Bioinspired approach toward molecular electrets: synthetic proteome for materials

Biomimicry, biomimesis and bioinspiration: what’s the difference?

As a first step toward learning lessons from nature, biomimicry involves sheer imitation [7]. Copying structural features, biomimicry leads to look-alike masques of biological systems that may or may not have the functionality of the imitated natural counterparts. Such examples encompass many of the first steps of humanity toward exploring the realms that are available for other species but inaccessible for us. For example, the first known human attempts to fly by the Andalusian polymath Abbas ibn Firnas, followed six centuries later by Leonardo da Vinci, involved designs of devices that appear to have had the perfect structural resemblance of bird wings [40]. Contrary to birds, however, such a set of wings alone could not provide the functionality needed to allow man to fly. Many of the purely structural imitations of life have remained in the realms of art. Nevertheless, they have played an important role in the human endeavors to reach beyond the horizons of what is thought as possible.

Biomimesis takes biomimicry to a deeper understanding of biological systems [7]. In addition to perfecting the structural imitations, biomimesis also aims at attaining functionalities comparable to those of the natural systems. The iterative processes of improving the features of the structural imitations in order to achieve the desired functionalities, provide invaluable lessons for learning how biological systems work. At systems level, the implementation of biomimesis allows for testing the hierarchy of importance of the various structure-function relationships in the natural systems. In the process of understanding the molecular and cellular aspects of living organisms, numerous examples of artificial enzymes [41], [42], [43], [44], synthetic de novo designs of proteins [45], [46], [47], [48], [49], [50], [51], [52], [53], [54], [55], [56], [57], [58], [59], artificial photo-driven proton pumps [60], [61], and artificial cells with designated functionalities [62], [63], [64], [65] have advanced the fields of biology, biochemistry and biophysics and provided unprecedented paradigms for bioengineering. At an organism-size level, examples for walking machines, which do not utilize wheels or roller, may seem somewhat impractical for the “optimal practical usage” of the current state of robotics. The process of designing and developing such walkers, however, plays a key role in advancing biomechanics and demonstrating its potential utility for rehabilitation medicine [66], [67], [68], [69], [70], and outside of biology [71], [72], [73], [74], [75], [76], [77]. Examples of taking structural features from nature and attaining the targeted functionalities include: (1) the invention of Velcro, for which George de Mestral used concepts from the burrs of the burdock plant [78], [79], and (2) superhydrophobic surfaces, which copy nanostructures of plant leaves capable of repelling water droplets [80], [81], [82], [83], [84], [85].

Going back to mimicking bird wings, airfoils designed with the understanding of Bernoulli’s principle involve a decrease in weight and an increase in lift, which allow objects to take flight provided that a form of thrust can be attained [86], [87]. Indeed, an increase in the wing area permitted the first human bird-like flights that involved gliding, using wind currents. The accounts about Br. Eilmer flying across river Avon in the 11^th century, and for Hezârfen Ahmed Çelebi flying from Europe to Asia crossing the Bosphorus in the 17^th century, testify to the importance of wings with large area in proportion to the carried weight [88].

Living systems have undergone epochs and eons of evolutionary optimization for surviving and flourishing in the dynamic and diverse environments on the Earth. Many of the biological systems, however, are frequently far from optimal for meeting the design demands and bare minimum requirements of viable energy technologies. Therefore, to truly benefit from the invaluable lessons from nature, it is essential to take them outside the constraints of the living systems and resort to biological inspiration.

Bioinspiraton goes beyond what nature has to offer: e.g. it uses concepts found in nature and applies them to synthetic systems [7]. For efficient engineering of bioinspired systems, it is essential to understand how nature works. Employing the fundamental understanding obtained from biology in manners most adequate for the targeted application that do not need to resemble the natural systems, opens up venues for unprecedented discoveries in science and developments in engineering. For example, the understanding of efficient propulsion to create thrust coupled with the lift characteristics of wings observed in birds have allowed man to take flight and led to modern aviation. As proposed and designed by Sir George Cayley in the 19^th century, going beyond mimesis by separating lift (i.e. fixed wings) from propulsion, and from vertical and horizontal control was the essential breakthrough that made modern aviation possible [89], [90], [91]. At molecular and cellular scales, taking ideas from natural photosynthesis for storing energy in the form of a pH gradient and employing non-native synthetic quinone structures allows for achieving unprecedented functionality of a photo-driven pump for calcium ions [92]. Seeking the functionality of the non-photochemical quenching in photosynthesis, without any true resemblance of the living systems, is another example of pushing the limits by using biological inspiration [93].

