KOtBu as a Single Electron Donor? Revisiting the Halogenation of Alkanes with CBr4 and CCl4

The search for reactions where KOtBu and other tert-alkoxides might behave as single electron donors led us to explore their reactions with tetrahalomethanes, CX4, in the presence of adamantane. We recently reported the halogenation of adamantane under these conditions. These reactions appeared to mirror the analogous known reaction of NaOH with CBr4 under phase-transfer conditions, where initiation features single electron transfer from a hydroxide ion to CBr4. We now report evidence from experimental and computational studies that KOtBu and other alkoxide reagents do not go through an analogous electron transfer. Rather, the alkoxides form hypohalites upon reacting with CBr4 or CCl4, and homolytic decomposition of appropriate hypohalites initiates the halogenation of adamantane.


Introduction
Recent reports have described transition-metal-free reactions, such as the coupling of haloarenes with arenes [1][2][3][4], that are promoted by a combination of KO t Bu and an organic additive. The reactions proceed by a radical mechanism, with a single electron transfer (SET) step initiating the formation of the radicals. It has previously been reported that KO t Bu is capable of reductively activating a range of substrates via single electron transfer [5,6], and this led some authors to propose that tert-butoxide anion is responsible for the initiation of these coupling reactions, either alone or in complexation with an additive, such as 1,10-phenanthroline, in the ground state [2,3,[7][8][9][10][11]. However, recent publications have shown that the KO t Bu base reacts with the organic additive to form an electron donor in situ. SET from this electron donor then initiates the radical chain mechanism, not SET from tert-butoxide anion [12][13][14]. The investigations begged the question: "under what circumstances would KO t Bu behave as a single electron donor?" To answer that question, we were drawn to the research of Schreiner and Fokin et al., who reduced CBr 4 using aqueous NaOH and a phase-transfer catalyst in a two-phase system that led to the bromination of adamantane [15,16]. They demonstrated that reaction of hydroxide anion with CBr 4 led to the tribromomethyl radical 4, and concluded that this reflected direct electron transfer from hydroxide to CBr 4 (Scheme 1). This tribromomethyl radical 4 undergoes a chain reaction, where it abstracts a hydrogen atom from adamantane 5 to form the alkyl radical 7 and bromoform 6. The adamantyl radical 7 abstracts a bromine atom from CBr 4 , forming adamantyl bromide 8 and regenerating the tribromomethyl radical 4 [15]. The reactions of the tribromomethyl radical show diagnostic high selectivities for abstracting tertiary CH hydrogen atoms (to ultimately form 1-bromoadamantane) over secondary (CH 2 ) counterparts (which would ultimately form 2-bromoadamantane).

Results
The analogy of the proposed reaction of KO t Bu with CBr 4 or CCl 4 (Scheme 2) to the Schreiner-Fokin reaction of NaOH with this halide is striking (Scheme 1). Upon donation of an electron, the alkoxide anion 9 would form the corresponding alkoxyl radical 10, which can undergo either hydrogen atom transfer (HAT) to form 11, or β-scission to form the derived ketone 12 (for appropriate R). Either the alkoxyl radical 10 or the methyl radical can abstract a hydrogen atom from adamantane; the adamantyl radical would then propagate the reaction as in Scheme 1. In order to detect the product from the β-scission of the alkoxyl radical 10, and bearing in mind the volatility of acetone from β-scission of tert-butoxyl radicals, we synthesized potassium 2-phenylpropan-2-olate 14 and used it as a base to explore a range of reaction conditions (Table 1).
Recently, we showed that KO t Bu reacts with CBr4 and adamantane (40 °C, 96 h) to afford bromoadamantane-in a reaction that appears to mirror the work of Schreiner and Fokin et al.-with NaOH, although phase-transfer conditions were not used [17]. The oxidation potential of KO t Bu in DMF (0.10 V vs. SCE) [8] is not too different from the reduction potential of CBr4 in DMF (−0.31 V vs. SCE) [18]. It was therefore proposed that KO t Bu might reduce CBr4 through an electron transfer mechanism, but the door was left open to further investigation. The aim of this paper is to explore the chemistry of KO t Bu and related tertiary alkoxides in this context. Scheme 1. The Schreiner-Fokin mechanism for the bromination of alkanes from reaction of hydroxide with carbon tetrabromide [1].

