Benzo[b]selenophene/thieno[3,2-b]indole-Based N,S,Se-Heteroacenes for Hole-Transporting Layers.

Two series of new N,S,Se-heteroacenes, namely, 6H-benzo[4',5']selenopheno[2',3':4,5]thieno[3,2-b]indoles and 12H-benzo[4″,5″]selenopheno[2″,3″:4',5']thieno[2',3'4,5]thieno[3,2-b]indoles, were successfully obtained using an effective strategy based on Fiesselmann thiophene and Fischer indole synthesis. The new molecules exhibit a large optical band gap (2.82 eV < E g opt < 3.23 eV) and their highest occupied molecular orbital (HOMO) energy formed by the plane π-core ranges between -5.2 and -5.6 eV, with the narrower optical band gap and lower HOMO level corresponding to selenated heteroacenes. In thin solid films of the heteroacenes, hole mobility measured using the conventional CELIV technique ranges between 10-5 and 10-4 cm2·V-1·s-1. All these make the proposed condensed-ring compounds a promising platform for the development of hole-transporting materials applicable in organic electronics.

Introduction

Organic π-conjugated ring-fused molecules, built up from both aromatic and heteroaromatic units, are key structure elements in photo- and electroactive materials. They found many applications as active components in organic light-emitting diodes,[1−3] organic field effect transistors,[4−6] and organic photovoltaics.[7−9] Thereby, the interest of researchers all around the world in such compounds, including their design, synthesis, and properties study, is steadily increasing at a rapid pace.[10−12] Among these molecules, π-excessive heteroacenes, bearing thiophene and pyrrole rings, have been widely shown as attractive p-type semiconductors exhibiting good charge-carrier mobility as well as better environmental stability, owing to their low-lying highest occupied molecular orbital (HOMO) energy levels, in comparison with their fully carbon-cored analogous acenes.[13,14] Furthermore, a number of fused thiophene/pyrrole molecules have been successfully used as electron-donating blocks in the construction of light-harvesting materials, both push–pull small molecules and polymers, for organic photovoltaics with high power conversion efficiency.[15−19]

At the same time, selenophene-based semiconductors exhibit a number of advantageous properties as compared to their thiophene-containing counterparts: narrower optical band gap, lower oxidation and reduction potentials, and well-organized solid-state packing due to strong intermolecular Se–Se interactions, all of which overall allow enhancing the performance of organic electronics devices based on such heteroacenes.[20−22] Taking into account these data, the incorporation of selenium atoms into π-conjugated frameworks and fused compounds, intended for the use in organic electronics, is clearly a promising molecular design strategy, allowing us to improve efficiency of these functional materials.[23,24] Therefore, research efforts aiming at elaboration of new types of selenophene-containing semiconductors as well as at studying their properties are fully justified from the perspective of further optoelectronic applications. Thus, various types of heteroacenes, containing thienopyrrole,[13,18] selenophenopyrrole,[25,26] or selenophenothiophene fragments,[27,28] have previously been synthesized and used as organic electronic materials. However, there are no reports on any fused molecules with a triple combination of pyrrole, thiophene, and selenophene rings to the best of our knowledge.

Taking into account rather strong interest into new ring-fused organic molecules, we report herein on a convenient synthesis of 6H-benzo[4′,5′]selenopheno[2′,3′:4,5]thieno[3,2-b]indole (BTTI) and 12H-benzo[4″,5″]selenopheno[2″,3″:4′,5′]thieno[2′,3′:4,5]thieno[3,2-b]indole (BSTTI) derivatives, depicted on Figure , as well as an investigation of their semiconductor properties.

Figure 1

Structures of N,S,Se-heteroacenes.

