Exploring Closed-Shell Cationic Phenalenyl: From Catalysis to Spin

Article pubs.acs.org/accounts

Exploring Closed-Shell Cationic Phenalenyl: From Catalysis to Spin Electronics Arup Mukherjee,† Samaresh Chandra Sau,‡ and Swadhin K. Mandal*,‡ †

Department of Chemistry, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, Massachusetts 02139, United States ‡ Department of Chemical Sciences, Indian Institute of Science Education and Research, Kolkata, Mohanpur 741246, India CONSPECTUS: The odd alternant hydrocarbon phenalenyl (PLY) can exist in three different forms, a closed-shell cation, an open-shell radical, and a closed-shell anion, using its nonbonding molecular orbital (NBMO). The chemistry of PLY-based molecules began more than five decades ago, and so far, the progress has mainly involved the open-shell neutral radical state. Over the last two decades, we have witnessed the evolution of a range of PLYbased radicals generating an array of multifunctional materials. However, it has been admitted that the practical applications of PLY radicals are greatly challenged by the low stability of the openshell (radical) state. Recently, we took a different route to establish the utility of these PLY molecules using the closed-shell cationic state. In such a design, the closed-shell unit of PLY can readily accept free electrons, stabilizing in its NBMO upon generation of the open-shell state of the molecule. Thus, one can synthetically avoid the unstable open-shell state but still take advantage of this state by in situ generating the radical through external electron transfer or spin injection into the empty NBMO. It is worth noting that such approaches using closed-shell phenalenyl have been missing in the literature. This Account focuses on our recent developments using the closed-shell cationic state of the PLY molecule and its application in broad multidisciplinary areas spanning from catalysis to spin electronics. We describe how this concept has been utilized to develop a variety of homogeneous catalysts. For example, this concept was used in designing an iron(III) PLY-based electrocatalyst for a single-compartment H2O2 fuel cell, which delivered the best electrocatalytic activity among previously reported iron complexes, organometallic catalysts for various homogeneous organic transformations (hydroamination and polymerization), an organic Lewis acid catalyst for the ring opening of epoxides, and transition-metal-free C−H functionalization catalysts. Moreover, this concept of using the empty NBMO present in the closed-shell cationic state of the PLY moiety to capture electron(s) was further extended to an entirely different area of spin electronics to design a PLY-based spin-memory device, which worked by a spin-filtration mechanism using an organozinc compound based on a PLY backbone deposited over a ferromagnetic substrate. In this Account, we summarize our recent efforts to understand how this unexplored closed-shell state of the phenalenyl molecule, which has been known for over five decades, can be utilized in devising an array of materials that not only are important from an organometallic chemistry or organic chemistry point of view but also provide new understanding for device physics.

■

INTRODUCTION

Historically, based on early Hückel calculations, Robert Haddon2 proposed that the presence of the NBMO in PLY could be advantageous in designing single-component neutral radical-based molecular conductors and superconductors. However, all initial experimental efforts to isolate the PLY radical in the solid state were ineffective due to kinetic instability, resulting in dimerization by σ-association.3 In 1999, Yamamoto and Nakasuji reported the successful synthesis and isolation of the first PLY radical in the crystalline state by introducing sterically bulky tert-butyl groups on the PLY scaffold (Chart 2,

Phenalenyl (PLY) is a thermodynamically stable, odd alternant tricyclic hydrocarbon with high symmetry (D3h). Its unique electronic structure has drawn attention in multiple areas.1 PLY exhibits amphoteric redox behavior and can exist in three redox species, namely, a closed-shell cation, an open-shell radical, and a closed-shell anion, with high thermodynamic stabilities (Chart 1a).1 The transformation from the cation to radical to anion progresses through successive reductions in an accessible nonbonding molecular orbital (NBMO). Hence, it does not greatly affect the stability of the resulting species. In the radical state, the spin density spreads over the α positions of the PLY skeleton (Chart 1b). © 2017 American Chemical Society

Received: March 22, 2017 Published: June 30, 2017 1679

DOI: 10.1021/acs.accounts.7b00141 Acc. Chem. Res. 2017, 50, 1679−1691

Article

Accounts of Chemical Research

material in secondary batteries.7 Furthermore, they synthesized an electronically stabilized phenalenyl radical (Chart 2, IV) with six methoxy groups introduced at all the α-carbon atoms functioning as a novel quantum spin simulator.8 Nakasuji and coworkers designed a tribenzodecacyclenyl (Chart 2, V) with extended conjugation containing three phenalenyl units, which behaved as a six-stage amphoteric redox compound.9 Kim, Wu, and co-workers reported a Ni-phenalenyl-fused porphyrinbiradical (Chart 2, VI).10 On the other hand, Haddon and coworkers took a unique approach to introduce extended electronic conjugation by connecting two or more PLY-based units via the central atom (Chart 2, VII and VIII).11 This approach has resulted in the development of the best organic molecular conductor,12 materials exhibiting simultaneous bistability in three physical channels,13 and realization of the elusive resonating-valence-bond (RVB) ground state11d earlier postulated by Pauling and Anderson.14 This intriguing area of openshell PLY chemistry has been reviewed recently by Morita, Takui and Hicks15,16 and also later by Kubo.17 Despite these developments, it has been admitted that the practical applications of PLY radicals are greatly challenged by their instabilities, which require manipulation in degassed solutions and under inert conditions, thus requiring special attention.15 This limits the wide applicability of the radical states of PLY-based molecules. To avoid this difficulty, we recently adopted an alternative approach to establish the utility of PLY-based molecules in diverse applications utilizing the closed-shell cationic state instead of the previously used open-shell (radical) state (Chart 1a). In this design, it was postulated that the closed-shell state of

Chart 1. (a) Three Stable Redox Species of PLY and (b) Resonance Structures of the PLY Radical1

