Probing the electronic origins of heterometallic tuning in platinum complexes by ¹⁹⁵Pt NMR spectroscopy

Abstract

Incorporation of secondary, effectively Lewis acidic metal cations into heterometallic complexes is a useful strategy for tuning chemical properties and reactivity, but addressing the origins of metal-driven tunability and quantitative correlation of metal-focused data with computational modeling have not received the attention they deserve. Here, the synthesis and characterization of a series of diamagnetic complexes containing platinum(ii) and secondary cations (K⁺, Sr²⁺, and Lu³⁺) are reported, along with nuclear magnetic resonance (NMR) measurements of ¹⁹⁵Pt chemical shifts (δ¹⁹⁵Pt). Similar to related species, the spectroscopic and electrochemical properties of the complexes depend on the Lewis acidity, or charge density, of the secondary cations; the change in δ¹⁹⁵Pt was found to be particularly well correlated with Lewis acidity and was therefore studied in detail. Computational investigations of the ¹⁹⁵Pt NMR data reveal that the secondary cations induce changes in electron donation to platinum from bridging phenoxide moieties that lead to the experimental chemical shifts. In sum, the results both (i) establish a quantitative model for understanding cation-driven tuning in complexes of this type as well as (ii) demonstrate the usefulness of ¹⁹⁵Pt NMR as tool for studying Pt-heterometallic compounds.

A joint experimental–theoretical study has been used to paint an electronic picture of the tuning induced by incorporation of secondary cations in a series of heterobimetallic complexes featuring Pt(ii) centers.

Introduction

Understanding how secondary metal cations drive the tunable properties of multimetallic coordination complexes represents an impactful topic, given the myriad of applications and possible uses of metal-containing materials with tailored properties.¹ For instance, heterometallic species typically afford access to properties that are inaccessible in homometallic analogues, making them attractive for development as homogeneous^2–4 or heterogeneous⁵ catalysts, as well as platforms for rational design of coordination complexes with particular properties.^6–8 Pioneering studies of compounds based on first-row transition metals established the ability of a range of cations to drive tunability and outlined the precise role(s) that secondary cations can play in heterometallic species.^9–12

The general principles of cation-driven tunability have been validated beyond first-row transition metals, setting the stage for more diverse work in the arena of heterometallic species.^13,14 For example, Blakemore and co-workers have demonstrated that U^VI/U^V reduction potentials can be tuned with secondary cations held in macrocyclic ligand environments,¹⁵ and that the symmetric [U–O_oxo] vibrational features in Raman spectra are correlated with cation identity as well.¹⁶ Non-macrocyclic uranium complexes also show tunability, as demonstrated by Ghosh and co-workers.¹⁷ Golwankar, Kumar, Day, and Blakemore found that the spectroscopic properties of macrocyclic palladium complexes can be tuned by secondary cations as well,¹⁸ building on older studies showing tunability in multicomponent systems containing ruthenium¹⁹ and palladium.²⁰ In the third row, a multicomponent system containing osmium,²¹ as well as defined complexes of iridium²² and rhenium²³ with mono- and divalent secondary cations have been reported.

Fewer studies have addressed the possible origins of the tuning effects from incorporating secondary metal cations into heterometallic species. Tsui, Agapie, and co-workers were, to the best of our knowledge, the first to show that shifts in reduction potentials were linearly correlated with the Lewis acidity of secondary metal cations incorporated into redox-active manganese compounds.²⁴ In this case, Lewis acidity was quantified by the pK_a values of the metal aqua complexes, [M(H₂O)_n]^m+, corresponding to the incorporated secondary cations.²⁵ Fukuzumi and co-workers have examined correlations between incorporated cations and spectroscopic signatures in electron paramagnetic resonance.²⁶ Yang, Alexandrova, and co-workers used theoretical calculations to address the cation-dependent optical absorption properties of bimetallic [M′,Na] and [M′,Ba] complexes (where M′ = Ni^II and Fe^III),²⁷ while Frontera and co-workers have examined electric field gradients in several compounds.²⁸ In order to obtain metal-specific information relevant to redox reactivity, Blakemore and co-workers examined the vibrational characteristics of heterobimetallic uranyl (UO₂²⁺)¹⁶ and vanadyl (VO²⁺)²⁹ complexes using Raman and infrared spectroscopy, respectively, finding strong linear correlations of [M–O_oxo] vibrational frequencies with Lewis acidity as estimated with the [M(H₂O)_n]^m+ pK_a values. Considering all the work to date, we hypothesized that charge density effects induced by secondary cations, readily parametrized by the [M(H₂O)_n]^m+ pK_a values reporting on effective Lewis acidity, give rise to the majority of cation-driven tunability effects measured to date. Obtaining evidence to support such a hypothesis is important to advance the field, however, because validation of this hypothesis would assist in the design of new tunable compounds and new applications of cation tuning.

Considering the potential impact of understanding the precise origins of cation-driven tuning effects, we envisioned that a study of platinum-based compounds would offer several appealing opportunities for the field of heterometallic chemistry. First, Pt(ii) compounds are typically diamagnetic and thus readily characterized by a variety of methods, including nuclear magnetic resonance (NMR). Second, platinum(ii) compounds tend to resist metal-centered oxidation, a feature that could contribute to stability in solution for extended spectroscopic investigations. Third and most importantly, ¹⁹⁵Pt is a naturally occurring (33.8% abundant) nuclide with both an appealing I = 1/2 nuclear spin and convenient Larmor frequency (ν_¹⁹⁵Pt = γ_¹⁹⁵PtB₀) of 107 MHz on a 11.74 T magnet (ν_¹H = γ_¹HB₀ = 500 MHz).^30,31 Direct observations of the NMR signals of transition metals are advantageous since the metal chemical shifts are very sensitive to chemical environment.³² Indeed, ¹⁹⁵Pt compounds display a wide chemical shift (δ) range,^31,33,34 spanning 10 000 ppm or more in prior work, making it potentially quite useful for studies of series of compounds that address environmental tuning.