Dipole effects on charge transfer

Local electric fields, originating from molecular dipoles, have profound effects on the electronic properties of the microenvironment. As a result, local electric fields from molecular dipoles and ions affect ET [94], [95], [96], [97], [98], [99], providing a promising means for increasing the efficiency of the desired CT processes while suppressing the undesired ones. Small dipolar moieties modify the electron affinity and the ionization energy of materials [100], [101], [102], [103], [104]. Because of their rectifying capabilities, dipolar π-conjugated molecules have been a focus for molecular and organic electronics [105], [106], [107]. At metal-semiconductor interfaces, monolayers of organic molecules containing polar functional groups alter the rectification characteristics of such Schottky junctions [108], [109], [110]. Push-pull conjugates within CT organic structures on semiconductor surfaces introduce molecular dipoles that not only affect the electronic properties of the material but also modify the kinetics of interfacial ET [111]. Similarly, at a molecular level, dipole-generated fields shift the energy levels of the frontier orbitals of the electron donors and acceptors [12].

In the search for achieving control over CT using local electric fields, molecular electrets, such as protein helices, present one of the best choices because of their large permanent electric dipoles. With intrinsic dipoles of about 5 Debyes (D) per residue that points from the C- to the N-terminus [8], [9], [10], [11], [12], [16], [17], [112], [113], [114], [115], protein α-helices generate in their vicinity fields in the order of GV/m. (The direction of electric dipoles is from their negative to positive poles [116]).

In the late 1990s, Galoppini and Fox demonstrated for the first time the effect dipoles have on photoinduced CT [117], [118], [119]. They employed 14-residue-long polypeptides that show high propensity for assuming helical secondary structures. Synthetic α-L-amino acids, containing donor and acceptor moieties attached to their side chains, are placed six-residues apart in the middle of the polypeptide sequences. Upon photoexcitation of the acceptor, an electron moves from the HOMO of the donor to fill the newly generated vacancy in the acceptor “HOMO.” (Strictly speaking, after photoexcitation, the HOMO becomes a singly occupied MO; and because the transfer is along the HOMOs of the CT species following the excitation of the acceptor [120], [121], [122], [123], [124], it can be viewed as hole transfer rather than electron transfer). The rates of CT are up to about 30 times larger when the acceptor is positioned close to the N-terminus than to the C-terminus of the helices [117]. That is, the rates are faster for electrons moving toward the positive poles of the dipole than toward the negative ones, which encompasses the most important discovery described in these first reports.

This effect of macromolecular dipoles on CT kinetics is ascribed to field-induced changes in the CT driving force, −ΔG_CT⁽⁰⁾ [12], [118]. The CS state is polarized and generates a dipole that points from the radical anion of the reduced acceptor to the radical cation of the oxidized donor. Opposing orientation between the CS dipole and the permanent electric dipole of the helix stabilizes the CS state and increases the CT driving force (if the system operates in the Marcus Normal region). Conversely, co-directional orientation of the CS and helix dipoles leads to a decrease in the CT rates. This dipole-induced Franck-Condon effect on the CT kinetics is especially prevalent for small −ΔG_CT⁽⁰⁾, and for systems with an identical electronic coupling between the donor and the acceptor for the CS states oriented along and against the permanent macromolecular dipole. When the structures with opposing dipole direction are not completely symmetrical, the differences in the electronic coupling can prevail over the dipole-induced effect for relatively large −ΔG_CT⁽⁰⁾ [125].

Since the beginning of the 21^st century, studies of dipole effects on CT have focused on systems comprising polypeptide helices [12], [95], [99], [126], [127], [128], [129], [130], [131], [132], [133], [134], [135], [136], [137], [138], [139], [140], [141]. Donor-bridge-acceptor (DBA) constructs, where the bridge is a helix, allow for testing dipole-induced charge-transfer rectification, i.e. the difference between the rates of ET along the dipole vs. ET against the dipole [137]. Self-assembled monolayers (SAMs) of polypeptide helices on gold surfaces show similar dipole-induced rectification of interfacial CT [95], i.e. of CT across a junction between liquid and a solid conductor. A junction between solid conductors interfaced with assemblies of polypeptide helices extends the studies of such dipole effects onto charge-transport currents [129], [130], [131]. The use of polypeptides composed of residues with chargeable side chains such as lysine, however, raises question about the effect of the counterions in the SAMs on the measured rectification, which may depend on the state of protonation of the polypeptides. Furthermore, while the use of gold interfaces provides important fabrication advantages, it also poses a potential for forming conducting pillar-like nanostructures within the bioorganic SAMs due to the mobility of gold atoms under the large field gradients inherent for such junctions. Such artifacts would interfere with analysis of the measured results. Nevertheless, the facile interfacing of gold with organic conjugates has made it possible to realize molecular junction and advance the fields not only of interfacial charge transfer, but also of molecular charge transport.

