Irradiation of chloroform solutions of LA using a Luzchem photoreactor equipped with UVA lamps led to a cloudy solution after 40 min of irradiation. For example, in a typical experiment, 4 mL samples of 1 mM and 5.6 mM LA both in chloroform, were irradiated in the photoreactor which was operated with 12 lamps, corresponding to an irradiance of ~ 25 W/m2 (Fig. S1 †ESI). While other solvents like acetonitrile, water-acetonitrile (1:1) and methanol also worked, chloroform proved to be convenient for the polymer synthesis since it leads to a cross-linked polymer, and most of this type of polymers are insoluble in different solvents given their linked chain structure which inhibits solubilisation [28, 29]. Figure 1 shows the spectra obtained after different times of UVA exposure.

LA spectra obtained at different irradiation times of 5.6 mM LA solution in chloroform with UVA light. Inset A shows solutions of LA 5.6 mM (left) and 1 mM (right) after 40 min UVA exposure, both solutions in chloroform

Using this UVA light source and conditions, led to the rapid bleaching of the band at 330 nm upon 20 min irradiation, with simultaneous increased absorption and light scattering (dashed lines in Fig. 1) indicating the formation of polymers or oligomers, consistent with the turbidity of the samples (Inset, Fig. 1). Based on Fig. 1, after 10 min of UVA irradiation, there is around 30% consumption of LA monomer. Visible light of ~ 400 (Fig. S2 †ESI) nm was also tested, and we found that prolonged irradiation affects the structure of LA, by breaking the S–S bond of the ring to produce a mixture of DHLA and PALA.

To establish the properties of the cross-linked polymer, the swelling behaviour was investigated. Dynamic Light Scattering (DLS) was used to relate the hydrodynamic diameter of the particles in solution to the polymer dimensions. Table 1 shows that PALA exhibited a typical swelling ability in nonpolar solvents (CH2Cl2, CHCl3, tetrahydrofuran, etc.) (Fig. S3 †ESI) and a low swelling ratio (shrink) in polar solvents like water.

DLS data are gathered in Fig. 2 and obtained by first measuring a sample of 1 mM LA in H2O (dashed black line), then after UVA irradiation of the sample for 40 min and upon the addition of DMSO to the aqueous system. The exponential curves in Fig. 2A show that the size in the LA solution increases after irradiation with UVA, given that the decay shifts to longer times, which reflects the cleavage of the S–S bond, and thus the initiation of polymerization. Further, in aqueous solution, the swelling ability was confirmed, because when adding DMSO, the curve shifts even more to the right, which is an indication that the mean size of the particles has increased.

DLS correlation function measured at 25℃ as a function of correlation time: 1 mM LA in aqueous solution, upon UVA exposure (Photoreactor) and after addition of DMSO (A); Number-weighted diameters of same samples (B)

Figure 2B also shows the particle size distribution which can be compared with the results of Table S1 (†ESI), wherein (i) the Z-average values show that longer irradiation times led to an increase in size, and (ii) that the identified particle population, which shows a peak around ~ 90 nm upon UVA irradiation, has a different peak after the addition in an organic solvent like DMSO (maximum around ~ 300 nm). This suggests that the penetration of an organic solvent into the polymer network can cause extensive swelling.

In addition, 1H-NMR offered evidence of the ring opening crosslinking driven by UVA irradiation. Figure 3 shows the 1H-NMR spectra of LA monomer and the polymer after 40 min of irradiation. For LA, (a1, a2) protons display resonances at δ 3.03–3.25 ppm, (b) protons show two peaks at δ 2.47 and 1.92 ppm, indicating the presence of the dithiolane ring. Additionally, the five-carbon straight-chain of LA show peaks at different resonances for the protons between δ 2.38 and 1.50 ppm. However, after UVA irradiation, the signals due to the hydrogens of LA decrease as the reaction progresses. Figure 3B displays that part of the signals, originally contributing to the region of protons (a) and (b), were converted to new signals at δ 3.11 ppm (a) and 2.01 ppm (b), respectively. This suggests that the ring of LA opened, and PALA had been obtained [30].Further, the location of signals for protons (d), (e), (f), (g) remain unchanged, but were broadened, which we attribute to reduced molecular mobility, compared with LA signals.

