We report on a method for “direct-write” conductive patterning via reduction of graphene oxide (GO) sheets using focused electron beam induced deposition (FEBID) of carbon. FEBID treatment of the intrinsically dielectric graphene oxide between two metal terminals opens up the conduction channel, thus enabling a unique capability for nanoscale conductive domain patterning in GO. An increase in FEBID electron dose results in a significant increase of the domain electrical conductivity with improving linearity of drain-source current vs. voltage dependence, indicative of a change of graphene oxide electronic properties from insulating to semiconducting. Density functional theory calculations suggest a possible mechanism underlying this experimentally observed phenomenon, as localized reduction of graphene oxide layers via interactions with highly reactive intermediates of electron-beam-assisted dissociation of surface-adsorbed hydrocarbon molecules. These findings establish an unusual route for using FEBID as nanoscale lithography and patterning technique for engineering carbon-based nanomaterials and devices with locally tailored electronic properties.

Graphene oxide (GO) is one of popular precursors for graphene-based devices and nanocomposite materials.1 It features the 2D graphitic structure with surface functionalities, such as epoxy and hydroxyl functional groups on the basal plane with a carboxyl group at the edges.2,3 It is a promising material, which can be utilized for various applications owing to its unique electronic/mechanical/chemical properties.1–3 Also, unlike other methods for obtaining graphene materials such as mechanically exfoliation or chemical vapor deposition (CVD), graphene oxide can be used as a primary precursor source for high-yield, low cost production of a large area graphene film via thermal, chemical, or electrochemical reduction, resulting in an increase of conductivity by transforming its electrical characteristic from insulating to semiconducting behaviors.1–5 

Conventional methodologies for reduction of graphene oxide are chemical treatments or thermal process.6 Although they have been widely accepted for reduction of bulk graphene oxide sheets, they are time-consuming and require hazardous chemicals such as borohydride7 and hydrazine8 or high temperature conditions,9 which are not environment-friendly and can be detrimental to electronic devices. Furthermore and perhaps more importantly, none of these batch techniques is suitable for patterned reduction of graphene oxide to “direct-write” high conductivity domains in the GO substrate. For chemical-free, patterned reduction of graphene oxide, electrochemical microstamping method,5 and an AFM tip-based thermochemical nanolithography10 have been developed. Especially, the tip-based thermochemical nanolithography enables nanoscale reduction of graphene oxide, resulting in improvements of its electrical conductivity. While this technique can be promising for controllable patterned reduction of graphene oxide sheets, it requires a custom experimental setup of AFM with heating of the tip over 600 °C and could also transfer undesired contaminations from the tip upon direct physical contact with a substrate.

In this work, we demonstrate a method to locally reduce graphene oxide sheets to form high electrical conductivity domains on nanoscale, using focused electron beam induced deposition (FEBID). The unique feature of our approach stems from the fact that FEBID is a non-toxic, chemical-free, low-temperature “direct-write” nanolithography technique, which allows fabrication of topologically complex nanostructures with a high degree of controls and resolution down to the diameter of a primary electron beam for suspended substrates (down to 1 nm).11–14 FEBID process involves interactions of primary electrons and adsorbed precursor molecules with a supporting dielectric or conductive substrate, affecting the deposition outcomes.12,15,16 Additionally, electron beam irradiation of graphene was shown to modify electronic properties of graphene (e.g., shift the Dirac point)17 as well as its chemical properties (e.g., strength of interactions with adsorbed species) by introducing structural defects into the 2D material structure.15 Therefore, there is an intriguing opportunity to modify the electronic properties of graphene oxide via application of a FEBID technique to enable the localized, high-resolution, controllable patterning for realization of interesting electronic functionalities. FEBID process can be easily implemented using an electron beam source in a scanning electron microscope (SEM) chamber. Thus, there is no need for specialized instrumentation and/or heterogeneous processing steps, and the entire device fabrication and characterization/imaging sequence can be accomplished within the same processing environment.18 Motivated by these compelling advantages, we have developed the FEBID process to form molecularly thin carbon film from a hydrocarbon precursor source on graphene oxide and explored the resulting electrical characteristics of graphene oxide modified via carbon deposition.

