Richard Chlebíka,
Erik Kertészb,
Milan Erben
a,
Aleš Růžička
a,
Roman Jambor
a,
Zoltán Benkő
*bc and
Libor Dostál
*a
aDepartment of General and Inorganic Chemistry, University of Pardubice, Studentská 573, CZ 532 10 Pardubice, Czech Republic. E-mail: libor.dostal@upce.cz
bDepartment of Inorganic and Analytical Chemistry, Faculty of Chemical Technology and Biotechnology, Budapest University of Technology and Economics, Műegyetem rkp. 3, H-1111 Budapest, Hungary. E-mail: benko.zoltan@vbk.bme.hu
cHUN-REN-BME Computation Driven Chemistry Research Group, H-1111 Budapest, Műegyetem rkp. 3, Hungary
First published on 8th September 2025
The coordination properties of the P,C,P-pincer ligand (Ar = 2,6-(tBu2PO)2C6H3) with organotin(IV) compounds were examined. For this purpose, a set of neutral compounds including ArSnPh2Cl (1), ArSnPhCl2 (2) and ArSnCl3 (3), ArSnBu3 (4) and the cations [ArSnPh2][BArF] (1+[BArF]−), [ArSnPhCl][BArF] (2+[BArF]−), [ArSnCl2][BArF] and (3+[BArF]−) ([BArF] = [3,5-(CF3)2C6H3]4B) were prepared and characterized by multinuclear NMR spectroscopy and single-crystal (sc) X-ray diffraction analysis (2, 3, 1+[BArF]− and 3+[BArF]−). This study revealed different types of ligand coordination, i.e. no P → Sn intramolecular interaction in 1 and 2, while one P atom is coordinated in 3 and both P atoms in tin cations 1+, 2+ and 3+. To further elucidate the strength of these P → Sn dative bonds, all compounds were reacted with [BH3(SMe2)] to prove whether it coordinates toward pendant P atoms or even de-coordinates those P atoms already connected to the tin atom. Thus, in 1, 2, and 4, both P atoms formed complexes with the borane, while in 3 only one phosphorus reacted with BH3, because the second remained bonded to the tin atom. Finally, even in the cation 1+ one of the P atoms could be blocked by borane leaving the tin atom four-coordinated, while it was not possible for 2+ and 3+. DFT calculations were used to gain a deeper insight into the P → Sn bonding interaction in the studied compounds.
In striking contrast, no complexes of p-block elements with classical P,C,P-ligands have been reported to the best of our knowledge so far. This is quite surprising due to the fact that various types of closely related N,C,N-, O,C,O- and O,C,N-ligands have been recognized as very useful structurally analogous platforms for p-block elements.4 Examination of their coordination capabilities using organotin(IV) compounds as suitable model species has been a common feature in their introduction into the chemistry of main group elements. The first examples of N,C,N-tin(IV) complexes were reported by van Koten in 1989,5 whereas the O,C,O-analogues were introduced by Jurkschat6 (1988) and later on (2002) by our group7 (Fig. 1B–D). Since these origins, the chemistry of all ligands has spread to other main group elements containing central atoms in various oxidation states and bonding situations. Consequently, interesting and relevant examples can be found with elements of Group 14,8–10 13,11,12 1513–15 and 16,16 underlining the exceptional utility of these coordination platforms.
In the present contribution, we introduce for the first time the P,C,P-pincer ligand (Ar = 2,6-(tBu2PO)2C6H3, Fig. 1E), which is so successful in transition metal chemistry,17 to the field of p-block elements. This study is aimed at validating a synthetic protocol based on the utilization of the lithiated precursor ArLi for the preparation of organotin(IV) compounds, while the tuning of the Lewis acidity of the central atom should allow us to obtain various coordination modes of the ligand. A complete set of both neutral and cationic tin compounds was synthesized and characterized for this purpose, while remarkable coordination variability of this ligand was obtained. The relative strength of intramolecular P → Sn interaction(s) was studied using multinuclear NMR spectroscopy in solution, single crystal (sc) X-ray diffraction analysis in the solid state and chemically by reacting isolated tin complexes with [BH3(SMe2)] aiming to block/de-coordinate accessible phosphorus function(s). The obtained coordination modes are also compared with those of the closest analogues shown in Fig. 1B and D. A detailed DFT investigation was also performed to acquire deeper insight into the nature of P → Sn bonds in target compounds.
