Beyond Tertiary Amines: Introducing Secondary Amines by Palladium/Norbornene-Catalyzed Ortho Amination

Abstract

Since the discovery of the palladium/norbornene (Pd/NBE)-catalyzed ortho amination in 2013, escaping the limitation of only yielding tertiary anilines has been a long-standing challenge in the past decade. Here, we describe that, by carefully choosing the phosphine ligand and NBE mediator, installation of a N-mono-alkylamino group becomes feasible. The reaction tolerates a wide range of aryl iodide substrates and various N-mono-tertiary alkylamine-derived electrophiles. Both ipso alkenylation and alkynylation can be realized. The synthetic utility of this method is exemplified in the formation of primary amino group via selective deprotection and streamlined access to N-heterocycles. Preliminary success of installing a bulky N-secondary alkylamino group and mechanistic understanding of the decomposition pathways of mono N-alkylamine electrophiles have been obtained.

Graphical Abstract

The first introduction of secondary amines at the aryl iodide ortho position by the palladium/norbornene (Pd/NBE) catalysis has been developed, which leads to versatile synthetic applications, such as primary amino group installation and concise heterocycle formations.

Owing to the prevalence of arylamines in bioactive molecules (Figure 1), such as natural products, pharmaceuticals, and agrochemicals,^[1] amination of aromatic compounds has been one of the pivotal tools for synthetic chemists. The most widely used approach is the ipso amination of arenes (Scheme 1a), in which an existing functional group is replaced by an amino group, such as S_NAr,^[2] Buchwald–Hartwig amination,^[3] Ullman–Goldberg^[4] and Chan–Evans–Lam^[5] couplings, etc.^[6] While direct substitution of a C(aryl)–H bond with a C(aryl)–N bond, namely C–H amination,^[7] is a very attractive approach, control of site-selectivity for electronically unbiased substrates mainly relies on the use of directing groups (Scheme 1b).^[8] On the other hand, the palladium/norbornene (Pd/NBE)-catalyzed ortho amination provides an alternative strategy to introduce amino groups to the ortho position of aryl halides (Scheme 1c).^[9] In this reaction, through the key aryl norbornyl palladacycle (ANP) intermediate,^[10] an amine electrophile, i.e., N-benzoyloxyamines,^[11] reacts with ANP to furnish ortho C–H amination and the catalytic cycle is completed by coupling a nucleophile (including alkenes) at the ipso position. As a result, it offers a complementary site-selectivity to the ipso amination approaches.^[9a]

Figure 1.

Selected pharmaceuticals that contain arylamines with secondary and primary amino groups.

Scheme 1.

Site-selective amination of arenes. a) Ipso amination via S_NAr or cross couplings; b) C–H amination often uses a directing group; c) Ortho C–H amination via the Pd/NBE catalysis; d) This work introduces secondary and primary amino groups via ortho C–H amination. Nu, nucleophile; Bz, benzoyl.

Since the Pd/NBE-catalyzed ortho amination was discovered in 2013,^[9a] a large variety of nucleophiles have been identified to be compatible for the ortho amination/ipso functionalization reactions.^[12] Its utility has been found in synthesizing indole derivatives,^[13] simplifying small-molecule drug preparation,^[12e,14] and enabling carbonyl 1,2-transposition.^[15] However, the main limitation lasting in the past decade is that this ortho amination approach can only introduce tertiary amino moieties, i.e., –NR₂ (R: alkyl). In particular, amino groups embedded in a six-membered ring, such as morpholinyl, are most common. Installation of secondary (–NHR) or primary (–NH₂) amino groups via the Pd/NBE catalysis remains unachieved to date (Scheme 1c). We envision that, by overcoming the limitation of tertiary amino groups, applications of the ortho amination can be further extended. For example, acylation of the resulting secondary and primary anilines can give amide products, and intramolecular cyclization with –NHR or –NH₂ groups can offer diverse heterocycles. Both applications are not easily realizable with the current ortho amination approach. Herein, we describe our preliminary efforts of introducing secondary amino groups via the Pd/NBE catalysis and primary amino groups via subsequent dealkylation (Scheme 1d).

