Enhanced Chemical Reactivity of Graphene by Fermi LevelModulationMyung Jin Park,*,†,⊥ Hae-Hyun Choi,‡ Baekwon Park,‡ Jae Yoon Lee,† Chul-Ho Lee,† Yong Seok Choi,‡

Youngsoo Kim,§ Je Min Yoo,‡ Hyukjin Lee,∥ and Byung Hee Hong*,‡

†KU-KIST Graduate School of Converging Science & Technology, Korea University, 145 Anam-ro, Seongbuk-gu, Seoul 02841,Korea‡Department of Chemistry, Seoul National University, 1 Gwanak-ro, Seoul 08826, Korea§Graphene Square Inc. Inter-University Semiconductor Research Center, Seoul National University, 1 Gwanak-ro, Seoul 08826,Korea∥College of Pharmacy, Ewha Womans University, 11-1 Daehyun-Dong, Seodaemun-Gu, Seoul 120-750, Korea⊥Department of Chemistry, University of Toronto, 80 St. George Street, Toronto, Ontario M5S 3H6, Canada

*S Supporting Information

ABSTRACT: Among various approaches to modify theelectronic and chemical properties of graphene, functionaliza-tion is one of the most facile ways to tailor these properties.The rearranged structure with covalently bonded diazoniummolecules exhibits distinct semiconducting property, and theattached diazonium enables subsequent chemical reactions.Notably, the rate of diazonium functionalization depends onthe substrate and the presence of strain. Meanwhile, accordingto the Gerischer-Marcus theory, this reactivity can be furthertuned by adjusting the Fermi level. Here, we preciselycontrolled the Fermi level of graphene by introducing the self-assembled monolayer (SAM) and investigated the degree ofchemical reactivity of graphene with respect to the dopingtypes. The n-doped graphene exhibited the highest reactivity not only for diazonium molecules but also for metal ions. Theincreased reactivity is originated from a remarkable electron donor effect over the entire area. In addition, the n-doped grapheneenabled spatially patterned functionalization of diazonium molecules, which was further utilized as a growth template for goldparticles that would be advantageous for enhanced electrochemical reactivity.

Providing new functions to graphene beyond its intrinsicproperties is important as it can broaden the potential

applications for electronic devices1,2 and biosensors.3,4 Studieson the chemical functionalization of graphene with diazoniummolecules have been exploited for bandgap tuning5−7 andfluorescence probe/biological sensor development.8,9 Forefficient chemical functionalization, proper understanding ofchemical reaction between this gapless substance and otherreactants must be preceded. In graphene, pi-electrons in theperpendicularly positioned pz orbital of sp2 hybridizedstructure can serve as an electron donor in chemical reactions.Previous works demonstrated that electron transfer fromgraphene to diazonium salt allowed covalent bond formationbecause the Fermi level of graphene and the vacant oxidationstate of diazonium salt are well-matched.11 Other studiessuggested that the chemical reactivity of graphene was moreactivated along the edge than the basal plane due to theexistence of dangling bonds.10,12,13 Notably, the rate ofgraphene functionalization can be enhanced by strain

engineering14 or by employing appropriate substrates withinhomogeneous charge fluctuations.15

Meanwhile, according to the Gerischer−Marcus theory, theelectron transfer rate is proportional to the allowed redoxpotential gap between different reactants.11,16 Given thatgraphene has a linearly dispersed band structure, the Fermilevel modulation by external condition can be achieved, whichindicates that the electron transfer rate from graphene can becontrolled. This unique band structure that shows limitationsin logic device applications is still worth investigating in termsof graphene chemistry. We noted that the Fermi level ofgraphene is susceptible to the substrates and/or the presenceof external molecules.17−20

Up to date, various doping methods influencing the Fermilevel of graphene have been reported including electric field,21

self-assembled monolayer (SAM),22,23 and photoinduced

Received: April 18, 2018Revised: August 1, 2018Published: August 1, 2018

Article

pubs.acs.org/cmCite This: Chem. Mater. XXXX, XXX, XXX−XXX

© XXXX American Chemical Society A DOI: 10.1021/acs.chemmater.8b01614Chem. Mater. XXXX, XXX, XXX−XXX

Dow

nloa

ded

via

Byu

ng H

ee H

ong

on A

ugus

t 21,

201

8 at

20:

39:3

5 (U

TC

).