In early work, Sadler and co-workers showed that δ¹⁹⁵Pt was linearly correlated with the number of amine (NH₃) ligands in cis- and trans-isomers of [Pt(NH₃)_x(H₂O)_4−x]²⁺ species, with more negative chemical shifts encountered for the species with more amine ligands (see Fig. 1).³⁵ Kaim and co-workers studied a series of aryl-ligated organoplatinum compounds, showing similar chemical shifts for related compounds but did not provide an analysis of trends, highlighting the challenge of correlating molecular properties with interpretation of δ¹⁹⁵Pt values.³⁷ Gilbert and Ziegler studied a series of halide-ligated species with both experiment and theory, demonstrating that the chemical shift tensor could be quantified with a series of individual contributions, including spin–orbit relativistic effects.³⁶ Autschbach and Le Guennic demonstrated early on that the key NMR parameters of ¹⁹⁵Pt in heterobimetallic complexes can be computationally determined with acceptable accuracy to allow detailed analysis of experimental results.^38,39 Most recently, Autschbach, Schurko, and co-workers used solid-state ¹⁹⁵Pt NMR data to analyze the relations between δ¹⁹⁵Pt tensors and Pt–ligand bonding for both Pt(ii) and Pt(iv).³³ As the state-of-the-art regarding ¹⁹⁵Pt NMR calculations have improved considerably in recent years,^33,40–42 we anticipated that a joint experimental-computational study of a series of Pt(ii)-containing heterobimetallic species featuring a range of secondary cations and for which ¹⁹⁵Pt NMR studies are conducted would likely afford insights into the origins of cation-driven tunability.

Fig. 1. Structures of selected compounds relevant to this study. Structures (a)–(d) show series of compounds studied by experimental ¹⁹⁵Pt NMR spectroscopy and/or computational approaches in prior work (reported in: (a) ref. 35, (b) ref. 36, (c) ref. 37, (d) ref. 34). Structure (e) shows the series of heterobimetallic platinum(ii) complexes studied in this report.

Herein, we report studies of a series of macrocyclic and heterobimetallic [Pt^II,M] compounds featuring a range of secondary Lewis acidic cations (M = K⁺, Sr²⁺, and Lu³⁺). Solution-phase spectroscopic and electrochemical measurements demonstrate that the secondary cations have a systematic influence on the Pt(ii)-containing macrocycle, including ¹⁹⁵Pt chemical shifts that are linearly correlated with Lewis acidity parametrized by [M(H₂O)_m]ⁿ⁺ pK_a values. Calculated ¹⁹⁵Pt NMR shifts and a localized molecular orbital (MO) analysis of the shielding constants identify how changes in electronic structure and bonding about the platinum centers give rise to the tunable spectroscopic properties. The experimental and computed properties reveal the particular importance of phenoxide moieties that bridge the metals. We thereby provide an experimentally validated molecular orbital-based view of cation-driven tuning, highlighting the future opportunities informed by atomistic understanding of Lewis acid effects on heterometallic species.

Results and Discussion

Synthesis and characterization

Macrocyclic Pt(ii) complexes remain rather uncommon,^43,44 with most prior examples supported by thiacrowns.^45–50 Golwankar and Ervin et al. detailed the preparation and isolation of a macrocyclic Pt(ii) complex capable of incorporating secondary metals, denoted [Pt].⁵¹ From this monometallic precursor, the desired heterobimetallic [Pt,M] complexes were synthesized by addition of 1 equiv. of the corresponding metal triflate (TfO⁻) salts into CH₃CN suspensions of [Pt] (see Fig. 2). Adducts of K⁺, Sr²⁺, and Lu³⁺ were prepared fully characterized with ¹H, ¹³C{¹H}, and ¹⁹F NMR spectroscopy, as well as elemental analysis (see Experimental section, pp. S5–S10). The full set of synthesized compounds feature secondary cations that span a wide range of Lewis acidity, with [M(H₂O)_m]ⁿ⁺ pK_a values varying from 16.0 for weakly acidic K⁺ to 7.9 for strongly acidic Lu³⁺.²⁵ Adducts of K⁺, Sr²⁺, and Lu³⁺ were studied both to span a wide range of cation properties (including both size and Lewis acidity) as well as to establish a homologous family of complexes suitable for comparisons to the prior series of [Ni,M] and [Pd,M] species^11,18 that have been reported. Detailed comparisons of all the complexes, including the new [Pt,M] series, will be discussed in a future report.⁵²

Fig. 2. Divergent synthetic strategy for preparation of the [Pt,M] complexes with anhydrous metal triflate salts.

Single crystal X-ray diffraction analysis

The structure of the monometallic [Pt] complex was previously reported,⁵¹ and this structure confirmed the binding of Pt(ii) in the [N₂,O₂] site of the Reinhoudt-type ligand (see Fig. 3). The Pt(ii) center templates the crown-ether-like site in [Pt], resulting in a rather planar arrangement of O1–O6 as quantified by the ω_crown parameter (see Table 1). The modest calculated ω_crown suggests prearrangement and readiness of the crown to bind secondary metal cations. Indeed, XRD analysis conducted on single crystals of [Pt,K], [Pt,Sr], and [Pt,Lu] confirmed binding of the secondary cations in the crown site as well as formation of the desired [Pt^II(μ₂-O_phenoxide)Mⁿ⁺] cores in each case (see Fig. 3). Metrics from the structures are presented in Table 1, and crystallographic details are given in the SI on pp. S25–S43.

Fig. 3. Solid-state structures (XRD) of (a) the monometallic [Pt] precursor and (b)–(d) the heterobimetallic [Pt,K], [Pt,Sr], and [Pt,Lu] complexes. For [Pt], all hydrogen atoms except those associated with the co-crystallized water molecule are omitted for clarity, and all atoms associated with minor components of disorder are omitted as well. For the heterobimetallic compounds, all hydrogen atoms, co-crystallized solvent molecules, and atoms associated with minor components of disorder are omitted for clarity. Displacement ellipsoids are shown at 20% probability level. The structure of [Pt] was first reported in ref. 51.

Table 1. Structural parameters for [Pt] and [Pt,M] from XRD analysis where M = K⁺, Sr²⁺, or Lu³⁺.

	[Pt]a	[Pt,K]	[Pt,Sr]b	[Pt,Lu]
C.N. of M	—	8	9	9
Ionic radius of M c (Å)	—	1.51	1.31	1.03
N1⋯N2 (Å)	2.60(4)	2.629(8)	2.667(16)	2.614(18)
O1⋯O2 (Å)	2.74(5)	2.683(6)	2.641(18)	2.531(12)
Pt⋯M (Å)	—	3.753(2)	3.658(1)	3.484(1)
Pt–Nimine (avg) (Å)	1.93(3)	1.952(6)	1.970(11)	1.940(12)
Pt–Ophenoxide(avg) (Å)	2.00(4)	1.985(4)	1.974(12)	1.976(8)
M–Ophenoxide (avg) (Å)	–	2.706(5)	2.558(11)	2.343(8)
τ 4  d	0.035	0.028	0.013	0.043
ω crown  e	0.204	0.210	0.052	0.465

^aFrom ref. 51.

^bValues are reported for the major component (74%) of this structure, which corresponds to the first of two cocrystallized/disordered species thar were found in the same volume of the asymmetric unit; comparisons of the major and minor components are given in Table S1.3.

