PID-1, 2, 3, and 4 were synthesised as shown in Figure 1A. Fischer indole synthesis afforded phenylindoles 4–6, followed by N-alkylation with bromoethane. Commercially available, 7, plus intermediate products 8–10 were dimerised via acetyl chloride-promoted Knoevenagel condensation (Noureddin et al, 2020). Their structures were confirmed by NMR analysis (see SI for the ¹H and ¹³C NMR data) with a singlet at approximately 5.16 ppm, highly characteristic of gem-disubstituted olefin resonances. The materials were robust solids, typically white or off-white in colour except for the nitro analogue which was yellow (as is typically found for many nitro aromatics) (Porter, 1955). Single-crystal X-ray diffraction of PID-1 confirmed the molecular structure (Figure 1B).

The switching behaviour of PID-1 was a serendipitous discovery, with its properties identified following the analysis of compound purity by thin-layer silica gel chromatography (TLC) and UV illumination (365 nm). Further analysis and synthesis showed that compounds PID-1, PID-2, and PID-3 all exhibited light switching/sensitivity and interestingly showed a range of colours (pink, purple, and orange) upon switching (Figure 2). With the removal of the UV light, the colours would disappear under ambient conditions via thermal relaxation. However, the p-nitrophenyl derivative, PID-4, was yellow before and after irradiation, with no apparent photo-switching character at this wavelength.

The application of PID-1 as a light-responsive ink was demonstrated by patterned illumination of PID-1 coated onto white printer paper. This was achieved by printing a black inverse mask onto an acetate transparency by inkjet printing which allowed light-induced patterning on the paper, with “pink/purple” images appearing, that remained visible for 15 min (Supplementary Figure S22). However, these patterns rapidly disappeared when irradiated at 525 nm. This process was fully reversible over many cycles without loss of colour intensity or photobleaching (Figure 8) (for video of photo-switching with photo and thermal reversibility, see Supplementary Video S1).

The pink/purple colour changes observed (see Figure 2) and absorbance measurements (showing a new band at approximately 515 nm) upon irradiation suggest that the product had an increased degree of conjugation. The fact that the reaction was light-induced suggested that the forward reaction could be a [2 + 2] cycloaddition, generating a strained cyclobutene unit upon reaction between the olefin and the indole (Scheme 1). This undergoes an allowed [2 + 2] reverse reaction when illuminated at 525 nm to regenerate the starting material. The room temperature (in the dark) reversion is slow, as a classical [2 + 2] reaction is “disallowed” explaining the relatively slow change seen. For example, when the compounds were illuminated as shown in Figure 2 in the dark, they took >15 min to return to the colourless state.

The proposed mechanism is in concurrence with evidence of other reported solid-phase photo-switches, such as diarylethenes (Kobatake et al, 1999) and nobornadienes (Bauer et al, 2019), with the transformation resulting in a relatively small change in volume of the molecule. It has been previously reported that small changes in volume between switched states is more favourable to solid-phase photo-switching due to the decreased steric hindrance of nearby molecules in the crystals (Gonzalez et al, 2020). We hypothesize that a reversible [2 + 2] cycloaddition is responsible for the photo-reversible reaction, but further studies are required to fully understand this reaction, perhaps by in situ illumination solid-state NMR spectroscopy. Other mechanistic options for the thermally reversible route include via a bis-radical or an indole cation/allylic anion.

Of the synthesised compounds, PID-1 showed the most promising properties due to its one-step synthesis, its stability in aerobic environments, and the large difference between the switched and unswitched states in the visible region of the absorption spectra. Its application as a photochromic ink was demonstrated with many cycles of writing (365 nm illumination) and erasing (535 nm illumination).

Possible future applications of these materials include writable filters or gratings. Within a biological context, PIDs could find applications in modulating drug–receptor interactions.

Irradiation of samples at 365 nm was performed using an ENF-240C E-Series UV Lamp (4 W, 120 V). Irradiation of samples at 525 nm was performed using a Thorlabs LIU525B-525 nm Green LED Array Light Source.

A solution of PID-1 (25 mM in acetone) was drop-coated onto A4 standard white printer paper (purchased from Banner). The solvent was evaporated under a nitrogen flow. Masks were prepared by inkjet printing patterns onto A4 acetate transparencies (purchased from Niceday) and cut to a size of 7.5 cm × 6 cm.

The paper and mask were sandwiched together and irradiated (365 nm, 15 s). The treated sample was either irradiated (525 nm, 5 s) or left in the dark (15 min).

Absorbance spectra were collected from 190 nm to 700 nm on an Agilent 8453 UV-visible absorbance spectrometer (integration time = 0.5 s). Thin films of the solid compound were prepared as follows: 50 µL of PID in DCM (38 mM) was placed onto one side of the 3-mL quartz cuvette (path length 1 cm) which was evaporated under N₂. Spectra for PID-1, 2, and 3 were collected with an increasing irradiation time (365 nm). Furthermore, spectra of PID-1 were collected before and after irradiation (365 nm, 2 min) and then either left in the dark for 15 min or irradiated at 525 nm for 15 s.

Cycles of 365 nm and 525 nm on thin films: Thin films of PID-1 were prepared as previously described, however with addition of 2 × 50 µL (38 mM, DCM) of PID-1. Spectra were collected before irradiation and then after 30 cycles of UV irradiation (365 nm, 5 s) and UV-visible irradiation (525 nm, 2 s).

ATR-FTIR spectroscopy was performed on a Perkin Elmer Spectrum Two (UAR Two) FTIR spectrometer. Spectra were taken on powder samples of PID-1 at 300 K before and after irradiation at 365 nm for 10 min.

Samples of PID-1 (4 mg) were added to glass vials (1.7 mL). Two samples were left in air, and two samples were purged and backfilled with N₂ (×3 times). The samples were irradiated at 365 nm for 60 s and then either left in the dark for 17.5 h or irradiated at 525 nm for 15 s.

PID-1 was recrystallised from a mixture of ethyl acetate and hexane by slow evaporation. A suitable crystal with dimensions 0.42 × 0.30 × 0.27 mm³ was selected and mounted on a MITIGEN holder in paratone oil on a Rigaku Oxford Diffraction Xcalibur diffractometer. The crystal was kept steady at T = 120.01 (10) K during data collection. The structure was solved using the SHELXS solution program by using direct methods and Olex2-1.5-beta as the graphical interface. The model was refined with SHELXL 2018/3 using full matrix least squares minimisation on F2. The crystallography data were uploaded to the Crystallography Open Database, under accession number 3000437.

For full synthetic procedure and characterisation by NMR spectroscopy and mass spectrometry, see the Supplementary Material section.