- Email: [email protected]
Dispersion studies and electronic transitions in nickel phthalocyanine thin films
ARTICLE IN PRESS
Optics & Laser Technology 37 (2005) 513–523 www.elsevier.com/locate/optlastec
Dispersion studies and electronic transitions in nickel phthalocyanine thin films M.M. El-Nahass, A.M. Farag, K.F. Abd El-Rahman, A.A.A. Darwish Faculty of Education, Ain Shams University, Heliopolis, Roxy, Cairo 11757, Egypt Received 29 September 2003; received in revised form 16 July 2004; accepted 27 August 2004
Abstract Thin films of the organic semiconductor nickel phthalocyanine (NiPc) are structurally investigated using X-ray diffraction and infrared light absorption. The optical absorption and dispersion studies of nickel phthalocyanine were investigated using spectrophotometric measurements of transmittance and reflectance at normal incidence in the wavelength range 190–2100 nm. The absorption spectra recorded in the UV-VIS region show two well-defined absorption bands of the phthacyanine molecules, namely, the Soret and the Q-band. The Davydove splitting of the main absorption peak in the metal phthalocyanines correlates with the relative tendencies of the metal to out-of-plane bonding. The refractive index n as well as the absorption index k were calculated and showed an independent of the film thickness in the film thicknesses range 400–770 nm. The refractive index n showed an anomalous dispersion in the absorption region as well as normal dispersion in the transparent region. Some of the important optical absorption such as the molar extinction coefficient, the oscillator strength, the electric dipole strength have been evaluated. The analysis of the spectral behavior of the absorption coefficient a in the absorption region revealed two indirect allowed transitions with corresponding energies 2.7770.03 and 1.6670.02 eV. An energy band diagram has been proposed to account for the optical transitions of NiPc thin film. All previous parameters were as well obtained for films annealed at 523 K for 2 h. Discussion of the obtained results and their comparison with the previously published data are also given. r 2004 Elsevier Ltd. All rights reserved.
1. Introduction Phthalocyanines are porphyrin derivatives, characterized with high symmetry, planarity and electron delocalization [1]. Phthalocyanines have attracted lots of attention for applications in organic optoelectronic devices [2]. Some of the important applications of phthalocyanines are in fabrication of solar cells, electronic displays, chemical sensors, photocopying and laser printers and optical date storage systems [3–5]. In addition to these applications, being a very good absorber of light in the UV-VIS region they are widely used as excellent laser dyes capable of optical amplification in the red region [6]. Phthalocyanines are generally p-type semiconductors and have the advantage Corresponding author.
E-mail address: [email protected] (A.M. Farag). 0030-3992/$ - see front matter r 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.optlastec.2004.08.016
of being sufficiently stable toward chemical and thermal treatments [7]. Most phthalocyanine compounds are insoluble in common organic solvents and so it is not possible to prepare them by solution casting techniques [8]. Hence, very often, thin films of phthalocyanines are made by a vacuum evaporation technique. This method has the advantage of producing high purity films without decomposition [7]. Growth and morphology of the various phthalocyanine films have been extensively studied [9–12]. Phthalocyanine films can exist in several crystalline polymorphs. Most common polymorphs are metastable a and stable b phases. The main differences between different polymorphs are the tilt angle of the molecules within the columns (stacks of molecules with molecular planes parallel to each other) and the mutual arrangement of the columns [2]. Transformations from a phase to b phase using successive sublimation [9,10] or annealing or deposition
ARTICLE IN PRESS 514
M.M. El-Nahass et al. / Optics & Laser Technology 37 (2005) 513–523
at higher temperatures [11,12] have been reported. The absorption spectra of different polymorphs of some phthalocyanine compounds show significant differences among each other [11]. There is a large portion of reports appearing, focusing on the optical properties of various metal substituted phthalocyanines. The visible and near ultraviolet absorption spectra have been measured previously by Davidson [13] for thin films of H2Pc, MgPc, FePc, CoPc, CuPc, and ZnPc. Schmeisser et al. [14] have found that the lowest absorption band at 1.73 eV due to transitions between the HOMO p and the LUMO p molecular orbitals, which due to the size of the 18-p electron systems are well separated in energy from molecular orbital structure of the benzene and aza group. They recorded a considerably smaller energy gap for PbPc than of H2Pc from the onset of the absorption spectrum to be 1.2 eV. According to this result, the band-to-band absorption has taken place. The UV photoemission spectrum, which was studied by Mockert et al. [15], shows a peak at 2.4 eV. This peak might originate from the direct band-to-band transition. Ukei [16] suggested a model where the electronic conduction is due to charge transfer from spz mixing orbit to the p electron system of the macrocyclic ring of a phthalocyanine. The direct electronic transition might be of the p to p type. The applicability of the band model to explain conduction in metal free phthalocyanine is suggested by the equality of the thermal activation energy for electrical conduction, the energy for first optical absorption, and the energy for the initiation of photoconductivity [17]. Further support for the band model is furnished by the similarities between phthalocyanine and anthracene since evidence has been advanced for the band-model approach in anthracene [18]. Ambily and Menon [19] have studied the optical properties of CuPc thin films. They suggested that the direct electronic transition from p2p orbital in the range 300–450 nm results in an intense band called the Soret-band which gives the absorption edge and the Qband in the 500–750 nm range yields two optical trapping levels. Bialek et al. [20] calculated HOMOLUMO gap of NiPc as 2.41 eV and suggested that orbitals next to the HOMO are separated from each other by 1.84, 2.44, 0.17, 0.28 and 0.09 eV, and the energy level of the second unoccupied MO is widely separated from the LUMO, and differences become smaller for the following orbitals. Although NiPc, like other metal-substituted phthalocyanines, is a subject of various investigations [21,22] considerably less attention has been paid on its structural and optical characterizations. A complete description of the optical properties requires a knowledge of a set of values at least two optical constants (n and k). In general, the optical and certain electronic properties are only specified if the optical constants are known. Hence phthalocyanines
have been investigated as organic laser materials; a knowledge of the refractive index is essential in determining the population inversion density required for lasing. Also, information about the actual spectral absorption characteristic of phthalocyanine pigments is required to analyze the action spectra of solar cells. The present work is an extension of these studies aimed at characterizing the optical properties of NiPc thin films. The specific aims are related to an improved method for the determination of the optical constants from the experimental transmittance (T) and reflectance (R) measurements, and also to further investigate the applicability of a band model approach for the description of this material.
2. Experimental details The nickel phthalocyanine (NiPc) powder used in this study is obtained from Kodak, UK. Thin NiPc films of different thicknesses are prepared by conventional thermal evaporation technique, using a high vacuum coating unit (Edwards type E 306 A, England). The films are deposited onto precleaned glass and optically flat quartz substrates. The films are vacuum deposited from a quartz crucible source heated by a tungsten coil in a vacuum of 105 Pa during deposition. The temperature of substrates were kept at room temperature and the deposition rate is controlled at 2 nm s1 using a quartz crystal thickness monitor (Model FTM6, Edwards, England). Thickness of the film is accurately determined by Tolnsky’s technique [23]. Some samples were heated after deposition to various temperatures up to 523 K for periods greater than 2 h for the purpose of identifying phase transformations within the film. The structural properties of NiPc in powder form and thin films were investigated using X-ray diffraction patterns and infrared spectroscopy techniques. A Philips X-ray diffractometer (model X’pert) was used for the measurements with utilized monochromatic Cu Ka radiation operated at 40 kV and 25 mA. The diffraction patterns were recorded automatically with a scanning speed of 21/min. The diffraction scans at 1/81/min scanning speed were acquired for the individual peaks to determine the mean crystallite size of the as-deposited and annealed films. Infrared spectroscopy on NiPc films was performed using ATI Mattson (Infinity series FTIR) infrared spectrophotometer in the infrared spectral range 1000–400 cm1. For this study 1 mg of NiPc powder was mixed with 50 mg of vacuum-dried IR grad KBr, as well as thin films of thickness 400 nm were depocited onto KBr optically flat substrate kept at room temperature. The transmittance, T and the reflectance, R of the films were measured at normal incidence in the spectral
ARTICLE IN PRESS M.M. El-Nahass et al. / Optics & Laser Technology 37 (2005) 513–523
range 190–2100 nm using a double beam spectrophotometer (JASCO model V-570 UV-VIS-NIR).
