In situ UV-visible spectroelectrochemistry in the course of oxidative monomer electrolysis

Accepted Manuscript Title: In situ UV-visible spectroelectrochemistry in the course of oxidative monomer electrolysis Author: D.V. Konev O.I. Istakova O.A. Sereda М.A. Shamraeva C.H. Devillers M.A. Vorotyntsev PII: DOI: Reference:

S0013-4686(15)01447-4 http://dx.doi.org/doi:10.1016/j.electacta.2015.06.076 EA 25201

To appear in:

Electrochimica Acta

Received date: Revised date: Accepted date:

1-3-2015 11-6-2015 18-6-2015

Please cite this article as: D.V.Konev, O.I.Istakova, O.A.Sereda, Mcy.A.Shamraeva, C.H.Devillers, M.A.Vorotyntsev, In situ UV-visible spectroelectrochemistry in the course of oxidative monomer electrolysis, Electrochimica Acta http://dx.doi.org/10.1016/j.electacta.2015.06.076 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

In situ UV-visible spectroelectrochemistry in the course of oxidative monomer electrolysis as a tool to characterize the molecular structure of poly(Mg(II)porphine) a,b,

a,b

c

c

d,

a,b,c,d,

D. V. Konev *, O. I. Istakova , О. А. Sereda , М. А. Shamraeva , C.H. Devillers *, M. A. Vorotyntsev

*

a

Institute for Problems of Chemical Physics of the Russian Academy of Sciences, Chernogolovka, Russia D. I. Mendeleev University of Chemical Technology of Russia, Moscow, Russia c M. V. Lomonosov Moscow State University, Moscow, Russia d ICMUB - UMR 6302 CNRS - Université de Bourgogne et Franche-Comté, Dijon, France b

Corresponding authors: D. V. Konev: [email protected] , M. A. Vorotyntsev [email protected] Abstract

C. H. Devillers [email protected] ,

Novel method to characterize the macromolecuar structure of an electroactive polymer deposited via electrooxidation of the corresponding monomer on the electrode surface has been proposed. It is based on experimental determination of the number of electrons spent for oxidation of an initially solute monomer species which is used to calculate the number of covalent bonds linking each monomer unit with neighboring units inside the polymer. The former parameter is found by tracing simultaneously the variations of instantaneous values of the solute monomer concentration and of the passed charge in the course of the monomer oxidation electrolysis. This monomer concentration is established with the use of its UV-visible absorbance in the vicinity of the monomer band(s) during this process on the basis of an original procedure to separate the absorbances due to the monomer and solute oxidation products for each instantaneous spectrum. Correctness of the proposed method (including its theoretical principles, home-made installation and experimental realization) has been examined for the polypyrrole deposition, with the conclusion on its (mostly) chain structure, which is in conformity with previous evidences. Then, our approach has been applied to the electrooxidative polymerization of (non-substituted) Mg(II) porphine, for which an indirect experimental information of its molecular structure has only been available earlier. The spectroelectrochemical method has resulted in the conclusion that the oxidation product, poly(Mg(II)porphine), consists of chains of single-bonded monomeric porphine macrocycles, with a relatively small number of interchain bonds.

Keywords Electroactive materials, electropolymerization, unsubstituted porphine, redox equivalent, conjugated polymer, solute oligomers

1. Introduction Electroactive materials containing the porphine macrocycle in their structure are actively studied [1,2] in view of their applied prospects in electrocatalysis [3-7], photocatalysis [8], sensors [9,10], solar energy conversion [11-13], non-linear optics [14,15], etc. Another research direction is oriented to synthesis of catalytic metal (M) containing centers of the MN4 type on the surface of carbon support for their application as platinum-free catalysts of oxygen reduction for fuel cells [4, 7, 16‒19]. Porphine unit is present in most systems as a pendant group of a polymer chain. There are a few examples of copolymers possessing a well-defined molecular structure of alternating porphyrin and (hetero)aromatic units which are mostly synthesized by organometallic catalysis, see e.g. refs [20-24]. On the contrary, molecular structures of polymer films deposited by direct oxidation of substituted porphyrins on electrode surface were not studied [25-28]. Our team succeeded in synthesis of the first homopolymer (pMgP-I) on the basis of non-substituted Mg(II) porphine (MgP, Fig. 1a) [29]. This polymer has been characterized by combination of methods of electrochemistry, spectroscopy and

microscopy which demonstrated that it represents a typical electroactive material, i.e. it can be charged and discharged in a quasi-reversible manner by means of potential variation (within certain limits), this variation of the oxidation state being accompanied by the drastic change of its redox, conductive and spectroscopic properties. Three characteristic states have been found for this polymer: 1) neutral state with a low redox activity and low electronic conductivity, 2) oxidized state (where the polymer matrix bears a potential-dependent positive charge) which possesses enhanced values of both the redox activity and electronic conductivity, 3) reduced state where the matrix is charged negatively, its properties being qualitatively similar to those for the oxidized state. Optical spectra for the oxidized state demonstrate the existence of delocalized low-energy electronic species which are expectedly responsible for the electronic conductivity.

Fig. 1. Mg(II) porphine (a) and model structures of its homopolymer of type I, pMgP-I, as a linear chain (b) or a 2D structure (c) IR spectra, XPS and electrochemistry data [29] testify in favor of conservation of both the porphine macrocycle as monomer unit and Mg cation inside this material. More recently, methods of ion exchange have been employed to replace Mg cations by protons and to replace protons by Zn or Co cations, thus obtaining the free-base polyporphine, pH2P-I, and pZnP-I or pCoP-I, of type I, respectively [30,31]. If the electrode potential exceeds a certain value these polyporphine films of type I (pMP-I, M = Mg, 2H, Zn or Co) are subject to the irreversible electrochemical oxidation, with formation of extra bonds between the monomer units, thus resulting in the corresponding polymers of type II, pMP-II [31,32]. One should emphasize that the ion-exchange process employed for the change of the central cation(s) does not modify the molecular structure, i.e. it is the same for all polyporphines of type I synthesized in our studies. However, the information of this structure has so far been incomplete and based on indirect data [29]. Surprising resemblance of the IR vibrational spectra for the monomer (MgP) and polymer (pMgP-I) films not only proves the conservation of the Mg porphine macrocycle inside the material but represents also an evidence in favor of a single bond between neighboring monomer units inside the polymer. Both quantum-chemical and experimental (reactivity) data show that such bonds for Mg(II) and Zn(II) porphines are formed in meso-positions of the macrocycle, see the discussion in ref [29]. As a whole this information results in the conclusion on the meso-meso bonded molecular structure of pMgP-I (and all other polyporphines of type I). However, the existence of four unoccupied meso-positions of the porphine macrocycle leads to the uncertainty on the molecular arrangement of these polymers, see hypothetic examples in Figs 1b and 1c. Substances extracted from this polymer, pMgP-I, by THF and characterized by MALDI-TOF spectroscopy [29] possess the masses corresponding to chain porphine oligomers of various lengths, with a single bond between the neighboring monomer units, i.e. corresponding to the structure in Fig. 1b. However, according to comparison of the UV-visible spectra for the film and for the extracted solution one can see that the principal fraction of the polymer material was left non-dissolved in THF, thus making this characterization of the structure not relevant to it. The situation was calling for finding an alternative method to determine the average number of covalent bonds formed by a monomer unit inside the polymer, pMgP-I. The goal of the actual study has been to get direct information on this key parameter of the polymer structure via a detailed study of the oxidative transformation of the monomer from its state in solution into the monomer unit inside the polymer (the number of such bonds is equal to 2 for the linear chain structure in Fig. 1b or to 4 for the alternative 2D structure in Fig. 1c). Our approach is based on fundamental principles of electrochemical polymerization processes, in particular, those resulting in deposition of electroactive polymer films on electrode surface, which relate the amounts of the consumed monomeric reactant, of the reaction product(s) and of the electric charge spent for this process via stoichiometric relations based on the general scheme of the electrodeposition process for a monomer, Mon:

