Investigation of time-resolved fluorescence lifetime of perylene dye
molecules embedded in silicon nanopillars
Sabriye Acikgoz
Received: 1 July 2014 / Accepted: 5 September 2014 / Published online: 11 October 2014 Ó Springer-Verlag Berlin Heidelberg 2014
Abstract The radiative decay rate of a perylene dye molecule attached to silicon nanopillar is investigated using a conventional time-correlated single photon count-ing technique. It is hard to produce a sustainable host with exactly the same dimensions all the time during fabrication to accommodate dye molecules for enhancement of spon-taneous emission rate. The laser-induced electrochemical anodization method allows us to have a control over size and shape of the silicon nanostructures. The effect of the silicon nanopillar on the radiative decay rate of the dye molecules is described by the Klimov’s prolate nanosphe-roid model. It is observed that the decay rate is significantly enhanced or inhibited due to plasmon resonance, depend-ing on whether the dipole is embedded closely right at the tip or at equator of the prolate nanospheroid. Both inhibi-tion and enhancement disappear when the distance between the dipole and prolate nanospheroid becomes large. Thus, the decay rate of the dye molecule approaches its natural value in the free space.
1 Introduction
Spontaneous emission or fluorescence is the process by which an atom in an excited state undergoes a transition to the ground state and the energy difference between the states transfers to a photon or a waveguide mode. The rate of the spontaneous emission could be altered or controlled by the modification of the electromagnetic vacuum field
that leads to inhibition or enhancement in the emission rate properties of molecules [1–3]. One way of controlling the radiative decay rate of a spontaneously emitted photon from a dye molecule is to embed it in the vicinity of a dielectric surface. The changes of the chemical (or pho-tonic) environment of the dye molecule can lead to large effects on the fluorescence lifetime. Understanding and controlling the emission properties of molecules in nano-structured geometries has a great potential for applications in the area of nano-optics, biochemistry and molecular biology [4–7].
The problem of radiative and nonradiative decay rates of a molecule in the presence of various dielectric nanobodies like spheroidal and cylindrical has been studied consider-ably over the past few decades. The most important question about this phenomenon to answer is to find an expressive relationship between the geometrical properties of the dielectric nanobodies in concern and the decay rates. Carminati et al. [8] have successfully derived an analytical procedure to explain the distance dependence of the radi-ative and nonradiradi-ative decay rates for spheroidal dielectric nanobodies. They demonstrated that the nonradiative decay rate follows an inverse sixth power dependence of the distance between the emitter and the center of the nano-particle at short range. Dielectric optical nanofibers and carbon nanotubes are important examples of the cylindrical nanocavities. The influence of such a cylindrical geometry is investigated for an atom being either confined inside or planted outside the cavity [9,10]. Another important the-oretical model to relate the geometrical parameters of various shapes to the decay rates of a fluorescence mole-cule is constructed by Klimov [11]. He investigated the optical properties of molecules in the vicinity of a prolate nanospheroid (dielectric or metallic), which is an inter-mediary geometry between an ideal nanosphere and S. Acikgoz (&)
Department of Material Science and Engineering, Faculty of Engineering, Karamanog˘lu Mehmetbey University,
Karaman 70100, Turkey
e-mail: sabriyeacikgoz@kmu.edu.tr DOI 10.1007/s00339-014-8771-y
nanocylinder. Klimov has succeeded in deriving some useful theoretical expressions for the radiative decay rates and showed that the molecular lifetime of the excited molecules strongly depends on the size and shape of the substrate particle as well as on the distance of the molecule to the nanospheroid axis. Apart from the nanospheroid size and the separation distance, the orientation of the molec-ular dipole with respect to the prolate nanospheroid surface is also significantly effective on the dipole’s radiative decay rate. It has been demonstrated that the radiative decay rate of a molecule located near a dielectric nano-spheroid is substantially different than the decay rate in a free molecule.
