Self-Assembly Properties of Solution Processable, Electroactive Alkoxy, and Alkylthienylene Derivatives of Fused Benzoacridines: A Scanning Tunneling Microscopy Study.

Self-organization in mono- and bilayers on HOPG of two groups of benz[5,6]acridino[2,1,9,8-klmna]acridine derivatives, namely, 8,16-dialkoxybenzo[h]benz[5,6]acridino[2,1,9,8-klmna]acridines with an increasing alkoxy substituent length and 8,16-bis(3- or 4- or 5-octylthiophen-2-yl)benzo[h]benz[5,6]acridino[2,1,9,8-klmna]acridines, i.e., three positional isomers of the same benzoacridine, is investigated by scanning tunneling microscopy. The layers were deposited from a solution of the adsorbate (in hexane or dichloromethane) and imaged ex situ at molecular resolution. In all cases, the resulting two-dimensional (2D) supramolecular organization is governed by the interactions between large, fused heteroaromatic cores that form densely packed rows separated by areas covered by substituents. In 8,16-dialkoxybenzo[h]benz[5,6]acridino[2,1,9,8-klmna]acridines, the alkoxy substituents, separating the rows of densely packed cores, are interdigitated. An increasing substituent length leads to an intuitively expected increase in this 2D unit cell parameter that corresponds to the orientation of the substituent in the monolayer. In the case of 8,16-bis(3- or 4- or 5-octylthiophen-2-yl)benzo[h]benz[5,6]acridino[2,1,9,8-klmna]acridine positional isomers, the self-assembly processes are more complex. Although the determined 2D unit cell is in all cases essentially the same, the role of alkylthienylene substituents in layer formation is distinctly different. Thus, the formation of monolayers and bilayers is very sensitive to isomerism. 8,16-Bis(5-octylthiophen-2-yl)benzo[h]benz[5,6]acridino[2,1,9,8-klmna]acridine is capable of forming the most stable monolayer and the most labile bilayer. In the case of 8,16-bis(3-octylthiophen-2-yl)benzo[h]benz[5,6]acridino[2,1,9,8-klmna]acridine, an inverse phenomenon is observed leading to the most labile monolayer and the most stable bilayer. These differences are rationalized in terms of dissimilar molecular geometries of the studied isomers and different interdigitation patterns in their 2D supramolecular structures.

Introduction

Nonlinear and fused azaacenes are promising materials for application in organic electronics as components of active layers in field effect transistors (FETs)[1−4] and light-emitting diodes (LEDs).[5] Among them, phenazine- or acridine-type azaacenes deserve special attention, not only because of their excellent properties but also due to the simplicity of their preparation. In our previous papers, we demonstrated that these derivatives can be relatively easily obtained from old and almost forgotten nitrogen atom-containing vat dyes such as indanthrone (6,15-dihydrodinaphtho[2,3-a:2′,3′-h]phenazine-5,9,14,18-tetraone) or flavanthrone (benzo[h]benz[5,6]acridino[2,1,9,8-klmna]acridine-8,16-dione) in a two-step, one-pot reaction involving reduction of their carbonyl group followed by O-alkylation or another type of substitution.[6−8] These derivatives are technologically attractive because they are solution processable and exhibit excellent photo- and electroluminescence.[7,9] In addition, they are electrochemically interesting showing electrochromic properties associated with the reversibility of their reduction. All of these electronic and optoelectronic properties facilitate their application in various types of organic electronic devices.[7,9,10]

Additionally, one of the main features of organic semiconductors, which is important for their application in organic electronics, is their self-assembly in thin films or even in monolayers. This is a consequence of the fact that the patterns formed by self-assembly of organic molecules usually determine electrical transport properties and other properties of anisotropic character in active organic layers or at the active layer–metal electrode interface. The effect of such self-assembly is hard to predict, especially in the case of organic molecules of complex molecular and electronic structures. Scanning tunneling microscopy (STM) is a very well suited scientific tool for this kind of investigation and has been frequently used to study various families of organic semiconductors (see, for example, representative review articles[11−21]). Recently, we have applied this technique to elucidate the self-organization of several semiconductors of the donor–acceptor–donor type, such as arylamine-substituted naphthalene,[22] oligothiophene-substituted tetrazine,[23] thiadiazole,[23−25] tetraalkoxy-substituted dinaphthophenazine,[6] and oligothiophene-substituted diketopyrrolopyrrole.[26] The reported microscopic studies have indicated that self-assembly of such molecules of complex electronic structure usually results from a complicated effect of mutual interactions between parts of neighboring molecules of a different chemical nature and by consequence different electron affinities, as well as interactions of the adsorbed molecules with the substrate. For semiconductors in which intermolecular interactions of various types are of comparable strength, the resulting balance can be delicate. Thus, a small change in molecular heterostructure (i.e., strengths of the acceptor or the donor units, their mutual positions, and the length of alkyl/alkoxy substituents) can induce drastic changes in the resulting supramolecular organization (see, for example, ref (23)).