Results and Discussion

Synthesis

For access to the aforementioned N,S,Se-heteroacenes, we constructed their thieno[3,2-b]indole (TI) part through three-step transition-metal-free way based on the Fiesselmann thiophene synthesis and the Fischer indole synthesis as key transformations. A similar synthetic plan was recently applied by us to the synthesis of thieno[3,2-b]indole-based N,S-heteroacenes (BTTI) 2(29) and 2-(hetero)aryl-substituted thieno[3,2-b]indoles 4(30) from the appropriate 1,3-C,C-dielectrophilic substrates, namely, 3-chlorobenzo[b]thiophene-2-carboxylates 1 and 2-bromo-3-(hetero)arylacrylates 3. They were used to obtain the intermediated 3-hydroxythiophen-2-carboxylates A via the Fiesselmann method, followed by conversion of compounds A to thiophen-3(2H)-ones B and the next treatment of ketones B with arylhydrazines to directly afford the target TI derivatives by Fischer indolization (Scheme ).

Scheme 1

Synthetic Way to TI Compounds, Based on the Fiesselmann and Fischer Methods

To this end, we chose a suitable benzo[b]selenophene (BS)-based source material to implement our synthetic method to the N,S,Se-heteroacene molecules. In this regard, our attention was drawn to a convenient one-pot synthesis of 3-bromobenzo[b]selenophene derivatives by the treatment of phenylacetylenes with SeO2 and 48% aq. HBr in the presence of cyclohexene, which was previously reported by Arsenyan and co-workers.[31,32] In particular, they demonstrated cyclization of ethyl phenylpropiolate to ethyl 3-bromobenzo[b]selenophene-2-carboxylate 5. The structure of the latter molecule contains the 1,3-dielectrophilic three-carbon fragment, formed by the ethoxycarbonyl group and the C-3 position with the bromine atom, which motivated us to utilize this BS compound as the starting substrate in our synthetic strategy. Thus, compound 5 prepared using the slightly modified literature procedure[31] was condensed with ethyl thioglycolate (2 equiv) in the presence of potassium tert-butoxide (4 equiv) in THF medium to form fused 3-hydroxythiophene-2-carboxylate 6 in 69% yield. Next, saponification of the latter ester with sodium hydroxide in aqueous DMSO solution, followed by acidic workup of a reaction mixture to perform in situ decarboxylation of the obtained acid, afforded thiophen-3(2H)-one 7 in 95% yield (Scheme ). In addition, tricyclic 3-hydroxyester 6 was also treated with triflic anhydride (Tf2O) in the presence of bases to form its triflate derivative 8, bearing a 1,3-C,C-dielectrophilic moiety, further needed to perform annulation of one more thiophene ring by the Fiesselmann reaction.

Scheme 2

Synthesis of Functional Benzo[4,5]selenopheno[3,2-b]thiophenes 6 and 7, 8

Synthesis of the BSTI molecules was successfully performed by the treatment of fused thienone 7 with hydrochloride of phenyl- 9a, 4-methylphenyl- 9b, 4-bromophenyl- 9c, or 1-naphtylhydrazine 9d (1.5 equiv) in the presence of anhydrous sodium acetate (1.5 equiv) in glacial acetic acid. Full indolization of initially formed arylhydrazones of ketone 7 proceeded when the reaction mixtures were being refluxed for 4 h. Thus, desired products 10a–d were isolated in 85–90% yields (Scheme ).

Scheme 3

Synthesis of BSTI Derivatives 10, Substrate Scope, and Yields

The same three-step strategy was applied for the construction of the BSTTI molecules bearing one extra thiophene unit in their frameworks. To this end, triflate derivative 8 was treated with ethyl thioglycolate (2 equiv) and potassium tert-butoxide (4 equiv) under the reaction conditions similar to the compound 6 synthesis, affording desired tetracyclic 3-hydroxyester 11. However, along with normal substitution of triflate in substrate 8, we observed that the undesirable reaction of its o-detriflylation was forming prior 3-hydroxyester 6. Nevertheless, pure 3-hydroxyester 11 was obtained in 30% yield by crystallization from toluene, while side product 6 was also isolated in 28% yield after dilution of the toluene mother liquor with ethanol. Next, 3-hydroxyester 11 was converted to corresponding thiophen-3(2H)-one 12 in almost quantitative yield. Heteroacenes 13a–d were obtained in 50–71% yield by the reaction of ketone 12 with arylhydrazines 9a–d in line with the Fischer indolization protocol (Scheme ).

