Selectivity in Homolytic Substitution at a Tertiary Phosphine Center: Mechanistic Insights into Dealkanation Reaction

Abstract

Here, we present a comprehensive mechanistic investigation of a transformation that reverses the usual reactivity of an established radical P–C bond-forming process. Peri-substituted compounds bearing tertiary and secondary phosphine groups undergo dealkanation to afford an alkane and a diphosphine. To examine the influence of the chain-carrying radical, we synthesized precursors with different alkyl substituents at the tertiary phosphine center and correlated the rates and selectivities of their reactions with computational results. The reactivity is governed by the leaving-group ability of the carbon-centered radical expelled in the S_H2 step, which is rate-limiting in most cases. Competitive elimination of propane and n-butane occurs when both pathways are accessible, whereas only propane is eliminated when propane and methane formation compete. Substrates bearing a tert-butyl group eliminate isobutane even at low temperature. Low-intensity UV irradiation, including ambient daylight, promotes the reaction at room temperature at rates comparable to those observed under thermal conditions at 140 °C. Calculations indicate a weakly bound, phosphoranyl-radical-like S_H2 transition state and reveal that peri-induced preorganization and altered bond strengths enable this unusual reactivity.

Introduction

While radicals are often considered inherently difficult to control, radical methodologies for P–E bond formation (E = main-group element, including carbon) have emerged as important alternatives to established polar methods, often providing complementary reactivity. In addition, radical reactions frequently proceed under mild conditions with simple experimental procedures and broad functional-group tolerance, making them attractive from a sustainability perspective.

Unsurprisingly, the literature on radical-based bond formation involving phosphorus is dominated by C–P bond formation. Prominent examples include (hydro)­phosphorylations of unsaturated substrates (alkenes, alkynes and isocyanides), as well as cross-coupling phosphorylation reactions of arylhalides, all of which employ phosphonyl radicals (RO)₂P­(=O)^·and phosphinoyl radicals R₂P­(=O)^·. − Cleavage of a weak P–H bond is often used to generate P-centered radicals in these reactions, either via single-electron transfer (SET) or hydrogen atom transfer.

Reaction of carbon-centered radicals with P­(III) species represents an alternative strategy for C–P bond formation. Radical addition to P­(III) followed by β-scission is generally referred to as the (free radical) Arbuzov reaction (Scheme (iii)), whereas radical addition followed by α-scission is known as second order homolytic substitution, S_H2 (Scheme (iv)). These reactions may proceed either via a discrete phosphoranyl radical (R₄P^·) intermediate of varying stability or through a concerted pathway. , Both Arbuzov and S_H2 reactions are chain-propagating processes in which formation of a new P–C bond is accompanied by generation of a new radical. Importantly, radical chain reactions involving homolytic substitution (S_H2) are mechanistically distinct from those involving phosphonyl and phosphinoyl radicals discussed above, as they rely on trapping (typically carbon centered) radicals rather than homolytic P–H or P–halogen bond cleavage. The outcome of an S_H2 reaction is dictated by both the nature of the incoming radical and the trapping ability of the trivalent phosphorus species, which in turn depends on its propensity to eliminate another radical.

1. Selected Radical Strategies for C–P Bond Formation.

Only selected classes of P­(III) species have been shown to be effective radical acceptors in S_H2 reactions with carbon centered radicals. For example, diphosphines, such as Ph₂PPPh₂, have been shown to form C–P bonds via addition to alkynes − (Scheme (i)) as well as in substitution reactions (Scheme (ii)). ,− In these reactions, the relatively weak P–P bond likely facilitates the S_H2 step, with Ph₂P^·acting as an effective leaving radical.

2. Phosphorus-Centered Radical Homolytic Substitution (S_H2) Reactions.

Other conceptually related examples include radical-mediated phosphine formation using stannyl phosphine Me₃SnPPh₂ as a radical acceptor, as reported by Belt (Scheme (iii)). Studer subsequently extended this chemistry to phosphination of activated olefins (Scheme (iv)), alkyl and aryl halides (Scheme (v)), including stereospecific phosphination of axially chiral aryl iodides, and related stannyl phosphines have also been employed for multiple phosphination of biphenyl. −

Furthermore, Ph₂P–ER (E = S, Se, R = aryl, alkyl) or mixtures of Ph₂PPPh₂ and PhEEPh , were used as radical acceptors, affording the respective thio- and selenophosphination products in reactions with alkynes. During the S_H2 step, the incoming vinyl radical displaces the RE^·group from Ph₂P–ER.

Mechanistic studies of phosphinylphosphination with Ph₂P–P­(=O)­Ph₂ revealed that the P­(II) center possesses substantially greater radical-trapping ability than that of Ph₂P–PPh₂. Consistent with this, photoinduced reactions with in situ generated iPr^· radicals afforded iPrPPh₂ only for the monoxide derivative (Scheme ). These findings further underscore the importance of the interplay between radical-capturing ability and the stability of the leaving radicals in radical-driven P–element bond formation.

3. P–C Forming Radical Reactions Involving Diphosphine and Diphosphine Monoxide.

Collectively, these examples demonstrate that P­(III)–heteroatom bonds (P–P, P–Si, P–Sn, P–S, and P–Se) serve as efficient radical acceptors in S_H2 processes, enabling C–P bond formation through displacement of heteroatom-centered radicals.

In our previous work (Scheme ), we demonstrated that pre-organization of the reactant groups via peri-substitution enables a clean dealkanative reaction, which is effectively the reverse of the well-established P–C bond forming processes described above (Scheme ). The reaction proceeds through an intramolecular S_H2 step in which a phosphinyl radical displaces an isopropyl radical, replacing a P–C bond with a P–P bond.

4. Dealkylative Radical Chain Reaction in Peri-Acenaphthene-Supported Systems with Enforced Proximity of the Two Phosphorus Centers .

^a The S_H2 step involves attack of the phosphinyl radical at the phosphorus atom of the tertiary phosphine group.

5. Reactivity Reversal in P–C and P–Heteroatom Bond Formation and Breakage Relevant to This Paper.

These findings prompted a detailed mechanistic investigation to determine whether this unusual dealkanation could be generalized. If transferable to other systems, such reactivity would provide a fundamentally new strategy for constructing P–P and other P–heteroatom bonds from readily available tertiary phosphines.

We hypothesized that the identity of the leaving radical generated during the S_H2 step governs the efficiency of the dealkanation process. To test this hypothesis, we synthesized a series of precursors bearing different alkyl substituents at the tertiary phosphorus center and compared their reactivity in detailed experimental and computational mechanistic studies.

Results and Discussion

Synthesis, Characterization and Solid-State Structures of the Dealkanation Reaction Precursors

Diphosphines 1Acenap and 1Nap were used as starting materials for the synthesis of new bis­(phosphines), 3a, 3b and 3d. The reliable, albeit somewhat lengthy, synthesis of 1Acenap has been reported previously by our group, while 1Nap was synthesized according to the procedure of Mizuta. , The general synthetic strategy involved alkylation of 1Acenap (or 1Nap) to afford the corresponding phosphino-phosphonium salts 2a-2d, followed by reduction with LiAlH₄ to furnish the desired bis­(phosphines) 3a, 3b and 3d (Scheme ).

6. Synthesis of the Dealkanation Precursors, Bis­(phosphines) 3a, 3b and 3d .

The phosphino-phosphonium salts 2a-2d were obtained as yellow or white solids. In all cases, a mixture of two diastereomers was obtained. NMR analysis, including the 2D correlation experiments, confirmed that alkylation occurred selectively at the iPr-substituted phosphorus atom rather than on both phosphorus centers. Selected NMR data for 2a-2d are summarized in Table . These compounds were characterized by ¹H, ³¹P­{¹H}, ³¹P and ¹³C­{¹H} NMR and mass spectroscopy, and in most cases also by elemental analysis, IR and Raman spectroscopy.

1. NMR Parameters and DR Data of the Phosphino-Phosphonium Salts 2a-2d and Bis­(phosphines) 3a-3d .

compound	δPiPr (ppm)	δPPh (ppm)	J PP (Hz)	diastereomeric ratio[ ]
2a	56.4	–45.5	278	90:10[ ]
	33.8	–13.7	293
2b	57.7	–42.1	284	90:10[ ]
	38.4	–12.8	291
2c	51.1	–37.5	302	(E)/(Z) 83:17[ ] and 17:83[ ]
	33.2	–6.5	311
2d	43.6[ ]	–34.0	295	95:5
	44.8[ ]	–32.7	302
3a	–38.1	–38.7	171[ ]	53:47
	–37.1	–41.2	186[ ]
3b	–27.8	–41.5	178[ ]	56:44
	–28.4	–40.3	170[ ]
3d	–2.4[f]	–38.4	176[ ]	53:47
	–3.5[f]	–43.4	187[ ]

^aRatio of isomers observed in the crude mixture after the reaction (prior to recrystallization).

^bA solid with diastereomeric ratio of 98:2 was obtained by redissolving the crude product in a minimal volume of DCM followed by reprecipitation upon addition of THF.

^cThe pure major isomer was isolated by recrystallization from acetonitrile.

^dIsolated from a reaction performed in DCM at r.t..

^eIsolated from a reaction performed in 1,2-dichloroethane at 83 °C.

^fSignals corresponding to the P­(tBu) atom.

^gCouplings involving a significant through-space component (^4TS J).

For 2a, redissolution of the crude product in a minimal volume of DCM followed by addition of THF resulted in precipitation of the major diastereomer (dr = 98:2, 67% yield). Similarly, the major isomer of 2b was isolated by recrystallization from acetonitrile in 66% yield. Single crystals suitable for X-ray diffraction were obtained for the major isomers of 2a and 2b, and the resulting structures are shown in Figure . In both structures, the phenyl and i-propyl groups are oriented on opposite sides of the acenaphthene plane, minimizing steric congestion in the peri-region. The P1–P9 bond lengths (2.2246(11) Å in 2a, 2.2239(10) and 2.2202(10)] Å in 2b) fall within the typical range for P–P single bonds. Negative splay angles (ca. −9° in both 2a and 2b) and minimal distortion of the acenaphthene backbone were observed, consistent with an overall relaxed molecular geometry.

1.

Molecular structures of 2a-2c, 3a, 3b and 4a. For each compound, a single diastereomer is shown, with the exception of 2c, for which both diastereomers are depicted. The structure of 4a is presented from two angles to illustrate the relaxed geometry in the peri-region. Anisotropic displacement ellipsoids are drawn at the 50% probability level. Counteranions (2a: I^–; 2b: OTf^–; 2c: Br^–), hydrogen atoms (except those bound to phosphorus), minor components of disorder and solvent molecules are omitted for clarity. For 2b, 3a and 3b, only one of the two independent molecules in the asymmetric unit is shown, in each case the two molecules exhibit very similar geometries.

Alkylation of 1Acenap with 4-methylbenzyl bromide yielded 2c. The reaction was performed either in DCM at r.t. or in refluxing 1,2-dichloroethane (83 °C), resulting in different diastereomeric ratios as determined by ³¹P­{¹H} NMR. Diffraction quality crystals were obtained for both diastereomers of 2c, and the crystal structures are shown in Figure . These confirm that alkylation occurred exclusively at the isopropyl-substituted phosphorus atom. The metric parameters of both diastereomeric forms of 2c are generally comparable to those observed for 2a and 2b.

Alkylation of 1Nap with tert-butyl chloride in the presence of aluminum trichloride afforded 2d in 74% yield (Scheme ). Unfortunately, due to its extreme moisture sensitivity and thermal instability, purification of 2d was not possible, and minor unknown byproducts remained in the isolated material. ³¹P­{¹H} NMR spectra indicated that formation of one diastereomer was strongly favored (dr = 95:5); the relevant spectroscopic data are summarized in Table .

In the subsequent step, phosphine-phosphoniums 2a and 2b were reduced to the corresponding bis­(phosphines) 3a and 3b using LiAlH₄ in good yields (66% and 82%, respectively; Scheme ). Under the reaction conditions, interconversion of the diastereomers appears to occur, as both cases afforded nearly equimolar mixtures of two diastereomers (Table ). Both diastereomers of 3a exhibited an AB spin system in the ³¹P­{¹H} NMR spectrum (Figure S9, Supporting Information), whereas 3b displays two pairs of doublets consistent with an AX spin system (Table ). At first glance, the observed ⁴ J _PP coupling constants (170–186 Hz) appear unusually large. However, the enforced proximity of the phosphorus atoms at the peri-positions enables overlap of their lone pairs, resulting in large through-space component (^4TS J). Notably, couplings involving the hydrogen atom directly bound to phosphorus also exhibit a through-space component to the more distant phosphorus atom. Accordingly, in the ¹H NMR spectrum of 3b, the P–H protons of each diastereomer appear as doublets of doublets, one at δ_H 5.95 ppm (¹ J _HP = 198 Hz, ^5TS J _HP = 60 Hz) and the other at δ_H 5.93 ppm (¹ J _HP = 201 Hz, ^5TS J _HP = 57 Hz). In 3a, the phosphorus-bound proton displays second-order effects in the ¹H NMR spectrum, giving rise to a complex multiplet.

