An Approach to the Synthesis of Functionalized Polycyclic Aromatic Hydrocarbons.

The application of a new benzannulation reaction for the regiocontrolled synthesis of functionalized chrysenes is reported. The initial benzannulation and the subsequent halogen displacement reactions are both highly regiospecific, which thereby enables the regiocontrolled synthesis of a variety of 4,10-disubstituted chrysenes from commercially available 1,5-dihydroxynaphthalene.

Introduction

Polyaromatic hydrocarbons (PAHs) have been a focus of research in the field of organic electronics since the first acene semiconductor devices were fabricated from materials such as pentacene and rubrene.1 Linear PAHs have since received the majority of attention, with many reporting high-mobility devices.2 More recently, numerous reports on the synthesis of linear acenes possessing heterocyclic structures have also appeared, and this area shows much promise for the development of new devices. In contrast, the chemistry of “nonlinear” or “angular” PAHs3,4a,4b is much less well developed despite the fact that these PAHs should be more stable than their linear analogues on the basis of Clar's paradigm.3,4c,4d Furthermore, few nonlinear PAHs have been investigated as potential semiconductor materials,5a–5c a situation which in part is due to the lack of general synthetic methods in this area.5d,5e

Results and Discussion

The synthesis of substituted acenes has focused, until now, on tandem aldol condensations6 to afford the corresponding acenequinones, which then may be derivatized and reduced to the acenes, most commonly their ethynyl derivatives.7 The development of new synthetic methods leading to the regiospecific synthesis of functionalized, nonlinear polyacenes is imperative if structure–function correlations are to be made for these compounds. In this paper, we present the results of an initial study directed towards the design and synthesis of polyaromatics made readily accessible through our newly developed benzannulation sequence, that is, the Bull–Hutchings–Quayle (BHQ) reaction. The BHQ reaction was discovered by Quayle et al. as a novel means to prepare benzannulated aromatics such as 1-chloronaphthalene (2) from 2-allylphenyl trichloroacetate (1, Scheme 1).8

Scheme 1

The BHQ reaction.

This new atom-transfer radical cyclization (ATRC)–benzannulation sequence is tolerant of a wide range of common functional groups and exhibits complete regiospecificity for the aryl chloride isomer shown. Moreover, allyl trichloroacetates such as 6 are readily available from phenols by using “textbook” chemistry, and the resulting four-step pathway from phenol 3 to benzannulated product 7 can be readily accomplished on a multigram scale over a short time period (Scheme 2).

Scheme 2

The pathway from a phenol to a polyaromatic aryl chloride.

Many aryl chlorides have been synthesized by using this methodology, including its extension to a “two-directional” double benzannulation sequence. This variant is most appealing if considering the development of new synthetic strategies for PAHs, which would become available from ubiquitous bis(phenols). In this regard, we have shown that 4,10-dichlorochrysene (12), a nonlinear phenacene with a novel substitution pattern (Scheme 3), is now readily available from bis(phenol) 8.

Scheme 3

From 1,5-dihydroxynaphthalene to 4,10-dichlorochrysene. Reagents and conditions: (a) allyl bromide (2.4 equiv.), K2CO3 (2.4 equiv.), acetone, 22 h, r.t.; (b) 210 °C, N2, neat; (c) Cl3CCOCl, pyridine, Et2O, 0 °C, 2 h; (d) CuCl (5 mol-%), diglyme, 162 °C, 2 h.

A range of conditions were investigated for the BHQ reaction of 11 leading to 12; gratifyingly, we found that 12 could be routinely obtained upon thermolysis of 11 in diglyme (b.p. 162 °C) containing CuCl (5 mol-%) in yields ranging from 38 to 43 % (Scheme 3).

