Enhanced magnetic spin-spin interactions observed between porphyrazine derivatives on Au(111)

The study addresses how functionalizing porphyrazine ligands with reactive thiadiazole rings enhances intermolecular and molecule–substrate interactions and thereby modifies magnetic behaviors relevant to spintronics and quantum information processing. Prior work on metal phthalocyanines (Pc) shows limited spin–spin interaction effects due to inert C–H-terminated perimeters. Replacing benzene rings with 1,2,5-thiadiazole units in tetrakis(thiadiazole)porphyrazine (TTDPz) is known to increase intermolecular interactions through S–N electrostatic attractions and extended π-electron delocalization. Vanadyl TTDPz (VOTTDPz) exhibits S = 1/2 spins and ferromagnetic ordering in one polymorph in bulk. The purpose here is to probe magnetic properties of VOTTDPz films on Au(111) using STM/STS, focusing on Kondo resonance features, spatial variations within single molecules, and the emergence of spin–spin interactions (RKKY) in closely packed molecular lattices. The work aims to clarify how ligand reactivity and adsorption geometry (VO-up flat-lying vs VO-down tilted) control Kondo behavior and spin coupling.

Metal phthalocyanines (Pc) have been extensively studied for spin properties, yet spin–spin interaction phenomena are scarce due to weak intermolecular and molecule–substrate interactions with inert perimeters. Stuzhin et al. synthesized tetrakis(1,2,5-thiadiazole) porphyrazines and metal complexes (MTTDPz, M = Mn, Fe, Co, Ni, Cu, Zn), highlighting enhanced reactivity and π-delocalization across thiadiazole rings. Suzuki et al. reported significant intermolecular interactions via S–N attractions between adjacent TTDPz molecules and electron-deficient peripheral rings promoting electronic delocalization. Prior work on VOTTDPz established two polymorphs (α: 1D, β: 2D) and S = 1/2 spin with ferromagnetic ordering in α-form. Kondo features in Pc systems have been attributed to either metal d-electron spins or ligand π-radicals, with varied spectral shapes (dip, staircase, symmetric peak). Manipulations such as dehydrogenation can shift spin polarization from metal centers to ligands, altering Kondo signatures. RKKY interactions are known to modify Kondo spectra, broadening or sharpening peaks depending on AFM or FM coupling, respectively, in various adsorbate systems including FePc lattices.

• Sample preparation: VOTTDPz synthesized per prior methods. Molecules sublimed onto clean Au(111) under ultra-high vacuum. Substrate cleaning, deposition, and low-temperature STM performed in UHV.
• STM/STS: Measurements at 4.7 K. STS acquired with lock-in detection using 1 mV modulation. Magnetic fields (0, 3, 5 T) applied perpendicular to the surface to study field dependence of Kondo features. Topographic imaging identified two ordered monolayer phases (I and II).
• Structural analysis: STM images of phase I and II used to define commensurate unit cells relative to Au(111) with matrix relations: phase I unit vectors s,t = [[9,9],[0,4]]·(a,b); phase II u,v = [[16,6],[0,4]]·(a,b). DFT (VASP) structural optimizations built models with VO-up flat-lying and VO-down tilted molecules to match STM contrast; Tersoff–Hamann method used to simulate STM images.
• DFT specifics: Spin-polarized calculations with plane-wave basis and PAW potentials; PBE GGA exchange-correlation; structures relaxed until forces < 0.03 eV/Å. Au(111) modeled as a three-layer slab with bottom layer fixed. Spin-resolved local density of states (LDOS) computed for V and ligand N atoms.
• Spectroscopic analysis: Kondo resonance characterized near EF. Background-subtracted dI/dV spectra fitted with Fano line shape: dI/dV ∝ (ε + q)^2/(ε^2 + Γ^2), where q is the Fano parameter, ε0 the resonance energy, and Γ the half-width. Spatially resolved STS collected at individual thiadiazole lobes within single molecules to probe local coupling variations.