See

http

s://p

ubs.

acs.

org/

shar

ingg

uide

lines

for

opt

ions

on

how

to le

gitim

atel

y sh

are

publ

ishe

d ar

ticle

s.

doping.24,25 Among these, the SAM method is particularlyuseful due to its stability even in the ambient condition. It canbe directly applied to the oxide substrate through silanizationand introduction of functional groups at the end of silane caneffectively influence the Fermi level of graphene. To decide theFermi level with respect to the doping type, substratemodifications were carried out with different SAMs: (1) ann-doped substrate by 3-aminopropyltriethoxysilane (APTES)with electron-donating lone pairs in −NH2, (2) a neutralsubstrate treated with octadecyltrichlorosilane (OTS) thatblocks the undesired charge, and (3) a pristine SiO2 substratethat has p-doping character caused by charged impurities.The chemical reaction of graphene shows distinguishable

behaviors for each doping type. The graphene on APTES (G/APTES) with the highest Fermi level exhibits significantlyhigher reactivity not only for the diazonium but also for thegold chloride compared with those of graphene on OTS (G/OTS) or SiO2 (G/SiO2). Our results suggest the potential ofgraphene to be utilized as a chemical platform in the 2Dmaterial chemistry.The electron transfer between graphene and diazonium is

thermodynamically allowed because the vacant oxidation stateof diazonium is below the Fermi level of graphene. Theelectron transfer rate (kET) can be defined by the equationfrom the modified Gerischer−Marcus theory,10,11,13−16

k E E W E E( )DOS ( ) ( )dE

E

ET red G oxredox

F,G∫ν ε=

where ν is the electron transfer frequency, EF,G is the Fermilevel of graphene, Eredox is the redox potential of reactant, εred isthe proportionality factor, DOSG is the density of state ofgraphene, and Wox(E) is the probability density function ofvacant states in the diazonium. From this equation, we assumethat the reaction rate of graphene is directly influenced by andproportional to the Fermi level.Schematic illustration in Figure 1a shows the representative

overall reaction process through the electron transfer in n-doped G/APTES. The neutral and p-doping types are decidedby with or without OTS treatment on the SiO2/Si substrate,respectively. Reaction for all the samples proceeded under thesame condition.To investigate the charge density of graphene corresponding

to the doping types, we analyzed the characteristics of thegraphene field effect transistors (FET) and Raman spectra.Figure 1b shows the FET characteristics for each graphene.The Dirac point of G/SiO2 appears at the positive Diracvoltage (Vdirac = +45 V), which is shifted to Vdirac = +5 V andVdirac = −165 V, respectively, for G/OTS and G/APTES. Thisindicates that the Fermi level of graphene is effectively adjustedby the underlying substrates. The spatial doping analysis of the

Figure 1. (a) Schematic illustration of the overall reaction process (i.e., reaction with diazonium and gold chloride) through the electron transferfrom n-doped G/APTES. Yellow region of the left image represents the SAM on SiO2/Si substrate. (b) Isd−Vg transport characteristics of graphenesupported on different doping inducing substrates. Inset images indicate graphene samples transferred on each substrate with different Fermi levels.(c) Correlation between G and 2D peaks of G/OTS (black symbols), G/SiO2 (blue symbols), and G/APTES (red symbols). Inset: arrows markedas eC, eT, eH, and en represent vectors affected by compressive strain, tensile strain, hole doping, and electron doping effect, respectively. Green dot(zero point) represents the strain and doping free graphene. Linear slopes are for the strain (black color), hole doping (red color), and electrondoping (dark yellow color), respectively, which were acquired from refs 21 and 26.

Chemistry of Materials Article

DOI: 10.1021/acs.chemmater.8b01614Chem. Mater. XXXX, XXX, XXX−XXX

B

Figure 2. (a) Representative Raman spectra of graphene that was supported on each substrate before and after diazonium functionalization. (b−g)Spatial Raman mapping images of D/G intensity ratio (ID/G) corresponding to graphene supported on each substrate before (b, d, f) and afterdiazonium functionalization (c, e, g). Mapping images were collected from 225 spectra points for each sample. Same regions were measured tocompare with before and after diazonium functionalization. All of the mapping areas are 30 μm × 30 μm, and scale bars are 10 μm.

Figure 3. Raman peak analysis extracted from mapping data before (O) and after (X) diazonium functionalization. (a) I2D/G as a function of G peakposition. (b) fwhm of 2D peak as a function of 2D peak position. (c) fwhm of G peak as a function of G peak position. Dotted line was obtainedfrom ref 21. (d) ID/G as a function of the modified Fermi level. Fermi level values for each sample are obtained from experimental Raman mappingdata of before functionalization and those values for G/OTS and G/SiO2 were adjusted by considering the electron−hole puddle model.