The 1H and 13C{1H} NMR spectra in C6D6 revealed an expected set of signals for both ligand and phenyl moieties attached to the tin atoms in 1 and 2. Signals found in the 31P{1H} NMR spectra of 1 (155.8 ppm) and 2 (160.6 ppm) are close to that of the starting ArBr (154.2 ppm, Table 1). The 119Sn{1H} NMR spectra exhibited one signal at −57.4/−44.7 ppm for 1/2, respectively, both being only marginally shifted compared to signals for the related Ph3SnCl (−48.0 ppm)19 and Ph2SnCl2 (−32.0 ppm).19 All these data suggest the absence of any significant intramolecular P → Sn interaction in benzene solution. Not surprisingly, the tin atom in the tetraorgano-derivative 4 also did not show any significant interaction with the phosphorus based on the obtained values of δ(31P) = 154.0 ppm and δ(119Sn) = −54.8 ppm.
δ(119Sn) | δ(31P) | 1JSn,P | δ(119Sn) | δ(31P) | 1JSn,P | ||
---|---|---|---|---|---|---|---|
a Acquired in C6D6.b Acquired in CDCl3.c Acquired in CD2Cl2.d Two sets of signals for 1/1′and 2/2′ detected.e Two signals detected.f The calculation was carried out without the counter anion.g For 4·AgOTf and 4·AgSbF6 the values correspond to 1J(109/107Ag,31P). | |||||||
ArBra | — | 154.2 (154.2/154.5) | — | 2(BH3)2a | −44.2 (−36.2) | 159.6 (164.0/160.8) | — |
ArBr(BH3)a | — | 155.6 (158.2) | — | 3a | −274.1 (−298.5) | 153.0/71.2e (150.3/51.0) | 308 (888) |
1a | −57.4 (−84.7) | 155.8 (160.6/149.3) | — | 3b | −275.0 | 153.9/72.0e | 323 |
1b,d | −57.3 | 157.3 150.5/83.4e | — | 3+[BArF]− c,f | −194.0 (−210.6) | 86.2 (71.5) | 88 (342) |
−203.4 | |||||||
1+[BArF]−c,f | −187.6 (−227.7) | 105.7 (105.6/88.7) | 835 (843/829) | 3(BH3)a | −278.5 (−299.0) | 152.8/70.6e (154.9/50.2) | 211 (807) |
1(BH3)2a | −57.4 (−79.8) | 155.8 (159.3/155.8) | — | 4a | −54.8 | 154.0 | — |
[1(BH3)]+ [BArF]−c,f | −88.5 (−108.9) | 156.0/116.2e (161.3/107.7) | 285 (733) | 4(BH3)2a | −46.7 | 154.5 | — |
2a | −44.7 (−24.5) | 160.6 (151.6/162.0) | — | 4·AgOTfc | −43.3 | 170.0 | 568/494g |
2b,d | −45.7 | 160.6 151.6/77.1e | — | 4·AgSbF6c | −42.4 | 169.7 | 567/497g |
−224.2d | |||||||
2+ [BArF]−c,f | −214.4 (−240.3) | 97.2 (83.4/83.3) | 640 (702) |
Contrary to the NMR spectra for all the compounds mentioned above, the 1H NMR spectrum of 3 in C6D6 showed two signals for tBu2P groups and three signals for aromatic CH groups of the Ar ligand pointing to a non-equivalence of ligand arms. This is supported by the observation of six signals for aromatic carbon atoms in the 13C{1H} NMR spectrum. Furthermore, the 31P{1H} NMR spectrum contained two signals, one observed close to the value for ArBr (153.0 ppm), while the second is significantly high-field shifted (71.2 ppm) and flanked by tin satellites (1J(119/117Sn,31P) = 308 Hz). This resonance pattern indicates tight coordination of the single ligand arm to the central tin atom, while the second phosphorus function remains pendant. This finding is further corroborated by the detection of a doublet at −274.1 ppm (1J(119Sn,31P) = 308 Hz) in the corresponding 119Sn{1H} NMR spectrum. In particular, the latter value indicates a tin atom more shielded compared to that in PhSnCl3 (cf. −63 ppm),19 which is consistent with coordination of phosphorus with tin.
