Photoelectrocatalytic Synthesis of Hydrogen Peroxide by Molecular Copper‐Porphyrin Supported on Titanium Dioxide Nanotubes

Abstract We report on a self‐assembled system comprising a molecular copper‐porphyrin photoelectrocatalyst, 5‐(4‐carboxy‐phenyl)‐10,15,20‐triphenylporphyrinatocopper(II) (CuTPP‐COOH), covalently bound to self‐organized, anodic titania nanotube arrays (TiO2 NTs) for photoelectrochemical reduction of oxygen. Visible light irradiation of the porphyrin‐covered TiO2 NTs under cathodic polarization up to −0.3 V vs. Normal hydrogen electrode (NHE) photocatalytically produces H2O2 in pH neutral electrolyte, at room temperature and without need of sacrificial electron donors. The formation of H2O2 upon irradiation is proven and quantified by direct colorimetric detection using 4‐nitrophenyl boronic acid (p‐NPBA) as a reactant. This simple approach for the attachment of a small molecular catalyst to TiO2 NTs may ultimately allow for the preparation of a low‐cost H2O2 evolving cathode for efficient photoelectrochemical energy storage under ambient conditions.

We report on as elf-assembled system comprising am olecular copper-porphyrin photoelectrocatalyst, 5-(4-carboxy-phenyl)-10,15,20-triphenylporphyrinatocopper(II) (CuTPP-COOH), covalently bound to self-organized, anodic titania nanotube arrays (TiO 2 NTs) for photoelectrochemical reduction of oxygen. Visible light irradiation of the porphyrin-covered TiO 2 NTsu nder cathodic polarization up to À0.3 Vv s. Normal hydrogen electrode (NHE) photocatalytically produces H 2 O 2 in pH neutral electrolyte, at room temperature and withoutn eed of sacrificial electron donors. The formation of H 2 O 2 upon irradiation is provena nd quantified by direct colorimetricd etection using 4nitrophenyl boronic acid (p-NPBA) as ar eactant. This simple approachf or the attachment of as mall molecular catalyst to TiO 2 NTsm ay ultimately allow for the preparation of al ow-cost H 2 O 2 evolving cathode for efficient photoelectrochemical energy storageunder ambient conditions. Two-electron oxygen reduction reaction (ORR) of dissolved oxygen (O 2 )i nw ater leads to formation of hydrogen peroxide (H 2 O 2 )w hich is av ersatile, high energy product, [1] capable of participating in numerousf urtherr edox reactions and is an active speciesi naplethora of biological processes. [2] Solardriven H 2 O 2 formation has been proposed for chemical energy storage. [1,[3][4][5] However, the widely used anthraquinone process for the formation of H 2 O 2 is known to be energy intensive. [6] For many decades, researchers have tried to address this issue and tackle the problem by introducing metal catalysts, [7][8][9][10][11][12] core-shell structures,m etal oxides, metal chalcogenides etc. [13][14][15] Additionally,p hotocatalytic reduction of O 2 to H 2 O 2 by inorganic semiconductors (e.g. are ZnO, CdS and TiO 2 )a nd organometallic complexes has been reported. [14,[16][17][18] Recently, metal-free carbon-based catalysts has been the focus for (photo)electrochemical reduction of dissolved O 2 .T his class mainly includes graphitic carbon nitrides (g-C 3 N 4 )a nd organic pigments. [19][20][21] However,a lmosta ll of these reactions require either acidic or basic conditions which maked aily applications challenging. Althought here are af ew examples, [22] the search for ac atalystw hich works under mild pH conditions is still in progress.
Here, we present ap hotoelectrode consisting of ap orphyrin derivative, namely CuTPP-COOH (Figure 1a), coated on TiO 2 nanotubes (NT) (TiO 2 NTs/CuTPP-COOH) for the reduction of O 2 to H 2 O 2 .T he introduction of ac arboxyl group enablest he attachment of the photoactive porphyrin onto the nanostructured TiO 2 NTs. [23] CuTPP-COOH wasc hosen owing to ease of its synthesis as well as the appropriate energy levels to reduce O 2 .W eh ave also utilized ZnTPP-COOHf or the same reaction; however the stability of this material was inferior.T he high surface area of TiO 2 NTs [24] increases the number of potential catalyticallya ctive sites .I na ddition, the amorphous structure of TiO 2 NTsh elps to anchor the CuTPP-COOH through the À COOH functional group. The reactiont akes place at neutral pH and ambient temperature (22 8C). By applying moderate negative potentials between 0.0 Va nd À0.3 Vv s. NHE (normal hydrogen electrode) andu pon photoexcitation of CuTPP-COOH at l > 395 nm, an exciton (electron-hole pair) is initially formed, as illustrated in Figure 1b.S ubsequently,t he newly created hole residing in the valence band of CuTPP-COOH is recombined with an electron suppliedf rom the externalc ircuit, while the electron in the conduction band is capable of reducing the dissolvedO 2 in water to H 2 O 2 .
