Asymmetric Catalysis Axially Chiral 1,1 ′ -Binaphthyl-2-Carboxylic Acid (BINA-Cox) as Ligands for Titanium-Catalyzed Asymmetric Hydroalkoxylation

: Axially chiral, enantiopure 1,1 ′ -binaphthyl-2-carb-oxylic acids (BINA-Cox) have recently been introduced as chiral ligands for transition metal catalysis. Together with equimolar, co-catalytic amounts of Ti(O i Pr) 4 and water they form an in situ catalyst that performs the asymmetric catalytic hydroalkoxylation of 2-allylphenols to 2-methylcoumarans at high temperature (240 °C, microwave heating). The synthesis of reference ligand 2 ′ -MeO-BINA-Cox ( L1 ) has been optimized and performed at molar scale. Synthetic routes have been developed to access a variety of substituted BINA-Cox ligands (>30 exam-Introduction of the high-temperature asymmetric catalytic cyclization of 2-allylphenols to 2-methyl-coumarans. The chiral carboxylic acid in titanium acid–water in situ preferra-bly an axially chiral biaryl-2-carboxylic acid with a methoxy group in the 2 ′ -position. Ligand structure variations for struc-ture-activity studies of the catalyst conveniently realized at the stage of enantiopure MeO-BINA-Cox ( L1 ) as platform chemical. The asymmetric synthesis of L1 by determining key reaction parameters in all steps and adapting the reaction conditions to molar scales with no for chromatography. A new protocol for saponification of sterically hindered, resilient esters in hot PEG-200 at elevated temperature was introduced, which permits the saponification of hindered esters in short reaction time with only a moderate excess of base. More than 30 novel axially chiral biaryl-2-carboxylic acids have been synthesized in enantiopure form and tested as ligands in the titanium-catalyzed asymmetric cycloisomerization of 2-allylphenol to 2-methylcoumaran. Compared with L1 , an increase of catalytic activity was observed in ligands having electron-donating and sterically demanding substituents at the remote 6 ′ -position ( L18 , L19 ). calcd.