^cFrom ref. 53 and 54

^dGeometry index for four-coordinate complexes; 1.00 indicates perfect tetrahedral geometry and 0.00 indicates rigorous square-planar geometry. Calculated as described in ref. 56.

^eDefined as the RMSD of the positions of atoms O1–O6 from the mean plane of their positions.

Based on the structural analysis, the secondary cations prefer to nestle into the binding site defined by the phenoxide oxygen atoms (O1 and O2). This tendency can be ascribed to the dual structural phenomena of rigidity in the Schiff-base site containing Pt(ii) and flexibility in the aliphatic polyether loop of the crown site. The Pt⋯M distance thus depends linearly on the Shannon ionic radius^53,54 of the secondary cations (see Fig. S1.40). Similarly, the M–O_phenoxide distance decreases monotonically across the series of complexes as well. Anionic character on O1 and O2 induced by deprotonation of the phenol moieties appears to drive the lateral movement of the cations (and their associated triflates) toward Pt(ii) in a manner dependent upon ionic radius. In line with the trend in Pt⋯M distance, the O1⋯O2 interatomic spacing (across the Pt(ii) site) decreases monotonically as a function of the ionic radius of Mⁿ⁺ (see Fig. S1.41). In line with the computational results (vide infra), the O1⋯O2 separations in the heterobimetallic species are all below the van der Waals limit for oxygen, a feature attributable to Pt–O_phenoxide covalency as well as distortion of O_phenoxide electron density by the Lewis acidic secondary cations. Nevertheless, the geometries about the Pt(ii) centers are mostly unperturbed by the secondary cations, with similar average Pt–N_imine, average Pt–O_phenoxide, and N1⋯N2 distances as well as similar τ₄ parameters across the series. These results underscore the apparent structural rigidity and high degree of covalency of the Pt(ii)-containing site, regardless of the identity of the secondary metal cation.⁵⁵

Spectroscopic and electrochemical results

The [Pt,M] species were found to display tunable properties that depend on the identity of the incorporated secondary metal cation. The tuning trends resemble those that we and others have observed in prior work, and therefore the results are described here only briefly. The tunability measured here sets the stage for establishing a frame of reference for interpreting the origins of the tunability through experimental/computational ¹⁹⁵Pt NMR analysis.

Optical spectroscopy

[Pt] features a moderately intense absorption feature with λ_max = 440 nm; this feature was assigned as a charge-transfer (CT) transition on the basis of its moderate molar absorptivity (ε) of 4350 M⁻¹ cm⁻¹ as well as based on prior assignments for related compounds (see Fig. 4, panel a, and Table 2). As observed previously, the most charge-dense cations induce the most significant spectral changes in the [Pt,M] complexes (see Fig. S1.18–S1.24); plotting the energy of the lowest-energy CT transition as a function of the [M(H₂O)_m]ⁿ⁺ pK_a values returns a slope of −45 ± 6 eV per pK_a (see Fig. S1.25). This slope is indistinguishable (within uncertainty) from those measured in related families of [Ni,M] (−34 ± 7 eV per pK_a) and [Pd,M] (−42 ± 6 eV per pK_a) species that have been prepared by Blakemore and co-workers,^11,18 suggesting similar effects are induced by the secondary cations in each case.⁵⁷

Fig. 4. Panel a: UV-visible absorption spectra of [Pt], [Pt,K], [Pt,Sr], and [Pt,Lu] recorded in CH₃CN solution. Panel b: Cyclic voltammograms of the platinum complexes; all scans were collected at a scan rate of 100 mV s⁻¹ in CH₃CN containing 0.1 M TBAPF₆; platinum concentration was 1 mM for [Pt] and 2 mM for the heterobimetallic complexes. Panel c: Plot showing the dependence of the optical CT transition energy as a function of the calculated HOMO–LUMO energy gap; data for all complexes plotted relative to values for [Pt]. Panel d: Plot showing the shift in reduction potential (E_p,c values, as discussed in the main text) as a function of the calculated LUMO energy; all data plotted relative to values for [Pt].

Table 2. Selected experimental parameters and calculated HOMO and LUMO energies of [Pt] and [Pt,M], where M = K⁺, Sr²⁺, or Lu³⁺.

	[Pt]	[Pt,K]	[Pt,Sr]	[Pt,Lu]
pKa of [M(H2O)m]n+ a	—	16.0	13.2	7.9
E p,c (V vs. Fc+/0)	–2.38b	–2.32	–2.14	–1.91
λ max (nm)	435b	425	406	381
λ max (eV)	2.81b	2.92	3.05	3.25
ε at λmax (M−1 cm−1)	4350b	4370	4470	4400
ε HOMO  c (eV)	–5.57	–5.84	–6.33	–7.13
ε LUMO  c (eV)	–1.74	–1.92	–2.19	–2.62
ΔεHOMO–LUMO c (eV)	3.83	3.92	4.14	4.51

^aFrom ref. 25.

^bFrom ref. 51.

^cCalculations were conducted at the same level of theory as used for the NMR analysis (vide infra).

Of relevance to the computational strategy used in this report, plotting the lowest-energy absorption energies of the [Pt,M] species (relative to λ_max of [Pt]) as a function of their calculated HOMO–LUMO gaps also returns a linear correlation (see Fig. 4, panel c), confirming satisfactory modeling of the complexes and the ability to use the calculations to guide understanding (vide infra). And, in accord with the latter linear relationships, plotting the HOMO–LUMO gap as a function of [M(H₂O)_m]ⁿ⁺ pK_a returns a linear correlation as well (see Fig. S2.3).