Having optimized the route to 12, we then wished to investigate its functionalization. Aryl chlorides usually react sluggishly in SNAr or metal-catalyzed displacement reactions,9 although we were encouraged by the initial reports of Scott et al.10 and more recently by Douglas et al.11 concerning the Kumada and Suzuki cross-coupling reactions of related chloroaromatics under the auspices of palladium catalysis.12 Significantly, the single-crystal X-ray structure of 12 has a distorted structure (Figure 1), an observation, which apparently was translated into enhanced reactivity of the C–Cl bonds, such that Pd-catalyzed cross-coupling, Cu-promoted coupling, and nucleophilic aromatic substitution reactions are also preparatively viable (Table 1).

Figure 1

X-ray crystal structures of compounds 12, 13, 14, 17, and 18 (hydrogen atoms omitted for clarity).

Table 1

Functionalization of 4,10-dichlorochrysene

Compound	R	Method[a]	Yield [%]
13	Me	a	78
14	Ph	a	91
15	p-MeOC6H4	a	69
16	1-naphthyl	b	60
17	3-thienyl	b	75
18	SPh	c	86
19	SNp	c	40
20	OPh	d	78
21	1-octyn-1-yl	e	61
22	H	f	–

Method a: RMgBr (4 equiv.), PEPPSI-IPr (5-mol-%), THF (4 mL per 100 mg of 12), r.t., 30 min. Method b: R(BOH)2 (2.2 equiv.), KOtBu (2.2 equiv.), PEPPSI-IPr (2 mol-%), EtOH (1 mL per 100 mg of 12), r.t., 30 min. Method c: PhSH (2.4 equiv.), K2CO3 (3 equiv.), DMF (1 mL per 100 mg of 12), 100 °C, 6 h. Method d: PhOH (2.4 equiv.), Cs2CO3 (3 equiv.), CuI (2.4 equiv.), diglyme (4 mL per 100 mg of 12), 150 °C, 48 h. Method e: 1-Octyne (2.2 equiv.), Cs2CO3 (2.4 equiv.), Pd(PPh3)2Cl2 (6 mol-%), PCy3 (15 mol-%), DMF (1 mL per 100 mg of 12), 110 °C, 40 h. Method f: Isolated as a byproduct of high-temperature coupling. PEPPSI = pyridine-enhanced precatalyst preparation, stabilization, and initiation, IPr = 1,3-bis(2,6-diisopropylphenyl)imidazol-2-ylidene, Cy = cyclohexyl, Np = naphthyl.

Dichloride 12 shows considerable stability to the harsh reaction conditions (>150 °C) of these reactions, and it is resistant towards decomposition even at reaction temperatures in excess of 200 °C for extended periods of time. In spite of this stability, most Pd-catalyzed coupling reactions can be satisfactorily performed without heating, although displacement with oxygen- and sulfur-centered nucleophiles requires somewhat harsher reaction conditions. We were able to prepare a small library of 4,10-disubstituted chrysenes, the structures of which were confirmed by single-crystal X-ray diffraction. Substitution of chrysene in this manner resulted in the molecule adopting a twisted geometry; 12 has a torsional bend of 7.24° about the central bond of the molecule. This bending is not symmetric in 12, 13, and 14 but is symmetric in 17 and 18.13 The substitution and the resultant twisting have a profound effect on the crystal packing of the compound; 12 assumes a 1D π–π stacking arrangement with an interlamellar distance of 3.775 Å, which is somewhat longer than the sum of the van der Waals radii. Bis(phenyl) derivative 14 develops a warped 2D brickwork motif dominated by H–π interactions with no observable π–π overlap, whereas both 17 and 18 adopt a familiar herringbone packing, and 18 has a π–π short contact of 3.325 Å. This small selection of chrysene derivatives exemplifies the complex relationship between molecular structure and solid-phase morphology that makes the predictive design of materials so difficult.

To determine the electronic properties of the compounds, they were analyzed by cyclic voltammetry (CV) and UV/Vis spectroscopy. The UV/Vis spectra are characterized by high-intensity absorption between 260 and 300 nm and then a series of lower-intensity bands between 300 and 350 nm (Figure 2). For molecules with more degrees of conformational freedom, these lower-intensity, lower-energy bands are absorbed into the more intense peak.

