The synthesis of brominated-boron-doped PAHs by alkyne 1,1-bromoboration: mechanistic and functionalisation studies

The synthesis of a range of brominated-Bn-containing (n = 1, 2) polycyclic aromatic hydrocarbons (PAHs) is achieved simply by reacting BBr3 with appropriately substituted alkynes via a bromoboration/electrophilic C–H borylation sequence. The brominated-Bn-PAHs were isolated as either the borinic acids or B-mesityl-protected derivatives, with the latter having extremely deep LUMOs for the B2-doped PAHs (with one example having a reduction potential of E1/2 = −0.96 V versus Fc+/Fc, Fc = ferrocene). Mechanistic studies revealed the reaction sequence proceeds by initial alkyne 1,1-bromoboration. 1,1-Bromoboration also was applied to access a number of unprecedented 1-bromo-2,2-diaryl substituted vinylboronate esters directly from internal alkynes. Bromoboration/C–H borylation installs useful C–Br units onto the Bn-PAHs, which were utilised in Negishi coupling reactions, including for the installation of two triarylamine donor (D) groups onto a B2-PAH. The resultant D–A–D molecule has a low optical gap with an absorption onset at 750 nm and emission centered at 810 nm in the solid state.


Introduction
The incorporation of main group elements into conjugated organic scaffolds is a powerful strategy to tune electronic properties. 1 Recently, this approach has been extended to the introduction of three-coordinate boron atoms into polycyclic aromatic hydrocarbons (PAHs). 1,2 In these "B-doped"-PAHs interaction between the empty p orbital on boron and the extended p system oen leads to molecules with low energy LUMOs. 2 B-doped PAHs now are being explored for a range of applications including organic light-emitting diodes, thin lm transistors, solar cells, lithium batteries and electrocatalysis. 2 However, two factors have limited the wider uptake of purely Bdoped PAHs (i.e. PAHs without co-doping elements e.g. N): (i) the limited simple synthetic routes to form B n-PAHs, particularly for n > 1, which generally have the deeper LUMOs; 2 (ii) the challenge associated with incorporating B n -doped-PAHs into more complex materials, e.g. by coupling reactions to access donor-acceptor (D-A) structures. 3 Some notable progress has been made addressing (i), with several routes to larger B n -PAHs, (n $ 2, see A-D, Fig. 1) including examples with low LUMO energies, recently reported. 4,5 However, the incorporation of B n -PAHs into more complex materials currently requires the post synthetic introduction of additional functionality onto the B n -PAH to enable subsequent steps (e.g. cross coupling). 6 This has associated challenges (e.g. selectivity in halogenation) along with step-economy issues. Therefore, a simple route to form a range of B n -doped PAHs with low LUMO energies that concomitantly installs a second useful functional group, such as halide, would be highly attractive and facilitate access to more complex materials with desirable properties.
In our previous work, a one-pot synthesis of 1-boraphenalenes (E) via sequential bromoboration/ intramolecular boron-Friedel-Cras reaction of 1-alkynylnaphthalenes ( Fig. 1, middle) was described. 7 We envisaged that this methodology could be extended by: (i) using larger (than naphthalene) p-scaffolds to obtain B 1 and B 2 doped PAHs possessing deeper LUMOs; (ii) utilising the bromo substituent to access more complex materials with desirable properties via coupling reactions. Herein these two objectives are realised enabling access to the lowest LUMO energy ambient stable B ndoped PAH reported to date, to the best of our knowledge, and a low-optical gap D-A-D material where the acceptor unit is a B 2 -PAH. In addition, mechanistic studies on the bromoboration/S E Ar process revealed it proceeds by alkyne 1,1-bromoboration, a process not previously observed. Therefore 1,1-bromoboration also was applied to simple diarylalkynes demonstrating its utility for forming 1-bromo-2,2-diaryl substituted vinylboronate esters that are otherwise challenging to access.

