Structural and morphological characterisation of ruthenium phthalocyanine films by energy dispersive X-ray diffraction and atomic force microscopy

Ž .Thin Solid Films 382 2001 74�80

Structural and morphological characterisation of rutheniumphthalocyanine films by energy dispersive X-ray diffraction and

atomic force microscopy

R. Caminitia, A. Capobianchib, P. Marovinoa, A.M. Paolettib,G. Padelettib, G. Pennesib, G. Rossib,�

a Chemistry Department, Instituto Nazionale per la Fisica della Materia, Uni�ersita di Roma ‘La Sapienza’, P. le Aldo Moro 5, Rome, Italy`b ( )Instituto di Chimica dei Materiali C.N.R. , Area della Ricerca di Roma, C.P. 10, 00016 Monterotondo Stazione, Italy

Received 15 February 2000; received in revised form 25 July 2000; accepted 22 September 2000

Abstract

Ž .Ruthenium phthalocyanine RuPc , deposited as a film by vacuum sublimation onto a variety of substrates, was structurally2Ž . Ž .and morphologically investigated by energy dispersive X-ray diffraction EDXD and atomic force microscopy AFM techniques.

The attempt to apply the EDXD technique to amorphous films was successful and the reported results show the preservation ofthe Ru-Ru bond between two phthalocyanines units while transferring from the bulk to the film. The unidimensionalarrangement of the dimeric phthalocyanine units in the film was observed to a larger extent than in the bulk; indeed 10 dimers

Ž .were superimposed along the stacking direction parallel to the Ru�Ru bonds , while six units were found in the bulk material,the higher order being also supported by conductivity measurements. The amorphous nature of the film, anomalous in respect ofthe generally microcrystalline films of other metal phthalocyanines, was not changed by annealing the films at 170�200�C formore than 10 h. Reproducible and uniform deposition is identified by both EDXD and AFM data. � 2001 Elsevier Science B.V.All rights reserved.

Keywords: Phthalocyanine; Film; Structural characterisation; X-Ray diffraction

1. Introduction

Molecular materials based on phthalocyanine unitshave been extensively studied due to the wide range ofparticular physico-chemical properties of their metalcomplexes. In the form of thin films, they are of great

� �interest for their electrical and optical properties 1 .Vacuum sublimation is a widely used technique forpreparing phthalocyanine films from most of the un-substituted phthalocyanine compounds. With this tech-nique, the physical and chemical properties of thesource materials are preserved, and possible decompo-sition during deposition is minimised. The morphology

� Corresponding author.

of the films, which depends on the sublimation condi-tions and strongly affects the performance of the mate-rial, can range from amorphous to highly crystalline.The deposition operation is a critical phase in theconstruction of a device. Despite extensive work car-ried out on the film characterisation and sensing be-

� �haviour of numerous metal phthalocyanines 2 , ruthe-nium phthalocyanine, which appears to be a promisingcompound due to its chemical�physical behaviour, re-mains largely unexplored in these areas. In a previous

� �study 3 , we showed the deposited ruthenium phthalo-Ž .cyanine acting as a semiconducting gas sensor NO ,x

with a significant increase in electrical conductivityvalues during the gas interaction process. In the papermentioned, the structural and morphological character-isation of the dimeric ruthenium phthalocyanine

0040-6090�01�$ - see front matter � 2001 Elsevier Science B.V. All rights reserved.Ž .PII: S 0 0 4 0 - 6 0 9 0 0 0 0 1 7 0 1 - 6

( )R. Caminiti et al. � Thin Solid Films 382 2001 74�80 75

Ž . Ž 2�.RuPc Pc�phthalocyaninato dianion C H N ,2 32 16 8processed by vacuum deposition, was considered.

� �Although, as demonstrated 4 , the evaporationshould not modify the structure of ruthenium phthalo-cyanine, it appeared essential, at first, to determine thechemical structure of the deposited material. This wasin order to understand any possible modification occur-ring when the film material is in contact with appropri-ate gases. In fact, the role of the metal and�or phtha-locyanine ring in the charge-transfer complex forma-tion is not clearly understood for this specific case.Since the amorphous nature of the sublimed materialmade it difficult to carry out the structural analysis bythe current X-ray techniques utilised for polycrystallinefilms, we attempted to examine structural features ofthe film by the energy dispersive X-ray diffractionŽ .EDXD technique, widely used for amorphous systems� �4�8 . The sample also underwent thermal treatment tocheck for any changes in microscopic structure, and

