Controlling Charged State Colors in Triphenylamine-Based Anodically Coloring Electrochromes

Abstract

In this work, we have designed a series of anodically coloring electrochromic (ACE) molecules comprised of thioalkyl-substituted 3,4-ethylenedioxythiophenes coupled to triphenylamine units (EDOT-TPA) that vary in the position and degree of electron richness of substituents, which influences the molecules geometric, electrochemical, optoelectronic, and excited-state properties. We evaluated their redox properties and discovered that modulation of both the first and second oxidation potential, formation of the cation radical and dication, can be varied from 0.03 to 0.18 V and 0.32 to 0.46 V versus Fc/Fc⁺ respectively. For the first time in ACE-based molecular systems, we demonstrated the ability to vary the electrochemical potential separation between successive charge states, which is directly involved in the generation of color. We use the chemical oxidant, Fe(OTf)₃, to visualize the saturation and contrast of the vibrantly colored cation radical solutions at 1 equivalent, followed by a second equivalents that opens a new and differing set in the color palette for the dication state. Optical transitions were probed during electrochemical oxidation using an optically transparent thin layer electrode (OTTLE) demonstrating selective control in generating successive charge states. We couple our findings with Density Functional Theory (TD-DFT) simulations to show how modulating electron richness and steric interactions control the optical transitions. Specifically, excited state analysis is performed to elucidate how substituent identity affects the neutral, cation radical, and dication transitions in the visible and near infrared, and thereby the resulting color.

Graphical Abstract

The ability to vary the electrochemical potential separation between successive charge states, which directly influences the generation of color, is reported for the first time in anodically coloring electrochromic molecules (ACE). A series of thioalkyl-substituted 3,4-ethylenedioxythiophenes coupled to triphenylamine units with variation in position and degree of electron richness of substituents, demonstrate changes in geometric, electrochemical, optoelectronic, and excited-state properties.

Introduction:

Controlled manipulation and separation of charges in molecular systems are at the forefront of many emerging technologies such as electrochromic displays^1–3, artificial photosynthesis^4–6, solar energy conversion^7–10, and molecular switches^11–13. The charge state of a molecule governs its physicochemical properties, such as conformation and reactivity.¹⁴ Specific interest has been focused on electro-active compounds due to the ability to undergo multiple and consecutive redox processes, with each redox state being distinguished by a specific property. In the field of electrochromism (EC), by controlling the redox states and the ability to switch between charge states, a range of optical (color) states can be attained and implemented into device applications.

Among the most representative examples of organic EC systems are viologen compounds and their analogues due to their advantageous dicationic oxidation state, excellent electron affinity, stability, and electrochemically reversible redox activity.¹⁵ Viologens (V²⁺) are planar, colorless, and transmissive dicationic π-conjugated acceptors that undergo two successive one-electron reductions yielding cation radicals (V^•-) and a neutral quinoid form (V⁰), both of which are colored species.^16–18 Many synthetic modifications have been applied to increase redox stability such as bridging groups in the bipyridine bay and incorporating substituents.^19–21

An alternative and promising approach to attaining fully clear to vibrantly colored systems consists of anodically coloring electrochromics (ACEs) where there has been significant progress in understanding how absorption in the charged state is controlled.^22–24 In 2019, we demonstrated that cross-conjugation can controllably tune the electronic energy levels of the cation radical independently of the neutral state by adjusting the electron rich character of substituents at the meta position of a bi-aryl 2-thiomethyl-3,4-ethylenedioxythiophene coupled to a 4-methoxybenzene derivative (EDOT-Ph).²⁵ In 2022, we reported on an alternative structural motif consisting of a thioalkyl-substituted bis(3,4-ethylenedioxythiophene)-1,4-phenylene (BEDOT-Ph) with varying alkoxy group placements. This deepened our understanding of how to synthetically control the optoelectronic properties as a function of charged state to fine-tune hue and saturation for the color blue.²⁶ Furthermore, we recently expanded our approach to vary the electronic and structural effects of BEDOT-Ph derivatives through substituent alterations, allowing us to access colors not previously attainable in ACE molecules.²⁷

ACE molecules differ from viologens in charged state formation, as illustrated in Figure 1A; where viologens are cathodic in nature, ACE molecules are anodic. A neutral ACE molecule is colorless and transmissive in the neutral state. Upon oxidation, a colored species is generated when forming a cation radical. Further oxidation results in the formation of a dicationic charged species, leading to a new sequential color. In our previous investigations, we uncovered using cyclic voltammetry, that sequential oxidation processes representing cation radical and dication formation are close in potential or lack a wide charge state separation, as depicted in Figure 1B. A motivation behind this present study is to increase the distance between these redox processes, as depicted in Figure 1C, where two reversible active species are generated at well-separated oxidation potentials. Accessing this will allow for direct control of the sequentially charged species formation and, thus, color control.