Results and discussion
Synthesis of boron-doped PAHs and 1,1-bromoboration studies Initially, 1-(pent-1-yn-1-yl)pyrene, 1, was combined with BBr 3 in the presence of one equivalent of 2,4,6-tri-tert-butylpyridine (TBP), targeting an analogous trans-haloboration/S E Ar process that previously afforded 1-boraphenalenes. 7 Post protection at boron by reaction with dimesitylzinc, purication afforded two isomeric compounds, 1a and 1b (Scheme 1) in low yield, partly due to low stability of these compounds to the column chromatography. To provide enhanced stability, the bulkier 2,4,6triisopropylphenyl (Tip) protecting group was used instead of mesityl (Mes). Using similar reaction conditions, compounds 1c and 1d were isolated in higher yields relative to 1a and 1b. B 2 -doped PAHs were next targeted through a two-fold bromoboration/S E Ar sequence starting from 1,6-di(pent-1-yn-1-yl) pyrene, 2. However, post borylation/protection at boron, compound 2a was the only fully fused B 2 -PAH isolable via column chromatography in our hands and it was obtained in a very low yield alongside 2b. Compound 2b presumably derives from protodeboronation of a vinyl C-B unit (analogous to that in 1a) followed by an E2 elimination reaction as previously observed for the 1-boraphenalenes (see Fig. 1, middle right). 7 Tip installation conditions also were applied to the product mixture derived from BBr 3 /2, however this led to the formation of compound 2c in 10% yield as the only isolable product in our hands. The structures of 1a, 2a and 2c were conrmed by single crystal X-ray diffraction analysis (see subsequent discussion). It should be noted that products derived from both 1,2 and 1,1haloboration were observed (e.g. the B-Br analogues of 2a/2c) and this may explain the complex mixtures observed starting from 2 with lower symmetry (than 2a/2c) borylated products also observed spectroscopically (presumably containing one 1,1-and one 1,2-bromoborated boracycle in the same molecule).
In previous work 1-(arylethynyl)naphthalenes reacted with BBr 3 to give different products to that starting from 1-(alkylethynyl)naphthalenes, with the former affording products from a 1,1-bromoboration/S E Ar cyclisation process (e.g. F, inset Scheme 2, is formed instead of an analogue of E, Fig. 1 middle), 7 with no 1,2-bromoboration derived products observed. Therefore, aryl-alkynyl-pyrene derivatives (e.g. 3) were utilised in an attempt to avoid forming mixtures derived from competing 1,1 and 1,2-bromoboration. Starting from 3, the desired Tip protected product 3a was isolated in 75% yield with negligible 1,2bromoboration products observed. In addition, the borinic acid 3b was accessible in good yield using simple aqueous workup conditions. 3b proved sufficiently stable for column chromatography and even towards 1 M HCl (aq.) for at least 0.5 h. Notably, all the aryl-substituted derivatives studied herein proved more resistant to protodeboronation than the alkyl congeners, with protection by mesityl also providing sufficient stabilisation for these derivatives. The NMR data for 3a/3b are comparable to that previously reported for the B-OH and B-aryl 1-boraphenalenes. 7 Earlier studies on the bromoboration of diphenylacetylene using BBr 3 by Lappert and co-workers suggested that only the 1,2-cis-bromoboration product (4, Scheme 3 inset) forms. 8 Furthermore, alkyne 1,1-haloboration using boron electrophiles was to the best of our knowledge unprecedented prior to this work, with 1,2-haloboration or 1,1-carboboration the general outcomes. 9 Thus, in our previous report on the bromoboration of 1-(arylalkynyl)-naphthalenes, 7 a mechanism was proposed for forming F that avoided 1,1-bromoboration. According to that previously proposed mechanism, compound 3c (inset Scheme 2) should be obtained starting from 3 via cleavage of the pyrene C1-C alkyne bond. However, X-ray diffraction analysis of the haloboration products 3a and 3b conrmed that boron was attached to C10 of pyrene and that the C1-C alkyne bond remains intact. These observations indicate the reaction most likely proceeds through 1,1-bromoboration followed by electrophilic C-H borylation (Scheme 2, bottom right). Therefore, the bromoboration of other diarylalkynes, including diphenylacetylene, using BBr 3 was revisited to probe the apparent discrepancy with earlier work.
