Thermally and Magnetically Robust Triplet Ground State Diradical

Abstract

High spin ( S = 1) organic diradicals may offer enhanced properties with respect to several emerging technologies, but typically exhibit low singlet triplet energy gaps and possess limited thermal stability. We report triplet ground state diradical 2 with a large singlet-triplet energy gap, Δ E_ST ≥ 1.7 kcal mol^-1, leading to nearly exclusive population of triplet ground state at room temperature, and good thermal stability with onset of decomposition at ∼160 °C under inert atmosphere. Magnetic properties of 2 and the previously prepared diradical 1 are characterized by SQUID magnetometry of polycrystalline powders, in polystyrene glass, and in other matrices. Polycrystalline diradical 2 forms a novel one-dimensional (1D) spin-1 ( S = 1) chain of organic radicals with intrachain antiferromagnetic coupling of J'/ k = -14 K, which is associated with the N···N and N···O intermolecular contacts. The intrachain antiferromagnetic coupling in 2 is by far strongest among all studied 1D S = 1 chains of organic radicals, which also makes 1D S = 1 chains of 2 most isotropic, and therefore an excellent system for studies of low-dimensional magnetism. In polystyrene glass and in frozen benzene or dibutyl phthalate solution, both 1 and 2 are monomeric. Diradical 2 is thermally robust and is evaporated under ultrahigh vacuum to form thin films of intact diradicals on silicon substrate, as demonstrated by X-ray photoelectron spectroscopy. Based on C-K NEXAFS spectra and AFM images of the ∼1.5 nm thick films, the diradical molecules form islands on the substrate with molecules stacked approximately along the crystallographic a-axis. The films are stable under ultrahigh vacuum for at least 60 h but show signs of decomposition when exposed to ambient conditions for 7 h.

Figure 1.

Thermally robust triplet ground state diradicals with onset of decomposition at T ≥ 160 °C based upon TGA.

Figure 2.

Top: Single crystal X-ray structure of diradical 2 with molecule A shown only; carbon, nitrogen, and oxygen atoms are depicted with thermal ellipsoids set at the 50% probability level. Bottom: Packing of molecules A and B into a one-dimensional S = 1 antiferromagnetic chain; nitrogens and oxygens with large positive spin densities and forming close intermolecular contacts, N2A∙∙∙N3B = 3.509 Å and O4B∙∙∙N1A = 3.335 Å, are emphasized as ball-and-stick. Additional information can be found in the Supporting Information (Figs. S1-S3 and Table S1).

Figure 3.

EPR (ν = 9.65 GHz) spectrum of 1.2 mM diradical 2 in 4:1 toluene/chloroform glass at 153 K. The Δm_s = 2 transition is shown as an inset. Simulation parameters: g_xx = 2.0072, g_yy = 2.0026, g_zz = 2.0052, |D/hc| = 8.08 × 10⁻³ cm⁻¹, |E/hc| = 1.17 × 10⁻³ cm⁻¹; linewidths: LW_x = 32.0 MHz, LW_y = 100.0 MHz, LW_z = 24.9 MHz. For EPR spectra (with simulations) of 2 in polystyrene matrix at 295 K, see: SI, Fig. S7.

Figure 4.

SQUID magnetometry of polycrystalline (solid) diradical 1: plots of χT vs. T and at χ vs. T at H = 5000 Oe in the warming mode. χT vs. T data, corrected for diamagnetism, are fit to a diradical model (eq. 1), using two variable parameters: singlet triplet energy gap, 2J/k, mean-field correction for intermolecular interactions between the radicals, θ. The values of standard error, SE, and parameter dependence, DEP, are provided; goodness of fit may be measured by standard error of estimate, SEE. Fitting parameters: 2J/k = 246 K (SE = 8), θ = −5.89 K (SE = 0.08), DEP = 0.0236, R² = 0.9941, SEE = 0.0175. Complete set of magnetic data (and fits) for solid diradical 1 may be found in the SI (Fig. S8 and Eq. S1).

Figure 5.

SQUID magnetometry of polycrystalline (solid) diradical 2: plots of χT vs. T and at χ vs. T at various applied magnetic fields, H = 300000 Oe and 5000 Oe in the warming mode and H = 500 Oe in the cooling mode. Numerical fits to the diradical model (eq. 1) for T = 70 – 320 K are carried out with two variable parameters: singlet triplet energy gap, 2J/k and mean-field correction for intermolecular interactions between the radicals, θ. Numerical fits to 1D chain (eq. 2) or dimer (eqs. S2A&B, SI) models for T = 1.8 – 70 K are carried out with three variable parameters: intermolecular Heisenberg exchange coupling constant, J’/k, weight factor, N, and weight factor for isolated S = 1 diradical, N_imp. The results for numerical fits are summarized in Table 2 and in the Supporting Information (Table S3 and Figs. S9 and S10).

Figure 6.

SQUID magnetometry of 30 – 40 mM diradicals 1 (A & B) and 2 (C & D) in polystyrene matrix. Plots A and C: M/M_sat vs. H/(T – θ) plots, where θ = –0.04 or –0.05 K, at T = 1.8 – 5 K (symbols) and the Brillouin curves corresponding to S = ½ - 3/2 (lines). Plots B and D: χT vs. T data at H = 5000 Oe in the warming mode were fit to a diradical model (eq. 1), using two variable parameters: singlet-triplet energy gap, 2J/k, mean-field correction for intermolecular interactions between the radicals, θ. The values of standard error, SE, and parameter dependence, DEP, coefficient of determination, R², and standard error of estimate, SEE. Fitting parameters: diradical 1: 2J/k = 182 K (SE = 5), N = 0.733 (SE = 0.001), DEP = 0.5311, R² = 0.9441, SEE = 0.0109; diradical 2: 2J/k = 800 K (SE = 20), N = 0.9278 (SE = 0.0008), DEP = 0.3336, R² = 0.746, SEE = 0.0066. Further details are reported in the SI: Table S3, Figs. S11 and S12.

Figure 7.

Thermogravimetric analysis (TGA) of diradicals 1 and 2 under N₂; heating rate = 5 °C min–1. Further details may be found in the SI (Figs. S4-S6).

Figure 8.

C 1s and N 1s core level XPS spectra of a multilayer of diradical 2 deposited on SiO₂/Si(111) wafers (top plots), compared to the powder spectra (bottom plots).

Figure 9.

Attenuation of the Si 2p XPS signal, normalized to the corresponding saturation signal, as a function of film nominal thickness, deposition at room temperature (top panel). A typical AFM image of a 1.4-nm nominally thick film (middle panel) and its averaged height profile (bottom panel).

Figure 10.

C-K NEXAFS spectra obtained from a 1.5-nm nominally thick film (left panel). The spectra were taken in grazing incidence and in normal incidence as indicated. Geometry of the experiment (right panel).

Scheme 1.

Synthesis of Diradical 2.