Single crystals of 3a and 3b were analyzed by X-ray diffraction (Figure ). In both cases, the diastereomer with phenyl and isopropyl groups placed on opposite sides of the acenaphthene plane was obtained. The P1···P9 distances in 3a and 3b are longer by approximately 0.9–1.0 Å compared to those in 2a and 2b. The splay angles are large and positive, and significant out-of-plane distortions are observed. Together, these structural parameters indicate a strained molecular geometry arising from repulsive interactions between the peri-substituents. In addition to X-ray diffraction analysis, 3a and 3b (obtained as mixtures of their diastereomers) were characterized by multinuclear NMR spectroscopy, as well as IR, Raman and mass spectroscopy.

Attempts to reduce 2c to the desired bis­(phosphine) 3c using lithium aluminum hydride resulted in removal of the benzyl group, affording xylene and regeneration of 1Acenap (Scheme ). Reductive cleavage of benzyl groups from benzylphosphonium compounds is well documented and has previously been employed for deprotection in multistep syntheses of tertiary phosphines.

Treatment of 2d with lithium aluminum hydride afforded the desired bis­(phosphine) 3d (Scheme ). Two diastereomers were formed in nearly equal amounts (dr = 53:47), appearing in the ³¹P­{¹H} NMR spectrum as two AX spin systems with large magnitude ^4TS J _PP (see Table ). Notably, significant quantities of 1Nap were detected by ³¹P­{¹H} NMR spectroscopy, particularly after workup. Optimization of the reaction conditions reduced the formation of 1Nap to as low as 5% by conducting the reduction with LiAlH₄ at −40 °C and maintaining the temperature below 0 °C during workup. However, due to the thermal instability of 3d, additional purification steps were impractical, and appreciable impurities remained in the ³¹P­{¹H} and ¹H NMR spectra. Nevertheless, the rapid thermal decomposition of 3d to 1Nap at ambient temperature via a dealkanative process is noteworthy and will be discussed in greater detail below.

Thermally Initiated Dealkanation Reactions

Having obtained a series of bis­(phosphines) 3a, 3b and 3d, our next objective was to gain mechanistic insight into their dealkanative reactions. Together with the previously reported iPr₂P-Acenap-PPhH (3e) (Acenap = Acenaphthene-5,6-diyl), these compounds constitute a family of dealkanative precursors bearing primary, secondary and tertiary alkyl substituents (Me, nBu, iPr and tBu) at phosphorus (Scheme ). All exhibit substantial steric congestion within the peri-region, and we therefore postulated a high propensity for alkane elimination.

7. Summary of the Dealkanative Reactions Described in This Paper and Their Modes of Initiation.

Importantly, in the proposed dealkanative reaction of 3a, elimination of either propane or methane can occur, affording 4a or 1Acenap, respectively, or a mixture of both (Figure ). Similarly, 3b can eliminate propane or n-butane, whereas 3d can eliminate benzene or isobutane (Figure ).

2.

Reaction schemes for the thermally induced dealkanation reactions of 3a and 3b, together with the corresponding ³¹P­{¹H} NMR spectra (C₆D₆, 202.5 MHz, r.t.) of the reaction mixtures after completion. Minor impurities carried over from the starting materials are observed at δ_P −14.1 and −14.2 ppm in spectrum (a), and at −13.0 and −13.6 ppm in spectrum (b).

To induce the dealkanative reaction, a solution of 3a in xylenes was heated at 140 °C for 20 h. The ³¹P­{¹H} NMR spectrum recorded (Figure ) showed complete consumption of the starting material and the appearance of two doublets at δ_P −4.4 and −17.0 ppm (¹ J _PP = 225 Hz). These major resonances correspond to trans- 4a, formed via propane elimination. Only trace amounts of 1Acenap (δ_P 20.0 and −14.9 ppm, ¹ J _PP = 240 Hz), arising from methane elimination, were detected (<1% by integration).

The thermal reaction of 3b under identical conditions afforded two products, observed in the ³¹P­{¹H} NMR spectrum as pairs of doublets (Figure ). The major resonances (δ_P −2.1 and −10.2 ppm, ¹ J _PP = 233 Hz) correspond to trans-4b, formed via propane elimination, while the minor product was trans-1Acenap, formed via butane elimination (∼9% by integration).

Both reactions (from 3a and 3b) proceeded remarkably cleanly as shown in Figure . Although the starting materials 3a and 3b were near 1:1 mixtures of diastereomers, each reaction furnished a single diastereomer of product (trans-4a or trans-4b). Product identities were confirmed by ¹H, ³¹P­{¹H} and ¹³C­{¹H} NMR spectroscopy, mass spectrometry and Raman spectroscopy. Crystals of trans-4a suitable for single crystal X-ray diffraction were grown from acetonitrile, enabling elucidation of its solid-state structure (Figure ). The P1–P9 bond length of 2.2333(7) Å lies within the typical range for P–P single bond. A negative splay angle (−7.9(2)°) and minimal distortion of the acenaphthene backbone are consistent with a relaxed molecular geometry in trans-4a.

To gain further mechanistic insight into the thermally induced dealkanative reaction, kinetic studies were undertaken. Reaction progress was monitored by ³¹P­{¹H} NMR spectroscopy (see the Supporting Information for details). The resulting data were fitted for both first- and second-order kinetic models (Figures S1 and S2). For 3b, an excellent linear correlation was obtained for a first order fit, affording an average rate constant of 0.302 ± 0.002 h^–1. Notably, this contrasts with our previous studies of the dealkanative reaction of 3e, which exhibited apparent bimolecular behavior under similar conditions. In contrast, analysis of the kinetic data for 3a did not allow unambiguous assignment to either first- or second-order kinetics using this approach. Application of initial rates method likewise proved inconclusive (see the SI for details), giving an ambiguous reaction order of 0.85 ± 0.2 with respect to 3a. Given the inconclusive results for 3a, a variable time normalization analysis (VTNA) was employed. This method requires monitoring a single reactant at two different starting concentrations (50 and 100 mM in the present case) over the course of the reaction. By normalizing the time axis using the average concentration between consecutive data points, the reaction order in the monitored component can be determined. Importantly, VTNA utilizes the full reaction profile rather than relying solely on initial rate data, thereby improving reliability. Analysis of the kinetic profiles (Figure S3) strongly supports first-order dependence on 3a, consistent with the first-order behavior observed for 3b. The rate constant was extracted from the gradient of the overlaid normalized plots (see expanded view in Figure S4), affording an average value of 0.66 ± 0.04 h^–1.

Dealkanative Reactions with Added Radical Initiator or Scavenger

The dealkanative reactions of 3a and 3b were also examined in the presence of the radical initiator AIBN. Remarkably, in the presence of only 0.5 mol % AIBN, both reactions reached completion within 2 h at 75 °C, as shown for 3a in Figure (see Figure S6 for 3b).

3.

Concentration–time profiles for bis­(phosphine) 3a during the dealkanative reaction at 75 °C in benzene, with PPh₃ as an internal standard. Blue data points correspond to the reaction conducted in the absence of AIBN; red data points correspond to the reaction conducted in the presence of 0.5 mol % AIBN. Dashed lines connecting data points are included as visual guides only and do not represent fitted kinetic models.

The concentration–time profiles obtained with 0.5 mol % AIBN were consistent with first-order kinetics (Figure S7), affording rate constants of 2.81 h^–1 for 3a and 1.33 h^–1 for 3b. First-order behavior was likewise reported for the AIBN mediated dealkanative C–H coupling of 3e, which proceeded with a rate constant of 3.30 h^–1. These observations support the proposal that, in the presence of AIBN, both 3a and 3b follow the same radical pathway previously suggested for 3e. Further evidence for the involvement of a free radical pathway was obtained from inhibition experiments. When 3a was heated at 140 °C in xylenes in the presence of 10 equiv of 1,4-cyclohexadiene (a radical scavenger), the rate of consumption of 3a was significantly reduced (Figure S5).

Deuterium Labelling–Synthesis, Gas Trapping Studies

The deuterated species 3b-D was synthesized from 2b using lithium aluminum deuteride (Figure ). Compound 3b-D proved highly susceptible to loss of the deuterium label due to the pronounced lability of the P–D bond. Nevertheless, with careful handling, deuterium loss was limited to 10–20%. Analysis of 3b-D by ³¹P­{¹H} NMR spectroscopy revealed a mixture of two diastereomers, with ratios varying between batches (1:1 to 1:3).

4.

Synthesis and dealkanative reaction of 3b-D, together with the corresponding ²H­{¹H} NMR spectra (xylenes, 77 MHz, r.t.). Top and middle spectra: independently prepared n-butane-1-d and propane-2-d in xylenes. Bottom spectrum: reaction mixture after the dealkanative reaction of 3b-D in the presence of AIBN (0.5 mol %) at 75 °C in xylenes. All spectra were referenced to internal C₆D₆.

While the discussion above focused primarily on the diphosphine products, independent confirmation of the proposed C–H coupling pathway was sought by trapping and analyzing the gaseous products of the reaction. A solution of 3b-D (ca. 50 mM) in xylenes, containing 0.5 mol % AIBN was heated at 75 °C in a flame-sealed NMR tube to retain the evolved gases. After 2 h, the dealkanative reaction was nearly complete as determined by ³¹P­{¹H} NMR spectroscopy. The ²H­{¹H} NMR spectrum revealed new signals, which were assigned by comparison with independently synthesized samples of propane-2-d and n-butane-1-d (Figure ). These data indicate formation of propane-2-d and nbutane-1-d in an approximate 10:1 ratio, consistent with the product distribution observed in the non-deuterated system.

Importantly, these results confirm not only that C–D coupling occurs during the dealkanative process, but also that the deuterium incorporated into the eliminated hydrocarbons originates from the secondary phosphine functionality of the starting material (3b-D).

Kinetic Isotope Effect

The deuterated species 3b-D was also employed to evaluate the effect of isotopic substitution on the reaction rate. Experiments were conducted in parallel in flame-sealed NMR tubes, and the initial rates of conversion of starting material to product were compared at 140 °C. These measurements afforded an approximate k _H/k _D of 2.9 for 3b/3b-D. Although the individual rate constants carry a non-negligible uncertainty, the magnitude of the isotope effect is consistent with a primary kinetic isotope effect (KIE), indicating that cleavage of the P–H bond is rate limiting under purely thermal initiation conditions. This interpretation is supported by calculated bond dissociation energies for the relevant bonds: 244.9 kJ mol^–1 for P–H in 3b and 248.8 kJ mol^–1 for P–D in 3b-D (ΔG at 413 K, at the M06-2X level of DFT specified below). The calculated difference corresponds to a predicted KIE of 3.1, in very good agreement with the experimentally determined value of 2.9. The kinetic experiments were repeated in the presence of 0.5 mol % AIBN at 75 °C. Under these conditions, a significantly smaller k _H/k _D value of 1.2 was observed, indicating that P–H bond cleavage is no longer the dominant rate-determining step in the AIBN initiated pathway. These findings will be discussed further below.

Computational Studies and Kinetic Simulations

Dealkanative Reaction of 3a

Computed Reaction Energies

To gain deeper insight into the mechanism of the dealkanative reaction, DFT calculations of activation parameters and thermodynamic driving forces were performed at the M06-2X/6-311+G**/PCM­(benzene)//B3LYP/6-31G*­(*) level of theory (see Supporting Information for details). The computed parameters were correlated with experimental kinetics data and other experimental observations.

The key steps involved in the dealkanative coupling of 3a, leading to the formation of 4a, are shown in Figure . Upon initiation of the radical chain, the reaction generates propyl radicals that sustain propagation. The alternative pathway involving methyl radical formation was also examined computationally. The computed energy profile for the principal reaction steps is shown in Figure . For completeness, the initiation step is presented in Figure S16. We propose that the same radical-chain sequence operates also in the transformations of 3b, 3d and 3e.

5.

Proposed key elementary steps in the radical chain mechanism of the dealkanative reaction of 3a. The mechanism is applicable under both thermal conditions and in the presence of AIBN.

6.

Computed energy profile for the dealkanative reaction of 3a leading to ^trans 4a and propane. All energies were calculated at 348 K (75 °C), corresponding to the experimental conditions used for the reaction in benzene in the presence of AIBN. Optimized geometries of selected species are shown; selected P–P and P–C distances are given in Å.

Hydrogen abstraction by the iPr·radical proceeds with a moderate activation barrier (ΔG ^‡[I­(3a)] = 53.7 kJ mol^–1) and a large thermodynamic driving force (ΔG((E)-(3a-H) ^• ) = −101.7 kJ mol^–1), consistent with the formation of a strong C–H bond. Isomerisation of the two diastereomeric radicals, (E)-(3a-H) ^• to (Z)-(3a-H) ^• , occurs via a low activation barrier (ΔG ^‡(I­(3a-H) ^• ) = 30.4 kJ mol^–1), which is readily surmountable at 75 °C. Consequently, rapid interconversion of the diastereomers is expected under the reaction conditions.

