Letter pubs.acs.org/JPCL
Photochemical Energy Storage and Electrochemically Triggered Energy Release in the Norbornadiene−Quadricyclane System: UV Photochemistry and IR Spectroelectrochemistry in a Combined Experiment Olaf Brummel,†,§ Fabian Waidhas,†,§ Udo Bauer,† Yanlin Wu,‡ Sebastian Bochmann,‡ Hans-Peter Steinrück,† Christian Papp,† Julien Bachmann,‡ and Jörg Libuda*,† †
Lehrstuhl für Physikalische Chemie II, Friedrich-Alexander-Universität Erlangen-Nürnberg, Egerlandstrasse 3, 91058 Erlangen, Deutschland ‡ Lehrstuhl für Anorganische und Allgemeine Chemie, Friedrich-Alexander-Universität Erlangen-Nürnberg, Egerlandstrasse 1, 91058 Erlangen, Deutschland S Supporting Information *
ABSTRACT: The two valence isomers norbornadiene (NBD) and quadricyclane (QC) enable solar energy storage in a single molecule system. We present a new photoelectrochemical infrared reflection absorption spectroscopy (PEC-IRRAS) experiment, which allows monitoring of the complete energy storage and release cycle by in situ vibrational spectroscopy. Both processes were investigated, the photochemical conversion from NBD to QC using the photosensitizer 4,4′-bis(dimethylamino)benzophenone (Michler’s ketone, MK) and the electrochemically triggered cycloreversion from QC to NBD. Photochemical conversion was obtained with characteristic conversion times on the order of 500 ms. All experiments were performed under full potential control in a thinlayer configuration with a Pt(111) working electrode. The vibrational spectra of NBD, QC, and MK were analyzed in the fingerprint region, permitting quantitative analysis of the spectroscopic data. We determined selectivities for both the photochemical conversion and the electrochemical cycloreversion and identified the critical steps that limit the reversibility of the storage cycle.
T
The release reaction can be initiated by oxidation, so that it proceeds via the radical cations QC•+ and NBD•+ as intermediates.26,28,29 Several strategies have been explored to trigger the energy release, for instance, involving the use of homogeneous metal complex catalysts,30−32 thermal activation33−35 or heterogeneous catalysts.36 The most common methods make use of mild oxidants,26,28,29 which implies, however, a poor level of control in most cases. Recently, we demonstrated that it is also possible to trigger the cycloreversion of QC to NBD electrochemically and showed that the reaction rate can be controlled via the electrode potential.37 The electrochemically triggered release can be understood as a first step toward a visionary concept in which the chemical energy could be released in the form of electrical energy.37 The latter would enable the design of an energy-storing solar cell, that is, a simple chemical system that combines conversion of solar energy to electricity and storage in a single device.
he conversion of solar energy can be achieved either via a photovoltaic or a photo(electro)chemical route.1,2 While the photovoltaic route permits efficient conversion to electrical energy, it does not allow for direct energy storage. In sharp contrast to this, photo(electro)chemical solar energy conversion yields the formation of a storable fuel.3 However, the conversion of these chemical fuels typically involves complex multistep reactions, often including several molecular species, rearrangement of multiple bonds, and/or the transfer of several electrons. Naturally, such complex reaction networks are prone to undesired side reactions and are difficult to control.1 An alternative and simpler concept is the use of intramolecular processes for energy storage such as valence isomerization. Such processes allow for energy conversion and storage in a single molecule and avoid intermolecular reactions.4−9 The most prominent example is the conversion of norbornadiene (NBD) to its metastable, strained, and energyrich isomer quadricyclane (QC).10−14 NBD and its derivatives4,15−19 are converted to QC photochemically via a tripletstate mechanism.20 Typically, the reaction is initiated by photosensitizers such as, for example, Michler’s ketone (MK), acetophenone, or others at wavelengths beyond 300 nm.20−24 Upon cycloreversion of QC back to NBD, the system can release up to 100 kJ/mol, which defines it as a solar fuel.11,25−27 © XXXX American Chemical Society
