Mixed-Orbital Charge Transport in N-Shaped Benzene- and Pyrazine-Fused Organic Semiconductors.

The hole-carrier transport of organic semiconductors is widely known to occur via intermolecular orbital overlaps of the highest occupied molecular orbitals (HOMO), though the effect of other occupied molecular orbitals on charge transport is rarely investigated. In this work, we first demonstrate evidence of a mixed-orbital charge transport concept in the high-performance N-shaped decyl-dinaphtho[2,3-d:2',3'-d']benzo[1,2-b:4,5-b']dithiophene (C10-DNBDT-NW), where electronic couplings of the second HOMO (SHOMO) and third HOMO (THOMO) also contribute to the charge transport. We then present the molecular design of an N-shaped bis(naphtho[2',3':4,5]thieno)[2,3-b:2',3'-e]pyrazine (BNTP) π-electron system to induce more pronounced mixed-orbital charge transport by incorporating the pyrazine moiety. An effective synthetic strategy for the pyrazine-fused extended π-electron system is developed. With substituent engineering, the favorable two-dimensional herringbone assembly can be obtained with BNTP, and the decylphenyl-substituted BNTP (C10Ph-BNTP) demonstrates large electronic couplings involving the HOMO, SHOMO, and THOMO in the herringbone assembly. C10Ph-BNTP further shows enhanced mixed-orbital charge transport when the electronic couplings of all three occupied molecular orbitals are taken into consideration, which results in a high hole mobility up to 9.6 cm2 V-1 s-1 in single-crystal thin-film organic field-effect transistors. The present study provides insights into the contribution of HOMO, SHOMO, and THOMO to the mixed-orbital charge transport of organic semiconductors.

Introduction

Organic semiconductors (OSCs) held by weak van der Waals intermolecular interactions are ideal materials for flexible, thin, and lightweight next-generation electronics.[1−4] In recent years, state-of-the-art hole-transporting p-type OSCs developed by rational molecular designs have demonstrated outstanding charge-carrier mobilities (μ) over 10 cm2 V–1 s–1 in solution-processed single-crystal organic field-effect transistors (OFETs),[5−9] well exceeding that of amorphous inorganic semiconductors. Among a handful of high-performance p-type OSCs that are reliably evaluated, decyl-dinaphtho[2,3-d:2′,3′-d′]benzo[1,2-b:4,5-b′]dithiophene (C10–DNBDT–NW)[10] with the N-shaped geometry demonstrates high hole μ up to 16 cm2 V–1 s–1, enabling the fabrication of high-frequency logic circuits using large-area solution-processed single-crystal thin films.[11−13] In light of the high performance of C10–DNBDT–NW, several experimental and theoretical studies have been carried out to understand its charge-transport properties. The band-like charge transport behavior is experimentally observed in C10–DNBDT–NW single-crystal thin films due to its effective intermolecular orbital overlaps.[14] The unconventional N-shaped geometry of C10–DNBDT–NW with a low degree of freedom compared to the linear-shaped π-electron systems likely results in a low degree of dynamic disorder, which has been revealed to be detrimental to charge transport.[15−17] A study reported by Troisi and co-workers suggests that C10–DNBDT–NW forming the two-dimensional (2D) herringbone molecular assembly in the solid state exhibits isotropic intermolecular orbital overlaps, quantified by balanced transfer integral (t) values, making it resilient to dynamic disorders compared to OSC systems with more unbalanced t.[18]

A recent report using the tight-binding approximation suggests that the high μ of [1]benzothieno[3,2-b][1]benzothiophene (BTBT)[7,1]Benzothieno[3,2-b]Benzothiophene (BTBT) Derivatives for High-Performance, Solution-Processed Organic Field-Effect Transistors. J. Am. Chem. Soc.. 2007 ">19] derivatives are associated with the contribution of their second highest occupied molecular orbitals (SHOMO) charge transport, in addition to the highest occupied molecular orbitals (HOMO) couplings.[20] This is attributed to the small energy difference of ca. 0.4 eV between the HOMO and SHOMO levels, and large t between HOMO–SHOMO electronic couplings. By just considering the HOMO–HOMO electronic couplings, octyl-substituted BTBT (C8–BTBT) exhibits a flat valence band maximum with small dispersion due to small t values of HOMO–HOMO in certain crystallographic directions, which is unfavorable for hole transports. When the SHOMO is incorporated in the band calculation, the curvature of the valence band maximum is drastically increased, resulting in a smaller effective mass (m*) of hole carriers than that without SHOMO couplings; m* is an important parameter for high-performance OSCs exhibiting band-like charge transport,[21−25] which is inversely proportional to μ (as well as t) described by the following equation, where q is the elemental charge and τ is the relaxation time.

