New Hypercoordinating Organostannanes for the Modular Functionalization of Mono- and Polystannanes: Synthetic and Computational Studies

A series of potentially hypercoordinate tin compounds derived from a substitutionally labile stannane was produced to gain access to a library of stannanes and polystannanes for structure/property investigations. Three model triphenylstannanes, containing either a propyloxybiphenyl (11), propylmethoxy (12) or propylthioester (13) group were synthesized in high yields via substitution reactions of the propyl tosylated stannane 4. Compounds 12 and 13 were converted to the appropriate mono-(14, 15) and dichlorido-(18, 19) stannanes via sequential chlorinations with HCl. Further transformation of 18 or 19 to the dihydridos (22, 23) was carried out with the use of an appropriate reducing agent. Structural characterizations by single crystal X-ray diffraction of 12, 14, 18 and 19 were also undertaken and are discussed. Several DFT methods were compared for accuracy in predicting the hypercoordinate geometries of these compounds. The relative energies of hypercoordinate conformers for the propylmethoxystannanes 12, 14, 18, and 22 were determined and the fractional abundance of each conformer in the gas and solution (CHCl3) phase was estimated. Relativistic DFT calculations of 119Sn NMR chemical shifts were carried out for a series of non-hypercoordinate reference compounds and the conformers, allowing the estimation of Boltzmann-averaged chemical shifts of the propylmethoxystannanes. A semi-crystalline homopolymer (25) was isolated from the dehydropolymerization of 22 using Wilkinson’s catalyst. Conversion of the liable tosylated polystannane (24) to a new partially substituted polystannane (28) via nucleophile displacement reactions was achieved. The structures of model stannanes, chlorinated stannanes, hydrido stannanes and polystannanes were confirmed via NMR (1H, 13C, 119Sn) spectroscopy, HRMS, and, in the case of the polymers (25, 28), also by elemental analysis, GPC, DSC and PXRD (25).


