1.

Introduction

This article describes the design rationale for oligomers and polymers with large permanent electric dipoles, i.e., molecular electrets, composed of anthranilamide (Aa) residues (Fig. 1).^1–3 A comparison with biopolymers illustrates the basis for referring to the Aa molecular electrets as bioinspired. For improved charge-transfer (CT) properties, the Aa electrets incorporate features from different biomolecular structures to form non-native systems beneficial for electronic and photonic materials for energy applications.

Fig. 1

Bioinspired molecular electrets composed of anthranilamide (Aa) residues: (a) a segment of a molecular electret—${\mathrm{Aa}}^{(\mathrm{i})}-{\mathrm{Aa}}^{(\mathrm{j})}-{\mathrm{Aa}}^{(\mathrm{k})}-{\mathrm{Aa}}^{(\mathrm{l})}-{\mathrm{Aa}}^{(\mathrm{m})}-{\mathrm{Aa}}^{(\mathrm{n})}$; (b) origin of the Aa dipole from ordered orientation of amide bonds and from the polarization upon hydrogen bonding.

Understanding and controlling CT at a nanometer scale is paramount for solar-energy harvesting, conversion, and storage.^4–8 Local electric fields, originating from molecular dipoles, can strongly affect the behavior of charge carriers. Molecular-scale engineering of field profiles along CT pathways provides an incomparable means for controlling the direction and the efficiency of electron and hole transduction. Therefore, molecular electrets can play a central role in the bottom-up designs of hybrid electronic materials and energy-conversion devices.

Oligomers and polymers of aromatic $\beta$-amino acids, i.e., Aa residues, have two uniquely combined features: (1) ordered amide and hydrogen bonds resulting in electric dipoles of about 3 D/residue, oriented from the N- to the C-termini (Fig. 1)⁹ and (2) extended $\pi$-conjugation along the macromolecular backbone that can prove essential for mediating long-range CT.¹⁰ Our computational and experimental studies demonstrated that Aa conjugates possess permanent electric dipoles, i.e., they are indeed molecular electrets.^2,9 Comparing the extent of deshielding (i.e., relative acidity) of the amide protons as determined from nuclear magnetic resonance (NMR) studies provided experimental confirmation that the total molecular dipoles are oriented form the N- to the C-termini of the Aa conjugates.²

We have demonstrated that an Aa residue can act as a molecular rectifier (or a diode), showing preference of electron transfer (ET) toward the C-terminus.¹ This emphasis on rectification of CT, rather than charge transport, allows for assessing the effects induced by small organic residues or short oligomers without the interference from the contacts with conducting and semiconducting substrates.^11,12 Interfacial charge transport and CT “hold” the key for improving energy-conversion efficiencies of materials and devices.^4,13,14 Molecular-level understanding of charge-transfer rectification, implemented with a rational control of surfaces and interfaces,^5,15–17 provides the venues for such efficiency improvements.

The magnitude of CT rectification for a single Aa residue is comparable with the rectification exerted by peptide helices comprising more than 10 amino acid residues.¹ The molecular dipoles have a dominating contribution toward this rectification effect. The asymmetry in electron-density distribution, induced by the positions of the substituents, and the molecular dynamics also contribute to the CT processes. The latter affects the photoinduced charge separation kinetics, while the former affects the consequent charge recombination.¹

To truly benefit from the promising CT properties of Aa, it is essential to achieve structural diversity of these electret conjugates, leading to wide variety of electronic functionalities. Extending principles from proteomics to the design of widely diverse molecular electrets places a demand on developing sets of their building blocks, i.e., the non-native Aa residues, with different electronic properties.

Herein, we describe the rationale behind bringing structural features from different classes of biomacromolecules (i.e., proteins and nucleic acids) to design Aa molecular electrets in the search of improved CT properties that are not inherent for living systems. We focus on a set of Aa residues that are chemically modified with electron-donating substituents (at $R_1$ and $R_2$, Fig. 1) for lowering the ionization energies and improving hole-transfer capabilities. Replacing the hydrogen at $R_2$ with three types of substituents with different electron-donating propensity causes negative shifts in the Aa reduction potential, amounting to about 0.1, 0.3, and 1 V. Moving an electron-donating group (EDG) from $R_2$ to $R_1$ results in a positive shift in the reduction potentials of about 0.1 to 0.3 V, which is indicative of an increase in the ionization energy. Discussion of their rectification and other CT properties illustrates the potential impact that the diversity in Aa molecular electrets can have on electronic materials and energy conversion.

