Introduction

The X-ray crystal structures of Mn2(CO)8 and Re2(CO)8 complexes by Dahl et al.1 confirmed the genuine existence of a metal-metal (M–M) single bond between two transition metal cations. This discovery spurred exploration of a variety of unusual M–M bonds, including Be–Be2, Mg–Mg3, and Zn–Zn4, as well as various higher-order M–M bonds5,6, such as double7, triple8, quadruple9, and quintuple10 (Fig. 1). These studies have deepened understanding of M–M bonding nature6,11,12 and revealed potential applications13,14,15,16. Many of these short M–M bonds are observed in bimetallic complexes, where the metal atoms are supported by bridging ligands. These ligands impart distinctive features, mainly (i) the close proximity of the two metal atoms and (ii) the preferred (bonding) angle between their lone pair orbitals. Several electronic factors, intrinsic to the metal, also influence the M–M bond lengths characteristic. Key features among these are the oxidation state of the metal atoms and the overall charge on the complex. In bimetallic complexes, the formal oxidation states of each metal typically range from 0 to +4, leading to net charges on the bimetallic core between 0 and +8. Lower oxidation states facilitate higher M–M bond orders by disfavoring binding to negatively charged ligands, thus allowing electrons on each metal atom to be available to form M–M bonds of higher order. Conversely, higher metal oxidation states tend to increase the distance between metal atoms due to electrostatic repulsion, resulting in longer M–M bond lengths17. Early-stage transition metals exhibit pronounced variations in ionic radius as oxidation state changes compared to late-stage transition metals. This enables early-stage metals to form shorter M–M bond lengths with high bond orders. For instance, chromium is well-known for forming exceptionally short Cr–Cr bonds, including the shortest CrII–CrII (quadruple) bond measuring 1.7293(12) Å18. In contrast to short M–M bonds formed by the metal ions with open-shell configuration, the metal ions with a closed-shell d10 electronic configuration, such as Cu+, Ag+, and Au+, do not engage in covalent M–M bonding6. Instead, they form metallophilic interactions, typically evidenced by short M–M bond lengths resulting from mixing a low-lying vacant ns orbital with the (n–1)dz2 orbital. In their M–M’ heterobimetallic complexes, a distinct energy mismatch exists among (n–1)dz2, ns, and npz orbitals of M vs. M’, significantly increasing or decreasing the potential for mixing of (n–1)dz2 orbitals of M with ns orbitals of M’ or (n–1)dz2 orbitals of M with npz orbitals of M’6. The magnitude of this heterobimetallic mixing is determined by the organic components attached to M and M’ cations. This mixing imparts bonding character to the dσ* antibonding molecular orbital, leading to heterobimetallophilic interactions such as argento-aurophilic (Fig. 1f)19 and cupro-aurophilic (Fig. 1g)20.

Fig. 1: Literature examples.
figure 1

Selected examples of a Mn–Mn single bond by Dahl et al.1, c Ni–Ni double bond by Koenig et al.7, d V–V triple bond by Cotton et al.8, b Re–Re quadruple bond by Cotton et al.9, e Cr–Cr quintuple bond by Nguyen et al.10, f Au–Ag single bond by Olmos et al.19, g Cu–Ag single bond by Desnoyer et al.20, and h I–Ag single bond of the present study.