We have recently described an intramolecular asymmetric catalytic hydroalkoxylation of 2-allylphenols (1) to 2-methylcoumarans (2) that is catalyzed by a peculiar titanium complex generated by mixing Ti(OR) 4 , the axially chiral carboxylic acid MeO-BINA-Cox (L1) and H 2 O in a 1:1:1 ratio. [19] The process is an example of asymmetric catalytic hydrofunctionalization, and a rare example of asymmetric catalysis with high enantioselectivity at the exceptionally high reaction temperature of 240°C (HOT-CAT, homogeneous thermal catalysis; Scheme 1). [20] Scheme 1. Asymmetric titanium-carboxylate catalyzed hydroalkoxylation of 2-allylphenol.
The ligand-effect in the titanium-carboxylate-catalyzed reaction is critical, and preliminary ligand variation studies covering a number of chiral O,O-, N,N-or O,N-potentially chelate-forming ligands failed to induce notable activity or any enantioselectivity. Another round of ligand screening that focused on combinations of titanium alkoxide with a variety of simple bifunctional chiral carboxylic acids (including proline, N-anisoylprolines (o-, m-, p-anisoyl isomers), mandelic acid, mandelic acid O-methyl ether, camphoric acid, camphoric acid monoamides) likewise failed to show catalytic activity. [21] After such unsuccessful forays into alternative basic ligand structures, it transpired that the 1,1′-biaryl-2-carboxylic acid skeleton should be conserved. A first successful ligand variation involved the substitution of H-6′ in 2-MeO-BINA-Cox (L1), by a tert-butyl group (L18), which slightly increased both activity and enantioselectivity of the model reaction. [19] The goal for further ligand variation studies was to retain the biaryl-2′-alkoxy-2-carboxylic acid substructure and substitute any available position. Since the number of readily accessible, enantiopure biaryl-carboxylic acids is limited, new synthetic routes had to be developed to access the desired products, either by de novo asymmetric synthesis, or by substitution of the more readily available enantiopure MeO-BINA-Cox (L1).
Here, we first present the various synthetic approaches that we have followed to prepare a variety of C 1 -symmetric axially chiral, enantiopure biarylcarboxylic acids. Next, their evaluation as ligands in titanium-catalyzed asymmetric hydroalkoxylation in the model cyclization of 2-allylphenol to 2-methylcoumaran will be compared; finally, we present studies towards the extension of the substrate range in the titanium-carboxylate catalyzed asymmetric catalytic hydroalkoxylation.
The Miyano synthesis of L1 involves a key S N Ar-reaction of 1-naphthylmagnesium bromide (9) with menthyl 1-menthyloxy-2-napthoate (7; Scheme 2), [25] in which the etheric menthyloxy leaving group is responsible for induction, [10] and the menthyl ester suppresses acyl substitution through shielding of the ester carbonyl. [35] The methodology has been applied by other groups, [14,16a-16c,17] and we also got satisfactory results at small scale. Various issues emerged upon scale-up, which were resolved step by step: Methylation of acid 3 to methoxyester 4 with methyl iodide in DMF [25] is uneconomic. Equally good results were obtained with Me 2 SO 4 -K 2 CO 3 in acetone (Scheme 2, a). The methoxy groups of 4 next are exchanged with (1R)menthol (5) under basic conditions. The reported procedure uses three molar equivalents of sodium menthoxide to push the alkoxy-exchange equilibrium [36] towards product 7. [25] The handling and use of NaH for generating the menthoxide base becomes unsafe and wasteful at large scale. We considered performing a catalytic alkoxide exchange with substoichiometric amounts of base, since the exchange product methoxide can be regenerated to sodium menthoxide by reaction with menthol and release of methanol. To drive the reaction towards product 7, methanol as the most volatile component may be removed from the reaction equilibrium.
In our experiments, sodium menthoxide (NaOMen) was generated by stirring 50 mol-% of sodium metal in excess (2.5 equivalents) molten menthol (5) at 190°C. [37] After cooling, the exchange reaction was performed in DMF solution by addition of 4. NMR analyses showed that a rapid transesterification to methoxy-menthyl ester 6 occurs, followed by the slower S N Ar-alkoxyexchange to 7. Dealkylation to 8 occurs as side-reaction at higher temperature, but remains insignificant (<1 mol-% 8) at ≤60°C. A dynamic vacuum (15 mbar) was applied to induce slow distillation of MeOH-DMF from the reaction mixture, presumably as an azeotrope. [38] Plenty of product 7 emerged in the process, but a portion of 6 remained unreacted (Scheme 2, I). Renewed addition of DMF to the concentrated reaction mixture, followed by a second dynamic vacuum distillation raised the conversion to 90 mol-% (Scheme 2, II). [39] Crystallization of the reaction mixture from ethanol gave pure 7 at molar scale in 78 % yield, matching the result of the reaction with excess base. [25] The precursor 1-bromo-2-methoxynaphthalene required for the key S N Ar-reaction (via Grignard reagent 9) has often been prepared by methyl iodide alkylation from commercial 1bromo-2-naphthol. [40] A more economic access at large scale is by bromination of the fragrant compound 2-methoxynaphthalene, which is high-yielding and selective when performed in acetic acid as solvent (see Table S3 for variations). [41] We were now in a position to approach the critical, diastereoselective S N Ar coupling step of 7 and Grignard reagent 9 to give (aS)-10. The scale-up of the reaction met with some difficulties, starting with the limited solubility of 9, which complicated its transfer to the reaction vessel and required using large amounts of solvent. In the actual S N Ar reaction with 7, incomplete conversion was often noted even after extending the reaction time to several days. Heating such reactions with the aim to raise the conversion of 7 induced dealkylation to 8 instead. Unfortunately, neither the original methodological work [25,42] nor later applications [14,16a-16c,17] reported on the impact of specific reaction parameters on the reaction. To learn about effects of specific variables on stereoselectivity and yield, data from published examples [17,25] was collected and supplemented with selected new experiments, in which we analyzed the composition of crude reaction mixtures by qNMR methods (Table S4; Scheme 2, c). The strategy of Miyano et al. to work in a lowpolarity medium (Et 2 O-PhH; Table S4, entry I, II) [25,43] at high dilution (0.05 M), while seemingly optimal to support the che-lated transition state of the stereoselective reaction, [25,26b] is inconvenient for scale-up considering the resulting large reaction volumes. Hoveyda et al. had obtained equally good results in THF-PhH at 0.25 M (entry III), [17] which implies that neither the low polarity of co-solvent ether nor high dilution are necessary. To circumvent solubility issues with Grignard reagent 9, we generated the latter in situ from aryl bromide and magnesium in the presence of substrate 7, but this resulted in a low yield of 10, and reductive C-O-cleavage in 7 to 8 became the major reaction pathway (Scheme 2, d; Table S4, entries 2, 3). [44] It emerged that reagent 9 is best prepared separately at 1 mol/L in THF-toluene (1:5) and transferred while still hot (at 50°C, to prevent crystallization) into a solution of 7 in toluene. Remarkable analytical yields of 99 % of 10 with a dr (aS:aR) of 97:3 were thus achieved at a reaction concentration of 0.5 M (based on initial 7) by applying a slight excess of 9 at 35°C (Scheme 2, e and Table S4, entries 5-8). Precipitation of the crude product and recrystallization gave very satisfactory yields of (aS)-10 (dr ≥99.8:0.2) at scales up to 0.4 mol with no need for chromatography (Scheme 2, e; Table S4, entry 8).
Finally, saponification of ester 10 to MeO-BINA-Cox (L1) with 50-70 equivalents of KOH in hot (80°C) ethanol is wasteful at large scale. To speed up hydrolysis, we intensified the reaction conditions by working in polyethylene glycol (PEG-200) at 150°C, which effected hydrolysis within a few hours with only 5 equivalents of base (Scheme 2, f; for additional tests of conditions, see Table S5). The product was precipitated by acidification and recrystallized to raise the ee of (aS)-L1 to ≥99.7 %, as determined by Fukushi′s 1 H NMR shift method with nicotine as chiral base. [45] This step was also readily scaled with one example performed at 0.25 mol and providing 80 g of L1. The overall yield of L1 from 3 was 51 % over 5 steps, and all purification steps are performed either by distillation or recrystallization, with no need for chromatography.

