Enantioenriched Boron C , N -Chelates via Chirality Transfer

: Molecules stereogenic only at tetrahedral boron atoms show great promise for applications, for example as chiroptical materials, but are scarcely investigated due to their synthetic challenge. Hence, this study reports a two-step synthesis of enantioenriched boron C , N -chelates. First, the diastereoselective complexation of alkyl/aryl borinates with chiral aminoalcohols furnished boron stereogenic hetero-cycles in up to 86% yield and d.r. > 98:2. Treatment of these O,N -complexes with chelate nucleophiles was surmised to transfer the stereoinformation via the ate-complex into the C , N -products. This chirality transfer succeeded by substitution of the O,N -chelates with lithiated phenyl pyridine to give boron stereogenic C , N -chelates in up to 84% yield and e.r. up to 97:3. The chiral aminoalcohol ligands could be recovered after isolation of the C,N -chelates. The chirality transfer tolerated alkyl, alkynyl and (hetero-)aryl moieties at boron and could be further extended by post-modification: transformations such as catalytic hydrogenations or sequential deprotonation/electrophilic trapping were feasible while maintaining the stereochemical integrity of the C,N -chelates. Structural aspects of the boron chelates were studied by variable temperature NMR and X-ray diffraction.

Most challenging is certainly the synthesis of compounds stereogenic only at boron and to the best of our knowledge only a handful of examples have been reported. [9,10,33,37]Among these, an efficient approach is the catalytic enantioselective desymmetrization via click-reaction or carbene insertion furnishing N,N-and C,N-chelates 8 and 10 (Scheme 1e). [9,10]Both methodologies use a Cu/BOX-ligand system and provide the enantioenriched boron-complexes 8 and 10 in high yields (up to 95 %) and enantioselectivities (up to 99 % ee).Notably, the N,N-chelates 8 exhibit circularly polarized luminescence (CPL) [9] useful for sensing applications. [38]Yet, these desymmetrization protocols are limited to triazole/alkyne or secondary carbons/ hydrogen peripheral groups attached to boron, although the post-functionalization of these groups is feasible.A more recent protocol in terms of the peripheral substituents is a catalytic enantioselective CÀ H arylation giving rise to boron-stereogenic BODIPYs which have then been applied in chiral recognition. [39]hese peripheral substituents influence the configurational stability of the chelate systems [25] and can even tune the optical properties as illustrated for phenylpyridine chelates (Scheme 2).The phenylpyridine unit is a prominent ligand system for boron-complexes due to its synthetic, [40,41] photoisomerization [20,42,43] and luminescent properties. [15,16,28,44]or phenylpyridine chelate dfppy (Scheme 2a), attaching triphenylamine moieties instead of phenyl groups to the boron atom lead to thermally activated delayed fluorescence (TADF), a property usually not observed for this type of C,N-chelates dfppy. [44]Introduction of this triphenylamine unit to a boronstereogenic chelate might establish another way to CPL-active TADF materials. [45,46]Moreover, the photoisomerization of phenylpyridine-coordinated borons such as in 1 and related C,N-chelates has been extensively studied (Scheme 2 b). [20,42,43]ere, the peripheral groups play a crucial role for the photoreactivity and the tailor-made introduction of these groups to enantioenriched boron chelates would expand the potential of the photoisomerization even further.In order to control the reactivity pattern and optical properties of boron C,N-chelates, the variation of the peripheral substituents is highly desirable.Consequently, we devised our own strategy towards enantioenriched C,N-chelates 18 where boron is the single stereogenic center with a special emphasis on the peripheral groups.
[49] We wondered if the versatility of the ate-complex could be expanded to the preparation of enantioenriched boron complexes 15.A hypothetical boron chelate 11 carrying two substituents of different size (R S , R L ) as well as a chiral O,D-Scheme 1. Known chiral tetrahedral boron compounds and their preparation strategies a) racemic mixtures of C,N-chelate 1 [20] and NHC-borane 2, [21] b) enantiopure boron complexes 3 [25] and 4 [28] by resolution of their racemic mixtures via chiral HPLC, c) enantioenriched boron complex 5 [8] via diastereoselective synthesis, d) enantioenriched boron complex 6 [37] via intramolecular chirality transfer, e) boron N,N-and C,N-chelates 8 [9] and 10 [10] via catalytic enantioselective desymmetrization.alkoxide-ligand would serve as a starting point.The chiral backbone of 11 should control the boron stereogenic center by steric factors (i.e. diastereoselective synthesis).Although this concept is seemingly trivial, it is, to the best of our knowledge, rarely investigated for bidentate boron chelates. [8,50,51]Even more, chiral alkoxide chelate ligands such as aminoalcohols have been so far only coordinated to symmetrical boron units (e. g.[54][55] Our devised strategy would then commence by addition of a chelating organometallic nucleophile 12 to O,D-chelate 11 to form ate-complex 13.[58] Hence, the chiral information of starting chelate 11 would be maintained in the boron-stereogenic ate-complex 13.If the alkoxide 14 would dissociate as a leaving group (LG) parallel to coordination of the donor atom D of the carbon nucleophile 12, the stereoinformation should be transferred to the final C,Dchelate 15.This proposed chirality transfer via the boron atecomplex 13 is, to the best of our knowledge, unprecedented and would give rise to enantioenriched chelates 15 stereogenic only at boron.
To realize this concept, two key challenges needed to be addressed, namely the diastereoselective complexation and subsequently the chirality transfer (Scheme 3b).The first task was to find a suitable ligand for the initial diastereoselective complexation.In this case, an alkyl (R 2 )/aryl (R 1 ) substituted boron alkoxide should be coordinated to readily available chiral aminoalcohol ligands and the diastereoselectivity of these O,Nchelates 16 should be evaluated.For the second task, the chirality transfer from O,N-chelates 16 to enantioenriched C,Nchelates 18, we chose the phenylpyridine unit 17 as a benchmark nucleophile due to its versatile applications discussed above.The envisaged chirality transfer via the ate-complex, its synthetic feasibility and the enantioselectivity of the C,Nchelates 18 are unexplored so far.In the current manuscript we disclose the realization of this novel approach towards boronstereogenic chelates.

