Search

The synthesis of graphene directly on Ge and on Ge deposited on Si provides a scalable route toward integrating graphene onto conventional semiconductors. This presentation will first survey the growth modes of graphene on Ge(001), Ge(011), Ge(111), Ge(112), Ge(001)-6°, and Ge(001)-9° via chemical vapor deposition (CVD), reporting on the effect of Ge surface orientation on graphene island formation and shape, strain in large-area graphene films, and the nanofaceting of the Ge below graphene.[1] We will then focus on the anisotropic growth of semiconducting graphene nanoribbons on Ge(001) and Ge(001)-9° and of nominally single crystal graphene on Ge(110). On Ge(001), we have discovered how to drive graphene crystal growth with a large shape anisotropy through control of kinetics.[2-6] This discovery enables the direct synthesis of narrow, armchair, semiconducting nanoribbons. The ribbons are self-orienting, self-defining, and have smooth edges. The ribbons exhibit excellent transport properties (e.g., high on-state conductance and on/off ratio at room temperature) and provide a promising pathway towards the direct integration of high-performance nanocarbon electronics onto conventional semiconductor wafer platforms. On Ge(001), nominally single crystal graphene has been reported in limited cases; however, conflicting studies have evidenced polycrystallinity. Here, the factors affecting the mosaicity of graphene on Ge(110) will be elucidated using low energy electron diffraction and microscopy data.[7] [1] R. M. Jacobberger, D. E. Savage, X. Zheng, P. Sookchoo, R. R. Delgado, M. G. Lagally, M. S. Arnold, SUBMITTED (2022). [2] R. M. Jacobberger, M. S. Arnold, et al., Direct Oriented Growth of Armchair Graphene Nanoribbons on Germanium, NATURE COMMUNICATIONS, 6, 8006 (2015). [3] B. Kiraly, M. S. Arnold, M. C. Hersam, N. P. Guisinger et al., Sub-5 nm, Globally Aligned Graphene Nanoribbons on Ge (001), APPLIED PHYSICS LETTERS, 108, 213101 (2016). [4] A. J. Way, R. M. Jacobberger, M. S. Arnold, Seed-Initiated Anisotropic Growth of Unidirectional Armchair Graphene Nanoribbon Arrays on Germanium, NANO LETTERS, 18, 898 (2018). [5] V. Saraswat, Y. Yamamoto, H.J. Kim, R.M. Jacobberger, K.R. Jinkins, A.J. Way, N.P. Guisinger, M.S. Arnold Synthesis of armchair graphene nanoribbons on germanium-on-silicon, THE JOURNAL OF PHYSICAL CHEMISTRY C 123 (30), 18445-18454 (2019). [6] A. J. Way, R. M. Jacobberger, N. P. Guisinger, V. Saraswat, X. Zheng, A. Suresh, J. H. Dwyer, P. Gopalan, Michael S. Arnold, SUBMITTED (2022). [7] R. M. Jacobberger, Z. Miao, T. Yu, V. Saraswat, M. G. Lagally, M. S. Altman, M. S. Arnold, SUBMITTED (2022).

This talk will present on strong light-matter interaction in optical microcavities incorporating dense semiconducting carbon nanotubes, characterizing hybridization of the bright and dark exciton states of nanotubes, mediated by a cavity photon plus exciton-phonon coupling. This talk will also present on the relaxation of optically excited, bright polariton modes into sub-radiant dark polaritons.

Selective deposition of semiconducting carbon nanotubes (s-CNTs) into densely packed, aligned arrays of individualized s-CNTs is necessary to realize their potential in semiconductor electronics. We report the combination of chemical contrast patterns, topography, and pre-alignment of s-CNTs via shear to achieve selective-area deposition of aligned arrays of CNTs. Alternate stripes of surfaces favorable and unfavorable to s-CNT adsorption were patterned with widths varying from 2000 nm down to 100 nm. When the chemical and topographical contrast patterns are reduced to less than the width of individual nanotubes (≤ 500 nm), confinement effects become dominant enabling the selective-area deposition of much more tightly aligned CNTs (~7 degrees). At a trench width of 100 nm, we demonstrate the lowest standard deviation in alignment degree of 7.6 ± 0.3° at a deposition shear rate of 4,600 s-1, while maintaining an individualized s-CNT density > 30 CNTs µm-1. Chemical contrast alone enables selective area deposition but chemical contrast in addition to topography enables more effective selective area deposition and stronger confinement effects, with the advantage of removal of nanotubes deposited in spurious areas via selective lift-off of the topographic features. These findings provide a methodology that is inherently scalable, and a means to deposit spatially selective, aligned s-CNT arrays for next-generation semiconducting devices.

Floating evaporative self-assembly (FESA) is a solution-based technique for fabricating highly-aligned semiconducting carbon nanotube films.1,2 Field-effect transistors (FETs) based on carbon nanotubes aligned via FESA have exhibited current densities exceeding that of GaAs and Si, demonstrating the promise of this technique.3 However, the global uniformity of FESA films is poor, as the films are composed of stripes of aligned and randomly oriented carbon nanotubes. Additionally, this technique is inherently difficult to scale. In the work presented here, we demonstrate a scalable technique to overcome the shortcomings of FESA and to achieve globally uniform films of aligned carbon nanotubes at desirable packing densities for FETs. In our technique, carbon nanotubes are continuously aligned across target substrates by lifting the substrates through a flowing nanotube ink/water interface. The ink/water interface collects and confines the nanotubes, which is vital to achieving a high degree of alignment. Polarized Raman spectroscopy mapping is used to quantify the degree of alignment in the resulting aligned nanotube films. The full width at half maximum (FWHM) of the alignment distribution within the films is 6.0° measured across a 1 mm2 area. We control the aligned carbon nanotube packing density from 0.2 to 59 µm-1 by varying the ink concentration from 0.1 to 100 µg mL-1. Using optimized conditions, we demonstrate the scalability of this technique by aligning carbon nanotubes across a 100 mm wide wafer. Charge transport measurements are performed using FETs with the aligned carbon nanotube films as the active channels, and the performance is compared to FETs based on FESA films. 1. Jinkins, K., R. et al. Nanotube Alignment Mechanism in Floating Evaporative Self-Assembly. Langmuir 33, 13407–13414 (2017). 2. Joo, Y., Brady, G. J., Arnold, M. S. & Gopalan, P. Dose-Controlled, Floating Evaporative Self-assembly and Alignment of Semiconducting Carbon Nanotubes from Organic Solvents. Langmuir 30, 3460–3466 (2014). 3. Brady, G. J. et al. Quasi-ballistic carbon nanotube array transistors with current density exceeding Si and GaAs. Sci. Adv. 2, (2016).