Scheme 4

Synthetic Way to BSTTI Derivatives 13, Substrate Scope, and Yields

Electronic Properties of Thin Solid Layers

In this study, we investigated electronic properties of synthesized BSTI and BSTTI derivatives 10a, d and 13a as well as previously reported thiophene/pyrrole-based analogues[29] of these N,S,Se-heteroacenes, such as compounds T-10a, d and T-13a (Figure ). Solubility of these compounds is low, so we could prepare solutions of N,S,Se-heteroacenes in tetrahydrofuran. The UV absorption spectra of the N,S,Se-heteroacenes in solution are presented in Figure a. Thin solid films of the new compounds were easily prepared using the thermal vacuum evaporation technique. Absorption spectra of the films are shown in Figure b. For defining the optical band gap, the absorption spectra plotted versus energy are presented in Figures S27-S29 as well. The narrower optical band gap and lower HOMO level corresponding to selenated heteroacenes (Table ) evidently originate from stronger intermolecular Se–Se interaction compared with that of for sulfur atoms.

Figure 2

Structures of studied N,S,Se-heteroacenes and their N,S-counterparts.

Figure 3

Absorption spectra of (a) diluted solutions of N,S,Se-heteroacenes and (b) solid layers of N,S,Se-heteroacenes and their N,S-counterparts.

Table 1

Characteristics of Thin Solid Films of N,S,Se-Heteroacenes and Their N,S-Counterpartsa

compound	λmax, nm	Egopt, eV	HOMO/LUMO energy level, eV (±1%)	dipole moment, D	hole mobility, ×10–5 cm2·V–1·s–1 (CIb 95%)
10a	362	3.00	–5.52/–2.52	1.73	1.6 ± 0.5
T-10a	356	3.24	–5.60/–2.36	1.60	1.8 ± 0.5
10d	340	2.83	–5.25/–2.42	1.55	3.9 ± 1.2
T-10d	338	3.00	–5.23/–2.23	1.42	3.3 ± 1.0
13a	360; 396	3.01	–5.20/–2.19	2.16	2.4 ± 0.7
T-13a	352; 394	3.04	–5.26/–2.22	2.31	2.0 ± 0.6

Absorption maximum λmax and optical band gap Egopt are from spectra in Figures and S27–S29. HOMO level is obtained from CV data (Figure S30). LUMO level is calculated from the HOMO using Egopt. Molecule dipole moment is from DFT calculations (Table S1). Hole mobility corresponds to that at an electric field of ca. 1 × 104 V cm–1.

Confidence interval calculated from 10 replicates.

The XRD spectra (Figure S26) show that solid layers of the N,S,Se-heteroacenes, and their N,S-counterparts consist of well-ordered domains visible as sharp diffraction peaks. Some compounds also possess an amorphous phase visible as a halo diffraction around 2θ = 24°. When looking more particularly at the five-ringed members of the series, they have larger mobilities than the four-ringed members, and we see that the N,S,Se-heteroacenes are more amorphous than their N,S-counterparts. We also see in Table that their mobilities are larger than the more crystalline N,S-heteroacenes. Generally, the HOMO (LUMO) energy levels in the ordered domains are higher (lower) compared to those in the corresponding molecules of the amorphous phase. In an entire layer, the ordered domains act as traps for holes (electron) when they do not form a continuous single crystalline network. That is how we interpret the origin of the low CELIV hole mobilities in the order of 10–5 cm2·V–1·s–1 (Table ) of the studied compounds and the differences between similar compounds; the electron mobility was lower by 1 order of magnitude at least.