I).4 This study paved the way to a new understanding of the nature of the intermolecular interactions between PLY radicals.5 For example, in a recent study, Kubo, Kertesz, and co-workers explored the behavior of PLY radicals (Chart 2, II) undergoing stacking via the formation of a nonclassical multicenter bond.6 Morita, Takui, and co-workers developed a tri-tert-butylated PLY radical (Chart 2, III), which was applied as an electroactive

Chart 2. Representative Examples of Structurally Characterized PLY Radicals

1680


Article


reduction processes according to Chart 1a, describing the three redox states of the PLY-based molecule. With this background in mind, the following possible resonance forms (A−F) of PLYcoordinated metal complexes may be considered. Representations A and B capture the PLY ligand as a simple uninegative ligand, which deprotonates upon coordination to the metal ion, and they can be combined into representation F. Representations C and D are adopted following the concept developed by Haddon and co-workers (Figure 1B), which can be translated into representation E showing a closed-shell cationic state of PLY that is generated upon coordination to a metal ion. Keeping this in mind, we explored the chemistry of the closed-shell cationic PLY unit in catalysis and further extended this concept to a completely different field of spin electronics.

PLY can readily accept free electrons in to its NBMO. This type of design avoids the synthesis of the unstable open-shell state but still can take advantage of it through in situ generation of the radical via external electron transfer into the empty NBMO. To the best of our knowledge, such an approach using the closedshell PLY system was unprecedented before our study. This Account will focus on our recent advancements using the closedshell cationic state of the PLY molecule and its application in broad multidisciplinary areas spanning catalysis to spin electronics. In this Account, we address three different types of PLY ligands (1−3) based on O,O-, N,O-, and N,N-donor atoms (Chart 3), which can be easily synthesized following previous literature methods.11a,18

■

Chart 3. PLY Ligand-Based Synthons

CONCEPTUAL DEVELOPMENT OF CLOSED-SHELL PLY AS A BUILDING BLOCK FOR ORGANOMETALLIC CATALYSTS For this work, we chose a redox-innocent metal ion, such as Al(III), to avoid redox complications that might arise from a transition-metal ion. It was postulated that coordination of PLY ligands with such a redox-innocent Al(III) ion would generate the cationic state of PLY. Such metal complexes of the PLY-based ligand systems were expected to behave as strong Lewis acids, and the initial interaction of the electron density of the approaching nucleophile could be supported by the presence of an energetically accessible NBMO controlled by the ligand system (Chart 4). Earlier, Mandal and Roesky documented a dramatic enhancement of the catalytic activity by increasing the Lewis acidity by grafting a Lewis acidic main-group metal fragment with a catalytically active metal ion through an oxygen atom.19 For this study, we chose the ring-opening polymerization (ROP) reactions of ε-caprolactone (ε-CL) and rac-lactide (racLA) monomers as model reactions to establish the proof of concept, since it is well-known that the key step in the ROP process involves nucleophilic attack of the oxygen atom of the monomer to the LUMO of the catalyst.20 Having this aim in mind, a series of organoaluminum complexes (Scheme 1; 4−10)

■

REPRESENTATION OF THE METAL COMPLEXES WITH PLY LIGANDS We postulate that metal coordination of these PLY-based synthons (1−3) might open its cationic state having an empty NBMO. Using the coordination of a PLY-based ligand to a metal ion [for example, M(III)], we drew the corresponding resonating structures to represent the PLY-based metal complexes (Figure 1A). This representation is reminiscent of the resonance structural drawings adopted by Haddon and co-workers in their spiro-conjugated boron PLY compounds, where they have envisaged a positive charge on two PLY units and a negative charge on the boron ion (as shown in Figure 1B) to make the molecule electroneutral (Figure 1B, I).11c This particular representation has advantages for visualizing the electrochemical

Figure 1. (A) The resonance structures (A−F) of the PLY-based metal complexes and (B) representation of Haddon’s spiro-biphenalenyl boron compounds. 1681


Article


nitrogen atmosphere, and the major highlights of the study are summarized in Table 1. Kinetic analyses suggested the

Chart 4. Substrate Activation through Charge Transfer from the HOMO of the Nucleophile to the LUMO of the Catalyst

Table 1. Comparative Rates of Homopolymerization of ε-CL and rac-LA with 4, 5, and 8 entry catalyst 1 2 3 4 5 6

4 5 8 4 5 8

substrate

[Sub]/[Cat]/ [BnOH]

time (min)

temp (°C)

conv (%)

ε-CL ε-CL ε-CL rac-LA rac-LA rac-LA

200/1/2 200/1/2 200/1/2 200/1/2 200/1/2 200/1/2

100 100 100 150 150 150

50 50 50 60 60 60

95 76 3 99 10 2

polymerization processes were well controlled. Under identical catalytic conditions, the catalytic activity for the homopolymerization of ε-CL and rac-LA was found to vary remarkably (Table 1) depending on the nature of the catalyst. In general, the catalytic activity followed the order of O,O-system (4) > N,Osystem (5) > N,N-system (8). This prompted us to probe the origin of such a drastic catalytic activity difference between these PLY-based complexes. To begin, we carried out density functional theory (DFT) calculations on some representative organoaluminum complexes bearing different donor atom combinations (4, 5, and 8). The calculations clearly revealed that the LUMO predominantly resides over the PLY part of the

bearing the PLY ligands were synthesized and structurally characterized by single-crystal X-ray diffraction studies. Compounds 4, 5, and 8 having O,O-, N,O-, and N,N-donor atom combinations were used for the catalytic ROP of ε-CL and racLA.21 Polymerization screenings were performed under a dry

Scheme 1. Syntheses and Representative Molecular Structures of the Organoaluminum Complexes 4−10

1682


Article

Accounts of Chemical Research molecule in each case (Figure 2). We found that the energy of the LUMO gradually increased from the O,O-system in 4 (−3.24

Figure 2. Computed LUMOs of 4, 5, and 8. Reproduced with permission from ref 21. Copyright 2012 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim.