Chemistry of Materials Article

DOI: 10.1021/acs.chemmater.8b01614Chem. Mater. XXXX, XXX, XXX−XXX

C

channel area for each sample was conducted by a Ramantemplate that elucidates the correlation between strain anddoping effects (Figure 1c).26 Hundreds of data points for Gand 2D modes were acquired by Lorentzian fitting. Whengraphene is under strain (compressive and tensile) or holedoping effect, both G and 2D peaks are upshifted with differentslopes; an uniaxial strain slope is (Δω2D/ΔωG)ε

uniaxial = 2.2,and a hole doping slope is (Δω2D/ΔωG)n

hole = 0.75. TheRaman points of G/SiO2 follow the uniaxial strain slope,moving toward the hole doping vector, which indicates thatgraphene is under p-doping (n: 4.8 × 1012 cm−2 ≈ Vdirac = 67V) effect with compressive strain (εc: 0.1%). Those of G/OTSare converged near 1582.5 ± 0.7 cm−2 and 2679.2 ± 1.1 cm−2,which correspond under relatively low p-doping (n: 1.8 × 1012

cm−2 ≈ Vdirac = 25 V) with negligible strain. On the contrary,the points of G/APTES are deviated downward from the holedoping slope. Considering that the 2D peak of graphene is lesssensitive to n-doping than p-doping and the peak isdownshifted under the high doping level, this is consistentwith the previous report.21 Overall, the spectral behaviors foreach sample coincide with the doping types, even though thecharge densities are somewhat larger than those values derivedfrom the FET measurement due to the gate-induced p-dopingeffect.20,27

To investigate the correlation between the chemical reactionand the Fermi level height of graphene, each sample withdifferent doping types was reacted in the 4-nitrobenzenedia-zonium tetrafluoroborate (4-NBD) solution at 30 °C for 4 hand analyzed through Raman peak information. In Figure 2a,differently doped samples had a sp3 bond related D peak withdifferent intensities in the order of G/APTES, G/SiO2, and G/OTS after functionalization, which was clearly verified inmapping images (Figure 2b−g and Figure S1). This indicatesthat electron transfer the most actively occurs in the n-dopedgraphene. However, the D peak intensity for G/SiO2 washigher than those of G/OTS, which is inconsistent with thereactivity being proportional to the Fermi level. This can beexplained by the electron−hole puddle model and will bediscussed later.We further investigated the spectral behavior with respect to

the reaction through the G and 2D peak information from themapping data in Figure 2. The 2D peak is noticeably sensitiveto the doping density, leading to decrease of its intensity andincrease of full width at half-maximum (fwhm) under highdoping density. This is originated from the increased electron−electron collision and inelastic scattering rates during thedouble resonance process.28−30 In addition, the G peakposition is upshifted proportionally to the doping density.Figure 3a is a graph for 2D/G intensity ratio (I2D/G) plottedagainst the G peak position. Before diazonium functionaliza-tion, the data points of I2D/G for each sample are decreased inthe order of G/OTS, G/SiO2, and G/APTES, which is similarbehavior to the Vdirac position in FET.After diazonium functionalization, I2D/G for all samples

decreased due to the p-doping effect of the nitro group indiazonium.14 The data points of the 2D peak fwhm are also inagreement with the doping effect, which follows the order ofdoping density initially and further increases after functional-ization. However, it is hard to differentiate the doping effectwith respect to functionalization that causes the D peak. This isdue to the physical adsorption tendency of diazoniummolecules being able to form the charge transfer complexeson the basal plane of graphene without breaking sp2 bonds. In

Figure 2a, the diazonium related peaks are definitely observedin all samples. This indicates that graphene can be doped byboth physical and chemical bond formation of diazonium.12

Figure 3c is a graph for the G peak fwhm plotted against the Gpeak position. The solid line indicates the doped graphenereferenced from ref 21 to differentiate doping and strain effect.The fwhm depends the doping effect before reaction, while it isinfluenced by the strain effect after reaction. This indicates thatthe surface of graphene lies on the mechanical deformationwhen graphene is exposed to diazonium.To investigate the surface morphology influenced by

diazonium, G/APTES in before and after reaction wereanalyzed using atomic force microscopy (AFM) (Figure S2).The thickness of chemical vapor deposition (CVD) graphenewas about 0.68 nm, which is close to the monolayer. Thethickness increased to 1.48 nm through the APTES SAMtreatment, and it was even more up to 4.02 nm by diazoniumfunctionalization (Figure S2b,c). Surface roughness was alsoincreased, which was confirmed by the height histogram for theentire region (Figure S2d). Interestingly, the diazoniumfunctionalization made the wrinkled regions clearer, resultingin surrounding areas to be protruded (Figure S2c). Thisinforms that the attachment of diazonium molecules is locallyactivated at the curvatured regions, leading to additionaldeformation.Figure 3d is a graph for D/G intensity ratio (ID/G) plotted

against the Fermi level. Fermi level was modified byconsidering the electron/hole puddle model to explain theunexpected behavior in Figure 2, where ID/G of G/SiO2 waslarger than that of G/OTS. Previous reports investigated thatrandomly fluctuated charge potential of SiO2 plays a negativerole with undesired doping effect in graphene devices.31,32 Thisrandom potential was explained by the electron/hole puddlemodel, which was introduced to calculate the practical Fermilevel used in the reaction.15

The modified Fermi level (EF,n) is derived from averageFermi level (EF,avg) (Figure S3). First, EF,avg is obtained fromthe IG/I2D and can be described by33

