Doping Graphene

Tuning the Doping Type and Level of Graphene with Different Gold Confi gurations

Yaping Wu , Wei Jiang , Yujie Ren , Weiwei Cai , Wi Hyoung Lee , Huifeng Li , Richard D. Piner , Cody W. Pope , Yufeng Hao , Hengxing Ji , Junyong Kang , * and Rodney S. Ruoff *

Au nanoparticles and fi lms are deposited onto clean graphene surfaces to study the doping effect of different Au confi gurations. Micro-Raman spectra show that both the doping type and level of graphene can be tuned by fi ne control of the Au deposition. The morphological structures of Au on graphene are imaged by transmission electron microscopy, which indicate a size-dependent electrical characteristic: isolated Au nanoparticles produce n-type doping of graphene, while continuous Au fi lms produce p-type doping. Accordingly, graphene fi eld-effect transistors are fabricated, with the in situ measurements suggesting the tunable conductivity type and level by contacting with different Au confi gurations. For interpreting the experimental observations, the fi rst-principles approach is used to simulate the interaction within graphene–Au systems. The results suggest that, different doping properties of Au–graphene systems are induced by the chemical interactions between graphene and the different Au confi gurations (isolated nanoparticle and continuous fi lm).

1. Introduction

Graphene with its unique electrical, thermal mechanical,

optical barrier, and chemical properties has aroused wide-

spread interest. [ 1–10 ] The quantum Hall effect, [ 2 ] extremely high

carrier mobility, [ 4 ] and tunable work function make graphene

an attractive material for the fabrication of electronic devices.

Accurate control of the electrical transport properties is con-

sidered critical for the development of graphene devices.

© 2012 Wiley-VCH Verlag Gmb

DOI: 10.1002/smll.201200520

Dr. Y. Wu, W. Jiang, Prof. W. Cai, Prof. J. KangDepartment of PhysicsXiamen UniversityXiamen 361005, People’s Republic of China E-mail: jykang@xmu.edu.cn

Dr. Y. Wu, Dr. Y. Ren, Dr. W. H. Lee, Dr. H. Li,Dr. R. D. Piner, C. W. Pope, Dr. Y. Hao, Dr. H. Ji, Prof. R. S. RuoffDepartment of Mechanical Engineering andthe Materials Science and Engineering ProgramThe University of Texas at AustinAustin, TX 78712, USAE-mail: r.ruoff@mail.utexas.edu

small 2012, 8, No. 20, 3129–3136

Surface decoration, using gas molecules or metal adatoms/

nanoparticles, can tune the band structure of graphene for

various electrical applications. [ 11–13 ] For example, several tran-

sition metals with different work functions, such as Ti, Fe, Cu,

Ag, Au, and Pt, [ 14 , 15 ] have been used for graphene band modu-

lation. Both n- and p-type conductivities can be obtained by

selecting metal dopants, and the carrier concentration can be

controlled by the quantity of metal depositied. [ 14 , 16 ] In partic-

ular, band arrangement is not only determined by work func-

tion, but also affected by the interfacial interactions, crucially

for such a thin single-layer graphene. [ 17 ] This means that the

structures of the metal would infl uence the charge transfer

between metal and graphene evidently. Therefore, fi nding

a way to control the interactions and further achieve either

doping type using only one type of metal will be meaning-

fully, especially for gold electrode materials. It would avoid

cross contamination and simplify the process. On the other

hand, the electronic properties and the interactions in the Au–

graphene systems are not fully understood so far, [ 18 , 19 ] e.g.,

the doping effect of Au on graphene is still debated, and both

n- and p-doping of graphene have been reported. [ 14 , 15 , 20–23 ]

Therefore, it is desirable to investigate the doping properties

and interactions between Au and graphene.

3129H & Co. KGaA, Weinheim wileyonlinelibrary.com

Y. Wu et al.

31

full papers

Figure 1 . a) The G- and b) 2D-peak positions in Raman spectra as a function of Au deposition time. The black arrow indicates the D ′ band position at 40 s.

2600 2650 2700 27501500 1550 1600 1650 1700

(b)

Raman shift (cm-1

)

70s

60s

40s

30s

20s

10s

Raman shift (cm-1

)

(a)

0s

Inte

ns

ity

(a

.u.)