A principal challenge of polypeptide CT systems composed of native amino acids is the inherent distance limitation for efficient CT. Along their backbones and hydrogen bonds, protein α-helices mediate ET via tunneling, the rate constants of which, k_et, fall off exponentially with an increase in the length of the ET pathways, r_et, i.e. k_et ∝ exp(−β r_et). Specifically, for protein α-helices, β is about 1.3 Å⁻¹ [21], [142], [143], [144]. The inherent presence of competing nanosecond processes, such as fluorescence, internal conversion and intersystem crossing, therefore, places a practical limit of about 2 nm for attaining photoinduced CT with acceptable efficiency in such bimolecular structures. Conversely, the inability of polypeptides, composed solely of native amino acids, to accept electrons or holes without undergoing reductive or oxidative degradation, prevents alternative mechanisms for achieving long-range CT. Another characteristic that limits the utility of biomolecules for materials applications is their conformational instability when placed outside of their native environment. Overall, the challenge for preserving the protein secondary structure, and thus the expected functionality, places additional constraints on the design and engineering of systems comprising biological and biomimetic macromolecular components.

Nevertheless, nature still provides some of the best examples for mediating efficient CT in low-dielectric-constant media, i.e. in proteins and across lipid bilayers. Proteins can mediate long-range CT via electron-hopping or hole-hopping along cofactors or redox-active residues (e.g. tyrosine and tryptophan [26]). Photosynthetic reaction centers and mitochondrial respiratory complex are some of the best examples for such electron-transfer chains [27].

Poly- and oligonucleotides present another example for mediating efficiently long-range CT at distances exceeding 2 nm [29], [30]. Short tunneling steps between the HOMOs of electron-rich bases in deoxyribonucleic acid (DNA) double strands lead to hole hopping that can be efficient at distances of several nanometers but limited by the dynamics (and the persistence length) of the macromolecule. Peptide nucleic acids (PNA) present a key alternative for achieving long-range hole hopping. As a result of their small helical twist, the PNA structure provides a better electronic coupling between neighboring bases in comparison with DNA [30]. Furthermore, the backbone of PNA contains secondary and tertiary amides. That is, unlike native polynucleotides, PNA strands do not have ionic charges along their backbones, which proves beneficial for examining dipole-generated local-field effects on CT kinetics, e.g. along single-strand PNA oligomers [145]. For observing dipole effects on long range CT mediated by such double-stranded biomolecules, however, the two biopolymers in the double helices have to be oriented in the same, rather than in opposite, direction. This structural requirement can place constraints on the utility of PNA as molecular electrets.

Bioinspired electrets

To utilize some of the best features of the natural systems, while overcoming their limitations, we undertake a bioinspired approach in the design of molecular electrets [7]. The basic motif of our design comprises a polypeptide backbone of ortho-aminobenzoic acids, i.e. of anthranilic acid derivatives. As we have demonstrated, these anthranilamide (Aa) structures are, indeed, molecular electrets [15], [146]. Much like protein helices, the Aa oligomers possess intrinsic dipole moments originating from ordered amide and hydrogen bonds along their backbones (Fig. 1) [15], [146]. Amides are polar groups with permanent dipoles of about 4 D [147]. Each amide bond, however, contributes about 1.8 D to the total dipole along the backbone of the Aa oligomer because of the angle between the amide dipoles and the backbone axis of the Aa polypeptide structures [15]. In addition, the polarization from the formation of each hydrogen bond contributes about an extra 0.9 D to the total Aa dipole along its main axis [15]. Conversely, while the intrinsic dipoles of protein α-helices and 3₁₀-helices point from their C- to their N-termini, the dipoles of Aa oligomers point from their N- to the C-termini, similar to those of polyproline type I helices, (Fig. 1b) [12], [15].

The pattern of directly linked aromatic moieties along the Aa backbone is the most important feature that sets apart the bioinspired Aa electrets from any protein structure. The aromatic moieties can act similar to the bases in oligonucleotides and provide pathways for long-range CT via a hopping mechanism. The two side chains of the non-native Aa residues, R₁ at position 4 and R₂ at position 5, provide a means for adjusting their electronic properties, and hence, for tuning the CT mechanism. Indeed, such adjustments of the electronic properties of the Aa structures by using different side chains also affect their dipole moments.