Furthermore, to elucidate a plausible influence of the solvent, 1H, 31P{1H} and 119Sn{1H} NMR spectra of 1–3 in CDCl3 were recorded as well. The 1H NMR spectra in all cases revealed a set of broad signals. The 31P{1H} spectrum of 3 again revealed two signals at 153.9 and 72.0 ppm, where the latter is flanked by tin satellites (1J(119/117Sn,31P) = 323 Hz), while one doublet is obtained in the 119Sn{1H} NMR spectrum at −275.0 ppm (1J(119/117Sn,31P) = 323 Hz). These data are almost identical to those found in C6D6 proving that the same structure for 3 exists in both solvents (Fig. 2A). Surprisingly, two sets of signals were observed in the 31P{1H} and 119Sn{1H} NMR spectra in CDCl3 for 1 and 2. The values of the major set are again closely related to the data described above in C6D6, i.e. one signal in the 31P{1H} NMR spectra at 157.3 (for 1) and 160.6 ppm (for 2) and one singlet in the 119Sn{1H} NMR spectra at −57.3 (for 1) and −45.7 ppm (for 2). However, the minor set of signals is indicative of the presence of second isomers 1′ and 2′ with one coordinated P donor group similar to the situation found in 3. Thus, the 31P{1H} NMR spectra consist of two broad signals at 150.5/83.4 ppm and 151.6/77.1 ppm for 1′/2′, respectively. Importantly, the 119Sn{1H} NMR spectra revealed a second signal in addition to that of 1 and 2 at −203.4 for 1′ and −224.2 ppm for 2′ (cf. −274.1 ppm for 3, Fig. 2B and C). In conclusion, whereas compound 3 exhibits the same structure with one coordinated P-donor in both C6D6 and CDCl3, compounds 1 and 2 show no tightly rigid coordination of P atoms in C6D6, but two isomers, i.e. 1/1′ and 2/2′, most probably coexist in CDCl3 solution.
![]() | ||
Fig. 2 Comparison of 119Sn{1H} NMR spectra of 1–3 acquired in C6D6 (blue) or CDCl3 (red) showing the same structure for 1 (A) and plausible formation of isomers 1′ (B) and 2′ (C) in CDCl3. |
The molecular structures of 2 and 3 determined by sc-XRD analysis are shown in Fig. 3 and are consistent with the proposed structures in C6D6 solution above. The Sn(1) atom lacks any close interaction with phosphorus donor atoms in 2 according to the Sn(1)–P(1/2) distances 4.8286(7)/4.7695(10) Å, respectively, while adopting a distorted tetrahedral geometry. The C(1)–Sn(1)–C(23) angle (128.86(8)°) represents the main deviation from the ideal shape. In sharp contrast, a strong Sn(1)–P(1) interaction (2.6313(9) Å, cf. ∑cov(P,Sn) = 2.51 Å (ref. 20)) is observed in 3 leaving the P(2) atom pendant (Sn(1)–P(2) 4.9105(11) Å). This leads to a distorted trigonal bipyramidal array around the Sn(1) atom (τ = 0.77;21 note: τ = 0 for ideal square pyramid geometry and τ = 1 for ideal trigonal bipyramid geometry) with C(1) and Cl(3) located at the apex positions (cf. C(1)–Sn(1)–Cl(3) 165.67(8)°).
The coordination behavior of the P,C,P-ligand in 1–3 deserves attention when compared to the previously reported N,C,N- and O,C,O-ligand counterparts (Fig. 3). No significant intramolecular P → Sn interaction was detected in 1 (in C6D6 solution based on NMR) and 2 (in C6D6 solution based on NMR and in the solid state based on sc-XRD) whereas in the N,C,N-chelated analogues [(2,6-(Me2NCH2)2C6H3)SnPh2]+ (A1+)22 and [2,6-(Me2NCH2)2C6H3]SnPhCl2 (A2)23 both nitrogen atoms coordinate the tin atom quite tightly, which even results in autoionization of the former allowing isolation of the whole set of tin cations related to A1+ with various counter anions.9l In the case of [2,6-(MeOCH2)2C6H3]SnPh2Cl (B1) and [2,6-(MeOCH2)2C6H3]SnPhCl2 (B2),7 the oxygen atoms are, albeit weakly, coordinated to the central atom as well. Similar autoionization (as in A1+) followed by elimination of an alkyl halide led to the isolation of a neutral compound [2-(OP(O)(OEt))-4-tBu-6-(P(O)(OEt)2)C6H2]SnPh2 (C1)24 in the case of Jurkschat's O,C,O-ligand (Fig. 3), while both oxygen atoms are coordinated in the diorgano-compound [2,6-(P(O)(OEt)2)2-4-tBu-C6H2]SnPhCl2 (C2) leading to a distorted octahedral geometry.6 Mono-organotin compounds [2,6-(Me2NCH2)2C6H3]SnBr3 (A3), [2,6-(MeOCH2)2C6H3]SnCl3 (B3) and [2,6-(P(O)(OEt)2)2-4-tBu-C6H2]SnCl3 (C3) again have both donor atoms coordinated with the tin atom giving an octahedral array around the tin atom, but the pincer ligand adopts either a pseudo-meridional (A39k and C324) or a pseudo-facial (B37) coordination mode. In contrast, in the case of 3, only one of the P atoms is sufficient for stabilization of the SnCl3 unit, while the second phosphorus remains pendant as proven both in solution (NMR) and in the solid state (sc-XRD). This results in a distorted trigonal bipyramidal array around the tin atom in 3 and underlines promising potential of the ligand to stabilize highly Lewis acidic species (vide infra).
