NPC-home page healing of defected...Created Date 7/9/2013 2:00:03 AM

Self healing of defected grapheneJianhui Chen, Tuwan Shi, Tuocheng Cai, Tao Xu, Litao Sun et al. Citation: Appl. Phys. Lett. 102, 103107 (2013); doi: 10.1063/1.4795292 View online: http://dx.doi.org/10.1063/1.4795292 View Table of Contents: http://apl.aip.org/resource/1/APPLAB/v102/i10 Published by the AIP Publishing LLC. Additional information on Appl. Phys. Lett.Journal Homepage: http://apl.aip.org/ Journal Information: http://apl.aip.org/about/about_the_journal Top downloads: http://apl.aip.org/features/most_downloaded Information for Authors: http://apl.aip.org/authors

Downloaded 09 Jul 2013 to 223.3.54.19. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://apl.aip.org/about/rights_and_permissions

Self healing of defected graphene

Jianhui Chen,1,2 Tuwan Shi,1,2 Tuocheng Cai,1,2 Tao Xu,3 Litao Sun,3 Xiaosong Wu,1,2,a)

and Dapeng Yu1,2,b)

1School of Physics, Peking University, Beijing 100871, People’s Republic of China2State Key Laboratory for Artificial Microstructure and Mesoscopic Physics, Peking University,Beijing 100871, People’s Republic of China3SEU-FEI Nano-Pico Center, Key Laboratory of MEMS of Ministry of Education, Southeast University,Nanjing 210096, People’s Republic of China

(Received 30 October 2012; accepted 28 February 2013; published online 13 March 2013)

For electronics applications, defects in graphene are usually undesirable because of their ability to

scatter charge carriers, thereby reduce the carrier mobility. It would be extremely useful if the damage

can be repaired. In this work, we employ Raman spectroscopy, X-ray photoemission spectroscopy,

transmission electron microscopy, and electrical measurements to study defects in graphene

introduced by argon plasma bombardment. We have found that majority of these defects can be cured

by a simple thermal annealing process. The self-healing is attributed to recombination of mobile

carbon adatoms with vacancies. With increasing level of plasma induced damage, the self-healing

becomes less effective. VC 2013 American Institute of Physics. [http://dx.doi.org/10.1063/1.4795292]

Graphene’s unique band structure endows it with extra-

ordinary electrical transport properties, owing to which it is

believed to find applications in the next generation electron-

ics and has thus drawn enormous attention.1 In particular,

graphene possesses a very high mobility, the highest one at

room temperature among all other materials.2 A high mobil-

ity means less resistive to electrical current and a faster

switching ability for transistors. Therefore, circuits built

from graphene can be potentially of low power consumption

and high speed.3 However, graphene’s mobility is affected

by various types of scatterers, i.e., charge impurities and

crystalline defects. Defects are especially harmful, because

they can backscatter electrons.4,5 To build either individual

devices or integrated circuits, many steps of lithographic

processes are required. It has been found that these processes

can cause damage to graphene.6–9 The mobility is then

reduced. To fully exploit the high mobility of graphene, it is

important not only to grow high quality materials but to de-

velop lithographic processes to maintain its original quality.

For the latter, an alternative approach is to cure the damage,

if it is difficult to avoid. In a related topic, methods have

been developed to reduce graphene oxide.10,11 However, the

main process there is reduction of oxide, e.g., removal of

functional groups. To study how structural defects in gra-

phene can be healed, it is much more relevant to study repair

of well defined defects. Ion bombardment has been exten-

sively used to study damage in graphite and graphene and

known to produce vacancies.12–20 Although repair of these

defects has been investigated in graphite,14,18 the defect there

can be more stable due to bond formation between layers,

hence more difficult to repair.15

In this work, we employ plasma bombardment to intro-

duce structural defects in monolayer graphene films. We per-

form thermal annealing to study healing effects on these

defects. By Raman, X-ray photoemission spectroscopy

(XPS), high resolution transmission electron microscopy

(HRTEM), and electrical transport measurements, it is found

that vacancies can be healed simply by thermal annealing.

The effect can be explained by annihilation of displaced car-

bon atoms with vacancies with assistance of thermal energy.

When the size of the vacancy increases, healing becomes

more difficult.