Full size image

Similarly to low oxidation state Ag+ and Au+ ions, removing an electron from a halogen atom results in halogen(I) cation (X+). These “naked” X+ halonium ions have a px2py2pz0 orbital arrangement, featuring a vacant pz0 lobe (the p-hole) and px2py2 electron-rich belt as illustrated in Fig. 221. The X+ itself is unstable and cannot be isolated as a single-atom cation, however, it can be stabilized by two Lewis bases (L), forming cationic [L–X–L]+ halonium complexes, featuring a 3-center 4-electron (3c4e) [L–I–L]+ bond and are defined as halogen-bonded22 halogen(I) complexes. The most prominent example is Barluenga’s reagent [bis(pyridine)iodine(I)]tetrafluoroborate, [C5H5N-I-NC5H5][BF4], (Fig. 2a)21. Within these complexes, halogen(I) cations remain highly reactive compared to, e.g., carbon-bound halogen atoms. Their role and applicability in organic synthesis have been well-documented23. Despite their reactivity, the [N–I–N]+ halogen bonds demonstrate bond strengths comparable to coordination bonds21. This has sparked significant interest in studying various [N–X–N]+ complexes, including those formed from different N-heterocycles, as well as larger supramolecular structures such as capsules24,25, macrocycles26, and halogen-bonded porous frameworks27,28. No matter the size of the N-heterocyclic ligands, these complexes are synthesized by adding elemental iodine to a solution containing the parent [N−Ag−N]+ complex to form the [N−I−N]+ complex along with AgI precipitate21. This process substitutes a two-coordinate silver(I) ion with iodine(I), resulting in the formation of the two-coordinate [N−I−N]+ complex and silver(I) iodide precipitation, typically separated through filtration. The presence of a nucleophilic electron belt encompassing the iodine(I) in Lewis base-stabilized [L–I–L]+ complexes definitively suggests its potential to act as an electron donor29,30,31, and this character could enable the formation of an I+–Ag+ non-metal-to-metal bond, with silver(I) acting as an electron acceptor. This interaction would mirror the well-established Pt2+–Ag+ coordination bonds observed in heterobimetallic complexes, where Pt2+ serves as the donor and Ag+ as the acceptor32,33.

Fig. 2: Comparison of 3-center-4-electron halogen-bonds.
figure 2

The representation of 3c4e XB bonding modes of a monodentate pyridine with iodine (I) and b polydentate pyridine N-oxide with iodine(I) cation. Abbreviation: e = electron.

Full size image

Results

In this work, we report the synthesis and X-ray crystal structure of a [Ag(m-O3SCF3)2{(4MePyNO)2I}]2 (1-AgI) complex, in which the iodine(I) displays a dual functionality by forming a [O−I−O]+ halogen-bonded complex by 4MePyNO34 and an I+–Ag+ coordination bond. Complex 1-AgI is derived from the preceding polymer silver(I) complex [Ag2(4MePyNO)2(OTf)2]∞ (1-Ag) (Fig. 3, see Figs. S2 and S10). Adding 1.0 eq of elemental iodine to 1-Ag converts the polymer into the discrete 1-AgI complex. The greatest challenge of this reproducible synthesis is associated with the stability of the in-situ formed 1-AgI crystals (for more details, see the experimental section). Note that reaction mixtures from which AgI precipitates were filtered off under an argon atmosphere did not yield crystals, possibly due to complex decomposition during the filtration process. Therefore, we opted to leave the reaction mixtures at −24 °C without filtering the AgI precipitates, which proved effective in obtaining crystals suitable for X-ray diffraction analysis. The crystals of 1-AgI exhibit significant instability, even when isolated under a nitrogen atmosphere at −40 °C for the single crystal X-ray diffraction analysis. This instability consequently restricted our current investigation to X-ray crystallography and Density Functional Theory (DFT) studies. It is worth emphasizing that the [N−Ag−N]+→[N−I−N]+ cation exchange reaction of pyridines occurs, transforming a discrete silver(I) complex to a discrete iodine(I) complex. This transformation is fundamentally driven by the precipitation of AgI, which facilitates the formation of linear two-coordinate silver(I) complexes. The parent silver(I) complex 1-Ag crystallizes from dichloromethane into a 1D polymeric structure, where the N-oxide oxygen coordinates to three silver(I) ions. The polydentate coordination nature of the N-oxide oxygen, in conjunction with bridging sulfonate anionic oxygens, accounts for the formation of the polymeric 1-Ag. The silver(I) complexes of N-oxides are predominantly polymeric, even when paired with weakly coordinating anions such as PF6 and SbF635,36. The 4MePyNO and AgOTf were mixed and subjected to crystallizations using six additional solvents, including a combination of dichloromethane and other solvents, to assess the robustness of 1-Ag and to investigate the potential formation of a discrete silver(I) complex (Table S1). These six experiments resulted in seven polymeric structures (Figs. S2–S12). Of these seven structures, five were identical to the one obtained from “pure” dichloromethane, while the remaining two were different from each other and different from the complex from dichloromethane. These silver(I) complexes do not incorporate solvent molecules. In summary, the fact that six out of seven crystallization experiments utilizing AgOTf yielded the formation of a polymeric 1-Ag complex suggests that 1-Ag forms a stable polymeric silver(I) network, which, upon the addition of elemental iodine, undergoes (partial) cation exchange to produce 1-AgI.