To enforce the obviously accessible intramolecular P → Sn interaction(s) (as found in 3), 1 was reacted with 1 eq. of AgSbF6 aimed at the production of the corresponding organotin cation [ArSnPh2][SbF6] (I, Fig. 4). However, in situ inspection of the reaction mixture by 31P{1H} NMR spectroscopy showed the formation of three species (Fig. 4), which is evidently the result of a high Ag+ ion affinity toward the phosphorus donors that hampered a clean abstraction of the chloride from the tin center. The target cation I revealed only one signal at δ(31P) = 105.8 ppm (1J(119/117Sn,31P) = 819 Hz), which is highly comparable with that of the subsequently isolated cationic pair 1+[BArF]− (cf. δ(31P) = 105.7 ppm, 1J(119/117Sn,31P) = 835 Hz, vide infra). The second species was tentatively assigned to a complex of I with incipient AgCl, i.e. compound II (Fig. 4), exhibiting two (1:
1) signals. The first signal, attributed to the phosphorus atom coordinated to tin, resonated at δ(31P) = 118.6 ppm (1J(119/117Sn,31P) = 239 Hz), closely resembling the value found for the ionic compound [1(BH3)]+[BArF]−, where again only one phosphorus coordinates to the tin center (cf. δ(31P) = 116.2 ppm, 1J(119/117Sn,31P) = 285 Hz, vide infra). The second signal at δ(31P) = 153.0 ppm was formed as a doublet of doublets, indicating a clear interaction with the silver atom as proven by 1J(109/107Ag,31P) = 584/500 Hz. Finally, the most intense signal in the mixture was tentatively assigned to a simple donor–acceptor complex of 1 with AgSbF6, i.e. {μ-(P,P)-Ag-[2,6-(tBu2PO)2C6H3]SnPh2Cl}2(SbF6)2 (III), observed as a doublet of doublets at 169.4 ppm (1J(109/107Ag,31P) = 571/495 Hz). To prove the proposed structure of the major product III, compound 4 lacking any chlorine available for the formation of AgCl, but containing pendant phosphorus donors at the same time, was treated with AgX (X = OTf or SbF6). In fact, this reaction readily provided the expected complexes {μ-(P,P)-Ag-[2,6-(tBu2PO)2C6H3]SnBu3}2(X)2 (X = OTf or SbF6), i.e. 4·AgOTf and 4·AgSbF6, according to Scheme 2 as direct analogues of proposed complex III.
![]() | ||
Fig. 4 31P{1H} NMR spectrum showing the reaction mixture after addition of AgSbF6 to compound 1. *Denotes unknown minor species. |
Both compounds revealed a set of expected signals in their 1H and 13C{1H} NMR spectra. The 119Sn{1H} NMR spectra showed one signal at −43.3/−42.3 ppm for 4·AgOTf/4·AgSbF6, respectively, close to that of the starting 4, whereas the 31P{1H} NMR spectrum contained a doublet of doublets for each complex at 170.0/169.7 ppm (1J(109/107Ag,31P) = 568/494 Hz for 4·AgOTf, 1J(109/107Ag,31P) = 567/497 Hz for 4·AgSbF6) almost identical to that of complex III. Finally, the molecular structures of both complexes were established by sc-XRD analysis for 4·AgOTf (Fig. 5) (Fig. S89 for 4·AgSbF6) and structurally they are closely related, thus only that of 4·AgOTf is described in more detail. The triflate anions are located outside the coordination sphere of silver atoms, while these are coordinated by two phosphorus from two different ligands, leading to a centrosymmetric dimeric dication (Fig. 5). The Ag(1)–P(1/2) bond lengths of 2.4044(4)/2.4017(4) Å correspond closely to the ∑cov(P,Ag) = 2.39 Å (ref. 20) while the P(1)–Ag(1)–P(2a) bonding angle is 166.05(2)°. The organotin fragments are directed away from the center of the molecule, and tin atoms adopt a distorted tetrahedral coordination geometry.