II

C f E( ( ) )G

2De ph F,avgγ ε= + | |‐

(1)

where γe‑ph (≈ 21 meV) is the average electron scatteringenergy of phonon emission at 514 nm and C (≈12.38 eV−1) isthe constant modified from ref 33. f is a function depending onthe dielectric constant (ε) and the calculated value is 0.075 forSiO2 (ε ≈ 4).28 EF,avg of each sample is adjusted as (−) for G/APTES and (+) for the others, respectively, which was decidedfrom the Vdirac position in the FET data. Modified Fermi level(EF,n) influenced by electron−hole puddle can be describedby15

E E ( )F,n F,avg 2Dα= + ΔΓ (2)

where α (= 0.08 eV cm) is a proportionality constant andΔΓ2D is the fwhm difference of the 2D peak between pristinegraphene (≈ 23 cm−2)26 and doped graphene. The modifiedFermi level (EF,n) was only applied to the G/SiO2 and the G/OTS because the G/APTES showed n-doping behavior(Figure S3). Consequently, the data in Figure 3d derivedfrom eqs 1 and 2 demonstrates that the reactivity isproportional to the Fermi level. The n-doped grapheneshowed catalytic character, which enabled selective function-alization (Figure 4). This selectivity clearly demonstrates that

Chemistry of Materials Article

DOI: 10.1021/acs.chemmater.8b01614Chem. Mater. XXXX, XXX, XXX−XXX

D

the introducing functional molecules that directly affects theFermi level height of graphene is more favorable to thechemical reaction than the pristine substrate. Selectivelyfunctionalized 2D region gives the versatility to the grapheneto be utilized in the vertical direction as well as in the lateraldirection.To further investigate the activated electron transfer of n-

doped graphene, we observed that novel metal ions with highoxidation number are reduced on graphene. For example,introducing gold particles to graphene is known to increase theFermi level of graphene.34 Assuming that gold particle isdirectly grown on graphene, it could provide additional n-doping. In fact, the gold particles were well-formed in allsamples without any addition of reducing agents (Figure 5 andSupporting Information).Since the redox potential of gold ion (−5.5 eV) is lower than

that of graphene (−4.4 eV), it is assumed that the electrontransfer from graphene to gold ions is energetically favorable.35

The gold particle concentration was the highest at G/APTES,which is analogous to diazonium functionalization. However,

the degree of concentration at G/OTS was comparable to G/SiO2 or even more distinguishably higher in some samples,which is the opposite behavior where G/SiO2 shows higherreactivity than G/OTS in the diazonium solution (Figure 5a−i). Thus, we extended the existing electron−hole puddle modelto account for this behavior (Figure 5j).In the diazonium reaction, each diazonium cation needs one

electron to make a bond with graphene. On the other hand, inthe gold ion reduction, the gold ion needs three electrons forthe reduction of one atom (Au3+(aq) → Au0(s)), which isdescribed by following formula.

AuCl (aq) 3e Au (s) 4Cl (aq)40+ → +− − −

In this reaction driven by multi-electron transfer, theuniformity of the Fermi level height can be more importantthan the electron−hole puddle with high amplitude. Assumingthat the distance between puddles for the electron transfer islarge, electrons cannot be effectively supplied to Au3+ (Figure5j). The uniformity of the Fermi level height was confirmed

Figure 4. Spatially controlled graphene functionalization. (a) Micropatterned OM image (50 × 50 μm2) fabricated by the electron-beam (E-beam)lithography. APTES is deposited on the micropatterned area (purple area). Subsequently, E-beam resist is removed by acetone. (b) OM image oftransferred CVD graphene on the selectively patterned APTES-SAM substrate. (c) Mapping image of G peak position before functionalization. Gpeak position is upshifted at patterned area. (d) Mapping image of 2D peak position before functionalization. Patterned image is not differentiatedbecause of the insensitivity of 2D peak under n-doping. (e) Mapping image of I2D/G before functionalization. The ratio is relatively lower along thepatterned area, which indicates that doping density of APTES is higher than that of SiO2. (f) Mapping image of ID/G before functionalization. Theratio is almost 0 over the entire area, which indicates that the defect density of CVD graphene is negligible. (g) Mapping image of ID/G afterfunctionalization. The ratio increases along the APTES patterned area. (h) ID/G height profile for a dotted line in (g). ID/G is higher along thepatterned area where graphene is functionalized with diazonium.