On combining diphenylacetylene with BBr 3 in heptane, crystals precipitated over the course of 5 minutes X-ray diffraction analysis on these revealed them to be the 1,2-cisbromoboration product, 4 (inset Scheme 3). NMR studies under the same conditions revealed that only one product is observed in solution, assigned as compound 4. Protodeboronation of this reaction mixture afforded 1-bromo-1,2-diphenylethene, 4-H with no 1,1-bromoboration isomer (5-H, Scheme 3) observed. While these observations are consistent with previous work, 8 upon heating the reaction mixture in heptane at 60 C an additional product, assigned as 5, was observed (by 13 C{ 1 H} and 11 B NMR spectroscopy). Protodeboronation of this reaction mixture afforded 1-bromo-2,2-diphenylethene, 5-H, derived from 1,1-haloboration as the major product along with minor amounts of 1-bromo-1,2-diphenylethene (4-H). Notably, when bromoboration was performed in DCM, both 1,2-and 1,1-bromoboration products were observed aer short reaction times at 20 C (ca. 10 minutes by NMR spectroscopy and by analysis of products post protodeboronation). These observations suggest that 1,1-bromoboration proceeds through a polar transition state(s), and potentially a vinyl cation type intermediate(s), more stabilised by the polar solvent DCM. Identical product distributions also were observed using 3 equiv. of BBr 3 in place of 1.5 equiv. of BBr 3 .
Similar outcomes were observed with di-p-tolylacetylene (6) and 1,2-bis(4-uorophenyl)ethyne with the 1,1-haloboration product being the major species observed in DCM in both cases. Exact ratios of the BBr 2 1,1 : 1,2-haloboration (both cis and trans isomers) products in DCM are challenging to determine in situ by NMR spectroscopy due to uncertainty in isomer assignment. Furthermore, ratios of BPin/vinyl C-H products post work-up and isolation are not indicative of in situ ratios due to competing pinacol protection/protodeboronation/ retrohaloboration (see ESI ‡ Section 4 for further discussion). For di-p-tolylacetylene a mixture of bromoboration products (ca. 3 : 10) was formed in DCM 10 minutes aer the addition of BBr 3 at room temperature. To determine the identity of the major isomer the fact that alkyne 1,2-bromoboration products (e.g. 4) are converted rapidly back into the starting alkyne on adding pyridine, 8 (concomitantly forming pyridine-BBr 3 ) was exploited. Upon addition of pyridine to the 3 : 10 bromoboration reaction mixture, the minor bromoboration product, assigned as 6a (Scheme 4), was converted into the starting alkyne and pyridine-BBr 3 , while the major bromoboration product, assigned as 6b, formed an adduct, 7, with pyridine at room temperature. Upon heating at 60 C, compound 7 also converted into di-p-tolylacetylene and pyridine-BBr 3 . Addition of excess BBr 3 to the mixture containing 7 to sequester all pyridine as pyridine-BBr 3 regenerated the mixture of 1,2-and 1,1-haloboration products in an identical 3 : 10 ratio to that originally observed. Combined these results indicate that 1,1-haloboration dominates in DCM, that this reaction reaches its equilibrium position rapidly in DCM, and that both 1,2-and 1,1bromoboration are reversible. DFT calculations (at the M06-2X/ 6-311G(d,p) PCM (DCM) level) revealed the 1,1-and 1,2-bromoboration products (e.g. 4 and 5) were effectively isoenergetic (all pairs of isomers had DE < 1.3 and DG < 0.8 kcal mol À1 ) consistent with the observation of mixtures on reaching equilibrium. Finally, 1,1-bromoboration is not limited to diarylalkynes, with 1-phenyl-1-propyne also leading to 1,1bromoboration products.
