[3] Diederik S.Wiersma, Paolo Bartolini, Ad Lagendijk & Roberto Righini, Nature, Vol.390(1997), p.671. [4] E.Yablonovich. J Modern Optics,Vol.41(1994), p.173.
Solid State Phenomena Vol. 94 (2003) pp 299-304 online at http://www.scientific.net © (2003) Trans Tech Publications, Switzerland
Spectral Properties of the Dye Activated Polymers with Added Fine Scattering Particles. E.Tikhonov1, Vasil P.Yashchuk2, O.Prygodjuk2, V.Bezrodny1, Yu.Filatov3, V.Bindjukevich2. 1
Physics Institute of the Ukrainian Academy of Science; 46, Nauky av., 02650 Kyiv, Ukraine; Physics Department of Kyiv Taras Shevchenko University; 6,Glushkov av., 03680 Kyiv, Ukraine; 3 Institute of Superhard Materials of the Ukrainian Academy of Science; 2, Avtozavodska St., 04074 Kyiv, Ukraine.
2
Keywords: scattering medium, rhodamine 6G, polymer, luminescence, dimers, associates.
Abstract. An investigation of the luminescence spectra (LS) and excitation spectra (ES) of R6G in polyvinylacetate (PVA) matrix with embedded fine-dispersed particles of SiO2 and Al2O3 was undertaken. The LS were decomposed on elementary component (EC) by means of Alentcev-Fock method. It is shown that the EC correspond to monomers, dimers and higher associates. The concentration ratio of these luminescence centres depends on particle concentration. The mechanism of the particle influence on the centre redistribution has been proposed. Introduction Investigation of the strong scattering media is of great interest because of recent observation of superluminescence [1,2] and photon localization [3], which occur at high particle concentration. Moreover, in the case of ordered media (photonic crystals) the stop-band effect is possible, which consists in a considerable decreasing of the transmittance and luminescence intensity in a certain spectral region [4-6]. To gain a complete understanding of these phenomena it is necessary to study the influence of the fine-dispersed particle embedded on the spectral properties of the luminescence medium. In the case of a random medium the influence can be conditioned by the particles’ influence on the creation and quantum yield of luminescence centres and by the interference of scattered waves. The investigations reported in literature were carried out with liquid dye solution and suspended scattering particles. Because of the probability of particles settling, using liquid solutions is complicated for research. Therefore, the use of solid matrices, which avoids the problem, is beneficial. The luminescence spectra (LS) and excitation spectra (ES) of rhodamine 6G (R6G) in solid polyvinylacetate (PVA) matrix with embedded fine particles SiO2 (d~0.5mm), and LS of the same samples with Al2O3 (d~2mm) were investigated using different particle concentrations. The particle weight concentration varied from 6% to 30%. In the subsequent discussion we will symbolize samples with particle concentration C by RsiC (rhodamine 6G with SiO2) and RalC (rhodamine 6G with Al2O3). The dye concentration was Cd=2*10-4M/l. The excitation beam was normal to sample surface. The luminescence was observed at 30o to the excitation beam in opposite direction. The luminescence was excited with a mercury lamp (l=546.1nm) and incandescent lamp (a band Dl~5nm was selected with double monochromator). Results and discussion Fig.1a, shows that LS depends considerably on particle concentration. The LS of each sample depends on the excitation wavelength (Fig.1b) and the ES depends on observation luminescence wavelength (Fig.1c). These results testify that LS are non-elementary.
All rights reserved. No part of contents of this paper may be reproduced or transmitted in any form or by any means without the written permission of the publisher: Trans Tech Publications Ltd, Switzerland, www.ttp.net. (ID: 130.203.133.33-15/04/08,01:17:50)
300
Interfacial Effects and Novel Properties of Nanomaterials
To determine the reason for this non-elementary nature and influence of the scattering particles on LS shape it is necessary to study the LS structure. Therefore it is beneficial to decompose these spectra on the elementary components (EC). The decomposition was carried out by means of the Alentcev-Fock method [7]. The method is suitable for the set of linearly independent spectra. The special convenience of this method is that any information about the shape of elementary
0,4 0,2
0,6 1
0,4 0,2
550
0,6 0,4 0,2 0,0
600
c) dE/dl, rel.un
0,8
1 2 3
l, nm 0,0
0,0
1,0
b)
0,8 dE/dl, rel.un
0,6
1,0
5 34 2
650
1 2
l,nm
550
700 1,0
d)
0,8
3
0,6 0,4 0,2 l,nm
400 450 500 550 600
dE/dl, rel.un
0,8
a) dE/dl, rel.un.