Electrochemistry

The monometallic [Pt] precursor⁵¹ and a series of analogous palladium complexes¹⁸ have been previously shown to undergo reduction processes, suggesting the opportunity here to study the redox chemistry of the [Pt,M] series of heterobimetallic compounds. Considering the similar resistance to reduction displayed by Pd(ii) and Pt(ii), we hypothesized that ligand-centered reduction processes would be measurable in the [Pt,M] series, and that incorporation of increasingly Lewis acidic secondary metal ions would shift these reductions anodically, perhaps in a similar manner to what was observed in the prior [Pd,M] series of complexes.¹⁸

The cyclic voltammetry (CV) for [Pt] displays a single reductive feature at E_p,c = –2.38 V vs. ferrocenium/ferrocene (denoted hereafter as Fc^+/0),⁵¹ a finding in line with that measured for the analogous Pd(ii) complex, [Pd], supported by the same Reinhoudt-type ligand scaffold (E_p,c = –2.40 V vs. Fc^+/0).¹⁸ Cyclic voltammetry of [Pt,K], [Pt,Sr], and [Pt,Lu] revealed reductive features at more positive potentials ranging from E_p,c = –2.32 V vs. Fc^+/0 for [Pt,K] to E_p,c = –1.91 V vs. Fc^+/0 for [Pt,Lu] (see Fig. 3, panel b). Notably the CV profiles for [Pt,Sr] and [Pt,Lu] display additional broad reductive features following the initial well-defined reduction wave, presumably due to follow-up chemical reactivity. Although these additional features were observed for the more Lewis acidic metal cations, and despite the clear chemical/electrochemical irreversibility of the voltammograms in all cases, the cathodic peak currents for the most positive features were found to vary linearly with scan rate, a behavior consistent with diffusion-controlled molecular behaviors of the oxidized species involved (see Fig. S1.26–S1.31).^58,59

The measured reduction waves represent a possible experimental read-out of LUMO energy, and in accord with this situation, plotting the shift of reduction potential (ΔE_p,c; calculated relative to the E_p,c of [Pt]) against the calculated LUMO energy (Fig. 4, panel d) reveals a strong linear correlation. When taken together with the satisfying correlation between the measured optical transitions and the calculated HOMO–LUMO gap (Fig. 4, panel c), this trend further supports the conclusion that the computational work (vide infra) faithfully models key features of the electronic structure of the [Pt,M] series of compounds.

Additionally, plotting the E_p,c for the [Pt,M] species as a function of the [M(H₂O)_m]ⁿ⁺ pK_a values revealed a slope of −50 ± 5 mV per pK_a (see Fig. S1.32). This result is consistent with prior observations of heterobimetallic Ni(ii) species studied by Karnes et al.¹¹ and Pd(ii) species studied by Golwankar et al.¹⁸ (−61 ± 12 and −46 ± 7 mV per pK_a, respectively). Although the slopes of these relationships are marginally distinguishable based on our uncertainty estimation, the closer agreement between the Pt and Pd cases underscores the assignment of ligand-centered reduction processes in both cases. We also note that the slopes of these trends appear close, in absolute value, to the Nernstian 59 mV per pH shift.^54,55 This alignment might be viewed as an encouraging use of the [M(H₂O)_m]ⁿ⁺ pK_a values as a descriptor of acidity in CH₃CN, although discussion of the nuances of this idea are beyond the scope of this study.

NMR measurements

NMR parameters have been shown to be very sensitive to changes in electronic structure, including the composition of bonding, anti-bonding, and non-bonding orbitals.^60–67 In the case of ¹⁹⁵Pt, joint experimental/computational work represents an opportunity to address which individual orbitals most strongly influence the ¹⁹⁵Pt magnetic shielding tensors, and by extension, their relationship to molecular-level structure and Pt–ligand bonding. For the [Pt,M] compounds studied here, ¹⁹⁵Pt NMR studies offer elementally specific insights into changes in Pt–ligand bonding induced by the secondary cations, and thereby a view of the origins of cation-driven tuning.^30,31,68,69

Prior to measuring the ¹⁹⁵Pt NMR spectra, ¹H NMR spectra were obtained in CD₃CN solution to use for indirect referencing of the ¹⁹⁵Pt NMR spectra (see the Experimental section on p. S4 for details). The ¹H NMR spectrum for [Pt] compares well to that found in the original report of this compound⁵¹ and features a single, well-defined set of resonances typical for a diamagnetic species; resonances span from 8.30 to 1.58 ppm, corresponding to the imine (H3b) and methyl (H1) protons, respectively (see Table 3 and Fig. S1.1). Incorporation of Lewis acidic secondary metal cations results in downfield shifts for all resonances. As in prior studies, these shifts display linear dependencies on [M(H₂O)_m]ⁿ⁺ pK_a values (see Fig. 5, panel a, and Fig. S1.11–S1.13). For example, the H3b imine proton resonances in the [Pt,M] species shift linearly with a slope of –0.012 ± 0.005 ppm per pK_a (Fig. 5, panel b). This slope is close to that found with our previously investigated [Pd,M] compounds (−0.017 ± 0.003 ppm per pK_a), but without additional data, no substantive insights can be obtained from these tabulated trends.¹⁸ Superficially, one might be tempted to interpret the spectra as indicating decreased shielding induced by the Lewis acidic metals, but no further details can be gleaned from the ¹H spectra on their own.

Table 3. ¹H (imine) and ¹⁹⁵Pt NMR chemical shifts of [Pt] and [Pt,M].

	pKa of [M(H2O)m]n+ a	δ 1Himine b	δ 195Pt d
[Pt]c	—	8.302	–1540
[Pt,K]	16.0	8.343	–1610
[Pt,Sr]	13.2	8.417	–1740
[Pt,Lu]	7.9	8.450	–1930

^aFrom ref. 25.

^bChemical shift values for the imine proton labeled H3b, see Fig. S1.1 for numbering scheme used for interpretation of NMR spectra. Uncertainty was calculated to be ±0.0003 ppm; this value corresponds to the arithmetic mean of the individual imine proton resonance linewidths for each of the [Pt,M] species; see Table S1.1 for details.

^cFrom ref. 51.

^dUncertainty on δ¹⁹⁵Pt was calculated to be ±10 ppm; this value corresponds to the arithmetic mean of the individual ¹⁹⁵Pt resonance linewidths for each of the [Pt,M] species; see Table S1.2 for details.

Fig. 5. Panel a: ¹H NMR spectra of the [Pt,M] compounds showing the resonances associated with protons in the two inequivalent imine moieties. Panel b: Plot of δ¹H across the [Pt,M] series for the proton resonance associated with imine carbons as a function of [M(H₂O)_m]ⁿ⁺ pK_a values; see Fig. S1.1 for numbering scheme; the proton associated with imine C1 is the more downfield resonance of the two imine resonances. Panel c: ¹⁹⁵Pt NMR spectra for the [Pt,M] species; ¹⁹⁵Pt NMR spectra were collected by direct detection. Panel d: Plot of δ¹⁹⁵Pt for the [Pt,M] compounds as a function of the corresponding [M(H₂O)_m]ⁿ⁺ pK_a values.

