Subscriber access provided by UNIV OSNABRUECK
Article
Nitrogen-Doped Graphene and Twisted Bilayer Graphene via Hyperthermal Ion Implantation with Depth Control Cory D. Cress, Scott W. Schmucker, Adam L Friedman, Pratibha Dev, James C. Culbertson, Joseph W. Lyding, and Jeremy T Robinson ACS Nano, Just Accepted Manuscript • DOI: 10.1021/acsnano.6b00252 • Publication Date (Web): 24 Feb 2016 Downloaded from http://pubs.acs.org on February 26, 2016
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
ACS Nano is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 31
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Nano
Nitrogen-Doped Graphene and Twisted Bilayer Graphene via Hyperthermal Ion Implantation with Depth Control Cory D. Cress,1†* Scott W. Schmucker,2† Adam L. Friedman,3 Pratibha Dev,2,4 James C. Culbertson,1 Joseph W. Lyding,5 and Jeremy T. Robinson1 1Electronics
Science and Technology Division, U.S. Naval Research Laboratory, Washington DC, 20375, USA
2National
Research Council Postdoctoral Research Associate residing at the U.S. Naval Research Laboratory, Washington DC, 20375, USA
3Materials
Science Division, U.S. Naval Research Laboratory, Washington DC, 20375, USA
4Department
5Department
of Physics and Astronomy, Howard University, Washington DC, 20059, USA
of Electrical and Computer Engineering, and the Beckman Institute for Advanced
Science and Technology, University of Illinois at Urbana-Champaign, Urbana, IL 61801, USA
*To whom correspondence should be addressed. E-mail
[email protected] †These authors contributed equally to this work.
1 ACS Paragon Plus Environment
ACS Nano
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ABSTRACT
We investigate hyperthermal ion implantation (HyTII) as a means for substitutionally doping layered materials such as graphene. In particular, this systematic study characterizes the efficacy of substitutional N-doping of graphene using HyTII over an N+ energy range of 25 eV to 100 eV. Scanning tunneling microscopy results establish the incorporation of N substituents into the graphene lattice during HyTII processing. We illustrate the differences in evolution of the characteristic Raman peaks following incremental doses of N+. We use the ratios of the integrated D and D` peaks, I(D)/I(D`) to assess the N+ energy-dependent doping efficacy, which shows a strong correlation with previously reported molecular dynamics (MD) simulation results and a peak doping efficiency regime ranging between approximately 30 eV to 50 eV. We also demonstrate the inherent monolayer depth control of the HyTII process, thereby establishing a unique advantage over other less-specific methods for doping. We achieve this by implementing twisted bilayer graphene (TBG), with one layer of isotopically-enriched 13C and one layer of natural 12C graphene, and modify only the top layer of the TBG sample. By assessing the effects of N-HyTII processing, we uncover dose-dependent shifts in the transfer characteristics consistent with electron doping and we find dose-dependent electronic localization that manifests in low-temperature magnetotransport measurements.
KEYWORDS Graphene, Hyperthermal Ion Implantation (HyTII), Nitrogen Doping, N-graphene, Twisted Bilayer Graphene (TBG), Raman.
