Synthesis of three anchoring group possessing porphyrin sensitizers

The zinc porphyrins used in this study Zn1T3A, Zn1U3A, Zn1S3A, Zn1TH3A, Zn1TC3A, Zn1BC3A, and Zn1TPA3A were synthesized by, firstly, mixed condensation,¹¹ then zinc metalation,¹² and, at last, base hydrolysis.¹³ The 5,10,15-meso-methylcarboxyphenyl ester substituted free base porphyrin are isolated from the same mixed condensation reactions as described in last chapter while their metalation and hydrolysis followed the same procedures as described in Chapter 3.

(I) Mixed condensation: Condensation of pyrrole, methyl 4-formylbenzoate, and the required aldehyde under Lindsey’s conditions catalyzed by boron trifluoride-diethyl etherate followed by subsequent oxidation by DDQ afforded the trimester derivatives of porphyrins in good yield along with mixture of five other porphyrins. The yields of the triester derivatives porphyrins

138 1T3E, 1U3E, 1TH3E, 1TC3E, 1BC3E, and 1TPA3E obtained from each separate reaction are reported in Table 1. The free base porphyrin 1TPA3E obtained in 10% yield is the precursor of the top performance Zn1TPA3A with a conversion efficiency of 5.26% (vide infra). In ATR-FTIR all the porphyrins shows one stretching frequency at around 1720 cm⁻¹ supporting the presence of the ester carbonyl group. In UV‒visible spectra the free base porphyrins shows single strong Soret band and four moderate Q bands.

aReaction conditions: (i) methyl 4-formylbenzoate (3.75 mmol), required-carboxyaldehyde (3.75 mmol), pyrrole (7.5 mmol), BF3·OEt2 (10 mol%), CH2Cl2 (750 ml), 1 h (ii) DDQ (7.5 mmol), 1 h. ^b Yields of analytically pure products.

(II) Zn metalation: The subsequent step of zinc metalation has been readily achieved in high yields by reacting free base porphyrin with zinc acetate. The yields of the zinc(II) porphyrins Zn1T3E, Zn1U3E, Zn1TH3E, Zn1TC3E, Zn1BC3E, and Zn1TPA3E are listed in Table 2.

The success of zinc metalation of all the porphyrin was confirmed through the complete disappearance of the NMR resonance of inner NH with slight upfield shifts for all remaining protons.

139

a Reaction conditions: free base porphyrin (60 mg), Zn(OAc)2·2H2O (1.5 equivalent), CH2Cl2/MeOH (3/1, v/v), reflux 1 h, ^b Yields of analytically pure products.

(III) Hydrolysis: Hydrolysis of metal complexes has been achieved straightforwardly by reacting metal complexes in a mixture solution of THF and methanol with excess aqueous KOH. The yields of final hydrolyzed products Zn1T3A, Zn1U3A, Zn1TH3A, Zn1TC3A, Zn1BC3A, and Zn1TPA3A are listed in Table 3. ATR-FTIR spectra of final acid products show shifting of carbonyl peaks in the range of 1675−1700 cm⁻¹ because of intermolecular hydrogen bonding. All of the porphyrins were fully characterized by optical spectroscopy, ATR-FTIR, nuclear magnetic resonance spectroscopy, and high-resolution mass spectrometry.

140

a Reaction conditions: Zn(II)porphyrin (50 mg), 1 M KOH (20 equivalent), THF/MeOH (3/1, v/v), reflux, 10 h. ^b Yields of analytically pure products.

Syntheses and Characterization data.

5,10,15-tris(4-methoxycarbonylphenyl)20-(4-methylphenyl)porphyrin (1T3E). mp ˃ 300

oC; ¹H NMR (300 MHz, CDCl3) δ = 9.91 (d, J = 4.7 Hz, 2H), 8.81 (m, 4H), 8.79 (d, J = 4.8 Hz, 2H), 8.44 (d, J = 8.2 Hz, 6H), 8.30 (d, J = 8.2 Hz, 6H), 8.09 (d, J = 8.1, 2H), 7.56 (d, J = 7.7 Hz, 2 H), 4.11 (s, 9H), 2.71 (s, 3 H), -2.71 (s, 2H); IR (Neat, cm^-1): 3307, 1719, 1605, 1434, 1370,1275, 1180, 1100, 1019, 964, 797; λ_abs/nm (CH₂Cl₂): 420, 516, 552, 590, 646;

HRMS-ESI calcd for C₅₁H₃₈N₄O₆ ([M+H]⁺): 803.2870, found 803.2877.