Results
The analogy of the proposed reaction of KO t Bu with CBr4 or CCl4 (Scheme 2) to the Schreiner-Fokin reaction of NaOH with this halide is striking (Scheme 1). Upon donation of an electron, the alkoxide anion 9 would form the corresponding alkoxyl radical 10, which can undergo either hydrogen atom transfer (HAT) to form 11, or β-scission to form the derived ketone 12 (for appropriate R). Either the alkoxyl radical 10 or the methyl radical can abstract a hydrogen atom from adamantane; the adamantyl radical would then propagate the reaction as in Scheme 1. In order to detect the product from the β-scission of the alkoxyl radical 10, and bearing in mind the volatility of acetone from β-scission of tert-butoxyl radicals, we synthesized potassium 2-phenylpropan-2-olate 14 and used it as a base to explore a range of reaction conditions (Table 1).
In the blank reaction (without CBr 4 ) ( Table 1, entry 3), bis((2-phenylpropan-2-yl)oxy)methane 20 is formed by the nucleophilic displacement of chloride from dichloromethane by two molecules of potassium 2-phenylpropan-2-olate 14 (Scheme 4) [23,24]. The fact that bis((2-phenylpropan-2yl)oxy)methane 20 is only observed in the absence of CBr 4 suggests that the reaction of potassium 2-phenylpropan-2-olate 14 with dichloromethane is slower than the reaction with CBr 4 .  Reflecting on the fact that potassium 2-phenylpropan-2-olate 14 did not give rise to halogenation of adamantane (Table 1), while halogenation was observed when KO t Bu was used as the alkoxide (the ratio of adamantane 18:1-bromoadamantane 31:2-bromoadamantane 51 was determined to be 39:3.3:1 [17]), this suggests that either the two alkoxide bases do not react with CBr4 through analogous mechanisms, or that the halogenation reaction was intercepted and inhibited when 14 was used. Interception could occur if methylstyrene 17-formed in the reaction from potassium 2phenylpropan-2-olate 14-shuts down the radical chain pathway. To probe this possibility, the reaction of KO t Bu with CBr4 and adamantane 18 was repeated in the presence of methylstyrene 17 (1 equiv., Scheme 5).
The addition of methylstyrene 17 completely inhibited the bromination of adamantane, which suggests that it can shut down the radical mechanism, perhaps by acting as a preferential source of hydrogen atoms, relative to adamantane 18, for the radical intermediates in the reaction. Computationally, the competition between methylstyrene and adamantane as a source of H atoms was modeled using CBr3 radicals, which showed that hydrogen atom abstraction by the CBr3 radical from methylstyrene 17 (ΔG ǂ = 10.7 kcal/mol and ΔGrxn = −7.7 kcal/mol) is thermodynamically and  Reflecting on the fact that potassium 2-phenylpropan-2-olate 14 did not give rise to halogenation of adamantane (Table 1), while halogenation was observed when KO t Bu was used as the alkoxide (the ratio of adamantane 18:1-bromoadamantane 31:2-bromoadamantane 51 was determined to be 39:3.3:1 [17]), this suggests that either the two alkoxide bases do not react with CBr4 through analogous mechanisms, or that the halogenation reaction was intercepted and inhibited when 14 was used. Interception could occur if methylstyrene 17-formed in the reaction from potassium 2phenylpropan-2-olate 14-shuts down the radical chain pathway. To probe this possibility, the reaction of KO t Bu with CBr4 and adamantane 18 was repeated in the presence of methylstyrene 17 (1 equiv., Scheme 5).