515
The program is intended to find the solution of the two simultaneous nonlinear equations f t ðn; kÞ ¼ T ðn;kÞ T exp ¼ 0;
(5)
f r ðn; kÞ ¼ Rðn;kÞ Rexp ¼ 0;
(6)
2.1. Method of calculation When a light beam impinges on a thin film deposited on a substrate, multiple reflections occur at the two boundary surfaces of the system. Assuming that the multi- reflections are coherent in the thin film and incoherent in the substrate, the expression of T and R; and R0 are given by [24] T¼
T f ð1 Rq Þ Tð1 Rq R0f Þ ; 0 ; Tf ¼ 1 Rq Rf ð1 Rq Þ T 2f Rq T 2f Rq ; R ¼ R ; f 1 Rq R0f 1 Rq R0f
(2)
Rq þ R0f ð1 2Rq Þ R0 Rq ; ; R0f ¼ 0 1 Rq Rf 1 Rq ðR0 2Þ
(3)
R ¼ Rf þ
R0 ¼
(1)
where R0 ; Rq ; Rf ; R0f and Tf, are the back reflections, the reflectance of the quartz substrate, the reflectance of the film, the back reflectance of the film and transmittance of the film, respectively. The calculated values of reflectance Rðn;kÞ and transmittance T ðn;kÞ of an absorbing film of thickness d on a non-absorbing substrate at normal incidence are expressed by Murmann’s exact equations [25]. In practice the equation giving Rðn;kÞ and T ðn;kÞ explicity in terms of the optical constants of the film and substrate are very complicated and have multiple solutions [26]. Many authors have applied different methods for determining n and k [27,28]. The most recent methods consist of computerized algorithms intended to solve these complex equations. Several modifications have been made for speeding up computation and enhancing accuracy. Bennett and Booty [26] used a computer program which is basically a univariate search technique. Their strategy is not a good one for general optimization problems. Abele and Theye [28] used the Newton–Raphson method to obtain the solution of the two equations T exp T ðn;kÞ ¼ 0; Rexp Rðn;kÞ ¼ 0;
where T ðn;kÞ and Rðn;kÞ refer to Murmann’s exact equations, i.e. the description of the iteration process is as follows: 1. The optical constants of the film material obtained from the literature are used as starting values, n0 and k0 : They are entered together with the values of Texp, Rexp and film thickness d: 2. Eqs. (5) or (6) is expanded in a Taylor series around the point ðn0 ; k0 Þ: Ignoring higher-order partial derivatives, the resulting equations take the form qf t qf t Dn þ Dk ¼ 0 (7) qn n0 k0 qk n0 k0
or qf r qn
Dn þ
n0 k 0
@f r @k
Dk ¼ 0:
(8)
n0 k 0
3. Divide the resulting linear Eqs. (7) or (8) by the largest nonzero coefficient to get an expression for the variable, corresponding to the coefficient, in terms of the other variable. 4. Similarly, the second linear equation is processed through steps 2 and 3 and the resulting expression for a variable is used in the first expression wherever the variable has appeared. 5. When the processing of the two expressions is terminated, the two variables will have numerical values (i.e. a solution).
(4)
where Texp and Rexp are the experimentally determined values of T and R; respectively. The main advantage of their method is that n and k are varied simultaneously in the search for the solution, and thus, the convergence is improved. In the present work the Newton–Raphson algorithm was modified; hence a new computer program was created for calculating n and k: The method was applied to determine the optical constants of NiPc thin films and the data were compared with those obtained by other authors.
Fig. 1. X-ray diffraction patterns for NiPc: (a) powder and (b) thin film.
ARTICLE IN PRESS M.M. El-Nahass et al. / Optics & Laser Technology 37 (2005) 513–523
516
6. If the difference between the solution obtained in step 5 and the starting point used in step 2 is less than or equal to the tolerance desired, then the solution is output, otherwise the initial guess is updated by the solution and steps 2–5 are repeated. For comparison, data for three different wavelengths were processed using the two modes of computation: namely, the method of Abele and Theye and our modified method. Identical values for n and k; accurate to the third decimal place, were obtained. The only difference was the processing speed. The run time taken by each mode to process the data was found to be 14 s for the Abele–Theye mode and 9 s for our mode.
3. Results and discussion
Fig. 2. A slow scan of the (1 0 0) plane reflection: (a) as-deposited film and (b) annealed film.
3.1. Structural characterization An X-ray diffraction traces derived from powder NiPc at room temperature is shown in Fig. 1a. An accurate determination of the peak intensity is obtained using the line profile analysis. The Bragg geometry is employed for the analysis. The scattered intensities are compared with the ICDD data. Indexing is carried out and peaks are identified. The lattice constants of NiPc are calculated and the obtained values are a=1.499 nm, b=0.47 nm, c=1.99 nm and b=121.151. Comparable unit cell data have been reported by Collins et al. [29] for b-NiPc and the structure is adjudged to be monoclinic. An X-ray diffraction traces of the as-deposited NiPc thin films can also be made. Fig. 1b shows an X-ray diffraction traces of the as-deposited NiPc thin film of thickness 400 nm as a representative example. As observed, there is only one significant peak at around 2y ¼ 6:81 implying a preferential orientation in the (1 0 0) direction. In general, the crystalline structure of our films assumes a similar configuration to those reported for b-Cu Pc [30–32] and for b-H2Pc [33] films. A slow scan of the (1 0 0) plane reflection of Fig. 1b is shown in Fig. 2a. The mean crystallite size, L for NiPc thin films may be estimated from the half-width value of the X-ray diffraction peak, using scherrer’s equation [30] L¼
K sl ; b0 cos y
(9)
where l is the X-ray wavelength (0.154 nm), b0 the width of a strong peak in radians at half-maximum intensity, and y the corresponding Bragg angle and Ks the scherrer’s constant. The value of Ks, in general, depends on the crystallite shape and normally of the order of unity for similar phthalocyanine films [33] using this value, and values of b0 and y derived from the peak of Fig. 2a, a value of L=25 nm is obtained. This value of L falls reasonably well within the expected values of micro
crystallite size of 20–100 nm reported in the literature for different phthalocyanine thin films [34,35]. Fig. 2b shows the effect of annealing temperature at 523 K for 2 h on the crystallite structure of NiPc films. As shown, there appears to be a strong preferential orientation in the same (1 0 0) direction with small shift in the peak 2y value. The figure also shows a tendency for the peak width to decrease and the maximum intensity to increase, both demonstrating an increase in the mean crystallite size which has been calculated as 36.8 nm. This behavior may be attributed to partial transformation of a-form of NiPc to b-form [35]. The infrared absorption technique has been applied to the identification of the polymorphs of the phthalocyanine molecules [30]. Fig. 3 shows infrared absorption spectra of different forms of NiPc. These relate to NiPc powder mixed with a vacuum dried IR-grad KBr (Fig. 3a), a thin film freshly deposited onto a KBr disc at room temperature (Fig. 3b), and the same film after heating at 523 K for 2 h (Fig. 3c). The main spectral features which distinguish between the different crystalline forms of NiPc were found to lie in the region 700–800/cm. These spectral differences indicated in Fig. 3 are attributed to the different crystalline packing of the NiPc molecules, especially those at 725 for the rich aform and at 730/cm for the rich b-form [30]. These two bands are absent in the alternative structure and may therefore, be used to identify the presence or absence of either form. The powder absorption spectrum shown in Fig. 3a was, therefore, identified as rich of b-NiPc, in comparison with the work of Assour [36] and of Sharp and Abkowitz [37], whereas that of a thin film thermally deposited at room temperature was identified with the rich of a-NiPc structure (Fig. 3b). Conversely the infrared spectrum of a thermally treated film of NiPc shows much less structure than was expected for the b-
ARTICLE IN PRESS M.M. El-Nahass et al. / Optics & Laser Technology 37 (2005) 513–523
Fig. 3. Infrared absorption spectra of NiPc: (a) as-deposited film and (b) annealed film.
polymorph (Fig. 3c). This can be ascribed to thinning of the film resulting from partial thermal sublimation of the NiPc layer due to the extended heating time at 523 K. However, the main spectral peaks of this sample define as rich in b-form Ni Pc. Complete conversion to the b-form is unlikely, however, as some bands characteristic of the a-form still apparent. Our results suggest that a complete phase conversion would not occur but only a partial transformation has taken place. These results give further support to the observations made by X-ray diffraction for the similar samples, as discussed above.
517
Fig. 4. The UV-visible spectrum observed for phthalocyanines originates from molecular orbitals within the aromatic 18p electron system and from overlapping orbitals on the central metal atom [38,39]. A close examination of the absorption band in the UV region which is known as the Soret (B) band, reveals three peaks and two shoulders. The broad absorption band in the ultraviolet region is preceded by the ultraviolet absorption band edge of the phthalocyanine molecule. The other well-known band of the phthalocyanine molecule, namely Q band appears in the region between 550 and 750 nm. As observed from Fig. 4 that the distinct characterized peaks for NiPc in visible region (Q-band) at 1.81 and 2.02 eV have generally been interpreted in terms of p2p excitation between bonding and antibonding molecular orbitals [40]. It can also be noticed that this band shows the characteristic splitting (Davidov splitting) in all phthalocyanine derivatives which has a value of 0.21 eV in agreement with other phthalocyanine molecules [40,41]. However, the Q-band consists of a high-energy peak at 2.02 eV and a lowenergy shoulder at 1.81 eV. The high-energy peak of the Q band has been assigned to the first p2p transition on the phthalocyanine macrocycle [40]. The low-energy peak of the Q band has been explained variously, as a second p2p transition, as an excitation peak [40] as a vibrational internal interval and as a surface state. In the high-energy region of the Soret band near 4.3 eV, the main suggestion of the large differences occurring in the absorption spectra of the phthalocyanines in this region indicate the presence of a d-band associated with the central metal atom. It is thought, that p2d transitions are involved since strong absorption occurs near 4.3 eV in NiPc and the other metal phthalocyanine derivatives [40]. This is because NiPc has partially occupied d bands. The absorption bands in the region of 4.5–5.9 eV may be due to d2p transition [40,41], which implies a broader d band. Similar behavior of the absorption spectra are obtained by Edwards and Gouterman [42] for NiPc thin films. The effect of annealing on the
3.2. Optical characterization 3.2.1. Absorption spectra The absorption spectra of the as-deposited NiPc thin films of thickness 403 nm and the film after annealing at 523 K for 2 h, as a representative example, is shown in
Fig. 4. The absorption spectra of NiPc: (a) as-deposited film and (b) annealed film.
ARTICLE IN PRESS 518
M.M. El-Nahass et al. / Optics & Laser Technology 37 (2005) 513–523
absorption spectra is also shown in Fig. 4. As observed there is little broadening of the Q and B bands for the film after annealing. In recent years there has been an interest in extending the absorption band of phthalocyanines into the near infrared region of the spectrum for a range of potential functional applications including optical data storage and security printing [43]. 3.2.2. Optical constants The spectral distribution of TðlÞ and RðlÞ at normal incidence in the wavelength range 190–2100 nm for the as-deposited and annealed NiPc films in the thicknesses range 230–770 nm are shown in Fig. 5. As a whole observation, it could be noted that at longer wavelength lo900 nm all films become transparent and no light is scattered or absorbed, i.e. T þ R ¼ 1 while the inequality R þ To1 at shorter wavelengths lo900 nm is due to the existence of absorption. Taking into account, the experimental error in measuring the film thickness to be72% and in n and k to be71%, the error in the calculated values of n and k was estimated to be73% and 2.5%, respectively, the optical constants were found to be independent of the film thickness for all the studied films.