70s

60s

40s

30s

20s

10s

0s

In this paper, we deposited Au on the surface of mono-

layer graphene at room temperature. Using Raman measure-

ments, the electrical properties of Au-doped graphene were

investigated, while the structures of Au on the graphene

were imaged at a series of deposition stages using transmis-

sion electron microscopy (TEM). In this way, the electrical

properties of the Au–graphene system were revealed to be

dependent on the size and confi guration of Au. Based on the

above results, fi eld-effect transistors (FET) were prepared,

and in situ measurements were performed in high vacuum

to systematically characterize the electrical transport in Au–

graphene systems. The results suggested that, both the doping

type and carrier density of graphene could be tuned by fi nely

controlling the Au deposition. To understand the experi-

mental observations, the fi rst-principles calculations were car-

ried out, and it was found that the charge transfer, and thus,

the doping type were closely associated with the interactions

between graphene and the different Au confi gurations. This

work may lead to possible applications in graphene-based

electronic devices.

2. Results and Discussion

2.1. Raman Scattering Spectra

For investigating the doping effect of different Au con-

fi gurations, monolayer graphene was synthesized and then

transferred onto 285-nm-SiO 2 /Si substrates. The quality

and number of stacking layers were evaluated by Raman

30 www.small-journal.com © 2012 Wiley-VCH Verlag GmbH & Co. KGaA,

spectroscopy. [ 24 ] The D - peak is almost

undetectable, indicative of the absence

of a signifi cant number of defects. The

2D band, centered at 2676 cm − 1 with a

full-width at half-maximum (FWHM) of

34 cm − 1 , shows a symmetric Lorentz shape,

and the intensity ratio of G/2D is about

0.35. These features are typical of high-

quality, monolayer graphene. [ 25 ]

Various amounts of Au were deposited

on the graphene surface to investigate

the size-dependent interaction between

Au and graphene. Raman spectroscopy is

highly sensitive to the electronic structure

of materials, and we used it to examine

the correlation between Au deposition

and doping properties. It is an effective

method for distinguishing the doping type

by comparing the positions of the G- and

2D-peaks between doped and pristine

graphene. [ 26 ] Because the SiO 2 layer under

the graphene fi lm is insulating, its effect on

the electron distribution of graphene is not

considered here. The evolution of Raman

spectrum for graphene with increasing

Au amount is displayed in Figure 1 , with

a comparison to the as-grown graphene.

When increasingly decorated by a small

amount of Au, the G - peak blue-shifts gradually from 1582

to 1587 cm − 1 while the 2D - peak red-shifts from 2676 to

2660 cm − 1 as a function of the increasing deposition time

(0 to ∼40 s), exhibiting typical n-type doping in the low-cov-

erage deposition regime. Reverse shifts from 1587 to 1578 cm − 1

for the G-peak and from 2660 to 2672 cm − 1 for the 2D-peak are

observed after the deposition time further increases, resulting

from a reduction of the n-doped effi ciency. As the deposition

time increases to 70 s, both the G- and 2D-peak show slight

blue-shifts to 1585 and 2682 cm − 1 , respectively, suggesting a

p-type doping. These Raman results indicate the controllable

n-type and p-type doping by depositing differing amounts of

Au. In addition, the defect-induced D ′ band at ∼ 1620 cm − 1

emerges as a shoulder peak of G band at 20 s [ 29 , 30 ] and obvi-

ously increases at 40 s (denoted by the black arrow in Figure

1 a), then slightly decreases at 60 and 70 s. Considering pre-

vious theoretical predictions, [ 17 ] we suggest that the evolve-

ment of the D ′ band in our Raman spectra results from the

defects created by the chemical interaction which breaks

the translational symmetry of the graphene lattice. It is also

responsible for charge-induced renormalization of the elec-

tronic and vibrational energies, leading to the broadening

of the 2D band. Consequently, the formation of a metal–

graphene interface not only results in a redistribution of the

electrons, but also to a metal–graphene chemical interaction.

2.2. TEM Morphology

Bright-fi eld TEM images of graphene/Au, [ 29 ] corresponding

to the deposition times of 10, 20, 30, 40, 60, and 70 s are

Weinheim small 2012, 8, No. 20, 3129–3136

Tuning the Doping Type and Level of Graphene with Different Gold Confi gurations

Figure 2 . a–f) Bright-fi eld TEM images of Au nanoparticles and fi lms on graphene for the deposition times of 10, 20, 30, 40, 60, and 70 s. The inset in (a) is a high-resolution TEM image of a single Au nanoparticle, which shows the lattice structure of the (111) crystal plane of Au.

shown in Figure 2 . With short time depositions (Figure 2 a–c),

the Au atoms assembled into isolated hexagonal nanoparti-

cles, [ 30 ] distributing on the graphene surfaces with relatively

uniform size and shape. The nanoparticles were imaged in

high-resolution TEM (inset of Figure 2 a). The sixfold sym-

metrical lattices have a fringe spacing of about 0.24 nm, thus

identifying the (111) crystal planes of Au, and showing the

high crystallinity of the Au nanoparticles. The diameters of

the nanoparticles are about 20, 27, and 32 nm, with the depo-

sition times of 10, 20, and 30 s, respectively. As the deposition

time is further increased to 40 s, where the Raman spectros-

copy shift reaches a maximum, the nanoparticles are about

45 nm in size and the morphology has evidently changed to

one without a uniform hexagonal shape (Figure 2 d). Fur-

ther deposition coalesces the Au nanoparticles (Figure 2 e),

leading to a large increase in their size. The G and 2D bands

back-shift (deposition time = 60 s) from the critical values

in the Raman spectra, suggesting a gradual reduction in the

level of n-type doping as the Au nanoparticles are coalescing.