The addition of methylstyrene 17 completely inhibited the bromination of adamantane, which suggests that it can shut down the radical mechanism, perhaps by acting as a preferential source of hydrogen atoms, relative to adamantane 18, for the radical intermediates in the reaction. Computationally, the competition between methylstyrene and adamantane as a source of H atoms was modeled using CBr3 radicals, which showed that hydrogen atom abstraction by the CBr3 radical from methylstyrene 17 (ΔG ǂ = 10.7 kcal/mol and ΔGrxn = −7.7 kcal/mol) is thermodynamically and Reflecting on the fact that potassium 2-phenylpropan-2-olate 14 did not give rise to halogenation of adamantane (Table 1), while halogenation was observed when KO t Bu was used as the alkoxide (the ratio of adamantane 18:1-bromoadamantane 31:2-bromoadamantane 51 was determined to be 39:3.3:1 [17]), this suggests that either the two alkoxide bases do not react with CBr 4 through analogous mechanisms, or that the halogenation reaction was intercepted and inhibited when 14 was used. Interception could occur if methylstyrene 17-formed in the reaction from potassium 2-phenylpropan-2-olate 14-shuts down the radical chain pathway. To probe this possibility, the reaction of KO t Bu with CBr 4 and adamantane 18 was repeated in the presence of methylstyrene 17 (1 equiv., Scheme 5).
The addition of methylstyrene 17 completely inhibited the bromination of adamantane, which suggests that it can shut down the radical mechanism, perhaps by acting as a preferential source of hydrogen atoms, relative to adamantane 18, for the radical intermediates in the reaction. Computationally, the competition between methylstyrene and adamantane as a source of H atoms was modeled using CBr 3 radicals, which showed that hydrogen atom abstraction by the CBr 3 radical from methylstyrene 17 (∆G } = 10.7 kcal/mol and ∆G rxn = −7.7 kcal/mol) is thermodynamically and kinetically more favorable than from adamantane 18 (∆G } = 13.9 kcal/mol and ∆G rxn = 1.1 kcal/mol) (see Supplementary Materials).
All yields were calculated using 1,3,5-trimethoxybenzene as the internal standard (10 mol %) in the 1 H-NMR. The yield of 18 was calculated based on recovery of adamantane 18, the yield of 16 and 19 was calculated based on CBr4. (Note that alkoxide (4 eq.) is capable of forming methylstyrene 17.) With the knowledge that methylstyrene 17 is capable of preventing bromination of adamantane, why does bromination succeed when KO t Bu 29 + CBr4 alone are used? The two hypobromites, tertbutyl hypobromite and 21, may undergo elimination at different rates; additionally, 2-methylpropene (b.p. −6.9 °C) exists as a gas at the reaction temperature. Such a volatile product would likely be found principally in the headspace above the reaction, rather than in solution, and therefore halogenation of adamantane 18 could occur.
As mentioned earlier, Wirth et al. [19] used tert-butyl hypobromite to achieve the bromination of alkanes (while Walling [20] used tert-butyl hypochlorite to achieve chlorination) and proposed that the mechanism was initiated via homolysis of the O-Br bond of tert-butyl hypobromite. Either the bromine radical or the tert-butoxyl radical could abstract the H atom from adamantane. Propagation could then occur when the adamantyl radical abstracts a Br atom from CBr4 [15] or from tert-butyl hypobromite [23].
To gain further information on mechanism, an alternative alkoxide, potassium triphenylmethanolate 30, was prepared and subjected to the reaction conditions with CBr4 and adamantane 18 (Table 2). Reaction of potassium triphenylmethanolate 30 with adamantane in the presence of CBr4 afforded products 1-bromoadamantane 31 (7%) and triphenylmethanol 32 (88%), as well as two additional products, benzophenone 33 (5%) and 4-benzhydrylphenol 34 (12%) ( Table 2, entry 1). When CBr4 was not present in the reaction mixture, triphenylmethanol 32 (81%) was isolated following workup, together with a trace (1%) of benzophenone 33 ( Table 2, entry 2). Further analysis of the starting material showed trace amounts of 33 present in the commercially supplied With the knowledge that methylstyrene 17 is capable of preventing bromination of adamantane, why does bromination succeed when KO t Bu 29 + CBr 4 alone are used? The two hypobromites, tert-butyl hypobromite and 21, may undergo elimination at different rates; additionally, 2-methylpropene (b.p. −6.9 • C) exists as a gas at the reaction temperature. Such a volatile product would likely be found principally in the headspace above the reaction, rather than in solution, and therefore halogenation of adamantane 18 could occur.