Why has a decade elapsed without addressing the limitation of the amino group scope in the Pd/NBE-catalyzed ortho amination reaction? This is because simply replacing the benzoyloxy dialkylamines with the corresponding mono N-alkylamine reagents under the identical reaction conditions led to no desired ortho amination product (Scheme 2). For example, when N-(n-butyl)-O-benzoylhydroxylamine (R = n-butyl) was used as the amination reagent, a complex mixture of unidentifiable compounds was obtained with no target product detected under the standard conditions reported for the ortho amination/ipso alkenylation^[12a] or alkynylation^[12d]. While the use of N-(tert-butyl)-O-benzoylhydroxylamine (R = t-butyl) also yielded no desired products, some side products (i-iv) could nevertheless be identified (Scheme 2, for details, see the Supporting Information), which provides some insights about the potential issues using mono N-alkylamine reagents. The main concern is associated with the acidic N–hydrogen for the primary amine-derived electrophiles. For example, after their reaction with ANP, deprotonation of the aniline moiety can trigger the reductive elimination to give the indoline side product (iii).^[16] In addition, nitrene is known to be generated from benzoyloxy mono N-alkylamines upon reacting with a transition metal,^[17] which can lead to decomposition of the amine electrophile and/or undesired pathways, including ortho arylation (for side product i), reductive elimination of ANP (for side product ii) and ipso amination (for side product iv). Moreover, phosphines are known to reduce related NH-OBz species to form iminophosphoranes,^[18] which is another pathway for decomposition of the amine electrophiles. Hence, these competing reactions underscore the challenges of introducing secondary or primary amino groups via the Pd/NBE-catalyzed ortho amination.

Scheme 2.

Challenges of ortho amination with mono N-alkylamine-derived electrophiles. E, electrophile.

To realize the desired ortho amination for preparing secondary anilines, the key is to avoid decomposition of the mono N-alkylamine reagents and to avoid NBE-mediated reductive elimination. We hypothesized that the use of an electron-deficient less-nucleophilic phosphine ligand could diminish unwanted reduction of the amine electrophile and the use of C2-substituted NBE should minimize NBE-mediated reductive elimination.^[19] To test this hypothesis, 2-iodoanisole (1a) was employed as the model substrate and the ortho amination/ipso Heck coupling reaction was tested using N-(tert-butyl)-O-benzoylhydroxylamine (2a) as the electrophile. Indeed, after examining a range of phosphine ligands and structurally modified NBEs (smNBEs),^[20] the electron-poor tris[3,5-bis(trifluoromethyl)phenyl]phosphine (L1) and C2-substituted NBE (N1)^[19] were found to be the optimal ligand/NBE combination, which ultimately delivered the desired secondary aniline product (4a) in 79% yield (Table 1, entry 1). Under the standard conditions, the previous side products (i-iv) were effectively inhibited, and the current side reaction is the direct ipso Heck. A number of control experiments were conducted to understand the role of the reactants. First, unsurprisingly, in the absence of Pd(OAc)₂, ligand or NBE, no desired product was observed (entries 2, 5, and 7). Other palladium pre-catalysts, such as PdCl₂ and Pd(COD)Cl₂, were somewhat less efficient (entries 3 and 4). Given the importance of phosphine ligands to the Pd/NBE catalysis as first demonstrated by Lautens,^[21] monodentate phosphine ligands with varying electronic properties were tested (entry 6). While simple PPh₃ gave almost no desired product with obvious formation of four-membered side product (ii), the use of the more electron deficient P(4-CF₃C₆H₄)₃ gave much improved yield, likely through inhibiting the iminophosphorane formation. The best result was obtained with tris[3,5-bis(trifluoromethyl)phenyl]phosphine (L1); in contrast, P(C₆F₅)₃ gave no reactivity, probably because it is too electron-deficient to promote oxidative addition of the aryl iodide substrate. The electronic effect of the benzoate leaving group on the electrophile was also assessed. While the electro-rich variant (2b) gave a comparable result, the more electro-deficient 2c, i.e. the one with a better leaving group, led to considerably lower yield, possibly owing to the decreased stability of 2c under the reaction conditions. On the other hand, the NBE effect is also evident (entry 8). Except C2-substituted NBEs (N1, isopropyl ester-derived N3^[22] and amide-derived N4^[23]), other smNBEs and simple NBE gave very low yield of the desired products and accompanied with some reductive elimination side products (ii and iii). THF proves to be the optimal solvent, and the use of less polar toluene or more polar acetonitrile was not effective (entries 9 and 10). Finally, replacing Cs₂CO₃ with K₂CO₃ led to significantly lower yield; however, the use of CsOAc afforded comparable efficiency (entries 11 and 12).

Table 1.

Control experiments.