Graphene oxide was synthesized from natural graphite flakes (325 mesh, 99.8% metal basis) purchased from Alfa Aesar, using the modified Hummer's method.4 Stable dispersion of graphene oxide in a solution mixture of methanol:water (5:1) was subjected to ultrasonication for 15 min, followed by centrifugation at 3000 rpm. Then, the graphene oxide sheets were transferred onto a 300 nm SiO2/Si substrate using Langmuir-Blodgett (LB) method at a room temperature, using a KSV 2000 LB minitrough.4 

Figure 1(a) shows the sequence of experimental steps used for this study. Lithographically defined drain-source metal contacts were fabricated on the deposited graphene oxide sheets supported by a SiO2/Si substrate. FEBID of hydrocarbon precursor was implemented using a FEI Quanta 200 Environmental SEM. The center of the graphene oxide sheets was irradiated by a 25 keV focused electron beam. Figures 1(b) and 1(c) show SEM images of the fabricated graphene oxide device before and after FEBID carbon deposition at the center of the conduction channel, respectively. Before/after FEBID carbon deposition, the direct current (DC) electrical characterization of the device was conducted in terms of two-terminal I-V curve, measuring the source-to-drain current vs. applied voltage Ids-Vds.

FIG. 1.

(a) Schematic showing the experimental steps used for studying the effect of FEBID carbon patterning on local modification of electronic properties of the GO sheets. SEM images of (b) as-fabricated graphene oxide device and (c) the device after FEBID carbon patterning (electron dose of ∼4.2 × 1018 e/cm2) at the center of the channel region (marked as a red dashed-line box).

FIG. 1.

(a) Schematic showing the experimental steps used for studying the effect of FEBID carbon patterning on local modification of electronic properties of the GO sheets. SEM images of (b) as-fabricated graphene oxide device and (c) the device after FEBID carbon patterning (electron dose of ∼4.2 × 1018 e/cm2) at the center of the channel region (marked as a red dashed-line box).

Close modal

Figure 2(a) shows the results of two-terminal I-V measurements as a function of the electron dose used in FEBID carbon deposition. As-fabricated GO device without FEBID treatment exhibits perfectly insulating behavior with no current flow at applied voltages up to ±4 V. After FEBID carbon film patterning with a low electron dose of ∼3.4 × 1017 e/cm2, the conduction channel opens up and current starts to flow between two terminals through the modified graphene oxide. Increasing the electron dose to ∼4.2 × 1018 e/cm2 results in both the improved linearity of the Ids-Vds curve and a dramatic increase of the measured current, by at least one order of magnitude. It is worth noting that relatively low source-to-drain current even after FEBID treatment is due to a relatively large in extent “untreated” regions of graphene oxide sheets which were not exposed to primary electrons during FEBID (Fig. 1(a)), especially near the source and drain metal contact areas, which are still electrically insulating.

FIG. 2.

(a) Ids-Vds measurements of the GO device as a function of the electron dose for FEBID carbon deposition, as compared to as-fabricated device without FEBID, indicating an increase of the channel conductivity with FEBID treatment with an increased electron irradiation dose. (b) Ids-Vds measurement of FEBID carbon film bridging the drain-source metal electrodes, deposited on a bare SiO2 substrate without GO sheets as a control experiment. Full area map AFM images and the corresponding cross-sectional profiles of (c) the GO device after FEBID with electron dose of ∼4.2 × 1018 e/cm2 and (d) the device without GO (as a control) after FEBID with electron dose of ∼1.7 × 1019 e/cm2.

FIG. 2.

(a) Ids-Vds measurements of the GO device as a function of the electron dose for FEBID carbon deposition, as compared to as-fabricated device without FEBID, indicating an increase of the channel conductivity with FEBID treatment with an increased electron irradiation dose. (b) Ids-Vds measurement of FEBID carbon film bridging the drain-source metal electrodes, deposited on a bare SiO2 substrate without GO sheets as a control experiment. Full area map AFM images and the corresponding cross-sectional profiles of (c) the GO device after FEBID with electron dose of ∼4.2 × 1018 e/cm2 and (d) the device without GO (as a control) after FEBID with electron dose of ∼1.7 × 1019 e/cm2.

