Palladium‐Catalyzed C(sp3)−H Arylation of Primary Amines Using a Catalytic Alkyl Acetal to Form a Transient Directing Group

Abstract C−H Functionalization of amines is a prominent challenge due to the strong complexation of amines to transition metal catalysts, and therefore typically requires derivatization at nitrogen with a directing group. Transient directing groups (TDGs) permit C−H functionalization in a single operation, without needing these additional steps for directing group installation and removal. Here we report a palladium catalyzed γ‐C−H arylation of amines using catalytic amounts of alkyl acetals as transient activators (e.g. commercially available (2,2‐dimethoxyethoxy)benzene). This simple additive enables arylation of amines with a wide range of aryl iodides. Key structural features of the novel TDG are examined, demonstrating an important role for the masked carbonyl and ether functionalities. Detailed kinetic (RPKA) and mechanistic investigations determine the order in all reagents, and identify cyclopalladation as the turnover limiting step. Finally, the discovery of an unprecedented off‐cycle free‐amine directed ϵ‐cyclopalladation of the arylation product is reported.


Introduction
Amines are crucial structural features in biological molecules, pharmaceuticals anda grochemical products. [1] Numerous simple aliphatic amines are commercially available and react at nitrogen with well-established transformations. In contrast, derivatization of amines by CÀCb ond formation at unactivated positions remainsafrontier challenge in synthesis. Transitionmetal-catalyzed C(sp 3 )ÀHf unctionalization offers huge potential for such value addingt ransformations of feedstock chemicals. [2] However,f ree amine substrates deactivate metalc atalysts by strong coordination, and readily undergo oxidation by b-hydride elimination, [3] typically preventing the desired transformation.I mportant recent developments have derivatized nitrogen with amide or sulfonamide directing groups to modify the coordinating ability of the amine and to direct CÀHf unctionalization. [4] This requires additionals teps for directing group installationa nd removal.V ery few C(sp 3 )ÀHf unctionalization methods have been successful on unprotecteda mines. Notably,G aunt developed palladium-catalyzed functionalization of hindered secondary amines ubstrates, [5] to avoid formation of inactive bis(amine) palladium complexes,i ncluding the arylation of tetramethylpiperidine derivatives (Scheme 1a). [5d] Shi reported acetoxylation of t-amylamine in AcOH with diacetoxyiodobenzene as the oxidant, minimizing inactive amine-Pd complexes by protonation. [6] Very recently,t ransient directing groups (TDGs) have been applied to the functionalization of C(sp 3 )ÀHb onds, [7] particularly arylation of aldehyde andk etones ubstrates via endo-imines formed in situ with amine additives. [8] The use of exo-imine directingg roups for CÀHf unctionalization of primary amines remainsr are, with examples using palladium catalysis. [9][10][11][12] In 2016 Dong used stoichiometric 8-formylquinoline to form an imine which effected arylation with bisaryliodonium salts. [9] At as imilar time, Ge reported the catalytic use of glycoxylic acid to form at ransienti mine directing group for arylation of tamylamine derivatives (Scheme 1b). [10] Yu reported catalytic 2hydroxynicotinaldehyde as ah ighly effective directing group, compatible with propylamine as well as a-a nd b-branched amines. [11] In 2017 Murakami used as toichiometrics alicaldehyde in separatei mine formation and hydrolysiss teps, completed in ao ne-pot sequence. [12] While preparing this manuscript, Young reported the use of carbon dioxide to promote CÀHa rylation of primary ands econdary aliphatic amines. [13] Here we report new transient directing groups for CÀHf unctionalization of amines with the first examples of alkyl aldehydes as transient directing groups (Scheme 1c). The use of a stable acetal precursor in lowc atalytic loadings promotes the g-arylation of primary amines with aryl iodides using palladium catalysis. Furthermore, we report detailed kinetic (RPKA) and mechanistic investigations. These studies provided important insights into the reaction, including reaction orders of all components and the identification of an unusual e-cyclometallation pathway of the arylated products.