Floating evaporative self-assembly (FESA) is a solution-based technique for fabricating highly-aligned semiconducting carbon nanotube films.1,2 Field-effect transistors (FETs) based on carbon nanotubes aligned via FESA have exhibited current densities exceeding that of GaAs and Si, demonstrating the promise of this technique.3 However, the global uniformity of FESA films is poor, as the films are composed of stripes of aligned and randomly oriented carbon nanotubes. Additionally, this technique is inherently difficult to scale. In the work presented here, we demonstrate a scalable technique to overcome the shortcomings of FESA and to achieve globally uniform films of aligned carbon nanotubes at desirable packing densities for FETs. In our technique, carbon nanotubes are continuously aligned across target substrates by lifting the substrates through a flowing nanotube ink/water interface. The ink/water interface collects and confines the nanotubes, which is vital to achieving a high degree of alignment. Polarized Raman spectroscopy mapping is used to quantify the degree of alignment in the resulting aligned nanotube films. The full width at half maximum (FWHM) of the alignment distribution within the films is 6.0° measured across a 1 mm2 area. We control the aligned carbon nanotube packing density from 0.2 to 59 µm-1 by varying the ink concentration from 0.1 to 100 µg mL-1. Using optimized conditions, we demonstrate the scalability of this technique by aligning carbon nanotubes across a 100 mm wide wafer. Charge transport measurements are performed using FETs with the aligned carbon nanotube films as the active channels, and the performance is compared to FETs based on FESA films. 1. Jinkins, K., R. et al. Nanotube Alignment Mechanism in Floating Evaporative Self-Assembly. Langmuir 33, 13407–13414 (2017). 2. Joo, Y., Brady, G. J., Arnold, M. S. & Gopalan, P. Dose-Controlled, Floating Evaporative Self-assembly and Alignment of Semiconducting Carbon Nanotubes from Organic Solvents. Langmuir 30, 3460–3466 (2014). 3. Brady, G. J. et al. Quasi-ballistic carbon nanotube array transistors with current density exceeding Si and GaAs. Sci. Adv. 2, (2016).

Recent discovery of graphene and its unique properties has attracted a great deal of interest in implementing this material in a variety of new devices, targeting applications in ultra-fast electronics, quantum information, carbon-free energy conversion, optoelectronics, and bio-integrated technologies [1,2]. However, graphene lacks a controlled non-zero bandgap, which prevents its large-scale application in electronic and energy conversion devices. To overcome this limitation, we exploit direct integration of III-V semiconductors on graphene through van der Waals epitaxy. Known for their efficient light emission and high charge carrier mobilities, III-V semiconductors are at the core of numerous technologies including high-efficiency solar cells, lasers, light emitting diodes, and ultra-fast transistors, to name a few [3]. Due to lack of dangling bonds on graphene’s surface, in epitaxial growth of III-V semiconductors on graphene the crystals follow the order of graphene lattice through van der Waals forces. Initially, two nucleation and growth processes were suggested for such van der Waals epitaxy: (1) nucleation starts at defect sites on graphene [4], and (2) nucleation starts on specific sites on graphene lattice, which are the favorable adsorption sites for III/V adatoms [5]. The latter growth mechanism limits the epitaxial growth of the III-V crystals in four different relative orientations (with different strain values) [5]. Recently, density functional theory calculations followed by homoepitaxial growth of InP, GaP, and GaAs through several layers of graphene suggested another determining factor in van der Waals epitaxial growth: long-range interaction between the III-V substrate under graphene and the III-V structures grown on graphene [6]. In this talk, we demonstrate the low-pressure metal-organic vapor phase deposition (MOCVD) of InAs, GaAs, InP, and GaP islands on single layer graphene sheets grown on copper foils and transferred on SiO2/Si (100), as well as graphene sheets directly grown on Ge (100), (110), and (111) substrates. We show that regardless of the nature of the substrate under graphene, III-V crystals grow mostly on graphene defect sites and their morphology can be somewhat controlled through controlling the growth parameters. We then show the effect of long-range forces between Ge and III-V structures on growth morphology, which results in significant coalescence and growth of tens of micron large islands on graphene/Ge as opposed to parasitic crystals grown on graphene/SiO2. In what follows, we provide a brief summary of our observations in both systems. All III-V semiconductors grown on graphene/SiO2/Si showed very similar growth morphology and crystal structure. Plan-view and cross-sectional scanning electron microscopy (SEM), micro-Raman spectroscopy, powder and high-resolution X-ray diffraction (XRD), along with cross-sectional transmission electron microscopy (TEM) and nanoscale cathodoluminescence analysis revealed formation of 1D, 2D, and 3D structures in zinc-blende, wurtzite, and polytypic (mixed zinc-blende/wurtzite) phases. We have recently demonstrated that polytypic InP crystals form a type II homojunction with potential applications in optoelectronic devices [7]. XRD analysis of the growth time evolution of texture, as well as growth rate and crystal size evolution analysis of the crystals using atomic force microscopy (AFM) suggested a growth mechanism similar to selected area MOCVD growth [7], which is widely used to grow nanostructures and devices. This enabled us to have some degree of control over the growth morphology (single crystals, polycrystals, nanowires) through controlling the growth parameters such as growth temperature, rate and time. Similar 1D, 2D, and 3D structure were grown on graphene/Ge (100), (110), (111). However, the size and crystalline quality of the islands have significantly improved for growth on Ge substrates. Moreover, two-step growth (nucleation at low temperature (450 °C) and growth at higher temperature (600 °C)) lead to coalescence of the crystals and formation of tens of microns large islands with smooth surface (roughness ~ 1nm from AFM data) on graphene/Ge, while no significant improvement in growth morphology was observed for growth on graphene/SiO2. These observations suggest a strong sensitivity of the morphology of the deposited films on the nature of the substrate below the graphene monolayers. This provides insight to control the properties of such hybrid systems and will pave the path to engineering a new class of electronic and optoelectronic devices by combining the advantages of semiconductors and graphene. References [1] K. S. Novoselov et al., Nature 490, 192 (2012). [2] A.K. Geim, IV Grigorieva, Nature 99, 419 (2013) [3] O. Moutanabbir, U. Gosele, Annu. Rev. Mater. Res. 40, 469 (2010). [4] Y. G. Hong et al., ACS Nano 5, 7576 (2011). [5] A. M. Munshi et al., NanoLetters 7, 713 (2013). [6] Y. Kim et al., Nature 544, 340 (2017). [7] S. Mukherjee et al., Adv. Funct. Mater. 28, 1705592 (2018). Figure 1

We have previously demonstrated efficient electron transfer from photoexcited small diameter (< 1 nm) semiconducting carbon nanotubes to C60 fullerene electron acceptors in planar, bilayer heterojunction diode devices, with internal quantum efficiency of 90%. However, the out-of-plane exciton diffusion length in planar nanotube thin films is poor (~ 5 nm), limiting the optical density of nanotube films that can be implemented in these devices, which in turn has limited external quantum efficiency to 50% at the nanotube bandgap and to << 50% off-resonance. Polymer solar cells also face a similar limitation (poor exciton diffusion length) and overcome this problem by blending the polymer and acceptor together into a bulk heterojunction. In this talk, our recent research aimed at addressing this shortcoming will be presented.