The ¹⁹⁵Pt spectrum of [Pt] revealed a single resonance with δ¹⁹⁵Pt of −1540 ppm, confirming the presence of a single Pt-containing species in solution as expected. The measured δ¹⁹⁵Pt value of −1540 ppm lies in the range typically observed for divalent platinum compounds.³⁰¹⁹⁵Pt NMR spectra were also collected on the full series of [Pt,M] complexes, and single resonances were observed in each case. δ¹⁹⁵Pt was found to shift to higher field (more strongly shielded, more negative δ¹⁹⁵Pt) upon incorporation of secondary metal cations, and plotting the observed δ¹⁹⁵Pt values as a function of [M(H₂O)_m]ⁿ⁺ pK_a revealed a linear correlation with a slope of 37 ± 2 ppm per pK_a (Fig. 5, panel d and Fig. S1.14–S1.17). In agreement with the δ¹H values, the more highly charged (more Lewis acidic) cations gave shifts of larger magnitude. However, incorporation of secondary cations resulted in upfield shifts of δ¹⁹⁵Pt, suggesting an increase in electronic shielding about platinum. As there have been no prior investigations of the relationship between δ¹⁹⁵Pt and Lewis acidity in heterometallic compounds, we pursued computational insights to guide interpretation of the experimental findings.

Relativistic density functional calculations

In agreement with the experimental observations, quantum chemical calculations^92–103 (see SI for details) with the PBE0 hybrid functional and the zeroth-order regular approximation (ZORA) relativistic Hamiltonian show that the increase of the charge of the secondary metal goes along with the ¹⁹⁵Pt NMR shift moving upfield, i.e., toward lower frequency (less positive/more negative chemical shift, δ) and toward increased isotropic absolute nuclear magnetic shielding, σ (see Fig. 6). It is worth noting that the calculated chemical shifts are in excellent agreement with the experimental data (see Fig. S2.4) given that the known Pt chemical shift range of diamagnetic complexes spans well over 10⁴ ppm.³³ As such, the calculations appear to be well suited to uncover the reasons for the upfield shift with increasing charge of the secondary metal.

Fig. 6. Conventions for δ (chemical shift) and σ (absolute shielding). The reference shielding is at 0 ppm on the chemical shift scale. The equation in the figure neglects a factor of 1/(σ^ref) ≅ 1 in the chemical definition.

The chemical shift and the shielding trends are opposite by definition. Therefore, compared to [Pt], the larger charge of the secondary metal is, the stronger is the shielding of Pt (the shift becomes increasingly negative). Since the electrons cause the shielding, it is tempting to assign the increasing Pt shielding along the series to variations in the electron density at the metal. For example, if the increasing secondary metal charge could somehow cause the ligands to donate more strongly to Pt, the increased electron density would then naively be associated with increased shielding. Such a simple explanation of the trend is unlikely to be correct, however, for reasons detailed shortly. Furthermore, an increasing ligand → Pt donation with increasing positive secondary metal charge defies chemical intuition.

The notion that the electrons cause the shielding is correct, of course. Broadly speaking, across the periodic table, it is also true that more electrons translate into stronger shielding and larger chemical shift ranges for heavy atoms such as Pt as compared to light atoms. However, when comparing the same nucleus in different chemical environments, more electron density associated with the probe atom does not necessarily mean stronger shielding, because the shielding trends are dominated by the magnetic coupling between occupied and unoccupied orbitals. (Also, the magnetic coupling contribution enters the paramagnetic term in the shielding which is typically negative.) Furthermore, for heavy elements such as Pt, relativstic effects such as the spin-orbit (SO) interaction are expected to be highly important. In other words, the fact that the NMR shielding is defined via the response of the system's electronic structure to the magnetic field means that shielding is not (or not only, to be precise) a function of the ground state electron density. See ref. 70 for examples from organic chemistry and carbon chemical shifts.

Table 4 summarizes the results of natural localized molecular orbital (NLMO) analyses of the absolute isotropic shielding for the Pt complexes. Table 5 provides the corresponding chemical shift data, using [Pt] as the NMR reference. The analysis provides contributions to the shielding for each occupied NLMO, with the magnetic-field response of the NLMO being absorbed into the result. The NLMOs are determined in scalar-relativistic (SR) calculations, i.e., neglecting the SO interaction, and they are quantum mechanical representations of bonding (BD), lone-pair (LP), and core (CR) electron pairs of the Lewis structure. The NLMOs incorporate delocalization if present in the system. Prior analyses of Pt NMR shielding and the associated chemical shifts^33,61,71 using the same techniques as employed herein showed that negative paramagnetic contributions to the Pt shielding from the occupied non-bonding 5d orbitals at platinum play a dominant role in the chemical shift, followed by contributions from Pt–ligand bonding NLMOs, Pt outer core contributions, shielding generated by orbitals centered on more distant atoms, and SO contributions from formally unoccupied NLMOs. The latter arise because the electronic structure including the SO interaction differs from the SR electronic structure. Consequently, contributions from formally unoccupied NLMOs enter the NMR shielding calculated at the SO level. (Note that an analysis of a SO NMR calculation in terms of SR orbitals does not constitute an approximation as long as all occupied and unoccupied NLMOs are used.)

Table 4. Shielding analysis in terms of NLMOs for select Pt complexes^a,b,c.

	[Pt]	[Pt,K]	[Pt,Sr]	[Pt,Lu]
Pt core	12 974	12 937	12 867	12 727
LP(Pt) dz2	−315	−318	−301	−342
LP(Pt)a + LP(Pt)b	−5513	−5429	−5364	−5170
LP(Pt) dxz	−1416	−1388	−1386	−1337
∑σ(Pt–N)d	−850	−858	−862	−921
∑σ(Pt–O)d	−797	−793	−741	−581
∑LP(O)	−91	−64	−45	−27
∑(others)e	−133	−144	−116	−88
SO unoccf	−427	−424	−407	−381
Total	3433	3520	3645	3879

^aIsolated molecular structures extracted from the refined crystal structures subjected to geometry optimization with a solvent model for acetonitrile, without the counter ions. ZORA(SO)/PBE0 calculations with basis set BS4 (see Table S2.1).

^bIn the NLMO analysis, the SO and paramagnetic contributions are combined into a single term, whereas purely paramagnetic (non-SO) effects can be obtained separately by running the analysis with only SR effects taken into account.

^cNLMOs listed as LP (lone-pair) are formally non-bonding. The orbitals labeled as LP(Pt)_a and LP(Pt)_b are linear combinations of platinum d_x²−y² and d_yz lone-pair NLMOs as explained in the main text.

^dSum from equivalent orbitals. Includes contributions from NLMOs with bonding and LP parent NBOs that exhibit ligand → Pt donation.

^eSum of shielding contributions from core, lone pairs and bonding NLMOs not listed in other table rows.

^fShielding contributions from SO induced mixing of unoccupied SR MOs into the SO-DFT ground state.

Table 5. Chemical shift analysis in terms of NLMOs, with [Pt] as the NMR reference^a,b,c.

