Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Characterization of poly(3-hexylthiophene) based Schottky diodes Dimopoulos, Alexandros Ioannis 2012

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2012_spring_dimopoulos_alexandros.pdf [ 2.47MB ]
Metadata
JSON: 24-1.0103463.json
JSON-LD: 24-1.0103463-ld.json
RDF/XML (Pretty): 24-1.0103463-rdf.xml
RDF/JSON: 24-1.0103463-rdf.json
Turtle: 24-1.0103463-turtle.txt
N-Triples: 24-1.0103463-rdf-ntriples.txt
Original Record: 24-1.0103463-source.json
Full Text
24-1.0103463-fulltext.txt
Citation
24-1.0103463.ris

Full Text

CHARACTERIZATION OF POLY(3-HEXYLTHIOPHENE) BASED SCHOTTKY DIODES by ALEXANDROS IOANNIS DIMOPOULOS B.Eng., University of Victoria, 2005  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE STUDIES (Electrical and Computer Engineering)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) April 2012 © Alexandros Ioannis Dimopoulos, 2012  Abstract This thesis describes the fabrication and electrical characterization of Schottky diodes based on the polymer semiconductor poly(3-hexylthiophene). Printed electronics may not be able to benefit from high vacuum processing, either for economic or technical reasons. The aim was to observe the effects on performance when Schottky diodes were built at atmospheric pressure. 200 nm thick films of poly(3-hexylthiophene) were formed on glass substrates by spinning a 1 wt% polymer solution in chloroform. Vacuum deposited aluminum and gold where used for the Schottky and ohmic contacts respectively. Two types of diodes were manufactured. One type (Au bottom) had its Schottky junction formed by evaporating aluminum onto the polymer under high vacuum. The other (Al bottom) had its Schottky junction formed by depositing the polymer onto aluminum at atmospheric pressure. The final yield of usable devices was 35% for Au bottom and 22% for Al bottom. The hole density and bulk mobility were derived from both DC and AC measurements. The bulk mobility was found to range from 2× 10!!   cm2 V -­‐1 s -­‐1 to 6×10!!   cm2 V -­‐1 s -­‐1 . The hole density was determined to be between 5×10!"   cm-­‐3 and 3×10!"   cm-­‐3 . DC measurements showed that Au bottom devices had a current rectification ratio of 2×10! at ±2  V, 100 times greater than Al bottom devices. The space charge limited current (SCLC) had to be considered to successfully model the DC behaviour. The small signal behaviour was modeled with a 2nd order series/parallel circuit, which was determined through impedance spectroscopy. Small signal performance of both device types was predicted to be poor. The corner frequency was determined to be less than 100 Hz for Al bottom devices, and less than 1 kHz for Au bottom devices. Large signal frequency performance of the diodes was tested with a half-wave peak rectifier. The maximum operating frequency was measured to be 40 kHz for Au bottom devices and 10 kHz for Al bottom devices.  ii  Table of Contents ABSTRACT .......................................................................................................... ii	
   TABLE OF CONTENTS ...................................................................................... iii	
   LIST OF TABLES ................................................................................................. v	
   LIST OF FIGURES .............................................................................................. vi	
   LIST OF ABBREVIATIONS ............................................................................... vii	
   ACKNOWLEDGEMENTS ................................................................................. viii	
   DEDICATION ...................................................................................................... ix	
   1	
   INTRODUCTION ............................................................................................. 1	
   1.1	
   THESIS ORGANIZATION .......................................................................................... 2	
   1.2	
   CONJUGATED POLYMERS ....................................................................................... 3	
   1.3	
   POLYTHIOPHENE .................................................................................................... 5	
   1.4	
   THE SCHOTTKY JUNCTION ..................................................................................... 7	
   1.5	
   CONCLUSIONS ....................................................................................................... 8	
    2	
   EXPERIMENTAL METHODS ....................................................................... 10	
   2.1	
   DESIGN ............................................................................................................... 10	
   2.2	
   FABRICATION ....................................................................................................... 12	
   2.3	
   MEASUREMENT .................................................................................................... 16	
   2.4	
   CONCLUSIONS ..................................................................................................... 17	
    3	
   DC MEASUREMENTS ................................................................................. 18	
   3.1	
   MEASUREMENT DESCRIPTION .............................................................................. 18	
   3.2	
   MODEL DESCRIPTION ........................................................................................... 19	
   3.3	
   MODEL FITTING ................................................................................................... 23	
   3.4	
   FIT RESULTS ....................................................................................................... 24	
   3.5	
   CONCLUSIONS ..................................................................................................... 29	
    4	
   SMALL SIGNAL AC MEASUREMENTS ..................................................... 30	
   4.1	
   DESCRIPTION OF THE C-V TECHNIQUE ................................................................. 30	
   4.2	
   GENERAL CONSIDERATIONS OF THE CV TECHNIQUE ............................................. 31	
   4.3	
   SPECIFIC CONSIDERATIONS WITH P3HT ............................................................... 32	
   4.4	
   MEASUREMENT DESCRIPTION .............................................................................. 33	
   4.5	
   DATA PREPROCESSING ........................................................................................ 33	
    iii  4.6	
   DATA PROCESSING .............................................................................................. 34	
   4.7	
   FIT RESULTS ....................................................................................................... 37	
   4.8	
   CONCLUSIONS ..................................................................................................... 41	
    5	
   PRACTICAL DIODE USES .......................................................................... 42	
   5.1	
   PEAK RECTIFIER .................................................................................................. 42	
   5.2	
   SMALL SIGNAL USE.............................................................................................. 45	
   5.3	
   CONCLUSIONS ..................................................................................................... 46	
    6	
   CONCLUSION .............................................................................................. 47	
   6.1	
   GENERAL CONCLUSIONS ...................................................................................... 47	
   6.2	
   FUTURE WORK .................................................................................................... 48	
    BIBLIOGRAPHY ................................................................................................ 50	
    iv  List of Tables TABLE 3.1: CURRENT RECTIFICATION RATIO AT ±2 V. FROM DATA. ................................................. 26	
   TABLE 3.2: DIODE IDEALITY FACTOR (N). FROM FITS. ...................................................................... 26	
   TABLE 3.3: RSH (ΩCM2). FROM FITS. .............................................................................................. 27	
   TABLE 3.4: JS (ACM-2). FROM FITS. ................................................................................................ 27	
   TABLE 3.5: HOLE MOBILITY (CM2V-1S-1). FROM FITS. ....................................................................... 28	
   TABLE 3.6: HOLE DENSITY (CM-3). INDIRECTLY FROM FITS. ............................................................. 28	
   TABLE 4.1: HOLE DENSITY FROM AC MEASUREMENTS (CM-3). ........................................................ 40	
   TABLE 4.2: BUILT IN VOLTAGE FROM AC MEASUREMENTS (V). ....................................................... 40	
   TABLE 5.1: FORWARD BIAS CUTOFF FREQUENCY (HZ). ................................................................... 45	
    v  List of Figures FIGURE 1.1: THE TWO EQUIVALENT FORMS OF TRANS-POLYACETYLENE. .......................................... 3	
   FIGURE 1.2: POLYTHIOPHENE. ........................................................................................................ 5	
   FIGURE 1.3: REGIOREGULAR POLY(3-HEXYLTHIOPHENE). ................................................................ 6	
   FIGURE 1.4: P3HT STACKING. ........................................................................................................ 6	
   FIGURE 1.5: THE IDEALIZED SCHOTTKY JUNCTION. .......................................................................... 7	
   FIGURE 2.1: SAMPLE LAYOUT. VERTICAL DIMENSIONS ARE NOT TO SCALE. ..................................... 11	
   FIGURE 2.2: METAL LIFT-OFF WITHOUT (LEFT) AND WITH (RIGHT) FIRST UNDERCUTTING THE PHOTORESIST. ...................................................................................................................... 13	
    FIGURE 2.3:SAMPLE HOLDER USED TO MAKE ELECTRICAL MEASUREMENTS OUTSIDE OF THE GLOVEBOX.  .......................................................................................................................... 17	
    FIGURE 3.1: TYPICAL J-V SCANS OF EXAMINED DEVICES. ............................................................... 19	
   FIGURE 3.2: SCHEMATIC OF THE MODIFIED SHOCKLEY EQUATION. ................................................. 20	
   FIGURE 3.3: DC MODEL USED. ...................................................................................................... 21	
   FIGURE 3.4: LOG-LOG J-V PLOT OF AN AL-BOTTOM DIODE (DEVICE B92L1Y) SHOWING DIFFERENT TRANSPORT REGIMES. THE CURRENT IS OHMIC AT LOW BIASES AND SPACE CHARGE LIMITED AT HIGH BIASES. ........................................................................................................................ 22	
    FIGURE 3.5:REPRESENTATIVE FIT RESULTS FOR AU BOTTOM (LEFT COLUMN), AU BOTTOM (F-SERIES) (CENTER COLUMN), AND AL BOTTOM (RIGHT COLUMN) DEVICES. FIT QUALITIES ARE BEST (TOP ROW), MEDIAN (MIDDLE ROW), AND WORST (BOTTOM ROW). ................................................... 25	
    FIGURE 4.1: A SMALL SIGNAL MODEL OF AN ORGANIC SCHOTTKY DIODE. ........................................ 32	
   FIGURE 4.2: EXAMPLE OF AC PRE-PROCESSING. ........................................................................... 34	
   FIGURE 4.3: IMPEDANCE SPECTRA OF A P3HT/AL SCHOTTKY DIODE (D93R1Z) BIASED AT 0.6 V (LEFT) AND -2 V (RIGHT). DATA PRESENTED HAS BEEN PREPROCESSED.................................. 35	
   FIGURE 4.4: FIT RESULT FOR DIODE F93R2Y BIASES AT -0.8 V. TOP: OBJECTIVE FUNCTION, BOTTOM: BODE PLOTS. ........................................................................................................................ 37	
   FIGURE 4.5: IMPEDANCE FIT RESULTS FOR DIODE F93R2Y. JUNCTION (Ÿ) AND BULK (+) VALUES ARE SHOWN................................................................................................................................. 38	
    FIGURE 4.6: DEPLETION CAPACITANCE FROM THE SUSCEPTANCE FOR DIODE F92R3Y. .................. 39	
   FIGURE 5.1: THE HALF-WAVE PEAK RECTIFIER. .............................................................................. 42	
   FIGURE 5.2: FREQUENCY RESPONSE OF RECTIFIER WITH AU BOTTOM (F-SERIES) DEVICES. ............ 43	
   FIGURE 5.3:FREQUENCY RESPONSE OF RECTIFIER WITH AL BOTTOM DEVICES. ............................... 44	