Direct synthesis of graphene nanoribbons on dielectric or semiconducting substrates offers a scalable route for integration of graphene-based devices into a conventional silicon-based technology.(1) So-far, wafer-scale synthesis of armchair graphene nanoribbons has been demonstrated on Ge (001) using chemical vapor deposition.(2) However, direct synthesis of graphene or graphene nanoribbons has not yet been achieved on CMOS compatible Si(001) due to formation of stable SiC at temperatures > 1000 K, which are needed to achieve synthesis of nanoribbons with smooth, armchair edges. A promising and realistic way to overcome this challenge is to synthesize graphene nanoribbons on CMOS compatible Ge(001)/Si(001) substrates. So far the synthesis of monolayer graphene on Ge(001)/Si(001) substrates has been demonstrated by CVD.(3, 4) In this study, we synthesized graphene nanoribbons on 3 µm epitaxial Ge(001)/Si(001) substrates, by ambient pressure CVD using methane as the carbon precursor. We show that growth kinetics of graphene nanoribbons on Ge(001)/Si(001) are comparable to that on Ge(001). By tuning the methane flow and growth time we are able to synthesize graphene nanoribbons ranging from 100 nm to 1 µm in length with high aspect ratios, whilst avoiding Si diffusion from the bulk. For instance, by restricting methane to < 6000 ppm and growth time to < 3h, graphene nanoribbons with widths < 10 nm can be synthesized with aspect ratios as high as 70. Such nanoribbons grown on Ge(001) have been shown to exhibit technologically relevant band-gaps as well as exceptional charge transport properties.(5) Furthermore, we also investigated the possible role of threading dislocations in Ge epilayer on nucleation or growth of nanoribbons and show that nanoribbons readily grow over the threading dislocations. Lastly, we studied the evolution of surface roughness with the nucleation density of nanoribbons. Our study provides valuable insight into the mechanism of graphene nanoribbon growth on Ge(001)/Si(001) substrates. From this information, it is expected that unidirectional graphene nanoribbons with rational placement and control over width poly-dispersity can be synthesized on Ge(001)/Si(001) platform akin to Ge(001), which has been demonstrated recently using seed-mediated growth.(6) This provides a scalable way for wafer scale integration of graphene nanoribbon arrays on Si and provides motivation for further research into this direction. References A. Khan et al., Direct CVD Growth of Graphene on Technologically Important Dielectric and Semiconducting Substrates. Advanced Science 5, (2018). R. M. Jacobberger et al., Direct oriented growth of armchair graphene nanoribbons on germanium. Nature Communications 6, (2015). I. Pasternak et al., Graphene growth on Ge(100)/Si(100) substrates by CVD method. Scientific Reports 6, (2016). M. Lukosius et al., Metal-Free CVD Graphene Synthesis on 200 mm Ge/Si(001) Substrates. Acs Applied Materials & Interfaces 8, 33786-33793 (2016). R. M. Jacobberger, M. S. Arnold, High-Performance Charge Transport in Semiconducting Armchair Graphene Nanoribbons Grown Directly on Germanium. Acs Nano 11, 8924-8929 (2017). A. J. Way, R. M. Jacobberger, M. S. Arnold, Seed-Initiated Anisotropic Growth of Unidirectional Armchair Graphene Nanoribbon Arrays on Germanium. Nano Letters 18, 898-906 (2018). Figure 1

We have previously demonstrated efficient electron transfer from photoexcited small diameter (< 1 nm) semiconducting carbon nanotubes to C60 fullerene electron acceptors in planar, bilayer heterojunction diode devices, with internal quantum efficiency of 90%. However, the out-of-plane exciton diffusion length in planar nanotube thin films is poor (~ 5 nm), limiting the optical density of nanotube films that can be implemented in these devices, which in turn has limited external quantum efficiency to 50% at the nanotube bandgap and to << 50% off-resonance. Polymer solar cells also face a similar limitation (poor exciton diffusion length) and overcome this problem by blending the polymer and acceptor together into a bulk heterojunction. These bulk heterojunctions are typically formed by co-casting the polymer with a solubilized fullerene derivative, phenyl-C61 (or 71) -butyric acid methyl ester (PBCM) to create the blend. However, this strategy has not yet worked well to create nanotube-based bulk heterojunction devices that are more efficient that bilayered devices. One problem is that the electron affinity of PCBM is shallower than the electron affinity of C60, which detrimentally impacts the nanotube to acceptor charge transfer efficiency. To address this shortcoming, new electron acceptors are needed. We have surveyed several new non-fullerene electron acceptors, including both small molecule and polymers, as potential C60 and PCBM replacements. In this talk, the results of this survey will be presented. Moreover, the use of the most promising new non-fullerene acceptors in bulk heterojunctions with semiconducting nanotubes will be discussed.

The wide scale utility of semiconducting carbon nanotubes (CNTs) depends on its electronic purity as well as development of methods to process them on the wafer-scale into the ideally organized and aligned arrays for integrated device structures. Now we can routinely achieve > 99.99% semiconducting CNTs with a suitable diameter distribution using conjugating polymers as sorting agents in organic solvents. However, the spatial localization using a scalable method still remains a challenge. This is important as now there are many methods available to get aligned CNTs, however an ideal array with the optimal spacing between the tubes remains elusive. An ideal CNT array microstructure for high-performance FET channels is predicted to consist of a sub-monolayer of electronic-type controlled semiconducting CNTs that are laterally aligned and evenly spaced with a packing density of 100-200 CNTs µm-1. In order to achieve this goal, we present our studies on understanding the multitude of factors from surface energy of the wafer, role of the polymer wrapper on the CNT, chemistry of the substrate, surface topography, chemical pattern, and the type of solvent used on the adsorption or desorption of the CNTs on substrates. These studies so far show that the static water contact angle of the substrate alone cannot fully explain the adhesion of CNTs to certain substrates. The surface roughness also plays a critical role. Through these studies we hope to lay down a rational basis for designing chemical patterns that can control the spatial localization of CNTs.

Graphene nanoribbons with sub-10 nm widths and smooth armchair edges possess exceptional electrical and thermal properties making them a material of interest for next-generation electronics. However, the difficulty of synthesizing ribbons with the qualities required for device applications has remained a challenge. Ribbons must have sub-10 nm widths, lengths greater than 100 nm, controlled orientation and placement, and a smooth, armchair edge structure. Top-down lithographic patterning of monolayer graphene into ribbons can address the issues of placement, orientation, and length. However, patterning resolution limits ribbons to greater than 10 nm widths and they suffer from poor edge structure, which degrades electrical and thermal properties. Bottom-up techniques such as the unzipping of graphite and nanotubes can result in narrow ribbons with smooth edges but do not provide control over placement, orientation, and edge chirality. Another method is polymerization, which results in a smooth, armchair edge structure and sub-10 nm widths but it lacks control over orientation and placement and the ribbons possess short lengths (~20 nm). Our group has shown the CVD growth of graphene on Ge(001) directly yields long, smooth armchair edged, semiconducting graphene nanoribbons [1-2]. The ribbons evolve from the bottom-up from CH4 due to highly anisotropic growth kinetics. The ribbons possess exceptional charge transport properties, as compared to others in literature, demonstrating an on/off conduction of 2x104 and an on-state conductance of 5 µS [3]. However, this fabrication method causes nanoribbons (with two orientations possible) to stochastically nucleate at random locations and times giving rise to length, width, and bandgap polydispersity. We have recently shown that seed-mediated growth of graphene nanoribbons on Ge(001) enables control over ribbon position and orientation, reduces length and width polydispersity, and enables the fabrication of nanoribbon arrays [4]. However, the diameter of our graphene seeds before growth was unknown, the effect of pitch (ribbon to ribbon distance) on ribbon growth was not studied, and sub-10 nm ribbon arrays were not achieved. Here, we present on the synthesis of sub-10 nm graphene nanoribbons from sub-10 nm seeds, at pitches varying from 50 to 500 nm, for the first time. We use lithography to create graphene seeds that vary from 20 to 55 nm in diameter and then etch the seeds in an Ar/H2 environment before nanoribbon synthesis to create sub-10 nm seeds. We find that the evolution of nanoribbons from the seeds proceeds similarly, regardless of pitch – indicating that growth is dictated by attachment limited processes that do not depend on inter-nanoribbon separation. These findings are used to fabricate tightly pitched arrays of registered, unidirectionally aligned nanoribbons, with sub-10 nm widths and lengths exceeding 200 nm, for the first time. These results show that seed-mediated nanoribbon growth is a viable route for creating dense arrays of nanoribbons for semiconductor electronics. [1] Jacobberger, R. M. et al. Nat. Commun. 2015, 6, 8006. [2] Kiraly, B. et al. Appl. Phys. Lett. 2016, 108, 213101. [3] Jacobberger, R. M. et al. ACS Nano 2017, 11, 8924−8929. [4] Way, A. J. et al. Nano Letters 2018, 18, 898.