The utilization of a low-nucleophilic anion silver salt for the abstraction of the chloride from 1 turned out to be non-selective and complicated due to the formation of a P–Ag complex, therefore the sodium salt Na[BArF] ([BArF] = [3,5-(CF3)2C6H3]4B) was used where significantly lower tendency of sodium ions to complex with pendant phosphorus atoms was expected. Indeed, this approach allowed smooth isolation of a full set of organotin(IV) cations (Scheme 1), i.e. [ArSnPh2][BArF] (1+[BArF]−), [ArSnPhCl][BArF] (2+[BArF]−) and [ArSnCl2][BArF] (3+[BArF]−), as crystalline solids in quantitative yields. The 1H and 13C{1H} NMR spectra in CD2Cl2 contained anticipated sets of signals for the ligand, phenyl moieties as well as for the [BArF]− anion for these compounds. The presence of the [BArF]− anion was also corroborated by 11B{1H} (signal at −7 ppm) and 19F{1H} (signal at −62.8 ppm) NMR spectra. The 31P{1H} NMR spectra revealed one signal for 1+/2+ at 105.7/97.2 ppm, both being significantly high-field shifted compared to starting 1 and 2 (Table 1). 3+ showed a signal at 86.2 ppm, indicating symmetric coordination of both donor atoms of the pincer ligand resembling the chemical shift value of the coordinated phosphorus atom in 3 (Table 1). 119Sn{1H} NMR signals were detected as triplets for 1+/2+ at −187.6/−214.4 ppm and these are high-field shifted compared to those for 1 and 2, indicating that both phosphorus atoms are coordinated to the central tin atom, unlike the neutral species, also leading to an increase in the coordination number of the tin atom to five (Table 1). In contrast, the chemical shift value of −194.0 ppm observed for 3+ is low-field shifted compared to that for 3 (−274.1 ppm). This finding probably reflects two contradictory factors that influence the shielding of tin in 3+, i.e. the positive charge on the central atom vs. the coordination of the second phosphorus atom, while the tin atom formally preserves its 5-fold coordination similarly to 3. A clear trend is also found among values of 1J(119/117Sn,31P) that amount to 835/640/88 Hz for 1+/2+/3+, respectively, i.e. becoming lower with increasing relative strength of the P → Sn interaction.
The structures of cations 1+ and 3+ are depicted in Fig. 6, but all attempts to determine the structure of 2+ resulted in heavily disordered structures only. The anions are situated outside the metal coordination sphere. The tin atom is tightly coordinated in both cationic parts by phosphorus atoms with the bond lengths Sn(1)–P(1)/(2) of 2.808(3)/2.7606(18) Å and 2.6689(7)/2.6647(7) Å for 1+/3+, respectively. The shorter distances detected in 3+ reflect the presence of a more Lewis acidic center (cf. ∑cov(P,Sn) = 2.51 Å (ref. 20)). The obtained values are still longer than in related highly Lewis acidic tin(IV) compounds, such as [SnMe3(PMe3)]+[AlCl4]− (2.5861(9) Å)25, [SnCl3(OTf)(PMe3)2] (2.5496(9)/2.5506(9) Å)26 or chelates [SnBu2Cl(Me2P(CH2)2PMe2)]+[AlCl4]− (2.5696(8)/2.7601(8) Å) and [SnBu2(Me2P(CH2)2PMe2)]2+{[AlCl4]−}2 (2.5654(9)/2.521(9) Å).25
The coordination polyhedron is described in both cases as only slightly distorted square-pyramid with the C(29) and Cl(2) atoms in the apical position with τ = 0.05/0.1521 for 1+/3+, respectively.
Some analogues of 1+ containing both N,C,N-, i.e. [(2,6-(Me2NCH2)2C6H3)SnPh2]+ (A1+)22 and O,C,O-ligands, i.e. [(2,6-(MeOCH2)2C6H3)SnPh2]+ (B1+)10b and [(2,6-(P(O)(OiPr)2)2-4-tBu-C6H2)SnPh2]+ (C1+)27 were structurally characterized showing two strong intramolecular interactions with donor atoms (Fig. 6), while the coordination polyhedron is somewhat more distorted in direction toward the trigonal bipyramid according to the τ21 value (cf. 0.45/0.38/0.55 for A1+/B1+/C1+, respectively vs. 0.05 in 1+). 3+ containing [ArSnCl2]+ (τ = 0.15) represents, to the best of our knowledge, the first example of such an organotin pincer cation reported to date.
The square-pyramidal arrangement in the solid state structures of 1+ and 3+ can be traced back to the presence of bulky tBu groups at the P centers. According to DFT calculations, the model analogues with methyl substituents instead of tBu reveal a significant increase in the τ parameters to 0.48/0.55 for the two hypothetical cations 1+(Me)/3+(Me), respectively, thus more resembling their N,C,N- and O,C,O-ligand counterparts. Note that the optimized geometries of the original 1+/3+ cations are highly similar to the solid-state structures with τ = 0.01/0.19, respectively.
