NANO LETTERS
Blue Luminescence Based on Quantum Confinement at Peptide Nanotubes
2009 Vol. 9, No. 9 3111-3115
Nadav Amdursky,† Michel Molotskii,‡ Daniel Aronov,‡ Lihi Adler-Abramovich, Ehud Gazit,† and Gil Rosenman*,‡ Department of Molecular Microbiology and Biotechnology, George S. Wise Faculty of Life Sciences, Tel AViV UniVersity, Tel AViV, 69978, Israel, and Department of Electrical EngineeringsPhysical Electronics, School of Engineering, Tel AViV UniVersity, Tel AViV, 69978, Israel Received March 15, 2009; Revised Manuscript Received August 4, 2009
ABSTRACT We report on observation of photoluminescence (PL) in blue and UV regions of exciton origin in bioinspired materialspeptide nanotubes (PNTs). Steplike optical absorption and PL measurements have allowed finding quantum confined (QC) phenomenon in PNTs. The estimations show that QC in these nanotubes occurs due to a crystalline structure of subnanometer scale dimension formed under the self-assembly process. Our new findings pave the way for the integration of PNT in a new generation of optical devices. A blue PL array of a PNT-patterned device is demonstrated.
Quantum confinement (QC) phenomenon has been intensively investigated in many solid state materials. The most common QC structures are quantum wells (QWs) mainly based on GaAs where two-dimensional double heterostructures comprising a thin layer of GaAs (in the order of ∼10 nm) surrounded by a bulk of AlGaAs or InGaAs possessing wider energy gap. QC structures produce remarkable changes in the optical properties of semiconductors and are widely used in photonic devices.1-3 Until now QC effects have been studied in inorganic materials where fabrication of QW is performed by conventional microelectronic technology. In this paper we report on observation of QC effects in bioinspired organic nanostructural material, peptide nanotubes (PNTs), where the self-assembly process of the dipeptides NH2-Phe-Phe-COOH (FF) leads to creation of crystalline structure in the subnanometer scale. We have found that this QC nanostructure demonstrates a steplike optical absorption which allows relating it to QWs. Strong photoluminescence in the blue and UV spectra of exciton origin has been observed, and a blue light source luminescence array is demonstrated. The FF PNTs were first discovered by Reches and Gazit.4 The discovery resulted from the determination of the smallest core recognition motif of amyloid beta (Aβ) peptide to be the diphenylalanine element. The Aβ peptide is the basic peptide of the Alzheimer disease fibrils. The single X-ray * Corresponding author,
[email protected]. † Department of Molecular Microbiology and Biotechnology, George S. Wise Faculty of Life Sciences, Tel Aviv University. ‡ Department of Electrical EngineeringsPhysical Electronics, School of Engineering, Tel Aviv University. 10.1021/nl9008265 CCC: $40.75 Published on Web 08/25/2009
2009 American Chemical Society
structure analysis of the FF PNT showed that the diphenylalanine monomers crystallize with hydrogen-bonded headto-tail chains in the shape of helices with four to six peptide molecules per turn.5,6 These PNTs have attractive properties such as high stability at high temperatures up to 300 °C and in various organic solvents.7 Moreover, they are extremely rigid, possessing a high aspect ratio of 19 GPa Young modulus,8 and have a defined crystalline structure.5,9 These properties differentiate them from other biological entities. Much recently, Ryu et al.10 showed the photoluminescence (PL) properties of FF PNT incorporated with photosensitizers and lanthanide ions. A new preparation process for making highly ordered alignments of FF PNT has been developed and used in our studies.11 This method allows covering any surface with dense, vertical, and aligned PNTs. A scanning electron microscopy image of a surface covered with aligned PNT can be seen in Figure 1b. The FF PNTs have a wide range of diameters, from 10 nm to 0.5 µm. The PNT deposition was performed onto quartz surfaces for detection of optical absorption and PL. We studied the optical properties of FF monomers and normally aligned PNTs. The PNT deposition was onto quartz surfaces for detection of optical absorption up to 210 nm. The measurements were conducted by the use of Cary 5000 UV-vis-NIR spectrophotometer (Varian, Inc. CA, USA) for the optical absorption, and FluoroMax-3 spectrofluorometer (Horiba Jobin Yvon, NJ, USA) for the PL properties. A DMRB fluorescence microscope (Leica, Germany) was used for fluorescence imaging.