While the different polypeptide helical structures of native α-amino acids possess intrinsic dipoles ranging from about 1.5 to 5 D per residue [12], the dipole of Aa oligomers composed of Ant, i.e. R₁=R₂=H (Fig. 1a), is about 2.7 D per residue. Adding electron-donating groups as an R₁ side chain causes polarization of the Aa residue, opposing the intrinsic dipole and decreasing it. Similarly, the polarization by an electron-donating group as an R₂ side chain increases the total dipole [15]. For example, the dipole of an oligomer of aminated Aa residues [R₁=H and R₂=N(R′)R″] is about 3.9 D per residue [15]. That is, placing an amine at position 5 causes more than a 40% increase in the total dipole.

The Aa dipoles affect the kinetics of CT. As we demonstrated for the first time, a single Aa residue rectifies not only the photoinduced CS, but also the consequent charge recombination [125]. For dyads composed of an Aa residue as the electron donor and a pyrene (Py) as the acceptor, CS is faster when the electron moves along the dipole (i.e. toward its positive pole) than when it moves against it (i.e. toward its negative pole), which perfectly agrees with the accepted notion for the dipole-induced Franck-Condon effect on the CT kinetics (Fig. 2) [12], [118]. The CR was also faster when the electron moved along the dipole than against it. While this result for CR might seem like an intuitively expected outcome, it contradicts the same accepted notion for the dipole effect on the Franck-Condon component of the CT kinetics.

Fig. 2:

Electron donor-acceptor dyads of an anthranilamide (Aa), 5Pip (Fig. 1c), as a donor, and pyrene (Py) as an acceptor [125]. (a) The acceptor is attached to the N- or the C-terminal amide of Aa, so that the electron moves along the dipole during CS in Aa-Py and CR in Py-Aa; and against the dipole during CS in Py-Aa and CR in Aa-Py. (b) Jablonski diagram for the Aa-Py and Py-Aa dyads, showing the effect of the permanent Aa dipole on the photoinduced CS and CR driving forces, −ΔG⁽⁰⁾.

During CS in the Aa-Py dyad, the electron moves along the Aa dipole forming an Aa˙⁺-Py˙⁻ CS state polarized oppositely to the permanent Aa dipole. Conversely, in the Py-Aa dyad, the electron moves against the dipole forming a Py˙⁻-Aa˙⁺ CS state polarized in the same direction as the Aa dipole (Fig. 2a). Therefore, the Aa dipole stabilizes the Aa˙⁺-Py˙⁻ CS state, lowering its energy, i.e. increasing the CS driving force, −ΔG_CS⁽⁰⁾, and increasing the CS rate when the electron moves along the dipole (Fig. 2b). In the same manner, the electret dipole destabilizes the Py˙⁻-Aa˙⁺ CS state, decreasing the −ΔG_CS⁽⁰⁾ and lowering the CS rate when the electron moves against the Aa dipole (Fig. 2b). For this dipole effect on the CS kinetics, the ET processes operate in the Marcus normal region.

Conversely, for CR, one ought to expect an opposite outcome (i.e. larger rates for an electron moving against the dipole than along it) because the large driving forces place CR mediated by these dyads in the Marcus inverted region [125], [148]. That is, when the polarization of the CS state opposes the Aa permanent dipole, i.e. when the CS state is stabilized, the CR driving force, −ΔG_CR⁽⁰⁾, decreases (Fig. 2b). For −ΔG_CR⁽⁰⁾ larger than the reorganization energy, Aa˙⁺-Py˙⁻, should undergo faster CR than Py˙⁻-Aa˙⁺, which is not the case.

Indeed, Aa-Py undergoes faster CS and slower CR than Py-Aa; and this trend has an excellent implication for the utility of molecular electrets, i.e. accelerating CS while suppressing CR. Nevertheless, the accepted notion for dipole effects on ET kinetics, focusing on the thermodynamic driving forces, cannot solely predict the trends observed for CR in these electret dyads. Therefore, we focus on the non-Franck-Condon aspects of the ET kinetics.

A further analysis reveals that differences in the electronic coupling between the donor and the acceptor cause the discrepancy in the CR kinetics observed for the Aa dyads [125]. For cases with small driving forces, the Franck-Condon effects dominate. Conversely, when the driving forces are large, the differences in the electronic coupling prevail. This synergetic control of the nuclear and the electronic components of the CT kinetics provides new principles for molecular design and engineering of electronic and energy-conversion systems.

