Angewandte
Chemie
tarry material of unknown composition could be isolated. The
absence of cyclodimers was surprising, as the same coupling
conditions worked well in most of our acetylene coupling
reactions and gave (with a similar but smaller substrate) high to
nearly quantitative product yields; see: N. Shabelina, S. Klyat-
et al. also obtained cyclic products when bipyridyl containing
bis(acetylene)s were oxidatively coupled under copper catalysis:
D. M. Opris, A. Ossenbach, D. Lentz, A. D. Schlꢁter, Org. Lett.
persistent macrocycles under copper catalysis; see: J.-K. Kim, E.
[22] In Figure 1, Figure A2(b–f), and Figure 2B(b,d), the unit cell
vectors along the lamellar directions are oriented with 9 Æ 28
towards the main axis of the HOPG substrate (see Supporting
Information), indicating surface induced chirality of the mole-
cules. For the alkoxy substituents, we assume the common
commensurability of all-gauche-constituted alkyl chains on the
graphene surface lattice.[23] We conclude that the angle between
the rigid PEB rods and the alkyl chains must be 81 Æ 28 for the
considered systems [5a]n and [6a]n. Clearly, this angle is not a
generally constant value for all alkoxy-substituted rigid systems,
but will rather strongly depend on the concrete PEB backbone
sequence. However, as individual short alkoxy chains (such as
OC6H13 in this case) cannot be resolved by STM under the
applied conditions, their orientation shall not be further
discussed herein.
[14] 1b and 4b were not investigated under the conditions of the
copper-catalyzed reaction.
[15] Only the cyclodimer [6b]2 was identified when 4b was coupled
under palladium catalysis; higher oligomers were not charac-
terized.
[23] For SAMs of alkanes on HOPG, see for example: T. Yang, S.
[16] Considering 1b and 4b with 2-ethylhexyloxy side chains,
solubility is clearly not responsible for the observed discrim-
ination between cyclic and acyclic products.
[24] The sizes of the STM images, tunneling parameters, applied
solutions, and annealing temperatures, and the dimensions of the
unit cells of the self-assembled monolayers of acyclic oligomers
[5a]2–6 are as follows: b) dimer [5a]2 (34.0 ꢆ 34.0 nm2, Vs =
À1.1 V, It = 150 pA, c = 10À5 molLÀ1, unit cell: a = 3.9 Æ 0.1 nm,
b = 2.7 Æ 0.1 nm, g = 68 Æ 28); c) trimer [5a]3 (32.4 ꢆ 32.4 nm2,
Vs = À1.35 V, It = 100 pA, c = 10À5 molLÀ1, annealed to 608C for
2 min, unit cell: a = 8.2 Æ 0.2 nm, b = 3.8 Æ 0.1 nm, g = 90 Æ 28);
d) tetramer [5a]4 (28.0 ꢆ 28.0 nm2, Vs = À1.1 V, It = 120 pA, c =
10À6 molLÀ1, annealed to 808C for 2 min, unit cell: a = 5.4 Æ
0.2 nm, b = 3.8 Æ 0.1 nm, g = 90 Æ 28); e) pentamer [5a]5 (24.3 ꢆ
24.3 nm2, Vs = À0.85 V, It = 11 pA, c = 10À5 molLÀ1, annealed to
808C for 2 min, unit cell: a = 13.5 Æ 0.2 nm, b = 3.8 Æ 0.1 nm, g =
90 Æ 28); f) hexamer [5a]6 (25.3 ꢆ 25.3 nm2, Vs = À0.3 V, It =
246 pA, c = 10À5 molLÀ1, annealed to 808C for 2 min, unit cell:
a = 8.0 Æ 0.2 nm, b = 3.8 Æ 0.1 nm, g = 90 Æ 28). The respective
parameter sets for cyclic oligomers [6a]2–6: b) dimer [6a]2 (20.9 ꢆ