The electrochemical behavior of the TiO 2 NTs/CuTPP-COOH photoelectrodes under applied potential in Ar-and O 2 -saturated conditions can be seen in Figure 2. In the absence of O 2 , the illumination led to no observable increaseinc urrent with a currentd ensity (j)o fa pproximately 1.3 mAcm À2 at À0.3 Vv s. NHE. However,u nder O 2 saturationa nd upon light illumination, the current value increased around 4f old and reached approximately 13 mAcm À2 at À0.3 V, signaling the reduction of dissolvedO 2 .
After chronoamperometry experiments,as eries of constant potentiale lectrolysis experiments were conducted to quantify the formation of H 2 O 2 .O ne of the reasonable ways for direct detection of H 2 O 2 ,i sa ni ndirect spectrophotometricm ethod for the quantification of the product, relying on as toichiometric reaction of arylboronic acids with newly generated H 2 O 2 under mild basic conditions to yield the respective photoactive phenolates. [26,27] In this work, p-nitrophenylboronic acid (p-NPBA) was used at pH 9, which was converted upon reaction with H 2 O 2 into p-nitrophenol (p-NP), for which absorption could be observed by using UV/Vis spectrophotometrya t 405 nm. Ac alibration curve forq uantitative determination of H 2 O 2 in ac oncentration range between 0.5 mm and 20 mm is shown in Figure 3. The detailed procedure for the preparation of the standards olutions can be found in the Supporting Information.
After each constant potentiale lectrolysis measurement, an aliquot of 100 mLw as pipetted from the electrolyte solution and then transferredi nto av ial containing the p-NPBAa nd carbonate buffer.A mounts of H 2 O 2 ,r eflected by those of newly formed p-NP,between 1.9 mm and 3.9 mm were observed     [28] and g-C 3 N 4 (4.25 mgmg cat À1 h À1 ) [19,29] Correspondingc ontrol experiments in which the O 2 -saturated solution was measured in the dark did not yield any detectable amount of H 2 O 2 .
To further evaluate the electrochemical characteristics of the O 2 reductiono nt he TiO 2 NTs/CuTPP-COOH photoelectrodes, we conducted potential-dependent electrochemical impedance spectroscopy (PEIS, Figure 4). Symbols represent experimental data and lines the best fits. The spectra were collected in the potentialr ange between 0.2 and À0.3 Vw ith as tep size of 0.1 V. Each potentialw as kept constantf or 10 min to ensure steady state conditions before the impedancem easurement, ranging from 100 kHz to 20 mHz, with ap eak amplitude of AE 10 mV.M easurements were performedu nder illumination in the Ar-and O 2 -saturated electrolyte solution containing 0.1 m Na 2 SO 4 .