	[Pt]	[Pt,K]	[Pt,Sr]	[Pt,Lu]
Pt core	0	37	107	247
LP(Pt) dz2	0	3	–14	27
LP(Pt)a + LP(Pt)b	0	–83	–149	–343
LP(Pt) dxz	0	–28	–30	–79
∑σ(Pt–N)d	0	8	13	72
∑σ(Pt–O)d	0	–4	–56	–63
∑LP(O)	0	–27	–46	–63
∑(others)e	0	10	–18	–45
SO unoccf	0	–4	–20	–47
Isotropic shift	0	–87	–212	–446
Experimental shift	0	–74	–196	–391

^aObtained from the data in Table 4 by subtracting the results for [Pt,K], [Pt,Sr], and [Pt,Lu], respectively, from those obtained for [Pt].

^bIn the NLMO analysis, the SO and paramagnetic contributions are combined into a single term, whereas purely paramagnetic (non-SO) effects can be obtained separately by running the analysis with only SR effects taken into account.

^cNLMOs listed as LP (lone-pair) are formally non-bonding. The orbitals labeled as LP(Pt)_a and LP(Pt)_b are linear combinations of platinum d_x²−y² and d_yz lone-pair NLMOs as explained in the main text.

^dSum from equivalent orbitals. Includes contributions from NLMOs with bonding and LP parent NBOs that exhibit ligand → Pt donation.

^eSum of shielding contributions from core, lone pairs and bonding NLMOs not listed in other table rows.

^fShielding contributions from SO induced mixing of unoccupied SR MOs into the SO-DFT ground state.

As expected, the largest contribution to the absolute shielding is from the Pt core (shells 1s through 5p, i.e., [Xe]). The Pt core also contributes to the chemical shift trend via polarization and weak core-valence hybridization of the outer core orbitals, presumably. The core contributions to the shift are not straightforward to rationalize because they are amplified by the inverse dependence of the NMR operators on the electron-nucleus distance.^60,61,72 As such, we note simply that the core counterbalances the effects from the Pt valence shells.

For the ensuing discussion, the equatorial Pt-coordinating ligand atoms are assumed to be in the xy plane, with the coordinating oxygen and nitrogen centers located in between the x and y axes. See Fig. 7–9, S2.1 and S2.2. This choice of coordinates means the 5d_xy orbital at Pt is the one whose lobes point at the coordinating equatorial ligands. In the MO-theoretical description of the complexes, there are occupied-bonding/unoccupied-antibonding MO pairs with large coefficients from 5d_xy. The antibonding combinations belong to the NLMOs labeled in the figures as σ*. It is important to note that the usual d-orbital labels d_xy, d_xz, d_yz, d_z², and d_x²−y² do not always apply to the d-orbitals obtained from calculations for systems that have mixed ligands and distortions from idealized coordination environments, as is the case here. For [Pt,Lu] (see Fig. 9), the expected set of non-bonding occupied 5d orbitals are generated among the NLMOs. In the chosen coordinate system, we can therefore assign them as a slightly distorted and ligand-mixed d_z², relatively pure d_x²−y² and d_yz, and a d_xz that shows some back-bonding into the imine π system of the ligand. For the other complexes, the d_x²−y² and d_yz appear as linear combinations instead, labeled LP(Pt)_a and LP(Pt)_b in the tables and figures. Fig. 8 shows for [Pt] that simple ± linear combinations of LP(Pt)_a and LP(Pt)_b restore the expected d_x²−y² and d_yz. Therefore, for comparison of the NMR analysis data among the systems, the sum of the LP(Pt)_a and LP(Pt)_b contributions to the shielding is used, which is equivalent to the sum of the d_x²−y² and d_yz contributions. There is likewise a difference in the two equatorial σ* NLMOs as generated in the calculations. For [Pt], the antibonding combinations accompanying the σ-bonding combinations of d_xy and 6s on Pt with ligand-centered combinations, respectively, are clearly visible in Fig. 7. For the other systems (Fig. 9, S2.1, and S2.2) we instead obtained two linear combinations that mix Pt 6s and 5d_xy hybrids with atomic orbitals of ligand atoms in trans position. Compare, for example, orbitals σ*(1) and σ*(2) for [Pt] (Fig. 7) with those of [Pt,Lu] (Fig. 9).

Fig. 7. Non-bonding 5d LPs and σ bonding/antibonding NLMOs (±0.03 isosurfaces) of [Pt]. The coordinate system chosen to label some of the orbitals is such that the z axis is perpendicular to the equatorial Pt coordination plane, pointing at the viewer. The orbitals labeled as LP(Pt)_a and LP(Pt)_b are linear combinations of platinum d_x²−y² and d_yz lone-pair (LP) NLMOs (as shown in Fig. 8).

Fig. 8. The orbitals labeled as LP(Pt)_a and LP(Pt)_b are linear combinations of the platinum d_x²−y² and d_yz (LP) NLMOs. The green/yellow color-coded orbitals in the bottom row were obtained as normalized ± linear combinations of the orbitals in the top row. The coordinate system chosen to label some of the orbitals is such that the z axis is perpendicular to the equatorial Pt coordination plane, pointing at the viewer.

Fig. 9. Non-bonding 5d and σ bonding/antibonding NLMOs (±0.03 isosurfaces) of [Pt,Lu]. The coordinate system chosen to label some of the orbitals is such that the z axis is perpendicular to the equatorial Pt coordination plane, pointing at the viewer. The orbitals labeled as LP(Pt)_a and LP(Pt)_b correspond to the platinum d_x²−y² and d_yz lone-pair (LP) NLMOs.