Since m* is obtained from either the valence maximum or the conduction band minimum, described aswhere ℏ is the Dirac’s constant, E is the valence band energy, and k is the wavevector, a change in the curvature of the valence band maximum would lead to a different m* of hole carriers.

In view of the energetic proximity of SHOMO and the third highest occupied molecular orbital (THOMO) levels of C10–DNBDT–NW[26] (ESHOMO – ETHOMO = 0.004 eV), we anticipate that the THOMO may also participate in the charge transport, in addition to HOMO and SHOMO. Using the tight-binding approximation, we discover that the curvature of the valence band maximum when the electronic couplings of HOMOs, SHOMOs, and THOMOs are taken into consideration, is notably increased compared to that with electronic couplings of only HOMOs (Figures a and S1). Our understanding of charge transport in OSC systems is largely based on frontier orbitals, where the hole transport is conducted via HOMO overlaps for p-type, and electron transport via the lowest unoccupied molecular orbital (LUMO) overlaps for n-type OSCs. The current results strongly suggest that the SHOMO and THOMO electronic couplings in C10–DNBDT–NW increase the curvature of the valence band maximum, leading to small m* and a mixed-orbital charge transport. Thus, electronic couplings involving SHOMOs and THOMOs should also be examined to fully understand the charge transport of OSCs.

Figure 1

(a) Molecular features and herringbone assembly of C10–DNBDT–NW, band structures (colored shades illustrate the bandwidths) and charge-transport capabilities estimated by the tight-binding approximation. (b) Molecular orbital distributions of C10–DNBDT–NW (based on the previously reported single-crystal structure) and Me–BNTP (optimized) calculated at the B3LYP/6-311G(d) level of theory.

In this work, we envisage a substitution of the central benzene ring of DNBDT with a pyrazine moiety for a new N-shaped bis(naphtho[2′,3′:4,5]thieno)[2,3-b:2′,3′-e]pyrazine (BNTP) π-core to induce enhanced electronic couplings of HOMOs, SHOMOs, and THOMOs (Figure b). The pyrazine moiety has been extensively employed as a π-linker for tuning the electronic property and intramolecular charge transfer.[27−29] Molecular design of various electron-transporting n-type OSCs also involves pyrazine to achieve electron affinities for n-type device operations.[30−32] However, the pyrazine unit has not been extensively explored for p-type OSC designs due to its electron-deficient properties. Although pyrazine is shown to be effective in tuning the HOMO–LUMO energy levels, its effect on HOMO, SHOMO, and THOMO engineering, as well as hole-carrier transports, remains elusive. From our preliminary calculations, the methyl-substituted BNTP (Me–BNTP) shows large HOMO, SHOMO, and THOMO coefficients and orbital distributions on the sulfur atoms and the dithienopyrazine (T2P) moiety (Figure b), which may be promising for inducing mixed-orbital charge transport. Herein, the novel R–BNTP π-core with different substituents is successfully synthesized, and a strong substituent effect is observed in their molecular assemblies. The decylphenyl-substituted BNTP (C10Ph–BNTP) shows large protruding HOMO, SHOMO, and THOMO coefficients on its sulfur atoms and the T2P moiety that are favorable for effective electronic couplings. The 2D herringbone molecular assembly of C10Ph–BNTP exhibits a pronounced increased curvature of the valence band maximum when the effects of HOMO, SHOMO, and THOMO electronic couplings are incorporated in the band estimation, leading to a small m*. The solution-processed thin film of C10Ph–BNTP demonstrates high μ up to 9.6 cm2 V–1 s–1, which is a new addition to the class of high-performance p-type OSCs that are applicable for high-end organic electronics. Our current results provide valuable insights into future molecular design and investigations of mixed-orbital charge transport in OSCs.