2.

CT Molecular Electrets: Practical Ideas from Structural Biology

Electrets are dielectrics with permanent electric polarization. The polarization can originate from volume-embedded or surface-trapped charges (i.e., real-charge electrets), or from codirectionally ordered electric dipoles (i.e., dipole-polarization or dipolar electrets).^18,19 The lack of directionality of electrostatic interactions and ionic bonds governs the dynamics and the complexity in the behavior of real-charge electrets that encompass “space-charge” and “surface-charge electrets.”^20,21 Conversely, the defined directionality of covalent bonds (which, indeed, are also electric in nature) provides a key advantage for arranging noncharged molecular dipolar moieties. For molecular dipoles, patterns of covalent bonding keep the displacement of the centers of positive and negative charges more or less permanently fixed. Therefore, dipolar electrets, which are electrostatic analogues of magnets,²² provide an important venue for the design and development of macromolecules and materials for electronic and energy applications.

Biology presents some of the best examples of molecular electrets and the utility of their electronic properties. Peptide bonds are aliphatic amides (with permanent dipoles of about 4 D)²³ that hold together the protein backbones. In protein $\alpha$-helices, the peptide bonds are codirectionally oriented along the helix axes (Fig. 2), resulting in substantial electrical polarization of these macromolecular secondary conformers,^24–26 i.e., protein $\alpha$-helices are biomolecular electrets.

Fig. 2

Protein $\alpha$-helix, ribbon diagram and ball and stick model with highlighted hydrogen-bond network.

A network of hydrogen bonds, which are also aligned along the helix axes supporting these protein structures (Fig. 2), further enhances the total macromolecular dipoles of protein α-helices. The hydrogen bonding between the carbonyl oxygens and amide hydrogens from peptide bonds on neighboring helix turns causes a collective shift in electron density from the oxygens to the hydrogens, i.e., it causes polarization that is in the direction of the amide dipoles. This effect, however, does not result in negatively charged hydrogens and positively charged oxygens. Rather, the oxygens in the peptide bonds are negatively polarized and upon hydrogen bonding, some of this partial negative charge shifts toward the N–H bond of an amide on the neighboring turn. The ordered hydrogen and peptide bonds in the $\alpha$-helices result in their intrinsic electric dipoles of about 5 D/residue, pointing from the C- to the N-termini (Fig. 2).²⁷

Other protein helices also exhibit dipole-induced polarization. Similar to $\alpha$-helices, alignment of hydrogen and amide bonds in $3_{10}$-helices results in intrinsic electric dipoles (4.5 D/residue) pointing from the C- to the N-termini.²⁷ In fact, $3_{10}$-helices are tight-wound analogues of $\alpha$-helices: i.e., $3_{10}$-helices contain three residues per turn and each hydrogen bond closes a backbone loop comprising 10 atoms. Indeed, protein $\alpha$-helices are ${3.6}_{13}$ helices.

Polyproline helices exhibit another interesting set of electronic properties. Proline is a native amino acid with a rigid structure and a secondary $\alpha$-amine. Therefore, polyprolines assume helical conformations without hydrogen bonding, making their intrinsic dipoles smaller than the dipoles of $\alpha$-helices.²⁷ In the two types of polyproline helices, type I and type II, the peptide bonds have different conformations and their dipoles are orientated differently. Type I comprises “cis” amides inducing a ground-state dipole of 4.1 D/residue (pointing from the N- to the C-terminus), and type II with “trans” amides has a dipole of 1.5 D/residue (pointing from the C- to the N-terminus). Therefore, conformational changes back and forth from type I to type II polyproline helices switch the direction of their polarization,²⁷ which is a unique feature of these biomolecular electrets.

The dipole-generated electric fields in the proximity of protein helices amounts to about ${10}^8$ to ${10}^9\mathrm{V}/\mathrm{m}$, which proves important for ET and ion-transport processes in living organisms.^28–31 In synthetic polypeptide helices, the intrinsic dipoles rectify the directionality of CT.^32–37 The broadband gap [or highest occupied molecular orbital (HOMO)–lowest unoccupied molecular orbital (LUMO) gap] of proteins, however, along with the inability to inject charges in them without chemically cleaving their backbones,^38–40 considerably limits their utility as electronic materials. Proteins and polypeptides (composed of native $\alpha$-amino acids and their derivatives) mediate ET solely via tunneling, the efficiency of which is limited to about 2 nm.^41–46 Conversely, arrays of cofactors and redox-active residues, such as tryptophan or tyrosine, make it possible to attain long-range ET.^47,48 That is, multiple short and efficient “electron hopping” steps along arrays of redox moieties allow ET to exceed the 2-nm tunneling limit.^48–52