As mentioned earlier, Wirth et al. [19] used tert-butyl hypobromite to achieve the bromination of alkanes (while Walling [20] used tert-butyl hypochlorite to achieve chlorination) and proposed that the mechanism was initiated via homolysis of the O-Br bond of tert-butyl hypobromite. Either the bromine radical or the tert-butoxyl radical could abstract the H atom from adamantane. Propagation could then occur when the adamantyl radical abstracts a Br atom from CBr 4 [13] or from tert-butyl hypobromite [23].
To gain further information on mechanism, an alternative alkoxide, potassium triphenylmethanolate 30, was prepared and subjected to the reaction conditions with CBr 4 and adamantane 18 (Table 2). Scheme 5. Using methylstyrene 17 to block the bromination of adamantane.
All yields were calculated using 1,3,5-trimethoxybenzene as the internal standard (10 mol %) in the 1 H-NMR. The yield of 18 was calculated based on recovery of adamantane 18, the yield of 16 and 19 was calculated based on CBr4. (Note that alkoxide (4 eq.) is capable of forming methylstyrene 17.) With the knowledge that methylstyrene 17 is capable of preventing bromination of adamantane, why does bromination succeed when KO t Bu 29 + CBr4 alone are used? The two hypobromites, tertbutyl hypobromite and 21, may undergo elimination at different rates; additionally, 2-methylpropene (b.p. −6.9 °C) exists as a gas at the reaction temperature. Such a volatile product would likely be found principally in the headspace above the reaction, rather than in solution, and therefore halogenation of adamantane 18 could occur.
As mentioned earlier, Wirth et al. [19] used tert-butyl hypobromite to achieve the bromination of alkanes (while Walling [20] used tert-butyl hypochlorite to achieve chlorination) and proposed that the mechanism was initiated via homolysis of the O-Br bond of tert-butyl hypobromite. Either the bromine radical or the tert-butoxyl radical could abstract the H atom from adamantane. Propagation could then occur when the adamantyl radical abstracts a Br atom from CBr4 [15] or from tert-butyl hypobromite [23].
To gain further information on mechanism, an alternative alkoxide, potassium triphenylmethanolate 30, was prepared and subjected to the reaction conditions with CBr4 and adamantane 18 (Table 2). Reaction of potassium triphenylmethanolate 30 with adamantane in the presence of CBr4 afforded products 1-bromoadamantane 31 (7%) and triphenylmethanol 32 (88%), as well as two additional products, benzophenone 33 (5%) and 4-benzhydrylphenol 34 (12%) ( Table 2, entry 1). When CBr4 was not present in the reaction mixture, triphenylmethanol 32 (81%) was isolated following workup, together with a trace (1%) of benzophenone 33 ( Table 2,  Reaction of potassium triphenylmethanolate 30 with adamantane in the presence of CBr 4 afforded products 1-bromoadamantane 31 (7%) and triphenylmethanol 32 (88%), as well as two additional products, benzophenone 33 (5%) and 4-benzhydrylphenol 34 (12%) ( Table 2, entry 1). When CBr 4 was not present in the reaction mixture, triphenylmethanol 32 (81%) was isolated following workup, together with a trace (1%) of benzophenone 33 (Table 2,  Background formation of benzophenone 33, in trace amounts from heterolytic fragmentation of similar tertiary alkoxides with expulsion of a phenyl anion, is also precedented [25,26]. The important difference between the reaction in the presence of CBr 4 and in its absence ( Table 2, entry 1 and entry 2, respectively) is the formation of 4-benzhydrylphenol 34, which can arise through the hypobromite 35 (Scheme 6). Potassium triphenylmethanolate 30 reacts with CBr 4 to generate the hypobromite 35. This hypobromite can react by three pathways (1) the OBr anion leaves to form the stabilized cation 36; (2) the O-Br bond fragments ionically with simultaneous migration of a phenyl moiety to form cation 38; or (3) the