    vi  List of Abbreviations AC  Alternating current  DC  Direct current  DI  Deionized  HMDS  Hexamethyldisilazane  HOMO  Highest occupied molecular orbital  LUMO  Lowest occupied molecular orbital  ME  Mobility edge  P3HT  Poly(3-hexythiophene)  PTFE  polytetrafluoroethylene  RMS  Root-mean-square  SCLC  Space charge limited current  TFSCLC  Trap free space charge limited current  TFT  Thin film transistor  vii  Acknowledgements As I sail away, there are a number of people whom I would very much like to thank for my adventure. My supervisor Dr. John Madden brought me to the strange new shores of conducting polymers with an enthusiasm that is difficult to match. On many occasions I felt as if I was following a path carved into an otherwise impenetrable wilderness by Dr. Arash Takshi. My colleagues and friends in AMPEL 341 kept me company around the campfire through many dark nights. I also happily acknowledge my patrons, my family, who funded my expedition with a bottomless reserve of love and patience, and even a little gold.  viii  Dedication  To adventurers everywhere  ix  1 Introduction Inorganic semiconductors are a ubiquitous class of materials. Their widespread use is primarily due to integrated circuit (IC) technology, which was first conceived in the late 1950s when Jack Kilby demonstrated that all the elements of an oscillator circuit could be constructed on a single silicon crystal. Robert Noyce also independently formulated the concept [1]. Since then, ICs of every increasing complexity have become ubiquitous. This has been due to the ability to form ever-larger perfect crystals as well as an exponential decrease in circuit feature size. The result has been a doubling in circuit density every 18 to 24 months in what is commonly known as Moore’s Law. As popular as these materials have become, there are some applications for which they are not ideally suited. High performance electronics demands that these semiconductors be used in crystalline form. This limits their usefulness in flexible structures. The size of the IC is also limited by the size of the crystal. In applications where performance is less demanding, the semiconductors can be used in amorphous or polycrystalline forms. This allows much larger circuits to be built, but the high temperatures needed to process them still generally exclude the use of flexible substrates. Some other drawbacks of silicon devices is that the bandgap is not controllable, a fact which presents difficulties for sensors, light emission, and photovoltaics. Antennas cannot be directly integrated into ICs, they must instead be externally affixed. This represents a fixed cost in manufacturing which does not scale with Moore’s law [2]. Various avenues are being examined to extend semiconductors past these limitations. Some are process based, such as slicing devices from crystalline wafers and transplanting them onto a flexible substrate [3], or finding ways to lower the temperature needed to form poly-Si [4]. Others involve using new materials such as graphene, engineered nanoparticles, or organic semiconductors. Organic chemistry is a well-established discipline, which can effectively design, synthesize, and purify new materials. This presents two interesting opportunities. First, properties such as the bandgap can be tuned. Second, the material can be produced inexpensively. Organic semiconductors can be divided into two groups: small molecules 1  and conjugated polymers. Many conjugated polymers have the advantage of being soluble in liquid solvents at room temperature. Thus films of these polymers can be formed at room temperature, which means that flexible substrates can be used. Polymer solutions could be used as inks and integrated into industrial printing processes, allowing arbitrarily large circuits to be produced inexpensively. Printing would also allow direct integration with elements which cannot be built in the semiconductor, such as antennas. For all of these interesting reasons, a conjugated polymer was used in this work. Specifically, the well-known polymer poly(3hexylthiophene) was used. The diode is a basic and versatile electronic device, with applications that include rectification, demodulation, mixing, and filtering. The vast majority of conjugated polymers are p-type semiconductors, which means that the Schottky diode is only diode that can be practically built with them. Though a number of examples of P3HT based Schottky diodes have been reported [5] [6] [7] [8] [9] [10], they are always formed under high vacuum. This is a perfectly reasonable approach when exploring material properties or as device demonstrations. This is however not ideal for considering practical manufacturing issues since a high vacuum may not be economical or compatible with processes such as roll-to-roll printing. A better analogue for a device built using “cheap and dirty” manufacturing would be one formed at atmospheric conditions. This thesis describes the construction and characterization of two types of Schottky diodes. The first type had its Schottky junction formed under high vacuum, and the second had its Schottky junction formed at atmospheric pressure.  1.1 Thesis Organization The rest of this introductory chapter presents an overview of background information on conjugated polymers in general, poly(3-hexylthiphene) in particular, and Schottky junction theory. The rest of this thesis is divided into four chapters. Chapter 2 covers the design and fabrication of the Schottky diodes as well as experimental methods. Chapter 3 discusses the DC measurements, modeling, and characterization. Chapter 4 does the same, but for small signal AC measurements. Chapter 5 briefly  2  considers the use diodes in electronic circuits. The final chapter concludes the work and presents recommendations for future work.  1.2 Conjugated Polymers Polymers are generally thought of as being electrically insulating, and are commonly used in this regard. For a solid to be electronically conductive, its electrons need to be delocalized. This is not the case for many organic molecules since their electrons are highly localized in σ orbitals [11]. A polymer that has alternating single and double bonds along its backbone is known as a conjugated polymer. The double bonds are due to π orbitals, which are delocalized. Polyacetylene is a good illustrative example. It was also the first polymer to be observed with a high conductivity, a discovery that lead to a Nobel Prize in Chemistry [12]. The trans form of this polymer has two energetically equivalent forms, which are shown in Figure 1.1. It is tempting to assign a resonance structure to transpolyacetylene and assume that conduction arises from a π orbital perfectly distributed over the entire molecule, as in benzene. In reality, the resonance form, which would result in metallic conduction, is unstable and polyacetylene takes a conjugated form. The double carbon bonds are stronger than the single bonds and are thus shorter. This causes a periodic distortion in the backbone, known as the Pieirls distortion, which results in a splitting of the energy band. This mechanism is present in all conjugated polymers and explains their semiconducting nature [11].  Figure 1.1: The two equivalent forms of trans-polyacetylene.  Conjugated polymers have wide bandgaps, polyacetylene for example has a bandgap of 1.7 eV [12]. Accordingly, they have low intrinsic conductivities. Analogously to inorganic semiconductors, they can be doped. The dopant species can either react with the polymer in the form of a redox reaction, or affect it through electron-electron repulsion [5]. In terms of energy structure, the result is a localized state in the bandgap. At high enough doping levels, the bandgap can be filled and the polymer will exhibit metallic conduction [11]. Molecular oxygen and water are known dopants for many 3  conjugated polymers [5]. This presents a problem since these polymers cannot be used in an ambient atmosphere without introducing uncontrolled doping. Thus far, only conduction along a single polymer chain has been considered. This is a one-dimensional description and is inadequate for a thin film. The carrier mobility in a thin film depends on the morphology since charges must travel between polymer chains as well as along them. The π orbitals of adjacent polymer chains can overlap if the chains are stacked together. This situation, known as π- π stacking, extends charge delocalization into a second dimension. This can lead to a semicrystalline film composed of π- π stacked domains separated by amorphous regions. While the mobility inside the domains may be large, the overall mobility is limited by inter-domain transport. The degree of π- π stacking, and thus the mobility, is influenced by the specifics of the deposition method and the surface energy of the substrate [13]. There is a low degree of π- π stacking in an amorphous film and charges must hop between polymer chains. There is evidence of correlation between chain length and hopping rate. A longer chain is more likely to have a region with a lowered barrier somewhere along its length [5]. Because the charge delocalization is limited, it is not strictly correct to refer to energy bands, as they may be very narrow or even nonexistent. Instead of a valence and conduction band, the analogous concepts of highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) are respectively used. The bandgap is thus the HOMO-LUMO energy difference [5]. Several mechanisms have been proposed to explain transport in conjugated polymers. For semi-crystalline films, a mobility edge (ME) model has been proposed and is considered promising [13] [14]. In this model, the ME is a defined energy that separates mobile states from localized states. The mobile states are considered to form a band and their density varies slowly at energies close to the ME. The localized states, which are associated with grain boundaries and disordered regions, extend away from the ME into the bandgap. Their density changes exponentially with energy [15]. The regions of localized states are known as the band tails and their widths are a measure of disorder in the semiconductor [14]. The model assigns a constant mobility to band states and a mobility of zero to tail states. Transport occurs when charges are thermally promoted from the tail to the band. Accumulating charge in the semiconductor shifts the 4  Fermi level towards the ME and increases the mobility [15]. This is done in a thin film transistor (TFT) by charging the gate; the result is the increased field-effect mobility. It can also be accomplished by doping the polymer [5]. Many conjugated polymers behave as single carrier materials. This is can be explained as an asymmetry in the band tails [5]. If for example, the tail extending from the LUMO is wider than the tail extending from the HOMO, electrons will be more localized and thus less mobile. The result is an intrinsically p-type material.  1.3 Polythiophene Polythiophene, shown in Figure 1.2, was first prepared in 1981 through electropolymerization of thiophene. Like polyacetylene, it is a planar molecule, it is a semiconductor, and it can be doped.  Figure 1.2: Polythiophene.  Also like polyacetylene, it is unfortunately insoluble. Conjugated polymers can be made soluble by substituting side groups longer than butyl (4 carbon atoms) [11]. A well-known example is the polymer used in this work, poly(3-hexylthiophene) (P3HT), shown in Figure 1.3. It is an intrinsically p-type semiconductor. Field-effect mobilities above 0.1  cm2 V -­‐1 s -­‐1 have been observed in TFTs [13], though the bulk mobility is commonly three to five orders of magnitude smaller [7] [16] [17] [8] [9]. P3HT has a bandgap of 1.7 eV and an electron affinity of 3.15 eV [18].  5  Figure 1.3: Regioregular poly(3-hexylthiophene).  Because the monomers are asymmetric, they can couple in several different orientations when forming the polymer. One of the couplings is particularly undesirable because neighbouring side chains interact with each other, causing the backbone to twist. This disrupts the π orbital and widens the bandgap. A high degree of regular monomer coupling, known as regioregularity, is thus desirable [11]. Solution deposited P3HT which is highly regioregular tends to self assemble in edge-on stacks [13], as shown in Figure 1.4. The mobility is anisotropic since the π- π stacking delocalizes charges in planes parallel to the substrate, which are separated from each other by the insulating hexyl groups.  Figure 1.4: P3HT stacking.  6  1.4 The Schottky Junction Depending on their relative Fermi levels, either a Schottky or ohmic junction may form when a metal and a semiconductor are brought into intimate contact. Charges will flow between the two until the semiconductor’s Fermi level is brought in line with the metal’s. The system will be in thermal equilibrium at this point [19]. Figure 1.5 shows the energy diagram of a Schottky junction to an idealized crystalline p-type semiconductor. There are no traps in the bandgap, nor are there any interfacial states in this case. The work function of the metal is smaller so holes flow from the semiconductor to the metal. The migration of holes out of the semiconductor results in a zone that is depleted of mobile charge known as the depletion region. This negative space charge causes the bands to bend down. To compensate this negative space charge, a sheet of positive charge has built up on the metal’s surface. As can be seen in Figure 1.5, holes passing from the metal to the semiconductor must overcome a potential barrier:  !! = ! +  !! − Φ! !  (1.1)  where ! is the semiconductor’s electron affinity, !! is the semiconductor bandgap, Φ! is the metal’s work function, and q is the electron charge. !"  )*)+,-./ )/)-&0  ! !"  !!  !# !&  ! '  !$ !%  !%'(  1.2(,(./ Figure 1.5: The idealized Schottky junction.  7  This barrier depends only on material properties and is not affected by an applied bias. Holes moving in the other direction must overcome the band bending, which at zero applied bias is characterized by the built-in voltage !!" . Since applying an external bias can both reduce and increase the band bending, the Schottky junction is rectifying [19]. If instead, the metal’s work function is larger than the semiconductor’s, an ohmic junction occurs. The semiconductor bands bend up and there is an accumulation of holes at the junction. No barrier is encountered if a bias is applied so that holes flow from the semiconductor to the metal. Unlike in the Schottky junction, current is not impeded when a reverse bias is applied. The accumulated charge acts like an anode and is able to provide a large number of holes. The current is thus limited by the semiconductor bulk [20]. The idealized assumptions made when describing the Schottky junction have to be modified when considering a conjugated polymer. Interface states will be present unless the semiconductor surface is entirely free of oxides and defect free. These states reduce the Schottky barrier’s size from the ideal and cause it to become bias-dependent [20]. In the idealized case, the band edges are sharp and holes come from a shallow acceptor level. In the polymer, the holes come from localized states in the band tail. Numerical simulations have shown that a wider tail reduces the built-in voltage [14]. The charge density in the depletion region is likely to be smaller than the dopant density since many holes are trapped in the tail [14]. The idealized case assumes that the space charge density and the dopant density are equal. The overall result is a shorter depletion width than predicted by the idealized case [14].  1.5 Conclusions Poly(3-hexylthiophene) has a lower mobility than crystalline semiconductors, but has the advantage of being solution processable at room temperature. It may not be possible to construct inexpensive printed diodes in a high vacuum. It would therefore be interesting to see how their performance is affected by building them at ambient pressures.  8  Schottky diodes are formed by selecting metals and semiconductors with appropriate differences in their work functions. The ideal Schottky junction theory can be used to approximately explain the diodes built in this work. However, certain limitations must be kept in mind. Interface states introduced into the junction by contaminants cause the Schottky barrier to be bias-dependent. The disordered nature of P3HT will likely cause the band bending to be overestimated. The discussion begins with experimental methods. The design and fabrication of the P3HT Schottky diodes is discussed, as well as measurement issues.  9  2 Experimental Methods This chapter begins by discussing the design of the Schottky diodes. Material selection is discussed as well as justifications for physical dimensions. Next, fabrication details are covered, including difficulties encountered. Methods to avoid unintentional doping of the semiconductor both during manufacturing and testing are also explained. The test fixture used in measurements is described at the end of this chapter, though specifics of each measurement are covered in appropriate chapters later in the thesis.  2.1 Design In its most basic configuration, a Schottky diode consists of a semiconductor sandwiched between two electrodes with differing work functions chosen to form a Schottky contact on one side, and an Ohmic contact on the other. The semiconductor, poly(3-hexylthiphene), is a well known soluble conjugated polymer. While the semiconductor was a solution processed organic material, the contacts were vacuum deposited metals. P3HT has a band gap of 1.7 eV and an electron affinity of 3.15 eV [18]. Thus under idealized conditions, the Schottky contact should have a work function of less than 4.85 eV, and the Ohmic contact, a work function larger than 4.85 eV. Aluminum (Al) and gold (Au) with work functions of 4.05 eV and 5.2 eV respectively [6], meet this criteria. This has in fact been extensively observed [6] [5] [7] [8] [9] [10]. Other metals, such as calcium and magnesium, have smaller work functions than aluminum. However, these materials are more reactive and hence more difficult to work with and diffuse more rapidly into the polymer than aluminum [5]. The Schottky diodes were built in a vertical configuration by laying thin films of the materials one on top of the other. The two types of devices were made by switching the stack order. Au bottom devices were built in the order of Au/P3HT/Al, and Al bottom devices in the order of Al/P3HT/Au. The substrate needed to be smooth and rigid since the P3HT was to be spincoated. Si wafers with a SiO2 coating were readily available and have the required physical properties. However, they have a parasitic capacitance large enough to prevent proper AC measurements [5]. Glass on the other hand does not have this problem. The devices were built on 1” square chips made by dicing large format 10  microscope slides measuring 3” by 2”. This chip size was small enough to get multiple electrode sets from a single glass slide, and large enough to comfortably manipulate in a glovebox. Each 1” chip was divided into a 3×3 grid using 3 top and 3 perpendicular bottom electrodes, to form 9 individually addressable diodes, as shown in Figure 2.1. A chip with a set of diodes will be referred to as a sample.  Figure 2.1: Sample layout. Vertical dimensions are not to scale.  Since the current direction in the P3HT film is normal to the substrate, a thinner film will reduce the series resistance caused by the semiconductor bulk. However, very thin films are more susceptible to pinhole defects, which short-circuit a diode. No special effort was made to minimize the semiconductor film thickness for fear of these defects. It is difficult to model the film thickness from process parameters. The thickness was assumed to be in the range of 100 nm to 250 nm based on published examples that employ the same process parameters [5] [4] [9]. Current through the electrodes is primarily parallel to the substrate, and thus thicker electrodes reduce the series resistance. A semiconductor film thicker than the underlying electrode may be problematic. The film could be discontinuous at the electrode edge, which would cause a short circuit. The electrode thickness was easily controlled as it was continuously measured during vacuum deposition. The electrode thickness was set to about 100 nm, which was thinner than the semiconductor. The 11  series resistance contribution from the electrodes should be no more than a few Ohms, and is thus negligible. The width of the electrode lines specifies the cross sectional area of the diodes. A large area increases the device current, making measurements easier. A smaller area reduces the chance of encountering a pinhole in the P3HT film. Electrical measurements of diodes of similar construction, and with a cross sectional area of 2×10!!   m! have been reported [5]. This set an effective lower limit on the area of the diodes since the same test equipment was used in this work. Such an area implies an electrode width of about 50  µμm, half the width of a human hair. A shadow mask could not be machined to that size in the available facilities. Features of this size would also make alignment by eye inside the glovebox difficult. The width of both the top and bottom electrodes was set at 1 mm.  2.2 Fabrication All bottom electrodes were patterned by photolithography instead of shadow masking in order to save time through a higher throughput. Photolithography was performed in the UBC nanofabrication facility which has a general cleanroom classification of 10 000, except for the lithography room which has a classification of 1000, and this room’s wet-benches which are classified at 100 [21]. As noted above, the substrates were glass microscope slides measuring 2” by 3” with a thickness of about 1 mm. The substrates were first cleaned in order to ensure proper adhesion of the photoresist. The cleaning sequence was: a boiling acetone bath, a boiling isopropyl alcohol bath, a deionized (DI) water rinse, drying with compressed nitrogen, and a two minute dehydration bake on a hotplate. Once the substrates were properly dried, they were primed with hexamethyldisilazane (HMDS) by spin coating in order to improve photoresist adhesion. The positive photoresist S1813 was then applied by spin coating at 5000 rpm for 50 s. The photomask was printed onto a Mylar sheet using an ordinary office laser printer with a resolution of 600 dpi. Metallization was done in a large electron beam evaporator capable of holding half a dozen substrates at a time. The unwanted portions of the newly deposited metal films were washed away with acetone, which dissolves the underlying photoresist. Once lift-off was complete, the 12  substrates were diced with a diamond saw and cleaned before being transferred into the glovebox. Cleaning was done with acetone, isopropyl alcohol, DI water, and finally an oxygen plasma treatment. Figure 2.2 depicts how undercutting the photoresist before metallization can improve lift-off results. Patterned photoresist always has sloping edges. This greatly increases the chances that metal deposited onto the bare substrate will be connected to the metal on top of the photoresist. When lift-off is performed, the edges of electrodes will be shorn off.  !"#$%## Figure 2.2: Metal lift-off without (left) and with (right) first undercutting the photoresist.  This would not noticeably affect the final size of the electrodes due to their very small aspect ratio. Sharp vertical protrusions do present a problem as they might reach through the P3HT layer and cause a short circuit. This was the suspected failure mechanism in early test devices. Undercutting the photoresist creates an overhanging structure that will cause a break in the subsequently deposited metal film, thus avoiding sheared edges. Photoresist undercutting was attempted via two approaches: dipping the exposed photoresist in toluene [22], and using an optical diffuser during exposure [23]. Neither technique proved effective. Profilometery of several completed electrodes revealed sheared edge spikes ranging from 50 nm to 200 nm beyond the electrode surface. The spikes were reduced to a few tens of nanometers after manual polishing of the electrodes. All samples were polished before use regardless of whether or not  13  undercutting was performed. Unfortunately, these complications cancelled the expected time savings. The resolution and contrast of the photomask were likely too low to define adequately sharp edges. A mask produced with a better printer may have yielded better results. There are also bilayer photoresists specifically for lift-off, but these were not readily available. The cleaned electrodes were immediately transferred to an argon-filled glovebox where the remaining manufacturing took place. The glovebox was outfitted with a built in thermal evaporator, and was also equipped with a hot plate, a spin coater, and an analytical balance. The argon used had a purity of 99.999%. The system included a solvent filter as well as a water and oxygen trap. However, neither oxygen nor moisture sensors were present so some impurities may have been present. Solution processing of conjugated polymers has long been considered the key to developing low-cost organic electronics. High-speed printing techniques are of particular interest, but are difficult to implement and beyond the scope of this work. Common processing methods used in device research are dip casting and spin coating. Dip casting involves dipping a substrate into the polymer solution and then slowly drawing it out. As the substrate is pulled away from the solution, the solvent on it evaporates, and a solid film is left behind. The process is capable of forming thin, well ordered films, but the drawing speed must be very slow. A typical drawing rate is 1 mm/min [24]. Also, enough solution must be prepared so that a sample may be totally submerged. In spin coating, a small volume of solution is deposited on a substrate, which is then spun at high speed to evenly spread the solution. The solvent quickly evaporates, leaving a solid film behind. The length of time needed depends largely on the evaporation time of the solvent used. It is much faster than dip coating since the entire substrate dries at once, typically less than one minute. Since the film dries more quickly than in dip casting, the polymer chains have less time to orient themselves, resulting is a less ordered film. This can be somewhat affected by selecting a solvent with a higher boiling point [25]. 