Figure 1.

A) Redox transformations undertaken by a generic ACE construct where upon two successive one-electron oxidations to form a cation radical and dication charged species occurs. B) Cyclic voltammogram depicting a current problem associated with the generation of separate charge states when difference in redox potentials is small and overlap. C) Depiction of an ideal cyclic voltammogram where two reversible redox active species have a large separation in charge state formation and form at differing oxidation potentials.

As mentioned, the ability to tailor and control colors, switching from colorless and transmissive to a high-contrast color, is a desirable feature for many applications.^28–32 The coloration mechanism in an ACE electrochromic system can be probed by comparing the evolution of the spectral response. As shown schematically in Figure 2, as the applied potential increases, the transition associated with the neutral species (also known as the π-π* transition) represented by the blue dotted line fully in the ultraviolet, will gradually decrease in intensity and new absorption peaks ascribed to the cation radical (indicated by the blue connected line) evolve with a high energy (HE) transition emerging in the visible region and a lower energy (LE) transition in the NIR. Further incremental increase in potential will generate the dication, represented by the green absorption peak. In our previous investigations^26–27, having sequential redox features that are not well-separated resulted in the simultaneous formation of the cation radical and dication species at specific high oxidation potentials, resulting in both charged states forming. This, in turn, lead to the inability to control the color attained, causing a mixture or blend of two charged species which creates a new color.

Figure 2.

Schematic diagram illustrating the spectra of three states in a generic ACE molecule where the neutral species (dotted blue line) will have a strong absorption in the UV. Upon oxidation, two major absorbance peaks (HE and LE) are generated (solid blue line) associated with the cation radical having two primary energy transitions in the visible and near infrared (NIR) region. Upon further oxidation, a single absorbance peak attributed to the dication species evolves primarily in the visible region.

In this report, we introduce new ACE molecule designs comprised of a triphenylamine donor coupled to an EDOT. Our goals are to maintain a highly transmissive and colorless neutral state while gaining a systematic understanding of how to control and increase the potential separation ($\Delta E)$ of each successive charged state. Fundamentally, it is essential to comprehend the dynamics of multiple redox reactions and the associated structural evolutions during the redox reactions. In addition, by controlling the dynamic formation of charged species, the generation of vibrantly colored species can be achieved and expand our ACE color palette. We have synthesized five new ACE electrochromes, systematically varying electron richness and steric interactions that are observed computationally by DFT calculations and experimentally by single crystal X-ray diffraction (hereafter, SCXRD). This approach enabled us to access both a cation radical and dication state using two chemical oxidants and maintain control over redox properties, separation and control of charge species, along with enhancing the color contrast and saturation in the cation radical state. Computational models and calculations were generated using TD-DFT at the mPW1PBE/cc-PVDZ level. These results are used in tandem with those produced experimentally to comprehend how modulating electron richness and steric interactions affects system color. Specifically, excited state analysis is performed to elucidate how substituent identity affects the neutral, cation radical and dication energy peaks in the visible and near infrared, and thereby the resulting color.

Results and Discussion:

Chromophore Design Rationale

Figure 3 shows the structures for five unsymmetrical substituted bi-aryl chromophores, each possessing varying degrees of electron density and different steric and electrostatic interactions. These variations are achieved by strategically placing methoxy substituents in specific locations, starting with the adjacent phenylene coupled to the EDOT and extending towards the outer phenylene pendant rings. Detailed synthetic and characterization information can be found in the SI and in Figures S1–S30. In brief, the synthesis relies on a two-fold Stille coupling process between a mono-brominated triphenylamine derivative and a 2-methylthio-3,4-dioxy-5-tributyltin-thiophene, conducted in the presence of Pd(PPh₃)₄.²⁵ This process results in the assembly of discrete molecules with moderate final step yields ranging from 43–63%. The methylthio (MeS) 2’ end cap serves to stabilize the cation radical and dication states. This is accomplished by inhibiting polymerization while simultaneously reducing the oxidation potential of the conjugated system via the resonance-donating effect from the sulfur atom.^33–34

Figure 3.

