- Email: [email protected]
Optical absorption in metal bisphthalocyanine sublimed films
Vacuum 61 (2001) 19}27
Optical absorption in metal bisphthalocyanine sublimed "lms A.K. Ray *, J. Exley , Z. Ghassemlooy , D. Crowther, M.T. Ahmet, J. Silver Physical Electronics & Fibre-optics Research Laboratories, School of Engineering, Information Technology, Shezeld Hallam University, City Campus, Pond Street, Shezeld S1 1WB, UK Department of Chemistry, Shezeld Hallam University, City Campus, Pond Street, Shezeld S1 1WB, UK Department of Chemistry and Biological Chemistry, University of Essex, Wivenhoe Park, Colchester, Essex CO4 3SQ, UK
Abstract Visible optical absorption and re#ection spectra are obtained for 50 nm thick sublimed "lms of heavy fraction rare-earth [HF(pc) (pc*)], gadolinium [Gd(pc) (pc*)] and thulium [Tm(pc) (pc*)] bisphthalocyanine compounds when they have undergone the post-deposition treatments of voltage-cycling to blue, voltage cycling to red and annealing at 393 K for 1 h. The di!erent post-deposition treatments produce di!erent e!ects on the absorption spectra; annealing generates phase changes in the "lms. The changes due to the voltage cycling are believed to be a result of oxidation/reduction processes taking place in the materials. The absorption data yield information on the dispersion of refractive index and dielectric constants within the optical frequency range. The energies of transition are found to be 1.9 and 2.8 eV, respectively for Q- and Soret bands of all untreated samples. The concentration of electrons involved in resonant oscillation is estimated to be in the order of 10 m\ for all types of samples, both fresh and treated. 2001 Elsevier Science Ltd. All rights reserved. Keywords: Phthalocyanines; Electrochromic; Optical absorption
1. Introduction Rare-earth element bisphthalocyanines have been the subject of intense investigation over the last 15 years primarily because of the potential use of these compounds as electrochromic materials [1,2]. These compounds can be fabricated as thin "lms using the Langmuir}Blodgett technique [3}5] and also by sublimation [6]. The applications of such electrochromics can be found in smart windows [7] and their possible uses in chemical sens-
* Corresponding author. Tel.: #44-114-2533409; fax: #44114-2533433. E-mail address: [email protected] (A.K. Ray).
ing have also been explored [8]. The neutral compounds are green when they are oxidised (one electron removed per molecule); they turn red and on reduction they turn blue "rst (one electron added per molecule), then purple (an additional electron added per molecule). The electrochromism is observed by voltage cycling of $1.2 V between these states. The changes in colour occur because of the di!erent number of electrons present in each of the species which causes the electronic orbitals on the phthalocyanine ring to change their energies and thus they have di!erent electronic absorption spectra. The electrical and optical properties of heavy fraction rare-earth bisphthalocyanine [HF(pc) (pc*)] sublimed "lms have recently been reported [9] (it is to be noted that
0042-207X/01/$ - see front matter 2001 Elsevier Science Ltd. All rights reserved. PII: S 0 0 4 2 - 2 0 7 X ( 0 0 ) 0 0 4 2 0 - 6
20
A.K. Ray et al. / Vacuum 61 (2001) 19}27
pc"phthalocyanato dianion, pc*"phthalocyanato mono radical anion, HF is a mixture of heavy fraction rare-earth element as found in partially re"ned heavy fraction ores [1]). Like monophthalocyanines, [HF(pc) (pc*)] "lms are found to be strongly absorbing materials having an absorption coe$cient in the order of 10 m\. The a !e type transitions give rise to the Q-band peak at about 1.9 eV. An extra peak referred to as a radical band, is observed within the energy window 2.2 and 2.4 eV and is denoted by an X-band. Further work is reported here on optical absorption and transmission through "lms of [HF(pc) (pc*)], gadolinium [Gd(pc) (pc*)] and thulium [Tm(pc) (pc*)] phthalocyanine compounds at room temperature in air within the visible range of incident photon energies. This paper examines the e!ects of post-deposition treatments on optical constants of these phthalocyanine molecules. Properties such as refractive index and the changes in refractive index consequent upon voltage cycling and annealing of the "lms are important for understanding the performance of materials when incorporated into working devices.
2. Experimental Using a Phillips PU8720 UV/Visible spectrophotometer, optical absorption spectra were obtained for deposited "lms of [HF(pc) (pc*)], [Gd(pc) (pc*)] and [Tm(pc) (pc*)] compounds at room temperature in air between 1.4 and 4 eV. Re#ectance measurements were carried out using a similar instrument equipped with an integrating sphere, enabling direct measurement of either diffuse or a combination of specular and di!use re#ection. For these investigations, the "lms were prepared on ITO glass substrates by vacuum sublimation. All samples were of the order of 50 nm thick. The preparation of the "lms and the measurement of the "lm thickness have been described [9]. A monochromatic beam of intensity I (photons/m s) was allowed to be incident perpendicular to the plane of the "lms. An uncoated substrate was used as a reference for absorption, and a surface silvered mirror for re#ection, so that the outputs I or I were solely in terms of the transR 0
mission characteristics of the "lms. Along with neutral (untreated) "lms of these compounds, measurements were also performed on "lms which had undergone the following post-deposition treatments: (i) the neutral "lm was voltage cycled "rst to red ("ve cycles at $ 1.2 V), (ii) the neutral "lm was voltage cycled to blue ("ve cycles at $ 1.2 V) and (iii) the neutral "lm was held at 393 K for 1 h to anneal the material. The combined specular and di!used re#ectance were measured with the sample in the 83 holder, followed by the di!use measurements only in the perpendicular holder. From these results, it is possible to establish the surface roughness from the ratio of specular to di!use re#ection.
