Article pubs.acs.org/JPCC
Highly Efficient Fluorescence Quenching with Graphene Amal Kasry,*,†,‡ Ali A. Ardakani,† George S. Tulevski,† Bernhard Menges,§ Matthew Copel,† and Libor Vyklicky† †
IBM T. J. Watson Research Center, 1101 Kitchawan Road, Yorktown Heights, New York 10598, United States Egypt Nanotechnology Center (EGNC), Smart Village, Giza 12577 Egypt § Max Planck Institute for Polymer Research, Ackermannweg 10, 55128 Mainz, Germany ‡
S Supporting Information *
ABSTRACT: Fluorescence quenching is a powerful technique used to obtain information about the dynamic changes of proteins in complex macromolecular systems. In this work, graphene is shown to be a very efficient quencher of fluorescence molecules where the quenching effect was one order of magnitude higher than that of gold. The fluorescence intensity was distance-dependent where increasing the distance between the fluorescence molecule and the graphene surface from 4 to 7 nm increased the fluorescence intensity by a factor of 7.5. This type of distance dependence suggests a nonradiative nature in the energy transfer between the graphene and the fluorophore due to the excitation of an exciton.
■
INTRODUCTION Graphene has attracted enormous interest for several applications due to its unique optical, electrical, and thermal properties.1−4 In one such application, graphene (typically graphene oxide or exfoliated graphene5−11) is employed as a substrate for biosensors, where energy transfer is the key sensing mechanism.12−14 Fluorescence quenching is one of the most powerful techniques for obtaining information about the conformational or dynamic changes of proteins in complex macromolecular systems.15,16 The emission of a fluorophore is perturbed by a quencher, which can be a substrate or another molecule, under certain conditions that allows for energy transfer between the fluorophore and the quencher. In the vicinity of a metal surface, the emission of the excited fluorophore is altered by the interaction with the electric field at the metal surface. The interactions between metal surfaces and fluorophores through radiative or nonradiative decay are well-understood.17−19 Graphene is a unique interface for fluorescence energy transfer because it is a nearly transparent semimetal. Graphene has a linear band dispersion around the corners of its Brillouin zone20 and a nearly constant optical absorption, and thus optically excited species can be quenched by resonant energy transfer via excitation of electron−hole pairs in the graphene. This behavior was observed between semiconducting nanocrystals and graphene.21 The rate of resonance energy transfer from the excited state of a dye to the π system of graphene was also theoretically studied.22,23 © 2012 American Chemical Society
In this work, graphene is shown to be a very efficient quencher for fluorescent molecules, as compared with metals. The fluorescence quenching is distance-dependent, which supports the nonradiative nature of the energy transfer between the graphene and the fluorophores. This effect can, potentially, be used to study protein−protein interactions and dynamic changes in biomolecules.
■
EXPERIMENTAL SECTION Graphene was grown on copper via a CVD method.24 The growth and the transfer method is explained elsewhere.25 Graphene was transferred to several substrates including gold, quartz, silicon, and 300 nm SiO2 on Si depending on the type of measurement that was performed. Fluorescence-labeled molecules were attached to graphene through biotinylated compounds of varying lengths (Scheme 1). The synthesis of compound I and its binding to graphene is explained elsewhere. 26 Streptavidin (SA) labeled with CY5 dye (Invitrogen) was bound to the modified graphene by incubating the substrate for 15 min at room temperature in a solution of the modified CY5 dye. This modification yields a modified graphene substrate where the labeled streptavidin is 4 nm above the substrate as measured by surface plasmon Received: August 18, 2011 Revised: December 13, 2011 Published: January 4, 2012 2858
dx.doi.org/10.1021/jp207972f | J. Phys. Chem. C 2012, 116, 2858−2862
The Journal of Physical Chemistry C
Article
Scheme 1a
static energy analyzer.27,28 Surface plasmon resonance (SPR) and surface plasmon fluorescence spectroscopy (SPFS) measurements were done using a home-built SPR setup using a 633 HeNe laser. The sample was mounted on the SPR set up in a Kretschman configuration.29−34 The biotin-based monolayer was assembled on the surface in situ using a 300 μL volume PDMS flow cell attached to the substrate. Binding of the SA to the surface was also monitored by SPR, where 1 μM SA was applied in situ to the surface. A reflectivity and fluorescence scan was performed after each layer was deposited. Fitting of the SPR data to determine the distance above the surface was done using Winspall software using a model based on the Fresnel and Maxwell equations. Fluorescence emission was also measured by direct excitation of the fluorophores using an ARAMIS Raman System from Horiba at 532 nm excitation wavelength.