Introduction
Polystannanes are main group polymers with a conjugated backbone of tin atoms prepared from the reductive coupling of dihalido-or dihydrido-diorganostannane monomers. 1,2 Sodium (Wurtz) [3][4][5] or electrochemical coupling 6,7 of X2SnRR' leads directly to the desired materials, while conversion of such materials to diorganostannanes, H2SnRR' followed by transition metal catalyzed dehydropolymerization [8][9][10] affords the target polystannanes. Key to both approaches is access to both homo-and hetero-diorganostannanes. This is traditionally achieved by the Kocheshkov reaction, a solventless redistribution of organo and halogen substituents at tetracoordinated tin centers to afford the required organohalo (R4-nSnCln) tin species. 1 Experimentally, alkyl substituents have been more difficult to selectively interchange than aryl ones (n-Bu < Me < Ph < Mes). 1 Interestingly, an increase of coordination number at tin from four to six has a notable impact on these redistribution reactions. Utilizing van Koten's N,C,N pincer ligand, 11 (2,6-(Me2NCH2)C6H3)-, Růžička and coworkers reported a partial redistribution at RT of an intramolecularly bound monochlorostannane (Figure 1, 1) in solvent to the dichlorostannane and the triphenylstannane after a 2 week time period. 12 Nechaev and co-workers later detailed the high yield redistribution reaction of six coordinate PhSn(κ 2 -O, N-OCH2CH2NMe2)2Cl (Figure 1, 2) with PhSnCl3 to the symmetrical dialkoxydichlorostannane and Ph2SnCl2 species. 13 In both examples, the presence of hypercoordinating interactions from the tethered donor side chains lead to a "reversed" redistribution of these organotin compounds. Functionalization of stannanes from chlorido or hydrido triorganostannanes has been demonstrated by the teams of Pace and Molloy, respectively. Tandem homogolation of triorganohalostannanes with modified methyl lithium reagents (e.g. LiCH2Cl) yielded α-functionalized monohalomethyl stannanes (3) in good yields (> 80%). 14 Molloy reported the preparation of dihalidodiorganostannanes 4 ( Figure 1) from an initial hydrostannylation of triphenylstannane with an allyl ether or allyl alcohol substituents, followed by stoichiometric bromination. 4 Our research 15 has demonstrated displacement reactions of the tosyl substituent in 5 ( Figure 1) with 4-hydroxyazobenzene under Williamson ether-like conditions. This approach may be an effective general method to prepare a variety of hypercoordinating stannane monomers for moisture and light stable polystannanes. Access to a wide variety of functional organic polymers from a single sacrificial macromolecular intermediate has been a successful strategy for the last three decades. 16 Post-polymerization modification of an intermediate using comparatively mild transformations enables the preparation of polymers that are otherwise hard to access, or sensitive to the polymerization conditions. Additionally, new polymers derived from a single, high yielding, low-cost macromolecular intermediate provides an opportunity to probe structure-property relationships between materials built from a common modular structural platform. Post-polymerization of main group and organometallic polymers is relatively unexplored with a few notable exceptions. Modification of poly(methylhydrosiloxanes), 6 ( Figure 2), with vinylsiloxanes via transition metal-catalyzed hydrosilylation 17 and the nucleophilic substitution of the halo substituents of thermally ring-opened hexachloropolyphosphazene 18 (7) or the closely related polythionylphosphazene (8) with alkoxy, amino or organo substituents has been established. 19 In the mid 1990's, Manners and coworkers also demonstrated the tunable nature of thermally or transition metal-catalyzed ring-opened chlorinated sila [1]ferrocenophanes (9a, b) and were able to prepare substituted poly(ferrocenylsilanes) with sp, sp 2 and sp 3 moieties. 20 In 2017, this same group was able to demonstrate an "inorganic click" reaction with the cycloaddition of cyclophosphines to a polymeric styrene nitrile to yield 10. 21 This approach provides access to a host of tailored polymers bearing inorganic moieties. To gain access to a broader selection of hypercoordinated stannanes and polystannanes, we have carried out a three-part study. First, we have investigated the substitution chemistry of a model tosylstannane with three nucleophiles, preparing a library of small molecules using reactions potentially extensible to post-polymerization modifications. Second, we have estimated the extent of hypercoordination in solution for a subset of these small molecules using DFT calculations of geometries and 119 Sn NMR chemical shifts, with comparison to experimental values. Finally, we disclose our first efforts to prepare tunable hypercoordinated polystannanes by post-polymerization modification of a previously known tosyl-containing polystannane. 17 Functionally similar polymeric materials were also prepared by dehydrocoupling, where appropriate, for comparison. The tosylstannane 4, 15 was prepared in two high yield steps following Yoshida's procedure without the need for column chromatography. 23 This air-and moisture-stable tosyl intermediate is also tolerant to acidic conditions, while substitutionally reactive under basic conditions. Conditions favourable for substitution at 4 ( Figure 3) with the methoxy, thioacetyl and biphenyloxy nucleophiles were explored. Williamson ether and Williamson ether-like conditions (Scheme 1) in acetone (Method A) were initially used to prepare compounds 11 and 13 in low yields (< 40%) but was unsuccessful in the case of 12. Higher yields of both 11 and 13 were obtained (> 75%) when the reaction was carried out in MeCN (Method B). High purity materials were isolated after extraction (without column chromatography) as determined by NMR spectroscopy ( 1 H, 13 C, 119 Sn), and for 11 was identical to material previously prepared by hydrostannylation protocols. 4 Compound 13 was recovered as a yellow-coloured clear oil, in contrast to solid 11. Compound 12 was readily prepared in good yield as a white coloured solid using NaOMe/MeOH in MeCN under mild conditions (< 2 h: Method C). As MeCN was also found to be preferable for substitutions of both the model stannanes and the tosyl polystannane, it was used exclusively thereafter. Compound 12 was recrystallized from DCM/hexanes and its solid-state structure established by X-ray crystallography ( Figure 4: top left). This model triphenylstannane assumes a distorted tetrahedral geometry (τ4 = 0.52) 24 around the Sn centre.  The new model stannanes 22 and 23 were prepared using established protocols. 19,20 The three step process involved sequential chlorination of the model tristannanes (12,13) to monochlorides (14,15), then dichlorides (18,19) and finally, reduction to the dihydrido (22,23) compounds with a mild hydride source (Scheme 2).