Alkoxy and alkylthienylene derivatives of flavanthrone described in this research, namely, 8,16-dialkoxybenzo[h] benz[5,6]acridino[2,1,9,8-klmna]acridines and 8,16-bis(3- or 4- or 5-octylthiophen-2-yl)benzo[h]benz[5,6]acridino[2,1,9,8-klmna]acridines, in addition to the previously mentioned excellent physical and chemical properties exhibit strong capability to form a highly ordered two-dimensional (2D) supramolecular structure, as demonstrated by preliminary STM studies.[7] In this paper, we present a detailed investigation of these self-assembly phenomena, analyzing the effect of the substituent length on the 2D self-organization of the studied alkoxy derivatives. We also compare the self-assembly patterns in monolayers of three positional isomers of octylthienylene-substituted fused benzoacridines differing in the position of the octyl substituent.

Experimental Section

8,16-Dialkoxybenzo[h]benz[5,6]acridino[2,1,9,8-klmna]acridines with increasing alkoxy substituent lengths (n-hexyl, n-octyl, n-decyl, and n-dodecyl) were synthesized following the procedure described in detail in ref (7). The syntheses of 8,16-bis(3 or 4 or 5-octylthiophen-2-yl)benzo[h]benz[5,6]acridino[2,1,9,8-klmna]acridines were carried out according to the procedure reported in ref (8). All synthesized compounds were identified by 1H and 13C NMR and HRMS, whereas their purity was determined by elemental analysis.

STM Investigations

For microscopic investigations, an STM system fabricated at the University of Bonn (Bonn, Germany) was used.[27] The monomolecular layers were prepared at room temperature by drop-casting from a solution (∼1–3 mg/L) of the investigated compounds in hexane (POCH) or dichloromethane (Sigma-Aldrich) on a freshly cleaved surface of HOPG (SPI Supplies). No specific solvent effect was noted. The surfaces of samples were imaged at molecular resolution after drying in air using mechanically cut Pt/Ir (80/20) tips. The control of the layer thickness, i.e., identification of monolayers or bilayers, was performed experimentally by correlation of the direct microscopic observation of the layer and adsorbate concentration in the solution used for the preparation. For each sample, the presented representative images were selected from a set of images collected for different surface areas. The proposed real-space models of monolayers were obtained by the correlation of the layer structure determined from the microscopic images and the molecular modeling of the adsorbate (HyperChem software package).

Density Functional Theory (DFT) Calculations

DFT calculations were carried out using Gaussian09 revision D.01[28] package and employing the hybrid B3LYP[29−31] exchange correlation potential combined with the 6-31G(d,p) basis set. Ground-state geometries were fully optimized until a stable local minimum was found, which was confirmed by normal-mode analysis (no imaginary frequencies were present). Initial structures were constrained to the C symmetry point group and then relaxed if a saddle point was found. The necessary data of DFT calculations were retrieved from output files using GaussSum 2.2.[32] All necessary initial geometries and final graphics (molecular orbitals) were generated in GaussView 5.0.[33]

Results and Discussion

The studied dialkoxy derivatives of benzo[h]benz[5,6]acridino[2,1,9,8-klmna]acridine, depicted in Figure a, are abbreviated as FOC6, FOC8, FOC10, and FOC12 hereafter, where 6, 8, 10, and 12 denote the number of carbon atoms in a given alkoxy substituent. Octylthienylene derivatives of dialkoxybenzo[h]benz[5,6]acridino[2,1,9,8-klmna]acridine are presented in Figure b and abbreviated as FT(3)C8, FT(4)C8, and FT(5)C8. The numerals 3–5 denote the position of the octyl substituent in the thiophene ring.

Figure 1

Chemical formulas of (a) 8,16-dialkoxybenzo[h]benz[5,6]acridino[2,1,9,8-klmna]acridines with different lengths of alkoxy substituents and (b) 8,16-bis(3- or 4- or 5-octylthiophen-2-yl)benzo[h]benz[5,6]acridino[2,1,9,8-klmna]acridines.