3. Results and discussions Fig. 1(a) displays four sets of absorption spectra for heavy fraction rare-earth bisphthalocyanine [HF(pc) (pc*)] sublimed "lms within the visible optical frequency range, one set for a particular type of treatment. As reported previously [9], the Q-band peak shown in Curve A of Fig. 1(a) for the [HF(pc) (pc*)] untreated sample is not sharp, probably due to the interaction of molecules in the solid phase. The di!erent post-deposition treatments produce di!erent e!ects on the intensity of absorption but, irrespective of the post-deposition treatment, the intensity of the Q-band absorption always remains higher than the X-band absorption. Upon annealing, the intensity of absorption increases by approximately a factor of two and, at the same time, the peak becomes narrower. On voltage cycling to blue, there is a noticeable broadening of the Q band with a shoulder and the peak intensity is increased by a factor of about 1.5 over that for the untreated sample. For the voltage cycling to red, the opposite e!ect is observed in terms of diminished Q-band peak. Annealing also increases the intensity of the X-band; cycling to blue causes the X band to almost disappear whereas cycling to red causes a much wider X band (2.3 to 2.7 eV). It may be inferred that the e!ect of voltage cycling on this
A.K. Ray et al. / Vacuum 61 (2001) 19}27
21
Fig. 1. Three groups of graphs showing the visible absorption spectra for sublimed "lms of (a) heavy fraction rare-earth [HF(pc) (pc*)], (b) gadolinium [Gd(pc) (pc*)] and (c) thulium [Tm(pc) (pc*)] phthalocyanine compounds. Measurements are performed when the "lms are (i) untreated, (ii) annealed at 393 K for 1 h, (iii) voltage-cycled "ve times to blue and (iv) voltage-cycled "ve times to red.
material is to cause a change in the oxidation states present in the return to `neutrala material. Fig. 1(b) gives the similar plots of optical absorption spectra for thulium phthalocyanine compounds [Tm(pc) (pc*)]. There are two interesting features in the spectrum: (i) a shoulder to the Q band at 2.1 eV and (ii) a pronounced absorption edge at 1.7 eV. The blue shoulder is attributed to the Q (1!0) vibronic band, generally more clearly resolved at lower temperatures. Like the other sample compounds, the tail towards red is due to smearing of many hot bands. Annealing of the sample does not appear to cause any noticeable shift of the Q-band peak along the energy scale, but the overall absorption is decreased. For 1.2 V volt-
age cycling to blue "ve times, the overall absorption increases and there is a very slight blue shift in the Q band. The intensity of absorption decreases slightly on voltage cycling to red and there is a pronounced e!ect on the shape of the shoulder at about 2.1 eV. For the sample which has been voltage-cycled to red, a signi"cant increase in absorption between the peak at 2.6 eV and the Q band is observed. As shown in Curve A of Fig. 1(c), the Q and B bands in the spectra for the untreated gadolinium phthalocyanine [Gd(pc) (pc*)] "lm are of approximately equal strength, in addition to the X band in the energy window centred at around 2.6 eV. The 5;1.2 V voltage cycling to blue causes a decrease
22
A.K. Ray et al. / Vacuum 61 (2001) 19}27
in the absorption at the Q and X bands but the Q band is now split, with one peak at an increased energy. Cycling at the same voltage to red causes a decrease in the peak absorption of the Q band; the X band is less signi"cantly decreased than for the voltage-cycled to blue. The annealing thus has a pronounced e!ect on the [Gd(pc) (pc*)] spectra: (i) an increase in the energy at which the Q band occurs in the spectrum, (ii) a decrease in the peak absorption and (iii) the broadening of the Q band. This suggests that the e!ect of the annealing process has been to promote a change in the structure of the "lm, possibly, a phase change in the "lm material. This is similar to previously published data for the transformation of -FePc to the form [10]. Table 1 summarises the main points of our observations which include the positions in the energy domain for the B, X and Q bands, together with their peak intensity absorbance values. The relative heights of the Q and X band peaks are modi"ed as a result of the post-deposition treatments. Fig. 2 shows the spectrum of optical re#ectivity R for the same batch of samples. The sharp changes
in re#ectance correspond to an electronic transition. It is observed on close examination (vide the insets) that the Soret band a Pe and Q-band a Pe transitions take place at 2.8 and 1.9 eV, respectively. The absorption coe$cient which is de"ned as the reciprocal of a distance over which the energy in the wave falls o! to exp(!1) can be determined in terms of I and I as [11] R I R "(1!R) exp(!t)/1!R exp(!2t), (1) I where t is the thickness of the "lm. The extinction coe$cient, , is, on other hand, related to the absorption coe$cient in the form " . 4
(2)
Using t"50 nm in Eq. (1), values of the absorption coe$cient are calculated for all three untreated samples from values of [I /I ] obtained from Fig. 1. R With a knowledge of the values of the absorption coe$cient , the value of the extinction coe$cient,
Table 1 Absorption peak heights and energies at which they occur for the three materials Compounds
B band
X band
Q band
and treatment
Energy (eV)
Height
Energy (eV)
Height
Energy (eV)
Height
HF ( pc) ( pc*) Untreated Annealed Voltage cycled (Red) Voltage cycled (Blue)
3.9 3.8 3.8 3.9
0.55 0.71 0.50 0.51
2.4 2.6 2.6 2.5
0.14 0.21 0.14 0.16
1.9 1.9 1.9 1.9
0.29 0.51 0.40 0.27
Tm( pc) ( pc*) Untreated Annealed Voltage cycled (Red) Voltage cycled (Blue)
3.6 3.8 3.7 3.9
0.67 0.83 0.88 0.92
2.7 2.6 2.7 2.6
0.41 0.32 0.38 0.35
1.9 1.9 1.9 1.9
0.84 0.78 0.99 0.75
Gd( pc) ( pc*) Untreated Annealed Voltage cycled (Red) Voltage cycled (Blue)
3.3 3.3 3.3 3.3
0.30 0.25 0.18 0.27
2.4 2.5 2.4 2.4
0.21 0.17 0.16 0.19
1.9 1.9 1.9 1.9
0.28 0.26 0.21 0.28
A.K. Ray et al. / Vacuum 61 (2001) 19}27
23
Similar behaviour is observed for monophthalocyanines [12]. For an absorbing medium such as these metal bisphthalocyanine "lms, the real part of refractive index may be found from the re#ectivity R using the following relation: (!I)# R" . (#I)#
(3)
Eq. (3) which is valid for monochromatic radiation at normal incidence on the "lm/air interface [13] is reduced to a set of quadratic equations in , when values of R and are substituted. The physically meaningful root of these equations gives the value of , the real part of the complex refractive index, or propagation constant. The dispersion behaviour of is shown for all three samples in Fig. 4. takes a value between 2 and 3 throughout the energy range for [HF(pc) (pc*)] and [Tm(pc) (pc*)], but for [Gd(pc) (pc*)], it is a little higher, approaching 4.1 at the maximum. The observed dispersive behaviour does not follow any polynomial of low order for the function (h) where h is the photon energy, and therefore the Cauchy relation is not expected to hold over the entire photon energy range. This anomalous behaviour is in keeping with the Debye theory [14]. The real and imaginary parts , of the dielectric constant * (" #j ) are related to and through the following equations: "! and "2.