Fig. 6 shows the dispersion of the refractive index nðlÞ in the wavelength range 190–2100 nm for the asdeposited and annealed NiPc films. It is observed that there is little variation in the position of peaks but the magnitude of the peaks are greater for the annealed films than that of as-deposited films. It is also observed that there is anomalous dispersion at lo900 nm as well as normal dispersion at l4900 nm: In anomalous dispersion region, there are four peaks and one shoulder. This behavior obeys a multioscillator model as follows [44]: 1 NP ¼ ¼ A þ Bu2 þ2 3 1X 1 1 þ Jj þ ; p j uoj u iGj uoj þ u þ iGj
ð10Þ
where G is the linewidth, u0j the resonance frequency and J is the intensity factor, which are related to the oscillator concentration N and the transition matrix element. These intensity factors are proportional to the integrated line intensity; P is the microscopical polarizability. The values A and B take the influence of resonance frequencies above 50,000 cm1 into account. The refractive index dispersion has been analyzed using
Fig. 5. Spectral distribution of TðlÞ and RðlÞ of NiPc: (a) as-deposited film and (b) Annealed film.
ARTICLE IN PRESS M.M. El-Nahass et al. / Optics & Laser Technology 37 (2005) 513–523
519
Fig. 6. Dispersion of the refractive index of NiPc: (a) as-deposited film and (b) annealed film.
the single oscillator model, in the normal dispersion region, developed by Toyode [45] and Wemple and DiDomenico [46]. They introduced an energy parameter Ed to describe the dispersion of the refractive index. In terms of this dispersion energy Ea and a single-oscillator energy E0, the refractive index at a frequency can be expressed in the form [45]: n2 ðnÞ 1 ¼ E d E 0 =½E 20 ðhnÞ2 :
(11)
It was found that the dispersion energy Ed, which is related to the interband transition strength, obeys an extraordinarily simple empirical relation for over 100 widely different ionic and covalent nonmetallic crystals and liquids [46]. Based on experimental observation, Wemple and DiDomenico [46] showed that: (i) Ed is independent of the absorption threshold (band gap) within experimental error, (ii) Ed is independent of the lattice constant within experimental error; and (iii) E0 is related to the lowest direct gap. In practice, the single-oscillator energy E0 and the dispersion energy Ed can be obtained from Eq. (11) by plotting ðn2 21Þ1 against E2. Fig. 7 shows an example of such a plot for the as-deposited and annealed NiPc thin films. It is observed that the plots are linear over the energy range from 0.77 eV to approximately 1.16 eV. The determination of E0 and Ed can be made from the plot, using the linear fit parameters which are calculated by the plotting program used. The values of Ed and E0 are obtained from the intercepts and the slopes of the curve as 3.570.03 and 8.0670.04 eV, respectively, for the as-deposited films and 3.5270.03 and 7.8470.04, respectively, for the annealed films. As observed there is a little effect of annealing on the oscillating and dispersion energies. A simple connection between the single-oscillator parameters E0 and Ed and the imaginary part of the dielectric constant 2 ðoÞ spectrum can be expressed in terms of moments of the 2 ðoÞ as [46] E 20 ¼ M 1 =M 3
(12)
Fig. 7. Plot of ðn2 1Þ1 versus ðhnÞ2 of NiPc:(a) as-deposited film and (b) annealed film.
and E 2d ¼ M 31 =M 3 :
(13)
The oscillator energy E 0 is independent of the scale of 2 and is consequently an ‘‘average’’ energy gap, whereas Ed depends on the scale of 2 and thus serves as an interband strength parameter. Since the 1 and 3 moments are involved in computation of E 0 and E d : The values of M1 and M3 are 2.3 and 0.188 eV2, respectively, for the as-deposited films and 2.23 and 0.179 eV2, respectively, for the annealed films. The relation between the infinitely high-frequency dielectric constant 1 ; wavelength l and refractive index (n) is given by [47] n2 ¼ 1 e2 =pc2 ðN=m Þl2 ;
(14)
where 1 is the infinitely high-frequency dielectric constant and N=m is the ratio of carrier concentration to the effective mass. The n2 versus l2 plot shown in Fig. 8 is linear verifying Eq. (14). The values of 1 and (N=m ) are determined from the extrapolation of these plots to l2 ¼ 0 and from the slopes of the graph. The obtained values are 3.71 and 1.2 1047 gm1 cm3 for the as-deposited films and 3.62 and 1.92 1048 gm1 cm3 for the annealed films. The results obtained here are in reasonably good agreement with previously reported results [47] for metal-free phthalocyanine. The spectral distribution of the absorption coefficient að¼ 4pk=lÞ for the investigated as-deposited and annealed NiPc films is shown in Fig. 9. As observed, the behavior of the spectra is similar to the absorption spectra, but has little effect of annealing on the behavior of the absorption coefficient. The optical absorption coefficient of a solid describes the exponential decay of light intensity with distance into the solid. It is the optical absorption coefficient of a solid that describes the exponential decay of light intensity with distance
ARTICLE IN PRESS 520
M.M. El-Nahass et al. / Optics & Laser Technology 37 (2005) 513–523
Fig. 8. Plot of n2 versus l2 of Nipc:(a) as-deposited film and (b) annealed film.
Fig. 10. Plot of emolar versus n¯ of NiPc.
The spectral parameters namely oscillator strength f, electric dipole strength q2 and absorption half-band width Dl are determined from the following expressions [47] using the fit calculation: Z 9 molar dn; (16) f ¼ 4:38 10 1 Dl molar : (17) 2500 l In phthalocyanine molecule,the oscillator strength of an interaction between the central metal atom (the metal atom is positioned at the center of the Pc ring) and the Pc ring. The calculated parameters at different band regions are calculated and collected in Table 1. A comparison of these results shows that all parameters varies in the sequence of band regions. The results obtained in Table 1 are in reasonably good agreement with previously reported results of other phthalocyanines [47]. There are different quantities frequently used to characterize the optical properties of the dielectric relaxation time t, which can be evaluated using the 1 1 tðsÞ ¼ ; (18) o2 q2 ¼
Fig. 9. Plot of a versus hn of Nipc: (a) as-deposited film and (b) annealed film.
into the solid. It is useful to relate a to the molar extinction coefficient emolar, which is often used to describe the absorption of light by nonsolid molecular media. If the solid has a concentration of N molecules per unit volume, each with molecular weight M; a and emolar are related by the expression [41] a¼
N r 103 lnð10Þmolar ¼ 103 lnð10Þmolar L M ¼ const molar ;
ð15Þ
where L is the Avogadro’s number, r the solid’s mass density, and emolar is in units of liters per mole cm. In our NiPc films, the reported density, r ¼ 1:59 g=cm3 ; M=571.23 and the constant in Eq. (15) becomes 6.41 mol/l. The spectral distribution of the molar extinction coefficient is shown in Fig. 10. Clearly, a Gaussian fit of the molar extinction coefficient of NiPc thin films leads to significantly good results, which is in good agreement with the behavior of the absorption coefficient.
where e1, e2 are the real and imaginary dielectric constants, respectively, and o is the frequency. The dielectric relaxation time represents the time of a single oscillation of a dipole in the potential well. Fig. 11 represents the dielectric relaxation time as a function of the photon energy hn for the as-deposited and annealed NiPc thin films. It is observed that the figure exhibit two absorption bands: the Soret band. In the near UV (3.77 and 3.59 eV for the as-deposited and annealed films respectively) and the Q-band in the visible region (2.13, 1.78 and 1.67 eV for the as-deposited films and 2.08, 1.81 and 1.65 eV for the annealed films). The peaks at 2.89 and 2.77 eV for the as-deposited and annealed films,
ARTICLE IN PRESS M.M. El-Nahass et al. / Optics & Laser Technology 37 (2005) 513–523
521
Table 1 The spectral parameters at different band regions E (eV)
a ( 105/cm)
molar ( 105) (mol/l cm)
f ( 104)
q2 ( 102) (A1)
1.81 2.01 3.85 4.45 5.8
1.31 1.82 3.4 2.3 2.72
0.23 0.31 0.56 0.396 0.45
0.97 1.32 2.45 1.72 1.95
0.91 1.8 4.05 2.3 3.28
Fig. 12. Plot of ðahnÞ1=2 versus hn of NiPc: (a) as-deposited film and (b) annealed film.
the analysis of the spectral dependence of the absorption near the fundamental absorption edges. In the above two regions the absorption coefficient a is well described by the relation [47]. ahn ¼ Bðhn E g Þm ;
Fig. 11. Plot of t versus hn of NiPc:(a) as-deposited film and (b) annealed film.