Graphene samples were prepared on SiO2 by mechanical

exfoliation. The thickness of graphene flakes was estimated

by optical images and confirmed by Raman spectra. Samples

were loaded into a Femto plasma system. 99.99% high impu-

rity argon gas was used to generate plasma. Low pressure

plasma (17 Pa) was used to increase the ion energy, thereby

enhance the bombardment effect. The plasma power was

20 W and the treatment time ranged from 5 to 60 s. The dos-

age per second is estimated to be 2� 1014=cm2 from the

power of the plasma generator and the area of the electro-

des.21 Thermal annealing was carried out in an argon envi-

ronment (300 Pa) in a tube furnace. The intake of the pump

was fitted with a custom made filter to reduce carbon con-

tamination. Raman spectra were taken on a Renishaw confo-

cal Raman system and the excitation was 514 nm. Samples

for HRTEM imaging were large area graphene films grown

by chemical vapour deposition on copper substrates. Pristine

graphene films were transferred to TEM copper grids, fol-

lowed by plasma treatment and annealing. The HRTEM

images were taken on Titan 80–300 TEM at an acceleration

voltage of 80 kV to avoid damage by electron beam irradia-

tion. For the transport study, samples were patterned into

1 lm� 6 lm wide Hall bars by e-beam lithography.

Electrical measurements were performed using a standard

lock-in method in an Oxford cryostat.

Graphene samples were subjected to argon plasma treat-

ment for various time durations. Raman spectra of pristine

graphene and treated graphene are shown in Fig. 1. A Dpeak, centered at 1350 cm�1, is significantly enhanced by

plasma treatment. The D peak is commonly known as a dis-

order peak, as it is associated with disorder in the sp2 bonded

carbon network.19,22 The development of a strong D peak

a)Electronic mail: [email protected])Electronic mail: [email protected].

0003-6951/2013/102(10)/103107/5/$30.00 VC 2013 American Institute of Physics102, 103107-1

APPLIED PHYSICS LETTERS 102, 103107 (2013)

Downloaded 09 Jul 2013 to 223.3.54.19. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://apl.aip.org/about/rights_and_permissions

indicates that substantial damage is introduced in graphene.

Two new peaks, appearing at 1623 cm�1 and 2939 cm�1,

are the D0 peak and the Dþ D0 peak. The amplitude of the

D0 peak is about one fifth of that of the D peak, indicating

that the type of the defects is not sp3-defects due to function-

alization but mainly vacancies.23,24 Moreover, we have also

taken spectra on bilayer regions after the plasma treatment.

As shown in Fig. 1, the Raman spectrum of a bilayer region

displays a strong D peak as monolayer regions do, indicating

that both layers experienced considerable damage. Note that

surface chemical modification will be very different for

monolayer and bilayer graphene.25 This provides evidence

that the damage is mainly caused by energetic argon ions,

and some ions are capable of penetrating through two

layers.18 We want to point out that the damage is severe,

because graphene films were completely removed after 60 s

plasma bombardment.

We then carried out thermal annealing on damaged

films. In Fig. 2(a), Raman spectra taken after annealing at

different temperatures are plotted. Starting from 300 �C, all

three disorder related peaks, D, D0, and Dþ D0, decrease as

the annealing temperature increases. For the highest anneal-

ing temperature, 900 �C, the two peaks that appeared after

plasma treatment, D0 and Dþ D0, almost completely disap-

pear, while the D peak is greatly reduced. This indicates that

large portion of the defects has been repaired. On the other

hand, the recovery of G and 2 D peaks signals that the 2D

sp2 bonded carbon network has been restored. The ratio

between the D peak intensity and the G peak intensity,

ID=IG, is inversely proportional to the domain size of gra-

phene.16,19,26 To quantitatively estimate the extent of repair,

we have calculated the domain size La as a function of the

annealing temperature from the ID=IG ratio, which are plot-

ted in Fig. 2(b). La increases from 10 nm to about 80 nm after

annealing.

XPS is employed to gain insight on the structure trans-

formation by annealing. C 1s spectra were taken for pristine

graphene, plasma treated graphene, and annealed graphene,

as seen in Fig. 2(c). The plasma treatment results in growth

of spectral weight on the high energy side of the C-C sp2

peak. By deconvolution, three peaks, centered at 286.6 eV,

288.1 eV, and 289.1 eV, are identified. The spectral weight

around 286.6 eV and 288.1 eV can be attributed to C-OH,

C¼O, or structural damage.27–30 We want to emphasize the

pronounced peak at 289.1 eV. It is assigned to COOH, which

only appears at graphene edges. In sharp contrast, this peak

FIG. 1. Raman spectra for graphene monolayer and bilayer films under 5 s,

10 s, and 15 s argon plasma treatment.

FIG. 2. Repair of defected graphene. (a) Raman

spectra for monolayer graphene samples sub-

jected to 5 s plasma treatment after annealing at

different temperatures. (b) Domain size La as a

function of the annealing temperature, showing

gradual healing of defects in the graphene film.

