldata were systematically analyzed and interpreted by the author and thendiscussed with Dr. Soltanian, Dr. Gholamkhass and Prof. Servati. Thesensor test setup as well as hydrothermal growth thermal bath were designedand built with the assistance of Dr. Soltanian. Most of the experiments inChapter 5, photovoltaic performance, were conducted mainly by Dr. Gho-lamkhas with the assistance of the author except the ZnO layer fabrication,characterization as well as device performance measurement that were per-formed by the author. All the results and discussion chapters, chapter 3 to5, are mainly based on the peer-reviewed journal papers resulted from thisviresearch.viiTable of ContentsAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiPreface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ivTable of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiiList of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiiList of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiiiList of Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . .xxivAcknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . .xxviDedication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xxvii1 Introduction and Background . . . . . . . . . . . . . . . . . . . 11.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Zinc Oxide Properties . . . . . . . . . . . . . . . . . . . . . . 31.3 Resistive Sensing Mechanism of ZnO . . . . . . . . . . . . . . 41.3.1 Effect of Downsizing . . . . . . . . . . . . . . . . . . . 5viiiTABLE OF CONTENTS1.3.2 Dielectrophoresis . . . . . . . . . . . . . . . . . . . . . 91.4 ZnO as Electron Selecting Electrode in Organic PhotovoltaicDevices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121.5 Objective of This Work . . . . . . . . . . . . . . . . . . . . . . 141.6 Thesis Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . 142 Experimental Setup and Device Preparation . . . . . . . . . 162.1 ZnO Nanostructure Synthesis . . . . . . . . . . . . . . . . . . 162.1.1 Chemical Vapour Deposition . . . . . . . . . . . . . . . Substrate preparation . . . . . . . . . . . . . Chemical Vapor Deposition System . . . . . . Evolution of CVD-grown NWs . . . . . . . . 192.1.2 ZnO Nanoparticle Film . . . . . . . . . . . . . . . . . . 262.1.3 Colloidal Solution of ZnO NP . . . . . . . . . . . . . . 262.1.4 Zinc Oxide Nanorods . . . . . . . . . . . . . . . . . . . 282.2 Dielectrophoresis . . . . . . . . . . . . . . . . . . . . . . . . . 302.3 Sensor Testing Setup . . . . . . . . . . . . . . . . . . . . . . . 372.4 Photovoltaic Device Fabrication and Characterization Setup . 393 Sensing Properties of ZnO Nanowires . . . . . . . . . . . . . . 413.1 Oxygen and Vacuum Sensing . . . . . . . . . . . . . . . . . . 413.1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . 413.1.2 Experimental Details . . . . . . . . . . . . . . . . . . . 423.1.3 Measurements and Analysis . . . . . . . . . . . . . . . 44ixTABLE OF CONTENTS3.1.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . 503.2 Ultra Violet (UV) Sensing . . . . . . . . . . . . . . . . . . . . 503.2.1 Nanowire Synthesis and Characterizations . . . . . . . 503.2.2 Device Fabrication . . . . . . . . . . . . . . . . . . . . 523.2.3 Results and Discussion . . . . . . . . . . . . . . . . . . 543.3 Relative Humidity Sensing . . . . . . . . . . . . . . . . . . . . 573.3.1 Nanowire Synthesis . . . . . . . . . . . . . . . . . . . . 593.3.2 Device Fabrication . . . . . . . . . . . . . . . . . . . . 603.3.3 Results and Discussion . . . . . . . . . . . . . . . . . . 623.3.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . 703.3.5 Correlation of UV and RH Sensing . . . . . . . . . . . 713.4 Oxygen Sensing . . . . . . . . . . . . . . . . . . . . . . . . . . 713.5 Carbon Monoxide Sensing . . . . . . . . . . . . . . . . . . . . 754 ZnO Sensors on Paper . . . . . . . . . . . . . . . . . . . . . . . 784.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 784.2 Device Fabrication . . . . . . . . . . . . . . . . . . . . . . . . 804.3 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . 814.3.1 Ultra Violet Sensing . . . . . . . . . . . . . . . . . . . 844.3.2 Carbon Monoxide Sensing . . . . . . . . . . . . . . . . 854.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 875 Photovoltaic Devices . . . . . . . . . . . . . . . . . . . . . . . . 895.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89xTABLE OF CONTENTS5.2 Experimental Section . . . . . . . . . . . . . . . . . . . . . . 925.2.1 Device Fabrication . . . . . . . . . . . . . . . . . . . . 925.2.2 Device Characterization . . . . . . . . . . . . . . . . . 935.3 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . 945.3.1 Optical Properties . . . . . . . . . . . . . . . . . . . . 945.3.2 Surface Morphology . . . . . . . . . . . . . . . . . . . . 965.3.3 Deposition of a Thin Film of Gold . . . . . . . . . . . . 975.3.4 Photovoltaic Performance . . . . . . . . . . . . . . . . 995.3.5 Device Stability . . . . . . . . . . . . . . . . . . . . . . 1045.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1075.5 Effect of Hydrothermally Synthesized Nanorods on PhotovoltaicPerformance . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1086 Conclusion and Future Work . . . . . . . . . . . . . . . . . . . 1116.1 Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . 1116.2 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116AppendicesA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129xiList of Tables2.1 The CVD growth parameters of different used recipes. . . . . . 202.2 Hydrothermal growth parameters used in different recipes. . . 305.1 Extracted parameters, short-current density (JSC), open-circuitvoltage (VOC), fill factor (FF) and power conversion efficiency(PCE) for P3HT:PCBM (50:50 wt.%) devices. Parenthesesdenote average performance obtained from 24 devices tested.The calculated JSC is based on the overlap integral betweenEQE and AM1.5G spectrum in the range of 330-800 nm. . . . 1005.2 Average photovoltaic parameters, short current density (JSC),open circuit voltage (VOC), fill factor (FF) and power conver-sion efficiency (PCE) for inverted devices. . . . . . . . . . . . 108xiiList of Figures1.1 Number of papers published on ZnO vs. year of publicationfrom ISI web of knowledge. . . . . . . . . . . . . . . . . . . . . 21.2 Wurtzite crystal structure of ZnO [1]. . . . . . . . . . . . . . . 41.3 Effect of adsorption/desorption of oxygen species on bandbending at the surface of ZnO. . . . . . . . . . . . . . . . . . . 61.4 The surface atoms percentage of Pd changes with the size ofcluster. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71.5 Effect of different forces in DEP process. . . . . . . . . . . . . 121.6 (a) Displays the schematic of the inverted OPV device used inthis thesis, and (b) shows energy level diagram of the invertedOPV device and schematic illustration of the photovoltaic con-version mechanism. . . . . . . . . . . . . . . . . . . . . . . . . 132.1 Gold nanoparticle distribution on (100) silicon substrate (a)before and (b) after optimization of HCl concentration. Im-ages with larger magnification are shown in the insets. Scalebars are 5 µm and 500 nm scale bar for images and insets,respectively. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18xiiiLIST OF FIGURES2.2 Horizontal tube furnace used for NW growth and its temper-ature profile. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192.3 (a) Schematics of the CVD system and (b) catalytic growthof ZnO NW. (c) SEM micrograph of a NW nucleated from anAu nanoparticle (NP at the base). . . . . . . . . . . . . . . . . 212.4 (a) and (b) show typical samples with low density of nanowires.Scale bar is 1 µm. (c) and (d) display samples with increaseddensity of nanostructures achieved by optimizing parametersto the values reported in Table 1 for recipe 1. Scale bar is 20µm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222.5 ZnO nano structures synthesized by recipe 2 deposed on (a)Si, (b) SiO2, (c) quartz and (d) aluminum foil. Scale bar is 5µm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232.6 ZnO nanobelts and nano-needles synthesized with zinc vaporand oxygen flow (recipe 3). Scale bar is 2 µm. . . . . . . . . . 232.7 ZnO nanowire structure at different downstream locations 6,8, 10, 12 and 14 cm from the centre of the furnace. Scale baris 5 µm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242.8 ZnO array on c-plane sapphire. Scale bar is 10 µm. . . . . . . 242.9 XRD pattern of three different samples: (a) grown by recipe2, (b) and (c) grown using recipe 4 . . . . . . . . . . . . . . . 252.10 ZnO nanoparticle film. Scale bar is 500 nm . . . . . . . . . . . 262.11 ZnO nanoparticle colloid synthesis process. . . . . . . . . . . . 27xivLIST OF FIGURES2.12 TEM image of solution processed NPs. The inset shows thesolution of NPs in chloroform (90 mg/mL). . . . . . . . . . . . 282.13 (a) Hydrothermal growth system used for the growth of ZnOnanorods, (b) the custom-made sample holder and (c) showshow the sample holder is placed in the solution. . . . . . . . . 302.14 SEM micrographs of three samples synthesized using differentrecipes for hydrothermal growth of ZnO NRs. Scale bar is 1µm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312.15 Dielectrophoresis set-up comprised of two probes connectingto a signal generator, under a stereo microscope. . . . . . . . . 322.16 Effect of Dielectrophoresis Force on NWs. Scale bar is 100 µm 332.17 (a)-(c) Heavy deposition of NWs without DEP optimization.(d)-(f) Single NW deposition when the DEP parameters areoptimized. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342.18 Effect of electrode geometry on DEP is shown for (a) two elec-trodes directly in front of each other, (b) two neighbour par-allel electrodes and (c) proposed geometry, two narrow elec-trodes in front to each other. The grey area between the elec-trodes represents the electric field and the darker areas showstronger field. . . . . . . . . . . . . . . . . . . . . . . . . . . . 362.19 (a) Probe station with a vacuum chamber, (b) Semiconductoranalyzer and (c) turbo vacuum pump with pressure monitor. . 372.20 Testing setup for gas and light sensing. . . . . . . . . . . . . . 38xvLIST OF FIGURES2.21 (a) Glovebox used for fabrication of the PV devices, (b) cus-tom made holder for OPV devices, (c) measurement systemfor characterization of the PV device and (d) Labview basedsoftware used to characterize the PV performance. . . . . . . . 403.1 (a) FE-SEM micrograph, (b) X-ray diffraction pattern, (c)electron diffraction pattern obtained from TEM, (d) and en-ergy dispersive X-ray spectroscopy of as-synthesized ZnO nanowires.Scale bar is 2 µm. . . . . . . . . . . . . . . . . . . . . . . . . . 443.2 (a) A schematic of a multi-fingered device structure showingthe placement of NWs and adsorption of oxygen species on thesurface of the NW during pressure experiments. (b) Current-voltage characteristics of a ZnO NW device under different airpressure. Scale bar is 2 µm. . . . . . . . . . . . . . . . . . . . 463.3 Relative sensitivity to pressure as a function of temperature.The inset illustrates the measured resistance of the NW deviceas a function of pressure of air, nitrogen and argon. . . . . . . 473.4 Current-voltage characteristics of a ZnO device with Schottkycontacts under different pressure levels, showing a change inthreshold for conduction as a function of pressure. The insetshows the band diagram at the contact vicinity after (solidlines). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49xviLIST OF FIGURES3.5 (a) low and (b) high magnification 45o-titled SEM micrographof vertically-aligned ZnO NWs (scale bar is 10 µm and 2 µmfor a and b respectively), and (c) X-Ray diffraction patternand (c) Photoluminescence spectrum of the synthesized ZnONWs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 523.6 Typical current-voltage characteristics of a working device. ADEP-assembled device with a NW captured between two goldelectrode as well as the schematics of a NW aligned across twometallic electrodes as are shown in the insets. . . . . . . . . . 543.7 (a) Current-voltage characteristics of the sensor under UV illu-mination with different intensities. The transient response ofthe sensor to (b) the UV pulses with different intensities, and(c) four successive UV pulses with the same intensity (1220µW/cm2: black, and 5 times attenuated: red) . . . . . . . . . 563.8 Transient response of the sensor to a 180 s UV pulse underatmospheric vs. reduced pressures. The inset shows the re-sistance of the sensor as a function of pressure under darkconditions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 583.9 (a) 45o titled SEM micrograph of vertically-aligned ZnO NWs(scale bar is 5µm), and (b) XRD measurements, (c) TEMphotomicrograph and crystalline diffraction pattern (d) andEDS elemental analysis results for synthesized ZnO NWs. . . . 61xviiLIST OF FIGURES3.10 Current voltage characteristics of a typical device at room tem-perature. Insets show SEM image of a single NW device andschematics of the experimental set up. Scale bar is 3 µm. . . . 623.11 (a) Steady state characteristics of the sensors at different RHvalues. The inset shows a semi-logarithmic graph of resistivityvs. RH. (b) Schematic of a single NW sensor and the changein electron density of NW in dry and humid air conditions. . . 673.12 (a) The changes in the measured current a ZnO NW devicesin response to exposure to pulses of 30% RH air between dryair flushes. (b) The measured current of the sensor, extractedfrom different RH pulses (17%, 25%, and 35%). (c) Arrheniusplot of the relative sensitivity of the sensor current to RH. . . 683.13 (a) Steady-state current-voltage characteristics of a sensor un-der various RH levels. The inset shows the transient responseto pulses of 86% RH between dry air purges. (b) The transientresponse of the sensor to increasing RH from dry to 86%. . . . 703.14 The transient response of the sensor to (a) the successive 80%RH air and dry air pulses under dark and UV irradiation, and(b) the successive UV pulses in the dry and 80% RH air ambient. 723.15 The transient response of the sensor to pulses of oxygen andargon under dark and UV irradiation at room temperature. . . 73xviiiLIST OF FIGURES3.16 (a) The transient response of the sensor to successive pulsesof oxygen and argon at different temperatures. (b) Sensitivityas a function of temperature, (c) and an Arrhenius plot of thesensor’s response. . . . . . . . . . . . . . . . . . . . . . . . . . 743.17 The transient response of the sensor to (a) pulses of pure COand air under dark and UV irradiation conditions, and (b)the successive 5 minute long CO and air pulses at differentambient temperatures 100 oC, 150 oC, 175 oC, and 200 oC. (c)The Arrhenius plot of the sensitivity of the NW sensor to pureCO. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 763.18 The transient response of the sensor to the successive pulses ofair having different concentrations of CO from 10% to 100%at 200 oC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 774.1 (a) Transparent ZnO NP solution in chloroform. (b) AFMimage of a spin-coated NP film. (c) High resolution SEMimage of a NP thin film and (d) TEM image of the ZnO NPs. 82xixLIST OF FIGURES4.2 (a) Sketching steps of a sensor on copy paper using calligraphyink. First step is stamping the electrode lines followed by stepII and III which are connecting the lines using a brush. Finallyin step IV the NP ink is painted on the drawn electrode lines,(b) A low magnification SEM micrograph of electrode linesand active channel of the device. (c) Higher magnificationSEM of the NPs on cellulose fibre matrix and (d) calligraphyink-drawn electrode comprised of carbon nano-clusters. . . . . 834.3 (a) Current-voltage characteristics of dark and photo-currentof a typical sensor. (b) sensor’s response to a 30 s pulse ofUV light and (c) device transient behaviour under various UVintensities. In the inset the signal from the sensor vs. theintensity of the UV is shown. . . . . . . . . . . . . . . . . . . 854.4 (a) Effect of 5 min CO pulses on two typical devices and copypaper. (b) 1-hour long pulse of CO (c) sensitivity under UVvs. dark. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 875.1 Inverted solar cell device structure (a) and approximate energylevel diagram (b). . . . . . . . . . . . . . . . . . . . . . . . . . 91xxLIST OF FIGURES5.2 Absorption spectrum of (a) ZnO on quartz substrates withthicknesses of 25 nm (dashed line), 50 nm (solid line) and 80nm (dotted line) and (b) P3HT:PCBM spin-coated from 1,2-DCB at 1000 rpm and slow dried before (solid line) and after(dashed line) thermal annealing at 150 oC for 1 h. . . . . . . . 955.3 SEM and AFM images of ZnO film surfaces on ITO used ashole-blocking layers in inverted OPV devices. ZnO film wasprepared by spin-coating zinc acetate solution in methanoland rapid heating to 350 oC for 5 min. For SEM images,thicknesses are (a) 25 and (b) 50 nm, and for AFM images (c)25, (d) 50 and (e) 80 nm for a 1 1 µm pixel. A 10 10 µmpixel of (e) is presented in (f). The inset in (b) shows a widerview (scale: 1 µm). . . . . . . . . . . . . . . . . . . . . . . . . 975.4 TEM images of P3HT:PCBM film used as photoactive lay-ers in inverted OPV devices (film thickness ∼ 150 nm) onZnO/PEDOT/ITO before (a) and after thermal annealing at150 oC for 1 h (b), and with 2-nm gold coating (c). Scale baris 100 nm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 985.5 A picture of the devices after PEDOT:PSS is spin-coated di-rectly on P3HT:PCBM (a and b) and on 5-nm gold coatedones (c and d). Notice that in the gold coated devices (c andd), the deposition of PEDOT:PSS layer is visually confirmedon gold coated area (marked by arrows and dotted lines). . . . 99xxiLIST OF FIGURES5.6 Current-voltage characterization of the inverted OPV deviceswith the highest efficiency among each series with (a) 2 and(b) 5 nm of gold coating on P3HT:PCBM/ZnO/ITO inter-face. Dashed and solid lines represent dark and photocurrent,respectively. The PV parameters for each device are: (a) VOC= 0.55 V, JSC = 13.1 mA/cm2, FF = 0.54 and PCE = 3.8%and (b) VOC = 0.55 V, JSC = 12.5 mA/cm2, FF = 0.54 andPCE = 3.6%. . . . . . . . . . . . . . . . . . . . . . . . . . . . 1035.7 External quantum efficiency (EQE) of the inverted OPV de-vices with a 50-nm ZnO layer for (a) 2-nm, (b) 5-nm and (c)without gold buffer layer, and (d) a normal P3HT:PCBM de-vice that is superimposed to those of (a) and (b), showing im-provement in EQE in 500-650 nm range of the spectrum con-sistent with the observed higher short circuit current. Filledand open circles represent the EQE value calculated based onincident power when ZnO/ITO and ITO substrates, respec-tively, were used for power measurements. . . . . . . . . . . . 105xxiiLIST OF FIGURES5.8 Upper panel: power conversion efficiency (PCE) of non-encapsulatedinverted OPV devices (a) stored in air under ambient condi-tions and (b) exposed to 100 mW/cm2 of incident white light.Lines and soft-edges are drawn as a guide for trends in PCEand its error margin, respectively. Lower panel: microscopicimages of a P3HT:PCBM film on 50 nm ZnO/ITO substratesfor (c) as-cast, (d) exposed to air (50 days) and white light(100 mW/cm2, 70 h), and (c) a thermally annealed film at150 oC for 1 h. Dark area in (d) is the device active area withgold coating. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1065.9 (a) Displays a NP layer of 25 nm thick and (b) NRs grownfor 5 min on top of the NP layer. IN the inset 45o tilted SEMmicrograph is shown. Scale bar is 1 µm (c) also shows thetransmittance with (solid line) and without (dash line) NRs . 109A.1 (a) Current voltage characteristics of different DEP-assembleddevices at room temperature. (b) characteristics of 5 almost-linear devices used for sensor testing . . . . . . . . . . . . . . 130A.2 Sensitivity of 8 DEP-assembled devices to 1220 µW/cm2 UVlight. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131xxiiiList of AbbreviationsAFM Atomic Force MicroscopeCVD Chemical Vapour DepositionDCB DichlorobenzeneDEP DielectrophoresisEDS Energy Dispersive SpectroscopyEDX Energy Dispersive X-rayEQE External Quantum EfficiencyFET Field Effect TransistorFF Fill FactorIQE Internal Quantum EfficiencyITO Indium Tin OxideNP NanoparticleNR NanorodNS NanostructureNW NanowireOPV Organic PhotovoltaicP3HT Poly(3-hexylthiophene)xxivLIST OF FIGURESPCE Power Conversion EfficiencyPCBM [6,6]-phenyl C61-butyric Acid Methyl EsterPEDOT:PSS Poly(3,4-ethylenedioxythiophene) Poly(styrenesulfonate)PPM Part Per MillionPV PhotovoltaicRH Relative HumiditySEM Scanning Electron MicroscopeTEM Tunneling Electron MicroscopeUV Ultra VioletXRD X-ray DiffractionxxvAcknowledgmentsThis work would have not been possible without the help and assistance ofmy supportive advisor, Dr. Peyman Servati, and my knowledgable colleaguesat Flexible Electronics and Energy Lab (FEEL). Great thanks go to ourresearch associate, Dr. Saeid Soltanian, and postdoctoral fellow, Dr. BobGholamkhass for their immense help and guidance.I would also like to thank to University of British Columbia (UBC), Nat-ural Sciences and Engineering Research Council of Canada (NSERC), andCanada Foundation for Innovation (CFI) for their financial support through-out my PhD studies.xxviDedicationTo my parents...xxviiChapter 1Introduction and Background1.1 MotivationIn 1959, when Richard Feynman presented his speech ”There’s Plenty ofRoom at the Bottom,” which is known as the origin of nanotechnology [2],nobody could even imagine the astonishing pace of advances in later yearsespecially from early 90s. Since then, the definition of nanotechnology hascontinuously evolved and expanded to a point that nowadays it applies toa wide range of research and technological fields that deal with dimensionssmaller than tenth of a micron. Downsizing many materials to nanoscaleleads to distinct properties not present in their bulk form. These propertiesare usually attributed to quantum confinement effects as well as large surface-to-volume-ratio of nanostructures. Among different materials, zinc oxide hasbeen studied extensively due to its unique electronic, optical and chemicalproperties with applications in large-area and flexible electronics, electronicskin, photonics, piezoelectric power generation and chemical sensing [3, 4, 5,6, 7, 8]. As shown in figure 1.1, the number of publications related to ZnOhas been exponentially increasing since early 90s. Although research on ZnO11.1. Motivationstarted from the first half of 20th century, its importance as a semiconductorstarted to gain attention in early 90s when new crystal growth methods weredeveloped to produce high quality and single crystalline ZnO in bulk amounts[1].1985 1990 1995 2000 2005 201002000400060008000  Number of PublicationsYearFigure 1.1: Number of papers published on ZnO vs. year of publication fromISI web of knowledge.21.2. Zinc Oxide Properties1.2 Zinc Oxide PropertiesZinc oxide is a II-VI compound semiconductor with a wide direct band gap(3.37 eV) that lies in the ultra-violet (UV) spectrum [9] and makes it at-tractive for opto-electronic applications such as solar cells and light emittingdiodes. It also has a large exciton binding energy (60 meV) [10] so that exci-tonic emission processes can take place at room and even higher temperatures[11]. ZnO crystallizes in a wurtzite structure (figure 1.2), and is available notonly as large bulk single crystals [11] but also can be easily synthesized invarious nanostructured morphologies such as nanorods, nanowires, tetrapods,nanoribbons and nanoparticles. ZnO is inexpensive, easy to synthesize andmore importantly, comprised of zinc and oxygen that are abundant materials.For many years, the use of ZnO as a functional semiconductor in elec-tronic devices such as transistors has been held back by the lack of controlover its electrical conductivity and its doping [1]. ZnO is usually n-type,the reason for this has been a controversial subject. Based on first princi-ples calculations using density functional theory (DFT) and all the extensiveexperimental investigations, it is known that the unintentional n-type dop-ing can be attributed to several factors such as native defects, impurities aswell as the surface of the metal oxide [1]. P-type doping has also remained achallenge and so far there has not been any reliable report on a stable and re-producible accepter incorporation into the ZnO crystal [10]. What is knownfor sure is that the conductivity and the carrier density of the semiconduc-31.3. Resistive Sensing Mechanism of ZnOtor can be tailored by the surrounding ambient which is quite desirable forapplications such as gas and environmental sensing.OZn[0001]Figure 1.2: Wurtzite crystal structure of ZnO [1].1.3 Resistive Sensing Mechanism of ZnOZnO nanostructures are appropriate candidates for sensing applications. Thepresence of oxygen vacancies on the surface [12, 13], coupled with highsurface-to-volume-ratio of nanostructures, leads to a superior sensitivity toenvironment, and the diversity of nano-structural forms (e.g., nanowires,nanorods, nanobelts, tetrapods, and nanosprings [14, 15, 16]), making ZnOsuitable for novel nanoscale sensing devices. One of the most convenient con-figurations and types of sensors is resistive sensor. Resistive sensors utilize41.3. Resistive Sensing Mechanism of ZnOvarious types of ZnO, where the electrical conductivity of the resistive chan-nel, ZnO, varies as a function of the type and amount of the adsorbed gasor the irradiated light on the ZnO surface [17]. The ZnO surface is usuallycovered with negatively charged oxygen species that are chemisorbed to theZnO surface and deplete the electrons from the region under the surface andcause a depletion region. Adsorption or desorption of these oxygen speciescan result in exchange of electrons between the metal oxide and the oxygenmolecule. In the case of gas sensing, these gases are adsorbed to the surfaceor react with the previously chemisorbed oxygen on the surface. This resultsin a change in the number of oxygen species present at the ZnO surface, thatshifts the Fermi level with respect to the valence and conduction bands asshown schematically in figure 1.3. The surface physics and chemistry involvedin the adsorption and desorption process at the surface of ZnO crystal arequite complex. There are plenty of factors for this complexity that includesthe number of polar and non-polar surfaces available, the presence of basic Oand acidic Zn sites, the complexity of defect chemistry, and the probabilityof both electron and oxygen atom transfer are factors that result in differenttypes of chemisorption [17].1.3.1 Effect of DownsizingDownsizing the crystal size results in significantly higher surface to volumeratio. In other words, by reducing the size of a crystal cluster, the ratioof surface atoms to the total number of atoms increases. This higher ratio51.3. Resistive Sensing Mechanism of ZnOEcEfEvEcEfEvO2-O2-O2-O2-O2-O2-O2-ZnOO2-O2-O2-O2-O2-O2-O2-O2-O2-O2-O2-O2-O2-O2-ZnOFigure 1.3: Effect of adsorption/desorption of oxygen species on band bend-ing at the surface of ZnO.boosts the sensitivity to the environment and consequently any change inthe surface can substantially influence the whole crystal’s electronic proper-ties. For instance, when the size of a Pd cluster [18] decreases from 100 nmto 10 nm, the percentage of surface atoms goes up from ∼ 1% to ∼ 10%.This percentage goes even higher to ∼ 30% when the size is about 3 nm asdisplayed in figure 1.4In addition to a higher surface atom ratio, in conventional thin film sen-sors or for micro scale sensors, the mechanism often used for describing thechange in conductance is the modification of the depletion region at the sur-61.3. Resistive Sensing Mechanism of ZnOFigure 1.4: The surface atoms percentage of Pd changes with the size ofcluster [18].face as the number of adsorbed oxygen molecules change, which implies alinear change of conductance with the presence of the stimulus gas. Thismechanism happens when the Debye screening length in the NW is smallerthan its radius. The Debye screening length gives us an idea about the dis-tance that charge imbalances are screened [19]. In larger crystals ( > 1µm)usually this length is much smaller than the size of the crystal and hencecharges at the surface can only affect the mobile electrons underneath of it.However, for some NW based sensors, an exponential increase in the NW71.3. Resistive Sensing Mechanism of ZnOconductivity was observed with increasing number of species at the surface,which suggests that the NW is operating in the subthreshold regime [20].In other words, current through the channel changes exponentially with thechange in gate potential (modified by adsorbed molecules), hence exhibitingmuch higher sensitivity to adsorbed species on the surface of the NW. Thisexponential behaviour can be attributed to the fact that the Debye screeninglength is larger than NW radius so that the whole NW volume is affectedby the gating of molecules on the surface. For the Debye screening length tobe at least two times the NW radius, the upper limit for the carrier concen-tration to meet the criteria can be estimated from the Debye screen lengthformula [19]:λZnO =√εZnOε0kBT/e2n (1.1)where εZnO is the dielectric constant of ZnO and ε0 the vacuum dielectricpermittivity, kB Boltzmann constant, T the absolute temperature, e theelementary charge and n the electron density. Knowing the electron mobilityand the NW conductivity without the presence of the stimulus, the carrierdensity, n, can be found from R=L/(µnε0pir02) where R is the resistanceof a cylindrical NW with the radius r0 and the length L (between the twoelectrodes), for a homogenous electron distribution. Once the carrier densityis smaller than the found upper limit from the equation 1.1, the conditionsare met and the ZnO NW is operating under sub threshold regime.81.3. Resistive Sensing Mechanism of ZnO1.3.2 DielectrophoresisThe biggest roadblock for bottom up integration of nanomaterials is howto manipulate these tiny structures and integrate them into devices withdifferent geometries, e.g., field effect transistors (FETs) and sensors. Thereare several expensive and complicated techniques such as e-beam lithographyfor fabrication of electrodes after locating a nanostructure or the use of nano-manipulator for placing a nanostructure on the contacts. In this work, weshow that dielectrophoresis (DEP) can be used as a low-cost, high yield,and more importantly, scalable technique for aligning nanowires betweentwo metallic contacts.Dielectrophoresis was defined by Pohl, the first who adopted it, as ”themotion of suspensoid particles relative to that of the solvent resulting frompolarization forces produced by an inhomogeneous electric field”[21]. Asit can be understood from the definition, when a suspension solution of aparticle (e.g. nanowire) is placed in an alternating electric field (e.g. be-tween two metallic electrodes connected to an AC signal) the now-polarizedparticles can be aligned to the applied field lines with a minimum energy re-configuration, such as along the field lines [22]. This alignment depends onthe electric field lines and their shape. The polarized particle is attracted tothe electrodes by means of dielectrophoretic forces, the magnitude of whichdepend on voltage, frequency, material properties and electrode geometrythat determines the pattern of electric field gradient in the solution. Thenanowire assembly process depends on the balance between many forces such91.3. Resistive Sensing Mechanism of ZnOas gravity, buoyancy, electrostatic repulsion, van der Waals, drag, and finallythe DEP force. Depending on the strength of these forces the direction ofNW manipulation can be determined. Once the DEP force along the z-axisis strong enough, there is a high chance of capturing a NW between the elec-trodes. Figure 1.5 schematically shows the NW alignment procedure as wellas the major forces applied to a NW during the DEP process.In order to calculate the DEP force, the well known equation for the forceexerted on a dipole in an electric field can be used:FDEP = (p.∇)E (1.2)From the equation, it is clear that in a uniform electric field the DEP forceis simply zero; in other words, a non uniform electric field can result in adielectrophoresis. However, this formula is only accurate when the magnitudeof the electric field does not change significantly across the dimension of theparticle [23]. In accordance to the mathematical proof, found in the reference[23], the time averaged DEP force is:< FDEP >=34εsRe[K(w)](p.∇)E2 (1.3)where Re(K(w)) is the real part of f the Clausius-Mossotti (CM) factor [24]101.3. Resistive Sensing Mechanism of ZnOthat can be defined as:Re[K(w)] = ( ε˜p − ε˜sε˜p + 2ε˜s) (1.4)where the subscripts εs and εp refer to relative complex permittivities of thesolvent and particle respectively.When the Clausius-Mossotti (CM) factor is positive, which means theparticle is more polarizable than the solvent, then DEP is going to be positiveand the particle will be attracted towards the maxima of the electric fieldgradient. Otherwise, the negative DEP happens and the particle is goingto be repulsed from the electrodes by the DEP force. Nevertheless, for aparticle like NW, depending on the orientation with respect to the electricfield gradient, CM factor along the long axis (length) or short axis (width)play different roles in the alignment of nanowires [25]. For instance, usuallyat low frequencies the DEP force is stronger but the long-axis component ismuch larger than the short-axis component that does not result in alignmentof a NW between two electrodes. However, at higher frequencies where theshort-axis CM factor is comparable to that of long-axis a higher yield ofcapturing NW can be achieved. Figure 1.5 schematically shows the importantforces applied to a NW during the DEP process.111.4. ZnO as Electron Selecting Electrode in Organic Photovoltaic Devices~FESFDEP FVDWMedium (solvent)Electric FieldNanowireElectrode ElectrodeFigure 1.5: Effect of different forces in DEP process.1.4 ZnO as Electron Selecting Electrode inOrganic Photovoltaic DevicesThe field of organic solar cells has been growing rapidly, showing a dra-matic increase in the reported numbers of published research papers, booksand patents as well as start-up companies [26]. Despite all the focused re-search on this topic, there are major issues that still need to be addressedfor commercialization of these devices. The main challenges are the powerconversion efficiency (PCE), life time (durability) and low-cost large-scalefabrication methods. Resolving these issues will pave the way for polymericsolar cells to become a viable economical solution in the renewable energymarket.121.4. ZnO as Electron Selecting Electrode in Organic Photovoltaic DevicesITOZnOP3HTPCBMPEDOT : PSSGlassGoldGold (2-5 nm)Figure 1.6: (a) Displays the schematic of the inverted OPV device used inthis thesis, and (b) shows energy level diagram of the inverted OPV deviceand schematic illustration of the photovoltaic conversion mechanism.One way to improve the stability of OPV device is to change the con-figuration to a less air-sensitive inverted structure that benefits from highwork function metals such as Ag or Au as the back electrode instead of theunstable Al, which is more common in regular configuration. In the inverteddevice, a layer of metal oxide such as TiO2 or ZnO is used on the front ITOelectrode to act as a hole blocking contact, as depicted in figure 1.6. More-over, this layer can behave as an oxygen barrier layer [27, 28] to protect thephotoactive organics, leading to a more air stable device. This could be thereason most of the large-scale roll-to-roll devices are fabricated based on thisstructure [26]. ZnO is an ideal candidate for this hole blocking layer, since itis transparent to visible light and compared to TiO2, has a higher electronmobility [29]. As shown in figure 1.6b, the energy levels of ZnO can facilitatethe transport of free electrons generated in the photoactive layer.131.5. Objective of This Work1.5 Objective of This WorkThe purpose of this work is primarily the synthesis of different ZnO nanos-tructures such as vertically aligned nanowires, hydrothermally grown nanorods,and solution processed nanoparticles as the main components of electronicdevices. Secondly, it investigates dielectrophoresis as an effective way tofabricate NW based devices with structure similar to field-effect transistors(FETs). Finally, this work studies different sensing properties (e.g. oxygen,humidity, carbon monoxide and ultra violet light) of these DEP-assembleddevices and compares them with similar devices using larger-size ZnO. Inaddition, the correlation between these sensing properties are examined andpossible ways to improve the selectivity of ZnO sensors are explored.In addition to the sensing properties of ZnO, application of these nanos-tructures in organic photovoltaic devices will be studied to investigate theeffect of ZnO nanostructures on efficiency as well as stability of the OPVdevices.1.6 Thesis OutlineThis thesis consists of 6 chapters. The current chapter introduces the moti-vation and objectives of this work as well as a brief background on the topics.More comparison with literature will be given in each of the chapters. Chap-ter 2 provides a description of all the experimental equipments and setupsthat were designed and made for this work. In addition, some details of the141.6. Thesis Outlineexperiments that are not mentioned in the results and discussion chapters,chapter 3 to 5, are briefly explained in chapter 2. The following three chap-ters are based on published and submitted journal papers (5 journal papers).In chapter 3, different sensing properties of DEP-assembled ZnO NW devicesare explained, which are based on three journal papers combined and mergedtogether. Chapter 4 introduces a low-cost and ubiquitous approach for fabri-cation of environmental sensors by simply sketching them on a piece of paper.Photovoltaic applications of ZnO nanostructures as electron selecting layersare investigated in an inverted organic solar cell configurations in chapter 5.Finally, chapter 6 presents a summary of the contributions and conclusion ofthe work and comments briefly on the future direction of this work.15Chapter 2Experimental Setup and DevicePreparationIn this chapter, a brief description of experimental apparatus, setups anddetails are explained. Primarily, various techniques and setups exploitedfor nanostructure synthesis are described followed by the device fabricationprocess. Finally, the setup developed for sensor sensing and the equipmentsfor photovoltaic device fabrication and measurements are explained.2.1 ZnO Nanostructure SynthesisThe first step towards fabrication of a ZnO nano-device is the synthesis ofthe nanostructures. Based on the type of application, different categories ofsynthesis methods such as chemical vapour deposition (CVD), hydrothermalgrowth, sol-gel process, and solution processed nanoparticle synthesis areused. In the following sections, a summary of various deposition conditionsand experimental steps are presented.162.1. ZnO Nanostructure Synthesis2.1.1 Chemical Vapour Deposition2.1.1.1 Substrate preparationSilicon wafers are attractive substrates due to the ease of processing and thepotential applications in fabrication of complex devices. Silicon (100) waferswith 300 nm thermal oxide were purchased and cut by a diamond saw intosmall pieces; followed by standard RCA1 and RCA2 cleaning methods. Goldnanoparticles were used as catalyst for zinc oxide growth. The particles weredeposited on substrate by a technique introduced by Woodruff et al. [30] forimmobilization of gold nanoparticles on silicon surface. The silicon dioxidewas etched by buffered HF 49% for 10 minutes so as to achieve a hydrogenterminated silicon surface, which not only preserves the etched surface fromoxidation for some time, but also improves the adhesion of gold nanoparticlesto the surface. The hydrogen terminated silicon substrates were immersedin a solution of 0.5 ml of 20 nm-gold nanoparticles in DI water (2 ml) and 1µl of HCl for 2-5 minutes, depending on the desired density of immobilizedparticles. The colloidal dispersed solution is citrate stabilized, i.e., negativelycharged citrate ions surround each nanoparticle, preventing agglomeration.Since silicon surfaces are known to have negative charge in water [18] thenthere is a repulsive force between the dispersed gold particles and silicon sur-face. However, adding acid to the colloidal solution of particles converts thecitrate ions to neutral citric acid due to the presence of protons. This pro-cess neutralizes gold nanoparticles, consequently leading to their deposition172.1. ZnO Nanostructure Synthesison silicon. Due to the tendency of particles to agglomerate, it is very criticalto add an optimum amount of HCl to the solution to prevent aggregation ofparticles on silicon. Figure 2.1 illustrates the scanning electron microscope(SEM) images of the immobilized particles before and after the optimizationof the HCl concentration. In addition to the substrate preparation described(a) (b)Figure 2.1: Gold nanoparticle distribution on (100) silicon substrate (a) be-fore and (b) after optimization of HCl concentration. Images with largermagnification are shown in the insets. Scale bars are 5 µm and 500 nm scalebar for images and insets, respectively.here, we used several other conventional deposition methods such as thermalevaporation of gold thin film on silicon as well as sapphire substrates whichwill be explained in the nanowire growth section. Chemical Vapor Deposition SystemIn this work, a conventional one-zone horizontal tube furnace was used as aCVD system, which is shown in figure 2.2. One end of the tube is connectedto mass flow controllers and the other end to a vacuum pump. Temperature182.1. ZnO Nanostructure Synthesisprofile is fairly constant (less than 6% temperature drop) within 10 cm fromthe center of the furnace and decreases dramatically outside this region, asillustrated in the figure.Vacuum Pump ArgonOxygenMFCs-30 -20 -10 0 10 20 3005001000Temperature (C)Distance from the center (cm)Figure 2.2: Horizontal tube furnace used for NW growth and its temperatureprofile. Evolution of CVD-grown NWsSince late 2009, different recipes have been used for ZnO nanowire growth.In our lab, we modified our system step-by-step toward getting vertically-aligned arrays of nanowires. Over time, we found the most favourable pre-cursor material mixture, source temperature as well as deposition tempera-ture, gas flow, and more importantly the process pressure. In this section,different recipes as well as the modifications which were done on the CVDsystem will be explained. Table 5.1 summarizes details of different growthrecipes used in this work.192.1. ZnO Nanostructure SynthesisTable 2.1: The CVD growth parameters of different used recipes.ID Precursor Evap.Temp.DepositionDistanceSubstrate Catalyst Gas Flow Vac. Fig.1 ZnO+C(1:1 w)950-1100oC5-10 cm SiliconQuartzAu NPs 30 sccmArNo 2.42 Zinc 600-700oC5-15 cm Silicon - 0-30 sccmArNo 2.53 Zinc 600-700oC5-15 cm Silicon Au NPs/1nmgold0-30 sccmAr + 3-5sccm O2No 2.64 ZnO+C(1:1 M)950 oC 12-16 cm SiliconSapphire100 nm1nm gold30 sccmAr + 3-5sccm O2Yes 2.7The general mechanism for nanowire growth is as follows: Zinc vapour isproduced by either thermal evaporation of zinc or reduction of ZnO powderat the centre of the furnace. Due to the carrier gas flowing and temperaturegradient, this vapour migrates toward the substrate which is placed down-stream of the furnace. At the substrate, zinc vapour diffuses into the AuNP and forms an alloy of Zn-Au. Once the alloy droplet is saturated, inthe presence of either oxygen or CO, a ZnO NW nucleates at the catalystNP. This process is shown schematically in figure 2.3. In the first recipe,we used equal weights of graphite and zinc oxide powder with a constant 30sccm flow of argon. Silicon substrates with 20 nm gold nanoparticles, wereplaced 5-10 cm downstream from the center, where the mixture of ZnO andgraphite precursor was placed at 950 oC. The tube furnace was at atmo-spheric pressure. ZnO powder is reduced by graphite forming zinc vapouras well as CO/CO2 at the center of the tube which has a higher tempera-202.1. ZnO Nanostructure SynthesisVapourTimeZnOAu NP(a)(b) (c)Figure 2.3: (a) Schematic of the CVD system and (b) catalytic growth ofZnO NW. (c) SEM micrograph of a NW nucleated from an Au nanoparticle(NP at the base).ture. Argon flow carries the zinc vapour to the gold particles forming alloydroplets that develop into supersaturated crystalline zinc oxide as a resultof the reaction between zinc and CO/CO2. As examples, SEM images ofsamples prepared with this recipe are shown in figure 2.4. Although somenanowires can be seen in the images, the resulting samples have small den-sities of NWs. Even by increasing the temperature to 1100 oC at the center,our furnace temperature limit, did not lead to any significant improvementin the resulting samples. Due to the low density of the NWs, as shown in thefigure 2.4, we decided to change the precursor to metallic zinc. Switching tothe zinc precursor resulted in a higher density of multi-wire structures suchas tetrapods with a few hundreds of nanometers in diameter, as shown infigure 2.5. Evaporation temperature was set to 600-700 oC with 0-20 sccm of212.1. ZnO Nanostructure Synthesis(a)Figure 2.4: (a) and (b) show typical samples with low density of nanowires.Scale bar is 1 µm. (c) and (d) display samples with increased density ofnanostructures achieved by optimizing parameters to the values reported inTable 1 for recipe 1. Scale bar is 20 µm.Ar flow and the substrates were placed 12-15 cm downstream from the cen-ter. The resulting structures on substrates with or without any gold catalystwere similar. We repeated the same experiments for other substrates suchas quartz, SiO2, and aluminium foil, and a similar behaviour was observed,as illustrated in figure 2.5. The next modification on the system was intro-ducing a small flow of oxygen (3-5 sccm) to the growth process. It resultedin a change in morphology of the nanostructure and instead of tetrapods,nanobelts, and nano-needle structures were observed in a catalytic growthas shown in figure 2.6. The nanostructures are mostly initiated from thegold particles. Since we were not able to obtain vertical arrays of ZnO NWs,we switched back to using zinc oxide and graphite powder. However, this222.1. ZnO Nanostructure Synthesis(a)Figure 2.5: ZnO nanostructures synthesized by recipe 2 deposed on (a) Si,(b) SiO2, (c) quartz and (d) aluminum foil. Scale bar is 5 µm.