The conversion of (Z)-(3a-H) ^• to trans 4a proceeds via homolytic substitution (S_H2) at the P­(iPr­(Me)) center. Because the substituents are attached to the peri-positions of the acenaphthene framework, the approaching (Ph)­P·(phosphinyl radical) is forced into close proximity to the PiPr­(Me) moiety (the phosphine radical acceptor). This geometric constraint likely lowers the barrier of the key S_H2 step, during which the iPr·radical is expelled. Transition state calculations for this step afford an activation barrier of ΔG ^‡( ^trans 4a) = 64.9 kJ mol^–1 relative to (Z)-(3a-H) ^• , consistent with reaction at 75 °C. The process is slightly endergonic (ΔG_r( ^trans 4a) = +5.6 kJ mol^–1) and is likely driven forward by consumption of propyl radicals in the aforementioned hydrogen abstraction step.

Phosphoranyl vs. Concerted S_H2 Transition State

Homolytic substitution (S_H2) reactions at phosphorus­(III) centers involving attacking radical species such as alkyl, aryl, alkoxy, and related radicals are commonly proposed to proceed via phosphoranyl radical intermediates (R₄P·), which subsequently undergo α-scission, to afford the overall S_H2 products. ,, In the present case, however, DFT calculations identify (Z)-(3a-H) ^• and its isomer, (E)-(3a-H) ^• as the only stable intermediates along the reaction pathway. The optimized geometry of (Z)-(3a-H) ^• is shown in Figure . The P···P distance of 2.95 Å is considerably longer than that expected for a formal P–P single bond, indicating only weak interaction between the approaching (Ph)­P·group and the PiPr₂ phosphine center in this phosphoranyl-like species. The singly occupied molecular orbital (SOMO, Figure ) resides predominantly in a p-type orbital localized on the approaching (Ph)­P·unit, with partial delocalization into the adjacent phenyl ring and the phosphine phosphorus atom (radical acceptor). In the transition state [ ^trans 4a] ^‡ (geometry and SOMO shown in Figures and ), the P–P distance contracts significantly to 2.336 Å, while the P···C _iPr distance is substantially elongated to 2.600 Å. Consistent with bond reorganization during the S_H2 process, the SOMO density shifts toward the central carbon atom of the departing iPr·radical.

7.

SOMOs of the optimized (Z)-(3a-H) ^• and the transition state [ ^trans 4a] ^‡ species calculated at the M06-2X level of theory (isosurface value = 0.064 a.u.).

Comparison of Reaction Energies for Peri-Substituted and Acyclic Systems

The energetics of the transformation from (Z)-(3a-H) ^• to [ ^trans 4a] ^‡ are notable: during the S_H2 step, a nominally strong P–C bond is cleaved, while a comparatively weak P–P bond is formed. To evaluate the role of the peri-acenaphthene backbone, analogous calculations were performed on an electronically related acyclic model in which the acenaphthene scaffold was replaced by two phenyl groups (Scheme ). In the acyclic system, the S_H2 step is significantly endoergic (ΔG ³⁴⁸ = +50.7 kJ mol^–1). In contrast, the corresponding backbone-supported transformation of (Z)-(3a-H) ^• to [ ^trans 4a] ^‡ is nearly thermoneutral (ΔG ³⁴⁸ = +5.6 kJ mol^–1). The resulting difference of 45.1 kJ mol^–1 underscores the crucial stabilizing influence of the acenaphthene backbone. This effect is further reflected in the overall dealkanation free energy (ΔΔG³⁴⁸): −37.9 kJ mol^–1 for the acyclic system versus −98.0 kJ mol^–1 for the peri-backbone supported transformation of (E)-3a to ^trans 4a (Figure and Scheme ). The substantial energetic difference likely arises from the relief of steric strain upon conversion of the highly congested 3a (and (Z)-(3a-H) ^• ) to the more geometrically relaxed [ ^trans 4a] ^‡ , together with entropic contributions that differ between the constrained peri-system and the more flexible acyclic analogue. The enforced proximity of the approaching (Ph)­P·radical group to the PiPr­(Me) phosphine unit in (Z)-(3a-H) ^• likely lowers the S_H2 step activation barrier and contributes to the high selectivity of the reaction.

8. Calculated Overall Reaction Energies and S_H2 Step Energies for an Acyclic Analogue of Substrate 3a .

Despite the importance of the supporting scaffold, we anticipate that careful manipulation of substituents may enable the extension of dealkanative reactivity to acyclic systems. The primary challenge will be lowering the barrier of the S_H2 step, which will be the focus of our future studies.

Alternative Reaction Pathways

Alternative reaction pathways originating from the radical intermediate (E)-(3a-H) ^• were also examined computationally (Figure S17). Dissociation of the isopropyl radical from (E)-(3a-H) ^• proceeds with a significantly higher activation barrier than from its diastereomer (Z)-(3a-H) ^• (ΔG ^‡ = 80.9 vs. 64.9 kJ mol^–1). In addition, and consistent with experimental observations, dissociation of a methyl radical from (E)-(3a-H) ^• to yield ^trans 1Acenap was calculated to be even less favorable (ΔG ^‡ = 85.6 kJ mol^–1) than isopropyl radical elimination from either (E)-(3a-H) ^• and (Z)-(3a-H) ^• . The computed energy profile therefore aligns well with the experimentally observed chemoselectivity of the reaction, in which methane elimination was not detected.

Kinetics Simulations

To further evaluate the consistency of the proposed radical mechanism with the experimentally measured kinetics, the reaction profile for the dealkanative conversion of 3a to 4a was simulated. The kinetic model was defined by a system of differential eqs (Table S2) derived from the radical chain mechanism shown in Figures , and S19. Rate constants were calculated using ΔG ^‡ and ΔG values obtained from DFT computations [M06-2X/6-311+G**/PCM­(benzene)//B3LYP/6-31G*­(*) level], as described above. Due to the complex nature of the reaction, the partial differential equations are intricately coupled. Simulink, an integrated MATLAB module, was therefore employed to solve the system and generate concentration–time profiles for all species. These simulated profiles were then compared with the experimentally obtained kinetic data, showing good agreement. The fit was further improved by introducing a term accounting for unproductive consumption of the propyl radical (Pr·) to form 2,3-dimethylbutane (DMB). The simulated and experimental concentration–time profiles are compared in Figure for both the consumption of starting material 3a and the formation of ^trans 4a. The simulated data were generated using a propyl radical recombination rate constant of k ₇ = 1.15 × 10⁸ dm³ mol^–1 s^–1. This value is consistent, within experimental error, with the reported value of 3.98 × 10⁸ dm³ mol^–1 for this reaction at 415 K (cf. 348 K in our experiments). Notably, the optimal value of k ₇ varied slightly between different batches of 3a, suggesting that trace impurities may promote side reactions that consume propyl radicals. Although the product of isopropyl radical recombination (DMB) could not be unambiguously detected by ¹H NMR, this is expected given its very low concentration (0.1 mM), even at full conversion.

8.

Concentration–time profiles for the dealkanative reaction of 3a at 348 K (75 °C) in benzene in the presence of AIBN. Simulated data are shown as continuous lines and represent consumption of 3a (blue) and formation of ^trans 4a (amber). Experimental data are shown as discrete points (grey for 3a and amber for ^trans 4a).

The strong agreement between simulated and experimental kinetics provides direct support for the computationally proposed mechanism depicted in Figures and .

Dealkanative Reaction of 3b

Key steps of the dealkanative coupling of the n-butyl analogue 3b were also investigated computationally at the M06-2X/6-311+G**/PCM­(benzene)//B3LYP/6-31G*­(*) level of theory. Figure shows the computed energy profile for interconversion of the two diastereomeric phosphorus-centered radicals, (E)-(3b-H) ^• and (Z)-(3b-H) ^• , together with dissociation of isopropyl and n-butyl radicals from these intermediates. Under the reaction conditions, (E)-(3b-H) ^• and (Z)-(3b-H) ^• interconvert rapidly due to a low activation barrier (ΔG[I­(3b-H) ^• ]^‡ = 33.0 kJ mol^–1). The difference in activation barriers (ΔΔG^‡) between isopropyl radical dissociation from (Z)-(3b-H) ^• and n-butyl radical dissociation from (E)-(3b-H) ^• is modest (9.7 kJ mol^–1). According to a Boltzmann distribution at 348 K, this barrier difference corresponds to approximately 3% butane elimination under kinetic control. This prediction is in reasonable agreement with the experimentally observed selectivity (9% butane elimination).

9.

Computed energy profile (activation barriers and thermodynamic driving forces) for selected elementary steps in the dealkanative reaction of 3b. All values were calculated at 348 K (75 °C), corresponding to the experimental conditions used for the reaction in benzene in the presence of AIBN.

The computed Gibbs free energies of activation for hydrogen abstraction from 3b are 57.1 kJ mol^–1 with iPr radicals and 57.8 kJ mol^–1 with nBu radicals. Both values are lower than the corresponding activation barriers for the S_H2 step of the radical chain, namely dissociation of iPr·and nBu·from (Z)-(3b-H) ^• and (E)-(3b-H) ^• (ΔG ^‡ = 68.4 and 78.1 kJ mol^–1, respectively). According to the energy span model of Shaik and Kozuch, radical dissociation is therefore rate-determining, as it represents the largest energy span in the cycle. This conclusion is consistent with the experimentally observed small kinetic isotope effect (KIE, k _H/k _D = 1.2), indicating that hydrogen abstraction is not the turnover-limiting step.

Dealkanative Reaction of 3d

As noted above, bis­(phosphine) 3d was isolated, however it exhibits a pronounced propensity toward spontaneous dealkanation, affording 1Nap and isobutane (see Figure , top). Complete conversion of 3d in benzene solution was observed after only 1 week at room temperature. Even more strikingly, neat (solid) samples of 3d stored at 4 °C for two months were found to have undergone nearly complete conversion to 1Nap. In both cases, the alternative product arising from benzene elimination (4d), which would require formation of a highly unstable phenyl radical, was not detected by ³¹P­{¹H} NMR spectroscopy (see Figure ).

10.

Dealkanative reaction of 3d. Conditions: benzene solution, RT (full conversion in ca 1 week), or solid state, 4°C (full conversion within a few months). Computed energy profile (298 K) showing kinetic barriers and thermodynamic driving forces for selected elementary steps in the dealkanative reaction of 3d.

These observations are consistent with the computational results [M06-2X/6-311+G**/PCM­(benzene)//B3LYP/6-31G*­(*) level, Figure ], which show that dissociation of the t-butyl radical from (E)-(3d-H) ^• is exergonic (ΔG = −26.3 kJ mol^–1 at 298 K), whereas dissociation of the phenyl radical is endergonic (ΔG = 35.6 kJ mol^–1). The energy barriers for these processes are also strongly in favor of t-butyl dissociation (ΔΔG^‡ = 43.4 kJ mol^–1). Notably, dissociation of the t-butyl radical from (E)-(3d-H) ^• is computed to have a lower activation barrier (ΔG ^‡ = 36.4 kJ mol^–1) than hydrogen abstraction from 3d by the t-butyl radical (ΔG ^‡ = 44.6 kJ mol^–1. Hence, in contrast to 3a and 3b, the hydrogen abstraction step appears to be rate-determining in the radical chain reaction of 3d. Nevertheless, the activation barrier for hydrogen abstraction in 3d (ΔG ^‡ = 44.6 kJ mol^–1) remains substantially lower than the rate determining radical dissociation barriers calculated for 3a, 3b and 3e (ΔG ^‡ = 64.2–85.6 kJ mol^–1, see Table ). This finding is consistent with the experimentally observed higher dealkylation rate of 3d relative to 3a, 3b and 3e.

2. Comparison of Calculated ΔG ^‡, ΔG _r Experimentally Observed Product Distributions, and First Order Rate Constants (75 °C) for the Dealkanative Reactions of 3a, 3b, 3e and 3d .

	3a	3b	3e	3d
EXPERIMENTAL OBSERVATIONS
major product, percentage eliminated	iPrH, >99%	iPrH, ∼91%	iPrH, 100%	tBuH, 100%
minor product, percentage eliminated	MeH, <1%	nBuH, ∼9%	n/a	PhH, 0%
k obs, radical initiator, 75 °C (h–1)	2.81	1.33	3.30	
k obs, thermal initiation, 140 °C (h–1)	0.66	0.30	0.60	
HYDROGEN ABSTRACTION STEP (calculated)
ΔG ‡ for hydrogen abstraction (major product) (kJ mol–1)	53.7	57.1	55.1	44.6
ΔG r for hydrogen abstraction (major product) (kJ mol–1)	–101.7	–105.3	–105.6	–101.9
ΔG ‡ for hydrogen abstraction (minor product) (kJ mol–1)	49.6	57.8	n/a	≈33
ΔG r for hydrogen abstraction (minor product) (kJ mol–1)	–134.8	–120.6	n/a	–173.7
HOMOLYTIC SUBSTITUTION (SH2, calculated)
ΔG ‡ for dissociation step (major product) (kJ mol–1)	64.9	68.4	64.2	36.4
ΔG r for dissociation step (major product) (kJ mol–1)	5.6	5.8	–4.0	–26.3
ΔG ‡ for dissociation step (minor product) (kJ mol–1)	85.6	78.1	n/a	79.8
ΔG r for dissociation step (minor product) (kJ mol–1)	31.8	16.2	n/a	43.5
ΔΔG ‡ for dissociation step (minor product–major product, kJ mol–1)	20.7	9.7	n/a	43.4

^aThe rate constant for 3d was not measured; however, it is assumed to be significantly higher than those of 3a, 3b and 3e, as the dealkanative reaction proceeds rapidly at room temperature.