In this presentation, I will present on 2 recent advances in nanocarbon electronics from my laboratory: (1) We have pioneered a scalable approach for assembling parallel arrays of ultrahigh purity (Fig. 1) semiconducting nanotubes. This approach has allowed us to create carbon nanotube field effect transistors (FETs) with current density that exceeds Si and GaAs, for the first time, which has been a goal of the nanoelectronics field for 20+ years. (2) We have discovered how to drive graphene crystal growth with a large shape anisotropy through control of kinetics on the surface of Ge(001) single crystal wafers via CH4 chemical vapor deposition. This discovery enables the direct synthesis of narrow, armchair, semiconducting nanoribbons. The ribbons are as narrow as 1.7 nm and are self-orienting, self-defining, and have smooth edges. The ribbons exhibit exceptional transport properties (e.g., high on-state conductance and on/off ratio at room temperature) and provide a promising pathway towards the direct integration of high-performance nanocarbon electronics onto conventional semiconductor wafer platforms. Both aspects of this work have implications towards extending Moore’s Law, creating ultra-low energy logic circuits, developing higher bandwidth RF communication devices, and realizing next-generation FET based sensors. [1] Joo et al. Langmuir (2014); Brady et al. APL (2014); Brady et al. ACS Nano (2014); Brady et al. Science Advances (2016); Jinkins et al. Langmuir (2017); Brady et al. J. Appl. Phys. (2017); Jinkins et al. In Preparation (2018). [2] Jacobberger et al. Nature Comm. (2015); Jacobberger et al. ACS Nano (2017); Way et al. Nano Lett. (2018); Fig. 1 SEM image of aligned semiconducting carbon nanotube (CNT) array FET. Figure 1

Individual semiconducting single-walled carbon nanotubes (s-SWCNTs) exhibit the high charge-carrier mobility, current carrying capacity, and sensitivity needed for applications based on thin-film transistors (TFTs) such as flexible electronics, displays, and sensors. However, assembling s-SWCNTs into aligned arrays to exploit the exceptional electronic properties of this material has been a major roadblock toward the integration of s-SWCNTs into electronic devices. Here, we are able to use confined shear-based alignment to fabricate aligned arrays of s-SWCNTs across arbitrarily large substrates. In this study, we demonstrate that large amounts of shear can be used to align s-SWCNTs from solution by forcing s-SWCNT ink at high volumetric flow rates through 20 - 100 µm tall channels on SiO2 substrates. We use Raman spectroscopy to quantify the ratio of the Raman G band (~1590 cm-1) to the Si peak (~520 cm-1) of our films. This ratio provides a measure of the amount of s-SWCNTs deposited. We study the deposition from three concentrations of ink (~70, 35, and 7 µg mL-1) while varying the volume of ink used at a set shear rate of 3.2*105 s-1. The G/Si ratio behavior is similar for each concentration. Initially, there is a steep increase in the G/Si ratio up to 100 µL of ink, after which, the G/Si ratio increases linearly with increasing volume of ink. Polarized Raman spectroscopy is used to quantify the disorder in the film. We show that for the s-SWCNT ink concentrations used in this study, the concentration does not affect the degree of alignment in the aligned s-SWCNT films and that the shear rate is the main factor affecting the alignment. Assuming the s-SWCNTs are oriented within a Gaussian angular distribution, from polarized Raman spectroscopy, the angular width of distribution was measured to be σ = 30° for a shear rate of 3.2*105 s-1 regardless of ink concentration and volume. This degree of alignment makes this technique non-ideal for fabricating devices with small channel lengths for high-performance applications such as logic or radio-frequency devices. However, for long-channel devices, there must be some amount of disorder in the s-SWCNT films to allow for charge percolation. This means this shear alignment technique is ideal for fabricating films for long-channel devices such as TFTs. Here, we fabricate transistors from films of shear-aligned s-SWCNTs and study the effect of degree of alignment on the transistor performance. Because of the simplicity of this technique, we are able to scale the deposition of aligned s-SWCNTs to arbitrarily large areas as long as the high shear rate is maintained. This makes the technique ideal for large-area scaling in industry applications. As a demonstration, we show the large-area, uniform alignment of s-SWCNTs via confined shear-based alignment across a 4” wafer.

We have discovered how to drive graphene crystal growth with a large shape anisotropy through control of kinetics on the surface of Ge(001) single crystal wafers via CH4 chemical vapor deposition. This discovery enables the direct synthesis of narrow, armchair, semiconducting nanoribbons. The ribbons are self-orienting, self-defining, and have smooth edges. The ribbons exhibit exceptional transport properties (e.g., high on-state conductance and on/off ratio at room temperature) and provide a promising pathway towards the direct integration of high-performance nanocarbon electronics onto conventional semiconductor wafer platforms. This talk will detail the synthesis of these ribbons (including seeded arrays of ribbons), the elucidation of their structure, and the characterization of their promising charge transport properties. M. Jacobberger, B. Kiraly, M. Fortin-Deschenes, P. L. Levesque, K. M. McElhinny, G. J. Brady, Richard Rojas Delgado, S. Singha Roy, A. Mannix, M. G. Lagally, P. G. Evans, P. Desjardins, R. Martel, M. C. Hersam, N. P. Guisinger, M. S. Arnold, Direct Oriented Growth of Armchair Graphene Nanoribbons on Germanium, NATURE COMMUNICATIONS, 6, 8006 (2015). Kiraly, A. J. Mannix, R. M. Jacobberger, B. L. Fisher, M. S. Arnold, M. C. Hersam, N. P. Guisinger, Sub-5 nm, Globally Aligned Graphene Nanoribbons on Ge (001), APPLIED PHYSICS LETTERS, 108, 213101 (2016). M. Jacobberger, M. S. Arnold, High Performance Charge Transport in Semiconducting Armchair Graphene Nanoribbons Grown Directly on Germanium. ACS NANO, 11, 8924 (2017). J. Way, R. M. Jacobberger, M. S. Arnold, Seed-Initiated Anisotropic Growth of Unidirectional Armchair Graphene Nanoribbon Arrays on Germanium, SUBMITTED (2017). S. Arnold, R. M. Jacobberger, Oriented Bottom-Up Growth of Armchair Graphene Nanoribbons on Germanium, U.S. PATENT 9,287,359 (2016). S. Arnold, A. J. Way, R. M. Jacobberger, Seed-Mediated Growth of Patterned Graphene Nanoribbon Arrays, U.S. PATENT 9,761,669 (2017).

