Constrained Phosphine Chalcogenide Selenoethers Supported by peri-Substitution

Abstract

A series of phosphorus and selenium peri-substituted acenaphthene species with the phosphino group oxidized by O, S, and Se has been isolated and fully characterized, including by single-crystal X-ray diffraction. The P(V) and Se(II) systems showed fluxional behavior in solution due to the presence of two major rotamers, as evidenced with solution NMR spectroscopy. Using Variable-Temperature NMR (VT NMR) and supported by DFT (Density Functional Theory) calculations and solid-state NMR, the major rotamers in the solid and in solution were identified. All compounds showed a loss of the through-space J_PSe coupling observed in the unoxidized P(III) and Se(II) systems due to the sequestration of the lone pair of the phosphine, which has been previously identified as the major contributor to the coupling pathway.

1. Dedication

This paper is dedicated to Professor J. Derek Woollins on the occasion of his well-earned retirement and for his outstanding contributions to main group chemistry.

2. Introduction

The selective, stepwise lithiation reaction of 5,6-dibromoacenaphthene allows synthetic access to heteroleptic bis(phosphino)acenaphthenes and has been used to synthesize bis(phosphine) A (Figure 1) [1]. Due to the inherent asymmetry of the heteroleptic phosphine groups, in the ³¹P{¹H} NMR spectrum, A shows two doublets of an AB spin system at δ_P −11.3 and −12.8 ppm, with a remarkably large ^4TSJ_PP of 180.0 Hz. This is attributed to the through-space coupling resulting from the overlap of the phosphorus lone pairs due to the constraints imposed by the rigid acenaphthene skeleton. Oxidation of the P(III) centers to P(V) with sulfur, atmospheric oxygen, or hydrogen peroxide results in a loss or significant decrease in the magnitude of the through-space J_PP coupling as the lone pairs of the phosphines are sequestered [1,2]. In only a handful of cases, where A acts as a bidentate ligand with MCl₂ (M = Zn, Cd, Hg), the magnitude of J_PP increases as the coupling is mediated by the large, diffuse s-character orbitals of the group 12 metals (e.g., A·HgCl₂ J_PP 309 Hz, Figure 1) [2].

Heteroleptic substitution is not limited to phosphorus substituents but can also involve other p-block and d-block heteroatoms (for some examples, see references [3,4,5,6,7,8,9,10,11]). Not only do heteroatoms present a challenging synthetic opportunity for peri-substitution, but they yield interesting NMR spectra when both nuclei are NMR active, as these nuclei can also experience through-space spin–spin coupling. When heavier nuclei are used, the orbitals are larger and more diffuse. As a result, through-space coupling can occur at longer peri-distances [12]. An excellent example of this is the series of phosphine–tin peri-substituted acenaphthene reported by Athukorala Arachchige et al. where ³¹P···¹¹⁹Sn J coupling can be observed [13]. The ¹¹⁹Sn isotope has I = ½ and a natural abundance of 8.6%, making it possible to observe J coupling with ³¹P (I = ½, 100%). In B (Figure 1), there is a direct P–Sn bond (2.815(3) Å) with ¹J_PSn 754 Hz, yet, in C, where there is no direct P–Sn bond but there is a sub-van der Waals P···Sn interaction (3.251(1) Å), a significant J_PSn of 373 Hz is still observed, demonstrating a clear 3c–4e type overlap of the phosphorus lone pair with the Sn–C_Ph σ* orbital. Other P/Sn acenaphthenes have also been reported with diphenylphosphino groups instead of diisopropylphosphino groups [14].

Woollins et al. previously published a series of naphthalene-based phosphine selenoethers [15]. In D, there is an efficient transfer of spin information between P and Se, as indicated by the ^4TSJ_PSe of 391 Hz (note, TS superscript indicates through-space coupling). When the P(III) center is oxidized with chalcogens to P(V) (compounds E), the magnitude of J_PSe diminishes to <24 Hz (Figure 1) [15]. An in-depth computational study has shown that the magnitude of J_PP and J_PSe in the related compound F has contributions from both through-space and through-bond pathways [16]. We recently reported a series of acenaphthene analogues (1) with various aryl groups bound to selenium [17]. As the electron-donating ability of the aryl group attached to selenium increases, so does the magnitude of J_PSe from 452 Hz, when R = phenyl, up to 545 Hz, when R = Mes* (2,4,6-tri-tert-butylphenyl).

3. Results and Discussion

3.1. Synthesis

Utilizing compound 1_Ph as our workhorse, we herein report the synthesis and characterization of the P(V) chalcogen oxidized species 1-O, 1-S, and 1-Se and the P(V)/Se(IV) species 1-O2 (Scheme 1).

Compound 1_Ph showed a singlet in the ³¹P{¹H} NMR spectrum at δ_P −6.5 ppm with ⁷⁷Se satellites giving J_PSe 452.2 Hz. This was complemented by a doublet at δ_Se 425.3 ppm, observed in the ⁷⁷Se{1H} NMR spectrum. Heating a solution of 1_Ph under reflux in toluene with one equivalent of gray selenium for 15 h, followed by purification, afforded 1-Se as a yellow microcrystalline powder (54% yield). 1-S was prepared in a similar manner but only required heating under reflux for 6 h for complete consumption of 1_Ph. After purification, 1-S was afforded as beige microcrystalline powder (86% yield).

The first attempt to produce 1-O using hydrogen peroxide resulted in the formation of a mixture of 1-O and 1-O2, as determined by solution ⁷⁷Se{¹H} NMR spectroscopy. Direct synthesis of 1-O2 was achieved by using an excess of hydrogen peroxide. After purification, 1-O2 was obtained as a white solid (45% yield). To control the oxidation of 1_Ph to selectively oxidize the phosphine, we attempted air oxidation by leaving a vigorously stirring solution of 1_Ph exposed to air; however, even after 24 h, no reaction had occurred, as judged by ³¹P{¹H} NMR spectroscopy. Instead, 1_Ph was stirred with one equivalent of H₂O₂·urea complex. The conversion was slow, but, as monitored by ³¹P{¹H} NMR spectroscopy, complete consumption of 1_Ph was observed after 72 h. After recrystallization, 1-O was isolated in a 39% yield. The mechanism of the phosphine oxidation was not studied in the scope of this work, as P(III) to P(V) oxidations by peroxides, cyclooctasufur (S₈), and gray selenium are well established from early thermochemical and mechanistic studies [18,19,20]. All compounds reported herein were found to be air stable, in the solid state, with no signs of degradation after twelve months.