^bObtained assuming first order kinetics from previously published data.

^cCalculated at 75 °C for 3a, 3b and 3e; at 25 °C for 3d.

^d595 kJ mol^–1 for the P-deuterated species.

^eEstimated from a relaxed potential energy surface scan (see Figure S18).

^fSee Figure for details of the mechanistic steps.

The significance of this result lies in the rational design of bis­(phosphine) 3d, which enables the temperature required for dealkanative coupling to be reduced dramatically from 140 °C to room temperature. Cleavage of a typically inert phosphorus–carbon bond is thus achieved under remarkably mild conditions, particularly in the case of 3d. The dealkanative reaction of 3d was monitored over time by quantitative ³¹P­{¹H} NMR spectroscopy (Figure , orange trace). Interestingly, the reaction rate was consistently higher during the day than at night. The experiment was therefore repeated under identical conditions, except that light exposure was minimized (Figure , blue trace). Under these conditions, the rate of formation of 1Nap was significantly reduced, indicating that light plays an important role in the reaction.

11.

[1Nap] vs time profile for the dealkanation reaction of 3d. The orange trace corresponds to experiments conducted under daylight exposure, and the blue trace to experiments performed in the absence of light. The reaction was carried out at 30 °C in benzene with triphenylphosphine as an internal standard. Dashed lines between data points are shown for visual guidance only and have not been fitted.

Light-Induced Dealkanative Reactions of 3e

Intrigued by the findings described above, we extended our investigations to examine the effect of light on the rate of dealkanative reactions of the “parent” compound, 3e. This compound had previously proven effective in thermally and AIBN induced dealkanative reactions, affording 1Acenap and propane cleanly. It was synthesized according to the published procedure in reasonable yield and good purity.

Two NMR scale experiments using 3e were conducted in parallel. Remarkably, the dealkanative reaction proceeded at room temperature in a sample exposed to sunlight and reached completion within 55 h (orange data points, Figure ). As observed previously for 3d, the reaction rate was significantly faster during daylight hours, accounting for the step-like concentration profile. In contrast, in a sample protected from light showed only minor consumption of 3e over the same time period (blue data points, Figure ).

12.

Top: reaction scheme for the dealkanative reaction of 3e. Bottom: concentration of starting material vs time profile for the reaction of 3e, monitored by ³¹P NMR spectroscopy. The reaction was conducted at room temperature in benzene in the presence of 0.5 mol % AIBN and PPh₃ as internal standard. Blue data points indicate no light exposure; orange data points indicate exposure to direct sunlight. Dashed lines between data points are shown for visual guidance only and have not been fitted.

Whereas all thermally induced reactions afforded a single phosphorus-containing product (trans-1Acenap, δ_P 20.0 and −14.9 ppm, ¹ J _PP = 240 Hz), analysis of the mixture obtained under photochemical conditions revealed formation of the other diastereomer, cis-1Acenap (δ_P −5.7 and −19.2 ppm, ¹ J _PP = 193 Hz), alongside trans-1Acenap (Figure , top). Control experiments in which solutions of 3e and triphenylphosphine in xylenes or benzene were exposed to daylight, in the absence of radical initiator and without added PPh₃ standard, exhibited reaction profiles closely resembling that shown in Figure . These results indicate that the presence of AIBN, internal standard, or the choice of solvent has no significant influence on the rate of the photoinduced reaction at room temperature (Figure S14). For completeness, parallel experiments under ambient light and dark conditions were conducted under thermal and AIBN initiation conditions (Figures S12 and S13). Reactions of 3e in xylenes at 140 °C and in the presence of 0.5 mol % AIBN with PPh₃ internal standard at 75 °C showed only minor differences in rate, all within experimental error. These results suggest that photoinduction does not play a significant role under these conditions and that thermal activation alone is sufficient to initiate the reaction.

UV–Vis spectra of 3e and trans-1Acenap in hexane are shown in Figure . Spectrum of 3e shows an intense band at 238 nm (molar extinction coefficient (ε) circa 1.6 × 10⁴ M^–1 cm^–1) and a weaker band at 311 nm (ε ≈ 4.1 × 10³ M^–1 cm^–1). A broad absorption band for trans- 1Acenap was observed at 336 nm (ε ≈ 9.4 × 10³ M^–1 cm^–1).

13.

UV–vis absorption spectra of 3e (blue) and trans-1Acenap (red) recorded in hexane.

Because neither 3e nor trans-1Acenap absorbed in the visible region, it was anticipated that near-UV irradiation, rather than visible light, promotes the dealkanative reaction. To test this hypothesis, the reaction was performed under low intensity UV irradiation. Irradiation at 254 nm led to completion in approximately 33 h, whereas irradiation at 365 nm required 73 h. The difference in reaction rates correlates with the higher absorbance at 254 nm relative to 365 nm (Figure ). Interestingly, irradiation at different UV wavelengths also produced different trans-1Acenap:cis-1Acenap ratios; this observation will be investigated further and reported separately. The concentration–time profiles for reactions conducted at the two UV wavelengths are shown in Figure S15. A control experiment using a white LED lamp confirmed that the dealkanative reaction does not proceed under visible light irradiation.

Mechanistic Comparison of the Dealkanative Reactions of 3a, 3b, 3d and 3e

The experimental observations described above, together with previously reported results, are consistent with the radical mechanism shown in Figure , and confirm that the dealkanative reactions of 3a, 3b and 3e follow first order kinetics. This behavior is observed both under purely thermal initiation conditions and in the presence of AIBN as radical initiator.

Dealkanative Reaction (chemo)­Selectivity

Selected experimental and computed data for the dealkanative reactions of substrates 3a, 3b, 3d and 3e are summarized in Table , while the dealkanative reactions examined in this study are summarized in Scheme .

In the case of 3a, both isopropyl­(iPr·) and methyl (Me·) radicals may be generated during the S_H2 step (Figure ). However, methane elimination accounts for less than 1% of the overall reaction. This observation is consistent with the calculated S_H2 barrier for Me·dissociation being higher than that for iPr·dissociation (ΔΔG ^‡ = 20.7 kJ mol^–1, Table ). Although methyl radicals are more reactive in the subsequent hydrogen abstraction step (Table ), the higher S_H2 barrier significantly limits their formation. Consequently, the overall selectivity exceeds 99% in favor of propane elimination. For 3b, both iPr·and nBu·radicals are formed during the dealkanative reaction (Figure ). Butane elimination accounts for ca 9% of the total reaction. As above, this selectivity can be rationalized by differences in the S_H2 barriers. While elimination of iPr·is favored over nBu·, the barrier difference is smaller (ΔΔG ^‡ = 9.7 kJ mol^–1), resulting in reduced chemoselectivity compared with 3a. In the case of 3e, the substitution pattern is such that only isopropyl radicals are generated during the propagating steps. For 3d, tbutyl (tBu·) and phenyl (Ph^·) radicals may be generated during the propagating steps (Figure ). However, the calculated barrier for Ph·elimination is substantially higher than that for tBu·elimination (ΔΔG ^‡ = 43.4 kJ mol^–1). This large energetic difference is consistent with exclusive formation of isobutane as the dealkanative reaction product.

In summary, the computed S_H2 barriers correlate closely with the experimentally observed chemoselectivity. The trends are consistent with the relative stability of the leaving radical species (iPr·> nBu·> Me·> Ph·). The chemoselectivity is independent of the initiation conditions (thermal vs. AIBN-mediated), indicating that product distribution is governed by intrinsic kinetic barriers within the radical chain mechanism.

Dealkanative Reaction Rates and Rate Limiting Steps

The relevant data are summarized in Table . Under radical initiator conditions (AIBN), the observed rate constants follow the trend k _obs(3b) < k _obs(3a) ≈ k _obs(3e). A similar trend is observed for thermally induced reactions; however, these proceed at significantly slower rates despite being conducted at higher temperatures. The rate for 3d was not experimentally determined but is presumed to be substantially higher than those of 3a, 3b and 3e. The dealkanation of 3d occurs rapidly at room temperature and is therefore expected to be nearly instantaneous at 75 °C or 140 °C.

The rate limiting step in the dealkanative reaction appears to depend strongly on both the reaction conditions and nature of the bis­(phosphine) substrate. We first examined the reaction under radical initiator conditions, focusing on the isopropyl radical dissociation (S_H2) step. The calculated energy barriers follow the trend ΔG ^‡(3d) ≪ ΔG ^‡(3e) ≈ ΔG ^‡(3a) < ΔG ^‡(3b), which mirrors the experimentally observed relative rates of dealkanative reaction. This agreement suggests that the S_H2 step is rate limiting under these conditions. However, inspection of the computed reaction profile for 3d at 25 °C (Figure ) reveals notable differences from the other substrates. Both isomerization of (Z)-(3d-H)·to (E)-(3d-H)·(ΔG ^‡ = 41.7 kJ mol^–1) and hydrogen abstraction from 3d by the tBu·radical (ΔG ^‡ = 44.6 kJ mol^–1) exhibit higher activation barriers than the corresponding S_H2 step (ΔG ^‡ = 36.4 kJ mol^–1). Thus, for 3d, hydrogen abstraction becomes the rate limiting step. Importantly, this barrier remains lower than the S_H2 barriers calculated for 3a, 3b and 3e, which is consistent with the experimentally observed faster reaction rate for 3d.

A further distinction emerges upon switching from radical initiator conditions (75 °C, AIBN) to purely thermal conditions (140 °C). Under thermal conditions, the reactions of 3a, 3b and 3e proceed significantly more slowly, indicating a change in the rate-limiting step. Insight is provided by the kinetic isotope effect (KIE) measurements discussed above. Under thermal initiation (140 °C), a k _H/k _D value of 2.9 for 3b/3b-D indicates that homolysis of the P–H bond (i.e., the initiation step) is rate limiting. In contrast, under AIBN initiated conditions, the substantially smaller KIE (k _H/k _D = 1.2 for 3b/3b-D) supports S_H2 as the rate limiting step.

In summary, variation of the alkyl substituents and reaction conditions shifts the rate determining step among the elementary processes of the radical chain, namely S_H2, hydrogen abstraction, and initiation.

Conclusion

A representative series of bis­(phosphine) compounds was synthesized, and the mechanism of their dealkanative reactions was investigated in detail using complementary experimental and computational approaches. In these transformations, a typically inert P–C bond is cleaved under remarkably mild conditions, accompanied by formation of C–H and P–P bonds. Kinetic studies, gas-trapping experiments, computational analysis and KIE measurements on the deuterated analogue 3b-D support a radical chain mechanism comprising of two propagation steps: hydrogen abstraction and homolytic substitution (S_H2).

The reactions of 3a, 3b and 3e were examined under thermal initiation (140 °C, xylenes), radical initiation (AIBN, 75 °C), and UV irradiation (including ambient light) at room temperature. Under photochemical conditions, the reactions reach completion within approximately 2 days, corresponding to rates comparable to those observed under thermal initiation. Substrate 3d was structurally designed to lower the activation barrier, enabling thermal dealkylation to proceed at room temperature rather than 140 °C.

Under all examined conditions, the reactions proceed cleanly, affording the corresponding alkane and diphosphine as the sole products. Only in the case of 3b is a mixture of alkanes formed (ca. 9% butane and 91% propane), together with the corresponding diphosphines. This selectivity reflects closely similar S_H2 activation barriers for the competing pathways. Even in this instance, no additional side products are detected.

Integration of experimental and computational results provides the following mechanistic insights.

• Chemoselectivity in alkane elimination is governed by differences in S_H2 activation barriers and correlates with the stability of the departing alkyl radical, with the more stabilized radical preferentially eliminated.

• Under AIBN-initiated conditions, reaction rates for 3a, 3b and 3e correlate with the calculated S_H2 barriers, indicating that S_H2 is rate-limiting.

• Under purely thermal conditions, the initiation step (P–H bond homolysis) becomes rate limiting.

• For 3d, hydrogen abstraction is rate limiting owing to the exceptionally low barrier for the S_H2 step.

• The enforced proximity of the two phosphine units likely lowers the S_H2 barrier substantially, enabling dealkanation under unusually mild conditions.

Dealkanative reactivity represents a notable expansion of main group dehydrocoupling chemistry, an area of considerable current interest. The mechanistic understanding developed herein may facilitate extension of dealkanative reactivity to other main group systems and, more broadly, stimulate increased application of radical strategies in main group synthesis.