Synthetic proteome for molecular electrets

In nature, the 22 proteogenic native amino acids are the key building blocks for proteomes that reflect the amazing diversity of life on Earth [149], [150], [151], [152], [153], [154], [155], [156], [157], [158], [159], [160]. Proteogenic residues are α-L-amino acids, differing only by a single side chain connected to their α-carbons, that can be genetically coded and expressed in a cell proteome (Fig. 3).

Fig. 3:

The 22 proteogenic native amino acids [150], [151], [152], [153], [154]. The state of protonation is depicted for physiological pH.

The bioinspired molecular electrets are polypeptides composed of non-native β-amino acids. Each of the electret residues has two side chains, R₁ and R₂, on the distal positions 4 and 5, respectively. In similarity with the native biomolecular structures, a set of a dozen or a couple of dozen non-native Aa amino acids (Fig. 1c) can provide countless CT functionalities when combined in different primary sequences. Indeed, the partial π-conjugation through the amide bonds, connecting the aromatic moieties, provides a strong electronic coupling needed for the short electron-hopping steps between the neighboring residues. The amide-mediated electronic coupling, however, is not strong enough to generate the formation of CT bands that can be detected in absorption or emission spectra, and the frontier orbitals tend to be relatively localized over individual residues [15], [146]. Therefore, characterizing the electronic and the photonic properties of individual Aa residues has a considerable predictive power of their characteristics when incorporate into Aa oligomers. The rational control of the primary sequence of the Aa oligomers provides a means for the design of CT pathways. Adding or removing charge traps and adding or removing potential barriers are some of the unexplored possibilities for the Aa electrets.

Much like amino acids, varying the two side chains opens up possibilities for making a diverse set of non-native amino acids for a “synthetic proteome” of molecular electrets with wide range of electronic properties and CT capabilities. Using three types of electron-donating substituents (alkyl, alkoxyl and amines), and placing them at the two distal positions, we vary the reduction potential for the oxidation of these residues over the range of one volt (Fig. 1d). It corresponds to adjusting the energies of the HOMOs of these residues between about −6.5 and −5.5 eV vs. the vacuum level (Fig. 1d), which give us an attractive range for hole-conducting materials. Another interesting finding involves a fine-tuning of the reduction potentials of the Aa residues by changing the positions of their side chains. It is important not only the type of a side chain (e.g. alkyl, alkoxyl and amine), but also where the side chain is located (i.e. position 4 vs. position 5). For example, shifting an amine substituent from position 4 to position 5 causes about a 200 mV negative shift in the reduction potential of oxidation.

A word of caution: published reports may casually refer to a “reduction potential of oxidation” as an “oxidation potential.” While this terminology is commonly used, it is fundamentally incorrect. The accepted nomenclature reports electrochemical potentials as “reduction potentials” [161], [162], i.e. the potentials for adding electrons, M+n e–⇄Mn–. Oxidation potentials, conversely, represent the opposite process, Mn–⇄M+n e–, and, thus, have opposite signs [163]. Therefore, reduction potentials characterize both oxidation and reduction processes. It is straightforward to call “reduction potential” the potential at which the reduction of species M with charge z occurs, i.e. Mz+n e–⇄Mz–n. For the oxidation, however, Mz+n+n e–⇄Mz, it is incorrect to call it an “oxidation potential.” The measured values represent the “reduction potential” of the oxidized species, M^z+n . Therefore, the reduction potentials of the one-electron oxidation of the Aa residues represents the reduction potentials of their radical cations, Aa˙⁺. Therefore, a decrease in the reduction potential of oxidation, i.e. a negative shift of its value, indicates a rise in the energy level of the HOMO, and improved electron-donating capabilities.

When designing hole-transfer systems, it is key to utilize components that can hold positive charges (i.e. holes) without undergoing irreversible transformations. Therefore, using electrochemical and computational analysis, we discovered two requirements for an Aa residue to maintain its chemical stability under hole-transfer conditions: (1) the reduction potential for oxidizing it cannot be more positive than a cutoff potential of about 1.2 or 1.5 V vs. SCE for acetonitrile and dichloromethane media, respectively, and (2) the strongest electron-donating side chain of the residue should be attached at position 5, i.e. at the para position to the N-terminal amide, in order to prevent the spin density distribution of the radical cation from extending over the C-terminal amide (Fig. 4).

Fig. 4:

Examples of spin-density distributions of the radical cations of Aa residues (Fig. 1c), with N- and C-termini capped as alkylamides (black – excess spin up, i.e. radical cation; and white, excess spin down) [32], [33], [35]. The residues are grouped according to the reversibility of their electrochemical oxidation. The reduction potentials for the oxidized Aa residues for the neat dichloromethane (DCM) and acetonitrile (MeCN) are obtained from extrapolation to zero electrolyte concentration [36], [37], [38], [39].