Results and Discussion

Synthesis

Strategies to synthesize pyrazine-containing π-electron systems often involve condensation reactions and Buchwald–Hartwig cross-coupling of diamine groups.[33,34] However, we found that these methods were not applicable for the synthesis of BNTP due to the unique reactivity of thiophene. Hence, we developed a novel synthetic strategy for the pyrazine-containing sulfur-bridged R–BNTP in the current work (Scheme ). From the previously synthesized compound 1, we performed a selective halogen exchange of the aryl iodide at −78 °C to generate the organozinc intermediate in situ. Then, a palladium-catalyzed Negishi cross-coupling using P(2-furyl)3 as the ligand[35] was done to generate the pyrazine compound 2 at room temperature in 67–77% yields. Bromination of the pyrazine core was difficult, and we discovered that the sequential use of the bulky tBu2Zn(TMP)Li[36] and BrCCl2CCl2Br[37] brominating reagent was able to selectively brominate the 2,5-positions of pyrazine to give compound 3 in 11–34% yields. Lithium-halogen exchange was performed at four aryl bromide sites of 3, and the intramolecular annulation reaction was completed via a nucleophilic attack by the tetralithiated intermediate on benzenesulfonic thioanhydride as the source of sulfur. C10– and TMS–BNTP derivatives were furnished in 52 and 57% yields, respectively. Further derivatization of the BNTP π-core started from TMS–BNTP, where the desilyliodination was directly performed by ICl at 80 °C to afford I–BTNP in 81% yield. Two more derivatives, Ph–BNTP and C10Ph–BNTP, were synthesized from the versatile I–BNTP precursor by Negishi cross-coupling using Pd(dppf)Cl2 in 70 and 84% yields, respectively.

Scheme 1

Synthetic Route of BNTP Derivatives

Reagents and conditions: (i) iPrMgBr, ZnCl2, −78 °C to r.t., LiCl, THF. (ii) 2,5-dibromopyrazine, Pd(dba)2, P(2-furyl)3, THF, r.t., 19 h. (iii) (1) tBuZn(TMP)Li, 0 °C, (2) BrCCl2CCl2Br, 0 °C, 20 h. (iv) (1) nBuLi, THF, −78 °C, (2) PhO2S-S-SO2Ph, THF, −78 °C to r.t., 12 h. (v) ICl, oDCB, −78 to 80 °C, 1.5 h. (vi) Ph–BNTP: (1) PhMgBr, ZnCl2, LiCl, 0 °C to r.t., (2) I–BNTP, Pd(dppf)Cl2, toluene, 110 °C, 12 h. C10Ph–BNTP: (1) C10PhBr, nBuLi, ZnCl2, −78 °C, (2) I–BNTP, Pd(dppf)Cl2, toluene, 110 °C, 12 h.

Fundamental Properties

The ionization potential (IP) of C10–BNTP is measured to be 5.81 eV by photoelectron yield spectroscopy (Figure S15), which is larger than that of C10–DNBDT–NW (5.24 eV)[10] due to the electron-withdrawing effect of the pyrazine moiety. The film-state UV–vis spectroscopy of C10–BNTP shows a red-shifted λmax of 478 nm (Figure S18) relative to C10–DNBDT–NW (445 nm).[9] The IP of Ph– and C10Ph–BNTP are measured to be 5.55 and 5.52 eV, respectively, which are smaller than that of C10–BNTP due to the substituent effect (Figures S16 and S17). Thin-film UV–vis absorption of Ph– and C10Ph–BNTP demonstrate λmax of 488 and 504 nm, respectively (Figures S19 and S20). The thermal stability of C10–BNTP is slightly reduced compared to C10–DNBDT–NW, demonstrated by their 5% weight loss (T95) of 377 °C (Figure S21) and 426 °C from the thermogravimetric analysis (TGA), respectively. However, by substituting the BNTP π-core with phenyl and decylphenyl groups, Ph– and C10Ph–BNTP exhibit increased T95 of 469 and 440 °C, respectively (Figures S22 and S23). From the differential scanning calorimetry (DSC) profiles, we observe a low-temperature phase transition peak at 49 °C for C10–BNTP, whereas C10Ph–BNTP demonstrates a stable crystalline phase up to 270 °C, and Ph–BNTP shows no phase transition peaks up to 350 °C (Figures S24–S26). The solubility of C10–BNTP in o-dichlorobenzene is 0.095 wt % at 60 °C, which is 2.2 times higher than that of C10–DNBDT–NW (0.042 wt %). C10Ph–BNTP also shows a higher solubility of 0.052 wt % than C10–DNBDT–NW, which indicates good solution processability of BNTP derivatives (Table S1).

