Comparison with Carbon Nanotube and Graphene - ACS Publications

Subscriber access provided by UNIV OF LOUISIANA

Functional Nanostructured Materials (including low-D carbon)

Edge-Functionalized Graphene Nanoribbon Chemical Sensor: Comparison with Carbon Nanotube and Graphene Kyeong Min Cho, Soo-Yeon Cho, Sanggyu Chong, HyeongJun Koh, Dae Woo Kim, Jihan Kim, and Hee-Tae Jung ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.8b16688 • Publication Date (Web): 13 Nov 2018 Downloaded from http://pubs.acs.org on November 13, 2018

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 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 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.

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 34 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 Applied Materials & Interfaces

Edge-Functionalized

Graphene

Nanoribbon

Chemical Sensor: Comparison with Carbon Nanotube and Graphene Kyeong Min Cho†‡ 〦 , Soo-Yeon Cho†‡ 〦 , Sanggyu Chong†, Hyeong-Jun Koh†‡, Dae Woo Kim†‡, Jihan Kim†, and Hee-Tae Jung*†‡ † Department

of Chemical and Biomolecular Engineering (BK-21 Plus), Korea Advanced Institute

of Science and Technology (KAIST), Daejeon 34141, Korea ‡ KAIST 〦 These

Institute for NanoCentury, Daejeon 34141, Korea authors contributed equally to this work

*E-mail : [email protected]

ACS Paragon Plus Environment

1

ACS Applied Materials & Interfaces 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

Page 2 of 34

ABSTRACT With growing focus on the use of carbon nanomaterials in chemical sensors, one-dimensional graphene nanoribbon (GNR) has become one of the most attractive channel materials, owing to its enhanced conductance fluctuation by quantum confinement effects and dense, abundant edge sites. Due to the narrow width of a basal plane with one-dimensional morphology, chemical modification of edge sites would greatly affect the electrical channel properties of a GNR. Here, we demonstrate for the first time that chemically functionalizing the edge sites with aminopropylsilane (APS) molecules can significantly enhance the sensing performance of a GNR sensor. The resulting APS-functionalized GNR has a sensitivity ((ΔR/Rb)max) of ~30% at 0.125 ppm nitrogen dioxide (NO2) and an ultra-fast response time (~6 s), which are, respectively, sevenfold and fifteen-fold enhancements compared to a pristine GNR sensor. This is the fastest and most sensitive gas-sensing performance of all GNR sensors reported. To demonstrate the superiority of the GNR-APS sensor, we compare its sensing performance with that of APS-functionalized carbon nanotube (CNT) and reduced graphene oxide (rGO) sensors prepared in identical synthesis conditions. Very interestingly, the GNR-APS sensor exhibited 30-fold and 93-fold enhanced sensitivity compared to the CNT-APS and rGO-APS sensors. This might be attributed to highly active edge sites with superior chemical reactivity, which are not present in CNT and rGO materials. Density functional theory clearly shows that the greatly enhanced gas response of GNR with edge functionalization can be attributed to the higher electron densities in the highest occupied molecular orbital levels of GNR-APS and incorporation of additional adsorption sites. This finding is the first demonstration of the importance of edge functionalization of GNR for chemical sensors.


2

Page 3 of 34 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


Keywords: graphene nanoribbon, carbon material, edge, defect, chemical sensor, density functional theory. Currently, solid-state gas sensors with low power consumption, miniaturization, device compatibility, multi-channel operation, and integrated circuit possibilities are widely utilized in various high-technology applications, including breath analysis for early diagnosis of patients, air pollution tracking, smart home systems, and industrial safety, owing to their ability to detect invisible harmful environmental variations.1–3 Various channel materials including metal oxide semiconductors,4 Si nanowires,5 two-dimensional (2D) materials,6 and carbon materials7 have been suggested for enhancing the performance of solid-state gas sensors. Among these materials, carbon nanomaterials including carbon black (CB), carbon nanofiber (CNF), carbon nanotubes (CNT), and graphene have displayed potential for next-generation autonomous sensing materials, because of their high specific surface area, high-quality crystal lattices avoiding grain boundary poisoning, high carrier mobility with superior conductivity, and ultra-thin-film processing.8 Among the various carbon nanomaterials, graphene nanoribbons (GNR), quasi-one-dimensional carbon structures with narrow widths (below tens of nanometers), exhibiting semiconductor characteristics via quantum confinement and edge effects are particularly promising for use in high-performance chemical sensing materials.9–13 This might be attributed to the unique properties of the GNR as a chemical sensing material compared to other carbon materials (i.e., CB, CNF, CNT, and graphene). First, GNR has the densest and most abundant edge defect sites with superior chemical reactivity within single-channel units.9 Due to the nonbonding of the localized states (each edge carbon atom possesses 0.14 unpaired electrons) and the closeness of the flat band to the Fermi level, the edge sites of GNR should resemble a radical, thus showing much more sensitive binding of analytes than pristine sp2-bonded graphene and a CNT channel.14 Furthermore,