Structural Modifications of MeO-BINA-Cox Ligands
The singular success of MeO-BINA-Cox (L1) as ligand in the titanium catalyzed intramolecular asymmetric hydroalkoxylation (cf. Scheme 1) [19] created a demand for incremental structure variations of the basic ligand structure, whose defining element is an axially chiral 1,1′-biaryl-2′-alkoxy-2-carboxylic acid. The following sections present various synthetic approaches towards such modified structures.
A few target structures with variations at C-3′ of the alkoxynaphthalene subunit were accessed following the Cram-Miyano-S N Ar route from 7 and the respective alkoxy-bromonaphthalene derived Grignard reagents. The syntheses, performed in analogy to that of L1, tended to proceed sluggishly and with lower stereoselectivity. Even so, the major diastereomers could be obtained in all cases and were saponified to the enantiopure target acids L2-4 (Scheme 3). The sparse results of those syntheses did not recommend further exploration of the de novo asymmetric synthesis approach. It was used once more to access the chiral carboxylic acid and NMR shift reagent MNCB (L5) as another potential ligand for catalytic hydroalkoxylation (Table 1), following Fukushi's synthetic route. [45]

Transformation of the 2′-Methoxy Group
Based on the well-developed route to L1, either the latter or its precursor 10 recommended themselves as synthetic platform for structure variations (Scheme 4). Dealkylation of 10 with BBr 3 at 0°C or r.t. returned lactone 12a, which suffers fast racemization at ambient temperature. [46] At -78°C the same reagent permitted demethylation to give diastereomerically and enantiomerically pure ester 12, whose successive Williamson etherification and saponification return various 2′-alkoxy-BINA-Cox ligands L23-25 (Scheme 4, b, c). Notably, alkylation with α,α′dibromoxylenes gave tethered bis-carboxylic acids L20 and L21. Phenoxy derivative L22 was accessed via Chan-Lam cou- pling, [47,48] after initial attempts at Ullmann coupling had failed. No epimerization occurred in the alkylation or arylation of 12, whose substitution products were diastereomerically pure. Absence of racemization under the conditions of saponification was further proven for the free acids by 1 H NMR spectroscopy with nicotine as chiral shift reagent. [45] Tosylate 13 derived from 12 was prepared with the intention to explore cross coupling approaches towards 2′-aryl-substituted derivatives of MeO-BINA-Cox. Nickel-catalyzed Suzuki coupling (Ni(COD) 2 , PCy 3 , K 3 PO 4 ) of 13 with phenyl boronic acid, [49] followed by saponification, initially returned hydroxyacid L27, besides target L26 (17 %) and hydro-de-metalation product (aS)-1,1′-binaphthyl-2carboxylic acid (L32; 6 %). Repeat experiments with careful exclusion of water provided the desired coupling product in superior selectivity. It was then saponified to L26 (Scheme 4, f and c). An analogous coupling with 3-methoxyphenylboronic acid gave L28, which reintroduces a methoxy group into the ligand periphery.

Ligand Syntheses via Metalation or Reduction of L1
According to Metz et al., H-3 of MeO-BINA-Cox (L1) was regioselectively lithiated with sec-BuLi-TMEDA [50] and then alkylated with MeI to give L13 (Scheme 4, h, i). [51] An analogous metalation followed by quenching with iodine gave L11, whose direct Suzuki coupling with phenyl boronic acid to L12 failed, but was realized at the stage of its methyl ester. Unfortunately, the resulting ester resisted standard saponification and thus was laboriously converted to L12 via LiAlH 4 reduction, IBX-oxidation and Lindgren [52] NaClO 2 oxidation. More efficiently, a phenyl group was introduced at C-3 of L1 following a protocol for Pd-catalyzed ortho-C-H-arylation [53] to give L12 in a single step (Scheme 4, p). Hydrogenation of L1 with in situ activated Raney-Nickel in aqueous 2-propanol [55] led to a 1:1 mixture of tetrahydro-and octahydrogenated derivatives L14/L14′, whose ratio remained unchanged when the mixture was exposed once more to the hydrogenation conditions. The two components could not be separated either by chromatography or fractional crystallization.