Fig. 3: Synthetic route to access I+−Ag+ coordination bond.
figure 3

Synthesis of complexes, 1-Ag and 1-AgI.

Full size image

Complex 1-AgI crystallized in the triclinic space group P-1, with the asymmetric unit consisting of two crystallographically independent centrosymmetric complexes and two dichloromethane solvent molecules (Fig. 4 and S13). The [O−I−O]+ XB bond lengths range from 2.175(14) to 2.205(15) Å and are ~0.05 Å shorter than in [N–I–N]⁺ halogen-bonded system formed by the parent 4MePy31, and ~0.01 Å longer than in the structurally similar [O–I–O]− motif formed by carboxylate37,38,39 and alkoxylate40 anionic oxygen atoms. Furthermore, these X-ray crystal structure I−O bond lengths are comparable to those reported for DFT optimized [O−I−O]+ complexes of 4MePyNO34. The O−I−O angle [178.5(5)°, and 178.6(6)°] in 1-AgI are analoguous with the linear 3c4e [N–I–N]+ halogen-bonded complexes. The pyridine N-oxide O···O distances in [O–I–O]⁺ [4.36(2) – 4.40(2) Å] are comparable to the pyridine N···N distances in [N–I–N]⁺ [4.372(2) – 4.592(4) Å]41 and carboxylate/alkoxylate O···O distances in [O–I–O]− [4.328(11) – 4.407(8) Å]37,38,39,40 halogen-bonded systems. The key structural feature of 1-AgI is the very short I+–Ag+ bond between the electron belt of the iodine(I) cation and the silver(I) cation (Fig. 4). The silver(I) cation in 1-AgI has a square pyramidal geometry, with four O-atoms from the triflate anion occupying the basal positions and the iodine(I) occupying the fifth apical position. The Ag···Ag argentophilic interactions (3.419(3) and 3.489(3) Å) between the two silver(I) ions are close to the sum of the van der Waals radii of silver(I) ions [3.44 Å]. However, the I+–Ag+ bonds are significantly shorter, measuring 2.8625(16) and 2.8673(17) Å, and are only slightly (0.12–0.14 Å) longer than those reported in Ag+-diiodine complexes42,43,44. The I+–Ag+ bond lengths in 1-AgI are comparable to known M–M’ bonds, e.g., Au+–Ag+ (2.8553(6) Å)45, Cu+–Ag+ (2.8616(9) Å)46, Pt2+–Ag+ (2.8602(4) Å)33, and Ni2+–Ag+ (2.9348(3) Å)47 in heterobimetallic complexes suggesting that I+–Ag+ has a coordination bond character. Importantly, previous I-Ag complexes29,30,31 reported I+···Ag+ bond lengths that correspond to the sum of the van der Waals radii of the involved cations (3.54 Å), significantly longer than the 2.86 Å reported herein. This suggests that these previously observed I+···Ag+ contacts that are shorter than the sum of the involved atom’s vdW radii occur in the solid-state between charged complexes, thus differing from the 1-AgI system with the observed I+–Ag+ coordination bond. The experimental I–Ag bond Raman frequencies have been reported in the 95–110 cm–1 range44,48,49. We carried out Raman spectroscopy on single crystals of the 1-AgI complex at –40 °C to characterize the I–Ag bond (Figs. S15–S19). However, no Raman signal for the I–Ag bond was observed. This can be attributed to several factors, including strong anharmonicity in the I–Ag bond leading to significant broadening of Raman peaks, as well as the bond’s highly ionic nature and involvement of heavy atoms, which result in weak polarizability variations and low Raman intensity which can not anymore be detected due to technical limitations in this area of the spectrum.