3


Page 4 of 34

the transition from a semimetal to a semiconductor results in reduced charge density and better current modulation by identical analyte adsorption.12,15 In addition, the narrow channel width of the GNR is more accessible to doping and chemical modification effects.16,17 Thus, conduction pathways within the narrow GNR channel can be easily affected by edge-decorated chemistry or dopants. Here, we significantly enhanced the gas-sensing performance of the GNR by edge functionalization with (3-aminopropyl)triethoxysilane (APS). Bundles of GNR having long and dense edge defects with active chemical states are synthesized using a CNT unzipping process and covalently bonded with APS with a primary amine. Then, we compared the chemical-sensing performances of the edge-functionalized GNR (GNR-APS) with those of pristine carbon nanomaterials (CNT, reduced graphene oxide (rGO), and GNR) and those of APS-functionalized materials in order to demonstrate the unique role of the edge sites of GNR. The GNR-APS sensors exhibit significantly enhanced chemical-sensing performances compared to pristine GNR sensors. In addition, the edge-functionalization effects of the GNR are superior to those of the CNT and rGO. To the best of our knowledge, this is the most sensitive gas-sensing performance of the GNRbased sensors. Notably, this is the first reported demonstration of the superior chemical-sensing performance of the edge-functionalized GNR. This is expected to illustrate the key factors that govern the chemical-sensing performance of carbon nanomaterials, including edge defects, semiconducting properties, and functionalization.


4

Page 5 of 34 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


RESULTS AND DISCUSSION GNR and GNR functionalized with aminopropylsilane (GNR-APS) were prepared using a scalable top-down process, as depicted in Figure 1a (detailed descriptions are in the Materials and Methods section). First, we synthesized the GNR via chemically oxidative unzipping of multiwall carbon nanotube (MWCNT) bundles. Due to the intercalation of sulfuric acid on layers of MWCNTs, one-dimensional (1D) nanotubes were unzipped longitudinally and the resultant graphene oxide nanoribbons (GONRs) were obtained with abundant oxygen functional groups, including hydroxyl (-OH), ketone (-C=O), and carboxylic (-COOH) groups.18,19 The GONR solution was chemically reduced using hydrazine as a reducing agent. To prepare GNR-APS, we grafted 3-aminopropyltriethoxysilane onto the GONRs at 85 °C under nitrogen (N2). To further reduce the GNR-based sensing channel, GONR-APS was chemically reacted with a hydrazine and ammonia (NH3) solution. Similarly, rGO functionalized with APS (rGO-APS) was prepared via functionalization of APS and additional chemical reduction by hydrazine. For the CNT functionalization, we oxidized the CNT surface using sulfuric acid and functionalized the resultant oxygen functional group with APS. The transmission electron microscopy (TEM) image clearly shows that the MWCNTs have a diameter of 10–20 nm with ~20 carbon layers (Figure 1b). After oxidative cleavage of the MWCNTs, we observed a sheet-like morphology instead of a layered nanotube structure. The nanoribbon has a quasi-1D structure with a width of 80-100 nm (Figure 1c). After functionalization of APS and partial reduction, the GNR-APS exhibits no obvious changes in the dimensions of the ribbon compared to GONR. In addition, grafted organic molecules covered the surface of the GNR-APS, and thick layers caused by aggregation and polymerization of APS cannot be observed (Figure 1d).