The solid line is guide to eye. In the inset,

logðL2a � L2

a0Þ is plotted against 1/T. The red line

is a linear fit excluding the lowest temperature

point. XPS spectra of a graphene film before (c)

and after (d) annealing at 750 �C.

103107-2 Chen et al. Appl. Phys. Lett. 102, 103107 (2013)

Downloaded 09 Jul 2013 to 223.3.54.19. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://apl.aip.org/about/rights_and_permissions

is much lower than C-OH and C¼O in graphene oxide.27,28

It is apparent that our samples possess significant amount of

vacancies, in agreement with the Raman results. The forma-

tion of COOH likely took place when damaged samples

were exposed to air. After annealing, the amplitudes of these

peaks are substantially reduced and the spectrum displays

only slight difference from that of the pristine graphene film,

seen in Fig. 2(d). It is evidenced that considerable amount of

vacancies have been repaired.

HRTEM images confirm the production and healing of

defects. Fig. 3 shows HRTEM images for samples before and

after plasma treatment and after annealing. Amorphous mate-

rials on the graphene surfaces are residue of poly(methyl

methacrylate) (PMMA) that were used for transfer. In the

clean region, the honeycomb lattice of graphene can be seen.

The pristine graphene film shows high crystalline quality with

no apparent defect. After the plasma treatment, a large amount

of vacancies were introduced. The average distance between

these defects is on the order of 10 nm, consistent with the

Raman result. Note that the distance is most likely shorter as

single vacancies can hardly be identified and part of the sur-

face is covered by residue. After annealing at 750 �C, the

defect density is substantially reduced. Only a few vacancies

can be unambiguously identified on the surface.

The restoration of the graphene structure is further sup-

ported by electrical transport measurements. The transport

data of a monolayer graphene sample are shown in Fig. 4.

The pristine sample exhibits characteristics of high quality

graphene, i.e., a weak temperature dependence of the resist-

ance and the half integer quantum Hall effect. The mobility

FIG. 3. HRTEM images for three samples (a), (b) pristine, (c), (d)after 5 s of plasma treatment and (e), (f) after annealing at 750 �C. The surfaces are partially

covered by PMMA residue. Red arrows indicate vacancies in the honeycomb lattice.

FIG. 4. Temperature dependence of resistance for a graphene film. (a) pris-

tine. (b) After plasma treatment. Inset: log q as a function of T�1=3. Red

solid line is a linear fit. (c) After thermal annealing at 500 �C.

103107-3 Chen et al. Appl. Phys. Lett. 102, 103107 (2013)

Downloaded 09 Jul 2013 to 223.3.54.19. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://apl.aip.org/about/rights_and_permissions

is 6100 cm2/Vs at room temperature. After 5 s plasma treat-

ment, the room temperature resistivity increased by over

20-fold to 10.8 kX/w. Furthermore, a strong temperature de-

pendence of the resistance is observed. The resistivity

diverges at low temperature, following a log q / T�1=3 law,

in agreement with the Mott variable range hopping behav-

iour. Thus, electrons in the system are strongly localized by

defects. A detailed calculation based on the fitting yields a

localization length of 21 nm.31,32 The localization length is

longer than the crystalline size La.33 We then annealed the

sample at a relatively low temperature, 500 �C, to prevent

metal electrodes from evaporation. After that, the electrical

transport was re-measured. It is found that the weak tempera-

ture dependence and metallic behaviour at high temperature

are recovered. The mobility is recovered to 910 cm2/Vs even

though the annealing condition is not ideal, suggesting that

thermal annealing along is very effective in healing defects.

The mobility will be further improved if the annealing tem-

perature can be increased.

Thermal annealing has also been performed for samples

with different levels of damage. Fig. 5 shows the Raman

spectra before and after 750 �C annealing for samples under

10 s and 15 s of plasma treatment. Although annealing

reduces the disorder related peaks, as the time increases, the

reduction of the D peak becomes more limited. It suggests

that when damage is severe, some is beyond repair.