Close modal

In an attempt to provide a direct evidence for reduction of graphene oxide by FEBID, Raman analysis of graphene oxide was performed before and after FEBID. However, presence of a thin amorphous carbon film on top of graphene oxide after FEBID masks the GO film and makes Raman measurement impractical in detecting any structural changes due to sensitivity limits. Instead, as a control experiment to establish that a change of electronic properties is indeed due to modification/reduction of graphene oxide and not simply due to an additional conduction path through the carbon film deposited atop of graphene oxide sheet, we implemented the same size of FEBID carbon patterning on a bare SiO2 substrate between source and drain metal electrodes but without a graphene oxide channel. FEBID carbon patterning was done using the same electron energy (25 keV) to that used in the graphene oxide device, even with much higher electron dose of ∼1.7 × 1019 e/cm2, resulting in a thicker FEBID carbon film than that in the GO device, which was deposited with electron dose of ∼4.2 × 1018 e/cm2, as shown in Figs. 2(c) and 2(d). Figure 2(b) shows the electrical measurement results for the control experiment, indicating that even thicker FEBID carbon film deposition alone is not sufficient to establish a conductive path for current flow. It is not surprising as by itself FEBID carbon is amorphous and highly insulating material.19,20 This result further supports the conclusion that FEBID carbon deposition must have resulted in doping modification/reduction of the graphene oxide substrate, thus increasing the channel conductivity observed in the experiments.

For greater mechanistic understanding of the experimentally observed effects, density functional theory (DFT) calculations were performed to simulate the FEBID process on the surface of graphene oxide. In DFT calculations, geometry optimization of “graphene oxide structure” and “graphene oxide + FEBID carbon species” were implemented, using Generalized Gradient Approximation (GGA) Perdew-Burke-Ernzerhof (PBE) functional,21–23 for the exchange correlation potential of interaction electrons with double numerical basis set in the DMol3.24 Self-consistent field (SCF) convergence, 10−5 Ha, was obtained at 9 × 9 × 1 Monkhorst-Pack (MP) k-point grid.15 For modeling graphene oxide, only epoxide and hydroxyl species were considered as surface functionalities of graphene oxide. The graphene oxide structure in Fig. 3(a) was obtained by the DFT calculations by positioning of four hydroxyl groups and one epoxide group on a 2 × 2 supercell of graphene (eight carbon atoms), which can lead to the most stable structure known as C8(OH)4O.25 Our DFT model prediction of the graphene oxide structure with 75% carbon atoms being oxidized fits well to the recent XPS results of the chemical composition of graphene oxide.4 The electronic band gap of the modeled graphene oxide was calculated to be ∼3.2 eV, which is in a good agreement with the literatures.25,26

FIG. 3.

(a) Graphene oxide model structure obtained by DFT calculations. (b) Optimized graphene oxide structures reacting with four possible CH4-derived intermediate species dissociated by electron beam during FEBID (grey: carbon, red: oxygen, and white: hydrogen).

FIG. 3.

(a) Graphene oxide model structure obtained by DFT calculations. (b) Optimized graphene oxide structures reacting with four possible CH4-derived intermediate species dissociated by electron beam during FEBID (grey: carbon, red: oxygen, and white: hydrogen).

Close modal

In order to describe the effect of FEBID process on the graphene oxide, we modeled four possible reactions of CH4–derived intermediate species (possible outcomes of electron beam induced dissociation of CH4 precursor molecule)15,27 with the graphene oxide surface, as shown in Fig. 3(b). Each species was positioned on the surface of the graphene oxide, and geometry optimization was implemented for each case to obtain the energetically most stable structures. Based on these calculations, addition of reactive species led to the most stable structure of graphene oxide by removing a surface functional group and forming CH3OH, CH2O, CHOOH, and CO groups, which are weakly bound to the graphene oxide surface and can be readily volatilized. These possible reactions are analogous to surface reaction of graphene oxide with highly reactive methane plasma generated by electron beam at low temperature, which (as consequence) reduces concentration of oxygen functional groups in a graphene oxide film.28 The difference between FEBID and the plasma treatment consists in whether electron beam dissociates precursor adsorbed on graphene oxide or pre-generated precursor plasma kinetically strikes graphene oxide. While kinetic energy of methane plasmas can facilitate reduction reactions with graphene oxide during plasma treatment, direct irradiation of electrons during FEBID can influence such reactions of dissociated species with functional groups of graphene oxide, which helps to overcome activation energies necessary for the reaction of highly reactive species with graphene oxide, similar to promoting chemisorption of FEBID intermediate species on graphene.15,29 Furthermore, the removal of functional groups and reduction of O:C ratio in GO via FEBID treatment are in a good agreement with the recently reported XPS results about the reduction of graphene oxide by irradiating high energy (10 MeV) electrons.30 Collectively, the DFT calculations and the relevant experimental results from the literatures provide a strong support for graphene oxide chemical modification/reduction facilitated by FEBID carbon deposition as a primary mechanism responsible for an increase of the local GO electrical conductivity and transition from its insulating to semiconducting behavior by FEBID treatment, as observed in our experiments upon an increase in the electron irradiation dose (Fig. 2(a)).