+α + Mon – nMon e = (1/n) (Mon )n + k H , nMon = k + α

(Scheme 1)

where the overall number of electrons, nMon, withdrawn from a monomer unit is equal to the sum of covalent bonds formed in average by a monomer unit inside the polymer, k, and the charging degree of the monomer unit under conditions of electrolysis, α. Reliable experimental determination of their parameters, k and α, may be used both to get conclusions on the molecular structure of the polymer and to estimate the current efficiency of the polymer film formation as well as the maximal degree of the polymer oxidation. The amount of oxidized monomer species may be determined from the temporal evolution of the spectrum of the electrolyzed solution in the UV-visible range. Its relation to the passed electric charge provides the information to calculate the number of chemical bonds that a monomer unit inside the polymer/oligomer structure forms with neighboring monomer units. This key parameter of the molecular structure of the oxidized product(s) has been established for this process for the Mg(II) porphine (MgP). Similar study of the pyrrole (Py) electropolymerization has been performed as a control experiment, aimed to check applicability of both the original home-made experimental setup and the above theoretical treatment of spectral data. The choice of the monomer amount as one of the principal variables of this study was dictated by the presence of several oxidation products in the course of the electrolysis: polymer in the form of the film on the working electrode and as colloid particles in solution as well as intermediate solute oxidation products (oligomers in various charging states).

2. Experimental objects and methods Our method for studying the electropolymerization process is based on the oxidative electrolysis (e.g. at constant potential) of the conjugated monomer solution, which is performed inside a special spectroelectrochemical cell, resulting in deposition of a polymer layer on the surface of anode. The UV-visible spectrum of the solution (in 200-900 nm wavelength range) is registered periodically, to trace the change of its composition. This measurement is synchronized with record of the chronoamperogram; its integration provides the temporal evolution of the passed oxidation charge. Spectroelectrochemical cell has been designed for studies under controlled atmosphere, in combination with the use of vacuum pumping for deoxygenation of the solution. Actual experiments were carried out with such pretreatment followed by Ar atmosphere with slight overpressure over solution. The cell is based on rectangular quartz spectrocuvette (Hellma, optical length: 10 mm) coated by hermetic cap with a hole for changing the atmosphere under the cap and current suppliers for three electrodes. Large area Pt gauze served in the actual study as working electrode (anode), Pt wire as counter electrode (CE) and Ag/0.01 M AgNO3 in acetonitrile (AN) as reference electrode (RE). Both CE and RE were separated from the principal solution by double frits with 0.1 M TBAPF6 + AN between them. Such separation was obligatory to realize the electrolysis inside the cell. Use of "true RE" (rather than the usually used "pseudoreference electrodes" like Ag or Pt wires) is necessary for the precise control of the potential imposed in the course of the electrolysis of Mg(II) porphine. Magnetic agitator is used to produce the convective motion of the solution of controlled intensity so that the monomer was transported towards the anode surface while the solute (intermediate) reaction products are distributed uniformly throughout the solution.. All potentials in the paper are given against this RE (Ag/0.01 M AgNO3 + AN). The potential of the Fc+/Fc couple in AN is 0.09 V with respect to this RE. Construction of the cell allows us to perform measurements of the spectrum either of the solution inside the cuvette or of the film (under control of its oxidation state by potentiostat) deposited on the optically transparent electrode (e.g. ITO). In the actual study the light beam passes inside the cuvette only the solution so that the measured absorbance is determined only by

components of this solution, in particular the monomer and solute reaction products. The spectrum prior to electrolysis is related solely to the monomer (as well as to other absorbing species, e.g. lutidine for the MgP oxidation). Since the principal object of this study, Mg(II) porphine (MgP), possesses very intensive absorption bands (absorption coefficients within the Soret and Q-bands are of the order of 105 and 104 M-1 cm-1) the absorbance within the principal bands of the monomer spectrum cannot be measured for the 10-mm optical path. The same problem takes place for the UV band of Py around 210 nm. Therefore, the optical path, Lopt, was shortened up to 2 mm by insertion of quartz cylindrical spacer (length along its axis: 8 mm) into the beam area inside the cuvette (thus replacing partially the solution inside this illuminated volume by quartz). This construction has allowed us to measure the whole spectrum of the MgP solution (approximate starting composition: 0.5 mM MgP + 1.5 mM lutidine + 0.1 M TBAPF6 in AN, see below), except for the Soret band, as well as the UV spectrum near 210 nm of the 0.5-1.0 mM Py solution. Prior to experiment the weight of the cell (with inserted electrodes and under Ar atmosphere) was measured. The cell was filled with monomer solution, which was deoxygenated by short-term vacuum pumping. Then, the weight of the cell with electrodes and solution was measured again for precise determination of the solution mass, msoln (with taking into account the solvent evaporation in the course of the vacuum treatment). This information gave us the volume of the solution: Vsoln = msoln / ρsoln, ρsoln being the density of 0.1 M TBAPF6 + AN solution, 0.79 g/cm3 (determined experimentally). The starting MgP concentration prior to electrolysis was found from its absorbance at the maximum (535 nm) of its principal Q-band (A535°), cMgP° = A535° / Lopt ε535, where the absorption coefficient, ε535 = 13737 ± 85 M-1 cm-1, was found by preliminary calibration of the concentration vs. absorbance relation in experiments without vacuum pumping. One can also determine the MgP concentration on the basis of the solution absorbance at another band maximum: ε308 = 18275 ± 100 M-1 cm-1. Similarly, the starting Py concentration prior to electrolysis was found from the absorbance at 210 nm, cPy° = A210° / Lopt ε210, ε210 = 6125 ± 25 M-1 cm-1. Then, the total amount of the monomer in solution prior to electrolysis was found on the basis of these data: NMgP° = cMgP° Vsoln or NPy° = cPy° Vsoln. For MgP this value was controlled by its comparison with the amount of the monomer dissolved in solution. The cell was placed into the spectrophotometer holder, its three electrodes being connected to contacts of the high-voltage potentiostat Elins P-30JM. The spectral and electric measurements were initiated synchronically. Monomer concentrations (about 0.5 mM or 0.1 mM) were chosen from the condition that the absorbances within the bands near 308 nm and 535 nm were in the linearity range of the spectrophotometer. Solvent (AN) and oxidation regime (potentiostatic, 0.35 V) for MgP were chosen on the basis of Ref. [29]. To increase the rate of the polymer deposition on the electrode surface as well as to increase the fraction of the oxidation charge resulting in the polymer film formation, strong proton acceptor, 2,6-dimethylpyridine (lutidine, Lu), was added to this solution, its concentration (about 3 equivalents of MgP) being optimal for this process [33]. Its choice among proton acceptors was determined by the absence of its coordination with the central ion, Mg(II), inside the porphine macrocycle.