Results and Discussion
We envisaged using alkyl aldehydes as TDGs for primary amine CÀHa rylation in order to aid reversible imine formationa nd open new possibilities for transientd irecting groupsi nC ÀH functionalization. This would enable as econd coordinating site to be included at the a-position of the aldehyde, which could be readily tuned to optimize the properties. The secondary coordinating groups were intended to provide ac helate to stabilize intermediates on the catalytic cycle, [14] limit imine isomerization, increasea ldehyde electrophilicity andp revent cyclometalation of the directing group. However,s uch functionalized aldehydes are typically unstable to storage, andt herefore we considered the use of acetals as more accessible alternatives, being easily prepared and commerciallya vailable.T his would be the first example of this type of transient directing group in any CÀHf unctionalization reaction.
After initial investigation, as creen of differentp otential TDGs was conducted using palladium acetate with silver trifluoroacetate in acetic acid and water as solvent (Scheme 2). Under these conditions, in the absence of any additional TDG using t-amylamine 1,t he free amine promoted the arylationi n 26 %y ield, which we aimed to improveb yi nvoking at ransient imine strategy.I nc omparison to our previousw ork using mono-N-Tse thylenediamine as aT DG for aldehydeC ÀHa rylation, [8c] we first considered 2-N-tosyl acetaldehyde acetal 3.A ddition of 0.15 equivalents of the acetal gave ad ecreased yield compared to the background reaction, presumably due to strong coordination to the palladiumc atalyst. However, an improvedy ield was achieved by using methoxyd erivative 4.T rifluoroethyl ether 5 had no beneficial effect on the reaction. A potentially tridentate TDG containing aM EM protected alcohol 6 improved the yield to 41 %, whereas the corresponding free alcohol 7 gave 32 %y ield. Pleasingly,c ommerciallya vailable phenoxyacetaldehyde dimethyl acetal 8 gave an improved 46 %y ield. Its corresponding aldehyde 9 gave as imilar result, consistentw ith hydrolysisu nder the reaction conditions. We then investigated variations to the aromatic ether. p-Methoxy substituted derivative 10 was less effective than the phenyl ether,a nd the more hindered 2,6-dimethoxy example 11 also gave al ower yield. Electron-poor 4-trifluoromethylphenyl derivative 12 gave the highest yield of 57 %. Further decreasing the electron density of the ring was detrimental (13,2 9%).
Further optimization of the reactionconditions was conducted with commercially available 8 (0.15 equiv), which improved the yield of 2a to 64 %( by 1 HNMR) and 59 %i solated yield, by using palladium pivalate( 10 mol %), AgTFA, and as olventm ixture of AcOH:HFIP:H 2 O. [15] The use of silver salts was essential. Notably,as imilar maximum yield wasa chieved with acetal 12. Using these conditions, the reactionwas shown to be compatible with ar ange of aryl iodides (Scheme 3).
Scheme3.Reactionscope of aryl iodides using acetals 8 and 12 as transient directing groups. Both directing groups 8 and 12 were investigated:a cetal 12 gave comparable and often improved yields especially for examplesw hen 8 was less effective. In all cases, the product amines were isolated in analyticallyp ure form by simple aqueous work-up, without chromatography.A ryl iodides bearing both electron-withdrawing ande lectron-donating functionality at the para-a nd meta-positions were successful to give amines 2a-l.H alogensw ere well tolerated, providing positions for further derivatization. Chlorophenyl derivative 2f was demonstrated on a2mmol scale with 46 %i solated yield, again involvingo nly as imple work up to provide the pure amine compound. Pleasingly,e ven ortho-substituted aryl iodides (Me, F and OMe), which are often incompatible with Pd-catalyzedC À Ha rylation,w eres uccessful giving 2m-o withoutasignificant reduction in yield, particularly using acetal 12.M ore complex aryl iodides were also tolerated (2p-s), including heterocyclic derivatives.
We then investigated the amine component with iodobenzene as the coupling partner (Scheme 4). Using derivatives with longer alkyl substituents gave similary ields and displayed selectivea rylation at the g-CH 3 position ( 14 and 15). Amine 16 with al ong hexyl chain gave al ower yield. Cyclohexyl and THP containing amines were suitable substrates for the arylation, affording 17 and 18 in good yields in this operationally simple one-pot process. Arylationo fg-methylene positionsw as not observedf or any example.