(a)Figure 2.6: ZnO nanobelts and nano-needles synthesized with zinc vapor andoxygen flow (recipe 3). Scale bar is 2 µm.time we did two modifications. First, we added a vacuum pump to maintainthe pressure at ∼ 1 mBar and second, equal molar ratios of graphite andZnO were used instead of equal weight ratios. We found the best positionfor growing vertical arrays of nanowires by inspecting samples synthesizedat different locations downstream from the center. Figure 2.7 displays SEM232.1. ZnO Nanostructure Synthesis(a)Figure 2.7: ZnO nanowire structure at different downstream locations 6, 8,10, 12 and 14 cm from the centre of the furnace. Scale bar is 5 µm.images of different samples at different positions. The same recipe, equalmolar ratios of ZnO and graphite powder at 950 oC (recipe 4), was also usedto synthesize nanowire arrays on sapphire substrate and consistent resultswere obtained as displayed in figure 2.8. In order to display the improvementFigure 2.8: ZnO array on c-plane sapphire. Scale bar is 10 µm.242.1. ZnO Nanostructure Synthesisof the synthesized nanowires over time, specifically their growth direction,three different samples and their XRD patterns are depicted in figure 2.9.Single XRD peak in figure 2.9c proves that all the NW on the substrate aregrown in the same crystal direction.(a)(c)(b)Figure 2.9: XRD pattern of three different samples: (a) grown by recipe 2,(b) and (c) grown using recipe 4.252.1. ZnO Nanostructure Synthesis2.1.2 ZnO Nanoparticle FilmA uniform layer of ZnO nanoparticles with an average size between 30-50 nmwas achieved for thin film solar cell as well as seed layer for nanorod growth,which will be discussed in chapter 5. 100 mg of zinc acetate dihydrate wasdissolved in 1 mL of methanol and spin coated on a glass substrate andthen baked at 350 oC, on a hot plate, under air ambient. The thermaltreatment results in decomposition of the zinc acetate and formation of zincoxide nanoparticles, displayed in the figure 2.10.Figure 2.10: ZnO nanoparticle film. Scale bar is 500 nm2.1.3 Colloidal Solution of ZnO NPFor some applications, it is desirable to have ZnO nanoparticles suspendedin a solution without requiring any thermal treatments after deposition, spe-cially for low cost and flexible substrates such as plastics and paper. For thispurpose, ZnO nanoparticles were synthesized based on the work of Weller etal. [31]. First, zinc acetate dihydrate (Sigma-Aldrich, 2.95 g) was dissolved262.1. ZnO Nanostructure Synthesisin methanol (125 mL) on a stirring hot plate at 120-150 oC, then a solutionof potassium hydroxide (Fischer Scientific, 1.48 g) in methanol (65 mL) wasadded slowly over 10 min to the zinc acetate solution under vigorous stirring.Zinc hydroxide formed and precipitated (figure 2.11a) but was dissolved again(figure 2.11b). After almost 1 hour the NPs start to precipitate (figure 2.11cand d); heating and stirring were stopped after 2 hours. Once NPs were allprecipitated, the NPs were washed with methanol 2 times and were dissolvedin 10 mL of chloroform, yielding an almost transparent suspension solutionof NPs (The inset of figure 2.12) with a concentration of ∼ 90 mg/mL, whichremains stable for several weeks. The NPs were characterized by TEM andthe micrograph (figure 2.12) confirms the size of the NPs to be ∼ 7-8 nm.KOH added After 7 mins After 1 hour After 1 hour and 10 mins(a) (b) (c) (d)Figure 2.11: ZnO nanoparticle colloid synthesis process.272.1. ZnO Nanostructure SynthesisFigure 2.12: TEM image of solution processed NPs. The inset shows thesolution of NPs in chloroform (90 mg/mL).2.1.4 Zinc Oxide NanorodsFor applications such as organic photovoltaic devices, applying nanowires andnanorods interestingly provide a higher surface area and can potentially har-vest more charge carriers once the exciton is dissociated at a donor-accepterinterface. However, exploiting CVD grown NWs is not quite compatible withthe OPV fabrication process due to high temperature fabrication process aswell as high aspect ratio of the CVD grown NWs which is not desirable forOPV devices that have thin film layers. An alternative method for OPVapplications would be using hydrothermally grown NWs, which require tem-peratures under water’s boiling temperature (100 oC) [32]. This technique is282.1. ZnO Nanostructure Synthesisa convenient method for growing short and dense NW arrays [33]. As shownin figure 2.13, the synthesis setup comprises of a thermal bath controlled by atemperature controller and a heater tape. Precursor materials are dissolvedin deionized (DI) water and then placed in the thermal bath to reach the finaltemperature. Once the temperature is stable, samples are loaded into the so-lution. Once the duration of the growth is over, samples are taken out, rinsedwith DI water, and blow-dried. The hydrothermal growth initiates from athin film of ZnO NP seed layer deposited using the method introduced insection 2.1.2. A mixture of a zinc salt, zinc nitrate hexahydrate (ZnNit), andan amine, hexamethylenetetramine (HMTA) or Tetramethylethylenediamine(TMEDA), was used as the precursor. Hydroxyl ions will be released as a re-sult of thermal degradation and react with Zn2+ ions to form ZnO [34]. Thechemical reactions for ZnNit and HMTA can be summarized as the followingequations [35]:(CH2)6N4 + 6H2O → 6HCHO + 4NH3, (2.1)NH3 +H2O → NH+4 +OH−, (2.2)2OH− + Zn2+ → ZnO(s) +H2O. (2.3)Similarly, the equations can be written with TMEDA as the amine. Nanorodswere grown with three different recipes as listed in table 2.2 and the resultingNRs are shown in figure Dielectrophoresis(a) (b) (c)Figure 2.13: (a) Hydrothermal growth system used for the growth of ZnOnanorods, (b) the custom-made sample holder and (c) shows how the sampleholder is placed in the solution.Table 2.2: Hydrothermal growth parameters used in different recipes.Precursor ReactionTemp.Seed Layer Substrate FigureZnNit 25 mM HMTA 25mM 90 oC 50 nm ZnO NPs Glass Figure 2.14aZnNit 25 mM HMTA 12.5 mM 90 oC 50 nm ZnO NPs Glass Figure 2.14bZnNit 25 mM TMEDA 25 mM 90 oC 50 nm ZnO NPs Glass Figure 2.14c2.2 DielectrophoresisAfter the nanowire growth process, the substrate is ultrasonicated (using abath ultrasonic) in ethanol for a few seconds to detach the NWs from the sub-strate and disperse them in a solvent such as ethanol. Sonication is usuallyfollowed by filtration to remove the large bulky particles from the solution.The nanowires were aligned between micro-patterned gold electrodes usingDEP, by applying a sinusoidal voltage with an amplitude of ∼ 5-10 V and afrequency of 0.5 to 2 MHz for 1-3 minutes, so as to generate an alternating302.2. Dielectrophoresis(a) (b)(c)Figure 2.14: SEM micrographs of three samples synthesized using differentrecipes for hydrothermal growth of ZnO NRs. Scale bar is 1 µm.electric field between the smallest gaps of the electrode pattern. The voltageamplitude of the DEP is tuned based on the gap size, in order to obtain ahigh probability for capturing a nanowire in the gap. It was observed thatthe change in the frequency of the AC signal, from a few 100 kHz to a fewMHz, did not have any significant effect on the DEP process. Similar to theliterature [36, 24, 23], a larger voltage amplitude results in a stronger DEPforce. Nanowire alignment is usually followed by a post annealing processat 300 oC for 5 minutes, in order to improve the integrity of electric con-tact between the electrode and NW, and the removal of the residual organic312.2. Dielectrophoresissolvent. Figure 2.15 illustrates the set-up used in the DEP process. Thereare two probes that apply the voltage to the gold electrodes and are con-nected to a signal generator. There is an oscilloscope to check the signal atthe probes, which provides an accurate measure of the potential between theelectrodes. To observe the effect of the DEP force, a few drops of the NWSignal generatorOscilloscopeStereo MicroscopeMicro-probesFigure 2.15: Dielectrophoresis set-up comprised of two probes connecting toa signal generator, under a stereo microscope.solution were placed for 10 mins on a patterned sample with and without theDEP force applied, and then checked under SEM. As shown in figure 2.16,there are a large number of NWs deposited at the electrode gaps only underDEP force, which proves the effect of the changing electric field in attractionof the NWs. The main challenge for the DEP process is to optimize theprocess in a way that would result in the deposition of a single NW at eachelectrode gap rather than having a group of NWs agglomerated together asshown in figure 2.17. This procedure involves considering various factors, as322.2. Dielectrophoresis(a)(b)Figure 2.16: Effect of dielectrophoresis force on NWs. A drop of NW solutionis placed on top of patterned electrodes (a) without DEP and (b) with DEPforce. Scale bar is 100 µm.mentioned in chapter 1, such as the electrode geometry, NW concentration,and voltage. In order to optimize the process, first a voltage amplitude,strong enough to capture the NWs, (10 V for 3 µm gaps and 7 V for 2 µmgap sizes) along with a frequency of 0.5 MHz are chosen, accordingly theconcentration of the NWs in the solution and the duration of the DEP areset to obtain single NW alignment at each electrode gap. Since finding the332.2. Dielectrophoresis(a) (b) (c)(d) (f)Figure 2.17: (a)-(c) Heavy deposition of NWs without DEP optimization.(d)-(f) Single NW deposition when the DEP parameters are optimized.exact concentration is not feasible, first, 5 µL of the NW solution is placedon the patterned electrodes and after running the DEP, based on the SEMimages, the duration as well as the concentration can be adjusted. Figure2.17a-c display devices after DEP when the process is not optimized and fig-ure 2.17d-f show devices with similar electrode geometry after optimization,which result in single NW deposition at each electrode gap. Although it isimportant to optimize the DEP parameters, not necessary all the resultingdevices show identical electrical behaviour. The electrical characterization ofthe DEP-assembled devices will be discussed in chapter 3. In addition, theelectrical behaviour (current-voltage characteristics) of several extra devicesare shown in Appendix.Figure 2.18 illustrates how the electrode geometry can affect the NW342.2. Dielectrophoresisdeposition. An analogy can be established by placing two electrodes (fingers)directly in front of each other, at a specified distance, as shown in figure 2.18a.The formation reveals that this positioning of the fingers is appropriate forcapturing multi NWs. In this situation, optimization is possible by changingthe density of the NWs and the duration of DEP for capturing a single NW.By repositioning the previous electrodes, as shown in figure 2.18b, so thatthe adjusted parallel electrodes still maintain the same distance from eachother. In this case, the interdigitated electrodes captivate a larger area ofthe electric field with a fairly high strength (darker areas), compared to theprevious case, and as a result, there is a higher chance of multi NWs beingcaptured. Furthermore, in some parts of the electrode gap, the distancebetween the fingers is larger than the length of the contained NWs andthis results in NW bridging, that introduces a new level of complicationand is not desirable for the purpose of this experiment. All in all, the bestgeometry proposed based on our experiences, is having thin, narrow, andstraight fingers, directly placed in front of each other (figure 2.18c), as wasdone in the first analogy. Although with this method the conditioning canbe idealized to capturing a single NW, the fabrication of these thin fingers isnot as straightforward as the others, and usually involves a more expensivephotolithography mask. Moreover, the process of making electrodes withsmall feature sizes, has a much lower yield than the ones with usual featuresizes (> 1 µm).352.2. Dielectrophoresis(a)(b)Au AuAuAuAu Au(c)Electric FieldFigure 2.18: Effect of electrode geometry on DEP is shown for (a) two elec-trodes directly in front of each other, (b) two neighbour parallel electrodesand (c) proposed geometry, two narrow electrodes in front to each other. Thegrey area between the electrodes represents the electric field and the darkerareas show stronger field.362.3. Sensor Testing Setup2.3 Sensor Testing SetupOnce the devices are made, they are tested using a probe station connectedto a semiconductor analyzer (Keithley 4200). The probe station is equippedwith a temperature controller and a heater as well as a turbo pump anda pressure gauge. Devices are tested at different hydrostatic pressures andtemperatures in this probe station chamber (all the measurements of section3.1 are done with this equipment).(a) (b)(c)Figure 2.19: (a) Probe station with a vacuum chamber, (b) semiconductoranalyzer and (c) turbo vacuum pump with pressure monitor.Although the probe station is very helpful for the characterization of372.3. Sensor Testing Setupthe fabricated devices, for many applications such as humidity sensing, CO,and oxygen gas sensing, it is required to have a test chamber, as small aspossible, in order to have minimum inertia for transient behaviours. TheSemiconductor AnalyzerChamberFlow Controller & ReadoutTemperature ControllerWirebonded SensorChamberUV lampBubblerBubblerMass Flow ControllersInputGasesExhaustFigure 2.20: Testing setup for gas and light sensing.382.4. Photovoltaic Device Fabrication and Characterization Setupsystem shown in figure 2.20 was developed for this purpose. As shown, adevice is wire-bonded to a prototype PCB board and placed in a custommade aluminium chamber. The inlet of the chamber is connected to a gasmixer which consists of 4 mass flow controllers (MFCs). One of the inputsof the mixer is humidified air supplied by a bubbler filled with DI water. Forhigher temperatures, the chamber is wrapped with heater tapes connected toa temperature controller (Digi Sense) and covered with a thermal insulator.The outlet of the chamber goes to a T shape connector that adds a humiditymeter (TPI 597) on the way to exhaust. Moreover, the sensors are electricallyconnected to a cable which goes to the semiconductor analyzer.2.4 Photovoltaic Device Fabrication andCharacterization SetupOrganic photovoltaic devices were fabricated on a 2×1 cm glass substratesin a glovebox as shown in figure 2.21. For measuring PV performance, thefabricated devices are placed in a custom made holder and the holder is placedin front of a 100 mW white light, passing through a AM1.5 filter. The devicesare connected to a source-measure Keithley 2400 that is controlled by aLabview software on a computer. The setup also has a monochromator whichis used for measuring the quantum efficiency by applying monochromaticlight and then measuring the current and voltage of the device.392.4. Photovoltaic Device Fabrication and Characterization SetupThermal EvaporatorWhite light MonochromatorSource Measure Unit(a) (b)(c) (d)Figure 2.21: (a) Glovebox used for fabrication of the PV devices, (b) custommade holder for OPV devices, (c) measurement system for characterizationof the PV device and (d) Labview based software used to characterize thePV performance.40Chapter 3Sensing Properties of ZnONanowires3.1 Oxygen and Vacuum Sensing3.1.1 IntroductionZinc oxide (ZnO) nanowires (NWs) and nanostructures have become a pointof interest in recent years due to unique electromechanical and chemical prop-erties with applications in large-area and flexible electronics, electronic skin,photonics, piezoelectric power generation and chemical sensing [3, 4, 5, 37, 7,8]. In particular, this metal oxide semiconductor is promising for gas sensingapplications due to the presence of oxygen vacancies on the surface [12, 13].The high surface-to-volume-ratio, which leads to a superior sensitivity to en-vironment, coupled with the diversity of structural forms (e.g., nanowires,nanorods, nanobelts, tetrapods, and nanospring [14, 15, 38, 16]) have madeZnO nanostructures a potential candidate for different nanoscale devices. Re-cently there have been numerous reports for the use of ZnO nanostructures413.1. Oxygen and Vacuum Sensingfor development of oxygen sensors and vacuum gauges that are essential partsof different equipment such as combustion engines, vacuum thin film deposi-tion systems, biomedical sensors and industrial process control systems. Oneof the challenges in fabrication of nanowire-based devices is the manipulationand alignment of a single nanowire in a bottom-up approach between metallicelectrodes. Recently, there have been reports on the use of dielectrophore-sis (DEP) as a high-precision bottom-up assembly technique with a yieldas high as 98.5% [36] (number of electrode gaps with only single nanowiresrelative to the total electrode gaps that was more than 16,000 over an areaof 400 mm2), signifying the potential of this method for efficient assemblyof transistors and sensors. In this work, we have used DEP for fabricationof ZnO nanowires on a planar network of micro-patterned gold electrodes.After describing experimental details in the following section, we present asystematic analysis of current-voltage characteristics of ZnO devices underdifferent levels of hydrostatic pressure of gases such as air, nitrogen and argonat different temperatures, so as to gain better understanding of the sensingmechanisms.3.1.2 Experimental DetailsP-type (100) oriented silicon was used as substrate for growth of ZnO nanowires.The substrates were cleaned in hot acetone and isopropyl alcohol followed bystandard RCA1 and 2. In order to remove the native oxide, hydrofluoricacid (HF) etching was used. Gold nanoparticles (NPs) were employed as423.1. Oxygen and Vacuum Sensingcatalysts for vapour-liquid-solid (VLS) growth of ZnO nanowires. Here, theNPs were immobilized on silicon by immersing hydrogen terminated waferin gold colloid solution for 2 min while 1µL of 49% HCL was added to thecolloidal solution resulting in an overall 2-3 pH [30]. Then, ZnO nanowireswere grown in a horizontal tube furnace using high purity, 99.999%, Zn pow-der as precursor. The furnace was heated up to 700 oC with a slope of 50oC/min and under a constant flow of 100 sccm comprising 2% oxygen and98% Argon. A zinc crucible was placed in the center of the tube and the sili-con substrate placed 10-20 cm away down in the tube. Figure 3.1a displays atypical SEM image of a synthesized nanowire. The XRD along with electrondiffraction pattern of these NWs reveal a single crystalline structure with apreferred c-axis growth. EDX spectrum of the sample demonstrates oxygen,zinc and silicon as the only elements in our sample. After growth, the sub-strate was ultrasonicated in ethanol for a few seconds followed by filtrationto remove the large bulky particles. Then, the dispersion was centrifuged toseparate the NWs, resulting in a uniformly dispersed solution of ZnO NWin ethanol. The nanowires were aligned between micro-patterned gold elec-trodes (Figure 3.2) using DEP, by applying a sine voltage with amplitudeof 4-8 V and frequency of 0.5 to 2 MHz for 1-2 minutes, so as to generatean alternating electric field between the smallest gaps of the electrode pat-tern. The amplitude and frequency were tuned based on the gap size andnanowire concentration in the solution, so as to obtain a high probability forcapturing a nanowire in the gap. A sudden increase in AC current indicates433.1. Oxygen and Vacuum Sensinga successful capture of a NW in the gap. Nanowire alignment is followed bya post annealing process at 300 oC for 5 minutes to improve the integrityof electric contact between the electrode and NW and removal of residualorganic solvent.3.1.3 Measurements and AnalysisThe measurements were performed in a vacuum probe station that was ini-tially pumped down for 2 hours to reach a vacuum of 0.01 mBar. Then,different gases (e.g., compressed air, nitrogen and argon) were injected toFigure 3.1: (a) FE-SEM micrograph, (b) X-ray diffraction pattern, (c) elec-tron diffraction pattern obtained from TEM, (d) and energy dispersive X-rayspectroscopy of as-synthesized ZnO nanowires. Scale bar is 2 µm.443.1. Oxygen and Vacuum Sensingthe system to reach different pressure levels, followed by measurement ofcurrent-voltage characteristics of the devices. We have observed Ohmic orSchottky contact properties for different devices, which is consistent withsimilar DEP-assembled devices in literature [39].Figure 3.2(b) illustrates the change in current as a function of the hydro-static pressure of the chamber for a device with an Ohmic contact, showingan increase in conductance with the decreasing pressure. This change canbe attributed to two mechanisms. The first mechanism is the adsorption ofreactive oxygen species such as O2−, O2− and O− [40] on the sensitive ZnOsurface facilitated by surface defects, in particular oxygen vacancies of metaloxides. The negatively charged species on the ZnO surface NW deplete theelectrons in the conduction channel, resulting in a lower conductivity and asmaller current. The other