S E Ar Functionalization of MeO-BINA-Cox Esters
Friedel-Crafts type functionalization at methoxynaphthalene C-6′ were performed on methyl ester 15 as substrate and succeeded in case of alkylation with tert-butyl chloride [56] or acylation with acetyl chloride [15] to provide L17 and L18, respectively, after saponification (Scheme 5, b-d). When the evaluation of ligand L18 in catalysis returned superior results over L1, we thought it worthwhile to study the effect of 1-adamantyl as typical dispersion energy donor substituent (DED [57] ). Its introduction at C-6′ was attempted through InCl 3 -catalyzed Friedel-Crafts alkylation of menthyl ester 10 with 1-bromoadamantane. [58] Curiously, adamantylated acid L19 was the major product from this reaction besides the expected ester. It appears that HBr, which is released in the alkylation step also cleaves the menthyl ester by dealkylation, and this was supported by detecting both menthyl-(δ H = 3.99) and neomenthyl bromide (δ H = 4.67) in the crude reaction mixture. Saponification was thus spared in the synthesis of L19. The absence of racemization under S E Ar reaction conditions was checked for L17 and L18 through 1 H NMR analysis with nicotine as shift reagent. [45]

Ligand Effects in Titanium-Catalyzed Asymmetric Hydroalkoxylation
The cycloisomerization of 2-allylphenol (1a) to 2-methylcoumaran (2a) served as reference reaction for comparing the performance of ligands L2-L34 in titanium-carboxylate catalyzed asymmetric hydroalkoxylation by determining ligand effects on catalyst activity and stereoselectivity (Table 1). [19] We opted for a short reaction time of 20 min, such that the analytical yields of 2a reflect relative catalytic activities. The reaction is generally sensitive to variations in the reaction temperature, the watercontent of substrates and solvent and the purity of starting allyl phenol 2a. As a means of quality control and to ascertain the integrity of the reaction setup, the standard reaction with L1 was repeated with each new experiment series. Any notable deviation in either yield or enantioselectivity of 2a pointed to problems with either reagents or the microwave unit. The temperature sensor of the latter was also regularly recalibrated.
In spite of the thermally forcing conditions of this reaction ("HOT-CAT", homogeneous thermal catalysis), initial blank experiments with titanium(IV)alkoxide in the absence of carboxylic acid ligand show very little substrate conversion (entry 1), and carboxylic acid L1 in the absence of titanium precursor gave no conversion at all (entry 2). Standard catalytic runs with L1 gave (2S)-2-methylcoumaran (2a) in yields that increased with the holding time at 240°C (entries 3a vs. 3b). The influence of polar functional groups in the ligand sphere upon catalytic activity was explored with suitably modified structures: Converting L1 to its sodium salt L10 quenches the catalytic activity of the in situ catalyst with Ti(OiPr) 4 and water (entry 4). This negative buffering effect points to the impor-tance of a minimal acidity level within the reaction system. Runs with variously functionalized 1,1-binaphthyl-derivatives (entries 2-15) imply that the presence of a single carboxyl (entry 8), or chelation by two coordinating donors (-CO 2 H, -OR, -NH 2 , -CONHR > CO 2 R; entries 5-7, 9, 12-15) are minimal require- EurJOC European Journal of Organic Chemistry ments for catalytic activity. However, enantioselectivity is only achieved with free carboxyls (entries 3,8,12), and high activity and selectivity are only reached by combining one carboxyl with a weakly coordinating 2′-methoxy group (entry 3). Replacing 2′-methoxy with sterically active non-donors retains some selectivity at low activity (entries 10,11). Shifting of the methoxy group from the 2′-position into the periphery by inserting a 2′-meta-anisyl group is ineffective (entry 11). Increasing the size of the 2′-alkoxy-group beyond methoxy successively reduces catalyst activity and selectivity (entries [16][17][18][19]. Attaching a group to the 3′-position (while retaining 2′-alkoxy) reduces catalyst activity and selectivity (entries 20-22). Substitution at C-3 in the naphthoic acid fragment has similar effects (entries 23-25), although the small methyl group boosts catalyst activity at somewhat reduced selectivity (entry 23). The increased activity of L13 might be a consequence of the σ-donor effect of methyl. By placing specific groups into the remote 6′-position of the MeO-BINA-Cox core structure, their electronic influence can be studied with minimal disturbance [a] Conditions: allylphenol (1.5 mmol), toluene (3 mL); Ti(OiPr) 4  of the coordination sphere around the metal. Entries [26][27][28][29][30][31] show that electron-rich +σ and +π groups induce high activity and enantioselectivity similar to, and sometimes surpassing that obtained with L1 (entries 29, 30). The most successful groups are the bulky ones, and thus a steric (or: dispersion donor) influence on catalyst properties cannot be discounted, besides the electron donor effect. In any case, the π-acceptor group of L17 completely suppresses catalyst activity (entry 26).
The in situ catalyst from Ti(OiPr) 4 , Ln and H 2 O is presumably a multinuclear titanium-μ-oxo species. [19] Tethered dicarboxylic acids L20 and L21 were prepared to potentially bridge metal centers more effectively than separate units of L1. Their low catalytic activity points to a steric misfit of the tethering unit, however (entries 33, 34). The partially hydrogenated ligand mixture L14/L14′ displayed lowered activity and selectivity (entry 35), whereas replacing the naphthoic with a dichlorobenzoic acid subunit (L5) [12,45] was well tolerated; the lower activity vs. L1 is consistent with assuming a deactivating σ-acceptor effect exerted by chlorine (entry 36).