The preparation of fully-substituted 1-bromo-1-alkenyl boronate esters currently is challenging (the major route to these types of compounds proceeds via hydroboration of alkynylhalides thus is limited to forming tri-substituted alkenes). 10 Indeed diaryl-derivatives (such as 8a-f Scheme 5), have not been previously reported to the best of our knowledge. While 1,1bromoboration represents a simple approach to these compounds to be useful protection at boron without (or with minimal) competing formation of the starting alkyne by retrobromoboration is required. In the presence of excess triethylamine, the 1,1-bromoboration products (e.g. 5) were converted into the pinacol boronate esters, 8a-g demonstrating that ortho, meta and para substituted aryl-alkynes are amenable to this procedure. The yields are moderate at best due to the presence of the 1,2-isomer in the reaction mixture and the propensity of the Br-vinyl-BBr 2 isomers (e.g. 4 and 5) to undergo retrohaloboration in competition to pinacol protection (see ESI ‡ for more details). Indeed, under these conditions all the 1,2-bromoboration products (and the mass balance of the 1,1haloboration products) were converted into the starting alkyne and Et 3 N-BBr 3 . Attempts at pinacol installation in the absence of base led to competitive protodeboronation (presumably due to the HBr by-product from pinacol installation onto the BBr 2 moiety) and lower yields of 8x in most cases. The pinacol boronate esters were fully characterised and a number conrmed by X-ray diffraction analysis (e.g. inset Scheme 5). It should be noted that substrates containing CF 3 were not amenable to this process due to B-F/C-Br bond formation on addition of BBr 3 . Furthermore, ether cleavage occurs concomitantly to haloboration, with substrate 8f isolated from the 4-methoxyaryl precursor. In addition, attempts to bromoborate 2,2 0 -bis(2thienyl)ethyne with BBr 3 were successful to some extent, but protection with pinacol under a range of conditions was unsuccessful. This instead led to protodeboronation under a range of conditions. Finally an electronically biased diaryl alkyne was subjected to haloboration in DCM, however pinacol protection revealed that multiple haloborated isomers, e.g. 8h, had formed that could not be separated in our hands.
The elongation of C1-C2 vs. C14-C15 may indicate some delocalisation of this p-electron density into the boron centre. 11a For compound 3b, the B-O bond length (1.369(4)Å) is close to the reported values for other related borinic acids, indicating some multiple bond character between boron and oxygen. The Mes/Tip and tert-butylphenyl groups are all almost orthogonal to the boracycle, which precludes close face to face p stacking involving the boracycles in the intermolecular structure (Fig. 2).
The solid-state structures of the B 2 -doped PAHs 2a, 2c, 9a, 10b and 12a also were determined (Fig. 3, 4 and ESI ‡). All ve compounds show planar p-conjugated cores with the peripheral substituents effectively orthogonal to the PAH core. As a result, no face to face p-stacking is observed in the extended structures for these compounds (excluding 10b, see ESI ‡). For example, the face to face intermolecular separation of the cores of two parallel molecules is found to be 6.67Å for 2c. The boracycle metrics are similar to the mono-borylated compounds with C-C bond length alternation within the boracycles also observed (for example a C11-C12 vinylic distance in 12a of 1.359(7)Å and long endocyclic B-C bonds 1.534(6) and 1.543(6) A indicate minimal p delocalisation consistent with previous NICS calculations). 7 Notably, for the diborylated pyrene compounds, the length of the C-C bonds in the pyrene K-region is longer than that in the mono-borylated compound (e.g. the bond length of C7-C10 is 1.379(3)Å for compound 2c). The elongation of C7-C10 bond length may indicate a greater delocalisation of these p-electrons into the boron centres for the di-borylated compounds relative to the mono-borylated compounds.

Cyclic voltammetry
The rst reduction potentials ranged from À1.91 V (3b) to À0.96 V (12a, Table 1). Within the scan window, the B 1 -doped PAHs showed one reversible reduction wave except for compound 11a, for which two reduction waves were observed. All B 2 -PAHs showed two reversible reduction waves; however, no oxidation waves were observed within the solvent window for these compounds. The hydroxyl group on boron reduced the electron-accepting abilities of the boron-doped PAHs compared with the Tip/Mes group as previously observed. 4c Also as noted by Würthner and co-workers, 4c the boron doping pattern also signicantly inuences the redox properties with the rst reduction potential of 9a more positive than that of 10a by 0.15 V (for 9b vs. 10b DE red1 1/2 ¼ 0.32 V). Notably, the reduction potential of 9a is more positive than bis-(BMes 2 )-pyrenes (4,9 Scheme 6 The formation of extended B n -PAHs by bromoboration/ S E Ar.   isomer E red1 1/2 ¼ À2.31 V, 1,6 isomer E red1 1/2 ¼ À1.81 V) 11 and the products from the electrophilic borylation of 1,6-(2-pyridyl) 2pyrene (when the boron moiety ¼ BPh 2 E red1 1/2 ¼ À2.08 V) 14 this is ascribed to greater p delocalization in 9a as a result of planarization and three-coordinate B centres. Finally, 9a has a more positive reduction potential than its all carbon analogue (for which E red1 1/2 ¼ À2.11 V), 15 further conrming boron-doping as an effective strategy to obtain molecules with deep LUMOs due to their electron-decient nature (B 2 -PAHs are isoelectronic to dicationic carbon analogues).