Ž .EDXD spectra, atomic force microscopy AFM imagesand electrical conductivity measurements were recordedbefore and after thermal treatment. The results ob-tained confirmed the organisation of ruthenium phtha-locyanine in a dimeric form with a double bond betweenthe ruthenium atoms, as was established for the bulk

� �phase 4 , with a similar columnar arrangement.The EDXD technique, used for the first time to

obtain structural information on amorphous films,proved to be an efficient tool for describing structuralfeatures. At the same time, vacuum evaporation provedto be a technique suitable to produce films with auniform material distribution. The most suitable start-ing material and the most appropriate substrates toobtain good quality films will also be discussed.

2. Experimental details

The starting material, bis-pyridine ruthenium phtha-� �locyanine, was synthesised as described by others 9 .

Films were deposited on different substrates by evap-oration from a molybdenum boat on an Edwards A 306coater, at 300�320�C and 10�6 torr. The substratesŽMylar, commercial adhesive tape, Platinum foil, Sili-

.con wafer, interdigital electrode , previously cleaned,were at room temperature. Film thickness, ranging

˚from 100 to 50 000 A, was monitored by an EdwardsŽ .FTM5 film thickness monitor density 1.6 . The evap-

oration apparatus allowed up to four samples to bedeposited simultaneously. To check any possible de-composition, the material deposited as film was solu-bilised in pyridine or sulfuric acid and qualitative andquantitative controls were carried out by UV-visiblespectroscopy. Transmission or total diffuse reflectanceUV-visible spectra were recorded on a Cary 5 spec-trophotometer.

2.1. X-Ray data

The EDXD technique supplies much of the informa-tion on amorphous or quasi-amorphous material and,since its discovery, it has been successfully employed infurnishing structural information for bulk materials.Recently it has been used for structural investigation

Ž . � � � �on RuPc 4 and rutenoxane 10,11 derivative2powders, both amorphous materials, obtaining exhaus-tive data for describing the structural feature of thecompounds examined. Following the same procedureand supported by the results obtained for both samples,we decided to undertake a structural study on films ofŽ .RuPc , an easily processible compound. Diffraction2,data were collected with a non-commercial diffrac-

� �tometer 5,7 equipped with an X-ray generatorŽ .water-cooled, W target , and a germanium solid-statedetector connected with a multichannel analyser. The

Ž .static structure function S q , where q indicates theŽ .scattering parameter i.e. q� 4��� sin� or the equiva-

Ž . Ž .lent q� 2�hc E sin� , where, in turn, 2� is the scat-tering angle, � the radiation wavelength, E the radia-tion energy, and c and h have their usual meanings,was used to extract all structural information. Since qdepends on both E and �, to execute a scan in the

Ž .reciprocal q space obtaining the diffraction pattern , itŽ .is possible at constant E to perform either an angular

Ž .or at constant � an energetic scan by means of apolychromatic X-ray beam and an energy-dispersive

Ž .solid sate detector SSD device. In the present case weused the second procedure, and the scattering intensitywas measured by the transmission method at the scat-

Ž .tering angles � of 0.5; 1.0; 1.5; 2; 3.0; 3.5; 5.0; 8.0;10.5; 15.5; 21.0; and 26.0�. The reflection method was

Ž .applied at properly chosen angles 1.8, 6.8� . In order toavoid mistakes in the data analysis for the energy-dis-persive method, the scattering angles were set so closelythat the two ends of the intensity curves obtained atadjacent angles could widely overlap each other. The

Ž .following working conditions were used: a alimenta-tion, high voltage 45 kV, current 35 mA, total power

Ž .1.575 kW; and b energy interval, 22.4�42.4 keV. Thecollected data were corrected for escape peak suppres-sion, normalisation to the incident radiation intensity,division by X-ray absorption and polarisation coeffi-cients, and elimination of inelastic scattering contribu-

Ž .tions. Then, from the observed intensity I E; � , theŽ .static structure function S q can be obtained as q �

Ž . Ž . � � Ž .S q �M q 12 . The Fourier transformation of the S qfunction gives the radial distribution function

qmax2 �1Ž . Ž . Ž . Ž .D r �4� r � �2 r� qS q M q sin rq dqH00

Ž . Ž . 2and it is represented as Diff r �D r �4� r � . Both0Ž . Ž .functions S q and Diff r were then compared with

( )R. Caminiti et al. � Thin Solid Films 382 2001 74�8076

Ž . Ž . Ž . Ž .Fig. 1. Observed qS q M q values for: a bulk; and b film, vs.Ž .Ž .q� 4��� sin� .