3.2. Crystallography

Crystals of 1-Se, 1-S, and 1-O were grown from a solution of CH₂Cl₂:hexane (1:3 v/v), and crystals of 1-O2 were grown from evaporation of a solution in CH₂Cl₂. The structures of 1-O, 1-S, and 1-Se are very similar with only minor differences due to the increased size of the chalcogen bound to phosphorus. The crystal structures are shown in Figure 2, and selected crystallographic data are presented in

Table 1. Selected bond lengths (ångströms (Å)) and angles (degrees, °) for 1-O, 1-S, 1-Se, and 1-O2.

Compound	1-O	1-S	1-Se	1-O2 [b]
peri-region bond distances
P1···Se1	3.322(2)	3.4863(5)	3.5012(7)	3.578(1) [3.610(1)]
P1-E	1.491(6)	1.9657(5)	2.1219(7)	1.487(4) [1.492(4)]
Se1···E	2.825(6)	3.2272(5)	3.2829(6)	2.646(3) [2.625(3)]
Se1-O1	−	−	−	1.669(3) [1.671(4)]
peri-region bond angles
C9-P1-E	110.7(4)	112.39(5)	112.40(8)	112.7(2) [113.1(2)]
P1-E-Se1	95.7(3)	80.36(2)	77.47(2)	117.1(2) [118.5(2)]
E-Se1-C19	165.4(3)	164.53(5)	166.29(8)	84.6(2) [86.4(2)]
O2-Se1-O1	−	−	−	169.3(1) [169.2(2)]
C1-Se1-C19	99.3(3)	97.01(6)	96.8(1)	97.7(2) [95.4(2)]
Splay [a]	18.5	19.9	20.2	28.1 [29.1]
dihedral angles
C9-C10-C5-C4	174.8(8)	173.7(1)	173.2(2)	178.3(5) [177.4(5)]
P1-C9···C1-Se1	22.2(5)	31.87(8)	32.0(1)	1.8(3) [1.7(3)]
out-of-plane displacements
P1	0.508	0.605	0.593	0.065 [0.007]
Se1	−0.393	−0.700	−0.725	0.006 [0.128]

^[a] splay angle = sum of the bay region angles—360. ^[b] values in square parentheses are for the 2nd molecule in the asymmetric unit.

.

The most notable differences between the structures of the precursor (1_Ph) [17] and the oxidized P(V) species (1-O, 1-S, and 1-Se) are in the peri-region. In 1_Ph, the P···Se distance is 3.055(1) Å; this increases to 3.322(2) Å in 1-O, 3.4863(5) Å in 1-S, and 3.5012(7) Å in 1-Se. Similarly, there are large increases in the splay angles (12.6° in 1_Ph to 32.0° in 1-Se), the P–C···C–Se dihedral angles, and the out-of-plane displacements of the P and Se atoms from the mean C₁₂ acenaphthene plane (see Table 1). These changes are expected due to the new steric demands placed on the molecule caused by the addition of another atom into the peri-gap when the iPr₂P group is oxidized to iPr₂P = E (where E = O, S, Se). When compared to the crystal structures of the naphthalene analogues (E, Figure 1), there are no significant differences [15]. The only dissimilarity observed between the acenaphthene and naphthalene analogues is that the absence of the ethylene bridge in the naphthalene structures results in a slightly decreased P···Se distance and slightly smaller splay angles. For example, in the naphthalene compound E (where E = Se), the P···Se distance is 3.278(2) Å (vs. 3.5012(7) Å for 1-Se), and the splay angle drops from 32.0° to 24.8°.

The structure of 1-O2 is very different from that of the other three due to the Se(IV) group (selenoxide) being present. The trends are similar to those observed in the Se(II) complexes, but taken to a new extreme. There is a greater steric demand on the molecule, as evidenced by the much larger splay angle of 28° (29° in the second molecule of the asymmetric unit) and peri-distance between the P and Se atoms of 3.578(1) Å (3.610(1) Å). Somewhat unexpectedly, the dihedral angle is much smaller at 1.8(3)° (1.7(3)°); however, this arises from the rotation around the C9–Se1 bond, such that the Se=O group points away from the peri-gap, which significantly reduces the steric crowding and the need for any out-of-plane deformations to relieve the steric strain.

3.3. NMR Spectroscopy of 1-Se

In solution, the precursor 1_Ph showed a sharp singlet in the ³¹P{¹H} NMR spectrum at δ_P −6.5 ppm with ⁷⁷Se satellites (7.6% natural abundance), giving ^4TSJ_PSe of 452.2 Hz. This was complemented by a doublet in the ⁷⁷Se{¹H} NMR spectrum centered at δ_Se 425.3 ppm (^4TSJ_SeP 452.8 Hz). Recently, we have shown that the large through-space coupling between ³¹P and ⁷⁷Se arises from the overlap of the phosphorus lone pair with the orbitals localized on the Se–C_Ph bond [17]. In this study, the lone pair of the phosphorus was sequestered by oxidation with a chalcogen atom in all complexes, and as expected, this significantly reduced the magnitude of J_PSe in all compounds. Based on the recent findings by Makina et al. [16], we assume the dominant pathway of coupling information being exchanged is through space in the unoxidized 1_Ph; therefore, we attribute the drop in magnitude to J_PSe to the loss of this pathway.