Opposite doping occurs when a thin continuous Au fi lm

starts to form (Figure 2 f); the corresponding G and 2D bands

in the Raman spectrum shift to 1585 and 2682 cm − 1 , respec-

tively (Figure 1 ). The Raman results, in conjunction with the

TEM images, demonstrate that the doping type of graphene

is clearly related to the confi guration of Au deposited on its

surface.

2.3. FET Measurements

Raman spectra and TEM images comfi rm that Au nanopar-

ticles and continuous fi lms produce n- and p-type doping

in graphene, respectively. This provides strong motivation

© 2012 Wiley-VCH Verlag Gmbsmall 2012, 8, No. 20, 3129–3136

to fabricate doping-type-tunable graphene FET devices by

controlling the Au deposition on the channel. An optical

image of a graphene FET device is shown in Figure 3 a and

b, and the schematic diagram of the device construction is

shown in Figure 3 d. The uniform color contrast of graphene

indicates uniform thickness. During the in situ FET tests, a

1.0 V direct current was applied between the source and

drain, and a variable gate bias, V g , ranging from − 100 to

+ 100 V was applied on the bottom of the SiO 2 /Si substrate.

Sequentially, the drain–source current ( I ds ) was monitored as

a function of V g through an electrical feedthrough. As-grown

graphene shows p-type behavior in the FET test due to the

pre-existing adsorbates (Figure 3 c). The I ds increases slowly

with decreasing V g . The minimum in the gate-dependent con-

ductivity defi nes the position of the Dirac point, which does

not show up within the measured range. After degassing, the

Dirac point moves to near the zero gate voltage, and the elec-

tron and hole conductions show much better symmetry in our

graphene FET devices. The slope corresponds to the mobility

( μ ) of charge carriers, which are deduced to be about 1400

and 2000 cm 2 V − 1 s − 1 for electrons and holes, respectively,

based on the equation: [ 15 ]

μ = L/(WCgVds)(�Ids/�Vg), (1)

where L and W are the channel length and width, respec-

tively, C g is gate capacitance per unit area, and V ds is drain–

source bias.

Using the same process, three FET devices A, B, and C

were prepared for electrical transport measurements. For

device A, we used a smaller heating current and longer dep-

osition distance to obtain small isolated Au nanoparticles

on the graphene channel (Figure 3 e). Figure 4 a shows the

3131www.small-journal.comH & Co. KGaA, Weinheim

Y. Wu et al.

3132

full papers

Figure 3 . a) Optical image of graphene FET device used for electrical transport measurements, with a channel of 5 mm in width and 1 mm in length. b) Optical microscopy image of monolayer graphene taken from the channel region shown in (a). c) I ds as a function of V g taken in high vacuum ( ∼ 1.0 × 10 − 8 mbar). d) Schematic diagram of the FET device constuction. e,f) Schematic diagrams of the FET devices with Au nanopaticles (e) and Au fi lms (f) on the channels.

gate-dependent drain–source current for various deposition

times of Au. The Dirac point shifts toward negative gate volt-

ages from 0 to –6.0 V as the deposition time increases from 0

to 25 s. This is because the Fermi level moves away from the

conical Dirac point for the case of graphene contacted with Au

nanoparticles, resulting in an effective increase of electron den-

sity and hence the n-type doping. For device B, a larger heating

current was used to deposit continuous Au fi lms onto the

graphene channel (Figure 3 f). The evolution of the I ds – V g curve

at a series of deposition times is shown in Figure 4 b. In contrast

www.small-journal.com © 2012 Wiley-VCH Ve

Figure 4 . a,b) The V g dependent I ds at various Au deposition times for devic,d) The deposition-time-dependent hole and electron mobilities for devic

-20 -10 0 10 205.0x10

-3

5.2x10-3

5.4x10-3

5.6x10-3

5.8x10-3

6.0x10-3

-10 05.9x10

-3

6.0x10-3

6.1x10-3

6.2x10-3

6.3x10-3

6.4x10-3

6.5x10-3

0 5 10 15 20 25

1800

1850

1900

1950

0 5 101070

1080

1090

1100

1110

1120(d)(c)

(b)

I ds (

A)