Figure 3. PL of FF PNTs (solid line) and FF monomers (dashed line). The excitation wavelength was 260 nm. Figure 1. (a) Schematic of the diphenylalanine monomer. (b) Scanning electron microscopy image of the normally aligned PNT. The inset shows higher magnification of the surface; scale bar of inset is 5 µm.
Figure 2. Optical absorption of FF PNTs (solid line) and FF monomers (dash line).
Figure 2 shows the absorption spectrum of the PNT in comparison to FF monomers in aqueous solution. The spectrum of the monomers caused only by the aromatic rings of the phenylalanine residues; thus, it possess the same absorption peak (located at 257 nm) as the phenylalanine residue.12,13 However, the absorption spectrum of the aligned FF PNT demonstrates significantly different properties. By looking at the absorption spectrum of the FF PNT (solid line, Figure 2) we notice two distinguished steps located at 245-264 and 300-370 nm, compared to a narrow absorption peak for FF monomers. The optical properties of semiconductors dramatically change for QC structures where the optical absorption coefficient is completely defined by the electron density of states (DOS). The DOS of a two-dimensional (2D) QW is characterized by a steplike behavior. The DOS properties cause the idealized plot of the absorption spectrum of a QW structure to adopt a steplike behavior as well. The recorded steplike optical absorption behavior (Figure 2) clearly indicates the existence of 2D-QW crystalline structures, embedded in the FF PNT during the self-assembly process of deposition. 3112
PNT structures also generate strongly different PL compared to the FF monomer. Figure 3 shows the PL of the FF PNT and FF monomers under excitation at 260 nm (the excitation wavelength of the phenylalanine residue). The FF monomers possess a single peak, located at 284 nm, that characterizes the phenylalanine residue.12,13 The main peak of the FF PNT at this excitation wavelength is a sharp peak, located at 305 nm. A second peak was found in the range of 400-500 nm. The first peak is red-shifted in 11 nm (300 meV). Red shifting of PL is already been observed at aromatic-based molecules upon aggregation.14 However, the red shift in our case is bigger in more than an order of magnitude from the examples that have been observed.14 As a consequence, we would like to distinguish the studied nanotubes from regular aggregation process and ascribe the observed red shift to the unique crystallization process described here. In order to investigate the PL nature of the PNTs, we measured the photoluminescence excitation (PLE) spectrum of the structures at two emission wavelengths, at the first peak of 305 nm (dashed line, Figure 4a) and at the second peak of 450 nm (solid line, Figure 4a). As we can see, the main origin of the 450 nm peak is located at 370 nm. At 260 nm the PLE intensity is about 15% from the intensity at 370 nm, hence the low intensity of the 450 nm PL peak as seen in Figure 3 (at excitation of 260 nm). The two PLE peaks fit perfectly to the red edges of the two absorption steps (solid line, Figure 2). Figure 4b shows the PL intensity of the 450 nm peak under excitation at 370 nm (solid line) in comparison to the intensity of the 305 nm peak under excitation at 260 nm (dashed line). The 450 nm peak is very intense, five times more than the 305 nm peak. Dramatic variation in the optical properties of PNT in comparison to FF-monomer showing a steplike optical absorption allows relating the phenomenon to a creation of two-dimensional crystallized regions in these biological nanotubes. This ordered 2D structure indicates an anisotropic self-assembly growth. The longitudinal size of PNT along the Z-axis reaches a few micrometers (Figure 1b). However, one of the transverse dimensions, the width of the QW, LZ, should be much smaller and may be reduced to nanoscale Nano Lett., Vol. 9, No. 9, 2009
of QW width based on experimentally measured optical properties. The size of LZ in QW structure defines the exciton binding energy ∆Eexc, its dimensionality, and basic optical properties due to enhanced the exciton oscillator strength in a lowdimensional structures. The LZ may be estimated from the experimentally found value of ∆Eexc. The implemented theoretical calculations19-21 showed that if LZ is sufficiently smaller than the effective Bohr radius of exciton, rB*, LZ , rB*, the 3D exciton degenerates into 2D dimension. The exciton binding energy critically depends on the well width: for LZ , rB* it reaches its maximum value of 4 Ry* and it is reduced to Ry* for a wide quantum well LZ . rB*, here Ry* is the effective Rydberg constant. For the well of a finite depth, the value of ∆Eexc may be found as Ry* < ∆Eexc < 4 Ry*. Increasing the exciton binding energy in confined quantum structures leads to pronounced exciton effects, observed in optical absorption and PL, not only at low but also at room and elevated temperatures22 (as seen in Figure 3). The values rB* and Ry* are defined as rB* )
Ry* ) Figure 4. (a) PLE spectrum of FF PNT while detecting the emission at two wavelengths, at 450 nm (solid line) and at 305 nm (dashed line). (b) PL spectrum of FF PNT at two excitation wavelengths, at 370 nm (solid line) and at 260 nm (dashed line).