respectively, may be attributed to band-to-band absorption because this maximum occure at hn E g which corresponds to the fundamental energy gap for most metal derivative phthalocyanines [19,29,48–50]. 3.2.3. Energy gap determination Fig. 9 indicates that NiPc thin films have a strong absorption band between 1.45 and 1.8 eV and another strong band starting at 2.7 eV which extends into the ultraviolet. The optical band gap was determined from
(19)
where hn is the energy of incident photons and Eg is the value of the optical band gap corresponding to transitions indicated by the value of m. The factor B depends on the transition probability and can be assumed to be constant within the optical frequency range. Plots of a1=2 versus hn near the absorption edge of the Q and B bands for the as-deposited and annealed films produce a linear fit over a wider range in hn as shown in Fig. 12. This is the characteristic behavior of indirect transitions in nanocrystalline materials. This type of transition is in agreement with Kumar et al. [47] for free and rare-earth phalocyanine and disagreement with Collins et al. [29] for PbPc thin films and Ambily and Menon [7] for CuPc thin films and El-Nahass et al. [50] for FePc thin films. The extrapolation of the straight line graphs to zero absorption will give the value of the optical band gap. The obtained energy band gaps are 1.6670.02 eV for the onset of the absorption spectrum and 2.7770.03 eV for the fundamental energy gap. As observed from Fig. 13 the band gap does not show any remarkable difference from its value in the as-deposited and the annealed films. This may be attributed to the two polymorphs that have the same system and differ only in the tilt angle of the molecules within the columns and the mutual arrangement of the columns [2]. The urbach tail was found to be related directly to similar exponential tail for the density of states of either one of the two band edges [51]. The origin of this exponential tail absorption probably arises from the random fluctuation of the internal fields associated with structural disorder which are considerable in an amorphous solid [51]. The width of the urbach tail is an indicator of disorder in the material. The following
ARTICLE IN PRESS M.M. El-Nahass et al. / Optics & Laser Technology 37 (2005) 513–523
522
Fig. 13. Plot of ln a versus hn: (a) as-deposited film and (b) annealed film.
relation was used to calculate the slope of the urbach tail [52] Fig. 14. Suggested energy band diagram of NiPc.
a ¼ ao e
hu=E t
;
(20) 4. Conclusion
where ao is the background absorption and Et is the parameter defining the slope (width) of the tail. Fig. 13 shows the linear dependence of the natural logarithm of the absorption coefficient of Fig. 12 (in the range 1.3–1.6 eV) versus the photon energy for the asdeposited and annealed films. The reciprocal of the slopes of each line yields the magnitude of Et. The obtained values of Et are 93.470.04 and 90.870.05 meV for the as-deposited and annealed films, respectively. The decrease of Et with annealing temperature indicates a decrease in the disorder in NiPc films. We have proposed a band structure of NiPc in which the bands are constructed from molecular orbitals of Ni dxz;yz and pyrrole carbon Pz electronic states and this is schematically shown in Fig. 14. In view of the ionicity of NiPc it is not clear how far this molecular orbital approximation can be used in this material. Since the band structure of NiPc has not been calculated, the model is only schematic and does not reflect the extent of electron overlap on the energy separation between different electronic states. Following this model we would suggest that the energy gaps of 1.66 and 2.77 eV, estimated from the energy dependence of the absorption coefficient (Fig. 9), could correspond to a transition involving the valence band (HOMO with eg symmetry) to the bottom of the conduction band (LUMO with a1u symmetry) and to the next unoccupied orbital. The highenergy peaks at 3.65, 4.5 and 5.9 eV shown in Fig. 9 may correspond to transitions involving the valence band to the split d band (b1g, eg and b1u). These results are consistent with other reported calculation on NiPc electronic structure [20].
X-ray powder diffraction patterns of Nickel phthalocyanine have shown that the initial powder is rich in bform. Films deposited at room temperature assumed rich in a-form with a preferential orientation in the (1 0 0) direction. When the films were heat-treated at about 523 K, X-ray diffraction analysis showed that they reverted partially to b-form, preferentially oriented in the same (1 0 0) direction. IR spectral analysis was employed to confirm the X-ray results and indicated that the annealed films did not totally transform to the b-form. The optical characterization of the as-deposited and annealed films have been studied in the spectra range 190–2100 nm. The stability in the peak positions in transmission and reflection spectra in absorbing region showed the stability of the structure of NiPc thin films. The refractive index as well as the absorption index are practically independent on the film thickness. The spectral distribution of the absorbance and the absorption coefficient of the investigated NiPc films characterized by distinct peaks in visible region have generally been interpreted in terms of p2p excitation. The absorption spectra occur in the high-energy region of the Soret band indicate the presence of d-band associated with the central metal atom. It is thought that p–d transitions are involved because NiPc has partially occupied d-band. Some of the important optical parameters such as the oscillating energy, the dispersion energy, the infinitely frequency dielectric constant, the molar extinction coefficient, oscillator strength, the electric dipole strength and the dielectric relaxation time have been evaluated. The mechanism of the electronic
ARTICLE IN PRESS M.M. El-Nahass et al. / Optics & Laser Technology 37 (2005) 513–523
transition can be understood in terms of the electron band structure of this material. Although no theoretical work has been done on the band structure of this material. No remarkable changes were observed in the parameters for the annealed samples compared to the as-deposited.
References [1] Leznoff CC, Lever ABP, editors. Phthalocyanines properties and applications. New York: VCH publishers Inc; 1996. p. 219. [2] Djurisic AB, Kwong CY, Lau TW, Guo WL, Li EH, Liu ZT, Kwok HS, Lam LSM. Opt Communications 2002;205:155–62. [3] Cook MJ. Pure Appl Chem 1999;71:2145. [4] Gregorg P. High-technology application of organic colorants. New York: Plenum Press; 1991. p. 759. [5] AO R, Kilmmert L, Haarer D. Adv Mater 1995;7:495. [6] Yoshino K, Hikida M, Tatsuno K, Kaneto K, Inushi Y. J Phys Soc Jpn 1973;34:44. [7] Ambily S, Menon CS. Thin Solid Films 1999;347:284. [8] Krishnakumar KP, Menon CS. J Solid State Chem 1997;128:27. [9] Brinkmann M, Wittmann JC, Chaumont C, Andre JJ. Thin solid Films 1997;292:192. [10] Berger O, Fischer WJ, Adolphi B, Tierbach S, Melev V, Schreiber J. J Mater Sci 2000;11:331. [11] Bayliss SM, Heutz S, Rumbles G, Jones TS. Mater Res Soc Symp Proc 1999;560:71. [12] Shihub SI, Gould RD. Thin Solid Films 1996;290:390. [13] Davidson AT. J Chem Phys 1992;77:168. [14] Schmeisser D, Rager A, Thonke K, Pilkuhn M, Frohlich D, Gaauglitz G, Schafer M, Oelkrug D. Synth Met 1991;41:1457. [15] Mochert H, Schmeisser D, Gopel W. Sensors and Actuators 1989;19:159. [16] Ukei K. J Phys Soc Jpn 1976;40:140. [17] Heilmeier GH, Warfield G. J Appl Phys 1962;34:2278. [18] Heilmeier GH, Warfield G, Harrison SE. Phy Rev Lett 1962;8:309. [19] Ambily S, Menon CS. Thin Solid Films 1999;347:284. [20] Bialek B, Kim IG, Lee JI. Synth Met 2002;129:151. [21] Achar BN, Jayasree PK. Synth Met 2000;114:219. [22] Chaabane RB, Guillaud G, Gamoudi M. Thin Solid Films 1997;296:145.
523
[23] Tolansky S. Multiple—beam interferometry of surfaces and films. New York: 1965. p. 55. [24] Laaziz Y, Bennouna A, Chahboun N, Outzourhit A, Ameziane A. Thin Solid Films 2000;372:149. [25] Heavens OS. Optical Properties of thin solid films. New York: Dover publication; 1965. p. 55. [26] Bennett JM, Booty MG. Appl Opt 1966;5:40. [27] Tomlin SG. J Phys D 1968;1:1667. [28] Abele F, Theye MI. Surg Sci 1966;5:325. [29] Collins RA, Krie A, Abass AK. Thin Solid Films 1993; 229:113. [30] Hassan AK, Gould RD. Phys Stat Sol (A) 1992;132:91. [31] Komiyama M, Sakaibara Y, Hirai H. Thin Solid Films 1987;151:109. [32] Debe MK, Kan K. Thin Solid Films 1990;186:298. [33] Shihub SI, Gould RD. Phys. Stat. Sol. (A) 1993;139:129. [34] Fan FR, Faulkner LR. J Chem Phys 1978;69:3334. [35] Boudgema B, Guilland G, Gamoudy M, Maitrot M, Andre JJ, Martin M, Simon J. J Appl Phys 1984;56:2323. [36] Assour JM. J Phys Chem 1965;69:2295. [37] Sharp JH, Abkowitz M. J Phys Chem 1973;77:477. [38] Stillman MJ, Nyokong TN. In: Leznott CC, Lever ABP, editors. Phthalocyanines: properties and applications, vol. 1. New York: VCH; 1989. p. 133 [Chapter 3]. [39] Krishnakumar KP, Menon CS. J Solid State Chem 1997; 27:128. [40] Davidson AT. J Chem Phys 1982;77:162. [41] Schechtman BH, Spicer WE. J Mol Spect 1970;33:28. [42] Edwards L, Gouterman M. J Mol Spect 1970;33:292. [43] Christie MR, Freer GB. Dyes Pigments 1994;24:113. [44] Stenzel O, Wilbrand S, Stendal A, Beckers U, Voigtsberger K, Borczyskowski CV. J Phys D 1995;28:2154. [45] Toyoda T. J Appl Phys 1988;63:5166. [46] Wemple SH, DiDomenico MD. Phys Rev B 1970;3:1338. [47] Kumar GA, Thomas J, George N, Kumar BA, Shnan PR, Poori VPN, Vallabhan CPG, Unnikrishnan NV. Phys Chem glasses 2000;41:89. [48] Bell RJ, Ordal MA, Alexender RW. Appl Opt 1985;24:3680. [49] El-Nahass MM, Bahabri FS, Al-Harbi SR. Egypt J Sol 2001;24:11. [50] El-Nahass MM, Soliman HS, Metwally HS, Farid AM, Farag AAM, El-Shazly AA. J Opt 2001;30:121. [51] Cody GD. J Non-Cryst Solids 1992;141:3. [52] Stutzmann M. In: Mahajan S, editor. Hand book on semiconductors, vol. 3A. Amsterdam: North-Holland; 1994. p. 657.