3. Results and discussion 3.1. Electropolymerization of pyrrole (Py) Spectrum of the initial Py solution in AN (Fig. 2b) is characterized by its absorption band inside the UV range (maximum: near 210 nm). Oxidative electrolysis at the potential, E = 0.8 V, results in progressive diminution of the absorbance within the whole band, which reflects the consumption of Py molecules by electrode reaction accompanied by a series of chemical and electrochemical steps. Short-length oligomers (in their neutral or charged states) are not accumulated in solution in significant amounts, as it follows from the absence of new absorption bands at longer wavelengths. There is no visible "colloidal absorption" throughout the whole range of wavelengths, i.e. polymer is mostly formed as a film on the working electrode

(which does not contribute to the spectral signal). By the end of the electrolysis the absorption band of Py in the UV range (dotted line) practically vanishes but some residual absorbance (with the amplitude monotonously increasing towards shorter wavelengths) remains. Electrolysis current diminishes monotonously in time (Fig. 2a). Its integration provides the temporal variation of the oxidation charge, Q(t), passed by a time moment, t (Fig. 2a). Then, the initial concentration of Py in the cell found from the maximal absorbance (Fig. 2b) and the mass of the solution (after its deoxygenation) were used to determine the total amount of Py prior to electrolysis, NPy°, as well as the normalization charge, QPy° = F NPy° = 0.16 C corresponding to one electron withdrawal from each monomer species.

Fig. 2. Oxidative electrolysis of Py solution. Potentiostatic regime (0.8 V vs. Ag/0.01 M AgNO3 + AN). Initial composition: 0.89 mM Py + 0.1 M TBAPF6 + AN. (a) Curve 1: chronoamperogram, I(t); curve 2: integrated charge, Q(t); (b) Progressive variation of the spectrum of the electrolyzed solution: three highlighted spectra correspond to the initial state prior to electrolysis (1), to 50% transformation of MgP (2) and to the end of this treatment (3). Spectra 4 correspond to various intermediate states in the course of the process. Optical path: Lopt = 2 mm. Plot 1 in Fig. 3 presents the relation between the instantaneous values of the absorbance at the band maximum, A210 (t) and of the oxidation charge, Q(t), in the course of the electrolysis, both quantities being normalized, A210/A210° (A210°, maximal absorbance before electrolysis) and Q/QPy°. Within the assumption that the monomer (Py) represents the only electroactive and the only light absorbing species, these quantities are related to the instantaneous amount of Py in solution at the same moment, NPy: Q = nPy F (NPy° - NPy), A210 = ε210 Lopt NPy ρs / ms, A210/A210° = 1 - (nPy)-1 Q/QPy°

(1)

where nPy is the average number of electrons withdrawn from one Py molecule in the course of its transformation from a solute species towards a monomer unit inside the polypyrrole (PPy) film. Then, plot 1 in Fig. 3 should be linear and its intercept at the horizontal axis (as well as its inverse slope) should be equal to nPy. The linearity of this plot in Fig. 3 (within the dispersion of experimental points) is confirmed for most of its range. The above extrapolation of the data gives the value: nPy = 2.8.

Fig. 3. Relation between normalized values of the maximal absorbance, A210/A210°, and the oxidation charge, Q/QPy°, in the course of the oxidative electrolysis of Py solution (corresponding to data in Fig. 2). Plot 1: no correction for the residual absorbance; plot 2: after correction for the residual absorbance; straight line 3: X = 1 - 0.405 Q / QPy° (linear fitting of plot 2). However, such consideration disregards the existence of the residual absorbance within the absorption range under study (around 210 nm) in Fig. 2b, even though the absorption band related to solute Py has disappeared by the end of the electrolysis. It means that the assumption of Eq (1) for A210 that only Py absorbs the light at 210 nm is not valid. The absorption of these extra species may change in the course of the electrolysis process. As a result, it is risky to correct Eq (1) simply by subtraction of the residual absorption, A210resid, from both the peak absorbance, A210(t), and its initial value, A210°. Therefore, we applied a novel method based on the difference between a peak-shape of the Py absorption band around 210 nm and an approximate linearity of the residual spectrum within this wavelength range (this procedure is described in more detail in section 3.3). Its application to each absorption spectrum in Fig. 2b provided the value of the instantaneous Py concentration normalized by its initial value, X(t). These values are shown in Fig. 3 (plot 2) versus the normalized oxidation charge, Q/QPy°. One can see that it is linear, in conformity with Eq (2) derived from Scheme 1: X = 1 - (nPy)-1 Q/QPy° It gives for the number of electrons consumed per Py species in the course of the electrolysis process: nPy = 2.4 - 2.5.

(2)

These electrons are spent not only for formation of covalent bonds between Py units inside the polymer but also for the polymer oxidation up to the state corresponding to the electrolysis potential. The latter was estimated from the redox charge passing during the CV response of the polymer layer deposited at the working electrode as a result of the electrolysis. Thus found value of α (about 0.3) is in conformity with conventional data for electrochemically deposited polypyrrole, 0.25 see e.g. [34]. One should keep in mind that the anodic limit in the CV experiment for PPy has to be limited to 0.6 V, while the polymerization potential was 0.8 V. The anodic limit cannot be seriously increased because of the film overoxidation, which takes place even if the film is kept at 0.6 V. Due to the latter problem one cannot use a much smaller scan rate, either, while the specific redox charge (charging degree of the monomer unit) found in the course of a fast potential scan (0.3) may be noticeably lower than the equilibrium one for the same potential (0.6 V). These factors suggest that the steady-state oxidation degree of PPy in the course of its polymerization, α, may be even a bit greater: α = 0.3 - 0.4. Its combination with the above value of nPy ≡ k + α = 2.4 - 2.5 leads to the conclusion that about 2 electrons per monomer (k = 2.0 - 2.2) are spent for formation of the polymer. It means that each monomer unit inside the film is linked in average by two bonds with neighboring units. This result corresponds well to the literature data that the polymer consists of chains of Py units (linked in 2,5-positions, according to previous studies of this process). Thus, one may conclude that the control experiment on application of the newly proposed method for characterization of the molecular structure has given a reasonable result for polypyrrole.

3.2. Electropolymerization of Mg(II) porphine (MgP) Oxidative electrolysis of MgP has been performed from 0.5 mM MgP + 1.5 mM Lu + 0.1 M TBAPF6 solution in AN under potentiostatic regime (0.35 V). Fig. 4a demonstrates the chronoamperogram, I(t), as well as the variation of the oxidation charge, Q(t). The former curve looks in average as a decaying exponent (with a significant background current, though). Current oscillations originate from hydrodynamic pulsations in the course of intensive agitation of the solution. These current pulses do not affect markedly the integrated curve, Q(t). Parallel evolution of the absorption spectrum of this solution measured periodically during the electrolysis is shown in Fig. 4b. The initial spectrum consists of several narrow peaks.