to support the found 2D-QC characterized by optical absorption and PL properties. The calculation of the QW dimensions is of fundamental interest and allows better understanding of the structure and dimensions of the elementary building blocks forming the PNT. One of the distinguished features of the QC phenomenon is the strengthening of Coulomb interaction “electron-hole” and the creation of excitons, whose binding energy exceeds the corresponding value for the 3D materials.1-3 The excitons are observed at the long-wavelength edge of the optical absorption spectra of QWs. Excitons in molecular solids possess physical properties which are intermediate between the ones described by Wannier and Frenkel models.14,15 However, due to its simplicity, the Wannier’s exciton model is widely applied for estimation of exciton binding energy, Eexc, in molecular solids.16-18 This model gives reasonable values in the cases when Eexc is sufficiently high and it exceeds by 1 to 2 orders of magnitude the exciton binding energy in covalent semiconductors. There is no published study on exciton electronic structure in organic molecular QW structures. In our case, the Wannier’s exciton model, which does not take into account some specific features of Frenkel’s exciton, was applied by us to organic molecular QW structures, where excitons possess intermediate properties characterized by Wannier and Frenkel models. According to that, it allowed us to estimate the parameters of localized excitons and to propose a new method for the determination Nano Lett., Vol. 9, No. 9, 2009
ε∞p2 µe02 µe04 2ε∞2p2
(1)
(2)
where e0 is the elementary charge, µ is the reduced mass of the exciton µ ) memh/(me + mh), where me and mh are the effective masses of electron and hole, respectively, and ε∞ is the high-frequency dielectric permittivity. It should be noted that if the exciton binding energy is less than the energy of phonons, the value of ε∞ should be changed to static dielectric permittivity ε0. For the sake of simplicity we consider the exciton in a potential well of infinite depth, when carrier wave function does not penetrate into the surrounding medium. Such a model was used in the work of Bastard et al.19 To calculate the electronic structure of the exciton in 2DQW, the data on electron and hole potential wells, effective masses and dielectric permittivity for the exciton frequency (ωexc ) ∆Eexc / p), corresponding to exciton binding energy, must be needed. Such data for peptide tubes are not available and any calculation may be considered as approximation. We will use the simple model by Bastard et al.,19 in our estimation, as well as our experimental observation of optical PNT properties. Our experimental measurements show steplike optical absorption (Figure 3) which might be understood as a transition between full and empty electronic states. Peaks localized at the “red” edge of these steps are related to direct optical exciton excitation. According to the data, the first absorption step is observed at Estep(1) ) 4.11 eV, the first exciton peak is found at the red edge of the first step, Eexc(1) ) 3.13 eV. Thus, the value of the exciton binding energy is ∆Eexc(1) ) Estep(1) - Eexc(1) ) 0.98 eV, which may be related to the states of the lowest subbands, with quantum numbers 3113
of transverse motion ne ) nh ) 1. The second exciton peak around 4.79 eV, which is clearly observed at the red edge of the second step, corresponds to ne ) nh ) 2. This step starts approximately at Estep(2) ) 5.51 eV. It should be noted that due to the quartz substance limitation, this energy region, which was used in our experiments, has a severe limitation to E ) 5.9 eV. The calculations, presented by Bastard et al.,19 of the electronic structure of excitons in GaAs QW, are in dimensionless units. The length is measured in units of Bohr’s radius and the value of energy is measured in Rydberg constant Ry* in a dielectric medium. Let us assume that dependence of dimensionless exciton binding energy ∆E/ Ry* versus dimensionless QW length LZ/rB* is the same as that obtained in ref 14. However, the basic parameters are taken from the experimental results with PNTs (Figure 3). The value ∆E was found from optical absorption data for PNT ∆Eexc(1) ) 0.98 eV. This value exceeds a maximum phonon energy, which allows using for ε∞, and its value for organic materials is about ε∞ ) 4. We should calculate two unknown variables, LZ and µ. Therefore we need to compose two equations. The first equation may be written as LZ µ LZ ) rB* ε∞rB m0
(3)
where m0 ) 9.1 × 10-28 g is the free-electron mass and rB ) 0.529 × 10-8 cm is Bohr’s radius in a hydrogen atom. Dimensionless exciton binding energy is given by m0 ∆Eexc(1) ) 1.152 Ry* µ