Comparing pyrene and perylene core structures the reduction potentials of the borylated perylene compounds are positively shied relative to the pyrene substrates. For example, the monoborylated compound 11a exhibited the rst reduction wave at À1.43 V more positive than 3a by 0.23 V. Indeed among all the compounds investigated, 12a shows the least negative rst reduction potential with E red1 1/2 ¼ À0.96 V, which is also much more positive than the related all carbon peropyrenes. 16 To the best of our knowledge, compound 12a has the least negative reduction potential of all ambient stable three coordinate at boron B-doped PAHs reported to date, with compound D being the ambient stable B n -PAH with the next least negative reduction potential (Table 1). 4c

Photophysical properties
By varying the polycyclic core and the boron substituent, a series of B n -doped PAHs (n ¼ 1, 2) are obtained forming solutions with colours ranging from yellow to dark blue. UV-vis absorption and uorescence spectra were recorded and the optical energy gap estimated ( Table 1). All compounds showed well resolved absorption/emission proles with small Stokes shis. Consistent with the observed solution colours, the obtained spectra showed the lowest energy absorption band spanning 449 nm (3b, light yellow solution) to 614 nm (12a, dark blue solution). The doubly borylated compound 9a showed l max at 567 nm, which is red-shied by 82 nm compared with the monoborylated compound 3a (l max ¼ 491 nm). The same trend was also observed for 11a/12a, due to the extended conjugation and double boracycle annulation leading to narrower HOMO-LUMO gaps. In addition, all borinic acids exhibited hypsochromic shis of their lowest absorption band by ca. 40 nm compared with the Tip/Mes protected analogues. The photoluminescence quantum yield (F PL ) of the brominated compounds studied herein were relatively low, which might be attributed to the heavy atom effect of bromine, as related heavy atom free analogues reported by Würthner and co-workers are notably more emissive (F PL s up to 95%). 4c

Computational studies
To provide more insight into the trends observed in the experimental data, DFT calculations were performed at the M06-2X/6-311G(d,p) level with a polarisable continuum model (PCM) of DCM, calculations emulated the full compound (except for 2c, see ESI ‡). The trend in the LUMO energies from the computational results were consistent with the trends in the cyclic voltammetry data with compound 3b, that has the most negative reduction potential, being calculated to have the highest LUMO energy (À2.05 eV) among the compounds studied, while compound 12a with the least negative reduction potential was predicted to have the deepest LUMO level (À3.08 eV, see Fig. 5). It is noteworthy that the computed LUMO energy level of 12a is lower than that of a perylene diimide (PDI, red, Fig. 5) calculated at the same level of theory. Indeed comparison of the CV data for 12a and a PDI compound (containing all C-H and two N-DIP groups), conrms that 12a has the deeper LUMO, 13 while 9a has very similar frontier orbital energies to this PDI (Table 1). This further indicates that both these diborylated pyrene and perylene compounds have potential as strong acceptor moieties in organic electronics.
As shown in Fig. 5 and the ESI ‡, the LUMO and HOMO of all molecules are delocalised over the whole p-backbone and involve boron contributions to both. Interestingly, the frontier molecular orbital distributions of 12a and 12b are very similar to that of PDI. Furthermore, the frontier orbitals of all the B 2 -PAHs studied herein (and indeed the related compounds reported by Würthner et al.) 4c are related closely to the calculated HOMO and LUMO of the parent p core molecule (e.g. the HOMO and LUMO of pyrene and perylene) 17 with some additional contributions from the appended B-C unit. For completeness, the anthracene-based compound G (blue, Fig. 5) was also calculated, which revealed that the HOMO and LUMO of this compound also was related in character to the HOMO and LUMO of anthracene. Note, attempts to synthesise G failed in our hands, possibly due to the additional peri C-H group inducing steric clash with the proximal aryl group (with some structural distortion observed in the calculated structure of G). It is notable that the trend in the relative HOMO/LUMO energy and optical gap of 9b, G and 12b mirrors that of the parent PAH core, 18 with perylene (deepest LUMO/highest HOMO) < anthracene < pyrene (highest LUMO/deepest HOMO) for optical gap. This indicates that access to even lower LUMO energy B 2 -PAHs for this series maybe possible starting from lower energy LUMO core PAH structures.