mechanism that is suggested to play a role is thepiezoelectric and piezoresistance effects in ZnO nanowires due to the asym-metric crystal structure. To identify these effects, we have evacuated thechamber and filled it with other gaseous species such as argon and nitrogenand monitored any change in current. Since these gases contain no oxygenspecies to interact with ZnO surface, only the piezoelectric effect is presentto impact conductance. For these gases, the changes in current were smallas compared to experiments with air, as depicted in the inset of figure 3.3and no meaningful dependence of current on pressure can be extracted con-sidering the measurement accuracy. These results are consistent with otherreports in literature [6, 41].453.1. Oxygen and Vacuum SensingFigure 3.2: (a) A schematic of a multi-fingered device structure showing theplacement of NWs and adsorption of oxygen species on the surface of theNW during pressure experiments. (b) Current-voltage characteristics of aZnO NW device under different air pressure. Scale bar is 2 µm.463.1. Oxygen and Vacuum SensingFigure 3.3 illustrates the relative sensitivity of devices to air pressure atdifferent temperatures. The relative sensitivity at temperature T is calcu-lated by (∆I/I)T when pressure changes from 1 mBar to 0.01 mBar and isnormalized by (∆I/I)300K at room temperature. While the current increasesat elevated temperatures due to the higher density of intrinsic carriers in ZnONWs, the devices show a higher sensitivity to pressure at higher tempera-tures, which is believed to be due to the type of the dominant reactive speciesat different temperatures [40], as well as the increased surface activity of theZnO nanowire at elevated temperatures [7]. At lower temperatures, oxygenFigure 3.3: Relative sensitivity to pressure as a function of temperature. Theinset illustrates the measured resistance of the NW device as a function ofpressure of air, nitrogen and argon.473.1. Oxygen and Vacuum Sensingmolecule absorbs one electron and form O2−. At temperatures higher than100 oC or 300 oC, oxygen molecule breaks down chemically and accepts moreelectrons to form 2O− or O2−, respectively [40]. Hence, an increase in tem-perature can effectively change the number of electrons per oxygen speciesadsorbed on the ZnO surface NWs, resulting in a higher sensitivity to pres-sure. It was also observed that the effect of pressure is more pronounced fordevices with Schottky contacts as shown in figure 3.4. In these devices, thereis a significant change in the threshold voltage of the device as a function ofchamber pressure. Here, the negatively charged oxygen species close to thecontact area cause a change in band bending at the contact vicinity, as de-picted in the inset of figure 3.4. Consequently, the height as well as the shapeand depth of the Schottky barrier between the metallic contact and semicon-ducting ZnO NW varies depending on the number of chemisorbed species[42]. As the density of adsorbed oxygen species increases, the change in thebarrier drastically affects electron tunneling and leads to a larger thresholdvoltage for the device. We believe the change in both height and the depthof the barrier increases the sensitivity of Schottky contact devices to pressurebeyond the effect due to the depletion of the channel from electrons by theoxygen molecules adsorbed on the surface observed in Ohmic devices. Thechemical adsorption and desorption of oxygen species at high and low pres-sures, respectively, show different time-dependence in our measurements. Inorder to observe the time-dependence of these processes, we evacuated thechamber from atmospheric pressure to 1 mBar and monitored the current of483.1. Oxygen and Vacuum SensingFigure 3.4: Current-voltage characteristics of a ZnO device with Schottkycontacts under different pressure levels, showing a change in threshold forconduction as a function of pressure. The inset shows the band diagram atthe contact vicinity after (solid lines).a device at a set bias point. The transient current shows a slow settlement(∼ 120 seconds) to 5% of the final current value at room temperature duringthe desorption process. On the other hand, by increasing the pressure from 1mBar to atmosphere, the current change was almost instantaneous for differ-ent devices and repeated measurements. This indicates that the desorptionof negatively charged species on NW surface is a slower process as comparedto the absorption, and is another piece ofµW evidence for the role of chemicaladsorption processes in the observed change in conductance, as opposed topiezoelectric effects.493.2. Ultra Violet (UV) Sensing3.1.4 ConclusionIn conclusion, we have demonstrated scalable fabrication of ZnO NW oxygensensors using a dielectrophoresis process. Devices with Ohmic or Schottkycontacts were observed, with distinctly different sensitivity to air pressure.By comparing the effect of pressure for different gases (nitrogen and argon),the underlying mechanism for the change in conductance is attributed to thechemical adsorption and desorption of oxygen species such as O2−, O2− andO− on the surface of the NW, as opposed to piezoelectric effects of the NWs.The temperature-dependent and time-dependent experiments demonstrateda higher sensitivity to pressure for devices at higher temperature and a slowerdesorption process as compared to absorption.3.2 Ultra Violet (UV) Sensing3.2.1 Nanowire Synthesis and CharacterizationsSynthesis of NWs was carried out in a horizontal tube of a CVD system.The NWs were grown on a (100) silicon wafer coated with a 100 nm ther-mally evaporated gold layer as substrate. Equal molar ratios of zinc oxide(99.999%, Alfa Aesar) and graphite powder (99.9%, Alfa Aesar) were usedas the precursor for NW growth. Horizontal tube furnace was heated up to950 oC and then a quartz boat containing the precursor mixture and silicongrowth substrates was loaded inside the furnace in such a way that the source503.2. Ultra Violet (UV) Sensingmaterial was located in the middle of the hot zone and the silicon substratesat 12-14 cm far from the source on downstream (∼ 800 oC). Furnace is con-nected to a vacuum pump that maintains the pressure level inside the tube at1 mbar under a constant flow of 30 sccm comprising 90% Ar and 10% oxygen.The growth process takes 2-3 hours depends on the desire length of the NWs.The as-grown NW arrays were characterized by scanning electron microscope(SEM), X-ray diffraction (XRD), and photoluminescence spectroscopy (PL).Figure 3.5 (a) illustrate a typical 45-degree tilted SEM micrograph of thesynthesized ZnO NW arrays, showing a high density of vertically alignedNWs grown uniformly, and (b) displays a higher magnification SEM imageconfirming NWs with average diameter of 80-130 nm with an approximatedensity of ∼ 25 NW/µm2. In addition, the strong intensity of ZnO (002)diffraction peak (Figure 3.5 c) confirms the vertical alignment of the singlecrystalline NWs. The photoluminescense properties of ZnO nanowire arrayswere examined at room temperature. As shown in figure 3.5d, a narrow bandultraviolet (UV) emission was observed in PL spectra at the wavelength ofabout 385 nm at room temperature using a pulsed laser (35 ps pulse dura-tion and 10 Hz repetition rate) with low excitation wavelength of 266 nmand intensity of 5 mW/cm2. The UV emission can be attributed to a nearband edge transition of ZnO NWs.513.2. Ultra Violet (UV) Sensing(a)(a)(d) 385nm(b)(c)Figure 3.5: (a) low and (b) high magnification 45o-titled SEM micrographof vertically-aligned ZnO NWs (scale bar is 10 µm and 2 µm for a and brespectively), and (c) X-Ray diffraction pattern and (c) Photoluminescencespectrum of the synthesized ZnO NWs.3.2.2 Device FabricationA 25 mm2 substrate onto which the NWs were grown was ultra-sonicated(using a bath ultrasonic) in 2 mL of ethanol for a few seconds to detachthe NWs from the substrate and disperse them in ethanol. N-type Si (100)wafers with 300 nm thermally grown silicon oxide were used as the substratefor device fabrication. Micro-patterned gold electrodes (100 nm thick and 3µm channel width) were developed by standard photolithography and lift-off process. ZnO NWs were then deposited between micro-patterned goldelectrodes using DEP by dropping 5 µL of NW suspension in ethanol and523.2. Ultra Violet (UV) Sensingapplying a sine voltage with amplitude of 10 V and frequency of 0.5 MHzfor 2 minutes. These values for amplitude and frequency were tuned basedon the gap size and NW concentration in the solution in order to obtain thehighest probability for capturing NWs in the gap. The AC signal generatesan alternating electric field in the gap between gold electrodes and induces adipole in ZnO NW so that the polarized NW is attracted towards electrodegap under a high electric field gradient [43]. Inset of the figure 3.6 shows SEMmicrographs of a device after DEP and capturing of NWs between metallicelectrodes. Figure 3.6 illustrates typical current-voltage characteristics ofa device at room temperature as well as schematic of a device. As seen,the device shows a resistive behavior with a relatively good Ohmic behaviorfor the applied voltage range of -10 to 10 V. The fabricated devices werewire-bonded to a PCB board and placed in a custom made aluminum en-vironmental chamber electrically wired to a semiconductor characterizationsystem (Keithley 4200-SCS). The chamber is connected to a gas mixer con-trolled with fully controlled MFCs (MKS 1179) and a digital readout (CCR400A) run by a computer. Moreover, the chamber as well as connected tubesand fittings were wrapped with heating tape connected to a temperaturecontroller equipped with a K-type thermocouple. The steady-state as wellas transient current-voltage characteristics are systematically measured andinvestigated under various conditions described in the following sections.533.2. Ultra Violet (UV) SensingAuAuZnO NW2 μmO2- O2- O2-O2- O2- O2-Depletion  Layer-10 -5 0 5 10-2-10123  Current (nA)Voltage (V)SiSiO2Figure 3.6: Typical current-voltage characteristics of a working device. ADEP-assembled device with a NW captured between two gold electrode aswell as the schematics of a NW aligned across two metallic electrodes as areshown in the insets.3.2.3 Results and DiscussionZinc oxide has a wide band gap (∼ 3.37 eV [44]) and as shown in the PL re-sults in figure 3.7d, the observed UV emission wavelength of 385 nm at roomtemperature corresponds to this band gap energy. Figure 3.7a illustrates thesensitivity of the current-voltage characteristics of the ZnO NWs to differ-ent intensities (from 0 to 1220 µW/cm2) of UV irradiation (peak intensityλmax of 365 nm). The intensity of the incident UV irradiation was measuredseparately using a photodetector (Newport 818-UV). As shown in the figure,543.2. Ultra Violet (UV) Sensingunder UV illumination, carriers are generated by a band-to-band transition,resulting in the observed dramatic increase in the current [45, 46, 47]. Forinstance, a UV intensity of 1220 µW/cm2 increases the electrical currentby more than four orders of magnitude, as compared to the dark condition(more samples were tested for UV sensing and the sensitivities are reportedin Appendix). Such a high sensitivity to UV irradiation in comparison toother ZnO UV detectors can be attributed to the low carrier density underdark conditions due to the oxygen rich CVD synthesis of our NWs 8. Fig-ure 3.7b depicts the transient response of the sensor to the UV irradiationpulses (300 s) with different intensities, showing the scaling of the currentwith the pulse intensity. Figure 3.7c demonstrates the transient responseof the sensor normalized to dark current to four pulses (120 s) of UV withthe same intensity 1220 µW/cm2 (and also for 5 times attenuated intensityof 250 µW/cm2), signifying the repeatability of the response. As observedfrom these two figures, the sensors demonstrate fast and repeatable transientresponses to UV pulses with a response time (10% to 90% of final value)of ∼ 38 seconds and a recovery time (90% to 10%) of 12 seconds for themaximum intensity. For the 20% intensity, the response and recovery timesare 21 and 6 s, respectively. The UV responses of the sensors are tested atthe atmospheric and reduced pressures by placing the devices in a vacuumchamber connected to a semiconductor analyzer. To ensure the removal ofhumidity, the chamber is purged and re-filled with the dry air (H2O < 10ppm). As seen in the inset of figure 3.8, the resistance of the sensor under553.2. Ultra Violet (UV) Sensing Dark Current 510  μW/cm2 850  μW/cm21220  μW/cm2 0 1000010000 Figure 3.7: (a) Current-voltage characteristics of the sensor under UV illumi-nation with different intensities. The transient response of the sensor to (b)the UV pulses with different intensities, and (c) four successive UV pulseswith the same intensity (1220 µW/cm2: black, and 5 times attenuated: red)dark condition decreases as the pressure reduces, due to the desorption ofpreviously chemisorbed oxygen species [48] from the surface of NWs thatleads to release of electrons contributing to improved conduction in the NW.Figure 3.8 illustrates the transient response of the sensor to a UV pulse (180s) at the atmospheric and 0.005 mbar pressures. Although the device showssimilar photocurrents under both pressure levels, it exhibits different recov-563.3. Relative Humidity Sensingery behaviors. It is deduced that the response and recovery of a ZnO NW toUV light is not solely due to bulk generation and recombination processes.Under UV irradiation the release of oxygen molecules from the surface canoccur due to the migration of holes toward the surface and reaction with thenegatively-charged chemisorbed oxygen species (h+ + O2− → O2) [49]. As aresult, when the UV is turned OFF, while the majority of electrons and holesrecombine rapidly, a residual electron population associated to the releasedoxygen molecules will remain until oxygen is reabsorbed on the surface. Asexpected this recovery is dramatically slow under vacuum conditions in com-parison to the atmospheric pressure. The same behavior, slower recoveryunder vacuum condition, is also observed previously [50, 51].3.3 Relative Humidity SensingHumidity monitoring and accurate control are essential for a broad rangeof applications and processes such as automotive, environmental, manufac-turing, medical, semiconductor and food industries [52, 53]. Over the pastyears, a number of sensing techniques were studied and used from wet anddry bulb thermometry to piezoelectric quartz thin films, as well as, capac-itive and resistive detectors [54, 55]. The mechanism behind the observedsensitivity is the change in properties of the sensitive material due to thechemical or physical reaction of the sensor materials in the presence of thewater molecules. Despite the diversity of the sensing mechanisms, most of573.3. Relative Humidity SensingAtmospheric PressureVacuumAtmospheric PressureVacuum1E-3 0.1 10 10001234Pressure (mBar)Relative Resistance(R/R0.005 mBar)Figure 3.8: Transient response of the sensor to a 180 s UV pulse underatmospheric vs. reduced pressures. The inset shows the resistance of thesensor as a function of pressure under dark conditions.the conventional techniques are deficient in satisfying the demand for accu-rate, fast, low cost, small footprint, chemically-stable and accurate sensingof relative humidity at low device operational temperature. Metal oxidenanostructures present suitable candidates for humidity sensing applicationas they are chemically and thermally stable [56], provide larger surface-to-volume ratio for interaction with the environment as compared to bulk thinfilm materials, and consequently, show high sensitivity with fast response. Inrecent years, metal oxide NWs and nanostructures such as ZnO [57, 48], tinoxide (SnO2) [58, 59], titanium oxide (TiO2) [60], indium oxide (In2O3) [61],583.3. Relative Humidity Sensingand tungsten oxide (WO3) [62] have been used for fabrication of a varietyof chemical sensors. Among these materials ZnO is promising as a sensitivematerial for humidity sensing due to the morphological diversity (e.g. NWs,nanorods, and nanobelts [63]), the variety of synthesis methods such as chem-ical vapour deposition (CVD) [64, 65], laser ablation [66], and hydrothermalmethod [67], and the low cost of the materials. One of the main obstacleson the way toward nano-scale devices fabrication is a low-cost manipulationand fabrication method for development of functional and accurate devices.Recent advancements in the dielectrophoresis (DEP) as a low-cost, scalableand efficient technique with a yield as high as 98.5% [36], proves the potentialof this approach for assembly of variety of devices such as transistors and sen-sors [48, 36]. In this work, we use CVD grown forests of ZnO NWs, preparea suspension of ZnO NWs in a solvent, and then apply DEP for placementand fabrication of a single NW between micro-patterned gold electrodes.3.3.1 Nanowire SynthesisSynthesis of NWs was carried out on a (100) p-type silicon substrate coatedwith a 300-nm thermal silicon dioxide layer, followed by a 100-nm thermally-evaporated gold film. Equal molar ratios of zinc oxide (99.999%, Alfa Aesar)and graphite powder (99.9%, Alfa Aesar) were used as the precursor for NWgrowth. Horizontal tube furnace was heated up to 950 oC with a constantrate of 100 oC/min and then a quartz boat containing the source powdermixture and silicon growth substrates was loaded inside the furnace in such593.3. Relative Humidity Sensinga way that the source material was located in the middle of the hot zoneand the silicon substrates at 12-15 cm far from the source on downstream.Furnace is connected to a vacuum pump that maintains the pressure levelinside the tube at 1 mbar under a constant flow of 30 sccm comprising 90% Arand 10% oxygen. The resulting NW arrays were characterized by scanningelectron microscope (SEM), X-ray diffraction (XRD), transmission electronmicroscope (TEM), as well as, energy dispersive spectroscopy (EDS). Figure3.9 (a) displays a typical SEM micrograph of the synthesized ZnO NW arrays,showing high density of vertically aligned NWs with average diameter andlength of 100 nm and 5 µm, respectively. The TEM image proves the singlecrystalline structure of these NWs and uniformity. In addition, the strongintensity of ZnO (002) diffraction peak confirms the vertical alignment of theNW arrays. As seen in this figure, the presence of the elements zinc, oxygen,gold, and silicon are confirmed by EDS measurements.3.3.2 Device FabricationThe synthesized NWs were sonicated in ethanol for a few seconds. Sonicationwas followed by filtration to remove any bulky particle resulted in a uniformlydispersed solution of ZnO NWs in ethanol. N-type Si (100) wafers with 300nm thermally-grown oxide were used as the substrate for device fabrication.100 nm-thick gold micro-patterned electrodes were fabricated by standardphotolithography and lift-off process, with a 3 m electrode gap as shown inthe inset of figure 3.10. ZnO NWs were deposited from the prepared suspen-603.3. Relative Humidity Sensing2 4 6 80 KeVOSiAuZnAuAuZnAu ZnCounts20004000(a)(b) (d)(c)Figure 3.9: (a) 45o titled SEM micrograph of vertically-aligned ZnO NWs(scale bar is 5µm), and (b) XRD measurements, (c) TEM photomicrographand crystalline diffraction pattern (d) and EDS elemental analysis results forsynthesized ZnO NWs.sion between micro-patterned gold electrodes using DEP by dropping 5 µLof NW suspension ink and applying a sine voltage with amplitude of 10 Vppand frequency of 0.5 MHz for 1-3 minutes. These values for amplitude andfrequency were tuned based on the gap size and NW concentration in thesolution in order to obtain the highest probability for capturing a NW in thegap. This AC signal generates an alternating electric field in the gap betweengold electrodes and induces a dipole in ZnO NW so that the polarized NWis attracted towards electrode gap which has a high electric field gradient.Dielectrophoresis is followed by a post annealing process at 250 oC for 10minutes in N2 ambient to evaporate any organic residues and improve elec-613.3. Relative Humidity Sensingtrical contact between the electrode and NW. Figure 3.10 illustrates typicalcurrent-voltage characteristics of the resistive device at room temperature,as well as, the schematic and field emission SEM micrograph of a single NWdevice. As seen, the device shows a resistive behavior with ohmic contactproperties for the applied voltage range of -5 to 5 V.Figure 3.10: Current voltage characteristics of a typical device at room tem-perature. Insets show SEM image of a single NW device and schematics ofthe experimental set up. Scale bar is 3 µm.3.3.3 Results and DiscussionFabricated devices were wire-bonded and placed in a custom made aluminumchamber connected to a semiconductor characterization system (Keithleymodel 4200-SCS). The chamber has inlet and outlet that is connected to an623.3. Relative Humidity Sensinginjection system that can provide dry and humidified air with controlled RHas monitored by an independent (0.1% RH accuracy) digital humidity meter(TPI 597) connected to the chamber. The chamber is filled with air with de-sired RH and the steady-state current-voltage characteristics were measuredat a stable RH reading. As depicted in figure 3.11, increasing RH significantlyincreases the current, measured when sweeping the applied voltage from -5V to 5 V, with a step of 0.25 V. In fact, as the inset of figure 3.11 shows in asemi-logarithmic plot, the NW resistance decreases with a high sensitivity asthe RH increases. This enhancement in conductivity in the presence of hu-midity is in agreement with previous reports [57, 68, 69]. This change can beattributed to the adsorption of water molecules on the ZnO surface NW simi-lar to other metal oxide semiconductors [53]. As seen in figure 3.10 and figure3.11, the current-voltage characteristics of the ZnO NWs in this work show apredominantly ohmic behaviour, while the presence of linear and non-linearcontact resistances cannot be ruled out. Due to the consistent measuredcurrent-voltage characteristics for several devices with different overlap areasbetween NW and contacts, we can attribute the observed response to thesurface of NW exposed to the humidity. Although water molecules can dis-sociate and donate electrons to ZnO through a chemisorption process, thisusually requires high energy and does not happen at temperatures below 100oC [70]. As a result, water molecules that are physically adsorbed on thesurface of the NW lead to a charge transfer on the surface that acts as