During the plasma treatment, energetic argon ions create

structural defects in graphene. The formation energy for a

single isolated vacancy in graphene is about 7.6 eV.34 For ar-

gon, only ions with sufficient energy (>47 eV) can displace

carbon atoms and lead to formation of vacancies in gra-

phene.12,15,16,18 Displaced carbon atoms are absorbed on gra-

phene film and diffuse. Because of the high energy cost of

the dangling bond, recombination between adatoms and

vacancies is energetically favourable. Assume that the

migration barrier for carbon atoms is Em and the vacancies

are fixed, the change rate of vacancies density n due to

recombination can be described by a bimolecular equation:

dn=dt ¼ �Cn2e�Em=kBT .35 Here, C is the frequency factor, kB

is the Boltzmann constant, and T is the temperature. Thus,

we have 1=n� 1=n0 ¼ Cte�Em=kBT , where n0 is the initial

defect density after plasma treatment. Assume n ¼ 1=L2a, it

becomes L2a � L2

a0 ¼ Cte�Em=kBT . By plotting L2a � L2

a0 as a

function of 1/T in a semilog scale, we obtain the migration

barrier Em ¼ 0:95 eV. Theoretical studies gives various

values for the barrier, from 0.47 eV,36 to 0.6–1 eV in carbon

nanotubes,37 to <1.5 eV.34 Recent studies predict a barrier of

0.53 eV38 on graphene, 0.25 eV inside carbon nanotubes.39

Previous experiments on graphite have indeed shown that

carbon adatoms can annihilate with vacancies at elevated

temperatures.14 Most recently, such self-healing processes

have been visualized by transmission electron microscope.40

Our experiments provide further evidence in a macroscopic

scale that healing of defects in graphene by thermal anneal-

ing is very effective.

In conclusion, we employ argon plasma bombardment

to produce structural defects in graphene and study healing

of defects by thermal annealing. By comparing the Raman,

XPS spectra, and HRTEM images before and after anneal-

ing, we show that the defect density is significantly reduced

due to a self-healing process. Electrical measurements

demonstrate that the healing process is very effective as the

mobility is recovered to 910 cm2/Vs when the annealing tem-

perature is only 500 �C. Repair of graphene without external

carbon source is advantageous in micro-fabrication in that

graphene is likely covered by an overlayer during the litho-

graphic processing, which blocks the feed pathway of any

external carbon source.

This work was supported by NSFC (Project No.

11074007) and MOST (Nos. 2012CB933404,

2009CB623703). We also acknowledge the International

Science & Technology Cooperation Program of China Sino

Swiss Science and Technology Cooperation Program

(SSSTC, 2010DFA01810).

1A. K. Geim and K. S. Novoselov, Nature Mater. 6, 183 (2007).2K. I. Bolotin, K. J. Sikes, J. Hone, H. L. Stormer, and P. Kim, Phys. Rev.

Lett. 101, 096802 (2008).3Y.-M. Lin, C. Dimitrakopoulos, K. A. Jenkins, D. B. Farmer, H.-Y. Chiu,

A. Grill, and P. Avouris, Science 327, 662 (2010).4E. McCann, K. Kechedzhi, V. I. Fal’ko, H. Suzuura, T. Ando, and B. L.

Altshuler, Phys. Rev. Lett. 97, 146805 (2006).5X. S. Wu, X. B. Li, Z. M. Song, C. Berger, and W. A. de Heer, Phys. Rev.

Lett. 98, 136801 (2007).6D. Teweldebrhan and A. A. Balandin, Appl. Phys. Lett. 94, 013101

(2009).7I. Childres, L. A. Jauregui, M. Foxe, J. Tian, R. Jalilian, I. Jovanovic, and

Y. P. Chen, Appl. Phys. Lett. 97, 173109 (2010).8B. Dlubak, P. Seneor, A. Anane, C. Barraud, C. Deranlot, D. Deneuve,

B. Servet, R. Mattana, F. Petroff, and A. Fert, Appl. Phys. Lett. 97,

092502 (2010).

FIG. 5. Annealing of graphene films with differ-

ent levels of damage. (a) 10 s plasma treatment.

(b) 15 plasma treatment.

103107-4 Chen et al. Appl. Phys. Lett. 102, 103107 (2013)

Downloaded 09 Jul 2013 to 223.3.54.19. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://apl.aip.org/about/rights_and_permissions

9J. Fan, J. Michalik, L. Casado, S. Roddaro, M. Ibarra, and J. De Teresa,

Solid State Commun. 151, 1574 (2011).10S. Pei and H.-M. Cheng, Carbon 50, 3210 (2012).11M. Cheng, R. Yang, L. Zhang, Z. Shi, W. Yang, D. Wang, G. Xie, D. Shi,

and G. Zhang, Carbon 50, 2581 (2012).12J. Hahn and H. Kang, Surf. Sci. 446, L77 (2000).13D.-Q. Yang and E. Sacher, Surf. Sci. 504, 125 (2002).14B. An, S. Fukuyama, K. Yokogawa, and M. Yoshimura, J. Appl. Phys. 92,

2317 (2002).15A. V. Krasheninnikov and F. Banhart, Nature Mater. 6, 723 (2007).16M. Lucchese, F. Stavale, E. M. Ferreira, C. Vilani, M. Moutinho, R. B.