5


Page 6 of 34

To confirm the unzipping of the CNTs and APS functionalization, we characterized the CNT, GNR, and GNR-APS using Raman spectroscopy (Figure 1e). The Raman spectra of the MWCNT show characteristic peaks at 1357 (D band), 1587 (G band), and 2697 cm−1 (2D band). The D and G bands correspond to the disorder of the graphene structure and in-plane vibration of sp2 carbon in graphitic materials, respectively. The D band observed in unzipped GNR was broadened compared to that in MWCNT as a result of edge and defect generation. In addition, the G bands of GNR were blue-shifted from 1587 cm−1 to 1605 cm−1, which was induced by chemical oxidation and exfoliation of layered.20-22 The 2D band arises from the overtone of the D band, which may be changed by the relative orientation and stacking order of the layered graphene. With oxidative unzipping of the MWCNT, the 2D band is noticeably broadened and decreased in the Raman spectra of the GNR, which is due to unzipping and exfoliation of the well-stacked layers of nanotubes, changing into loosely stacked nanoribbons.23 Thus, the differences in the Raman spectra clearly indicate the formation of GNR via oxidative unzipping of MWCNT. On the other hand, GNR-APS exhibits no obvious changes compared to GNR. To further confirm the functionalization of APS on GNR, we conducted X-ray photoelectron spectroscopy (XPS). From the XPS survey, GNR-APS contains substantial amounts of silicon (Si) and nitrogen (N) originating from the aminosilane group. The GNR and GNR-APS, respectively, contain 3.36% and 6.93% of nitrogen, while CNT has no nitrogen species (Figure 1f). In the XPS N 1s spectra of GNR, we attribute the peak at lower binding energy (398.5 eV) to pyridinic N, which was inserted into defects in the basal plane during the chemical reduction with hydrazine (N2H4). Specifically, in the XPS N 1s of GNR-APS, a peak at higher binding energy (399.7 eV) was observed, which is assigned to the amino groups (-NH2) of functionalized aminosilane (Figure 1g). From the XPS C 1s spectrum of GNR, we observed an oxygen-functional group of the carbon–


6

Page 7 of 34 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


oxygen bond (C–O, 286 eV) and a carboxylic group (C(O)–O, 288 eV). After aminefunctionalization of GNR, the new peak at 285.5 eV corresponding to the C–N bond is generated (Figure S1).24,25 Accordingly, amino groups are successfully functionalized on the GNR.


7


Page 8 of 34

Figure 1. (a) Schematic of the synthesis of graphene nanoribbon (GNR) and GNR functionalized with aminopropylsilane (GNR-APS). TEM images of (b) CNTs, (c) a GNR, and (d) GNR-APS (magnified images of the white-dashed box regions are shown in inset). (e) Raman spectrum, (f) XPS survey, and (g) N 1s of the CNT, GNR, and GNR-APS.


8

Page 9 of 34 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 conducted Fourier-transform infrared (FT-IR) spectroscopy of GNR-APS to identify the chemical attachment of APS onto the GNR (Figure 2a). The spectra for GONR exhibited peaks at 1030, 1224, 1418, and 1715 cm−1, which are assigned to the abundant oxygen functional group including epoxide (C–O–C), carbonyl (C–O), hydroxyl (OH), and ketone group (C=O).26 These peaks indicate that the prepared GONR has the abundant oxygen functional group due to the chemically oxidative unzipping of the CNT. In the spectra of GNR-APS, we observe prominent peaks at 750, 910, and 1030 cm−1 associated with Si–O, Si–OH, and Si–O–Si bonding indicative of APS functionalization (Figure S2). In addition, the new band of -NH2 bonding emerged at 1524 cm-1. The results indicate that the alkoxy group (–R) of APS was hydrolyzed and the silane group reacted to the oxygen functional group of GONR and amine group was functionalized onto surface of the GNR (Figure 2b).27,28 To clearly verify that the APS molecules were chemically attached at the edge sites of the GNR, we conducted elemental mapping of C, N, and Si on GNR-APS using energy-dispersive X-ray (EDX) analysis via TEM. In the EDX mapping images, the carbon (red), nitrogen (green), and silicon (blue) elements are clearly visible on the GNR (Figure 2c). The carbon spread throughout the GNR surface originated from the sp2 carbon lattice of the GNR. Interestingly, the Si atoms are concentrated at edge and defect sites where the abundant oxygen functional groups are located on the GONR. In the FT-IR results, the silane group of APS chemically bonded to the GNR, suggesting that the location of a Si atom indicates a functionalization site. Accordingly, APS molecules are predominantly functionalized on the edge of the GNR. The majority of nitrogen atoms are dispersed near the Si atoms associated with the amine group of functionalized APS. Some of the nitrogen atoms are placed in the basal plane by chemical doping with hydrazine reduction (Figures 2d and 2e). As the amount of loaded APS was increased, N and Si were