Ž .those obtained for the bulk material Figs. 1 and 2 . Adetailed and more accurate description of the appara-

� �tus and technique is given in previous papers 5,7,8 .Various substrates have been examined in order to

select those which do not interfere in the range ofreciprocal space examined; the peak intensities as afunction of thickness were monitored. To this aim,

˚films with a thickness ranging from 100 to 21 000 A

Ž . Ž . 2 Ž .Fig. 2. Experimental function Diff r �D r �4� r � for: a bulk;0Ž .and b film.

Ž .Fig. 3. Series of diffraction patterns for RuPc films with increasing2thickness; channels reported in abscissa correspond to a � E of thetotal energy interval utilised.

were prepared, and the respective curves of intensityvs. thickness are reported in Fig. 3.

2.2. AFM experimental

AFM analysis was performed in air with a Nanos-cope IIIa Digital Instruments microscope equipped withan optical deflection system in combination with siliconcantilevers and tips working in tapping mode. With thistechnique the cantilever is oscillated near its frequencyas it is scanned over the sample surface. The frequencyranges from 250 to 390 kHz. In tapping mode thecantilever is excited into resonance oscillation with apiezoelectric driver. The oscillation amplitude is usedas a feedback signal to measure topographic variationsin the samples. In phase imaging, the phase lag of thecantilever oscillation relative to the signal sent to thepiezo driver is monitored and recorded at the sametime.

The phase lag is very sensitive to variation in theproperties of the materials. Phase imaging can also actas a real-time contrast enhancement technique. Asphase imaging highlights edges and is not affected bylarge-scale differences, it provides for clearer observa-tion of fine features, such as grain edges, which can beobscured by rough topography. Topographic imageswere recorded over scanned areas ranging from 500�500 nm2 up to 10�10 m2, each with a resolution of256�256 data points.

The surfaces were characterised by means of thepick-valley excursion registered in the scanned areaŽ . Ž .Z , by means of the mean roughness Ra of thersurface relative to a reference centre plane: R �a

( )R. Caminiti et al. � Thin Solid Films 382 2001 74�80 77

� Ž . � Ž .�1�LxLyHH f x, y d xd y where f x, y is the surfacerelative to the centre plane and Lx and Ly are thedimension of the surface, and by means of the standarddeviation of Z values within the given area: R �qŽ Ž .2 .1�2Ý Z �Z �N where Z is the average of thei ave aveZ values within the given area, Z is the current Zivalue, and N is the number of points within the givenarea.

2.3. Electrical measurements

The material was deposited on interdigital electrodescontaining 40 interdigital pairs of gold fingers on sili-con�silicon oxide and connected to a computer-aidedapparatus with a Keithely 236 Source measuring unit.Current measurements as a function of time wererecorded at room temperature. The film surfaces were

Ž .cleaned by a flow of inert gas N .2

3. Results and discussion

3.1. Film preparation

Ž .PcRu as a bulk material can be obtained by ther-2Žmal treatment of PcRuL L�pyridine, dimethyil-2

. Ž �2 .sulphoxide, or butylamine under vacuum 10 mmHg� �and at a relatively high temperature of 320�330�C 9 .

During the vacuum pyrolysis the two axial ligands wereremoved and a rearrangement of RuPc monomericunits occurred as the red-brown polycrystalline samplesturned into a dark-green amorphous material. Thematerial was fully characterised by X-ray structural

Ž .investigation large-angle X-ray scattering , spectro-� �scopy and magnetic studies 4,11 . The presence of a

˚Ž .double bond 2.40 A between two ruthenium atomswas found to give rise to a dimeric unit staked in aunidimensional array. In the dimer formed the twophthalocyanines rings are not perfectly planar, as themacrocycles assume a staggered position with a rota-tion angle of ca. 45�. This was the first example of anamorphous dimeric species in the field of metal phtha-locyanines. A similar arrangement was found for a

� �ruthenium octaethylporphyrin complex 13 . Since thesolid dimer is stable and can be sublimed, it was firstused as the starting material for obtaining films byvacuum evaporation, but several experiments led us torealise that this procedure frequently yields an oxygen-contaminated material. Evidence of this fact was foundby dissolving the sublimed film in sulfuric acid. Therespective UV-visible spectra showed a small amountof the species with a broad absorption at 680 nm, thusindicating the presence of an associated oxygen-con-

� �taining product 10 , probably originating from a negli-gible amount of oxygen present in the vacuum cham-

Ž .ber. When RuPc py was used as the source material,2reproducible pure samples were obtained. In the evap-

oration process the pyridine was eliminated first; then,at ca. 400�C, evaporation of the compound occurred.

