APPLIED PHYSICS LETTERS 91, 063124 共2007兲

Making graphene visible

P. Blakea兲 and E. W. Hill
Department of Computer Sciences, University of Manchester, Manchester M13 9PL, United Kingdom
A. H. Castro Neto
Department of Physics, Boston University, 590 Commonwealth Avenue, Boston, Massachusetts 02215
K. S. Novoselov, D. Jiang, R. Yang, T. J. Booth, and A. K. Geim
Department of Physics and Astronomy, University of Manchester, Manchester M13 9PL, United Kingdom
共Received 1 May 2007; accepted 14 June 2007; published online 10 August 2007兲
Microfabrication of graphene devices used in many experimental studies currently relies on the fact
that graphene crystallites can be visualized using optical microscopy if prepared on top of Si wafers
with a certain thickness of SiO2. The authors study graphene’s visibility and show that it depends
strongly on both thickness of SiO2 and light wavelength. They have found that by using
monochromatic illumination, graphene can be isolated for any SiO2 thickness, albeit 300 nm 共the
current standard兲 and, especially, ⬇100 nm are most suitable for its visual detection. By using a
Fresnel-law-based model, they quantitatively describe the experimental data. © 2007 American
Institute of Physics. 关DOI: 10.1063/1.2768624兴

Since it was reported in 2004,1 graphene—a one-atom- So far, this detection technique has been demonstrated
thick flat allotrope of carbon—has been attracting increasing and widely used only for a SiO2 thickness of 300 nm
interest.1–3 This interest is supported by both the realistic 共purple-to-violet in color兲, but a 5% change in the thickness
promise of applications and the remarkable electronic prop- 共to 315 nm兲 can significantly lower the contrast.2 Moreover,
erties of this material. It exhibits high crystal quality, ballistic under nominally the same observation conditions, graphene’s
transport on a submicron scale 共even under ambient condi- visibility strongly varies from one laboratory to another 共e.g.,
tions兲 and its charge carriers accurately mimic massless see images of single-layer graphene in Refs. 1 and 4兲, and
Dirac fermions.2–4 Graphene samples currently used in ex- anecdotal evidence attributes such dramatic differences to
periments are usually fabricated by micromechanical cleav- different cameras, with the cheapest ones providing better
age of graphite: a euphemism for slicing this strongly layered imaging.8 Understanding the origin of this contrast is essen-
material by gently rubbing it against another surface.5 The tial for optimizing the detection technique and extending it to
ability to create graphene with such a simple procedure en- different substrates, aiding experimental progress in the re-
sures that graphene was produced an uncountable number of search area.
times since graphite was first mined and the pencil invented In this letter, we discuss the origin of this optical contrast
in 1565.6 and show that it appears due not only to an increased optical
Although graphene is probably produced every time one path but also to the notable opacity of graphene. By using a
uses a pencil, it is extremely difficult to find small graphene model based on the Fresnel law, we have investigated the
crystallites in the “haystack” of millions of thicker graphitic dependence of the contrast on SiO2 thickness and light wave-
flakes which appear during the cleavage. In fact, no modern length ␭, and our experiments show excellent agreement
visualization technique 共including atomic-force, scanning- with the theory. This understanding has allowed us to maxi-
tunneling, and electron microscopies兲 is capable of finding mize the contrast and, by using narrow-band filters, to find
graphene because of their extremely low throughput at the graphene crystallites for practically any thickness of SiO2
required atomic resolution or the absence of clear signatures and also on other thin films such as Si3N4 and polymethyl
distinguishing atomic monolayers from thicker flakes. Even methacrylate 共PMMA兲.
Raman microscopy, which recently proved itself as a power- Figure 1 illustrates our main findings. It shows graphene
ful tool for distinguishing graphene monolayers,7 has not yet viewed in a microscope 关Nikon Eclipse LV100D with a
been automated to allow search for graphene crystallites. Un- 100⫻, 0.9 numerical aperture 共NA兲 objective兴 under normal,
til now, the only way to isolate graphene is to cleave graphite white-light illumination on top of a Si wafer with the stan-
on top of an oxidized Si wafer and then carefully scan its dard 300 nm thickness of SiO2 关Fig. 1共a兲兴. For comparison,
surface in an optical microscope. Thin flakes are sufficiently Fig. 1共c兲 shows a similar sample but on top of 200 nm SiO2,
transparent to add to an optical path, which changes their where graphene is completely invisible. In our experience,
interference color with respect to an empty wafer.1 For a only flakes thicker than ten layers could be found in white
certain thickness of SiO2, even a single layer was found to light on top of 200 nm SiO2. Note that the ten-layer thick-
give sufficient, albeit feeble, contrast to allow the huge ness also marks the commonly accepted transition from
image-processing power of the human brain to spot a few graphene to bulk graphite.2 Top and bottom panels in Fig. 1
micron-sized graphene crystallites among copious thicker show the same samples but illuminated through various
flakes scattered over a millimeter-sized area. narrow-band filters. Both flakes are now clearly visible. For
300 nm SiO2, the main contrast appears in green
a兲
Electronic mail: [email protected] 关see Fig. 1共b兲兴, and the flake is undetectable in blue light. In