Scheme 2:
Preparation of functional stannanes used in this study.
Stannane 14 was obtained through an initial chlorination of 12 (>99% yield), with further conversion (>95%) to the dichlorinated stannane 18 after a second treatment with chlorinating reagent. Both crude chlorinated products were washed with a minimal volume of hexane, recrystallized from a solution of DCM and hexane and recovered as white crystalline solids. Similar transformations were achieved with compounds 15 and 19 in good yields and purity with isolation as a yellow-coloured semi-solid and orange-coloured crystals, respectively. The dihydridos (22,23) were isolated as white coloured semi-solids and kept at low temperature (-20 C) and in the absence of light. Large upfield 119 Sn NMR chemical shifts for these species were observed, consistent with a pseudo-tetrahedral, rather than trigonal bipyramidal geometry around tin. Further structural identification of the chlorinated methoxy-containing compounds 14 and 18 ( Figure 4: top right, and bottom left respectively) by single crystal X-ray analysis reveals significant hypercoordination between the methoxy substituent and tin centres, similar to that observed previously by Lebl

DFT Geometry Optimizations and Structural Comparisons
In two previous studies, 23,24 we compared a total of 19 X-ray crystallographic structures of Sn hypercoordinate compounds with those calculated by a selection of computational DFT methods to assess which level of theory provided the best structural predictions. Although comparisons between experimental solid state and calculated gas phase structures are problematic for a variety of reasons, 23,25,26 there are few other practical options. 27 In general, adding empirical dispersion to the computational methods significantly improved the accuracy of the calculated structures. The B3PW91 28 and PBE0 ("PBE1PBE") 29 functionals supplemented with Grimme's D3 empirical dispersion function and Becke-Johnson damping (GD3BJ) 30 were the best of ten methods studied; 23 the M052X 31 method with Grimme's D3 empirical dispersion (GD3) performed comparably. 24 In this study, nine crystallographic structures were compared using the three best DFT methods from the previous work as well as the MN15 32,33 functional with GD3BJ empirical dispersion (see SI, Table S-1 for details). The four DFT methods used in geometry optimizations gave similar results, with the PBE0-GD3BJ method generally providing the best prediction of solid-state molecular geometries (Table S-1), as previously observed. 23,24 The M05-2X method also performs well.