5,10,15-tris(4-methoxycarbonylphenyl)20-(undecyl)porphyrin (1U3E). mp ˃ 300 ^oC; ¹H NMR (300 MHz, CDCl₃) δ = 9.48 (d, J = 4.8 Hz, 2H), 8.86 (d, J = 4.7 Hz, 2H), 8.76 (s, 4H), 8.46- 8.42 (m, 6H), 8.30-8.26 (m, 6H), 4.95 (t, J = 8.0, 2H), 4.14 (s, 6H), 4.11 (s, 3H), 2.53 (p, J = 7.4 Hz, 2H) 1.78 (p, J = 7.3 Hz, 2H), 1.51 (p, J = 7.4 Hz, 2H), 1.37-1.26 (m, 12 H), 0.87 (t, J = 6.8 Hz, 3H), -2.73 (s, 2H); IR (Neat, cm^-1): 3313, 1719, 1606, 1434, 1271, 1226, 1192,

141 1177, 1099, 1021, 962; λabs/nm (CH2Cl₂): 418, 516, 552, 592, 648; HRMS-ESI calcd for C₅₅H₅₄N₄O₆ ([M+H]⁺): 867.4122, found 887.4146.

5,10,15-tris(4-methoxycarbonylphenyl)20-(4-hexyl-2-thienyl)porphyrin (1TH3E). mp ˃ 300 ^oC; ¹H NMR (400 MHz, CDCl₃) δ = 9.17 (d, J = 4.12 Hz, 2H), 8.79 (m, 6H), 8.44 (d, J = 1722, 1608, 1550, 1481, 1363, 1274, 1263, 1180, 1110, 975, 918, 875, 800, 763, 750; λabs/nm (CH2Cl2): 421, 517, 554, 591, 650; HRMS-ESI calcd for C68H57N5O6S ([M+H]⁺): 1072.4108,

142

5,10,15-tris(4-methoxycarbonylphenyl)20-(undecyl)porphyrinato zinc(II) (Zn1U3E). mp

˃ 300 ^oC; ¹H NMR (400 MHz, CDCl3) δ = 9.48 (d, J = 4.8 Hz, 2H), 8.87 (d, J = 4.6 Hz, 2H), 1697, 1606, 1552, 1475, 1363, 1271, 1176, 1114, 1101, 1073, 995, 975, 910, 877, 808, 792, 761; λabs/nm (CH2Cl2): 424, 552, 595; HRMS-FAB calcd for C68H55N5O6SZn ([M+H]⁺):

1134.3243, found 1134.3274.

143

5,10,15-tris(4-carbonylphenyl)20-(undecyl)porphyrinato zinc(II) (Zn1U3A). mp ˃ 300 ^oC;

1H NMR (300 MHz, DMSO-D₆) δ = 13.24 (s, 3H), 9.69 (d, J = 4.5 Hz, 2H), 8.81 (d, J = 4.4

1.59-144

5,10,15-tris(4-carbonylphenyl)20-(3,6-di-tert-butyl-9-(thiophen-2-yl)-9H-carbazole)porphyrinato zinc(II) (Zn1TC3A). mp ˃ 300 ^oC; ¹H NMR (400 MHz, DMSO-D6) δ = 13.25 (s, 3H), 9.30 (d, J = 4.7 Hz, 2H), ), 8.87 (d, J = 4.7, 2H), 8.79 (s, 4H), 8.38-8.36 (m,

The UV‒visible absorption spectra of the studied zinc(II) porphyrin dyes displayed in Figure 1a show one strong Soret band and two Q bands. The peak positions and their molar absorption coefficients (ε) are summarized in Table 4. The shiftings in absorption bands of porphyrins are substituent-dependent. Porphyrins Zn1T3A and Zn1U3A display Soret band