Self-Organization of 8,16-Dialkoxybenzo[h]benz[5,6]acridino[2,1,9,8-klmna]acridines: Effect of the Solubilizing Substituent Length

A representative STM image of FOC8 molecules deposited on the HOPG surface is presented in Figure a. The adsorbate molecules form a densely packed well-ordered monolayer consisting of clearly distinguishable, parallel rows. In particular, rows characterized by bright stripes are separated by darker zones. This image contrast reflects local differences in the monolayer electrical conductivity in the direction perpendicular to the substrate (HOPG) surface. It can therefore be postulated that the bright rows correspond to closely packed, fused azaacene cores (see Figure a) that are separated by dark areas where nonconductive alkoxy substituents are interdigitated. Higher-resolution images provide more information, especially concerning the inner structure of these bright stripes. First, as shown clearly in Figure b, each stripe consists of two rows of well-resolved bright spots. This peculiar inner structure is related to two principal factors, namely, (i) arrangement of the adsorbate in the molecular row and (ii) the electronic structure of its conjugated fused aromatic–azaacene core. At this point, it is worth analyzing the shape and localization of the frontier orbitals, i.e., HOMO and LUMO of FOC8 (see panels b and d of Figure ). There are some distinct differences in the spatial distribution of these orbitals. Although both HOMO and LUMO occupy the central part of the molecule, i.e., its fused aromatic–azaacene core, HOMO is preferentially located on the sides of this unit with the largest contribution from areas close to the oxygen atoms of the alkoxy substituents and with zero contribution from the two central carbon atoms of the conjugated core. Thus, two electron-rich parts of the molecule are clearly separated and symmetrically located. The distribution of the LUMO frontier orbitals is more diffuse comprising a strong contribution from the central part of the core and a still strong impact from the aforementioned oxygen and nitrogen atoms. It is worth emphasizing a very good correlation between the shapes of HOMO and LUMO orbitals and the observed positions and differences in contrast of bright stripes in the STM images obtained under scanning conditions corresponding to the electron tunneling in two opposite directions. In the case of a positively polarized STM probe, each bright stripe is visible as a set of two rows of spots (Figure a). This inner contrast is well focused with characteristic dark areas between the rows of each stripe. The observation mentioned above indicates that the central part of each FOC8 molecule, to be more precisely its fused aromatic–azaacene core, is visible in a submolecular contrast as two bright spots separated by a well-resolved darker area. This inner contrast is in accordance with the distribution of HOMO frontier orbitals with their higher degree of occupation near the oxygen atoms of the alkoxy substituents. Let us remind the reader that in the case of positive polarization of the probe, electrons tunnel from the deposited molecules to the probe and a higher tunneling current should be detected in areas of higher density of occupied states. Reversing the polarity leads to a different STM image. When the probe is negatively polarized, each bright stripe is more diffuse, showing significant brightness at the central part of the molecule. Again, a good correlation between the localization of LUMO and the STM image can be found (Figure c). The observed changes in the submolecular contrast of FOC8 molecule arise from a comparison of images collected one by one with the same STM tip, i.e., by performing experiments that exclude the influence of the tip shape and the measurement conditions on the resulting images. Their reproducibility was qualitatively confirmed by several independent experiments. In the conclusion of this part of the research, it should be stated that the submolecular contrast in the observed STM images of FOC8 monolayers is dependent on the bias voltage and is mainly determined by the electronic structure of the molecule fused aromatic–azaacene core.

Figure 2

(a–c) STM images and (d) corresponding model of adsorption geometry of the FOC8 monolayer on HOPG. Scanning area and parameters: (a) 15 nm × 12 nm, (b) 5 nm × 5 nm, It = 0.5 nA, Vtip = 1 V; (c) 5 nm × 5 nm, It = 2 nA, Vtip = 0.8 V.

Figure 3

(a and c) Higher-resolution STM images of the FOC8 monolayer on HOPG obtained for the opposite bias voltage polarity and (b and d) corresponding frontier molecular orbital plots [(b) HOMO, isosurface value of 0.03; (d) LUMO, isosurface value of 0.03]. Scanning area and parameters: (a) 6 nm × 6 nm, It = 1 nA, Vtip = 1 V; (c) 6 nm × 6 nm, It = 1 nA, Vtip = −0.8 V.