DFT Calculation of 119 Sn NMR Chemical Shifts for Hypercoordinated Stannanes Conformers in Solution
Extending the computational modelling of these flexible stannane compounds to the calculation of solution-state properties presents three major challenges. First, due to the propyl linkages and hypercoordinate interactions, each of these compounds has several dozen stable conformations. Determining the 119 Sn NMR chemical shift for a compound requires calculating the chemical shift for each geometrically optimized conformer, multiplying this by the conformer's Boltzmann-weighted fractional contribution at the measurement temperature and then summing these. A second challenge is due to the solution state model. Implementation of a polarizable continuum model (PCM) for the solvent with quasi-ideal-gas statistical mechanics calculations of thermodynamic properties causes significant overestimation of translational and rotational entropies and thus systematic errors in Gibbs energy estimates. 34 Finally, it is not known which DFT functionals most accurately predict energies for these conformers in a PCM calculation.
In this study, we adopt some simplifying approximations to make the problem more tractable. First, we focus on the methoxystannane series 12, 14, 18, and 22, because these compounds only have one potentially hypercoordinating substituent (the methoxyl O-atom) and are the simplest structurally of the model stannanes in this study ( Figure 3). Second, we assume that the most stable of the non-hypercoordinating conformers is the one with a fully extended propyl linker in an antimethoxyl conformation. Third, the calculated solution state Gibbs energies are used for estimating Boltzmann averaged properties for a set of conformers for a given compound. Because the conformers are substantially similar, their translational and rotational entropies are nearly identical and the overestimation cancels out in the Boltzmann averaging. On that basis, we consider the accuracy of the relative Gibbs energies in the solution calculations to be comparable to the gas phase ones. Fourth, we use the PBE0-GD3BJ DFT method, given its good performance in predicting solid state geometries. Each of the compounds has multiple hypercoordinate conformers. For 14, 18, and 22, the Sn substituent transto the hypercoordinating O-atom can vary (Ph and Cl or H) and there are also stable conformers with differing phenyl ring orientations. In hypercoordinate geometries, the methoxy-methyl can adopt an antior gaucheconformation with respect to C-2 of the propyl linkage. Only one gaucheconformer is observed because the other would bring the methyl group and Sn center into close proximity. The hypercoordinate and fully extended conformers for each of the methoxystannanes (12, 14, 18, and 22) were determined by geometric optimization without and with solvation (chloroform) and verified with frequency calculations. As expected, hypercoordinate conformers are the lowest energy species, with trans-Cl stabilized relative to trans-Ph in 14 and 18 and trans-Ph stabilized relative to trans-H in 22. The fractional abundance of each conformer for a given compound was estimated from the Gibbs energy differences. The amounts of the extended non-hypercoordinate conformers are negligible for 14 and 18, and minor for 12 and 22, suggesting that it is reasonable to ignore other higher energy nonhypercoordinate conformers. See SI and Table S-2 for details. The solvation model significantly alters the geometry of the hypercoordinate conformers, notably by reduction of the Sn-O distance (Table S-3). The most stable conformer remains the same for each of the compounds, but the relative energies and fractional abundances of the conformers change (Table S-2). The most stable hypercoordinate conformers for 12 and 22 are less stabilized relative to the extended conformers, while those of 14 and 18 are more so. Although the distribution of fractional abundances for the various conformers is similar with or without solvation, the differences are significant enough to influence the calculation of the 119Sn NMR chemical shifts.