2 ACS Paragon Plus Environment
Page 2 of 31
Page 3 of 31
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Nano
The ability to modify material properties at the atomic scale remains at the forefront of nanoscience. In particular, atomic-scale modifications through impurity doping allows device designers to control the electron and hole carrier concentrations, and concomitantly the Fermilevel, giving rise to built-in junctions that limit or permit current flow under the appropriate external perturbations.1 It is common practice to employ ion implantation to precisely dope bulk semiconductors, with spatial and dose control in bulk semiconductor materials. Recently, researchers extended the use of ion implantation to 2-dimensional (2-D) materials, successfully demonstrating the incorporation of N and B substituents in graphene for ions in the hyperthermal energy range (~15-200 eV).2-4 Hyperthermal ion implantation (HyTII) doping offers a number advantages over growth-based doping methods that can achieve substitutional doping by introducing the appropriate impurity precursor during chemical vapor deposition (CVD) growth of graphene.5-7 In particular, HyTII can employ a wide range of dopant species and the ability to spatially control the concentration of one or more dopants within a single graphene monolayer, potentially leading to the formation of intra-layer p-n junctions. HyTII based doping of other 2-D nanomaterials (e.g., h-BN, transition metal dichalcogenides – TMDs, phosphorene) is a natural extension of the technique,8 and may have the depth control needed to dope individual monolayers of van der Waals solids by performing alternating growth - doping steps. In theory, 2-D materials like graphene are ideal candidates for direct substitutional doping via ion implantation, as molecular dynamics (MD) simulation results predict a doping probability exceeding 50% for both N and B ions with the appropriate ion energy (near 45 eV).9 At these hyperthermal energies, the implanted ion possesses enough energy to displace a carbon atom from the lattice, but after transferring the energy, no longer has enough energy to escape the attractive forces stemming from the newly created vacancy, where it ultimately comes to rest. For incident ∗ ions with a kinetic energy that results in a C recoil with kinetic energy equal to the threshold
3 ACS Paragon Plus Environment
ACS Nano
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
∗ energy for displacement (i.e., =
Page 4 of 31
= where and are the masses
of the two atoms involved), only perfect collisions would lead to a displacement, but essentially ∗ 100% would result in a substitutional replacement event. Using =22 eV for graphene,10 =
22.13 eV for 14N and 22.04 eV for 11B in 12C graphene (22.03 eV and 22.15 eV for 13C graphene, ∗ respectively). Ions with energies below cannot directly displace C atoms, but the incident
ionic species are reactive,11 and in the case of N+, may incorporate into the graphene lattice in the form of pyridinic N (1 N substituent adjacent to a vacancy) similar to low-energy plasma ∗ treatments.12 For ions a few eV above , the additional energy has little effect on the probability
that the ion remains in the lattice following collision since the cohesive energy per atom is ~7 eV.13 But, the additional energy increases the ion displacement cross section since the energy transferred ∗ , and thereby increases the in imperfect, or slightly misaligned, collisions may still exceed
overall fraction of incident ions that remain in the lattice. For much higher ion energies, displacement events that yield ion and recoil trajectories that result in a large in-plane momentum component (i.e., motion perpendicular to the incident ion direction) become more probable and more likely to yield defects since the more energetic ion rapidly escapes the attractive forces of the most distant nearest neighbor carbon atom(s). Overall, one predicts a narrow window of hyperthermal ion energies that balances the competing outcomes of ion reactions, direct substitutional doping, and vacancy formation. The efficacy of doping graphene with N or B via HyTII remains largely unexplored. In one report, low-energy ion implantation of graphene was studied with ions ranging in energy from 25 eV to 100 eV, and a positive correlation was found between N concentration and ion energy for nominally equivalent doses.2 However, this study utilized ions from a sputter gun that yields both N+, and N2+ ions with a wide energy range, potentially 5 eV or greater. For optimal substitutional doping using HyTII, the use of mass-selected ions with a tightly controlled energy spread is critical
4 ACS Paragon Plus Environment
Page 5 of 31
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Nano
since the energy window for vacancy-free doping may be quite narrow. In fact, using mass-selected N+, Bangert et al. demonstrated the ability to dope graphene with 25 eV to 35 eV ions based on high resolution scanning transmission electron microscopy.3 More recently, the same system was used to dope epitaxially-grown graphene on SiC substrates with detailed scanning tunneling microscopy (STM),4 and magneto-transport measurements8 substantiating both the impact of substitutional dopants on the local electronic structure and long-range charge transport. In this study, we experimentally investigate the efficacy of N doping of graphene on Cu substrates and on device structures on SiO2/Si as a function of hyperthermal ion energy in the range of 25 eV to 100 eV. Using STM, we confirm the presence of N substituents for optimal ion energies, and compare the results with simulated STM images computed using density functional theory (DFT). Using characteristic peaks in the Raman spectra, namely the D, D`, G, and 2D, we correlate the nature of the modification observed in the Raman spectra over the hyperthermal ion energy range with previously reported MD simulation results, showing a strong correlation between experiment and simulation. We also employ 13C isotopically enriched graphene and combine it with natural 12C graphene to form 13C/12C (and 12C/13C) twisted bilayer graphene (TBG) films that display characteristic Raman peaks unique to each layer. These samples allow us distinguish the extent of doping/disorder independently in both layers, which we ultimately use to demonstrate the inherent mono-layer doping control of N-HyTII processing of 2-D materials. Our study ends with an analysis of the impact that N-doping of graphene has on the field-effect and magnetotransport properties of samples doped using N-HyTII, demonstrating its effectiveness to controllably modify the properties of graphene.