Fig. 4: X-ray crystallography.
figure 4

The X-ray crystal structure of 1-AgI with the thermal displacement parameter at 50% probability level. The second symmetry-related complex and dichloromethane solvent molecules are omitted for clarity. The italicized bond parameters correspond to the second symmetry-related complex. Color key: Purple = iodine, ash white = silver, yellow = sulfur, lime green = fluorine, red = oxygen, blue = nitrogen, gray = carbon, white = hydrogen.

Full size image

To visualize the positive and negative regions of the [O−I−O]+ halogen-bonded motif, the molecular electrostatic potential surface (MEPS) of [4MePyNO−I − ONPyMe4]+ halogen-bonded complex was calculated at the B3LYP/aug-cc-pVTZ-PP level of theory using the crystal structure coordinates of 1-AgI. The MEPS of [4MePyNO−I−ONPyMe4]+ (Fig. 5a) was compared with the MEPS calculated from the crystal structure coordinates of the [N–I–N]⁺ halogen-bonded motif of a 4-methylpyridine (4MePy) derived from [{4MepyN−I+−NpyMe4}{4MepyN−Ag+−NpyMe4}](PF6)231 (4MePy-AgI) (Fig. 5c). The MEPS analysis revealed that the negative electrostatic potential at the [O−I−O]+ motif is much larger than that of the [N–I–N]⁺ motif, suggesting that the iodine(I) in [4MePyNO−I−ONPyMe4]+ has much stronger electron-donating ability than in [4MepyN−I−NpyMe4]+. This can be attributed to the observation that in the [N–I–N]⁺ system, a strong interaction of the iodine(I) p-hole and N-atom lone pairs, as well as π-electron delocalization across the co-planar [4MepyN−I−NpyMe4]+ structure is facilitated by the direct coordination of 4MePy N-atoms to iodine(I)21. The electron density of the iodine(I) in [N–I–N]⁺ motif is enhanced due to this charge delocalization over a larger surface area, stabilizing the iodine(I) cation. The non-co-planar [4MePyNO−I−ONPyMe4]+ halogen-bonded complex does not experience such a π-electron delocalization as the pyridine N-oxide O-atoms are not part of the π-system and the electron-pull property of the N-oxide oxygen50. As a result, the electron density of the N-oxide oxygen atoms would “concentrate” on the iodine(I), increasing its nucleophilicity and facilitating the donation of the antibonding π* HOMO orbitals to the silver(I) cation. Consequently, the iodine(I) and silver(I) form a much stronger I+–Ag+ coordinative bond (2.86 Å) in 1-AgI. However in the solid state only weak I+···Ag+ contacts (e.g., 3.5184(7) and 3.5372(6) Å) in the [{4MepyN−I+−NpyMe4}{4MepyN−Ag+−NpyMe4}](PF6)2 complex) have been observed in other [{L−I−L}{L−Ag−L}]2+ complexes29,30,31.

Fig. 5: Density functional studies.
figure 5

For clarity, only half of the DFT complexes are displayed in (d–f). a Computed molecular electrostatic potential surfaces (MEPS) for the [4MepyN–I–NpyMe4]+ complex in 4MePy-AgI and the [4MePyNO−I−ONPyMe4]+ complex in 1-AgI mapped on the 0.0004 e/bohr3isosurface of the electron density. The chemical structures of b (4MePy-Ag)2, c 4MePy-AgI, d ½(1-Ag), e ½(1-AgI), and f 1/2(1*-AgI) displaying the AIM analysis data. The italicized numbers close to atoms/ions represent atomic charges in a.u. All calculations utilized the B3LYP functional, with Dunning’s cc-pVDZ basis set applied to hydrogen, carbon, and fluorine atoms, cc-pVTZ for nitrogen, oxygen, and sulfur, and aug-cc-pVTZ-PP basis sets for silver and iodine. Abbreviation: e = electron.