145 absorption at 425 nm, whereas in the cases of porphyrins Zn1TH3A, Zn1TC3A, Zn1BC3A and Zn1TPA3A with electron donating substituents the Soret bands shift to 426‒427 nm. The overall highest molar absorption coefficient (ε) has been observed for Zn1U3A. It is interesting to observe that the porphyrins Zn1TH3A with relatively high molar absorption coefficient (ε) in Q band region also displays higher efficiency. Although the molar absorption coefficient (ε) is not high for ZnTPA3A, the low absorbance might be compensated by a broaden absorption and better electron injection to result in the highest overall conversion efficiency (vide infra). The emission spectra of all the porphyrins excited at Soret band displayed in Figure 1b show two peaks. The emission spectra of Zn1THA3A and Zn1TPA3A are significantly red shifted compared to remaining porphyrins.

Table 4. Optical and Electrochemical Data of Porphyrins Dye λabsa/nm were estimated from the intersection of the absorption and emission spectra. ^eExcited state oxidation potentials approximated from Eox and E0-0.

To understand the adsorption behavior of porphyrins on TiO2,the thin film absorption spectra of these porphyrins were studied. The UV‒visible spectra of studied porphyrins on TiO2

146 Figure 1. (a) UV‒visible spectra of Zn1T3A, Zn1U3A, Zn1S3A, Zn1TH3A, Zn1TC3A, Zn1BC3A, and Zn1TPA3A. Inset displaying enlarged spectra of longer wavelength. (b) Fluorescence spectra of studied porphyrins.

displayed in Figure 2 show shifting in peak positions with broadenings in shape. Porphyrins Zn1T3A and Zn1S3A broaden in shape with no significant shifting in peak positions, whereas TiO2 films absorbing Zn1TH3A, Zn1TC3A, Zn1BC3A, or Zn1TPA3A stacked end-to-end to produce J-aggregation with 4‒5 nm red shifts.¹⁴ Porphyrin Zn1U3A split into a 15 nm blue shifted band and 3 nm red shifted shoulder compared to THF solution of Zn1U3A owing to both H and J-aggregation.¹⁵ Under the current experimental conditions porphyrins Zn1TH3A and Zn1TPA3A shows highest absorbance on TiO₂.

Figure 2. UV‒visible spectra of porphyrin sensitizers Zn1T3A, Zn1U3A, Zn1S3A, Zn1TH3A, Zn1TC3A, Zn1BC3A, and Zn1TPA3A on TiO₂.

147 4.5 Electrochemical properties and energy levels

To determine the first oxidation potential (EOX) of porphyrins cyclic voltammetry measurements containing 0.1 M tetrabutylammonium hexafluorophosphate ([TBA]PF6) were performed in degassed THF. Figure 3 shows representative cyclic voltammograms of Zn1U3A, Zn1TC3A, Zn1BC3A and Zn1TPA3A. All of the porphyrins show well resolved and reversible wave for the first oxidation couple. The values of oxidation potential as depicted in

Figure 3. Cyclic voltammogram of Zn1U3A, Zn1TC3A, Zn1BC3A and Zn1TPA3A.

Table 4 indicate that the peripheral substituents on meso positions moderately alter the oxidation potentials. The lowest first oxidation potential 1.05 V has been observed for Zn1U3A. Porphyrin Zn1TPA3A shows two successive oxidations at 0.88 V and 1.05 V for the oxidation processes at triphenylamino group and porphyrin core, respectively.¹⁶ The oxidation potentials of all studied porphyrins dyes are greater than that of I^‒/I3‒ couple assuring the regeneration of oxidized dyes. The excited-state oxidation potentials (EOX*) are obtained from the equation EOX* = Eox1 ‒ E0–0 where Eox1 is the first oxidation potential of a porphyrin dye and E0–0 is the zero-zero excitation energy obtained from the intersection of the absorption and emission spectra. The calculated EOX* energy levels of these porphyrins depicted in Table 4 are more negative than the conduction band of TiO2 (‒0.50 V vs. NHE), exhibiting the sufficient driving force for electron injection from the excited state of the dye to the CB of

148 TiO₂. The systematic energy level diagrams of studied porphyrins are shown in Figure 4 demonstrating feasibility of dye regeneration with electrolyte and electron injection to conduction band.⁶ It is observed from Figure 4 that electron injection and dye regeneration is more favorable in the cases of Zn1U3A and Zn1TPA3A which exhibit a positive impact contributing to their higher conversion efficiencies of η = 3.45% and η = 5.26%, respectively.