3.2. Structural features of e�aporated films

˚ŽThe X-ray transmission spectra of the film 50 000 A.on Mylar substrate were typical of amorphous material

� �and similar to that registered for the bulk 4 . Thestructural investigation was carried out by comparingthe bulk and film data; the structure functions

Ž . Ž .qS q M q for the film and the bulk are reported inFig. 1a,b. This figure shows that the functions are

˚�1similar in the range q�4 A , as some differences canbe noted for lower q values. Considering in detail the

Ž .peaks of S q for the film, it was noted that they are˚�1exactly located at q�0.45 and 1.875 A . They present

a higher intensity compared to the bulk, the first peakintensity being even tripled. These values correspond,

˚in real space, to distances of 15.9 and 3.36 A, respec-tively. Following the model used for the bulk, the

˚15.9-A distance is ascribed to the distance between two˚closely staked columns, while 3.36 A corresponds to the

average distance between the molecular planes ofruthenium phthalocyanine. These considerations derive

Ž .from the experimental radial function Diff r in Fig. 2,where the peak distribution is reported, generated bythe interaction of atoms at different distances, referredto the real and not reciprocal space. It can be notedthat the oscillations of the two curves have the samephase, while they are more intense, narrow and well-

˚ Ž .defined up to 50 A in the curve b . This indicates thepresence of a greater order than in the bulk. Indeed,the number of the superimposed dimeric rutheniumphthalocyanine molecules in the film should be higherthan in the bulk. Further support for the presence of a

˚greater order is also given by the peak at 15.9 Apresent in the film curve, while only a shoulder wasobserved in the powder diffraction spectra. The struc-

Ž .ture function S q and the radial distribution functionŽ .Diff r were compared with the theoretical structure

function and its corresponding Fourier transformation,obtained by using the model built for fitting the bulkexperimental data. The same model, applied to repro-duce experimental film data, allowed us to affirm thatthe ruthenium phthalocyanine deposited as a film con-

Ž .sists of dimeric molecules, i.e. PcRu with ten molec-2ular units superimposed along the staking directionŽ .parallel to the Ru�Ru bonds and with chains ar-

˚ranged side by side at a distance of 15.9 A. Intramolec-ular distances are the same as reported for the bulkmaterial.

The structural study also included the growth analy-sis of the principal peak present in spectra of the film

˚Ž .q�0.45 and 1.875 A and of the first and secondŽ .florescence lines of the Ru atom K , K , located at� �

19.405 and 21.375 keV, respectively, as a function of

( )R. Caminiti et al. � Thin Solid Films 382 2001 74�8078

thickness. To this aim, 20 films with thickness ranging˚from 100 to 20 000 A were fabricated and experimental

data were recorded by the reflection method, which isconsidered more appropriate than the transmissionmethod to investigate film surfaces. All samples wereexamined by putting them through the X-ray beams,with each one manually centred, and the respectivespectra registered. The results obtained for three peaks

˚Ž .q�0.45 A, K , K of five samples are shown in Fig.� �

3; an increase in intensity, proportional to the thicknessfor all peaks considered is observed. Furthermore, bymoving from one point to another on the surface, thesame peak intensity for ruthenium florescence lineswas recorded for each sample. This last result is evi-dence of equal amounts of ruthenium atoms at differ-ent surface points and shows good uniformity of thedeposited material for all thickness values considered.As discussed above, the reflection method is useful forstudying surface modifications, but the substrate con-

Žtribution must be minimised it was impossible to to-tally eliminate the substrate absorption, owing to un-

.reachable grazing angles on the surface . Substratesperfectly transparent in the range considered wererequired, unlike the transmission method, where thesubstrate contributions are subtracted. In our case, twosubstrates were used: Mylar sheets for studying the

˚�1peak located at 0.45 A , and commercial adhesivetape. This latter is the only one we found without peaks

˚ Žin the neighbourhood of the 1.875-A peak results not.shown .

Since it is known that slight changes in evaporationconditions cause variation in the degree of crystallinityor state of aggregation and, therefore, considerablechanges in electrical properties, exact control of theevaporation conditions is a very important requirementto obtain reproducible layers, as mentioned above.Thus, an attempt was made to change the evaporationgeometry by varying the distance between the substrateand the source material.