Received: April 24, 2017 Accepted: May 31, 2017 Published: May 31, 2017 2819
DOI: 10.1021/acs.jpclett.7b00995 J. Phys. Chem. Lett. 2017, 8, 2819−2825
Letter
The Journal of Physical Chemistry Letters Toward the design of such a system, an in-depth understanding of the underlying reaction mechanisms and kinetics is essential. In our previous work, we showed that it is possible to follow the electrochemically triggered energy release in situ by electrochemical IR reflection absorption spectroscopy (ECIRRAS). In this work, we go one step further. We present a new in situ spectroscopy experiment that allows for monitoring of the complete storage cycle, that is, the photochemical conversion of NBD to QC and the electrochemically triggered cycloreversion of QC to NBD. To this end, we combine ECIRRAS with a photochemical setup in a single experiment. The experimental setup can be easily implemented as an add-on to most electrochemical IR spectroscopy systems and allows for studying of a wide range of photoswitchable systems in contact with well-defined electrodes. To the best of our knowledge, this is the first publication that describes such a photoelectrochemical infrared reflection absorption spectroscopy (PEC-IRRAS) experiment, and it is also the first in situ spectroscopy study that addresses both the energy storage and the electrochemically triggered energy release in the NBD/QC system. In the present study, the experiment allows us to probe the reversibility of the storage cycle and to identify critical steps in the process. The efficiency of the experiment marks a breakthrough in molecular photoelectrochemistry at welldefined electrodes and will foster future development of new molecular photoelectrochemical energy storage systems. The schematic layout of the PEC-IRRAS experiment is shown in Figure 1. The setup is based on the concept of electrochemical IR spectroscopy in a thin-layer configuration.38,39 Specifically, we use a home-built electrochemical PTFE cell with a three-electrode setup and a Pt(111) single crystal as the working electrode (see previous publications37,40,41 for details). During the measurements, the crystal is pressed against the flat side of a hemispherical CaF2 window. Thereby, a thin electrolyte layer with a thickness of a few micrometers is formed,42 which contains the dissolved NBD. This thin layer is sufficiently transparent in the IR range. The IR beam is reflected at the working electrode with an angle of incidence of 60° (with respect to the sample normal) and, therefore, propagates twice through the layer (see Figure 1a). Note that diffusion in the thin layer is largely decoupled from the bulk solution.43 The IR spectra are recorded as difference spectra upon alteration of the electrode potential. For further details about the technique, we refer to the literature.38,39,42 For the present experiment, the vacuum FTIR spectrometer (Bruker Vertex 80v) with the EC-IRRAS setup described above was equipped with an additional UV source (see Figure 1a). To this end, a high-power UV LED (Seoul Viosys) on a home-built Cu cooling stage was mounted inside of the vacuum optics directly underneath of a CaF2 hemisphere. Note that the CaF2 hemisphere is transparent both in the UV and in the IR range (see Figure 1b). The UV LED provides a radiant flux of 2.35 W at a wavelength of 365 nm. The latter fits the absorption characteristics of MK (4,4′-bis(dimethylamino)-benzophenone) perfectly, which exhibits an absorption maximum at 366 nm.22 Noteworthy, MK is also among the best studied photosensitizers for the NBD/QC system.20,34,44 With this setup, we estimate that approximately 32% of the emitted UV light is focused onto the surface of the working electrode with an area of 0.785 cm2, which yields a photon flux density of 1.8 × 1018 photons·s−1·cm−1 (see the Supporting Information). Thus, we obtain an outstandingly efficient photoconversion setup, as we will demonstrate below. As the MCT detector of
Figure 1. (a) Schematic representation of the combined PEC-IRRAS experiment; (b) design of the IR optics, UV source, and electrochemical cell in the measurement compartment of the FTIR spectrometer; (c) schematic representation of the overall experimental setup.
