β C–H di-halogenation via iterative hydrogen atom transfer† †Electronic supplementary information (ESI) available. CCDC 1581032. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c8sc01214h

A radical relay strategy for mono- and di-halogenation (iodination, bromination, and chlorination) of sp3 C–H bonds has been developed. This first example of double, geminal C–H functionalization is enabled via iterative, hydrogen atom transfer (HAT) by in situ generated imidate radicals.


Introduction
The halogenation of an sp 3 C-H bond 1 enables direct conversion of an inert motif into a versatile synthetic handle that permits broad reactivity via cross-coupling and substitution. 2 Generally, C-H halogenation occurs by radical-mediated 3 or organometallic 4 mechanisms. Each approach exhibits complementary reactivity and selectivityespecially for incorporation of the most versatile halide: an iodide (Fig. 1a). In the realm of metal-mediated sp 3 C-H iodination, there are just a few methods that can install this reactive handle; they are stoichiometric 5 or catalytic 6 in Pd. In the latter cases, only Yu and Rao have reported directed sp 3 C-H iodinationemploying oxazolines, amides, or oximes as directing groups (Fig. 1b). 6 These Pd-catalyzed methods exclusively effect primary C-H conversion to a terminal mono-iodide, which is deactivated to further reactivity. In this mechanism, a second iodination at a distal, primary C-H affords a 1,3-di-iodide. 7 Alternatively, radical mechanisms can promote efficient iodination of various types of sp 3 C-H bonds via hydrogen atom transfer (HAT). 8 Moreover, intramolecular HAT provides unique, d selective C-H functionalizations. 9 Yet, non-directed methods 10 surpass the few, pioneering examples of d (or g) C-H halogenation. 11 Notably, a directed C-H iodination has yet to be developed, despite the key intermediacy of a distal iodide in several d C-H aminations (or etherication) mediated by 1,5-HAT. 12 Due to the penchant for iodide displacement, intercepting this alkyl iodide intermediate is challenging. As an alternate strategy, we proposed a cascade mechanism -involving abstraction of the adjacent, a-iodo C-Hmight enable geminal C-H di-iodination (Fig. 1c).
We noted that Suárez observed a minor di-iodide byproduct upon intramolecular d amination of 8-membered lactams. 13 Fig. 1 Directed, mono-and di-iodination of sp 3 C-H bonds.
Benzylic tri-iodination mechanisms have also been proposed, 14 but no method yet exists to isolate them.
Given the limited synthetic accessibility (and potential pharmacological value 15 ) of gem-di-iodidesan important, versatile motif (previously only accessible from hydrazones or vinyl iodides) 16we sought to design a strategy to harness a directed, iterative HAT mechanism to introduce geminal dihalides at remote carbons. Notably, this new type of double C-H iodination at a single carbon atom is complementary to Pdcatalyzed methods and uniquely possible via a radical mechanism (Fig. 1).
To develop a versatile b C-H di-iodination via iterative, intramolecular HAT and sequential iodination, we chose to employ imidates as readily accessible, radical relay precursors (Fig. 1c). In our proposed di-iodination mechanism, we envisioned that in situ formation of a weak imidate sp 2 N-I bond would enable its rapid homolysis by visible light. Selective translocation of the ensuing N-centered radical to a b Cc can occur via thermodynamically favored 1,5-HAT. Finally, either radical recombination with Ic (derived from the initial N-I homolysis), or homolytic substitution by I 2 (or N-I), can afford a reactive b iodide. However, we were cognizant of two major challenges ( Fig. 1d) for trapping the d iodide intermediate of HAT mechanisms, including its reactivity: (1) as a leaving group, and (2) towards further oxidative decomposition.
Whereas, we previously observed weak C-H bonds (e.g. benzyl, allyl) provide activated iodides that are rapidly displaced (in a formal C-H amination), 17 secondary (2 ) C-H bonds yield complete decomposition. Given our knowledge that I 3 À efficiently mediates HAT of 2 C-H bonds, 18 we hypothesized a b iodide intermediate is formed, yet is prone to further Ioxidation. In this case, decomposition may ensue from the resulting sp 3 hypervalent iodide, which is an excellent nucleofuge for elimination or cyclization. 19 Instead, to enable access to gem-di-iodides, we proposed an alternate N-selective oxidation may promote a second HAT of the slightly weaker b C-H (103 vs. 105 kcal mol À1 ). 20 Importantly, however, this iterative HAT mechanism for directed, di-functionalization is only possible if N-oxidation is more rapid than the previously observed, Ioxidation pathway.