■
RESULTS AND DISCUSSION To study the effect of graphene on fluorescence intensity, we excited fluorophores in the total internal reflection mode using SPFS and by direct excitation using a spectrometer equipped with a 532 nm laser source and a cooled CCD detector. SPFS utilizes the evanescent electromagnetic field of a surface plasmon to excite fluorophores in close proximity to the surface.32−35 Graphene was transferred to a glass substrate coated with 2 nm Cr/50 nm Au to use the plasmon mode in the gold layer. For comparison, similar measurements were performed on bare gold. Nonspecific binding to a bare graphene surface does not occur as previously demonstrated,26 so the collected fluorescence signal was only from the bound species at the defined distance and not due to any false signal from nonspecific binding. There are two key parameters in this experiment: the distance between the metal and the fluorescence molecule and the surface density of the fluorescence molecules. Both should be accurately calculated to eliminate the quenching effect of the metal layer and measure only the quenching effect of graphene. To calculate the distance and the surface density, SPR reflectivity scans were performed and fitted by a model based on the Fresnel equations. To determine the fitting parameters, the thickness and the optical constants (n and k) of the graphene layer were obtained. The graphene thickness was determined from the MEIS measurements. In this system, protons with energy of 100 keV are incident along a major crystallographic direction in the solid. Energy- and angleresolved detection of backscattered ions provide surface structural and compositional information. Figure 1a shows the MEIS results indicating approximately one layer of graphene, which corresponds to a thickness of 0.4 nm.27 Refractive index and extinction coefficient (n and k) were measured by ellipsometry. Fitting the ellipsometer results gave n and k values for the graphene layer of 2.7 and 2, respectively, at 630 nm (Figure 1b). The diazotized compound I (Scheme 1 a) was first bound to the graphene surface, the binding was monitored in real time, followed by binding of the SA labeled with Cy5 dye. The binding of the SA was also monitored in real time by measuring the fluorescence signal. Reflectivity and fluorescence scans were performed after each binding step. The results obtained from MEIS and ellipsometry were used to fit the SPR reflectivity scan data to determine the thickness of each layer. Figure 1c shows the shift in the resonance angle after adding each layer to the substrate. After the biotin compound and labeled SA were
a
(a) Compounds used for self-assembly on graphene. Compound I was converted to a diazonium salt prior to reacting with graphene. The reaction of compounds II and III yields the hydroxamic acid IV. (b,c) Diagrams showing the self-assembly of different chain length compounds on graphene and on gold.
resonance (SPR).26 This distance was increased by modifying graphene with compound (IV). (For the synthetic method, see Supporting Information.) Compound (IV) was assembled on graphene by incubating overnight and then washing with ethanol to remove the unbound compound. NHS-Biotin(III) was applied to the surface overnight, followed by washing with DMF and ethanol, and SA labeled with Alexa-Fluor dye was applied for 15 min; the longer spacer achieves a distance of ∼7 nm after the SA binding. For ellipsometery measurements, a J. A.Woollam VUV-VASE ellipsometer VU-302 was used to determine the optical constants (refractive index (n) and extinction coefficient (k)) of a single-layer graphene transferred to a quartz substrate. Graphene samples were measured in the wavelength range of 150−800 nm, and the data (with backside correction applied for quartz substrates) were fitted to obtain the n and k values by building a general oscillator model. The thickness of the graphene layer was determined by medium energy ion scattering (MEIS) measurements.27 The graphene layer was transferred to a Si substrate, and the MEIS measurements were performed using an HVEE system (100 KeV protons). Backscattered protons were analyzed with a toroidal electro2859
dx.doi.org/10.1021/jp207972f | J. Phys. Chem. C 2012, 116, 2858−2862
The Journal of Physical Chemistry C
Article
Figure 1. (a) MEIS measurements showing the backscattered spectra for a graphene film on silicon. The fitted data give ∼1.2 ML of graphene corresponding to a thickness of 0.4 nm. (b) Ellipsometry measurements showing the refractive index and extinction coefficient of graphene. (c) SPR reflectivity scans as a function of the incident angle of graphene on 50 nm gold before (black curve) and after modification with the diazotized compound I (red curve) and after binding with streptavidin (blue curve).