Separation of electronically pure, narrow dispersed, pristine semiconducting single walled carbon nanotubes (S-CNT) from a heterogeneous as-synthesized mixture is essential for various semiconducting technologies and biomedical applications. While conjugated polymers are essential for this sorting step, it is highly desirable to remove any organic residues from the resulting devices. We report here the design and synthesis of a mild acid degradable π-conjugated polyimine polymer (PFO-N-BPy) that is structurally analogous to the commonly used and commercially available poly[(9,9-dioctylfluorenyl-2,7-diyl)-alt-co-(6,6’-(2,2’-bipyridine))] (PFO-BPy). An acid cleavable imine link (-HC=N-) was introduced in the PFO-N-BPy backbone to impart degradability, which is absent in PFO-BPy. PFO-N-BPy was synthesized via a metal catalyst free Aza-Wittig reaction in high yields. PFO-N-BPy with a degree of polymerization of just ~10 showed excellent (> 99% electronic purity) selectivity for both large diameter (1.3-1.7 nm) arc-discharge S-CNTs and smaller diameter (0.8-1.2 nm) HiPCO S-CNTs. Overall, selectivity for semiconducting species is similar to that of PFO-BPy but with an advantage of complete depolymerization under mild acidic conditions into recyclable monomers. We further show by UV-Vis, X-ray photoelectron spectroscopy (XPS), and SEM that the PFO-N-BPy wrapped S-CNTs can be aligned into a monolayer array on gate dielectrics using a floating evaporative self-assembly process from which the polymer can be completely removed. Short channel FETs were fabricated from the polymer-stripped aligned S-CNT arrays which further confirmed the semiconducting purity on the order of 99.9% or higher. We will present the recent optimized FET device results using this removable polymer.

Semiconducting single-walled carbon nanotubes (s-SWCNTs) have attracted significant attention as a photoactive component in thin film photovoltaics due to their strong absorptivity and high charge mobility. However, the external quantum efficiency (EQE) of s-SWCNT/C60 heterojunction solar cells is limited by poor exciton diffusion to the heterointerface. Exciton diffusion is believed to be strongly dependent on the length between defects, which quench excitons. Here, we systematically study the influence of defects on s-SWCNT/C60 planar heterojunction photovoltaic devices. We use covalent sidewall functionalization via diazonium chemistry to introduce sp3 defects to the s-SWCNTs at varying concentrations. We generate a defect-adding model to describe the position of defects on the surface of the nanotube, and use it to successfully reproduce the time-resolved exciton decay rate measured with transient absorption spectroscopy. We fabricate s-SWCNT/C60 heterojunction photovoltaic cells and characterize the EQE as a function of defect concentration. We develop a diffusion limited contact quenching Monte Carlo model to reproduce the EQE trends we observe. The model allows us to predict the behavior of theoretical s-SWCNT devices with varying length distribution and defect concentration, toward the limit of long length and zero defects. This work guides the field of nanotube photovoltaics and illuminates the promise and limitations of s-SWCNT solar cell devices.

Calculations have indicated that aligned arrays of semiconducting carbon nanotubes (CNTs) promise to outperform conventional semiconducting materials in short-channel, aggressively scaled field effect transistors (FETs) like those used in semiconductor logic and high frequency amplifier technologies. These calculations have been based on extrapolation of measurements of FETs based on one CNT, in which ballistic transport approaching the quantum conductance limit of 2Go =4e2/h has been achieved. However, constraints in CNT sorting, processing, alignment, and contacts give rise to non-idealities when CNTs are implemented in densely-packed parallel arrays, which has resulted in a conductance per CNT far from 2Go . The consequence has been that it has been very difficult to create high performance CNT array FETs, and CNT array FETs have not outperformed but rather underperformed channel materials such as Si by 6x or more. Here, we report nearly ballistic CNT array FETs at a density of 50 CNTs um-1, created via CNT sorting, wafer-scale alignment and assembly, and treatment. The on-state conductance in the arrays is as high as 0.46 Go per CNT, and the conductance of the arrays reaches 1.7 mS um-1, which is 7x higher than previous state-of-the-art CNT array FETs made by other methods. The saturated on-state current density reaches 900 uA um-1 and is similar to or exceeds that of Si FETs when compared at equivalent gate oxide thickness, off-state current density, and channel length. The on-state current density exceeds that of GaAs FETs, as well. This leap in CNT FET array performance is a significant advance towards the exploitation of CNTs in high-performance semiconductor electronics technologies. *Brady GJ, Way AJ, Safron NS, Evensen HT, Gopalan P, Arnold MS, Quasi-ballistic carbon nanotube array transistors with current density exceeding Si and GaAs, SCIENCE ADVANCES, 2 (9), e1601240 (2016)

We study photophysics and free carrier recombination in (7,5) semiconducting single walled carbon nanotube (s-SWCNT)/C60 fullerene heterojunction diodes over the temperature range 4.5-300 K. We characterize the temperature dependence of the magnitude and spectral peak position of the external quantum efficiency. Appling the Marcus theory description of energy transfer at the interface, we estimate that the reorganization energy and change in Gibbs free energy are relatively equal in magnitude, and predict the high exciton dissociation rate at the s-SWCNT/C60 heterojunction to be relatively resistant to changes in temperature. We show that the current-voltage characteristics in forward bias at all temperatures are best fit to a two-diode model in which traps dictate recombination at the heterointerface. From the diode ideality factor, we calculate the trap depth for holes in (7,5) s-SWCNT films to be 150 meV. We measure the temperature dependence of the open-circuit voltage, short-circuit current density, and fill factor of the devices. Open-circuit voltage strongly depends on temperature, rising to a maximum of 0.82 V at 20 K. Meanwhile, fill factor decreases precipitously below 150 K to a fluence-independent value of 28%. We are able to verify our two-diode model by successfully recreating the temperature and fluence dependence of the open-circuit voltage, and suggest that decreasing fill factor is consistent with a temperature-dependent polaron pair recombination rate at the heterointerface. Finally, we calculate the interfacial energy gap of the s-SWCNT/C60 heterojunction using two methods and find agreement, estimating 0.90 eV by examining the current-voltage characteristics in the dark and 0.86 eV by extrapolating the open-circuit voltage to its 0 K value. These insights provide a greater understanding of the processes that take place at the heterointerface of s-SWCNT/C60 solar cells, guiding further research to improve the efficiency of these devices.