Optics & Laser Technology 37 (2005) 513–523 www.elsevier.com/locate/optlastec
Dispersion studies and electronic transitions in nickel phthalocyanine thin films M.M. El-Nahass, A.M. Farag, K.F. Abd El-Rahman, A.A.A. Darwish Faculty of Education, Ain Shams University, Heliopolis, Roxy, Cairo 11757, Egypt Received 29 September 2003; received in revised form 16 July 2004; accepted 27 August 2004
Abstract Thin films of the organic semiconductor nickel phthalocyanine (NiPc) are structurally investigated using X-ray diffraction and infrared light absorption. The optical absorption and dispersion studies of nickel phthalocyanine were investigated using spectrophotometric measurements of transmittance and reflectance at normal incidence in the wavelength range 190–2100 nm. The absorption spectra recorded in the UV-VIS region show two well-defined absorption bands of the phthacyanine molecules, namely, the Soret and the Q-band. The Davydove splitting of the main absorption peak in the metal phthalocyanines correlates with the relative tendencies of the metal to out-of-plane bonding. The refractive index n as well as the absorption index k were calculated and showed an independent of the film thickness in the film thicknesses range 400–770 nm. The refractive index n showed an anomalous dispersion in the absorption region as well as normal dispersion in the transparent region. Some of the important optical absorption such as the molar extinction coefficient, the oscillator strength, the electric dipole strength have been evaluated. The analysis of the spectral behavior of the absorption coefficient a in the absorption region revealed two indirect allowed transitions with corresponding energies 2.7770.03 and 1.6670.02 eV. An energy band diagram has been proposed to account for the optical transitions of NiPc thin film. All previous parameters were as well obtained for films annealed at 523 K for 2 h. Discussion of the obtained results and their comparison with the previously published data are also given. r 2004 Elsevier Ltd. All rights reserved.
1. Introduction Phthalocyanines are porphyrin derivatives, characterized with high symmetry, planarity and electron delocalization [1]. Phthalocyanines have attracted lots of attention for applications in organic optoelectronic devices [2]. Some of the important applications of phthalocyanines are in fabrication of solar cells, electronic displays, chemical sensors, photocopying and laser printers and optical date storage systems [3–5]. In addition to these applications, being a very good absorber of light in the UV-VIS region they are widely used as excellent laser dyes capable of optical amplification in the red region [6]. Phthalocyanines are generally p-type semiconductors and have the advantage Corresponding author.
E-mail address: [email protected] (A.M. Farag). 0030-3992/$ - see front matter r 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.optlastec.2004.08.016
of being sufficiently stable toward chemical and thermal treatments [7]. Most phthalocyanine compounds are insoluble in common organic solvents and so it is not possible to prepare them by solution casting techniques [8]. Hence, very often, thin films of phthalocyanines are made by a vacuum evaporation technique. This method has the advantage of producing high purity films without decomposition [7]. Growth and morphology of the various phthalocyanine films have been extensively studied [9–12]. Phthalocyanine films can exist in several crystalline polymorphs. Most common polymorphs are metastable a and stable b phases. The main differences between different polymorphs are the tilt angle of the molecules within the columns (stacks of molecules with molecular planes parallel to each other) and the mutual arrangement of the columns [2]. Transformations from a phase to b phase using successive sublimation [9,10] or annealing or deposition
ARTICLE IN PRESS 514
M.M. El-Nahass et al. / Optics & Laser Technology 37 (2005) 513–523
at higher temperatures [11,12] have been reported. The absorption spectra of different polymorphs of some phthalocyanine compounds show significant differences among each other [11]. There is a large portion of reports appearing, focusing on the optical properties of various metal substituted phthalocyanines. The visible and near ultraviolet absorption spectra have been measured previously by Davidson [13] for thin films of H2Pc, MgPc, FePc, CoPc, CuPc, and ZnPc. Schmeisser et al. [14] have found that the lowest absorption band at 1.73 eV due to transitions between the HOMO p and the LUMO p molecular orbitals, which due to the size of the 18-p electron systems are well separated in energy from molecular orbital structure of the benzene and aza group. They recorded a considerably smaller energy gap for PbPc than of H2Pc from the onset of the absorption spectrum to be 1.2 eV. According to this result, the band-to-band absorption has taken place. The UV photoemission spectrum, which was studied by Mockert et al. [15], shows a peak at 2.4 eV. This peak might originate from the direct band-to-band transition. Ukei [16] suggested a model where the electronic conduction is due to charge transfer from spz mixing orbit to the p electron system of the macrocyclic ring of a phthalocyanine. The direct electronic transition might be of the p to p type. The applicability of the band model to explain conduction in metal free phthalocyanine is suggested by the equality of the thermal activation energy for electrical conduction, the energy for first optical absorption, and the energy for the initiation of photoconductivity [17]. Further support for the band model is furnished by the similarities between phthalocyanine and anthracene since evidence has been advanced for the band-model approach in anthracene [18]. Ambily and Menon [19] have studied the optical properties of CuPc thin films. They suggested that the direct electronic transition from p2p orbital in the range 300–450 nm results in an intense band called the Soret-band which gives the absorption edge and the Qband in the 500–750 nm range yields two optical trapping levels. Bialek et al. [20] calculated HOMOLUMO gap of NiPc as 2.41 eV and suggested that orbitals next to the HOMO are separated from each other by 1.84, 2.44, 0.17, 0.28 and 0.09 eV, and the energy level of the second unoccupied MO is widely separated from the LUMO, and differences become smaller for the following orbitals. Although NiPc, like other metal-substituted phthalocyanines, is a subject of various investigations [21,22] considerably less attention has been paid on its structural and optical characterizations. A complete description of the optical properties requires a knowledge of a set of values at least two optical constants (n and k). In general, the optical and certain electronic properties are only specified if the optical constants are known. Hence phthalocyanines
have been investigated as organic laser materials; a knowledge of the refractive index is essential in determining the population inversion density required for lasing. Also, information about the actual spectral absorption characteristic of phthalocyanine pigments is required to analyze the action spectra of solar cells. The present work is an extension of these studies aimed at characterizing the optical properties of NiPc thin films. The specific aims are related to an improved method for the determination of the optical constants from the experimental transmittance (T) and reflectance (R) measurements, and also to further investigate the applicability of a band model approach for the description of this material.
2. Experimental details The nickel phthalocyanine (NiPc) powder used in this study is obtained from Kodak, UK. Thin NiPc films of different thicknesses are prepared by conventional thermal evaporation technique, using a high vacuum coating unit (Edwards type E 306 A, England). The films are deposited onto precleaned glass and optically flat quartz substrates. The films are vacuum deposited from a quartz crucible source heated by a tungsten coil in a vacuum of 105 Pa during deposition. The temperature of substrates were kept at room temperature and the deposition rate is controlled at 2 nm s1 using a quartz crystal thickness monitor (Model FTM6, Edwards, England). Thickness of the film is accurately determined by Tolnsky’s technique [23]. Some samples were heated after deposition to various temperatures up to 523 K for periods greater than 2 h for the purpose of identifying phase transformations within the film. The structural properties of NiPc in powder form and thin films were investigated using X-ray diffraction patterns and infrared spectroscopy techniques. A Philips X-ray diffractometer (model X’pert) was used for the measurements with utilized monochromatic Cu Ka radiation operated at 40 kV and 25 mA. The diffraction patterns were recorded automatically with a scanning speed of 21/min. The diffraction scans at 1/81/min scanning speed were acquired for the individual peaks to determine the mean crystallite size of the as-deposited and annealed films. Infrared spectroscopy on NiPc films was performed using ATI Mattson (Infinity series FTIR) infrared spectrophotometer in the infrared spectral range 1000–400 cm1. For this study 1 mg of NiPc powder was mixed with 50 mg of vacuum-dried IR grad KBr, as well as thin films of thickness 400 nm were depocited onto KBr optically flat substrate kept at room temperature. The transmittance, T and the reflectance, R of the films were measured at normal incidence in the spectral
ARTICLE IN PRESS M.M. El-Nahass et al. / Optics & Laser Technology 37 (2005) 513–523
range 190–2100 nm using a double beam spectrophotometer (JASCO model V-570 UV-VIS-NIR).
515
The program is intended to find the solution of the two simultaneous nonlinear equations f t ðn; kÞ ¼ T ðn;kÞ T exp ¼ 0;
(5)
f r ðn; kÞ ¼ Rðn;kÞ Rexp ¼ 0;
(6)
2.1. Method of calculation When a light beam impinges on a thin film deposited on a substrate, multiple reflections occur at the two boundary surfaces of the system. Assuming that the multi- reflections are coherent in the thin film and incoherent in the substrate, the expression of T and R; and R0 are given by [24] T¼
T f ð1 Rq Þ Tð1 Rq R0f Þ ; 0 ; Tf ¼ 1 Rq Rf ð1 Rq Þ T 2f Rq T 2f Rq ; R ¼ R ; f 1 Rq R0f 1 Rq R0f
(2)
Rq þ R0f ð1 2Rq Þ R0 Rq ; ; R0f ¼ 0 1 Rq Rf 1 Rq ðR0 2Þ
(3)
R ¼ Rf þ
R0 ¼
(1)
where R0 ; Rq ; Rf ; R0f and Tf, are the back reflections, the reflectance of the quartz substrate, the reflectance of the film, the back reflectance of the film and transmittance of the film, respectively. The calculated values of reflectance Rðn;kÞ and transmittance T ðn;kÞ of an absorbing film of thickness d on a non-absorbing substrate at normal incidence are expressed by Murmann’s exact equations [25]. In practice the equation giving Rðn;kÞ and T ðn;kÞ explicity in terms of the optical constants of the film and substrate are very complicated and have multiple solutions [26]. Many authors have applied different methods for determining n and k [27,28]. The most recent methods consist of computerized algorithms intended to solve these complex equations. Several modifications have been made for speeding up computation and enhancing accuracy. Bennett and Booty [26] used a computer program which is basically a univariate search technique. Their strategy is not a good one for general optimization problems. Abele and Theye [28] used the Newton–Raphson method to obtain the solution of the two equations T exp T ðn;kÞ ¼ 0; Rexp Rðn;kÞ ¼ 0;
where T ðn;kÞ and Rðn;kÞ refer to Murmann’s exact equations, i.e. the description of the iteration process is as follows: 1. The optical constants of the film material obtained from the literature are used as starting values, n0 and k0 : They are entered together with the values of Texp, Rexp and film thickness d: 2. Eqs. (5) or (6) is expanded in a Taylor series around the point ðn0 ; k0 Þ: Ignoring higher-order partial derivatives, the resulting equations take the form qf t qf t Dn þ Dk ¼ 0 (7) qn n0 k0 qk n0 k0
or qf r qn
Dn þ
n0 k 0
@f r @k
Dk ¼ 0:
(8)
n0 k 0
3. Divide the resulting linear Eqs. (7) or (8) by the largest nonzero coefficient to get an expression for the variable, corresponding to the coefficient, in terms of the other variable. 4. Similarly, the second linear equation is processed through steps 2 and 3 and the resulting expression for a variable is used in the first expression wherever the variable has appeared. 5. When the processing of the two expressions is terminated, the two variables will have numerical values (i.e. a solution).