applied to the graphene surface, the distance above the metal was found to be 4.5 nm, which includes the graphene layer. To apply a similar system above the gold layer, we used Anti-SA to be directly physisorbed to the gold surface, and then CY5labeled SA was bound to it; this system gave a similar distance of 4.5 nm. The surface density was determined to be 1.184 and 2.122 ng/mm2 on both graphene and gold, respectively. The surface densities were calculated from the SPR shift.36 Fluorescence scans as a function of incident angle were performed after applying the CY5-SA layer in both cases. The results are shown in Figure 2. The fluorescence intensity on the graphene/gold system is almost one order of magnitude less than that on gold despite having the fluorescence molecule at the same distance above both substrates. Additionally, the surface density is 1.8 times less than that on gold. Both measurements were performed under the same conditions of laser intensity and binding time. This large reduction in the fluorescence intensity clearly indicates the highly efficient fluorescence quenching of graphene. Anomalously high efficiency quenching was also observed with semiconducting nanocrystals on graphene.21 At a distance of 4 nm above graphene, one cannot assume the quenching is due to charge transfer where the charge transfer takes place at very small distances.37 Another possible mechanism is the nonradiative decay of the fluorescence molecule via creation of an electron−hole pair.17,38 When fluorophores are in close proximity to metals, the metal layer alters the way an excited fluorophore loses its energy.32,39 There are additional decay channels that are contributing to the decrease in the radiative quantum yield of the fluorophores. They take place at different characteristic fluorophore-metal distances. At a distance less than 10 nm of the metal surface, the
Figure 2. (a) Reflectivity and fluorescence scans as a function of the incident angle of both gold and graphene on gold after binding of the Cy 5 labeled Streptavidin to the linker attached to both surfaces. The graphs show a reduction of the fluorescence intensity on the graphene surface due to its quenching effect. The surface densities in both cases calculated from the angle shift were found to be 1.184 ng/mm2 on graphene and 2.122 ng/mm2 on gold.
nonradiative decay of fluorescence is the dominating process. The excitation is assumed to be due to dipole−dipole interaction if the origin is excitation of an electron−hole pair (exciton) in the metal. The energy-transfer model (Förster model)40 gives an R−6 dependence of the transfer rate to the separation distance due to the coeffect from both near fields of the donor molecule and the acceptor molecule. However, the distance dependence of energy transfer can be greatly compromised to R−3 ≈ R−4 due to the integration over the enlarged number of effective acceptor sites. At an intermediate distance regime (a few nanometers up to ∼20 nm), a significant 2860
dx.doi.org/10.1021/jp207972f | J. Phys. Chem. C 2012, 116, 2858−2862
The Journal of Physical Chemistry C
Article
order of magnitude due to doping,22,23 which translates to a decrease in the fluorescence intensity. The doping effect needs to be further studied.
fraction of the excited fluorescence couples back to surface plasmon polaritons of the metal by fulfilling momentummatching conditions, and it represents a significant loss of fluorescence yield. At sufficient separation distances (>20 nm), free emission of the fluorophore dominates. These kinds of interactions between a fluorophore and a metal are theoretically calculated and quantitatively measured.41 To compare these effects of metals on fluorescence emission to that of graphene, we measured the fluorophore at different distances above a single graphene layer transferred to 300 nm SiO2 on Si substrates and measured the fluorescence intensity by direct excitation. In this case, SA labeled with Alexa Fluor 532 was used. To increase the distance, we applied an 11aminoundecylhydroxamic acid molecule (for the synthetic method, see the Supporting Information) overnight to the surface, NHS-biotin was bound by incubating overnight, followed by washing with DMF and ethanol; then, labeled SA was applied. Both systems should achieve an estimated total distance of 4 and 7 nm, respectively. Applying the hydroxamic acid to metal oxides forms a chain tilt of 20°,42 where the distance between the fluorophore and the graphene is ∼6.8 nm. The results of the fluorescence intensity above the graphene at different distances are shown in Figure 3. These results show an
■
CONCLUSIONS Graphene is shown to be a very efficient fluorescence quencher as compared with metals. The quenching mechanism of graphene is due to the nonradiative decay of the fluorophore and not due to charge transfer. The fluorescence intensity above graphene is distance-dependent, where increasing the distance between the fluorescence molecules and the graphene from 3 to 6.8 nm increased the fluorescence intensity by a factor of 7.5. Other factors are contributing to the fluorescence intensity like substrate roughness and doping, which create new channels for the fluorescence intensity to be decayed; these effects require further investigation.