Figure 4. Molecular orbital energy diagrams indicating the HOMO− LUMO energy gaps between the organoaluminum catalysts and the LA monomer. Reproduced with permission from ref 21. Copyright 2012 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim.

eV) to the N,O-system in 5 (−2.87 eV) to the N,N-system in 8 (−2.51 eV). Additionally, the calculations revealed that the energy differences between the LUMO and LUMO + 1 in these complexes were quite substantial, showing the highest difference in 4 (2.08 eV) and the lowest in 8 (1.85 eV). This observation implies the presence of an energetically accessible empty NBMO to accept electrons in these PLY-based organoaluminum complexes, which makes 4 as the most Lewis acidic. The electrochemical studies on these complexes (4, 5, and 8) established two one-electron reduction processes, which correspond to the sequential generation of the anionic radical and dianionic species (Figure 3).11,12 This electrochemical result confirms our earlier assumption that metal-ion coordination to the PLY ligand will open the closed-shell cationic state of the PLY moiety. Furthermore, the end group analysis of the obtained polymer by 1H NMR spectroscopy suggested that the ROP of the cyclic esters proceeded by a monomer-activated mechanism. Calculations on the transition states associated with this mechanism revealed that, during the initial interaction, the cyclic ester monomer transfers electron density to the catalyst. The calculated charge distribution of PLY in the probable transition state involving the monomer and catalyst suggested that the charge on the PLY unit decreases from +0.43 (in 4) au to +0.17 au. This result clearly supports that the PLY unit acts as a Lewis acceptor. Furthermore, the energy comparison between the LUMO of the catalyst and the highest occupied molecular orbital (HOMO) of the model monomer (rac-LA) yielded differences of 286, 322, and 356 kJ/mol for 4, 5, and 8, respectively (Figure 4). This suggests that the activity of the catalyst was largely controlled by the difference in the HOMO−LUMO energy gap between the nucleophilic monomer and the electrophilic catalyst. The energetically most accessible LUMO of the catalyst (observed in 4) resulted in the best catalytic activity, while the

least accessible LUMO (observed in 8) yielded the lowest catalytic activity in the ROP process. We further demonstrated that these PLY-based organoaluminum complexes followed a similar trend in their catalytic activity for a different transformation, the intermolecular hydroamination of carbodiimides with aromatic amines to produce substituted guanidines.22 Additionally, we also synthesized and structurally characterized a series of Zn complexes with the PLY ligands. These organozinc complexes (Chart 5; 11−19) Chart 5. Organozinc Complexes (11−19) of the PLY Ligands

were tested for ROP of cyclic esters and intramolecular hydroamination, where the activities correlated largely to the electron acceptance capability of these PLY-based organometallic complexes and the HOMO−LUMO gaps between the substrate and catalysts.23,24

Figure 3. Cyclic voltammetry of 4 in acetonitrile revealing two successive one-electron reductions. Reproduced with permission from ref 21. Copyright 2012 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim. 1683


Article


■

APPLICATION OF CLOSED-SHELL PLY IN ORGANOCATALYSIS In the previous section, we demonstrated that the closed-shell cationic state of the PLY unit can be generated by coordination with a metal ion and that the catalytic activity can be tuned depending on the energy of the LUMO of the catalyst. Next, we sought to explore the cationic state of the PLY unit under metalfree conditions for homogeneous organic transformations. This section of the Account will discuss organic Lewis acid catalysts developed with the PLY unit that exist in their cationic forms without requiring any metal-ion coordination. By definition, a Lewis acid possesses a low-lying LUMO, which can accept a pair of electrons. In the literature, most Lewis acids contain metal ions or elements that are electronically unsatisfied and have easily accessible orbitals (for example, boron- and aluminum-based Lewis acids).25 However, the LUMOs of most organic molecules are antibonding in nature, which makes it difficult to access them energetically during catalysis. The major advantage associated with the use of Lewis acid catalysts having cationic PLY units is the easy accessibility of the empty NBMO (Chart 6) without

the optimized conditions, the reactions proceeded smoothly and gave the corresponding β-hydroxy amines in good to excellent yields (Scheme 2). To check the possibility of a Brønsted acid catalyzed reaction, we performed a control experiment considering that catalyst 20 or 21 may in situ generate HBF4 upon reaction with the amine substrates. When we physically mixed catalyst 20 and an amine (1:1) and recorded the 1H NMR spectrum, we did not observe any significant change in the chemical shifts from either of these two individual moieties, which ruled out the possibility of any chemical reaction forming HBF4, or in other words, it indicated that the catalysis was not driven by H+ formed in situ. Interestingly, the catalytic performance of 20 was found to be considerably higher than that of 21 in every case (Scheme 2).26 DFT calculations noted that the LUMO energy levels in these PLY units could be tuned by changing the substitution from O,O- (20) to N,O- (21) (Figure 5), establishing a relatively

Chart 6. Cationic Phenalenyl 20 and 21 and a Schematic Description of the NBMO of Cationic PLY Supporting the Interaction with a Nucleophilic Substrate (Nu)

Figure 5. (a) Computed LUMO of catalyst 20 and (b) molecular orbital energy diagrams of catalysts 20 and 21. Reproduced with permission from ref 26. Copyright 2014 American Chemical Society.

lower energy LUMO for catalyst 20. An electrochemical study was carried out to further understand the electron acceptance properties of these organic Lewis acids, indicating the presence of two one-electron reduction processes making the singly occupied MO and doubly occupied MO to generate the radical and anion, respectively (Figure 6). Comparing the first reduction potentials of 20 and 21 [20 (E11/2 = −0.408 V) and 21 (E11/2 = −0.931 V)], the LUMO of 20 was more accessible for reduction compared to that of 21, as also revealed by DFT calculations (Figure 5b). These observations also agree with our previous findings with organoaluminum complexes.21,22 To gain mechanistic understanding of the catalytic reaction, we carried out computational studies (Figure 7). The calculations revealed that the LUMO of catalyst 20 remained 0.72 eV above the HOMO of the amine, while for the HOMO of the epoxide, the LUMO of 20 was located 1.62 eV above. Thus, the comparatively smaller