1,0
600
650
4 3 2
1
0,0
l,nm
550
600
650
Fig. 1 LS of R6G in PVA matrix under the different SiO2 particle concentration: (1) – sample without particles; (2) - Rsi06, (3) – Rsi18, (4) – Rsi24, (5) - Rsi30. b) LS of Rsi24 at the excitation wavelength: lex»551nm(1), 530nm(2), 475nm(3). c) ES of Rsi30 at the luminescence wavelength: llum»580nm(1), 630nm(2), 615nm(3). d) LS of the polymer sample with R6G concentration CR6G=2*10-4M/l without particles at the excitation wavelength: lex»495nm(1), 518nm(2), 555nm(3); and LS of the sample with R6G concentration CR6G=10-5M/l(4).
components unnecessary. This is of considerable importance for dye spectra, because dye spectra, as well as their associate spectra, are asymmetric. So elementary components of the spectra being studied should also be asymmetric. If it is supposed, that particles addition influences the ratio of the different EC (but does not influence their shape), then LS presented on Fig.1a will form a series of linearly independent spectra. Using this method, qualitative conclusions concerning the LS structure can be done by analysis of different LS ratio gij=Ii(l)/Ij(l) (Fig.2), where the Ii and Ij are LS of the samples with different particle concentration Ci and Cj. Referring to Fig.2, all curves can be divided into three specific portions. Two of them lie on the wings of the LS (l605nm). Ratio gij does not depend on wavelength over these ranges (so we will name them as “horizontal” parts). The third range coincides with the central part of the LS (l»555¸605nm). Over this spectral range curves gij(l) depend on wavelengths in different ways. Some of them have an extremum and the other ones changing monotonically. This provides the possibility to suggest that there are three EC and their locations coincide with these specific portions of gij(l). If spectral bands of the EC are symbolized as L1, L2, L3 (according to the order of maximum wavelength), the LS is exhibited as:-
Solid State Phenomena Vol. 94
301
3
aik Lk (l ) .
Ii =
(1)
k =1
In terms of Eq.1 gij(l) can be determined as 3
I (l ) g ij (l ) = i = I j (l )
a ik Lk (l ) k =1 3
,
(2)
a jk Lk (l ) k =1
dE/dl, normalised data
Spectra relation g(l),rel.un.
where coefficients aik are amplitudes of Lk in the spectrum Ii(l). So, these “horizontal” parts of gij correspond to the spectral region where components Lk are not overlapped. According to the Alentcev-Fock method the Lk(l) shapes can be obtained by consecutive elimination of the other components Lm (m¹k) from the LS. 1,6 The elimination can be carried out by 1,5 4 means of linearly independent spectra 1,4 subtraction. To eliminate component 1,3 Lm from the spectrum Ii(l), the other 1,2 spectrum Ij(l) multiplied by 1,1 coefficient gijm should be subtracted 1,0 1 from the Ii(l). This coefficient gijm is 0,9 2 equal to “horizontal” portion value gij 0,8 3 over the Lm location range. 0,7 0,6 l,nm The Lk spectral shapes so obtained are 0,5 520 560 600 640 given in Fig.3. To receive three components only three linearly independent spectra are necessary. Thus, the other LS can be used to Fig. 2 Spectral dependence of the ratio of the samples LS: Rsi06 and Rsi18(1), Rsi30 and Rsi24(2), Rsi24 and receive all of the components from the Rsi06(3); and LS of Rsi30(4). other spectra combinations (Fig.3). The EC received by different ways are similar. This similarity confirms the L2 L1 L3 reliability of the obtained 0,8 decomposition and indicates that all LS consist of the EC with the same parameters. So the component’s 56 parameters are independent of the 0,4 particle concentration. Consequently, 12 the influence of interference and 34 reabsorption on EC spectral shape is 0,0 l,nm inessential, because this influence 520 560 600 640 should depend on the particle concentration by means of varying of the phase shift between scattered Fig.3 Comparison of the EC shapes, obtained from waves and changing the effective path different spectra combinations: 1 —L1 (Rsi24, of photons through the sample. Rsi18), 2— L1 ( Rsi18, L2, L3); 3— L2 (Rsi18, According to the observed data, Rsi06), 4 — L2 (Rsi30, Rsi24); 5 — L3 (Rsi24, shown in Fig.3, the decomposition Rsi06), 6 — L3 (Rsi18, Rsi12, L2). errors of the band peak wavelengths
302
Interfacial Effects and Novel Properties of Nanomaterials
are dlmax»5nm, 1nm and 3nm, of the bandwidth are 3nm, 1nm and 6nm for L1, L2, L3 respectively. To carry out the decomposition of LS we used the averaged envelopes of the component bands. According to the results, deviations of calculated LS shapes (Eq.1) from the experimental do not exceed 1-2% over the whole wavelength interval. Such a small deviation value is conditioned by the particularly cancellation of component parameters errors under EC summarizing in Eq.1.