0003-6951/2007/91共6兲/063124/3/$23.00 91, 063124-1 © 2007 American Institute of Physics

Downloaded 10 Oct 2010 to 117.195.224.168. Redistribution subject to AIP license or copyright; see http://apl.aip.org/about/rights_and_permissions
 063124-2 Blake et al. Appl. Phys. Lett. 91, 063124 共2007兲

and characterized by a complex refractive index n3共␭兲 that,

importantly, is dependent on ␭ 关for example, n3共␭
= 400 nm兲 ⬇ 5.6− 0.4i兴.9 The SiO2 layer is described by
thickness d2 and another ␭-dependent refractive index n2共␭兲
but with a real part only9 关n2共400 nm兲 ⬇ 1.47兴. We note that
these n2共␭兲 and n3共␭兲 accurately describe the whole range of
interference colors for oxidized Si wafers.10 Single-layer
graphene is assumed to have a thickness d1 equal to the
extension of the ␲ orbitals out of plane11 共d1 = 0.34 nm兲 and
a complex refractive index n1共␭兲. While n1共␭兲 can be used in
our calculations as a fitting parameter, we avoided this un-
certainty after we found that our results were well described
by the refractive index of bulk graphite n1共␭兲 ⬇ 2.6− 1.3i,
which is independent of ␭.9,12 This can be attributed to the
FIG. 1. 共Color online兲 Graphene crystallites on 300 nm SiO2 imaged with fact that the optical response of graphite with the electric
white light 共a兲, green light and another graphene sample on 200 nm SiO2 field parallel to graphene planes is dominated by the in-plane
imaged with white light 共c兲. Single-layer graphene is clearly visible on the electromagnetic response.
left image 共a兲, but even three layers are indiscernible on the right 共c兲. Image Using the described geometry, it is straightforward to
sizes are 25⫻ 25 ␮m2. Top and bottom panels show the same flakes as in 共a兲
and 共c兲, respectively, but illuminated through various narrow bandpass filters
show that the reflected light intensity can be written as:13
with a bandwidth of ⯝10 nm. The flakes were chosen to contain areas of
different thickness so that one can see changes in graphene’s visibility with I共n1兲 = 兩共r1ei共⌽1+⌽2兲 + r2e−i共⌽1−⌽2兲 + r3e−i共⌽1+⌽2兲
increasing numbers of layers. The trace in 共b兲 shows steplike changes in the
contrast for 1, 2, and 3 layers 共trace averaged over 10 pixel lines兲. This + r1r2r3ei共⌽1−⌽2兲兲 ⫻ 共ei共⌽1+⌽2兲 + r1r2e−i共⌽1−⌽2兲
proves that the contrast can also be used as a quantitative tool for defining
the number of graphene layers on a given substrate.
+ r1r3e−i共⌽1+⌽2兲 + r2r3ei共⌽1−⌽2兲兲−1兩2 , 共1兲
where
comparison, the use of a blue filter makes graphene visible
even on top of 200 nm SiO2 共see lower panels兲. n0 − n1
r1 = ,
To explain the observed contrast, we consider the case of n0 + n1
normal light incidence from air 共refractive index n0 = 1兲 onto
a trilayer structure consisting of graphene, SiO2, and Si 共see
inset of Fig. 2兲. The Si layer is assumed to be semi-infinite n1 − n2
r2 = ,
n1 + n2