Results and discussion
To our delight, adaptation of our radical relay strategy allowed us to intercept the 2 b iodide intermediate for the rst time to access both mono-and di-b C-H iodides. The key factors that enabled discovery of these new reactions included judicious choice of oxidant, increased reaction concentration, and shorter reaction durationall essential to limit product decomposition. Notably, NIS oxidant was found to favor b mono-iodide 1 formation, while a combination of NaI and PhI(OAc) 2 provides desired b di-iodide 2-17. For the latter, a strong solvent effect was also observed, wherein greater solubility of NaI (in HFIP or CH 2 Cl 2 ) affords less product (3, <30%), while more polar, but less solubilizing MeCN affords a higher yield of b di-iodide 3 (58%). Ultimately, a 3 : 1 mixture of CH 2 Cl 2 : MeCN was found to provide the gem-di-iodide most efficiently (3, 88%, 83% isolated yield) (see ESI † for full details of optimization). Having developed the rst method for b C-H di-iodination, we next investigated the generality of this radical-mediated transformation with a variety of imidatesderived from baseinduced addition of alcohols into Cl 3 C-CN. In all cases, we observed efficient formation of b di-iodides with greater than 20 : 1 regioselectivity (Table 1).
Except for the NIS-based conditions that afford mono-iodide 1, di-iodide is always the major product, typically isolated in high yields (2)(3). Interestingly, this reaction is tolerant of steric congestion (4-5) and remains b selective even in the presence of weaker C-H bonds adjacent to arenes, halides, ethers, esters, and amides at the g or d positions (6-11). Secondary alcohols are also amenable to this di-iodination with selectivity observed for secondary over primary C-H bonds (12)in contrast to Pdmediated pathways. 6 While acyclic 2 alcohols efficiently yield di-iodide (13), cyclic alcohols afford a 2 : 1 mixture of di-and mono-iodide (14)illustrating conformational constraints for the HAT mechanism. Similarly, an estradiol-derived imidate affords a 1 : 1 mixture of mono-and di-iodide (15). Imidates derived from cholic acid and amino acid, valine, yield gem-diiodides (16-17) efficiently.
Cognizant of the synthetic utility of gem-di-halides, we sought to extend this unique di-iodination mechanism to other halides. To this end, we found that the use of NaBr or NaCl (instead of NaI) affords analogous b halogenation ( Table 2). These new transformations require slight deviation from standard reaction conditions since NaBr and NaCl are less soluble. In these cases, increased halide concentration via phase transfer catalysts (Bu 4 N + X À ) and a more solubilizing solvent mixture (3 : 1 HFIP : CH 2 Cl 2 ) are the key factors that enable these new reactions.
Notably, a stronger N-Cl intermediate requires UV light (300 nm) for initiation of the radical relay. It is also noteworthy that C-H chlorination ceases aer the rst halogenation despite a relative similarity in the a-Cl and a-Br C-H bond strengths (AE1 kcal). 21 The scope is as general as the iodination, with three representative examples shown for each halide (18)(19)(20)(21)(22)(23). X-ray crystallographic analysis of di-bromide 18 conrms the structure of these distal geminal halides.
Interested in further understanding this exceptionally efficient sequential di-iodination (which provides orthogonal reactivity and selectivity to Pd catalysis), we sought to explore our hypothesis that the weaker a-iodo C-H bond enables this transformation. First, a kinetic study by 1 H NMR illustrates a rapid conversion of the mono-iodide intermediate to the diiodide product (Fig. 2). Aer an initial induction period (ca. 10 min), mono-iodide 24 is formed in $30% yield, before rapid conversion to di-iodide 2.
In separate experiments, initial rates of formation of monoiodide 24 and di-iodide 2 were independently measured from their respective starting materials (Fig. 3a), using 1 equiv. of oxidant, for more accurate measurements. A relative rate of 2.2 was observed in the second iodination, supporting the expectation it is more rapid than the rst due to a weaker C-H bond. In the course of our studies, we were also interested in comparing the relative rates of reactivity among the various halides. To this end, we performed competition experiments between NaI & NaBr/NaCl (Fig. 3b). In the I/Br competition, a statistical mixture of products is formed (1 : 1 : 2 di-iodide 4 : di-bromide 18 : mixed 25)suggesting both reaction rates are comparable. On the other hand, an I/Cl competition provides greater selectivity. Only mono-and di-iodide products (4) are observed with visible light irradiation (since chlorination requires UV light); yet UV irradiation (which unproductively consumes iodinated species) exclusively affords chlorination (19). Lastly, we exploited the difference in halide reactivity to enable a synthetically useful, iterative C-H halogenation (Fig. 3c). In the sequence, mono C-H chlorination (26) and subsequent C-H iodination affords b geminal halide 27 that contains two different halides (Cl, I).
Equipped with the rst method to access b gem-di-halides via C-H functionalization, we sought to elucidate the synthetic utility of these versatile handles. Fig. 4 illustrates ve post-

Conclusions
In summary, a radical relay strategy has enabled the one-step conversion of imidates to mono-or di-halides via iterative b C-H halogenation. In particular, synthetic access to the versatile, geminal di-halides is uniquely facilitated by an imidate radical-based 1,5-HAT mechanism. By developing a new strategy to bypass oxidative decomposition pathways, reactive alkyl halide intermediates of a radical relay reaction mechanism were intercepted. Along with new methods for mono-and di-C-H halogenation (X ¼ I, Br, Cl), competitive rates and kinetic proles have also been investigated. Finally, the versatility of the b di-iodides is showcased in the synthesis of functionally rich moleculesuniquely enabled by an HATbased b C-H functionalization mechanism.

Conflicts of interest
There are no conicts to declare.