The second cause of concentration dependency of total energy ES(C) is a redistribution of luminescence centre concentrations with increased particle concentration C. The effect is significant at high particle concentration. While energy of L2 and L3 increase, the L1 component energy decreases. This dependency can be
dE/dl,rel.un.
Energy, rel.un.
As can be seen from the Fig.1d, the LS of pure R6G sample is non-elementary too, because its shape depends on the excitation wavelength. The decomposition of this spectrum gives the same components as the components of the samples with embedded particles. Thus all the spectra (including the pure one) contain the same components Lk and the particles influence is in changing the concentration ratio of different 70 luminescence centres. So the EC are connected 60 with the luminescence centres, that are formed by Total energy 50 dye molecules in pure polymer matrix [8]. Thus, they probably are monomers, dimers and 40 associates of higher degree. The accordance of the L3 30 luminescence centres with elementary components 20 as well as the way that particles influence the LS L2 can be found by analysis of the energy dependence 10 L1 on SiO2 concentration (Fig.4). As can be seen, the concentration, % 0 total energy curve (ES) increases approximately 0 6 12 18 24 30 two times when the least particle concentration (6%) is embedded. Then ES increases slightly as Fig. 4 Total and EC energies versus the concentration increases from 6% to 24%. When SiO2 concentration the particle concentration is 30%, the total energy increases 1.4 times. The behaviour of the energy dependence proves that there are at least two reasons for the total energy increasing. The first most probable causes of this energy increase is multiple scattering of radiated and excitation light. Multiple scattering of excitation light increases the portion of the absorbed radiation by means of the photon pass increasing. So, the optical density of the sample rises which is confirmed by the flat-topped change of the excitation spectra shape with embedding of the particles. According to [9], this is typical for such situations. Multiple scattering of radiated light increases portion of the luminescence photons observed in backward direction in accordance with L1 L2 L3 the experimental setup. For these / 1,0 2 photons the effect of multiple 2 / 1 1 scattering is analogous to diffuse reflector with reflection coefficient about unity. 0,5 3
3
/
0,0 550
600
650
Fig.5 Comparison of the component shapes Lk {k=1¸3} obtained for the samples with SiO2 particles (1,2,3) and with Al2O3 particles (1/, 2/, 3/)
Solid State Phenomena Vol. 94
303
explained by the increased concentration of the second and the third luminescence centre types, which correspond to the components L2 and L3, while the first centre type (associated with L1) concentration decreases. Energy values of the second and third components are higher than the first component one. This can be explained by difference in centre concentrations. Because L1 energy decreases with particle concentration growth, the L1 should be correspond to the monomer, and the other two components (as their energies increase) correspond to dimer and higher associate.
dE/dl,rel.un.