In this work, an alternative experimental way of studying inhibition and enhancement in the spontaneous emission rate is displayed by means of placing fluo-rescing perylene dye molecules into silicon nanopillars. These nanostructures are grown by laser-induced elec-trochemical anodization of p-type silicon wafers at low current densities in a hydrofluoric acid solution. Low cost and easy fabrication process make silicon nanostructures a particularly attractive material which can be used to modify the properties of emitters, by filling the pores with other materials such as liquids, dyes or polymers. With its nanoscale dimensions, silicon has become in-dispensible material for many optoelectronics and light-wave applications, such as solar cells, nanoscale elec-tronic devices, new lasers, waveguides, and chemical and biosensors [12–15]. The modification of the radiative decay rate of perylene dye molecules embedded in sili-con nanopillars is experimentally demonstrated and the distance-dependent interaction between dye molecules and nanostructures is explained using Klimov’s prolate nanospheroid model.
2 Experimental section
2.1 Growth of the silicon nanopillars
Silicon nanopillars are prepared by electrochemical anod-ization of p-type silicon wafers at low current densities in HF:C2H5OH (1:1) solutions under illumination of blue
pulsed diode laser head with wavelength 470 nm (Pico-quant, LDH-C-D-470). Laser power density of pulsed laser is 50.0 mW/cm2. The illumination area of the laser beam is different for each laser used, but it is approximately 11.8–12.6 mm2, which is sufficient to capture the general trend in the nanostructure size gradient. The clean silicon wafers are first cut into pieces, and aluminum contacts are coated as thin films at the back of the samples by evapo-ration in Edwards Coating System, E306A. Copper wires are attached to the aluminum films at the back of the silicon
samples by silver paste. The silicon samples are then immersed into HF:C2H5OH solution. The copper wire is
connected to the positive terminal of the power supply, and the stainless steel is connected to the negative terminal as shown in Fig.1. The current is kept constant during anodization. The current density and also the time it takes for creating the best samples with the most intense pho-toluminescence (PL) are noted, and all samples are pre-pared accordingly. The current density passing through the silicon sample and the etching time are 100 mA/cm2 and 45 min, respectively.
2.2 Time-resolved lifetime and fluorescence intensity measurements
Time-resolved fluorescence lifetime and fluorescence intensity measurements are taken using a Picoharp 300 photon counting instrument (Picoquant, GmbH) and a fiber-optic spectrometer (USB4000-VIS-NIR Ocean Optics), respectively [16]. The optical setup is shown in Fig.2. The excitation source used in the experiment is a pulsed diode laser head with a wavelength of 405 nm (LDH-C-D-405 Picoquant, GmbH). Laser head is driven by a diode driver (PDL 800-B Picoquant, GmbH) at a repe-tition rate of 40 MHz. The separation of the fluorescence emission and the excitation occurs at a dichroic mirror. The dichroic mirror, which is oriented at 45° to laser head, reflects the excitation light vertically toward a microscope objective. The excitation light is focused onto the sample using this microscope objective of 0.70 numerical aperture with a working distance of 10.1 mm (Nikon ELWD 100X). The time-resolved fluorescence spectra are monitored using an avalanched photodiode detector (Micro Photon Devices, SPAD). Appropriate spectral filter is used to block any stray light from the excitation beam. A confocal pin-hole, which has a diameter 75 lm, is placed in the focal Fig. 1 Laser-induced electrochemical anodization setup
plane to exclude out-of-focus background fluorescence. The optical system used in our experimental work is based on a confocal light detection scheme via a 75-lm pinhole in the setup, which allows monitoring the reflected light coming from the very center of the small focused area only. In other words, the possibility of getting illuminations apart from the focal center is eliminated by this pinhole. In the optical setup, piezo-scanner from Piezosystem Jena is used for silicon nanosurface scanning.
For multi-exponential fluorescence decay fitting, FluoFit computer program (Picoquant, GmbH) is used. The fluo-rescence intensity decays are recovered from the time-domain data in terms of a multi-exponential model,
I tð Þ ¼X
n
i¼1
Aiexpðt=siÞ ð1Þ
where Aiis the amplitude of each component and siis its
lifetime. The fractional contribution of each component to the steady-state intensity is described by
fi¼
Aisi
P
jAjsj
: ð2Þ
The intensity-weighted average lifetime is represented as
hsi ¼X
i
fisi: ð3Þ
3 Results and discussion
The size measurements of silicon nanopillars are accom-plished by means of an atomic force microscope (AFM) and scanning electron microscope (SEM). Figure3 shows AFM and SEM micrographs of top view of the conical nanopillars. Bulk crystalline silicon has an indirect energy gap, which prevents radiative recombination under normal circumstances. However, silicon nanostructures show strong PL characteristics in the visible to near-infrared wavelength ranges at room temperature owing to the recombination of quantum confined excitations [17–19]. It is demonstrated that both PL properties of silicon nano-structures are dramatically affected by various factors of the fabrication parameters, such as HF concentration, cur-rent density, etching time and the light source used for illumination [20–22].