Single-Crystal Structures

The solution-grown single crystals of C10–BNTP appear as thin yellow needles. C10–BNTP crystallizes in the monoclinic C2/c space group, and in contrast to C10–DNBDT–NW, the former assumes a one-dimensional (1D) π–π stacking motif with standing alkyl chains (Figure a). The C10–BNTP π-core shows a completely planar geometry, different from the bent-shaped geometry of C10–DNBDT–NW, and a large core-to-core intermolecular distance of 3.42 Å in the π–π stacking direction. The alkyl chains of C10–BNTP form a large torsion angle of 110.1° with the BNTP π-core and disrupt intermolecular orbital overlaps in the π–π stacking direction.

Figure 2

Molecular geometry, bent angle, packing motif (substituents omitted for clarity), and interlayer molecular assembly of (a–d) C10–, Ph- (1D and 2D), and C10Ph–BNTP, respectively, in the single-crystal structures obtained at ambient temperatures.

By changing the substituents on BNTP, we observe polymorphism on Ph–BNTP that forms both needle- and plate-like single crystals. The isolated needle-like crystals (Ph–BNTP-1D) again exhibit the 1D π–π stacking motif, with slightly improved edge-to-face C–H···π interactions (Figure b). Interestingly, the plate-like crystals of Ph–BNTP (Ph–BNTP-2D) on the other hand, form a bent-shaped geometry with a dihedral angle of 6.5° between the planes of naphthalene and pyrazine. The dihedral angles between the phenyl ring and the BNTP π-core are measured to be 41.8 and 26.5° for Ph–BNTP-1D and Ph–BNTP-2D, respectively. The molecular assembly of Ph–BNTP shows a 2D herringbone assembly, where the angle between planes of dimers (defined as the herringbone angle) is measured to be 38° (Figure c).

After examining the above single-crystal structures of BNTP derivatives, it is apparent that the 2D herringbone assembly can be induced by enhanced intermolecular interactions. We subsequently employed the decylphenyl substituent to stabilize the herringbone assembly and obtained isomorphic crystal structure by increasing the total van der Waals forces from the introduction of alkyl chains.[25] Indeed, C10Ph–BNTP forms exclusively large yellow plate-like single crystals, and the structure exhibits a bent geometry, with a bent angle of 7.0°, likely induced by the enhanced intermolecular interactions (Figure d). The phenyl rings of C10Ph–BNTP form a small dihedral angle of 14.9° with the bonded naphthalene, and the decyl alkyl chains are laterally extended, which is drastically different from C10–BNTP. The 2D herringbone motif is adopted by C10Ph–BNTP, with a herringbone angle of 49°.

Estimation of Transfer Integrals

From the molecular assemblies of R–BNTP, t values of HOMO–HOMO electronic couplings are calculated at the PBEPBE/6-31G(d) level to estimate their charge-transport capabilities.[38] The 1D π–π stacking of C10–BNTP shows a large core-to-core distance in the b-axis direction that results in a t value of −20 meV and no significant intermolecular orbital overlaps in the c-axis direction, indicating a 1D charge-transport behavior (Figure a). Similarly, the major charge transport of Ph–BNTP-1D is contributed by the face-to-face π–π stacking along the a-axis direction with a t value of +14.9 meV, while t values from the edge-to-face interactions are −2.23 and −0.236 meV (Figure b). Charge transport of the herringbone molecular assembly can be quantified aswhere d is the lattice constant in the column direction and tcolumn and |ttrans| are the transfer integrals in the column and transverse directions, respectively. Here, the sign of tcolumn is important, with the positive t value (overlaps with the opposite phases of molecular orbitals) being favorable for charge transport and vice versa, while only the absolute value of ttrans is important. For Ph–BNTP-2D and C10Ph–BNTP that form the herringbone assembly, we define their tcolumn as t1,4, and their |ttrans| as |t2,3,5,6| as shown in Table S3. Since single-crystal structures of Ph–BNTP-2D, C10Ph–BNTP, and C10–DNBDT–NW have similar lattice constants in the column direction, ranging from 6.11 to 6.51 Å, their charge transports can be effectively studied with t values.