Fig. 2. A set of three graphs showing the dependence of re#ectivity R on incident wavelength () for untreated sublimed "lms of (a) heavy fraction rare-earth [HF(pc) (pc*)], (b) gadolinium [Gd(pc) (pc*)] and (c) thulium [Tm(pc) (pc*)] phthalocyanine compounds.
(which is the imaginary part of the complex refractive index N"#j) may be evaluated from Eq. (2). Fig. 3 shows the variation of with respect to the incident photon energy. The variations of extinction coe$cient follow a similar pattern to those for absorbance, as expected from the relationship between the two parameters [vide Eq. (2)].
(4)
Resulting dependences of , on photon energy h are presented for untreated samples in Fig. 5. For all "lms, the variation of the real part broadly follows the same pattern as the imaginary part, both of which resemble the envelope of the absorption spectrum. In general, values of the real part are appreciably higher. The obvious peaks are seen at an energy between 1.5 and 2.2 eV, depending upon the type of material. The (h) curves are believed to be caused by the electron oscillation with resonant frequency corresponding to the N peak. The electron density N can be determined using classical oscillation theory [15]: m ( ) N" N . e
(5)
24
A.K. Ray et al. / Vacuum 61 (2001) 19}27
Fig. 3. A set of three curves showing the variations of extinction coe$cient (imaginary part of refractive index) for untreated sublimed "lms of (a) heavy fraction rare-earth [HF(pc) (pc*)], (b) gadolinium [Gd(pc) (pc*)] and (c) thulium [Tm(pc) (pc*)] phthalocyanine compounds with respect to incident photon energy h within the visible range.
Fig. 4. A set of three graphs showing the variations of optical propagation constant n (real part of refractive index) for untreated sublimed "lms of (a) heavy fraction rare-earth [HF(pc) (pc*)], (b) gadolinium [Gd(pc) (pc*)] and (c) thulium [Tm(pc) (pc*)] phthalocyanine compounds with respect to incident photon energy h within the visible range.
It is found that values of N lie in the order of magnitude of 10 m\. Calculations have been performed for the samples which have undergone post-deposition treat-
ments and the results are summarised in Table 2. The e!ects of the post-deposition treatments are apparent in the values of the propagation constant in all cases. The energy shifts in the peaks of the
A.K. Ray et al. / Vacuum 61 (2001) 19}27
25
Fig. 5. A set of three graphs showing the dependences of real ( ) and imaginary ( ) parts of the dielectric constant (*) for untreated sublimed "lms of (a) heavy fraction rare-earth [HF(pc) (pc*)], (b) gadolinium [Gd(pc) (pc*)] and (c) thulium [Tm(pc) (pc*)] phthalocyanine compounds on incident photon energy h within the visible range. Circles and squares symbols are used for and , respectively.
propagation constants observed in the calculated spectra after treatments are marked for the X band; additionally the intensities of the two peaks for [HF(pc) (pc*)] are approximately reversed, indicating the phase change induced by the annealing process. The upper and lower limits of the propagation constant, however, remain reasonably constant throughout the treatments. The e!ects broadly follow the absorption spectra. The modi"cation of energy values at which the peaks occur in the , and spectra are generally more pronounced than those visible in the absorbance data. In particular, the e!ects of annealing upon the X or characteristic band centre are very pronounced for [HF(pc) (pc*)] compounds. For the other three treatments, the
major peak is the original characteristic (X) band location for the untreated material. The process of annealing causes a diminution of the X band at this energy location to a great extent, and an increase of the intensity on the other part of the split peak corresponding to the phase. Thus, the calculation of fundamental optical constants provides a less ambiguous assessment of the structural and chemical changes occurring in the "lms during treatment. Surface roughness of the "lms can be given in terms of the exponential dependence of irregularity height on incident wavelength :
4 RJR exp ! ,
(6)
26
A.K. Ray et al. / Vacuum 61 (2001) 19}27
Table 2 Peak values of real part ()of refractive index, real ( ) and imaginary ( ) parts of dielectric constant (*) with their corresponding incident energy
Compound and treatment
Peak
Energy (eV)
Peak
Energy (eV)
Peak
Energy (eV)
N (;10 m\)
[HF( pc) ( pc*)] Untreated Annealed Voltage cycled (Red) Voltage cycled (Blue)
2. 7 2.7 1.9 2.8
2.0 2.1 2.8 1.9
7.2 9.5 7.9 7.6
1.9 1.5 1.8 1.9
1.7 3.3 1.8 2.4
1.9 1.5 1.8 1.9
1.7 1.1 1.4 1.0
[Gd( pc) ( pc*)] Untreated Annealed Voltage cycled (Red) Voltage cycled (Blue)
4.0 3.0 2.7 3.0
1.6 2 1.9 1.9
7.7 9.0 8.0 8.0
2.1 1.9 1.9 1.9
1.4 1.8 2.3 2.5
2.1 1.9 1.9 1.9
1.9 1.0 1.6 1.6
[Tm( pc) ( pc*)] Untreated Annealed Voltage cycled (Red) Voltage cycled (Blue)
2.8 2.8 2.7 2.8
2.2 2.2 2.0 2.2
8 7.9 5.7 7.9
2.2 2.2 2.2 2.2
2.1 1.3 1.2 1.1
2.2 2.2 2.2 2.2
1.3 0.5 0.6 0.6
N is the concentration of electrons at resonant frequency of oscillation.