Deoxyribonucleic acid (DNA) strands represent a class of biopolymers that can mediate CT at distances exceeding 2 nm.^53–57 Short efficient tunneling steps between closely situated bases, with relatively negative reduction potentials of their radical cations, mediate hole hopping. The lack of detectable distance dependence of the CT rates (for a donor–acceptor separation larger than about 1 nm) is an indication for the hopping mechanism. (In a case of tunneling, the CT rates, $k_{\mathrm{CT}}$, fall off exponentially with donor–acceptor distance, $r(\mathrm{DA})$, i.e., $k_{\mathrm{CT}}\propto {\epsilon }^{-\beta r(\mathrm{DA})}$, while in a case of hopping, they fall off as an inverse of the number of hopping sites, $N$, i.e., $k_{\mathrm{CT}}\propto 1/N^{\eta }$.)^49–51

As an alternative to DNA, peptide nucleic acid (PNA) strands present certain advantages for long-range hole hopping.^58–61 The smaller helical twist of PNA, in comparison with DNA, improves the electronic coupling between neighboring bases.⁵⁸ Furthermore, PNA derivatives are not acids (despite the name) and they do not contain ionic charges along their backbones. Instead, PNA backbones comprise primary and secondary amides resulting in the intrinsic electric dipole reported for single-stranded structures.⁶² The double-stranded PNA, however, has certain conformational rigidity and an advantage for CT over the single-stranded one. To attain permanent intrinsic dipoles in double-stranded structures, the PNA chains have to assemble in a parallel manner. Conversely, the antiparallel double-stranded PNA is slightly more stable than the parallel one.^63,64

Considering the electronic features of the various natural macromolecular conjugates, we undertake a bioinspired approach to the development of molecular electrets that can mediate CT (Fig. 1).³ Our de novo designs are based on derivatives of anthranilic acid, which is an aromatic β-amino acid. Similar to protein $\alpha$- and $3_{10}$-helices, Aa has ordered arrangements of amide and hydrogen bonds that result in permanent dipoles in the order of 4 to 5 D/residue pointing from the N- to the C-terminus (Fig. 1).^2,9 Unlike protein helices, however, the aromatic residues can provide sites for charge hopping along the Aa backbones. In DNA and PNA, the bases providing the charge hopping sites are electronically coupled via noncovalent face-to-face $\pi$-stacking. In Aa, the aromatic moieties compose the backbones and the partially $\pi$-conjugated covalent amide linkers ensure considerable overlap between the frontier orbitals on the neighboring residues.^9,65 Furthermore, the ability to adjust the electronic properties of each residue by synthetically altering the distal substituents ($R_1$ and $R_2$, Fig. 1) provides an incomparable unique means for tuning the CT mechanism from tunneling to hole hopping and to electron hopping.

3.

Principles of Biodiversity Extended to Electronic Materials

The diversity of life on Earth is made possible through a variety of protein structures and functionalities that stem from only 20 common native $\alpha$-amino acids. The sequence of amino acids in a protein chain directs the folds toward the preferred secondary conformations, tertiary structures, and quaternary self-assemblies that result in various protein functionalities. Cofactors and “noncommon” proteinogenic amino acids, such as selenocysteine and pyrrolysine, may appear as the centerpieces of observed protein activity. It is the whole protein microenvironment, however, with nanometer scale structural features that tunes the vital functionalities. Overall, the properties of the single side chain of each residue and its place in the protein sequence determine the emergence of the structure–function relationships at a macromolecular level. That is, altering the sequences comprising a few “common” residues that differ only by a single side chain leads to countless functionalities.

Can electronic and photonic material achieve the same diversity by employing similar principles? The bioinspired molecular electrets are composed of $\beta$-amino acids, i.e., Aa residues (Fig. 1). Each Aa residue has two side chains, $R_1$ and $R_2$ (Fig. 1), that can be synthetically modified (Fig. 3). The availability for variations at two sites, $R_1$ and $R_2$, provides a certain advantage over the native $\alpha$-amino acids that contain only a single side chain each.

Fig. 3

Aa residues, with alkyl-terminated C- and N-termini, used for spectroscopic and electrochemical studies.