3.3.1. Fluxionality in Solution

The ³¹P{¹H} NMR spectrum of 1-Se revealed two singlets, at δ_P 86.3 and 58.4 ppm, when only one was anticipated (Figure 3, top). The signal at δ_P 58.4 ppm was accompanied by a broadened set of ⁷⁷Se satellites with ¹J_PSe of ca. 690 Hz. This was indicative of a P = Se double bond and closely resembled those reported in the literature (cf. Ph₃PSe; ¹J_PSe 730 Hz) [21]. The signal at δ_P 86.3 ppm showed significant broadening and no resolvable ⁷⁷Se satellites. The presence of two rotational conformers was confirmed by ⁷⁷Se{¹H} NMR with the spectrum showing two broad singlets at δ_Se 426.2 and 419.0 ppm corresponding to the selenoether, as well as two doublets at δ_Se −358.5 and −451.0 ppm (¹J_SeP ca. 696 and 693 Hz, respectively) corresponding to the phosphine selenide (Figure 3, bottom). Initially, the presence of two sets of peaks suggested the occurrence of a side reaction; however, upon further investigation using Variable-Temperature NMR, it was concluded they were due to the very large steric bulk around the peri-region exhibiting fluxional behavior in solution. Acquisition of a ¹H–³¹P HMBC spectrum (Figure S1) showed a strong correlation between both ³¹P signals and the hydrogen atoms in the isopropyl groups, strongly supporting the idea of fluxional behavior. This was unexpected as the naphthalene equivalent (E) was reported as showing one sharp singlet in the ³¹P{¹H} NMR spectrum, which does not suggest fluxional behavior [15].

To obtain further evidence of the fluxional behavior, a one-dimensional exchange spectroscopy (EXSY) NMR experiment was performed on 1-Se. This showed magnetization transfer within the NMR timescale at 253 K. This indicated an exchange between two magnetic environments in two different isomers and confirmed the presence of fluxional behavior in solution. Therefore, the two sets of signals observed in the ambient temperature (293 K) ³¹P{¹H} and ⁷⁷Se{¹H} NMR spectra arose from two different rotational conformations present in solution. The notion of different rotational conformations in solution in peri-substituted naphthalenes has been reported by Woollins previously [22]. The species Nap(POCl₂)(PCl₂) (Nap = naphthalene-1,8-diyl), with a P(III)/P(V) peri-substitution, was demonstrated to have two rotamers in solution, with the ³¹P{¹H} NMR spectrum at 233 K showing two similar signals for the PCl₂ group (δ_P 145.52 and 145.50 ppm) and one signal for the POCl₂ group (δ_P 42.9 ppm). At 298 K, the signals at δ_P 145.52 and 145.50 ppm were not observed. Kilian et al. reported that these results are interpreted as “the hindered rotation around the P-C_(Nap) bonds, resulting the presence of two conformers whose interconversion is slow on the NMR time scale”.

Variable-temperature ³¹P{¹H} and ⁷⁷Se{¹H} NMR experiments were carried out using 1-Se; however, to overcome the coalescence point, a high-boiling solvent was needed. For the elevated-temperature experiments, d₅-bromobenzene (boiling point 156 °C, 429 K) was used as the NMR solvent. For low-temperature experiments, d-chloroform (melting point −64 °C, 209 K) was used. At 253 K, fully resolved signals of the two conformations with observable satellites were observed. The ³¹P{¹H} NMR spectrum at 253 K showed two singlets at δ_P 86.2 and 58.0 ppm, with ¹J_PSe of 682.9 and 681.3 Hz, respectively (Figure 4). The ⁷⁷Se{¹H} NMR spectrum at 255 K showed two singlets at δ_Se 422.8 and 415.3 ppm corresponding to the selenoether and two doublets at δ_Se, −362.6 and −452.4 ppm, with ¹J_SeP 681.5 and 683.6 Hz, corresponding to the phosphine selenide (Figure 5). The singlet at δ_Se 415.3 ppm also showed a ^5TSJ_SeSe coupling of 182.0 Hz as ⁷⁷Se satellites (Figure 5). Even in the slow-motion regime at 253–255 K, no through-space coupling was present between the phosphorus and the selenoether. This was expected as the phosphorus lone pair was sequestered in the P = Se bond and, hence, was no longer available to overlap with the SePh orbitals. The ³¹P{¹H} NMR spectrum acquired at 363 K showed one broad singlet at δ_P 68.3 ppm as the energy barrier between the two conformations had been overcome, but the speed of the exchange was only marginally faster than the NMR timescale. At 363 K, the ⁷⁷Se{¹H} NMR spectrum showed one singlet at δ_Se 433.0 ppm, corresponding to the selenoether. The upfield signal attributed to the P = Se group was not observed, likely due to the fact that 363 K is close to the coalescence temperature. Due to limitations with the equipment, we could not acquire any data at temperatures exceeding 368 K.

As the coalescence was observed in the ³¹P{¹H} and the ⁷⁷Se{¹H} NMR spectra, the coalescence method could be used to estimate the rotational barrier (ΔG^‡) of 1-Se, assuming the coalescence followed typical Eyring behavior.

$$\Delta G^{‡}=a{T}_C\left[9.972+log\left(\frac{T_c}{\Delta \nu }\right)\right]$$

Using this equation, the temperature of coalescence (T_C = 363 K) and the largest separation between the signals of the two conformers obtained from the lowest-temperature ³¹P{¹H} VT NMR spectra (Δν = 5715 Hz); ΔG^‡ was estimated as 61 kJ mol⁻¹.

3.3.2. Solid-State NMR of 1-Se

To corroborate the large coupling values observed and to confirm the number of conformations in the solid state, ³¹P{¹H} (Figure S2) and ⁷⁷Se{¹H} SS-MAS NMR spectra of 1-Se were acquired (Figure 6). In the ³¹P{¹H} SS-MAS NMR spectrum, a singlet at δ_P 60.2 (with ⁷⁷Se satellites giving ¹J_PSe = 699.2 Hz) was observed, with spinning sidebands. In the ⁷⁷Se{¹H} SS-MAS NMR spectrum, there were two signals: a singlet at δ_Se 431.6 ppm and a doublet at δ_Se −353.9 ppm, with ¹J_SeP of 700.8 Hz (Figure 6). The singlet corresponds to the selenoether and the doublet to the phosphine selenide environment. The large ¹J_SeP was still present in the solid state, albeit with a slightly larger magnitude than in the solution state.

The key finding is that no other signals were present for each ³¹P and ⁷⁷Se environment, indicating that while the bonding environments and connections were the same in both the solution and the solid state, only one conformer was present in the solid state. If two conformers were present in the solid state, two isotropic signals (with spinning side bands) would be expected in both the upfield and downfield regions of the ⁷⁷Se{¹H} MAS spectrum. It is likely that the dominant isomer corresponded to the conformation elucidated by the crystal structure (Figure 2). However, it is also possible that some solvates were formed, as demonstrated recently [23].