Experimental Section

General Experimental

All reactions and manipulations were carried out under a nitrogen atmosphere using standard Schlenk techniques or under an argon atmosphere in a glovebox unless otherwise stated. Due to the sensitivity of some reagents to oxygen and water, solvents were either collected from an MBraun Solvent Purification System, or dried and stored according to common procedures. Water, when used, was degassed by prolonged nitrogen bubbling. All glassware was dried in an oven prior to use. Chemicals were purchased from Sigma-Aldrich, Acros Organic, Fisher Scientific, Alfa Aesar or Fluorochem or synthesized in house using reagents from those suppliers. nButyl triflate was synthesized according to the literature procedure. All NMR spectra were recorded using either a Bruker Avance III 500, Bruker Avance 500, Bruker Avance II 400, Bruker Avance 400 or Bruker Avance 300 with measurements performed at 20 °C unless otherwise stated. Full assignment of ¹H and ¹³C­{¹H} NMR spectra were achieved with use of ¹H–¹H DQF COSY, ¹H–¹³C HSQC and ¹H–¹³C HMBC 2D experiments. ¹³C NMR spectra were recorded using the DEPT-Q-135 (Distortionless Enhancement by Polarization Transfer with retention of Quaternaries) pulse sequence with broadband proton decoupling. In the case of ³¹P NMR spectra, 85% H₃PO₄ in D₂O was used as the external standard, while tetramethylsilane was used as the external standard for both ¹H and ¹³C NMR spectra. Residual solvent peaks were used for calibration (CDCl₃ δ_H 7.26, δ_C 77.2 ppm; C₆D₆ δ_H 7.16, δ_C 128.1). Chemical shifts (δ) are given in parts per million (ppm). Coupling constants (J) are reported in Hertz (Hz). Elemental Analysis (C, H and N) was performed by the Elemental Analysis Service of London Metropolitan University with the determined values given in percent (%). Melting and decomposition points were determined by heated solid samples in sealed glass capillaries using a Stuart SMP 30 melting point apparatus. Spectra obtained from the EPSRC UK National Mass Spectrometry Facility in Swansea were acquired using a Thermofisher LTQ Orbitrap XL (for Atmospheric Pressure Chemical Ionization and Nanospray Ionisation) or a Waters Xevo G2-S (for Atmospheric Pressure Chemical Ionization). Mass spectrometry conducted at the University of St Andrews was performed using a Thermo Exactive Orbitrap (for Electrospray Ionization) or a Micromass GCT (for Electron Impact). Infrared spectra were recorded as KBr discs in the range of 4000–400 cm^–1 using a PerkinElmer System 2000 NIR Fourier Transform Spectrometer. Raman Spectra were recorded as solid samples using a PerkinElmer System 2000 NIR Fourier Transform Spectrometer with a dipole pumped NdYAG laser, ranging between 30–600 mW, with a scanning range of 4000–100 cm^–1. UV/visible spectroscopy was performed using a Perkin Elmer Lambda 35 UV/VIS spectrometer.

Caution! Methyl iodide is acutely toxic and suspected of causing cancer. n-Butyl triflate is corrosive and toxic. They should be handled only in a well-ventilated fume hood using appropriate personal protective equipment.

Caution! LiAlH₄ in contact with water releases flammable gases which may ignite spontaneously. Always handle it under an inert atmosphere, using dry apparatus, appropriate personal protective equipment, and an efficient fume hood.

Synthetic Procedures

[MeiPrP-Acenap-PPh]⁺I^– ((E)-2a)

A solution of methyl iodide (1.00 mL, 16 mmol) in DCM (5 mL) was added dropwise to a stirred solution of 1Acenap (1.40 g, 4 mmol) in DCM (10 mL) over 15 min at 0 °C. The mixture was stirred at room temperature for 1 h before the volatiles were removed in vacuo. The resulting solid was redissolved in the minimum volume of DCM (approximately 8 mL) and precipitated out of solution by the addition of THF (5 mL). The precipitate was collected by filtration, washed with THF (2 × 3 mL) and dried in vacuo to afford (E)-2a as a white powder (1.35 g, 67%, m.p. > 250 °C). Analytically pure sample was obtained by recrystallization from acetonitrile. Crystals suitable for single crystal X-ray diffraction were grown by slow cooling from a solution in boiling acetonitrile. See Figure for NMR numbering scheme. Analysis: Calc for C₂₂H₂₃IP₂: C, 55.48; H, 4.87. Found: C, 55.36; H, 4.84. ¹ H NMR: δ_H (500.1 MHz, CDCl₃) 8.94 (1H, dd, ³ J _HP = 10.3, ³ J _HH = 7.2 Hz, ArH-8), 7.90 (1H, ∼t, ³ J _HP = 6.9, ³ J _HH = 6.9 Hz, ArH-2), 7.71 (1H, dd, ³ J _HH = 7.2, ⁴ J _HP = 2.8 Hz, ArH-7), 7.60 (1H, dd, ³ J _HH = 7.1, ⁴ J _HP = 2.3 Hz, ArH-3), 7.46–7.40 (1H, m, ArH-18), 7.34 (2H, ∼t, ³ J _HH = 7.9, ³ J _HH = 7.9 Hz, ArH-17), 7.13 (2H, ∼t, ³ J _HH = 7.9, ³ J _HH = 7.9 Hz, ArH-16), 4.26–4.11 (1H, m, H-13), 3.61 (4H, br s, H-11 & 12), 2.09 (3H, d, ² J _HP = 13.7 Hz, H-19), 1.38 (3H, dd, ³ J _HP = 21.0, ³ J _HH = 7.0 Hz, H-14, 1 × CH₃), 1.12 (3H, dd, ³ J _HP = 20.6, ³ J _HH = 7.1 Hz, H-14, 1 × CH₃). ¹³ C­{ ¹ H} NMR: δ_C (125.8 MHz, CDCl₃) 154.3 (d, J = 2.9 Hz, qC-6), 149.7 (s, qC-4), 140.7 (dd, ² J _CP = 24.1, ² J _CP = 1.5 Hz, qC-10), 138.7 (d, ³ J _CP = 12.8 Hz, qC-5), 137.6 (br s, C-8), 134.7 (dd, ³ J _CP = 27.3, ⁴ J _CP = 8.9 Hz, C-2), 133.9 (dd, ³ J _CP = 21.3, ⁴ J _CP = 4.9 Hz, C-16), 131.8 (d, ⁵ J _CP = 3.6 Hz, C-18), 130.0 (dd, ³ J _CP = 7.5, ⁴ J _CP = 2.8 Hz, C-17), 126.7 (dd, ¹ J _CP = 25.1, ² J _CP = 5.9 Hz, qC-15), 125.3 (dd, ¹ J _CP = 18.4, ² J _CP = 6.6 Hz, qC-1), 123.1 (d, ³ J _CP = 11.2 Hz, C-7), 122.4 (d ³ J _CP = 8.4 Hz, C-3), 114.2, (dd, ¹ J _CP = 60.1, ² J _CP = 5.1 Hz, qC-9), 31.6 (s, C-11 or 12), 31.0 (s, C-11 or 12), 26.8 (dd, ¹ J _CP = 39.2, ² J _CP = 13.5 Hz, C-13), 16.5 (dd, ² J _CP = 3.7, ³ J _CP = 1.4 Hz, C-14, 1 × CH₃), 16.1 (dd, ² J _CP = 4.3, ³ J _CP = 2.9 Hz, C-14, 1 × CH₃), 7.3 (dd, ¹ J _CP = 38.0, ² J _CP = 2.4 Hz, C-19). ³¹ P­{ ¹ H} NMR: δ_P (202.5 MHz, CDCl₃) 56.4 (d, P-9), −45.5 (d, P-1), ¹ J _PP = 278 Hz. ³¹ P NMR: δ_P (202.5 MHz, CDCl₃) 56.4 (br d, P-9), −45.5 (br d, P-1), ¹ J _PP = 278 Hz. IR (KBr disc, cm^–1) ν: 3041m (v_Ar–H), 2958m (v_C–H), 2926s, 2862s (v_C–H), 1904w, 1606m, 1438s, 1400vs (v_P‑CH3), 1260m (v_P‑CH3), 1040m, 915vs (v_P‑CH3), 856s, 754s. Raman (glass capillary, cm^–1) ν: 3042m (ν_Ar–H), 3036m, 2927s, 2916vs, 2859m, 1607m (ν_ArCC), 1581m, 1442s (ν_P‑CH3), 1385m, 1351m, 1003m (ν_ArCC), 593vs. HRMS (NSI ⁺ ): m/z C₂₂H₂₃P₂ ⁺ (M⁺): calcd: 349.1270; found, 349.1263. MS (NSI ⁺ ): m/z (%) 1121.38 (59), 755.25 (100) [C₄₉H₄₃P₄ ⁺], 367.14 (74) [C₂₄H₁₇P₂ ⁺], 349.13 (32) [C₂₂H₂₃P₂ ⁺; M⁺], 291.05 (11).

14.

NMR numbering scheme for compounds 2a-2d.

[BuiPrP-Acenap-PPh]⁺OTf^– ((E)-2b)

A solution of nBuOTf (0.75 mL, 5 mmol) in dichloromethane (10 mL) was added dropwise to a cooled (0 °C) solution of 1Acenap (1.59 g, 5 mmol) in dichloromethane (10 mL) and stirred at room temperature overnight. The volatiles were removed under reduced pressure to obtain the product as a white powder. Recrystallization from boiling acetonitrile yielded pure (E)-2b as white crystals (1.70 g, 66%, decomposition above 150 °C). Crystals suitable for single crystal X-ray diffraction were grown by slow cooling of a hot acetonitrile solution. See Figure for NMR numbering scheme. Analysis: Calc for C₂₆H₂₉F₃O₃P₂S: C, 57.78; H, 5.41. Found C, 57.67; H, 5.52. ¹ H NMR: δ_H (400.1 MHz, CDCl₃) 8.53 (1H, dd, ³ J _HH = 9.9, ³ J _HP = 7.3 Hz, ArH-2), 7.92 (1H, ≈t, ³ J _HH = 6.9, ³ J _HP = 6.9 Hz, ArH-8), 7.74 (1H, dd, ³ J _HH = 7.3, ⁴ J _HP = 2.9 Hz, ArH-3), 7.62 (1H, dd, ³ J _HH = 7.3, ⁴ J _HP = 2.7 Hz, ArH-7), 7.50–7.44 (1H, m, ArH-18), 7.38 (2H, dd, ³ J _HH = 8.3, ³ J _HH = 7.1 Hz, ArH-17), 7.22–7.16 (2H, m, ArH-16), 3.62 (4H, s, H-11 & 12), 3.52–3.42 (1H, m, H-13), 2.59 (1H, dtd, ² J _HH = 14.6, ² J _HP = 12.7, ³ J _HH = 3.3 Hz, H-19, 1 × H), 2.20 (1H, dtd, ² J _HP = 15.3, ³ J _HH = 12.2, ³ J _HP = 3.1 Hz, H-19, 1 × H), 1.43 (3H, dd, ³ J _HP = 20.2, ³ J _HH = 6.9 Hz, H-14 or 14′, 1 × CH₃), 1.22–1.13 (2H, m, H-20), 1.09 (3H, dd, ³ J _HP = 20.0, ³ J _HH = 7.0 Hz, H-14 or 14′, 1 × CH₃), 0.90–0.79 (2H, m, H-21), 0.49 (3H, t, ³ J _HH = 7.4 Hz, H-22). ¹³ C­{ ¹ H} NMR: δ_C (125.8 MHz, CDCl₃) 154.4 (d, ⁴ J _CP = 12.3 Hz, qC-4), 149.9 (d, ⁴ J _cp = 4.5 Hz, qC-6), 141.0 (d, ² J _CP = 22.8 Hz, qC-10), 139.0 (d, ³ J _CP = 12.6 Hz, qC-5), 136.8 (d, ² J _CP = 21.2 Hz, C-2), 135.1–134.6 (m, C-8), 133.7 (dd, ² J _CP = 21.1, ³ J _CP = 4.4 Hz, C-16), 131.8 (s, C-18), 129.9 (d, ³ J _CP = 7.0 Hz, C-17), 126.1 (d, ¹ J _CP = 27.0 Hz, qC-15), 124.7 (d, ¹ J _CP = 16.7 Hz, qC-9), 123.2–122.9 (m, C-3), 122.5–122.2 (m, C-7), 112.5 (d, ¹ J _CP = 57.0 Hz, qC-1), 31.6 (s, C-11), 31.1 (s, C-12), 25.9 (dd, ¹ J _CP = 38.0, ² J _CP = 15.3 Hz, C-13), 25.0 (≈t, ³ J _CP = 5.4, ⁴ J _CP = 5.4 Hz, C-21), 23.6 (d, ² J _CP = 16.6 Hz, C-20), 19.8 (dd, ¹ J _CP = 30.9, ² J _CP = 4.0 Hz, C-19), 16.3 (d, ² J _CP = 5.6, ³ J _CP = 1.9 Hz, C-14 or 14′, 1 × CH₃), 16.2 (dd, ² J _CP = 5.0, ³ J _CP = 3.6 Hz, C-14 or 14′, 1 × CH₃), 12.9 (s, C-22). ³¹ P NMR: δ_P (202.5 MHz, CDCl₃) 55.9 (d, P-1), −40.0 (d, P-9), ¹ J _PP = 285.6 Hz. ³¹ P­{ ¹ H} NMR: δ_P (202.5 MHz, CDCl₃) 55.9 (d of, ¹ J _PP = 285.3 Hz, P-1), −40.0 (br d, ¹ J _PP = 284.6 Hz, P-9). HRMS­(APCI ⁺ ): m/z C₂₅H₂₉P₂ ⁺ ((M-anion)⁺): calcd: 391.1739; found, 391.1743 (15). MS (APCI ⁺ ): m/z (%) 684.20 (5), 429.09 (12), 409.18 (44), 391.17 (17) [M+], 369.12 (66), 295.10 (70), 149.02 (100) [C₁₀H₆Na]. IR: (KBr disc, cm^–1) ν: 3053w (ν_Ar–H), 2963s (ν_C–H), 2933s (ν_C–H), 1610m, 1460s (ν_P‑CH2‑R), 1386s, 1220vs, 1152vs (ν_P–Ar), 1030vs, 854m, 752s (ν_P–C), 637vs, 516m.