Fig. 6. A set of three graphs showing the wavelength dispersion of logarithmic normalised re#ectance for untreated sublimed "lms of (a) heavy fraction rare-earth HF(pc) (pc*)], (b) gadolinium [Gd(pc) (pc*)] and (c) thulium [Tm(pc) (pc*)] phthalocyanine compounds.
where R is re#ection from a smooth surface. This formula is applicable for our case since the angle of acceptance is small [16].
As shown in Fig. 6, the plots of ln(R/R ) as a function of the inverse of the square of incident wavelength are linear allowing a value of to be
A.K. Ray et al. / Vacuum 61 (2001) 19}27
determined from the slope. Using data from the graphs, values of 18.5 nm for HF(pc) (pc*), 26 nm for Gd(pc) (pc*) and 12.1 nm for Tm(pc) (pc*) have been obtained. These values are very high relative to the "lm thickness, although a result of this magnitude might be expected for a thin, vacuum-evaporated layer which, in these cases, is visibly inhomogeneous. It is also hypothesised that this may be largely a factor consequence of the roughness of the relatively thick ITO "lm between the pc and the glass substrate, the thin phthalocyanine "lm following the contours of this.
4. Concluding remarks The e!ects of the various post-deposition treatments upon the optical frequency dielectric properties of these three materials have been examined. For phthalocyanines, electronic transitions occur notably in the 1.5}2.0 eV visible region (the Q band) and 3}4 eV (the B or Soret band), both being due to }* transitions. The Q band is highly localised on the phthalocyanine ring, and is reported for many phthalocyanines to be highly susceptible to a number of macrostructural factors, particularly to the geometry of nearest-neighbour molecules in the structure [17]. This in the light of the results appears to be true for the three compounds investigated and signals that the phase changes are promoted by annealing. This also suggests that no structural damage to the phthalocyanine rings has occurred during the treatment. The e!ect of annealing upon the Q band intensity is di!erent for [HF(pc) (pc*)] compared with the other two compounds. In the case of [HF(pc) (pc*)] there is an increase in absorption but the other two compounds exhibit a decrease in absorption. This e!ect must be related to di!erences in the structure of [HF(pc) (pc*)] compared to the other materials; [(Lu(pc) (pc*)] is reported to form liquid crystals, so the annealing process causes closer ordering in this material than is noticeable in the raw state. Annealing [HF(pc) (pc*)] shows a narrowing of the Q band and an increase in its intensity; this may be due to a reordering of molecules in the structure, leading to a more highly ordered phase. The reverse e!ect on the other two compounds
27
indicates a more random ordering illustrated by the peak broadening. The reversible redox processes resulting from voltage cycling are reduction or voltage cycling to blue giving the derivative [M(pc) ]\, and oxida tion or voltage cycling to red giving [M(pc*) ]>. For N, the number of electrons per unit volume participating in oscillation at peak frequency, values in the order of 10 m\ are obtained for all materials, irrespective of treatment. Acknowledgements The authors are grateful to Dr. J.R. Travis for his encouragement and fruitful discussions. One of us (JE) thanks the Science and Engineering Research Council, UK for his studentship. References [1] Frampton CS, O'Connor JM, Peterson J, Silver J. Displays Tech Appl 1988;9:174. [2] Nicholson MM. In: Lezno! CC, Lever ABP, editors. Electrochromism and display devices of Phthalocyanines, Properties and Applications, vol. 3. NewYork: VCH Publishers, 1993, p. 77 [Chapter 2]. [3] Petty M, Lovett DR, Townsend P, O'conor JM, Silver J. J Appl Phys 1989;22:1604. [4] Petty M, Lovett DR, O'Connor JM, Silver J. Thin Solid Films 1989;79:387. [5] Petty M, Lovett DR, Miller JR, Silver J. J Mater Chem 1991;1:971. [6] Song SA, O'Connor JM, Barber DJ, Silver J. J Crystal Growth 1988;88:477. [7] Silver J, New Scientist, 30th September 48, (1989). [8] Souto J, Aroca R, Desaja JA. J Raman Spectroscopy 1991;22:349. [9] Exley J, Ray AK, Ahmet MT, Silver J. J Mater Sci Mater in Electron 1991;5:59. [10] Silver J, Lukes P, Houlton A, Howe S, Hey P, Ahmet MT. J Mater Chem 1992;2:849. [11] Swan R, Ray AK, Hogarth CA. Phys Stat Sol(a) 1991;127:555. [12] Collins RA, Krier A, Abass AK. Thin Solid Films 1993;229:113. [13] Moss TS, Burrell GJ, Ellis B. Semiconductor opto-electronics. London: Butterworth, 1973. p. 10. [14] Elliott SR. Advances in Physics 1987;36(2):138. [15] Cook BE, Spear WE. J Phys Chem Solids 1969;30:1125. [16] Gupta P, Maiti B, Maity AB, Chaudhuri S, Pal AK. Thin Solid Films 1995;260:75. [17] Ho ZZ, Ju CY, Hetherington III WM. J Appl Phys 1987;62:716.