Full size image

Atoms in molecules (AIM) analysis have been performed for five DFT calculated complexes, 4MePy-Ag, 4MePy-AgI, 1-Ag, 1-AgI, and a 1*-AgI complex formed by 4-methylphenolate, to determine whether the interatomic interaction between Ag+ and I+ is ionic, covalent or van der Waals like (Fig. 5). The parameters51 evaluated at the bond critical points (BCP) of the I+–Ag+ and Ag+–Ag+ bonds, including electron density (ρBCP), the Laplacian of the electron density (∇2ρBCP), the electron localization function (ELFBCP), and the ratio of the potential and kinetic energy density (|V|/G)BCP are compared in Fig. 5. These parameters for the Ag–O/N and I–O/N/C bonds are given in the Supplementary Information (Figs. S20–S25). Generally, a low value of ρBCP, a large positive value of ∇2ρBCP, and small values of ELFBCP and (|V|/G)BCP indicate an ionic interaction. A large ρBCP, a large negative ∇2ρBCP, high ELFBCP, and (|V|/G)BCP above 2.0 indicate a covalent bond. The values for the I+–Ag+ and Ag+–Ag+ bonds clearly indicate the absence of covalency, unlike, e.g., in the C–O bond of 4-methylphenolate (ρBCP = 2.008 Å−3, ∇2ρBCP = − 16.084 Å−5, ELFBCP = 0.66, (|V|/G)BCP = 2.61).

Upon examining five DFT calculated complexes, it is evident that the Ag–O/N coordination bonds and I–O/N/C XBs have large values of ρBCP and ∇2ρBCP, indicating notably stronger interactions with ionic character. The larger ρBCP values for Ag+–Ag+ bonds in 4MePy-Ag and smaller ρBCP values for the I+–Ag+ bonds in 4MePy-AgI complexes suggest the latter type is weak. In contrast, the 1-Ag and 1-AgI complexes display the opposite trend, smaller ρBCP values for Ag+–Ag+ bonds and larger for Ag+–I+ bonds. This observation suggests that the Ag+–Ag+ in 1-Ag and I+–Ag+ in 1-AgI, which are formed with 4MePyNO, are stronger compared to those observed in 4MePy-Ag and 4MePy-AgI, which use the parent 4MePy. The trends of ∇2ρBCP and ELFBCP values align with those of ρBCP. The ρBCP and ∇2ρBCP values for the I+–Ag+ in 1-AgI are ~4 times larger and ELFBCP is ~3 times larger than for the long I+···Ag+ contacts observed in 4MePy-AgI, reinforcing the notion of a stronger chemical bond with ionic character. Additionally, the energy required to break one Ag–I bond, that is, removing one [O−I−O]+ unit, from the 1-AgI assembly is ~409 kJ/mol while breaking one Ag–O(OTf) bond, which entails removing half an OTf unit, requires ~567 kJ/mol. When comparing the 1-AgI and 1*-AgI systems, the C–O bond in 1*-AgI exhibits a significantly more covalent character than the N–O bond in 1-AgI. The charge difference between Ag+ and I+ ions in 1-AgI is 0.33 a.u., and in 1*-AgI, it is 0.27 a.u., leading to a decrease of ∇2ρBCP in 1*-AgI. Nevertheless, the ionic characteristics of I+–Ag+ coordination bonds and I–O XBs in 1-AgI and 1*-AgI remain similar.