In addition, the influence of temperature was studiedby post-annealing the amorphous films at 170�200�C.

˚Ž .To this aim, films 1000 A deposited on Pt and Si werefabricated and heated to 200�C in an oxygen-free nitro-gen environment for over 20 h. X-Ray spectra wererecorded before and after heating, showing that boththe position and intensity of all existing peaks remainidentical. The amorphous nature of the samples wasmaintained and no crystalline transitions were observed.This result is somewhat surprising, since it is knownthat phthalocyanine films commonly yield, upon ther-mal treatment, different polymorphic forms, also

� �changing the arrangement of their macro-order 14 . Inaddition, these materials have been rarely found in theamorphous phase when sublimed on substrates held at

� �room temperature 15 .

A convincing explanation for this unusual behaviouris not easily found; in our opinion, the amorphousnature of the source material and the fact that theevaporation process does not influence the moleculararrangement on moving from the bulk to the film, maypartly account for the experimental findings. Up to nowthe amorphous state of the bulk has not been ex-plained, and it is well known that the similarly synthe-

� �sised porphyrine analogue is not amorphous 13 . Wecan only ascribe this behaviour to the more favourablethermodynamics of the amorphous state.

Our structural findings would appear to disagree� �with those found by Gopel et al. 16 on microscopic¨

modifications of Ru phthalocyanine films. They re-ported strong intermolecular interactions between Ru

˚Žphthalocyanine molecules in thin films 70�400-A.growth in ultra-high vacuum conditions as deduced by

Ž .X-ray photoelectron spectroscopy XPS data and sug-gested the formation of dimers associated with a loss ofone central ruthenium atom. In fact, the EDXD tech-nique did not allow an investigation of such a thin film,but, as the structural data recorded in the transmissionmethod with subtraction of substrate contribution donot show any particular features connected with demet-allation of the molecule, we reasonably think thatdimers of ruthenium phthalocyanines with a Ru-Rudouble bond are also formed in the first monolayers.

3.3. AFM characterisation

In order to give a further insight into the film charac-terisation, AFM investigations were carried out on the

Ž .two different substrates Si and Pt , as well as on thefilm surface before and after treatment at 200�C. Fig.

Ž .4a,b shows the images obtained on RuPc �Si before2and after thermal treatment, respectively. In Fig. 4a,the surface is covered by particles that are elliptic in

Ž . Ž .shape, with a size ranging from 55�35 to 142�113nm. Annealing produces a more ordered surface struc-ture; in addition, the particles covering the whole sur-face are now homogeneous and increased in size, rang-

Ž . Ž .ing from 234�430 to 410�3519 nm. This processis effective, as it also produces a reduction in theroughness parameter values, as shown in Table 1. This

Ž .trend is also confirmed in the RuPc �Pt system. The2considerable differences in parameter values betweenthe two films are due to the different roughness andhomogeneity of the two substrates. The result obtainedfor this system, with the phase imaging on the annealedfilm, allows some interesting considerations. The film

Ž .texture in the height image Fig. 5a , related to thetopography, shows a granular structure with a verydifferent grain size. It is not possible to distinguish if acoalescence of small particles took place or not. On the

Ž .contrary, phase imaging Fig. 5b reveals that the cover-

( )R. Caminiti et al. � Thin Solid Films 382 2001 74�80 79

age is homogeneous, and is constituted of particlesŽdistributed all over the sample surface size ranging

.from 322�234 to 156�127 nm . The phase imagingpermits highlighting of the grain edges and is notaffected by large-scale differences on the sample sur-face. These observations indicate that a homogeneousdeposition is obtained by working with the illustratedevaporation conditions, in agreement with X-ray inves-tigation results.

3.4. Electrical measurement

Electrical conductivity measurements recorded onfilms as a function of time at room temperature gave avalue of �1�10�4 ��1 cm�1, one order of magni-

Ž .tude higher than found for the RuPc pressed powder2� �4 . This result is consistent with structural evidence,which suggests a longer chain of columnar-stakedŽ . Ž .RuPc dimers 10 instead of six dimers found in the2

Ž .film. The thermal treatment 200�C of samples pro-Žduces a slight increase in conductivity value �5�

Table 12AFM roughness parameters for 5�5 m of scanned areas of

Ž .RuPc films and substrates2

R R Za q rŽ . Ž . Ž .nm nm nm

Si substrate 0.73 1.6 73Ž .RuPc on Si 2.7 5.7 1032

Ž .Annealed RuPc on Si 1.13 1.9 312Pt substrate 12.7 16 209Ž .RuPc on Pt 21 27 1842

Ž .Annealed RuPc on Pt 6.8 10.3 1402

�4 �1 �1.10 � cm , possibly due to the larger particlessize obtained, as shown by AFM images.