Capaz, C. Achete, and A. Jorio, Carbon 48, 1592 (2010).17E. H. M. Ferreira, M. V. O. Moutinho, F. Stavale, M. M. Lucchese, R. B.

Capaz, C. A. Achete, and A. Jorio, Phys. Rev. B 82, 125429 (2010).18I. N. Kholmanov, J. Edgeworth, E. Cavaliere, L. Gavioli, C. Magnuson,

and R. S. Ruoff, Adv. Mater. 23, 1675 (2011).19L. G. Cancado, A. Jorio, E. H. M. Ferreira, F. Stavale, C. A. Achete, R. B.

Capaz, M. V. O. Moutinho, A. Lombardo, T. S. Kulmala, and A. C.

Ferrari, Nano Lett. 11, 3190 (2011).20F. Banhart, J. Kotakoski, and A. V. Krasheninnikov, ACS Nano 5, 26

(2011).21The effective dosage per second is much less than 2� 1014=cm2 due to

three factors. First, the power that is converted to the ion energy in the

dark sheath is over-estimated because of the transfer efficiency. Second, in

17 Pa, the mean free path of Arþ ions is about 1 mm, much shorter than

the dark sheath width. Multiple collisions that ions experience effectively

reduce their energy. Third, the energy threshold for an Arþ ion to displace

a carbon atom is 47 eV. In our setup, Arþ ions of energy less than 30 eV

dominate. Therefore, majority of ions have no impact on graphene.22A. C. Ferrari, Solid State Commun. 143, 47 (2007).23P. Venezuela, M. Lazzeri, and F. Mauri, Phys. Rev. B 84, 035433 (2011).

24A. Eckmann, A. Felten, A. Mishchenko, L. Britnell, R. Krupke, K. S.

Novoselov, and C. Casiraghi, Nano Lett. 12, 3925 (2012).25S. Ryu, M. Y. Han, J. Maultzsch, T. F. Heinz, P. Kim, M. L. Steigerwald,

and L. E. Brus, Nano Lett. 8, 4597 (2008).26F. Tuinstra and J. L. Koenig, J. Chem. Phys. 53, 1126 (1970).27H.-K. Jeong, H.-J. Noh, J.-Y. Kim, M. H. Jin, C. Y. Park, and Y. H. Lee,

Europhys. Lett. 82, 67004 (2008).28D. Yang, A. Velamakanni, G. Bozoklu, S. Park, M. Stoller, R. D. Piner,

S. Stankovich, I. Jung, D. A. Field, C. A. Ventrice, Jr., et al., Carbon 47,

145 (2009).29D.-Q. Yang and E. Sacher, Langmuir 22, 860 (2006).30G. Speranza, L. Minati, and M. Anderle, J. Appl. Phys. 102, 043504

(2007).31N. Mott, J. Non-Cryst. Solids 1, 1 (1968).32N. Mott, M. Pepper, S. Pollitt, R. H. Wallis, and C. J. Adkins, Proc. R.

Soc. Lond. A 345, 169 (1975).33A. Lherbier, S. M.-M. Dubois, X. Declerck, S. Roche, Y.-M. Niquet, and

J.-C. Charlier, Phys. Rev. Lett. 106, 046803 (2011).34L. Li, S. Reich, and J. Robertson, Phys. Rev. B 72, 184109 (2005).35J. W. Corbett, R. B. Smith, and R. M. Walker, Phys. Rev. 114, 1460

(1959).36P. O. Lehtinen, A. S. Foster, A. Ayuela, A. Krasheninnikov, K. Nordlund,

and R. M. Nieminen, Phys. Rev. Lett. 91, 017202 (2003).37A. V. Krasheninnikov, K. Nordlund, P. O. Lehtinen, A. S. Foster,

A. Ayuela, and R. M. Nieminen, Phys. Rev. B 69, 073402 (2004).38H. Zhang, M. Zhao, X. Yang, H. Xia, X. Liu, and Y. Xia, Diamond Relat.

Mater. 19, 1240 (2010).39Y. Gan, J. Kotakoski, A. V. Krasheninnikov, K. Nordlund, and F. Banhart,

New J. Phys. 10, 023022 (2008).40R. Zan, Q. M. Ramasse, U. Bangert, and K. S. Novoselov, Nano Lett. 12,

3936 (2012).

103107-5 Chen et al. Appl. Phys. Lett. 102, 103107 (2013)

Downloaded 09 Jul 2013 to 223.3.54.19. This article is copyrighted as indicated in the abstract. Reuse of AIP content is subject to the terms at: http://apl.aip.org/about/rights_and_permissions