The covalency magnitude of an interaction is represented by the (|V|/G)BCP ratio, an energetic descriptor used to differentiate between two types of closed-shell bonding52,53. When (|V|/G)BCP < 1, the kinetic energy density becomes the predominant factor, resulting in electron destabilization at the BCP. This scenario indicates an absence of covalency and is characterized as pure closed-shell interactions. Conversely, when (|V|/G)BCP > 2, it signifies electron sharing, a.k.a. a covalent bond, with a high potential energy density and electron stabilization at the BCP. For values between 1 and 2, the bonds are classified as intermediate bonds with covalent and ionic character. The (|V|/G)BCP values of Ag+–I+ bonds in 1-AgI and 1*-AgI are slightly above one and notably larger than those in 4MePy-AgI. This observation implies that the I+–Ag+ bonds in 1-AgI and 1*-AgI can be described as closed-shell type but exhibit considerable electron sharing and covalent character. Additionally, the atomic charge on the silver(I) ion coordinated by N-oxide oxygens in 1-Ag is +0.79, whereas the iodine(I) atom in 1-AgI carries a charge of +0.45. This difference indicates greater electron donation from the oxygen atoms to iodine(I) than to silver(I), as well as greater electron delocalization onto the iodine center. These findings suggest that the I–O interactions are stronger and more covalent in nature, while the Ag–O interactions are comparatively weaker and less covalent.

To futher describe the nature of I–Ag bond, we employed the extended transition state-natural orbital for chemical valence (ETS-NOCV) method, which previously has been used for describing the donor-acceptor behavior in non-metal-metal systems54,55,56. While energy decomposition analysis dissects the interaction between two fragments into electrostatics, Pauli repulsion, and orbital contributions, ETS-NOCV focuses on the orbital term, further decomposing it into chemically meaningful contributions57. This approach, based on pairs of bonding and antibonding orbitals, enables the extraction of a “diatomic-like” picture of chemical bonding. Additionally, the NOCV-pair contributions to the deformation density often correspond to electron donation and back-donation between fragments. For our analysis, the 1-AgI system was divided into three fragments (Fig. S26). The total orbital interaction energy, obtained by summing all NOCV contributions, is –38.8 kcal/mol, underscoring the substantial orbital involvement in I–Ag bond formation. The total electron transfer amounts to 3.7 electrons, close to the four electrons expected for two coordination bonds.

The six most energetically significant NOCV pairs are summarized in Table 1 (energies > 1 kcal/mol and electron transfer > 0.1 e) and illustrated in Fig. 6. The first two pairs, nearly degenerate, represent electron donation from I- to Ag-atom, with a combined orbital interaction energy of –14.8 kcal/mol. Iodine’s donor orbitals exhibit 90% p-character, while the acceptor orbitals on Ag-atom are primarily s-type (85%) with a minor p-character contribution (15%), classifying them as p(I) → s(Ag) charge transfers. Pairs 3 and 4, also nearly degenerate, correspond to Ag → I back-donation, with a total orbital interaction energy of –3.4 kcal/mol. The donor orbitals on Ag-atom exhibit dominant d-character (60%), while the acceptor orbitals on I-atom are primarily p-type (86%), indicating a d(Ag) → p(I) back-donation. Pairs 5 and 6 also represent back-donation but with lower eigenvalues (0.11) and a higher combined orbital interaction energy of –4.6 kcal/mol. In this case, the donor orbitals on Ag include 40% d-character, while s and p contributions together account for 60%. The iodine acceptor orbitals remain predominantly p-type (60%). This NOCV analysis reveals that (i) the coordination bond character of the I–Ag bond, marked by substantial orbital contributions and significant electron transfer, and (ii) a dominant I → Ag donation accompanied by a noteworthy Ag → I back-donation. In donor–acceptor interactions, the minimum of electron density (ED) along the bond path typically lies closer to the electron acceptor, while the minimum of the electrostatic potential (ESP) is closer to the electron donor. We examined the ED and ESP along the I–Ag bond path to verify the I → Ag electron transfer. As shown in Fig. S27, the ESP minimum is closer to the iodine atom with respect to the ED minimum (bond critical point)58. This observation aligns with the ETS-NOCV results and reinforces the role of iodine as the primary electron donor in I–Ag bond.