4. Conclusions

Ž .Films of dimeric RuPc can be obtained by vacuum2evaporation if great care is devoted to the workingconditions. The presence of oxygen during the sublima-

Ž . Ž . Ž . Ž . Ž .Fig. 4. AFM height images of: a RuPc film on Si substrate as grown Z scale �48 nm ; and b after annealing Z scale �20 nm .2

Ž . Ž . Ž . Ž .Fig. 5. AFM images Z scale �20 nm of annealed RuPc film on Pt substrate: a height image; and b phase image.2

( )R. Caminiti et al. � Thin Solid Films 382 2001 74�8080

tion process is likely to cause the formation of oxidisedderivatives. The evaporation method is indeed a usefulway to obtain reproducible and homogenous films, ashas been shown by AFM and EDXD studies. The datapresented clearly indicate that, in the passage from thebulk to the film, the dimeric structure of the compoundexamined is maintained, and the columnar packing ofdimeric molecules is more extended than in the bulk.When the films were heated to 200�C, structural modi-fications were not observed; only an increasing orderwas evidence as an effect of thermal treatment. TheEDXD technique, used here for the first time, is apromising tool for structural investigation of amor-phous films. Furthermore, this technique, by observingone peak-diffraction in different angle positions, issensitive to structural modifications, due, for instance,to thermal stress, which cause even a slight shift inpeak position.

References

� � Ž .1 C.C. Leznoff, A.B.P. Lever Eds. , Phthalocyanines: Propertiesand Applications, vols 1�4, VCH, Weinhem, 1989�1996.

� �2 G. Guillaud, J. Simon, J.P. Germain, Co-ord. Chem. Rev. 180Ž .1998 1433.

� �3 A. Capobianchi, A.M. Paoletti, G. Pennesi, G. Rossi, in: C. DiŽ .Natale, A. D’Amico Eds. , Proceedings of the First Italian

Conference on Sensors and Microsystems, Rome, Italy, 19�20February 1996, Sensors and Microsystems, World ScientificPublishing, Singapore, 1996, p. 50.

� �4 A. Capobianchi, A.M. Paoletti, G. Pennesi, G. Rossi, R.Ž .Caminiti, C. Ercolani, Inorg. Chem. 33 1994 4635.

� �5 R. Caminiti, C. Sadun, V. Rossi, F. Cilloco, R. Felici, Italianpatent No RM�93 01261484, 23 June 1993.

� � Ž .6 G. Fritsch, D.A. Keimel, Mater. Sci. Eng. A134 1991 888.� � Ž .7 M. Carbone, R. Caminiti, C. Sadun, J. Mater. Chem. 6 1996

1709.� �8 R. Caminiti, M. Carbone, G. Mancini, C. Sadun, J. Mater.

Ž .Chem. 7 1997 1331.� � Ž .9 M. Hanack, S. Osio-Barcina, E. Witke, Synthesis 1992 211.

� �10 A. Capobianchi, G. Pennesi, A.M. Paoletti, G. Rossi, R.Ž .Caminiti, C. Sadun, C. Ercolani, Inorg. Chem. 35 1996 4643.

� �11 L. Alagna, A. Capobianchi, P. Marovino, A.M. Paoletti, G.Ž .Pennesi, T. Prosperi, G. Rossi, Inorg. Chem. 38 1999 3688.

� � Ž .12 K. Nishikava, T. Itjima, Bull. Chem. Soc. Jpn. 57 1984 1750.� �13 J.P. Collman, C.E. Barnes, P.N. Swepston, J.A. Ibers, J. Am.

Ž .Chem. Soc. 106 1984 3500.� � Ž . Ž14 R.D. Gould, Co-Ord. Chem. Rev. 156 1996 237 and refer-

.ences therein .� �15 N.B. McKeown, Phthalocyanine Materials, Cambridge Univer-

sity Press, Cambridge, UK, 1998.� �16 D. Schmeiber, J. Pohmer, M. Hanack, W. Gopel, Synth. Met.¨

Ž .61 1993 115.