Vg (V)

0s

5s

10s

15s

20s

25s

(a)

I ds (

A)

V

µh

ole

(cm

2V

-1s-1

)

Time (s)

µele

c(c

m2V

-1s-1

)

Ti

to the case of Au nanoparticles on device A, the continuous

Au fi lm, acting as an acceptor on device B, produces p-type

doping in the graphene. As a result, the Dirac point shifts from

0 to + 11 V as the deposition time increases from 0 to 25 s. The

evolution of the hole and electron mobilities for devices A and

B are ploted in Figure 4 c and d, respectively. It is clearly seen

that the hole mobility of graphene gradually decreases from

1960 to 1796 cm 2 V − 1 s − 1 after p-doping whereas the electron

mobility reduces from 1118 to 1075 cm 2 V − 1 s − 1 after n-doping.

The slight suppression of carrier mobility results from the addi-

rlag GmbH & Co. KGaA,

ces A (a) and B (b). es A (c) and B (d).

10 20

15 20 25

g (V)

0s

5s

10s

15s

20s

25s

me (s)

tional impurity scattering caused by the Au

deposition. [ 31 ]

Based on the opposite doping types

of devices A and B, device C was gradu-

ally deposited with Au nanoparticles until

a continuous Au fi lm was formed, using

the same condition as that for the sample

in Raman measurements. The amount of

charge transfer was tuned by controlling

the deposition time, and the shift of the

I ds – V g curve was systematically monitored

by the corresponding FET transport meas-

urements after each deposition. The device

was found to be doped stepwise from n- to

p-type, as shown in Figure 5 a; the Dirac

point fi rst shifts to –4, –12, –17, –22, and

–26 V, and then gradually shifts back to

–25, –24, –20, –16, –8 V, passes through 0

V, and then to + 2 and + 6 V. The maximum

of Dirac point shift corresponds to the fre-

quency shift in the Raman spectroscopy

at time, t = 40 s, likewise the signifi cant

change of Au morphology can be observed

in the TEM image at this time mentioned

above. Hole and electron mobilities based

on the deposition time for device C are

Weinheim small 2012, 8, No. 20, 3129–3136

Tuning the Doping Type and Level of Graphene with Different Gold Confi gurations

Figure 5 . a) The V g dependent I ds at various Au deposition times for device C. b) The deposition-time-dependent carrier densities for devices A, B, and C. c,d) The deposition-time-dependent hole (c) and electron mobilitiy (d) for device C.

9.0x10-4

1.0x10-3

1.1x10-3

1.2x10-3

-60 -40 -20 0 209.0x10

-4

1.0x10-3

1.1x10-3

0 10 20 30 40 50 60 70 80

-1.0x1012

-5.0x1011

0.0

5.0x1011

1.0x1012

1.5x1012

2.0x1012

0 10 20 30 40115

120

125

130

135

40 45 50 55 60 65 70 75110

115

120

125

130

135

(a)

(d)(c)

(b)

I ds

(A

)

0s

5s

10s

20s

30s

40s

I ds

(A

)

Vg(V)

40s

45s

50s

55s

60s

65s

70s

75s

Device A

Device B

Device C

Ele

ctr

on

de

ns

ity

(c

m-2

)

Deposition time (s)

µh

ole

(cm

2V

-1s-1

)

Time (s)

µele

c(c

m2

V-1

s-1

)

Time (s)

ploted in Figure 5 c and d, respectively. Although the mobili-

ties of device C are much lower than that of devices A and B,

the evolution of the hole and electron mobilities are similar

with the two devices that, electron mobility decreases after

p-doping while hole mobility decreases after n-doping.

According to the Dirac point shifts, the carrier density

can be estimated from the equation:

n = CgVDirac/e, (2)

where V Dirac is the Dirac point shift relative to that of

pristine graphene, and e is the electron charge. The carrier

densities as a function of deposition time for devices A, B,

© 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Figure 6 . a) Top and b) side views of the atomic structures of graphene–Au nanoparticle system. c) Top and d) side views of the atomic structures of graphene–Au fi lm. The gray balls represent the carbon atoms in graphene, while the yellow, green, and purple balls represent the fi rst, second, and third layer, respectively, of the Au atoms adsorbed on top of the graphene, with the [111] direction perpendicular to the graphene surface.

small 2012, 8, No. 20, 3129–3136

and C are shown in Figure 5 b. The induced

electron concentration is estimated at

–1 × 10 12 to ∼ 2 × 10 12 cm − 2 . Minus values

indicate a concentration of hole carriers

rather than of electron carriers. Based

on the expression of the change in Fermi

energy: [ 26 ]

�EF(n) = h̄|vF|√

πn, (3)

where the Fermi velocity | v F | is 1.1 ×

10 6 ms − 1 ; the Fermi energy in graphene

can be adjusted between –0.09 and ∼ 0.12

eV, with the sign depending on downshift

or upshift. These electrical transport prop-

erties allow us to tune the Fermi level of

graphene from below to above the Dirac

point; that is, the conductivity type and

carrier density can be selected by control-

ling the Au deposition.