Chemical structure of the five unsymmetrical substituted bi-aryl (2-MeS-EDOT-TPA) chromophores with varying degrees of electron density, steric, and electrostatic interactions. The shortcut designated name and number is highlighted yellow and will be used throughout the text.

Incorporating a triphenylamine unit is of particular interest due to its excellent electron-donating nature leading to ease of oxidation attributed to the conjugation of the nitrogen electron lone pair with the phenyl rings.³⁵ Additionally, there are prevalent strategies for stabilizing TPA generated cation radicals such as 1) adding substituents at the para-position and 2) planarizing the propeller-shaped TPA with a bridge atom. Stability is enhanced in the latter case because the N-centered cation radical is delocalized over the entire planarized π-system.^36–37 By using these strategies and using triphenylamine as a building block, we have an opportunity to gain a deeper understanding of the electronic structure and the generation of charged species.

Structural Analysis:

To elucidate the impact of structural modification on the properties of these systems, DFT calculations at the mPW1PBE/cc-PVDZ level (previously the best pairing from laborious benchmarking) with inclusion of dichloromethane through the conductor polarizable continuum model (CPCM) was performed.³⁸ Optimized geometries for the neutral, cation radical as well as the dication were generated and analyzed. Figure 4, shows the results of these optimizations along with the stabilizing (S▪▪▪O, H▪▪▪O) and destabilizing (O▪▪▪O) non-covalent bonding interactions which were less than or equal to the van der Waals radii (O▪▪▪H ~2.60 Å, O▪▪▪S ~3.3 Å, O▪▪▪O ~2.8 Å) are shown.³⁹

Figure 4.

The neutral (0), cation radical (+1) and dication (+2) geometric results calculated for all TPA systems with representative non-covalent bonding interactions provided in constructs above table.

In all cases, oxidation produces a more planar geometry, which is evidenced by the dihedral angles as well as reducing the distance between the atoms involved in non-covalent bonding. The planarity ranking was TPA 3 > TPA 2 > TPA 1 ~ TPA 5 > TPA 4. The unsubstituted central phenyl ring for both TPA 1 and TPA 5 produced nearly the same central geometries. On the other hand, the torsional strain provided by the two methoxy groups next to the EDOT ring coupled with the destabilizing O▪▪▪O interaction resulted in a large out-of-plane twist for TPA 4. Both TPA 2 and TPA 3 produced nearly identical geometries with only slightly more planar neutral geometry for TPA 3 likely due to the electron density provided by the additional methoxy group.

Single crystals suitable for X-ray structure determination were grown by slow liquid-liquid diffusion from a binary solvent by dissolving the compounds in dichloromethane and layering a poor solvent of hexanes. The crystals formed fine-needle rods or blocks that maintained a pale yellow color (see Tables S1–S6 for crystallographic data, data collection, refinement parameters, bond distances, and bond and torsional angles). As shown in Figure 5, the structures obtained are mostly ordered (see supporting information for further details). Figure 5A displays TPA 1 where the structure contains a steric interaction between the phenylene hydrogen and sulfur atom on the EDOT aryl rings resulting in a torsional angle of 37.8° (which corresponds to an angle of 142.2° in the computations) and an O---H interatomic distance to be 2.923 Å. Figure 5B displays TPA 3 which contains two crystallographically independent molecules in the asymmetric unit resulting in the −O1X−C5X−C6X−O2X− moiety to be disordered over two orientations. For the compound A variant in Figure 5B, the torsional twist is 22.7° and two interactions occur, one being at S1A–O6A resulting in an atomic distance measurement of 2.718 Å, and another on the EDOT ethylene bridge (O---H) measured at 2.245 Å. The compound B variant has a torsional twist of 30.5° with two interactions occurring, one being at S1B–O6B resulting in an atomic distance measurement of 2.771 Å, and another on the EDOT ethylene bridge consisting of an O—H interaction measured at 2.392 Å. This decreases the distance due to stabilizing noncovalent steric interactions. In Figure 5C, the compound TPA 4 displays an asymmetric unit containing one partially occupied lattice DCM solvent molecule that is found at one site of threefold axial symmetry (and thus must be disordered). The compound has the largest torsional strain due to the repulsive interaction between the adjacent oxygen atoms on the phenylene ring and the ethylene bridge of EDOT. This results in a greater torsional strain of 46.7° and a O---S interatomic distance of 3.152 Å while the steric interaction between O---O is shorter at 2.906 Å (this influences the behavior of this molecule discussed later). In comparison to using DFT as a guide, there is an overall calculation variation of 15–20° in the torsional angle and 0.1–0.4 Å in the atomic distance, which may be due to packing influences in the solid state.