Substrate Scope of Asymmetric Cycloisomerization
The cyclization of 2-allylphenols (1) to 2-methylcoumarans (2) was the original assay used for the discovery of the Ti(OiPr) 4 - Table 3. Extended substrate range of intramolecular asymmetric catalytic hydroalkoxylation. [a] [a] Conditions: allylphenol (1.5 mmol), toluene (3 mL); Ti(OiPr) 4  Eur. J. Org. Chem. 2020, 2062-2076 www.eurjoc.org L1-H 2 O catalyst system; results with various core-substituted 2-allylphenols were already reported in our communication [19] and are included in Table 2 for completeness. The catalytic runs for the substrate scope were typically performed with L1 and additional selected examples with L18, or other ligands. The heating phase was extended to 50 minutes in order to approach complete conversion of substrates.

Further Extension of the Substrate Range
To probe further variations of the substrates for titanium-catalyzed (asymmetric) hydroalkoxylation we next concentrated on allylphenols bearing additional substituents within the alkene unit. Crotylphenol (1l) showed little conversion under standard conditions with L1, and lower enantioselectivity ( Table 3, entry  1); 1m with a further extended allyl chain failed to cyclize entirely (entry 2). Core-methylation of crotylphenol to 1n likewise did not improve reactivity (entry 3). Based on those results, we assumed that there is a detrimental steric effect of alkene substitution but were next surprised to find catalytic activity restored with the still higher substituted 2,3-dimethyl-allyl-substrates 1o and 1p (entries 4, 5). The presence of inseparable side products prevented a reliable determination of the ee of the reaction products. Cyclization of 1q and 1r took a different path in that 6-endo-trig cyclization to chromans was preferred (entries 6, 7); chroman 2q is achiral. Aliphatic alkenol 1s cyclized only with difficulty (entry 8). With substrate 1t we found a new type of substrate for the reaction, whose asymmetric 5-exo-trig cyclization also illustrates compatibility of tertiary alkanol substrates (entries 9). Finally, a 6-exo-trig cyclization could also be realized asymmetrically with 2′-vinyl-1,1′-biphenyl-2-ol 1u, which bears an alkene and a phenolic moiety on separate phenyl groups; product 2u was obtained in limited yield and enantioselectivity, however (entries 10).

Substrates Failing to Cyclize
Additional substrates that failed to cyclize will be shortly discussed: 2-Allylphenols with substituents ortho to hydroxyl fail in the model reaction ( Figure 2). A methyl group is sufficient to suppress the catalysis (1aa), disregarding more hindered substrates (1ab, 1ac). The presence of polar, coordinating or πacceptor groups, even remote ones, is another limitation, judging from the inactivity of substrates 1ad and 1af-1ai. The case of 2-prop-1-enylphenol (1aj) is of interest, since this compound could principally be formed by isomerization from 1a. The 5endo-trig cyclization of 1aj to 2a was not observed under catalytic reaction conditions, even though the starting material was fully converted. This points to a relatively fast polymerization of 1aj under reaction conditions, which also explains why the latter is not observed as side-product in the model cyclization 1a→2a, where it escapes analytical detection through fast polymerization. The group of ortho-alkenyl-benzyl alcohols (1t, 1ak-1ao) has their reactive centers homologously shifted relative to allylphenols. Unlike for α,α-dimethylcarbinol 1t, which cyclized to coumaran 2t under standard conditions (cf. Table 3, entries 9), experiments with 1ak-1am returned only starting material. The formal introduction of a -methyl group into 1t prevented cyclization in the resulting 1ao, which suffered elimination of water to give a dialkenylbenzene instead (Scheme S2). The 2-styryl alcohols 1ap and 1aq failed to cyclize under conditions of the model catalysis. In another attempt at converting aliphatic alkenols (cf. 1s in Table 3, entry 8), both substrates 1ar and 1as failed to cyclize.