agate controlling the electron density in the core. Since our experiments were633.3. Relative Humidity Sensingcarried out at room temperature, the mechanism for change in current is thecompetitive replacement of the chemically pre-adsorbed oxygen species onthe NW surface exposed to dry air [40], with physisorbed water molecules asshown in figure 3.11b. The result is the release of electrons in the NW core,which promotes the conductivity [59, 70]. In conventional thin film sensors,the mechanism often used for describing the change in conductance is themodification of the depletion region at the surface as the number of adsorbedmolecules change, which implies a linear change of conductance with RH. In aNW sensor, this mechanism happens where the Debye screening length in theNW is smaller than its radius. However, in our devices, as shown in the insetof figure 3.11, the exponential increase in the NW conductivity with increas-ing RH suggests that the NW is operating in the subthreshold regime [20],where current through the channel changes exponentially with the change ingate potential (modified by physisorbed water molecules), hence exhibitingmuch higher sensitivity to adsorbed species on the surface of the NW. Thisexponential behavior can be attributed to the fact that the Debye screeninglength is larger than NW radius so that the whole NW volume is affected bythe gating of molecules on the surface. For the Debye screening length to beat least two times the NW radius (50 nm), the carrier concentration in theNW should be lower than 1015 cm−3, estimated fromλZnO =√εZnOε0kBT/e2n (3.1)643.3. Relative Humidity Sensingwhere εZnO is the dielectric constant of ZnO and ε0 the vacuum dielectricpermittivity, kB Boltzmann constant, T the absolute temperature, e theelementary charge and n the electron density. Knowing the NW conductivityunder dry air and assuming an electron mobility of 10 we can find n =2.2×1013 from R=L/(µnε0pir02) where R is the resistance of a cylindricalNW with the radius r0 and the length L (between the two electrodes), for ahomogenous electron distribution, which satisfies the condition for the Debyescreen length to be ∼ 16 times of the NW radius. Poisson’s equation incylindrical coordinates with the main z axis at the center of NW reads1rddr (rdϕdr ) = −eεZnO∆n (3.2)where ϕ represents the potential in the NW, r the radial distance from thecore of NW, and ∆n the change in the electron density at a location in theNW. By solving this equation, the change in the electron density and thus theconductance can be related to the change in the surface potential ∆ϕ due toadsorption of molecules on the surface of the NW. As mathematically shownin Ref. [20], considering the large Debye length for the electron density inthe NW, the NW resistance R2 after adsorption of water molecules on itssurface can be written in terms of its initial resistance R1 as,R2/R1 = e−(e∆ϕ/kBT ) (3.3)By fitting the previous formula to the inset of the figure 3.11a, ∆ϕ can be653.3. Relative Humidity Sensingextracted as a function of RH as ∆ϕ=a×RH where a is a constant with thevalue of 470 mV and RH attains 0 to 1. In addition, as the water moleculesare absorbed to the surface and the number of oxygen atoms masked on thesurface increases, there is a possible enhancement in mobility of electronsin the NW [51] since the Coulomb interactions between electrons and fixedcharges attenuates. Moreover, our devices take advantage of maximum ex-posure to environment as the NW is free standing and surrounded with air,resulting in a higher sensitivity.As shown in the inset of figure 3.11a, ZnO NW resistance in dry air (< 10ppm water concentration) is 88 times higher than that in 30% RH and almost5 orders of magnitude higher than the resistance in 60% RH air. The absolutesensitivity, defined as Rdry/Rhumid , i.e., Ihumid/Idry under constant voltage,is about 4000 for 50% RH which is significantly higher than that previouslyreported for ZnO nanowires and nanostructures [57, 69, 71, 72, 73] and iscomparable to the sensitivity of traditional porous ceramic composite filmsat 90% RH [74]. Based on our previous work [48], changing the environmentfrom pure nitrogen to dry air leads to three times change in resistance, whichis much lower than the observed response in this work when exposed tohumid air, thus ruling out the influence of change in oxygen concentrationas a dominant mechanism for the observed change in resistance. In order tocheck the response, recovery, and reproducibility of these devices, pulses ofhumid air were applied to the chamber and transient response of the sensorswere recorded as depicted in figure 3.12a. Pulses of 30% RH air were supplied663.3. Relative Humidity Sensing(a)(b) Dry Air:Humid Air:Figure 3.11: (a) Steady state characteristics of the sensors at different RHvalues. The inset shows a semi-logarithmic graph of resistivity vs. RH. (b)Schematic of a single NW sensor and the change in electron density of NWin dry and humid air conditions.to the chamber for 150 s, whereby the response time for the measured sensor’scurrent (10% to 90% of final value) was about 60 s. Then, a dry air pulse673.3. Relative Humidity Sensingwas supplied for another 150 s, which showed in the sensor’s current recoverytime (90% to 10%) of ∼ 3 s. To check the sensor stability, this sequence wasrepeated and the response for four RH pulses is illustrated in figure 3.12a.The experiment was also carried out for different RH conditions (17%, 25%and 35%) and the response was monitored for 1500 s and displayed in figure3.12b, indicating a stable behavior over this time. The measurements wereperformed after 1 day, 1 week, and 2 weeks and almost similar results wereobtained.Time (s)(b)(a)30%dry air(c)Figure 3.12: (a) The changes in the measured current a ZnO NW devicesin response to exposure to pulses of 30% RH air between dry air flushes.(b) The measured current of the sensor, extracted from different RH pulses(17%, 25%, and 35%). (c) Arrhenius plot of the relative sensitivity of thesensor current to RH.683.3. Relative Humidity SensingFigure 3.12c also illustrates the Arrhenius plot of the relative sensitiv-ity of the sensor at different temperatures. The relative sensitivity at tem-perature T is calculated by absolute sensitivity for 30% RH, Rdry/R30%,at temperature T normalized by the absolute sensitivity for the same RHat room temperature, which is 88. The higher sensitivity at elevated tem-peratures could be caused by higher surface activity with respect to theadsorption/desorption processes [75]. The semi-logarithmic plot of the rel-ative sensitivity vs. reciprocal temperature (1/T) is in agreement with thethermionic emission relation [76], in which the current of a heated materialexponentially increases with temperature. The effective activation energy for30% RH, Ea, can be extracted from the slope (slope =Ea/kB, where kB isthe Boltzmann constant) Ea = 0.6 eV which is consistent with similar workson TiO2 nanocomposites under 30% RH (Ea = 0.56 eV) [77].In order to investigate the response of NW-based sensors to step-like in-creasing RH level, similar devices (as shown in figure 3.5) were fabricatedand tested for relative humidity sensing. Initially, the steady state and tran-sient responses of the ZnO NWs are investigated under humidified air withcontrolled RH as shown in figure 3.13a. In the inset of figure 3.13a, thetransient response of the sensor to 86% RH pulses is shown, which signifies afast, and repeatable response. Figure 3.13b illustrates the transient responseof the sensor to step-like changes of RH from dry to 86% (10 min long steps)showing a stable operation at room temperature.693.3. Relative Humidity SensingFigure 3.13: (a) Steady-state current-voltage characteristics of a sensor undervarious RH levels. The inset shows the transient response to pulses of 86%RH between dry air purges. (b) The transient response of the sensor toincreasing RH from dry to 86%.3.3.4 ConclusionWe have demonstrated a nanoscale ultra-sensitive relative humidity sensorbased on CVD grown single-crystalline zinc oxide NWs. These devices illus-trate exponential change in its resistance in excess of 5 decades in response toa change of relative humidity from a dry air to 60% RH air at room temper-ature, associated to a subthreshold carrier modulation in the NW core, highsurface- to-volume ratio of the NW and complete exposure of the NW surfaceto air, indebted to the its free standing structure. These sensors demonstratestable behavior, reproducible switching response, with 60 s rise time and 3s recovery time in response to 30% RH pulses between dry air flushes andpronounced sensitivity at elevated temperatures.703.4. Oxygen Sensing3.3.5 Correlation of UV and RH SensingFigure 3.14 depicts the correlation of the effect of UV and RH on the sensorresponse. Figure 3.14a illustrates the transient response to the successivepulses of 80% RH air and dry air under both dark and UV irradiation con-ditions. In contrast, figure 3.14b shows the response to UV pulses underdry and 80% RH conditions. From these two measurements it is observedthat not only the sensitivity to RH drops in the presence of UV but also thesensitivity to UV diminishes substantially in the presence of humidity. Forexample the sensor’s sensitivity to 80% RH pulses with respect to dry airdecreases from 8000 to 3.5 under UV irradiation. Similarly, the sensitivity tothe UV pulses drops from 2000 to 5 in an 80% RH ambient. As mentionedearlier, both UV and humidity lead to an increased number of carriers. Asa result, in the presence of UV or humidity, the sensitivity to the other islower due to the existing carriers in the NW.3.4 Oxygen SensingIn addition to sensitivity to vacuum, adsorption and desorption of oxygenon the surface of the NW can be used for sensing the oxygen content of theambient. Devices are fabricated similar to the ones used in section 3.2.1.Figure 3.15 illustrates the response of a ZnO NW sensor to repetitive pulses(300 s) of pure oxygen and argon in the dark and under UV irradiation.As seen, the sensor response is weak in the dark and at room temperature,713.4. Oxygen SensingFigure 3.14: The transient response of the sensor to (a) the successive 80%RH air and dry air pulses under dark and UV irradiation, and (b) the suc-cessive UV pulses in the dry and 80% RH air ambient.since the surface of NWs are highly saturated with oxygen. However, whenthe sensor is under UV irradiation, the current of the sensor changes about723.4. Oxygen Sensing250% when the surrounding oxygen atmosphere is replaced with argon. Thisaffirms our previous discussion that the UV illumination facilitates removalof some chemisorbed oxygen species, thus increasing sensitivity to oxygenin the ambient. In particular, UV irradiation forms weakly-bound photo-induced oxygen ions [78], which require less energy for desorption at roomtemperature. Thermal energy can be used to overcome the energy barrierFigure 3.15: The transient response of the sensor to pulses of oxygen andargon under dark and UV irradiation at room temperature.of oxygen adsorption and desorption instead of UV irradiation [79]. Figure3.16 displays the transient response of the ZnO NWs to pulses of oxygen andargon at different temperatures (100 oC, 150 oC, and 175 oC). The increasedsensitivity at higher temperatures can be attributed to the increased surface733.4. Oxygen Sensingactivity as well as the additional role of some reactive oxygen species thatoccur at higher temperatures. It is known that at room temperature, anoxygen molecule accepts one electron from the NW to forms O2−. However,at temperatures higher than 100 oC, the oxygen molecule dissociates andaccepts another electron to form 2O− [40]. From the Arrhenius plot, theactivation energy of this reaction is estimated to be 146 meV for pure oxygenambient for 100 oC < T < 200 oC [80].Figure 3.16: (a) The transient response of the sensor to successive pulses ofoxygen and argon at different temperatures. (b) Sensitivity as a function oftemperature, (c) and an Arrhenius plot of the sensor’s response.743.5. Carbon Monoxide Sensing3.5 Carbon Monoxide SensingFigure 3.17 demonstrates the transient responses of ZnO NW sensors (De-vices are fabricated similar to the ones used in sec 3.2.1) to pulses of CO.Oxygen species on the surface of NW can react with CO to form carbondioxide and release an electron in the NW. Under dark conditions, the sen-sitivity of the sensor to pure CO gas is low, as shown in figure 3.17a, dueto the high activation energy of the proposed reaction. However, under UVirradiation, the sensitivity increases to 75% due to the formation of reactivephoto-induced oxygen ions on the surface of the NW [78]. The improvementin the sensitivity, however, is not as remarkable as the case of oxygen. Thisdifference can be attributed to the partial release of the chemisorbed oxygenspecies from the surface of the NWs under UV irradiation, which is not de-sirable for CO sensing. Figure 3.17b illustrates the sensitivity of sensor tothe CO pulses at different ambient temperatures (i.e., 100 oC, 150 oC, 175oC and 200 oC). As seen, at the elevated temperatures, the sensitivity sig-nificantly improves due to the enhanced surface activity similar to the caseof oxygen. In addition, as reported by Takata et al. [40], at temperaturesbelow 100 oC, the dominant reactive oxygen ions are O2−, as compared tothe more reactive O− species at temperatures between 100 and 300 oC [81].Figure 3.17(c) depicts the Arrhenius plot of the sensitivity to pure CO, sig-nifying a linear behavior in agreement with the thermionic emission model[76]. An effective activation energy Ea of 180 meV for pure CO ambient can753.5. Carbon Monoxide SensingFigure 3.17: The transient response of the sensor to (a) pulses of pure COand air under dark and UV irradiation conditions, and (b) the successive 5minute long CO and air pulses at different ambient temperatures 100 oC, 150oC, 175 oC, and 200 oC. (c) The Arrhenius plot of the sensitivity of the NWsensor to pure CO.be extracted (slope = Ea/kB, where kB is the Boltzmann constant) for 100oC < T < 200 oC. Fig 3.18 shows the transient response of the sensor at200 oC ambient temperature to the successive pulses of air having differentconcentrations of CO (i.e., 10%, 20%, and 100%), which displays a mono-tonic dependence of the current to the concentration of CO. When the COpercentage varies from 10% to 100% the response times vary from 50 s to763.5. Carbon Monoxide Sensing160 s, respectively, showing a slower response at larger changes in the con-centrations of CO. However, the recovery times remain almost independentof the CO concentration and between 87 and 97 s.Figure 3.18: The transient response of the sensor to the successive pulses ofair having different concentrations of CO from 10% to 100% at 200 oC.77Chapter 4ZnO Sensors on Paper4.1 IntroductionIn the last decade, due to the soaring demand for low-cost, fast, and flexibleelectronic devices, micro and nanotechnology experienced a significant shiftin terms of substrate selection from expensive and rigid materials to recentlow-cost and flexible substrates [82]. This evolution has not only lowered theproduction cost of the electronic devices, but also has significant role in de-velopment of simplified fabrication processes. Paper is one of the most recentemerging substrates that is gaining large amount of interest due to its avail-ability and unique properties [83]. It is a low-cost and ubiquitous materialwhich can be available in various forms and costs even less than one tenthof plastic substrates [84]. Paper is recyclable and friendly to environment,also adaptable and compatible with a variety of printing as well as paintingtechniques [85, 86]. As a substrate, it provides a high surface area due to itsfibrous composition that makes it suitable for sensing applications. More-over, paper’s cellulose fibrous matrix shows strong adhesion and bonding toa broad range of materials and nanostructures [87, 88]. A nanostructured784.1. Introductionfeature is an essential requirement for fabrication of gas and light sensors,since it offers higher surface-to-volume ratios. For instance, a nanoparticle(NP) of 100-nm size is comprised of about 1% surface atoms while this ratiogoes up to 10% for a 10 nm NP [18], thus the down-sizing can significantlyboost the sensitivity to the environment. Metal oxide nanostructures, includ-ing zinc oxide (ZnO), are well-known for their sensing properties due to thepresence of surficial oxygen vacancies [48, 43]. ZnO nanostructures are at-tractive for sensor applications not only due to their morphological diversity(e.g., nanowires, nanorods, and nanoparticles), but also the variety of synthe-sis methods (e.g., such as chemical vapour deposition (CVD), laser ablation,hydrothermal, and sol-gel process) and the low cost of precursor materials.One of the main obstacles in materialization of commercial devices basedon nanostructures is the complexity and cost of fabrication methods usedfor development of functional and accurate devices. Almost all the methodsinvolve sophisticated and expensive patterning techniques such as electronbeam lithography or have fairly low yield thus the resulting devices lack astable and repeatable performance, which is a crucial factor for any commer-cial device. In this work, we present our latest results on sketching low-cost,ubiquitous, and yet accurate sensing devices prepared on regular paper usingsolution-processed ZnO nanoparticle inks. The electrodes of the sensor aredrawn by hand, using painting brush as well as a home-made stamp, ex-ploiting a custom made calligraphy ink. We believe this work in addition tosimilar reports [83, 87, 84, 88, 89], pave the way toward low-cost fabrication794.2. Device Fabricationof practical electronic and sensing devices, which can be extended to broadrange of substrates and materials [89].4.2 Device FabricationZnO nanoparticles were synthesized based on the work of Weller et al. [31].First, zinc acetate dihydrate (Sigma-Aldrich, 2.95 g) was dissolved in methanol(125 mL) on a stirring hot plate at 120 0C, a solution of potassium hydrox-ide (Fischer Scientific, 1.48 g) in methanol (65 mL) was added slowly over10 min to the zinc acetate solution under vigorous stirring. Zinc hydrox-ides precipitated but dissolved again. After almost 2 hours the NPs startto participate then heating and starring are stopped. Once NPs are all pre-cipitated, the NPs were washed with methanol 2 times and in the end theNPs were dissolved in 10 ml of chloroform yielding an almost transparentsuspension solution of NPs, shown in figure 4.1a, with an approximate con-centration of 90 mg/mL, which remains stable for several weeks. The NPswere characterized by AFM and SEM as shown in figure 4.1b and c, illus-trating the uniformity of the size of NPs and finally the TEM micrographconfirms the size of the NPs to be ∼ 7-8 nm. Since the resistance of theNP layer is in the order of ∼ Gohm even an electrode with resistance of upto a few Mohm does not affect the total electronic performance of the sens-ing device. A variety of electrodes were exploited such as evaporated silver,silver paint, pencil-drawn line, as well as carbon paste and calligraphy ink804.3. Results and Discussionand illustrated similar behaviours. We chose a custom made calligraphy inkbased on a traditional recipe by using equal volume ratios of lampblack asthe black pigments and gum arabic (Sigma Aldrich), which not only helpsthe carbon particles stabilize but also makes the ink more viscous and resultsin adhesion of the carbon particles to the paper. First the gum Arabic wasdissolved in deionized water to form a thick syrup-like solution then lamp-black was added to it. Due to the low cost, strong adhesion and bindingto paper and compatibility with drawing, painting, stamping and printingtechniques calligraphy ink was preferred for this work.Moreover, lampblack is known to be composed of as small as 10 nm carbonparticles that fuse into eachother during the combustion and can form stablenano-clusters of 100-300 nm in size [90], which was also confirmed by SEMmicrograph as shown in the figure 4.2d.4.3 Results and DiscussionThe regular copy paper is extremely porous and composed of randomly ori-ented micrometer-sized cellulose fibers, which is not suitable for many elec-tronic devices as it is hard to do various fabrication processes on such roughsurfaces. However, it is observed to be a suitable substrate for our applica-tion as it provides a porous matrix which holds the NP ink and increase theinterface area to the surrounding environment that promotes the sensitivity.As displayed in figure 4.2a, sketching steps of a sensor on copy paper are;814.3. Results and Discussionfirst using a custom made stamp with 6 1-inch long and 1-mm wide electrodelines, which have 1-mm spacing. Then three of the lines are connected usingthe ink and a brush. Although all the electrode lines can be drawn by abrush, we used a stamp to have devices with a similar size to be able tooptimize the whole process. Each stamped line has a resistance of 150-250KΩ. Once the electrode lines are drawn, then ZnO NP ink is deposited usinga painting brush. Although precise and accurate control over the amount of(d)(a) (100 nm(c)10 nmFigure 4.1: (a) Transparent ZnO NP solution in chloroform. (b) AFM imageof a spin-coated NP film. (c) High resolution SEM image of a NP thin filmand (d) TEM image of the ZnO NPs.824.3. Results and DiscussionZnO deposition was not possible as the sensors were prepared by hand, fairlyconsistent results (20% tolerance) with 3-6 times of brushing the device area.In addition, it was observed by weighing the device before and after brushingthat by 3-6 times brushing, ∼ 3-7.5 mg of ZnO is deposited on the wholedevice.