Fig. 4. Oxidative electrolysis of 0.5 mM MgP + 1.5 mM Lu + 0.1 M TBAPF6 solution. Potentiostatic regime (0.35 V vs. Ag/0.01 M AgNO3 + AN). (a) Curve 1: chronoamperogram, I(t); curve 2: integrated charge, Q(t); (b) Progressive variation of the spectrum of the electrolyzed solution. Three highlighted spectra correspond to the initial state prior to electrolysis (1), to 50% transformation of MgP (2) and to the end of this treatment (3). Spectra 4 correspond to various intermediate states in the course of the process. Directions of the absorbance variation at characteristic wavelengths are shown by arrows. Absorbances in the central part of the Soret band (over 2.0) are beyond the measurement range of the spectrophotometer, they are not used for the further analysis. Optical path: Lopt = 2 mm. For their attribution the spectrum of pure Lu (with various acid additions) solutions in AN have been recorded (Fig. 5). One can see that the absorption is practically absent for the whole wavelength range over 300 nm. Therefore, all bands within this interval observed in Fig. 4b should be attributed to the monomer, MgP (initial spectrum), as well as to its soluble oxidation products. On the contrary, all Lu solutions in Fig. 5 possess an absorption band in the range between 250 and 280 nm which increases its intensity with initial acid additions because of the Lu protonation, with saturation of the absorption starting from the threefold excess of the acid where practically all Lu species have been protonated (LuH+).

Fig. 5. Absorption spectra of 0.2 mM Lu solution in AN with various additions of trifluoroacetic acid, TFA. Lu-to-TFA molar ratios: (1) 1 : 0, (2) 1 : 1, (3) 1 : 2, (4) 1 : 3, (5) 1 : 4, (6) 1 : 5. Optical length: Lopt = 2 mm.

Therefore, the further analysis of the spectra in Fig. 4 will be limited to the range over 300 nm. Within this interval one can note the Soret band around 400 nm (not used for analysis here because of the spectrophotometer saturation effect), several Qbands (the most intensive one lies at 535 nm) and the band around 308 nm related to electronic transitions inside porphine macrocycle. During the electrolysis the intensities of all MgP bands diminish progressively because of the monomer oxidation while the Lu band grows in intensity owing to additional protonation of Lu by protons which are liberated in the course of bond formation between monomeric and/or oligomeric porphine blocks. More complicated (non-monotonous) evolution of the solution spectrum is observed in some other ranges, e.g. around 438 and 565 nm. Here the absorbance is increasing in the initial period of the process, e.g. a new narrow band is visible in the middle of the electrolysis time. Continuation of the oxidation process inverts this tendency, and the absorbance becomes much less intensive by termination of the electrolysis. This behavior allows us to attribute these bands to soluble intermediate products, with their further oxidative transformation into the polymer. According to our previous data on the MALDI-TOF analysis of species extracted from deposited poly(Mg(II)porphine (pMgP-I) films by THF [29], porphine oligomers starting from their trimer are poorly soluble in the deposition mediun (AN solution) so that the porphine dimer, (MgP)2, should expectedly play the key role among these AN-soluble products, see below. Besides the above mentioned bands, one can note a progressive increase of "colloidal absorption" in the range over 600 nm that should be attributed to light scattering by polymer particles dispersed in solution by its intensive agitation. As a whole the absorption spectrum in Fig. 4b for any moment, t, of the electrolysis, Aλ(t), represents the sum of contributions due to absorption by the monomer (MgP), solute oxidation products, polymer colloid particles as well as non-protonated and protonated forms of lutidine, Lu and LuH+ (near 270 nm). All these features are in perfect agreement with the previous study of this process for MgP in CH2Cl2 [35] as well as with observations for Mg(II) and Zn(II) porphines doubly coordinated with pyridine molecules [36] and for 5,10,15-substituted Zn(II) porphyrin [37] in DMF.

Fig. 6. Relation between normalized values of the maximal absorbance, A535 (1) or A308 (2), and the oxidation charge, Q, in the course of the oxidative electrolysis of MgP solution based on experimental data: (a) in Fig. 4, (b) in Fig. 8, without correction for the residual absorbance. Points 3: relation for the normalized amount of the monomer, X = NMgP/NMgP° calculated with correction for absorbance of oxidation products in solution, see section 3.3. Straight lines 4 (passing through the initial point: X = 1, Q = 0) illustrate the extrapolation procedure. Similar to Fig. 3, plots in Fig. 6 show the relation between the instantaneous values of the absorbance at the band maximum, A308 (2, blue line) or A535 (1, red line), and of the oxidation charge, Q(t), in the course of the electrolysis, both quantities being normalized, A308/A308° or A535/A535° (A308° or A535°, maximal absorbance before electrolysis at 308 nm or 535 nm) and Q/QMgP° (QMgP° = F NMgP°, NMgP° is the total amount of MgP in the starting solution). Analog of Eq (1) can be derived for each absorption peak within the assumption that the absorbance at this wavelength is solely due to the monomer, MgP: A308/A308° = 1 - (nMgP)-1 Q/QMgP° or A535/A535° = 1 - (nMgP)-1 Q/QMgP°

(3)

Data in Fig. 6a are in conformity with predictions of Eqs (3) on the linearity of plots for each band (slight deviation from the linearity visible for the 535 nm band within the early stage of the process, especially in Fig. 6b, originates from the dominant formation of the porphine dimer in solution as the oxidation product, its increasing absorbance compensating a significant part of the monomer absorbance diminution, see Fig. 7 below) as well as on the proximity of the normalized absorbances for various bands, A308/A308° and A535/A535° (for the same charge). Extrapolation of plots 1 and 2 towards the horizontal axis

results in the values for the number of electrons per monomer, nMgP, equal to 3.0 or 3.3 for the 308 nm or 535 nm bands, respectively. This result is based on the hypothesis that only MgP absorbs within these bands. However, the absorbance does not vanish even for an extended electrolysis period, i.e. there should be other solution components which are optically active for these wavelengths. The oxidative electropolymerization of various conjugated monomers passes through formation of their shortchain oligomers as intermediate products which are subject to further oxidation into the corresponding polymer. Therefore, it is natural to expect a similar process for MgP resulting in Mg(II) oligoporphines. Unfortunately, to our best knowledge we did not find literature data on such species. In this context we had to address to the spectra of short-chain meso-meso bonded oligomers of 5,15-substituted Zn porphyrin [38] in view of the well-established similarity between properties of porphyrins with coordinated either Mg(II) or Zn(II) cations, in particular their tendency to form meso-meso bonds between monomer units. It has been revealed that the Soret band of such a Zn monomer is split into two narrow bands for its oligomers, one of them retaining the wavelength of the monomer while the other band being shifted to a longer wavelength (the shift depends on the oligomer length). Similar splitting of the Soret band of the monomer in its dimer has been observed for 5,10,15-substituted Zn porphyrin (Fig. 7). Another evidence on the nature of soluble oxidation product(s) is provided by our results [29] on the MALDI-TOF analysis of the composition of the solution obtained via extraction by THF from the pMgP-I polymer film. Its spectrum revealed Mg porphine oligomers with a chain structure, [MgP]n, n being 3 or greater while the dimer peak was absent. Since the film was deposited from the AN solution it implies that the porphine dimer possesses a sufficient solubility in AN to stay in the solute state while the porphine trimer or tetramer etc. are much less soluble in AN so that they are trapped by the film. This conclusion of the dimer as the principal soluble oxidation product is in conformity with the results of other electrolysis studies of Mg(II) and Zn(II) porphine/porphyrins [35-37]. Taken together, these findings testify in favor of the Mg porphine dimer as the principal intermediate oxidation product soluble in AN. Since the Soret band of such species should consist of two narrow bands, one may attribute the peak at 438 nm (compare Figs. 4 and 7) to the second (shifted) Soret band of the dimer. Since the other component of this band of the dimer should be located closely to the Soret wavelength of the Mg porphine monomer one can conclude of a strong contribution of the dimer around 400 nm, thus explaining the failure to liquidate this peak of the monomer by electrolysis because of the dimer presence. Superposition of the oligomer and monomer absorption may also be expected for other bands of the monomer. In particular, it has been established that the 5,15-substituted Zn porphyrin dimer (see Fig. 10 in Ref. [39]) and 5,10,15-substituted Zn porphyrin [37] (Fig. 7) possess a strong absorption in the vicinity of the principal Q-band of the corresponding monomer, with its maximum shifted towards longer wavelengths by 17 nm. This result matches well to our observation of a significant residual absorption within the MgP band around 535 nm as well as of the increasing absorption band around 565 nm, related to MgP oligomers.