DFT 119 Sn Chemical Shift calculations
The calculation of 119 Sn relative chemical shifts by DFT methods has been extensively studied 35 (for selected examples see Ref. [36][37][38][39][40][41][42] ). Non-relativistic calculations using the Gauge-Including Atomic Orbitals (GAIO) method and the allelectron IGLO II/III basis set are accurate for Sn compounds lacking other heavy atoms. 36,[38][39][40] Unfortunately, the IGLO II basis set for Sn does not appear to be in the public domain. Relativistic calculations using the zeroth order regular approximation with spin-orbit coupling (ZORA-SO) also give very good agreement with experimental results, even for Sn compounds with heavy atoms such as Br or I. 37 In this work, we apply the ZORA-SO method for 119 Sn chemical shift calculation to the conformer geometries optimized using the PBE0-GD3BJ level of theory. Using different levels of theory for geometrical optimization and chemical shift calculations is acceptable because the chemical shift is essentially a nuclear property influenced by electronic distributions rather than an intrinsically electronic property. Figure 5 shows the correlation between calculated and experimental 119 Sn chemical shifts for a series of non-hypercoordinate reference compounds with structural elements related to the methoxystannane substituents in this study and with a wide range of chemical shifts. The calculated geometries and 119 Sn NMR chemical shifts are for gas-phase molecules while the experimental chemical shifts were measured in a non-coordinating solvent (usually chloroform). A good linear agreement is obtained with a slope of 1.012, and a correlation coefficient of 0.9966. Because both experimental and calculated chemical shifts are relative to the chemical shift of SnMe4 defined as 0 ppm, the regression must pass through the origin. Using the same methodology, 119 Sn NMR chemical shifts were calculated for each of the methoxystannane conformers, as geometrically optimized in both gas and solvent phases ( Using the estimated fractional abundances of the conformers (Table S-2) and the calculated chemical shifts (Table S-3), Boltzmann-averaged chemical shifts for compounds 12, 14, 18 and 22 were calculated ( Figure 5, Table 1). Due to the relatively low Gibbs energy differences for the extended conformers of 12 and 22 in the solvent phase compared to the most stable hypercoordinate conformer, the corresponding calculated 119 Sn NMR chemical shifts are more approximate; higher energy non-hypercoordinate conformers would also need to be considered in the average. The Boltzmann averaged 119 Sn NMR chemical shifts for the gas phase geometries agree well with the experimental values, falling essentially on the linear correlation ( Figure 5). The values for the compounds calculated with the chloroform solvent model are significantly lower. On this basis, the PBE0-GD3BJ gas phase model appears to more accurately predict the conformer geometries/relative energies than the solution phase PCM: the condensed phase calculations seemingly overestimate the extent of hypercoordination and systematically underestimate the Sn-O distances. This points to a potential weakness in the "structural benchmarking" approach that has not been previously identified. The PBE0-GD3BJ method was originally selected on the basis of giving the best structural predictions for Sn hypercoordinate compounds in comparison with solid-state X-ray crystallographic structures. This approach may inadvertently factor in the effects of a condensed phase on relatively compressible interactions, such as the Sn-O hypercoordination, such that if the gas-phase calculation approximates it, a solution-phase calculation with the same functional will overestimate the strength and underestimate the distance. This suggests that alternative approach(s) are required for accurately modelling hypercoordinate interactions in solution. The tosyl-containing polystannane 24, as previously described, was prepared via a transition metal-catalyzed dehydrocoupling of 20 in a foil-wrapped Schlenk flask under a N2 atmosphere (Scheme 3). These precautions are to (i) avoid the possibility of light degrading the monomer, (ii) suppress homolytic cleavage of the resulting polystannanes and (iii) prevent the formation of stannoxanes from nucleophilic attack by moisture, respectively. Polymer 24 was purified by precipitation of a THF solution (3 mL) containing crude 24 into an excess of hexanes (100 mL) and thereafter recovered as an orange-coloured solid. GPC analysis in THF revealed a molecular weight (Mw) of 17 kDa and a Ð of 1.26. The tosylated polymer 24 was also characterized by both 1 H and 119 Sn NMR (C6D6) spectroscopy and was in agreement with the reported literature values. 17 A similar dehydrocoupling reaction of 22 was successfully carried out to yield polymer 25 which was recovered as a yellow coloured powder in moderate yields (64 %). A molecular weight determination of 25 after purification proved difficult as the polymer was almost completely insoluble in THF, the solvent required for analysis. Attempts to redissolve purified 25 in the original reaction solvent also proved futile, and all analyses were thus carried with crude 25 due to the apparent highly crystalline nature of the resulting polymer. NMR ( 119 Sn) analysis of the crude polystannane 25 prior to purification reveal evidence of partial hypercoordination with two nearly equally intense resonances at δ = -193 and -238 ppm ( Figure S53), similar to what was observed for the phenyl propyl hydroxyl stannane polystannane, 27 ( Figure 5). 17 These two signals represent both the "open" and "closed" configurations of the polymer where the methoxy group can potentially interact with a Sn centres. After a single purification, the polymer is notably less soluble and displays only a single 119 Sn resonance at -237.9 ppm ( Figure S54), suggesting that the polymer likely migrates to a fully "closed" form. DSC analysis of homopolymer 25 reveals reversible semi-crystalline/highly crystalline character with Tm = 127 °C, and Tg = 91 °C ( Figure S63), similar to that for 27 (Tm = 110 °C, and Tg = 49 °C) as previously reported. 17 The PXRD diffractogram of 25 ( Figure 6  Attempts to prepare the thioacetyl homopolymer 26 from 23 under similar protocols were unsuccessful owing largely to the extreme sensitivity of the polymer to light, which changed from a light yellow to a darker insoluble brown solid after a minimal exposure to light. Further explorations of this material were therefore abandoned.