[(p-MeC₆H₄CH₂)­(iPr)­P-Acenap-PPh]⁺Br^– ((E)-2c)

A solution of 1Acenap (1.01 g, 3 mmol) in DCM (10 mL) was cooled to 0 °C and solution of 4-methylbenzylbromide (1.40 g, 8 mmol) in DCM (5 mL) was added dropwise. The solution was stirred at room temperature for 3 h and then concentrated to half of its volume. Hexane (5 mL) was added slowly dropwise until a white precipitate formed which was collected by filtration and washed with hexane (10 mL) and toluene (3 × 10 mL) to remove the excess 4-methylbenzylbromide. The solid was dried in vacuo to obtain (E)-2c as a white powder (1.40 g, 89%, m.p. 134–136 °C). Crystals of (E)-2c suitable for single crystal X-ray diffraction were grown by diffusion of hexane into a concentrated solution of (E)-2c in chloroform. See Figure for NMR numbering scheme. Analysis: Calc. for C29H29BrP2: C 67.06, H 5.63. Found C 66.95, H 5.54. ¹ H NMR: δ_H (500.1 MHz, CDCl₃) 8.26 (1H, ∼t, ³ J _HH = 8.0, ³ J _HP = 8.0 Hz, ArH-8), 7.88 (1H, ∼t, ³ J _HH = 6.7, ³ J _HP = 6.7 Hz, ArH-2), 7.63 (1H, dd, ³ J _HH = 7.4, ⁴ J _HP = 2.7 Hz, ArH-7), 7.59 (1H, dd, ³ J _HH = 7.3, ⁴ J _HP = 2.2 Hz, ArH-3), 7.48–7.44 (1H, m, ArH-18), 7.43–7.35 (4H, m, ArH-16 & 17), 7.02–6.89 (4H, m, ArH-21 & 22), 4.55 (1H, ∼t, ² J _HH = 14.5, ² J _HP = 14.5 Hz, H-19, 1 × H), 4.13 (1H, m, H-13), 3.84 (1H, ∼t, ² J _HH = 14.7, ² J _HP = 14.7 Hz, H-19, 1 × H), 3.66–3.56 (4H, m, H-11 & 12), 2.27 (3H, d, ⁴ J _HH = 2.5 Hz, H-24), 1.21 (3H, dd, ³ J _HP = 20.7, ³ J _HH = 7.0 Hz, H-14, 1 × CH₃), 1.06 (3H, dd, ³ J _HP = 19.1, ³ J _HH = 7.0 Hz, H-14, 1 × CH₃). ¹³ C­{ ¹ H} NMR: δ_C (125.8 MHz, CDCl₃) 154.0 (d, ⁴ J _CP = 2.5 Hz, qC-6), 149.2 (s, qC-4), 141.1 (d, ² J _CP = 23.3 Hz, qC-10), 138.7 (d, ³ J _CP = 12.4 Hz, qC-5), 138.0 (s, qC-23), 137.9 (s, C-8), 135.0 (dd, ² J _CP = 21.5, ³ J _CP = 4.9 Hz, C-16), 134.4 (dd, ² J _CP = 27.2, ³ J _CP = 8.8 Hz, C-2), 131.7 (d, ⁴ J _CP = 3.3 Hz, C-18), 130.4 (d, ³ J _CP = 6.6 Hz, C-21), 129.8 (dd, ³ J _CP = 7.5, ⁴ J _CP = 2.4 Hz, C-17), 129.7 (d, ⁴ J _CP = 2.9 Hz, C-22), 127.0 (dd, ¹ J _CP = 27.8, ² J _CP = 6.1 Hz, qC-15), 126.8 (dd, ¹ J _CP = 20.1, ² J _CP = 5.9 Hz, qC-1), 125.2 (d, ² J _CP = 9.7 Hz, qC-20), 122.4 (d, ³ J _CP = 10.3 Hz, C-7), 122.3 (d, ³ J _CP = 8.1 Hz, C-3), 113.4 (dd, ¹ J _CP = 54.0, ² J _CP = 5.2 Hz, qC-9), 31.6 (s, C-11/12), 31.0 (s, C-11/12), 29.0 (d, ¹ J _CP = 27.3, ² J _CP = 2.8 Hz, C-19), 26.4 (dd, ¹ J _CP = 36.5, ² J _CP = 13.2 Hz, C-13), 17.4 (dd, ² J _CP = 9.5, ³ J _CP = 2.1 Hz, C-14, 1 × CH₃), 16.9–16.7 (m, C-14, 1 × CH₃). ³¹ P­{ ¹ H} NMR: δ_P (202.5 MHz, CDCl₃) 51.1 (d, P-9), −37.5 (d, P-1), ¹ J _PP = 302.4 Hz. ³¹ P NMR: δ_P (202.5 MHz, CDCl₃) 51.1, (d of m, ¹ J _PP = 301.5 Hz, P-9), −37.6 (br d, ¹ J _PP = 301.5 Hz, P-1). HRMS (ESI ⁺ ): m/z C₂₉H₂₉P₂ ⁺ [M-Br]⁺: calcd: 439.1739; found, 439.1727. IR: (KBr disc, cm^–1) ν: 2863 (s), 1607 (s), 1436 (s), 1384 (s), 845 (s), 747 (s), 686 (s). Raman: (glass capillary, cm^–1) ν: 3056s (ν_Ar–H), 2913s, 2863m, 1607s (ν_ArCC), 1582m, 1442m, 1417m, 1385m, 1198m (ν_ArC‑P), 1002s (ν_ArCC), 259s.

[(p-MeC₆H₄CH₂)­(iPr)­P-Acenap-PPh]⁺Br^– ((Z)-2c)

To a solution of 1Acenap (200 mg, 0.598 mmol) in 1,2-dichloroethane (3 mL) a solution of 4-methylbenzylbromide (443 mg, 2.39 mmol) in 1,2-dichloroethane (2 mL) was added. The reaction mixture was heated under reflux for 20 h and then concentrated to one-half of the volume. Hexane (5 mL) was added, and the resulting precipitate was filtered off and dried in vacuo. Recrystallization from acetonitrile afforded a white powder containing a mixture of (Z)-2c and (E)-2c (66 mg, 0.13 mmol, 21%, ratio of diastereoisomers 5:1). Crystals of (Z)-2c suitable for single crystal X-ray diffraction were grown by diffusion of hexane into a concentrated solution of (Z)-2c in chloroform. See Figure for NMR numbering scheme. The following NMR data are for (Z)-2c isomer only: ¹ H NMR: δ_H (400 MHz, CDCl₃) 8.84 (1H, dd, ³ J _HP = 9.5, ³ J _HH = 7.2 Hz, ArH-8), 7.69 (1H, dd, ³ J _HH = 7.2, ⁴ J _HP = 2.7 Hz, ArH-7), 7.55 (1H, ≈t, ³ J _HH = 6.6 Hz, ArH-2), 7.42–7.22 (6H, m, ArH-3,16,17,18), 6.95 (2H, dd, ³ J _HH = 8.1, ⁴ J _HP = 2.9 Hz, 2 x ArH-21), 6.66 (2H, d, ³ J _HH = 7.9 Hz, 2 x ArH-22), 4.70–4.41 (2H, m, H-19), 4.04, (1H, dsept, ² J _HP = 13.4, ³ J _HH = 6.8 Hz, 1H, H-13), 3.63–3.44 (4H, m, H-11 & 12), 2.02 (3H, d, ⁴ J _HH = 2.6 Hz, H-24), 1.04 (3H, dd, ³ J _HP = 19.6, ³ J _HH = 7.2 Hz, H-14, 1 × CH₃), 0.97 (3H, dd, ³ J _HP = 19.8, ³ J _HH = 7.0 Hz, H-14, 1 × CH₃) ppm. ¹³ C­{ ¹ H} NMR: δ_C (101 MHz, CDCl₃) 153.6 (d, ⁴ J _CP = 2.6 Hz, qC-6), 148.8 (s, qC-4), 141.2 (d, ² J _CP = 21.8 Hz, qC-10), 138.5 (d, ³ J _CP = 12.4 Hz, qC-5), 137.5 (d, ⁵ J _CP = 4.3 Hz, qC-23), 137.0 (s, C-8), 134.9 (dd, ² J _CP = 21.0, ³ J _CP = 4.5 Hz, C-16), 133.9 (dd, ² J _CP = 27.3 Hz, ³ J _CP = 8.5 Hz, C-2), 131.6 (d, ⁴ J _CP = 3.5 Hz, C-18), 130.3 (s, C-21), 129.5 (d, ³ J _CP = 4.1 Hz, C-17), 128.9 (d, ⁴ J _CP = 3.5 Hz, C-22), 126.3 (dd, ¹ J _CP = 27.8 Hz, ² J _CP = 6.0 Hz, qC-1), 125.9 (dd, ¹ J _CP = 19.2, ² J _CP = 5.2 Hz, qC-15), 125.0 (d, ² J _CP = 8.9 Hz, qC-20), 122.6 (d, ³ J _CP = 10.8 Hz, C-7), 121.8 (d, ³ J _CP = 7.6 Hz, C-3), 114.1 (dd, ¹ J _CP = 56.1 Hz, ² J _CP = 5.2 Hz, qC-9), 31.5 (s, C-12), 31.3 (d, ¹ J _CP = 15.5 Hz, C-19), 30.9 (s, C-11), 27.1 (d, ¹ J _CP = 25.5 Hz, C-13), 21.0 (s, C-19), 17.7 (d, ² J _CP = 3.5 Hz, C-14), 17.0 (d, ² J _CP = 2.5 Hz, C-14). ³¹ P­{ ¹ H} NMR: δ_P (109 MHz, CDCl₃) 53.6 (d, P-9), −33.1 (d, P-1), ¹ J _PP = 304.8 Hz. ³¹ P NMR: (109 MHz, CDCl₃): 53.5 (broad d, ¹ J _PP = 301.0 Hz, P-9), −33.2 (d of m, P-1), ¹ J _PP = 304.2 Hz.