the FTIR spectrometer is sensitive to UV light, it was protected by an automatic shutter during UV irradiation. The potentiostat (Gamry Reference 600), the UV LED power supply, and the UV shutter were triggered by the FTIR spectrometer in an automatized data acquisition procedure (see Figure 1c). Further details on the setup and the experimental procedure are provided in the Experimental Section and the Supporting Information. The results of a photoelectrochemical in situ spectroscopy experiment on the conversion of NBD to QC are displayed in Figure 2. A 0.1 M solution of NBD in acetonitrile with 20 mM MK as the photosensitizer and 0.1 M Bu4NClO4 as the supporting electrolyte was irradiated at an electrode potential of −0.08 Vfc (note that in this work all potentials Vfc are referred to the ferrocene couple; see the Experimental Section). At this potential, no cycloreversion from QC to NBD is expected. All IR spectra were recorded using nonpolarized light. The 2820
DOI: 10.1021/acs.jpclett.7b00995 J. Phys. Chem. Lett. 2017, 8, 2819−2825
Letter
The Journal of Physical Chemistry Letters
Figure 2. In situ IR spectroscopy during photochemical conversion of NBD to QC (0.1 M NBD with 20 mM MK and 0.1 M Bu4NClO4 in acetonitrile) at an electrode potential of −0.08 Vfc; (a) schematic representation of the experimental procedure applied; (b) structures of NBD, QC, and MK; (c) reference transmission spectra of a mixture of NBD (blue) and QC (orange) as well as pure MK (green); (d) IR difference spectra taken before (red) and after UV irradiation (black, pulse length 250 ms); (e) changes of the concentration of QC and NBD as a function of UV irradiation time; (f) selectivity for QC as a function of UV irradiation time.
reference spectrum was measured before UV irradiation, and the stability of the system was tested by measuring another IR spectrum before starting the UV irradiation. In this spectrum (Figure 2d, red), no changes could be observed, showing that the system is indeed stable before starting the kinetic measurements. Subsequently, the sample was exposed to UV pulses of 250 ms duration, and IR spectra were recorded after each pulse. The corresponding procedure is illustrated in Figure 2a. To assign the IR bands, we also recorded reference spectra in transmission of a mixture of NBD and QC and of pure MK. The data are presented in Figure 2c. The IR bands of NBD and QC are in perfect agreement with our previous work,36,37 and the bands of MK are in accordance with literature.45 The assignments to different species and vibrational modes are summarized in Table 1. In the following, the IR bands attributed to NBD, QC, and MK are marked with the color code given in Figure 2b (NBD: blue; QC: orange; MK: green). During irradiation (see Figure 2d), positive bands of NBD appear at 1543, 1314, and 1206 cm−1 and positive bands of MK appear at 1322, 1289, 1183, and 1171 cm−1. At 1230 cm−1, a positive band is observed that contains contributions from both NBD and MK. Simultaneously, negative bands of QC arise at 1257 and 1240 cm−1. Note that all IR spectra are difference spectra using the system before irradiation as a reference. Therefore, positive bands correspond to species that are consumed, whereas negative bands correspond to species that are formed. Accordingly, the bands observed indicate
Table 1. Experimentally Observed IR Band Positions for NBD, QC, and MK and Calculated Extinction Coefficients νexp [cm−1]
species
description of vibrational modesa
εExp [L·mol−1·cm−1]
1543
NBD
36
1322 1314
MK NBD
1289 1257
MK QC
1240
QC
1230 1227
MK NBD
1206
NBD
1183 1171
MK MK
asymmetric stretching vibration of the two double bonds36,37 stretching vibration of N−Cring45 symmetric C3−C4−C5 deformation/ stretching mode coupled with its C2− C1−C6 counterpart36,37 stretching vibration C−Cring45 asymmetric coupled C7−C1 and C7−C4 stretching modes and symmetric C−C mode of two cyclopropane rings36,37 asymmetric coupled C2−C3 and C5−C6 stretching modes36,37 stretching vibration Carom45 asymmetric stretching modes of C7−C1 and C7−C436,37 asymmetric deformation/stretching modes of C3−C4−C5 and C2−C1− C636,37 C−Hring deformation mode45 C−Hring deformation mode
408 95
445 9.7
93 76
8.5
533 318
a
See Figure 2b for the numbering of the carbon atoms in NBD and QC.