n2 − n3
r3 = 共2兲
n2 + n3
are the relative indices of refraction. ⌽1 = 2␲n1d1 / ␭ and ⌽2
= 2␲n2d2 / ␭ are the phase shifts due to changes in the optical
path. The contrast C is defined as the relative intensity of
reflected light in the presence 共n1 ⫽ 1兲 and absence 共n1 = n0
= 1兲 of graphene,

I共n1 = 1兲 − I共n1兲
C= . 共3兲
I共n1 = 1兲
For quantitative analysis, Fig. 2 compares the contrast
observed experimentally with the one calculated by using
Eq. 共3兲. The experimental data were obtained for single-layer
graphene on top of SiO2 / Si wafers with three different SiO2
thicknesses by using 12 different narrow-band filters. One
can see excellent agreement between the experiment and
theory. The contrast reaches up to ⯝12%, and the peaks in
graphene’s visibility are accurately reproduced by our
model.14 Note, however, that the theory slightly but system-
atically overestimates the contrast. This can be attributed to
deviations from normal light incidence 共because of high NA兲
and an extinction coefficient of graphene, k1 = −Im共n1兲, that
may differ from that of graphite. k1 affects the contrast both
by absoption and by changing the phase of light at the inter-
FIG. 2. Contrast as a function of wavelength for three different thicknesses
faces, promoting destructive interference. To emphasize the
of SiO2. Circles are the experimental data; curves the calculations. Inset: the important role played by this coefficient, the dashed line in
geometry used in our analysis. Fig. 2共c兲 shows the same calculations but with k1 = 0. The
Downloaded 10 Oct 2010 to 117.195.224.168. Redistribution subject to AIP license or copyright; see http://apl.aip.org/about/rights_and_permissions
 063124-3 Blake et al. Appl. Phys. Lett. 91, 063124 共2007兲