To further experimentally elucidate the strength of the P → Sn interaction(s) in the studied compounds, all were treated with [BH3(SMe2)] to determine whether they coordinate with the phosphorus donor atom(s). Not surprisingly, in the case of ArBr (taken as a model) and compounds 1, 2 and 4, both phosphorus atoms are smoothly coordinated toward the borane as no significant P → Sn interaction was observed in the parent compounds (Scheme 3). Thus, the compounds ArBr(BH3)2, 1(BH3)2, 2(BH3)2 and 4(BH3)2 could be isolated and characterized using multinuclear NMR spectroscopy (see the SI) and sc-XRD analysis in the case of ArBr(BH3)2, 1(BH3)2 and 2(BH3)2 (Fig. 7).
![]() | ||
Scheme 3 Structures of isolated borane adducts obtained by reaction of parent compounds with 2 eq. of [BH3(SMe2)], along with isolated yields given in parentheses. Note: [BArF] = [3,5-(CF3)2C6H3]4B. |
In the case of 3, only one of the phosphorus atoms could be blocked by BH3 giving complex 3(BH3), while using an excess of borane did not alter the result of the reaction. This fact is reflected by the observation of three signals for aromatic CH groups of the Ar ligand in the 1H NMR spectrum pointing to a non-equivalence of ligand arms. Furthermore, two signals were detected in the 31P{1H} NMR spectrum, i.e. for the phosphorus atom bound to the tin atom at 70.6 ppm (1J(119/117Sn,31P) = 211 Hz) and for the second one at 152.8 ppm, which is broadened due to the coupling with boron nuclei. The 11B{1H} NMR spectrum revealed a signal at −37.8 ppm for the coordinated BH3. The value of δ(119Sn) = −278.5 ppm is only marginally shifted compared to that for 3 (cf. −274.1 ppm) indicating that the tin atom retains its coordination number of five as in the parent compound.
Treatment of 1+[BArF]− with [BH3(SMe2)] quite surprisingly also gave a 1:
1 complex with BH3, i.e. [1(BH3)]+[BArF]−, despite the fact that both phosphorus atoms were quite tightly coordinated to the tin atom in the starting compound. In contrast, no reaction was obtained for 2+[BArF]− and 3+[BArF]− reflecting the presence of more Lewis acidic tin centers that prevent de-coordination of phosphorus donor atoms from the metal center (vide infra). The 1H and 13C{1H} NMR spectra of the cation [1(BH3)]+ proved the nonequivalence of both ligand arms. The 31P{1H} NMR spectrum contained two signals: one at 116.2 ppm (1J(119/117Sn,31P) = 285 Hz) and the second at 156.0 ppm and, similarly, two signals were detected in the 11B{1H} NMR spectrum for the [BArF]− anion at −7.2 ppm and for the coordinated BH3 at −40.1 ppm. The signal for the tin atom in [1(BH3)]+ at −88.5 ppm is low-field shifted compared to that for the parent 1+ (−187.6 ppm), reflecting the absence of one of the phosphorus donors in the tin coordination sphere after being blocked by the borane. The molecular structure is shown in Fig. 7 and, as expected, the P(2) is coordinated by the borane unit (P(2)–B(1) 1.917(5) Å). The P(2)–Sn(1) bond length is very short at 2.5825(9) Å, approaching the value of ∑cov(P,Sn) = 2.51 Å (ref. 20) and, not only is it shorter than that in 1+, it is even shorter than the bond length in the above described cation 3+. It is also comparable to the value found in [SnMe3(PMe3)]+[AlCl4]− (2.5861(9) Å).25 This evidently results only from the four-coordinated tin cation that adopts a distorted tetrahedral geometry in [1(BH3)]+.
Starting from the solid-state structures, we conducted conformational analysis searches for each of the compounds. After a set of low-energy isomers was located, the geometries were further optimized at the DFT level. Among the several isomers considered, the structures that were similar to those determined by sc-XRD were always proven to be the most stable. In most cases, it was possible to optimize geometries with three different bonding motifs, that is, 2, 1, or none of the phosphorus atoms establish dative bonds with the Sn centre, as unequivocally characterized by bond-critical points between the phosphorus and the tin centres. Moreover, in the case of complexes with no or one P → Sn dative bond, the spatial arrangement of the uncoordinated P centre was also tested, but we found that it only has a negligible influence on the stability of such complexes (with respect to the phenyl group, the in-plane position is more stable than the out-of-plane position). For more details and geometrical parameters, see the SI.