14  The P3HT films were deposited by spin coating since a spin coater was available. It is also a better analogue for high speed manufacturing processes due to the shorter processing time. P3HT was obtained from Sigma-Aldrich and used as received. The polymer was in the form of crystalline granules; it had a number average molecular weight of 64 kDa, and a regioregularity of over 98.5%. Chloroform was used as the solvent as it is known to dissolve P3HT well [5] [7] [9]. Concentrations of about 1% by weight yield films of around 120 nm in thickness [5] [9]. A concentration of 15 mg/mL was used, corresponding to almost exactly 1% by weight. The solvent and polymer were placed in a sealed vial and stirred for at least two hours at 40 °C. The solution was then passed through a 0.45 μm polytetrafluoroethylene (PTFE) syringe filter into a clean receptacle and immediately used in order to remove any remaining particles. Using a pipette, each sample was entirely coated with the solution and then spun at 1000 rpm for 40 s. This process was immediately repeated to form a double-layer P3HT film. Adding this second layer reduced the number of short-circuited devices from about 50% to 30%. This second layer likely plugged pinholes in the underlying film. The samples were then placed on a hotplate set to 120 °C for 10 minutes to drive out any remaining solvent. The film thickness was later determined by profilometry to be about 230 nm. The final manufacturing step was to deposit the top electrodes. The samples were fitted with a shadow mask and placed in the glovebox’s built-in thermal evaporator. The deposition rate was kept to about 1  Ås -­‐1 . It has been suggested that a low deposition rate may limit metal diffusion into the polymer and thus reduce the occurrence of short circuits [5]. The electrodes were deposited to a thickness of 100 nm at a pressure of 2×10!!   Torr. One set of samples (the F-series) had their top electrodes deposited at a pressure of 1.5×10!!   Torr, due to a fault in the evaporator. The resulting devices still functioned as diodes, but their reduced performance was not recognized until the data was analyzed.  15  A total of 11 Au bottom, and 12 Al bottom samples, each with 9 devices, were manufactured. 30% of both Au bottom and 40% of Al bottom devices were shortcircuited. Eliminating devices which had an extremely poor DC reverse characteristic or which short-circuited during measurement further reduced the yield. The final yield for Au bottom devices was 35% (35 devices) and 22% (24 devices) for Al bottom devices.  2.3 Measurement Measurement leads were limited to 30 cm by performing electrical measurements outside of the glove box. A special sample holder was prepared so that finished samples could be taken out of the glovebox without exposure to oxygen or water. A small polypropylene box with a gasket was fitted with wires and a set of gold plated spring clips to hold one sample at a time. The assembly is shown in Figure 2.3. The sample holder was placed inside of a custom-made grounded aluminum shield box in order to reduce interference during measurements. Relevant measurement details are covered in the next chapters. The thicknesses of the P3HT films were measured by profilometry once electrical measurements were completed.  16  Figure 2.3:Sample holder used to make electrical measurements outside of the glovebox.  2.4 Conclusions P3HT based Schottky diodes were successfully manufactured using two different sequences. Devices with Au as the bottom layer had a final yield of 35 %. Devices with Al on the bottom had a yield of 22%. The higher throughput of the lift-off process did not translate into saved time due to difficulties encountered. Patterning by shadow mask would also have been effective.  17  3 DC Measurements A current density-voltage (J-V) scan is the most basic way to examine a diode. In this chapter, a DC model of the P3HT/Al diodes is described and fitted to measurements. The model parameters reveal information about the quality of the diodes as well as the bulk hole mobility in the P3HT film. The hole density is also derived.  3.1 Measurement Description Measurements were made with the Model 6430 source measurement unit (SMU) from Keithley Instruments Inc. [26]. The I-V scans were performed outside of the glovebox, as described in chapter 2. The bias was applied as a linear staircase sweep with a step size of 10 mV. There was a delay time of 100 ms between the rising edge and the start of a measurement. Each measurement was averaged over one power line cycle, or 17 ms. The temperature was not controlled and was assumed to be 296 K. Each device was measured multiple times over periods of time stretching from several hours to about 10 days. Typical results are shown in Figure 3.1. The F-series devices are clearly inferior to the rest of the Au bottom devices. They are included as they were also used for small signal tests presented in chapter 4.  18  ï1  10  ï2  10  ï3  10  Current density (Aucmï2)  ï4  10  ï5  10  Au bottom Au bottom (F series) Al bottom  ï6  10  ï7  10  ï8  10  ï9  10  ï10  10  ï2  ï1  0  1 Applied bias (V)  2  3  4  Figure 3.1: Typical J-V scans of examined devices.  3.2 Model Description The current density of a Schottky diode is ideally modeled by the Shockley equation  ! = !! exp  !" − 1 = !! exp !" − 1 !"#  (3.1)  where ! is the ideality factor and !! is the reverse saturation current density. The exact expression of !!   depends on the transport mechanism over the barrier, which is not exactly understood and is beyond the scope of this work. A diffusion mechanism is expected due to the low mobility of carriers in P3HT, causing !! to depend on both temperature and bias [5] [27]. As can be seen in Figure 3.1, the forward current of the measured devices is not purely exponential. The recorded data cannot be modeled by 19  an ideal diode alone, and a more elaborate model is instead needed. Figure 3.2 depicts the equivalent circuit of the modified Shockley equation, which is commonly used in photovoltaics. The series resistor !! accounts for the voltage drop across the semiconductor bulk, and the shunt resistance !!" models the bias dependence of !! . This model was not used as it has been shown to be inadequate for modeling a polymer based Schottky diode [28].  Figure 3.2: Schematic of the modified Shockley equation.  The space charge limited current (SCLC) needs to be considered due to P3HT’s low mobility. This model describes transport in low conductivity materials and is typically applied to insulators, including depleted semiconductors. A density of injected charge exceeding the material’s intrinsic carrier density causes the formation of a space charge region. As a result, the current is no longer ohmic. SCLC becomes dominant in P3HT at electric fields exceeding 10!   V∙cm-­‐1 [7]. This should occur at an applied bias of about 2 V for the majority of devices examined in this work. This effect can be accounted for by adding a non-ohmic impedance element in parallel with !! , as depicted in Figure 3.3. Here, the semiconductor bulk is represented by the parallel combination of the resistor !! and the SCLC impedance element. The actual junction is modeled with the ideal diode ! and the resistor !!" .  20  Figure 3.3: DC model used.  The applicability of this model to the current work is suggested by its successful application in similar published work [7]. Examining the I-V data on a log-log plot, as in Figure 3.4, lends further support. Linear regions in the plot indicate regions where the current density is a power function of bias, with the exponent equal to the slope of the line. The unity slope at low biases indicates ohmic conduction. The slope at high biases is 2, which indicates a trap free space charge limited current (TFSCLC) [16]. This means that all the traps in the bulk have been filled [16] and that the current density can be expressed as !=  9 !! !! ! ∙ !! = !! ! 8 !!  (3.2)  where ! is the thickness of the semiconductor film, !! !! is the permittivity of the semiconductor, and ! is the bulk mobility [7]. Given this dependence, the bulk mobility can be easily extracted from the J-V data, provided that data is recorded up to biases in the SCLC regime.  21  data slope = 1.03 slope = 2.16  ï3  10  Current Density (Aucmï2)  5  ï1  Eïfield ~ 10 Vucm  ï4  10  ï5  10  ï6  10  ï7  10  ï2  ï1  10  0  10  10 Applied Bias (V)  Figure 3.4: Log-log J-V plot of an Al-bottom diode (device B92L1Y) showing different transport regimes. The current is ohmic at low biases and space charge limited at high biases.  Three equations govern the DC model. The voltage across the entire device is given by !! = !! + !!  (3.3)  where !! is the voltage dropped across the semiconductor bulk, and !! is the voltage across the junction. The current density through the semiconductor bulk is described as !=  !! + !!!! !!  (3.4)  where ! is proportional to the bulk mobility, as described in (3.2). The current density across the junction is equal to the current density through the bulk and is given by !=  !! + !! exp !!! − 1 !!"  (3.5)  22  where ! includes the ideality factor !, which is a measure of the quality of the rectifying junction, as described in (3.1). A perfect junction has an ideality of 1; an increase in ! corresponds to a decrease in quality.  3.3 Model Fitting Determining the model parameters is not a straightforward process. They cannot be directly extracted from measured data since the model depends on the hidden variables !! and !! . The strategy employed was to iteratively solve the set of equations (3.3)-(3.5) using the nonlinear least-squares facilities in MATLAB. The algorithm is outlined here: 1. Solve (3.4) and (3.5) over the forward bias for the parameters. • Assume that !! and !! are correctly known. • Update the parameter values. 2. Solve the system of equations at each bias point for variables !! and !! . • Assume that the parameters are correctly known. • Update the voltage values. 3. Check the parameters for convergence. • Repeat if needed. Convergence was defined as a relative change of less than 10!!   between iterations. The maximum number of iterations was set at 100. There is no guarantee that the solver would converge to the desired solution, or even to any solution, when starting from an arbitrary initial guess. It is therefore important to start with an initial guess as close as possible to the final solution. This was done by dividing the data into three bias ranges and assuming different model elements dominated in each one. The resistors were assumed to dominate over reverse biases, the SCLC element at high biases, and the diode and resistors over the intermediate range. As discussed earlier, the bias at which the SCLC begins to dominate can be found on a log-log plot. The SCLC parameter, !, was estimated from a linear fit to the high bias data on a log-log scale. For the other parameters, a first coarse estimate was refined by least 23  squares fitting. !! was coarsely estimated from the physical dimensions of the P3HT film and an assumed resistivity of 1.47×10!   Ωcm based on literature values [29].   !!" was coarsely estimated by performing a linear fit over the reverse bias range and subtracting the coarse value of !! .The contributions of !! and !!" over the intermediate bias region were cancelled, and then !! and ! were estimated from a linear fit in a semilog scale. A least squares fit over the intermediate bias range was then performed in order to refine the initial values of all parameters but !. Initial estimates of the internal voltages were then found by solving (3.4) and (3.3) for !! and !! respectively.  3.4 Fit Results Each device had been measured multiple times. The DC model was fitted to each resulting data set, with varying results. Only the model parameters from the best fits were considered. This required a “quality of fit” metric  !"# =  1 !  !  !!!  !! − !! !!  !  (3.6)  where !! is the measured current density and !! is the fitted current density. Good fits will have small QoF values. For each device, only the fit with the best QoF was retained. Figure 3.5 shows some representative results. Fitting was most successful with Al bottom devices, least successful for Au bottom devices excluding the F-series, and F-series fits falling in between. The fitter may have had more difficulty with very low current data, as was present in most Au bottom devices. A finer tuning of the parameter bounds imposed on the fitter may correct this. The results presented below are from a further reduced set of fits. For each device type, only fits with an above median QoF were retained. The minimum, median, and maximum values listed are from this reduced set of fits.  24  ï1  ï1  10  QoF = 1.3Eï1  ï7  10  ï10  1  2 bias (V)  3  2 bias (V)  3  1  2 bias (V)  3  10  ï6  10 10  4  1  2 bias (V)  3  ï6  J (Aucmï2)  10  ï8  10  1 bias (V)  1.5  2  1 bias (V)  1.5  2  1 bias (V)  1.5  2  QoF = 1.8Eï2  ï4  10  ï6  10  10  4  ï1  10  QoF =1.7  0.5  ï8  0  0  0.5  ï2  10  QoF = 1.4Eï1  ï3  J (Aucmï2)  ï4  0  ï2  10  QoF = 8.1Eï2  ï8  0  ï6  10  10  4  J (Aucmï2)  J (Aucmï2)  J (Aucmï2)  ï8  ï10  J (Aucmï2)  1  ï3  10  ï4  10  ï8  0  ï1  10  QoF = 3.1Eï1  ï5  10  ï6  10 10  4  10  10  10  ï8  0  ï3  10  QoF = 8.4Eï3  J (Aucmï2)  ï4  10  10  ï2  10  QoF = 3.0Eï2  ï3  J (Aucmï2)  J (Aucmï2)  10  10  ï5  10  QoF = 1.4Eï1  ï4  10  ï10  10  ï7  0  1  2 bias (V)  3  4  10  ï7  0  1  2 bias (V)  3  4  10  0  0.5  Figure 3.5:Representative fit results for Au bottom (left column), Au bottom (F-series) (center column), and Al bottom (right column) devices. Fit qualities are best (top row), median (middle row), and worst (bottom row).  25  The first results, presented in Table 3.1, are directly from the data instead of the fit results. These are the current rectification ratios at ±2 V. At 2 V, the bias is large enough for the device to be well out of the ohmic region, but the current should not yet be dominated by the SCLC mechanism. The Au bottom devices, excluding the F-series, demonstrated a rectification of 1.8×10! , which is in line with reported values [16] [4]. The Au bottom devices outperformed the Al bottom devices by two orders of magnitude. Unexpectedly, the F-series Au bottom devices had the smallest rectification ratio of all. Device type  Min  Median  Max  Au bottom  3.3×10!  1.8×10!  5.6×10!  Au bottom (F-series)  1.1×10!  2.0×10!  1.8×10!  Al bottom  7.3×10!  2.2×10!  7.5×10!  Table 3.1: Current rectification ratio at ±2 V. From data.  The ideality factors of the diodes were extracted from the fits by assuming a temperature of 296 K (a thermal voltage of 26 mV) and are presented in Table 3.2. The results from the Au bottom devices are comparable to published results [10] [28] [7] [5]. The Al bottom devices could not be compared to reported examples since ideality factors for such a device geometry have not previously been reported. Device type  Min  Median  Max  Au bottom  1.8  3.4  4.4  Au bottom (F-series)  3.5  9.9  12  Al bottom  3.1  4.4  6.5  Table 3.2: diode ideality factor (n). From fits.  !!" and !! are also directly related to the quality of the junction and are presented in Table 3.3 and Table 3.4.  26  Device type  Min  Median  Max  Au bottom  5.3×10!  3.7×10!  2.6×10!  Au bottom (F-series)  2.5×10!  3.8×10!  8.2×10!  Al bottom  7.9×10!  1.9×10!  5.5×10!  2  Table 3.3: RSH (Ωcm ). From fits.  Device type  Min  Median  Max  Au bottom  2.1×10!!"  7.9×10!!"  6.1×10!!!  Au bottom (F-series)  7.0×10!!  3.1×10!!  1.3×10!!  Al bottom  6.2×10!!  1.3×10!!  2.5×10!!  -2  Table 3.4: Js (Acm ). From fits.  The model parameters which relate to the junction are generally consistent with the observed diode rectification. That is, the Au bottom devices, excluding the F-series, have the superior junctions, and the F-series devices have the worst junctions. The !!" values are however inconsistent, indicating that the Al bottom devices have the worst junctions by a factor of two. Small signal junction resistances derived from AC measurements, presented in chapter 4 also indicate that the F-series devices have the least ideal junctions. It was expected that the Al bottom devices would be the worst performing. The Al electrodes of these devices were exposed to oxygen before deposition of the P3HT layer, which undoubtedly led to the formation of an interfacial oxide layer. In fact, the worst performing devices were the Au bottom F-series devices. The Schottky junctions of these devices were formed at a higher pressure than the rest of the Au bottom devices. A higher level of contamination could therefore be anticipated. The fact that these devices had an even poorer performance than the Al bottom devices suggests a large amount of contamination, possibly oil from the molecular diffusion pump. Results for both hole mobility and hole density fall within the range of reported values [16] [17] [4] [7] [9]. The mobility depends only on the model 27  parameter !. The hole density cannot be determined from a single model parameter. It can only be extracted from !! together with the mobility. Device type  Min  Median  Max  Au bottom  4.4×10!!  2.5×10!!  5.2×10!!  Au bottom (F-series)  1.0×10!!  2.3×10!!  1.0×10!!  Al bottom  8.0×10!!  5.8×10!!  1.2×10!!  2  -1 -1  Table 3.5: Hole mobility (cm V s ). From fits.  Device type  Min  Median  Max  Au bottom  1.6×10!"  1.3×10!"  9.7×10!"  Au bottom (F-series)  1.8×10!"  1.0×10!"  6.1×10!"  Al bottom  2.9×10!"  8.0×10!"  2.0×10!"  -3  Table 3.6: Hole density (cm ). Indirectly from fits.  Hole mobility values extracted from AC fits in chapter 4 agree with values presented here and can be considered correct. On the other hand, hole densities extracted from AC measurements are an order of magnitude larger than those derived from DC fits. The values presented here should be considered underestimated. The hole mobility derived from AC measurements depends on the hole density from those same measurements. If the mobility is correct, then the hole density should also be correct. This means that the DC fits overestimated the size of !! by an order of magnitude.  28  3.5 Conclusions J-V scans of the diodes demonstrated that the Au bottom devices outperformed the Al bottom devices in current rectification by two orders of magnitude. The F-series Au bottom devices had a much poorer performance than expected. This emphasized the extreme sensitivity of the Schottky junction to contamination when being formed. A high quality vacuum is essential for forming a high performance Schottky junction. A proper DC model of the P3HT/Al Schottky diode must take the space charge limited current into account. The model used adequately explained the observed behaviour. Better fits could be achieved by applying a finer control on the algorithms used. A more sophisticated model, that for example incorporates a field-dependent mobility, may also yield better fits. The bulk hole mobility in P3HT was found to be in the range of 2×10!!   cm2 V -­‐1 s -­‐1 to 6×10!!   cm2 V -­‐1 s -­‐1 . This is in agreement with reported values as well as results from AC measurements presented in this work. The hole density in P3HT was found to range from 10!"   cm-­‐3 to 10!"   cm-­‐3 . While this agrees with reported values, it is likely an order of magnitude too small. This may be due to an overestimation of the bulk resistance. The hole density and bulk mobility values measured in this chapter were compared to values extracted from AC measurements. These AC measurements and results are completely independent of the DC methods described in this chapter, and are presented next in chapter 4.  29  4 Small Signal AC Measurements In this chapter, small signal measurements are used to determine the carrier density, and thus the trap density, in the P3HT film. The bulk mobility is also determined from these measurements. The conventional method for determining a semiconductor’s carrier profile is not applicable to P3HT and a modified technique is used.  4.1 Description of the C-V Technique The capacitance-voltage (C-V) technique is a well-established method of determining the carrier profile, which has been used with Schottky diodes, as well as pn diodes, MOS capacitors, and MOSFETs [30]. The method is based on measuring the variation of the junction’s depletion capacitance with bias. The differential capacitance at a given bias is measured by applying a small amplitude AC signal over a DC bias. The C-V technique is dependent on several assumptions: that the depletion approximation applies, that the carrier density is invariant over a short segment of the depletion width, and that the AC signal is small enough for the differential capacitance to be approximately linear. As long as these assumptions hold, it has been shown that the differential capacitance at a given bias ! can be modeled as a parallel plate capacitor !=  !!! ! !  (4.1)  where ! is the differential capacitance, ! is the cross-sectional area of the device, ! is the relative dielectric constant of P3HT, and ! is the width of the depletion region when the junction is biased at !  [31] . This allows a direct measurement of the depletion depth since the spread of the depletion region in a Schottky junction is limited on one side by the metal. It has also been shown that the carrier density at a given ! can be related to the capacitance ! ! =  −2 ! !"!! !! !" ! !!  (4.2)  30  where ! ! is the carrier density in the semiconductor at a depth of !from the metallurgical junction, and ! is the elementary charge [30]. In the case of a spatially uniform carrier density, a plot of ! !! against ! is linear, and the depletion width can be described as !=  2!!! !" !!" − ! − !" !  (4.3)  where !!" is the built in potential of the junction (also called the diffusion potential) [19]. In this case, the capacitance can be related to the bias by !" 2 !!" − ! − ! 1 = !! !"!! !  (4.4)  [19].  4.2 General Considerations of the CV Technique The spatial resolution of the carrier density extracted by the C-V technique is characterized by the Debye length !! =  !!! !" !! !(!)  (4.5)  . One of the simplifying assumptions made under the depletion approximation is a sharp boundary between the depletion and bulk regions of the semiconductor. In reality, carriers diffuse over a distance characterized by !! . This means that carrier density variations over distances less than !! are effectively meaningless. The depletion approximation is satisfied if the step size in a C-V scan is no less than two to three times !! [30]. The minimum probing depth occurs at zero applied bias. The maximum depth is limited by the semiconductor’s breakdown electric field strength [30]. The differential capacitance must behave approximately linearly if the parallel plate capacitor model is to be valid. In order to guarantee that this is the case, the AC amplitude must be kept small relative to the DC bias [31]. Typical values range from 10 mV to 20 mV [30]. 31  4.3 Specific Considerations with P3HT The usual procedure is to apply an AC signal of constant frequency while sweeping the applied DC bias. The measured capacitance will be almost identical to the depletion capacitance so long as the leakage current and the series resistance are quite small. A typical measurement frequency is 1 MHz [30]. The high mobilities in materials such as crystalline Si and GaAs means that free carriers provided by dopant atoms will easily respond an excitation at such a frequency. Using a single probing frequency with the devices in this work is problematic. Several transport mechanisms have been proposed for polymeric semiconductors [15] [32] [16]. They have in common a dependence on trap states that have an associated time constant, which depends on their energetic position. This would imply a frequency dependent differential capacitance, which has in fact been observed [33] [34]. If a lump element circuit with frequency-independent components can adequately describe the AC behaviour, then the C-V technique can still be used to extract the carrier profile. Such a model exists [33] [34] [35] and is depicted in Figure 4.1  Rj  Rb  Cj  Cb  Figure 4.1: A small signal model of an organic Schottky diode.  The model elements correspond to physical features of the diode. !! and !! are the junction resistance and capacitance, which account for the leakage current and the depletion capacitance. !! and !! represent transport through the bulk and charge stored in the bulk. Thus the depletion capacitance can be indirectly acquired by fitting this model to the impedance spectrum of the diode. By fitting spectra at different biases, the dependence of the depletion capacitance on DC bias, and consequently the carrier profile, can be observed. !! should be linearly dependent on the depletion width, or equivalently, inversely dependent on !! . 32  4.4 Measurement Description A Solartron 1260A impedance/gain-phase analyzer by Solartron Analytical [36] was used to measure the impedance spectra. Measurements were carried out with a 10 mV amplitude signal. The frequency was swept from 1 Hz to 100 kHz with 50 steps per decade. Spectra were measured at DC biases ranging from 2 V to -5 V in 200 mV increments. Each measurement point was averaged over a period of up to 10 s before being recorded by the impedance analyzer. Measurements were completed outside of the glovebox, as described in chapter 2.  