O-Br bond undergoes homolysis to form alkoxyl radical 40 and a bromine radical, for which there is literature precedent [23]. If pathway (1) is followed, the carbocation 36 is attacked by another molecule of potassium triphenylmethanolate 30 and, due to steric effects, the alkoxide attacks at the para position of one of the benzene rings. In doing so, the species 37 is formed, which, following tautomerism and hydrolytic workup, leads to 34. Alternatively, 36 affords triphenylmethanol 32 on workup. Pathways (2) and (3)  Molecules 2018, 23, x FOR PEER REVIEW 6 of 16 triphenylmethanol 32. Background formation of benzophenone 33, in trace amounts from heterolytic fragmentation of similar tertiary alkoxides with expulsion of a phenyl anion, is also precedented [25,26]. The important difference between the reaction in the presence of CBr4 and in its absence ( Table 2, entry 1 and entry 2, respectively) is the formation of 4-benzhydrylphenol 34, which can arise through the hypobromite 35 (Scheme 6). Potassium triphenylmethanolate 30 reacts with CBr4 to generate the hypobromite 35. This hypobromite can react by three pathways (1) the OBr anion leaves to form the stabilized cation 36; (2) the O-Br bond fragments ionically with simultaneous migration of a phenyl moiety to form cation 38; or (3) the O-Br bond undergoes homolysis to form alkoxyl radical 40 and a bromine radical, for which there is literature precedent [23]. If pathway (1) is followed, the carbocation 36 is attacked by another molecule of potassium triphenylmethanolate 30 and, due to steric effects, the alkoxide attacks at the para position of one of the benzene rings. In doing so, the species 37 is formed, which, following tautomerism and hydrolytic workup, leads to 34. Alternatively, 36 affords triphenylmethanol 32 on workup. Pathways (2) and (3)   The study above provides evidence for significant formation of hypohalites from alkoxides 14 and 30. However, it might be possible for the O t Bu anion in KO t Bu to behave differently. The absence of electron-withdrawing aryl groups would render it more electron-rich, and, therefore, it might be a better candidate than the other alkoxides to undergo electron transfer. of electron-withdrawing aryl groups would render it more electron-rich, and, therefore, it might be a better candidate than the other alkoxides to undergo electron transfer.
To probe the ability of alkoxides to donate a single electron, computational modeling was implemented to calculate the energy profile for the SET from both KO t Bu and potassium 2-phenylpropan-2-olate 14 to CBr 4 in dichloromethane ( Figure 1) [33]. Our previous calculations had been based on the classical Nelsen four-point method [34], but, since that time, we published a more accurate complexation method [33] for predicting the activation free energy and the relative free energy of reactions. (The calculations were conducted using the M06-2X functional [35,36] with the 6-311++G(d,p) basis set [37][38][39][40][41] on all atoms, except for the bromine. Bromine was modeled with the MWB28 relativistic pseudo-potential and associated basis set [42]. All calculations were carried out using the C-PCM implicit solvent model [43,44] with the dielectric constant for dichloromethane (ε = 8.93) or carbon tetrachloride (ε = 2.228) as appropriate. All calculations were performed in Gaussian09 [45]). energy of reactions. (The calculations were conducted using the M06-2X functional [35,36] with the 6-311++G(d,p) basis set [37][38][39][40][41] on all atoms, except for the bromine. Bromine was modeled with the MWB28 relativistic pseudo-potential and associated basis set [42]. All calculations were carried out using the C-PCM implicit solvent model [43,44] with the dielectric constant for dichloromethane (ε = 8.93) or carbon tetrachloride (ε = 2.228) as appropriate. All calculations were performed in Gaussian09 [45]).