Optical absorption in metal bisphthalocyanine sublimed "lms A.K. Ray *, J. Exley , Z. Ghassemlooy , D. Crowther, M.T. Ahmet, J. Silver Physical Electronics & Fibre-optics Research Laboratories, School of Engineering, Information Technology, Shezeld Hallam University, City Campus, Pond Street, Shezeld S1 1WB, UK Department of Chemistry, Shezeld Hallam University, City Campus, Pond Street, Shezeld S1 1WB, UK Department of Chemistry and Biological Chemistry, University of Essex, Wivenhoe Park, Colchester, Essex CO4 3SQ, UK
Abstract Visible optical absorption and re#ection spectra are obtained for 50 nm thick sublimed "lms of heavy fraction rare-earth [HF(pc) (pc*)], gadolinium [Gd(pc) (pc*)] and thulium [Tm(pc) (pc*)] bisphthalocyanine compounds when they have undergone the post-deposition treatments of voltage-cycling to blue, voltage cycling to red and annealing at 393 K for 1 h. The di!erent post-deposition treatments produce di!erent e!ects on the absorption spectra; annealing generates phase changes in the "lms. The changes due to the voltage cycling are believed to be a result of oxidation/reduction processes taking place in the materials. The absorption data yield information on the dispersion of refractive index and dielectric constants within the optical frequency range. The energies of transition are found to be 1.9 and 2.8 eV, respectively for Q- and Soret bands of all untreated samples. The concentration of electrons involved in resonant oscillation is estimated to be in the order of 10 m\ for all types of samples, both fresh and treated. 2001 Elsevier Science Ltd. All rights reserved. Keywords: Phthalocyanines; Electrochromic; Optical absorption
1. Introduction Rare-earth element bisphthalocyanines have been the subject of intense investigation over the last 15 years primarily because of the potential use of these compounds as electrochromic materials [1,2]. These compounds can be fabricated as thin "lms using the Langmuir}Blodgett technique [3}5] and also by sublimation [6]. The applications of such electrochromics can be found in smart windows [7] and their possible uses in chemical sens-
* Corresponding author. Tel.: #44-114-2533409; fax: #44114-2533433. E-mail address: [email protected] (A.K. Ray).
ing have also been explored [8]. The neutral compounds are green when they are oxidised (one electron removed per molecule); they turn red and on reduction they turn blue "rst (one electron added per molecule), then purple (an additional electron added per molecule). The electrochromism is observed by voltage cycling of $1.2 V between these states. The changes in colour occur because of the di!erent number of electrons present in each of the species which causes the electronic orbitals on the phthalocyanine ring to change their energies and thus they have di!erent electronic absorption spectra. The electrical and optical properties of heavy fraction rare-earth bisphthalocyanine [HF(pc) (pc*)] sublimed "lms have recently been reported [9] (it is to be noted that
0042-207X/01/$ - see front matter 2001 Elsevier Science Ltd. All rights reserved. PII: S 0 0 4 2 - 2 0 7 X ( 0 0 ) 0 0 4 2 0 - 6
20
A.K. Ray et al. / Vacuum 61 (2001) 19}27
pc"phthalocyanato dianion, pc*"phthalocyanato mono radical anion, HF is a mixture of heavy fraction rare-earth element as found in partially re"ned heavy fraction ores [1]). Like monophthalocyanines, [HF(pc) (pc*)] "lms are found to be strongly absorbing materials having an absorption coe$cient in the order of 10 m\. The a !e type transitions give rise to the Q-band peak at about 1.9 eV. An extra peak referred to as a radical band, is observed within the energy window 2.2 and 2.4 eV and is denoted by an X-band. Further work is reported here on optical absorption and transmission through "lms of [HF(pc) (pc*)], gadolinium [Gd(pc) (pc*)] and thulium [Tm(pc) (pc*)] phthalocyanine compounds at room temperature in air within the visible range of incident photon energies. This paper examines the e!ects of post-deposition treatments on optical constants of these phthalocyanine molecules. Properties such as refractive index and the changes in refractive index consequent upon voltage cycling and annealing of the "lms are important for understanding the performance of materials when incorporated into working devices.
2. Experimental Using a Phillips PU8720 UV/Visible spectrophotometer, optical absorption spectra were obtained for deposited "lms of [HF(pc) (pc*)], [Gd(pc) (pc*)] and [Tm(pc) (pc*)] compounds at room temperature in air between 1.4 and 4 eV. Re#ectance measurements were carried out using a similar instrument equipped with an integrating sphere, enabling direct measurement of either diffuse or a combination of specular and di!use re#ection. For these investigations, the "lms were prepared on ITO glass substrates by vacuum sublimation. All samples were of the order of 50 nm thick. The preparation of the "lms and the measurement of the "lm thickness have been described [9]. A monochromatic beam of intensity I (photons/m s) was allowed to be incident perpendicular to the plane of the "lms. An uncoated substrate was used as a reference for absorption, and a surface silvered mirror for re#ection, so that the outputs I or I were solely in terms of the transR 0
mission characteristics of the "lms. Along with neutral (untreated) "lms of these compounds, measurements were also performed on "lms which had undergone the following post-deposition treatments: (i) the neutral "lm was voltage cycled "rst to red ("ve cycles at $ 1.2 V), (ii) the neutral "lm was voltage cycled to blue ("ve cycles at $ 1.2 V) and (iii) the neutral "lm was held at 393 K for 1 h to anneal the material. The combined specular and di!used re#ectance were measured with the sample in the 83 holder, followed by the di!use measurements only in the perpendicular holder. From these results, it is possible to establish the surface roughness from the ratio of specular to di!use re#ection.