(vide infra)

Figure 4. Energy level diagram of studied porphyrin dyes 4.6 ATR-FTIR Spectroscopy

We have used ATR-FTIR spectroscopy to determine the number and mode of meso carboxyphenyl groups attaching on TiO2. To compare the differences, ATR-FTIR spectra of neat porphyrin samples were contrasted with spectra of porphyrins adsorbed on TiO2.^5,16-19 The representative comparative spectra of (a) Zn1U3A compared with Zn1U3A/TiO2,(b) 1S3A compared with Zn1S3A/TiO2, (c) Zn1TC3A compared with Zn1TC3A/TiO2, and (d) Zn1TPA3A compared with Zn1TPA3A/TiO2 are shown in Figure 5. The ATR-FTIR spectra of Zn1U3A, Zn1S3A, Zn1TC3A and Zn1TPA3A show strong ν(C=O) stretching at 1685, 1680, 1722 and 1686 cm^-1 respectively, whereas νsym(COO¯) and νasym(COO¯) stretching are expected to be observed at ~1400 cm^-1 and 1600 cm^-1, respectively. In the comparative spectra of Zn1S3A/TiO2, Zn1U3A/TiO2, ν(C=O) stretching peak decreased slightly with noticed increase in the stretching peaks of ν_sym(COO¯) and ν_asym(COO¯) at ~1400 cm^-1 and 1600 cm

-1, respectively. The observations noted above indicate that two out of three p-carboxyphenyl group on meso-position of porphyrin are utilized for bonding on TiO₂. Same observations are noticed in the remaining porphyrins Zn1T3A, Zn1TH3A, and Zn1BC3A(SI Figure 11).

149 Figure 5. ATR-FTIR spectra of porphyrins Zn1U3A and Zn1U3A/TiO2, (b) Zn1S3A and

Zn1S3A/TiO₂, ATR-FTIR spectra of porphyrins on TiO₂ are normalized for comparison.

Although our current results using ATR-FTIR as the probe cannot distinguish two possible dual anchoring modes, 5,10-cis anchoring and 5,15-trans anchoring, it is expected that cis anchoring will be much stable and more likely considering a relative flat TiO₂ surface relative to the size of the porphyrin dyes. Based on the observed ATR-FTIR results, the likely attaching mode of the studied three p-carboxyphenyl groups possessing porphyrins is shown in Figure 6.

Figure 6. Possible modes of attachment of studied porphyrins on TiO₂ are illustrated. For demonstration purpose only, relative sizes of molecules and nanoparticles are not correlated in real dimensions.

150 4.7 Density Functional Theory Calculations

To gain insight of electronic structures for the frontier orbitals of the porphyrins, the density functional theory calculations of studied porphyrins were done at B3LYP/6-31G level. Figure 7 illustrates the electron density distributions of studied porphyrins in their respective HOMO-1, HOMO, LUMO, and LUMO+1 molecular orbitals. Analyzing the orbitals in detail, for HOMO-1 and HOMO orbitals, the electron density distributions for porphyrin dyes without an electron donating amino moiety are distinctively different from those with an amino moiety.