20.9 nm2, Vs = À0.42 V, It = 14 pA, c = 10À5 m, unit cell: a = 3.8 Æ
0.1 nm, b = 2.7 Æ 0.1 nm, g = 69 Æ 28); c) trimer [6a]3 (30.8 ꢆ
30.8 nm2, Vs = À1.1 V, It = 5 pA, c = 10À6 molLÀ1, annealed to
808C for 2 min, unit cell: a = 10.1 Æ 0.2 nm, b = 4.7 Æ 0.1 nm, g =
44 Æ 28); d) tetramer [6a]4 (26.0 ꢆ 26.0 nm2, Vs = À1.43 V, It =
14 pA, c = 10À5 molLÀ1, annealed to 808C for 2 min, inset:
12.1 ꢆ 12.1 nm2, Vs = À1.20 V, It = 10 pA, unit cell: a = 7.6 Æ
0.2 nm, b = 2.7 Æ 0.1 nm, g = 79 Æ 38); e) pentamer [6a]5 (50 ꢆ
50 nm2, Vs = À0.87 V, It = 43 pA, c = 10À5 molLÀ1, annealed to
808C for 2 min, amorphous); f) hexamer [6a]6 (40 ꢆ 40 nm2, Vs =
À0.6 V, It = 14 pA, c = 10À5 molLÀ1, annealed to 808C for 2 min,
expected unit cell: a = 11.5 nm, b = 2.7 nm, g = 838). Note: a and
b denote the long and short unit cell vectors, respectively. By
definition, the unit cells for (sufficiently large) odd acyclic
oligomers [5a]n (n = 3, 5) are larger than for the respective even
oligomers (n = 4, 6). All images were calibrated in situ using the
HOPG substrate as reference grid (see the Supporting Informa-
tion).
[17] Proton NMR studies (see Supporting Information) showed that
all oligomers obtained from the copper-promoted coupling
reactions still contain ethynyl end groups, whereas all separated
fractions of the palladium-catalyzed coupling reactions did not
exhibit ethynyl protons. Furthermore, the peak patterns of the
aromatic protons in the spectra are different for both reactions.
HRMS data for [5a]2 and [6a]2 show the exact masses for the
proposed structures.
[18] The difference in conformational freedom between the acyclic
and cyclic oligomers is also reflected by their thermal behavior.
Although the acyclic dimers [2a]2 and [5a]2 have melting points
of 718C and 958C, the cyclic analogue [3a]2 melts at 2218C and
[6a]2 decomposes at > 3258C; see: H. A. Staab, H. Brꢂumling,
[19] For theoretical descriptions of the contrast mechanism in STM,
see for example: a) R. Lazzaroni, A. Calderone, J. L. Brꢅdas, J. P.
[20] As visualized by the inset of the molecular model in Fig-
ure 2A(b), the distance of the terminal ethynyl units is smaller
than the lateral changes of the tunneling current. Therefore, we
cannot clearly resolve the difference of the cyclic and acyclic
dimer, [5a]2 and [6a]2, by microscopy. Nevertheless, the presence
of a cyclic dimer [6a]2 as impurity in [5a]2 could be excluded by
GPC, HRMS, and 1H NMR analysis.[17] Furthermore, the
presence of two distinguishable O- and S-shaped polymorphs
of the dimer [14a]2 on HOPG is rather expected instead of
peculiar.
[21] STM image size, tunneling parameters, concentration, and unit
cell dimensions for the SAM of the monomer 4a on HOPG are:
24.1 ꢆ 24.1 nm2, Vs = À0.95 V, It = 6 pA, c = 10À4 molLÀ1
;
O-shaped polymorph: a = 3.8 Æ 0.1 nm, b = 2.7 Æ 0.1 nm, g =
68 Æ 28; S-shaped polymorph: a = 3.6 Æ 0.1 nm, b = 2.7 Æ
0.1 nm, g = 90 Æ 28.
Angew. Chem. Int. Ed. 2010, 49, 6101 –6105
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6105