Conclusions
The present work extends our studies of the high-temperature asymmetric catalytic cyclization of 2-allylphenols to 2-methylcoumarans. [19] The chiral carboxylic acid ligand in the novel titanium alkoxide-carboxylic acid-water in situ catalyst is preferrably an axially chiral biaryl-2-carboxylic acid with a methoxy group in the 2′-position. Ligand structure variations for structure-activity studies of the catalyst were conveniently realized at the stage of enantiopure MeO-BINA-Cox (L1) as platform chemical. The asymmetric synthesis of L1 was improved by determining key reaction parameters in all steps and adapting the reaction conditions to molar scales with no need for chromatography. A new protocol for saponification of sterically hindered, resilient esters in hot PEG-200 at elevated temperature was introduced, which permits the saponification of hindered esters in short reaction time with only a moderate excess of base. More than 30 novel axially chiral biaryl-2-carboxylic acids have been synthesized in enantiopure form and tested as ligands in the titanium-catalyzed asymmetric cycloisomerization of 2-allylphenol to 2-methylcoumaran. Compared with L1, an increase of catalytic activity was observed in ligands having electron-donating and sterically demanding substituents at the remote 6′-position (L18, L19).
Extensions of the substrate scope of the asymmetric catalytic hydroalkoxylation reaction with non-activated alkenes have been explored. Alkylated and halogenated allylphenols, as well as allylphenols having aryl or alkyl groups attached to the alkene unit or to the α-allylic (benzylic) position were tolerated in the catalytic reaction, although with various levels of success regarding the yield and enantiomeric excess of the products. New examples of asymmetric catalytic hydroalkoxylation reactions have been identified, such as the cyclization of 2′-vinylphenyl-(1,1-dialkyl)methanol (cf. 1t) or of 2′-vinyl-1,1′-biphenyl-2-ol (1u) to 1,3-dihydrobenzofuran (2t) or 6H-benzo[c]chromene (2u), respectively.
Based on the ready access to L1, we plan to isolate and study the chiral titanium carboxylate complexes that appear to catalyze the hydroalkoxylation reaction. In combination with the findings from the substrate structure variations, we hope to gain insight into the mechanism of this peculiar reaction.

Experimental Section
General remarks. Unless otherwise specified, all reagents were obtained from commercial suppliers and used without further purification. 2-Allylphenol (1a) was distilled in high vacuum (short-path distillation) and stored under exclusion of air in the dark; typical water content 200 ppm. K 2 CO 3 was dried in high vacuum with heating to 120°C. K 3 PO 4 was finely powdered and dried in high vacuum at 100°C. Commercial KOH flakes (85 % content) were ground to a fine powder before use. Solvents for synthesis were commercially obtained and used without purification. Solvents for column chromatography were of technical grade and used after distillation. Water-free solvents were obtained by passing commercial solvents through a column of dry Al 2 O 3 and storing under argon over 4 Å molecular sieves. Residual water was analyzed by coulometric Karl-Fischer titration.
Chromatography. Column chromatography (CC) was performed on silica gel 60 (35-70 μm particle size) with 0.2 bar positive air pressure. Thin layer chromatography (TLC) was performed on glass plates coated with silica gel 60 F 254 and visualized with UV light (254 nm) and by staining with Mostain [10 g (NH 4 ) 6  Microwave Syntheses were carried out in an Anton Paar Monowave 300 reactor equipped with a MAS 24 autosampler. The temperature was monitored by an external IR thermometer, which was regularly calibrated against an internal optical ruby thermo-probe. Specified reaction times correspond to the holding time at target temperature.

General procedures
General procedure for saponification of biarylcarboxylic acid esters with KOH in EtOH (GP-1A): To a solution of 1.00 equiv. ester and KOH (85 % content) in EtOH (5-10 mL/mmol), a little water was added (ca. 0.2 mL/mmol) and the reaction mixture was heated to reflux overnight (bath 100°C). After cooling to r.t. and addition of H 2 O and Et 2 O, the layers were separated. The aqueous layer was acidified using aqueous 2 M HCl and extracted with several portions of Et 2 O. The organic layers were combined and washed with aqueous 2 M HCl and saturated aqueous NaCl. After drying over Na 2 SO 4 and filtration, the solvent was evaporated giving the crude reaction product.