Single walled carbon nanotubes (SWCNTs) exhibit extraordinarily high current carrying capacity and suitable band gaps for logic and thin film field-effect transistors (FETs). Progress in honing the exceptional properties of SWCNTs at the multi-tube level, however, has been dampened by incremental advances in the materials science of electronic type sorting and assembly of SWCNTs. To address these challenges we leveraged the exceptional semiconducting sorting fidelity of polyfluorene polymers and we pioneered an alignment technique known as Floating Evaporative Self-Assembly (FESA).1 Recently, we demonstrated individually placed and uniformly pitched purely semiconducting SWCNT arrays using these techniques, allowing for negligible current lost to scaling when the arrays were implemented as the channel in FETs. The SWCNT array FETs exhibited the highest on-state conductance of 261 μS/μm with a simultaneous on/off conductance ratio exceeding 105 for a multi-tube SWCNT FET.2 Here we present our recent optimization of the contact resistance at the palladium-SWCNT interface, which enables a substantial increase in device on-current density compared to our previous report. For SWCNTs, the quality of the contact-semiconductor interface is highly dependent on the SWCNT packing density, film cleanliness, and surface energy, where any non-idealities may lead to poor palladium film adhesion, or even complete film delamination.3 To address the contact challenge we perform surface treatments on the SWCNT array prior to depositing palladium contacts, which involves rinsing the SWCNT arrays in specific solvents and annealing the films in vacuum. In XPS and optical absorbance spectra, we observe significant mass reduction of the polyfluorene wrapper while maintaining the quality and alignment of the SWCNT array. AFM height maps of the palladium contacts deposited on top of the surface treated SWCNTs demonstrate nearly ideal conformation of the palladium on the SWCNT array. The surface-treated SWCNT array FETs exhibit more than 10-fold improvement in the device on-conductance compared to non-treated samples. [1] Y. Joo, G. J. Brady, M. S. Arnold, and P. Gopalan, Langmuir 30 (12), 3460 (2014). [2] G. J. Brady, Y. Joo, M. Y. Wu, M. J. Shea, P. Gopalan, and M. S. Arnold, ACS Nano 8 (11), 11614 (2014). [3] V. Perebeinos and J. Tersoff, Physical Review Letters 114 (8), 4 (2015).

In order to widely use single-walled carbon nanotubes (SWCNTs) as the semiconducting material in electronic devices, it is essential to separate the desirable semiconducting SWCNTs from as-synthesized electronically heterogeneous mixtures of metallic (m-) and semiconducting (s-) SWCNTs. Among the sorting methods developed, polyfluorene polymers have been studied as semiconducting-selective agents with selectivity for chirality, diameter and electronic type. Developing an understanding for the factors that lead to strong selective interactions between the conjugated polymer and the semiconducting tubes is evolving but far from complete. We choose a commonly used commercially available polyfluorene derivative PFO-BPy as a model polymer to directly probe the effect of chain rigidity on the wrapping/unwrapping process on the s-SWCNTs. Our choice of this model system is based on its commercial availability, outstanding sorting selectivity, availability of functional groups in the backbone to alter rigidity, and recent device results. Recently, we demonstrated the extraordinarily electronic-type selectively of the polyfluorene derivative PFO-BPy by measuring the on/off ratio of FETs, where zero metallic nanotubes were encountered in the measurement of more than 5,000 s-SWCNTs.1-3 Using the similar polymer poly(9,9-dioctylfluorene-2,7-diyl) (PFO), Bindl et al., demonstrated that even after several aggressive rinsing steps to remove free or excess polymer using ultracentrifugation a significant amount of PFO remained bound to the SWCNT. 4 These remaining polymer residues are expected to increase the contact resistance at the metal-nanotube interface of SWCNT FETs, ultimately limiting the conductance of these FETs at sub-100 nm channel lengths. A more complete removal of adsorbed polymer residues from the SWCNT surface is one possible approach towards improving contact resistance and reducing variation in device performance.5 Recently we demonstrated6 an effective yet simple and mild method to remove post-sorting, and the wrapping PFO-BPy copolymer from the surface of s-SWCNTs by switching the backbone rigidity. We used chelation chemistry to complex pentacarbonylrhenium chloride (Re (CO)5Cl) to the bipyridine (BPy) moiety in the wrapping copolymer backbone. We showed the effectiveness of the chelation chemistry on both large and small diameter tubes. The chain-stiffness of the PFO-BPy changes upon complexation, likely providing the driving force to overcome the π-π and electronic interactions7 which adhere the polymer to the nanotube. Optical absorbance, X-ray photoelectron spectroscopy (XPS), Raman spectroscopy, and large-scale molecular dynamics simulations were used to characterize the extent of PFO-BPy removal and configurational changes to the PFO-BPy after Re(CO)5Cl complexation (PFO-BPy:Re).The ability to switch the rigidity of the PFO-BPy in order to unwrap from the nanotubes opens up the door to new design strategies where specific functional groups can be placed on the polymer backbone to chemically alter the rigidity. Here, we present recent results on direct experimental measurements of changes in stiffness of the polymer backbone and its aggregation state upon metal complexation. 1. Brady, G. J.; Joo, Y.; Roy, S. S.; Gopalan, P.; Arnold, M. S. Appl. Phys. Lett. 2014, 104, (8), 083107. 2. Joo, Y.; Brady, G. J.; Arnold, M. S.; Gopalan, P. Langmuir 2014, 30, (12), 3460-3466. 3. Mistry, K. S.; Larsen, B. A.; Blackburn, J. L. ACS nano 2013, 7, (3), 2231-2239. 4. Bindl, D. J.; Shea, M. J.; Arnold, M. S. Chem. Phys. 2013, 413, 29-34. 5. Wang, W. Z.; Li, W. F.; Pan, X. Y.; Li, C. M.; Li, L. J.; Mu, Y. G.; Rogers, J. A.; Chan-Park, M. B. Adv. Funct. Mater. 2011, 21, (9), 1643-1651. 6. Joo, Y.; Brady, G. J.; Shea, M. J.; Oviedo, M. B.; Kanimozhi, C.; Schmitt, S. K.; Wong, B. M.; Arnold, M. S.; Gopalan, P. ACS Nano 2015, 9, (10), 10203-10213. 7. Nish, A.; Hwang, J. Y.; Doig, J.; Nicholas, R. J. Nature nanotechnology 2007, 2, (10), 640-6.