While electron density mainly localizes on the porphyrin core for HOMO-1 and HOMO orbitals in Zn1T3A, Zn1U3A, Zn1S3A and Zn1TH3A, majority of electron density localizes on TPA or carbazole and its linking phenylene or thiophene in Zn1TC1A, Zn1BC3A and Zn1TPA3A. Noticeably, electron density localizes near exclusively on electron donating meso substituent for HOMO orbitals in the cases of Zn1TC1A, Zn1BC3A and Zn1TPA3A. While porphyrin core is still the main electron population site, in both LUMO and LUMO+1 orbitals, significant amount of electron density extensively populates toward the meso carboxyphenyl substituents. Likely owing to the different molecular symmetry, Zn1T3A, Zn1U3A, Zn1BC3A, and Zn1TPA3A have electron density localizing at 5- and 15-carboxyphenyl for LUMO orbitals and at 10-carboxylphenyl for LUMO+1 but the opposite distribution is observed for the rest zinc porphyrin dyes. Given that our ATR-IR analyses demonstrate a cis two arms anchoring with one dangling carboxyphenyl groups for these tricarboxyphenyl containing porphyrin dyes, it is likely that LUMO and LUMO+1 work synchronize to pump electron density into the conduction band of TiO₂. Nevertheless, it is clearly evidenced from the electron density distribution that the porphyrin dyes containing an electron-donating meso substituent provides better charge separation which slows down the charge recombination and results a more efficient electron injection.

151

Dye HOMO-1 HOMO LUMO LUMO+1

Zn1T3A

Zn1U3A

Zn1S3A

Zn1TH3A

Zn1TC3A

Zn1BC3A

Zn1TPA3A

Figure 7. The molecular orbital diagrams of studied porphyrins obtained from DFT calculations

152 4.8 Photovoltaic measurements

The photovoltaic performances of porphyrin sensitizers are strongly influenced by various factors such as immersion solvent, immersion time, CDCA amount, and the type of electrolyte, etc. The performance of studied dyes has been optimized systematically by analyzing the effect of immersion solvents and CDCA equivalence on the representative dye, Zn1TPA3A.

Through optimizing the DSSC conditions, we summarize the ideal conditions and apply to all efficiency measurements. The efficiency data from the solvent and CDCA ratio screening are relatively low because of inferior quality of TiO2 photoanode films but we were able to observe meanful trends and choose the best conditions.

To obtain optimizing conditions for efficiency measurements, ideal solvent and CDCA ratio for DSSCs using 1TPA3A as the sensitizing dye have been examined. All conditions have been repeated four times using different DSSC assemblies and the conclusion was drawn from the comparisons on the best one from four repeats. Table 5 and 6 represent the photovoltaic measurement data using different solvent and CDCA ratio.

Table 5. Photovoltaic measurement data of Zn1TPA3A in studied solvents.

Solvent J_SC (mA cm^-2) V_OC (V) FF (%) η (%)

153

Table 6. Photovoltaic measurement data of Zn1TPA3A in with studied CDCA ratio in THF CDCA:dye ratio JSC (mA cm^-2) VOC (V) FF (%) η (%)

As shown in Table 5, the efficiency of Zn1TPA3A is checked in different immersion solvents such as THF, THF/EtOH (2:8, v/v), THF/MeOH (2:8, v/v), and THF/toluene (2:8, v/v). Polar solvent THF alone or a micture of THF with MeOH, EtOH, or toluene were selected for screening. Zn1TPA3A with a concentration of 2 × 10^-4 M was immersed in studied solvents for 0.5 h at 50^oC. The condition of 0.5 h absorption was adopted directly from our immersing time

154 screening on dual anchoring porphyrins described in last chapter. Comparable efficiency 2.33% and 2.47% has been observed for THF and THF/toulene (2:8, v/v). For the reason of better solubility, THF has been used for the next CDCA ratio optimizations.

The conversion efficiency of Zn1TPA3A has been examined using 2 × 10^-4 M Zn1TPA3A in THF with the addition of 1, 2, 3, 4 and 10 equiv. of CDCA as the immersion solution. In this particular case, the film obtained from solution containing 1 equiv. of CDCA is found to be the most suitable condition for preventing dye aggregation and demonstrates the best performance with a 2.87% efficiency. In brief, the optimized conditions revealed for Zn1TPA3A are 2 × 10

-4 M dye with 2 × 10^-4 M CDCA immersed in THF at 50^oC for 0.5 h. The photovoltaic performances of all multi-anchoring porphyrin dyes have been checked with these optimized conditions from devices using Zn1TPA3A as the sensitizer and are summarized in Table 7.