Charge carrier mobility in organic materials is governed by electron hopping between neighboring molecules. It increases with the wave-function overlap between the nearest neighbors. The larger the conjugated molecule is, the larger the wavefunction overlap will be, which will lead to larger mobilities. This justifies that 10a and T-10a molecules, with five condensed rings versus six ones for the other molecules, exhibit the smallest mobilities. Similarly, 10d and T-10d show extended π-electron conjugation with respect to 13a and T-13a, which also justifies their relatively larger mobilities. Moreover, the dipole moment of the 10d and T-10d molecules is lower than that in other molecules, and this fact agrees well with the correlation established for molecular glasses: the lower molecule dipole moment, the lower amplitude of the disorder, and hence, the larger charge carrier mobility of the film.[33] For each couple of selenated and sulfurated molecules, electron density distribution on HOMO levels are similar (for more details, see Table S1) and the somewhat difference in hole mobility is within the experimental error.

The difference between selenated molecules 10d and 13a and the sulfurated ones T-10d and T-13a can also be attributed to π-conjugation, as revealed by the UV spectra in Figures and S28–S29, showing that the spectra of selenated molecules are red-shifted.

One may conclude that the larger the planar area of the studied molecules is, the better the intermolecular π–π stacking interactions are, and as a consequence, better is the intermolecular electron transfer. Nevertheless, the apparently low mobilities measured using the technique of CELIV can be attributed to the disorder in the polycrystalline thin films under study.[34] The CELIV mobility of the staple-compound spiro-MeOTAD, which is well known as a hole transport layer (HTL), measured under similar conditions was 8.5 × 10–7 cm–2·V–1·s–1,[35] whereas that measured under the ambient condition was 7.2 × 10–4 cm–2·V–1·s–1[36] (for comparison, the mobility obtained from space charge-limited current measurements in diodes was found to be equal to 4 × 10–5[37] and 1.6 × 10–4 cm–2·V–1·s–1[38]). This makes the reported compounds realistic candidates for HTL in solid-state perovskite based devices as the HOMO-level positions of the heteroacenes match well to the valence band of methylammonium lead iodide (MAPbI3), for instance. Moreover, the lowest unoccupied molecular orbital (LUMO) of each compound is higher than the edge of the conduction band which ranged within −3.7 and −4.0 eV for MAPbI3 and its analogues. Thus, the studied compounds possess the electron blocking capability, another critical parameter of HTL.

Conclusions

In summary, we presented an effective synthetic approach toward two classes of N,S,Se-heteroacenes with the combination of pyrrole, thiophene, and selenophene rings in their fused scaffolds, namely, BSTI and BSTTI derivatives. Indeed, construction of these fused molecules was readily performed from ethyl 3-bromobenzo[b]selenophene-2-carboxylate, available BS-cored material, via the sequence of transition-metal-free reactions, including the Fiesselmann thiophene and Fischer indole synthesis. Semiconductor properties of some herein-obtained BSTI and BSTTI derivatives as well as their analogues, bearing benzo[b]thiophene instead of benzo[b]selenophene in the structure, were studied by preparing thin solid films and measuring the charge carrier mobility. CELIV hole mobility of the order of 10–5 cm2·V–1·s–1 estimated for electric field of 104 V·cm–1 makes them potential materials for use as a hole-transporting layer in organic optoelectronic devices. Because their HOMO energy levels match well to the valence band of methylammonium lead iodide and similar perovskites, one may conclude that the studied heteroacenes can be used as HTL in perovskite-based solar cells, photodiodes, and LEDs.

Experimental Section

Analytical studies were carried out using the equipment of the Center for Joint Use “Spectroscopy and Analysis of Organic Compounds” at the Postovsky Institute of Organic Synthesis of the Russian Academy of Sciences (Ural Division). Melting points were determined on combined heating stages and are uncorrected. Elemental analysis was carried on an automated CHN analyzer. Mass spectrometry was performed using a high-resolution Q-TOF LC–MS/MS spectrometer. NMR measurements were performed on NMR spectrometers at 400 and 500 MHz for 1H, 126 MHz for 13C and 471 MHz for 19F spectra in DMSO-d6 or CDCl3, with tetramethylsilane as an internal standard for 1H, 13C spectra and perfluorobenzene for 19F ones. The13C NMR spectra of some BSTTI (13b–d) could not be determined because of poor solubility of these compounds in a majority of deuterated solvents. Unless otherwise stated, all reagents were purchased from commercial sources and used without further purification.