The energy barriers for SET from either KO t Bu or potassium 2-phenylpropan-2-olate 14, to a molecule of CBr4 were calculated to be ΔG ǂ = +35.4 kcal/mol and +36.1 kcal/mol, respectively (while the corresponding ΔGrxn values were +13.9 and +18.9 kcal/mol) (Figure 1). When the energy barrier was calculated for the reactions with CCl4, a similar energy profile was obtained for the SET from either of the two alkoxides to a molecule of CCl4 (see Supplementary Materials). The energy barriers calculated were ΔG ǂ = 42.5 kcal/mol and 44.5 kcal/mol for SET to CCl4 from KO t Bu and potassium 2phenylpropan-2-olate 14, respectively. These barriers for reactions of both CBr4 and CCl4 are not accessible for a reaction performed at 40 °C, even as an initiation step. These computationally derived energy profiles for the SET step indicate that the initiation for this halogenation of adamantane 18 is not via SET from the alkoxide; the likely alternative radical pathway arises through homolysis of the hypohalite intermediate [19]. (Importantly, our computation predicts a barrierless, but endergonic, profile for homolysis of the O-Br bond in t BuO-Br with ΔGrxn = +27.6 kcal/mol [34]. This is a highly endergonic reaction, but as it is simply an initiation step, few t BuO-Br molecules are required to undergo homolysis).  The energy barriers for SET from either KO t Bu or potassium 2-phenylpropan-2-olate 14, to a molecule of CBr 4 were calculated to be ∆G } = +35.4 kcal/mol and +36.1 kcal/mol, respectively (while the corresponding ∆G rxn values were +13.9 and +18.9 kcal/mol) (Figure 1). When the energy barrier was calculated for the reactions with CCl 4 , a similar energy profile was obtained for the SET from either of the two alkoxides to a molecule of CCl 4 (see Supplementary Materials). The energy barriers calculated were ∆G } = 42.5 kcal/mol and 44.5 kcal/mol for SET to CCl 4 from KO t Bu and potassium 2-phenylpropan-2-olate 14, respectively. These barriers for reactions of both CBr4 and CCl 4 are not accessible for a reaction performed at 40 • C, even as an initiation step. These computationally derived energy profiles for the SET step indicate that the initiation for this halogenation of adamantane 18 is not via SET from the alkoxide; the likely alternative radical pathway arises through homolysis of the hypohalite intermediate [19]. (Importantly, our computation predicts a barrierless, but endergonic, profile for homolysis of the O-Br bond in t BuO-Br with ∆G rxn = +27.6 kcal/mol [34]. This is a highly endergonic reaction, but as it is simply an initiation step, few t BuO-Br molecules are required to undergo homolysis).

General Experimental Information
All reagents were bought from commercial suppliers and used without further purification unless stated otherwise. All the reactions were carried out under argon atmosphere. Diethyl ether, tetrahydrofuran, dichloromethane and hexane were dried with a Pure-Solv 400 solvent purification system, marketed by Innovative Technology Inc. (Herndon, VA, USA). Organic extracts were, in general, dried over anhydrous sodium sulfate (Na 2 SO 4 ). A Büchi rotary evaporator was used to concentrate the reaction mixtures. Thin layer chromatography (TLC) was performed using aluminium-backed sheets of silica gel and visualized under a UV lamp (254 nm). The plates were developed using phosphomolybdic acid or KMnO 4 solution. Column chromatography was performed to purify compounds by using silica gel 60 (200-400 mesh).
The electron transfer reactions were carried out within a glove box (Innovative Technology Inc.) under nitrogen atmosphere, and performed in an oven-dried or flame-dried apparatus using anhydrous solvents, which were degassed under reduced pressure, then purged with argon and dried over activated molecular sieves (3 Å), prior to being sealed and transferred to the glovebox. All solvent or samples placed into the glovebox were transferred through the port, which was evacuated and purged with nitrogen 10 times before entry. When the reaction mixtures were prepared, the reaction vessel was removed from the glove box and the rest of the reaction was performed in a fumehood.
Proton ( 1 H) NMR spectra were recorded at 400.13, 400.03 and 500.16 MHz on Bruker AV3, AV400 and AV500 spectrometers, respectively. Carbon ( 13 C) NMR spectra were recorded using broadband decoupled mode at 100.61, 100.59 and 125.75 MHz on Bruker AV3, AV400 and AV500 spectrometers, respectively. Spectra were recorded in either deuterated chloroform (CDCl 3 ) or deuterated dimethyl sulfoxide (d 6 -DMSO), depending on the solubility of the compounds. The chemical shifts are reported in parts per million (ppm), calibrated on the residual non-deuterated solvent signal, and the coupling constants, J, are reported in Hertz (Hz). The peak multiplicities are denoted using the following abbreviations: s, singlet; d, doublet; t, triplet; q, quartet; sx, sextet; m, multiplet; br s, broad singlet; dd, doublet of doublets; dt, doublet of triplets; td, triplet of doublets.