General procedure for saponification of biarylcarboxylic acid esters with KOH in EtOH (GP-1B):
To a solution of 1.00 equiv. ester and solid KOH (85 % content) in EtOH (5-10 mL/mmol), a little water (0.2 mL/mmol) was added and the reaction mixture was heated to reflux overnight (bath 100°C). After cooling to r.t. and addition of H 2 O and Et 2 O, the layers were separated. The organic layer was extracted with saturated aqueous LiOH, then discarded. The combined aqueous layers were acidified with aqueous 2 M HCl and extracted with several portions of Et 2 O. The combined extracts were dried with Na 2 SO 4 . After filtration, the solvent was evaporated to give the crude product. Note: use of aqueous LiOH can be beneficial for bringing carboxylate into the aqueous layer, in case the potassium salt is partially soluble in the organic layer. EurJOC European Journal of Organic Chemistry the aqueous layer was extracted with EtOAc (2-3 ×). The combined organic phases were washed with H 2 O (5 ×), dried with MgSO 4 and filtered. After removal of the solvent in vacuo, the crude residue was dissolved in a small amount of CH 2 Cl 2 and a defined amount of trichloroethene was added as an internal standard for qNMR analysis. After the NMR analysis, solvent and internal standard were removed in vacuo to give the crude reaction product.
General procedure for determining the enantiomeric excess of biarylcarboxylic acids (GP-2). The enantiomeric excess of chiral carboxylic acids was determined by an NMR chiral shift method using (-)-nicotine as chiral base according to Fukushi: [45] A sample of the chiral carboxylic acid (15-30 μmol, 1.0 equiv.) was dissolved in CDCl 3 and (-)-nicotine (10 μL, 60 μmol, 2-4 equiv.) was added. The 1 H NMR spectrum was recorded using a relaxation delay (d1) of 20 seconds. For MeO-BINA-Cox derivatives, the methoxy signals for enantiomeric anions appear at different chemical shifts (e.g., Δδ H 0.05 for L1). Integration of the methoxy singlets -by means of deconvolutive peak analysis, if necessary -gave the relative amounts of diastereomeric ion pairs (dr), from which the ee of the acid is derived.
General procedure for asymmetric catalytic hydroalkoxylation (GP-3). Under argon, the ligand (0.05 equiv.) was combined with titanium(IV) isopropoxide (0.05 equiv.; a stock solution in toluene may be used) in a borosilicate glass vial. H 2 O (0.05 equiv.) was added to the lower vessel wall by micro-syringe, followed by dry toluene (3 mL). The resulting mixture was stirred for 10 min at 60°C. The substrate (1.00 equiv.) was added and the mixture was heated in a microwave reactor to the target temperature, where it was held for the indicated reaction time. After cooling, an internal standard (tetradecane) was added to the crude reaction mixture and an aliquot was removed for qNMR analysis. The reaction mixture was placed on top of a solvent-filled silica gel column for purification by CC. Enantiomeric excess was determined by chiral HPLC analysis of chromatographically purified reaction product.