Reproduced with permission from ref 26. Copyright 2014 American Chemical Society

affecting the stability of the transition states, which is unlike the case when a molecule utilizes a formally antibonding orbital. To test this hypothesis, we chose a model reaction, namely, an epoxide ring-opening reaction involving an amine in the presence of the organic PLY cations as catalysts (Scheme 2).26 The PLY-based cations (Chart 6), 9-methoxy-1-ethoxyphenalenium tetrafluoroborate (20, O,O-system) and 9-methylamino-1-ethoxyphenalenium tetrafluoroborate (21, N,O-system), were synthesized starting from 1 and 2 by reaction with Et3OBF4, following a previously reported procedure.11a The catalytic reactions with 20 and 21 were carried out at 40 °C under solvent free conditions with 2.5 mol % catalyst loadings. Under

Scheme 2. Aminolysis Reaction of Epoxides in the Presence of 20 or 21

1684


Article


Figure 6. Cyclic voltammograms of (a) 20 and (b) 21, revealing (c) two successive one-electron reductions. Reproduced with permission from ref 26. Copyright 2014 American Chemical Society.

Figure 7. Computed potential energy surface (PES) for the aminolysis reaction in the presence of catalyst 20. Reproduced with permission from ref 26. Copyright 2014 American Chemical Society.

■

TRANSITION-METAL-FREE C−H FUNCTIONALIZATION USING A CLOSED-SHELL CATIONIC PLY COMPLEX Next, we explored a different area of transition-metal-free C−H functionalization of aryl halides with a closed-shell cationic PLYbased complex (Scheme 3).27 Transition-metal-catalyzed arylation of aromatic compounds via C−H functionalization has long been a matter of interest. However, there has been burgeoning interest for the development of catalysts with moreabundant, less-precious, and biocompatible metals avoiding

HOMO−LUMO gap for the adduct formed between the catalyst and amine resulted in a stronger donor−acceptor interaction. Furthermore, to understand the superior activity of catalyst 20 over catalyst 21, it was found by DFT calculations that the free energy of activation required for the rate-determining step (shown in Figure 7 through TS3, A to B) in the presence of catalyst 21 was higher in energy than the respective barrier in the presence of catalyst 20, which originated from the lower energy LUMO present in 20. 1685


Article

Accounts of Chemical Research Scheme 3. (A) Synthesis of Biaryls in the Presence of 2 as a Catalyst and (B) Plausible Reaction Mechanism


loadings (20−40 mol %) for other transition-metal-free catalytic C−H functionalization processes.28,29Additionally, the catalytic reaction was accomplished at a loading down to 0.5 mol % but in a lower yield (36%). To understand the mechanistic cycle, we performed a stoichiometric reaction to isolate the catalytically active intermediate. Experimental and computational studies suggested the formation of an in situ generated tBuOHcoordinated potassium complex of PLY (Scheme 3B, 23). This type of metal-ion coordination can generate the cationic state of the phenalenyl ligand, as documented from our earlier studies.21,22 Consequently, complex 23 with an empty NBMO can readily interact with the in situ generated tert-butanol resulting in a SET to form a PLY-radical-based transient complex 24 (Scheme 3B). Next, the singly occupied molecular orbital (SOMO) of the phenalenyl-based radical 24 can transfer an electron to the aryl halide, forming a highly reactive aryl radical, which can subsequently couple with the arene partner to yield a biaryl product, as depicted in Scheme 3B.

transition-metal ions. In this context, a transition-metal-free single-electron transfer (SET) approach with the assistance of an organic ligand appeared to be the most promising choice. An initial report by Itami, prompted the development of a transitionmetal-free arylation protocol using various organic ligands.28,29 However, these catalytic protocols typically suffer from long reaction times, limited substrate scopes, and most critically, the use of semistoichiometric catalyst loadings (20−40 mol %).28,29 Thus, it has been a challenge to carry out C−H functionalization with a typical transition-metal-based catalyst loading (1−5 mol %). To address this problem, phenalenyl motifs in coordination with a metal ion can effectively accommodate incoming electron in their NBMOs to stabilize the radical state, and thus, it can act as a mediator of SET during the arylation process under low catalyst loading conditions. In this work,27 we hypothesized that the donor centers (O and N) present in the PLY motif will be appropriate to form a chelate complex by coordination with the alkali-metal ion present in the base (potassium tert-butoxide), facilitating the SET process. To our delight, the catalytic reaction was successfully accomplished with only 5 mol % loading of PLY catalyst 2 and provided several biaryl motifs covering a broad substrate scope with satisfactory yields (43−95%, Scheme 3A). Both electrondonating and electron-withdrawing groups on the aryl moieties survived under the reaction conditions. Most importantly, this catalytic reaction could be carried out at a much lower catalyst loading of 5 mol %, as opposed to the generally observed catalyst

■

CLOSED-SHELL PLY−Fe COMPLEX AS AN ELECTROCATALYST FOR H2O2 FUEL CELLS Our search to find applications of the closed-shell cationic state of the PLY unit in diverse disciplines prompted us to apply it as an electrocatalytic material (as a cathode) in a one-compartment hydrogen peroxide (H2O2) fuel cell. The one-compartment H2O2 fuel cell has evolved as a simple, compact, efficient, and environmentally benign power source.30 The H2O2 fuel cell 1686