The development of the electret “synthetic proteome” aims at diversity of electronic properties and charge-transfer functionalities. The position of the side chain of the Aa residues has a compounded effect on these key features: it affects their total electric dipoles, reduction potentials and stability against oxidative degradation. Indeed, multifaceted analyses of structure-function relationships and holistic design approaches are essential for the development of the Aa bioinspired molecular electrets. This complexity offers unexplored possibilities for the search of emerging properties essential for electronic materials and energy-conversion systems.

Backbone vs. side-chain amides as sites for hole injection

The injection of a hole into an Aa residue involves an extraction of an electron from its HOMO. Indeed, “wiring” an Aa oligomer via its backbone, which involves attaching electrodes or an auxiliary donor and acceptor to the C- or N-terminal amides, appears structurally as quite a feasible approach. For chemical stability, however, the spin-density distribution of the radical cation of a residue, Aa˙⁺, which closely resembles the distribution of the HOMO of the ground-state electroneutral species, Aa, should not extend over its C-terminal amide (Fig. 4) [32], [33]. Therefore, a covalent attachment via the N-terminal amide of such Aa moiety provides a stronger electronic coupling for hole injection than attachment via the C-terminal amide.

For example, in a Feb dimer the HOMOs are predominantly localized over each of the amino acids (Fig. 5). The intrinsic dipole removes the degeneracy between the frontier orbitals of neighboring residues, making the one localized on the N-terminal Feb a HOMO and one on the C-terminal Feb a HOMO-1 (Fig. 5b,c). The HOMO extends over the N-terminal amide, ensuring a good electronic coupling with moieties attached to it. Conversely, the HOMO-1 does not truly extend over the C-terminal amide, compromising the electronic coupling needed for hole injection through it. This difference between the quality of the electronic coupling via the C-terminal and N-terminal amides can prevail over the dipole effects on HT kinetics. Furthermore, hole injection into the residue on the C-terminal end of an Aa oligomer provides the benefits of the dipole-enhanced rates of HT toward the N-terminus because the hole travels preferentially toward the negative pole of the dipole. Yet the small electronic coupling via the C-terminal amide does not appear to provide the best pathway for CT within these molecular electrets.

Fig. 5:

Dimer of Aa residues, Feb–Feb, with its highest occupied molecular orbitals. (a) Structure of Feb–Feb with indication of the points for hole injection via its C- and N-termini. (b) The highest occupied molecular orbital (HOMO) of Feb–Feb is localized on its N-terminal residue and extends over the N-terminal amide providing a good electronic coupling for hole injection (i.e. extraction of an electron from the HOMO) at that site. (c) The second highest occupied molecular orbital (HOMO-1) of Feb–Feb is localized on its C-terminal residue and does not extend over the C-terminal amide.

An examination of the HOMO distributions reveals that it pronouncedly extends over the R₂ side chains at position 5 (Fig. 5b,c). Therefore, a hole injection through a side chain at position 5 on the C-terminal residue, rather than through the C-terminal amide, provides a strong electronic coupling for initiating HT toward the negative pole of the Aa oligomer dipole (Fig. 1a). That is, “wiring” the Aa oligomers via the side-chain at position 5 of the C-terminal residues will contribute beneficially to the electronic coupling and Franck-Condon components of the HT kinetics (Fig. 2).

To eliminate the difference in the electronic coupling when injecting holes in the C- and N-terminal residues, we resort to a 5-amidoanthranilamide (Aaa) derivative with an amide side chain connected via its nitrogen to the carbon 5 of the aromatic ring of the amino acid (Fig. 6). Due to their partial π-conjugation between the nitrogen and the carbonyl carbon, amide bonds tend to be planar with their trans conformers more stable than their cis ones. Strictly speaking, despite its common use for amide and peptide bonds, cis/trans nomenclature is not quite applicable for these partially π-conjugated systems because the four substituents (on the nitrogen and on the carbonyl carbon) are different. The relative position of the two carbons, attached to the nitrogen and to the carbonyl carbon, provides the basis for assigning the isomers of secondary amides as cis or trans. Conversely, according to the applicable E/Z nomenclature, a trans secondary amide is Z and the cis one is E.