Figure 5.

X-ray crystal structures of (A) 2-MeS-EDOT-TPA-pOCH₃ (TPA 1) and (B) 2-MeS-EDOT-2,5-dimethoxy-TPA-pOCH₃ (TPA 3-variant A & B depicted), and C) 2-MeS-EDOT-2,6-dimethoxy-TPA-pOCH₃ (TPA 4) determined at 110.00(10) K. Selected atoms, the O▪▪▪H, C▪▪▪H, and O▪▪▪S bond distances (yellow line with black numerical value), and the torsion angles (green atoms highlighted with blue value) are shown. Disorder and lattice solvent molecule (for C only) were removed for clarity.

Electrochemical Evaluation

The position of a methoxy substituent on the chromophore not only impacts its structure but also subsequently influences its redox properties. This stems from both the impact on electron density and steric interactions that can significantly affect the energy required for the formation of oxidized species. To understand the redox process, solution electrochemistry was performed in 0.5 M tetrabutylammonium hexafluorophosphate (NBu₄PF₆) / dichloromethane (DCM) solution, described in the Supporting Information sections. Cyclic voltammetry (CV) and differential pulse voltammetry (DPV) were used to probe the onset of oxidation by extracting the faradaic current and extracting the half-wave potentials (E_1/2).

Well-resolved quasi-reversible events were observed in the formation of the cation radical and dication for all five chromophores via CV seen in Figure S31. Figure 6A displays the CV and the inset contains the DPV (Figure S32) of TPA 1 with the extracted E_1/2 of the cation radical and dication species generated in Figure 6B, which summarizes the electrochemical results for all five chromophores showing the potential difference between the two redox processes. TPA 1 represents the standard in the set where substitution at the para position of the pendant phenylene rings remains constant with a methoxy substituent. This serves to add electron density and cap the reactive end inhibiting radical coupling. TPA 1 oxidizes to the cation radical at an E_1/2 of 0.33 V Ag/Ag⁺, while the dication forms at 0.62 V Ag/Ag⁺ with the highest potential separation of the two redox processes at 0.29 V. TPA 2 and TPA 3 both possessed the lowest E_1/2 to form the cation radical at 0.26 V where the non-covalent bonding interaction (O···S) results in conformational locking. It is interesting to note that adding an additional O···H intramolecular interaction between the methoxy in the 5’ position on the phenylene and aryl pendant on the triphenylamine lowers the E_1/2 of the dicationic species to 0.49 V. When we shift the design series to incorporating repulsive (O···O) steric interactions found in TPA 4, the E_1/2 of the cation radical rises to 0.38 V and the dication lowers to 0.52 V resulting in the smallest charge state separation in the defined set at 0.14 V. Increasing the electron richness of the substituent at the meta position of the triphenylamine (TPA 5) resulted in increasing E_1/2 to form the cation radical species to the highest at 0.39 V, but had no substantial impact on the dication. This demonstrates the importance of how controlled structures tailors redox potentials. Strategic substituent placement, such as attractive interactions that involve conformational locking, can result in lowering the oxidation potential of both charged species. Interactions that involve repulsive steric interactions, and lead to a torqued system, raises the oxidation potential to form the cation radical, with less effect on dication formation leading to a smaller potential difference between the two redox processes. In order to find the highest potential separation of charged states, a balance between electron density and steric interactions must be considered.

Figure 6.

A) Cyclic voltammogram of EDOT-TPA-pOCH₃ with extracted half-wave potentials (E_1/2) of the cation radical (blue) and dication (red). Inset is the abstracted differential pulse voltammogram. B) Table below the cyclic voltammogram summarizes the electrochemical properties of all five chromophores with $\Delta E$ representing potential difference between first and second oxidation peaks. C) DPV solution-phase oxidation potential (E_ox) compared to DFT simulated adiabatic ionization potential (A_IP) for the first oxidation state representing the cation radical.