2.4. First-Principles Calculations

To interpret the doping effect of Au nano-

particles/fi lm on graphene, the interac-

tions between graphene and different

confi gurations of Au adsorbates were simulated by the

fi rst-principles calculations. Based on the optimized models

in Figure 6 , the equilibrium separations between Au and

graphene, the binding energies, and the work functions were

calculated, as listed in Table 1 . Compared with the graphene–

Au fi lm, the graphene–Au nanoparticle system has smaller

equilibrium separation and larger binding energy, which also

implies stronger chemical interactions. [ 17 ]

Using the calculated work functions and equilibrium dis-

tances in Table 1 , the band alignments of both contact models

are ploted and shown in Figure 7 a and b, with the vacuum

level and the orginal work function (or Fermi energy) of Au

W Au fi xed as a reference. In the graphene–Au particles contact

system, the difference between W G-Au particle

and W Au defi nes a potential change Δ V of

0.83 eV, which is mainly generated by the

interfacial chemical interaction. [ 17 ] Com-

pared with the orginal work function of

graphene W G , the value W G-Au particle is

lower for about 0.42 eV ( W G-Au particle – W G =

–0.42 eV), leading to an upshift ( Δ E F ) for

the graphene Fermi energy. Hence, the

graphene is n-doped by Au particle. As

graphene contacts with Au fi lms, Δ V is only

0.21 eV deduced from the smaller differ-

ence between W G-Au fi lm and W Au . A 0.20 eV

increase of W G-Au fi rm than W G ( W G-Au fi lm –

W G = 0.20 eV) determines a downshift of

graphene Fermi energy, and deminstrates

that the graphene is p-doped by Au fi lm. It

could be found from the two illustrations

that the chemical interaction induced Δ V

3133www.small-journal.com

Y. Wu et al.

31

full papers Table 1. Calculated equilibrium separations, binding energies, and work functions of graphene (G)–Au nanoparticle/fi lm systems.

G Au G–Au nanoparticle

G–Au fi lm

Equilibrium

separation d [Å]

3.10 3.58

Binding energy Δ E [eV] 4.690 5.102 0.036 0.015

Work function W [eV] 4.275 4.893

plays a negative role on the work-function-difference-induced

electron transfer between Au and graphene. As mentioned

above, graphene has stronger chemical interactions with the

Au particles, which creates a much higher Δ V in the graphene–

Au particle model than in the graphene–Au fi lm, and causes

these two systems to have opposite doping types.

The directions of charge transfer is also confi rmed by the

effective potentials of the whole systems. The effective poten-

tials between graphene and Au comprises three terms: [ 32 ]

υeff (r ) = −∫

dr ′ n+(r ′)|r − r ′| +

∫dr ′ n−(r ′)

|r − r ′| + υxc(r ), (4)

where r is the atomic radius, the fi rst two terms represent the

electrostatic potentials caused by the positive background

and by the electronic density, and v xc (r) is the exchange-cor-

relation potential. Figure 7 c and d show the average effective

potentials along the [0001] direction for the graphene–Au

nanoparticle and fi lm confi gurations, respectively. The

34 www.small-journal.com © 2012 Wiley-VCH Ve

Figure 7 . a,b) The band alignment illustrations of graphene–Au nangraphene–Au fi lm (b) systems, using the calculated work functions and eqlisted in Table 1 . The vacuum level and the orginal work function of Au (reference. c,d) Average effective potentials in graphene–Au nanoparticle (cfi lm (d) systems along graphene [0001] direction.

macroscopic averages of each layer were performed using

macroscopic average techniques, which allows a general

description of the effects of interfacial perturbations. [ 33 , 34 ]

In each curve, the left valley represents the potential of Au,

while the right valley represents the potential of graphene.

As shown in Figure 7 c, the effective potential is inclined in

the graphene. Driving from the potential difference, the elec-

trons transfer from the Au nanoparticles to graphene, which

creates a built-in electric fi eld in the same direction. The

donated electrons will raise the Fermi level, inducing n-type

doping in the graphene, which agrees well with the band illus-

tration of Figure 7 a. In contrast, graphene exhibits a relatively

higher effective potential when contacting Au fi lm, as shown

in Figure 7 d. The electrons are inclined to transfer from

graphene to the Au fi lm, resulting in a built-in electric fi eld

directed to the Au from graphene. The Fermi level will down-

shift due to the injected holes, which produces p-type doping

in the graphene, and is consistent with the band structure in

Figure 7 b. These theoretical calculations reveal a close rela-

tionship between the doping type and the interactions within

the graphene–Au systems; qualitatively our experimental

results could be interpreted as graphene being n-doped by

isolated Au nanoparticles and p-doped by a continuous fi lm.