The primary differences in isotropic shielding among the complexes, other than the aforementioned outer core contributions, arise from the deshielding caused by the Pt LP 5d atomic orbitals (AOs), as well as from valence ligand orbitals that donate electron density to Pt. In a planar coordination environment, assuming the z direction is normal to the plane, the occupied lone pair 5d_z², contributes only weakly to the shielding generated by a magnetic field in the z direction. This is because a pure AO of this type is an eigenfunction of the z-component of the angular momentum operator, which is up to a constant the same as the electron spin independent part of the non-relativistic electron magnetic moment operator. The occupied 5d_x²−y² orbital LP, however, is strongly perturbed by a magnetic field along the z direction. As explained elsewhere,^33,34,73 a magnetic field perpendicular to the plane effectively rotates the d_x²−y² orbital by 45°, resulting in the overlap of all four lobes with the aforementioned σ* orbitals. This is another way of saying that the occupied d_x²−y² is able to couple magnetically with σ* orbitals resulting from the equatorial ligand dative bonding, which produces paramagnetic ‘ring currents’ in the 5d shell and contributes to the (negative) paramagnetic component of the shielding. Similarly, the d_xz and d_yz orbitals also undergo magnetic field-induced mixing with σ* MOs, though to a lesser degree than d_x²−y². This is mostly caused by the absence of σ* MOs in the z direction in planar Pt(ii) coordination environments, which diminishes the induced magnetic coupling, since only two of the lobes of the ‘rotated’ AOs overlap with antibonding σ* MOs (instead of all four lobes in the case of d_x²−y²). As mentioned, the d_xz orbital is delocalized into the imine moieties of the ligands, constituting π back bonding. This back bonding results in less pure Pt 5d character of d_xz, which reduces the contribution toward the Pt shielding compared to a purely non-bonding orbital. Table 5 summarizes the chemical shift analysis in terms of NLMOs, with respect to the internal reference [Pt], and it shows that the major contribution towards the upfield ¹⁹⁵Pt NMR trend is driven by LP(Pt)_a and LP(Pt)_b, along with σ(Pt–O)^(LP), a σ-type Pt–O bonding NLMO derived from an O-centered LP. The following discussion is facilitated by the Pt character of these orbitals along with the occupation of their ‘parent’ NBOs as collected in Table 6. The relevant non-bonding Pt orbitals have increasingly large delocalization the more these metrics deviate from the idealized Pt character (which would be very close to 100% for truly non-bonding LPs) and parent NBO occupations of 2. For ligand-centered lone-pairs donating to Pt, the delocalization by which electron density is donated to Pt is stronger when the Pt weight is larger and the parent NBO occupation deviates more from 2.

Table 6. %Pt character of selected NLMOs and their parent natural bond orbitals (NBO) occupations^a,b.

	%Pt				NBO
	[Pt]	[Pt,K]	[Pt,Sr]	[Pt,Lu]	[Pt]	[Pt,K]	[Pt,Sr]	[Pt,Lu]
LP(Pt) dz2	99.2	99.0	99.1	98.8	1.98	1.98	1.98	1.98
LP(Pt)a	98.4	98.4	98.4	98.3	1.97	1.97	1.97	1.97
Lp(Pt)b	98.0	97.9	97.8	97.7	1.96	1.96	1.96	1.95
LP(Pt) dxz	93.0	93.2	93.6	94.3	1.87	1.87	1.88	1.89
σ(Pt–O)(LP)	10.2	10.2	9.7	8.6	1.69	1.70	1.72	1.73
σ*(1)	52.9	67.0	67.3	67.6	0.93	0.32	0.30	0.29
σ*(2)	80.5	66.7	67.1	67.2	0.38	0.32	0.30	0.29

^a σ(Pt–O)^(LP) corresponds to a donating oxygen lone pair, the values presented are averaged over both the oxygen donors.

^bThe orbitals labeled as LP(Pt)_a and LP(Pt)_b are linear combinations of platinum d_x²−y² and d_yz lone-pair (LP) NLMOs.

It is interesting to note that the 5d occupied nonbonding LPs (Pt)_a and (Pt)_b, the major contributors to the upfield NMR trend, hardly undergo any change in %Pt character as well as the parent NBO occupancy/population upon increasing the charge on the secondary metal. However, even minor trends in these orbitals can potentially cause substantial changes in the shielding. When only a subset of Pt 5d orbitals are occupied, and when the 5d orbitals are relatively pure AOs, this causes a strong paramagnetic deshielding from the 5d shell for two main reasons:³⁴ (i) When these orbitals have little delocalization beyond the Pt atom, they contribute the most to the Pt shielding because of the inverse dependence of the shielding contributions on the electron-nucleus distance. (ii) The magnetic coupling between occupied and empty d-orbitals generates the deshielding contributions in the first place. The complexes studied herein have formal Pt 5d⁸ configurations, with the formally empty orbital being 5d_xy. As seen in the data, the strong paramagnetic negative shielding contributions are evident in the analysis. An important counteracting influence is the equatorial ligand donation into the 5d_xy orbital, which reduces the availability of this AO for paramagnetic coupling with the other 5d orbitals, all else being equal. The extreme limit of this scenario would be the full occupation of 5d_xy by donation, which would presumably eliminate the 5d intra-shell paramagnetic contributions to the Pt shielding in the resulting hypothetical 5d¹⁰ configuration.

The interpretation of the NLMO data needs to take these influences into account. In Table 6, a clear and monotonic trend that might provide a clue about the present systems is the Pt contribution and parent NBO occupation of the orbitals listed as σ(Pt–O)^(LP), which are oxygen LPs that display dative bonding (donation) to Pt. These orbitals become less donating the stronger the charge of the secondary metal is, as evidenced by the rather pronounced decrease in the Pt weights along the series of complexes (along with increasing parent NBO occupations). This finding is chemically intuitive, because the presence of the secondary cation is expected to render the Pt-coordinating oxygen atoms less donating because of inductive effects. The combined Pt weights of the antibonding σ* orbitals display a corresponding slight increase along with an accompanying decrease in the combined parent NBO occupations. This is to say that the weaker oxygen donation with stronger secondary metal charge renders the 5d Pt orbitals more pure. As a result, one would expect the combined paramagnetic deshielding from LP(Pt)_a and LP(Pt)_b to become stronger along the series. However, the trend obtained in the analysis is weaker deshielding (increased overall shielding), in line with the overall trend in the chemical shift.

Another well-known influence that modulates the paramagnetic shielding is via the orbital energies. All else being equal, the paramagnetic deshielding is expected to become stronger/weaker with a smaller/larger energy gap between the occupied and unoccupied orbitals that couple magnetically. The analysis in terms of orbital energy differences is only rigorous in terms of the so-called canonical MOs (CMOs), which are the delocalized Kohn–Sham orbitals generated by the DFT calculations, as explained elsewhere.⁷³ Nonetheless, it is to be expected that larger energy gaps between occupied and unoccupied NLMOs also translate into weakened paramagnetic shielding contributions. In Table S2.2, the average energy of the pair of σ* NLMOs changes from 4.2 to 3.6 eV along the series, while the average energies of the LP(Pt)_a and LP(pt)_b NLMOs decreases from −8.5 eV for [Pt] to −9.7 eV in [Pt,Lu]. The increasing NLMO energy gap with increasing charge of the secondary metal is therefore likely one of the reasons for the overall diminishing paramagnetic contributions from LP(Pt)_a and LP(Pt)_b, resulting in the increased shielding observed experimentally. It is also worth noting that the highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) stabilize in energy along the series (because of the increasing positive charge), while the HOMO–LUMO gap increases from 3.83 eV for [Pt] to 4.51 eV for [Pt,Lu] (Table S2.3).