Methyl 1-methoxy-2-naphthoate (4):
A three-necked 4 L roundbottom flask was charged with 1-hydroxy-2-naphthoic acid (3; 400 g, 2.13 mol, 1.00 equiv.), acetone (2 L) and dimethyl sulfate (424 mL, 4.48 mol, 2.10 equiv.; CAUTION). a) To the mechanically stirred suspension, b) potassium carbonate (648 g, 4.69 mol, 2.20 equiv.) was added in portions over the course of 5 h. The internal reaction temperature was initially kept at 20°C by an external water bath, to which ice was added as needed. After half the amount of the base had been added, the reaction mixture was warmed by the reaction heat, and the water bath was additionally heated to 50°C. c) After completion of the base addition, the reaction mixture was stirred for 2 h at 50°C, when TLC indicated consumption of both starting material (3) and the intermediary methyl 1-hydroxynaphthoate (R f 0.49; EtOAc-hexanes, 1:10). After cooling to r.t., aqueous 25 % NH 3 (100 mL) and H 2 O (1.2 L) were added slowly with continued stirring. d) The top organic layer was removed e) from the aqueous layer. The aqueous layer was extracted with Et 2 O (2 ×) and the combined organic layers were washed with saturated aqueous NaCl (2 ×). After drying (MgSO 4 ) and filtration, solvents were evaporated. The crude oil was distilled (118-122°C, oilpump vacuum, ca 0.1 mbar) to give bright-yellow liquid (442 g, 96 %). Notes: a) Safety measures for the case of spilling of dimethyl sulfate or bursting of the reaction vessel were taken. The reaction vessel was placed into a water bath in a metallic pan, and aqueous 25 % ammonia was kept in reach for decontamination of spills. b) Motor-driven mechanical stirring is required at large scale. c) Stirring of the heated suspension proved to be considerably easier Eur. J. Org. Chem. 2020, 2062-2076 www.eurjoc.org than of the cooled reaction mixture. External ice-cooling may not be necessary at all, if K 2 CO 3 is added at a rate to keep the temperature of the water bath below the boiling point of the reaction mixture. d) The addition of ammonia (EXOTHERM!) is a safety-measure to quench excess dimethyl sulfate by alkylation, which renders the ensuing work-up more safely. e) Since no sufficiently large separatory funnel was available, phase separation was effected by transfer of the upper organic layer through PTFE tubing under a positive nitrogen pressure. The lower aqueous phase was extracted by mechanical stirring with new solvent added to the reaction vessel. R f 0.35 (EtOAc-hexanes, 1:10)  Under argon, a solution of 1-bromo-2-methoxynaphthalene (119 g, 500 mmol, 1.25 equiv.) in dry toluene (460 mL) was added to magnesium (14.6 g, 600 mmol, 1.50 equiv.) in dry THF (90 mL) in portions; initiation of the reaction was assured after the first addition. The reaction temperature was kept at 40-50°C by means of an external water bath to prevent either over-reaction at elevated or crystallization of the Grignard reagent at lower temperature. After completion of the addition, the reaction solution was heated to 55°C for another 1 h. The resulting Grignard solution was transferred (while warm) continuously or in several portions through PTFE tubing into a solution of (1R)-menthyl 1-(1R)-menthyloxynaphthyl-2-carboxylate (7; 186 g, 400 mmol, 1.00 equiv.) in dry toluene (220 mL) kept at r.t. by a water bath, also ensuring that no magnesium-metal was transferred.  (12). Under argon, a solution of (1R)-menthyl (aS)-2′-methoxy-(1,1′binaphthyl)-2-carboxylate (10; 2.53 g, 5.32 mmol, 1.00 equiv.) in dry CH 2 Cl 2 (60 mL) was cooled to -78°C. BBr 3 (1 M in CH 2 Cl 2 ; 10.8 mL, 10.8 mmol, 2.00 equiv.) was added dropwise over 15 min and the reaction mixture was stirred at -78°C for 5 h. Saturated aqueous LiOH (15 mL) was added and the mixture warmed to r.t. After addition of H 2 O (20 mL), the layers were separated. The aqueous layer was extracted with CH 2 Cl 2 (50 mL) and the combined organic layers were washed with saturated aqueous NaCl (100 mL). After drying (Na 2 SO 4 ) and filtration, the solvent was removed in vacuo. The crude product was purified by CC (SiO 2 , EtOAc-hexanes, 1:40→1:4) to give 2.01 g (82 %) colorless solid. 1    thyl (aS)-2′-tosyloxy-[1,1′-binaphthyl]-2-carboxylate (13; 300 mg, 495 μmol, 1.00 equiv.). The reaction mixture was heated to 45°C for 48 h. After filtration, EtOAc (25 mL) and H 2 O (25 mL) were added to the filtrate and the layers were separated. The aqueous layer was extracted with EtOAc (25 mL) and the combined organic layers were washed with saturated aqueous NaCl (2 × 25 mL). After drying (Na 2 SO 4 ) and filtration, the solvent was removed in vacuo. The crude product was purified by CC (SiO 2 , EtOAc-hexanes, 1:20) to give 172 mg (68 %) colorless solid. 1 13 13 [53] To a mixture of L1 (328 mg, 1.00 mmol, 1.00 equiv.), Ag 2 CO 3 (303 mg, 1.10 mmol, 1.10 equiv.), K 2 CO 3 (138 mg, 1.00 mmol, 1.00 equiv.), Pd(OAc) 2 (22.5 mg, 0.10 mmol, 0.10 equiv.) and N-acetylglycin (23.4 mg, 0.20 mmol, 0.20 equiv.) under argon, iodobenzene (1.30 mL, 12.0 mmol, 12.0 equiv.) and HOAc (1.00 mL, 18.0 mmol, 18.0 equiv.) were added in one portion each at r.t. The reaction was heated to 90°C for 3 d. After cooling to r.t., aqueous 1 M HCl (5 mL) was added and the mixture was filtered through celite, followed by washing of the filter cake with EtOAc (3 × 25 mL). After separation of the layers, the aqueous layer was extracted with EtOAc (2 × 25 mL) and the combined organic phase was washed with saturated aqueous NaCl (10 mL), dried (MgSO 4 ), and filtered. Purification of the crude product by CC (SiO 2 , EtOAc-hexanes, 1:10 + 1 % HOAc) and recrystallization from MeOH (1.5 mL) gave 206 mg (51 %) yellow solid (≥95 % ee by GP-2). 1