3. Conclusion

Au nanoparticles and fi lms were deposited onto graphene

surfaces at room temperature. Raman spectra indicated that

graphene was n- and p-doped by low and high coverage of

rlag GmbH & Co. KGaA

oparticle (a) and uilibrium distances W Au ) are fi xed as a ) and graphene–Au

Au, respectively. The morphology of Au on

graphene imaged by TEM suggested that

isolated Au nanoparticles yielded n-type

doping while continuous Au fi lms resulted

in p-type doping. Accordingly, graphene

FETs were fabricated, in which both the

doping type and carrier density could

be tuned by fi nely controlling the Au

deposition. The interactions between the

graphene and Au nanoparticle/fi lm were

simulated by fi rst-principles calculations.

The results revealed that the different

doping properties observed in the experi-

ments were derived from the effective

potential difference related to the interfa-

cial interaction within graphene–Au nano-

particle and fi lm systems. This work is of

signifi cant importance for graphene band

modulation and device design.

4. Experimental Section

The fi rst-principles calculations were performed in the framework of density-functional theory (DFT), [ 35 , 36 ] within the generalized gradient approximation (GGA), using the Vienna Ab-initio Simulation Package (VASP). [ 37 ] Calculation models were periodically repeated by an atomic slab consisting of one graphene monolayer

, Weinheim small 2012, 8, No. 20, 3129–3136

Tuning the Doping Type and Level of Graphene with Different Gold Confi gurations

(128 carbon atoms, lattice constant of 2.46 Å), with a Au nanoparticle or one continuous fi lm located on the top, as shown schematically in Figure 6 a and c, respectively. The triangular lattice of Au(111) surface matches the honeycomb lattice of graphene in the unit cells, which was proven to be the most stable confi guration of graphene with Au contact. [ 17 ] An 18 Å thick vacuum layer was applied for separating the slabs to form a surface. The lateral distance between neighboring par-ticles was about 19.8 Å, which was large enough to avoid the overlap of their electronic states in the periodically expanded model. Au 5d6s was treated explicitly as valence. [ 38 , 39 ] Ions and electrons interactions were approximated by ultrasoft pseudopotentials, and the plane-wave cut-off energy was set to 330 eV. A (8 × 8 × 1) Monkhorst–Pack mesh was used to sample the Brillouin zone. A conjugate gradient algorithm was used to optimize the structures by relaxing all the atomic geometries, [ 40 ] with forces on all the atoms converged to within 0.01 eV/Å. Although the part of van der Waals forces between Au atoms and graphene are not involved in DFT calculation, there should be some direct hybridization in our models, espectially between Au 6s orbital and graphene states, because the lowest layer of Au atoms mostly locate on the T site of graphene. [ 17 , 41 ]

Large-area monolayer graphene was grown by chemical vapor deposition on 25- μ m-thick Cu foils (Alfa Aesar, item No. 13382), [ 42 ] and transferred onto 285-nm-SiO 2 /Si substrates ( p + doping, ρ = ∼ 0.002–0.005 Ω cm) using a process similar to that described pre-viously. [ 43 ] We also transferred the as-grown graphene onto 300-mesh TEM copper grids (with lacey carbon supporting fi lms, Ted Pella, Inc.), using the same transfer process. Au deposition was performed in a home-built vacuum system with a base pressure of 1.0 × 10 − 8 mbar. A high-purity Au (99.9999%) source was affi xed to a tungsten fi lament in the upper part of the vacuum chamber, and Au could be evaporated by direct current resistive heating to a temperature around 1000 ° C. By adjusting the heating current and the distance between graphene and the Au source, the deposition rate, and hence the coverage and size of Au nanoparticles could be controlled. The evolution of Raman spectrum for different deposi-tion states was detected by a confocal microprobe Raman system (WITec Alpha300, 532 nm excitation laser, ∼ 6 mW). [ 25 , 27 ] The cor-responding bright-fi eld TEM images were obtained on a JEOL2010F TEM operating at an accelerating voltage of 200 keV.

For fabricating the graphene FET devices, ∼ 500 nm thick Au fi lms were deposited on the graphene surface as source and drain electrodes, and consequently defi ned a transport channel of 5 mm in width and 1 mm in length between the two electrodes. The FET transport measurements were performed in the vacuum chamber where the graphene FET device was attached to a ceramic heater with a temperature tunable up to 150 ° C. Before the deposition and tests, the devices were degassed for 2 h by in situ annealing at 100 ° C to eliminate the pre-existing adsorbates (e.g., H 2 O, O 2 and/or other molecules), and then cooled down to room temperature. Using a similar deposition process, three FET devices A, B, and C were pre-pared with controlled heating currents and deposition distances.