Here, we show all t values of Ph–BNTP-2D, C10Ph–BNTP, and C10–DNBDT–NW (as reference) involving HOMO, SHOMO, and THOMO (Table S3) to investigate their mixed-orbital charge-transport capabilities. Based on the previous report, the sign of transfer integrals in the transverse direction becomes important in investigating the mixed-orbital charge transport.[20] Although Ph–BNTP-2D exhibits a 2D herringbone assembly, the small herringbone angle of 38° in combination with its HOMO distribution results in ineffective intermolecular orbital overlap in the column direction (c-axis) that is indicated by t1,4 of +6.90 meV, but the transverse direction (b-axis) shows large |t2,3,5,6| of 47.9 meV. Ph–BNTP-2D shows a much smaller t value of +0.131 meV from HOMO–SHOMO couplings in the column direction, suggesting that the SHOMO may have small contributions to its charge transport. By calculating t values from HOMO–THOMO couplings, t1,4 increases to −8.93 meV in the column direction, t2,5 and t3,6 also show large values of −42.9 and −42.3 meV, respectively, in the transverse directions. The t values of HOMO, SHOMO, and THOMO of Ph–BNTP-2D here suggest that the electronic couplings from these orbitals should be fully investigated for understanding the charge transport of OSCs.

The 2D charge-transport capability is apparent in C10Ph–BNTP as the column direction (c-axis) shows an enhanced t1,4 of +16.6 meV, and the transverse direction also exhibits large |t2,3,5,6| of 28.8 meV (Table S3). Similar to Ph–BNTP-2D and C10–DNBDT–NW, C10Ph–BNTP shows smaller t values in the column directions from HOMO–SHOMO couplings than those from HOMO–HOMO couplings. However, C10Ph–BNTP demonstrates t1,4 of −14.9 meV from HOMO–THOMO couplings. In addition, the transverse directions of C10Ph–BNTP exhibit large t2,5 and t3,6 of −46.3 and −49.0 meV from HOMO–THOMO compared to its HOMO–HOMO couplings in the transverse directions. The results herein indicate a significant contribution of HOMO, SHOMO, and THOMO electronic couplings in Ph–BNTP-2D and C10Ph–BNTP.

The fascinating electronic couplings of current BNTP derivatives may be potentially attributed to two reasons. For instance, the small energy differences between HOMO–SHOMO and HOMO–THOMO levels of C10Ph–BNTP are calculated to be 0.51 and 0.64 eV, respectively, based on the geometry of its single-crystal structure (Figure ). We propose that OSCs with an energy difference around 0.6 eV between HOMO, SHOMO, and THOMO may be considered for the mixed-orbital charge transport. Note that the orbital energies calculated using the single-crystal structure may be different from those of the optimized structure due to the effect of molecular packings. Similar to C10–DNBDT–NW, the energy difference between SHOMO and THOMO levels of C10Ph–BNTP is only 0.13 eV. We postulate that such energetic proximity of these three occupied molecular orbitals allows them to participate in electronic couplings and contribute to the charge transport. Another factor may be attributed to the molecular orbital distributions of BNTP. The single-crystal structure of C10Ph–BNTP shows large HOMO distribution on the sulfur atoms as well as the T2P moiety, which may strongly contribute to the electronic couplings in both the column and transverse directions.[9] Large SHOMO and THOMO distributions are also observed on these units that participate in the electronic couplings through effective intermolecular orbital overlaps. Similarly, C10–DNBDT–NW and Ph–BNTP both exhibit large molecular orbital distributions on sulfur and central BDT and T2P moieties (Figure S27).