conversion of NBD to QC. Furthermore, the positive bands of MK show that the photosensitizer is not fully stable but 2821
DOI: 10.1021/acs.jpclett.7b00995 J. Phys. Chem. Lett. 2017, 8, 2819−2825
Letter
The Journal of Physical Chemistry Letters partially consumed in the reaction. The experiment demonstrates that we can indeed follow the photochemical conversion of NBD to QC and unwanted side reactions in situ in the PECIRRAS experiment. For quantitative analysis, we determined the extinction coefficients of selected NBD and QC bands. To that end, a NBD solution was partially converted to QC in a separate photochemical experiment, and the resulting concentrations were quantified by NMR using p-dicyanobenzene as an internal standard (see the Supporting Information for details on the procedure). The extinction coefficients obtained are summarized in Table 1. Using these data, we calculated the change of the concentration of NBD and QC as a function of irradiation time. Specifically, we used the IR bands of NBD at 1543 cm−1 and of QC at 1240 cm−1, which do not overlap with each other or with the bands of MK (see the Supporting Information for details). The resulting data are displayed in Figure 2e and demonstrate the high efficiency of the photoconversion setup. An exponential fit yields a characteristic photoconversion time τ of 560 ms for the consumption of NBD and a final conversion level of 80% (see the Supporting Information). The fact that the NBD conversion levels off at this value is attributed to cycloreversion catalyzed by Pt,36 spatial inhomogeneity of the UV flux density, and diffusion of NBD from the cell reservoir into the thin layer (even if diffusion in this thin layer is slow42,43). Depletion of the photosensitizer can be excluded as the corresponding IR bands indicate a loss of only 25% of the initial MK concentration after 2 s of UV irradiation. From the photon flux density and the concentration change, we can estimate an initial effective external quantum efficiency of the experiment, defined as the number of QC molecules produced per incident photon, of approximately 1.5%. This value is much lower than the internal quantum efficiency of the NBD/QC system and mainly limited by the low UV absorption in the thin film (thickness of approximately 4 μm; see the Supporting Information).24 Regarding potential applications of the NBD/QC system, undesired side reactions are the most critical issue.20,46 Here, the PEC-IRRAS experiment allows for determination of the selectivity (SQC = −ΔcQC/ΔcNBD) for QC directly during photochemical conversion. The corresponding data are displayed in Figure 2f. We calculate an initial selectivity SQC of 95% (±5%), which decreases to 87% at 80% conversion. The decreasing selectivity suggests that at a high conversion level the selectivity is limited by a follow-up reaction, which is most likely between QC and the photosensitizer in the triplet state,47,48 which leads to decomposition of QC. This assumption is confirmed by findings of Gorman et al. that the exited benzophenone, the parent compound of MK, reacts selectively with QC but not with NBD.48 In the second part of the experiment, we combined the photochemical conversion of NBD to QC with the electrochemically triggered cycloreversion in the same experiment. The procedure is shown in Figure 3a. A solution of 0.1 M NBD, 5 mM MK, and 0.1 M Bu4NClO4 in acetonitrile was irradiated for 1 s at an electrode potential of −0.08 Vfc. Before UV irradiation, an IR background was taken, and all spectra are referred to this reference. Further IR spectra were recorded immediately before and after UV irradiation. Subsequently, the potential was increased stepwise to +1.82 Vfc (0.1 V steps) with an IR spectrum recorded after each potential step. The resulting data are shown in Figure 3b.
Figure 3. In situ IR spectroscopy during photochemical conversion of NBD to QC and electrochemically triggered cycloreversion to NBD (0.1 M NBD, 5 mM MK, and 0.1 M Bu4NClO4 in acetonitrile): (a) schematic representation of the experimental procedure applied. (b) IR difference spectra taken before UV irradiation (red), after UV radiation (purple, at −0.08 Vfc), and upon a stepwise change of the electrode potential (black, IR reference taken at −0.08 Vfc); IR bands of NBD are marked in blue, QC is in orange, and MK is in green. (c) Calculated concentration changes of NBD and QC.
Similar to the experiment in Figure 2, we observe the characteristic bands upon UV irradiation, which indicate the conversion of NBD to QC and the partial consumption of MK. After irradiation, a small fraction (