Article


that the PLY moiety played a key role in determining the final power output of the iron(III)complex. A combined experimental and computational study was carried out to gain insight into the mechanistic aspects of the catalytic reaction. The CV of complex 26 revealed a multiredox process in solution (Scheme 5a). Electronic structure calculations by DFT on 26 revealed that the electron density on the β-LUMO had appreciable metal contribution (∼70%, Scheme 5b), which suggested that the first reduction of 26 to 26−1 was predominantly metal centered. Moreover, the electron densities on the β-LUMO (Scheme 5c) and β-LUMO + 1 (Scheme 5d) of 26−1 were completely based on the PLY moiety, which suggested that the second (26−1 to 26−2, E1/2 2 = −1.38 V) and third (26−2 to 26−3, E1/2 = −1.79 V) reductions of 26 were 3 exclusively PLY based, creating open-shell states of the phenalenyl ligand, as explained in Scheme 5e. Based on this study, we postulated the mechanism as depicted in Scheme 5f. Initially, oxidation of H2O2 at the anode produced electrons and H+, which moved toward the cathode to reduce 26 into 26−1. Complex 26−1 transferred electrons to H2O2, which in combination with diffused H+ produced H2O (route i) and regenerated catalyst 26. Since H2O2 reduction is a two-electron process, the reduction depicted in this pathway requires two molecules of 26−1. Thus, route i could be considered as a general one to any Fe(III)-based molecule based on the Fe(III)/Fe(II) oxidation−reduction process. Alternatively, a one-step twoelectron reduction of H2O2 could be considered, as shown in route ii, involving the two-electron reduced species 26−2, which is supported by the CV observations establishing multielectron acceptance by complex 26, thus enhancing greatly the efficiency of catalyst 26 compared to earlier reported Fe(III) complexes.

constitutes a benign alternative to fossil fuels, generating H2O and O2 as byproducts. The H2O2 fuel cell operates with the controlled oxidation of H2O2 at the anode, giving O2 and 2H+, and reduction of H2O2 at the cathode, producing 2H2O (Scheme 4A). Most of the efficient cathode materials for one-compartment H2O2 fuel cells have been based on precious metals, such as Pt. In the present section, we discuss the application of a cationic PLY-based FeIII(PLY)3 complex (26) as a cost-effective electrocatalytic material for a one-compartment membraneless H2O2 fuel cell.31 Complex 26 was synthesized and structurally characterized (Scheme 4B). The performance test of the H2O2 fuel cell was carried out using an electrode modified with 26 as a cathode and Ni as an anode with 300 mM H2O2 in aqueous 0.1 M H2SO4 (Figure 8) and 0.1 M HClO4 solutions. Gratifyingly, the power

Figure 8. Performance test of the single-compartment H2O2 fuel cell using 26 as a catalyst. Reproduced with permission from ref 31. Copyright 2015 American Chemical Society.

■

acquired with 26 was found to be nearly 30-fold greater than that observed with an earlier report using [FeIII(Pc)Cl] (Pc = phthalocyanate, 10 μWcm−2) under identical conditions.31 To determine the participation of the PLY unit, a model complex, FeIII(acac)3 (acac = acetylacetonate), containing redox inactive acac ligands was chosen, which is structurally reminiscent of FeIII(PLY)3 (26). Under identical conditions, the best result of FeIII(acac)3 + 10 wt % carbon provided a maximum power density of 0.089 mW cm−2, which was significantly lower (nearly 15-fold) compared to the power output obtained using complex 26. Moreover, the maximum power density of 26 was improved to 1.43 mW cm−2 by adding 10 wt % carbon, which was found to be 143 times higher than the most efficient neutral mononuclear Fe(III)complex previously reported.32 This result clearly noted

APPLICATION OF CLOSED-SHELL PLY IN SPIN ELECTRONICS Despite the tremendous promise of PLY-based radicals for future device and spin-based electronic applications as documented by Miller, Morita, Takui, and Hicks, most early attempts in constructing spin-electronic devices were ineffective.15,16,33 However, our study in catalysis established that the generation of a closed-shell cationic PLY-based molecule could serve as an alternative to avoid isolating the unstable radical but still allow the empty NBMO of these molecules to form in situ the radical state in solution. Interestingly, transforming such a concept into the solid state might lead to generation of novel spin-electronic materials.

Scheme 4. (A) Electrode Reactions in a H2O2 Fuel Cell and (B) Synthesis and Molecular Structure of 26

1687


Article


Scheme 5. (a) CV of 26, Isodensity Plots of the (b) β-LUMO of 26, (c) β-LUMO of 26−1, and (d) β-LUMO + 1 of 26−1, (e) Three Successive One-Electron Reductions Showing the Formation of Phenalenyl-Centered Radical Anions, and (f) Plausible Routes for the Electrocatalytic Reduction of H2O2 Using 26


netic Co layer was essential to generate such a MR effect. This unprecedented MR signal was further explained by considering a spin-filtration mechanism, which originated from the capability of 27 to accept the injected spin from the bottom ferromagnetic layer. A detailed computational (DFT) modeling study confirmed that the Co layer transferred its spin into the first adsorbed molecular layer of 27 via charge transfer through hybridization. Due to the planar structure of molecule 27, the pz atomic-type orbitals hybridized strongly with the d-states (mainly with the dz2, dxz and dyz orbitals) of the Co atoms forming organometallic hybrid molecule−metal pz−d interface states. As a result, the first zinc-PLY molecular layer developed a spin-imbalanced electronic structure and acquired a net moment of 0.11 μB oriented antiparallel to the moment of the hybridized surface layer of Co atoms. The PLY ring of the zinc-PLY complex having a partial radical state (as a donor layer) further interacted with the PLY ring of the second layer of complex 27 (acceptor layer with empty NBMO) via an intermolecular π−π interaction, giving rise to a weak donor−acceptor-type π-dimer formation with a ∼3.26 Å interlayer separation, which is smaller than the sum of the van der Waals radii of the carbon atoms (Figure 10). Although the calculations indicated that the second molecular layer of 27 possessed no magnetic moment, the π−π interactions between the molecules created a spin-imbalanced electronic structure. The first energy level above the Fermi energy (LUMO) level was observed to be spin-split with a difference of ∼0.14 eV, responsible for the spin bias, which resulted in a spin-filtration property responsible for this unprecedented MR signal.