Acknowledgements

We appreciate support from the National “973” Program of China (Grant Nos. 2012CB619301 and 2011CB925600), the National

© 2012 Wiley-VCH Verlag Gmbsmall 2012, 8, No. 20, 3129–3136

[ 1 ] K. S. Novoselov , A. K. Geim , S. V. Morozov , D. Jiang , Y. Zhang , S. V. Dubonos , I. V. Grigorieva , A. A. Firsov , Science 2004 , 306 , 666 .

[ 2 ] Y. B. Zhang , Y. W. Tan , H. L. Stormer , P. Kim , Nature 2005 , 438 , 201 . [ 3 ] K. S. Novoselov , A. K. Geim , S. V. Morozov , D. Jiang ,

M. I. Katsnelson , I. V. Grigorieva , S. V. Dubonos , A. A. Firsov , Nature 2005 , 438 , 197 .

[ 4 ] X. Du , I. Skachko , A. Barker , E. Y. Andrei , Nat. Nanotechnol. 2008 , 3 , 491 .

[ 5 ] J. S. Bunch , A. M. van der Zande , S. S. Verbridge , I. W. Frank , D. M. Tanenbaum , J. M. Parpia , H. G. Craighead , P. L. McEuen , Sci-ence 2007 , 315 , 490 .

[ 6 ] B. Standley , W. Bao , H. Zhang , J. Bruck , C. N. Lau , M. Bockrath , Nano Lett. 2008 , 8 , 3345 .

[ 7 ] J. R. Williams , L. DiCarlo , C. M. Marcus , Science 2007 , 317 , 638 . [ 8 ] M. D. Stoller , S. Park , Y. Zhu , J. An , R. S. Ruoff , Nano Lett. 2008 , 8 ,

3498 . [ 9 ] X. Huang , Z. Yin , S. Wu , X. Qi , Q. He , Q. Zhang , Q. Yan , F. Boey ,

H. Zhang , Small 2011 , 7 , 1876 . [ 10 ] S. Watcharotone , D. A. Dikin , S. Stankovich , R. Piner , I. Jung ,

G. H. B. Dommett , G. Evmenenko , S.-E. Wu , S.-F. Chen , C.-P. Liu , S. T. Nguyen , R. S. Ruoff , Nano Lett. 2007 , 7 , 1888 .

[ 11 ] F. Yavari , C. Kritzinger , C. Gaire , L. Song , H. Gullapalli , T. Borca-Tasciuc , P. M. Ajayan , N. Koratkar , Small 2010 , 6 , 2535 .

[ 12 ] F. Schedin , A. K. Geim , S. V. Morozov , E. W. Hill , P. Blake , M. I. Katsnelson , K. S. Novoselov , Nat. Mater. 2007 , 6 , 652 .

[ 13 ] O. Leenaerts , B. Partoens , F. M. Peeters , Phys. Rev. B 2008 , 77 , 125416 .

[ 14 ] K. Pi , K. M. McCreary , W. Bao , W. Han , Y. F. Chiang , Y. Li , S. W. Tsai , C. N. Lau , R. K. Kawakami , Phys. Rev. B 2009 , 80 , 075406 .

[ 15 ] Y. Ren , S. Chen , W. Cai , Y. Zhu , C. Zhu , R. S. Ruoff , Appl. Phys. Lett. 2010 , 97 , 053107 .

[ 16 ] S. Xiao , J.-H. Chen , S. Adam , E. D. Williams , M. S. Fuhrer , Phys. Rev. B 2010 , 82 , 041406 .

[ 17 ] G. Giovannetti , P. A. Khomyakov , G. Brocks , V. M. Karpan , J. van den Brink , P. J. Kelly , Phys. Rev. Lett. 2008 , 101 , 026803 .

[ 18 ] R. Zan , U. Bangert , Q. Ramasse , K. S. Novoselov , Small 2011 , 7 , 2868 .

[ 19 ] P. A. Pandey , G. R. Bell , J. P. Rourke , A. M. Sanchez , M. D. Elkin , B. J. Hickey , N. R. Wilson , Small 2011 , 7 , 3202 .

[ 20 ] K. S. Subrahmanyam , A. K. Manna , S. K. Pati , C. N. R. Rao , Chem. Phys. Lett. 2010 , 497 , 70 .