(a) (b)(c) (d)50 μm 500 nm800 μm 2 cmNP on cellulose fiberElectrodeFigure 4.2: (a) Sketching steps of a sensor on copy paper using calligraphyink. First step is stamping the electrode lines followed by step II and IIIwhich are connecting the lines using a brush. Finally in step IV the NPink is painted on the drawn electrode lines, (b) A low magnification SEMmicrograph of electrode lines and active channel of the device. (c) Highermagnification SEM of the NPs on cellulose fibre matrix and (d) calligraphyink-drawn electrode comprised of carbon nano-clusters.834.3. Results and Discussion4.3.1 Ultra Violet SensingZnO is well-known for its ultra-violet (UV) detection properties originatedfrom electronic transition within its wide band gap (∼ 3.37 eV) which liesin the ultraviolet region of electromagnetic spectrum. A device made onpaper was exposed to a UV lamp with a peak wavelength at 365 nm andpower intensity of 1220 µW/cm2) at 4 cm distance measured by a Newport818-UV photo-detector. The steady state current-voltage characteristic wasmeasured before and after UV exposure by a semiconductor analyzer (Keith-ley 4200). As shown in figure 4.3a current increases ∼ 2000 times that can beattributed to two different mechanisms, first the sequential absorption of theUV photon, a band to band transition and electron-hole pair (EHP) genera-tion (as shown in the inset of figure 4.3a) and second, migration of the holestoward the surface of NPs due to the band bending and release of previouslyadsorbed oxygen at the surface of NPs [50]. These two mechanisms result ina significant boost in conductivity of the devices which is in agreement withprevious reports on the UV-sensing property of ZnO [88, 91, 46].In order to study the transient behavior of these devices to UV irradiation,a 30 s-long UV pulse was illuminated to sensors and the current was measuredas shown in figure 4.3b. The response of a sensor illustrates a fast behaviourwith response (10% to 90% of final value) and recovery time (90% to 10%) of6 s and 3 s respectively which is a quite fast response attributed to the sizeof these NPs. More importantly, as shown in figure 4.3c different intensitiesof UV light were applied to the devices and the device transient current was844.3. Results and Discussionmeasured accordingly. As observed, sensors respond according to the UVintensities as displayed in the inset of figure 4.3c.Figure 4.3: (a) Current-voltage characteristics of dark and photo-current ofa typical sensor. (b) sensor’s response to a 30 s pulse of UV light and (c)device transient behaviour under various UV intensities. In the inset thesignal from the sensor vs. the intensity of the UV is shown.4.3.2 Carbon Monoxide SensingZinc oxide like other metal oxides is sensitive to carbon monoxide. Resistivesensing of CO works based on a reaction between the gas and the negatively854.3. Results and Discussioncharged oxygen species at the surface of the metal oxide, in this case ZnO,which usually involves high temperature or alternative methods such as sur-face doping of the NW to facilitate the process [92, 93, 94]. However, byusing the ZnO NPs it was observed that these devices show fairly fast andstrong response to CO environment as shown in the figure 4.4. This interest-ing behaviour is attributed to the size of NPs (∼ 7-8 nm) and the fact that alarge proportion of atoms are surface atoms and therefore any change at thesurface, even small one, can considerably influence the electronic behaviourof the whole NP. As displayed in figure 4.4a, exposure to CO environmentresult in 200% increase in the current of the device after 5 min. In addition,it was observed that by purging the chamber with dry air the current recov-ered back to its original value that confirm the repeatability of the response.However, by applying UV light to the devices not only the sensitivity (S= ICO/Iair) increases, but also the response and specifically recovery timeimproves as shown in the Figure 4b. Under UV incident light with an ap-proximate intensity of 250 µW/cm2), the sensitivity increases more than 37times (after 5 mins of exposure to CO) compared to the one under darkcondition. This is attributable to the formation of reactive photo-inducedoxygen ions at the ZnO surface [78, 95] that reacts faster with CO.864.4. ConclusionFigure 4.4: (a) Effect of 5 min CO pulses on two typical devices and copypaper. (b) 1-hour long pulse of CO (c) sensitivity under UV vs. dark.4.4 ConclusionIn this work we have developed a low-cost, fast and simple approach for fabri-cation of UV and CO sensors based on solution processed ZnO nanoparticles874.4. Conclusionink, which can be extended to a variety of solution-processed nanostructureinks. Electrodes were sketched using the carbon calligraphy ink and the ZnOlayer was deposited using a painting brush on a regular copy paper. Thesedevices show more than 2000 times improvement in conductivity when ex-posed to 1220 µW/cm2 UV irradiation with a maximum wavelength of 365nm, and illustrate a linear response to incident UV power. In addition, whenthe devices are exposed to a 30-s UV pulse, they exhibit 6 s and 3 s responseand recovery times, respectively. Moreover, it was shown that the sensors,exposed to a CO environment for 5 mins, show about 3 times improvement inconductivity. Moreover, it was observed that applying UV irradiation resultsin not only a faster but also 37 times more sensitive response to CO.88Chapter 5Photovoltaic Devices5.1 IntroductionSolar energy has been regarded as a clean and renewable resource for genera-tion of electricity. Polymer solar energy devices have increasingly become thefocus of research in recent years [96, 97, 98, 99], since they are amenable to lowcost roll-to-roll manufacturing processes by using organic or inorganic-organichybrid inks. The power conversion efficiency (PCE) of these organic photo-voltaic (OPV) devices remains low; currently, the highest recorded efficiencyis approximately 8% [100]. In addition to low PCE, the short life-time ofOPV devices, mainly due to the degradation of organic/conducting materialinterfaces, undermines their long term cost effectiveness [101]. Bulk hetero-junction (BHJ) OPV device structures [96] based on blends of a donor poly-mer poly(3-hexylthiophene) (P3HT) and an acceptor fullerene [6,6]-phenylC61-butyric acid methyl ester (PCBM) are considered a technological bench-mark against which OPVs are often measured [102, 103, 104, 105]. In thesesystems, the electron-donating polymer and electron-accepting fullerene forma nanometer-sized, interpenetrating network [106, 107]. Incident light is ab-895.1. Introductionsorbed within the photoactive BHJ layer, generating bound electron-holepairs, i.e., excitons. Limited to their lifetime, the excitons can diffuse to theinterface of the donor and acceptor polymers, where they separate into chargecarriers. The diffusion length is generally in the order of tens of nanometres[108], giving rise to a device performance that depends heavily on the nano-morphology of the BHJ layer. Major improvements in terms of stability havebeen achieved by developing inverted OPV devices [109, 110, 111, 112]. Ina regular device structure, holes and electrons are injected into transpar-ent indium-tin oxide (ITO) and counter electrode (e.g., Al), respectively. Incontrast, in an inverted device structure electrons are injected into the ITOwhile holes are collected by the top electrode [113], which can be a less airsensitive, high work-function metal such as silver or gold. At the ITO in-terface, metal oxides such as TiO2 [114] and ZnO [115, 29, 116, 117, 118]have been applied as an electron selective contact with PCE of over 3%. Thehigh electron mobility of ZnO makes it an ideal electron selective contactlayer for inverted devices [119]. It has been shown that thin films of ZnOnanoparticles can be solution-processed on ITO/glass or flexible substrates atlow temperature [29], an essential requirement for all-printed devices. Thesedevices as well as those fabricated by a high temperature sol-gel process onITO/glass is reported to exhibit a maximum PCE of 3.6% [29], the highestvalues reported for inverted OPVs. In the present study, we have focusedon optimizing both electron- and hole-collecting electrodes for an invertedsolar cell to demonstrate high efficiency (3.8%) and improved air stability,905.1. Introductionas schematically illustrated in figure 5.1. We have achieved transparent ZnOthin films by spin coating of zinc acetate dehydrate, Zn(OAc)2.2H2O solu-tions. Moreover and in order to obtain an efficient inverted OPV device, wehave paid particular attention to the development of the hole transport layer,i.e., poly(3,4-ethylenedioxythiophene) poly(styrenesulfonate) (PEDOT:PSS)that is coated on top of the low surface energy P3HT:PCBM photoactivelayer. A thin film of gold (2-5 nm) was thermally deposited on organic layerso as to improve wettability of the photoactive layer towards the aqueous PE-DOT:PSS dispersions that present large surface tensions. The photovoltaiccharacteristics and stability of devices are investigated as a function of ZnOand gold contact layers.Figure 5.1: Inverted solar cell device structure (a) and approximate energylevel diagram (b).915.2. Experimental Section5.2 Experimental Section5.2.1 Device FabricationAs shown in figure 5.1, the inverted OPV devices had ITO/ZnO/P3HT:PCBM/Au(2-5 nm)/PEDOT:PSS/Au (100 nm) structure with an active area of 0.2cm2. The fabrication steps are as follows. ITO was patterned by maskingwith tape and etching the exposed area with conc. HCl at 60 oC. Aftersequential ultra-sonic cleaning in H2O2/NH3/H2O (1:1:5 v/v), deionized wa-ter, acetone and 2-propanol the substrates were dried at room temperature.The zinc acetate dehydrate Zn(OAc)2.2H2O (Sigma-Aldrich), 50, 100 and125 mg/mL in methanol was then spin coated on ITO at a rate of 1000 rpmfor 30 s, followed by annealing at 150-350 oC for 10 min to afford a ZnOlayer with thicknesses of 25, 50 and 80 nm, respectively. Bulk heterojunctionfilms were deposited from solutions of PCBM (American Dye Source Inc.)and P3HT (regioregularity ∼ 95%, Reike Metals) in 1,2-dichlorobenzene (1,2-DCB; 99%, Sigma-Aldrich), filtered through a 0.45 m polytetrafluoroethylene(PTFE) membrane filter (Pall Co.) prior to mixing. The blend solution wasspin-cast on ZnO/ITO substrates in a glove box with a sequential rate of 100rpm (1 s), 300 rpm (5 s) and 1000 rpm (20 s). The films were left in a coveredPetri dish to control the solvent evaporation rate (drying time: 20-40 min).A thin film of gold (2-5 nm) was thermally deposited through a shadow maskonto P3HT:PCBM film using a MBraun glovebox integrated thermal evapo-rator system at a 0.1 A˚/s deposition rate. A PEDOT:PSS (Clevios P VP AI925.2. Experimental Section4083, H.C. Starck Inc.) solution mixed with equal amount of 2-propanol wasspin coated with a sequential rate of 300 rpm (3 s), 700 rpm (60 s) and 5000rpm (30 s). Dilution with 2-propanol is necessary in order to improve itsadhesion. The PEODT:PSS coated substrates were transferred back to theglove box and thermally annealed at 140 oC for 10 min. Finally, gold (100nm) was thermally deposited (at a vacuum level of ∼ 10−6 torr) through ashadow mask to complete the PV device.5.2.2 Device CharacterizationCurrent-voltage measurements of devices were conducted on a computer-controlled Keithley 2400 Source Meter. A xenon lamp (150 W, NewportCo.) equipped with an AM1.5G filter was used as the light source. Theoptical power was 100 mW/cm2, measured using a broadband power meter(Newport Co.) equipped with a thermopile detector head. The externalquantum efficiency (EQE) was measured using monochromatic light from amonochromator (Newport Co.) and the incident power was measured using apower/energy meter (Newport Co.) that is also used for collecting transmit-tance/absorption spectra of films on ITO substrates. Bright-field transmis-sion electron microscopic (TEM), scanning electron microscopic (SEM) andatomic force microscope (AFM) images were recorded using Hitachi 8100,Hitachi S-4700 and MFP3D-SA Asylum Research systems, respectively. Anoptical microscope equipped with a CCD camera was used to examine mor-phology of the photoactive layer. Film thicknesses were measured using an935.3. Results and DiscussionAlpha-Step IQ profilometer (KLA-Tencor). For the purpose of TEM analy-sis, a 40-nm layer of PEDOT:PSS was first spin coated at 5000 rpm on ITOand baked at 140 oC for 10 min in an ambient atmosphere. P3HT:PCBMlayer was deposited on a ZnO/ITO substrate (with or without gold coating).The ITO substrates were immersed in a dilute HCl (5%) solution causing theZnO layer to dissolve thus enabling the remaining polymer film to be peeledoff. The floating films were picked up using 500-mesh TEM copper grids.5.3 Results and Discussion5.3.1 Optical PropertiesFor the preparation of ZnO contact layer, as shown in figure 5.1, a spin coatedzinc acetate solution on ITO/glass substrates was converted into ZnO by arapid heating at 150-350 oC on a pre-heated hot plate at ambient conditions.At these temperatures ZnO film is practically transparent. Depending onthe concentration of Zn(OAc)2.2H2O solution in methanol (50, 100 and 125mg/mL), the ZnO thicknesses were varied from 25 to 50 and 80 nm. Figure5.2 shows the visible spectra of ZnO with transmittance of near 100% at390-800 nm while at shorter wavelengths (320-390 nm) the transmittance ismore than 75%. The wide bandgap of ZnO ensures high transmittance inthis wide spectral range [120]. When a slow heating rate is applied the ZnOfilms became less transparent due to formation of nano-ripples as reportedrecently [116]. The UV-Vis absorption spectra of a blend of P3HT and945.3. Results and DiscussionFigure 5.2: Absorption spectrum of (a) ZnO on quartz substrates with thick-nesses of 25 nm (dashed line), 50 nm (solid line) and 80 nm (dotted line) and(b) P3HT:PCBM spin-coated from 1,2-DCB at 1000 rpm and slow dried be-fore (solid line) and after (dashed line) thermal annealing at 150 oC for 1h.PCBM (50:50 wt.%), before and after thermal annealing are shown in figure5.2b. The absorption spectra provide evidence of extensive stacking of thepolythiophene backbone, as indicated by absorption peaks at 525, 560, and610 nm and an extension of the absorption band to 650 nm [121]. Uponthermal annealing at 150 oC for 1 h, the absorption spectra remain largelyunchanged that is due to a slow drying process used to develop the films. Thedeveloped morphology is the kinetically stable one and thermal annealingmay not necessarily contribute in enhancing the optical density [122].955.3. Results and Discussion5.3.2 Surface MorphologyThe SEM and AFM images of ZnO/ITO substrates with different thicknessesof ZnO layer are illustrated in figure 5.3. The ZnO nanoparticles with arelatively uniform diameter of 30-50 nm are formed and fused together afterthermal treatment as described above. The surface roughness of the ZnO filmappears to be a function of the thickness. It is evident that the agglomerationof ZnO nanoparticles is far more extensive in figure 5.3b for a thicker filmmade by spin coating of higher concentration of zinc acetate solution incomparison to that in figure 5.3a. This is also supported by the AFM imagespresented in figure 5.3c-f, depicting the increased porosity and roughness ofthe ZnO layer with increasing thickness. A remarkable difference for thethicker (80-nm) ZnO film is the formation of large ZnO crystals rather thanfused nanoparticles.The photoactive BHJ blend of P3HT and PCBM (50:50 wt.%) was spin-coated on the ZnO/ITO interface using 1,2-DCB, a solvent with a high boilingpoint. As a result, the slow drying process enables the polymer to slowlycrystalize and form an optimized morphology resulting in improved opticaland electronic properties [123] and [122]. The morphology of P3HT:PCBMin the solid state, examined by TEM, showed evidence of a binary networkof P3HT and fullerene. Bright-field TEM images of the P3HT:PCBM films,before and after thermal annealing at 150 oC, are shown in figure 5.4. Inthe TEM image of the as-cast samples, the morphology is well developedand the donor and acceptor domains show a typical feature size of 10-20 nm.965.3. Results and DiscussionFigure 5.3: SEM and AFM images of ZnO film surfaces on ITO used ashole-blocking layers in inverted OPV devices. ZnO film was prepared byspin-coating zinc acetate solution in methanol and rapid heating to 350 oCfor 5 min. For SEM images, thicknesses are (a) 25 and (b) 50 nm, and forAFM images (c) 25, (d) 50 and (e) 80 nm for a 1 1 µm pixel. A 10 10 µmpixel of (e) is presented in (f). The inset in (b) shows a wider view (scale: 1µm).After thermal annealing at 150 oC for 1 h the domain sizes are practically thesame. The TEM images reveals that thermal annealing improves the purityof the donor and acceptor network, i.e., the phase contrast becomes clearerand easily observable with a slightly larger domain size as compared to theas-cast films.5.3.3 Deposition of a Thin Film of GoldAn important technical issue in inverted OPV devices is that aqueous PE-DOT:PSS does not wet the P3HT:PCBM surface effectively, even after di-975.3. Results and DiscussionFigure 5.4: TEM images of P3HT:PCBM film used as photoactive layersin inverted OPV devices (film thickness ∼ 150 nm) on ZnO/PEDOT/ITObefore (a) and after thermal annealing at 150 oC for 1 h (b), and with 2-nmgold coating (c). Scale bar is 100 nm.lution with 2-propanol (50:50 vol.%). In order to fabricate an inverted de-vice free of pin holes and short contacts, we paid particular attention tothe deposition of PEDOT:PSS layer. To improve the wettability of thephotoactive layer, a thin film of gold (2-5 nm) was thermally depositedon P3HT:PCBM/ZnO/ITO interface. The TEM of the gold thin film onP3HT:PCBM is included in figure 5.4. As observed, the gold film is composedof nano-dots with a size ranging from 5 to 15 nm that are not connected.The height of the gold film was varied between 2 and 5 nm by controlling thedeposition time at a rate of 0.1 A˚/s. The advantage of the thin gold film isto improve the wettability of the photoactive layer for aqueous PEDOT:PSSthat present large surface tensions [124]. Figure 5.5 displays a picture of thedevices after PEDOT:PSS is spin-coated on P3HT:PCBM without (devicesa and b) and with 5-nm gold coating (devices c and d). Successful depositionof PEDOT:PSS layer is visually confirmed on gold coated area. While thegold thickness has a minor impact on PV property (see below), it turned out985.3. Results and Discussionthat devices with thicker gold coating (5 nm) maintain their performanceeven after a long period (7 weeks in this study).Figure 5.5: A picture of the devices after PEDOT:PSS is spin-coated directlyon P3HT:PCBM (a and b) and on 5-nm gold coated ones (c and d). Noticethat in the gold coated devices (c and d), the deposition of PEDOT:PSS layeris visually confirmed on gold coated area (marked by arrows and dotted lines).5.3.4 Photovoltaic PerformanceThe PV properties of inverted devices, exposed to ambient atmosphere,were characterized by current-voltage and external quantum efficiency (EQE)measurements. For comparison, a series of normal devices with a struc-ture of glass/ITO/PEODT:PSS/P3HT:PCBM/Ca/Al was also fabricatedand tested as described previously [125]. Table 1 shows the highest andaverage PV performances for normal and inverted OPV devices fabricatedin this work. The maximum PCE of 3.8% and 3.6% were achieved for in-verted devices with a gold buffer layer of 2 and 5 nm, respectively, deposited995.3. Results and Discussionin between the photoactive layer and PEDOT:PSS. Without the gold bufferlayer PCEs are in the range of 1.9-3.0%. Fabricating the same active layerin the normal (non-inverted) device architecture using Ca (20 nm) and Al(80 nm) as the top electrode yields a PCE of 2.5%. For the invertedTable 5.1: Extracted parameters, short-current density (JSC), open-circuitvoltage (VOC), fill factor (FF) and power conversion efficiency (PCE) forP3HT:PCBM (50:50 wt.%) devices. Parentheses denote average performanceobtained from 24 devices tested. The calculated JSC is based on the overlapintegral between EQE and AM1.5G spectrum in the range of 330-800 nm.DeviceTypeZnO JSCCalculated(mA/cm2)JSCObserved(mA/cm2)VOC FF PCE(%)Inverted 2-nm Au25-nm50-nm80-nm13.2 11.5(11.1)13.1(13.0)13.1(10.3)0.62(0.60)0.55(0.55)0.45(0.43)0.31(0.30)0.54(0.45)0.39(0.37)2.2(2.0)3.8(3.2)2.1(1.6)Inverted 5-nm Au25-nm50-nm80-nm13.3 11.2(10.5)12.5(12.0)12.3(10.9)0.59(0.58)0.55(0.52)0.50(0.51)0.41(0.33)0.54(0.50)0.50(0.48)3.2(2.1)3.6(3.3)3.1(2.9)InvertedwithoutAu25-nm50-nm80-nm12.3 11.1(10.9)12.5(2.1)10.9(10.8)0.52(0.52)0.50(0.47)0.44(0.43)0.47(0.42)0.52(0.45)0.40(0.39)2.7(2.4)3.0(2.5)1.9(1.8)Normal N/A 9.8 9.8(9.6) 0.61(0.60) 0.45(0.40) 2.5(2.3)devices, the fill factor (FF) varies between 0.30 and 0.54 with the highervalues matching the ones reported for structurally similar inverted devices[115, 29, 116, 117, 118]. Increasing the thickness of the ZnO (from 25 nm)results in a typically enhanced PCE (50 nm) followed by a reduction in PCE(80 nm): for 2-nm gold coated devices, PCE of 2.2 increases to 3.8 and dropsto 2.1% and for 5-nm gold coated ones, PCE of 3.2 increases to 3.6 and thendrops to 3.1%. Similar trends are also observed for the short-circuit current1005.3. Results and Discussiondensity (JSC) and FF (Table 5.1) that point to a trade-off between completecoverage of ITO (i.e., a low density of pin-holes) and an increase in seriesresistance. One of the few disadvantages of ZnO is the restriction to a ratherthin film due to its low conductivity [126]. Thin layer of ZnO provides amuch better electron-selective interface and ohmic cathode contact [29, 127].Based on the result of this study, an optimized performance is achieved fora ZnO film thickness of 50±10 nm.VOC = (1/e)(|EP3HT (HOMO)| − |EPCBM(LUMO)|)− 0.3V (5.1)Here, e is the elementary charge and EP3HT (∼ 5.2 eV) and EPCBM (∼ 4.3eV) are the HOMO and LUMO levels of each component, respectively. Thevalue of 0.3 V, an empirical factor discussed in details by Scharber et al.