The probable mechanism of the scattering influence on the redistribution is the reduction of the dimension geometry of the dye molecule location. This change results in modification of the functional relationship between the dye molecular concentration n and the distance between these molecule r (r~n-1/3 for 3D 1 1,0 geometry and r~n-1 for quasi-linear 23 0,8 geometry). This assumption is supported by 4 results obtained for the porous glass [10]. The 0,6 location geometry of high concentrated samples is not identical to geometry in porous 0,4 glass. But if the particle concentration grows, the local geometry of dye molecule location 0,2 between particles approaches to quasi-linear 0,0 (or quasi-planar) geometry. Thus, increasing l,nm particle concentration under the same value n 550 600 650 leads to decreasing of the effective distance reff between molecules. This decrease, firstly promotes the dimers creation and then higher Fig. 6 LS of R6G in PVA under the different associates one. According to this assumption, scatterer materials: CeO2(1), synthetic diamond component L3 should be matched with (2), Al2O3(3), SiO2(4). dimers, because its energy increase begins at lower particle concentration than for L2. Consequently, the second component L2 is connected with higher associate radiation. The correspondence between the components and luminescence centers is confirmed by the short-wave shift (to the L1 spectral position) of the pure sample LS under the dye concentration decreasing (Fig.1d). An additional effect of the embedded particles on the LS is the EC frequency shift, which depends on the particle material. This effect is demonstrated by decomposition of LS of the R6G in PVA samples with embedded Al2O3 particles. The surface activity of Al2O3 is considerably lower than the SiO2 surface one. According to the data (Fig.5) the EC of both samples have a similar shape, but the all EC of the LS of the samples with corundum particles are short-wave-shifted. This shift value is greater than the decomposition error. This effect substantiates the influence of the particle surface on the luminescence centers. Moreover, LS of R6G in the polymer samples depends on the particle material (under the same weight concentration). In according with Fig.6, the dye LS maximum can be considerably shifted (30nm) by varying of the embedded particle material. The probable reason for this shift is a redistribution of the monomer, dimer and higher associate concentrations in the sample. From our point of view, it is conditioned by different interspace between particles and, as a consequence, by different dimensions of the location geometry of the dye molecule in these interspaces. The SiO2 density r and particle size d are considerably lower than the corundum, diamond and CeO2 ones (r = 2.65, 4.0, 3.5, 6.26gr/sm3, d =0.5, 2, 7, 1 mm respectively). So the effective interspace size in the samples with SiO2 is considerably smaller than in the other samples under the same weight concentration. The dependence of the LS maximum of R6G in a polymer matrix on the concentration, size, and material of the scattering particles may be used to redistribute the emitted energy within LS spectral region.
304
Interfacial Effects and Novel Properties of Nanomaterials
Conclusions This paper is devoted to the investigation of the structure of the luminescence spectra (LS) of R6G in multiple scattering polymer matrix. It is found that LS of R6G in polymer matrix with embedded fine-dispersed particles are complicated, spectrally inhomogeneous, and consist of three elementary components. They seemingly correspond to the luminescence of monomers, dimers, and higher degree associates. The increasing of scattering particle concentration redistributes the luminescence centre concentration in favor of dimers and higher associates. The probable mechanism of this redistribution consists in variation of location geometry of the dye molecules and its interaction within the interspace between the scattering particles. Moreover, there is interaction between the particle surface and dye molecules, which manifest itself in spectral shift of the elementary components. Thus, the wavelength of luminescence spectrum maximum can be varied by means of variation of the concentration, size and material of the scattering particles. References [1] N.M.Lawandy, R.M.Balachandran, A.S.L.Gomes, E.Sauvain, Nature, Vol.368(1994), p.436 [2] A.Z.Genack & J.M.Drake, Nature, Vol.368(1994), p.400 [3] Diederik S.Wiersma, Paolo Bartolini, Ad Lagendijk & Roberto Righini, Nature, Vol.390(1997), p.671 [4] E.Yablonovich. J Modern Optics,Vol.41(1994), p.173 [5] . . , . . , . . , . . , , Vol.63(1996), p. 496 [6] Hery P. Schremer, Henry M.van Driel, A Femius Koenderik and Willem L.Vos, Phys Rev. A, Vol.63, 011801(R )(2000), p.1 [7] . ., . , Vol.59(1972), p.3 [8] . ., . ., . ., , Vol.54(1991), 3, .418 [9] . ., . ., . ., . . ., Vol.45(1986), 4, p.612 [10] . ., . . ., . ., Vol.66(1989), 1, .120.
!
&
'
%
"
!
#
!
+
'
,
-
$
(
)
,
*
$
-
$