The second detail of the monolayer STM image is an inner structure of its darker areas. It can be noticed in Figure c that these areas are periodically divided by brighter lines oriented in one direction (marked by an arrow) that finally link conjugated cores of FOC8 molecules from two adjacent rows. As already stated, darker stripes in the image correspond to the areas of interdigitated alkoxy substituents originating from two adjacent rows formed by the molecules. The observed brighter lines in these dark stripes can tentatively be assigned to locally occurring differences in the density of alkoxy chains in the interdigitation area or, alternatively, to partially nonplanar arrangement of these chains with respect to the monolayer surface. It should be additionally stated that the direction of the observed lines corresponds to the orientation of alkoxy substituents in the monolayer. This interpretation of the obtained STM images allows us to postulate a plausible model of the adsorption geometry (Figure d). The lattice constants estimated from the image are 0.9 nm (a), 2.3 nm (b), and 96° (α). An interesting feature here is the nearly rectangular shape of the unit cell that is in contradiction with the 3-fold symmetry of the HOPG surface. This can be taken as an indication that the observed 2D organization is a result of direct interactions between the adsorbate molecules that dominate over the adsorbate–substrate interactions. The shorter dimension of the unit cell (0.9 nm) corresponds to the distance between adjacent molecules in the same row. It correlates well with the size of the fused aromatic–azaacene core of the molecule and confirms dense packing of these cores in each row of the monolayer. No correlation of this type can be found in the case of the longer dimension of the unit cell (2.3 nm), which is significantly shorter than the dimension of the molecule in its extended conformation (∼3.2 nm measured between terminated hydrogen atoms of two alkoxy substituents located on opposite sides of the molecule). The divergence is a consequence of interdigitation of the alkoxy substituents of molecules located in two adjacent rows. It is worth emphasizing that according to the postulated model of adsorption geometry the molecules in a particular row are oriented in the same direction in such a way that two nitrogen atoms of their fused aromatic–azaacene unit are located nearly along the short axis of the unit cell (atoms of the molecule marked in black in Figure d). A very similar structural arrangement was observed in a monolayer of tetraalkoxydinaphthophenazine molecules deposited on HOPG.[9,34] For these compounds, it was experimentally confirmed that the orientation of this adsorbate on HOPG and, as a consequence, the direction the nitrogen atoms pair in the conjugated core are forced by perpendicularity of the dialkoxyanthracene segments with respect to one of the graphite axis.

Statistical inspection of the FOC8 monolayer over larger areas allows us to distinguish six domain orientations, which are mutually rotated by a fixed angle of 30°. They are schematically indicated by white lines in Figure a. The orientations of three domains shown in the image are marked by d1–d3. At this point, the rather peculiar shape of the fused aromatic–azaacene core of FOC8 should be mentioned and as a consequence the whole investigated molecule. These kinds of molecules can adsorb on one of two opposite sides leading to two pseudoenantiomorphic forms in the plane of the monolayer surface. Therefore, six different orientations of domains should be considered as a set of three pairs of mirror symmetric directions with respect to each crystallographic axis of the HOPG substrate (three graphite axes are additionally presented in the image by dotted white lines). Each domain from one pair consists of molecules that are adsorbed by the opposite sides of the fused aromatic–azaacene core. This arrangement is schematically presented in Figure b. Two domains indicated in this scheme are rotated by ±15° with respect to axis (1), which represents one of the axes of the HOPG substrate surface. Because molecules from two different domains, from the left and right sides of this axis, are adsorbed by their opposite sides, their arrangement versus the directions of the formed rows exhibits mirror symmetry.

Figure 4

(a) STM large-scale image of the FOC8 monolayer on HOPG (six possible adsorbate domain orientations and HOPG substrate axes are marked by solid and dotted lines, respectively) and (b) proposed adsorption geometry in two domains with mirror symmetric directions vs one of the HOPG substrate axis (±15°). Scanning area and parameters: (a) 60 nm × 60 nm, It = 1 nA, Vtip = 1 V.