Table 7. Photovoltaic properties of studied porphyrins

Depending on substituents on their meso-positions, the performance of porphyrins could be classified into four groups. First, the Zn1T3A and Zn1U3A porphyrins having no electron donating substituent on their meso-positions exhibit moderate efficiencies of 4.12% and 3.45%, respectively. Secondly, the Zn1S3A and Zn1TH3A dyes with the presence of a heavier sulfur atom show diverse efficiencies of 2.44% and 5.36%, respectively. Thirdly, the Zn1TC3A and Zn1BC3A bearing the same carbazole as electron donating unit but have different connecting units, thienyl and phenylene, show comparably efficiencies of 4.49% and 4.40%, respectively.

Finally, the porphyrin dye Zn1TPA3A having a strong electron donating triarylamino group

155 demonstrates 5.26% efficiency. The unexpected high efficiency of device using Zn1TH3A as the dye might be attributed to its higher absorbance, broader absorption wavelengths, and decreases dye aggregation. The high efficiency of Zn1TPA3A is mostly because of its effective charge separation. The large phenyl groups on amino nitrogen might also shield the approaching of electrolyte toward TiO₂ layer and retard the back electron transfer.

Figure 8. I‒V curves of studied porphyrins sensitized solar cells under one sun illumination.

The IPCE spectra of studied porphyrins are shown in Figure 9. The IPCE values of studied porphyrins show strong correlation to its meso-substituents. The IPCEmax at Soret band decreased in the order of Zn1TPA3A (53%) > Zn1TH3A (48%) > Zn1BC3A (47%) >

Zn1TC3A (46%) > Zn1T3A (45%) >Zn1U3A (41%) > Zn1S3A (40%). The IPCEmax observed here are roughly consistent with the efficiency performance. The higher IPCEmax at Soret band observed for Zn1TPA3A and Zn1TH3A have the highest efficiencies whereas lowest IPCEmax

at Soret band is observed for Zn1S3A having lowest efficiency. The dye density of porphyrins on TiO2 films are examined and summarized in Table 7. There is no obvious correlation between dye densities and efficiencies of porphyrins. Despite of low dye densities Zn1TPA3A and Zn1TH3A are able to give better performance because of higher J_SC and V_OC. In contrast, Zn1T3A, Zn1TC3A and Zn1BC3A have relative higher dye densities but only exhibit moderate efficiencies. It is noticed that regular desorption solvent, 0.1 M KOH(aq) in THF (2:8, v/v) failed to completely desorb these three carboxyphenyl substituted sensitizing dyes with even a 24 h exposure. It requires 0.2 M KOH(aq) in THF (2:8, v/v) and a 1 h exposure time to completely desorb the dye attached on TiO₂ surface. The use of high concentration of

156 base to desorb the dye is one of the additional supportings to prove the higher stability of dual anchoring group possessing porphyrins than single anchoring group possessing porphyrins.

Figure 9. The IPCE curves of studied porphyrin dyes 4.9 Stability study

For the commercialization of DSSCs, a long term stability is required along with a high efficiency. We have used a straightforward method reported in literature to examine the stability of studied dyes.^20,21 To obtain the stability informaiton, absorption spectra of three carboxyphenyl groups possessing porphyrins Zn1TPA3A, Zn1BC3A, and Zn1TC3A are compared with a single anchoring group possessing porphyrin Zn3TPA1A after irradiation.

The TiO2 plates with Zn1TPA3A, Zn1BC3A, Zn1TC3A, and Zn3TPA1A as the sensitizer were irradiated under standard one sun illumination and monitored the absorption curves after of irradiated for 0, 5, and 30 min to give the plot as shown in Figure 10. The observed results suggest that the absorbances of dyes decrease in various rate but without any shifting in the wavelength of maximum absorbance. The decrease in the absorbance is significantly higher for Zn3TPA1A compare to Zn1TPA3A, Zn1BC3A, and Zn1TC3A suggesting that the porphyrins possessing multi-anchoring group are more stable compare to porphyrins possessing single anchoring group.

157 Figure 10. Absorption spectra of Zn1TCA3A, Zn1BC3A, Zn1TPA3A, and Zn3TPA1A

adsorbed on TiO₂ films irradiated for 0 min, 5 min, and 30 min.