Interestingly,w hen using neopentylamine under our standard conditions arylated pivaldehyde derivatives were detected due to competingo xidation processes. [15] Therefore, milder reactionc onditions were developedf or this family of amines using ar educed reaction time and silver trifluoroacetatel oading (3 h, 1.5 equivA gTFA) to reduce oxidation. These afforded an improved isolatedy ield for arylated neopentylamine 19 (30 %, combined mono andd iarylation). a,b-Branched amines were also arylated under these modifiedc onditions to afford products 20 and 21,both in 33 %yield. When using acommercial mixture of cis-a nd trans-2-methylcyclohexanamine the only product corresponded to trans-22 (33 %y ield);n oa rylation occurred on the cis-diastereomer indicating ah igh selectivity.I nc omparison, under the same conditions Yu's 2-hydroxy-nicotinaldehyde transient directing group [11] gave am ixture of products consisting of both trans-a nd cis-arylation as well as further arylation on other g-positions.
Next, structuralc omparisons were undertaken to indicate the structural features of the acetal which were important in promoting the arylation (Scheme 5). Ketal 23 did not improve the yield above the background reaction, suggesting restricted imine formation. Hydrocinnamaldehyde acetal 24,w ithoutt he ether oxygen, gave al ower yield than in absence of any additive, indicating the importance of the OAr group in forming the putative bidentate TDG. In each case, the corresponding carbonyl species 25 and 26 gave similar yields, agains upporting the rapid hydrolysis of the acetals under the reactionc onditions. Catalytic acetal 8 was shown by 1 HNMR to readily hydrolyse to the aldehyde under the reaction conditions. The use of ethers 27-29,w ith similars teric and coordinating properties to acetal 8 were unable to enhancethe yield, again supporting the need for imine formation.
Based on these resultsa nd prior studies we propose the reaction to be accelerated by formation of at ransienti mine directing group. Unfortunately,a ttempts to isolate Pd-complexes with the imine were unsuccessful. Therefore, we examined the kinetics of the reactiont op robe the detailso ft he dual catalytic cycle. To date, there are no reports of kinetic analysiso n transientC ÀHf unctionalization reactions. Blackmond pioneeredr eactionp rogressk inetic analysis( RPKA) to make use of the kinetic information in the full reaction profile, [16] as well as the use of visual comparisons of concentration profiles. [17] BurØs' recent advances have provided tools to extend the visual comparison methodst od etermine reactiono rder for catalyst [18] and other reactionc omponents from reaction profiles. [19] Here, due to the heterogeneousn ature of the reactions, data was generated through the quenching of individual reactions, monitoring formation of arylated amine 2a.
The overall reaction profile using acetal 8 indicated that the reactionw as rapid, with maximum conversion occurring by 3h and little subsequent change. Unreacted t-amylamine made up the mass-balance.U sing the BurØsm ethod, the reactionw as shown to be first order in Pd, consistent with am onomeric Pd catalyst. Surprisingly,v ariation in the loading of acetal 8 had little effect:i dentical rates were observed at 5, 15 and 25 mol %i ndicating az ero order (Figure 1). [15,20] The same rate was observed with acetal 12,w hereas the reactionw as considerably slower in the absence of the acetal (with am uch lower final yield) suggesting ad ifferent mechanism. Different excess experiments for amine 1,P hI and AgTFA, [21] also indicated zero order in thesec omponents.W ea lso found identical reaction rates with different aryl iodide species ( 4-iodoanisole, iodobenzene, and 4-chloroiodobenzene). [15] The degree of stirring had little effect on the reactionr ate indicating mass transport is not rate limiting.
To probe the CÀHa ctivation step, amine 1 was subjected to the CÀHa rylationc onditions with PhI using AcOD-D 4 as the solvent. No D-incorporation was observed at the benzylic position of the arylated product, suggestingi rreversible CÀHa ctivation (Scheme 6a). Furthermore, comparison of the rate of formationo famines 15 and D 4 -15 showed af aster rate for the proteo-substrate, and as ignificant kinetici sotope effect (k H / k D = 2.26, Scheme 6b). These kinetic andd euteration studies all indicatec yclopalladationa st he turnover limiting step in the reaction.