9


Page 10 of 34

distributed uniformly onto the whole surface. As a result, at first, APS was grafted onto the edge and defect sites of the GONR where many oxygen functional groups were presented. With additional APS molecules, polymerized APS clusters covered the GNR surface (Figure S3).29 Based on FT-IR spectroscopy and elemental mapping, the APS molecules are uniformly functionalized onto the edge of the GNR by reaction between the silane group of hydrolyzed APS and the oxygen functional group of the GONR.


10

Page 11 of 34 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


Figure 2. (a) FT-IR spectrum of GONR and GNR-APS. (b) Suggested chemical structure of GNRAPS. (c) TEM EDX mapping for carbon (red), nitrogen (green), and silicone (blue) on GNR-APS. (d) EDX mappings for each element. (e) Line profile of GNR-APS.


11


Page 12 of 34

Figure 3a shows the construction of the sensing devices. The highly thin and uniform films used for the channel materials were fabricated using vacuum filtration with a porous alumina membrane filter. We then transferred the filtrated films onto Si/SiO2 substrates printed with an interdigitated micro-electrode.30 Optical microscopy (OM) and top-view scanning electron microscopy (SEM) images show that the ultra-thin functionalized GNR-APS films are uniformly integrated with the 100-μm Cr/Au electrode. Surface SEM images of the CNT and rGO sensors before and after APS functionalization are shown in Figure S4. We can clearly see that highly thin and uniform surfaces are maintained even with APS functionalization, indicating that APS is finely chemically bonded with channel materials without any chunks or polymerizations. To investigate the contact resistance at the junction between the CNT, rGO, GNR, and GNRAPS channels and metal electrodes, we obtained current−bias (IDS−VDS) curves using a probe station (Figure 3b). The IDS−VDS curves of the CNT, rGO, GNR, and GNR-APS channels all exhibit good ohmic contact characteristics with linear IDS−VDS curves, indicating that contact resistance is not a dominant factor and that the channels are well connected to the pre-deposited Au/Cr electrode.25 In particular, the GNR and GNR-APS devices have IDS−VDS curves with low gradients, indicating higher channel resistance between the source and drain due to the semiconductor characteristics and quantum confinement effects of the GNR band structure (magnified IDS−VDS curves of the GNR and GNR-APS devices are shown in the inset, rightbottom).31 Figure 3c shows the baseline channel resistance of the sensors. The CNT and rGO sensors exhibit low channel resistance on the order of tens of ohms due to the good transduction properties of the materials. However, the GNR sensor exhibits a significant increase in resistance by three orders of magnitude. When the GNR is functionalized with APS, channel resistance increases by


12

Page 13 of 34 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


one order of magnitude due to conduction pathway hindrance between GNR junctions by APS molecules.25 These devices were simultaneously loaded on a homemade gas-sensing chamber, and the sensing signals of each film were measured using multichannel sensing systems (a detailed schematic of the entire gas delivery system is shown in Figure S5). We applied a constant drain bias to the two-probe sensor, and the change in electrical resistance of the sensor upon exposure to analytes was monitored and recorded as the sensing signal.32 Figure 3d shows the real-time gas response behavior of the CNT, rGO, GNR, and GNR-APS sensors to repetitive exposure of 5-ppm NO2. First, the CNT and rGO sensors display slight decreases in resistance of 0.7% and 1.5% (magnifications of the real-time response behavior are shown in the right graph of Figure 3d). Typically, the sensors based of carbon material, including CNT and rGO, displayed negative responses to the NO2 molecules, as the pristine carbon materials most often exhibit p-type semiconductor properties in ambient conditions because of the adsorption of oxygen, water, solvent molecules, and other impurities.25,33 Electron transfer from the CNT and rGO channels to NO2 effectively increases the hole carrier concentration of the channel, thereby decreasing electrical resistance. Because of the unzipping of the CNT, the GNR sensor exhibits a significant response to the same analytes with a scale of the order of 10%, which is approximately a twenty-fold enhancement. Accordingly, the real-time signal of the CNT and rGO sensors cannot even be seen at the same signal scale as for the GNR sensor. The significant improvements in the response amplitude can be attributed to (i) highly dense and abundant edge defects in the narrow channel and (ii) the reduced charge density and better current modulation by the analytes.9,12,14 Interestingly, the gas-sensing performance of the GNR is significantly improved with APS functionalization in terms of both response amplitude (S) and