Entry[a]	Variation from the “standard conditions”	Yield of 4a (%)[b]	Yield of 4a’ (%)[b]
1	none	79	12
2	No Pd	0	0
3	PdCl2 instead of Pd(OAc)2	56	16
4	Pd(COD)Cl2 instead of Pd(OAc)2	52	20
5	No L1	0	0
6	other ligands instead of L1	listed above
7	No N1	0	<5
8	other norbornenes instead of N1	listed above
9	toluene instead of THF	16	15
10	CH3CN instead of THF	trace	trace
11	K2CO3 instead of Cs2CO3	5	28
12	CsOAc instead of Cs2CO3	62	16

^[a]Reaction conditions: 1 (0.10 mmol), 2 (0.2 mmol), 3 (0.11 mmol), Pd(OAc)₂ (0.01 mmol), L1 (0.025 mmol), N1 (0.10 mmol), Cs₂CO₃ (0.25 mmol), THF (1.0 mL), 100 °C, 12 h.

^[b]Yield was determined by ¹H NMR using dibromomethane as the internal standard. COD, 1,5-cyclooctadiene.

With the optimized reaction conditions in hand, the scope of the ortho amination/ipso Heck reaction was explored (Table 2). Regarding the aryl iodide substrates,^[24] those featuring diverse electronic properties, encompassing electron-donating or electron-withdrawing groups (4b–4t), all furnished desired products in good yield. This method exhibited excellent tolerance toward various functional groups (FGs), including fluoro (4b), bromo (4c, 4g and 4j), chloro (4k), ester (4h and 4p), silyl ether (4i), secondary aniline (4l), tertiary amine (4m), free alcohol (4n), amide (4o), cyclopropane (4p), epoxide (4q), and terminal olefin (4r). In addition, aryl iodides with diverse fused structures, such as 2,3-dihydrobenzofuran (4t), naphthalene (4u), and dibenzofuran (4v), also afforded the desired products in decent yield. Moreover, pyridine-based substrates (4w, 4x and 4y) also worked well. Ph-Davephos^[15,25] worked better for the more electron-rich substrate (4s). Furthermore, the method was amenable to complex substrates derived from natural products and pharmaceuticals (4z, 4aa and 4ab), indicating potential for late-stage modifications. Simple para-substituted aryl iodides gave a di-ortho amination product (4ac) under the current conditions, albeit in lower yield.

Table 2.

Reaction scope of the ortho amination/ipso Heck.^[a]

^[a]Reaction conditions: 1 (0.20 mmol), 2 (0.40 mmol), 3 (0.22 mmol), Pd(OAc)₂ (0.02 mmol), L1 (0.05 mmol), N1 (0.20 mmol), Cs₂CO₃ (0.5 mmol), THF (2.0 mL), 100 °C, 12 h. Isolated yield.

^[b]Ph-Davephos was used instead of L1.

^[c]4-Iodoanisole 1ac (0.20 mmol) was used as the substrate, and 0.8 mmol of 2a was used.

^[d]0.4 mmol of N1 was used. TBS, tert-butyldimethylsilyl; Boc, tert-butyloxycarbonyl.

Besides tert-butyl acrylate, a variety of Michael acceptors, including methyl acrylate (4ad), 2-(trimethylsilyl)ethyl acrylate (4ae) and acrylamides (4af and 4ag), were found to be effective for the ipso alkenylation. The scope of the amine electrophile was also evaluated using various N-benzoyloxy mono-alkyl amines (4ah-4aq). While at this stage only N-tertiary alkylamine-based reagents afforded good yield (vide infra, Scheme 3), a variety of interesting scaffolds derived from carbocycles (4al), heterocycles (4am-4ao), and [2.2.2] bicycles (4ap), along with meaningful FGs (4ah-4ak), can be tolerated. In addition, the ortho amination followed by intramolecular ipso Heck (4aq) can also be realized with the alkene-tethered amine-electrophile. Note that it is not possible to introduce the N-tertiary alkylamino group by the hypothetical two-step procedure: ortho amination with the N-benzyl-N-tertiary alkylamine reagent, followed by benzyl deprotection (Eq. 1), likely due to the steric hindrance of the electrophile.

Scheme 3.

Control experiments. a) Secondary alkyl amine-derived electrophiles can also be used; b) Di-amination and ipso amination have been obtained with the amine electrophile.

Apart from the Heck quench, alkynylation also proved to be feasible for the ipso functionalization (Table 3).^[26] Simple (triisopropylsilyl)acetylene (5a)^[27] was found to be the most effective coupling partner for this reaction, whereas other reagents, such as phenylacetylene, 2-methyl-4-phenyl-3-butyn-2-ol, and phenylpropiolic acid, were not effective.