RESULTS AND DISCUSSION
5 ACS Paragon Plus Environment
ACS Nano
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
We begin our analysis with an investigation of the atomic-scale substitution process of nitrogen in highly-ordered pyrolytic graphite (HOPG) and monolayer graphene on the Cu growth substrate. Figure 1(a,b) surveys the spatial distribution of N substituents to the 100 nm scale. Shown are topographic scanning tunneling microscopy (STM) images of 45 eV N+ hyperthermal ion implanted (N-HyTII) HOPG measured with a +0.5 V sample bias. In Figure 1(a) the bright regions correspond with N substituents, which appear to be randomly distributed about the surface across multiple terraces of the HOPG surface. Subjectively, we observe no evidence of substituent clustering on flat terraces, along step edges, or in the vicinity of surface contaminants. In Figure 1(b), we observe the triangular symmetry of the enhanced charge density surrounding the graphitic N substituents, with dopants present within both sub-lattices, based on the 180° rotation of the triangles outlined in red and blue. For N-HyTII processed HOPG, we observe no evidence of graphite sub-lattice selectivity, as has been observed previously for CVD grown N-doped graphene.14 We performed STM on several N-HyTII processed samples on Cu, but the inherent surfaced roughness of polycrystalline Cu foil prevented dopant-resolved wide-area imaging.14,15 Locally, however, we observed evidence of graphitic N doping of the graphene lattice following 45 eV N-HyTII processing, as depicted in Figure 1(c). With a of -1 V applied to the sample, electrons tunnel from filled electronic states within the sample, and the image relates to filled states with energy > − . As previously reported,16 the N substituent appears as a topographic depression in the filled-states STM image, relative to its nearest neighbor C atoms. The N atom rests at the lattice site within the small dashed triangle. The vertices of the small triangle sit on the three nearest neighbor carbons, each showing enhanced charge density. The outer triangle is provided as a guide to the eye and terminates at the higher charge density regions between the C-C bonds on the opposite side of the three adjacent six-membered rings, which are illustrated in the simulated STM image provided in Figure 1(d). The agreement of topographic STM images with
6 ACS Paragon Plus Environment
Page 6 of 31
Page 7 of 31
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Nano
simulation supports our assignment of these triangular surface features as N substituents, a conclusion that will be further verified by Raman spectroscopy. Our DFT calculations provide additional information regarding the changes in the electronic structure of the graphene resulting from graphitic N dopants. In agreement with prior work,16 we observe an enhanced projected density of states (PDOS) from the nearest neighbor carbon atoms and the graphitic nitrogen at the graphene Fermi level consistent with electron doping (see Supp. Info., Figure S1). Graphitic N introduces a nitrogen-2pz derived impurity state that hybridizes with the bulk π* state. In contrast, the 2px- and 2py-derived states of nitrogen and carbon participate in σ-bonding, lying deep in the valence band, and are not considered in Figure 1(d). The triangular symmetry of the STM image simulated for the graphitic-nitrogen dopant closely matches the experimental results, as well as other theoretical results (for example, see Ref. [16]). The main contribution to the tunneling current comes from carbon-atoms located in vicinity of the dopant, lying within two to three lattice constants from the defect center of the N substituent. This analysis, in conjunction with a growing body of STM4,8,14 and scanning transmission electron microscopy (STEM)3,7 studies of N-doped graphene, strongly support the predictions of MD simulations9,17 that hyperthermal ion implantation leads to substitutional N-doping of graphene under the appropriate conditions. Raman spectroscopy further supports the identification of graphitic N substituents in CVD graphene films, and lends insight into the nature of the bonding after HyTII processing. Subsequent to our STM study, the 45 eV graphene on Cu sample was transferred to a SiO2 substrate for Raman mapping. Additional graphene monolayers were grown, HyTII processed with N+ energies ranging from 25 eV to 100 eV, and equivalently transferred to SiO2. In Figure 2(a) we compare the evolution in Raman spectra with dose for graphene samples processed with 25 eV, 45 eV, and 100 eV N+. Transferring graphene from the Cu substrate to the SiO2 substrate likely leads to contamination of
