Well-ordered molecular heterojunction of epitaxial C₆₀ on single-crystal dinaphtho[2,3-b:2',3'-f]thieno[3,2-b]thiophene (DNTT)

"Band transport" in organic semiconductors requires a well-ordered molecular arrangement, which is essential for organic electronic devices potentially competing with inorganic counterparts. ^1–4) The material used in this study, dinaphtho[2,3-b:2',3'-f]thieno[3,2-b]thiophene (DNTT, Fig. 1 inset), is a prominent example, having considerable charge carrier mobility surpassing 1 cm²Vs^{−1 5–8)} and widely-dispersed intermolecular electronic bands. ^9,10) Therefore, it has been used as a benchmark molecular semiconducting material in the development of advanced electronic devices. ^11–15) For optoelectronic devices such as organic photovoltaics, which generally require heterojunctions of donor- and acceptor-type compounds, DNTT is used as a donor molecule. ^16–21) For instance, the combination of DNTT with C₆₀ as a representative acceptor molecule is known to exhibit photovoltaic characteristics. ^16,18) Compared to unsubstituted donor acenes, DNTT is regarded as more advantageous, owing to its chemical stability against (photo-)oxidation reactions. ^22,23)

Zoom In Zoom Out Reset image size

In this study, well-ordered molecular heterojunctions were constructed by stacking C₆₀ molecules on the single-crystal (001) surface of DNTT. Despite significant dissimilarities in crystallographic symmetry and lattice constants, the hexagonal arrangement of C₆₀ in the (111) orientation of its bulk face-centered cubic (fcc) crystal structure grows epitaxially on a rectangular surface lattice of monoclinic DNTT, which can be understood from paradigms known as "van der Waals epitaxy" or "weak epitaxy." ^24–26) Precise analysis using surface X-ray diffraction techniques reveals a unique inter-lattice relationship at this molecular heterojunction, in which the nearest-neighbor molecular direction of C₆₀ is aligned parallel to the a -axis of DNTT. A plausible factor for this alignment of C₆₀ molecules is similar to the previously reported epitaxial C₆₀ and perfluoropentacene on the single-crystal pentacene. ^27–29)

Single crystals of DNTT were produced using a horizontal physical vapor transport technique in a nitrogen stream at atmospheric pressure. ³⁰⁾ Plate-shaped crystals were collected under an ambient atmosphere and fixed on Si wafer pieces by electrostatic force to prepare "single-crystal substrates." C₆₀ was deposited onto the single-crystal DNTT under ultra-high vacuum conditions (typically 1 × 10⁻⁷ Pa). The deposition rate of C₆₀ was finely controlled at 0.02 nm s⁻¹. Although the sample temperature was not intentionally controlled during C₆₀ deposition, it was considered to be RT. The surface morphology of the heterojunction samples was observed using non-contact mode atomic force microscopy (nc-AFM).

The crystallographic structure of the heterojunctions was elucidated using out-of-plane X-ray diffraction (XRD) and grazing-incidence X-ray diffraction (GIXD) at BL19B2, BL46XU, and BL13XU of SPring-8. Two types of X-ray detectors were used for the experiments: a two-dimensional pixel detector (PILATUS-300K) for two-dimensional GIXD (2D-GIXD) and a scintillation counter for out-of-plane XRD and GIXD spot profile analyses. The X-ray glancing angle was set to 0.12° for the GIXD experiments. For the 2D-GIXD measurements, diffraction images were collected during azimuthal rotation of the sample over 360°, and the 2D detector was placed perpendicular to the incident X-rays at approximately 175 mm from the sample rotation center. Details of the measurement setup can be found elsewhere. ²⁷⁾ In these experiments, the X-ray wavelength was fixed at 0.100 nm and the measurements were conducted under an ambient atmosphere at RT.