Chemistry of Materials Article

DOI: 10.1021/acs.chemmater.8b01614Chem. Mater. XXXX, XXX, XXX−XXX

E

from the Dirac point distribution in FET. We conducted thestatistical analysis by measuring 12 FET devices per eachdoping type (Figure S4a). The Fermi level derived from FETmeasurement is described by E nF Fν π= ℏ , where ℏ is theplank constant, νF is the Fermi velocity, and n is the chargecarrier density. The charge carrier density (n) is given by n =αVdirac, where α is 7.2 × 1010 cm−2 V−1 and Vdirac is the Diracvoltage.21 In Figure S4a, the Fermi level of G/SiO2 has a broaddistribution compared to the other devices, which indicatesthat G/SiO2 is under the inhomogeneous charge fluctuation.This spatially fluctuated potential results in suppression of bothelectron and hole conductances, reducing the correspondingmobility (μ) (Figure S4b).38 In addition, we found that thereducing agent enhances the growth of gold particles. In FigureS5, the gold particles were further grown with dendrite shapeby addition of ascorbic acid.To confirm the electrochemical reaction of graphene with

respect to the Fermi level, we assessed the hydrogen evolutionreaction (HER). Graphene transferred on the doping inducingsubstrate was masked by epoxy resin, leaving active area (1 × 1cm2) and connecting as a working electrode to the standardthree electrode system in a 0.5 M H2SO4 electrolyte. Inpolarization curves, the overpotential was lowered inverselyproportional to the height of the Fermi level, in which the n-doped graphene with gold particles had the lowest value out ofthe samples (Figure S6a). This is attributed to the freeelectrons of gold particles that trigger additional n-doping,which facilitates electron transfer from graphene by loweringpotential barrier for hydrogen reduction. Charge transferimpedance behaviors were also consistent with the polarizationplots (Figure S6b). Electrocatalytic performance was notcomparable to recent reports because we focused more on thedoping dependent behaviors rather than overall catalyticperformance. However, we believe that it has a potential tobe improved through combining other functional materials

with the n-doped graphene that serves as a catalytic growthflatform.36,37

In conclusion, we demonstrated the chemical reaction of thegraphene with respect to the Fermi level. The Fermi level wassuccessfully modulated by applying self-assembled monolayerswith different doping agents onto underlying SiO2 substrates.The degree of chemical reactivity on the graphene surface wasdirectly governed by the type of doping. Among the sampleswith different doping types, the n-doped graphene exhibitedthe most remarkable reactivity toward organic molecules (4-NBD) as well as metal ions (AuCl4

−). This indicates thatelectron-rich n-doped graphene plays a pivotal role in thechemical reduction due to increased electron donating effect.Raman and FET analysis demonstrated that n-doped grapheneis more reactive than graphene supported on SiO2 with chargefluctuation, which allowed the selective functionalization.Furthermore, gold particles formed on n-doped grapheneenabled additional n-doping, which was demonstrated by theHER measurement. The n-doped graphene will provide theway to be utilized as a platform for the growth of functionalmaterials. Our study suggests that the CVD graphene withmodulated Fermi level can broaden the range of chemicalapplications by serving as an atomically thin catalytic platformin the 2D material chemistry.

■ METHODSChemicals. Poly(methyl methacrylate) (PMMA), ammonium

persulfate (APS), 3-aminopropyltrimethoxysilane (APTES, −NH2end group), octadecyltrichlorosilane (OTS, −CH3 end group), 4-nitrobenzenediazonium tetrafluoroborate (4-NBD), sodium dodecylsulfate (SDS), and gold(III) chloride trihydrate from Sigma-Aldrich;acetone from Samchun Pure Chemicals; deionized water fromPurescience were used.

CVD Graphene Growth and Transfer. The graphene wassynthesized on 23 um thick copper foil by chemical vapor deposition(CVD) method in the mixture of CH4 (40 sccm) and H2 (4 sccm)

Figure 5. Optical microscopy (OM) and scanning electron microscopy (SEM) images for the reduction of gold ion on the graphene: (1) Brightfield (a, d, g), (2) dark field (b, e, h), and (3) SEM (c, f, i) images. (a−c) are for the G/SiO2. (d−f) are for the G/OTS. (g−i) are for the G/APTES. Scale bar: 50 μm (bright field and dark field images), 10 μm (SEM images), and 2 μm (inset images). (j) Schematic illustration ofextended electron−hole puddle model in the gold ion/graphene reaction. Dirac cone images with different Fermi levels represent the doping typefor each sample. Solid lines indicate the average Fermi level, and solid curves represent the degree of charge fluctuation. Graphene with high Fermilevel in the low charge fluctuation is appropriate to the multielectrons transfer for the reduction of metal ion with high oxidation number. On theother hand, graphene with low Fermi level in the high charge fluctuation is inappropriate for the reduction of gold ions due to the long distance forthe multielectrons transfer.