■
ASSOCIATED CONTENT
S Supporting Information *
Synthesis of Compound 4 and surface density of the SA on the graphene modified with long and short chain linkers. This material is available free of charge via the Internet at http:// pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected];
[email protected].
■
ACKNOWLEDGMENTS We would like to thank Dr. Dennis Newns and Dr. Rzvan Nistor from IBM research for the useful discussions. This work was partially funded by the 2008 joint development agreement between IBM Research and the Government of the Arab Republic of Egypt through the Egypt Nanotechnology Center (EGNC); http://www.egnc.gov.eg
■
Figure 3. Fluorescence intensity measured by direct excitation showing the increase by a factor of 7.5 with increasing the distance above the graphene from 4 to 6.8 nm.
REFERENCES
(1) Geim, A. K.; Novoselov, K. S. Nat. Mater. 2007, 6, 183−191. (2) Geim, A. K. Science 2009, 324, 1530−1534. (3) Zhang, Y.; Tan, J. W.; Stormer, H. L.; Kim, P. Nature 2005, 438, 201−204. (4) Novoselov, K. S.; Geim, A. K.; Morozov, S. V.; Jiang, D.; Zhang, Y.; Dubonos, S. V.; Grigorieva, I. V.; Firsov, A. A. Science 2004, 306, 666−669. (5) Yang, W.; Ratinac, K. R.; Ringer, S. P.; Thordarson, P.; Gooding, J. J.; Braet, F. Angew. Chem., Int. Ed. 2010, 49, 2114−2138. (6) Lu, Y.; Goldsmith, B. R.; Kybert, N. J.; Johnson, A. T. C. Appl. Phys. Lett. 2010, 97, 083107−083108. (7) Liu, J.; Tao, L; Yang, W.; Li, D.; Boyer, C.; Wuhrer, R.; Braet, F.; Davis, T. P. Langmuir 2010, 26, 10068−10075. (8) Nair, R. R.; Blake, P.; Blake, J. R.; Zan, R.; Anissimova, S.; Bangert, U.; Golovanov, A. P.; Morozov, S. V.; Geim, A. K.; Novoselov, K. S.; Latychevskaia, T. Appl. Phys. Lett. 2010, 97, 153102−153104. (9) Chen, W.; Wee, T. S.; Loh, K. P. J. Am. Chem. Soc. 2008, 130, 14392−14393. (10) Ohno, Y.; Maehashi, K.; Matsumoto, K. J. Am. Chem. Soc. 2010, 132, 18012−18013. (11) Lu, C. H.; Yang, H. H.; Zhu, C. L.; Chen, X.; Chen, G. N. Angew. Chem., Int. Ed. 2009, 48, 4785−4787. (12) Loh, K. P.; Bao, Q.; Eda, G.; Chhowalla, M. Nat. Chem. 2010, 2, 1015−1024. (13) Chang, H.; Tang, L.; Wang, Y.; Jiang, J.; Li, J. Anal. Chem. 2010, 82, 2341−2346.
increase in the fluorescence intensity by a factor of 7.5 with increasing the distance by almost 3 nm, although the surface density of labeled SA in the case of the long-chain modification was almost half that on the short chain one. The surface density was measured by modifying graphene surfaces with the short and long chains, binding of SA-coated magnetic beads (from Mitenyi Biotec), and imaging with SEM. (See the Supporting Information.) This increase in the fluorescence intensity at higher distances is less than what is theoretically expected.40,41 This can be due to several factors; one of them is the surface roughness that could decrease the fluorescence intensity due to scattering. Further decrease in the fluorescence intensity was observed on rougher surfaces (data not shown). Another important factor is the degree of doping of the graphene. Modification by charge transfer as in our method causes the graphene to be doped.43 Doped graphene introduces new charge carriers that can change the rate of exciton creation and consequently affects the energy transfer. According to Swathi and Sebastian,22 doping does not modify the rate for electronic excitation energy transfer significantly. However, in the case of vibrational transfer, which could be the case in our system due to dipole−dipole interaction, the rate is found to increase by an 2861