For hole-transducing molecular electrets, we prepared Aa residues with EDGs as the $R_1$ and $R_2$ substituents (Fig. 3). Replacing the hydrogen at $R_2$ with a strong EDG, such as dialkylamine, causes about 1 V negative shift in the reduction potentials [compare 5Pip and Ant, Figs. 3 and 4(b)]. This finding is in an agreement with the 1-eV elevation of the energy levels of the HOMOs induced by similar dialkylamine substituents,⁹ i.e., such EDGs considerably improve the capabilities of the Aa residues to mediate hole transduction. The electrochemical reduction potentials of singly oxidized organic species, i.e., of ${\mathrm{Aa}}^{•+}+e^{-}\to \mathrm{Aa}$, correlate with their ionization energies and hence, with the energy levels of their HOMOs.^66–68 Alkyl and alkyloxy substituents that are weaker EDGs than the amines also cause negative shifts in the reduction potential of Aa but not as large as the one observed for $R_2=\text{dialkylamine}$ [Fig. 4(b)].

Fig. 4

Photophysical and electrochemical properties of different Aa residues (Fig. 3). (a)  UV/visible absorption and fluorescence spectra of 5Pip for dichloromethane (DCM) (for fluorescence, ${\lambda }_{\mathrm{ex}}=310\mathrm{nm}$). The wavelength, ${\lambda }_{00}$, at the crossing point between the intensity-normalized spectra provides an estimate for the zero-to-zero energy, ${\mathcal{E}}_{00}=hc/{\lambda }_{00}$, which represents the highest occupied molecular orbital (HOMO)–lowest unoccupied molecular orbital (LUMO) optical gap.^5,65,69–71 Inset: cyclic voltammogram of 5Pip for DCM in the presence of 200 mM tetrabutylammonium hexafluorophosphate ($\text{scan}\text{rate}=200\mathrm{mV}/\mathrm{s}$). (b) Reduction potentials of singly oxidized Aa residues, $E^{(0)}$ (${\mathrm{Aa}}^{•+}+e^{-}\to \mathrm{Aa}$), obtained from cyclic voltammetry for solvents with different polarity (as represented by the solvent static relative dielectric constants, $\epsilon$). The data points (the circles) represent $E^{(0)}$ for four neat solvents, obtained from extrapolation to zero electrolyte concentration.^72,73 The dashed lines are least-square data fits using $E=a+b/\epsilon$, as expected from the Born solvation energy.⁷⁴ (c) Zero-to-zero energy of Aa residues. The circles are data points for four different solvents and dotted lines are linear data fits. [For (b) and (c), the four solvents with different $\epsilon$ are DCM, benzonitrile, acetonitrile, and propylene carbonate.]

Another feature of the electronic properties of the Aa residues is the significance of the exact position of the substituent; i.e., it is important not only what the substituent is but also where it is in the aromatic ring, i.e., $R_1$ versus $R_2$. Moving dialkylamine from the $R_2$ to the $R_1$ position causes about a 0.2-V positive shift in the reduction potential [Fig. 4(b)], making the Aa residue a slightly worse mediator of hole transduction.

The optical properties of the residues also manifest a strong dependence not only on the type of substituent but also on its position [Fig. 4(c)]. Thus, a single substituent for the two side chains, $R_1$ and $R_2$, of the Aa residues provide a means for tripling the diversity in properties compared to native amino acids with single side chains. For an electron-donating or electron-withdrawing substituent, G, the three possible Aa residues, $R_1=\mathrm{G}$, $R_2=\mathrm{G}$ and $R_1=R_2=\mathrm{G}$, would have different electronic properties. This trend demonstrates one of the key advantages of bioinspired approaches versus the biomimetic and biological ones.³