The silicon nanostructures are impregnated by perylene dye molecules. Perylene is a brown crystalline polycyclic aromatic hydrocarbon with the chemical formula C20H12. It
is soluble in most of the organic solvents, and its blue fluorescence in solution can be seen even in visible light. High fluorescence quantum yield, photostability and effi-cient carrier mobility also make perylene an excellent candidate for use in organic electroluminescent diodes [23]. The fluorescence emission spectrum of a dilute
Fig. 2 Time-resolved fluorescence lifetime spectroscopy setup
Fig. 3 aAFM and b SEM image of the silicon nanopillars
perylene solution is characterized by an ensemble of three major vibronic bands with well-defined peaks at 450, 480 and 510 nm, respectively. This spectrum is essentially a structured mirror image of the absorption spectrum, as would be expected for relaxation from the excited singlet states of isolated molecules. Perylene is dissolved in methanol, and the dilute solution is deposed on the surface of a sample by a syringe. After wetting, the penetration of dye molecules is studied by spectroscopy, and the results show that the dried dye uniformly covers the nanopillars. The PL spectrums of perylene and silicon nanostructures are given in Fig.4.
The time-resolved fluorescence lifetime of the perylene dye molecule is measured using the Picoharp 300 system based on the time-correlated single photon counting (TCSPC) method. In this method, the time between the excitation laser pulse (start signal) and the detected single
photon of the fluorescence (stop signal) is measured. The measured data are plotted as a fluorescence lifetime his-togram. Decay parameters are determined using the dou-ble-exponential tailfit model, and the best fits are obtained by minimizing v2values. In order to compare the fluores-cence dynamics of perylene dye on the nanostructured silicon with the respective dynamics on non-nanostructured silicon surface, the same perylene solution is also deposed on bulk silicon wafer as a thin film. At dilute concentration, perylene exhibits monomer emission, and the intensity-weighted fluorescence lifetime (s0) of perylene monomer
band is measured as 3.631 ns as shown in Fig.5. There-fore, the decay rate (c0) of the emission which is defined as
the inverse of the fluorescence lifetime becomes 0.28 ns-1. The nanostructured silicon sample is placed on a piezo-nanopositioning stage to allow an excitation from the side section of the sample. Firstly, the top surface of the sample is illuminated by pulsed diode laser, and the fluorescence intensity of a perylene dye molecule is also controlled using a fiber-optic spectrometer. Our optical system is based on a confocal light detection scheme via a pinhole in the setup, which allows monitoring the reflected light coming from the very center of the small focused area only. In other words, the possibility of getting illuminations apart from the focal center is eliminated by this pinhole. When the tip of a silicon nanopillar is illuminated, the fluores-cence intensity reaches a maximum value. The value of y axis of the piezo-nanopositioner is recorded. In this work, a nanopillar is treated as nanospheroid, and the position of its equator can be determined using the AFM image. Then, the y axis is adjusted to equator of the silicon Fig. 4 Photoluminescence spectrums of the perylene and silicon
nanopillars
nanospheroid, and the fluorescence intensity measurement is used to adjust z axis of the nanopositioner. Therefore, the radiative decay time of perylene dye molecules which are attached to different locations of the silicon nanostructures are measured by changing the z axis of the nanopositioner. It is observed that the presence of semiconductor interfaces influences the radiative transition frequencies and decay rates. As can be seen in Fig.6, the modification rate of the fluorescence lifetime of perylene dye molecules depends on the position of fluorescent molecule with respect to semiconductor surface.