Fig. 7. UV-visible absorption spectra of the 5,10,15-substituted Zn(II) monomer and its meso-meso bonded dimer in CH2Cl2 (adapted from reference [37]). Similar reasoning is also relevant for the evolution of absorption within the monomer band around 308 nm. However, no significant shift of the band takes place for meso-meso bonded oligomers of 5,15-substituted Zn porphyrin, compared to the corresponding monomer [38,39]. In conformity with this feature the band around 308 nm is decreasing monotonously as a whole (outside the range affected by the growing 270 nm band related to LuH+) in the course of the electrolysis.

One should keep in mind that the concentration of the dimer (as it is evident from the variation of the intensities at 438 nm and 565 nm) changes in time, its growth at the initial stage being replaced by its diminution later. Therefore, similar to the case of Py above, one cannot apply a simple procedure of subtraction of the residual absorption at the end of the electrolysis process, Aλresid.

Fig. 8. See legend to Fig 4, 0.1 mM MgP + 0.3 mM Lu + 0.1 M TBAPF6 solution. Optical path: Lopt = 10 mm With the hope to diminish their effect by reducing the dimer-to-monomer concentration ratio, the electrolysis experiment has been repeated for a much lower monomer concentration in solution, 0.1 mM MgP. The ratio of the starting MgP and Lu concentrations was kept at its optimal value, 1 : 3. Since the specific absorbance of all monomer peaks becomes smaller the electrolysis was performed in the cell without spacer, i.e. for the 10 mm optical path, in order to facilitate the agitation of solution as a whole and to avoid a non-uniform concentration distribution in bulk solution. Fig. 8 presents experimental data for variation of the current, oxidation charge (Fig. 8a) and spectrum (Fig. 8b). Then, the absorbances for the same wavelengths, 308 nm and 535 nm, have been used to get plots for the relations between the normalized values of these absorbances and the oxidation charge in the course of the electrolysis (Fig. 6b). All qualitative features were the same as those in Figs. 4 and 6a, in particular the shape of plots in Figs. 6b. However, the value of the principal parameter, nMgP, turned out to be about 4, i.e. markedly different from its value found from the data in Fig. 6a. It confirms once again that the correct value of this key parameter requires developing a special procedure of taking the absorption of soluble oxidation products (and that of the polymer colloid) into account in a proper way.

3.3. Procedure for separation of absorbances due to Mg(II) porphine (MgP) and its oxidation products Here, a more elaborated procedure is proposed which is designed for determination separately the contributions of both the monomer, MgP, and all other components into the absorption spectrum of the solution measured at any moment of the electrolysis, Aλ(t). Our approach is based on a characteristic qualitative feature of the relation between the spectra of the monomeric and dimeric porphyrin species. Namely, as discussed above the principal Q-band of the dimer is shifted slightly (less than 20 nm) towards longer wavelengths with respect to that of the corresponding monomer [37,39]. This shift is smaller than the sum of the band widths for them so that these bands are strongly overlapping. On the other hand, the maximum of the monomer band (λ = 535 nm) is located on the left slope of the dimer band. As a result, the tangent of the former plot is changing much faster in the vicinity of its maximum than that of the dimer spectrum, i.e. the latter (together with the absorbances due to the polymer colloid and Lu in both forms) may be approximated by a straight line in the interval between 520 and 550 nm, corresponding to the slowly varying slope of the dimer absorption spectrum: Adim+colloid+Lu+LuH+ ≅ a + b (λ - λ535)

(4)

where a priori unknown parameters a and b depend on the time moment, t. The experimentally measured absorbance, Aλ(t), is equal to the sum of contributions of the monomer and of the other species: Aλ(t) ≡ X(t) Aλ° + Adim+colloid+Lu+LuH+

(5)

where Aλ° is the monomer absorption prior to electrolysis, X(t) = NMgP(t)/NMgP°, fraction of the initial amount of the monomer left by the moment, t, X(t) being a priori unknown, either. The procedure below is designed to find this unknown value of X(t) on the basis of the spectral curve, Aλ(t), measured for any time moment, t. We discuss first its principles, then its practical realization.

Let us consider the experimental absorption spectrum, Aλ(t), of the solution in Fig. 4b or Fig. 8b measured at any particular time moment, t. According to Eqs (5) and (4), rapid change of the tangent of this function of the wavelength near λ = 535 nm is solely due to the first term, X(t) Aλ°. One should choose a set of test X values, Xtest, between 0 and 1, e.g. 0, 0.01, 0.02… 0.98, 0.99, 1. For each test value, Xtest, one should calculate the "difference function", Aλ(t) - Xtest Aλ°, for all wavelengths in the interval between 520 and 550 nm. For the initial test value, Xtest = 0, this difference function is equal to X(t) Aλ° + Adim+colloid+Lu+LuH+ (X(t) being the unknown true X value) and its second derivative with respect to λ has a large negative value at λ = 535 nm because of the first term, X(t) Aλ°. Increase of Xtest results in the diminution of the coefficient in front of Aλ°: Aλ(t) - Xtest Aλ° ≡ [X(t) - Xtest] Aλ° + Adim+colloid+Lu+LuH+, i.e. to a progressive increase of its second derivative with respect to λ at λ = 535 nm. For the value of Xtest equal to the real amount of monomer at this moment (Xtest = X(t)) the difference function becomes close to a straight line, according to Eqs (5) and (4): Aλ(t) - Xtest Aλ° = Adim+colloid+Lu+LuH+ ≅ X(t) Aλ° + a(t) + b(t) (λ λ535). As soon as the test value, Xtest, overcomes this critical point (Xtest > X(t)) the second derivative of the difference function, Aλ(t) - Xtest Aλ° ≡ [X(t) - Xtest] Aλ° + Adim+colloid+Lu+LuH with respect to λ at λ = 535 nm becomes positive. Thus, one can determine the true value of the remaining fraction of the monomer, X(t), via analyzing the sign of the 2nd derivative of the difference function at 535 nm for the set of Xtest values and choosing the Xtest value where this derivative is equal to 0 (then, Xtest = X(t)). For the practical realization of the procedure it is too time consuming to calculate the difference functions for all test values. Instead of it one can divide the initial interval (between 0 and 1) in two subintervals by a somehow chosen test value, Xtest (e.g. 0.5 or Aλ(t)/Aλ°) and calculate the 2nd derivative of the difference function for this Xtest value. Then, one may only consider the subinterval, in which the 2nd derivative changes its sign, and separate it again in two shorter subintervals, etc. These theoretical predictions are illustrated in Fig. 9 with the use of a particular absorption curve, Aλ(t), from Fig. 4. Difference functions for this curve, Aλ(t) - Xtest Aλ°, are shown for three test values, Xtest, which are chosen to demonstrate the change of the shape of the difference function within the whole range of wavelengths, depending on the relation between Xtest and the true value of X for this moment, X(t) (which was determined from this procedure). Detailed evolution of this spectrum, Aλ(t) - Xtest Aλ°, within the principal range between 500 and 580 nm is given in inset of Fig. 9. One can see that for the value of Xtest equal to 0.9 X(t), a hump around 535 nm (related to the monomer peak in Aλ°) is still retained in the corresponding difference function while this hump disappears if the test value of X reaches its true value: Xtest = X(t). Further increase of Xtest (Xtest = 1.1 X(t)) results in formation of a pit near 535 nm. Moreover, the difference function for this Xtest value within the whole wavelength range (Fig. 9) becomes negative around 300 nm which means that the real value of X is smaller: X(t) < Xtest.