dx.doi.org/10.1021/jp207972f | J. Phys. Chem. C 2012, 116, 2858−2862
The Journal of Physical Chemistry C
Article
(14) Dong, H.; Gao, W.; Yan, F.; Ji, H.; Ju, H. Anal. Chem. 2010, 82, 5511−5517. (15) Mátyusa, L.; Szöllősia, J.; Jeneia, A. J. Photochem. Photobiol., B 2006, 83, 223−236. (16) Zhuang, X.; Ha, T.; Kim, H. D.; Centner, T.; Labeit, S.; Chu, S. Proc. Natl. Acad. Sci. U.S.A. 2000, 9, 14241−14244. (17) Persson, B. N. J.; Lang, N. D. Phys. Rev. B 1982, 26, 5409−5415. (18) Lakowicz, J. R. Anal. Biochem. 2004, 324, 153−169. (19) Chhabra, R.; Sharma, J.; Wang, H.; Zou, S.; Lin, S.; Yan, H.; Lindsay, S.; Liu, Y. Nanotechnology 2009, 20, 485201−485210. (20) Neto, A. H. C.; Guinea, F.; Peres, N. M. R.; Novoselov, K. S.; Geim, A. K. Rev. Mod. Phys. 2009, 81, 109−162. (21) Chen, Z.; Berciaud, S.; Nuckolls, C.; Heinz, T. F.; Brus, L. E. ACS Nano 2010, 4, 2964−2968. (22) Swathi, R. S.; Sebastian, K. L. J. Chem. Phys. 2008, 129, 0547031−054703-9. (23) Swathi, R. S.; Sebastian, K. L. J. Chem. Phys. 2009, 130, 0861011−086101-3. (24) Li, X. S.; Cai, W. W.; An, J. H.; Kim, S.; Nah, J.; Yang, D. X.; Piner, R. D.; Velamakanni, A.; Jung, I.; Tutuc, E.; Banerjee, S. K.; Colombo, L.; Ruoff, L. R. S. Science 2009, 324, 1312−1314. (25) Kasry, A.; Kuroda, M. A.; Martyna, G. J.; Tulevski, G. S.; Bol, A. A. ACS Nano 2010, 4, 3839−3844. (26) Kasry, A.; Ardakani, A. A.; Oida, S.; Han, S. J.; Menges, B.; Tulevski, G. S. Chem. Mater. 2011, 23, 4879−4881. (27) Copel, M. IBM J. Res. Dev. 2000, 44, 571−528. (28) Copel, M.; Oida, S.; Kasry, A.; Bol, A. A.; Hannon, J. B.; Tromp, R. M. Appl. Phys. Lett. 2011, 98, 113103−113105. (29) Otto, A. Z. Phys. 1968, 216, 398−410. (30) Raether, H. Surface Plasmons on Smooth and Rough Surfaces and on Gratings. In Springer Tracts in Modern Physics; Hoehler, G., Eds.; Springer: Berlin, 1988; Vol. 111. (31) Knoll, W. Annu. Rev. Phys. Chem. 1998, 49, 569−638. (32) Liebermann, T.; Knoll, W. Colloids Surf., A 2000, 171, 115−130. (33) Stengel, G.; Knoll, W. Nucleic Acids Res. 2005, 33, e69−e69. (34) Dostálek, J.; Knoll, W. Biointerphases 2008, 3, 12−22. (35) Kasry, A.; Knoll, W. Appl. Phys. Lett. 2006, 89, 101106-1− 101106-3. (36) Yu, F.; Persson, B.; Lofås, S.; Knoll, W. J. Am. Chem. Soc. 2004, 126, 8902−8903. (37) Nistor, R. A.; Newns, D. M.; Martyna, G. J. ACS Nano 2011, 5, 3096−3103. (38) Weitz, D. A.; Garoff, S.; Hanson, C. D.; Gramila, T. J.; Gersten, J. I. Opt. Lett. 1982, 7, 89−91. (39) Barnes, W. L. J. Mod. Opt. 1998, 45, 661−699. (40) Lakowicz, J. R. Principles of Fluorescence Spectroscopy, 2nd ed.; Kluwer: New York, 1999. (41) Vasilev, K.; Knoll, W.; Kreiter, M. J. Chem. Phys. 2004, 20, 3439−3445. (42) Folkers, J. P.; Gorman, C. B.; Laibinis, P.; Buchholz, S.; Whitesides, G. M. Langmuir 1995, 11, 813−824. (43) Farmer, D. B.; Mojarad, R. G.; Perebeinos, V.; Lin, Y. M.; Tulevski, G. S.; Tsang, J. C.; Avouris, P. Nano Lett. 2009, 9, 388−392.
2862
dx.doi.org/10.1021/jp207972f | J. Phys. Chem. C 2012, 116, 2858−2862