Figure 2

Normalized UV/Vis spectra of selected chrysene derivatives.

Oxidation potentials were measured by using a ferrocene internal standard, in which Fc/Fc+ was taken to be 0.64 V. From this, the energy of the HOMO may be estimated relative to the ferrocene HOMO level of –4.8 eV. The chrysenes described herein display high-energy band-gap transitions and high oxidation potentials, which are typical for nonlinear PAHs. Derivatization of chrysene, as illustrated by this study, has a marked effect upon the electronics of the molecule; values of λmax for these compounds range from 270 nm for 13 to 292 nm for 21 (see the Supporting Information for the full data). The oxidation potentials seem to display a relationship with the electronic nature of the substituents, as expected (Table 2). The graphs in Figure 3 show the experimentally acquired UV/Vis absorption curves overlaid with those derived from DFT calculations (Supporting Information). The calculations14 suggest that there is no significant delocalization from the chrysene core to the periphery, except in 18 and 19.

Table 2

Oxidation potentials and HOMO energy levels

R	Oxidation potential[a]	HOMO level
	[V]	Exp.[b]	Calcd.[c]
Cl	1.83	–5.99	–5.92
Me	1.61	–5.77	–5.55
Ph	1.72	–5.88	–5.58
SPh	1.62	–5.78	–5.38

Oxidation potential relative to standard hydrogen electrode (SHE) by using ferrocene as an internal standard for which V(Fc/Fc+) = 0.64.

Experimental HOMO levels estimated relative to the Fc/Fc+ couple for which HOMO(Fc) = –4.8 eV.

B3LYP/6-311G(d,p) calculations based on 6-31G(d,p) geometries.

Figure 3

Comparison of the experimentally acquired (gray) and computed (black) UV spectra. (a) 12 (R = Cl), (b) 13 (R = Me), (c) 14 (R = Ph), (d) 18 (R = SPh).

When 12 was spin-coated onto glass from anisole, a noncrystalline film was observed. When spin-coated as a 80:20 w/w blend with amorphous statistical copolymers of styrene (PS) and 4-methoxystyrene (MeOPS; Mn = 6 kDa), crystallization of the organic semiconductor was induced (Figure 4).

Figure 4

Optical micrograph (200 × 200 μm) of thin films of 12 coated as a 80:20 w/w blend with amorphous statistical copolymers of PS-MeOPS (Mn = 6 kDa). (a) Pure PS, (b) 1:1 mol ratio PS/MeOPS, and (c) 1:3 mol ratio PS/MeOPS.

Although the band gap of 12 is too large to be a useful material for use in a field-effect transistor, it is interesting to note the effect of an amorphous binder on the crystallization of 12. We ascribe this effect to favorable interactions between slightly electron-deficient 12 and the more electron-rich statistical copolymer of styrene and 4-methoxystyrene. This has previously been reported for linear acenes,15 but the effect of fine-tuning the electron density of the binder has previously not been disclosed.

Conclusions

This preliminary investigation serves as proof of principle and demonstrates that the BHQ benzannulation reaction can be used to generate novel scaffolds, which enables easy access to a range of chrysene derivatives that are not readily accessible by existing methods.16 The reactivity of the C–Cl bonds in 4,10-dichlorochrysene enables derivatization of the chrysene core. Application of this basic benzannulation sequence to the synthesis of higher PAHs and investigations into their electronic properties is now in progress. The potential to induce crystallization in thin films of PAHs by judicious blending with amorphous polymeric binders opens up opportunities in controlled crystal engineering and control of the resultant electronic properties.

CCDC-933450 (for 12), -CCDC-933451 (for 17), -CCDC-933452 (for 14), -CCDC-933453 (for 13), and -CCDC-933454 (for 18) contain the supplementary crystallographic data for this paper. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.

Supporting Information (see footnote on the first page of this article): Full experimental details, including the synthesis, and UV data.