Herein, we discuss the closed-shell PLY unit, created upon coordination with a zinc metal ion as a noninnocent building block for the construction of a spin-memory device.34 In this study, we used an organozinc complex 27 with a PLY backbone. Complex 27 was synthesized by treating ZnMe2 with 1 under ambient conditions (Scheme 6) and was structurally characterized by single-crystal X-ray diffraction. Scheme 6. Synthesis of Organozinc Complex 27 for Applications in Spin Electronics

The molecular device with 27 for application in spintronics was fabricated by a vapor deposition technique using thin films, with copper (Cu) serving as the top contact electrode and cobalt serving as the bottom ferromagnetic electrode for spin injection34 with layer thicknesses of Co (8 nm)/27 (40 nm)/ Cu (12 nm) (Figure 9). Before the experiment, the device was cooled down to 4.2 K. To our surprise, measurements of the magnetotransport properties of this device demonstrated a spindependent resistance resulting in an interfacial magnetoresistance (IMR). A large magnetoresistance (MR) signal near 25% was observed when the Co magnetization switched at its coercive field. A control experiment showed that the direct contact between the zinc-PLY molecule 27 and the ferromag-

■

SUMMARY AND CONCLUSIONS In summary, we documented our recent efforts to capture the chemistry of the unexplored closed-shell cationic state of the PLY 1688


Article


Figure 9. (a) Charge transfer from the Co-ferromagnetic layer creating the PLY-based radical in situ in the solid state and (b) magnetoresistance measurements of a device with the single ferromagnetic electrode [Co (8 nm)/ZMP (40 nm)/Cu(12 nm)]. Reproduced with permission from ref 34. Copyright 2013 Macmillan Publishers Limited.

Biographies Arup Mukherjee obtained his Ph.D in 2013 from IISER, Kolkata (India), under the supervision of Prof. Swadhin K. Mandal. Then, he moved to the Weizmann Institute of Science (Israel) to work with Prof. David Milstein as a postdoctoral fellow. Since 2015, he has been working with Prof. Christopher C. Cummins as a Fulbright-Nehru postdoctoral research fellow at MIT (USA). His research interests include maingroup chemistry and organometallic catalysis. Samaresh Chandra Sau obtained his Ph.D in 2015 from IISER, Kolkata (India), working under the supervision of Prof. Swadhin K. Mandal. Subsequently, he is working as a Research Scientist in collaboration with the drug discovery company Invictus Oncology Pvt. Ltd. under Prof. Swadhin K. Mandal.

Figure 10. Computed diagram of the relaxed and spin filtered molecules on the surface of the device. Reproduced with permission from ref 34. Copyright 2013Macmillan Publishers Limited.

Swadhin K. Mandal is currently an Associate Professor at the Indian Institute of Science Education and Research, Kolkata, in the area of Chemical Sciences. He obtained his doctoral degree under the supervision of Prof. S. S. Krishnamurthy at the Indian Institute of Science, Bangalore. He has been a postdoctoral fellow in the Department of Chemistry at the University of California, Riverside, with Prof. Robert C. Haddon and an Alexander von Humboldt fellow at the University of Göttingen with Prof. Herbert W. Roesky. His current research interests include a wide range of areas, such as organometallic catalysis using base metals and main-group metals, metal-free CO2 fixation and its conversion into fuels, alternative approaches for C−H bond functionalization using radicals, the development of spinelectronic materials, and the discovery of new anticancer drugs based on supramolecular assembly. He served as an international editorial advisory member for the ACS journal Organometallics from 2013 to 2015. He was a recipient of the Young Investigators Meeting Boston award in 2012 for his contributions in Organometallic Catalysis.

molecule. This concept of using the empty NBMO present in the closed-shell cationic state of the PLY moiety to capture electron(s) was utilized to design a variety of catalysts. For example, this concept was used in designing organometallic catalysts for various homogeneous organic transformations (polymerization and hydroamination), an organic Lewis acid catalyst for the ring opening of epoxides, and transition-metalfree C−H functionalization catalysts. An iron(III) PLY-based electrocatalyst for a single-compartment H2O2 fuel cell was developed, which delivered the best electrocatalytic activity among previously reported iron-based complexes. Furthermore, we extended this concept of electron acceptance by the closedshell PLY molecule to the solid state to develop a spin-memory device, which operated via a spin-filtration mechanism using an organozinc compound of a PLY-based molecule. Given the evolving importance of multidisciplinary science, we believe that this is the appropriate time for this Account to be published, as it will be of interest to a broad range of the scientific community.

■

■

AUTHOR INFORMATION

ACKNOWLEDGMENTS

S.K.M. gratefully acknowledges the hard work by all the coworkers whose names are mentioned in the relevant references. Furthermore, S.K.M. thanks DST (Grant No. SR/FT/CS-020/ 2008), SERB (Grant No. SR/S1/IC-25/2012), CSIR (Grant No. 01(2369)/10/EMR-II), and IISER, Kolkata, for financial support. A.M. thanks the Fulbright-Nehru Postdoctoral Fellowship for support. S.C.S. thanks Invictus Oncology Pvt. Ltd. for a Research Scientist position.