Chemistry of Materials Article

DOI: 10.1021/acs.chemmater.8b01614Chem. Mater. XXXX, XXX, XXX−XXX

F

gases with vacuum pumping at 1000 °C for 30 min. One side ofgraphene was protected by PMMA coating, and the other side wasetched by reactive ion etcher (RIE). The Cu foil was etched with 80mM of APS solution. After rinsing with distilled water, the graphenewas transferred on target substrates and then dried under 80 °C for 5h.Silanization and Device Fabrication. SiO2/Si substrates were

under UV−ozone treatment to make silanol on the surface.Subsequently, substrates were immersed in APTES and OTS mixedtoluene solution with different ratios (SAM:tolutene), 1:50 forAPTES, and 1:100 for OTS, for 2 h, respectively. The reaction wasconducted in the glovebox. After several rinsing with toluene, PMMAcoated graphene was transferred on silanized substrates, subsequently,PMMA was removed by acetone. For the fabrication of grapheneFET, Cr/Au (5 nm/40 nm) electrodes were deposited on thesubstrate using the prepatterned stencil mask. Channel areas (50 ×250 μm2 (length × width)) were isolated by e-beam lithography.Drain−source current (Ids) was measured at constant drain−sourcevoltage (Vds = 10 mV).Patterning of APTES-SAMs. The SiO2/Si substrate was

patterned by e-beam lithography. Silanization is conducted indeionized water (DW). After several rinses with DW, the e-beamresist is removed by acetone, followed by graphene being transferredon the prepatterned substrate.Diazonium Functionalization. Graphene transferred on differ-

ent substrates were immersed in 4-NBD aqueous solutions (10 mM 4-NBD with 0.5 wt % SDS) at 30 °C for 4 h. Figure 4 was reacted at 50°C for 30 min to compare the reactivity. After several rinses withdistilled water, samples were dried by nitrogen.Growth of Gold Particles on Substrates. Each sample (G/

APTES, G/OTS, G/SiO2) was soaked in gold chloride solution (0.25mM of HAuCl4·3H2O in 20 mL of distilled water) at 30 °C for 1 h.Then, they were rinsed several times in distilled water and dried bynitrogen gas. Ascorbic acid was used as a mild reducing agent to seethe additional reduction effect after the gold particle formation on then-doped graphene.Sample Preparation for the Hydrogen Evolution Reaction.

Graphene transferred on the SAM treated SiO2 substrate wasprepatterned to deposit Cr/Au (5 nm/40 nm) electrode, leavingthe active area (1 × 1 cm2). After electrode deposition, the entireregion except for the active area was masked by epoxy resin (Loctite9340).Characterization. Raman spectra was obtained by Renishaw 2000

(Excitation at 514.5 nm from an Ar ion laser). The power wasmaintained below 2 mW to avoid sample damage during measure-ment. AFM images were obtained by Park System XE-100 withnoncontact mode. Agilent 2602 was used to measure the electricalproperties of the FET devices in the ambient condition. SEM imageswere obtained by SUPRA 55VP field-emission scanning electronmicroscopy (Carl Zeiss). Bright-field and dark-field images werecaptured using optical microscopy (Nikon, Eclipse LV100ND).Electrochemical experiments were performed by an Ivium potentio-stat (Ivium Technologies; Eindhoven, Netherlands, Compact-stat)with a three-electrode system using a Pt mesh as counter electrodeand a saturated calomel reference electrode in a 0.5 M H2SO4standard electrolyte solution. Electrochemical impedance spectrosco-py (EIS) was conducted at −0.25 V vs RHE from 350 kHz to 0.1 Hz.

■ ASSOCIATED CONTENT*S Supporting InformationThe Supporting Information is available free of charge on theACS Publications website at DOI: 10.1021/acs.chemma-ter.8b01614.

Additional experimental details and figures (PDF)

■ AUTHOR INFORMATIONCorresponding Authors*E-mail: byunghee@snu.ac.kr (B.H.H.).

*E-mail: hanson2525@gmail.com (M.J.P.).ORCIDHyukjin Lee: 0000-0001-9478-8473Byung Hee Hong: 0000-0001-8355-8875NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSThis work was supported by the Basic Science ResearchProgram (2012M3A7B4049807) and the KU-KIST SchoolProject.