7 ACS Paragon Plus Environment
ACS Nano
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
the graphene surface by carbonaceous residues. However, such residues should not affect our analysis here, which primarily uses peak-intensity ratios that are sensitive to defects in sp2-carbon instead of peak positions that are sensitive to doping. For all three panels, the dose increases from the bottom to the top and each spectrum has approximately the same integrated D peak to integrated 2D peak ratio (I(D)/I(2D)) across the three energies for each dose level. We use the I(D)/I(2D) ratio as an internal reference for dose, instead of the I(D)/I(G) ratio, since the former is more sensitive to increasing disorder,18 and was recently correlated with disorder resulting from Ndoping.19 Subtle differences in the evolution of the Raman spectra reveal differences in bonding following HyTII processing. In particular, we measure a comparatively larger D` peak in all 45 eV NHyTII spectra and a more pronounced D peak in the 25 eV and 100 eV spectra. Additionally, the maximum I(D)/I(G) ratio approaches 3 for the 25 eV and 100 eV spectra, consistent with previous reports,20,21 while the maximum I(D)/I(G) ratio never exceeds 2.2 for the 45 eV N+ samples. Inspection of the peak positions, which informs on doping and/or strain, shows essentially no shift in the G peak frequency (1585 cm-1) with dose for the 25 eV samples, while the 45 eV samples shift from 1585 cm-1 to 1593 cm-1 and the 100 eV sample shifts from 1585 cm-1 to 1590 cm-1. The D, D`, and 2D remain fixed at 1343 cm-1, 1624 cm-1, and 2681 cm-1 for the 25 eV and 100 eV samples, while the 45 eV sample gradually shifts with dose ending at 1348 cm-1, 1624 cm-1 (no shift), and 2685 cm-1 for the same peaks. We estimate an effective electron doping of 7×1012 cm-2 and 4×1012 cm-2 in the 45 eV and 100 eV samples, based on the blue shift of ~8 cm-1 and 5 cm-1, respectively, in the G peak frequency.22 We note that the latter estimate is based on Raman spectra collected from electrostatically doped graphene and therefore is not convoluted with other factors that could introduce shifts (e.g., the larger mass of 14N over 12C, vacancy defects, sample dependent surface contamination, etc.) in our N-doped samples. The I(D)/I(D`) intensity ratio provides a measure of the nature of the defects,23 while the I(D)/I(2D) correlates with the disorder (resulting from defects or nitrogen doping), thus serving as 8 ACS Paragon Plus Environment
Page 8 of 31
Page 9 of 31
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Nano
an internal dose reference. Therefore, in Figure 2(b) we plot the I(D)/I(D`) ratio as a function of the I(D)/I(2D) to observe how the nature of the HyTII-induced graphene modification evolves with dose. The I(D)/I(D`) of ~2 observed in the 45 eV N-HyTII processed sample agrees well with the value measured for N-graphene doped during CVD growth.19 Moreover, the ratio remains nearly constant with increasing I(D)/I(2D) (i.e., dose) indicating that N substitution into the graphene lattice continues even as the N dose increases. The 100 eV N-HyTII processed sample begins with an I(D)/I(D`) of ~5.5, which is consistent with a combination of vacancy type defects ((I(D)/I(D`) of 7)23 and N substituents ((I(D)/I(D`) of 2), yet the presence of pyridinic-nitrogen defects cannot be ruled out.24 However, the magnitude of this ratio reduces with increasing dose indicating that the relative fraction of vacancies to N substituents decreases. Similarly, the 25 eV samples initially begin with I(D)/I(D`) ~7 and therefore may comprise vacancies, N-adatoms with I(D)/I(D`) of ~13, and N substituents, and the relative fraction of N substituents appears to increase with increasing dose. In Figure 2(c) we plot the Raman I(D)/I(D`) of all N-HyTII processed samples as a function of the N+ incident energy and overlay the probability of N-adatom formation, substitutional N substitution, and single vacancy (SV) formation, based on MD simulations adapted from ref. [9]. We note that the simulations from ref. [9] are for free-standing graphene and may not be directly comparable to our experiments with copper-supported graphene due to contributions from the substrate.25 From this plot we observe a remarkable correlation between I(D)/I(D`) ratio and the MD probabilities, where in the energy region between 35 eV and 65 eV the I(D)/I(D`) is approximately two (indicating N-doped graphene), which matches well with the predicted MD simulation results for the highest probability of N-doping. For ion energies outside this window, we observe a larger I(D)/I(D`) consistent with adatom formation at the lower energies (25 eV) and vacancy formation at the higher energies (100 eV). From these results, we conclude that the N+ energies between 30-50 eV are optimal for N-HyTII processing of graphene on Cu substrates since 9 ACS Paragon Plus Environment