Interestingly,t he addition of further acetal or Pd at later time points in the reaction did not improvec onversion, [22] posingaquestion about what determined the final yield obtained.T he use of a' same excess' experiment, developed by Blackmond, [16] provided important insight. This experiment uses reduced concentrations of the reactants to mimic starting the reaction at al ater time point (here40% conversion;1hr eaction time), and can highlight catalystd eactivation or product inhibition. On time adjustment of the profiles, the traces did not overlap, indicating ar eduction in rate under the standard conditions at later times due to one of these effects (Figure 2). [23] Adding the expected product concentration from the start of the 'same excess' experiments gave ap rofile that overlaid very well with that using the standard conditions, showingt he rate reduction was due to product inhibition.
While product inhibition might be expected based on equilibriumc onsiderations, the pronouncede ffect suggesteda n additional explanation. Therefore, we examined the interaction of the product with palladium acetate in AcOD-D 4 ,e xpecting a preferential complexation of the product amine with Pd over the starting material. Mixing amine 2f with Pd(OAc) 2 in a2 :1 ratio at rt gave three amine species, assigneda st he protonated amine, am onoamino-Pd complex, and ap alladacycle corresponding to e-cyclometallation (Scheme7). [15,24] On heating at the reactiont emperature for 30 min, the aromatic signals changed to a2 Hs inglet corresponding to deuteration of the ortho-positions of the aryl group. This surprisingly facile and reversible cyclopalladation process out-competes coordination of Figure 1. Zero order rate dependence of acetal 8.Y ields determined by 1 HNMR using 1,3,5-trimethoxybenzene as an internal standard on discrete reactions.
Based on these studies, ap roposed mechanism for the dual catalytic system is given in Scheme 8. Under the reaction conditions, acetal 8 is hydrolysed to the corresponding aldehyde 9.T he aldehyde then condenses with the free amine 1,f rom a pool of the protonated amine species, to form imine I.C omplexation to the palladium catalystw ith loss of ac arboxylate ligand would give ap ositively charged imine-Pd species II which undergoes turnover-limiting cyclopalladation to afford the cyclometalated imine intermediate III.I ti sl ikely that deprotonation or ligand exchange occurs to form an eutral species that then undergoes oxidative addition to Pd IV intermediate IV.C ÀCb ond formation by reductive elimination would afford the product imine V andr egenerate the Pd II catalyst. The transient directing group is turned over by hydrolysis with water,o rd irect transamination, generating the arylated product 2.T he product amine inhibits the reaction by competitive coordination to the catalysta nd facile e-cyclometallation, removingp alladium from the productive cycle. The observedr eaction profile can be rationalized by the following assumptions. In the earlys tages of the reaction, when the product concentration is low,l ittle to no Pd-product amine complexation occurs, and so [cat] t = [cat] 0 and the increasei np roduct concentration is linear with zeroo rder kinetics. However,a ta critical higher product concentration (at approx.4 0% conversion), inactive and thermodynamically favourable Pd-product complexes are formed, removing active catalyst from the reac-tion and causingareduction in rate. Eventually,a ll of the catalyst is trapped in these complexes, causing the yield to plateau. Addition of more catalyst (10 mol %) at t = 1h didn ot lead to increased conversion;presumably also due to rapid formationo ft he inactive complexes at high product concentration, or due to breakdown of the transienta ldehyde species. [15] Conclusion We have described the development and mechanistic investigation of an ew transienta ctivator in palladium catalyzed CÀH arylation of primarya mines.T he use of ab ench stable alkyl acetala cts through formation of at ransient imine directing group, with ac lear influence of the (masked) aldehyde and the aryl ether as as econd coordinating group. Commerciallya vailable (2,2-dimethoxyethoxy)benzenep rovides af acile and rapid reaction of branched primary amines with al arge range of aryl iodides. Improved yields can be obtained in some cases with (2,2-dimethoxyethoxy)-4-trifluoromethylbenzene as am odified TDG. The CÀHa ctivation step was found to be turnover limiting, determined by az ero-orderr ate dependence observed in all reactants as well as ap ronouncedH /D kinetic isotope effect.T he reactionw as first order in palladium catalyst. Furthermore, we have uncovered the unexpected formation of a 7-membered palladacycle, activating an e-C(sp 2 )ÀHb ond of the product, thus inhibiting the reaction. We expect thesei nsights will also feed into future design of improvedd irecting groups,and aid the continued rapid progress of the field.