Diastereoselective borinate complexation with chiral aminoalcohol ligands
The synthesis of aminoalcohol chelates 16 started with treatment of diisopropylphenyl boronate 20 a with methylmagnesium bromide in Et 2 O to generate an intermediate borinate PhMeBOiPr (Scheme 4a), followed by in situ complexation with the respective aminoalcohol 19 in Et 2 O/EtOH to provide the desired O,N-chelates 16 in > 90 % purity.The presence of a tetrahedral BÀ N boron was confirmed by 11 B NMR (δ(16 a-l) = 7.0-11.1 ppm) [52] and the relative stereochemistry was elucidated using NOESY (for details see Supporting Information, chapter 1.5).

Screening of aminoalcohol complexes in the chirality transfer reaction
The obtained set of O,N-chelates 16 a-l was then treated with the lithiated phenylpyridine 21 to furnish the C,N-chelate 18 a and the results were evaluated regarding 1 H NMR yields and enantiomeric ratios (Scheme 5).The reaction conditions obtained after initial optimization (SI, Table S1) consisted of lithiation of 2-(2-bromophenyl)pyridine 21 with t-BuLi (2.0 equiv.) in THF, addition of the O,N-chelate 16 at À 90 °C and warming to À 20 °C.Chelates 16 a,c,d,j,k with free NH functionalities were supposed to be incompatible with the strongly basic aryllithium and together with quinine-compound 16 l only gave poor results (< 10 % yield).Thus, these chelates are not discussed further with regards to the chirality transfer.
For NH 2 -valinol complex 16 b the desired C,N-chelate 18 a was formed in 13 % yield and e.r.88 : 12. Albeit suffering from a low yield, the concept of chirality transfer could be successfully demonstrated: starting from a diastereomeric ratio > 98 : 2 the stereochemical information at boron was preserved, yet with a lower enantioselectivity (e.r.88 : 12).An improved yield of 42-62 % could be obtained for the monosubstituted dimethylaminoalcohol ligands 16 e,f,g.Unfortunately, the chirality transfer was less efficient for these monosubstituted NMe 2 -ligands 16 e,f,g with e.r.'s ranging from virtually racemic (16 e, e.r.Here, the same enantioselectivity as for 16 b was observed and the enantiodivergent chirality transfer of 16 b,h,i and 16 e-g is part of ongoing investigation.An enantioenriched analytical sample of 18 a (e.r.88 : 12) could be further recrystallized from toluene to e.r.> 99.5 : 0.5.Single crystals from this sample investigated by X-ray diffraction revealed the major enantiomer of 18 a to be (S)-configured at boron (for details see structural study below).