An important feature revealed by the experimental findings is the difference in the extent with which the substituents affect the electrochemical and optical properties of the Aa residues. The reduction potentials of ${\mathrm{Aa}}^{•+}$, which correlate with the ionization energy of the residues, pronouncedly depend on the electron-donating strength of the substituents. The exact position of the substituents further tunes the ionization energy but it does not dominate the observed trend. That is, the negative shift in the reduction potential, $E^{(1/2)}(\mathrm{Ant})>E^{(1/2)}(\mathrm{Met})>E^{(1/2)}(\mathrm{Hox})>E^{(1/2)}(4\mathrm{Pip})>E^{(1/2)}(5\mathrm{Pip})$, follows the electron-donating strength of the substituent, $\mathrm{H}<{\mathrm{CH}}_3<{\mathrm{OC}}_6{\mathrm{H}}_{13}<\mathrm{N}{({\mathrm{CH}}_2)}_5$ [Fig. 4(b)]. The optical properties of the Aa residues, on the other hand, manifest a considerably stronger sensitivity to the position of the substituents than to their electron-donating strength or $\pi$-conjugating propensity. For the residues with substituents at position 5, Met, Hox and 5Pip (Fig. 3), the zero-to-zero energy, ${\mathcal{E}}_{00}$ (which is equivalent to optical band gap for molecular spectroscopy) decreases as $R_2$ alters from H to ${\mathrm{CH}}_3$, ${\mathrm{OC}}_6{\mathrm{H}}_{13}$, and $\mathrm{N}{({\mathrm{CH}}_2)}_5$ [Fig. 4(c)]. This trend follows the $\pi$-conjugating propensity of the substituents. Shifting the piperidine substituent from position 5 to 4 (i.e., 5Pip to 4Pip), increases ${\mathcal{E}}_{00}$ by about 0.6 eV, making it equivalent to ${\mathcal{E}}_{00}$ of Met [Fig. 4(c)]. Indeed, these trends are based on examining five conjugates, and their universality is still to be tested by expanding the set of Aa residues. Nevertheless, they reveal key design considerations for electronic and photonic organic materials.

Inducing CT rectification is the most important electronic feature of the molecular electrets. The intrinsic dipoles remove the degeneracy between the frontier orbitals on neighboring Aa residues resulting in “cascade” energy configurations for driving hole transfer or ET.⁹ As the length of the molecular electrets increases above 10 nm or so, the energy of each of the frontier orbitals (e.g., HOMOs and LUMOs) should asymptotically approach constant values toward one of the termini due to dipole-induced charge displacement, making the “cascade” CT profiles nonlinear.

Once charge carriers are injected in a sequence of Aa residues, the intrinsic dipoles can ensure a preference for hole transfer toward the N-termini along the HOMOs of the residues, and for ET toward the C-termini along the LUMOs. The capability of Aa conjugates to act as molecular rectifiers (or diodes) can prove immensely beneficial for enhancing forward ET or hole transfer and suppressing undesired charge recombination. These rectifying properties can be attained with homomeric molecular electrets based on sequences of the same Aa residue selected for its ionization energy or electron affinity.

Conversely, at donor–acceptor interfaces, diversity in the electronic properties of the Aa residues can significantly aid the efficacy of the initial processes that convert the absorbed light energy into charge carriers. Coulomb trapping of charge-separation states at donor–acceptor interfaces is a principal source of losses in organic photovoltaic and other electronic devices.^14,75–83 It also imposes a need for larger driving forces than that required by the fundamental limits (decreasing, for example, the open circuit voltages, $V_{\mathrm{OC}}$) in order to provide excess energy for the generated charges to escape the traps.^14,84–87 Even in the presence of favorably oriented dipoles, these trapping Coulomb interactions can extend over a nanometer or two from the donor–acceptor interfaces, and the depth of the Coulomb traps can be a few hundred millielectronvolts. Hence, because of the presence of such traps for the electret dipoles to exert a dominant effect, the electron or the hole has to migrate a couple of nanometers away from the countercharge at the donor–acceptor interface. Because the increment of Aa polyamides is about $0.45\mathrm{nm}/\text{residue}$, the first two-to-four residues from the donor–acceptor interface are most important for eliminating undesired Coulomb trapping. Considering, for example, hole hopping (i.e., on-resonance hole transfer) along Aa sequences, selecting the first few residues attached to the electron acceptor to have slightly higher ionization energies than the residues in the rest of the polyamide chain can decrease the depth of the Coulomb traps and even completely eliminate them. Molecular electrets with widely diverse sequences will allow for implementing this beneficial CT feature to different electronic materials and energy-conversion systems. In parallel with proteins and their amino-acid building blocks, such diversity in molecular electrets can originate from a few Aa residues with different ionization energies and electron affinities.

4.

Experimental Methods

5.

Conclusions

Inducing rectification and managing charge traps (and potential barriers along CT pathways) represent only a couple of examples of the promising features of molecular electrets. Sets of non-native Aa residues, therefore, can serve as principal tools for the development of widely diverse molecular electrets and provide venues for bottom-up designs of functional electronic and photonic materials for energy applications.