Attempted Macromolecular substitution
Substitution of the tosyl substituent of 24 for a methoxy functionality was attempted (Scheme 4). After several attempts (Table S48), the optimized conditions for an SN2 type substitution of 24 utilizing a 10:1 mixture of toluene and MeOH and one equivalent of NaOMe stirring at RT for 1 week were identified. The crude yellow coloured polymer, 28, was gravity filtered to remove the resulting salts and the volatile components were then removed under reduced pressure. The substituted product, 28, was further purified by redissolving the crude polymer in a minimum of THF and induced precipitation from this solution by adding it to an excess of cold hexanes. NMR spectroscopic analysis ( 1 H, 119 Sn) of the substituted polymer suggests that 28 is ≈ 50% methoxy substituted. The 119 Sn NMR ( Figure S56) reveals chemical shift values typically of both the homopolymer 25 in open and closed conformations, as well as a resonance for the unsubstituted polymer 24. A GPC analysis of the partially substituted polymer was not performed, as purified 28 was found to be insufficiently soluble in THF.

Conclusion
Nucleophilic substitution of a labile model tosylated triphenylstannane 4 to either aryloxy, alkyoxy or thioacetate bearing stannanes in good yields was demonstrated. DFT optimization of individual molecules using the PBE0-GD3BJ method generally provided the best predictions of solid-state molecular properties. Modelling of hypercoordinate conformations for the methoxy-series 12, 14, 18, and 22 with PBE0-GD3BJ showed that inclusion of the PCM solvent model predicts significantly shorter Sn-O distances than in the gas phase. Relativistic DFT calculations of 119 Sn chemical shifts for these conformations with Boltzmann averaging closely approximated experimental values when gas phase geometries and conformer distributions were used but performed poorly with solvent phase ones. The selection of the PBE0-GD3BJ method by "structural benchmarking" gas-phase geometries to solid-state structures may have effectively factored in the impact of a condensed phase on the hypercoordinate interactions, causing the solvation model to overestimate these. A new light and moisture stable semicrystalline polystannane, 25, was prepared by the dehydropolymerization in the presence of Wilkinson's catalyst. Evidence for extensive hypercoordination was detected by 119 Sn NMR spectroscopy. Access to the similar, but only partially substituted methoxy stannane, 28, via an SN2 reaction involving a tosylated polymeric intermediate 24 in good yields has also been demonstrated using mild reaction conditions. Despite the sensitivity of the Sn-Sn bonds of the liable polystannane, prolonged reaction periods (> 1 week) of tosyl containing polystannane 24 with a nucleophile in the absence of light can allow access to functionally stable new polystannanes.

Experimental
General remarks 1 H NMR (400 MHz), 13  Time-of-flight mass spectrometry analyses were performed at the AIMS Mass Spectrometry Laboratory, University of Toronto using a JMS-T1000LC mass spectrometer (JEOL Inc., Peabody, MA) equipped with a Direct Analysis in Real Time (DART) ionization source (DART-SVP, Ionsense Inc., Saugus, MA). The DART source was operated with He gas and the temperature was adjusted in the range 100-400 C. Isotopic distributions for the observed ionic species were calculated using the Mass Center utility (JEOL) and were in good agreement with the measured mass spectra. Absolute molecular weights of the polymers were determined by GPC using a Viscotek Triple Model 302 Detector system. GPC columns were calibrated versus polystyrene standards (American Polymer Standards). A flow rate of 1.0 mL/min. was used with ACS grade THF as the eluent. GPC samples were prepared using 3-10 mg of each polymer per mL THF, and filtered using a 0.45 μm filter. All samples were run with and without UVA (conc. ≈ 0.001 M) for comparison. The X-ray diffraction data for compounds 12, 14, 18, and 19 were collected on a Bruker Kappa APEX-DUO diffractometer using monochromated Mo-K〈 radiation (Bruker Triumph) and were measured using a combination of  scans and  scans. The data were processed using APEX2 and SAINT programs. Absorption corrections were carried out using SADAB. The structures were solved using SHELXT 47 and refined using SHELXL-2013 47 for full-matrix least-squares refinement that was based on F 2 . For all structures, H atoms were included in calculated positions and allowed to refine in a ridingmotion approximation with Uiso tied to the carrier atom. All reactions were carried out under a nitrogen atmosphere using Schlenk techniques unless otherwise described. LiAlH4 (1.0 M in Et2O), NaBH4, 1.0 M solution of HCl in Et2O, anhydrous CaCl2, anhydrous MgCl2 and Wilkinson's catalyst were purchased commercially and used without further purification. Solvents were dried by standard procedures prior to use.