(4)
where Texp and Rexp are the experimentally determined values of T and R; respectively. The main advantage of their method is that n and k are varied simultaneously in the search for the solution, and thus, the convergence is improved. In the present work the Newton–Raphson algorithm was modified; hence a new computer program was created for calculating n and k: The method was applied to determine the optical constants of NiPc thin films and the data were compared with those obtained by other authors.
Fig. 1. X-ray diffraction patterns for NiPc: (a) powder and (b) thin film.
ARTICLE IN PRESS M.M. El-Nahass et al. / Optics & Laser Technology 37 (2005) 513–523
516
6. If the difference between the solution obtained in step 5 and the starting point used in step 2 is less than or equal to the tolerance desired, then the solution is output, otherwise the initial guess is updated by the solution and steps 2–5 are repeated. For comparison, data for three different wavelengths were processed using the two modes of computation: namely, the method of Abele and Theye and our modified method. Identical values for n and k; accurate to the third decimal place, were obtained. The only difference was the processing speed. The run time taken by each mode to process the data was found to be 14 s for the Abele–Theye mode and 9 s for our mode.
3. Results and discussion
Fig. 2. A slow scan of the (1 0 0) plane reflection: (a) as-deposited film and (b) annealed film.
3.1. Structural characterization An X-ray diffraction traces derived from powder NiPc at room temperature is shown in Fig. 1a. An accurate determination of the peak intensity is obtained using the line profile analysis. The Bragg geometry is employed for the analysis. The scattered intensities are compared with the ICDD data. Indexing is carried out and peaks are identified. The lattice constants of NiPc are calculated and the obtained values are a=1.499 nm, b=0.47 nm, c=1.99 nm and b=121.151. Comparable unit cell data have been reported by Collins et al. [29] for b-NiPc and the structure is adjudged to be monoclinic. An X-ray diffraction traces of the as-deposited NiPc thin films can also be made. Fig. 1b shows an X-ray diffraction traces of the as-deposited NiPc thin film of thickness 400 nm as a representative example. As observed, there is only one significant peak at around 2y ¼ 6:81 implying a preferential orientation in the (1 0 0) direction. In general, the crystalline structure of our films assumes a similar configuration to those reported for b-Cu Pc [30–32] and for b-H2Pc [33] films. A slow scan of the (1 0 0) plane reflection of Fig. 1b is shown in Fig. 2a. The mean crystallite size, L for NiPc thin films may be estimated from the half-width value of the X-ray diffraction peak, using scherrer’s equation [30] L¼
K sl ; b0 cos y
(9)
where l is the X-ray wavelength (0.154 nm), b0 the width of a strong peak in radians at half-maximum intensity, and y the corresponding Bragg angle and Ks the scherrer’s constant. The value of Ks, in general, depends on the crystallite shape and normally of the order of unity for similar phthalocyanine films [33] using this value, and values of b0 and y derived from the peak of Fig. 2a, a value of L=25 nm is obtained. This value of L falls reasonably well within the expected values of micro
crystallite size of 20–100 nm reported in the literature for different phthalocyanine thin films [34,35]. Fig. 2b shows the effect of annealing temperature at 523 K for 2 h on the crystallite structure of NiPc films. As shown, there appears to be a strong preferential orientation in the same (1 0 0) direction with small shift in the peak 2y value. The figure also shows a tendency for the peak width to decrease and the maximum intensity to increase, both demonstrating an increase in the mean crystallite size which has been calculated as 36.8 nm. This behavior may be attributed to partial transformation of a-form of NiPc to b-form [35]. The infrared absorption technique has been applied to the identification of the polymorphs of the phthalocyanine molecules [30]. Fig. 3 shows infrared absorption spectra of different forms of NiPc. These relate to NiPc powder mixed with a vacuum dried IR-grad KBr (Fig. 3a), a thin film freshly deposited onto a KBr disc at room temperature (Fig. 3b), and the same film after heating at 523 K for 2 h (Fig. 3c). The main spectral features which distinguish between the different crystalline forms of NiPc were found to lie in the region 700–800/cm. These spectral differences indicated in Fig. 3 are attributed to the different crystalline packing of the NiPc molecules, especially those at 725 for the rich aform and at 730/cm for the rich b-form [30]. These two bands are absent in the alternative structure and may therefore, be used to identify the presence or absence of either form. The powder absorption spectrum shown in Fig. 3a was, therefore, identified as rich of b-NiPc, in comparison with the work of Assour [36] and of Sharp and Abkowitz [37], whereas that of a thin film thermally deposited at room temperature was identified with the rich of a-NiPc structure (Fig. 3b). Conversely the infrared spectrum of a thermally treated film of NiPc shows much less structure than was expected for the b-
ARTICLE IN PRESS M.M. El-Nahass et al. / Optics & Laser Technology 37 (2005) 513–523
Fig. 3. Infrared absorption spectra of NiPc: (a) as-deposited film and (b) annealed film.
polymorph (Fig. 3c). This can be ascribed to thinning of the film resulting from partial thermal sublimation of the NiPc layer due to the extended heating time at 523 K. However, the main spectral peaks of this sample define as rich in b-form Ni Pc. Complete conversion to the b-form is unlikely, however, as some bands characteristic of the a-form still apparent. Our results suggest that a complete phase conversion would not occur but only a partial transformation has taken place. These results give further support to the observations made by X-ray diffraction for the similar samples, as discussed above.
517
Fig. 4. The UV-visible spectrum observed for phthalocyanines originates from molecular orbitals within the aromatic 18p electron system and from overlapping orbitals on the central metal atom [38,39]. A close examination of the absorption band in the UV region which is known as the Soret (B) band, reveals three peaks and two shoulders. The broad absorption band in the ultraviolet region is preceded by the ultraviolet absorption band edge of the phthalocyanine molecule. The other well-known band of the phthalocyanine molecule, namely Q band appears in the region between 550 and 750 nm. As observed from Fig. 4 that the distinct characterized peaks for NiPc in visible region (Q-band) at 1.81 and 2.02 eV have generally been interpreted in terms of p2p excitation between bonding and antibonding molecular orbitals [40]. It can also be noticed that this band shows the characteristic splitting (Davidov splitting) in all phthalocyanine derivatives which has a value of 0.21 eV in agreement with other phthalocyanine molecules [40,41]. However, the Q-band consists of a high-energy peak at 2.02 eV and a lowenergy shoulder at 1.81 eV. The high-energy peak of the Q band has been assigned to the first p2p transition on the phthalocyanine macrocycle [40]. The low-energy peak of the Q band has been explained variously, as a second p2p transition, as an excitation peak [40] as a vibrational internal interval and as a surface state. In the high-energy region of the Soret band near 4.3 eV, the main suggestion of the large differences occurring in the absorption spectra of the phthalocyanines in this region indicate the presence of a d-band associated with the central metal atom. It is thought, that p2d transitions are involved since strong absorption occurs near 4.3 eV in NiPc and the other metal phthalocyanine derivatives [40]. This is because NiPc has partially occupied d bands. The absorption bands in the region of 4.5–5.9 eV may be due to d2p transition [40,41], which implies a broader d band. Similar behavior of the absorption spectra are obtained by Edwards and Gouterman [42] for NiPc thin films. The effect of annealing on the
3.2. Optical characterization 3.2.1. Absorption spectra The absorption spectra of the as-deposited NiPc thin films of thickness 403 nm and the film after annealing at 523 K for 2 h, as a representative example, is shown in
Fig. 4. The absorption spectra of NiPc: (a) as-deposited film and (b) annealed film.
ARTICLE IN PRESS 518
M.M. El-Nahass et al. / Optics & Laser Technology 37 (2005) 513–523
absorption spectra is also shown in Fig. 4. As observed there is little broadening of the Q and B bands for the film after annealing. In recent years there has been an interest in extending the absorption band of phthalocyanines into the near infrared region of the spectrum for a range of potential functional applications including optical data storage and security printing [43]. 3.2.2. Optical constants The spectral distribution of TðlÞ and RðlÞ at normal incidence in the wavelength range 190–2100 nm for the as-deposited and annealed NiPc films in the thicknesses range 230–770 nm are shown in Fig. 5. As a whole observation, it could be noted that at longer wavelength lo900 nm all films become transparent and no light is scattered or absorbed, i.e. T þ R ¼ 1 while the inequality R þ To1 at shorter wavelengths lo900 nm is due to the existence of absorption. Taking into account, the experimental error in measuring the film thickness to be72% and in n and k to be71%, the error in the calculated values of n and k was estimated to be73% and 2.5%, respectively, the optical constants were found to be independent of the film thickness for all the studied films.