• Two ordered phases identified:
  • Phase I: Flat-lying VO-up molecules. DFT indicates two thiadiazole groups bend toward Au (stronger coupling) and two toward vacuum (weaker coupling), breaking fourfold symmetry.
  • Phase II: Alternating flat-lying VO-up and tilted VO-down molecules within a commensurate lattice. STM simulations reproduce bright (tilted) and dark (flat) contrasts.
• Spin distribution: Spin-polarized LDOS shows V 3d (dz2) SOMO and spin density delocalized onto the porphyrazine ligand; ligand N also exhibits SOMO just below EF with opposite-spin counterpart above EF.
• Magnetic-field dependence (phase I, ligand site): After background normalization, ΔG/G0 increases with B: 0.38 (0 T), 0.86 (3 T), 1.40 (5 T). Fano fits show ε0 shifts toward EF with B: −12.0 mV (0 T), −5.5 mV (3 T), −2.5 mV (5 T). Peak widths: ~9.0 mV (0 T) to ~7.8 mV (5 T). Peaks become stronger and more symmetric with B.
• Regime crossover: Data consistent with a magnetic-field-induced crossover from mixed valence (MV) at low B (broader, off-EF) to Kondo regime at higher B (sharper, closer to EF). Zeeman splitting not resolved due to intrinsic broadening comparable to expected splitting.
• Intramolecular spatial variation (phase I): STS on individual thiadiazole lobes reveals Kondo peak on lobes bent toward vacuum and Fano dip on lobes bent toward substrate, indicating position-dependent molecule–substrate coupling. Fano fit parameters: for position B (peak) q = −0.08 ± 0.01, Γ = 7.7 ± 1.5 mV; for position IV (dip) q = 467 ± 1, Γ = 8.6 ± 1.7 mV.
• Phase II spin-coupling effects: Clear Kondo peaks at b1, d1, d2; suppressed/absent Kondo at b2. Spin-polarized DFT shows all molecules retain spin, but b1 couples ferromagnetically (FM) with neighbors (d1, d2) while b2 couples antiferromagnetically (AFM). Observed Kondo behavior aligns with RKKY interaction: FM coupling yields sharper, stronger Kondo peak; AFM coupling broadens/weakens and can suppress the peak.
• Supporting evidence: Preliminary XMCD indicates weaker net magnetization in VOTTDPz films than VOPc, consistent with partial AFM ordering in phase II due to enhanced intermolecular interactions from thiadiazole ligands.

Functionalizing porphyrazine ligands with thiadiazole units strengthens molecule–molecule and molecule–substrate interactions, reducing intermolecular spacing and enabling significant spin–spin coupling on Au(111). This enhanced interaction manifests in two ways: (1) within single molecules, thiadiazole groups with stronger substrate coupling produce Fano dips while weaker coupling yields Kondo peaks, demonstrating that Fano line shapes can probe local chemical environments at submolecular resolution; (2) across molecules, the compact phase II lattice supports RKKY-mediated interactions where FM-coupled sites maintain strong Kondo resonances and AFM-coupled sites exhibit suppressed/broadened features. The magnetic-field-driven evolution of the resonance toward EF with increased intensity and reduced width suggests a crossover from mixed valence to the Kondo regime, rather than a simple gapped Zeeman splitting scenario, due to significant intrinsic broadening. Collectively, the results address the central question of how ligand substitution tunes spin–spin interactions: thiadiazole functionalization enhances RKKY effects that can modulate or quench the Kondo state depending on local coupling, offering a route to engineer spin interactions in molecular films for spintronic applications.

This work demonstrates that vanadyl tetrakis(thiadiazole) porphyrazine on Au(111) forms compact, commensurate monolayers with two distinct phases that exhibit rich spin physics. In phase I (flat VO-up), Kondo resonances appear at both metal center and ligand, with intramolecular crossover between Kondo peak and Fano dip reflecting local molecule–substrate coupling at individual thiadiazole groups. In phase II (alternating flat VO-up and tilted VO-down), spin-polarized DFT and STS reveal site-dependent Kondo behavior governed by RKKY interactions: FM-coupled sites show pronounced Kondo peaks while AFM-coupled sites show suppressed features. Magnetic fields tune the spectra from mixed valence toward the Kondo regime. These findings highlight that reactive thiadiazole ligands enhance spin–spin interactions and that Fano/Kondo line-shape analysis can serve as a sensitive probe of local chemical and magnetic environments. Future work should clarify the microscopic mechanism of the field-induced crossover (including potential magneto-mechanical distortions), include dispersion-corrected DFT for more accurate adsorption geometries and charge transfer, and employ element-specific magnetometry (e.g., XMCD) to map FM/AFM ordering across phases.

• DFT employed semi-local PBE without explicit dispersion corrections, introducing uncertainty in molecule–surface distances and charge transfer for weak van der Waals interactions.
• Zeeman splitting of the Kondo resonance could not be resolved at available fields due to intrinsic spectral broadening.
• The proposed field-induced crossover from mixed valence to Kondo regime and potential magneto-mechanical distortion mechanisms require further investigation.
• Assignment of tilted configurations in phase II relies on STM contrast, steric considerations, correspondence to bulk β-phase, and simulated STM images rather than direct structural determination.
• XMCD evidence for AFM ordering is preliminary and comparative; comprehensive magnetic characterization remains to be done.