Per Table 7, the assigned reduced O → Pt donation along the series goes along with an increasing positive partial charge of Pt and an increasing negative partial charge of the coordinating oxygen atoms. Wiberg and Mayer bond orders obtained from the NBO calculations likewise indicate reduced dative (covalent) bonding along the series. Oxygen also becomes less donating towards the σ*(Pt–N) orbitals trans to the oxygen donors. The corresponding NBO delocalization energies (E⁽²⁾, in kcal mol⁻¹) confirm this trend, going as [Pt] (−121.5) vs.[Pt,K] (−101.3) vs.[Pt,Sr] (−97.2) vs.[Pt,Lu] (−86.0). In turn, this renders the covalent (Pt–N) bonding stronger, via reduction of the trans influence from the O–Pt bonds.

Table 7. Charge analysis (NPA) and the bond orders (WBI/Mayer) of the complexes in this study^a.

	NPA				WBI	Mayer	WBI	Mayer
	[Pt]	[M]	[O]avg	[N]avg	σ(Pt–O)(LP)	σ(Pt–O)(LP)	σ(Pt–O)(BD)	σ(Pt–O)(BD)
[Pt]	0.887	0.000	–0.703	–0.435	0.364	0.523	0.486	0.651
[Pt,K]	0.896	0.959	–0.739	–0.528	0.353	0.499	0.492	0.652
[Pt,Sr]	0.901	1.917	–0.812	–414	0.331	0.446	0.505	0.652
[Pt,Lu]	0.909	2.376	–0.824	–0.396	0.295	0.366	0.525	0.651

^aσ(Pt–O)^(LP) corresponds to the LP(O), the values presented are the averaged values obtained from both the oxygen donors. Similarly, the average values from both N atoms and σ(Pt–O)^(BD) are presented. Data obtained from SR DFT calculations.

The corresponding trend is most pronounced in the Wiberg bond orders, whereas the Mayer bond orders remain similar across the complexes. The fact that the O → Pt donation influences the covalent Pt–N bonding also has implications regarding the Pt shielding, of course. Namely, the O–Pt trends noted earlier are partially counter-balanced by the Pt–N trends, which then leaves the NLMO energy gap argument as the one that would seem to be the most relevant for the Pt NMR shift trend. We note in passing that the NBO E⁽²⁾ associated with the back-donation from d_xz into the π* orbitals of the imine moiety of the ligand, follows a weak trend, going as [Pt] (−12.9) vs.[Pt,K] (−12.8) vs.[Pt,Sr] (−12.6) vs.[Pt,Lu] (−12.2, all values in kcal mol⁻¹). This weakening of back-donation reflects the increasing %Pt character of the d_xz LP as the charge on the secondary metal increases. Counter-intuitively, the paramagnetic deshielding contribution from this orbital decreases in magnitude along the series from [Pt] to [Pt,Lu], which means that the orbital energy gap argument likely prevails here as well.

To summarize the computational results, the NLMO analysis of the Pt shielding in the set of complexes [Pt], [Pt,K], [Pt,Sr], and [Pt,Lu] indicates weakened O → Pt donation with increasing secondary metal charge and increased energy gaps between relevant occupied and unoccupied orbitals that define the shielding trends. The overall increasing shielding of Pt with increasing charge of the secondary metal ion is a result of weakened negative paramagnetic contributions, which are most likely a result of increased energy gaps between occupied and unoccupied orbitals that couple magnetically. The analysis highlights the fact that the NMR shielding is not simply a property of the ground state, and that more electron density at a given atom does not necessarily cause higher shielding, even in cases for which such a correlation is present.⁷⁰ Our findings underscore the importance of differentiating between the unique influences of diamagnetic contributions (in the ground state) and paramagnetic contributions (from excited states), which is typically a challenging task. These separate contributions were laid out already in Ramsey's nonrelativistic formalism for calculating the magnetic shielding of nuclei in molecules,⁷⁴ and our treatment here is framed nicely by the broad usefulness of this classic approach.⁷⁵ The state-of-the-art in the computational approach has obviously made great strides since Ramsey's work, such that a detailed analysis including relativistic effects can be carried out.

Conclusions

In this work, a series of macrocyclic platinum-containing heterobimetallic complexes were prepared, isolated, and studied to obtain structural, spectroscopic, and electrochemical insights into the origins of the tuning of the complexes by incorporation of mono-, di-, and tri-valent secondary metal cations. This study demonstrates the ability to systematically modulate the electronic structure and reductive electrochemical properties of third-row transition metal complexes through the strategic inclusion of Lewis acidic cations. Here, we employed ¹⁹⁵Pt NMR spectroscopy as a previously unexplored spectroscopic probe of the effects of Lewis acidic secondary metal cations on the platinum-containing complexes. Although the change in δ¹⁹⁵Pt defies naïve chemical intuition regarding nuclear magnetic shielding, a per-orbital analysis of the calculated ¹⁹⁵Pt NMR chemical shifts reveals that the measured trend can be faithfully reproduced and explained. The increasing charge of the secondary cation reduces the electron donation to platinum from bridging phenoxide moieties, highlighting the synthetic opportunity of influencing multimetallic compounds through the identity of the key bridging ligands.²⁹ This goes along with an increase in the Pt shielding primarily because of an increase in the energy gaps between the non-bonding Pt 5d orbitals and antibonding 5d-ligand combinations, which reduces the magnitude of negative (paramagnetic) shielding contributions from the non-bonding 5d orbitals.

The results given here are reminiscent of those obtained in analogous cases that have been studied in the past few years, in the sense that we have found that incorporation of cations resulted in systematic changes in the ¹⁹⁵Pt NMR spectra. The work described here demonstrates both the sensitivity of this spectroscopic technique to Lewis acid effects as well as the satisfying nature of the model, further refined here with detailed, orbital-specific insights, that is now available to explain these effects. Taken together, the findings described here thus offer a platform for understanding the electronic consequences of Lewis acid incorporation on heterobimetallic systems and suggest opportunities for future development in this area.

Author contributions

R. R. G., A. C. E., and S. R. D. contributed equally to this work. The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting the findings of this work are available in the main text or supplementary information (SI). Supplementary information: NMR spectra, electronic absorption spectra, electrochemical data, X-ray crystallographic information, and computational details (PDF); experimental XRD structures (CIF); Cartesian coordinates for experimental XRD structures with minor components of disorder and any co-crystallized outer-sphere solvent removed (XYZ). See DOI: https://doi.org/10.1039/d6dt01469k.

Ref. 79–103 are cited in the SI.

CCDC 2492675, 2492676 and 2492677 contain the supplementary crystallographic data for this paper.^76–78