ACS Nano
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
within this energy window, N-substitution occurs with high probability and with minimal adatom or vacancy formation. The assembly of heterostructures from 2D materials (e.g., graphene/h-BN/MoS2, etc.) is of considerable interest in order to form van der waals (vdW) solids with novel properties not found in the stand-alone constituents.26,27 Building on this model, substitutional doping of individual layers within a 2D material stack could provide even greater flexibility within artificially assembled vdW solids. We explore the use of N-HyTII processing to achieve both depth- and spatiallycontrolled doping within bilayer stacks comprising one layer of natural 12C graphene and one layer of isotopically enriched 13C graphene. Since the starting films are polycrystalline (grain size ~100 µm), the bilayer graphene samples display a variety of stacking orientations ranging from 0° to 30° and we refer to these films as twisted bilayer graphene (TBG).28,29 As illustrated in the inset of Figure 3(a), the various stacking orientations of TBG give rise to unique optical properties that depend on the relative twist angle between layers.30-34 Similar twist-angledependent signatures also appear in the relative intensities of the characteristic Raman bands. By using a more massive 13C graphene layer in the bilayer films the characteristic Raman peaks (D, G, and 2D) red-shift by ~48 cm-1, 56 cm-1, and 95 cm-1, respectively,30,33,34 which enables unique characterization of each layer. For example, in Figure 3(a) the bottom two Raman spectra highlight different stacking sequences (13C/12C (red) and 12C/13C (black)), where the “top” layer has a larger 2D peak intensity due to reduced doping from the underlying SiO2 substrate.30 The dependence of twist angle on the Raman spectra is also shown (labeled “resonant”, “small”, and “large”), which further confirms that the interface between the layers can be atomically clean to allow for direct electronic coupling. To simplify the analysis here, we focus our subsequent investigations on the “large” twist angles (>16° here), since they represent the most electronically independent regions.
10 ACS Paragon Plus Environment
Page 10 of 31
Page 11 of 31
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Nano
In Figure 3(b,c), we show Raman spectra of 12C/13C TBG N-HyTII processed with 46 eV and 200 eV N+, respectively, for a range of doses. The 46 eV N+ energy has a high probability for substitutional doping for single monolayer films (Fig. 2c). Focusing in the region near the D peak (~1347 cm-1) we observe a single peak emerge for the 12C graphene layer, which is the top layer of the TBG stack. Conversely, we observe the opposite response for sister samples comprised of 13C/12C
TBG processed under similar conditions (see Supp. Info., Figure S3). For the 200 eV N+ HyTII
treatment, both graphene layers display an approximately equal intensity D peak, indicating that the more energetic ions penetrate further into the TBG stack. We do not observe the substantial preference for defects within the top layer that was reported previously for bilayer graphene irradiated with 100 keV Ar+.18 In that study, a 30% reduction in defects was observe in the bottom layer. We observe a small difference in D/2D ratio, with the bottom layer showing fewer defects, but the difference is significantly less than 30%. We now transition our focus to investigate how N-HyTII impacts the transport properties of graphene using both in situ FET measurements and ex situ magnetotransport measurements. In both cases, we perform N-HyTII processing after we transfer graphene to a SiO2/Si substrate. In Figure 4(a) we show the transfer characteristics of a wire-bonded graphene test structure measured in situ during N-HyTII doping with 40 eV N+. The Dirac point (!, ) initially at ! ~ 0 V, shifts towards negative gate bias with increasing N+ dose. Concurrently, the drain current of the electron channel (i.e., for ! > !, ) increases (consistent with an increase in electron concentration and/or increased electron effective mobility), while the hole branch displays a degradation in drain current with increasing dose, and the minimum drain current (at ! = !, ) decreases within increasing dose. The inset in Figure 4(a) shows the extracted electron and hole field effect mobilities, µFE, plotted with respect to the extracted Dirac point minimum. The monotonic degradation in hole µFE is consistent with Coulomb scattering resulting from trapped holes in the underlying oxide.35-37 This is also consistent with the negative shift in the transfer 11 ACS Paragon Plus Environment