maximized for any SiO2 thickness by using appropriate fil-

ters. Our work establishes a quantitative framework for de-
tecting single and multiple layers of graphene and other two-
dimensional atomic crystals5 on top of various substrates.
The authors thank I. Martin for illuminating discussions
and C. Luke from Nikon UK for the loan of the monochrome
camera.8 The Manchester work was supported by EPSRC
共UK兲, and one of the authors 共A.H.C.N.兲 by NSF under
Grant No. DMR-0343790. After our letter was submitted,
four preprints15–18 discussing the same topic appeared on the
cond-mat arXiv.
FIG. 3. 共Color online兲 Color plot of the contrast as a function of wavelength 1
K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang, S. V.
and SiO2 thickness according to Eq. 共3兲. The color scale on the right shows
Dubonos, I. V. Grigorieva, and A. A. Firsov, Science 306, 666 共2004兲.
the expected contrast. 2
A. K. Geim and K. S. Novoselov, Nat. Mater. 6, 183 共2007兲.
3
A. H. Castro Neto, F. Guinea, and N. M. R. Peres, Phys. World 19, 33
latter curve does not bare even a qualitative similarity to the 共2007兲.
4
experiment, which proves the importance of opacity for the Y. Zhang, J. W. Tan, H. L. Stormer, and P. Kim, Nature 共London兲 438,
201 共2005兲.
visibility of graphene. 5
K. S. Novoselov, D. Jiang, F. Schedin, T. J. Booth, V. V. Khotkevich, S. V.
To provide a guide for the search of graphene on top of Morozov, and A. K. Geim, Proc. Natl. Acad. Sci. U.S.A. 102, 10451
SiO2 / Si wafers, Fig. 3 shows a color plot for the expected 共2005兲.
6
contrast as a function of SiO2 thickness and wavelength. This H. Petroski, The Pencil: A History of Design and Circumstance 共Knopf,
plot can be used to select filters most appropriate for a given New York, 1989兲, Chap. 4, pp. 36–47.
7
A. C. Ferrari, J. C. Meyer, V. Scardaci, C. Casiraghi, M. Lazzeri, F. Mauri,
thickness of SiO2. It is clear that by using filters, graphene S. Piscanec, D. Jiang, K. S. Novoselov, S. Roth, and A. K. Geim, Phys.
can be visualized on top of SiO2 of practically any thickness, Rev. Lett. 97, 187401 共2006兲.
except for ⬇150 nm and below 30 nm. Note, however, that 8
Filtered light images are taken with a Nikon DS-2MBWc monochrome
the use of green light is most comfortable for eyes that, in camera. White light images are taken with a Nikon DS-2Mv color camera.
our experience, become rapidly tired with the use of high- Cheaper cameras are more likely to do extensive postprocessing of images
in firmware or software that could enhance contrast.
intensity red or blue illumination. This makes SiO2 thick- 9
Handbook of Optical Constants of Solids, edited by E. D. Palik
nesses of approximately 90 and 280 nm most appropriate 共Academic, New York, 1991兲, 2, pp. 457–458.
10
with the use of green filters as well as without any filters, in J. Henrie, S. Kellis, S. Schultz, and A. Hawkins, Opt. Express 12, 1464
white light. In fact, the lower thickness of ⯝90 nm provides 共2004兲.
11
Linus Pauling, The Nature of the Chemical Bond 共Cornell University
a better choice for graphene’s detection 共see Fig. 2兲, and we
Press, Ithaca, 1960兲, Chap. 7, pp. 234–235.
suggest it as a substitute for the present benchmark thickness 12
In Ref. 9, the refractive index of bulk graphite is within 5% of 2.6− 1.3i
of ⯝300 nm. between 300 and 590 nm. At 630 nm, the extinction coefficient jumps to
Finally, we note that the changes in the light intensity 1.73, but this coincides with a change of reference in the handbook, which
due to graphene are relatively minor, and this allows the we have chosen to ignore in our model.
13
H. Anders, Thin Films in Optics 共Focal, London, 1967兲, Pt. 1, pp. 18–48.
observed contrast to be used for measuring the number of 14
The experimental contrast was found by computer analysis of the images
graphene layers 共theoretically, multilayer graphene can be obtained using a monochrome camera Ref. 8. The thickness of SiO2 usu-
modeled by the corresponding number of planes separated ally differs by up to 5% from nominal values provided by suppliers and,
by d1兲. The trace in Fig. 1共a兲 shows how the contrast changes accordingly, in our theoretical calculations in Fig. 2, the following values
with the number of layers, and the clear quantized plateaus for d2 were used to acheive the best fit: 共a兲 290 nm, 共b兲 190 nm, and 共c兲
88 nm.
show that we have regions of single, double, and triple layer 15
I. Jung, M. Pelton, R. Piner, D. A. Dikin, S. Stankovich, S. Watcharotone,
graphene. Furthermore, by extending the same approach to M. Hausner, and R. S. Ruoff, e-print arXiv:cond-mat/0706.0029.
16
other insulators, we were able to find graphene on 50 nm C. Casiraghi, A. Hartschuh, E. Lidorikis, H. Qian, H. Harutyunyan, T.
Si3N4 using blue light and on 90 nm PMMA using white Gokus, K. S. Novoselov, and A. C. Ferrari, e-print arXiv:cond-mat/
light. 0705.2645.
17
D. S. L. Abergel, A. Russell, and V. I. Fal’ko, e-print arXiv:cond-mat/
In summary, we have investigated the problem of visibil- 0705.0091.
ity of graphene on top of SiO2 / Si wafers. By using the 18
S. Roddaro, P. Pingue, V. Piazza, V. Pellegrini, and F. Beltram, e-print
Fresnel theory, we have demonstrated that contrast can be arXiv:cond-mat/0705.0492.

Downloaded 10 Oct 2010 to 117.195.224.168. Redistribution subject to AIP license or copyright; see http://apl.aip.org/about/rights_and_permissions