3.4. NMR Spectroscopy of 1-S and 1-O

For other chalcogen-oxidized compounds of 1_Ph, the change in rotational barrier was expected to follow the trend 1-Se > 1-S > 1-O, as the larger atomic radius of selenium provides a greater barrier to the rotation of the molecule (single-bond covalent radii Se 1.16 Å; S 1.03 Å; O 0.63 Å) [24].

The lighter congeners, 1-S and 1-O were prepared; as with 1-Se, the solution state ³¹P{¹H} NMR spectrum of 1-S acquired at ambient conditions was notably broad with two signals at δ_P 82.8 and 65.1 ppm, neither of which showed any ⁷⁷Se satellites (Figure 7, left). The solution-state ⁷⁷Se{¹H} NMR spectrum mirrored the observations of the ³¹P{¹H} spectrum with two broad singlets present at δ_Se 422.9 and 418.2 ppm (Figure 7, right). To determine the coalescence temperature and thus determine the rotational energy barrier, VT NMR experiments were performed on 1-S (Figure 7). The ³¹P{¹H} NMR spectra showed coalescence was reached at 368 K, although the signal at δ_P 71.9 ppm was still observed as a reasonably broad singlet, while completely resolved signals of the two rotamers in the slow-motion regime were observed at 253 K, showing two singlets at δ_P 82.6 and 64.7 ppm. Comparatively, the ⁷⁷Se{¹H} NMR spectra showed fast free rotation was achieved at 368 K with a sharp singlet observed at δSe 433.2 ppm and verified the full resolution of signals in the slow-motion regime at 255 K (δ_Se 418.9 and 413.2 ppm).

Similar observations were made for 1-O. The ³¹P{1H} NMR spectrum at ambient conditions revealed two broad singlets at δ_P 55.4 and 54.3 ppm, with the ⁷⁷Se{¹H} NMR spectrum showing two singlets at δ_Se 436.6 and 400.9 ppm (Figures S3 and S4). Additionally, VT NMR studies were carried out, with the coalescence observed at 323 K in the ³¹P{¹H} NMR spectrum with complete sharpening of signals observed at 373 K. Two fully resolved singlets were observed in both the ³¹P{¹H} and ⁷⁷Se{¹H} spectra at 255 K (δ_P 56.1 and 55.2 ppm; δ_Se 430.4 and 397.0 ppm). Using Δν = 3617 Hz from the 255 K spectra for 1-S and Δν = 184 Hz from the 255 K spectra for 1-O, ΔG^‡ was estimated to be ca. 62 kJ mol⁻¹ for 1-S and 63 kJ mol⁻¹ for 1-O (at 323 K), which, somehow contrary to expectations, was marginally higher than for 1-Se (61 kJ mol⁻¹). However, one needs to realise that the coalescence method is only an approximation, and that the ΔG^‡ values were obtained for different temperatures. Full Erying analysis was not possible as the spectrometer could not exceed 373 K, meaning complete sharpening of the peaks was never observed. As the van der Waals radii of Se is larger than that of S, which is larger than that of O (1.93, 1.85, 1.37 Å, respectively) [25], one may expect the rotational barriers to follow this order; therefore, it is likely that other steric and electronic effects were dominant here.

3.5. NMR Spectroscopy of 1-O2

The solution-state ³¹P{¹H} NMR spectrum of 1-O2 showed one sharp downfield-shifted singlet at δ_P 56.4 ppm with the ⁷⁷Se{¹H} NMR spectrum also showing a sharp downfield-shifted singlet at δ_Se 896.4 ppm (cf. 1_Ph δ_P −6.5 ppm; δ_Se 425.3 ppm) (Figure S5). The large downfield shift of both peaks was consistent with the oxidation of both the iPr₂P and SePh moieties to the P(V) and Se(IV) species, iPr₂P(O) and Se(O)Ph. No other signals were present in the NMR spectra, unlike for 1-O, 1-S, and 1-Se, indicating that only one rotational conformation was present in solution, presumably due to the increased steric bulk in the peri-region causing extremely hindered rotation of the iPr₂P(O) and Se(O)Ph groups. Due to the phosphorus lone pair being sequestered, as well as one of the selenium lone pairs, no J_PSe couplings were observed in either spectra.

3.6. Computational Studies

To complement these findings, we performed calculations at the B3LYP-D3/6-311+G(d,p)/CPCM(C₆H₅Br)//B3LYP-D3/6-31+G(d,p) level of density functional theory (DFT). Starting from the conformation observed in the solid, selected rotamers were constructed by rotating the SePh and iPr₂P(Se) moieties about the C(acenaphthene)–E bonds (E = P, Se). The resulting optimized structures are shown in Figure 8, and computed relative energies are collected in

Table 2. Computed ^a relative energies ΔE, enthalpies ΔH, and free energies ΔG^a for selected rotamers in kJ mol⁻¹ relative to 1-Se(A), as well as computed ^a and experimental ^b (in italics) δ(³¹P) and δ(⁷⁷Se) chemical shifts of 1-Se(A) and 1-Se(D).

Molecule	ΔE	ΔH298	ΔG298	δ(P)	δ(Se = P)	δ(SePh)
1-Se(A)	0	0	0	61.7 58.0 b	−389.9 −362.6 b	385.1 415.3 b
1-Se(B)	19.0	19.7	23.5
1-Se(C)	32.8	33.8	34.3
1-Se(D)	−2.6	−2.4	−2.4	89.1 86.2 b	−555.7 −452.4 b	433.8 422.8 b

^a Energies at B3LYP-D3/6-311+G(d,p)/CPCM(C₆H₅Br)//B3LYP-D3/6-31+G(d,p) level, thermodynamic corrections from B3LYP/6-31+G(d,p); chemical shifts at GIAO-B3LYP/ILGO-II`/CPM(CHCl₃)//B3LYP-D3/6-31+G(d,p) level ^b 235 K in CDCl₃ (this work).

.