Functionalisation of boron doped PAHs by cross coupling
One attraction of the compounds reported herein is the potential utility of the bromide unit, derived from the concomitant installation of B/Br in the initial step of B n -PAH formation. While cross coupling reactions can be performed using the C-B units as the nucleophilic coupling partner, 19 this removes the electronic effect of B-doping, thus utilising the C-Br unit in cross-coupling reactions offers a way to construct complex functional electronic materials that retain the C 3 B units that impart the deep LUMO character. 3 Initially compound 3a containing the sterically more demanding Tip unit, was explored and found to be amenable to Negishi cross coupling reaction conditions. Specically, in the presence of Pd(PPh 3 ) 4 , 3a reacted with the in situ generated aryl-zinc reagent affording the desired product 13 (Scheme 7) in moderate 33% isolated yield. We attribute the modest yield to the highly sterically encumbered vinyl-Br position present in these B-PAHs, thus the B-Mes derivatives were explored.
The same methodology could be applied to B-Mes derivatives to access a D-A-D molecule with a narrow HOMO-LUMO energy gap. Using the procedure established for 13, 9a was coupled with the organozinc compound derived in situ from 4bromotriphenylamine to construct compound 14 (Fig. 6). The desired double cross coupled product, 14, was isolated by column chromatography in 21% yield. Cyclic voltammetry measurements on compound 14 showed one oxidation wave within the solvent window (peak-current E ox ¼ 0.49 V vs. Fc/Fc + ) and two reduction processes (E red1 1/2 ¼ À1.23 V, E red2 1/2 ¼ À1.62 V). The rst reduction process is shied cathodically by 0.2 V relative to 9a, presumably due to the replacement of inductively electron-withdrawing bromine for the triphenylamine groups. DFT calculations (Fig. 6, bottom) were performed on compound 14 at a lower theory level due to its size (M06-2x/6-31G(d)), 9a was calculated at the same level for direct comparison (which conrmed effectively identical HOMO/LUMO distributions for 9a using the two basis sets). Consistent with a D-A molecule the HOMO/LUMO are spatially separated, with the HOMO being predominantly localised on the triaryl amine unit thus being 0.91 eV higher in energy than the HOMO of 9a. The LUMO in 14 is localised on the B 2 -PAH acceptor unit and is closely related in character to the LUMO of 9a. Notably, the calculated energy for the LUMO of 14 is 0.19 eV higher than that for 9a in excellent agreement with the CV data; again this is attributed to the exchange of inductively withdrawing Br groups for triarylamine units.
The photophysical properties of 14 in solution, solid-state and dispersed in a PMMA lm were investigated. Compound 14 displayed a broad absorption band stretching up to 750 nm in dichloromethane, with the onset corresponding to an optical gap of 1.65 eV (comparable to DE redox determined by electrochemistry). Thus, the strong acceptor character of the B 2 -PAH unit combined with the triphenylamine donor units results in a small HOMO-LUMO energy gap. We employed timedependent DFT (TDDFT) calculations using the Tamm-Dancoff approximation at the PBE0/6-31G(d,p) level of theory to assign the spectral features observed in the absorption spectrum of 14. The lowest energy absorption features found within the tail of the band showing a maximum at 530 nm and a shoulder at 600 nm are assigned to be charge transfer (CT) states from the Ph 2 NPh group to the B 2 -PAH core (S 0 / S 1 , S 0 / S 2 ). The principal distinguishable absorption features at 444 nm originate from mixed CT/LE (LE ¼ locally excited) transitions, but this time the CT contribution implicates the mesityl groups as donors to the B 2 -PAH core (S 0 / S 3 , S 0 / S 4 , see ESI ‡ for more details); the LE contribution is localized on the B 2 -PAH acceptor. Higher energy bands involve increasing contributions of LE transitions of the B 2 -PAH core. Recognizing the propensity of large acenes to aggregate in solution, even at very dilute concentrations we next investigated the effect of concentration on both the absorption and emission spectra. We note that both the 3 values and the spectral features in the absorption spectra were independent of concentration, precluding aggregation of the molecule in its ground state in DCM at the concentrations studied (see ESI ‡). In contrast, the emission in DCM was highly concentration dependent. At 0.15 mmol mL À1 the emission is broad and unstructured and centred at 479 nm. As the concentration is increased to 63 mmol mL À1 , there is a decrease in intensity of this band coupled with a progressive and signicant red-shi of the emission. At the highest concentrations studied, the emission narrows but remains unstructured (Fig. 7). We attribute this complex behaviour to the formation of mixtures of different excimers as the concentration increases, presumably with higher order aggregates forming at the higher concentrations.