3. Results and discussions Fig. 1(a) displays four sets of absorption spectra for heavy fraction rare-earth bisphthalocyanine [HF(pc) (pc*)] sublimed "lms within the visible optical frequency range, one set for a particular type of treatment. As reported previously [9], the Q-band peak shown in Curve A of Fig. 1(a) for the [HF(pc) (pc*)] untreated sample is not sharp, probably due to the interaction of molecules in the solid phase. The di!erent post-deposition treatments produce di!erent e!ects on the intensity of absorption but, irrespective of the post-deposition treatment, the intensity of the Q-band absorption always remains higher than the X-band absorption. Upon annealing, the intensity of absorption increases by approximately a factor of two and, at the same time, the peak becomes narrower. On voltage cycling to blue, there is a noticeable broadening of the Q band with a shoulder and the peak intensity is increased by a factor of about 1.5 over that for the untreated sample. For the voltage cycling to red, the opposite e!ect is observed in terms of diminished Q-band peak. Annealing also increases the intensity of the X-band; cycling to blue causes the X band to almost disappear whereas cycling to red causes a much wider X band (2.3 to 2.7 eV). It may be inferred that the e!ect of voltage cycling on this
A.K. Ray et al. / Vacuum 61 (2001) 19}27
21
Fig. 1. Three groups of graphs showing the visible absorption spectra for sublimed "lms of (a) heavy fraction rare-earth [HF(pc) (pc*)], (b) gadolinium [Gd(pc) (pc*)] and (c) thulium [Tm(pc) (pc*)] phthalocyanine compounds. Measurements are performed when the "lms are (i) untreated, (ii) annealed at 393 K for 1 h, (iii) voltage-cycled "ve times to blue and (iv) voltage-cycled "ve times to red.
material is to cause a change in the oxidation states present in the return to `neutrala material. Fig. 1(b) gives the similar plots of optical absorption spectra for thulium phthalocyanine compounds [Tm(pc) (pc*)]. There are two interesting features in the spectrum: (i) a shoulder to the Q band at 2.1 eV and (ii) a pronounced absorption edge at 1.7 eV. The blue shoulder is attributed to the Q (1!0) vibronic band, generally more clearly resolved at lower temperatures. Like the other sample compounds, the tail towards red is due to smearing of many hot bands. Annealing of the sample does not appear to cause any noticeable shift of the Q-band peak along the energy scale, but the overall absorption is decreased. For 1.2 V volt-
age cycling to blue "ve times, the overall absorption increases and there is a very slight blue shift in the Q band. The intensity of absorption decreases slightly on voltage cycling to red and there is a pronounced e!ect on the shape of the shoulder at about 2.1 eV. For the sample which has been voltage-cycled to red, a signi"cant increase in absorption between the peak at 2.6 eV and the Q band is observed. As shown in Curve A of Fig. 1(c), the Q and B bands in the spectra for the untreated gadolinium phthalocyanine [Gd(pc) (pc*)] "lm are of approximately equal strength, in addition to the X band in the energy window centred at around 2.6 eV. The 5;1.2 V voltage cycling to blue causes a decrease
22
A.K. Ray et al. / Vacuum 61 (2001) 19}27
in the absorption at the Q and X bands but the Q band is now split, with one peak at an increased energy. Cycling at the same voltage to red causes a decrease in the peak absorption of the Q band; the X band is less signi"cantly decreased than for the voltage-cycled to blue. The annealing thus has a pronounced e!ect on the [Gd(pc) (pc*)] spectra: (i) an increase in the energy at which the Q band occurs in the spectrum, (ii) a decrease in the peak absorption and (iii) the broadening of the Q band. This suggests that the e!ect of the annealing process has been to promote a change in the structure of the "lm, possibly, a phase change in the "lm material. This is similar to previously published data for the transformation of -FePc to the form [10]. Table 1 summarises the main points of our observations which include the positions in the energy domain for the B, X and Q bands, together with their peak intensity absorbance values. The relative heights of the Q and X band peaks are modi"ed as a result of the post-deposition treatments. Fig. 2 shows the spectrum of optical re#ectivity R for the same batch of samples. The sharp changes
in re#ectance correspond to an electronic transition. It is observed on close examination (vide the insets) that the Soret band a Pe and Q-band a Pe transitions take place at 2.8 and 1.9 eV, respectively. The absorption coe$cient which is de"ned as the reciprocal of a distance over which the energy in the wave falls o! to exp(!1) can be determined in terms of I and I as [11] R I R "(1!R) exp(!t)/1!R exp(!2t), (1) I where t is the thickness of the "lm. The extinction coe$cient, , is, on other hand, related to the absorption coe$cient in the form " . 4
(2)
Using t"50 nm in Eq. (1), values of the absorption coe$cient are calculated for all three untreated samples from values of [I /I ] obtained from Fig. 1. R With a knowledge of the values of the absorption coe$cient , the value of the extinction coe$cient,
Table 1 Absorption peak heights and energies at which they occur for the three materials Compounds
B band
X band
Q band
and treatment
Energy (eV)
Height
Energy (eV)
Height
Energy (eV)
Height
HF ( pc) ( pc*) Untreated Annealed Voltage cycled (Red) Voltage cycled (Blue)
3.9 3.8 3.8 3.9
0.55 0.71 0.50 0.51
2.4 2.6 2.6 2.5
0.14 0.21 0.14 0.16
1.9 1.9 1.9 1.9
0.29 0.51 0.40 0.27
Tm( pc) ( pc*) Untreated Annealed Voltage cycled (Red) Voltage cycled (Blue)
3.6 3.8 3.7 3.9
0.67 0.83 0.88 0.92
2.7 2.6 2.7 2.6
0.41 0.32 0.38 0.35
1.9 1.9 1.9 1.9
0.84 0.78 0.99 0.75
Gd( pc) ( pc*) Untreated Annealed Voltage cycled (Red) Voltage cycled (Blue)
3.3 3.3 3.3 3.3
0.30 0.25 0.18 0.27
2.4 2.5 2.4 2.4
0.21 0.17 0.16 0.19
1.9 1.9 1.9 1.9
0.28 0.26 0.21 0.28
A.K. Ray et al. / Vacuum 61 (2001) 19}27
23
Similar behaviour is observed for monophthalocyanines [12]. For an absorbing medium such as these metal bisphthalocyanine "lms, the real part of refractive index may be found from the re#ectivity R using the following relation: (!I)# R" . (#I)#
(3)
Eq. (3) which is valid for monochromatic radiation at normal incidence on the "lm/air interface [13] is reduced to a set of quadratic equations in , when values of R and are substituted. The physically meaningful root of these equations gives the value of , the real part of the complex refractive index, or propagation constant. The dispersion behaviour of is shown for all three samples in Fig. 4. takes a value between 2 and 3 throughout the energy range for [HF(pc) (pc*)] and [Tm(pc) (pc*)], but for [Gd(pc) (pc*)], it is a little higher, approaching 4.1 at the maximum. The observed dispersive behaviour does not follow any polynomial of low order for the function (h) where h is the photon energy, and therefore the Cauchy relation is not expected to hold over the entire photon energy range. This anomalous behaviour is in keeping with the Debye theory [14]. The real and imaginary parts , of the dielectric constant * (" #j ) are related to and through the following equations: "! and "2.