4.5 Data Preprocessing The impedance analyzer generated a separate data file for each DC bias setting. All data pre-processing was done with MATLAB and applied independently to each recorded spectrum. The parasitic effects of the measurement setup were measured and subtracted from the data as a first pre-processing step. This had a minimal effect since the parasitic elements were negligible over the frequency band of interest. Curiously, a phase discontinuity was observed in all measurements. The discontinuity always occurred at 66.07 kHz and ranged in size from approximately 1° to 20°, increasing as the biased moved from away from 0 V. This was assumed to be due to an instrument fault and was subtracted from the data. The final pre-processing step was to apply some noise reduction filters. First, a median filter was used to eliminate outlying data points; next a running average filter was used to smooth the noise. The real and imaginary parts of the impedance were filtered separately.  33  8  10  measured preprocessed  |Z| (1)  6  10  4  10  2  10 0 10  1  2  10  3  10 10 frequency (Hz)  4  10  5  10  50 measured preprocessed e (!)  0  ï50  ï100 0 10  1  2  10  3  10 10 frequency (Hz)  4  10  5  10  Figure 4.2: Example of AC pre-processing.  4.6 Data Processing The applicability of the small signal model depicted in Figure 4.1 was confirmed by examining the measured data. Figure 4.3 shows two representative impedance spectra for a device biased at 0.6 V (left) and -2 V (right). A first order circuit could account for the reverse bias response, but a second order circuit is needed to explain the forward bias response. Nonliear least-squares curve fitting utilities in the MATLAB Optimization Toolbox were used to fit the small signal model. This was not entirely straightforward since the tools only work with real valued functions. To get around this, various real-valued functions derived from the impedance, such as the magnitude and phase, were fitted. The best results were found when fitting the sum of the resistance and reactance. To clarify, the impedance of the small signal model is 34  ! = ! + !" =  !! !! + 1 + !"!! !! 1 + !"!! !!  (4.6)  And the chosen objective function was ! =!+!  5  (4.7)  8  10  10  |Z| (1)  |Z| (1)  6  4  10  10  4  10 3  2  10 0 10  10 0 10  5  10 f (Hz)  5  10 f (Hz)  0  0  ï20  e(Z) (!)  e(Z) (!)  ï20  ï40  ï40 ï60 ï80  ï60 0 10  5  10 f (Hz)  ï100 0 10  5  10 f (Hz)  Figure 4.3: Impedance spectra of a P3HT/Al Schottky diode (D93R1Z) biased at 0.6 V (left) and -2 V (right). Data presented has been preprocessed.  Picking a good starting point for fitting routine increases the chances of convergence on a proper solution. With this in mind, an iterative approach was taken to selecting the initial guess values of the parameters. The impedance spectra can be fairly well described by a first order parallel RC circuit. By assuming that the junction 35  rather than the bulk dominates the diode’s response, the impedance can be roughly described as !! =  !! 1 + !"!! !!  (4.8)  The reactance of equation (4.8) reaches a minimum of − !! 2 at the corner frequency. This allowed easy extraction of initial guess values for !! and !! . Obtaining the initial guess values of !! and !! was less obvious. The initial values of !! were observed to satisfy equation (4.4). This allowed an initial approximation of the carrier density and the depletion width to be determined. These results along with the measured P3HT film thickness and a guessed bulk mobility of 10!! cm2V-1s-1 [5] led to an initial guess of !! . There was unfortunately no clear way of extracting an initial guess for !! from the impedance data. It was calculated in two steps. First, a rough guess was calculated by assuming a !! !! corner frequency of 10! Hz, a value determined from data inspection. Next a refined initial guess was found by a least squares fit to the full small signal model using the initial guesses of all parameters and leaving only !! free. Once initial values of the parameters were obtained, a least squares fit was performed with all parameters left free. It was assumed that the initial guesses were not too far from the solution point. The fit was repeated 100 times with different initial conditions in order to increase the chances of finding the global minimum. The initial point was perturbed by up to 50% and a sliding window was applied to the bounding region.  36  4.7 Fit Results 6  x 10  20  data fit  R+X (1)  15 10 5 0 ï5 0 10  1  10  2  3  10 10 frequency (Hz)  8  10  4  10  5  10  0 ï20  10  e (!)  |Z| (1)  6  4  10  ï40 ï60 ï80  2  10 0 10  5  10  ï100 0 10  frequency (Hz)  5  10 frequency (Hz)  Figure 4.4: Fit result for diode F93R2Y biases at -0.8 V. Top: objective function, bottom: Bode plots.  Figure 4.4 shows a typical fit result. As can be seen, the fit to the chosen objective function was successful. However, the Bode representation of the data might bring into question the validity of the model since the fit resembles the response of a first order circuit. The issue is clarified by examining the fit results of individual spectra over the entire bias range, as shown in Figure 4.5. !! , !! , and !! exhibit a bias dependence, whereas !! is scattered and far too large.  37  ï7  10  7  10  6  10  ï8  Resistance (1)  Capacitance (F)  10  ï9  10  5  10  4  10  ï10  10  3  10  ï11  10  ï5  2  ï4  ï3  ï2  ï1  bias (V)  0  1  2  10 ï5  ï4  ï3  ï2  ï1  0  1  2  bias (V)  Figure 4.5: Impedance fit results for diode F93R2Y. Junction (Ÿ) and bulk (+) values are shown.  The solver failed to converge to a reasonable solution for !! in most data sets; especially those recorded when a reverse bias was applied. The conclusion drawn is that !! is too small to affect the impedance over the recorded frequencies. The diode impedance would have to be measured at frequencies above 100 kHz if !! were to be properly determined. A lack of information about !! does not affect the characterization of the diode since !! is required for determining the carrier profile and !! for determining the mobility. As can be seen in Figure 4.5, !! decreases with increasing reverse bias. This is typical of a spatially uniform carrier density, which is also localized in energy. However, !! starts to increase at biases below -3.4 V. This behaviour was observed in all the measured devices when bias was lowered to this region. This does not appear to be an artefact caused by the solver. The capacitance extracted directly from the impedance spectra by assuming a first order circuit response is a reasonable approximation of !! , especially in reverse bias. As can be seen in Figure 4.6, the increase in capacitance at 3.4 V is also seen directly in the data. This is not likely due to the semiconductor being fully depleted since the capacitance should be constant for biases beyond full depletion. 38  With amorphous semiconductors, ! !! is not necessarily a linear function of reverse bias. The capacitance may also increase with reverse bias if the density of states varies rapidly near the mobility edge [20]. It is possible that the increase in capacitance is due to localized states at the metal interface or inside the semiconductor. These states may be aligned with the Fermi level when the applied bias is around -3.4 V, thus accounting for the sudden increase in capacitance. If this interpretation is correct, the capacitance should once again decrease past a certain reverse bias and continue along the previous trend. A number of diodes were damaged during DC measurements when large reverse biases were applied. The reverse bias was limited to -5 V during the AC measurements for this reason. ï10  8  x 10  10 kHz 1 kHz 100 Hz  7  B/t (F)  6  5  4  3  2 ï5  ï4.5  ï4  ï3.5  ï3  ï2.5 bias (V)  ï2  ï1.5  ï1  ï0.5  0  Figure 4.6: Depletion capacitance from the susceptance for diode F92R3Y.  Impedance spectra with variations in temperature as well as bias would be needed to map the energetic structure of the diode [35] [20]. The results in Figure 4.5 are typical and !!!! was observed to be linear with reverse bias between 0 V and the bias at which !! reached a minimum. Without a better 39  understanding of the energetic structure, it can be assumed that most transport occurs near the mobility edge and that the C-V technique yields a reasonable idea of the carrier density. Through linear regression and equation (4.4), the hole density and built in voltage was extracted from !! . Three Al bottom devices from three different samples were measured. Four F-series Au bottom diodes from two separate samples were also measured. The results are summarized in Table 4.1 and Table 4.2. Device Type  Min  Median  Max  Au bottom (F-series)  6×10!"  3×10!"  6×10!"  Al bottom  2×10!"  5×10!"  1×10!" -3  Table 4.1: Hole density from AC measurements (cm ).  Device Type  Min  Median  Max  Au bottom (F-series)  0.3  0.5  0.7  Al bottom  0.1  0.1  0.6  Table 4.2: Built in voltage from AC measurements (V).  The hole density values are within the range of reported values, though on the high end [16] [35] [9] [17]. These values are likely overestimated since they depend on the relative permittivity of P3HT. Many sources use a value of 3 [16] [9], which was also adopted in this work. However, this value is derived from optical measurements [37] and it has been shown that the relative permittivity actually increases with decreasing frequency. The value at 100 kHz is about 4 and increases to 9 at 100 Hz [29]. The larger hole density in the Au bottom electrode devices may be attributed to variations in the spin coating process. These devices were all made in a single production run and the Al bottom devices represent two production runs. The sample size was too small to understand all variations that might result in manufacturing. A less ordered film with more traps might lead to a larger hole density measure. The built in voltage for the Al bottom electrode diodes was found to be smaller than for the Au bottom electrode diodes. This may be attributed to an increased number of interface states. For the first group of devices, the Al electrodes were exposed to air 40  before being transferred to the glovebox. For the second group, the Al electrodes were evaporated onto the P3HT, forming a more intimate contact. The Al in the first group may have formed a thin oxide layer while exposed to air, which resulted in an increased number of interface states. The band bending and thus the built in voltage can be decreased if these states have a net positive charge [20]. The bulk mobility was extracted from the fitted values of !! and the extracted values for the hole density and built in voltage. Physically, !! is described by !! =  !−! !"#$  (4.9)  where t is the total P3HT thickness. By substituting equation (4.3), !! can be related to the applied bias !! =  ! 2!!! − ! −! !"#$ !! !! !! ! ! !"  (4.10)  thus making it possible to find the mobility. The extracted bulk mobility values ranged from 4×10!!   cm2 V -­‐1 s -­‐1 to 6×10!!   cm2 V -­‐1 s -­‐1 . There was no noticeable difference between the two types of diodes, except for one of the samples with an Al bottom electrode, which had a mobility five times larger. These values are within the range of reported bulk mobility [16] [17] [4] [7].  4.8 Conclusions A small signal model of the Schottky diodes was fitted to impedance spectroscopy data. Meaningful results for the bulk capacitance were difficult to obtain. Extending measurements to at least 1 MHz may correct this. The hole density and mobility in the P3HT films were extracted from the fitted model parameters. The results agreed with reported values. The carrier profile suggests trap states are uniformly distributed in space but not energy. Extending measurements to higher reverse biases, as well as performing measurements at varied temperatures would clarify this. The discussion is continued in the next chapter, where the results in Chapters 3 and 4 are used to examine the feasibility and challenges associated with practical use of the P3HT diodes demonstrated in this work.  41  5 Practical Diode Uses The discussion up to this point has mainly been concerned with material properties rather than engineering applications. This section briefly considers some practical applications.  5.1 Peak Rectifier One of the envisioned applications of conjugated polymers is in the construction of passive wireless devices such as RFIDs [2]. These devices have no on-board energy storage and must scavenge their power from radio waves. The incoming AC power needs to be rectified for digital components. One of the simplest rectifiers is the halfwave peak rectifier circuit, depicted in Figure 5.1. The circuit is simply a first order low pass filter in series with a diode [38]. As long as the filter’s corner frequency is below the input frequency, the circuit will produce a fairly stable DC voltage. It also has applications in signal processing as a peak detector, and in particular as an amplitude demodulator [38].  Figure 5.1: The half-wave peak rectifier.  