In contrast, cationic complexes 1+–3+ differ markedly from neutral congeners. In these, the P → Sn interaction has a remarkable effect on the geometry and stability of the complexes. In all of the cases, both P centres establish interaction with the Sn centre. If one of the P → Sn interactions is absent, a significant destabilization is observable in terms of relative energy compared to the isomer with two P → Sn bonds (ΔE = 7.3/13.9/17.6 kcal mol−1, for 1+/2+/3+, respectively; see Tables S6–S8). This trend corresponds well with the finding that only one of the phosphorus atoms in 1+ is blocked by borane leading to [1(BH3)]+. Cleaving the second P → Sn bond results in very strained structures, which are highly destabilized with relative energies of ΔE = 39.5–61.4 kcal mol−1, indicating the importance of the dative interactions for stabilization of the cations. However, in these geometries, a dative bond between the O and Sn centres forms in line with the high Lewis acidity of the tin centre (see the SI).
We also used relaxed scan calculations to screen the energy dependence upon changing the Cipso–Cortho–O–P dihedral angle. The rotation barrier is significantly lower in the neutral complexes (11.3–12.9 kcal mol−1), than in their cationic counterparts (15.9–27.9 kcal mol−1), in line with weaker P → Sn interactions in the former (Table S10).
In the following, we will discuss only the isomers having the lowest energies, which also resemble the solid-state structures.
To quantitatively assess the Lewis acidity of the Sn centre and the stabilization effects offered by the P → Sn interaction, the energy of the model complex formation reaction below (termed as interaction energy, ΔEint) has been calculated for x = 0, 1, 2, 3 and 4, as well as for the mono-cationic counterparts (x = 0 to 3). The complex formation reactions utilizing the model phosphine (PhO)Ph2P for obtaining the interaction energies ΔEint:
Importantly, these model reactions deliver the binding energy of a phosphorus centre to the tin acceptor in the absence of any geometrical constraints of the chelating backbone and thus describe the inherent strength of the P → Sn interaction in general. It should be mentioned that simpler phosphines such as PH3 or PH2(OMe) were also tested in addition to PPh2(OPh); however, in most cases, the adducts with them underwent spontaneous dissociation during the optimisation runs. Nevertheless, the PPh2(OPh) as a Lewis-base seems to be a realistic model for quantifying the strength of the P → Sn interactions (Table 2).
The computed reaction energies for complex formation show that in neutral complexes the formation of the P → Sn bond is somewhat exothermic, but endergonic. This is in line with the moderate strength of the P → Sn interactions and explains the small relative energy difference between the isomers and the high degree of conformational flexibility of the complexes. In the case of the SnPh3Cl and the SnPh4, not even an adduct with P(Ph)2(OPh) could be accessed (no minimum on the potential energy surface could be located), while for the other stannanes, the interaction energy is practically independent of the number of Cl and Ph centres (slightly decreases with the number of Cl centres). These results indicate that the interaction between the P and Sn centres is of a weak, non-covalent nature. In contrast to the neutral analogues, in the case of the cations [SnPhxCl3−x]+ the forming P → Sn interactions effectively stabilize the complexes with significant reaction energies (−56.8 to −93.9 kcal mol−1). These complex formation reactions gradually become more exothermic and exergonic with the number of Cl centres in the tin model cations, in line with the increasing Lewis acidity of these centres.
Analogous to the experiments mentioned above, we also obtained the borane affinities (ΔEBA) of the complexes 1–3 and 1+–3+ as the energy of their reactions with BH3 leading to the corresponding borane adduct. Borane affinities indirectly quantify the strength of the P → Sn interaction (P → B bonds are expected to be similar in the formed adducts). The calculated interaction energies and borane affinities show a clear trend: the more negative the interaction energies (ΔEint), the less prone the P centre is to cleavage of the P → Sn interaction and the less negative the borane affinity of the complex (Fig. 8 and Table S11). Again, the neutral complexes are rather similar to each other, their borane affinities are rather low (exothermic reactions), and the corresponding interaction energies are insignificant. This indicates a high tendency to cleave the P → Sn bond, which is in good agreement with the observation that all neutral congeners may capture one (3) or even two BH3 Lewis acids. In contrast, in the case of cationic complexes, a monotonous increase in borane affinities can be observed in the direction 1+ → 2+ → 3+. This means that changing the P → Sn bond to a P → B interaction becomes gradually more difficult in this direction. Nevertheless, the negative borane affinities corresponding to complexes 2+ and 3+ would suggest that the attack of the borane is thermodynamically feasible. However, the activation barrier to cleave the P → Sn bond is large for these complexes (23.3 and 27.9 kcal mol−1, respectively, unlike the other cases) and thus, borane addition is kinetically hampered. This observation is in agreement with the outcome of the experiments, in that [1(BH3)]+ could only be obtained by coordination of one molecule of borane to 1+. In general, the Lewis acidity of the tin centre decreases with the increasing number of the phenyl groups (while the steric crowding increases), leading to the weakening of the P → Sn interaction.