In summary, we explored the effect of FEBID treatment on the localized electrical property modification of a graphene oxide based electronic device using an electron beam with ∼100 nm resolution. A substantial increase in the graphene oxide channel conductivity was experimentally demonstrated with FEBID process upon exposure of GO conduction channel to high energy electrons, with transition to nearly linear Ohmic I-V behavior after irradiation with a higher electron beam dose. A fundamental mechanism of graphene oxide electrochemical transformation facilitated by electron beam irradiations was supported by comprehensive DFT calculations, whose results suggest that FEBID carbon doping and GO chemical interactions with highly reactive intermediate products of the electron-beam dissociation of the adsorbed hydrocarbon precursor induce local reduction of graphene oxide, resulting in substantially increased electrical conductivity by at least one order of magnitude. These findings establish the foundation for a physico-chemical route of localized reduction of graphene oxide by electron beam with direct-write patterning capability on nanoscale, which enables FEBID engineering of electrical properties of graphene oxide for electronic device applications unachievable by conventional batch-scale reduction techniques.

This work was supported by Semiconductor Research Corporation GRC Contract No. 2011-OJ-2221 and AFOSR BIONIC Center Award Nos. FA9550-09-1-0162 and FA9550-14-1-0269. Test structure microfabrication was performed using NSF supported NNIN facilities at Georgia Tech's Institute for Electronics and Nanotechnology (IEN). Electrical measurements were performed using experimental facilities in Professor Bakir's laboratory (ECE, Georgia Tech). Technical assistance with Raman measurements of graphene oxide was provided by S. Malak.