A rectifier based on the P3HT diodes would need to operate at frequencies greater than 135 kHz to be used in an RFID [2]. Voltage rectification is a large signal operation so the AC measurements performed in chapter 4 cannot be applied here. To measure the frequency performance, the rectifier was built by wiring a conventional discreet resistor and capacitor to a P3HT diode. A 1  MΩ resistor and a 1  µμF capacitor were used for !! and !! , resulting in a filter corner frequency of 0.16 Hz, well below the measurement range of 10 Hz to 500 kHz. The 10 Vpp sinusoidal input was provided by an arbitrary waveform generator (Agilent 33220A) [39]. The output voltage was measured with an oscilloscope (Tektronix TDS2000C) [40]. The rectifier 42  was kept in the same aluminum shield enclosure used during other measurements in this work. Results for four Au bottom devices (F-series) from two different samples are shown in Figure 5.2. Results for three Al bottom devices from two different samples are shown in Figure 5.3. The output voltages were normalized to their low frequency values so that the 3 dB drop could be easily identified. 5  Normalized Output Voltage (dB)  F91 1X F91 1Y F93 3Y F93 3Z  0  ï3  ï5  ï10 1 10  2  10  3  10 Frequency (Hz)  4  10  5  10  Figure 5.2: Frequency response of rectifier with Au bottom (F-series) devices.  43  5  Normalized Output Voltage (dB)  E91 3X E92 1X E92 3X  0  ï3  ï5  ï10 1 10  2  10  3  10 Frequency (Hz)  4  10  5  10  Figure 5.3:Frequency response of rectifier with Al bottom devices.  The sample with the higher performing Al bottom devices (E92) received a heat treatment prior to this test. It was placed on a hotplate set to 120 °C for ten minutes. This treatment is reported to improve the mobility of P3HT films [24]. The order of magnitude difference in performance is encouraging. It indicates that poorly performing devices built at an ambient pressure can be easily improved. Even so, the Al bottom devices had corner frequencies about four times lower than the Au bottom devices. The best Au bottom device had a maximum frequency of 40 kHz, falling well short of passive RFID requirements. The output voltage was about half the input voltage amplitude. This is suggestive of a diode resistance of around 1  MΩ. This requires a P3HT resistivity of around 5×10!   Ωcm, which corresponds to the range of hole density and mobility values measured in chapters 3 and 4. Reducing the diode resistance could boost the output voltage and corner frequency of the rectifier. Most simply, the semiconductor could be made thinner. P3HT films 20 nm thick can be produced through spin coating [25]. This represents a reduction of film thickness, and thus resistance, of 90% from those made in this work. Though, evaporating metal onto such a thin polymer film without significantly damaging 44  it may be difficult. The depletion width would also extend through the entire thickness of such a thin film. Such a dramatic reduction in film size may not be needed. The corner frequency of a similar rectifier has been reportedly increased by an order of magnitude by halving the thickness of the semiconductor [8]. The semiconductor could be made more crystalline by using a different solvent [25]. This has been reported to boost the field effect mobility by an order of magnitude [25]. It should also have a positive effect on the bulk mobility, but it has not been reported. The carrier density could also be increased by controllably doping the semiconductor [41].  5.2 Small Signal Use The nonlinearity of a diode can also be exploited to build frequency multipliers. Resistive diode multipliers use forward biased diodes as nonlinear resistors, and reactive diode multipliers use reverse biased diodes as voltage controlled capacitors (varactors) [42]. The small signal frequency performance of a Schottky diode can be characterized by the forward bias cutoff frequency !!! =  1 2!!! !!  (5.1)  where !! and !! are the diode’s resistance and capacitance at a bias of 0.1 V to the flat band condition [19]. The utility of this metric lies in being able to compare devices with different built in voltages. !!! was computed using the junction resistance and capacitance, !! and !! , extracted from AC measurements in chapter 4. Results for Au bottom (F-series) and Al bottom devices are listed in Table 5.1. Bottom electrode  Min  Median  Max  Al  28  63  120  Au (F-Series)  65  420  2800  Table 5.1: forward bias cutoff frequency (Hz).  The small signal performance of Au bottom (F-series) devices is better than that of Al bottom devices, but still far too low for practical use. Increasing this performance is not straight forward, especially for the Al bottom devices, since it would involve increasing the quality of the Schottky junctions, by ensuring a clean interface. 45  5.3 Conclusions The applicability of P3HT based Schottky diodes to voltage rectification and frequency mixing was considered in this chapter. Al bottom devices showed consistently inferior performance. The best performing devices had poor enough performance to make them impractical for the suggested applications. Heat treatment appears promising for boosting performance. Better performance may also be achieved by using thinner and better ordered P3HT layers. Improvements of almost two orders of magnitude may be possible. A corner frequency of above 1 MHz has been reported for a similar P3HT/Al diode based rectifier [8].  46  6 Conclusion In this thesis, two types of P3HT/Al Schottky junction diodes were fabricated, characterized, and compared. Carrier density and mobility were derived from DC as well as impedance measurements. Comparisons made included Schottky junction quality, frequency performance in a rectifier, and small signal frequency performance.  6.1 General Conclusions Both varieties of diodes comprised stacks of the same materials: gold, poly(3hexylthiophene), and aluminum. Their difference was in the formation of the Schottky junction. The first type, designated Au bottom, had its Schottky junction formed in a high vacuum by evaporating aluminum onto a film of P3HT. The second type, designated Al bottom, had its Schottky junction formed by depositing a film of P3HT onto aluminum at a standard atmospheric pressure. Though all steps involving P3HT were carried out in an argon-filled glovebox, the aluminum electrodes used in the Al bottom diodes had been previously exposed to air. The final yield of usable devices was 35% for Au bottom, and 22% for Al bottom. Due to an equipment fault, one set of Au bottom devices was manufactured at an elevated pressure. Difficulties were also encountered during the manufacturing of the bottom electrodes by lift-off. The cause was likely a photolithographic mask of inadequate quality. Current-voltage scans of the diodes showed that Au bottom devices had a current rectification ratio of 2×10! , 100 times greater than for Al bottom devices. The DC behaviour was properly explained by a model that took into account the space charge limited current. Model fitting was done by least-squares and required an iterative approach. The bulk mobility was extracted from fits and found to range from 2× 10!!   cm2 V -­‐1 s -­‐1 to 6×10!!   cm2 V -­‐1 s -­‐1 and is in line with reported values. The hole density was also calculated from the fit parameters to be between from 10!"   cm-­‐3 and 10!"   cm-­‐3 . These values also agree with reported numbers. Impedance spectroscopy was used to model the diodes’ small signal behaviour with a 2nd order series/parallel RC circuit. The presence of a bulk capacitance could be seen in the recorded spectra, but exact values proved difficult to determine. Fitted values of the junction capacitance suggest a spatially uniform hole density of between 5×10!"   cm-­‐3 and 3×10!"   cm-­‐3 . While these values are in line with those reported in 47  literature, they are an order of magnitude larger than those derived from DC measurements. These density values were used to extract mobility from the fitted bulk resistance. The range of these mobilities was from 4×10!!   cm2 V -­‐1 s -­‐1 to 6×10!!   cm2 V -­‐1 s -­‐1 . In this case the agreement between AC and DC derived mobilities suggests that the values are correct. This implies that the AC hole densities are correct and DC hole densities are underestimated by an order of magnitude and means that the series resistance in the DC model was consistently overestimated. Fitted junction capacitance and resistance values suggest poor small signal performance for both device types. The maximum frequency is below 100 Hz for Al bottom devices, and an order of magnitude larger for Au bottom devices. The frequency performance of a half-wave peak rectifier built using the P3HT diodes and conventional components was measured. The maximum operating frequency with Au bottom diodes was 40 kHz, and 10 kHz with Al bottom diodes. Heating the Al bottom diodes for 10 minutes at 120°C seemed to increase their maximum frequency from 1 kHz to 10 kHz. The Au bottom diodes consistently outperformed the Al bottom devices, demonstrating the large sensitivity of the Schottky junction to contamination. This suggests that practical diodes based on P3HT must be entirely manufactured in an inert atmosphere, and that the rectifying junction may need to always be made in a high vacuum. This is problematic if printing methods are to be employed since vacuum deposited metals will have to be replaced with solution processable materials.  6.2 Future Work The small signal measurements should be extended in the following ways. Measurement frequencies should be extended to at least 1 MHz to improve model fitting. The reverse bias range should be doubled and temperature variation should be introduced in order to probe the energetic structure of the Schottky junction. The reverse breakdown voltage should be measured. This should be done after all other electrical measurements are complete as it may irreversibly affect the diode.  48  Performance gains may be achieved by adjusting the fabrication. Bottom electrodes should be entirely fabricated inside the glovebox and never exposed to air. A solvent with a higher boiling point than chloroform should be used. Completed devices should be heat-treated. Major changes are needed if flexible and printable Schottky diodes are to be explored further. First the glass substrate should be substituted with a flexible substrate. Most significantly, the vacuum deposited metal electrodes need to be replaced with materials which can be solution processed. Finally, devices should be entirely printed.  49  Bibliography [1]	
  The	
  Royal	
  Swedish	
  Academy	
  of	
  Sciences.	
  (2012,	
  April)	
  The	
  2000	
  Nobel	
  Prize	
  in	
  Physics	
   -­‐	
  Scientific	
  Background.	
  [Online].	
   http://www.nobelprize.org/nobel_prizes/physics/laureates/2000/advanced.html	
   [2]	
  V	
  Subramanian,	
  "Radio	
  Frequency	
  Identity	
  Tags,"	
  in	
  Organic	
  Field	
  Effect	
  Transistors,	
  Z	
   Bao	
  and	
  J	
  Locklin,	
  Eds.	
  Boca	
  Raton,	
  FL,	
  USA:	
  CRC	
  Press,	
  2007,	
  pp.	
  489-­‐504.	
   [3]	
  D.-­‐H	
  Kim	
  et	
  al.,	
  "Stretchable	
  and	
  Foldable	
  Silicon	
  Integrated	
  Circuits,"	
  Science,	
  vol.	
  320,	
   no.	
  5875,	
  pp.	
  507-­‐511,	
  Apr	
  2008.	
   [4]	
  Michael	
  G	
  Kane	
  et	
  al.,	
  "100-­‐MHz	
  CMOS	
  circuits	
  directly	
  fabricated	
  on	
  plastic	
  using	
   sequential	
  laterally	
  solidified	
  silicon,"	
  Journal	
  Of	
  The	
  Society	
  For	
  Information	
  Display,	
   vol.	
  15,	
  no.	
  7,	
  pp.	
  471-­‐478,	
  Jan	
  2007.	
   [5]	
  A	
  Takshi,	
  "Organic	
  metal-­‐semiconductor	
  field-­‐effect	
  transistor	
  (OMESFET),"	
   Department	
  of	
  Electrical	
  and	
  Computer	
  Engineering,	
  University	
  of	
  British	
  Columbia,	
   Vancouver,	
  PhD	
  dissertation	
  2007.	
   [6]	
  L	
  Burgi,	
  TJ	
  Richards,	
  RH	
  Friend,	
  and	
  H	
  Sirringhaus,	
  "Close	
  look	
  at	
  charge	
  carrier	
   injection	
  in	
  polymer	
  field-­‐effect	
  transistors,"	
  Journal	
  of	
  Applied	
  Physics,	
  vol.	
  94,	
  no.	
  9,	
   pp.	
  6129-­‐6137,	
  Jan	
  2003.	
   [7]	
  Michele	
  Giulianini,	
  Eric	
  R	
  Waclawik,	
  John	
  M	
  Bell,	
  and	
  Nunzio	
  Motta,	
  "Current-­‐voltage	
   characteristics	
  of	
  poly(3-­‐hexylthiophene)	
  diodes	
  at	
  room	
  temperature,"	
  Applied	
   Physics	
  Letters,	
  vol.	
  94,	
  no.	
  8,	
  p.	
  083302,	
  Jan	
  2009.	
   [8]	
  Chan-­‐mo	
  Kang,	
  Seohee	
  Kim,	
  Yongtaek	
  Hong,	
  and	
  Changhee	
  Lee,	
  "Frequency	
  analysis	
   on	
  poly(3-­‐hexylthiopene)	
  rectifier	
  using	
  impedance	
  spectroscopy,"	
  Thin	
  Solid	
  Films,	
   vol.	
  518,	
  no.	
  2,	
  pp.	
  889-­‐892,	
  Jan	
  2009.	
   [9]	
  VR	
  Nikitenko,	
  H	
  Heil,	
  and	
  H	
  von	
  Seggern,	
  "Space-­‐charge	
  limited	
  current	
  in	
  regioregular	
   poly-­‐3-­‐hexyl-­‐thiophene,"	
  Journal	
  of	
  Applied	
  Physics,	
  vol.	
  94,	
  no.	
  4,	
  pp.	
  2480-­‐2485,	
  Jan	
   2003.	
   [10]	
  HL	
  Gomes	
  and	
  DM	
  Taylor,	
  "Schottky	
  barrier	
  diodes	
  from	
  semiconducting	
  polymers,"	
   Iee	
  Proceedings-­‐Circuits	
  Devices	
  And	
  Systems,	
  vol.	