Fig. 9. Illustration of the treatment of one of the instantaneous spectra in Fig. 4, Aλ(t) (measured after t = 20 min of electrolysis): Plots of the difference functions, Aλ(t) - Xtest Aλ°, for three Xtest values equal to 0.9 X(t), X(t) and 1.1 X(t) where X(t) = 0.67 (determined from this fitting procedure). Inset: fragments of these graphs in the wavelength range of the principal Q-bands for MgP and its dimer. The above criterion for determination of X(t) on the basis of the second derivatives at 535 nm is not suitable for its practical use because of the inevitable presence of experimental noise which strongly affects the values of the 2nd derivative. Therefore, more reliable results are obtained on the basis of another condition related to Eq (4): the value of the fitting parameter, Xtest, is chosen from the condition that the difference function, Aλ(t) - Xtest Aλ°, should be as close to a straight line as possible in the integral meaning, i.e. its mean-square deviation from an (a priori unknown) straight line should pass a minimum at Xtest = X(t). Such minimization procedure has been applied to all experimental spectra in the course of the electrolysis in Fig. 4b and in Fig. 8b. Thus found X(t) values for all moments of the process, t, are plotted in Figs. 6a and 6b (curves 3) versus the normalized

oxidation charge, Q/QMgP°, for the same time moments, t. Then, these data were interpreted with the use of Eq (3'), analogous to Eq (2) for Py, where the number of electrons withdrawn from one MgP species, nMgP, corresponds to Scheme 1: X = 1 - (nMgP)-1 QMgP/QMgP°

(3')

Extrapolation of their linear sections to the horizontal axis (or determination of the slopes of these lines) gives nMgP = 2.3±0.1. One may note a deviation of corrected plots in Figs 6a,b from linearity for the final stage of the electrolysis, the deviation starting earlier for the lower monomer concentration (Fig. 6b). First, such deviations of plots from the theoretical prediction at a later stage of the process are a typical feature of numerous electrolyses which originate from the fraction of the current spent for secondary reactions (oxidation of contaminations, etc.) growing in time because of diminution of the reactant (monomer) concentration and correspondingly of its oxidation current. Besides, it is well known that the exposure of the polymer in contact with the monomer-free solution to the polymerization potential results in the slow but progressive polymer overoxidation where an extra charge is spent for irreversible oxidation of the "normally deposited polymer". One may expect that the situation during the later stage of the electrolysis is close to the above one since the monomer concentration in solution becomes very low and an increasing fraction of the current is related to the overoxidation of the earlier deposited polymer. On the contrary, the same polymerization potential does not result in such polymer overoxidation in the course of its long-term deposition from a solution of a "normal" (sufficiently high) monomer concentration. These reasons resulted in our determination of slopes of plots 3 in Figs. 6 within their linearity ranges, without use of their long-term experimental points. Correctness of the proposed procedure for determination of X(t) values has been verified via calculation of the corresponding absorption spectra for all solution components, except for the monomer, with the use of Eq (6): Adim+colloid+Lu+LuH+(t) = Aλ(t) - X(t) Aλ°

(6)

Thus calculated Adim+colloid+Lu+LuH+(t) spectra are given in Figs. 10a and 10b for the initial MgP concentration equal to 0.5 mM or 0.1 mM, respectively, representative ones for 3 characteristic time moments (t1 < t2 < t3) being set off. For each system this absorbance increases in the course of the electrolysis process due to several factors: 1) protonation of Lu (peak around 270 nm), 2) accumulation of the dimer, (MgP)2 (peaks around 308, 438 and 565 nm), 3) formation of the polymer colloid (it determines the absorbance for wavelengths over 650 nm but it should give a contribution within the whole range of wavelengths). One should note that the ratios of the intensities of these three peaks for the dimer are retained for various moments, ti, and that their values are close to each other for Figs.10a and 10b. Moreover, the shape of thus found dimer spectrum corresponds well to that for dimeric porphyrins [35-37,39]. More quantitative comparison between all these values is not justified since the spectra in Figs. 9 in the range over 300 nm include not only the dimer absorbance but also a contribution of the polymer colloid.

F Fig. 10. Difference functions, Adim+colloid+Lu+LuH+(t) = Aλ(t) - X(t) Aλ° (for X(t) values found via the procedure in Fig. 9), demonstrating the total absorption by all solution components, except for the MgP monomer, for various time moments in the course of the electrolysis process. Initial MgP concentration: (a) 0.5 mM, (b) 0.1 mM. X(t) values correspond to data in Figs. 6a and 6b, respectively. The difference functions are not shown in the range between 390 and 415 nm because of too high absorbance, Aλ(t), within the Soret band of the monomer. Absorbance is stronger for all wavelengths in Fig. 9b (despite expectedly lower concentrations of the monomer for this system), in comparison with Fig. 9a, since the optical path in Fig. 9b was 5 times longer, 10 mm. 3.4. Estimation of the contribution due to the redox charge of the polymer The charge needed for oxidation of a monomer species up to a monomer unit inside the polymer is mostly spent for formation of covalent bonds between neighboring units, each bond requiring to withdraw 2 electrons, i.e. 1 electron from each bonded unit. Besides, an additional charge is spent to oxidize the polymer up to the state corresponding to the electrolysis potential. To estimate the latter contribution the working electrode coated by the polymer film after termination of the electrolysis was placed into a conventional three-electrode cell in contact with background (TBAPF6 +AN) solution. Redox responses of

polymer films deposited for two initial MgP concentrations are shown in Fig. 11. Their shape corresponds well to our previous observations for the same polymer, poly(Mg(II)porphine) of type I (pMgP-I), deposited at a Pt or GC disc electrode [29]. The redox response is stable upon cycling within this potential range. The oxidation and reduction charges (Qredox) are stable and well balanced, being equal to 7.5 mC or 2.5 mC for the films deposited via electrolysis of the 0.5 mM or 0.1 mM MgP solution, respectively. Taking into account the total electrolysis charges (Qtot), 193 mC and 51.5 mC, as well as the number of electrons spent for oxidation of a monomer species, nMgP = 2.4 - 2.5, one can get an upper estimate for the number of monomer units inside the film (in moles): NMgP(polym) ≤ Qtot / nMgP F. The number of electrons per monomer unit spent for the polymer oxidation from its neutral state, α, is determined by relations: α = Qredox / F NMgP(polym) ≥ nMgP Qredox / Qtot