13


Page 14 of 34

response time (τ90). We calculate the response time (τ90) as the time taken by the sensor resistance to reach 90% of its minimum value upon exposure to gas. Before APS functionalization, the pristine GNR sensor exhibited S ≈ 28% and τ90 ≈ 2.6 min. Surprisingly, after APS functionalization, both S and τ90 were significantly enhanced to respectively 93% and 6 s, which are 3.3- and 26-fold improvements, respectively. This response time on the order of a few seconds at parts-per-million level of NO2 is the fastest among GNR-based sensors and is also hard to achieve with CNT and graphene-based gas sensors (Table S1).10–13 This might be due to the strong adsorption properties of NO2 by the primary amine (RNH2) of APS. The primary amine of APS can strongly adsorb the NO2 with a chemisorption process forming a nitrate compound.2 Thus, GNR-APS can attract and adsorb greatly enhanced amounts of NO2 within a much shorter time. As described in the Introduction, one of the strong points of using a GNR as a sensing channel is that it is more accessible to doping and functionalization effects owing to its narrow channel width and highly dense edge defect sites. To verify this point, an identical APS functionalization process (APS mixing at 85°C for 24 h, stirring) was applied to a CNT and rGO and the sensing performance of these sensors were simultaneously compared (Figure 3e). For the CNT sensor, there were negligible improvements to both S and τ90, indicating that APS is difficult to bond to CNT surfaces with low numbers of defect sites.14 In addition, the sp2 defect sites of CNT have much lower chemical reactivities compared to those of GNR edge defects. Accordingly, APS is difficult to attach to sp2 defect sites of a CNT under identical functionalization conditions as for the GNR. In contrast, for the rGO sensor, the response amplitude is dramatically enhanced by APS functionalization, rising from S ≈ 1.5% to S ≈ 12.18%. Similar to the preparation of the GNR, a large number of hydroxyl groups are provided on the rGO surface during chemical exfoliation of graphite with strong acid and oxidizing agents.34 Thus, APS can be easily bonded to the hydroxyl


14

Page 15 of 34 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


groups of rGO, resulting in greatly enhanced NO2 adsorption on the rGO surfaces. Meanwhile, rGO-APS did not display a significant enhancement in response time, unlike the GNR-APS cases. The τ90 of rGO improved from 4.1 min to 3.8 min, which is only a 1.07-fold enhancement, while GNR-APS had a 26-fold enhancement. This is due to the difference in accessibility of analytes between the GNR and rGO nanostructures.16,17 Since GNR width is less than a few tens of nanometers, the primary amine sites of APS can be saturated with NO2 adsorption in a very short time and the strongly adsorbed NO2 can quickly induce charge transfer along the whole region of channel materials. For these reasons, the τ90 of the GNR can be significantly improved with APS functionalization. In contrast, for rGO with a relatively larger width than the GNR (a few micrometers in width), it should take a much longer time to transfer charge variation due to NO2 adsorption of APS. Thus, the τ90 of rGO enhancement is negligible with APS functionalization, even though the NO2 adsorption is greatly enhanced with APS. Accordingly, we can conclude that the GNR is a unique and very suitable carbon nanostructure for highly efficient doping or functionalization effects on gas-sensing performances. To further study the effect of APS functionalization on the chemical-sensing behavior of resulting sensors, we performed DFT calculations. Details of the calculations can be found in the Materials and Methods section. First, we obtained energy-minimized configurations for an armchair-GNR model with ten carbon atoms in the lateral direction (10AGNR) before and after APS functionalization onto a COOH edge defect. Then, we introduced a single NO2 molecule into the system and performed a second energy relaxation to study the gas adsorption behavior. We chose NO2 for computational validations as it yielded the greatest enhancement in device sensitivity experimentally. We studied three different adsorption scenarios: adsorption onto the primary amine of 10AGNR-APS, adsorption onto the ribbon surface of 10AGNR-APS, and