Fig. 2. A set of three graphs showing the dependence of re#ectivity R on incident wavelength () for untreated sublimed "lms of (a) heavy fraction rare-earth [HF(pc) (pc*)], (b) gadolinium [Gd(pc) (pc*)] and (c) thulium [Tm(pc) (pc*)] phthalocyanine compounds.
(which is the imaginary part of the complex refractive index N"#j) may be evaluated from Eq. (2). Fig. 3 shows the variation of with respect to the incident photon energy. The variations of extinction coe$cient follow a similar pattern to those for absorbance, as expected from the relationship between the two parameters [vide Eq. (2)].
(4)
Resulting dependences of , on photon energy h are presented for untreated samples in Fig. 5. For all "lms, the variation of the real part broadly follows the same pattern as the imaginary part, both of which resemble the envelope of the absorption spectrum. In general, values of the real part are appreciably higher. The obvious peaks are seen at an energy between 1.5 and 2.2 eV, depending upon the type of material. The (h) curves are believed to be caused by the electron oscillation with resonant frequency corresponding to the N peak. The electron density N can be determined using classical oscillation theory [15]: m ( ) N" N . e
(5)
24
A.K. Ray et al. / Vacuum 61 (2001) 19}27
Fig. 3. A set of three curves showing the variations of extinction coe$cient (imaginary part of refractive index) for untreated sublimed "lms of (a) heavy fraction rare-earth [HF(pc) (pc*)], (b) gadolinium [Gd(pc) (pc*)] and (c) thulium [Tm(pc) (pc*)] phthalocyanine compounds with respect to incident photon energy h within the visible range.
Fig. 4. A set of three graphs showing the variations of optical propagation constant n (real part of refractive index) for untreated sublimed "lms of (a) heavy fraction rare-earth [HF(pc) (pc*)], (b) gadolinium [Gd(pc) (pc*)] and (c) thulium [Tm(pc) (pc*)] phthalocyanine compounds with respect to incident photon energy h within the visible range.
It is found that values of N lie in the order of magnitude of 10 m\. Calculations have been performed for the samples which have undergone post-deposition treat-
ments and the results are summarised in Table 2. The e!ects of the post-deposition treatments are apparent in the values of the propagation constant in all cases. The energy shifts in the peaks of the
A.K. Ray et al. / Vacuum 61 (2001) 19}27
25
Fig. 5. A set of three graphs showing the dependences of real ( ) and imaginary ( ) parts of the dielectric constant (*) for untreated sublimed "lms of (a) heavy fraction rare-earth [HF(pc) (pc*)], (b) gadolinium [Gd(pc) (pc*)] and (c) thulium [Tm(pc) (pc*)] phthalocyanine compounds on incident photon energy h within the visible range. Circles and squares symbols are used for and , respectively.
propagation constants observed in the calculated spectra after treatments are marked for the X band; additionally the intensities of the two peaks for [HF(pc) (pc*)] are approximately reversed, indicating the phase change induced by the annealing process. The upper and lower limits of the propagation constant, however, remain reasonably constant throughout the treatments. The e!ects broadly follow the absorption spectra. The modi"cation of energy values at which the peaks occur in the , and spectra are generally more pronounced than those visible in the absorbance data. In particular, the e!ects of annealing upon the X or characteristic band centre are very pronounced for [HF(pc) (pc*)] compounds. For the other three treatments, the
major peak is the original characteristic (X) band location for the untreated material. The process of annealing causes a diminution of the X band at this energy location to a great extent, and an increase of the intensity on the other part of the split peak corresponding to the phase. Thus, the calculation of fundamental optical constants provides a less ambiguous assessment of the structural and chemical changes occurring in the "lms during treatment. Surface roughness of the "lms can be given in terms of the exponential dependence of irregularity height on incident wavelength :
4 RJR exp ! ,
(6)
26
A.K. Ray et al. / Vacuum 61 (2001) 19}27
Table 2 Peak values of real part ()of refractive index, real ( ) and imaginary ( ) parts of dielectric constant (*) with their corresponding incident energy
Compound and treatment
Peak
Energy (eV)
Peak
Energy (eV)
Peak
Energy (eV)
N (;10 m\)
[HF( pc) ( pc*)] Untreated Annealed Voltage cycled (Red) Voltage cycled (Blue)
2. 7 2.7 1.9 2.8
2.0 2.1 2.8 1.9
7.2 9.5 7.9 7.6
1.9 1.5 1.8 1.9
1.7 3.3 1.8 2.4
1.9 1.5 1.8 1.9
1.7 1.1 1.4 1.0
[Gd( pc) ( pc*)] Untreated Annealed Voltage cycled (Red) Voltage cycled (Blue)
4.0 3.0 2.7 3.0
1.6 2 1.9 1.9
7.7 9.0 8.0 8.0
2.1 1.9 1.9 1.9
1.4 1.8 2.3 2.5
2.1 1.9 1.9 1.9
1.9 1.0 1.6 1.6
[Tm( pc) ( pc*)] Untreated Annealed Voltage cycled (Red) Voltage cycled (Blue)
2.8 2.8 2.7 2.8
2.2 2.2 2.0 2.2
8 7.9 5.7 7.9
2.2 2.2 2.2 2.2
2.1 1.3 1.2 1.1
2.2 2.2 2.2 2.2
1.3 0.5 0.6 0.6
N is the concentration of electrons at resonant frequency of oscillation.