Silicon nanopillars can be treated as almost prolate nanospheroid structures as depicted in Fig.7a. Solution of the quasi-static problem allows one to express the total decay rate of an atom, which is placed near a prolate nanospheroid, as the sum of radiative, waveguided and nonradiative components. In the radiative decaying pro-cess, the excitation energy is converted to a photon, which may escape away from the surrounding body. If the exci-tation energy is transferred to a photon that is localized inside the body, guided decay rate changes. Nonradiative decay rate becomes effective for those dielectrics which have a complex dielectric permittivity, in which case
excitation energy transforms to the thermal heating as Joule loss. In the presence of a prolate nanospheroid with a real permittivity, only radiative decay rate shows a remarkable change. Klimov has successfully derived an analytical procedure to explain the distance dependence of the spontaneous emission decay rate for prolate nanospheroid nanostructures [11]. The geometry of the problem is given in Fig.7b. In this theoretical model, Cartesian coordinates x, y and z are transformed to n, g and u coordinate system, where n C 1Cg C -1 and 0 B u B 2p.
Radiative decay rate of an atom in the vicinity of this prolate nanospheroid depends on several parameters: ori-entation of the atom with respect to the nanospheroid’s surface, the distance between the atom and prolate nano-spheroid axis (n), and the distance between the tip of the nanospheroid and its equator (g). Radiative decay rates for different dipole orientations are expressed as,
c c0 n ¼ n 2 1 n2 g2 g2 1þ G 10 d dnQ1ð Þn 2 " þ 1 g 2 n ffiffiffiffiffiffiffiffiffiffiffiffiffi n2 1 p þ G11 d dnQ 1 1ð Þn !23 5 ð4Þ Fig. 6 a Fitting and calculation of decay parameters of perylene
located on (blue) the equator of silicon nanopillar and on (green) the tip of silicon nanopillar. Solid lines indicate multi-exponential fitting
curve. Residuals for fittings on b the equator of silicon nanopillar, cthe tip of silicon nanopillar
c c0 g ¼ 1 n2 g2 g2 n2 1 1þ G11 Q1 1ð Þn ffiffiffiffiffiffiffiffiffiffiffiffiffi n2 1 p 2 2 4 þ 1 g 2n2 1þ G10 Q1ð Þn n 2# ð5Þ c c0 / ¼ 1 þ G11 Q1 1ð Þn ffiffiffiffiffiffiffiffiffiffiffiffiffi n2 1 p 2 ð6Þ where Q1ð Þ ¼n n2lnnþ1n1 1 and Q11ð Þ ¼n ffiffiffiffiffiffiffiffiffiffiffiffiffi n2 1 p dQ1ð Þn dn are
the Legendre functions of the second order. The coeffi-cients of G10and G11 are defined as
G10¼ e 1 ð Þn0 n0dnd0Q1ð Þ eQn0 1ð Þn0 G11¼ e 1 ð Þn0 ffiffiffiffiffiffiffiffiffiffiffiffiffi n20 1 q d dn0Q 1 1ð Þ enn0 0dnd0Q1ð Þn0 ð7Þ
where e is the dielectric permittivity constant, n0¼
a. ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffipa2 b2 and n¼ z. ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffipa2 b2. Since the above
radi-ative decay rate equations are quite complicated, these equations may be simplified for the following particular cases:
1. if the atom is located on the axis of the spheroid, g = 1,
2. if the atom is located near the equator of the spheroid, g = 0. Thus, c c0 n;g¼1 ¼ 1 þ G11 d dnQ1ð Þn 2 ð8Þ c c0 g;g¼1 ¼ 1 þ G11 d dnQ1ð Þn 2 ð9Þ c c0 n;g¼0 ¼ 1 þ G11 ffiffiffiffiffiffiffiffiffiffiffiffiffi n2 1 p n d dnQ 1 1ð Þn 2 ð10Þ c c0 g;g¼0 ¼ 1 þ G10 Q1ð Þn n 2 ð11Þ
A relationship between the relative decay rate and the distance of the dye from the surface of the dielectric nanospheroid is explained with the model given in the theory part above. Klimov’s theoretical model can be simplified for two special cases of the dye molecule with respect to the prolate nanospheroid vertical axis. While dye molecule is located on the equator of the nanospheroid for g = 0 case, the molecule is located on its tip for g = 1 case. In our experiments, both cases are experimentally studied, and a fairly good agreement is demonstrated between the theory and experimental results. As depicted in Fig.7, a perylene dye molecule located nearby a distance n according to the axis of the nanospheroid. The perylene molecule consists of two naphthalene molecules connected by a carbon–carbon bond at the 1 and 8 positions on both molecules, with usually long bonds of 0.15 nm in length observed between the two halves of the molecule. All of the carbon atoms in perylene are sp2-hybridized. One can notice from the chemical structure of the perylene that the dye molecules are physically attached to silicon nanopil-lars, and they can be easily embedded into the deeper parts of the pillars. The distance n is assumed to be from the center of mass (CM) of the perylene molecule to the sur-face of the spheroidal silicon nanopillar which ranges from 1.1 up to 2.5. The distance between the axis of the nano-spheroid and its surface, which is defined as n0, is
calcu-lated from the AFM picture of our porous silicon sample shown in Fig.3. Prolate spheroid semi-axes a and b (a [ b) are measured to be approximately 41.0 and 16.4 nm, respectively. Thus, n0 is calculated to be 1.1.