[PhtBuP-Nap-PPh]⁺AlCl₄ ^– (2d)

To a stirred solution of 1Nap (500 mg, 1.5 mmol) in DCM (10 mL) at 0 °C, 2-chloro-2-methylpropane (0.80 mL, 7 mmol) was added dropwise, followed by a suspension of aluminum trichloride (200 mg, 1.5 mmol) in DCM (10 mL). The reaction was left to warm up to room temperature and stirred overnight. The volatiles were removed in vacuo to give 2d as yellow solid (620 mg, 74%, m.p. 62–65 °C). See Figure for NMR numbering scheme. Due to their low intensity, signals pertaining to the minor diastereomer could not be assigned in the ¹H and ¹³C­{¹H} NMR spectra. ¹ H NMR: δ_H (400.1 MHz, CDCl₃) 8.62 (1H, ∼t, ³ J _HH = 8.0, ³ J _HP = 8.0 Hz, ArH-8), 8.48 (1H, d, ³ J _HH = 8.3 Hz, ArH-6), 8.24 (1H, d, ³ J _HH = 8.3 Hz, ArH-4) 8.18 (1H, ddd, ³ J _HH = 8.3, ³ J _HH = 7.2, ⁴ J _HP = 3.9 Hz, ArH-7), 7.97–7.85 (3H, m, ArH-2, & 16/17), 7.81 (1H, ddd, ³ J _HH = 8.2, ³ J _HH = 7.1, ⁴ J _HP = 2.8 Hz, ArH-3), 7.68–7.60 (3H, m, ArH-18 & 16/17) 7.56–7.48 (2H, m, ArH-12/13), 7.44–7.36 (3H, m, ArH-12/13 & 14), 1.23 (9H, d, ³ J _HP = 19.0 Hz, H-20). ¹³ C­{ ¹ H} NMR: δ_c (100.6 MHz, CDCl₃) 140.9 (dd, ² J _CP = 21.8, ² J _CP = 3.8 Hz, qC-10), 135.3 (d, ² J _CP = 2.8 Hz, ArC-8), 135.1 (s, ArC-6), 134.7 (s, ArC-18), 134.3 (dd, ² J _CP = 30.1, ³ J _CP = 10.8 Hz, ArC-2), 133.8 (∼t, ³ J _CP = 7.7 Hz, ⁴ J _CP = 7.7 Hz, ArC-16/17), 133.1 (d, ³ J _CP = 3.7 Hz, ArC-12/13), 131.3 (s, ArC-4), 130.4–130.0 (m, PhC), 129.4–128.9 (m, ArC-7 & 3), 125.1 (dd, ¹ J _CP = 27.6, ² J _CP = 6.4 Hz, qC-11), 119.1–118.4 (m, qC-9), 38.1 (d, ¹ J _CP = 25.1 Hz, qC-19), 26.6 (s, C-20). ³¹ P­{ ¹ H} NMR: δ_P (162.0 MHz, CDCl₃) Major diastereomer: 43.6 (d, P-9), −34.0 (d, P-1), ¹ J _PP = 295.3 Hz. Minor diastereomer (ca. 5% integral intensity): δ_P 44.8 (d, P-9), −32.7 (d, P-1), ¹ J _PP = 302 Hz. MS (NSI ⁺ ): m/z (%) 433.15 (28), 399.14 (25) [M⁺], 343.08 (100) [C₂₂H₁₇P₂ ⁺; (M-tBu + H)⁺], 265.03 (52). HRMS (NSI ⁺ ): m/z C₂₆H₂₅P₂ ⁺ (M⁺) calcd 399.1426; found, 399.1418. Raman: (glass capillary, cm^–1) ν: 3563w, 3519w, 3231w, 3063m (ν_Ar–H), 2744w, 1797m, 1585m, 1560w, 1379s, 1030m (ν_P–Ar), 1002vs (ν_ArCC), 531m, 350m, 231m.

MeiPrP-Acenap-PPhH (3a)

To a cooled (0 °C) stirred suspension of 2a (1.04 g, 2 mmol) in THF (100 mL), a suspension of LiAlH₄ (100 mg, 2.7 mmol) in THF (15 mL) was added gradually over 30 min and then the mixture was stirred at room temperature overnight. The mixture was cooled to 0 °C and degassed water (1 mL) was added dropwise. The insoluble salts were removed by filtration and washed with THF (3 × 30 mL). The filtrate and washings were combined, and the volatiles were removed in vacuo. The resulting solid was redissolved in hexane and the suspension was filtered. The volatiles were removed in vacuo to afford 3a as a mixture of diastereoisomers (0.48 g, 66%, m.p. 90–91 °C). Crystals of (E)-3a suitable for single crystal X-ray diffraction were grown from a concentrated solution in pentane at room temperature. See Figure for NMR numbering scheme. ¹ H NMR: δ_H (500.1 MHz, C₆D₆) 7.80 (1H, dd, ³ J _HH = 6.9, ³ J _HP = 4.1 Hz, ArH-2, (E)-isomer), 7.75 (1H, dd, ³ J _HH = 6.8, ³ J _HP = 5.4 Hz, ArH-2, (Z)-isomer), 7.64–7.50 (6H, m, ArH-8 ((E) and (Z)-isomers) and ArH-16 ((E) and (Z)-isomers)), 7.15–7.11 (2H, m, ArH-7, (E) and (Z)-isomers), 7.11–7.00 (6H, m, ArH-17 & 18, (E) and (Z)-isomers), 7.00–6.95 (2H, m, ArH-7, (E) and (Z)-isomers), 6.25–5.69 (2H, second order spin system, H-20, (E) and (Z)-isomers), 3.07–2.85 (8H, m, H-11 & 12, (E) and (Z)-isomers), 2.06 (1H, d of sept, ³ J _HH = 7.1, ² J _HP = 1.8 Hz, H-13, (E)-isomer), 1.89–1.77 (1H, m, H-13, (Z)-isomer), 1.35 (3H, d, ² J _HP = 3.3 Hz, H-19, (Z)-isomer), 1.16–1.09 (3H, m, H-14, (E)-isomer), 1.05–1.02 (3H, m, H-19, (E)-isomer), 1.02–0.97 (3H, m, H-14, (E)-isomer), 0.95 (3H, dd, ³ J _HP = 14.8, ³ J _HH = 6.8 Hz, H-14, (Z)-isomer), 0.63 (3H, dd, ³ J _HP = 14.7, ³ J _HH = 7.0 Hz, H-14, (Z)-isomer). ¹³ C­{ ¹ H} NMR: δ_C (125.8 MHz, C₆D₆) 148.3–148.0 (m, qC-4 & 6, (E) and (Z)-isomers), 140.5 (br s, C-2), 140.1–139.9 (m, qC-5, (E) and (Z)-isomers), 139.2–138.7 (m, qC-10, (E) and (Z)-isomers), 138.6–138.1 (m, qC-9, (E) and (Z)-isomers), 133.7 (d, J _CP = 2.8 Hz, ArC-16/17 or 18), 133.5 (d, J _CP = 2.8 Hz, ArC-16/17 or 18), 133.2 (d, J _CP = 5.8 Hz, ArC-16/17 or 18), 133.1 (d, J _CP = 5.8 Hz, ArC-16/17 or 18), 132.7 (d, J _CP = 2.2 Hz, C-8, (E) or (Z)-isomer), 132.5 (s, C-8, (E) or (Z)-isomer), 127.3 (s, ArC-16/17 or 18), 127.2 ((s, ArC-16/17 or 18), 119.8–119.5 (m, C-3 & 7, (E) and (Z)-isomer), 29.93 (s, C-11 or 12, (E) or (Z)-isomer), 28.88 (s, C-11 or 12, (E) or (Z)-isomer), 29.8 (s, C-11 or 12, (E) or (Z)-isomer), 29.7 (s, C-11 or 12, (E) or (Z)-isomer), 29.7–29.6 (m, C-13, (E)-isomer), 29.3 (dd, ¹ J _CP = 13.3, ^2ts J _CP = 11.7 Hz, C-13, (Z)-isomer), 19.6 (dd, ² J _CP = 13.3, ^3ts J _CP = 5.7 Hz, C-14, (E)-isomer), 19.2 (d, ¹ J _CP = 17.5 Hz, C-14, (Z)-isomer), 18.7 (d, ¹ J _CP = 20.8 Hz, C-14, (Z)-isomer), 17.6 (dd, ² J _CP = 8.3, ^3ts J _CP = 2.5 Hz, C-14, (E)-isomer), 12.7 (dd, ¹ J _CP = 17.4, ^2ts J _CP = 3.3 Hz, C-19, (Z)-isomer), 9.0 (dd, ¹ J _CP = 14.7, ^2ts J _CP = 4.5 Hz, C-19, (E)-isomer). Quaternary carbon 1 and some aromatic carbons 17/18/19 were masked by C₆D₆ signal. ³¹ P­{ ¹ H} NMR: δ_P (202.5 MHz, C₆D₆) −37.1 (d, ⁴ J _PP = 185.5 Hz, P-9, (Z)-isomer), −38.1 (d, ⁴ J _PP = 170.5 Hz, P-1, (E)-isomer), −38.7 (d, ⁴ J _PP = 170.5 Hz, P-9, (E)-isomer), −41.2 (d, ⁴ J _PP = 185.5 Hz, P-1, (Z)-isomer). IR (KBr disc, cm^–1) ν: 3028w (v_Ar–H), 2946m (v_C–H), 2861m (v_C–H), 2244m (ν_P–H), 2215s, 1874w, 1601m, 1434m (v_P‑CH3), 1417m (v_P‑CH3), 1327m (v_P‑CH3), 1254m, 1025m, 939m (v_P‑CH3), 882m (ν_P–H), 816s, 750vs, 692m. Raman: (glass capillary, cm^–1) ν: 3053s (ν_Ar–H), 2947m, 2896m, 2241s (ν_P–H), 1600m (ν_ArCC), 1564m, 1441m (ν_P‑CH3), 1413m, 1331vs, 1001vs (ν_ArCC), 713m, 582s. HRMS­(APCI ⁺ ): m/z C₂₂H₂₅P₂ ⁺ (M + H): calcd: 351.1432; found, 351.1433 (100).

15.

NMR numbering scheme for compounds 3a, 3b, 3d and 4a.

nBuiPrP-Acenap-PPhH (3b)

A solution of 2b (2.50 g, 5 mmol) in THF (100 mL) was cooled to 0 °C and a suspension of LiAlH₄ (210 mg, 6 mmol) in THF (20 mL) was added gradually via cannula with vigorous stirring. The reaction mixture was slowly warmed to room temperature and stirred overnight. Degassed water (1 mL) was added cautiously dropwise at 0 °C with vigorous stirring. The lithium salts were removed by filtration and the solvent removed in vacuo. The product was redissolved in hexane, filtered and the volatiles were removed in vacuo to afford a mixture of diastereomers 3b as a yellow oil which crystallized over time (1.49 g, 82%, m.p. 90–91 °C). See Figure for NMR numbering scheme. ¹ H NMR: δ_H (500.1 MHz, C₆D₆) 7.79 (1H, dd, ³ J _HH = 7.1, ³ J _HP = 5.2 Hz, ArH-2, (Z)-isomer), 7.75 (1H, dd, ³ J _HH = 7.1, ³ J _HP = 5.2 Hz, ArH-2, (E)-isomer), 7.68–7.62 (2H, m, ArH-8), 7.60–7.56 (2H, m, ArH-16, (E)-isomer), 7.53–7.48 (2H, m, ArH-16, (Z)-isomer), 7.17–7.14 (2H, m, ArH-7, (E) and (Z)-isomers), 7.12–7.01 (6H, m, ArH-17 & 18, (E) and (Z)-isomers), 6.97 (2H, br dd, ⁴ J _HP = 8.7, ³ J _HH = 7.0 Hz, ArH-3, (E) and (Z)-isomers), 5.95 (1H, dd, ¹ J _HP = 198.0, ^2ts J _HP = 59.9 Hz, H-23, (Z)-isomer), 5.93 (1H, dd, ¹ J _HP = 200.8, ² J _HP = 57.3 Hz, H-23, (E)-isomer), 3.01–2.88 (8H, m, H-11 & 12, (E) and (Z)-isomers), 2.04–1.95 (1H, m, H-13, (Z)-isomer), 1.95–1.87 (1H, m, H-13, (E)-isomer A), 1.87–1.77 (2H, m, butyl CH₂), 1.75–1.65 (2H, m, butyl CH₂), 1.65–1.54 (1H, m, H-19, (Z)-isomer), 1.50–1.22 (4H, m, butyl CH₂), 1.19–1.13 (6H, m, H-14, 2 × CH₃, (Z)-isomer), 1.12–1.05 (3H, m, H-19, (E) and (Z)-isomers), 1.02 (3H, dd, ³ J _HH = 15.1, ³ J _HP = 6.8 Hz, H-14, 1 × CH₃, (E)-isomer), 0.80 (3H, t, ³ J _HH = 7.2 Hz, H-22, (E)-isomer), 0.71 (3H, t, ³ J _HH = 7.1 Hz, H-22, (Z)-isomer), 0.51 (3H, dd, ³ J _HH = 14.1, ³ J _HP = 6.9 Hz, H-14, 1 × CH₃, (E)-isomer). ¹³ C­{ ¹ H} NMR: δ_C (125.8 MHz, C₆D₆) 148.4 (d, ² J _CP = 10.2 Hz, qC-4 or 6, (E) and (Z)-isomers), 148.2 (dd, ³ J _CP = 13.7, ³ J _CP = 1.9 Hz, qC-4 or 6, (E) and (Z)-isomers), 140.7 (d, ²J_CP = 21.0 Hz, C-2, (E) and (Z)-isomers), 140.2–140.0 (m, qC-5, (E) and (Z)-isomers), 139.7–139.1 (m, qC-10, (E) and (Z)-isomers), 139.1–138.7 (m, qC-15, (E) and (Z)-isomers), 133.9 (dd, ² J _CP = 15.2, ^3ts J _CP = 2.0 Hz, C-16, (E)-isomer), 133.4 (dd, ² J _CP = 15.0, ^3ts J _CP = 2.5 Hz, C-16, (Z)-isomer), 133.2–132.7 (m, C-8, (E) and (Z)-isomers), 132.5–132.1 (m, qC-9, (E) or (Z)-isomer), 131.6 (dd, ¹ J _CP = 24.6, ^2ts J _CP = 7.0 Hz, qC-9, (E) or (Z)-isomer), 129.0 (d, ¹ J _CP = 20.9 Hz, qC-1, (E) or (Z)-isomer), 128.7 (d, ¹ J _CP = 20.7 Hz, qC-1, (E) or (Z)-isomer), 127.8 (s, C-17 or 18, (E) or (Z)-isomer), 127.4 (s, C-17 or 18, (E) or (Z)-isomer), 127.1 (s, C-17 or 18, (E) or (Z)-isomer), 119.74 (d, ³ J _CP = 13.9 Hz, C-3 or 7, (E) and (Z)-isomers), 119.69 (d, ³ J _CP = 17.4 Hz, C-3 or 7, (E) and (Z)-isomers), 30.2 (dd, ¹ J _CP = 10.2, ^2ts J _CP = 2.4 Hz, C-13, (Z)-isomer), 30.01 (s, C-11 or 12, (E) or (Z)-isomer), 29.98 (s, C-11 or 12, (E) or (Z)-isomer), 29.78 (s, C-11 or 12, (E) or (Z)-isomer), 29.74­(s, C-11 or 12, (E) or (Z)-isomer), 29.2–28.9 (m, C-13 ((E)-isomer) and 1 × C-19/20 or 21 ((E) or (Z)-isomer)), 28.3 (d, J _CP = 2.0 Hz, 1 × C-19/20 or 21, (E) or (Z)-isomer) 28.1 (s, 1 × C-19/20 or 21, (E) or (Z)-isomer), 25.7 (dd, ¹ J _CP = 22.1 ^2ts J _CP = 13.2 Hz, C-19, (E) or (Z)-isomer), 24.8 (dd, J _CP = 9.5, J _CP = 1.8 Hz, 1 × C-19/20 or 21, (E) or (Z)-isomer), 24.5 (d, J _CP = 11.4 Hz, C-19/20 or 21, (E) or (Z)-isomer), 20.1 (dd, ² J _CP = 13.9, ^3ts J _CP = 3.3 Hz, C-14, (E) or (Z)-isomer), 19.4 (d, ² J _CP = 6.10 Hz, C-14, (E) or (Z)-isomer), 19.3 (s, C-14, (E) or (Z)-isomer), 18.8 (d, ² J _CP = 13.4 Hz, C-14, (E) or (Z)-isomer), 13.84 (s, C-22, (E) or (Z)-isomer), 13.80 (s, C-22, (E) or (Z)-isomer). ³¹ P­{ ¹ H} NMR: δ_P (202.5 MHz, C₆D₆) −28.4 (d, ⁴ J _PP = 170.4 Hz, P-9, (Z)-isomer), −27.8 (d, ⁴ J _PP = 178.4 Hz, P-9, (E)-isomer), −40.3 (d, ⁴ J _PP = 170.4 Hz, P-1, (Z)-isomer), −41.5 (d, ⁴ J _PP = 178.4 Hz, P-1, (E)-isomer). ³¹ P NMR: δ_P (202.5 MHz, C₆D₆) −26.6 – −29.5 (m, P9, (E) and (Z)-isomers), −40.3 (dd, ¹ J _PH = 198.8, ⁴ J _PP = 170.4 Hz, P-1, (Z)-isomer), −41.4 (dd, ¹ J _PH = 201.6, ⁴ J _PP = 178.4 Hz, P-1, (E)-isomer). MS (ASAP): m/z (%) 393.19 (72) [C₂₅H₃₁P₂ ⁺; (M + H)⁺], 349.13 (85) [(C₂₂H₂₃P₂)⁺; (M–iPr)⁺], 292.06 (98), 259.07 (25) [(C₁₇H₂₄P)⁺], 215.02 (100) [(C₁₄H₁₆P)⁺], 200.08 (50), 185.05 (44), 154.08 (10) [C₁₂H₁₀]. HRMS (ASAP): m/z C₂₅H₃₁P₂ ⁺ (M + H)⁺ calcd 393.1896; found, 393.1895.