Figure 1(a) shows the X-ray intensity of a DNTT single-crystal sample covered with 10 nm thick C₆₀ plotted as a function of the out-of-plane scattering vector q_z . Sharp peaks at q_z = 3.88 and 7.75 nm⁻¹ correspond to the 001 and 002 reflections, respectively, of the single-crystal DNTT, ⁵⁾ indicating that the single-crystal DNTT substrates are in c *-axis orientation. It is worth noting that the interlayer spacing (1.622 nm) derived from the 004 spot agrees with the reported bulk crystal structure of DNTT, and any additional features assignable to the thin-film phase of DNTT ³¹⁾ are not found (Supplemental Materials Fig. S1). Compared to the 001 peak, the spot profile of the 002 reflection exhibits a bell-shaped curve. This relatively broad feature centered at q_z = 7.68 nm⁻¹ is attributable to the 111 reflection of C₆₀ in the fcc structure. ³²⁾ This indicates that fcc-C₆₀ crystallites grew in the (111) orientation on the (001) surface of the DNTT single crystals. As shown in Fig. 1(a), the C₆₀-derived feature is well reproduced by a Gaussian curve with a FWHM of 0.64 nm⁻¹. This width corresponds to a crystalline coherent length of 10.0 nm based on a formula given in Ref. 33. In addition, the C₆₀ 111 reflection is accompanied by several side peaks, which can be reproduced by the one-dimensional Laue function of the crystalline repeating number N = 10 [Fig. 1(b)]. These results indicate that crystalline coherence is fulfilled over the entire thickness range of the 10 nm thick C₆₀ overlayers on DNTT, which is the case for RT-grown C₆₀ on single-crystal pentacene, ³⁴⁾ and the crystalline thickness is quite uniform over the surface, as also suggested by AFM observations (Fig. S2). It is worth noting that, as suggested in Fig. 1(c), the out-of-plane crystalline coherent length of C₆₀ is shorter than the nominal thickness for a 50 nm thick C₆₀-covered sample, similar to the case of a rubrene derivative on single-crystal rubrene, ³⁵⁾ while the substantial agreement of these two is satisfied for a 20 nm thick sample (Fig. S3).

Figure 2(a) shows the 2D-GIXD results for the 10 nm C₆₀-covered DNTT single-crystal sample, for which the image was obtained by integrating data from 720 2D-GIXD snapshots taken at 0.5° intervals of the sample azimuthal angle during 360° rotation, and the background signal was subtracted for enhancement of the image contrast (see Fig. S4). While most of the intense signals can be assigned to diffractions spots of the (001)-oriented DNTT (indicated by diamond-shaped marks in Fig. S5), several additional spots are attributable to C₆₀, e.g., spots at (q_xy , q_z ) = (12.5 nm⁻¹, 0 nm⁻¹) indicated with an arrow with a circle mark corresponds to six $2\bar{2}0$-equivalent diffractions of the (111)-oriented fcc-C₆₀. ²⁸⁾ This also confirms the growth of C₆₀ crystallites in the (111) orientation on DNTT.

Zoom In Zoom Out Reset image size

The 2D-GIXD signal intensities of 3 × 25 pixels including the (q_xy , q_z ) position (10.16 nm⁻¹, 0.44 nm⁻¹) and those of 5 × 13 pixels including (q_xy , q_z ) = (12.46 nm⁻¹, 0 nm⁻¹) are plotted as a function of the in-plane azimuthal angle ϕ in Fig. 2(b), where the former corresponds to the 100 and $\bar{1}00$ diffractions of DNTT [indicated with an arrow with a diamond mark in Fig. 2(a)] and the latter corresponds to the $2\bar{2}0$-equivalent diffractions of C₆₀. The sample azimuthal orientation where the DNTT 100 diffraction is detected is defined to be ϕ = 0° hereafter. C₆₀ $2\bar{2}0$-equivalent spots are present only six times at 60° intervals. Common to all the five individual samples examined in the present study, the C₆₀ $2\bar{2}0$-equivalent spots appear at ϕ = (−1.15 ± 0.5)° + n × 60° (n = 1, 2, ⋯, 6) as summarized in Fig. S6. The C₆₀-derived diffraction intensity appears only at specific azimuthal orientations at 60° intervals, indicating that the C₆₀ crystallites are uniformly aligned along a specific axis of the DNTT (001) surface; in other words, C₆₀ grew epitaxially on the single-crystal DNTT substrates. ^27,28)