Table 3.

Ortho amination with ipso alkyne termination.^[a]

^[a]Reaction conditions: 1 (0.20 mmol), 2 (0.40 mmol), 3 (0.22 mmol), Pd(OAc)₂ (0.02 mmol), L1 (0.05 mmol), N1 (0.40 mmol), Cs₂CO₃ (0.5 mmol), THF (2.0 mL), 100 °C, 12 h. Isolated yield. TIPS, triisopropylsilyl.

A series of additional control experiments were carried out to gain insight into the reactivity of the amine electrophile. As shown in Scheme 3a, while the isopropylamine-derived reagent (2e) failed to give any desired ortho amination product, the use of a bulkier secondary alkylamine (2f) afforded the desired product 4ar in 28% yield, accompanied with forming imine side-product 7 in about 40% yield (based on the initial amount of 2f). The reactivity difference between α-tertiary and secondary alkylamine-derived electrophiles and the steric effect observed with different secondary alkylamine-derived electrophiles could be possibly explained by the relative stability of the amine electrophiles and/or aminopalladium intermediates under the reaction conditions. The formation of imine 7 indicates an E2 elimination or a Pd-mediated β-H elimination pathway with secondary alkylamine-derived reagents. Interestingly, in the absence of the nucleophile, 1,2-di-amination product 8 was isolated in 12% yield under the standard condition (Scheme 3b). This is a surprising result because intermolecular ipso amination has not been successful by directly using amines as the nucleophile.^[10d,28] In addition, our further control experiments show that the Pd-catalyzed ipso amination cannot take place with the free amine under the standard condition (in the absence of nucleophile and NBE).^[29] In contrast, replacement of the free amine with the electrophilic 2a led to the formation of the ipso amination product in 22%. Although the exact source of the reductant in the latter reaction remains to be uncovered, these unusual outcomes suggest that the observed ipso amination may not go through the normal Buchwald–Hartwig pathway. This, on the other hand, implies a Pd-catalyzed nitrene-mediated ipso amination,^[30] as it is known that nitrene can be generated from this type of N–O reagents.^[17] Altogether, these experiments suggest that β- and/or α-elimination of N-mono alkylamine reagents can compete with the desired ortho amination pathway.

One benefit of introducing the t-butylamino group by this method is that the t-butyl group can be easily removed under acidic conditions in high yield (Scheme 4a).^[31] As a result, this method offers a convenient approach to access primary anilines, which can be further transformed to other important moieties, such as amides, diarylamines, heterocycles, etc. For example, the primary amine generated from the ortho amination product can cyclize with the enoate group to afford a 2-hydroxyquinoline product (11) upon treatment with hydrochloric acid (Scheme 4b). This transformation can be operated in either two steps or a one-pot manner, which involves t-butyl group removal, alkene isomerization and lactam bond formation. Notably, the previous route to access 2-hydroxyquinoline 11 requires five steps from aniline 12;^[32] here our method offers a single-step approach to prepare this important heterocycle directly from the commercially available aryl iodide (1b). Finally, a unique application of the ortho amination/ipso alkynylation is the rapid preparation of aza-indole products from the corresponding simple aryl iodides (Scheme 4c). The triisopropylsilyl (TIPS) group of the reaction intermediate can be removed in situ by adding tetrabutylammonium fluoride (TBAF), and the resulting terminal alkyne in compound 13 can undergo a gold (III)-catalyzed annulation^[33] to deliver 5-aza-indole 14.

Scheme 4.

Synthetic applications. a) Introducing primary amines via N-deprotection; b) Efficient preparation of functionalized 2-hydroxyquinoline; c) Synthesis of 5-aza-indole. TFA, trifluoroacetic acid.

In summary, beyond forming tertiary anilines, we have developed the first method to directly prepare secondary amines via the Pd/NBE catalysis. Diverse aryl substrates and N-tertiary-alkyl amine-derived electrophiles are identified as suitable coupling partners, leading to versatile synthetic applications, such as primary amino group installation and concise heterocycle formations. While it remains challenging to introduce other types of amino groups at this stage, they can potentially be accessed via the primary amine intermediate. In addition, important mechanistic insights have been obtained, and preliminary success to install a bulky secondary alkyl amino group has been achieved. Efforts on new reagent/catalyst design to further expand the scope of the nitrogen-containing moieties that can be introduced, as well as deeper mechanistic understanding of these transformations, are ongoing.