Fig. 6:

Electronic features of Aaa obtained from DFT calculations, optimized for chloroform media. (a) Structures of the two conformers of Aaa with different orientation of the side chain amide at position 5. E and Z conformers are ascribed according to the partially conjugated bond between the aromatic carbon at position 5, C⁽⁵⁾, and the amide nitrogen of the side chain, N⁽⁵⁾. (b) Spatial distribution of the HOMOs of the conformers, with indication of the sites for hole injection. (c) Optimized structures of the conformers with their permanent dipole moments. (d) Spin density distribution (SDD) for the radical cations of the two conformers (black – excess spin up, i.e. radical cation; and white, excess spin down).

All three amides in Aaa are in their Z conformations. The hydrogen bonding between the proton of the N-terminal amide and the oxygen of the C-terminal locks the conformation of the amides on positions 1 and 2 in relation to the plane of the aromatic ring. Partial π-conjugation with the carbon 5 of the aromatic ring ensures a planar conformation in that region of the Aaa molecule. That is, the amide side chain, R₂ at position 5, is preferentially in the same plane with the aromatic ring, providing possibility for E and Z conformers determined by the N⁽⁵⁾–C⁽⁵⁾ bond (Fig. 6a).

For chloroform, DFT calculations reveal that the spatial HOMO distribution is quite similar for both, the E and Z, conformers (Fig. 6b). Furthermore, the HOMOs extend over the N-terminal and the side-chain amide, but not over the C-terminal one (Fig. 6b). Therefore, hole-injection pathways via the N-terminus and via the side chain should provide stronger electronic coupling than via the C-terminus.

Conversely, the E conformer of Aaa has a larger permanent electric dipole, 10.7 D, than the Z conformer, which is 4.57 D (Fig. 6c). Also, the Z conformer is about 100 meV more stable than the E conformer, i.e. at room temperature, only about 2% of Aaa exists as the E chloroform when dissolved in chloroform. Proton NMR spectrum of Aaa for CDCl₃ does not reveal any splitting of the side-chain amide proton, a_S, or any other protons (Fig. 7). It indicates that either Aaa exists predominantly as only one of the two conformers, or the transition between the E and Z isomers is fast on the NMR acquisition time scales at room temperature due to relatively low activation energy. Indeed, the former appears consistent with the computational findings.

Fig. 7:

One-dimensional ¹H NMR spectrum of Aaa (500 MHz, CDCl₃), with the assignments of the protons in the molecular structure to the chemical shifts, and the measured relative integration shown under the principle axis. No splitting is apparent for the side-chain amide proton, a_S, at 7.73 ppm. Because two methylene protons c₂, n₂, and s₂, are diastereotopic, they exhibit different chemical shifts even when attached to the same carbon. The same diastereotopicity, however, does not cause detectable splits in chemical shifts of c₃, n₃, and s₃.

The spin-density distribution of the radical cation of Aaa does not extend over its C-terminal amide (Fig. 6d), which precludes that the oxidized form of this residue can be stable. The other requirement for attaining stability of Aaa˙⁺ is for its reduction potential to not exceed about 1.2–1.5 V vs. SCE (depending on the media). Electrochemical analysis reveals that Aaa oxidizes at 1.52 and 1.44 V vs. SCE for neat dichloromethane and acetonitrile (Fig. 8), respectfully, which is right at the threshold for which its radical cation may be unstable and undergo oxidative degradation [32]. The cyclic voltammograms exhibit chemical reversibility for the oxidation of Aaa in DCM and irreversibility in MeCN (Fig. 8a,b), which is consistent with our predictions about the border-line stability of Aaa˙⁺. While the reduction potential is less positive for Aaa oxidation in the more-polar solvent, MeCN, the positively charged radical cation is more stable in DCM, wich has a larger polarizability, α, than MeCN, i.e. α_DCM is about 6.5 ų, and α_MeCN – about 4.5 ų [164]. Conversely, MeCN is more polar than DCM, i.e. the Onsager function for solvent polarity, f_O(ε, n²), yields 0.61 and 0.43 for MeCN and DCM, respectively, where f_O(ε, n²)=f_O(ε)–f_O(n²), and f_O(x)=2(x−1)/(2x+1). Therefore, the stability of Aaa˙⁺ in MeCN and DCM should not differ much based on the solvent polarity and polarizability.

Fig. 8:

Electrochemical oxidation of Aaa. (a, b) Cyclic voltammograms of the first oxidation waves at different concentrations (C_el) of the supporting electrolyte, NBu₄PF₆., for (a) dichloromethane, DCM, and (b) acetonitrile, MeCN. (c) Dependence of the half-wave potentials, E^(1/2), on C_el, and extrapolation to zero electrolyte concentration to obtain the values of the reduction potentials for neat solvents [36], [37], [38], [39].

