Growth of an Ordered Crystalline Organic Heterojunction**

Organic semiconductors have gained tremendous attention over the last two decades due to their potential for use in thin film transistors (TFTs), sensors, solar cells and organic light emitting diodes The best device performances (particularly for TFTs) have often been achieved using free standing single crystals. However, in general the size, thickness, and shape of such crystals are not easily controlled. Many researchers have therefore focused on methods for producing highly aligned crystalline organic thin films through substrate templating. In this work, we demonstrate the quasi-epitaxial growth of an ordered heterojunction between two archetypal organic molecular layers consisting of copper phthalocyanine (CuPc) and 3,4,9,10-perylenetetracarboxylic dianhydride (PTCDA) on a single crystal KBr substrate using organic vapor phase deposition (OVPD). The surface mesh for single layer growths of PTCDA and CuPc, monitored in situ and in real time by high pressure reflection high energy electron diffraction (HP-RHEED), were found to orient over extended substrate distances (∼ 2 cm) along the [100] and [110] KBr lattice directions, respectively, for film growth of up to several hundred Ångstroms thick. Furthermore, highly ordered PTCDA films are shown to be a suitable surface for subsequent growth of ordered films of CuPc. The two archetypal molecules, PTCDA and CuPc, have both been proven useful in organic electronic applications. Also, both have been quasi-epitaxially grown on weakly binding metal surfaces, and on highly oriented pyrolytic graphite (HOPG). Growth on alkali-halide (AH) substrates with a clear relationship between film and substrate has been used to achieve epitaxial growth, although ordered films of only a few monolayers (ML) thickness could be grown prior to structural collapse due to large accumulated strain energies. Additionally, the organic-organic heteroepitaxy of MLs of PTCDA grown on MLs of other organic molecules has been demonstrated. While many of these studies have revealed the fundamental differences between the forms of organic and inorganic epitaxy, nearly all have been limited to growth in ultrahigh vacuum, and involve only a few MLs, neglecting the issues of structural evolution into more practically useful thicker films. The few reports of quasi-epitaxial evolution of thick films have been limited to growth conditions where discontinuous, strain-relieved 3D islanding occurs after the formation of an ultrathin (1–2 ML) oriented wetting layer, following the Stranski-Krastanow growth mode. Moreover, we are unaware of demonstrations of long range azimuthal order greater than the microscopic image field size of (1 nm–1 lm) for several hundred Ångstrom thick film growth. For example, although PTCDA has been grown on a number of AH substrates including NaCl, KCl, KI, and KBr, to our knowledge, no evidence of quasiepitaxial alignment has been reported for growth on KBr. The HP-RHEED pattern for the initial KBr substrate is shown in Figure 1a, indicating a flat and ordered surface for growth. Subsequent HP-RHEED patterns for 400 Å thick films of PTCDA grown at a substrate temperature of Tg = –15 °C (see Experimental Sec.) are shown in Figure 1 for a 0.1 × 20 mm incident beam parallel to the [100] (see Fig. 1b), and [110] (Fig. 1c) KBr crystal directions. Scanning electron microscopy (SEM) images of the corresponding PTCDA surface are shown for Tg = –15 °C and 75 °C (Fig. 1d and e). At both temperatures, PTCDA films were found to stack in the [102] direction of the a-phase (see Fig. 2), as determined from X-ray data. In the [100] direction, two peaks are assigned to the (20) and the (02) surface mesh directions of the flat-lying PTCDA (see Fig. 1b). This index pair results from the 4-fold symmetry of the KBr substrate, where a rotation of 90° is energetically equivalent for the PTCDA unit cell alignment. Along the [110] direction, several orders of (n n) and (n 2n) are observed, with no evidence for the (0 n) or (n 0). This suggests that while there are equivalent, 90° rotated domains, all domains align either along [100] or [010]. This symmetry also explains, for example, why both the (n n) and (n 2n) appear simultaneously, where a rotation of the surface mesh by 90° results in a near overlap, in angle, of these two reciprocal planes. Several orders of the (n 4n) are also observed, but are difficult to distinguish, for example, from the (2n 3n). There is negligible change in the fitted lattice parameters for switching the (1 4) index to (2 3); the former is slightly more consistent with the d-spacings where dmeas = 4.7 ± 0.1 Å, d14 = 4.62 ± 0.01 Å, C O M M U N IC A IO N