Fig. 6. A set of three graphs showing the wavelength dispersion of logarithmic normalised re#ectance for untreated sublimed "lms of (a) heavy fraction rare-earth HF(pc) (pc*)], (b) gadolinium [Gd(pc) (pc*)] and (c) thulium [Tm(pc) (pc*)] phthalocyanine compounds.
where R is re#ection from a smooth surface. This formula is applicable for our case since the angle of acceptance is small [16].
As shown in Fig. 6, the plots of ln(R/R ) as a function of the inverse of the square of incident wavelength are linear allowing a value of to be
A.K. Ray et al. / Vacuum 61 (2001) 19}27
determined from the slope. Using data from the graphs, values of 18.5 nm for HF(pc) (pc*), 26 nm for Gd(pc) (pc*) and 12.1 nm for Tm(pc) (pc*) have been obtained. These values are very high relative to the "lm thickness, although a result of this magnitude might be expected for a thin, vacuum-evaporated layer which, in these cases, is visibly inhomogeneous. It is also hypothesised that this may be largely a factor consequence of the roughness of the relatively thick ITO "lm between the pc and the glass substrate, the thin phthalocyanine "lm following the contours of this.
4. Concluding remarks The e!ects of the various post-deposition treatments upon the optical frequency dielectric properties of these three materials have been examined. For phthalocyanines, electronic transitions occur notably in the 1.5}2.0 eV visible region (the Q band) and 3}4 eV (the B or Soret band), both being due to }* transitions. The Q band is highly localised on the phthalocyanine ring, and is reported for many phthalocyanines to be highly susceptible to a number of macrostructural factors, particularly to the geometry of nearest-neighbour molecules in the structure [17]. This in the light of the results appears to be true for the three compounds investigated and signals that the phase changes are promoted by annealing. This also suggests that no structural damage to the phthalocyanine rings has occurred during the treatment. The e!ect of annealing upon the Q band intensity is di!erent for [HF(pc) (pc*)] compared with the other two compounds. In the case of [HF(pc) (pc*)] there is an increase in absorption but the other two compounds exhibit a decrease in absorption. This e!ect must be related to di!erences in the structure of [HF(pc) (pc*)] compared to the other materials; [(Lu(pc) (pc*)] is reported to form liquid crystals, so the annealing process causes closer ordering in this material than is noticeable in the raw state. Annealing [HF(pc) (pc*)] shows a narrowing of the Q band and an increase in its intensity; this may be due to a reordering of molecules in the structure, leading to a more highly ordered phase. The reverse e!ect on the other two compounds
27
indicates a more random ordering illustrated by the peak broadening. The reversible redox processes resulting from voltage cycling are reduction or voltage cycling to blue giving the derivative [M(pc) ]\, and oxida tion or voltage cycling to red giving [M(pc*) ]>. For N, the number of electrons per unit volume participating in oscillation at peak frequency, values in the order of 10 m\ are obtained for all materials, irrespective of treatment. Acknowledgements The authors are grateful to Dr. J.R. Travis for his encouragement and fruitful discussions. One of us (JE) thanks the Science and Engineering Research Council, UK for his studentship. References [1] Frampton CS, O'Connor JM, Peterson J, Silver J. Displays Tech Appl 1988;9:174. [2] Nicholson MM. In: Lezno! CC, Lever ABP, editors. Electrochromism and display devices of Phthalocyanines, Properties and Applications, vol. 3. NewYork: VCH Publishers, 1993, p. 77 [Chapter 2]. [3] Petty M, Lovett DR, Townsend P, O'conor JM, Silver J. J Appl Phys 1989;22:1604. [4] Petty M, Lovett DR, O'Connor JM, Silver J. Thin Solid Films 1989;79:387. [5] Petty M, Lovett DR, Miller JR, Silver J. J Mater Chem 1991;1:971. [6] Song SA, O'Connor JM, Barber DJ, Silver J. J Crystal Growth 1988;88:477. [7] Silver J, New Scientist, 30th September 48, (1989). [8] Souto J, Aroca R, Desaja JA. J Raman Spectroscopy 1991;22:349. [9] Exley J, Ray AK, Ahmet MT, Silver J. J Mater Sci Mater in Electron 1991;5:59. [10] Silver J, Lukes P, Houlton A, Howe S, Hey P, Ahmet MT. J Mater Chem 1992;2:849. [11] Swan R, Ray AK, Hogarth CA. Phys Stat Sol(a) 1991;127:555. [12] Collins RA, Krier A, Abass AK. Thin Solid Films 1993;229:113. [13] Moss TS, Burrell GJ, Ellis B. Semiconductor opto-electronics. London: Butterworth, 1973. p. 10. [14] Elliott SR. Advances in Physics 1987;36(2):138. [15] Cook BE, Spear WE. J Phys Chem Solids 1969;30:1125. [16] Gupta P, Maiti B, Maity AB, Chaudhuri S, Pal AK. Thin Solid Films 1995;260:75. [17] Ho ZZ, Ju CY, Hetherington III WM. J Appl Phys 1987;62:716.