Figure 6C compares the calculated Adiabatic Ionization Potential (AIP) with the electrochemical measurement of the ionization potential attained from DPV measurements (Figure S32), demonstrating there is a strong correlation between theory and experiment for the first oxidation (cation radical state). However, we did not observe this same trend for AIP modeling of the dication state. It has been reported that dications require sophisticated treatment of electron correlation in order to accurately compute interaction energies for dication complexes without overestimation.^40–41

Cyclic stability tests on all five chromophores, as shown in Figure S33, were performed up to 1000 cycles by at a scan rate of 150 mV/s between the potentials of −0.4 to 1.0 V vs Ag/Ag⁺. Satisfactory stability is demonstrated by the near identical CV profiles of all chromophores with no more than 5–15% current loss occurring at higher applied oxidation potentials (> 0.8 V).

Chemical Generation of the Cation Radical and Subsequently the Dication

To elucidate the origin of the spectral peaks for each of the species, a time-dependent DFT treatment was performed in which the lowest lying 15 excited states were calculated. As an illustration, Figure 7 shows UV-Vis spectra for A) TPA 1 and B) TPA 4 as well as C) the excited states for each oxidized species. Both set of spectra show a neutral state with dominant absorption in the UV, along with two cation radical peaks and a dominant dication peak through the visible and near IR. These same trends are also demonstrated by the other three systems (S34–S48 and Table S7–11). A close examination of each charged state in the following discussion provides insight into how the location and number of methoxy substituent impact the spectra which leads to changes in color.

Figure 7.

A) Overlayed simulated UV-Vis spectra of 2-MeS-EDOT-TPA-pOCH₃ (TPA 1), and B) 2-MeS-EDOT-2,6-dimethoxy-TPA-pOCH₃ (TPA 4) for the neutral (blue line), radical cation (red line), and dication (black line). C) Excited analysis results for all three charges of all systems under examination. In each case, the excited state wavelength and corresponding (oscillator strength) is provided.

In the case of the neutral species, there were two significant transitions present (Figure 7A–B (blue trace) and 7C neutral column in table). The most prominent peak in all cases resulted from a transition from the highest occupied molecular orbital (HOMO, H) to the lowest unoccupied molecular orbital (LUMO, L). The wavelengths corresponding to this state ranged from 339 nm (−2,6-OCH₃-, referred to as TPA 4) which possessed the most twisted out-of-plane geometry to 402 nm (−2,5-OCH₃-, TPA 3) the most planar system. The second excited state, which was primarily the result of the H → L + 2 transition only differed by 9 nm. The exception was produced by the unsubstituted (-H-, TPA 1) system which was 6 nm blue-shifted relative to the other systems and was due to the H → L + 4 transition. In all cases, except for the TPA 3 system which has a first excited state at 402 nm, the calculated neutral state was outside the visible range as desired for the compounds to appear transparent.

Cation radical excited states involve the excitation of an alpha (α) electron or a beta (β) electron and as such each electron type will be identified for a given excitation. In all cases, the first cation radical excited state occurred from a singly occupied molecular orbital (SOMO) β electron to the L_β level and ranged from 1044 nm (−2-OCH₃-, (TPA 2)) to 1125 nm (−2,6-OCH₃-, TPA 4). The small difference in this transition for -H- (TPA 1) and −2,5-OCH₃- (TPA 3) suggested that there would ultimately be little difference (~ 6 nm) in the color these molecules would exhibit as cation radicals. Additionally, they exhibited a third transition produced primarily from S_β - 4 → L_β with a large contribution from S_α → L_α in which Δλ was only 22 nm. For the remaining three, the higher energy peak was produced by the S_α → L_α which ranged from 445 nm to 473 nm and was at most 60 nm lower than the other set. Finally, there was another excited present for the TPA 1, TPA 3, and TPA 4 species which possessed lower oscillator strengths, but nonetheless resulted in a widening of the higher energy peak. Overall, the differences between these peaks, coupled with those contributed from lower energy transitions, suggested that there would be a difference in coloration within this set of structures.

Finally, we computationally observed the presence of a dication for all systems. For the lower energy peak, an electron was excited from the HOMO to the LUMO level in which the wavelength ranged from 792 nm (-H-, TPA 1) to 864 nm (−3,4,5-OCH₃-, TPA 5). The trimethoxy species exhibited a close lying lower energy excitation at 825 (f = 0.4926) which resulted from a H - 1 → L. The other four species also exhibited this transition which ranged from 612 nm (-H-) to 729 nm (2,5-OCH₃-, TPA 3). There were additional excited states which contributed to the overall width of the spectral peaks. For TPA 1 – TPA 3, there was little difference in their dication spectra (Δλ_max = 22 nm) hence the color difference in the dication state is expected to be small. On the other hand, the λ_max difference for TPA 4 and TPA 5 relative to TPA 1 were 60 nm, and 72 nm, respectively. Hence, there could be a color difference for these two relative to the other three. To verify the validity of these predictions experimental studies were performed.