An important consequence of this molecular arrangement in the monolayer and the molecular row directions is the orientation of alkoxy substituents that are interdigitated along the two remaining axes of the HOPG substrate surface [marked by (2) and (3)]. The proposed model is consistent with the information published previously, concerning the adsorption of organic semiconductors of fused aromatic type and their derivatives on graphite substrates. The first characteristic point concerns the 2D enantiomorphic assembly, described by the “chirality” effect. This phenomenon is frequently observed and described even in fully achiral systems (adsorbate and substrates) for low- and high-molecular mass organic semiconductors.[35−37] The second one is associated with the preferential orientation of alkoxy substituents in the monolayer of FOC8 that are interdigitated exactly along one of the graphite axes. This type of linear hydrocarbon alignment is frequently observed on HOPG surfaces in the case of adsorption of saturated linear hydrocarbons[38−40] or fused aromatic or heteroaromatic molecules containing linear alkyl/alkoxy substituents.[41,42]

To determine the effect of alkoxy substituent length on the 2D supramolecular organization of fused benzoacridines, we performed STM investigations of a series of monolayers formed on HOPG by 8,16-dialkoxybenzo[h]benz[5,6]acridino[2,1,9,8-klmna]acridines of increasing substituent length [FOC6, FOC8, FOC10, and FOC12 (see Figure a)]. The unit cell parameters determined from the corresponding STM images are 0.9 nm, 2 nm, and 91°; 0.9 nm, 2.3 nm, and 96°; 1 nm, 2.6 nm, and 95°; and 1.1 nm, 3 nm, and 96° (a, b, and α, respectively) for FOC6, FOC8, FOC10, and FOC12, respectively. Figure shows representative images of monolayers of four investigated derivatives and the corresponding models of adsorption geometry obtained for molecules with the shortest (C6) and longest (C12) substituents. This comparison demonstrates that the adsorbed molecules are organized into the same row-like 2D structure, independent of the length of their alkoxy groups; i.e., their supramolecular organization is analogous to that determined for FOC8 (vide supra). It is therefore expected that unit cell parameter a should remain constant or fairly independent of the substituent length whereas its b parameter, reflecting the separation between the rows of closely packed fused aromatic–azaacene cores, should increase with extension of the substituent length. Indeed, the main difference between the presented images of monolayers of FOC6 and FOC12 is the width of the darker stripes, corresponding to the areas of interdigitated alkoxy chains of molecules from two adjacent rows. As a consequence, the size of the unit cell in this direction (longer dimension) increases from 2 nm for FOC6 to 3 nm for FOC12. When mutual interdigitation of the substituents in the layer is taken into account, the obtained difference correlates well with the length of the C–C bond in saturated hydrocarbon chains (∼0.14 nm). On the other side, the shorter dimension of the unit cell is nearly the same, varying from 0.9 to 1.1 nm for the series studied. This distance corresponds to the separation between fused aromatic–azaacene cores of two adjacent molecules in the same molecular row, which are of the same size, independent of the substituent. The postulated model of adsorption geometry is in accordance with the findings described above (see Figure e,f).

Figure 5

STM images of (a) FOC6, (b) FOC8, (c) FOC10, anbd (d) FOC12 monolayers on HOPG and corresponding models of their adsorption geometry of (e) FOC6 and (f) FOC12. Scanning area and parameters: (a, b, and d) 9 nm × 9 nm, It = 1 nA, Vtip = 1 V; (c) 9 nm × 9 nm, It = 0.8 nA, Vtip = 1 V.

To conclude this part of the paper, on the basis of the experimental STM data supported by modeling of the adsorption geometry, we can state that the 2D supramolecular organization of the derivatives studied is governed by direct interactions of their fused aromatic–azaacene cores, which are essentially independent of substituent. The only observed substituent effect is the change in the unit cell dimension in one particular direction corresponding to the orientation of these substituents in the monolayer. On the other hand, it should be emphasized that the alkoxy chains influence the formation of this type of organization due to their strong tendency to interdigitate. The simplest way to balance the core–core and substituent–substituent interactions is to force the same orientation of the molecules and organize them into a 2D system of parallel molecular rows.

2D Supramolecular Organization in Monolayers of 8,16-Bis(octylthiophen-2-yl)benzo[h]benz[5,6]acridino[2,1,9,8-klmna]acridines: Effect of the Solubilizing Alkyl Substituent Position

The second group of the investigated molecules involved three positional isomers of fused benzoacridines of the same benzo[h]benz[5,6]acridino[2,1,9,8-klmna]acridine core and octyl thienylene substituents with the solubilizing alkyl group located at position 3, 4, or 5 of the thienylene ring (Figure b). The position of the octyl group significantly changes the molecule geometry because the substituent can be directed either out of the molecule core in two different directions (position 4 vs position 5) or toward it (position 3 vs position 5). These should lead to a significantly different supramolecular organization. In Figure , STM images of monolayers of the three studied isomers are presented, together with the proposed adsorption geometry: FT(3)C8, FT(4)C8, and FT(5)C8. It should be noted that the images of these molecules observed with submolecular resolution are somehow similar to those recorded for the derivatives containing alkoxy substituents, which may indicate that the contribution of the thienylene ring to the STM image of the core is minimal. This is a direct consequence of similar distributions of frontier orbitals (HOMO and LUMO) in both types of molecules (see the corresponding data presented in refs (7) and (8)).