(7)

The above results for the charges give a lower estimate for α: 0.10±0.01. The real value of this parameter is greater since a marked fraction of the oxidized monomer is accumulated in the course of the electrolysis in the form of the dimer and the polymeric colloid. Since the shape of the redox response of the polymer layer in Fig. 11 corresponds well to that for pMgP-I films deposited under identical conditions for the same solution composition inside the conventional electrochemical cell one may assume that the upper estimate for the redox charge per monomer unit, α > 0.15 [33], also valid for the actual study. In section 3.1 for the PPy deposition we discussed two factors which might affect the estimate of the charging degree of the polymer: 1) difference between the polymerization potential and the anodic limit of the CV scan used to measure the redox charge, 2) lower redox charge given by the CV study compared to its equilibrium value for the anodic limit of the scan. In the case of the MgP polymerization these factors act in the opposite directions since the anodic limit of the scan (0.4 V) is higher than the polymerization potential (0.35 V) which should counterbalance the incomplete charging of the polymer at 0.4 V during the potential scan.

Fig. 11. CV response of the working electrode (coated with polymer film in the course of the MgP oxidative electrolysis) in contact with TBAPF6 + AN solution. Initial MgP concentration prior to electrolysis: 0.5 mM (20 mV/s, 193 mC electrolysis charge, curve 1) or 0.1 mM (100 mV/s, 51.5 mC electrolysis charge, curve 2). One should keep in mind that 2 electrons per monomer unit represent an absolute minimum for formation of a polymer material. This value corresponds its chain structure where each unit forms in average only two bonds with its neighbors. This structure includes linear chains as well as various zigzag or branching ones. Such structural elements as cycles or/and 2D lattices require to spend more than 2 electrons per unit for their formation from monomers. The experimentally found number of electrons spent for formation of covalent bonds between the monomer units, k, is probably slightly greater than 2 (k = 2.1 - 2.3). This excess may mean the presence of some number of intermolecular bonds between different polymer chains or cycle elements. This value of k matches well with our results on the effect of Lu additions into the MgP solution for electropolymerization. Namely, it was found that the rates of both the monomer oxidation and the polymer film deposition are accelerated strongly in the presence of this active proton acceptor, parallel to increase of the Lu-to-MgP molar ratio up to 3 : 1, while a further increase of the Lu concentration does not lead to a further effect. One may expect that the strong acceleration of the process takes place for Lu concentrations which are comparable with the concentration of protons liberated due to the polymer bond formation, k = 2.1 - 2.3 per monomer species. Therefore, values of the Lu-to-MgP molar ratio over 3 : 1 correspond already to excessive amounts of the proton acceptor, in conformity with our experimental observations.

4. Conclusions

Novel method to characterize the molecular structure of an electrochemically deposited conjugated polymer has been proposed on the basis of the evolution of the UV-visible spectrum of the monomer solution in the course of its oxidative electrolysis. Necessity of a proper subtraction of the contribution due to absorption by various oxidation products generated inside the electrolyzed solution has been revealed. Original procedure to carry out such subtraction for each instantaneous spectrum during the electrolysis has been proposed, with its application to the polymerization process of the Mg(II) porphine as an illustration of its capacities. It has been demonstrated that such a treatment allows one to determine both the amount of the monomer (left in solution at any time moment) and the absorption spectrum of all oxidation products in solution (also for a set of time moments). Application of this method has shown that about two electrons are consumed for formation of the polymer structure both in polypyrrole and in poly(Mg(II)porphine). It means that the average number of covalent bonds between neighboring monomer units inside these polymer films is approximately equal to the number of monomer units, i.e. they possess a chain molecular structure, without a great number of intermolecular links.

5. Acknowledgements: The study has been carried out under financial support of the Russian Science Foundation (grant 1413-01244).

References 1. Kadish K.M., Smith K.M. and Guilard R. The Porphyrin Handbook. World scientific, 2010-2011. V. 1– 20. 2. Zhao H.Z., Chang Y.Y. and Liu C., Electrodes modified with iron porphyrin and carbon nanotubes: application to CO reduction and mechanism of synergistic electrocatalysis, J. Solid State Electrochem., 17, 2013, 1657-1664. 3. Jahan M., Bao Q. and Loh K.P., Electrocatalytically Active Graphene–Porphyrin MOF Composite for Oxygen Reduction Reaction, J. Am. Chem. Soc. 134, 2012, 6707- 6713. 4. Han A., Jia H., Ma H., Ye S., Wu H., Lei H., Han Y., Cao R. and Du P., Cobalt porphyrin electrode films for electrocatalytic water oxidation, Phys. Chem. Chem. Phys. 16, 2014, 11224-11232. 5. He Q., Mugadza T., Hwang G. and Nyokong T., Mechanisms of Electrocatalysis of Oxygen Reduction by Metal Porphyrins in Trifluoromethane Sulfonic Acid Solution, Int. J. Electrochem. Sci. 7, 2012, 7045-7064. 6. Liu B., Brückner C., Lei Y., Cheng Y., Santoro C. and Li B., Cobalt porphyrin-based material as methanol tolerant cathode in single chamber microbial fuel cells (SCMFCs), J. Power Sources. 257, 2014, 246-253. 7. Hijazi I., Bourgeteau T., Cornut R., Morozan A., Filoramo A., Leroy J. and Campidelli S., Carbon Nanotube-Templated Synthesis of Covalent Porphyrin Network for Oxygen Reduction Reaction, J. Am. Chem. Soc. 136, 2014, 6348-6354. 8. Ladomenou K., Natali M., Iengo E., Charalampidis G., Scandola F. and Coutsolelos A.G., Photochemical hydrogen generation with porphyrin-based systems, Coord. Chem. Rev. 2014. DOI:10.1016/j.ccr.2014.10.001. 9. Chandra S., Mende C., Bahadur D., Hildebrandt A. and Lang H., Fabrication of a porphyrin-based electrochemical biosensor for detection of nitric oxide released by cancer cells, J. Solid State Electrochem. 19, 2015, 169-177. DOI 10.1007/s10008-014-2583-z 10. Giancane G. and Valli L., State of art in porphyrin Langmuir–Blodgett films as chemical sensors, Adv. Colloid Interface Sci. 171, 2012, 17-35. 11. Deng L., Gu Z., Cao Z., Shen P., Li X., Zhang L. and Tan S. , Synthesis and characterization of novel D– A porphyrin-containing copolymers for polymer solar cells, Mater. Sci. Eng.: B. 177, 2012, 16411648. 12. Cao J., Liu J.C., Deng W.T., Li R.Z. and Jin N.Z., Effect of a novel self-assembly based on coordination polymer with zinc porphyrin in supramolecular solar cells, Org. Electronics. 14, 2013, 2713-2720.