The rational synthesis of graphene nanoribbons that are semiconducting with sub-10 nm width, controlled crystallographic orientation, and well-defined edges on non-metallic substrates has been a significant challenge. The growth of nanoribbons on metal substrates precludes their direct use in semiconducting electronics due to the conductive substrate, and the direct synthesis of nanoribbons in solution is complicated by challenges of their post-synthetic assembly. In this talk, we demonstrate the direct, scalable synthesis of graphene nanoribbons via chemical vapor deposition (CVD) on Ge(001).[1] Low energy electron diffraction (LEED) and scanning tunneling microscopy (STM) show that the ribbons are self-orienting ±2.9° from the Ge[110] directions and are self-defining. The nanoribbons have predominately smooth armchair edges that give rise to electron interference patterns indicative of high quality edges. By tuning the precursor flux, growth time, and growth temperature, the ribbon anisotropy and growth kinetics can be tailored to yield ribbons with controlled width < 10 nm and aspect ratio > 60. Compared to previous low aspect ratio crystals of graphene obtained on Ge, we find that in order to realize high aspect ratio nanoribbons, it is critical to operate in a regime in which the growth rate is especially slow, on the order of 5 nm/h in the width direction. This work is important because unlike continuous two-dimensional graphene, which is semimetallic, one-dimensional graphene nanoribbons can be semiconducting, allowing for the substantial modulation of their conductance and enabling their application in semiconductor logic, optoelectronics, photonics, and sensors. Moreover, the direct synthesis of ultranarrow and smooth graphene nanoribbons on Ge demonstrated here provides a scalable, high throughput pathway for integrating semiconducting graphene directly on conventional large-area semiconductor wafer platforms that are compatible with planar processing. [1] Jacobberger RM, Kiraly B, Fortin-Deschenes M, Levesque PL, McElhinny KM, Brady GJ, Rojas Delgado R, Singha Roy S, Mannix A, Lagally MG, Evans PG, Desjardins P, Martel R, Hersam MC, Guisinger NP, Arnold MS, Direct Oriented Growth of Armchair Graphene Nanoribbons on Germanium, NATURE COMMUNICATIONS, 6, 8006 (2015)

Deformable field-effect transistors (FETs) are expected to facilitate new technologies like stretchable displays, conformal devices, and electronic skins. Thin film percolating networks of electronic-type controlled semiconducting carbon nanotubes are highly intriguing options for the active channel of stretchable FETs due to the excellent mechanical resilience of individual nanotubes, their possibility to accommodate large strain in thin film form via nanotube–nanotube sliding and buckling, and their exceptional charge transport properties. We demonstrate stretchable FETs based on buckled thin films of polyfluorene-wrapped semiconducting single-walled carbon nanotubes as the channel, buckled metal films as electrodes, and unbuckled flexible ion gel films as the dielectric.[1] The buckled thin film morphology is induced by depositing the nanotubes onto a pre-strained polydimethylsiloxane (PDMS) substrate and then releasing it, greatly improving the device stretchability. The FETs are stretchable up to 50% without appreciable degradation in performance before failure of the ion gel film. We furthermore show that by buckling the ion gel, the integrity and performance of the nanotube FETs are extended to nearly 90% elongation, limited by the stretchability of the elastomer substrate.[2] The FETs maintain an on/off ratio of >104 and a field-effect mobility of 5 cm2 V−1 s−1 under elongation and demonstrate invariant performance over 1000 stretching cycles. [1] Xu F, Wu M-Y, Safron NS, Singha Roy S, Jacobberger RM, Bindl DJ, Seo J-H, Chang T-H, Ma Z, Arnold MS, Highly Stretchable Carbon Nanotube Transistors with Ion Gel Gate Dielectrics, NANO LETTERS, 4 (2), pp 682–686 (2014). [2] Wu M-Y, Zhao J, Xu F, Chang T-H, Jacobberger RM, Ma Z, Arnold MS, Highly Stretchable Carbon Nanotube Transistors Enabled by Buckled Ion Gel Gate Dielectrics, APPLIED PHYSICS LETTERS, 107, 053301 (2015).

We have pioneered a scalable approach for depositing aligned arrays of ultrahigh purity semiconducting SWCNTs (prepared using polyfluorene-derivatives) called floating evaporative self-assembly (FESA). In this talk, we will first present on the scaling and physics of FESA. We will then present on high performance field effect transistors (FETs) fabricated from the arrays. FESA is exploited to create FETs with exceptionally high combined on-conductance and on-off ratio of 261 μS/μm and 2x105, respectively, for a channel length of 240 nm. This is 1400x greater on-off ratio than SWCNT FETs fabricated by other methods, at comparable on-conductance per width of ~250 µS/µm, and 30-100x greater on-conductance per width, at comparable on-off ratio of 105-107.

The rational synthesis of graphene nanoribbons that are semiconducting with sub-10 nm width, controlled crystallographic orientation, and well-defined edges on non-metallic substrates has been a significant challenge. The growth of nanoribbons on metal substrates precludes their direct use in semiconducting electronics due to the conducting substrate, and the direct synthesis of nanoribbons in solution is complicated by challenges of their post-synthetic assembly. In this talk, we demonstrate the scalable synthesis of graphene nanoribbons from the bottom-up via chemical vapor deposition (CVD) on Ge(001). Low energy electron diffraction (LEED) and scanning tunneling microscopy (STM) show that the ribbons are self-orienting ±2.9° from the Ge[110] directions and are self-defining. The nanoribbons have predominately smooth armchair edges that give rise to electron interference patterns that are indicative of the high quality of the edges. By tuning the precursor flux, growth time, and growth temperature, the ribbon anisotropy and growth kinetics can be tailored to yield ribbons with controlled width < 10 nm and aspect ratio > 60. Compared to previous reports of the growth of low aspect ratio crystals of graphene on Ge, we find that in order to realize high aspect ratio nanoribbons, it is critical to operate in a regime in which the growth rate is especially slow, on the order of 5 nm/h for growth in the width direction. Scanning tunneling spectroscopy shows that the ribbons have electronic structures that are consistent with semiconductors with bandgaps that are > 500 meV and that vary inversely with width. This work is important because unlike continuous two-dimensional graphene, which is semimetallic, one-dimensional graphene nanoribbons can be semiconducting, allowing for the substantial modulation of their conductance and enabling their application in semiconductor logic, optoelectronics, photonics, and sensors. Moreover, the direct synthesis of ultranarrow and smooth graphene nanoribbons on Ge demonstrated here provides a scalable, high throughput pathway for integrating semiconducting graphene directly on conventional large-area semiconductor wafer platforms that are compatible with planar processing.

Carbon nanotubes are attractive materials for the light absorbing active layers of photovoltaic devices because of their tunable bandgaps, strong optical absorptivity, and ultrafast intra-tube charge and exciton transport. Also of critical importance in carbon nanotube devices are inter-tube charge and exciton transport processes. However, these are generally less well understood. We have developed two-dimensional white light spectroscopy (2D WL) as a novel probe of coupled thin films of semiconducting nanotubes and many other systems in the solar energy sciences. The advantage of exploiting 2D WL over typical 2D electronic spectroscopies is that the spectral bandwidth produced from supercontinuum generation is significantly broader than that accessible from an optical parametric amplifier. Thus, we are able to cover both S22 and S11 ranges with both pump and probe. By studying the evolution (as a function of waiting time) of crosspeaks between the Sii states for different bandgap nanotubes coupled together in thin film networks, we are able to map out Sii-Sjj inter-tube energy transfer pathways and rates and obtain maps of instantaneous photoexcitation distributions. Several surprising results will be reported.