Fig. 6 shows the dispersion of the refractive index nðlÞ in the wavelength range 190–2100 nm for the asdeposited and annealed NiPc films. It is observed that there is little variation in the position of peaks but the magnitude of the peaks are greater for the annealed films than that of as-deposited films. It is also observed that there is anomalous dispersion at lo900 nm as well as normal dispersion at l4900 nm: In anomalous dispersion region, there are four peaks and one shoulder. This behavior obeys a multioscillator model as follows [44]: 1 NP ¼ ¼ A þ Bu2 þ2 3 1X 1 1 þ Jj þ ; p j uoj u iGj uoj þ u þ iGj
ð10Þ
where G is the linewidth, u0j the resonance frequency and J is the intensity factor, which are related to the oscillator concentration N and the transition matrix element. These intensity factors are proportional to the integrated line intensity; P is the microscopical polarizability. The values A and B take the influence of resonance frequencies above 50,000 cm1 into account. The refractive index dispersion has been analyzed using
Fig. 5. Spectral distribution of TðlÞ and RðlÞ of NiPc: (a) as-deposited film and (b) Annealed film.
ARTICLE IN PRESS M.M. El-Nahass et al. / Optics & Laser Technology 37 (2005) 513–523
519
Fig. 6. Dispersion of the refractive index of NiPc: (a) as-deposited film and (b) annealed film.
the single oscillator model, in the normal dispersion region, developed by Toyode [45] and Wemple and DiDomenico [46]. They introduced an energy parameter Ed to describe the dispersion of the refractive index. In terms of this dispersion energy Ea and a single-oscillator energy E0, the refractive index at a frequency can be expressed in the form [45]: n2 ðnÞ 1 ¼ E d E 0 =½E 20 ðhnÞ2 :
(11)
It was found that the dispersion energy Ed, which is related to the interband transition strength, obeys an extraordinarily simple empirical relation for over 100 widely different ionic and covalent nonmetallic crystals and liquids [46]. Based on experimental observation, Wemple and DiDomenico [46] showed that: (i) Ed is independent of the absorption threshold (band gap) within experimental error, (ii) Ed is independent of the lattice constant within experimental error; and (iii) E0 is related to the lowest direct gap. In practice, the single-oscillator energy E0 and the dispersion energy Ed can be obtained from Eq. (11) by plotting ðn2 21Þ1 against E2. Fig. 7 shows an example of such a plot for the as-deposited and annealed NiPc thin films. It is observed that the plots are linear over the energy range from 0.77 eV to approximately 1.16 eV. The determination of E0 and Ed can be made from the plot, using the linear fit parameters which are calculated by the plotting program used. The values of Ed and E0 are obtained from the intercepts and the slopes of the curve as 3.570.03 and 8.0670.04 eV, respectively, for the as-deposited films and 3.5270.03 and 7.8470.04, respectively, for the annealed films. As observed there is a little effect of annealing on the oscillating and dispersion energies. A simple connection between the single-oscillator parameters E0 and Ed and the imaginary part of the dielectric constant 2 ðoÞ spectrum can be expressed in terms of moments of the 2 ðoÞ as [46] E 20 ¼ M 1 =M 3
(12)
Fig. 7. Plot of ðn2 1Þ1 versus ðhnÞ2 of NiPc:(a) as-deposited film and (b) annealed film.
and E 2d ¼ M 31 =M 3 :
(13)
The oscillator energy E 0 is independent of the scale of 2 and is consequently an ‘‘average’’ energy gap, whereas Ed depends on the scale of 2 and thus serves as an interband strength parameter. Since the 1 and 3 moments are involved in computation of E 0 and E d : The values of M1 and M3 are 2.3 and 0.188 eV2, respectively, for the as-deposited films and 2.23 and 0.179 eV2, respectively, for the annealed films. The relation between the infinitely high-frequency dielectric constant 1 ; wavelength l and refractive index (n) is given by [47] n2 ¼ 1 e2 =pc2 ðN=m Þl2 ;
(14)
where 1 is the infinitely high-frequency dielectric constant and N=m is the ratio of carrier concentration to the effective mass. The n2 versus l2 plot shown in Fig. 8 is linear verifying Eq. (14). The values of 1 and (N=m ) are determined from the extrapolation of these plots to l2 ¼ 0 and from the slopes of the graph. The obtained values are 3.71 and 1.2 1047 gm1 cm3 for the as-deposited films and 3.62 and 1.92 1048 gm1 cm3 for the annealed films. The results obtained here are in reasonably good agreement with previously reported results [47] for metal-free phthalocyanine. The spectral distribution of the absorption coefficient að¼ 4pk=lÞ for the investigated as-deposited and annealed NiPc films is shown in Fig. 9. As observed, the behavior of the spectra is similar to the absorption spectra, but has little effect of annealing on the behavior of the absorption coefficient. The optical absorption coefficient of a solid describes the exponential decay of light intensity with distance into the solid. It is the optical absorption coefficient of a solid that describes the exponential decay of light intensity with distance
ARTICLE IN PRESS 520
M.M. El-Nahass et al. / Optics & Laser Technology 37 (2005) 513–523
Fig. 8. Plot of n2 versus l2 of Nipc:(a) as-deposited film and (b) annealed film.
Fig. 10. Plot of emolar versus n¯ of NiPc.
The spectral parameters namely oscillator strength f, electric dipole strength q2 and absorption half-band width Dl are determined from the following expressions [47] using the fit calculation: Z 9 molar dn; (16) f ¼ 4:38 10 1 Dl molar : (17) 2500 l In phthalocyanine molecule,the oscillator strength of an interaction between the central metal atom (the metal atom is positioned at the center of the Pc ring) and the Pc ring. The calculated parameters at different band regions are calculated and collected in Table 1. A comparison of these results shows that all parameters varies in the sequence of band regions. The results obtained in Table 1 are in reasonably good agreement with previously reported results of other phthalocyanines [47]. There are different quantities frequently used to characterize the optical properties of the dielectric relaxation time t, which can be evaluated using the 1 1 tðsÞ ¼ ; (18) o2 q2 ¼
Fig. 9. Plot of a versus hn of Nipc: (a) as-deposited film and (b) annealed film.
into the solid. It is useful to relate a to the molar extinction coefficient emolar, which is often used to describe the absorption of light by nonsolid molecular media. If the solid has a concentration of N molecules per unit volume, each with molecular weight M; a and emolar are related by the expression [41] a¼
N r 103 lnð10Þmolar ¼ 103 lnð10Þmolar L M ¼ const molar ;
ð15Þ
where L is the Avogadro’s number, r the solid’s mass density, and emolar is in units of liters per mole cm. In our NiPc films, the reported density, r ¼ 1:59 g=cm3 ; M=571.23 and the constant in Eq. (15) becomes 6.41 mol/l. The spectral distribution of the molar extinction coefficient is shown in Fig. 10. Clearly, a Gaussian fit of the molar extinction coefficient of NiPc thin films leads to significantly good results, which is in good agreement with the behavior of the absorption coefficient.
where e1, e2 are the real and imaginary dielectric constants, respectively, and o is the frequency. The dielectric relaxation time represents the time of a single oscillation of a dipole in the potential well. Fig. 11 represents the dielectric relaxation time as a function of the photon energy hn for the as-deposited and annealed NiPc thin films. It is observed that the figure exhibit two absorption bands: the Soret band. In the near UV (3.77 and 3.59 eV for the as-deposited and annealed films respectively) and the Q-band in the visible region (2.13, 1.78 and 1.67 eV for the as-deposited films and 2.08, 1.81 and 1.65 eV for the annealed films). The peaks at 2.89 and 2.77 eV for the as-deposited and annealed films,
ARTICLE IN PRESS M.M. El-Nahass et al. / Optics & Laser Technology 37 (2005) 513–523
521
Table 1 The spectral parameters at different band regions E (eV)
a ( 105/cm)
molar ( 105) (mol/l cm)
f ( 104)
q2 ( 102) (A1)
1.81 2.01 3.85 4.45 5.8
1.31 1.82 3.4 2.3 2.72
0.23 0.31 0.56 0.396 0.45
0.97 1.32 2.45 1.72 1.95
0.91 1.8 4.05 2.3 3.28
Fig. 12. Plot of ðahnÞ1=2 versus hn of NiPc: (a) as-deposited film and (b) annealed film.
the analysis of the spectral dependence of the absorption near the fundamental absorption edges. In the above two regions the absorption coefficient a is well described by the relation [47]. ahn ¼ Bðhn E g Þm ;
Fig. 11. Plot of t versus hn of NiPc:(a) as-deposited film and (b) annealed film.