In our previous studies^25–27,42 to assess the transmissivity of the neutral state and color vibrancy of the charged state, ACE chromophores were chemically oxidized using iron (III) trifluoromethanesulfonate (Fe(OTf)₃) to generate the cation radical. We follow this standard but add 1 and 2 equivalents of Fe(OTf)₃ to reach the cation radical and dication state. Each chromophore was dissolved at a concentration of 250 μM in acetonitrile, titrated with the respected equivalents of Fe(OTf)₃, and measured after 1 hour. The stock solution was diluted and measured at 63 μM for 1 equivalent of Fe(OTf)₃ and 30 μM for the addition of 2 equivalents of Fe(OTf)₃, photographs depict 63 μM for each charged state color representation. The neutral spectra of all chromophores show excellent correlation between the computationally determined λ_max and those determined experimentally, Figure S49. The experimental results, which include sequential solution oxidation and the photographed color transitions are shown in Figure 8.

Figure 8.

A) 2-MeS-EDOT-TPA-pOCH₃ (TPA 1), B) 2-MeS-EDOT-2-methoxy-TPA-pOCH₃ (TPA 2), C) 2-MeS-EDOT-2,5-dimethoxy-TPA-pOCH₃ (TPA 3), and D) 2-MeS-EDOT-2,6-dimethoxy-TPA-pOCH₃ (TPA 4) overlay UV-Vis spectra of 1 eq. Fe(OTf)₃ (black line) followed by 2 eq. Fe(OTf)₃ (red line). Inserted pictures demonstrate color transition from neutral to each respected charged state.

All chromophores exhibit a neutral absorption λ_max in the UV (attributed to the π-π* transition), resulting in a highly transmissive and colorless appearance (Figure S49). Upon the addition of an oxidant, the neutral absorbance depletes while new transitions are generated, resulting in highly vibrant and bright colors. Figure 8 and Figure S50 display the absorption profiles after the initial titration of 1 equivalent of Fe(OTf)₃ (black curve), along with an additional equivalent of Fe(OTf)₃ -total of 2 eq.- (red curve). When comparing the cation radical spectra, TPA 1 (λ_max = 491 and 558 nm), TPA 2 (λ_max = 482 and 534 nm), TPA 3 (λ_max = 483 and 537 nm), and TPA 5 (λ_max = 500 and 553 nm) all exhibit double absorbance peaks with slight variations in peak height, resulting in a range of vibrant and bright magenta/pink/red, except for TPA 4 displaying the color orange and exhibiting a single absorbance peak at λ_max = 481, which is discussed later. Regarding the LE transition in the cation radical charged state, all molecules display at least one LE absorbance peaks in the NIR region, as well as harboring a shoulder, with variations in peak intensity compared to the HE. Shifting the LE peak to the NIR is essential, as it allows for a dominant HE transition in the visible region of the spectrum to predominantly influence the perceived color.

An additional equivalents of Fe(OTf)₃ generates a new palette of colors shown in Figure 8 and Figure S50. All of the chromophores undergo a drastic reduction in the intensity of the long-wavelength NIR absorption, and two dominant absorption peaks emerge in the visible region, which we attribute to the formation of dication states. Comparing these spectra for TPA 1 (λ_max = 728 nm), TPA 2 (λ_max = 697 and 776 nm), TPA 3 (λ_max = 637 and 801 nm), TPA 4 (λ_max = 713 and 805 nm), and TPA 5 (λ_max = 494 nm) these dominant and relatively broad absorbances are found in the range of 400–800 nm. While the cation radicals of each system (except for the TPA 4 derivative), all attain similar colors, this is not the case for the dications where there is a much greater color diversity as the spectra are significantly different.

Electrochromism

To demonstrate the electrochromic activity of these molecules, spectroelectrochemistry was conducted using an optically transparent thin layer electrode (OTTLE). All measurements were carried out at a concentration of 250 μM in 0.5 M tetrabutylammonium hexafluorophosphate (NBu₄PF₆) / dichloromethane (DCM) electrolyte. The results are depicted in Figure 9 for TPA 1, TPA 2, TPA 3 and TPA 4, while Figure S51 displays TPA 5. The photographic insets demonstrate the colors generated at specific applied potentials.