Several schemes have been demonstrated to successfully extract semiconducting (s-) single-walled carbon nanotubes (SWCNTs) from heterogeneous as-grown mixtures containing metallic (m-) SWCNTs. The polyfluorene family of conjugated polymers, specifically poly(9,9-dioctylfluorene-2,7-diyl) (PFO), have been shown to extract s-SWCNTs from mixtures of s- and m-SWCNTs with high selectivity (>99% s-SWCNT) by preferentially wrapping the s-SWCNT components. In addition to electronic type, the selectivity is heavily dependent on the nanotube diameter and chiral angle. To date, there is an inadequate understanding of the polymer-CNT interactions that lead to the high preference for s- over m-SWCNTs in the wrapping process. It is vital to understand these parameters to design separation schemes to reduce the proportion of m-SWCNT impurities to technologically relevant ppm and ppb levels. In this work, we explore the chemistry of non-covalent nanotube-polymer wrapping processes and describe the effects of surface coverage and binding configuration on the separation efficiency. We disperse SWCNTs grown by the HiPCO process, which contain many chiral species of SWCNT with diameters near 1 nm, in toluene solutions of varying PFO concentration. The resulting dispersions are analyzed with optical absorption spectroscopy, excitation-emission photoluminescence spectroscopy, and photoluminescence anisotropy. First, we experimentally quantify the amount of PFO wrapping the SWCNTs by measuring exciton energy transfer from PFO to SWCNTs via photoluminescence spectroscopy. We characterize the dependence of the surface coverage on PFO concentration, finding that the surface coverage increases by a factor of 7 as PFO concentration increases from 0.2 to 2 mg/mL. Second, we demonstrate that the wrapping of nanotubes by PFO is described by a Langmuir adsorption isotherm, in which surface coverage of PFO on the nanotubes follows an S-shaped curve as a function of PFO concentration. The shape of the curve is related to the polymer-nanotube binding energy, which depends on the nanotube electronic type, diameter, chiral angle, solvent, and polymer molecular weight. Third, we determine the binding configuration of the PFO on the nanotube surface as a function of PFO concentration and demonstrate that the PFO becomes more ordered as the surface coverage decreases. In the highly-ordered PFO regime we estimate the wrapping angle of individual PFO strands around nanotubes in solution. Finally, we find that the metal-semiconductor separation efficiency is lowest when the concentration of PFO is high, surface coverage is high, and surface ordering is low. Understanding the PFO wrapping process will be an important step toward attaining high-purity s-SWCNT from this and similar sorting procedures for a range of polymer-CNT systems. These advances will enable further development of technologically relevant high purity s-SWCNTs for next-generation electronic devices.

Single-walled carbon nanotubes (SWCNTs) have many attractive properties for conversion of sunlight into electricity or solar fuels. In solar conversion schemes where SWCNTs are used as absorptive charge donors, it is critical to understand the processes by which initially created excitons are dissociated to produce free charges. In this presentation, I’ll discuss time-resolved spectroscopy studies exploring how photogenerated excitons produce charges in both neat SWCNT films and bilayer heterojunctions designed to dissociate photogenerated SWCNT excitons. In neat SWCNT films, we find charge generation yields of several percent even in the absence of an intentional exciton dissociation interface. We use a thin film of C60 to dissociate SWCNT excitons by interfacial electron transfer from films containing a variety of SWCNT chirality combinations. In these “bilayer” PV active layers, exciton dissociation efficiency varies with the thermodynamic driving force for electron transfer, as well as a number of morphological bilayer properties. Charge mobility and recombination also vary dramatically with bilayer morphology and the diameter distribution of SWCNTs within the film. I will discuss our current understanding of the mechanisms for charge generation and recombination in both neat and bilayer films, and how this understanding helps us envision ways to improve upon SWCNT PV active layers.

Semiconducting single-walled carbon nanotubes (s-SWCNTs) are promising materials for charge transport in field-effect transistor (FET) devices. In order to develop short channel s-SWCNT FETs with high on/off ratios relevant for logic applications, ultra-pure s-SWCNTs of 99.99999% semiconducting purity and higher are needed. There have been many recent studies on the sorting of carbon nanotubes, but most reports have only reached semiconducting purity levels near 99%. In addition, aligned arrays of s-SWCNTs have the potential to improve the mobility and current output of FETs. However, unacceptable semiconducting purity levels (~99%), inter-nanotube charge screening, and inter-nanotube bundling have limited the optimization of such technologies in the short channel regime.1,2 Polyfluorene polymers have been shown to selectively wrap s-SWCNTs on the basis of electronic type, chirality, and diameter. Here we assess the purity and electronic transport properties of aligned arrays of polyfluorene-sorted s-SWCNTs in FET devices. We obtained high purity s-SWCNTs by dispersing nanotube powder in solutions of poly[(9,9-dioctylfluorenyl-2,7-diyl)-alt-co-(6,6’-{2,2’-bipyridine)] (PFO-BPy) in toluene following procedures similar to Mistry et al.3 We fabricated well-aligned s-SWCNT arrays by dispersing droplets of of s-SWCNTs at an air water interface, leading to diffusion and self-assembly of nanotubes into well-aligned “stripes” spanning the entire substrate. Finally, we implemented the aligned films as the channel material in field-effect transistors at varying channel lengths ranging from 9 μm to 300 nm. We characterized the films using Raman spectroscopy, AFM and SEM, and found a high degree of alignment, monolayer thickness, and packing density of 50 tubes/micrometer. Previous works on aligned nanotube FETs have used nanotube densities of 200 tubes/micrometer; however, the proximity of nanotubes in these films cause inter-nanotube screening effects resulting lower conductance and mobility values than expected.1 The polyfluorene polymer plays an important role in isolating s-SWCNTs by preventing direct contact, preventing inter-nanotube screening effects.4 The electronic properties of the aligned film FETs were measured demonstrating high values of on/off ratio, mobility, and conductance. In previous studies, device on/off ratios degraded substantially to less than 103 at channel lengths shorter than 1 micrometer, indicating the presence of metallic nanotubes.5 Here we observe a different trend; as the channel length decreases from 9 μm to 300 nm the on/off ratio only decreases from 3x107 to 2x105. This data suggests that m-SWCNTs are far scarcer in our films than in other short channel s-SWCNT FET devices. We estimated purity by counting the number of nanotubes per channel in direct transport regime devices that have channel lengths of 500 nm and shorter. In fact, of the devices measured, encompassing approximately 1,000 nanotubes, zero exhibited metallic character (defined as having an on/off ratio of at least 103), suggesting a s-SWCNT purity of greater than 99.9%. We achieved consistently high performance devices indicated by high mobility in devices that also have high on/off ratios. The highest mobility was 45 cm2/Vs for a 2 μm channel length device with on/off ratio 2x105. We observed high mobility values ranging from 16-45 cm2/Vs across a range of channel lengths. Previous works were limited by purity, leading to trade offs between high on/off ratio and mobility, especially in the direct transport regime.5 Here we demonstrate a route to achieve both a large on/off ratio and high mobility, which is particularly significant for short channel devices in which nanotubes directly span the channel. References 1 Q. Cao, et al., Nature Nanotechnology 8 (3), 180 (2013). 2 M. Engel, et al., ACS Nano 2 (12), 2445 (2008). 3 K S. Mistry, et. al., ACS Nano 7 (3), 2231 (2013). 4 F. Léonard, Nanotechnology 17 (9), 2381 (2006). 5 Nima Rouhi, et al., ACSNANO 5 (11), 8471 (2011).

not Available.

not Available.

not Available.

not Available.

not Available.