Figure 3

Distribution of HOMO, SHOMO, and THOMO of C10Ph–BNTP calculated at the B3LYP/6-311G(d) level of theory using the single-crystal structures. Orbital composition analysis was performed with the Mulliken partition using the Multifwn program.[39]

Estimation of Effective Masses

To further investigate the charge-transport capability of BNTP derivatives, we performed band structure and m* (in units of m0: the rest mass of an electron) calculations via the plane-wave basis.[40] Derivatives C10–BNTP and Ph–BNTP-1D with 1D assemblies give negligible valence band dispersions and large m* values that are unfavorable for charge transport (Figure S31). Both Ph–BNTP-2D and C10Ph–BNTP with the 2D herringbone assembly demonstrate large curvature of valence bands at the Γ point, resulting in small m*|| values of 1.54 and 1.43 m0, respectively, in the column direction, and m*⊥ of 1.70 and 1.91 m0, respectively, in the transverse direction (Figure a,b). The small m* values of Ph–BNTP-2D and C10Ph–BNTP [comparable to those of C10–DNBDT–NW (Figure a)] are intriguing considering that their t values from HOMO–HOMO couplings are not as large as C10–DNBDT–NW. Since both derivatives exhibit large SHOMO and THOMO bandwidths and overlaps between these bands (Figure a,b), we postulate that their band structures estimated from the plane-wave basis contain significant contributions from SHOMO and THOMO in addition to HOMO.

Figure 4

Band structures (colored shades illustrate the bandwidths) and effective masses of (a) Ph–BNTP-2D and (b) C10Ph–BNTP obtained using the plane-wave (c* and b* correspond to the column and transverse directions in the herringbone packing) and tight-binding approximations calculated at the PBEPBE/6-31G(d) level of theory.

With the tight-binding approximation, we are able to separately examine the effects of only HOMO couplings, as well as combined HOMO, SHOMO, and THOMO couplings on the curvature of valence band maxima. For C10–DNBDT–NW, the curvature of the valence band maximum from only HOMO–HOMO electronic couplings at the Γ point is different from that when all HOMO, SHOMO, and THOMO electronic couplings are taken into consideration in the tight-binding approximation. The effect of orbital hybridization can be preliminarily expressed by the energy difference (ΔE, determined by both the sign and magnitude of transfer integrals)[20] of the bands at the Γ point, which is 15 meV for C10–-DNBDT–NW (Figure a), but it is minuscule for Ph–BNTP-2D (Figure a). In the case of C10Ph–BNTP, incorporating the HOMO, SHOMO, and THOMO electronic couplings in the calculation drastically increases the curvature of the valence band maximum at the Γ point, which is accompanied by a significant increase in the ΔE of 31 meV (Figure b), suggesting an effect of the mixed-orbital charge transport.

We observe that the m* values of Ph– and C10Ph–BNTP in the column direction (Γ–Z) are significantly reduced from the increased curvature of their valence band maxima as the electronic couplings of SHOMO and THOMO are considered (Table S4). A similar reduction of m* in the column direction is observed in C10–DNBDT–NW with the incorporation of orbital hybridizations (Table S4), though to a lesser degree than that of C10Ph–BNTP. The tight-binding approximation suggests that the mixed-orbital phenomenon is present in C10–DNBDT–NW, Ph–BNTP-2D, and C10Ph–BNTP, while C10Ph–BNTP shows the strongest contribution from the electronic couplings of HOMO, SHOMO, and THOMO to its charge-transport capability.

OFET Performances

Bottom-gate/top-contact OFETs are fabricated with R–BNTP derivatives as the active layer to experimentally evaluate their charge-transport capabilities. The single-crystal thin films of C10–BNTP (Figure S28) fabricated by the edge-casting method[41] show no apparent semiconductor behavior (Figure S32a,c), which is reasonable considering its unfavorable molecular assembly, intermolecular orbital overlaps, and the negligible dispersion of valence band structure. Interestingly, Ph–BNTP-2D crystals can be exclusively obtained by the physical vapor transport (PVT) method, and the plate-like crystals are manually laminated on OFET substrates. The OFET channels are constructed along the column direction (c-axis) (Figure S29) of the Ph–BNTP-2D single crystals, and the highest μ is measured to be 0.8 cm2 V–1 s–1 (Figure S32b,d). Despite the 2D herringbone assembly, Ph–BNTP-2D shows small t values along the column direction (channel direction), which explains its low μ.