[128], is the difference between the built-in potential and the VOC that isoriginated by the dark I-V curve of the photodiode. Here, we postulate alikely scenario for a thicker ZnO film, where an additional PN heterojunctionis present for charge separation and photocurrent generation. Namely, theP3HT:ZnO heterojunction rather than P3HT:PCBM is gradually becomingthe dominant one in defining the VOC when the porosity of the ZnO layeris increased (figure 5.3). In a similar study, low VOC of hybrid P3HT/ZnOnanofiber PV devices was attributed to the presence of midgap states on theZnO surface resulting in surface pinning of the Fermi level [129]. Based onthe approximate energy diagram (figure 5.1), this assumption yields a VOC1015.3. Results and Discussionof 0.5 V for the inverted devices (free of pin-holes and defects) using:VOC = (1/e)(|EP3HT (HOMO)| − |EZnO(CB)|)− 0.3V (5.2)where EZnO(CB) is the conduction band energy of ZnO (∼ 4.4 eV). Opticalinterference and light scattering effects, strongly depending on the ZnO layerthickness, can also contribute in reducing the intensity of light reaching thephotoactive layer [130]specifically, in the wavelength range of 300-400 nmwhere ZnO absorbs/scatters the light (figure 5.2) reducing the VOC evenfurther. figure 5.6 depicts the I-V characteristics for two devices with thehighest PCE in this work. The observed JSC and PCE for these devices areamong the highest reported values for an inverted OPV device (2-nm goldfilm: JSC = 13.1 mA/cm2, PCE = 3.8% and 5-nm; JSC = 12.5 mA/cm2, PCE= 3.6%). The increase in JSC is responsible for a remarkable improvement inPV performance of these devices compared to that of regular or the invertedones without gold buffer layer (Table 1). It must be noted that for mostof the devices the observed JSC ’s closely match the calculated values fromthe overlap integral between EQE and AM1.5G solar spectrum within 0.2mA/cm2.The EQE of the inverted and normal devices are shown in figure 5.7.For all of these devices, an efficient capturing of photo-generated excitonswithin the range of 400-600 nm is confirmed with EQE reaching a maximumvalue of 0.6 between 500 and 550 nm. EQEs are calculated based on incident1025.3. Results and DiscussionFigure 5.6: Current-voltage characterization of the inverted OPV deviceswith the highest efficiency among each series with (a) 2 and (b) 5 nm ofgold coating on P3HT:PCBM/ZnO/ITO interface. Dashed and solid linesrepresent dark and photocurrent, respectively. The PV parameters for eachdevice are: (a) VOC = 0.55 V, JSC = 13.1 mA/cm2, FF = 0.54 and PCE =3.8% and (b) VOC = 0.55 V, JSC = 12.5 mA/cm2, FF = 0.54 and PCE =3.6%.power reaching the active layer when ZnO/ITO and ITO substrates wereused for power measurements. As it is seen in figure 5.7, the apparent EQE(filled circles) which is calculated based on a monochromatic light intensityafter it passes through a ZnO/ITO substrate, and reaches the active layer, ishigher. A detailed inspection of these spectra reveals that for all the invertedOPV devices, the EQE value in the 300-400 nm range is lower than that ofa normal device that is in agreement with the transmittance spectra of ZnOfilms (figure 5.2a). Therefore, inverted OPV devices with a thicker ZnOdo suffer from reduced incident power. However, the reduction in photo-generated current is estimated at ∼ 4% based on the overlap integral withinthe range of 300-400 nm. On the other hand, the EQE values of inverted1035.3. Results and Discussiondevices with 2 and 5 nm of gold layer are larger than that of a normal deviceat wavelengths longer than 400 nm (figure 5.7d) resulting in an increase of ∼8% in photocurrent in the 400-800 nm range. The improved performance ofinverted OPV devices (in terms of increased short circuit current and PCE) istherefore attributed to an efficient charge collection at both top and bottomcontact layers. The optimization of processing parameters and formation ofa transparent ZnO thin film is the key issue in developing high efficiencyinverted OPV devices.5.3.5 Device StabilityThe devices were subjected to a two-stage long-term stability test. In thefirst stage, devices were exposed to air for 7 weeks, during which their per-formance was tested periodically. After 7 weeks, in the second stage, theywere illuminated with uninterrupted white light at 100 mW/cm2 of incidentpower in air using a tungsten lamp filtered through AM1.5G filter. The tem-perature of the device holder was also monitored during light exposure (∼38 ± 2 oC). Figure 5.8 illustrates the PCE of the inverted OPV devices with2- and 5-nm gold layer during the long-term test, displaying a significantlybetter stability for the latter in the first stage of experiment when they areexposed to ambient conditions. Namely, the PCE of the device with 5-nmgold under the PEDOT:PSS remains unchanged (at ∼ 3.6%) after 50 days.The stability of PCEs monitored under white light exposure, however, areless promising. After 50 and 70 h of light exposure, the PCE of 2- and 5-1045.3. Results and DiscussionFigure 5.7: External quantum efficiency (EQE) of the inverted OPV deviceswith a 50-nm ZnO layer for (a) 2-nm, (b) 5-nm and (c) without gold bufferlayer, and (d) a normal P3HT:PCBM device that is superimposed to thoseof (a) and (b), showing improvement in EQE in 500-650 nm range of thespectrum consistent with the observed higher short circuit current. Filledand open circles represent the EQE value calculated based on incident powerwhen ZnO/ITO and ITO substrates, respectively, were used for power mea-surements.nm gold layer devices dropped to < 0.1 and 1.3%, respectively. This twostage experiment helps to separate the role of ambient and light exposure onstability of the device.As it has been reported in the literature, thin films of P3HT:PCBM un-1055.3. Results and DiscussionFigure 5.8: Upper panel: power conversion efficiency (PCE) of non-encapsulated inverted OPV devices (a) stored in air under ambient conditionsand (b) exposed to 100 mW/cm2 of incident white light. Lines and soft-edgesare drawn as a guide for trends in PCE and its error margin, respectively.Lower panel: microscopic images of a P3HT:PCBM film on 50 nm ZnO/ITOsubstrates for (c) as-cast, (d) exposed to air (50 days) and white light (100mW/cm2, 70 h), and (c) a thermally annealed film at 150 oC for 1 h. Darkarea in (d) is the device active area with gold coating.dergo gross phase separation at elevated temperatures as illustrated by thegrowth of large PCBM crystals-a process that reduces the interface betweenthe donor polymer and acceptor fullerene [125, 131, 132, 133, 134]. Figure 5.8also shows microscopic images of a P3HT:PCBM film on 50 nm ZnO/ITOsubstrates before and after it is irradiated with white light as well as a ther-mally annealed film (at 150 oC for 1 h). Under white light illumination,the thermal energy transferred to the blend accelerates the phase separationprocess causing macro-phase segregation of the blend constituents, i.e., for-1065.4. Conclusionmation of PCBM crystals (figure 5.8d). This process is very similar but notas catastrophic as in samples annealed thermally at 150 oC for 1 h (figure5.8e). The outcome is a slow degradation of the performance that needs tobe addressed in order to improve the long-term stability of these invertedOPV devices.5.4 ConclusionElectronic contact layers are critically important in determining how an op-toelectronic device operates (in terms of electrical and optical performance),and as such, play a critical role in the photovoltaic performance of OPVdevices. This paper presents inverted OPV devices with significantly im-proved performance in terms of efficiency (3.8%) and stability, by systematicdevelopment of contact layers. ZnO nano-porous films were investigated aselectron-selective contact layer on the ITO transparent electrode. While someabsorption loss in high energy photons are observed, the short-circuit currentdensity is significantly improved (∼ 13 mA/cm2) for a ZnO contact layer witha thickness of 50 nm. Thicker ZnO films suppress the performance due to thelow conductivity of this contact layer, increased loss of high energy photons,and reduced open circuit voltage possibly indebted to the parasitic role ofZnO:P3HT heterojunctions. A thin gold film deposited under PEDOT:PSShas shown to be highly effective in improving the wettability of BHJ layerfor PEDOT:PSS deposition, resulting in improved efficiency and stability of1075.5. Effect of Hydrothermally Synthesized Nanorods on Photovoltaic Performancethe device. These results highlight the significant role of contact layers anddevice engineering in increasing the efficiency and stability of OPV devices5.5 Effect of Hydrothermally SynthesizedNanorods on Photovoltaic PerformanceIn addition to the work published in Organic Electronics [135], more effortswere put for exploring effect of NRs instead of ZnO NPs as electron selectinglayer. However, the PV performance improvement was not as good as theones reported earlier in this chapter. For this purpose, first a thin film ofNP layer (25 nm) were deposited using zinc acetate, then followed by ahydrothermal growth of ZnO NRs. Nanorods were grown using a solutionof 25 mM zinc nitrate hexahydrate, Zn(NO3)2.6H2O (ZnNit), and 25 nMhexamethylenetetramine (HMTA) in DI water at 90 oC in a thermal bath for5 mins. The other fabrication steps were similar to the inverted devices insection 5.2.1 made with and without NRs as shown in figure 5.9. Then deviceswere similarly characterized and the the average parameters are reported inthe table 5.2.Table 5.2: Average photovoltaic parameters, short current density (JSC),open circuit voltage (VOC), fill factor (FF) and power conversion efficiency(PCE) for inverted devices.Type of Device JSC (mA/cm2) VOC FF PCE (%)Inverted 25 nm ZnO NP 10.0 0.52 0.42 2.2Inverted 25 nm ZnO NP + ZnO NR 10.7 0.52 0.44 2.51085.5. Effect of Hydrothermally Synthesized Nanorods on Photovoltaic Performance(a) (b)Figure 5.9: (a) Displays a NP layer of 25 nm thick and (b) NRs grown for5 min on top of the NP layer. In the inset 45o tilted SEM micrograph isshown. Scale bar is 1 µm (c) also shows the transmittance with (solid line)and without (dash line) NRs1095.5. Effect of Hydrothermally Synthesized Nanorods on Photovoltaic PerformanceAs expected, devices with NR display higher short current density as wellas power conversion efficiency which can be attributed to a more efficientcharge collection due to higher surface area of the NRs. Moreover, the NRsprovide shorter path for electrons to reach the electrodes that also improvesthe quantum efficiency. However, the results of using nanorods are not asgood as the ones with NPs reported earlier in this chapter and still requiremore characterization followed by trouble shooting to pave the way for moreefficient inverted devices.110Chapter 6Conclusion and Future WorkThis work focuses on a variety of sensing and photovoltaic applications ofZnO nanostructures. In this thesis, synthesis methods for various types ofnanostructure including CVD grown nanowires, hydrothermally synthesizednanorods, as well as solution processed nanoparticles, and incorporating suchstructures in electronic and opto-electronic device geometries, have been de-scribed. Various sensing properties of ZnO nanostructures such as oxygen,carbon monoxide, relative humidity, and ultra-violet light were investigated.In the last part of the thesis, the application of ZnO nanostructures in or-ganic photovoltaic devices is studied as an efficient electron selecting layerthat preserves the organic layer and results in air stability.6.1 Contributions• In comparison to conventional e-beam lithography fabrication of NWdevices, dielectrophoresis assembled devices provide a lower cost andless labour intensive route for device fabrication. In chapter 3, I con-tributed in demonstrating the highest sensitivity (5 orders of magnitude1116.1. Contributionsimprovement in the conductivity of the device) for relative humiditysensing using ZnO NW devices. The highest value for ZnO NW hu-midity sensors prior to our work is reported by Fang et al. [73] withless than 4 orders of magnitude (3000) improve in the conductivity forthe similar change in RH (60%), by using arc-discharge NWs humiditysensors. The reason for observing such a high sensitivity for our sensorsis the small size with low carrier density of the NWs that result in aDebye screening length larger than the diameter of the NWs. This typeof sensitivity has only been shown for silicon NW biosensors by Gao etal. [20]. This small footprint of the sensor, having one single NW asthe sensitive block, compared to previous NW-based RH sensors witha forrest of NWs bridging into each other (hundreds of µm2 in Changet al. [57] work, a few mm2 sized devices in Zhang et al. [69], Qi et al.[72] and Fang et al. [73] works) is important for embedding the sen-sors into fuel cells and complex machineries. This work is accompaniedby oxygen, carbon monoxide and ultra violet sensing properties of theDEP-assembled ZnO NW devices. In addition, the effect of tempera-ture and UV on sensing behaviour of ZnO was investigated in chapter3.• In chapter 4, as another contribution of this thesis, we have demon-strated a fully sketched ZnO based carbon monoxide and ultra violetsensor on a piece of paper. Prior to this work, there were three reportson using ZnO on paper substrate for sensing applications. Gimenez1126.1. Contributionset al. [88] studied the effect of UV by dropping ZnO solution on apaper substrate, Gullapalli et al. [136] investigated strain sensing onZnO that has been deposited on paper by dip coating, and the lastone, by VithalGhule et al. [87], is about the antibacterial activity ofZnO coated on paper by using a sonicator. In comparison with theseworks, our method has a simpler approach and uses low-cost nanos-tructured electrodes as well as the sensitive layer, ZnO (< 10 nm), thatboost the sensitivity to environmental elements such as carbon monox-ide. Due to the ease of processing and the low cost of fabrication, thisapproach can be done on variety of substrates by sketching, painting,and stamping the electrodes and the sensitive material. One example issketching a CO sensor that everyone can put on top of their fireplaceswith minimum cost.• In chapter 5, what we have shown is an inverted organic photovoltaicdevice that makes use of nanostructured ZnO and Au NPs that showexceptional short current density (∼ 13 mA/cm2 compared to 11.2mA/cm2 reported by White et al. [109] and 11.5 mA/cm2 by Hau et al.[29]), signifying efficient charge collection at e- and h- selecting layers.Moreover, compared to previous air-stable devices with less than 20%drop in power conversion efficiency in 40 days, our devices show lessthan 5% decrease in efficiency after 7 weeks of air exposure, attributingto a more stable PEDOT:PSS layer. Deposition of the thin layer of AuNPs has shown to be highly effective in improving the wettability of1136.2. Future Workthe BHJ layer for PEDOT:PSS deposition and results in a more stablelayer.6.2 Future WorkAlthough the use of ZnO in many electronic applications has been hindereddue to the unknown origins of its doping, this has never caused an obstaclefor sensing applications. However, gaining more knowledge and insight onthe surface science of ZnO as well as the chemistry behind chemisorptionand physisorption of different molecules on the surface, can open the doorsfor tailoring the sensitivity and controlling the selectivity of the metal oxideto various stimuli. Despite the ultra-sensitive behaviour of ZnO to environ-mental factors, the main issue that remains unsolved is the selectivity. Thiscan be addressed not only by increasing our knowledge about the surfaceof the metal oxide, as explained earlier, but also by exploring techniques topronounce the sensitivity to one stimulus and diminishing the response tothe others present in the environment. As shown in this work, UV light isone way to improve the sensitivity to oxygen as opposed to relative humid-ity. Finding more alternatives for improving the selectivity can significantlyincrease the future applications of these devices as fast, accurate, and smallenvironmental sensors. One of the blocking obstacles for commercializationof the bottom-up approach in nano-devices is still the high cost and com-plexity of the whole device fabrication process. In addition to small and1146.2. Future Workfast devices benefiting from single NW structure, for many applications, itis not necessary to have that much accuracy and complications for devicefabrications. It is critical to develop techniques for making devices that ex-ploit nanostructures with minimum cost and hassle. This involves cheaperapproaches for synthesis of nanostructures such as low temperature and so-lution processable fabrication process. Furthermore, the simplicity of devicefabrication could be achieved with low-cost and ubiquitous substrates suchas paper, plastic, and textile.One way to improve organic photovoltaic devices’ stability, as shown inthis thesis, is using an inverted structure that takes advantage of high workfunction metals as the back electrode. ZnO is shown to be a suitable candi-date for the electron selecting layer due to its energy levels, similar to TiO2,and its transparency to visible light. The remaining source of chemical insta-bilities in the inverted structure is PEDOT:PSS. Since 2006, when Shrotriyaet al. [137] reported the use of transition metal oxides such as MoO3 andV2O5 as the hole selecting layer, there has been a significant amount of in-terest in replacing PEDOT:PSS with stable metal oxides. Moreover, with re-cent advances in solution processing of the transition metal oxides [138, 139],which is an advantage for low-cost and large-scale roll-to-roll fabrication ofsolar cells, it is more plausible and cost-effective to utilize metal oxides forboth e- and h- selecting layers.115Bibliography[1] A. Janotti and C. G. Van de Walle, “Fundamentals of zinc oxide asa semiconductor,” Reports on Progress in Physics, vol. 72, no. 12,p. 126501, 2009.[2] R. P. Feynman, “There’s plenty of room at the bottom,” Engineeringand Science, vol. 23, no. 5, pp. 22–36, 1960.[3] S. H. Ko, I. Park, H. Pan, N. Misra, M. S. Rogers, C. P. Grigoropou-los, and A. P. Pisano, “ZnO nanowire network transistor fabricationon a polymer substrate by low-temperature, all-inorganic nanoparticlesolution process,” Applied Physics Letters, vol. 92, no. 15, pp. 154102–154102, 2008.[4] Y.-Y. Noh, X. Cheng, H. Sirringhaus, J. I. Sohn, M. E. Welland, andD. J. Kang, “Ink-jet printed ZnO nanowire field effect transistors,”Applied Physics Letters, vol. 91, no. 4, pp. 043109–043109, 2007.[5] X. Wang, X. Wang, C. J. Summers, and Z. L. Wang, “Large-scale hexagonal-patterned growth of aligned ZnO nanorods for nano-optoelectronics and nanosensor arrays,” Nano Letters, vol. 4, no. 3,pp. 423–426, 2004.[6] X. Wang, X. Wang, J. Zhou, J. Song, J. Liu, N. Xu, and Z. L. Wang,“Piezoelectric field effect transistor and nanoforce sensor based on asingle ZnO nanowire,” Nano Letters, vol. 6, no. 12, pp. 2768–2772,2006.[7] Q. Wan, Q. Li, Y. Chen, T. Wang, X. He, J. Li, and C. Lin, “Fabrica-tion and ethanol sensing characteristics of ZnO nanowire gas sensors,”Applied Physics Letters, vol. 84, no. 18, pp. 3654–3656, 2004.[8] Z. Fan and J. G. Lu, “Gate-refreshable nanowire chemical sensors,”Applied Physics Letters, vol. 86, no. 12, pp. 123510–123510, 2005.[9] D. Thomas, “The exciton spectrum of zinc oxide,” Journal of Physicsand Chemistry of Solids, vol. 15, no. 1, pp. 86–96, 1960.[10] D. C. Look, “Recent advances in ZnO materials and devices,” MaterialsScience and Engineering: B, vol. 80, no. 1-3, pp. 383–387, 2001.116Bibliography[11] D. C. Reynolds, D. C. Look, and B. Jogai, “Optically pumped ultra-violet lasing from ZnO,” Solid State Communications, vol. 99, no. 12,pp. 873–875, 1996.[12] E. Comini, G. Faglia, G. Sberveglieri, Z. Pan, and Z. L. Wang, “Sta-ble and highly sensitive gas sensors based on semiconducting oxidenanobelts,” Applied Physics Letters, vol. 81, no. 10, pp. 1869–1871,2002.[13] E. Comini, “Metal oxide nano-crystals for gas sensing,” AnalyticaChimica Acta, vol. 568, no. 1, pp. 28–40, 2006.[14] M. H. Huang, Y. Wu, H. Feick, N. Tran, E. Weber, and P. Yang, “Cat-alytic growth of zinc oxide nanowires by vapor transport,” AdvancedMaterials, vol. 13, no. 2, pp. 113–116, 2001.[15] X. Y. Kong and Z. L. Wang, “Polar-surface dominated ZnO nanobeltsand the electrostatic energy induced nanohelixes, nanosprings, andnanospirals,” Applied Surface Science, vol. 84, no. 6, pp. 975–977, 2004.[16] Y. Leung, A. Djuriˇsic´, J. Gao, M. Xie, Z. Wei, S. Xu, and W. Chan,“Zinc oxide ribbon and comb structures: synthesis and optical proper-ties,” Chemical Physics Letters, vol. 394, no. 4, pp. 452–457, 2004.[17] P. Cox, The surface science of metal oxides. Cambridge universitypress, 1994.[18] C. Nu¨tzenadel, A. Zu¨ttel, D. Chartouni, G. Schmid, and L. Schlapbach,“Critical size and surface effect of the hydrogen interaction of palladiumclusters,” The European Physical Journal D, vol. 8, no. 2, pp. 245–250,2000.[19] B. G. Streetman and S. Banerjee, Solid state electronic devices, vol. 4.Prentice Hall New Jersey, 2000.[20] X. P. Gao, G. Zheng, and C. M. Lieber, “Subthreshold regime has theoptimal sensitivity for nanowire fet biosensors,” Nano Letters, vol. 10,no. 2, pp. 547–552, 2009.[21] H. A. Pohl, “The motion and precipitation of suspensoids in divergentelectric fields,” Journal of Applied Physics, vol. 22, no. 7, pp. 869–871,1951.[22] M. C. Wang and B. D. Gates, “Directed assembly of nanowires,” Ma-terials Today, vol. 12, no. 5, pp. 34–43, 2009.[23] H. Morgan and N. G. Green, AC electrokinetics: colloids and nanopar-ticles. No. 2, Research Studies Press, 2003.117Bibliography[24] P. Hesketh, M. Gallivan, S. Kumar, C. Erdy, and Z. Wang, “Theapplication of dielectrophoresis to nanowire sorting and assembly forsensors,” in Intelligent Control, 2005. Proceedings of the 2005 IEEEInternational Symposium on, Mediterrean Conference on Control andAutomation, pp. 153–158, IEEE, 2005.[25] S. Raychaudhuri, S. A. Dayeh, D. Wang, and E. T. Yu, “Precise semi-conductor nanowire placement through dielectrophoresis,” Nano Let-ters, vol. 9, no. 6, pp. 2260–2266, 2009.[26] F. C. Krebs, Polymeric solar cells: materials, design, manufacture.DEStech Publications, Inc, 2010.[27] K. Lee, J. Y. Kim, S. H. Park, S. H. Kim, S. Cho, and A. J. Heeger,“Air-stable polymer electronic devices,” Advanced Materials, vol. 19,no. 18, pp. 2445–2449, 2007.[28] A. Hayakawa, O. Yoshikawa, T. Fujieda, K. Uehara, and S. Yoshikawa,“High performance polythiophene/fullerene bulk-heterojunction solarcell with a TiOx hole blocking layer,” Applied Physics Letters, vol. 90,no. 16, pp. 163517–163517, 2007.[29] S. K. Hau, H.-L. Yip, N. S. Baek, J. Zou, K. OMalley, and A. K.-Y.Jen, “Air-stable inverted flexible polymer solar cells using zinc oxidenanoparticles as an electron selective layer,” Applied Physics Letters,vol. 92, p. 253301, 2008.[30] J. H. Woodruff, J. B. Ratchford, I. A. Goldthorpe, P. C. McIntyre,and C. E. Chidsey, “Vertically oriented germanium nanowires grownfrom gold colloids on silicon substrates and subsequent gold removal,”Materials Today, vol. 7, no. 6, pp. 1637–1642, 2007.[31] C. Pacholski, A. Kornowski, and H. Weller, “Self-assembly of ZnO:From nanodots to nanorods,” Angewandte Chemie International Edi-tion, vol. 41, no. 7, pp. 1188–1191, 2002.[32] L. Vayssieres, K. Keis, A. Hagfeldt, and S.-E. Lindquist, “Three-dimensional array of highly oriented crystalline ZnO microtubes,”Chemistry of Materials, vol. 13, no. 12, pp. 4395–4398, 2001.[33] L. Vayssieres, “Growth of arrayed nanorods and nanowires of ZnO fromaqueous solutions,” Advanced Materials, vol. 15, no. 5, pp. 464–466,2003.