In the conformer found in the solid (structure 1-Se(A) in Figure 8), the two Se atoms displayed sub-van der Waals contact (Se···Se distance 3.29 Å and 3.28 Å from B3LYP-D3 and XRD, respectively), and the Se–Ph group was oriented along the Se···Se axis and anti with respect to the Se atom on the phosphine. Rotating either the SePh group or the iPr₂P(Se) group such that the Se atoms were still in contact but the SePh group was roughly perpendicular to the Se···Se axis afforded two minima (1-Se(B) and 1-Se(C), respectively, in Figure 8) which were significantly higher in energy than conformer 1-Se(A) (by ca. 19–34 kJ mol⁻¹, see Table 2). Further rotating the iPr₂P(Se) moiety such that the Se atom on the phosphine was pointing away from the other Se atom in the SePh substituent afforded a new minimum (1-Se(D) in Figure 8) which was slightly more stable than conformer 1-Se(A) (by ca. −2 to −3 kJ mol⁻¹, see Table 2). One of the isopropyl groups was also rotated to minimize steric clash between a methyl group and the Se(Ph) atom in rotamer 1-Se(D). These results are fully compatible with the observation of a mixture of two slowly interconverting isomers. Based on the comparison of computed and observed ³¹P and ⁷⁷Se chemical shifts (see Table 2), we assigned the more deshielded ³¹P resonance, and the more “extreme” ⁷⁷Se shifts (i.e., the most deshielded and the most shielded one), to rotamer 1-Se(D). From the observed relative intensities of these two sets of signals (Figure 4 and Figure 5), it appears that it was indeed rotamer 1-Se(A) that was more abundant, i.e., more stable. This assignment also agrees with the comparison of the ⁷⁷Se resonances observed in the solid (Figure 6), arguable arising from 1-Se(A), and those of the more abundant form in solution. In addition, only for 1-Se(A), a notable indirect J_SeSe spin–spin coupling constant was computed (145 Hz, with 182 Hz observed), whereas that in 1-Se(D) was negligibly small. The reason why the computed relative stabilities of 1-Se(A) and 1-Se(D) were reversed is not clear at the moment. Indeed, switching the solvent model to CHCl₃, or the functional to M06-2X, which has performed very well for energetics in other related systems [5], did not change the relative sequence of both.

The reason for the apparent stability of rotamer 1-Se(D) seems to be more the relief of Se···Se repulsion rather than Se···P bonding interactions; the optimized Se···P distance in 1-Se(D) was 3.72 Å. This is close to the sum of the van der Waals radii of 4.09 Å; consequently, only a very small Wiberg bond index of 0.01 was obtained between these two atoms.

4. Materials and Methods

4.1. General Considerations

All synthetic manipulations were performed under an atmosphere of dry nitrogen using standard Schlenk techniques or under an argon atmosphere in a Saffron glove box. However, all compounds reported herein were found to be air stable, so repeated reactions were performed under air with no detrimental effects. All glass apparatus were stored in a drying oven (ca. 120 °C) prior to use. Dry solvents were collected from an MBraun Solvent Purification System and stored over appropriate molecular sieves. Water used in experiments was subject to nitrogen sparging and stored under nitrogen prior to use. Chemicals were taken from the laboratory inventory and used without further purification. Infrared Spectra were acquired using a Nicolet 308 FT-IR (Thermo Fisher Scientific, Oxford, UK) with Specac ATR attachment, recorded between 4000 and 500 cm⁻¹.

All solution-state NMR spectra were recorded using either a Bruker Avance III (500 MHz) or Bruker Avance III-HD (500 MHz) spectrometer operating at a magnetic field strength of 11.7 Tesla at 20 °C, unless otherwise specified. Assignments of ¹H and ¹³C spectra were made in conjunction with appropriate 2D spectra. ¹³C NMR spectra were recorded using the DEPTQ pulse sequence with broadband proton decoupling. The following external standards were used: ¹H and ¹³C NMR, tetramethylsilane; ³¹P NMR, 85% H₃PO₄ in D₂O; ⁷⁷Se NMR, dimethyldiselende (Me₂Se₂) and diphenyldiselenide (Ph₂Se₂) as a secondary reference at 463.0 ppm. Residual solvent peaks were also used for secondary calibration (CDCl₃ δ_H 7.260 ppm; δ_C 77.160 ppm; C₆D₅Br δ_H 7.300, 7.019, 6.946 ppm; δ_C 130.900, 129.339, 126.162, 122.181 ppm). Chemical shifts (δ) are given in parts per million (ppm) relative to the residual solvent peaks where possible. Coupling constants (J) are quoted in Hertz (Hz). The NMR numbering scheme for all compounds is shown in Figure 9.

Solid-state ³¹P{¹H} and ⁷⁷Se{¹H} NMR (SS-MAS NMR) measurements were performed using a Bruker Avance III 400 MHz spectrometer operating at a magnetic field strength of 9.4 T. Experiments were carried out using a conventional 4 mm MAS probe with a MAS rate of 14 KHz for ³¹P{¹H} and 10 kHz for ⁷⁷Se{¹H}. The ⁷⁷Se{¹H} cross-polarization MAS experiments (using ramped contact pulse durations of 5 ms and TPPM 1H decoupling) were carried out with signal averaging for 2048 transients with a recycle interval of 3 s. Chemical shifts are reported in ppm, relative to Me₂Se at 0 ppm, using the isotropic resonance of solid H₂SeO₃ at 1288.1 ppm as a secondary reference. The position of the isotropic resonance within the spinning sideband patterns was unambiguously determined by recording a second spectrum at a different MAS rate.

Melting and decomposition points were determined by heating solid samples in sealed glass capillaries using a Stuart SMP30 Melting Point Apparatus. High-Resolution Mass Spectrometry of 1-S and 1-Se was performed by the EPSRC UK National Mass Spectrometry Facility (NMSF) at Swansea University using a Thermofisher LTQ Orbitrap XL (Atmospheric-Pressure Chemical Ionization). Mass Spectrometry on 1-O and 1-O2 was performed at the University of St Andrews using a Micromass LCT (Electrospray Ionization) from solutions of the analyte in methanol or acetonitrile. Elemental Analysis was performed by the EA Service at London Metropolitan University.