respectively, may be attributed to band-to-band absorption because this maximum occure at hn E g which corresponds to the fundamental energy gap for most metal derivative phthalocyanines [19,29,48–50]. 3.2.3. Energy gap determination Fig. 9 indicates that NiPc thin films have a strong absorption band between 1.45 and 1.8 eV and another strong band starting at 2.7 eV which extends into the ultraviolet. The optical band gap was determined from
(19)
where hn is the energy of incident photons and Eg is the value of the optical band gap corresponding to transitions indicated by the value of m. The factor B depends on the transition probability and can be assumed to be constant within the optical frequency range. Plots of a1=2 versus hn near the absorption edge of the Q and B bands for the as-deposited and annealed films produce a linear fit over a wider range in hn as shown in Fig. 12. This is the characteristic behavior of indirect transitions in nanocrystalline materials. This type of transition is in agreement with Kumar et al. [47] for free and rare-earth phalocyanine and disagreement with Collins et al. [29] for PbPc thin films and Ambily and Menon [7] for CuPc thin films and El-Nahass et al. [50] for FePc thin films. The extrapolation of the straight line graphs to zero absorption will give the value of the optical band gap. The obtained energy band gaps are 1.6670.02 eV for the onset of the absorption spectrum and 2.7770.03 eV for the fundamental energy gap. As observed from Fig. 13 the band gap does not show any remarkable difference from its value in the as-deposited and the annealed films. This may be attributed to the two polymorphs that have the same system and differ only in the tilt angle of the molecules within the columns and the mutual arrangement of the columns [2]. The urbach tail was found to be related directly to similar exponential tail for the density of states of either one of the two band edges [51]. The origin of this exponential tail absorption probably arises from the random fluctuation of the internal fields associated with structural disorder which are considerable in an amorphous solid [51]. The width of the urbach tail is an indicator of disorder in the material. The following
ARTICLE IN PRESS M.M. El-Nahass et al. / Optics & Laser Technology 37 (2005) 513–523
522
Fig. 13. Plot of ln a versus hn: (a) as-deposited film and (b) annealed film.
relation was used to calculate the slope of the urbach tail [52] Fig. 14. Suggested energy band diagram of NiPc.
a ¼ ao e
hu=E t
;
(20) 4. Conclusion
where ao is the background absorption and Et is the parameter defining the slope (width) of the tail. Fig. 13 shows the linear dependence of the natural logarithm of the absorption coefficient of Fig. 12 (in the range 1.3–1.6 eV) versus the photon energy for the asdeposited and annealed films. The reciprocal of the slopes of each line yields the magnitude of Et. The obtained values of Et are 93.470.04 and 90.870.05 meV for the as-deposited and annealed films, respectively. The decrease of Et with annealing temperature indicates a decrease in the disorder in NiPc films. We have proposed a band structure of NiPc in which the bands are constructed from molecular orbitals of Ni dxz;yz and pyrrole carbon Pz electronic states and this is schematically shown in Fig. 14. In view of the ionicity of NiPc it is not clear how far this molecular orbital approximation can be used in this material. Since the band structure of NiPc has not been calculated, the model is only schematic and does not reflect the extent of electron overlap on the energy separation between different electronic states. Following this model we would suggest that the energy gaps of 1.66 and 2.77 eV, estimated from the energy dependence of the absorption coefficient (Fig. 9), could correspond to a transition involving the valence band (HOMO with eg symmetry) to the bottom of the conduction band (LUMO with a1u symmetry) and to the next unoccupied orbital. The highenergy peaks at 3.65, 4.5 and 5.9 eV shown in Fig. 9 may correspond to transitions involving the valence band to the split d band (b1g, eg and b1u). These results are consistent with other reported calculation on NiPc electronic structure [20].
X-ray powder diffraction patterns of Nickel phthalocyanine have shown that the initial powder is rich in bform. Films deposited at room temperature assumed rich in a-form with a preferential orientation in the (1 0 0) direction. When the films were heat-treated at about 523 K, X-ray diffraction analysis showed that they reverted partially to b-form, preferentially oriented in the same (1 0 0) direction. IR spectral analysis was employed to confirm the X-ray results and indicated that the annealed films did not totally transform to the b-form. The optical characterization of the as-deposited and annealed films have been studied in the spectra range 190–2100 nm. The stability in the peak positions in transmission and reflection spectra in absorbing region showed the stability of the structure of NiPc thin films. The refractive index as well as the absorption index are practically independent on the film thickness. The spectral distribution of the absorbance and the absorption coefficient of the investigated NiPc films characterized by distinct peaks in visible region have generally been interpreted in terms of p2p excitation. The absorption spectra occur in the high-energy region of the Soret band indicate the presence of d-band associated with the central metal atom. It is thought that p–d transitions are involved because NiPc has partially occupied d-band. Some of the important optical parameters such as the oscillating energy, the dispersion energy, the infinitely frequency dielectric constant, the molar extinction coefficient, oscillator strength, the electric dipole strength and the dielectric relaxation time have been evaluated. The mechanism of the electronic
ARTICLE IN PRESS M.M. El-Nahass et al. / Optics & Laser Technology 37 (2005) 513–523
transition can be understood in terms of the electron band structure of this material. Although no theoretical work has been done on the band structure of this material. No remarkable changes were observed in the parameters for the annealed samples compared to the as-deposited.
References [1] Leznoff CC, Lever ABP, editors. Phthalocyanines properties and applications. New York: VCH publishers Inc; 1996. p. 219. [2] Djurisic AB, Kwong CY, Lau TW, Guo WL, Li EH, Liu ZT, Kwok HS, Lam LSM. Opt Communications 2002;205:155–62. [3] Cook MJ. Pure Appl Chem 1999;71:2145. [4] Gregorg P. High-technology application of organic colorants. New York: Plenum Press; 1991. p. 759. [5] AO R, Kilmmert L, Haarer D. Adv Mater 1995;7:495. [6] Yoshino K, Hikida M, Tatsuno K, Kaneto K, Inushi Y. J Phys Soc Jpn 1973;34:44. [7] Ambily S, Menon CS. Thin Solid Films 1999;347:284. [8] Krishnakumar KP, Menon CS. J Solid State Chem 1997;128:27. [9] Brinkmann M, Wittmann JC, Chaumont C, Andre JJ. Thin solid Films 1997;292:192. [10] Berger O, Fischer WJ, Adolphi B, Tierbach S, Melev V, Schreiber J. J Mater Sci 2000;11:331. [11] Bayliss SM, Heutz S, Rumbles G, Jones TS. Mater Res Soc Symp Proc 1999;560:71. [12] Shihub SI, Gould RD. Thin Solid Films 1996;290:390. [13] Davidson AT. J Chem Phys 1992;77:168. [14] Schmeisser D, Rager A, Thonke K, Pilkuhn M, Frohlich D, Gaauglitz G, Schafer M, Oelkrug D. Synth Met 1991;41:1457. [15] Mochert H, Schmeisser D, Gopel W. Sensors and Actuators 1989;19:159. [16] Ukei K. J Phys Soc Jpn 1976;40:140. [17] Heilmeier GH, Warfield G. J Appl Phys 1962;34:2278. [18] Heilmeier GH, Warfield G, Harrison SE. Phy Rev Lett 1962;8:309. [19] Ambily S, Menon CS. Thin Solid Films 1999;347:284. [20] Bialek B, Kim IG, Lee JI. Synth Met 2002;129:151. [21] Achar BN, Jayasree PK. Synth Met 2000;114:219. [22] Chaabane RB, Guillaud G, Gamoudi M. Thin Solid Films 1997;296:145.
523
[23] Tolansky S. Multiple—beam interferometry of surfaces and films. New York: 1965. p. 55. [24] Laaziz Y, Bennouna A, Chahboun N, Outzourhit A, Ameziane A. Thin Solid Films 2000;372:149. [25] Heavens OS. Optical Properties of thin solid films. New York: Dover publication; 1965. p. 55. [26] Bennett JM, Booty MG. Appl Opt 1966;5:40. [27] Tomlin SG. J Phys D 1968;1:1667. [28] Abele F, Theye MI. Surg Sci 1966;5:325. [29] Collins RA, Krie A, Abass AK. Thin Solid Films 1993; 229:113. [30] Hassan AK, Gould RD. Phys Stat Sol (A) 1992;132:91. [31] Komiyama M, Sakaibara Y, Hirai H. Thin Solid Films 1987;151:109. [32] Debe MK, Kan K. Thin Solid Films 1990;186:298. [33] Shihub SI, Gould RD. Phys. Stat. Sol. (A) 1993;139:129. [34] Fan FR, Faulkner LR. J Chem Phys 1978;69:3334. [35] Boudgema B, Guilland G, Gamoudy M, Maitrot M, Andre JJ, Martin M, Simon J. J Appl Phys 1984;56:2323. [36] Assour JM. J Phys Chem 1965;69:2295. [37] Sharp JH, Abkowitz M. J Phys Chem 1973;77:477. [38] Stillman MJ, Nyokong TN. In: Leznott CC, Lever ABP, editors. Phthalocyanines: properties and applications, vol. 1. New York: VCH; 1989. p. 133 [Chapter 3]. [39] Krishnakumar KP, Menon CS. J Solid State Chem 1997; 27:128. [40] Davidson AT. J Chem Phys 1982;77:162. [41] Schechtman BH, Spicer WE. J Mol Spect 1970;33:28. [42] Edwards L, Gouterman M. J Mol Spect 1970;33:292. [43] Christie MR, Freer GB. Dyes Pigments 1994;24:113. [44] Stenzel O, Wilbrand S, Stendal A, Beckers U, Voigtsberger K, Borczyskowski CV. J Phys D 1995;28:2154. [45] Toyoda T. J Appl Phys 1988;63:5166. [46] Wemple SH, DiDomenico MD. Phys Rev B 1970;3:1338. [47] Kumar GA, Thomas J, George N, Kumar BA, Shnan PR, Poori VPN, Vallabhan CPG, Unnikrishnan NV. Phys Chem glasses 2000;41:89. [48] Bell RJ, Ordal MA, Alexender RW. Appl Opt 1985;24:3680. [49] El-Nahass MM, Bahabri FS, Al-Harbi SR. Egypt J Sol 2001;24:11. [50] El-Nahass MM, Soliman HS, Metwally HS, Farid AM, Farag AAM, El-Shazly AA. J Opt 2001;30:121. [51] Cody GD. J Non-Cryst Solids 1992;141:3. [52] Stutzmann M. In: Mahajan S, editor. Hand book on semiconductors, vol. 3A. Amsterdam: North-Holland; 1994. p. 657.