13. Chao Y.I.H., Jheng J.F., Wu J.S., Wu K.Y., Peng H.H., Tsai M.C. and Hsu C.S., Porphyrin-Incorporated 2D D–A Polymers with Over 8.5% Polymer Solar Cell Efficiency, Adv. Mater. 2014, 5205-5210. DOI: 10.1002/adma.201401345. 14. Screen T.E.O., Thorne J.R., Denning R.G., Bucknall D.G. and Anderson H.L., Two methods for amplifying the optical nonlinearity of a conjugated porphyrin polymer: transmetallation and selfassembly, J. Mater. Chem. 13, 2003, 2796-2808. 15. Screen T.E.O., Lawton K.B., Wilson G.S., Dolney N., Ispasoiu R., Goodson T. and Anderson H.L., Synthesis and third order nonlinear optics of a new soluble conjugated porphyrin polymer, J. Mater. Chem. 11, 2001, 312-320. 16. Morozan A., Jousselme B. and Palacin S., Low-platinum and platinum-free catalysts for the oxygen reduction reaction at fuel cell cathodes, Energy Environ. Sci. 4, 2011, 1238-1254. 17. Vengatesan S., Cho E. and Oh I. H., Development of non-precious oxygen reduction reaction catalyst for polymer electrolyte membrane fuel cells based on substituted cobalt porphyrins, Korean J. Chem. Eng. 29, 2012, 621-626.. 18. Cheng N., Kemna C., Goubert-Renaudin S. and Wieckowski A., Reduction Reaction by PorphyrinBased Catalysts for Fuel Cells, Electrocatal. 3, 2012, 238-251. 19. Tarasevich M.R., Beketaeva L.A., Efremov B.N., Zagudaeva N.M., Kuznetsova L.N., Rybalka K.V. and Sosenkin V.E., Electrochemical Properties of Carbon Black AD-100 and AD-100 Promoted with Pyropolymer of Cobalt Tetra(p-methoxyphenyl)porphyrin, Russ. J. Electrochem. 40, 2004, 542-551. 20. Martin R.E. and Diederich F., Linear Monodisperse π-Conjugated Oligomers: Model Compounds for Polymers and More, Angew. Chem., Int. Ed. 38, 1999, 1350-1377. 21. Anderson H.L., Building molecular wires from the colours of life: conjugated porphyrin oligomers, Chem. Commun., 1999, 2323-2330. 22. Harvey P.D., Recent Advances in Free and Metallated Porphyrin Assemblies and Arrays; A photophysical Behavior and Energy Transfer Perspective, in: The Porphyrin Handbook, K.M. Kadish, K.M. Smith and R. Guilard (Eds.), vol. 18, Academic Press, 2003, ch. 113, pp. 63-250. 23. Liddell P.A., Gervaldo M., Bridgewater J.W., Keirstead A.E., Lin S., Moore T.A., Moore and Gust D., Porphyrin-Based Hole Conducting Electropolymer, Chem. Mater. 20, 2008, 135-142. 24. Umeyama T., Takamatsu T., Tezuka N., Matano Y., Araki Y., Wada T., Yoshikawa O., Sagawa T., Yoshikawa S. and Imahori H., Synthesis and Photophysical and Photovoltaic Properties of Porphyrin−Furan and −Thiophene AlternaRng Copolymers, J. Phys. Chem. C, 113, 2009, 1079810806. 25. Bedioui F., Trevin S., Albin V., Villegas M. G. G. and Devynck J., Design and characterization of chemically modified electrodes with iron(III) porphyrinic-based polymers: study of their reactivity toward nitrites and nitric oxide in aqueous solution, Anal. Chim. Acta 341, 1997, 177-185. 26. Bedioui F., Devynck J. and Bied-Charreton C., Immobilization of metalloporphyrins in electropolymerized films: design and applications, Acc. Chem. Res. 28, 1995, 30-36. 27. Rault-Berthelot J., Paul-Roth C., Poriel C., Juillard S., Ballut S., Drouet S. and Simonneaux G., Comparative behaviour of the anodic oxidation of mono-, di- and tetra-arylporphyrins: Towards new electroactive materials with variable bandgaps, J. Electroanal. Chem. 623, 2008, 204-214. 28. Paul-Roth C.O., Letessier J., Juillard S., Simonneaux G., Roisnel T. and Rault-Berthelot J., Synthesis, solid-state molecular structure and polymerization of a trans-substituted meso-porphyrin with thienyl pendant arms, J. Mol. Struct. 872, 2008, 105-112. 29. Vorotyntsev M.A., Konev D.V., Devillers C.H., Bezverkhyy I. and Heintz O., Magnesium(II) polyporphine: The first electron-conducting polymer with directly linked unsubstituted porphyrin units obtained by electrooxidation at a very low potential, Electrochim. Acta. 55, 2010, 6703-6714. 30. Konev D.V., Devillers C.H., Lizgina K.V., Zyubina T.S. , Zyubin A.S., Valkova L.A. and Vorotyntsev M.A., Synthesis of new electroactive polymers by ion-exchange replacement of Mg(II) by 2H+ or Zn(II) cations inside Mg(II) polyporphine film, with their subsequent electrochemical transformation to condensed-structure materials, Electrochim. Acta, 122, 2014, 3-10. 31. Konev D.V., Devillers C.H., Lizgina K.V., Baulin V.Е., Kalashnikova I. P. and Vorotyntsev M.A., Novel Family of Porphine-Based Electroactive Materials. Electrocatalysis of Oxygen Reduction Reaction, Proceedings of the 65th Annual Meeting of the International Society of Electrochemistry, ise142347, 2014, Lausanne, Switzerland.

32. Vorotyntsev M.A., Konev D.V., Devillers C.H., Bezverkhyy I. and Heintz O., Electroactive polymeric material with condensed structure on the basis of magnesium(II) polyporphine, Electrochim. Acta, 56, 2011, 3436-3442. 33. Konev D.V., Devillers C.H., Lizgina K.V., Baulin V.E. and Vorotyntsev M.A., Electropolymerization of non-substituted Mg(II) porphine: Effects of proton acceptor addition, J. Electroanal. Chem. 737, 2015, 235-242. 34. Diaz A.F., Castillo J.I., Logan J.A. and Lee W.Y., Electrochemistry of conducting polypyrrole films, J. Electroanal. Chem. 129, 1981, 115-132. 35. Devillers C.H., Lucas D., Dime A.K.D., Rousselin Y. and Mugnier Y., Exploring the redox reactivity of magnesium porphine. Insight into the origins of electropolymerisation, Dalton Trans. 39, 2010, 2404-2411. 36. Devillers C.H., Dime A.K.D., Cattey H. and Lucas D., Crystallographic, spectroscopic and electrochemical characterization of pyridine adducts of magnesium(II) and zinc(II) porphine complexes, C. R. Chimie, 16, 2013, 540-549. 37. Dimé A.K.D., Devillers C.H., Cattey H., Habermeyer B. and Lucas D., Control over the oxidative reactivity of metalloporphyrins. Efficient electrosynthesis of meso,meso-linked zinc porphyrin dimer, Dalton Trans. 41, 2012, 929-936. 38. Osuka A. and Shimidzu H., meso, meso-Linked Porphyrin Arrays, Angew. Chem. Int. Ed. Engl. 36, 1997, 135-137. 39. Miyahara T., Nakatsuji H., Hasegawa J., Osuka A., Aratani N. and Tsuda A., Ground and excited states of linked and fused zinc porphyrin dimers: Symmetry adapted cluster (SAC)—configuration interaction (CI) study, J. Chem. Phys., 117, 2002, 11196-11207.