ACS Nano
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
characteristics. Typically, the electron field effect mobility follows a similar monotonic behavior as the hole field effect mobility, but in this experiment µFE-electron increases with N+ dose to over 1×1013 N+cm-2. We attribute this increasing µFE-electron with nitrogen substitutional doping, which provides excess electrons for transport. This process, however, is in competition with mobility degradation due to Coulomb scattering and defect scattering (at higher concentrations) that ultimately causes the electron mobility to saturate.36,37 We note that incident energetic electrons ( LD= 6 nm for the 0.96 sample and Lφ=660 nm> LD= 16 nm for the 0.09 sample. Nitrogen is an n-type dopant, and the strong localization behavior is consistent with other n-type dopants, such as hydrogen.41,42 The presence of strong localization as opposed to weak localization may provide further evidence of successful doping rather than simply the creation of defects, as strong localization is expected for impurity scattering and weak localization is more typical of defective 12 ACS Paragon Plus Environment
Page 12 of 31
Page 13 of 31
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Nano
graphene.43 Characteristic with localization, we also observe a complete loss of negative MR at room temperature. These results indicate that the HyTII process is capable of N-doping graphene and reliably altering its electronic properties. They also suggest that HyTII doping can offer a greater deal of electronic property control than can comparable methods that dope graphene during growth, as the amount of localization can be controlled directly and precisely by the doping process. CONCLUSIONS We have demonstrated a route to substitutionally N-dope graphene on Cu and SiO2/Si substrates using hyperthermal ion implantation (HyTII). By performing a systematic energy/dose study of HyTII processed graphene on Cu we determined the energy range, 30 eV to 50 eV over which N-doping graphene yields the highest doping fraction while minimizing direct knock-on or reaction-based vacancy formation. We demonstrated the depth controlled doping capability inherent to HyTII by doping the top layer of a twisted bilayer graphene sample comprised of isotopically labeled 13C graphene, which provides the depth-dependent discrimination when characterized using Raman spectroscopy. The viability of directly doping electrical test structures was confirmed. Using in situ electrical testing, we observed a negative shift in the graphene transfer characteristics and a concomitant increase in electron field effect mobility to a dose of 1.4×1013 N+cm-2. Similarly, we observed the dose-dependent changes in magnetoresistance (MR) of N doped graphene test structures, as evidenced by a dose-dependent saturation of the low temperature negative MR stemming from N substituent induced strong localization. These results establish the utility and define the operating range of N-HyTII processing for efficiently, and selectively, substitutionally doping graphene with Nitrogen. We envision HyTII processing of graphene and other 2-D materials to work in concert with in situ growth based approaches to doping to further our ability to control these materials at the atomic scale.
13 ACS Paragon Plus Environment
ACS Nano
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
EXPERIMENTAL METHODS Sample growth, transfer, and HyTII processing: We grew graphene by low-pressure chemical vapor deposition on copper foil using a methane precursor following the procedure detailed by Li et al.44 To grow 13C isotopically enriched graphene, we replaced standard methane with 13C labeled methane (99.5 atomic % 13C, Sigma-Aldrich) while all other procedures remained unaltered. We use these films to form isotpically labeled twisted bilayer graphene (TBG) following the procedure we have detailed previously,28 with 13C graphene comprising either the top or bottom layer. We hyperthermal ion implant (HyTII) these samples along with CVD graphene layers transferred to SiO2, CVD graphene on the Cu growth surface, and freshly cleaved highly ordered pyrolitic graphite (HOPG), with N+ ranging in energy from 25 eV to 200 eV using a Colutron Model G-2-D ion gun with an initial base pressure
