Deconstructive Cycloaromatization Strategy towards N,O-Bidentate Ligands and Their Four- Coordinate Organoboron Complexes


 The innovative construction of novel N,O-bidentate ligands and N,O π-conjugated four-coordinate organoboron complexes represent a long-standing challenge for chemists. Here, we report an unprecedented and straightforward approach for the construction of N,O-bidentate ligands and their organoboron complexes via the merge of ring deconstruction with cycloaromatization of indolizines and cyclopropenones. Without any catalysts, our method is able to deliver a series of polyaryl 2-(pyridin-2-yl)phenol ligands, N,O π-conjugated organoboron (BF2 and BAr2) complexes with good functional-group compatibility which are difficult or even impossible to synthesize with previous methods. Importantly, the formed N,O-bidentate ligands were easy to scale up and derive with valuable drugs and active molecules. In addition, the photoluminescence measurements and the HOMO/LUMO gap have been investigated, the results have revealed that N,O π-conjugated tetracoordinate boron complexes display bright fluorescence, large Stokes shifts, and good quantum yields (Φlum = 0.15–0.45). The method proposed by the paper will inspire the development of various N,O-bidentate metal and boron complexes, which is expected to move the area of catalysis chemistry and material science forward.


Results And Discussion
Condition optimization.
We have long been committed to developing new methods to construct nitrogen-containing heterocyclic derivatives. [69][70][71][72][73] Thus, easily available 2-phenylindolizine and diphenylcyclopropenone were chose as the model substrates to conduct the reaction under 120 o C in DCM for 18 h (Table 1). Surprisingly, the sequence deconstructive recycloaromatization product was secured with a prominent yield (95%) without any catalysts and additives ( Performing the reaction at nitrogen atmosphere affected the reaction e ciency slightly and afforded the polyaryl 2-(pyridin-2-yl)phenol in 88% yield (Table 1, entry 4). Finally, the radical inhibitor TEMPO and BHT were added to the reaction, respectively, the hardly changed yields indicated that a free radical process was not involved in the transformation.

Synthesis of polyaryl N,O-bidentate ligands.
With the optimal conditions in hands, we began to evaluate the substrate adaptability of this deconstructive-cycloaromatization reaction, as summarized in Scheme 2. Firstly, various substituted indolizines were applied to get the target products under the optimized conditions. The 2-phenylindolizine with substituents at its pyridine ring, such as methyl, methoxy, ethyl, and ester, could give the desired N,Obidentate ligands 3b-3e in 68%-93% yields. To our delight, products 3f and 3g were also obtained with high yields when alkyls replaced aryl group at C-2 position of indolizines. Both electron-donating and electron-withdrawing groups at para-position of benzene were well tolerated to deliver the corresponding N,O-bidentate ligands 3h-3r with moderate to high yields, in which product 3r was detected by singlecrystal X-ray diffraction to gure out its concrete structure (CDCC 2101314). Similarly, when the metaand ortho-positions of phenyl ring were equipped with various substituents, regardless of whether monoor di-substituted, products 3s-3aa were all achieved in satisfactory yields (70%-91%). And the reaction was not severely affected by the service of 2-(4-uorophenyl)-7-methylindolizine and 2-(4-chlorophenyl)-8methylindolizine (3ab, 85%; 3ac, 86%). It was found that thienyl-, furyl-, and pyridyl-substituted indolizines was effectively leading to 70-81% yields of 3ad-3af. Then several cyclopropenones were entrusted with the mission to complete this deconstructive-cycloaromatization reaction. In comparison to the model substrate, cyclopropenones with methyl and methoxyl group on phenyl ring also offered the ideal products 3ag-3ai in nice yields. When it came to dithienylcyclopropenone, the reaction result made us amused with 82% yield of 3aj. It was worth mentioning that dithienylcyclopropenone and 2-(thiophen-2yl)indolizine proved to be successful substrates for the procedure of deconstructive-cycloaromatization under metal-and oxidant-free conditions, affording N,O-bidentate ligand 3ak in 60% yield. Similarly, different cyclopropenones were proved to be reliable candidates, furnishing the BF 2 complexes 4g and 4h in moderate yields. To our delight, when BPh 3 participated in this reaction with model reactants, the targeted N,O-bidentate organic BPh 2 complexes 5a were achieved in nice yields without any other catalysts and additives. Also, the indolizines bearing various groups, such as different-substituted phenyl, furyl, and pyridyl, could also provide the corresponding BPh 2 complexes 5b-5f in good yields (65%-82%). More importantly, we found that a variety of boric acid can also directly participate in the reaction as a boron source, leading to four-coordinate organoboron complexes by one-pot method. The results proved that different substituted phenylboronic acids were tolerable in the presence of K 3 PO 4 and the expected diarylboron complexes (5g-5m) were afforded in good yields (77%−90%). Besides phenyl, other arylboronic acids, such as thienyl, furyl, naphthyl, and carbazolyl boronic acid, could enable to forge target products 5n-5q in 77%-86% yields. The crystal structure of 5q was analyzed (CDCC 2121318). Next, we introduced triphenylamine groups to indolzine and used it to get the corresponding BPh 2 -complexs 5r and 5s successfully with good yields. Finally, bisindolizine derivative prepared via Suzuki-coupling of iodoindolizine and 1,4-phenylenebisboronic acid nished the deconstructive-cycloaromatization and provided product 5t in 84% yield. Notably, compound 5t contains two four-coordinate organoboron centers.