On the other hand, C10Ph–BNTP with long alkyl chains offers excellent solution processability, where large single-crystalline domains are easily formed via the solution-processed edge-casting method (Figure a). Its thin-film structure is consistent with its bulk single-crystalline structure, which allows us to correlate our mixed-orbital charge-transport estimations with its device performance. The OFETs with channels constructed along the column direction (c-axis) (Figure S30) of C10Ph–BNTP exhibit excellent p-type OSC behavior (Figure b,c) with small hysteresis and Vth of −18 V, and the highest μ is measured to be 8.6 cm2 V–1 s–1 (Figure d) on the parylene diX-SR insulating layer (averaged μ of 6.3 cm2 V–1 s–1 over 17 devices (Figure S33)). C10Ph-BNTP OFETs using the β-phenethyltrichlorosilane (β-PTS) dielectric layer with the same device architecture demonstrate an improved highest μ of 9.6 cm2 V–1 s–1, with a Vth of −19 V (Figure S34). We summarize the fundamental properties and OFET performances of C10–DNBDT–NW and R–BNTP derivatives in Table S5 for comparison. The μ value of C10Ph–BNTP is considerable among the high-performance thienoacene-type organic semiconductors.[42] Although the current single-crystalline OFET performances of C10Ph–BNTP exhibit relatively large threshold voltages and bias stress (Figure S35), further optimization of the device fabrications needs to be carried out in future work.

Figure 5

(a) Polarized optical microscopic image of the channel; (b) transfer characteristic, the gray dashed line illustrates the fit to |ID|1/2, from which the μ was estimated; (c) output characteristic; and (d) gate voltage-dependent μ of single-crystalline thin-film OFET fabricated with C10Ph–BNTP on the parylene diX-SR insulating layer.

Despite having relatively small t values from HOMO–HOMO couplings, C10Ph–BNTP with large SHOMO and THOMO coefficients in the central T2P moiety exhibits large t values from HOMO–SHOMO and HOMO–THOMO electronic couplings. The plane-wave basis suggests a small m* of C10Ph–BNTP, and we have identified significant contributions from SHOMO and THOMO electronic couplings based on the tight-binding approximation, which indicates a mixed-orbital charge transport. The estimated charge-transport capability of C10Ph–BNTP is further substantiated by the high μ in solution-processed single-crystalline thin-film OFETs.

Conclusions

In this work, we report the molecular design for mixed-orbital charge transport via HOMO, SHOMO, and THOMO engineering with a pyrazine-containing N-shaped BNTP π-core. A robust synthetic strategy is developed to synthesize various R–BNTP derivatives, which is also applicable for the synthesis of other pyrazine-containing π-electron systems. The single-crystal structure of C10-BNTP adopts a 1D π–π stacking that results in poor charge transport. The favorable 2D herringbone assembly is induced by introducing phenyl and decylphenyl substituents. Ph– and C10Ph–BNTP both demonstrate relatively small t values from HOMO–HOMO couplings compared to the high-performance C10–DNBDT–NW. However, the curvature of valence band maxima of Ph– and C10Ph–BNTP are drastically increased when SHOMO and THOMO electronic couplings are incorporated in the band estimation, compared to those where only HOMO–HOMO couplings are considered. C10Ph–BNTP with evidence of effective mixed-orbital charge transport demonstrates μ as high as 9.6 cm2 V–1 s–1 in solution-processed single-crystalline thin films, which is promising for fabricating high-end organic electronics. The present study suggests that the orbital hybridization of HOMO, SHOMO, and THOMO may contribute to the formation of the valence band of OSCs, and the mixed-orbital effect should be investigated to gain a more complete understanding of the charge transport of OSCs.