  144,	
  no.	
  2,	
  pp.	
  117-­‐122,	
  Jan	
  1997.	
   [11]	
  M	
  MacLachlan,	
  "Course	
  notes	
  for	
  Chemistry	
  527,"	
  Chemistry,	
  The	
  University	
  of	
  British	
   Columbia,	
  Vancouver,	
  September	
  2006.	
   [12]	
  The	
  Royal	
  Swedish	
  Academy	
  of	
  Sciences.	
  (2012,	
  April)	
  The	
  Nobel	
  Prize	
  in	
  Chemistry	
   2000	
  -­‐	
  Scientific	
  Background.	
  [Online].	
   http://www.nobelprize.org/nobel_prizes/chemistry/laureates/2000/advanced.html	
   [13]	
  A	
  Salleo,	
  "Charge	
  transport	
  in	
  polymeric	
  transistors,"	
  Materials	
  Today,	
  vol.	
  10,	
  no.	
  3,	
   pp.	
  38-­‐45,	
  Mar	
  2007.	
   [14]	
  Arash	
  Takshi,	
  Milad	
  Mohammadi,	
  and	
  John	
  D	
  Madden,	
  "Study	
  the	
  effect	
  of	
  distribution	
   of	
  density	
  of	
  states	
  on	
  the	
  depletion	
  width	
  of	
  organic	
  Schottky	
  contacts,"	
  Solid	
  State	
   Electronics,	
  vol.	
  52,	
  no.	
  11,	
  pp.	
  1717-­‐1721,	
  Nov	
  2008.	
   [15]	
  A	
  Salleo,	
  T.	
  W	
  Chen,	
  A.	
  R	
  Völkel,	
  and	
  R.	
  A	
  Street,	
  "Intrinsic	
  hole	
  mobility	
  and	
  trapping	
  in	
   a	
  regioregular	
  poly(thiophene),"	
  Physical	
  Review	
  B,	
  vol.	
  70,	
  no.	
  11,	
  p.	
  115311,	
  Sep	
   2004.	
   [16]	
  Z	
  Chiguvare	
  and	
  V	
  Dyakonov,	
  "Trap-­‐limited	
  hole	
  mobility	
  in	
  semiconducting	
  poly(3-­‐ hexylthiophene),"	
  Physical	
  Review	
  B,	
  vol.	
  70,	
  no.	
  23,	
  p.	
  235207,	
  Jan	
  2004.	
   50  [17]	
  J	
  Li,	
  A	
  Nardes,	
  Z	
  Liang,	
  and	
  S	
  Shaheen,	
  "Simultaneous	
  measurement	
  of	
  carrier	
  density	
   and	
  mobility	
  of	
  organic	
  semiconductors	
  using	
  capacitance	
  techniques,"	
  Organic	
   Electronics,	
  Jan	
  2011.	
   [18]	
  AJ	
  Cascio	
  et	
  al.,	
  "Investigation	
  of	
  a	
  polythiophene	
  interface	
  using	
  photoemission	
   spectroscopy	
  in	
  combination	
  with	
  electrospray	
  thin-­‐film	
  deposition,"	
  Applied	
  Physics	
   Letters,	
  vol.	
  88,	
  no.	
  6,	
  p.	
  062104,	
  Jan	
  2006.	
   [19]	
  S.	
  M.	
  Sze,	
  Physics	
  of	
  Semiconductor	
  Devices,	
  2nd	
  ed.	
  New	
  York:	
  John	
  Wiley	
  &	
  Sons,	
  1981.	
   [20]	
  E.	
  H.	
  Rhoderick	
  and	
  R.	
  H.	
  Williams,	
  Metal-­‐Semiconductor	
  Contacts,	
  2nd	
  ed.	
  New	
  York:	
   Oxford	
  University	
  Press,	
  1988.	
   [21]	
  AMPEL	
  Nanofabrication	
  Facility.	
  (2012,	
  April)	
  About	
  the	
  facility.	
  [Online].	
   http://nanofab.ubc.ca/content/about-­‐facility	
   [22]	
  The	
  UCSB	
  Nanofabrication	
  Facility.	
  (2012,	
  April)	
  Lithography.	
  [Online].	
   http://www.nanotech.ucsb.edu/index.php?option=com_content&view=article&id=45 &Itemid=27	
   [23]	
  HS	
  Lee	
  and	
  JB	
  Yoon,	
  "A	
  simple	
  and	
  effective	
  lift-­‐off	
  with	
  positive	
  photoresist,"	
  Journal	
   of	
  Micromechanics	
  and	
  Microengineering,	
  vol.	
  15,	
  no.	
  11,	
  pp.	
  2136-­‐2140,	
  Jan	
  2005.	
   [24]	
  Shinuk	
  Cho	
  et	
  al.,	
  "Thermal	
  annealing-­‐induced	
  enhancement	
  of	
  the	
  field-­‐effect	
  mobility	
   of	
  regioregular	
  poly(3-­‐hexylthiophene)	
  films,"	
  Journal	
  of	
  Applied	
  Physics,	
  vol.	
  100,	
  no.	
   11,	
  p.	
  114503,	
  Jan	
  2006.	
   [25]	
  JF	
  Chang	
  et	
  al.,	
  "Enhanced	
  mobility	
  of	
  poly(3-­‐hexylthiophene)	
  transistors	
  by	
  spin-­‐ coating	
  from	
  high-­‐boiling-­‐point	
  solvents,"	
  Chemistry	
  Of	
  Materials,	
  vol.	
  16,	
  pp.	
  4772-­‐ 4776,	
  Jan	
  2004.	
   [26]	
  Keithley	
  Instruments	
  Inc.	
  (2012,	
  April)	
  6430.	
  [Online].	
   http://www.keithley.com/products/dcac/sensitive/highresistance/?mn=6430	
   [27]	
  A	
  Assadi,	
  C	
  Svensson,	
  M	
  Willander,	
  and	
  O	
  Inganäs,	
  "Properties	
  of	
  the	
  planar	
  poly	
  (3-­‐ octylthiophene)/aluminum	
  Schottky	
  barrier	
  diode,"	
  Journal	
  of	
  Applied	
  Physics,	
  Jan	
   1992.	
   [28]	
  S	
  Tagmouti	
  et	
  al.,	
  "Electrical	
  characteristics	
  of	
  W/P3MT/Pt	
  diodes,"	
  Thin	
  Solid	
  Films,	
   vol.	
  379,	
  no.	
  1-­‐2,	
  pp.	
  272-­‐278,	
  2000.	
   [29]	
  J	
  Obrzut	
  and	
  K	
  Page,	
  "Electrical	
  conductivity	
  and	
  relaxation	
  in	
  poly	
  (3-­‐ hexylthiophene),"	
  Physical	
  Review	
  B,	
  Jan	
  2009.	
   [30]	
  Dieter	
  K.	
  Schroder,	
  Semiconductor	
  Material	
  and	
  Device	
  Characterization.	
  New	
  York:	
   John	
  Wiley	
  &	
  Sons,	
  Inc.,	
  1990.	
   [31]	
  Peter	
  Blood	
  and	
  John	
  W.	
  Orton,	
  The	
  electrical	
  characterization	
  of	
  semiconductors:	
   majority	
  carriers	
  and	
  electron	
  states,	
  N.	
  H.	
  March,	
  Ed.	
  London,	
  UK:	
  Academic	
  Press,	
   1992,	
  vol.	
  14.	
   [32]	
  M	
  Vissenberg	
  and	
  M	
  Matters,	
  "Theory	
  of	
  the	
  field-­‐effect	
  mobility	
  in	
  amorphous	
  organic	
   transistors,"	
  Physical	
  Review	
  B,	
  vol.	
  57,	
  no.	
  20,	
  pp.	
  12964-­‐12967,	
  1998.	
   [33]	
  D.	
  M	
  Taylor,	
  "Space	
  charges	
  and	
  traps	
  in	
  polymer	
  electronics,"	
  Ieee	
  Transactions	
  On	
   Dielectrics	
  And	
  Electrical	
  Insulation,	
  vol.	
  13,	
  no.	
  5,	
  pp.	
  1063-­‐1073,	
  Jan	
  2006.	
   [34]	
  S	
  Karg,	
  M	
  Meier,	
  and	
  W	
  Riess,	
  "Light-­‐emitting	
  diodes	
  based	
  on	
  poly-­‐p-­‐phenylene-­‐ vinylene.1.	
  Charge-­‐carrier	
  injection	
  and	
  transport,"	
  Journal	
  of	
  Applied	
  Physics,	
  vol.	
  82,	
   no.	
  4,	
  pp.	
  1951-­‐1960,	
  Jan	
  1997.	
   51  [35]	
  P	
  Stallinga,	
  HL	
  Gomes,	
  M	
  Murgia,	
  and	
  K	
  Müllen,	
  "Interface	
  state	
  mapping	
  in	
  a	
  Schottky	
   barrier	
  of	
  the	
  organic	
  semiconductor	
  terrylene,"	
  Organic	
  Electronics,	
  vol.	
  3,	
  no.	
  1,	
  pp.	
   43-­‐51,	
  2002.	
   [36]	
  Solartron	
  Analytical.	
  (2012,	
  April)	
  Model	
  1260A	
  Impedance-­‐Gain/Phase	
  Analyzer.	
   [Online].	
  http://www.solartronanalytical.com/Pages/1260AFRAPage.htm	
   [37]	
  RHM	
  van	
  de	
  Leur,	
  B	
  de	
  Ruiter,	
  and	
  J	
  Breen,	
  "Dielectrical	
  and	
  dynamic	
  mechanical	
   properties	
  of	
  three	
  poly	
  (3-­‐N-­‐Alkaylthiophene)	
  s,"	
  Synthetic	
  Metals,	
  vol.	
  57,	
  no.	
  2-­‐3,	
  pp.	
   4956-­‐4961,	
  1993.	
   [38]	
  A	
  Sedra	
  and	
  K	
  C	
  Smith,	
  Microelectronic	
  Circuits,	
  4th	
  ed.	
  New	
  York,	
  NY,	
  USA:	
  Oxford	
   University	
  Press	
  Inc.,	
  1998.	
   [39]	
  Agilent	
  Technologies.	
  (2012,	
  April)	
  33220A	
  Function/Arbitrary	
  Waveform	
  Generator.	
   [Online].	
  http://www.home.agilent.com/agilent/product.jspx?nid=-­‐ 536902257.536883183.00&cc=CA&lc=eng	
   [40]	
  Tektronix,	
  Inc.	
  (2012,	
  April)	
  TDS	
  2000.	
  [Online].	
   http://www.tek.com/oscilloscope/tds2000	
   [41]	
  Y	
  Chen,	
  I	
  Shih,	
  and	
  S	
  Xiao,	
  "Effects	
  of	
  FeCl3	
  doping	
  on	
  polymer-­‐based	
  thin	
  film	
   transistors,"	
  Journal	
  of	
  Applied	
  Physics,	
  vol.	
  96,	
  no.	
  2,	
  pp.	
  454-­‐458,	
  July	
  2004.	
   [42]	
  D.	
  M.	
  Pozar,	
  Microwave	
  Engineering,	
  3rd	
  ed.	
  Hoboken,	
  NJ,	
  USA:	
  Wiley,	
  2005.	
    52  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

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"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0103463/manifest

Comment

Related Items