Substrate scope of pseudoephedrine borinates with alkyl/aryl substitution
After proof of concept for the chirality transfer, we investigated the scope and limitations of the developed two-step sequence.First, O,N-chelates 16 based on the pseudoephedrine-ligand 19 a with peripheral boron-substituents (R 1 and R 2 ) other than methyl or phenyl were examined.By using the previously established Grignard method (see Scheme 3, 16 i) for the intermediate borinate generation (PhEtBOiPr), the desired ethylsubstituted boron heterocycle 16 ia was formed in 75 % yield and d.r.93 : 7 (Scheme 6, method A).
However, the use of other Grignard reagents than MeMgBr or EtMgBr resulted in complex product mixtures und consequently, another protocol now employing organolithium reagents was developed.The respective lithium reagents R-Li were commercially available, or prepared by lithium-halogen exchange or deprotonation of alkynes (Scheme 7, method B).Phenyl 20 a or TMS-alkynyl boronates 20 b were then added to generate the intermediate ate-complexes, and the borinates were liberated by quenching with acetyl chloride. [60,61]The following in situ complexation with aminoalcohol 19 a proceeded smoothly in THF/EtOH mixtures.By this method B, butyl and ethynyl substituents were installed starting from diisopropylphenyl boronate 20 a in yields of 72 % (16 ib) and 37 % ( 16 ic Gratifyingly, the devised borinate complexation also allowed a one-pot preparation of the O,N-chelates 16 (Scheme 8, method C).For this, an aryllithium was mixed with B(OiPr) 3 and treated with trimethylsilyl trifluoromethane sulfonate (TMSOTf) [62] to furnish a boronate R 1 B(OiPr) 2 .
Subsequent addition of methyllithium followed by decomposition of the ate-complex with acetyl chloride furnished the intermediate borinates R 1 (Me)BOiPr which were then coordinated in situ to ligand 19 a.The protocol allowed the introduction of alkylated (16 ik), fluorinated (16 il), trifluoromethylated (16 im) and π-extended or electron rich aryl moieties ( 16 in,o).The yields ranging from 62-86 % were reasonably high for a one-pot assembly and moreover, the diastereoselectivities were satisfying throughout (d.r.92 : 8-94 : 6).The obtained set of aminoalcohol borinates was now to be probed in the substitution towards C,N-chelates.