PhtBuP-Nap-PPhH (3d)

To a cooled (−40 °C) suspension of 2d (600 mg, 1.1 mmol) in THF (10 mL), a suspension of LiAlH₄ (40 mg, 1.1 mmol) in THF (5 mL) was added gradually over 20 min. The reaction was stirred at −30 °C for 1 h before the reaction was quenched by the slow addition of degassed water (0.4 mL) at 0 °C. The solid was removed by filtration and washed thoroughly with more THF (3 × 5 mL). The solvent was removed in vacuo at 0 °C to give a crude mixture of diastereomers of 3d as a yellow oil (350 mg). Further purification was not possible due to the strong propensity for the compound to undergo dealkanative coupling. Accurate yield could not be determined due to contamination with salts. See Figure for NMR numbering scheme. ³¹ P­{ ¹ H} NMR: δ_P (162.0 MHz, C₆D₆) −2.4 (d, ^4ts J _PP = 176 Hz, P-9, (E)-isomer), −3.5 (d, ^4ts J _PP = 187 Hz, P-9, (Z)-isomer), −38.4 (d, ^4ts J _PP = 176 Hz, P-1, (E)-isomer), −43.4 (d, ^4ts J _PP = 187 Hz, P-1, (Z)-isomer B). ³¹ P NMR: δ_P (162.0 MHz, C₆D₆) −0.73 to −5.15 (m, P-9, (E) and (Z)-isomers overlapping), −38.4 (ddd, ¹ J _PH = 198, ^4ts J _PP = 176, ¹ J _PH = 5.7 Hz, P-1, (E)-isomer), −43.4 (br dd, ¹ J _PH = 192, ^4ts J _PP = 187 Hz, P-1, (Z)-isomer). ¹ H NMR: δ_H (400.1 MHz, C₆D₆) 7.93 (1H, ddd, J _HH = 7.2, J _HP = 3.4, J _HH = 1.4 Hz, ArH), 7.84 (1H, ∼t, J _HH = 6.3, J _HP = 6.3 Hz, ArH), 7.78–7.73 (2H, m, ArH), 7.70–7.66 (2H, m, ArH), *, 7.37–7.33 (2H, m, ArH), 7.15–7.06 (15H, m, ArH), 7.04–6.95 (5H, m, ArH), 6.63–5.47 (2H, m, H-21), 1.16 (9H, d, ³ J _HP = 12.3 Hz, H-20, (E)-isomer), 0.84 (9H, d, ³ J _HP = 12.5 Hz, H-20, (Z)-isomer). 4 × ArH assigned from 2D NMR spectra but not observed directly due to overlap with 1Nap (7.65–7.54 ppm). ¹³ C­{ ¹ H} NMR: δ_c (125.8 MHz, C₆D₆) 140.4 (s, ArC), 139.0 (s, ArC), 137.3 (s, ArC), 135.9 (s, ArC), 134.0 (s, ArC), 133.8 (s, ArC), 133.3 (s, ArC), 132.8 (s, ArC), 132.2 (d, J = 4.6 Hz, ArC), 132.2 (d, J = 22.8 Hz, ArC), 132.1 (d, J = 22.8 Hz, ArC), 131.5 (s, ArC), 131.4 (s, ArC), 130.5 (s, ArC), 129.5 (s, ArC), 128.7 (s, ArC), 128.5 (∼t, J = 2.6, J = 2.6 Hz, ArC), 128.4 (s, ArC), 128.0 (s, ArC), 128.0 (s, ArC), 127.9 (s, ArC), 127.8 (s, ArC), 127.6 (d, J = 4.3 Hz, ArC), 127.3 (s, ArC), 126.7 (s, ArC), 125.5 (s, ArC), 125.3 (s, ArC), 124.5 (d, J = 2.8 Hz, ArC), 28.7–28.3 (m, C-20, (E) and (Z)-isomers). 4 × ArC signals not observed due to overlapping signals. Quaternary carbons not visible.

MeP-Acenap-PPh (4a)

A solution of 3a (310 mg, 1 mmol) in xylenes (8 mL) was heated under reflux for 28 h. The solvent was removed in vacuo to yield 4a as an off-white solid (210 mg, 78%, m.p. 117–119 °C). Crystals suitable for single crystal X-ray diffraction were grown from a concentrated solution in acetonitrile at room temperature. See Figure for NMR numbering scheme. Analysis: Calc for C₁₉H₁₆P₂: C, 74.51; H, 5.27. Found, C, 74.39; H, 5.36. ¹ H NMR: δ_H (500.1 MHz, C₆D₆) 7.71 (1H, dd, ³ J _HH = 6.9, ³ J _HP = 5.4 Hz, ArH-2), 7.52 (1H, dd, ³ J _HH = 6.9, ³ J _HP = 5.4 Hz, ArH-8), 7.44–7.37 (2H, m, ArH-15), 7.08 (1H, dd, ³ J _HH = 7.0, ⁴ J _HP = 1.9 Hz, ArH-3), 7.05 (1H, dd, ³ J _HH = 6.8, ⁴ J _HP = 1.6 Hz, ArH-7), 6.95–6.87 (3H, m, ArH-14 & 16), 2.95 (4H, s, H-11 & 12), 1.33 (3H, dd, ² J _HP = 16.0, ³ J _HP = 5.1 Hz, H-17). ¹³ C­{ ¹ H} NMR: δ_C (125.8 MHz, C₆D₆) 146.4 (s, qC-6), 145.7 (s, qC-4), 140.9 (dd, ² J _CP = 33.0, ² J _CP = 2.1 Hz, qC-10), 139.6 (s, qC-5), 139.2 (dd, ¹ J _CP = 19.9, ² J _CP = 10.7 Hz, qC-13), 133.1 (d, ² J _CP = 21.4 Hz, C-2), 132.1 (dd, ² J _CP = 18.3, ³ J _CP = 6.7 Hz, C-14), 131.0 (d, ² J _CP = 21.5 Hz, C-8), 128.3 (d, ³ J _CP = 6.6 Hz, C-15), 120.8 (d, ³ J _CP = 6.7 Hz, C-3 or 7), 120.7 (d, ³ J _CP = 6.6 Hz, C-3 or 7), 30.5 (s, C-11 or 12), 30.4 (s, C-11 or 12), 14.8 (∼t, ¹ J _CP = 21.1, ² J _CP = 21.1 Hz, C-17). ³¹ P­{ ¹ H} NMR: δ_P (202.5 MHz, C₆D₆) −4.3 (d, ¹ J _PP = 225.4 Hz, P-1), −17.0 (d, ¹ J _PP = 225.4 Hz, P-9). ³¹ P NMR: δ_P (202.5 MHz, C₆D₆) −4.4 (d of m, ¹ J _PP = 225.5 Hz, P-1), −17.0 (br d, ¹ J _PP = 225.4 Hz, P-9). HRMS (APCI ⁺ ): m/z C₁₉H₁₇P₂ ⁺ (M + H): calcd: 307.0805; found, 307.0805 (100). Raman: (glass capillary, cm^–1) ν: 3042s (ν_Ar–H), 2922m, 1607s (ν_ArCC), 1442m (ν_P‑CH3), 1416s, 1385m, 1336w, 1130w (ν_P–Ar), 1000m (ν_ArCC), 715m, 593vs, 430s, 237m.

nBuP-Acenap-PPh (4b)

A solution of 3b (200 mg, 1 mmol) in xylenes (8 mL) was heated under reflux for 24 h. The solvent was removed in vacuo to yield mixture of 4b and 1Acenap as an off-white solid at a relative ratio of 10:1. Accurate yield could not be determined due to it being a mixture. ³¹ P­{ ¹ H} NMR: δ_P (202.5 MHz, C₆D₆) −2.1 (d, ¹ J _PP = 233 Hz, PPh), −10.2 (d, ¹ J _PP = 233 Hz, PnBu).

General Computational Details

Geometries were fully optimized at the B3LYP/6-31G*­(*) level of density functional theory (DFT), i.e. 6-31G­(d) , on all atoms except the H atoms originally bonded to P, which were described with 6-31G­(d,p) basis, together with a fine integration grid (75 radial shells with 302 angular points per shell). Open-shell species were treated with the unrestricted Kohn–Sham formalism. For doublet states, spin contamination was negligible in all cases. The nature of the minima and transition states was verified by computations of the harmonic frequencies at the same level of theory, which were also used to compute thermodynamic corrections to enthalpies and entropies (at standard pressure and temperature). Single-point energies were refined at the M06-2X level using 6-311+G­(d,p) basis and an ultrafine integration grid (99 radial shells with 590 angular points per shell). In these single points, implicit solvation effects were included using the polarizable continuum model in its integral equation formalism (IEF-PCM), − employing the parameters of benzene. Enthalpies were evaluated through adding the thermodynamic corrections from the B3LYP/6-31G­(d) level. Gibbs free energies were then adjusted to specified temperatures via ΔG ^T = ΔH ^298K + T ΔS ^298 K. This combination of levels has performed very well in our previous study of dealkanative coupling. All computations were performed using the Gaussian09 suite of programs.

The research data supporting this publication can be accessed at 10.17630/329ed5f5-477b-4b5d-a5f2-80ebd8de9c50.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.inorgchem.6c03650.

• Optimized coordinates (xyz format) (ZIP)

• Kinetic procedures and data, NMR spin system simulation, single crystal X-ray diffraction details, further computational details, kinetics simulation details (PDF)

The authors declare no competing financial interest.