The inter-lattice relationship between the epitaxial C₆₀ and the single-crystal DNTT (001) surface can be deduced from the ϕ-dependence of the C₆₀ $2\bar{2}0$-equivalent spots based on a procedure reported previously. ³⁶⁾ In brief, an angle difference δ from the reference axis of the epitaxial C₆₀ with respect to that of DNTT can be deduced from the ϕ value where a C₆₀-derived diffraction spot appears as δ + φ'_C − φ'_D = −ϕ. Here, φ'_C is an in-plane angle of the incident X-ray wave vector at which the C₆₀-derived topical spot fulfills the diffraction condition, and φ'_D is that for a DNTT-derived diffraction spot. The incident X-ray orientation angle φ' is taken with respect to the reference axis of each material, and φ' for a diffraction spot of q ≡ (q_xy cosα, q_xy sinα, q_z ) is estimated as φ' = α − cos⁻¹[| q |²/2kq_xy ], where α is the argument angle of the reciprocal point corresponding to the topical diffraction and k is the magnitude of the X-ray wave vector (in this case k = 2π/0.100 nm⁻¹). In this study, DNTT 100 and C₆₀ $2\bar{2}0$-equivalent diffraction spots were used for the analyses, and the [100] and [$2\bar{1}\bar{1}$] directions are defined as the reference axes for the DNTT (001) surface and (111)-oriented C₆₀ crystallites, respectively, denoted as a _D and x _C hereafter. From the above equation, φ'_D is calculated to be −85.356° from α = 0°, | q | = 10.165 nm⁻¹, and q_xy = 10.155 nm⁻¹ for the DNTT 100 diffraction. For C₆₀ $2\bar{2}0,$ φ'_C is calculated to be −114.308° from α = −30° and | q | = q_xy = 12.463 nm⁻¹. Therefore, the experimental ϕ value of (−1.15 ± 0.5)° for the C₆₀ $2\bar{2}0$ diffraction leads to δ = (30.1 ± 0.5)°. This strongly suggests that the angle difference between a _D and x _C is approximately 30°; in other words, the [$1\bar{1}0$] direction of C₆₀ is aligned with the a -axis of the single-crystal DNTT. The inter-lattice relationship derived from the above discussion is illustrated in Fig. 3(a).

Zoom In Zoom Out Reset image size

On the (111) surface of fcc-C₆₀, the molecules form a hexagonal arrangement within the first molecular layer, and the nearest-neighbor directions of the C₆₀ molecules correspond to the 〈$1\bar{1}0$〉 axes. In the case of the epitaxial growth of C₆₀ and perfluoropentacene on the single-crystal pentacene surface, ^28,29) the nearest-neighbor directions of the overlayer molecules align along the [$\bar{1}10$] axis of the pentacene, which is predicted to be the direction where surface diffusion of the adsorbed molecules occurs most frequently. ³⁸⁾ Although the migration behavior of C₆₀ on the DNTT single-crystal surface is unknown, frequent surface diffusion along the [100] direction seems likely because of the atomic-scale ridge and trough structure running in this direction, as illustrated in Fig. 3(b). Therefore, it is suggested that the present crystallographic orientation of the epitaxial C₆₀ on single-crystal DNTT is determined by the same principle as that described previously for C₆₀ on pentacene.

Finally, possible photo-response of the topical molecular junction is discussed. Mori and Takimiya reported a superior power conversion efficiency of solar cells with an exciton dissociating interface between crystalline thin films of DNTT and C₆₀ overlayers produced by sequential vacuum deposition in comparison to pentacene-C₆₀ solar cells fabricated in the same manner. ¹⁶⁾ They attributed a plausible origin of the better cell performance to a higher crystalline order of the DNTT thin-film which leads to more efficient charge carrier transport characteristics. Considering that not only the single-crystal DNTT yields an ideal hole transporting efficiency for this species but also much improvement of electron conduction can be expected for C₆₀ in a uniform crystalline orientation, the present epitaxial C₆₀/DNTT heterojunction has the potential to lead to the development for highly-efficient photovoltaic applications. On the other hand, to achieve a mature understanding of solar cell performance, knowledge of the electronic energy level alignment at the exciton dissociating interface is also indispensable. While photoemission experiments unveiled the energy level diagram and the fulfillment of the Schottky–Mott limit for the case of epitaxial C₆₀ on the single-crystal pentacene, ³⁹⁾ it remains a challenge for demonstration of the photoelectron spectroscopy measurements on the epitaxial C₆₀ on the single-crystal DNTT samples.

In summary, the epitaxial growth of C₆₀ on single-crystal DNTT (001) surfaces was demonstrated using surface XRD measurements. The out-of-plane XRD results indicate that C₆₀ forms (111)-oriented fcc crystallites with a uniform coherent thickness on DNTT. In-plane azimuthal analysis of the 2D-GIXD data revealed an inter-lattice relationship between the epitaxial C₆₀ and the single-crystal DNTT (001) surface in which the nearest-neighbor direction of the C₆₀ molecules uniquely aligns along the crystalline a -axis of DNTT.

Acknowledgments