[ 21 ] K. M. McCreary , K. Pi , A. G. Swartz , W. Han , W. Bao , C. N. Lau , F. Guinea , M. I. Katsnelson , R. K. Kawakami , Phys. Rev. B 2010 , 81 , 115453 .

[ 22 ] S. Huh , J. Park , K. S. Kim , B. H. Hong , S. Bin Kim , ACS Nano 2011 , 5 , 3639 .

[ 23 ] E. J. H. Lee , K. Balasubramanian , R. T. Weitz , M. Burghard , K. Kern , Nat. Nanotechnol. 2008 , 3 , 486 .

[ 24 ] H. Yufeng , W. Yingying , W. Lei , N. Zhenhua , W. Ziqian , W. Rui , K. Chee Keong , S. Zexiang , J. T. L. Thong , Small 2010 , 6 , 195 .

[ 25 ] A. C. Ferrari , J. C. Meyer , V. Scardaci , C. Casiraghi , M. Lazzeri , F. Mauri , S. Piscanec , D. Jiang , K. S. Novoselov , S. Roth , A. K. Geim , Phys. Rev. Lett. 2006 , 97 , 187401 .

[ 26 ] A. Das , S. Pisana , B. Chakraborty , S. Piscanec , S. K. Saha , U. V. Waghmare , K. S. Novoselov , H. R. Krishnamurthy , A. K. Geim , A. C. Ferrari , Nat. Nanotechnol. 2008 , 3 , 210 .

[ 27 ] L. M. Malard , M. A. Pimenta , G. Dresselhaus , M. S. Dresselhaus , Phys. Rep. 2009 , 473 , 51 .

[ 28 ] S. Ryu , J. Maultzsch , M. Y. Han , P. Kim , L. E. Brus , ACS Nano 2011 , 5 , 4123 .

[ 29 ] G. Yanjie , S. Litao , F. Banhart , Small 2008 , 4 , 587 .

Natural Science Foundation of China (Grant Nos. 90921002 and 60827004), and the National Science Foundation (Grant No. 1006350), the Offi ce of Naval Research, SWAN NRI.

3135www.small-journal.comH & Co. KGaA, Weinheim

Y. Wu et al.

313

full papers

[ 30 ] R. Anton , P. Kreutzer , Phys. Rev. B 2000 , 61 , 16077 . [ 31 ] D . B. Farmer , R. Golizadeh-Mojarad , V. Perebeinos , Y. Lin ,

G. S. Tulevski , J. C. Tsang , P. Avouris , Nano Lett. 2008 , 9 , 399 .

[ 32 ] M. Quijada, R. D. Muiño , P. M. Echenique , Nanotechnology 2005 , 16 , S176 .

[ 33 ] M. Peressi , N. Binggeli , A. Baldereschi , J. Phys. D: Appl. Phys. 1998 , 31 , 1273 .

[ 34 ] X. G. Lin , J. C. Li , J. Y. Kang , Semicond. Sci. Technol. 2010 , 25 , 045012 .

[ 35 ] G. Kresse , D. Joubert , Phys. Rev. B 1999 , 59 , 1758 . [ 36 ] G. Kresse , J. Furthmuller , Phys. Rev. B 1996 , 54 , 11169 . [ 37 ] Y. Wu , J. Kang , F. Liu , J. Mater. Res. 2008 , 23 , 3347 . [ 38 ] Y. Wu , Y. Zhou , C. Zhou , H. Zhan , J. Kang , J. Chem. Phys. 2010 ,

133 , 124706 .

6 www.small-journal.com © 2012 Wiley-VCH Ve

[ 39 ] A. V. Krasheninnikov , P. O. Lehtinen , A. S. Foster , P. Pyykko , R. M. Nieminen , Phys. Rev. Lett. 2009 , 102 , 126807 .

[ 40 ] M. C. Payne , M. P. Teter , D. C. Allan , T. A. Arias , J. D. Joannopoulos , Rev. Mod. Phys. 1992 , 64 , 1045 .

[ 41 ] K . T. Chan , J. B. Neaton , M. L. Cohen , Phys. Rev. B 2008 , 77 , 235430 .

[ 42 ] X. Li , W. Cai , J. An , S. Kim , J. Nah , D. Yang , R. Piner , A. Velamakanni , I. Jung , E. Tutuc , S. K. Banerjee , L. Colombo , R. S. Ruoff , Science 2009 , 324 , 1312 .

[ 43 ] X. Li, X. Li , Y. Zhu , W. Cai , M. Borysiak , B. Han , D. Chen , R. D. Piner , L. Colombo , R. S. Ruoff , Nano Lett. 2009 , 9 , 4359 .

Received: March 8, 2012 Revised: June 11, 2012Published online: July 24, 2012

rlag GmbH & Co. KGaA, Weinheim small 2012, 8, No. 20, 3129–3136