Detailed analysis of Nyquistplots under O 2 saturation reveals the presence of three, not fully developeds emi circles ( Figure S5, Supporting Information) The first semi-circle (I) at high frequencies between 4.5 kHz and 200 Hz waso bservable in all spectra and may describe the interfacialT iO 2 /CuTPP-COOH charget ransfer.Asecond semi-circle (II) at medium frequencies between 65 Hz and 1.4 Hz is also observable in all spectra and may represent the resistance for electron transport along the TiO 2 NTsa nd the corresponding surface capacitance. [30] The development of an additional semi-circle (III) at potentials below 0.0 Vand lower frequenciesb etween 0.94 Hz and 20 mHz is observable only if the electrolyte is saturated with O 2 .T his may correspond to the charge transfer resistance of the O 2 reduction reaction. At wo-step reactionp rocess is expectedt ob et he reason for the occurrence of semi-circle (III), for example an intermediate state that is involved. [31] For furtherq uantification of the measuredE IS data, correspondinge lectronic elements were determined by fitting the experimental spectra to the proposed equivalent circuit depicted in Figure 4d.T he proposed equivalent circuit is am odified version of the equivalentc ircuit introduces by Kçleli et.al.f or CO 2 reduction on polyaniline-coated electrodes. [31] An additional R/C element at high frequenciesh as been added to account for the TiO 2 /CuTPP-COOH interface at the nanostructured support-electrodes,p artly adopted from the transmission line model originally introduced for nanostructured TiO 2 hybrid solar cells. [30] The real capacitors C nt and C r are modeled with CPEs to account for the non-ideal behavior (i.e. depressed semi-circle) of the capacitive part at medium and low frequencies. [32] The finite length Warburg impedance (ZW) is used to describe the transport phenomena of O 2 into the porphyrin film and the transport of reduction products out of the film. The parallel configuration of the R tr /CPE nt and R r / CPE r elementsm ay be justifiedo wing to the inhomogeneity (porosity) of the CuTPP-COOH covered TiO 2 NTs. From the EIS data we concluded that the ohmic resistance of the electrolyte solution (R s )i sa lmostc onstant at all potentials,f luctuating slightly between 17 and 20 W.T he R f (interfacial electron charge transfer resistance) is relativelyh igh at positive potentials with 111.6 kW at 0.2 Va nd decreases significantly to 3.1 kW at À0.3 V. This suggests enhanced charge transfer over the TiO 2 /CuTPP-COOH interface with applied negative bias. R tr , which describes the resistance for electron transport along the TiO 2 NTs, decreases only slightly with the applied potential from initially 2.4 kW at 0.2 Vt o1 58 W at À0.3 V. This characteristic behavior of R tr is expected for relatively highly doped nanotubes suggesting as mall variation of the carrierd ensity with bias (unless full depletion is obtained). [30] The charge transfer resistance related to the O 2 reduction reaction (R r ) could not be determined for positive potentials of 0.2 and 0.1 V, respectively,s ince the correspondings emi-circle was not developed in the measured frequency limit (20 mHz). Therefore, it was sufficient to fit the electrochemical impedance spectroscopy data at 0.2 Vand 0.1 Vw ithoutthe electronic elements used for describing the O 2 diffusion and reduction reaction (ZW, R r and CPE r ). At 0.0 Vthe occurrence of semi-circle (III) becomes notable and R r was determined with 214 kW. R r then decreaseds ignificantly to about2 .3 kW at À0.3 V, suggesting enhanced O 2 reduction at lower potentials. This is congruent with the observed characteristics from chronoamperometry experiments (Figure2). Overall,t he authors are fully aware that the proposed equivalent circuit mayn ot cope with the complexity of the investigated system and was introduced only as an initial attempt to describe the measured EIS data. Also one has to point out that the EIS measurements were not performed under diffusion controlled conditions (i.e. by ar otating disk electrode), rendering its interpretation challenging. Nonetheless, the proposed equivalent circuit demonstrated good fitting congruency in the Nyquista nd Bode plots (Figure 4b and Figure S6, SupportingI nformation), with, fore xample, am ean square deviation (X 2 /Z) of 0.3 %f or the EIS data recorded at À0.3 Vu nder O 2 saturation (Figure 4b). Ad etailed summary of all fitting parameters and their corresponding mean square deviations is given in Ta ble S1, and ac omparison of all impedance measurements under Ar and O 2 saturation is shown in Figure S7, in the Supporting Information.
In summary,w ehave demonstrated an ovel photocathode capable of reducing dissolved O 2 to H 2 O 2 with evolutionr ates ranging between2and 13 mg H2O2 mg cat -1 h -1 .B ya ttaching the CuTPP-COOH catalyst onto TiO 2 NTst hrough its carboxyl group, we created ah eterogeneousm olecular catalyst where the formed photocathodei sc onvenientf or use in aqueous mediuma nd inherits as ignificantly higher surface area over planar electrodes owing to self-organized nanostructured TiO 2 NTs. The TiO 2 nanostructures were previously used as catalytic moieties together with sensitizers such as porphyrins and phthalocyanines. However,t he use of such molecular porphyrins as photoelectrocatalysts is not common.W eh ave also demonstrated that our system is capable of driving the aforementioned reactionu nder pH neutral conditions which is expected to reduce the technical complications originating from high acidic or alkaline media.

Experimental Section
Experimental Details such as the synthesis and the details of the electrochemical setup as well as electrochemical impedance spectroscopy can be found in Supporting Information.