Fig. 6: The ETS-NOCV analysis in 1-AgI.
figure 6

The deformation energy plots (Δρ) of the dominant pairs of NOCVs along with the orbital interaction energies (in kcal/mol) for degenerate pairs 1&2 (a), 3&4 (b), and 5&6 (c) in 1-AgI. The isosurface value is 0.0008 a.u. The green/blue regions indicate the accumulation/depletion of the electron density and the arrows indicate the electron flow.

Full size image
Table 1 The six most energetically significant NOCV pairs are summarized, providing the interaction energy (kcal/mol), electron flow, eigenvalue, and the % of s, p, d orbital contributions at I- and Ag-atoms
Full size table

Discussion

The X-ray crystal structure of 1-AgI, viz, [Ag(m-O3SCF3)2{(4MePyNO)2I}]2, revealed, together with a paddle-wheel-type dianion [Ag2(m-O3SCF3)4]2–, a [O–I–O]+ halogen-bonded moiety, featuring two symmetric I+–Ag+ coordination bonds. The orthogonality of [O–I–O]+ halogen bonds and I+–Ag+ coordination bonds describes a scenario where these two bonding interactions coexist and operate independently within a 1-AgI molecular framework. Their orthogonality results from the spatial arrangement generated by the p-hole and electron-belt surrounding the iodine(I) cation. The [O–I–O]+ halogen bonds are stabilized by the 4-methylpyridine N-oxide system. Notably, the single crystal X-ray structure of the [MePyNO–I–ONPyMe]+ provides the atom-precise experimental structural proof of this [O–I–O]+ halogen-bonded system. The I+–Ag+ bond length of 2.8625(16) Å in 1-AgI is typical among genuine M–M bonds. DFT studies revealed that the I+–Ag+ bond is a closed-shell interaction with considerable electron sharing and covalent character. Our future studies will explore the coordination behavior of I+ with other metals, expand the scope to diverse N-oxides and mixed-ligand systems, and seek to isolate new crystalline 1-AgI analogs to investigate their solid-state and solution-phase reactivity toward various halogens and halogentated systems in different solvents.

Method

4-Methylpyridine N-oxide (99%, lot no # 14227CAV) was purchased from Sigma Aldrich, whereas silver(I) trifluoromethanesulfonate (>99%, lot no # DQDPC-EO) and elemental iodine (>99%, lot no # Z5T8NRB) were purchased from TCI Chemicals Europe. Dichloromethane from SPS system dried over 4 Å molecular sieves were used for synthesis. The 4-methylpyridine N-oxide was vacuum-dried overnight before usage.

X-ray crystallography

The X-ray crystal structure data for 1-Ag (crystallized from dichloromethane) and 1-AgI was collected at 100 K, using a Bruker D8 Venture diffractometer equipped with a CMOS area detector and Mo-Kα (λ = 0.71073 Å) radiation. APEX5 (version v2023.9-2) was used for the data collection and reduction. The X-ray crystal structure data for all other silver(I) complexes were collected at 120 K using an XtaLAB Synergy R four-circle diffractometer equipped with a Hybrid Pixel Array Detector (detector type: HyPix-Arc 100). The diffraction source type is PhotonJet R (Cu, λ = 1.54184 Å) X-ray source. CrysAlisPro (version 1.171.42.89a) was used for the data collection and reduction, and the intensities were absorption corrected using a Gaussian face index absorption correction method. The intensities were absorption-corrected using a multi-scan absorption correction method. The structures were solved by intrinsic phasing (SHELXT)59 and refined by full-matrix least-squares on F2 using the OLEX260, utilizing the SHELXL-2015 module61,62.