Moreover, the dielectric constant of nanospheroid surface is taken as 4.5. The distance (n - n0) is the critical
parameter by which one can monitor the change in the radiative decay rate for the dipole–nanospheroid system. For n = n0 distance, the influence of the prolate
nano-spheroid on the radiative decay rate of perylene molecule is Fig. 7 aCross-sectional view
of a silicon nanopillar shows how it is treated as a prolate nanospheroid. b Geometry of the problem
the greatest. As the distance n increases, the influence of the prolate nanospheroid on the radiative decay rate decreases and the decay rate value approaches to that of the free space one. For g = 0 case, decay parameters of per-ylene dye molecules are analyzed using double-exponential fitting model, and calculated lifetime values are summa-rized in Table1. It is found that the decay rate of the g-oriented perylene molecules with respect to the spheroid surface is inhibited up to 0.15 ns-1. Figure 8 shows experimental results and calculations for relative linewidth c=c0 versus distance (n - n0) for g-oriented dipole of the
dye molecules with respect to spheroids surface, using Eq.11. A comparison is given with various values of (n - n0) distance, ranging from 0 to 2. It is clearly seen
that our experimental results for the decay rates of the perylene dye molecules placed on the equator of the prolate spheroid are in good agreement with the Klimov’s theo-retical model.
To ensure the appropriateness of the Klimov’s theoret-ical model for our experimental results, the dye molecules are also located on the tip part of the prolate nanospheroid, that is, g = 1 case. To demonstrate such an experimental configuration, the dye molecules are restricted to be on the tip of the prolate nanospheroid by means of filling the silicon pores with a polymer layer, like polyethylene glycol (PEG), and then coat the surface of this polymer with the dye molecules using a dilute chemical solution. It is
observed that the decay rate of perylene dye molecules which are closely located from the tip of the nanospheroid is enhanced up to 1.01 ns-1. The calculated decay lifetimes of these molecules are summarized in Table2. As can be seen in Fig. 9, the radiative decay rate of perylene dye molecules located on the tip of the prolate nanospheroid is enhanced, and our experimental results are convenient with Eq. 6. After filling the silicon pores with PEG polymer, results confirm that the perylene dye molecules seem to become n-oriented with respect to prolate nanospheroid.
4 Conclusion
Radiative decay rate of perylene dye molecules attached to silicon nanopillars is investigated by means of the time-resolved fluorescence lifetime spectroscopy method. Nanopillars are prepared by laser-induced electrochemical anodization of p-type silicon wafers, and these nanostruc-tures are assumed to have almost prolate nanospheroids. The ellipticity of a spheroid is used to effectively control the spectroscopic characteristics of atoms placed near them. It is experimentally demonstrated that the radiative decay rate of the perylene dye molecules placed in the close vicinity of a prolate nanospheroid may increase or decrease, depending on the dipole orientation and its position to the axes of the nanospheroid. Because of the
Fig. 8 Comparison of the results with the Klimov’s prolate nano-spheroid model for g = 0 case
Table 2 Decay parameters of perylene dye molecules for g = 1 case
z (nm) n n - n0 s (ns) c (ns-1) c=c 0 45 1.20 0.10 0.99 1.01 3.61 55 1.46 0.36 1.89 0.53 1.89 65 1.73 0.63 2.65 0.38 1.35 75 1.99 0.89 3.22 0.31 1.11 95 2.53 1.43 3.61 0.28 0.99
Fig. 9 Comparison of the results with the Klimov’s prolate nano-spheroid model for g = 1 case
Table 1 Decay parameters of perylene dye molecules for g = 0 case
z (nm) n n - n0 s (ns) c (ns-1) c=c0 42 1.12 0.02 6.50 0.15 0.55 52 1.38 0.28 4.55 0.22 0.78 62 1.65 0.55 3.98 0.25 0.90 72 1.91 0.81 3.85 0.26 0.93 92 2.45 1.35 3.67 0.27 0.97
fact that the size of our nanopillar and distance between dye and the spheroid are much smaller than the radiation wavelength, our results can be explained using the quasi-static approach. By comparing our data with a theoretical model, the mechanisms responsible for the observed changes can be identified. The changes in the spontaneous emission rate appear due to plasmon resonance described by the Klimov’s model. The radiative decay rate of the dye molecules attached to the equator part of the silicon nanopillars is inhibited. On the other hand, the decay rate is enhanced for the dye molecules placed on the tip of the nanopillars.