General Procedure for Cyclic Voltammetry

In thin layers of the N,S,Se-heteroacenes, HOMO energy levels were determined by cyclic voltammetry (CV). The CV experiment was carried out at the scan rate of 20 mV/s in a three-electrode three-compartment electrochemical cell in the glove box with a dry argon atmosphere. Platinum sheets served as the working and counter electrodes. A 20 nm solid layer of an examined material was preliminarily deposited onto the working electrode by thermal evaporation of the compound under 10–6 mbar vacuum at the rate of 1 Å/s. A 0.2 M solution of tetrabutylammonium hexafluorophosphate (NBu4PF6) in acetonitrile (HPLC-grade) was used as an electrolyte. An Ag wire immersed into the electrolyte solution with the addition of 0.1 M AgNO3 was used as a pseudo reference electrode (Ag/Ag+). It was calibrated against ferrocene/ferricenium couple (0.039 V vs Ag/Ag+), and its potential was recalculated to the energy scale using 4.988 eV value for Fc/Fc+ in acetonitrile reported in the literature.[39] Thus, the energy level of Ag/Ag+ (EAg/Ag) in this case is 5.027 eV. Taking into account the accuracy of the CV experiment (±0.02 V), this value should be rounded to 5.03 eV. For all the substances, we failed to determine the LUMO level because of the rather wide optical band gap that ranged around 2.4–2.6 eV. Therefore, taking into account the determined electrochemical response of the HOMO, the response of the LUMO should lie in the range of potentials near −2.4 V, which is close to the cathodic limit of the electrochemical stability window of acetonitrile.

Charge Carrier Mobility Measurements

In thin solid layers, the charge carrier mobility was measured using the technique of charge extraction by linearly increasing voltage (CELIV).[40] Metal-insulator-semiconductor (MIS) diode structures, similar to those described earlier,[41,42] were prepared. Onto the ITO (indium-tin oxides) electrode of the ITO/glass substrate, a SiO2 insulator layer of 70 nm in thickness was deposited by conventional magnetron sputtering. Then, a 100 nm layer of the studied N,S,Se-heteroacenes and a 80 nm layer of Au electrode were deposited successively by thermal evaporation of the material under 10–6 mbar vacuum at the rate of 1 Å/s. The CELIV setup included a digital USB-oscilloscope (DL-Analog Discovery, Digilent Co.), which played roles of the master pulse generator and transient current pulse monitor. RC constants were at least a factor of 20 smaller than the time scales of interest. Small charge extraction regime was used, that is, no injection (offset) voltage was applied in the experiments. The SiO2 layer blocked injection of charge carries from the ITO electrode.

For all the studied devices, the displacement current j(0) was observed to exceed the maximum drift current Δj, that is, Δj≪j(0), as seen in a typical transient curve presented in Figure . Therefore, the charge carrier mobility μ was estimated in accordance with eq (40)where d is the film thickness equal to 100 nm, A is the applied voltage ramp in the range 4/10 kV/s, and tmax is the time corresponding to the maximum of the current j(tmax) = Δj + j(0). The mobility data range was 95% confidence interval calculated from 10 replicates for each sample.

Figure 4

Charge carrier mobility measurements using the CELIV method. Transient current of holes in an ITO/SiO2/10d/Au structure at A = 5.6 × 103 V/s, tmax = 16.7 μs.

Computational Methods

The computations of frontier molecular orbitals, the HOMO and LUMO energy levels, and the dipole moment values were based on restricted density functional theory (DFT).[43−46] Computing was carried out in the Orca 4.0.1 software package using the DFT B3LYP, 6-311G* method.[47]