Figure 6

STM images and corresponding models of the adsorption geometry of (a, d, and g) FT(3)C8, (b, e, and h) FT(4)C8, and (c, f, and i) FT(5)C8 monolayers on HOPG. Scanning area and parameters: (a–c) 14 nm × 14 nm, (d–f) 8 nm × 8 nm, It = 1 nA, Vtip = 1 V.

It is instructive to start the discussion with a comparison of the supramolecular organization of the FT(5)C8 isomer (Figure c,f,i) with the corresponding octyloxy derivative, discussed above [FOC8 (Figure )]. The only difference in the molecular structure of these adsorbates is the type of linker between the octyl substituent and the molecule core, i.e., oxygen and thienylene ring in the cases of FOC8 and FT(5)C8, respectively. Both molecules are organized qualitatively in a very similar way, forming on HOPG row-like 2D structures. Also, the submolecular contrast in their STM images is nearly the same. In both cases, an individual molecule is visible as a set of two bright spots localized in the same areas that correspond to two parts of the fused aromatic–azaacene core directly connected to alkoxy-type oxygen (FOC8) or the thienylene ring [FT(5)C8]. The distinguishable quantitative differences between their supramolecular organizations are the STM-determined lattice constant values: 0.9 nm, 2.3 nm, and 96° and 1.2 nm, 3.2 nm, and 105° (a, b, and α, respectively) for FOC8 and FT(5)C8, respectively. The most pronounced difference is an increase in the longer unit cell dimension (b) from 2.3 to 3.2 nm when the octyloxy substituent is replaced with octylthienylene one. This increase can be considered as a manifestation of the substituent length increase. The shorter dimension of the unit cell changes to a lesser extent and stays around 1 nm. As already stated, this distance results from dense packing of the fused aromatic–azaacene cores of molecules constituting an individual row, which is essentially the same for both compounds.

The next problem to be discussed is the effect of the octyl substituent position in the thienylene ring on the 2D supramolecular organization in the monolayer of three positional isomers studied. Direct inspection of their STM images (Figure ) leads to an unexpected conclusion that the monolayers formed by these distinctly different isomers are characterized by the same unit cell parameters: 1.1 nm, 3.2 nm, and 105°; 1.1 nm, 3.1 nm, and 105°; and 1.2 nm, 3.2 nm, and 105° (a, b, and α, respectively) for FT(3)C8, FT(4)C8, and FT(5)C8, respectively. This indicates that the 2D supramolecular order is, in all three cases, governed by intermolecular interactions of fused aromatic–azaacene cores as they keep the same molecular arrangement in an individual row, independent of the isomeric form of the adsorbate. However, if it is assumed that the substituent alkyl chains are located in the same plane as the molecular cores (i.e., in the plane of the monolayer), the same dimensions of the unit cell, determined for different isomers, should have an important consequence for their supramolecular arrangement. Depending on the isomer, they have to adopt different arrangements in the monolayer to yield the same unit cell. This becomes evident by a comparison of the adsorption geometry models postulated for monolayers of these three isomers (Figure g–i). To rationalize the experimental findings, i.e., very similar unit cells obtained experimentally for all three isomers studied, despite distinctly different orientations of their alkyl chain with respect to the fused core, different extents of alkyl substituent interdigitation must be assumed, from full interdigitation [FT(5)C8 (Figure i)] to negligible [FT(3)C8 (Figure g)]. The supramolecular organization in the monolayer of FT(4)C8 is an intermediate case in which alkyl chains partially interdigitate (Figure h). It seems that these distinct differences in the supramolecular arrangement of the studied isomers should have an impact on their monolayer stability that should increase with an increasing degree of alkyl substituent interdigitation, i.e., from FT(3)C8 to FT(5)C8.