(4)
The second curve in Figure 2 links expressions 3 and 4. The second equation may be composed from the energy difference of the first and second steps in optical absorption spectrum (Figure 3). The starting point of a step with quantum numbers n ) ne ) nh is equal to En ) U +
π2p2n2 2µLZ2
(5)
where U is the constant value which disappears in the final equation. The difference between energies of the second and first steps does not depend on U and may be written as ∆E12 ) Estep(2) - Estep(1) )
3π2p2 2µLZ2
(6)
Expressions 3 and 6 depend on µ and LZ. In eq 3 one can find µLZ, while in eq 6 we have µLZ2. Therefore, using our experimental values of the exciton binding energy, ∆Eexc(1) ) 0.98 eV and ∆E12 ) 1.64 eV, we may find both µ and LZ. Substitution of the experimental data allowed to get the values of µ ) 0.86m0 and LZ ) 9 Å. 3114
The obtained data allow defining the exciton properties in self-assembly quantum confined bioinspired PNT material. The exciton has a high binding energy, ∆Eexc(1) ) 0.98 eV, which significantly exceeds other known semiconductor materials due to much stronger confinement. The found width of QW in PNT, LZ ) 9 Å, is a bit smaller than the calculated effective Bohr’s diameter 2rB* ) 13.8 Å, obtained for µ ) 0.86m0. The found value of exciton binding energy in QW ∆Eexc ) 0.98 eV exceeds the binding energy of 3D exciton ∆E3D ) Ry* ) 0.43 eV by approximately 2 times; however it is 2 times less than the binding energy in 2D space ∆E2D ) 4E3D ) 1.72 eV. Therefore, we may relate the exciton to an intermediate deformed one, occupying the position between 2D and 3D excitons. For structural validation of our calculation, we refer to the pioneer works made by Go¨rbitz on the crystal structure of the FF PNT.5,6 The FF PNTs crystallize to a noncentrosymmetric P61 space group. The unit cell dimension of the nanotubes suggest two equal sidewalls (a and b) at the size of 24.07 Å and a c sidewall at the size of 5.45 Å.5 Our calculated confinement size, LZ, is in the order of the c sidewall. Moreover, Go¨rbitz suggests a model for the inner surface of the nanotubes, containing multiple hydrophilic/ hydrophobic channels, aligned parallel to the main axis of the tube (Figure 3 of ref 6). From this model we can locate the confined axis to be between two adjacent channels. The size between two adjacent channels is in the order of 10 Å, which corresponds with our calculations. In this case the confinement is being created by structural borders, unlike the common semiconductors QW structure where the confinement is created by changes in the band gaps. As is well-known, excitons are effectively captured by shallow traps. Such exciton localization allows observing enormous growth of the oscillation strength, known as Rashba effect.23 It has been found that localization of the exciton in a QW, which does not contain point defects, noticeably increases the oscillator strength of radiation transitions. Dramatic growth by 3-4 orders of magnitude of the oscillator strength may be observed in the case of the weak exciton localization at numerous shallow traps existing in QW. Such an effect leads to fast exciton decay due to a strong growth of the radiative recombination rate and PL intensity. The same reason causes a sharp increase optical absorption. Another exciton-related effect is PL. The PL from PNT observed at room temperature demonstrates two peaks which may be ascribed to radiative decay of excitons, at 305 nm, located at the UV range, and at 450 nm, located at the blue range of the visible spectrum. The found “red” shift (compared to the location of exciton optical adsorption peaks) may be related to the Stocks effect. The full width at halfmaximum (fwhm) values of the PL peaks are ∆λ1 ) 33 nm and ∆λ2 ) 78 nm for the first and second PL peaks, respectively. Such a wide PL peak (the second peak) may be ascribed to electron-phonon interactions and/or influence of structural defects generated during the forming process of the PNT. The defects may lead to numerous overlapping Nano Lett., Vol. 9, No. 9, 2009
luminescence array. Moreover, self-assembly quantum confined bioinspired nanostructures may lead to a new class of biolasers. Acknowledgment. We thank D. Huppert for assisting in the spectroscopic measurements and fruitful discussion. References
Figure 5. Fluorescence microscopy image of a patterned surface of FF PNT on silicon under excitation at 340-380 nm. The blue squares are the PL from the FF PNTs; the purple circle is a result of the reflections from surface of the excitation beam.