Synthesis of polyaryl phenolic esters.
Then, we turned our attention to put acetic acid into the reaction edi ce, the assumed polyaryl 2-(pyridin-2-yl)phenol derivatives were produced in one-pot (Scheme 4). For example, when acetic acid was used, essential 2-(pyridin-2-yl)phenol derivatives 6a-6d were got in 60%-75% yields with variant indolizines, and the crystal structure data of 6c has arrived in our hand (CDCC 2101315). This novel deconstructive aromatization strategy provides a convenient, direct and practical access to synthesize complex N,Obidentate ligand and their derivatives in one-pot. metal-and oxidant-free conditions.
As we know, phenol derivatives are not only indispensable components of many drugs, but also good synthetic intermediates. Therefore, the polyaryl 2-(pyridin-2-yl)phenols synthesized by the present method were easily incorporated into several drug molecules and bioactive units via esteri cation process under mild conditions (Scheme 4). We rstly employed a simple pyrazolecarboxylic acid, which is an intermediate of YW2065 (anti-CRC molecule), and fortunately, the desired esteri cation product 6e was yielded in 82% yield. Then two approved drugs indometacin and febuxostat with complicated structure could combine with polyaryl 2-(pyridin-2-yl)phenol, generating the corresponding derivatives 6f and 6g in high yields, respectively. Herein, we developed a rapid method for the construction of complex framework polyaryl 2-(pyridin-2-yl)phenols and successfully applied it to the modi cation of several pharmaceuticals and biologically active molecules. In order to further evaluate the practicability and expansibility of this deconstructive aromatization reaction, we expanded the scale of this metal and additive-free transformation to 8 mmol (Scheme 5).
Surprisingly, we easily and e ciently acquired the nal product 3a in 81% yield (2.6 g) by direct lter and wash with MeOH. The further transformations were studied. When halogenating agents, Nuorobenzenesulfonimide (NFSI), 1,3-dichloro-5,5-dimethylhydantoin (DCDMH), and N-bromosuccinimide (NBS), were stirred with N,O-bidentate ligand at 80 o C respectively, the corresponding halogenated products 7a-7c were given with 78%-96% yields, which could be puri ed readily by recrystallization (hexane and DCM). The obtained halogenated N,O-bidentate ligands were easily ordinated with arylboronic acids to produce the corresponding four-coordinate organoboron complexes (8a-8c). Next, three different boric acids were coupled with brominated N,O-bidentate ligand, generating 9a-9c with good yields. Reaction could also occur between N,O-bidentate ligand and per uorotriphenylborane to cater for product 10 in 90% yield.

Photophysical property investigations
To prove the promising applications as optical material of the four-coordinate organoboron complexes, we next investigated the photophysical properties of N,O-bidentate BPh 2 complexs prepared by this method (Scheme 6 and Table 2). Clearly, the selected four-coordinate organoboron complexes (5a, 5i, 5q, 5t, and 8a) display bright blue uorescence in DCM, while compounds 9b and 5s display light green uorescence and 5r displays intensive yellow uorescence under UV light irradiation (365 nm) (Scheme 6a). The detailed spectra datas were shown in Scheme 6. Absorption spectra of 5a in different solvent shows that DCM is the best solvent for these N,O-bidentate BPh 2 complexes (Scheme 6b). The absorption maximum wavelengths of the selected comlexes range from 311 nm to 347 nm in DCM solution (Scheme 6c). In uorescence emission spectra, their maximum photoluminescence wavelength are concentrated at 473 nm to 570 nm (Scheme 6d). Different substituent groups weakly affect on emission wavelength and uorescence intensity, while compounds 9b, 5s, and 5r have remarkable red shifts. It is worthy to note that compound 5r displays the emission maxima at 570 nm (Scheme 6d). Interestingly, these organoboron complexs exhibits large Stokes shifts ranging from 139 nm to 223 nm, and such large Stokes shifts makes them appealing in valuable applications (Table 2). Finally, density functional theory (DFT) calculations were conducted at the B3LYP/6-31G(d) level using the Gaussian 03 package for further study the optoelectronic properties. [74] The calculated HOMO and LUMO energy levels of selected compounds (5a, 5r, 5s, 5t, and 9b) were summarized at Scheme 6e and Table 2. The results disclosed that the HOMO−LUMO gaps of these four-coordinate organoboron complexes are in the range from 2.99 to 3.70 eV.