Infrared spectra were recorded on an ATR-IR spectrometer. Melting points were determined on a Gallenkamp melting point apparatus. The mass spectra were recorded by either gas-phase chromatography (GCMS) or liquid-phase chromatography (LCMS), using various ionization techniques, as stated for each compound: atmospheric pressure chemical ionization (APCI), electron ionization (EI), electrospray ionization (ESI). GCMS data were recorded using an Agilent Technologies 7890A GC system coupled to a 5975C inert XL EI/CI MSD detector. Separation was performed using the DB5MS-UI column (30 m × 0.25 mm × 0.25 µm) at a temperature of 320 • C, using helium as the carrier gas. LCMS data were recorded using an Agilent 6130 Dual source mass spectrometer with Agilent 1200, Agilent Poroshell 120ECC 4.6 mm × 75 mm × 2.7 um column. High-resolution mass spectrometry (HRMS) was performed at the University of Wales, Swansea, in the EPSRC National Mass Spectrometry Centre. Accurate mass was obtained using atmospheric pressure chemical ionization (APCI), chemical ionization (CI), electron ionization (EI), electrospray ionization (ESI) or nanospray ionization (NSI) with a LTQ Orbitrap XL mass spectrometer.

Conclusions
This study has led to a revision of earlier thoughts on the mechanism for the halogenation of adamantane 18 by using a combination of KO t Bu and CBr 4 . It is proposed that the mechanism does not occur through SET, as was previously believed, but proceeds through hypobromite intermediates (Scheme 7). The alkoxide in the reaction mixture, 53, forms a hypobromite in the presence of CBr 4 . The hypobromite 54 can undergo reaction along two pathways. The first option is an elimination reaction to form the alkene 57, which reacts with carbenes formed in the reaction to afford final product 58. The second pathway is the O-Br bond homolysis of the hypobromite 54. This forms the alkoxyl radical 55 and a bromine radical. These radical intermediates perform a hydrogen atom abstraction from adamantane 18 to form adamantyl radical 56 (alternatively, the hydrogen atom abstraction may occur at the C-2 position to ultimately give the 2-Br isomer). The radical 56 will abstract a bromine from a hypobromite molecule 54, or from CBr 4 , to form 31 and an alkoxyl radical 55, or CBr 3 radical. The radical formed will propagate the chain pathway by hydrogen atom abstraction from adamantane 18, thus creating a radical chain mechanism. Thus, it appears that the search for reactions where ground-state KO t Bu behaves as an electron donor must continue. The outcomes of this project have led us to reexamine the remaining claims [3,4] of this phenomenon. 147.0 triphenylmethanol 32 [49]. These signals are consistent with the literature values and reference samples.

Conclusions
This study has led to a revision of earlier thoughts on the mechanism for the halogenation of adamantane 18 by using a combination of KO t Bu and CBr4. It is proposed that the mechanism does not occur through SET, as was previously believed, but proceeds through hypobromite intermediates (Scheme 7). The alkoxide in the reaction mixture, 53, forms a hypobromite in the presence of CBr4. The hypobromite 54 can undergo reaction along two pathways. The first option is an elimination reaction to form the alkene 57, which reacts with carbenes formed in the reaction to afford final product 58. The second pathway is the O-Br bond homolysis of the hypobromite 54. This forms the alkoxyl radical 55 and a bromine radical. These radical intermediates perform a hydrogen atom abstraction from adamantane 18 to form adamantyl radical 56 (alternatively, the hydrogen atom abstraction may occur at the C-2 position to ultimately give the 2-Br isomer). The radical 56 will abstract a bromine from a hypobromite molecule 54, or from CBr4, to form 31 and an alkoxyl radical 55, or CBr3 radical. The radical formed will propagate the chain pathway by hydrogen atom abstraction from adamantane 18, thus creating a radical chain mechanism. Thus, it appears that the search for reactions where ground-state KO t Bu behaves as an electron donor must continue. The outcomes of this project have led us to reexamine the remaining claims [5,6] of this phenomenon. Scheme 7. The modified mechanism for halogenation of adamantane.
Supplementary Materials: The following are available online, (i)Additional computational analysis and xyz coordinates relating to all computed structures (ii) Table S1 "The reaction of tert-butyl hypochlorite 46 in dichloromethane or carbon tetrachloride as solvent" (iii) Table S2. tert-Butyl hypobromite 50 in bromination of adamantane 18; (iv) Experimental procesdures and spectroscopid data in support of the experiments discussed in those Tables (v) NMR spectra of key products.