optical transitions and provide significant widening of PL spectral bands. The high intensity of the blue PL peak encouraged us to image the fluorescence of the structures. Figure 5 shows a fluorescence microscopy image of a patterned sample of FF PNT under excitation at 340-380 nm. We can clearly see the blue PL from the PNT patterning (in comparison to the dark purple reflection of the excitation beam, located in the center of the sample). QC phenomena such as steplike optical absorption and PL of exciton origin have been observed in bioinspired PNTs. It has been found that highly ordered subnanocrystalline 2DQW structures are embedded by a self-assembly process along the bio-organic FF PNT. This self-assembled quantum confined structure is highly unique, not like in other aggregation processes of aromatic organic molecules.14 PL in UV and blue regions is paving the way for the use of FF PNT in the field of new generation of environmentally clean optical materials for photonic short wavelength devices. As shown, the FF PNT surface can be easily patterned which promotes the use of FF PNTs in a novel organic blue
Nano Lett., Vol. 9, No. 9, 2009
(1) Bastard, G. WaVe Mechanics Applied to Semiconductor Heterostructure; Wiley-Interscience: New York, 1991. (2) Mitin, V. V.; Kochelap, V. A.; Stroscio, M. A. Quantum Heterostructures: Microelectronics and Optoelectronics; Cambridge University Press: Cambridge, 1999. (3) Weisbuch, C.; Vinter, B. Quantum Semiconductor Structures; Academic Press: New York, 1991. (4) Reches, M.; Gazit, E. Science 2003, 300 (5619), 625–627. (5) Gorbitz, C. H. Chem.sEur. J. 2001, 7 (23), 5153–5159. (6) Gorbitz, C. H. Chem. Commun. 2006, (22), 2332–2334. (7) Adler-Abramovich, L.; Reches, M.; Sedman, V. L.; Allen, S.; Tendler, S. J. B.; Gazit, E. Langmuir 2006, 22 (3), 1313–1320. (8) Kol, N.; Adler-Abramovich, L.; Barlam, D.; Shneck, R. Z.; Gazit, E.; Rousso, I. Nano Lett. 2005, 5, 1343–1346. (9) Reches, M.; Gazit, E. Nat. Nanotechnol. 2006, 1 (3), 195–200. (10) Ryu, J.; Lim, S. Y.; Park, C. B. AdV. Mater. 2009, 21 (16), 1577– 1581. (11) Adler-Abramovich, L.; Aronov, D.; Beker, P.; Yevnin, M.; Stempler, S.; Buzhansky, L.; Rosenman, G.; Gazit, E. Nat. Nanotechnol. 2009, in press. (12) Mihalyi, E. J. Chem. Eng. Data 1968, 13 (2), 179–182. (13) van Holde, K. E.; Johnson, C.; Ho, P. S. Principles of Physical Biochemistry, 1st ed.; Prentice Hall: Upper Saddle River, NJ, 1998. (14) Pope, M.; Swenberg, C. E. Annu. ReV. Phys. Chem. 1984, 35, 613– 655. (15) Gumbs, G.; Mavroyannis, C. Solid State Commun. 1982, 41 (3), 237– 240. (16) Heeger, A. J. ReV. Mod. Phys. 2001, 73 (3), 681–700. (17) Moses, D.; Wang, J.; Heeger, A. J.; Kirova, N.; Brazovski, S. Synth. Met. 2001, 125 (1), 93–98. (18) Zheng, G.; Clark, S. J.; Brand, S.; Abram, R. A. J. Phys.: Condens. Matter 2004, 16 (47), 8609–8620. (19) Bastard, G.; Mendez, E. E.; Chang, L. L.; Esaki, L. Phys. ReV. B 1982, 26 (4), 1974–1979. (20) Ralph, H. I. Solid State Commun. 1965, 3 (10), 303–306. (21) Shinada, M.; Sugano, S. J. Phys. Soc. Jpn. 1966, 21 (10), 1936–1946. (22) Schmittrink, S.; Chemla, D. S.; Miller, D. A. B. AdV. Phys. 1989, 38 (2), 89–188. (23) Rashba, E. I. SoV. Phys. Semicond. 1975, 8, 807–816.
NL9008265
3115