Method B.
In a 100 mL round bottom flask, compound 4 (0.50 g, 0.89 mmol), K2CO3 (0.16 g, 1.11 mmol) and 4phenylphenol (0.19 g, 1.11 mmol) were dissolved in ACN. The reaction was allowed to reflux for 72 h, after which the volatile components were removed under reduced pressure. The resulting product was redissolved in DCM and washed with 3 aliquots (50 mL) of a 5% aq. Na2CO3 solution. The organic phase was dried with anhydrous MgSO4 and filtered, and the volatile components removed under vacuum to afford an off-white gel. The crude product was tritrated with MeOH (3 × 20 mL) to yield the pure product 11 as a white powder. Yield: 0.49 g, 98%. The NMR spectrum obtained agreed with previously reported data. 26

3-(methoxypropyl)triphenylstannane 12
In a 100 mL round bottom flask, 4 (0.50 g, 0.91 mmol) and 25% w/v NaOMe/MeOH (0.70 mL, 2.64 mmol) were combined and dissolved in 50 mL ACN. The solution was heated under reflux for 24 h, after which the ACN was removed under reduced pressure. The product was then redissolved in DCM and washed with aliquots (3 × 50 mL) of a 5% aq. Na2CO3 solution. The organic later was dried with anhydrous MgSO4, filtered, and the DCM removed under reduced pressure. The resulting solution was recrystallized in heptane to yield the purified product as white crystals. Yield: 0.34 g, 89%.

Dichloro((3-phenylstannyl)propyl) methylbenzenesulfonate 16
In a 100 mL round bottom flask, 4 (1.13 g, 2.01 mmol) was dissolved in C6H6 (50 mL). 1M HCl in Et2O (2.01 mL, 2.01 mmol) was added to the stirring solution, and the mixture was allowed to react at RT. After 1h, 1M HCl in Et2O (2.01 mL, 2.01 mmol) was added and the reaction was allowed to stir for an additional hour. The volatile components were removed under reduced pressure and the resulting product recrystallized in heptane to yield the purified product as a white powder. Yield: 0.93 g, 97%. The NMR spectrum obtained agreed with previously reported data. 22

3-(phenylstannyl)propyl methylbenzenesulfonate 20
In a 100 mL Schlenk flask, 16 (0.50 g, 1.09 mmol) was dissolved in dry EtOH (12 mL) and allowed to stir on ice. In a 50mL Schlenk flask, NaBH4 (0.35 g, 9.36 mmol) was dissolved in dry EtOH (6 mL) and allowed to stir on ice. The solution containing NaBH4 was cannula transferred into the flask containing 16. The mixture was allowed to react at 0 °C for 1 h, after which the reaction was quenched with degassed water. The product was extracted with hexane, which was then washed once with degassed water (25 mL). The organic layer was dried with anhydrous MgSO4, filtered and the volatile components were removed under reduced pressure to yield the purified product as an off-white gel. Yield: 0.31 g, 69%.
The NMR spectrum obtained agreed with previously reported data. 22

S-(3-(phenylstannyl)propyl) ethanethioate 23
In a 100 mL Schlenk flask, 19 (0.19 g, 0.49 mmol) was dissolved in dry EtOH (12 mL) and allowed to stir on ice. In a 50 mL Schlenk flask, NaBH4 (0.17 g, 4.45 mmol) was dissolved in dry EtOH (6 mL) and allowed to stir on ice. The solution containing NaBH4 was cannula transferred into the flask containing 19. The mixture was allowed to react at 0 °C for 1 h, after which the reaction was quenched with degassed water. The product was extracted with hexane, which was then washed once with degassed water (25 mL