Acknowledgments This work was supported by TUBITAK under contract number 106T011 and Karamanog˘lu Mehmetbey Universities Research Funds under contract numbers 01-M-13.
References
1. G.S. Solomon, M. Pelton, Y. Yamamoto, Phys. Rev. Lett. 86, 3903 (2001)
2. P.T. Worthing, W.L. Barnes, J. Opt. A: Pure Appl. Opt. 1, 501 (1999)
3. K.H. Drexhage, J. Lumin. 1, 693 (1970)
4. M.Y. Berezin, S. Achilefu, Chem. Rev. 110, 2641 (2010) 5. S. Astilean, S. Garrett, P. Andrew, W.L. Barnes, J. Mol. Struct.
651, 277 (2003)
6. E.C. Wu, J.H. Park, J. Park, E. Segal, F. Cunin, M.J. Sailor, ACS Nano 2, 2401 (2008)
7. R. Bardhan, N.K. Grady, J.R. Cole, A. Joshi, N. Halas, J. ACS Nano 3, 744 (2009)
8. R. Carminati, J.-J. Greffet, C. Henkel, J.M. Vigoureux, Opt. Commun. 261, 368 (2006)
9. C. Dekker, Phys. Today 52, 22 (1999)
10. V.V. Klimov, M. Ducloy, Phys. Rev. A 69, 013812 (2004) 11. V.V. Klimov, M. Ducloy, V.S. Letokhov, Eur. Phys. J. D 20, 133
(2002)
12. V. Sivakov, G. Andra, A. Gawlik, A. Berger, J. Plentz, F. Falk, S.H. Christiansen, Nano Lett. 9, 1549 (2009)
13. K.S. Dancil, D.P. Greiner, M.J. Sailor, J. Am. Chem. Soc. 121, 7925 (1999)
14. N.M. Park, T.S. Kim, S.J. Park, Appl. Phys. Lett. 78, 2575 (2001) 15. A.G. Cullis, L.T. Canham, P.D.J. Calcott, J. Appl. Phys. 82, 909
(1997)
16. S. Acikgoz, I. Sarpkaya, P. Milas, M.N. Inci, G. Demirci, R. Sanyal, J. Phys. Chem. C 115, 22186 (2011)
17. L.T. Canham, Appl. Phys. Lett. 57, 1046 (1990)
18. N.P. Mandal, S. Dey, S.C. Agarwal, Thin Solid Films 451, 375 (2004)
19. X.J. Li, Y.H. Zhang, Phys. Rev. B 61, 12605 (2000) 20. Z. Gaburro, H. You, D. Babic, J. Appl. Phys. 84, 6345 (1998) 21. M.C. Cheung, P.J.R. Roche, M.H. Hassan, A.G. Kirk, Z. Mi, V.P.
Chodavarapu, J. Nanophoton 5, 53503 (2011)
22. N. Chiboub, N. Gabouze, J.N. Chazalviel, F. Ozanam, S. Moulay, A. Manseri, Appl. Surf. Sci. 256, 3826 (2010)
23. B. Ding, W. Zhu, X. Jiaing, Z. Zhang, Solid State Commun. 148(5), 226 (2008)