Self-Organization of 8,16-Bis(octylthiophen-2-yl)benzo[h]benz[5,6]acridino[2,1,9,8-klmna]acridines in a Bilayer

Different types of arrangements in monolayers of the three studied isomers raise a question concerning the self-organization and its stability in their multilayers. Figure presents a set of two images and profiles of selected lines of a layer of FT(4)C8 on either individual molecule (Figure a,c) or whole molecular rows (Figure b,d) of the second layer. The crucial point in these images is a significant difference in the brightness of the molecules of the first layer and the second layer. This results from a different vertical position of the STM tip with respect to each layer. In the case of the large-area image (Figure a), individual molecules nucleating the second layer are observed as bright spots. They are randomly distributed on darker rows corresponding to orderly arranged molecules of the first layer. It is worth reminding the reader that each molecule in the submolecular resolution appears as a set of two bright spots. Consequently, in the image presented in Figure b, which was collected with higher magnification, the single molecular row consists of two rows of spots. Hence, darker rows correspond to the first layer whereas brighter ones to the second layer. The presented height profiles of selected lines from both images unequivocally confirm similar thicknesses of molecules in the first layer and in the second layer. An inspection of presented images leads to the conclusion that individual molecules as well as molecular rows in the second layer are located directly above the molecular rows in the first layer. The same behavior was also observed for the remaining studied isomers. This is logical, taking into account the adsorption geometry of molecules in the monolayer, i.e., flat orientation of fused aromatic–azaacene cores with respect to the substrate surface leading to mutual π–π interactions. As a consequence, adsorption of molecules in the second layer directly above the molecules in the first layer is the favored interaction because it leads to π–π stacking oriented perpendicular to the substrate surface.

Figure 7

STM images and height profiles of FT(4)C8 (a and c) monolayers and (b and d) bilayers on HOPG. (c and d) h0 = 0.36 nm, which within an experimental error corresponds to the thickness of one graphene layer and was determined from the edge of the uncovered HOPG surface. Scanning area and parameters: (a) 60 nm × 60 nm, (b) 30 nm × 30 nm, It = 1 nA, Vtip = 1 V.

Figure presents a set of four images of FT(3)C8 layers obtained by drop-casting from dichloromethane solutions of different concentrations (which varied from 1.5 to 3 mg/L). As expected, an increase in the adsorbate concentration in the solution used for drop-casting results in more complete coverage of the first layer by the second one as manifested by a larger number of molecular rows in the second layer [i.e., brighter rows in the image (Figure a–d)]. The set of presented images clearly shows the mechanism of growth of the second layer. For a lower concentration of FT(3)C8 in casting solutions, randomly distributed rows of the second layer appear on the first layer. A gradual increase in the adsorbate concentration leads to the progressive filling of empty spaces between the rows by molecules up to the complete coverage. This implies a typical layer-by-layer growth (Frank–van der Merwe mode). When this observation is taken into account, one can expect that FT(3)C8 molecules are capable of forming well-ordered multilayer films.

Figure 8

STM large-scale images of FT(3)C8 layers on HOPG prepared with a solution used for drop-casting of different adsorbate concentrations: (a) 1.5, (b) 2.2, and (c and d) 3 mg/L. The coexisting areas of molecular rows in monolayers and bilayers are noted, distinguished as darker and brighter rows, respectively. Scanning area and parameters: (a–c) 30 nm × 30 nm, (d) 120 nm × 120 nm, It = 1 nA, Vtip = 1 V.

The same procedure was applied for the fabrication of FT(4)C8 and FT(5)C8 multilayers, revealing, however, significant difficulties in the formation of their ordered films. The difficulties are significantly higher for FT(5)C8 than for FT(4)C8. For both adsorbates, the applied procedure leads to ordered films up to incomplete double layers. Moreover, its stability is significantly lower and the layer can be relatively easily destroyed on a local scale during scanning by the microscopic probe. When the preference of the formation of “missing rows” in the second layer, which strongly depends on the isomeric form of the adsorbate molecules, is taken into account, it seems that the observed differences are of thermodynamic rather than kinetic origin.