4.2. Synthetic Procedures and Analytical Data

4.2.1. Synthesis of 1-O

A solution of 1_Ph (500 mg, 1.18 mmol) in dichloromethane (25 mL) was prepared. To this, a solution of hydrogen peroxide urea adduct (111 mg, 1.18 mmol) in water (100 mL) was added in one batch. The solution was stirred vigorously for three days. The organic layer was separated and dried over magnesium sulfate. The volatiles were removed in vacuo to afford the crude product. Recrystallization from dichloromethane:n-hexane (1:4 v/v) at −20 °C afforded analytically pure crystals of 1-O (200 mg, 40%) (melting with decomp. 170–173 °C). These crystals were suitable for single-crystal X-ray diffraction.

¹H NMR: (500.1 MHz, C₆D₅Br, 368 K) δ_H 7.84 (1H, d, ³J_HP 7.3 Hz, H-2), 7.20 (1H, d, ³J_HH 7.2 Hz, H-7), 7.19–7.13 (2H, m, H-17), 7.00–6.90 (4H, m, H-3, 18, 19), 3.15–3.03 (4H, m, H-11,12), 1.33 (6H, dd, ³J_HP 15.2, ³J_HH 6.9 Hz, H-14/15), 1.00 (6H, ³J_HP 15.5, ³J_HH 6.9 Hz, H14/15). ¹³C DEPTQ NMR: (125.8 MHz, C₆D₅Br, 368 K), δ_C 151.1 (d, ⁴J_CP 2.4 Hz, qC-6), 147.4 (s, qC-4), 140.6 (d ³J_CP 8.2 Hz, qC-5), 140.3 (s, C-2), 138.2 (s, qC-16), 135.9 (d, ²J_CP 23.0 Hz, qC-10), 131.2 (s, C-18), 130.4 (s, qC-1/8), 128.8 (s, C-17), 126.1 (s, C-19), 125.3 (s, qC-9), 120.7 (s, C-3), 118.5 (d, ³J_CP 11.6 Hz, C-7), 30.0 (s, C-11/12), 29.7 (s, C-11/12), 29.5 (d, ¹J_CP 67.0 Hz, C-13), 17.3 (d, ²J_CP 3.4 Hz, C-14/15), 16.9–16.8 (m, C-14/15). ³¹P{¹H} NMR: (202.4 MHz, CDCl₃, 253 K) δ_P 56.1 (s), 55.2 (s). ³¹P{¹H} NMR: 202.4 MHz, CDCl₃, 295 K) δ_P 55.4 (br s), 54.3 (s). ³¹P{¹H} NMR: (202.4 MHz, C₆D₅Br, 373 K) δ_P 53.0 (s). ⁷⁷Se{¹H} NMR: (95.4 MHz, CDCl₃, 253 K) δ_Se 430.4 (s), 397.0 (s). ⁷⁷Se{¹H} NMR: (95.4 MHz, CDCl₃, 295 K) δ_Se 436.6 (s), 400.9 (s). ⁷⁷Se{¹H} NMR: (95.4 MHz, C₆D₅Br, 363 K) no signals observed. IR: ν_max ATR/cm⁻¹ 3067w (ν_CH), 2963w (ν_CH), 1576m (ν_C=C), 1138s (ν_P=O), 851m, 733s, 691s. HRMS: (ES+): m/z (%) Cacld. for C₂₄H₂₇POSeNa: 465.0857, found: 465.0842 (100) [M+Na].

4.2.2. Synthesis of 1-O2

A solution of 1_Ph (500 mg, 1.18 mmol) in dichloromethane (20 mL) was prepared. To this, 30% aqueous hydrogen peroxide (0.25 mL, 2.47 mmol) was added dropwise over five minutes with vigorous stirring. The solution was stirred at ambient conditions for a further six hours. The organic layer was separated and dried over magnesium sulfate. The volatiles were removed in vacuo to afford 1-O2 as a pale orange solid (240 mg, 45%) (melting with decomp. 208–214 °C). The aqueous layer was quenched with aqueous sodium metabisulfite before disposal. Crystals of 1-O2 suitable for single-crystal X-ray diffraction were grown from a dichloromethane/n-hexane vapor diffusion set up at ambient conditions.

¹H NMR (500.1 MHz, CDCl₃) δ_H 8.45 (1H, dd, ³J_HH 7.5 Hz, H-8), 8.05–7.98 (2H, m, H-18), 7.57 (1H, dd, ³J_HP 13.7, ³J_HH 7.3, H-2), 7.37 (1H, d, ³J_HH 7.6 Hz, H-7), 7.32 (1H, d, ³J_HH 7.3 Hz, H-3), 7.23 (3H, m, H-17,19), 3.34–3.21 (4H, m, H-11,12), 2.54–2.36 (2H, m, H-13,13′), 1.28–1.14 (9H, m, H-14/14′/15/15′, 3 × CH₃), 0.88 (3H, dd, ³J_HP 15.3, ³J_HH 7.1 Hz, H-14/14′/15/15′, 1 × CH₃). ¹³C DEPTQ (125.8 MHz, CDCl₃) δ_C 152.8 (s, qC-4), 151.0 (s, qC-6), 148.3 (s, ¹J_CSe 127.9 Hz, C-16), 140.3 (d, ³J_CP 8.8 Hz, C-5), 139.8 (d, ³J_CP 3.3 Hz, qC-9), 134.1 (d, ²J_CP 11.7 Hz, C-1), 132.3 (s, C-8), 132.2 (d, ²J_CP 4.6 Hz, qC-10), 129.5 (s, C-19), 128.5 (s, C-17), 127.7 (s, C-18), 121.5 (s, C-7), 119.2 (d, ¹J_CP 84.8 Hz, qC-1), 118.6 (d, ³J_CP 13.0 Hz, C-3), 30.6 (s, C-11/12), 29.5 (s, C-11/12), 28.8 (d, ¹J_CP 64.1 Hz, C-13/13′), 26.5 (¹J_CP 68.3 Hz, C-13/13′), 17.1 (s, C-14/15, 1 × CH₃), 16.1 (d, ²J_CP 2.1 Hz, C-14/15, 1 × CH₃), 15.9 (d, ²J_CP 3.5 Hz, C-14/15, 1 × CH₃), 15.6 (d, ²J_CP 1.8 Hz, C-14/15, 1 × CH₃). ³¹P{¹H} NMR (202.5 MHz, CDCl₃) δ_P 56.4 (s). ⁷⁷Se{¹H} NMR (95.4 MHz, CDCl₃) δ_Se 869.4 (s). IR: ν_max ATR/cm⁻¹ 3047w (ν_CH), 2962w (ν_CH), 1597m (ν_C=C), 1437m (ν_C=C), 1149s (ν_P=O), 818vs (ν_Se=O), 752s, 690s. HRMS: (ES+): m/z (%) Cacld. for C₂₄H₂₈PO₂Se: 459.0987, found: 459.0973 (100) [M+H].