[34] L. Schmidt-Mende and J. L. MacManus-Driscoll, “ZnO–nanostructures, defects, and devices,” Applied Physics Letters,vol. 10, no. 5, pp. 40–48, 2007.118Bibliography[35] S. Baruah and J. Dutta, “Hydrothermal growth of ZnO nanostruc-tures,” Science and Technology of Advanced Materials, vol. 10, no. 1,p. 013001, 2009.[36] E. Freer, O. Grachev, X. Duan, S. Martin, and D. Stumbo, “High-yieldself-limiting single-nanowire assembly with dielectrophoresis,” NatureNanotechnology, vol. 5, no. 7, pp. 525–530, 2010.[37] Z. L. Wang and J. Song, “Piezoelectric nanogenerators based on zincoxide nanowire arrays,” Science, vol. 312, no. 5771, pp. 242–246, 2006.[38] J. Grabowska, K. Nanda, E. McGlynn, J.-P. Mosnier, and M. Henry,“Studying the growth conditions, the alignment and structure of ZnOnanorods,” Surface and Coatings Technology, vol. 200, no. 1, pp. 1093–1096, 2005.[39] C. S. Lao, J. Liu, P. Gao, L. Zhang, D. Davidovic, R. Tummala, andZ. L. Wang, “ZnO nanobelt/nanowire Schottky diodes formed by di-electrophoresis alignment across Au electrodes,” Nano Letters, vol. 6,no. 2, pp. 263–266, 2006.[40] M. TAKATA, D. TSUBONE, and H. YANAGIDA, “Dependence ofelectrical conductivity of ZnO on degree of sintering,” Journal of theAmerican Ceramic Society, vol. 59, no. 1-2, pp. 4–8, 1976.[41] S. Chang, T. Hsueh, C. Hsu, Y. Lin, I. Chen, and B. Huang, “AZnO nanowire vacuum pressure sensor,” Nanotechnology, vol. 19, no. 9,p. 095505, 2008.[42] T.-Y. Wei, P.-H. Yeh, S.-Y. Lu, and Z. L. Wang, “Gigantic enhance-ment in sensitivity using Schottky contacted nanowire nanosensor,”Journal of the American Chemical Society, vol. 131, no. 48, pp. 17690–17695, 2009.[43] N. Mohseni Kiasari, S. Soltanian, B. Gholamkhass, and P. Servati,“Room temperature ultra-sensitive resistive humidity sensor basedon single zinc oxide nanowire,” Sensors and Actuators A: Physical,vol. 182, no. 0, pp. 101–105, 2012.[44] F. Shan and Y. Yu, “Band gap energy of pure and Al-doped ZnOthin films,” Journal of the European Ceramic Society, vol. 24, no. 6,pp. 1869–1872, 2004.[45] C. Soci, A. Zhang, B. Xiang, S. Dayeh, D. Aplin, J. Park, X. Bao,Y. Lo, and D. Wang, “ZnO nanowire UV photodetectors with highinternal gain,” Nano Letters, vol. 7, no. 4, pp. 1003–1009, 2007.119Bibliography[46] T. Moon, M. Jeong, W. Lee, and J. Myoung, “The fabrication and char-acterization of ZnO UV detector,” Applied Surface Science, vol. 240,no. 1, pp. 280–285, 2005.[47] J. Zhou, Y. Gu, Y. Hu, W. Mai, P. Yeh, G. Bao, A. Sood, D. Polla, andZ. Wang, “Gigantic enhancement in response and reset time of ZnOUV nanosensor by utilizing Schottky contact and surface functionaliza-tion,” Applied Physics Letters, vol. 94, no. 19, pp. 191103–191103–3,2009.[48] N. M. Kiasari and P. Servati, “Dielectrophoresis-assembled ZnOnanowire oxygen sensors,” Electron Device Letters, IEEE, vol. 32, no. 7,pp. 982–984, 2011.[49] D. Melnick, “Zinc oxide photoconduction, an oxygen adsorption pro-cess,” The Journal of Chemical Physics, vol. 26, p. 1136, 1957.[50] Q. Li, T. Gao, Y. Wang, and T. Wang, “Adsorption and desorptionof oxygen probed from ZnO nanowire films by photocurrent measure-ments,” Applied Physics Letters, vol. 86, no. 12, pp. 123117–123117–3,2005.[51] Q. Li, Q. Wan, Y. Liang, and T. Wang, “Electronic transport throughindividual ZnO nanowires,” Applied Physics Letters, vol. 84, no. 22,pp. 4556–4558, 2004.[52] B. Kulwicki, “Humidity sensors,” Journal of the American CeramicSociety, vol. 74, no. 4, pp. 697–708, 1991.[53] E. Traversa, “Ceramic sensors for humidity detection: the state-of-the-art and future developments,” Sensors and Actuators B: Chemical,vol. 23, no. 2-3, pp. 135–156, 1995.[54] Z. Rittersma, “Recent achievements in miniaturised humidity sensors–areview of transduction techniques,” Sensors and Actuators A: Physical,vol. 96, no. 2-3, pp. 196–210, 2002.[55] J. Fraden, Handbook of modern sensors: physics, designs, and applica-tions. Springer Verlag, 2004.[56] T. Nitta and T. Seiyama, “Development and application of ceramichumidity sensors,” Chemical Sensor Technology, vol. 1, pp. 57–78, 1988.[57] S. Chang, S. Chang, C. Lu, M. Li, C. Hsu, Y. Chiou, T. Hsueh, andI. Chen, “A ZnO nanowire-based humidity sensor,” Superlattices andMicrostructures, vol. 47, no. 6, pp. 772–778, 2010.120Bibliography[58] A. Kolmakov, D. Klenov, Y. Lilach, S. Stemmer, and M. Moskovits,“Enhanced gas sensing by individual SnO2 nanowires and nanobeltsfunctionalized with Pd catalyst particles,” Nano Letters, vol. 5, no. 4,pp. 667–673, 2005.[59] Q. Kuang, C. Lao, Z. Wang, Z. Xie, and L. Zheng, “High-sensitivity hu-midity sensor based on a single SnO2 nanowire,” Journal of the Amer-ican Chemical Society, vol. 129, no. 19, pp. 6070–6071, 2007.[60] R. Wu, Y. Sun, C. Lin, H. Chen, and M. Chavali, “Composite of TiO2nanowires and nafion as humidity sensor material,” Sensors and Actu-ators B: Chemical, vol. 115, no. 1, pp. 198–204, 2006.[61] A. Vomiero, S. Bianchi, E. Comini, G. Faglia, M. Ferroni, N. Poli, andG. Sberveglieri, “In2O3 nanowires for gas sensors: morphology andsensing characterisation,” Thin Solid Films, vol. 515, no. 23, pp. 8356–8359, 2007.[62] C. Rout, K. Ganesh, A. Govindaraj, and C. Rao, “Sensors for the nitro-gen oxides, NO2, NO and N2O, based on In2O3 and WO3 nanowires,”Applied Physics A: Materials Science & Processing, vol. 85, no. 3,pp. 241–246, 2006.[63] Z. Wang, “Zinc oxide nanostructures: growth, properties and applica-tions,” Journal of Physics: Condensed Matter, vol. 16, p. R829, 2004.[64] Y. Kong, D. Yu, B. Zhang, W. Fang, and S. Feng, “Ultraviolet-emittingZnO nanowires synthesized by a physical vapor deposition approach,”Applied Physics Letters, vol. 78, p. 407, 2001.[65] N. M. Kiasari, J. Shen, B. Gholamkhass, S. Soltanian, and P. Servati,“Well-aligned zinc oxide nanowire arrays for transparent electrode ap-plications,” in Photonics Conference (PHO), 2011 IEEE, pp. 561–562,2011.[66] A. Hartanto, X. Ning, Y. Nakata, and T. Okada, “Growth mecha-nism of ZnO nanorods from nanoparticles formed in a laser ablationplume,” Applied Physics A: Materials Science & Processing, vol. 78,no. 3, pp. 299–301, 2004.[67] B. Liu and H. Zeng, “Hydrothermal synthesis of ZnO nanorods in thediameter regime of 50 nm,” Journal of the American Chemical Society,vol. 125, no. 15, pp. 4430–4431, 2003.[68] Q. Wan, Q. Li, Y. Chen, T. Wang, X. He, X. Gao, and J. Li, “Positivetemperature coefficient resistance and humidity sensing properties ofCd-doped ZnO nanowires,” Applied Physics Letters, vol. 84, p. 3085,2004.121Bibliography[69] Y. Zhang, K. Yu, D. Jiang, Z. Zhu, H. Geng, and L. Luo, “Zinc ox-ide nanorod and nanowire for humidity sensor,” Nanomaterials andInterfaces, vol. 242, no. 1-2, pp. 212–217, 2005.[70] Z. Chen and C. Lu, “Humidity sensors: a review of materials andmechanisms,” Sensor Letters, vol. 3, no. 4, pp. 274–295, 2005.[71] Y. Qiu and S. Yang, “ZnO nanotetrapods: Controlled vapor-phasesynthesis and application for humidity sensing,” Advanced FunctionalMaterials, vol. 17, no. 8, pp. 1345–1352, 2007.[72] Q. Qi, T. Zhang, Q. Yu, R. Wang, Y. Zeng, L. Liu, and H. Yang,“Properties of humidity sensing ZnO nanorods-base sensor fabricatedby screen-printing,” Sensors and Actuators B: Chemical, vol. 133, no. 2,pp. 638–643, 2008.[73] F. Fang, J. Futter, A. Markwitz, and J. Kennedy, “UV and humid-ity sensing properties of ZnO nanorods prepared by the arc dischargemethod,” Nanotechnology, vol. 20, no. 24, p. 245502, 2009.[74] N. Yamazoe and Y. Shimizu, “Humidity sensors: principles and appli-cations,” Sensors and Actuators, vol. 10, no. 3-4, pp. 379–398, 1986.[75] Q. Wan, Q. Li, Y. Chen, T. Wang, X. He, J. Li, and C. Lin, “Fabrica-tion and ethanol sensing characteristics of ZnO nanowire gas sensors,”Applied Physics Letters, vol. 84, p. 3654, 2004.[76] Z. Fan, D. Wang, P. Chang, W. Tseng, and J. Lu, “ZnO nanowire field-effect transistor and oxygen sensing property,” Applied Physics Letters,vol. 85, no. 24, pp. 5923–5925, 2004.[77] P. Su and L. Huang, “Humidity sensors based on TiO2 nanopar-ticles/polypyrrole composite thin films,” Sensors and Actuators B:Chemical, vol. 123, no. 1, pp. 501–507, 2007.[78] T. Barry and F. Stone, “The reactions of oxygen at dark and irradiatedzinc oxide surface,” Proceedings of the Royal Society of London. SeriesA. Mathematical and Physical Sciences, vol. 255, no. 1280, pp. 124–144,1960.[79] J. Chang, H. Kuo, I. Leu, and M. Hon, “The effects of thickness andoperation temperature on ZnO: Al thin film CO gas sensor,” Sensorsand Actuators B: Chemical, vol. 84, no. 2, pp. 258–264, 2002.[80] P. Clifford and D. Tuma, “Characteristics of semiconductor gas sensorsII. Transient response to temperature change,” Sensors and Actuators,vol. 3, pp. 255–281, 1983.122Bibliography[81] S. Morrison, “Mechanism of semiconductor gas sensor operation,” Sen-sors and Actuators, vol. 11, no. 3, pp. 283–287, 1987.[82] J. A. Rogers, T. Someya, and Y. Huang, “Materials and mechanicsfor stretchable electronics,” Science, vol. 327, no. 5973, pp. 1603–1607,2010.[83] J. P. Rolland and D. A. Mourey, “Paper as a novel material platformfor devices,” MRS Bulletin, vol. 38, no. 04, pp. 299–305, 2013.[84] D. Tobjrk and R. sterbacka, “Paper electronics,” Advanced Materials,vol. 23, no. 17, pp. 1935–1961, 2011.[85] M. C. Barr, J. A. Rowehl, R. R. Lunt, J. Xu, A. Wang, C. M. Boyce,S. G. Im, V. Bulovi, and K. K. Gleason, “Direct monolithic integra-tion of organic photovoltaic circuits on unmodified paper,” AdvancedMaterials, vol. 23, no. 31, pp. 3500–3505, 2011.[86] A. C. Siegel, S. T. Phillips, M. D. Dickey, N. Lu, Z. Suo, and G. M.Whitesides, “Foldable printed circuit boards on paper substrates,” Ad-vanced Functional Materials, vol. 20, no. 1, pp. 28–35, 2010.[87] A. VithalGhule, “Preparation and characterization of ZnO nanoparti-cles coated paper and its antibacterial activity study,” Green Chem-istry, vol. 8, no. 12, pp. 1034–1041, 2006.[88] A. J. Gimenez, J. Yanez-Limon, and J. M. Seminario, “ZnO paperbased photoconductive UV sensor,” The Journal of Physical ChemistryC, vol. 115, no. 1, pp. 282–287, 2010.[89] K. A. Mirica, J. G. Weis, J. M. Schnorr, B. Esser, and T. M. Swager,“Mechanical drawing of gas sensors on paper,” Angewandte ChemieInternational Edition, vol. 51, no. 43, pp. 10740–10745, 2012.[90] J. R. Swider, V. A. Hackley, and J. Winter, “Characterization of chi-nese ink in size and surface,” Journal of Cultural Heritage, vol. 4, no. 3,pp. 175–186, 2003.[91] K. Chen, F. Hung, S. Chang, and S. Young, “Optoelectronic character-istics of UV photodetector based on ZnO nanowire thin films,” Journalof Alloys and Compounds, vol. 479, no. 1, pp. 674–677, 2009.[92] S. Chang, T. Hsueh, I. Chen, and B. Huang, “Highly sensitive ZnOnanowire CO sensors with the adsorption of Au nanoparticles,” Nan-otechnology, vol. 19, no. 17, p. 175502, 2008.[93] R. Joshi, Q. Hu, F. Alvi, N. Joshi, and A. Kumar, “Au decorated zincoxide nanowires for CO sensing,” The Journal of Physical Chemistry.C, Chemical Reviews, vol. 113, no. 36, p. 16199, 2009.123Bibliography[94] V. Sysoev, B. Button, K. Wepsiec, S. Dmitriev, and A. Kolmakov, “To-ward the nanoscopic ”electronic nose”: Hydrogen vs carbon monox-ide discrimination with an array of individual metal oxide nano-andmesowire sensors,” Nano Letters, vol. 6, no. 8, pp. 1584–1588, 2006.[95] N. Kiasari, S. Soltanian, B. Gholamkhass, and P. Servati, “Environ-mental gas and light sensing using zno nanowires,” Nanotechnology,IEEE Transactions on, vol. 13, pp. 368–374, March 2014.[96] J. Peet, A. J. Heeger, and G. C. Bazan, “”Plastic” solar cells: Self-assembly of bulk heterojunction nanomaterials by spontaneous phaseseparation,” Accounts of Chemical Research, vol. 42, no. 11, pp. 1700–1708, 2009.[97] S. Gu¨nes, H. Neugebauer, and N. S. Sariciftci, “Conjugated polymer-based organic solar cells,” Chemical Reviews, vol. 107, no. 4, pp. 1324–1338, 2007.[98] F. Yang and S. R. Forrest, “Photocurrent generation in nanostructuredorganic solar cells,” ACS Nano, vol. 2, no. 5, pp. 1022–1032, 2008.[99] F. C. Krebs, “All solution roll-to-roll processed polymer solar cells freefrom indium-tin-oxide and vacuum coating steps,” Organic Electronics,vol. 10, no. 5, pp. 761–768, 2009.[100] Y. Liang, Z. Xu, J. Xia, S.-T. Tsai, Y. Wu, G. Li, C. Ray, andL. Yu, “For the bright future-bulk heterojunction polymer solar cellswith power conversion efficiency of 7.4%,” Advanced Materials, vol. 22,no. 20, pp. E135–E138, 2010.[101] B. Gholamkhass and S. Holdcroft, “Enhancing the durability of poly-mer solar cells using gold nano-dots,” Solar Energy Materials and SolarCells, vol. 95, no. 11, pp. 3106–3113, 2011.[102] K. Sivula, C. K. Luscombe, B. C. Thompson, and J. M. Fre´chet, “En-hancing the thermal stability of polythiophene: fullerene solar cells bydecreasing effective polymer regioregularity,” Journal of the AmericanChemical Society, vol. 128, no. 43, pp. 13988–13989, 2006.[103] G. Li, V. Shrotriya, J. Huang, Y. Yao, T. Moriarty, K. Emery, andY. Yang, “High-efficiency solution processable polymer photovoltaiccells by self-organization of polymer blends,” Nature Materials, vol. 4,no. 11, pp. 864–868, 2005.[104] K. M. Coakley and M. D. McGehee, “Conjugated polymer photovoltaiccells,” Chemistry of Materials, vol. 16, no. 23, pp. 4533–4542, 2004.[105] H. Hoppe and N. S. Sariciftci, “Organic solar cells: An overview,”Journal of Materials Research, vol. 19, no. 07, pp. 1924–1945, 2004.124Bibliography[106] H. Hoppe, M. Niggemann, C. Winder, J. Kraut, R. Hiesgen, A. Hin-sch, D. Meissner, and N. S. Sariciftci, “Nanoscale morphology of con-jugated polymer/fullerene-based bulk-heterojunction solar cells,” Ad-vanced Functional Materials, vol. 14, no. 10, pp. 1005–1011, 2004.[107] S. Barrau, V. Andersson, F. Zhang, S. Masich, J. Bijleveld, M. R. An-dersson, and O. Inganas, “Nanomorphology of bulk heterojunction or-ganic solar cells in 2D and 3D correlated to photovoltaic performance,”Macromolecules, vol. 42, no. 13, pp. 4646–4650, 2009.[108] P. E. Shaw, A. Ruseckas, and I. D. Samuel, “Exciton diffusion mea-surements in poly (3-hexylthiophene),” Advanced Materials, vol. 20,no. 18, pp. 3516–3520, 2008.[109] M. White, D. Olson, S. Shaheen, N. Kopidakis, and D. S. Gin-ley, “Inverted bulk-heterojunction organic photovoltaic device usinga solution-derived ZnO underlayer,” Applied Physics Letters, vol. 89,no. 14, pp. 143517–143517, 2006.[110] Y. S¸ahin, S. Alem, R. de Bettignies, and J.-M. Nunzi, “Developmentof air stable polymer solar cells using an inverted gold on top anodestructure,” Thin Solid Films, vol. 476, no. 2, pp. 340–343, 2005.[111] M. Yeon Song, K.-J. Kim, and D. Y. Kim, “Enhancement of photo-voltaic characteristics using a pedot interlayer in TiO2/MEHPPV het-erojunction devices,” Solar Energy Materials and Solar Cells, vol. 85,no. 1, pp. 31–39, 2005.[112] A. Watanabe and A. Kasuya, “Effect of atmospheres on the open-circuit photovoltage of nanoporous TiO2/poly(3-hexylthiophene) het-erojunction solar cell,” Thin Solid Films, vol. 483, no. 1, pp. 358–366,2005.[113] L.-M. Chen, Z. Hong, G. Li, and Y. Yang, “Recent progress in polymersolar cells: manipulation of polymer: fullerene morphology and theformation of efficient inverted polymer solar cells,” Advanced Materials,vol. 21, no. 14-15, pp. 1434–1449, 2009.[114] G. K. Mor, K. Shankar, M. Paulose, O. K. Varghese, and C. A. Grimes,“High efficiency double heterojunction polymer photovoltaic cells usinghighly ordered TiO2 nanotube arrays,” Applied Physics Letters, vol. 91,no. 15, pp. 152111–152111, 2007.[115] C. Waldauf, M. Morana, P. Denk, P. Schilinsky, K. Coakley, S. Choulis,and C. Brabec, “Highly efficient inverted organic photovoltaics usingsolution based titanium oxide as electron selective contact,” AppliedPhysics Letters, vol. 89, no. 23, pp. 233517–233517, 2006.125Bibliography[116] D. C. Lim, W. H. Shim, K.-D. Kim, H. O. Seo, J.-H. Lim, Y. Jeong,Y. D. Kim, and K. H. Lee, “Spontaneous formation of nanoripples onthe surface of ZnO thin films as hole-blocking layer of inverted organicsolar cells,” Solar Energy Materials and Solar Cells, vol. 95, no. 11,pp. 3036–3040, 2011.[117] T. T. Larsen-Olsen, E. Bundgaard, K. O. Sylvester-Hvid, and F. C.Krebs, “A solution process for inverted tandem solar cells,” OrganicElectronics, vol. 12, no. 2, pp. 364–371, 2011.[118] P. de Bruyn, D. Moet, and P. Blom, “A facile route to inverted polymersolar cells using a precursor based zinc oxide electron transport layer,”Organic Electronics, vol. 11, no. 8, pp. 1419–1422, 2010.[119] H.-L. Yip, S. K. Hau, N. S. Baek, and A. K.-Y. Jen, “Self-assembledmonolayer modified ZnO/metal bilayer cathodes for polymer/fullerenebulk-heterojunction solar cells,” Applied Physics Letters, vol. 92,p. 193313, 2008.[120] U. Ozgur, Y. I. Alivov, C. Liu, A. Teke, M. Reshchikov, S. Dogan,V. Avrutin, S.-J. Cho, and H. Morkoc, “A comprehensive review ofZnO materials and devices,” Advanced Functional Materials, vol. 98,no. 4, pp. 041301–041301, 2005.[121] C. Winder and N. S. Sariciftci, “Low bandgap polymers for photon har-vesting in bulk heterojunction solar cells,” Journal of Materials Chem-istry, vol. 14, no. 7, pp. 1077–1086, 2004.[122] G. Li, Y. Yao, H. Yang, V. Shrotriya, G. Yang, and Y. Yang,“”solvent annealing” effect in polymer solar cells based on poly (3-hexylthiophene) and methanofullerenes,” Advanced Functional Mate-rials, vol. 17, no. 10, pp. 1636–1644, 2007.[123] W. Ma, C. Yang, X. Gong, K. Lee, and A. J. Heeger, “Thermallystable, efficient polymer solar cells with nanoscale control of the in-terpenetrating network morphology,” Advanced Functional Materials,vol. 15, no. 10, pp. 1617–1622, 2005.[124] F. C. Krebs, “Air stable polymer photovoltaics based on a process freefrom vacuum steps and fullerenes,” Journal of Applied Physics, vol. 92,no. 7, pp. 715–726, 2008.[125] S. Ebadian, B. Gholamkhass, S. Shambayati, S. Holdcroft, andP. Servati, “Effects of annealing and degradation on regioregularpolythiophene-based bulk heterojunction organic photovoltaic de-vices,” Solar Energy Materials and Solar Cells, vol. 94, no. 12,pp. 2258–2264, 2010.126Bibliography[126] T. Stubhan, H. Oh, L. Pinna, J. Krantz, I. Litzov, and C. J. Brabec,“Inverted organic solar cells using a solution processed aluminum-doped zinc oxide buffer layer,” Organic Electronics, vol. 12, no. 9,pp. 1539–1543, 2011.[127] S. H. Park, A. Roy, S. Beaupre´, S. Cho, N. Coates, J. S. Moon,D. Moses, M. Leclerc, K. Lee, and A. J. Heeger, “Bulk heterojunc-tion solar cells with internal quantum efficiency approaching 100%,”Nature Photonics, vol. 3, no. 5, pp. 297–302, 2009.[128] M. C. Scharber, D. Mu¨hlbacher, M. Koppe, P. Denk, C. Waldauf,A. J. Heeger, and C. J. Brabec, “Design rules for donors in bulk-heterojunction solar cells-towards 10% energy-conversion efficiency,”Advanced Materials, vol. 18, no. 6, pp. 789–794, 2006.[129] D. C. Olson, S. E. Shaheen, R. T. Collins, and D. S. Ginley, “The effectof atmosphere and ZnO morphology on the performance of hybrid poly(3-hexylthiophene)/ZnO nanofiber photovoltaic devices,” The Journalof Physical Chemistry C, vol. 111, no. 44, pp. 16670–16678, 2007.[130] L. Koster, V. Mihailetchi, R. Ramaker, and P. Blom, “Light intensitydependence of open-circuit voltage of polymer: fullerene solar cells,”Applied Physics Letters, vol. 86, no. 12, pp. 123509–123509, 2005.[131] C. H. Woo, B. C. Thompson, B. J. Kim, M. F. Toney, and J. M.Fre´chet, “The influence of poly (3-hexylthiophene) regioregularity onfullerene-composite solar cell performance,” Journal of the AmericanChemical Society, vol. 130.[132] B. J. Kim, Y. Miyamoto, B. Ma, and J. M. Frechet, “Photocrosslinkablepolythiophenes for efficient, thermally stable, organic photovoltaics,”Advanced Functional Materials, vol. 19, no. 14, pp. 2273–2281, 2009.[133] S. Miyanishi, K. Tajima, and K. Hashimoto, “Morphological stabiliza-tion of polymer photovoltaic cells by using cross-linkable poly (3-(5-hexenyl) thiophene),” Macromolecules, vol. 42, no. 5, pp. 1610–1618,2009.[134] B. Gholamkhass and S. Holdcroft, “Toward stabilization of domains inpolymer bulk heterojunction films,” Chemistry of Materials, vol. 22,no. 18, pp. 5371–5376, 2010.[135] B. Gholamkhass, N. M. Kiasari, and P. Servati, “An efficient invertedorganic solar cell with improved ZnO and gold contact layers,” OrganicElectronics, vol. 13, no. 6, pp. 945–953, 2012.127Bibliography[136] H. Gullapalli, V. S. Vemuru, A. Kumar, A. Botello-Mendez, R. Vajtai,M. Terrones, S. Nagarajaiah, and P. M. Ajayan, “Flexible piezoelec-tric ZnO-paper nanocomposite strain sensor,” small, vol. 6, no. 15,pp. 1641–1646, 2010.[137] V. Shrotriya, G. Li, Y. Yao, C.-W. Chu, and Y. Yang, “Transitionmetal oxides as the buffer layer for polymer photovoltaic cells,” AppliedPhysics Letters, vol. 88, no. 7, pp. 073508–073508, 2006.[138] J.-S. Huang, C.-Y. Chou, M.-Y. Liu, K.-H. Tsai, W.-H. Lin, and C.-F. Lin, “Solution-processed vanadium oxide as an anode interlayer forinverted polymer solar cells hybridized with ZnO nanorods,” OrganicElectronics, vol. 10, no. 6, pp. 1060–1065, 2009.[139] S. R. Hammond, J. Meyer, N. E. Widjonarko, P. F. Ndione, A. K.Sigdel, A. Garcia, A. Miedaner, M. T. Lloyd, A. Kahn, D. S. Ginley,J. J. Berry, and D. C. Olson, “Low-temperature, solution-processedmolybdenum oxide hole-collection layer for organic photovoltaics,”Journal of Materials Chemistry, vol. 22, no. 7, pp. 3249–3254, 2012.128Appendix ADuring my PhD about 70 NW-based devices were fabricated using dielec-trophoresis. The fabricated devices were tested and characterized and dif-ferent characteristics of such devices are shown in figure A.1. As shown infigure A.1a, the NW devices, depending on the area and quality of the con-tacts between ZnO and gold electrode at both ends of the NW, display fromrectifying behaviour to almost linear and ohmic characteristics. Except thesection 3.1 where we used both linear and rectifying devices for oxygen sens-ing, for the rest of the sections in chapter 3 we only tested the almost-lineardevices (5 samples are shown in A.1b) for sensor experiments.The steady-state ultra violet sensitivity of 7 extra samples in additionto the one reported in section 3.2 are displayed in figure A.2. Devices weretested under 1220 µW/cm2 of UV irradiation and the measured sensitivities(IUV /Idark) at room temperature and showing an average sensitivity of ∼15000.129Appendix A.-5.0 0.0 5.0-15-10-5051015-5 0 5-2024Voltage (V)Current (nA)Current (nA)Voltage (V)(a)(b)LinearRectifyingFigure A.1: (a) Current voltage characteristics of different DEP-assembleddevices at room temperature. (b) Characteristics of 5 almost-linear devicesused for sensor testing.130Appendix A.1 2 3 4 5 6 7 8103104105Sensitivity (I UV /I Dark)Sample NumberFigure A.2: Sensitivity of 8 DEP-assembled devices to 1220 µW/cm2 UVlight.131


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items