Synthesis of 4-substituted azopyridine-functionalized Ni(II)-porphyrins as molecular spin switches

We present the synthesis and the spin switching efficiencies of Ni(II)-porphyrins substituted with azopyridines as covalently attached photoswitchable ligands. The molecules are designed in such a way that the azopyridines coordinate to the Ni ion if the azo unit is in cis configuration. For steric reasons no intramolecular coordination is possible if the azopyridine unit adopts the trans configuration. Photoisomerization of the azo unit between cis and trans is achieved upon irradiation with 505 nm (trans→cis) and 435 nm (cis→trans). Concurrently with the isomerization and coordination/decoordination, the spin state of the Ni ion switches between singlet (low-spin) and triplet (high-spin). Previous studies have shown that the spin switching efficiency is strongly dependent on the solvent and on the substituent at the 4-position of the pyridine unit. We now introduced thiol, disulfide, thioethers, nitrile and carboxylic acid groups and investigated their spin switching efficiency.


Synthesis of 3-(3-bromophenylazo)-4-mercaptopyridine (12)
In flame dried glassware under an atmosphere of nitrogen hexamethyldisilathiane (64.2 mg, 360 µmol, 7) was dissolved in 7.6 mL of dry tetrahydrofuran and cooled to 0 °C. n-Buthyllithium (144 µL 2.5 M in hexane, 360 µmol) was slowly added and stirred for 30 min at 0 °C and further 30 min at ambient temperature. Separately, 3-(3-bromophenylazo)-4-chloropyridine (106 mg, 358 µmol, 10) was dissolved in 3 mL of dry tetrahydrofuran. Both solutions were cooled to −78 °C, combined and stirred for 2 h at this temperature in darkness. Afterwards, the mixture was allowed to warm to ambient temperature and stirred for further 3 h in darkness. The mixture was diluted with ethyl acetate (30 mL), a saturated solution of sodium chloride (10 mL) and deionized water (20 mL). Layers were separated and the aqueous layer was extracted with ethyl acetate (7 × 20mL). The combined organic layers were dried over magnesium sulfate and the solvent was removed in vacuo. The crude product was purified by column chromatography (silica, cyclohexane/ethyl acetate 1.
For the molecules already described in literature [3] the values were revised due to results of Gutzeit et al. [6].

Determination of the cis-trans ratio in 1 H NMR
The cis-trans-ratio for 1f and 1i was analysed by integration of H-11 because this signal suffers least from paramagnetic line broadening and overlap with other signals [4]. The results are given in Figure S14.

Determination of extinction coefficients of 1e, 1g, 1h, 1j
The extinction coefficient of 1e, 1g, 1h and 1j in acetone in the pure diamagnetic state was determined, by protonation of the pyridine to prevent coordination. Towards this end an excess of trifluoracidic acid was added and the sample was diluted to different concentrations. UV-vis spectra were recorded and the absorbance at λmax of the paramagnetic Soret peak was plotted as a function of the concentration ( Figure   S15.1-4).

Determination of the switching efficiency by UV-vis spectroscopy
The percentage of paramagnetic record player molecules was determined from UVvis spectra and calculated by equation [7]: ). %para percentage of paramagnetic record player molecules in %.
%dia percentage of diamagnetic record player molecules in %.

Aobs.
observed absorbance at λmax of the paramagnetic Soret peak.
ε(para) extinction coefficient of completely paramagnetic record player molecules at λmax of the paramagnetic Soret peak in Lmol -1 cm -1 .
ε(dia) extinction coefficient of completely diamagnetic record player molecules at λmax of the paramagnetic Soret peak in Lmol -1 cm -1 .
c molar concentration of record player molecules in molL -1 .

DFT calculations on disulfide 1g
General: All calculations were performed with Turbomole7.2 [8], the m4 grid (in Turbomole nomenclature) and the usage of resolution-of-identity (RI) with multipole accelerated RI-J (marij) to speed up the calculations.
All energy values are based on geometry optimizations using density functional theory with the standard GGA functional PBE [9] and Ahlrichs double zeta basis def2SVP [9] followed by single point calculations the hybrid B3LYP [11] (parametrization as implemented in Turbomole7.2) and the triple zeta basis def2TZVP [11]. The B3LYP/def2TZVP/PBE/DZ level of theory performed very well for the prediction of the association enthalpies of pyridines to Ni(II)-pentafluorophenylporphyrins and is therefore a good assumption to describe record player-like systems quantum chemically [5].

Results:
The conformational analysis of the disulfide record player system 1g reveals that the double coordinated trans state is favored by more than 20 kcal mol -1 (see Table   S2) over the decoordinated species. Because of the large difference in energy the equilibrium is strongly driven towards the paramagnetic state which explains the UV/vis measurements (trans isomer >80% paramagnetic).

Switching and stability
A solution of porphyrin 1e, 1f, 1h, 1i, 1j in acetone was alternately irradiated with 505 nm and 435 nm into the PSS. UV-vis spectra were recorded after the irradiation cycle and the absorbance of λmax. of the paramagnetic (diamagnetic for 1i) Soret peak was measured (Figure S18.1-S18.8).

Spin switching in an aqueous environment
A solution of record player 1e, 1h, 1j in acetone was diluted with oxygen free water and acetone to obtain a ratio of 1:9 for acetone:water. UV-vis spectra in the PSS at 505 nm and 435 nm were recorded. Afterwards an excess of Cs2CO3 was added for deprotonation and UV/vis spectra in the PSS were recorded.

Hammett Plot
The equilibrium constant (K1) in the PSS at 505 nm (Table S3) for the coordination of the axial ligand (CN4  CN5) was determined from 1 H NMR spectroscopy (Equation

S3
), which is a powerful tool to obtain the ratio of coordinated/decordinated cis isomer (Equation S1).