![]() | ||
Fig. 8 BSSE-corrected interaction energy and borane affinity values calculated at the ωB97X-D/def2-TZVP//ωB97X-D/def2-SVP level of theory. |
Finally, to further supplement the observations from the heteronuclear NMR experiments, we simulated the 31P and 119Sn chemical shifts in these complexes (Tables S12 and S13). Nuclear shielding parameters were calculated considering the scalar and spin–orbit effects. In the case of 119Sn chemical shifts the standard is SnMe4 (δref(119Sn) = 0 ppm), while for the 31P chemical shifts ArBr employed in the experiments was chosen (δref(31P) = 154.2 ppm). The latter was selected, as PH3, which is typically applied as standard in DFT computations, was found to be a less suitable reference due to the complexity of the P,C,P-chelate complexes and the strongly differing chemical environment. As the NMR spectra were recorded in non-coordinating and lower polarity solvents (C6D6, CD2Cl2 and CDCl3), the data values calculated without solvation models show good agreement with the experimental ones. In general, the calculated 31P and 119Sn chemical shifts show excellent correlation with the experimental data (R2 = 0.989, see Fig. S90). However, for the latter, a precise description of the spin–orbit coupling is required (R2 = 0.979, see Fig. S91).
The 119Sn NMR chemical shifts and 1J(119Sn,31P) coupling constants are also practical indicators for elucidating the geometries of organotin compounds, especially with regard to the coordination sphere around the tin centres. In complexes 1 and 2, the 119Sn NMR shifts reveal four-coordinated tin centres, in line with the singlet resonances lacking any visible 1J(119Sn,31P) coupling in the experimental spectra. In contrast, the significantly more shielded tin centre in complex 3 signifies a hypervalent tin centre. Based on the computationally obtained geometries, a penta-coordinated or a hexa-coordinated (with four stronger and two weaker donations) motif seems possible. The former is consistent with the geometries obtained by sc-XRD. In the case of the cationic complexes 1+–3+, the signals are considerably shifted up-field, indicating structures with penta-coordinated tin centers, stabilized by two strong P → Sn interactions. A further trend can be observed in the 31P NMR chemical shifts corresponding to the coordinated P centres, which are more shielded compared to ArBr. In cationic complexes, with an increasing number of chlorine substituents at the Sn centre, the 31P NMR chemical shifts gradually decrease (105.7/97.2/86.2 ppm). This is in good agreement with the intensification of Lewis acidity of the Sn centres exemplified by the complex formation reaction energies (−37.2/−43.7/−57.4 kcal mol−1) and with the increase of WBI indices (0.45/0.51/0.54). The lower 31P NMR chemical shifts indicate a stronger dative bond, with a higher covalency triggered by the more Lewis acidic tin centre.
In general, the trend of the calculated 1J(119Sn,31P) coupling constants follows that of the experimentally obtained values (see Table 1). In certain cases, the numerical agreement is excellent, whereas for several compounds, significant deviations can be found. The estimation of 1J(119Sn,31P) values is especially challenging due to the presence of the heavy element. It should be noted that the coupling constants are in general strongly influenced by minor changes in geometrical parameters (see Fig. S92–S94), often resulting in substantial deviations between the experimental and calculated values.29 Indeed, the coupling constant shows a strong dependency on the Sn–P distance and dihedral angles (see Fig. S95 and S96).
Owing to the fact that p-block complexes with O,C,O-, but especially N,C,N-pincer ligands,4,10,13 have for a long time been among the most successful, interesting, and continuously explored main group compounds, the introduction of totally unexplored heavier P,C,P-analogues is, in our opinion, a highly desirable step forward in this field. This work should help in this effort, and exciting developments in this direction are envisaged in the near future.
CCDC 2469044 (2), 2469048 (3), 2469047 (4·AgOTf), 2469052 (4·AgSbF6), 2469053 (1[BArF]), 2469049 (3[BArF]), 2469051 (1(BH3)2), 2469046 (2(BH3)2), 2469050 ([1(BH3)][BArF]) and 2469045 (ArBr(BH3)2) contain the supplementary crystallographic data for this paper.30a–j
This journal is © The Royal Society of Chemistry 2025 |