Corresponding Author

*E-mail: [email protected]. ORCID

Swadhin K. Mandal: 0000-0003-3471-7053 Notes

The authors declare no competing financial interest. 1689


Article


■

(15) Morita, Y.; Suzuki, S.; Sato, K.; Takui, T. Synthetic Organic Spin Chemistry for Structurally Well-defined Open-shell Graphene Fragments. Nat. Chem. 2011, 3, 197−204. (16) Hicks, R. G. Switchable Materials: A New Spin on Bistability. Nat. Chem. 2011, 3, 189−191. (17) Kubo, T. Phenalenyl-Based Open-Shell Polycyclic Aromatic Hydrocarbons. Chem. Rec. 2015, 15, 218−232. (18) Haddon, R. C.; Rayford, R.; Hirani, A. M. 2-Methyl- and 5methyl-9-hydroxyphenalenone. J. Org. Chem. 1981, 46, 4587−4588. (19) Mandal, S. K.; Roesky, H. W. Assembling Heterometals through Oxygen: An Efficient Way to Design Homogeneous Catalysts. Acc. Chem. Res. 2010, 43, 248−259. (20) Nomura, N.; Akita, A.; Ishii, R.; Mizuno, M. Random Copolymerization of ε-Caprolactone with Lactide Using a Homosalen−Al Complex. J. Am. Chem. Soc. 2010, 132, 1750−1751. (21) Sen, T. K.; Mukherjee, A.; Modak, A.; Ghorai, P. K.; Kratzert, D.; Granitzka, M.; Stalke, D.; Mandal, S. K. Phenalenyl-Based Molecules: Tuning the Lowest Unoccupied Molecular Orbital to Design a Catalyst. Chem. - Eur. J. 2012, 18, 54−58. (22) Mukherjee, A.; Sen, T. K.; Ghorai, P. K.; Mandal, S. K. The Noninnocent Phenalenyl Unit: An Electronic Nest to Modulate the Catalytic Activity in Hydroamination Reaction. Sci. Rep. 2013, 3, 2821. (23) Sen, T. K.; Mukherjee, A.; Modak, A.; Mandal, S. K.; Koley, D. Substitution Effect on Phenalenyl Backbone in the Rate of Organozinc Catalyzed ROP of Cyclic Esters. Dalton Trans. 2013, 42, 1893−1904. (24) (a) Mukherjee, A.; Sen, T. K.; Ghorai, P. K.; Samuel, P. P.; Schulzke, C.; Mandal, S. K. Phenalenyl-Based Organozinc Catalysts for Intramolecular Hydroamination Reactions: A Combined Catalytic, Kinetic, and Mechanistic Investigation of the Catalytic Cycle. Chem. Eur. J. 2012, 18, 10530−10545. (b) Mukherjee, A.; Sen, T. K.; Ghorai, P. K.; Mandal, S. K. Organozinc Catalyst on a Phenalenyl Scafold for Intramolecular Hydroamination of Aminoalkenes. Organometallics 2013, 32, 7213−7224. (25) Ishihara, K. Lewis Acids in Organic Synthesis; Yamamoto, H., Ed.; Wiley-VCH: Weinheim, Germany, 2000; Vol 1, p 135. (26) Roy, S. R.; Nijamudheen, A.; Pariyar, A.; Ghosh, A.; Vardhanapu, P. K.; Mandal, P. K.; Datta, A.; Mandal, S. K. Phenalenyl in a Different Role: Catalytic Activation through the Nonbonding Molecular Orbital. ACS Catal. 2014, 4, 4307−4319. (27) Paira, R.; Singh, B.; Hota, P. K.; Ahmed, J.; Sau, S. C.; Johnpeter, J. P.; Mandal, S. K. Open-Shell Phenalenyl in Transition Metal-Free Catalytic C−H Functionalization. J. Org. Chem. 2016, 81, 2432−2441. (28) Yanagisawa, S.; Ueda, K.; Taniguchi, T.; Itami, K. Potassium tButoxide Alone Can Promote the Biaryl Coupling of Electron-Deficient Nitrogen Heterocycles and Haloarenes. Org. Lett. 2008, 10, 4673−4676. (29) Sun, C.-L.; Li, H.; Yu, D.-G.; Yu, M.; Zhou, X.; Lu, X. Y.; Huang, K.; Zheng, S. F.; Li, B.-J.; Shi, Z.-J. An Efficient Organocatalytic Method for Constructing Biaryls through Aromatic C-H Activation. Nat. Chem. 2010, 2, 1044−1049. (30) (a) Yamazaki, S.; Siroma, Z.; Senoh, H.; Ioroi, T.; Fujiwara, N.; Yasuda, K. A Fuel Cell with Selective Electrocatalysts Using Hydrogen Peroxide as Both an Electron Acceptor and a Fuel. J. Power Sources 2008, 178, 20−25. (b) Fukuzumi, S.; Yamada, Y. Hydrogen Peroxide used as a Solar Fuel in One-Compartment Fuel Cells. ChemElectroChem 2016, 3, 1978−1989. (c) Yamada, Y.; Yoneda, M.; Fukuzumi, S. High and robust performance of H2O2 fuel cells in the presence of scandium ion. Energy Environ. Sci. 2015, 8, 1698−1701. (d) Mase, K.; Yoneda, M.; Yamada, Y.; Fukuzumi, S. Seawater usable for production and consumption of hydrogen peroxide as a solar fuel. Nat. Commun. 2016, 7, 11470. (31) Pariyar, A.; Vijaykumar, G.; Bhunia, M.; Dey, S. K.; Singh, S. K.; Kurungot, S.; Mandal, S. K. Switching Closed-Shell to Open-Shell Phenalenyl: Toward Designing Electroactive Materials. J. Am. Chem. Soc. 2015, 137, 5955−5960. (32) Yamada, Y.; Yoshida, S.; Honda, T.; Fukuzumi, S. Protonated Iron−phthalocyanine Complex Used for Cathode Material of a Hydrogen Peroxide Fuel Cell Operated Under Acidic Conditions. Energy Environ. Sci. 2011, 4, 2822−2825. (33) Miller, J. S. Bistable Electrical, Optical, and Magnetic Behavior in a Molecule-Based Material. Angew. Chem., Int. Ed. 2003, 42, 27−29.

DEDICATION S.K.M. dedicates this Account to Prof. Robert C. Haddon (University of California, Riverside, USA) and deeply mourns the sudden death of Prof. Haddon.