■ REFERENCES(1) Elias, D. C.; Nair, R. R.; Mohiuddin, T.; Morozov, S.; Blake, P.;Halsall, M.; Ferrari, A.; Boukhvalov, D.; Katsnelson, M.; Geim, A.;Novoselov, K. S. Control of graphene’s properties by reversiblehydrogenation: evidence for graphane. Science 2009, 323, 610−613.(2) Nair, R.; Sepioni, M.; Tsai, I.-L.; Lehtinen, O.; Keinonen, J.;Krasheninnikov, A.; Thomson, T.; Geim, A.; Grigorieva, I. Spin-halfparamagnetism in graphene induced by point defects. Nat. Phys. 2012,8, 199−202.(3) Huang, Y.; Dong, X.; Shi, Y.; Li, C. M.; Li, L.-J.; Chen, P.Nanoelectronic biosensors based on CVD grown graphene. Nanoscale2010, 2, 1485−1488.(4) Jung, Y.; Moon, H. G.; Lim, C.; Choi, K.; Song, H. S.; Bae, S.;Kim, S. M.; Seo, M.; Lee, T.; Lee, S.; et al. Humidity-Tolerant Single-Stranded DNA-Functionalized Graphene Probe for Medical Applica-tions of Exhaled Breath Analysis. Adv. Funct. Mater. 2017, 27,1700068−1700076.(5) Niyogi, S.; Bekyarova, E.; Itkis, M. E.; Zhang, H.; Shepperd, K.;Hicks, J.; Sprinkle, M.; Berger, C.; Lau, C. N.; Deheer, W. A.; et al.Spectroscopy of covalently functionalized graphene. Nano Lett. 2010,10, 4061−4066.(6) Zhang, H.; Bekyarova, E.; Huang, J.-W.; Zhao, Z.; Bao, W.;Wang, F.; Haddon, R. C.; Lau, C. N. Aryl functionalization as a routeto band gap engineering in single layer graphene devices. Nano Lett.2011, 11, 4047−4051.(7) Bekyarova, E.; Itkis, M. E.; Ramesh, P.; Berger, C.; Sprinkle, M.;de Heer, W. A.; Haddon, R. C. Chemical modification of epitaxialgraphene: spontaneous grafting of aryl groups. J. Am. Chem. Soc. 2009,131, 1336−1337.(8) Kasry, A.; Ardakani, A. A.; Tulevski, G. S.; Menges, B.; Copel,M.; Vyklicky, L. Highly efficient fluorescence quenching withgraphene. J. Phys. Chem. C 2012, 116, 2858−2862.(9) Kasry, A.; Afzali, A. A.; Oida, S.; Han, S.-J.; Menges, B.; Tulevski,G. S. Detection of biomolecules via benign surface modification ofgraphene. Chem. Mater. 2011, 23, 4879−4881.(10) Sharma, R.; Baik, J. H.; Perera, C. J.; Strano, M. S. Anomalouslylarge reactivity of single graphene layers and edges toward electrontransfer chemistries. Nano Lett. 2010, 10, 398−405.(11) Paulus, G. L.; Wang, Q. H.; Strano, M. S. Covalent electrontransfer chemistry of graphene with diazonium salts. Acc. Chem. Res.2013, 46, 160−170.(12) Koehler, F. M.; Jacobsen, A.; Ensslin, K.; Stampfer, C.; Stark,W. J. Selective Chemical Modification of Graphene Surfaces:Distinction Between Single-and Bilayer Graphene. Small 2010, 6,1125−1130.(13) Nair, N.; Kim, W.-J.; Usrey, M. L.; Strano, M. S. A Structure−Reactivity relationship for single walled carbon nanotubes reactingwith 4-hydroxybenzene diazonium salt. J. Am. Chem. Soc. 2007, 129,3946−3954.(14) Bissett, M. A.; Konabe, S.; Okada, S.; Tsuji, M.; Ago, H.Enhanced chemical reactivity of graphene induced by mechanicalstrain. ACS Nano 2013, 7, 10335−10343.(15) Wang, Q. H.; Jin, Z.; Kim, K. K.; Hilmer, A. J.; Paulus, G. L.;Shih, C.-J.; Ham, M.-H.; Sanchez-Yamagishi, J. D.; Watanabe, K.;Taniguchi, T.; et al. Understanding and controlling the substrate