The F PL value in DCM for 14 at 0.52 mmol mL À1 is 4% and this value is not sensitive to oxygen. The excited state emission decays measured at both dilute (0.52 mmol mL À1 ) and concentrated (33 mmol mL À1 ) solutions are multi-exponential in nature but remain in the nanosecond regime, consistent with a uorescence mechanism. In order to mitigate these non-radiative pathways, we next investigated the solid-state photophysical properties as thin lms in PMMA at 0.1, 1 and 10 wt% doping concentrations, and as neat lms. Generally, the same evolution in emission proles is observed in the solid-state as was observed in DCM. As a 0.1 wt% PMMA doped lm two unstructured emission bands are observed, one of higher relative intensity at high energy at l PL ¼ 442 nm that we ascribe to the same monomer emission observed in dilute DCM, and a lower intensity and broader low energy band at l PL ¼ 638 nm Scheme 7 Negishi cross coupling to form compound 13. that we ascribe to excimeric emission. The F PL of this lm is 5%, which is essentially identical to that measured in DCM. This suggests that during the spin-coating process aggregation remains strong even at this low doping concentration. As a 1 wt% PMMA doped lm, two unstructured emission bands remain, one at l PL ¼ 472 nm and one at l PL ¼ 726 nm; here, we note the large red-shi and signicant enhancement in intensity of the low-energy band. The F PL of this lm is slightly lower at 3%. The 10 wt% doped PMMA lm also shows two emission bands where the relative intensity of the high energy band at l PL ¼ 434 nm is reduced further compared to the low energy band at l PL ¼ 793 nm. The F PL of this lm is <1%. The neat lm shows only a single very weak and red-shied emission band at l PL ¼ 824 nm with a F PL < 1%. There is a systematic shortening of the s PL with increasing doping concentrations that mirrors the decrease in F PL (see ESI ‡). Irrespective of the complexity of the emission spectra the successful formation of the low optical gap material 14 demonstrates the utility of these dibrominated B 2 -PAHs in accessing complex organic materials.

Conclusions
In conclusion, sequential bromoboration/intramolecular electrophilic C-H borylation enables formation of a range of brominated B n -doped PAHs in useful yields via an operationally simple, one-pot route from readily available precursors (alkynes and BBr 3 ). Cyclic voltammetry and photophysical studies revealed that these molecules have very low LUMO energies thus are attractive acceptor units for use in organic electronic applications. In particular, compound 12a has the least negative reduction potential among all reported ambient stable B-doped PAHs to the best of our knowledge. The C-Br units in the B n -PAHs can be utilised directly in Negishi cross-coupling reactions, which enabled formation of a donor-acceptor-donor molecule displaying solution absorption up to 750 nm. Finally, mechanistic studies on the bromoboration reaction indicated it proceeds through the 1,1-bromoboration of the diarylalkynes, a reaction not previously observed using just boron electrophiles. 1,1-Bromoboration can be applied to functionalise other diarylalkynes, enabling access to unprecedented 1-bromo-2,2diaryl substituted vinylboronate esters direct from internal alkynes. The utility of the 1,1-bromoboration reaction is being studied further in our laboratory as are the brominated-B 2 -PAHs, particularly to access useful functional materials.

Conflicts of interest
There are no conicts to declare.