Scope of chirality transfer to enantioenriched boron C,Nchelates
Exploring the devised chirality transfer required further synthetic optimization (for details see Supporting Information, Table S2) which included the lithiation of 2-(2bromophenyl)pyridine 21 at À 90 °C for 1 h instead warming up to 0 °C (Scheme 9).This finally enabled isolation of methyl/ phenyl substituted 18 a in 78 %.For 18 a, the crude e.r. of 87 : 13 (see Supporting Information, chapter 1.2) was improved by recrystallization to e.r.91 : 9. Given the diastereomeric ratio of  the O,N-chelated precursor 16 i (d.r.92 : 8), the chiral information at the boron atom was transferred to the C,N-chelated 18 a (e.r.91 : 9, crude e.r.87 : 13) in a sufficient manner.It should be noted that the chiral aminoalcohol 19 a could be recovered by simple column chromatography (details see Supporting Information, chapter 1.2).
However, this conservation of stereochemical information did not succeed for substituents larger than a methyl group: ethyl-and vinyl-C,N-chelates 18 b,c could be isolated in 75 % (18 b) and 37 % (18 c) yield but gave virtually racemic mixtures (e.r.54 : 46) in both cases.Thus, the chirality transfer seems quite sensitive to the size of the smaller peripheral substituent (R 2 ).On the other hand, by retaining the methyl-substituent and varying the aryl group we did not encounter such problems.For example, m-tolyl-and p-F-phenyl-substituted C,N-chelates 18 d,e could be isolated in 66 % (18 d) and 57 % (18 e) yield and enantioselectivities of e.r.88 : 12 and 97 : 3. Additionally, a bistrifluoromethylated aryl group 18 f was tolerated in 38 % yield and e.r.88 : 12.Moreover, we were pleased to see that naphtyl-functionalized C,N-chelate 18 g or triphenylamine-equipped 18 h with bulky aromatic moieties could be obtained in high yields of 82 % (18 g) and 80 % (18 h) and selectivities (e.r.92 : 8 (18 g), 85 : 15 (18 h)).Replacing the methyl-group with a TMS-protected alkyne provided the C,Nchelate 18 i in 64 % and e.r.79 : 21.Recrystallization improved the e.r. from 79 : 21 to 91 : 9 (37 % yield).Other functionalized alkynes equipped with either a phenyl or hexyl group (18 j,k) could be employed as well with yields of 67 % and 81 %.The e.r.'s of the TMS-(18 i, e.r.79 : 21 prior to recrystallization) and Ph-capped (18 j, e.r.84 : 16) examples were slightly lower than the hexyl-functionalized compound 18 k (e.r.89 : 11).Interestingly, pairing the TMS-alkyne with a p-MeO-functionalized phenyl ring (18 l) improved the yield to 74 % and selectivity to e.r.88 : 12. Finally, triphenylamine-functionalized 18 n was isolated in 61 % but in contrast to methyl-chelate 18 h, the determination of the e.r. was not possible by our repertoire of chiral HPLC or NMR shift reagents.However, TMS-deprotection of 18 n to alkyne 18 p enabled chiral HPLC separation and indicated the e.r. of 18 n to be � 81 : 19 (18 p, Scheme 10).The major limitation for aromatic groups represented the benzofur-ane-complex 18 m which could be isolated only in 20 % yield and e.r.58 : 42.
The moderate enantiomeric ratios with larger groups than methyl or alkynyl could be tackled by post-functionalization of the obtained C,N-chelates (Scheme 10).First, the TBAF-mediated deprotection of TMS-alkynes 18 i and 18 n proceeded smoothly to give 18 o and 18 p in 92-95 % and e.r.90 : 10 (18 o) and 81 : 19 (18 p).Hydrogenation of 18 o with Pd/C provided ethyl-derivative 18 b in 91 % yield.Gratifyingly, the enantiomeric ratio of the starting TMS-alkyne 18 i (e.r.91 : 9) was completely retained over two steps (18 b e.r.91 : 9).Attempted hydroboration of alkyne 18 o with HBpin revealed a particular reactivity in form of concomitant deborylation which furnished the unsubstituted alkene 18 c in 58 % yield (94 % brsm) and e.r.90 : 10 (for details of deborylation see Supporting Information, chapter 1.3).Finally, alkyne 18 o survived deprotonation with n-BuLi and electrophilic trapping gave rise to ester 18 q in 54 % and e.r.92 : 8.It should be emphasized, that neither hydroboration nor deprotonation/electrophilic trapping affected the boron stereocenter's integrity.dynamics has been described for comparable NMe 2 -B-dative bonds and the barriers reported were in a range of 48-76 kJ mol -1 . [52,63,64]The acyclic isomer 23 could not be observed according to 11 B NMR with a single signal at δ(16 f) = 11.1 ppm.

Structural study of boron O,N-and C,N-chelates by NMR and X-ray diffraction
Cooling the sample to À 80 °C resulted in splitting of the NMe 2 -signal together with appearance of a second signal set in a ~55 : 45 ratio ( 1 H NMR, Figure 1b).The splitting of the NMe 2signal is explained by freezing the BÀ N bond dynamics whereas the second signal set originates from another dynamic process.
Unfortunately, the overlapping signals of both processes prevented the exact barrier determination of the B-N bond strength by VT NMR (see Supporting Information, Figure S7).The second process is supposedly a conformational barrier resulting from a densely substituted oxaborolidine and such a conformational freeze of related six-membered boron-O,Nchelates has been reported. [53]The hypothesis of conformers was further supported as the iPr-substituted NH 2 -chelate 16 b with a stronger BÀ N bond [52] (and hence, a slower BÀ N bond dynamics) displayed similar doubling of signals at À 80 °C (Figure S1, Supporting Information).On the other hand, the BÀ N bond for the enantioenriched C,N-chelate 18 a proved to be far more stable as illustrated by a racemization experiment in heptane at 95 °C (Figure 1 c, Supporting Information chapter 1.6).The racemization barrier ΔG ¼ 6 rac of 18 a was 125 kJ mol À 1 and thus is higher than in comparable NMe 2 -donor-complexes such as 3 (94-116 kJ mol À 1 ) [25] but lower than respective N,Nchelates such as 8 (137-145 kJ mol À 1 ). [9]n addition to solution studies, the solid-state structures of O,N-chelates 16 e, 16 i and of C,N-chelate 18 a were obtained from single crystal X-ray diffraction (Figure 2).For more details see Supporting Information, chapter 5.The single crystals were grown by slow evaporation of heptane/CH 2 Cl 2 (16 e), Et 2 O (16 i) or toluene (18 a) solutions.These solid-state structures revealed the absolute stereochemistry of the respective compounds.
The monosubstituted O,N-chelate 16 e displayed a BÀ N bond length of 1.689(3) Å similar to the disubstituted 16 i (1.683(3) Å).The BÀ N length of 16 i is shorter than its corresponding BPh 2 -derivative (1.74(1) Å), [52] likely due to steric reason (Figure 2 left).On the other hand, C,N-chelate 18 a has a comparable BÀ N-distance of 1.620(2) Å as its BPh 2 -and B(biphenyl) pendants (1.618(3) Å and 1.617(5) Å). [41,43] The tetrahedral character (THC) according to Höpfl [65] is derived from all bond angles around the boron atom and, as a measure of the boron sp 3    This set of O,N-chelates 16 was subjected to the chirality transfer reaction and gave rise to the C,N-phenylpyridinechelates 18 a-o in an enantioenriched fashion.A variety of peripheral groups such as methyl and alkynyl on the one side, and bulky and heteroatom-substituted aryl moieties on the other side could be introduced in yields up to 84 % and up to e.r.97 : 3. Notably, the aminoalcohol 19 a used for chirality transfer could be recovered by simple column chromatography.The post-functionalization enabled alkyne derivatization and introduction of larger alkyl and alkenyl substituents without affecting the boron stereocenter.Gratifyingly, our devised concept of chirality transfer via the ate-complex for enantiocontrol at boron atoms succeeded and consequently expands the space of this synthetic handle from so far carbon-based stereoselective transformations.Further, the possibility to address the peripheral substitution complements the few hitherto known catalytic enantioselective approaches to B-stereogenic chelates, which predominantly focused on the boron chelate ligand.The tolerance of the presented chirality transfer regarding other types of chelate ligands and insights into the origin of stereoselectivity will be reported in due course.

