*3.1. Synthesis of A1 and A2*

The same general procedure was employed for the synthesis of A1 and A2. The synthesis of A1 is described as an example. To 3.76 g (0.01 mmol) of AB-NH<sup>2</sup> [39] dissolved at 70 ◦C in 30 mL of dry THF 3.87 g (0.02 mmol) of 4-(diethylamino)-2-hydroxybenzaldehyde was added under stirring. After 1 h, the crude product precipitated at room temperature. The compound was crystallized by dichloromethane/hexane and furtherly purified by washing with hot acetone. T<sup>m</sup> <sup>=</sup> <sup>280</sup> ◦C; T<sup>d</sup> = 330 ◦C. <sup>1</sup>H NMR (400 MHz, DMSO *d*6, 25 ◦C, ppm): 1.12 (t, 12H), 3.46 (m, 8H), 4.00 (s, 6H), 6.21 (d, 4H), 6.44 (d, 2H), 7.46 (m, 8H), 7.65 (m, 2H), 8.89 (s, 2H), 10.65 (s, 2H). Elemental analysis calculated (%) for C42H46N8O4: C, 69.40; H, 6.38; N, 15.42; found: C, 69.60; H, 6.88; N, 15.88. MALDI-TOF of A1 m/z: 727.39 (M + H).

Chromophore A2 was obtained using 1,1,2,2-tetrachloroethane as the solvent and furtherly purified by washing with hot acetone and then hot dioxane: T<sup>m</sup> = 327 ◦C; T<sup>d</sup> = 330 ◦C. <sup>1</sup>H NMR (400 MHz, DMSO-d6, 25 ◦C, ppm): 4.14 (s, 6H), 6.65 (m, 4H), 7.20–8.80 (m, 12 H), 9.20 (s, 2H), 12.55 (s, 2H). Elemental analysis calculated (%) for C34H26N8O8: C, 60.53; H, 3.88; N, 16.61; found: C, 60.50; H, 3.09; N, 16.80. MALDI-TOF of A2 m/z: 675.16 (M + H).

#### *3.2. Theoretical Calculations*

Quantum-mechanical calculations were performed with the Jaguar package, Schrödinger Release 2017-4 [58], on the theoretical level DFT/B3LYP, and the molecular geometry was optimized with functional B97-D3 [59]. Charges were determined using the NBO approach. Dunning's correlation-consistent triple-ζ basis set cc-pVTZ (-f), which includes a double set of polarization functions, was used for single-point calculations on optimized geometries. TD-DFT and Tamm-Dancoff [60] approximations were used to perform calculations at neutral compound geometry to extract absorption values from vertical excitation energies. The solvent was simulated using Poisson Boltzmann Solver (PBF) [61]. The stacked dimers shown in Figure S2 were constructed with the Jaguar package from the minimized structures. Each dimer was initially minimized by simulated annealing using AMBER15FB [62] forcefield and then optimized at the B3LYP level, maintaining the dimer starting structure.

Computed redox data was used to calculate "scaled" HOMO and LUMO energies, through the following equations:

$$\text{Absolute Electrical Potential} = \text{Electrode Potential} + \text{NHE Energy} \tag{1}$$

$$\text{Orbital Energy} = \text{Redox Potential} + \text{Absolute Electrical Potential} \tag{2}$$

where "NHE Energy" represents the energy of the NHE electrode in water (−4.28 V) and "Electrode Potential" represents the potential of the chosen electrode relative to NHE.

#### **4. Conclusions**

The spectroscopic behavior of two novel azobenzene chromophores with a symmetrical skeleton was systematically examined. Their coloring ability was tested in solution as neat crystalline samples and dissolved in a polymeric amorphous matrix. The doped PVC blends are an example of stable dyed polymers. In both cases, the dyes A1 and A2 are soluble up to 30 wt% in PVC matrix with CIE coordinates in the orange to red region. The films of both neat and blended dyes kept over three months under natural light at room temperature in the air perfectly retain structural and optical characteristics. No swelling nor release was detected in distilled water for 30 days.

On the other hand, the effort to obtain good steady PL response from azobenzene scaffolds is justified by the poor scientific documentation about this versatile and tunable functional group. A systematic study of the optical performance of the two differently substituted azobenzene dyes was performed, and DFT computational study gave information on HOMO-LUMO localization. The

azo-based chromophores are emissive in solution and the solid-state, claiming a role as dye-dopants for emissive layers. Low-doped blends were obtained in PS, PVK, and PFO. PLQYs measured on the emissive blends are above 10% in all cases. In particular, PLQY of the orange-red A1-PFO blend (57%) represents a remarkable result for azobenzene based material, able to modulate the optoelectronic characteristics of the blue emissive PFO. The simple synthetic procedure, the affordability, solubility and processability of the dyes and the doped blends make the azo-dyes rare examples of azobenzene based materials potentially employable both as dyes and as fluorophores.

**Supplementary Materials:** Table S1. Electro-optical properties calculated on A1 and A2 in vacuum and associated with PFO. Table S2. Reduction potential of similar azobenzenes [2,3]. Figure S1. Energies of orbital levels of PFO, A1 and A2. The arrows show possible electronic transitions. Figure S2. Frontiers orbitals HOMO and LUMO calculated for the dimers PFO-A1 (above) and PFO-A2 (below). Figure S3. Pictures of A1 and A2 crystalline films under polarized light. Figure S4. Emission spectra of PVK (orange line) and PFO (blue line) films, excited on the absorption maxima.

**Author Contributions:** Conceptualization, U.C. and B.P.; data curation and formal analysis, R.D.; funding acquisition, U.C.; investigation, U.C.; methodology, R.D. and B.P.; project administration, U.C.; resources, U.C.; writing—original draft, R.D., S.P., S.C, performed the experiments; analyzed the data S.C., R.D. and R.S.; writing—review & editing, B.P. and U.C. All authors have read and agreed to the published version of the manuscript.

**Funding:** We gratefully acknowledge the financial aid provided by the Italian Ministry of Education, University and Research (MIUR) under grants PON PANDION 01\_00375.

**Acknowledgments:** We thank Francesco Marrafino for assistance with TDDFT calculations and for comments that significantly improved the manuscript.

**Conflicts of Interest:** There are no conflicts to declare.