Figure 9.

Spectroelectrochemistry of A) 2-MeS-EDOT-TPA-pOCH₃ (TPA 1), B) 2-MeS-EDOT-2-methoxy-TPA-pOCH₃ (TPA 2), C) 2-MeS-EDOT-2,5-dimethoxy-TPA-pOCH₃ (TPA 3), and D) 2-MeS-EDOT-2,6-dimethoxy-TPA-pOCH₃ (TPA 4) in 250 μM concentration with applied potentials from 0 – 1 V with 50 mV increments held for 3 minutes each. Photo demonstrates colored charge species forming on the platinum net at specified potentials.

For all compounds, as the potential is increased, the π-π* transition for the neutral state (not shown) diminishes, while absorption peaks attributed to the cation radical as the HE and LE transitions emerge in the visible and near infrared regions, respectively. During the initial steps of applied potential, the spectra resemble those generated via initial chemical oxidation (Figure 8) as the cation radical is formed. Notably, two absorption peaks in the visible region, accompanied by a broad absorption in the NIR, are observed for TPA 1 (516, 710, and 1200 nm), TPA 2 (500, 542, and 1160 nm), TPA 3 (460, 556, and 1205 nm), and TPA 5 (510, 580, and 1220 nm). In contrast, a single absorption peak in the visible region, accompanied by a broad absorption in the NIR, is observed for TPA 4 (490 and 1200 nm). Photographs of the OTTLE cells during oxidation reveal distinct color changes from a colorless and transparent solution. Specifically, TPA 1, TPA 3, and TPA 5 exhibit a visible pink/magenta hue on the platinum mesh, with color intensification at higher potentials. On the other hand, TPA 2 displays a higher saturated red color, while TPA 4 appears as an orange-yellow.

As previously investigated²⁷, at higher applied potentials (typically > 0.50 V), the dication absorption band is expected to emerge, indicated by a new absorption peak growing between 700–850 nm, while the absorption peak in the NIR gradually decreases. Reflecting back on the spectra in Figure 8 from chemical oxidation using Fe(OTf)₃, the dication absorption emerges around 600–800 nm. According to Krauss et al⁴³, certain triarylamine based cation salts (e.g., MeoTPD and spiro-MeOTAD reacted with NOPF₆) exhibit an absorption at $\approx$ 700 nm due to localized HOMO-LUMO transitions of the triphenylamine moiety. In addition, according to Corrente et al⁴⁴, Heckmann et al⁴⁵, and Hankache et al⁴⁶, the pi-conjugated bridging phenylene unit between the two redox active centers of EDOT and arylamine, plays a role in optically inducing the intervalence charge transfer transition (IVCT). It is possible that chemical oxidation and electrochemical oxidation generate differing valence properties but further experimentation would be required to validate this claim.

To validate the controlled formation of charge, an additional investigation involved observing the spectroelectrochemistry at no applied potential, at the two E_1/2 values, and at a high potential (1.0 V) over a 10 minute duration. Figure 10 and Figure S52 display stacked plots showing evolution of absorption for the three states of TPA 1 compared to TPA 4 . Based on the information in Figure 6, which illustrates our electrochemical redox evaluation, it is evident that TPA 1 displayed the largest separation of charge ($\Delta E=0.29V$) between the cation radical and dication, while TPA 4 displayed the smallest separation ($\Delta E=0.14V)$. In Figure 10, we begin with the neutral species without any applied potential and all absorption is in the UV (salmon colored band). At the first E_1/2 = 0.33 V for TPA 1 , 47% of the neutral absorption intensity remains while the cation radical becomes evident in the gray band. Examining the results for TPA 4, at 0.38 V, 59% of the absorption remains. At the second E_1/2, both chromophores have nearly completed oxidation of the neutral species, resulting in the coexistence of cation radical and dication (yellow band) species. For TPA 1 at 0.62 V, 4% of the neutral species remains, while 42% of cation radicals and 68% dications are present. Conversely, for TPA 4 at 0.52V, we observe 20% remaining neutral species, 26% cation radicals, and 54% dications. At 1 V, no neutral species are evident for either compound. At this potential, TPA 1 retains 10% of the cation radicals with dications dominating as the primary charged species. In contrast, TPA 4 at 1 V, 20% of the cation radicals remains, while 80% of the charged species are dications. These results thus show that a larger separation of charge significantly influences the formation of each charged species in a controlled manner.

Figure 10.