4.2.3. Synthesis of 1-S

A suspension of 1_Ph (1.50 g, 3.52 mmol) and sulfur (177 mg, 3.65 mmol) in toluene (30 mL) was heated under reflux for six hours. The solution was cooled to ambient conditions and all volatiles removed in vacuo to afford the crude product. Recrystallization from dichloromethane:n-hexane (1:3 v/v) at −20 °C afforded white analytically pure crystals of 1-S (1.38 g, 86%) (melting with decomp. 232–237 °C). These crystals were suitable for single-crystal X-ray diffraction. Elemental Analysis: Cacld. (%) for C₂₄H₂₇PSSe: C 63.01, H 5.95, found: C 62.89, H 6.03. ¹H NMR (500.1 MHz, C₆D₅Br, 363 K) δ_H 7.98 (1H, d, ³J_HH 7.2 Hz, H-8), 7.15 (1H, d, ³J_HH 7.3 Hz, H-3), 7.03–6.98 (2H, m, H-18), 6.93 (1H, d, ³J_HH 7.2 Hz, H-3), 6.96–6.87 (3H, m, H-17,19), 3.47 (2H, br s, H-13,13′), 3.09–3.01 (4H, m, H-11,12), 1.34 (6H, dd, ³J_HP 17.1, ³J_HH 6.9 Hz, H-14,14′), 1.04 (6H, dd, ³J_HP 17.7, ³J_HH 6.9 Hz, H-15/15′). ¹³C DEPTQ NMR (125.8 MHz, C₆D₅Br, 368 K) δ_C 151.4 (d, ⁴J_CP 2.5 Hz, qC-4), 148.2 (s, qC-6), 142.2 (s, C-8), 141.0 (d, ³J_CP 8.6 Hz, qC-5), 135.8 (s, qC-10), 130.0 (s, C-18), 128.8 (s, C-17), 125.9 (s, C-19), 124.1 (d, ¹J_CP 64.2 Hz, C-1), 120.8 (s, C-7), 118.3 (d, ³J_CP 12.3 Hz, C-3), 30.1 (d, ¹J_CP 50.2 Hz, C-13,13′), 29.8 (s, C-11/12), 29.8 (s, C-11/12), 17.8 (s, C-14,14′), 17.5 (s, C-15/15′). ³¹P{¹H} NMR (202.4 MHz, CDCl₃, 253 K) δ_P 82.6 (s), 64.7 (s). ³¹P{¹H} NMR (202.4 MHz, CDCl₃, 295 K) δ_P 82.8 (s), 65.1 (s). ³¹P{¹H} NMR (202.4 MHz, C₆D₅Br, 363 K) δ_P 71.9 (s). ⁷⁷Se{¹H} NMR (95.4 MHz, CDCl₃, 253 K) δ_Se 418.9 (s), 413.2 (s). ⁷⁷Se{¹H} NMR (95.4 MHz, CDCl₃, 293 K) δ_Se 422.9 (br s), 418.2 (br s). ⁷⁷Se{¹H} NMR (95.4 MHz, C₆D₅Br, 368 K) δ_Se 433.2 (s). IR: ν_max ATR/cm⁻¹ 3055w (ν_CH), 2958w (ν_CH), 1578m (ν_C=C), 1477m (ν_C=C), 1022m, 744s, 687vs (ν_P=S). HRMS (APCI+): m/z (%) Cacld. for C₂₄H₂₈PSSe: 459.0815, found: 459.0814 (100) [M+H].

4.2.4. Synthesis of 1-Se

A suspension of 1_Ph (1.50 g, 3.52 mmol) and selenium (276 mg, 3.50 mmol) in toluene (30 mL) was heated under reflux for fifteen hours. The solution was cooled to ambient conditions and all volatiles removed in vacuo to afford the crude product. Recrystallization from dichloromethane:n-hexane (1:3 v/v) at −20 °C afforded yellow analytically pure crystals of 1-Se (1.10 g, 62%) (melting with decomp. 237–242 °C). These crystals were suitable for single-crystal X-ray diffraction. Elemental Analysis: Cacld. (%) for C₂₄H₂₇PSe₂: C 57.15, H 5.40, found: C 56.93, H 5.36. ¹H NMR (500.1 MHz, C₆D₅Br, 363 K) δ_H 8.02 (1H, d, ³J_HH 7.2 Hz, H-8), 7.12 (1H, d, ³J_HH 7.4 Hz, H-3), 7.00–6.92 (3H, m, H-7,18), 6.92–6.85 (3H, m, H-17,19), 3.55 (2H, br s, H-13,13′), 3.16–2.90 (4H, m, H-11,12), 1.33 (6H, dd, ³J_HP 17.5, ³J_HH 6.8 Hz, H-14,14′), 1.04 (6H, dd, ³J_HP 18.3, ³J_HH 6.9 Hz, H-15,15′). ¹³C DEPTQ NMR (126.8 MHz, C₆D₅Br, 363 K) δ_C 151.6 (d, ⁴J_CP 2.6 Hz, qC-4), 148.3* (s, qC-6), 142.4 (s, C-8), 141.0 (d, ³J_CP 8.2 Hz, qC-5), 135.8 (s, qC-10), 129.8 (s, C-18), 128.8 (s, C-17), 125.9 (s, C-18*), 123.4 (s, qC-9), 121.9* (s, qC-1), 120.9 (s, C-7), 118.3 (d, ³J_CP 12.4 Hz, C-3), 29.8 (s, C-11/12), 29.6 (s, C-11/12), 29.3 (d, ¹J_CP 42.5 Hz, C-13,13′), 18.8* (s, C-14,14′), 18.7* (s, C-15,15′). ³¹P{¹H} NMR (202.5 MHz, CDCl₃, 253 K), δ_P 86.2 (s, ¹J_PSe 682.9 Hz), 58.0 (s, ¹J_PSe 681.3 Hz). ³¹P{¹H} NMR (202.5 MHz, CDCl₃, 295 K) δ_P 86.3 (br s), 58.4 (br s). ³¹P{¹H} NMR (202.5 MHz, C₆D₅Br, 363 K), δ_P 68.3 (br s). ³¹P{¹H} SS-MAS NMR (162.0 MHz) 60.2 (s, ¹J_PSe 699.2 Hz). ⁷⁷Se{¹H} NMR (95.4 MHz, CDCl₃, 255 K) δ_Se 422.8 (s), 415.3 (s, ^5TSJ_SeSe 182.0 Hz), −362.6 (d, ¹J_SeP 681.5 Hz), −452.4 (d, ¹J_SeP 683.6 Hz). ⁷⁷Se{¹H} NMR (95.4 MHz, CDCl₃, 293 K) δ_Se 426.2 (br s), 419.0 (br s), −358.5 (d, ¹J_SeP 696.4 Hz), −451.0 (d, ¹J_SeP 693.4 Hz). ⁷⁷Se{¹H} NMR (95.4 MHz, C₆D₅Br, 363 K) δ_Se 433.0 (s). ⁷⁷Se{¹H} SS-MAS NMR (76.3 MHz) 431.6 (s), −353.9 (d, ¹J_PSe 700.8 Hz). IR: ν_max ATR/cm⁻¹ 3047w (ν_CH), 2962w (ν_CH), 1601m (ν_C=C), 1473m (ν_C=C), 1018m, 744s, 636s. HRMS (APCI+): m/z (%) Cacld. for C₂₄H₂₈PSe₂: 507.0263, found: 507.0264 (100) [M+H]. Note: ¹³C signals denotated with * were observed in the 2D ¹H–¹³C HMBC only.