1.
K.
Hu
,
D. D.
Kulkarni
,
I.
Choi
, and
V. V.
Tsukruk
,
Prog. Polym. Sci.
39
,
1934
(
2014
).
2.
W.
Gao
,
L. B.
Alemany
,
L.
Ci
, and
P. M.
Ajayan
,
Nat. Chem.
1
,
403
(
2009
).
3.
Y.
Zhu
,
S.
Murali
,
W.
Cai
,
X.
Li
,
J. W.
Suk
,
J. R.
Potts
, and
R. S.
Ruoff
,
Adv. Mater.
22
,
3906
(
2010
).
4.
D. D.
Kulkarni
,
S.
Kim
,
M.
Chyasnavichyus
,
K.
Hu
,
A. G.
Fedorov
, and
V. V.
Tsukruk
,
J. Am. Chem. Soc.
136
,
6546
(
2014
).
5.
K.
Hu
,
L. S.
Tolentino
,
D. D.
Kulkarni
,
C.
Ye
,
S.
Kumar
, and
V. V.
Tsukruk
,
Angew. Chem., Int. Ed.
52
,
13784
(
2013
).
6.
D. R.
Dreyer
,
S.
Park
,
C. W.
Bielawski
, and
R. S.
Ruoff
,
Chem. Soc. Rev.
39
,
228
(
2010
).
7.
A. B.
Bourlinos
,
D.
Gournis
,
D.
Petridis
,
T.
Szabo
,
A.
Szeri
, and
I.
Dekany
,
Langmuir
19
,
6050
(
2003
).
8.
M. C.
Kim
,
G. S.
Hwang
, and
R. S.
Ruoff
,
J. Chem. Phys.
131
,
064704
(
2009
).
9.
E. C.
Mattson
,
H.
Pu
,
S.
Cui
,
M. A.
Schofield
,
S.
Rhim
,
G.
Lu
,
M. J.
Nasse
,
R. S.
Ruoff
,
M.
Weinert
,
M.
Gajdardziska-Josifovska
,
J.
Chen
, and
C. J.
Hirschmugl
,
ACS Nano
5
,
9710
(
2011
).
10.
Z.
Wei
,
D.
Wang
,
S.
Kim
,
S. Y.
Kim
,
Y.
Hu
,
M. K.
Yakes
,
A. R.
Laracuente
,
Z.
Dai
,
S. R.
Marder
,
C.
Berger
,
W. P.
King
,
W. A.
Heer
,
P. E.
Sheehan
, and
E.
Riedo
,
Science
328
,
1373
(
2010
).
11.
S. J.
Randolph
,
J. D.
Fowlkes
, and
P. D.
Rack
,
Crit. Rev. Solid State Mater. Sci.
31
,
55
(
2006
).
12.
W. F.
van Dorp
and
C. W.
Hagen
,
J. Appl. Phys.
104
,
081301
(
2008
).
13.
W. F.
van Dorp
,
X.
Zhang
,
B. L.
Feringa
,
T. W.
Hansen
,
J. B.
Wagner
, and
J. Th. M.
De Hosson
,
ACS Nano
6
,
10076
(
2012
).
14.
J. C.
Meyer
,
C. O.
Girit
,
M. F.
Crommie
, and
A.
Zettl
,
Appl. Phys. Lett.
92
,
123110
(
2008
).
15.
S.
Kim
,
D. D.
Kulkarni
,
R.
Davis
,
S. S.
Kim
,
R. R.
Naik
,
A. A.
Voevodin
,
S. S.
Jang
,
V. V.
Tsukruk
, and
A. G.
Fedorov
,
ACS Nano
8
,
6805
(
2014
).
16.
J.
Bishop
,
C. J.
Lobo
,
A.
Martin
,
M.
Ford
,
M.
Phillips
, and
M.
Toth
,
Phys. Rev. Lett.
109
,
146103
(
2012
).
17.
G.
Liu
,
D.
Teweldebrhan
, and
A. A.
Balandin
,
IEEE Trans. Nanotechnol.
10
,
865
(
2011
).
18.
A.
Fedorov
,
S.
Kim
,
M.
Henry
,
D.
Kulkarni
, and
V. V.
Tsukruk
,
Appl. Phys. A: Mater. Sci. Process.
117
,
1659
(
2014
).
19.
D. D.
Kulkarni
,
S.
Kim
,
A. G.
Fedorov
, and
V. V.
Tsukruk
,
Adv. Funct. Mater.
22
,
2129
(
2012
).
20.
N.
Kurra
,
V. S.
Bhadram
,
C.
Narayana
, and
G. U.
Kulkarni
,
ACS Appl. Mater. Interfaces
4
,
1030
(
2012
).
21.
E. J. G.
Santos
,
A.
Ayuela
, and
D.
Sanchez-Portal
,
New J. Phys.
14
,
043022
(
2012
).
22.
D. W.
Boukhvalov
and
M. I.
Katsnelson
,
J. Phys. D: Appl. Phys.
43
,
175302
(
2010
).
23.
J. P.
Perdew
,
K.
Burke
, and
M.
Ernzerhof
,
Phys. Rev. Lett.
77
,
3865
(
1996
).
24.
B.
Delley
,
J. Chem. Phys.
113
,
7756
(
2000
).
25.
D. W.
Boukhvalov
and
M. I.
Katsnelson
,
J. Am. Chem. Soc.
130
,
10697
(
2008
).
26.
R. J. W. E.
Lahaye
,
H. K.
Jeong
,
C. Y.
Park
, and
Y. H.
Lee
,
Phys. Rev. B
79
,
125435
(
2009
).
27.
D. A.
Alman
,
D. N.
Ruzic
, and
J. N.
Brooks
,
Phys. Plasmas
7
,
1421
(
2000
).
28.
M.
Baraket
,
S. G.
Walton
,
Z.
Wei
,
E. H.
Lock
,
J. T.
Robinson
, and
P.
Sheehan
,
Carbon
48
,
3382
(
2010
).
29.
J. D.
Jones
,
K. K.
Mahajan
,
W. H.
Williams
,
P. A.
Ecton
,
Y.
Mo
, and
J. M.
Perez
,
Carbon
48
,
2335
(
2010
).
30.
E.
Jin
,
J.
He
,
K.
Sheng
,
Z.
Zhang
,
G.
Shi
, and
Q.
Zheng
,
Acta Mater.
61
,
6466
(
2013
).