Tosyl homopolystannane 24
In a foil wrapped 100 mL Schlenk flask, a solution containing RhCl(PPh3)3 (0.027 g, 0.0291 mmol) in toluene (10 mL) and stirred at 0 o C for 15 min to activate the catalyst. Afterwards, 20 (0.30 g, 0.730 mmol) in toluene (10 mL) was added dropwise over 10 min to the catalyst solution and stirred at 0 o C for 4 h. The mixture was reduced in volume and dissolved again in dry THF. The solution was transferred dropwise to a 100 mL Schlenk flask containing a stirring solution of cold dry hexane (65 mL). A yellow orange precipitate was observed immediately and stirred for 5 min before settling. Solvent decanting was done using a Pasteur pipet, after which the residue of 24 was dried in vacuo to obtain a yellow orange solid. Yield: 0.20 g (67%). Mw = 17 kDa. Ɖ = 1.26.

Propylmethoxy Polystannane 25
In a 100 mL Al foil wrapped Schlenk flask, Wilkinson's Catalyst (0.008 g, 0.008 mmol) was added and dissolved in toluene (30 mL). The mixture was allowed to stir on ice for 20 min. In a separate 100 mL Al foil wrapped Schlenk flask 22 (0.12 g, 0.44 mmol) was dissolved in toluene. The solution containing 22 was added to the stirring solution containing Wilkinson's Catalyst dropwise. The mixture was allowed to react at RT for 4 h, after which the toluene was removed under reduced pressure. The crude polymer 25 was redissolved in minimal THF (2 mL) and precipitated in a stirring solution of cold hexane (50 mL

Preparation of a Substituted MeO Polystannane 28
In a 100 mL Al foil wrapped Schlenk flask, homopolymer 24 (0.45 g, 1.10 mmol) was added and dissolved in a 10:1 solution of toluene/MeOH (20 mL). Once dissolved, 25% w/v NaOMe/MeOH (0.23 mL, 1.06 mmol) was added to the solution. The mixture was allowed to stir at RT for 1 wk, after which the solution was filtered, and the volatile components removed. The crude polymer was dissolved in a minimal amount of THF (1-2 mL) and precipitated into a stirring solution of cold hexane (50 mL). The top layer of hexane was decanted, and the residual solvent was removed under reduced pressure to afford 28 as an off-white powder. Yield: 0.170 g, 57%.

Computational Details
The Gaussian 16 suite of programs was used for all geometric optimizations and frequency calculations (G16 Rev C.01). 48 DFT was implemented with four density functionals (PBE0 "PBE1PBE", M05-2X, B3-PW91, and MN15) supplemented with Grimme's D3 empirical dispersion function and, with the exception of M05-2X, Becke-Johnson damping. 30 The LANL08d basis set (obtained from the Basis Set Exchange) 49 was used for Sn, and the 6-31+G(d,p) basis set for all other atoms, as a reasonable compromise between accuracy and computational cost. 50,51 Tight convergence criteria and superfine integration grids were used for all geometry optimizations, in each case with verification by frequency calculation that a potential energy minimum had been located. Counterpoise corrections in calculations of dimers were applied during geometrical optimizations and frequency calculations. Cartesian coordinates for experimental structures were extracted from crystallographic information files (.cif) with OpenBabelGUI. 52 Solvation was implemented with the default Gaussian 16 method (the Polarizable Continuum Model using the integral equation formalism variant (IEF-PCM). 53 Conformer fractions for compounds were calculated by determining qi =αe -∆G/RT for each conformer i, where α is the total degeneracy of the conformer. Dividing qi by the sum of all q gives the fraction of conformer i. The Amsterdam density functional code (ADF/2017.207) 54 was used to calculate single point energies by the relativistic zeroth-order regular approximation (ZORA) method with optimized TZP basis sets and the PBE0 method with 50% Hartree-Fock exchange. 119 Sn chemical shifts were determined with the ADF nmr property module 55 at the PBE0-50/TZP spin-orbit level. SnMe4 (geometrically optimized as above) was used as the reference to determine relative chemical shifts. Boltzmann-averaged 119 Sn chemical shifts were determined by summing the product of the chemical shift and the conformer fraction for each conformer.

Conflicts of interest
"There are no conflicts to declare".