4.3. Crystallographic Details

The crystallographic data for 1-O were collected using a Rigaku XtaLAB P200 diffractometer using multi-layer mirror monochromated Mo Kα radiation at −180 °C (±1). The crystallographic data for 1-S were collected using a Rigaku XtaLAB P100 diffractometer using multi-layer mirror monochromated Cu Kα radiation at −100 °C (±1). The crystallographic data for 1-O2 and 1-Se were collected using a Rigaku SCX mini diffractometer using graphite monochromated Mo Kα radiation at −100 °C (±1) (Mo Kα = λ = 0.71073 Å; Cu Kα = λ = 1.54184 Å).

Intensity data were collected using ω steps accumulating area detector frames spanning at least a hemisphere of reciprocal space. All data were corrected for Lorentz, polarization, and long-term intensity fluctuations. Absorption effects were corrected on the basis of multiple equivalent reflections. The structures were solved by direct methods [26]. Non-hydrogen atoms were refined anisotropically, and hydrogen atoms were refined using the riding model.

The crystal structures were refined by full-matrix least squares against F2 (SHELXL) [27,28] using the CrystalStructure GUI [29]. Searches of the Cambridge Structural Database (CSD) were performed using the webCSD [30]. Images and manipulations of crystal structures and computed rotamers were obtained using OLEX-2 [31].

4.4. Computational Details

Geometries were fully optimized at the B3LYP level [32,33] (using a fine integration grid, i.e., 75 radial shells with 302 angular points per shell) with Curtis and Binning’s 962(d) basis [34] on Se and 6-31+G(d,p) elsewhere. The solid-state structure was used as starting point for the optimizations of conformer 1-Se(A). The nature of the stationary points was verified by computation of the harmonic frequencies at the same level of theory, which were also used to compute thermodynamic corrections to obtain enthalpies and free energies (standard pressure and temperature). The structures were then re-optimized at the dispersion-corrected B3LYP-D3 [35] level using the same basis set and Becke–Johnson damping [36,37]. Single-point energies were refined for the B3LYP-D3 structures at the B3LYP-D3 level using 962+(d,f) basis on Se, i.e., including the recommended [38] diffuse s and p set and the f-function, and 6-311+G(d,p) elsewhere; an implicit solvent model was used in these single-point calculations, namely the Conductor-Like Polarizable Continuum Model (CPCM) [39,40], using the default settings in Gaussian09 and the parameters of bromobenzene. Wiberg bond indices (WBIs) were computed at that level from natural bond orbital (NBO) analysis. The WBI is a measure for the covalent character of a bond and adopts values close to 1 and 2 for true single and double bonds, respectively [41]. This and similar levels have performed well in previous studies of related acenaphthene chalcogen and pnictogen compounds [2,3,4,5,6,7,13,15,17]. Magnetic shieldings and spin–spin coupling constants (SSCCs) were computed at the GIAO-B3LYP level using IGLO DZ basis on H atoms and IGLO-basis II everywhere else (denoted ILGO-II), which was designed for computation of magnetic properties [42] and the CPCM model with the parameters of chloroform. The relative ⁷⁷Se shifts were referenced relative to Me₂Se (computed σ = 1652.1 ppm at the same level). Because the experimental standard for ³¹P NMR, concentrated phosphoric acid, is difficult to model computationally, chemical shifts were first referenced to Ph₃PSe (computed σ = 235.1 ppm) and converted to the usual δ scale using the experimental chemical shift of that compound in CDCl₃, 43.2 ppm [43]. In the computations of SSCCs, the basis set was uncontracted for evaluating the Fermi contact contribution (keyword NMR = (Spin–Spin, Mixed) in Gaussian). All computations were performed using the Gaussian09 suite of programs [44].

5. Conclusions

A series of phosphorus and selenium peri-substituted acenaphthenes with the phosphorus atom oxidized by oxygen, sulfur, and selenium was synthesized and characterized by single-crystal X-ray diffraction and multinuclear NMR spectroscopy. For the Se(II) species, there were two major rotational conformers in solution, as identified by Variable-Temperature NMR experiments and supported with DFT calculations. Only one of these conformations was present in the solid state, as verified by X-ray crystallography and solid-state NMR spectroscopy.