To understand the origin of the observed difficulties, it is instructive to carefully analyze the growth of FT(4)C8 films, for which the limitations are more pronounced. Representative images are presented in panels a and b of Figure . As already stated, this isomer forms ordered double layers of the incomplete second layer coverage and showing limited stability. In this respect, the crucial information can be extracted from the image presented in Figure b. This image shows the growth of a second layer in two adjacent empty rows surrounded by already formed rows of the second layer. This growth process occurs in two rows from two opposite directions (marked by white arrows). As evidenced by this image, the growth is stopped when the molecules of the two growing rows come into mutual contact. This effect, frequently observed in different areas of the sample surface, indicates steric limitations of the second layer formation. As a consequence, the molecules in the second layer most frequently fill two rows that are separated by an empty row. This steric limitation is even more pronounced in the case of multilayers of FT(5)C8. For this adsorbate, it is possible to obtain incomplete ordered double layers in which molecules of the second layer occupy every second molecular row only (Figure c,d). For higher coverages, the layer becomes unstable and can be easily disturbed by the scanning microscopic probe. The obtained results are intriguing in the sense that the FT(5)C8 isomer that produces the most stable monolayer with fully interdigitated alkyl groups is incapable of forming a stable bilayer. The observed bilayers of this isomer suffer from incomplete coverage of the bottom layer as well as from serious steric limitations in the self-ordering process. On the contrary, the FT(3)C8 isomer, whose monolayer is the least stable, shows strong ability to form a well-ordered bilayer via layer-by-layer growth. The presented comparison can be considered as an evident indication that the self-organization processes in 2D and 3D systems can be governed by different conditions and requirements. A question concerning the identification of the factors that make the behavior of both isomers different arises. These factors selectively affect the self-organization of molecules placed either on an atomically flat HOPG surface or on a chemically less homogeneous surface of the first molecular layer. Additional investigations lead to conclusions that the molecular shape can be considered as one of these factors. Figure shows the molecular geometries in vacuum determined by DFT calculations. It follows that the alkyl substituent tilt angle with respect to the fused heteroaromatic core (angle marked by α) is the largest for FT(3)C8 (∼61°), making this isomer the least planar. This is a direct consequence of the fact that in this isomer the substituents are positioned more closely to the molecular core as compared to the cases of two other isomers studied. Thus, for FT(3)C8, more pronounced intramolecular interactions are expected between these different parts of the molecule. Although FT(4)C8 and FT(5)C8 are not planar either, their calculated tilt angles are evidently smaller (37° and 41°, respectively). This dissimilarity in the shape of individual molecules can be important from the point of view of surface diffusion and self-organization in the second layer. In general, due to the tendency of the rather extended cores of FT(4)C8 and FT(5)C8 to strongly interact via π–π stacking, limited surface diffusion of the molecules in the second layer is expected, which hinders the ordering process in this layer. As expected, this limitation should be less pronounced for FT(3)C8 because of its nonplanar shape. Therefore, this isomer should be more able to form ordered multilayers.

Figure 9

STM images of (a and b) FT(4)C8 and (c and d) FT(5)C8 bilayers on HOPG. Scanning area and parameters: (a) 30 nm × 30 nm, (b) 16 nm × 16 nm, (c) 120 nm × 120 nm, (d) 30 nm × 30 nm, It = 1 nA, Vtip = 1 V.

Figure 10

Comparison of (a) FT(3)C8, (b) FT(4)C8, and (c) FT(5)C8 molecular geometries determined by DFT calculations (top images, top views; bottom images, side views).

Conclusions

These STM investigations of the supramolecular organization in mono- and bilayers of two series of alkoxy and alkylthienylene derivatives of fused benzoacridines, formed on HOPG by casting from dichloromethane solutions, enabled us to draw the following conclusions.

(i) All investigated derivatives exhibited a strong tendency to self-assemble, forming highly ordered monolayers. The observed 2D aggregations were characterized by clearly distinguishable, parallel molecular rows. A clear dependence of the submolecular STM contrast on the tunneling current direction (polarity of bias voltage) was noted. It could be directly correlated with differences in the location of electron-poor and electron-rich parts of the molecules, reflected by the distribution of the molecular orbitals (LUMO and HOMO) in their central fused aromatic–azaacene cores.

(ii) Comparative investigations of alkoxy derivatives indicated that an increasing alkoxy substituent length affected only one dimension of the unit cell, i.e., that which corresponded to the separation between the molecular rows consisting of densely packed aromatic–azaacene cores. This observation confirmed a similar orientation of the alkoxy substituents with respect to the molecule core and their interdigitation as the main self-assembly pattern.

(iii) The same supramolecular organization (unit cell parameters) was found for the investigated alkylthienylene derivatives (three positional isomers differening in the alkyl substituent position). Despite this similarity, the role of alkylthienylene substituents in the formation of monolayers of each of the investigated isomer was distinctly different. Depending on the alkyl group position in the thienylene ring, the studied isomers had to adopt different arrangements in the monolayer to yield the same unit cell. As a consequence, the formation and stability of monolayers and bilayers were found very sensitive to isomerism. These differences were rationalized in terms of dissimilar molecular geometries and different interdigitation patterns.