Spectroelectrochemistry of A) 2-MeS-EDOT-TPA-pOCH₃ (TPA 1) and B) 2-MeS-EDOT-2,6-dimethoxy-TPA-pOCH₃ (TAP 4) in 375 μM concentration specific applied potentials (with low, the two E_1/2, and a high potential increments held for 10 minutes each. Salmon color regime represents neutral, gray band represents cation radical, and yellow band represents dications species formation.

Lowering the Oxidation State: A Summary of the ACE Generations

By completing this new family of TPA containing ACE electrochromes, we have now developed a collection of EDOT-based systems,^25–27 and compare their electrochemical oxidation potentials in Figure 11. As we consider the potential use of these molecules in electrochromic devices, multiple key considerations include: 1) having wide optical gaps that predominately absorb in the UV in the neutral state to create full transparency in the off-state, 2) a variety of hues and degrees of saturation to be accessed to obtain vibrant colors in the charged on-states, and 3) low oxidation potential to achieve stable cation radical, and possibly even dication species, for long-term redox switching. Considering sustainability, which involves energy efficiency and savings,^47–48 ideal material integration properties include minimum power consumption and small switching voltages. As demonstrated through each new generation of electrochromes, the oxidation potential (E_ox) (V vs Fc/Fc⁺) is varied over a range from 0.0 V to a high of 0.7 V. For diaryl ring EDOT-phenylene molecules (Generation 1), the E_ox was highest between 0.4 to 0,6 V. Transitioning to the three-ring BEDOT-phenylene molecules (Generation 2), the E_ox is observed over a wider range based on the interplay of electron rich character and steric interactions. For example, we have shown that the amine substituted BEDOT-DAB reached an E_ox near 0.0 V; however, we find the charged state to be quite unstable. By coupling the EDOT with the TPA moiety in this study (Generation 3), we find that we attain oxidation potentials between 0.0 and 0.1 V and bring desirable switching stability (1,000) cycles.

Figure 11.

Oxidation potential of all 3 generation sets of EDOT based ACEs compiled recently. Generation 1 (circled red 1–4) represents EDOT-phenylene (Reference 25), Generation 2 (circled green 5–11) represent Bis-EDOT phenylene (Reference 26 and 27), and Generation 3 (circled blue 12–16) represents this study comprising of EDOT-TPA motif.

Conclusion and Perspective

In summary, we compare the geometric, electrochemical, spectral, and colorimetric properties of a series of ACE molecules based on thioalkyl-substituted 3,4-ethylenedioxythiophene coupled to triphenylamine units (EDOT-TPA). The phenylene ring coupled adjacently to the EDOT unit has a range of electron rich character brought by methoxy substituents, which influence both steric and electrostatic interactions. Single crystal structures allow us to determine the torsional angles and non-covalent bonding interactions. Incorporating methoxy substituents in the 2’ position or 2’,5’ results in attractive S---O and O---H interactions, which favors planarity. Conversely, incorporating a methoxy in the 2’,6’ position results in a repulsive interaction between O---O resulting in torsional strain.

Chemical and electrochemical oxidation demonstrate that these class of molecules are fully transparent in the neutral state and generate vibrantly colored solutions of the cation radical and dication species. Chemical oxidation using 1 equivalent of Fe(OTF)₃ generates the cation radical while 2 equivalents generates the dication. Vibrant and varying saturation and contrast of the color magenta/pink/red is accessed for most of the molecules in the cation radical state. However, 2-MeS-EDOT-2,6-dimethoxy-TPA-pOCH₃ (TPA-4), which demonstrates repulsive strain, exhibited a blue-shifted orange color. Interestingly, further oxidation to access the dication state resulted in each molecule exhibiting new colors in the RGB color space. Electro-oxidation successfully generated the cation radical with higher potentials resulting in the generation of the dication species. We demonstrate selective generation and color of the cation radical and dication species by increasing the potential difference between the two charged species in a given molecular system, thus, minimizing intermixing of colors.

Overall, this investigation serves as a comprehensive guide towards designing ACE molecules and controlling their electronic and steric properties that allows tuning of the charged species redox and optical properties. Examining the field of molecular electrochromism, accessing and controlling charge separation is well-understood in cathodically controlled viologen systems, but has not been demonstrated prior to this work for anodic coloring systems. Overall, we summarize the evolution of ACE constructs (Figure 11) and demonstrate that we can lower the ACE motifs oxidation potential leading to enhanced redox switching stability.