■

REFERENCES

(1) Reid, D. H. The Chemistry of the Phenalenes. Q. Rev. Chem. Soc. 1965, 19, 274−302. (2) Haddon, R. C. Design of Organic Metals and Superconductors. Nature 1975, 256, 394−396. (3) Beer, L.; Mandal, S. K.; Reed, R. W.; Oakley, R. T.; Tham, F. S.; Donnadieu, B.; Haddon, R. C. The First Electronically Stabilized Phenalenyl Radical: Effect of Substituents on Solution Chemistry and Solid-state Structure. Cryst. Growth Des. 2007, 7, 802−809. (4) Goto, K.; Kubo, T.; Yamamoto, K.; Nakasuji, K.; Sato, K.; Shiomi, D.; Takui, T.; Kubota, M.; Kobayashi, T.; Yakusi, K.; Ouyang, J. A Stable Neutral Hydrocarbon Radical: Synthesis, Crystal Structure, and Physical Properties of 2,5,8-Tri-tert-butyl-phenalenyl. J. Am. Chem. Soc. 1999, 121, 1619−1620. (5) (a) Anamimoghadam, O.; Symes, M. D.; Long, D.-L.; Sproules, S.; Cronin, L.; Bucher, G. Electronically Stabilized Nonplanar Phenalenyl Radical and Its Planar Isomer. J. Am. Chem. Soc. 2015, 137, 14944− 14951. (b) Cui, Z.-h.; Lischka, H.; Beneberu, H. Z.; Kertesz, M. Rotational Barrier in Phenalenyl Neutral Radical Dimer: Separating Pancake and van der Waals Interactions. J. Am. Chem. Soc. 2014, 136, 5539−5542. (6) Mou, Z.; Uchida, K.; Kubo, T.; Kertesz, M. Evidence of σ- and πDimerization in a Series of Phenalenyls. J. Am. Chem. Soc. 2014, 136, 18009−18022. (7) Morita, Y.; Nishida, S.; Murata, T.; Moriguchi, M.; Ueda, A.; Satoh, M.; Arifuku, K.; Sato, K.; Takui, T. Organic Tailored Batteries Materials Using Stable Open-shell Molecules with Degenerate Frontier Orbitals. Nat. Mater. 2011, 10, 947−951. (8) Ueda, A.; Suzuki, S.; Yoshida, K.; Fukui, K.; Sato, K.; Takui, T.; Nakasuji, K.; Morita, Y. Hexamethoxyphenalenyl as a Possible Quantum Spin Simulator: An Electronically Stabilized Neutral π Radical with Novel Quantum Coherence Owing to Extremely High Nuclear Spin Degeneracy. Angew. Chem., Int. Ed. 2013, 52, 4795−4799. (9) Kubo, T.; Yamamoto, K.; Nakasuji, K.; Takui, T.; Murata, I. Hexatert-butyl tribenzodecacyclenyl: A Six-Stage Amphoteric Redox System. Angew. Chem., Int. Ed. Engl. 1996, 35, 439−441. (10) Zeng, W.; Lee, S.; Son, M.; Ishida, M.; Furukawa, K.; Hu, P.; Sun, Z.; Kim, D.; Wu, J. Phenalenyl-fused Porphyrins with Different Ground States. Chem. Sci. 2015, 6, 2427−2433. (11) (a) Mandal, S. K.; Itkis, M. E.; Chi, X.; Samanta, S.; Lidsky, D.; Reed, R. W.; Oakley, R. T.; Tham, F. S.; Haddon, R. C. New Family of Aminophenalenyl-based Neutral Radical Molecular Conductors: Synthesis, Structure, and Solid State Properties. J. Am. Chem. Soc. 2005, 127, 8185−8196. (b) Pal, S. K.; Itkis, M. E.; Tham, F. S.; Reed, R. W.; Oakley, R. T.; Haddon, R. C. Trisphenalenyl-Based Neutral Radical Molecular Conductor. J. Am. Chem. Soc. 2008, 130, 3942−3951. (c) Chi, X.; Itkis, M. E.; Patrick, B. O.; Barclay, T. M.; Reed, R. W.; Oakley, R. T.; Cordes, A. W.; Haddon, R. C. The First Phenalenyl-Based Neutral Radical Molecular Conductor. J. Am. Chem. Soc. 1999, 121, 10395−10402. (d) Pal, S. K.; Itkis, M. E.; Tham, F. S.; Reed, R. W.; Oakley, R. T.; Haddon, R. C. Resonating Valence-Bond Ground State in a PhenalenylBased Neutral Radical Conductor. Science 2005, 309, 281−284. (12) Mandal, S. K.; Samanta, S.; Itkis, M. E.; Jensen, D. W.; Reed, R. W.; Oakley, R. T.; Tham, F. S.; Donnadieu, B.; Haddon, R. C. Resonating Valence Bond Ground State in Oxygen Functionalized Phenalenyl-based Neutral Radical Molecular Conductors. J. Am. Chem. Soc. 2006, 128, 1982−1994. (13) Itkis, M. E.; Chi, X.; Cordes, A. W.; Haddon, R. C. Magneto-optoelectronic Bistability in a Phenalenyl-based Neutral Radical. Science 2002, 296, 1443−1445. (14) (a) Pauling, L. The Metallic State. Nature 1948, 161, 1019−1020. (b) Anderson, P. W. Resonating Valence Bonds: A New Kind of Insulator? Mater. Res. Bull. 1973, 8, 153−160. 1690


Article

Accounts of Chemical Research (34) Raman, K. V.; Kamerbeek, A. M.; Mukherjee, A.; Atodiresei, N.; Sen, T. K.; Lazic, P.; Caciuc, V.; Michel, R.; Stalke, D.; Mandal, S. K.; Blugel, S.; Munzenberg, M.; Moodera, J. S. Interface-engineered Templates for Molecular Spin Memory Devices. Nature 2013, 493, 509−513.

1691


Exploring Closed-Shell Cationic Phenalenyl: From Catalysis to Spin

Recommend Documents