Scheme 2 .
Scheme 2. Influence of peripheral boron groups on the emission (a) and reactivity (b) of phenylpyridine chelates dfppy and 1.
16 k, d.r.83 : 17 and 16 l, d.r.88 : 12) compared to mono-/disubstituted examples 16 a-j (up to d.r.> 98 : 2), indicating that additional bulk does not improve the diastereoselectivity.The major factor governing the boron stereo-chemistry of the O,N-chelates 16 is steric hindrance: the B-phenyl groups and the (more bulky) chiral substituent of the O,N-ligand 19 adapt a trans-relation on the oxazaborolidines 16.

Scheme 5 .Scheme 6 .
Scheme 5. Screening of O,N-chelates 16 in the chirality transfer reaction to C,N-chelate 18 a.Reactions were performed on 0.2 mmol scale.Results were examined by 1 H NMR yield using 1,3,5-trimethoxybenzene as internal standard and enantiomeric ratios (e.r.) were determined by chiral HPLC.

Scheme 9 .
Scheme 9. Scope of C,N-chelates 18 a-n prepared by the chirality transfer reaction from O,N-complexes 16 ia,id-io (including respective d.r.).Enantiomeric ratios (e.r.) were determined by chiral HPLC.[a] Enantiomeric ratio of the crude product was e.r.87 : 13. [b] e.r. could not be determined by our means of chiral HPLC or NMR shift reagents.Instead, the e.r. of postfunctionalized 18 p could be determined (e.r.81 : 19).
Scheme 10.TMS-deprotection of 18 i and 18 n and functionalization of alkyne 18 o to access C,N-chelate 18 b, 18 c and 18 q.
-hybridization, was 74 % for O,N-16 e, 76 % for O,N-16 i and 66 % for C,N-18 a. Conclusion The enantioselective synthesis of boron C,N-chelates 18 stereogenic only at the B-atom was realized by chirality transfer from O,N-to C,N-chelates.Initial borinate complexation with chiral aminoalcohols 19 provided O,N-chelates 16 in high diastereose-

Figure 1 .
Figure 1.a) BÀ N bond dynamics of aminoalcohol complex 16 f at room temperature.b) 1 H NMR spectra of 16 f in CD 2 Cl 2 (500 MHz) at 25 °C and À 80 °C, c) racemization of C,N-chelate 18 a in heptane at 95 °C and racemization barrier ΔG ¼ 6 rac .

Figure 2 .
Figure 2. Solid-state structures of 16 e, 16 i and 18 a including selected bond lengths and their tetrahedral character (THC).Structures represent absolute configurations.H-atoms omitted for clarity.Grey: carbon, red: oxygen, blue: nitrogen, green: boron.