15


Page 16 of 34

adsorption onto the ribbon surface of 10AGNR. The final relaxed configurations of three adsorption schemes are presented in Figures S10-12. Subsequently, total density of states (TDOS), NO2 adsorption energies, and charge density difference plots were obtained for the resulting AGNR systems. The TDOS plots of the resulting systems shown in Figure S6 confirm that the sensor devices are all p-type semiconductors, as the highest occupied molecular orbital (HOMO) levels are found just below the Fermi level in all cases. Interestingly, the TDOS plots do not show a significant difference between the energy gaps of the modelled sensor systems, with all three systems exhibiting an energy gap of around 1.0 eV. The negligible differences in the energy gaps between the tested sensor devices hint that selective detection cannot be newly achieved with APS functionalization, which is indeed the case in our experiments. However, the APS-functionalized systems show higher electron densities in the HOMO levels. Considering that real APSfunctionalized devices would have higher rates of APS functionalization compared to the model systems, differences in HOMO densities could be even higher, which could improve the sensitivity of APS-functionalized devices. We calculated the binding energy of NO2 on each model system to determine whether the primary amines of APS can serve as competent additional binding sites for the target adsorbates (Figure 3f). The results, presented in Figure 3f, show that the amine group of APS does exhibit a comparable binding energy with binding sites on the ribbon surface. Figure 3g (left) shows that the amine site of 10AGNR-APS is located away from the basal plane of GNR, which makes it a separate binding site from the ribbon surfaces. Hence, the gas molecules would bind to both the ribbon surface and the primary amines of the APS functional groups, leading to an enhanced sensitivity of the APS-functionalized device. We also plotted charge density differences to see


16

Page 17 of 34 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


whether adsorption to the APS amine site would cause sufficient change in the charge density of the GNR plane to produce a detectable signal. The charge density difference plot for 10AGNRAPS shown in Figure 3g (right) shows significant changes in the charge density of GNR, though NO2 is adsorbed onto the primary amine site located away from the basal plane. Furthermore, these changes are comparable to that induced by NO2 adsorption directly onto the GNR surface as shown in Figure S9. Overall, the DFT calculations clearly demonstrate that functionalization of APS onto the GNR can significantly enhance the sensitivity of the resulting sensor devices by providing additional binding sites found in the primary amines of APS.


17


Page 18 of 34

Figure 3. (a) OM and SEM images of the GNR-APS sensor structure. (b) IDS–VDS curves of the CNT, rGO, GNR, and GNR-APS devices showing ohmic contact behavior. (c) Channel resistances of each device. (d) Real-time sensing performance of the sensors to repetitive exposure of 5-ppm NO2 (right: magnified scale for the CNT and rGO sensors). (e) Real-time sensing performance of the CNT and rGO sensors before and after APS functionalization. (f) NO2 adsorption energies of different 10AGNR systems. (g) Left: final configuration of 10AGNR-APS with NO2 adsorbed to amine (grey: C; white: H; blue: N; red: O; yellow: Si). Right: charge density difference plot of 10AGNR-APS with NO2 adsorbed on amine (yellow: positive charge density, cyan: negative charge density).


18

Page 19 of 34 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


Figure 4a shows the real-time NO2 sensing behavior of the sensors from 0.125 ppm to 5 ppm. A magnified graph with a smaller scale y-axis showing the CNT and rGO sensing performance is shown in Figure 4b. It can be clearly seen that both the GNR and GNR-APS sensors have a higher gas response to various concentrations of NO2, which is barely apparent for the CNT and rGO sensors. In particular, only the GNR-APS sensor produces a significant signal for the lowest concentration of NO2 (0.125 ppm) with a 30% resistance variation and an ultra-fast response time of

Comparison with Carbon Nanotube and Graphene - ACS Publications

Recommend Documents