Chemistry of Materials Article

DOI: 10.1021/acs.chemmater.8b01614Chem. Mater. XXXX, XXX, XXX−XXX

G

effect on graphene electron-transfer chemistry via reactivity imprintlithography. Nat. Chem. 2012, 4, 724−732.(16) Bard, A. J.; Faulkner, L. R.; Leddy, J.; Zoski, C. G.Electrochemical methods: fundamentals and applications; John Wiley& Sons: New York, 1980.(17) Geim, A. K.; Novoselov, K. S. The rise of graphene. Nat. Mater.2007, 6, 183−191.(18) Xue, J.; Sanchez-Yamagishi, J.; Bulmash, D.; Jacquod, P.;Deshpande, A.; Watanabe, K.; Taniguchi, T.; Jarillo-Herrero, P.;LeRoy, B. J. Scanning tunnelling microscopy and spectroscopy ofultra-flat graphene on hexagonal boron nitride. Nat. Mater. 2011, 10,282−285.(19) Kalbac, M.; Reina-Cecco, A.; Farhat, H.; Kong, J.; Kavan, L.;Dresselhaus, M. S. The influence of strong electron and hole dopingon the Raman intensity of chemical vapor-deposition graphene. ACSNano 2010, 4, 6055−6063.(20) Ryu, S.; Liu, L.; Berciaud, S.; Yu, Y.-J.; Liu, H.; Kim, P.; Flynn,G. W.; Brus, L. E. Atmospheric oxygen binding and hole doping indeformed graphene on a SiO2 substrate. Nano Lett. 2010, 10, 4944−4951.(21) Das, A.; Pisana, S.; Chakraborty, B.; Piscanec, S.; Saha, S.;Waghmare, U.; Novoselov, K.; Krishnamurthy, H.; Geim, A.; Ferrari,A.; Sood, A. K. Monitoring dopants by Raman scattering in anelectrochemically top-gated graphene transistor. Nat. Nanotechnol.2008, 3, 210−215.(22) Lee, W. H.; Park, J.; Kim, Y.; Kim, K. S.; Hong, B. H.; Cho, K.Control of Graphene Field-Effect Transistors by Interfacial Hydro-phobic Self-Assembled Monolayers. Adv. Mater. 2011, 23, 3460−3464.(23) Yan, Z.; Sun, Z.; Lu, W.; Yao, J.; Zhu, Y.; Tour, J. M.Controlled modulation of electronic properties of graphene by self-assembled monolayers on SiO2 substrates. ACS Nano 2011, 5, 1535−1540.(24) Ju, L.; Velasco, J., Jr; Huang, E.; Kahn, S.; Nosiglia, C.; Tsai, H.-Z.; Yang, W.; Taniguchi, T.; Watanabe, K.; Zhang, Y.; et al.Photoinduced doping in heterostructures of graphene and boronnitride. Nat. Nanotechnol. 2014, 9, 348−352.(25) Park, M. J.; Kim, Y.; Kim, Y.; Hong, B. H. Continuous Films ofSelf-Assembled Graphene Quantum Dots for n-Type Doping ofGraphene by UV-Triggered Charge Transfer. Small 2017, 13,1603142.(26) Lee, J. E.; Ahn, G.; Shim, J.; Lee, Y. S.; Ryu, S. Opticalseparation of mechanical strain from charge doping in graphene. Nat.Commun. 2012, 3, 1024.(27) Novoselov, K. S.; Geim, A. K.; Morozov, S. V.; Jiang, D.;Zhang, Y.; Dubonos, S. V.; Grigorieva, I. V.; Firsov, A. A. Electric fieldeffect in atomically thin carbon films. Science 2004, 306, 666−669.(28) Basko, D.; Piscanec, S.; Ferrari, A. Electron-electroninteractions and doping dependence of the two-phonon Ramanintensity in graphene. Phys. Rev. B: Condens. Matter Mater. Phys. 2009,80, 165413.(29) Ni, Z. H.; Yu, T.; Luo, Z. Q.; Wang, Y. Y.; Liu, L.; Wong, C. P.;Miao, J.; Huang, W.; Shen, Z. X. Probing charged impurities insuspended graphene using Raman spectroscopy. ACS Nano 2009, 3,569−574.(30) Berciaud, S.; Li, X.; Htoon, H.; Brus, L. E.; Doorn, S. K.; Heinz,T. F. Intrinsic line shape of the Raman 2D-mode in freestandinggraphene monolayers. Nano Lett. 2013, 13, 3517−3523.(31) Cullen, W. G.; Yamamoto, M.; Burson, K. M.; Chen, J.; Jang,C.; Li, L.; Fuhrer, M. S.; Williams, E. D. High-fidelity conformation ofgraphene to SiO2 topographic features. Phys. Rev. Lett. 2010, 105,215504.(32) Zhang, Y.; Brar, V. W.; Girit, C.; Zettl, A.; Crommie, M. F.Origin of spatial charge inhomogeneity in graphene. Nat. Phys. 2009,5, 722−726.(33) Casiraghi, C. Doping dependence of the Raman peaks intensityof graphene close to the Dirac point. Phys. Rev. B: Condens. MatterMater. Phys. 2009, 80, 233407.

(34) Huh, S.; Park, J.; Kim, K. S.; Hong, B. H.; Kim, S. B. Selectiven-type doping of graphene by photo-patterned gold nanoparticles.ACS Nano 2011, 5, 3639−3644.(35) Zaniewski, A. M.; Trimble, C. J.; Nemanich, R. J. Modifying thechemistry of graphene with substrate selection: A study of goldnanoparticle formation. Appl. Phys. Lett. 2015, 106, 123104.(36) Li, Y.; Wang, H.; Xie, L.; Liang, Y.; Hong, G.; Dai, H. MoS2Nanoparticles Grown on Graphene: An Advanced Catalyst for theHydrogen Evolution Reaction. J. Am. Chem. Soc. 2011, 133, 7296−7299.(37) Huang, X.; Zeng, Z.; Bao, S.; Wang, M.; Qi, X.; Fan, Z.; Zhang,H. Solution-phase epitaxial growth of noble metal nanostructures ondispersible single-layer molybdenum disulfide nanosheets. Nat.Commun. 2013, 4, 1444.(38) Farmer, D. B.; Golizadeh-Mojarad, R.; Perebeinos, V.; Lin, Y.-M.; Tulevski, G. S.; Tsang, J. C.; Avouris, P. Chemical Doping andElectron−Hole Conduction Asymmetry in Graphene Devices. NanoLett. 2009, 9, 388−392.

Chemistry of Materials Article

DOI: 10.1021/acs.chemmater.8b01614Chem. Mater. XXXX, XXX, XXX−XXX

H