Density functional theory studies

Bonding analysis according to the atoms in molecules (AIM) scheme was performed based on single point DFT calculations of molecular entities cut out from the crystal structures. The 1-Ag structure was obtained from the 1-AgI structure by replacing the two iodine atoms with silver atoms (for 1*-AgI, all four nitrogen atoms were replaced with carbon atoms) without reoptimizing the structure. All calculations employed the B3LYP functional. Dunning’s cc-pVDZ basis set63 was used for hydrogen, carbon, and fluorine atoms, and Dunning’s cc-pVTZ basis set63,64 for nitrogen, oxygen, and sulfur. For silver and iodine, aug-cc-pVTZ-PP basis sets65,66,67 with corresponding effective ([Ar]3d10) core potentials66,67,68 were applied. All calculations were performed with the Gaussian16 program69. Atomic charges were calculated from natural atomic orbitals with the NBO 7.0 program70. Bader analysis was performed with the MultiWFN program71.

Raman spectroscopy

Raman measurements for 1-AgI and 4MePyNO were recorded using a Bruker MultiRAM II spectrometer equipped with a low-temperature Ge detector (1064 nm, 100–180 mW, 4 cm−1 resolution). Spectra of single crystals of 1-AgI were recorded at −196 °C using the Bruker RamanScope III. Raman measurements for 1-Ag, AgOTf, diiodine were carried out at room temperature using a Thermo Scientific DXR Raman Microscope equipped with a 532 nm laser at 10 mW of power and a 10× microscope objective, resulting in a 2 µm spot size. The Raman measurements were carried out using 2.0 s of exposure.

Synthesis of 1-Ag

Silver(I) trifluoromethanesulfonate (47.1 mg, 0.183 mmol, 1.0 eq) and 4-methylpyridine N-oxide (1) (20 mg, 0.183 mmol, 1.0 eq) were mixed and dried under vacuum overnight. Subsequently, dichloromethane (5 mL), and the mixture stirred until solids formed. The solids were separated from the solution through filtration, and the filtrate was stored at −20 °C, resulting in colorless crystals suitable for X-ray diffraction examination analysis. The identical crystal structure can be obtained by repeating the experiment without vacuum drying the solids and allowing the solvent to gradually evaporate at the room temperature. Yield 62 mg (91%) Elemental analysis: calculated for (C6H7NO)(AgOTf): C, 22.97; H, 1.93; N, 3.83. Found for (C6H7NO)(AgOTf)(H2O): C, 21.89; H, 2.35; N, 3.66. Raman (cm−1): 3096, 2933, 1627, 1395, 1223, 1184, 1052, 1032, 854, 761, 665, 573, 514, 481, 354, 321. 206.

Synthesis of 1-AgI

Silver(I) trifluoromethanesulfonate (9.5 mg, 0.366 mmol, 1.0 eq) was added to the 5 mL dichloromethane solution of 4-methylpyridine N-oxide (1) (40 mg, 0.366 mmol, 1.0 eq). The mixture was shaken to dissolve the components. Diiodine (5.6 mg, 0.220 mmol, 0.6 eq) was added to the solution. The silver (I) iodide precipitates were not filtered. The sample was stored at −24 °C. Colourless crystals formed overnight on the top of the silver(I) iodide principates. Notes: (i) The synthesis was performed in a Schlenk tube with a PTFE valve. (ii) The first attempt at 1-AgI synthesis was carried out in the glovebox. The crystallization experiment was conducted twice under the exclusion of moisture and oxygen using the Schlenk procedure. The glovebox and the Schlenk procedures resulted in crystals with identical crystal structures. (iii) The sample was carried to the diffractometer chamber by placing the Schlenk tube in a Dewar filled with −40 °C ethanol. (iv) The crystals for X-ray diffraction analysis were selected at −40 °C under a nitrogen environment. (v) Crystals decompose during the selection procedure. Elemental analysis could not be obtained due to crystals instability. Raman (cm−1): 1506, 1388, 1314, 1220, 735, 578, 384, 294, 181.