Another important characteristic of chlorinated solvents is their relatively high electron affinities in comparison with non-halogenated solvents. That is, DCM is a better electron acceptor than MeCN. Thus, DCM may provide a sufficiently oxidative environment to stabilize the Aaa radical cations. In addition, DCM also oxidizes at a more negative potential than MeCN. In comparison with MeCN, therefore, the HOMO of DCM is closer in energy to the SOMO of Aaa˙⁺, which may contribute to electronic stabilization of the solvated radical cation,

Indeed, the solvent polarity, polarizability, electron affinity and ionization energy may explain the reversibility of Aaa oxidation for DCM, while it is irreversible for MeCN (Fig. 8a,b). In our opinion, however, the difference between the proton affinities of MeCN and DCM can very well be the underlying reason for the observed solvent effects on the reversibility.

A loss of a proton is a common pathway for decomposition of radical cations [165]. As we reported, a loss of an amide proton from Aa˙⁺ is one of the pathways leading to oxidative degradation of Aa residues, resulting in irreversible voltammograms [32]. MeCN has a higher proton affinity (or basicity) than DCM. Hence, at the relatively high potentials for oxidation of Aaa, MeCN would aid the abstraction of a proton from the formed Aaa˙⁺. Based on the solvent basicity, therefore, the propensity for oxidative degradation of Aaa should be higher for MeCN media than for DCM solutions, which is consistent with the reversibility patterns (Fig. 8a,b).

The photophysical properties of Aaa show small to practically negligible solvent dependence (Fig. 9, Table 1), which is similar to what we observe for the other Aa residues [31], [34]. We focus on aprotic solvents to preserve the integrity of the hydrogen bond between the N-terminal amide proton and the C-terminal carbonyl oxygen. Aaa is a moderately strong fluorophore with emission quantum yield of about 0.1–0.15 and excited-state lifetime ranging between 1.5 and 2 ns for the different solvents (Table 1).

Fig. 9:

Steady-state absorption and fluorescence of Aaa (λ_ex=325 nm). The molar extinction coefficients for the different solvents are calculated from a series of absorption measurements of Aaa samples with concentrations between 10 and 200 μM. The linear dependence of the measured absorbance on the sample concentration suggests no aggregation (within the examined concentration range).

Overall, Aaa presents a good choice for hole injection in Aa oligomers: i.e. via its N-terminal amide (when Aaa is at the N-terminus of an oligomer), and via its side-chain amide (when Aaa is at the C-terminus of an oligomer). While the radical cation of Aaa exhibits border-line stability, it should be emphasized that this characterization is based on electrochemical oxidation. For charge-transfer systems, such as the Aa molecular electrets, mediating hole-hopping in the pico and subpicosecond time scales, the residence of the positive charges on the terminal Aaa residues will be at least 10 orders of magnitude shorter than the time scales for cyclic voltammetry measurements. Therefore, Aaa presents a nice compromise in electret design, providing optimal electronic coupling for the most important charge-separation (or hole-injection) step and acceptable stability for mediating the hole transfer to the neighboring anthranilamide residue.

Materials

General synthesis information

All chemicals were used as received unless otherwise specified. Aaa, i.e. N,N′-(2-(heptan-4-ylcarbamoyl)-1,4-phenylene)bis(2-propylpentanamide), was prepared in three steps (Scheme 1), employing adopted procedures for making anthranilamide derivatives [31], [33].

Scheme 1:

Synthesis of Aaa.

The ¹H NMR and ¹³C NMR spectra were recorded on 125, 300 or 500 MHz spectrometers at ambient temperature using CDCl₃ or DMSO-d₆ as solvents. Chemical shifts are reported in parts per million relative to the solvent peaks of CHCl₃ in CDCl₃ (¹H, δ=7.24), CDCl₃ (¹³C, δ=77.23), DMSO-d₅ in DMSO-d₆ (¹H, δ=2.50) or DMSO-d₆ (¹³C, δ=39.51). Data for ¹H NMR are reported as follows: chemical shift, integration, multiplicity (s=singlet, d=doublet, t=triplet, q=quartet, p=pentaplet/quintet, h=hextet/sextet, e=eptet(from επτά)/heptet m=multiplet), integration and coupling constants. All ¹³C NMR spectra were recorded with complete proton decoupling. High- resolution mass-spectra were obtained on a Q-TOF mass spectrometer. Analytical thin layer chromatography (TLC) was performed using 0.25 mm silica gel 60-F plates. Flash chromatography was performed using 60 Å, 32–63 μm silica gel. Yields refer to chromatographically pure materials, unless otherwise stated.

Methods