1H NMR spectra of LA 5 mM in CDCl3 (A), and PALA upon 40 min UVA irradiation (B)

Kisanuki et al. [31] provided NMR evidence for the polymer obtained by a thermal method, wherein even though the signals are more intense due to the high concentrations used, they showed results of the thermal polymerization of LA in the presence of DHLA, in which the polymer yield was suppressed significantly in the presence of a small amount of the thiol compound, indicating that the thiol acts as an effective chain transfer agent in the polymerization of LA. In this case, when UVA light is involved in the polymerization process, it is highly possible that DHLA is being formed after the initial rupture of the disulfide bond. DHLA proton signals can be identified when performing 1H-NMR of LA at higher concentrations than 5 mM following irradiation times of 40 min or longer. In fact, in Fig. 3, low resonances at δ 2.9–3.0 ppm, δ 1.8 ppm and δ 1.30 ppm show evidence of protons (a), (b), (c) and those from the -S–H- bonds of DHLA, respectively [32].This suggests the presence of DHLA obtained after irradiating LA can contribute to the polymerization of LA.

Different investigations about the structure of PALA have been recently reported, wherein a polycatenane structure is proposed [33], and it could be achieved by shorter polymerization times, which leads to low polydispersity as well as narrow size distribution. Based on the UV–Vis and NMR results, we assume that PALA might have a cross-linked structure, consistent with the swelling observed upon DMSO addition.

To elucidate further the structure of the reaction product after UVA light exposure, HPLC–ESI–MS and GPC measurements were performed. Figure S4 (†ESI) shows the mass spectrum of LA which revealed that the protonated molecular ion (m/z 205) was partially fragmented to m/z 171 [M-H2S-H]−. However, after using UVA light for 40 min a 5.6 mM sample of LA (Fig. 4A) shows charged species with m/z 410.6, 616.8 and 1033.8 detected in the negative ionization mode, due to the presence of the dimer, trimer and tetramer of LA; higher oligomers would not be detectable in our instrument since the masses exceed the detection limit. We speculate that 40 min of UVA light, under the conditions described above, might be enough to produce a polymer with at least 5 or 6 units of LA. Regarding the structure of the polymer, a polycatenane structure can be suggested; the structure is not always recognizable due to the complexity of polymer network [34] and analytical technique limitations.

ESI- MS spectrum of PALA in acetonitrile after 40 min UVA irradiation of a 5.6 mM LA solution (A), GPC traces for LA, 40 min and 20 h UVA irradiation (B) in tetrahydrofuran (THF) HPLC grade as solvent

Gel Permeation Chromatography (GPC) was also used to measure the molecular weight of PALA. Figure 4B compares three GPC curves of PALA produced after UVA irradiation of LA 5.6 mM for 40 min, about 20 h and LA before irradiation. Clearly, even after 20 h, some LA or low oligomers are still part of the solution, since the peak for the irradiated samples extends up to ~ 11.5 min, where the band due to LA is ending before any exposure. This, is a confirmation that LA can undergo structural changes under UVA irradiation for at least 40 min. In fact, evidence that LA is being polymerized upon exposure of UVA light can be confirmed from Fig. S5, which shows the UV–Vis profile of the peak at 9.5 min for the LA sample that was irradiated for 40 min, and the peak around 10.5 min for the 5.6 mM LA sample of the GPC curves described in Fig. 4B. The typical band of LA around ~ 330 nm, due to its ring, is present before irradiation (Fig. S4.B) and upon 40 min of UVA exposure, such band has disappeared (Fig. S4. A).

The peak shape for both samples in the GPC curves (Fig. 4B) is narrow suggesting a low polydispersity. By using the calibration curve with standard polystyrenes (Fig. S6, †ESI), molecular weight of PALA samples was determined to be ~ 1316 g mol−1 for the sample irradiated for 40 min and ~ 582 g mol−1 for the sample after 20 h of UVA exposure. It is worth noting that this does not lead to an absolute value, but to an estimated molecular weight that was validated with the m/z values obtained from HPLC–ESI–MS (Fig. 4A). For the small peak around ~ 12 min of Fig. 4B, it corresponds to an impurity and therefore it should be disregarded.

Interestingly, when using light as part of the polymerization process of LA, 40 min of irradiation for a concentration of 5.6 mM might seem enough to at least produce PALA of 5–6 units of monomer. However, increasing the irradiation time up to 20 h helps to produce more polymer, but this is not reflected in the molecular weight determined by GPC, since the peak for the sample irradiated up to 20 h shifts to longer times with a maximum at ~ 9.6 min and thus a lower molecular weight (~ 582 g mol−1). This may seem unusual, yet, we speculate that prolonged irradiation (20 h) leads to increased crosslinking, limiting the solubility in the THF mobile phase and as a result the molecular weight of the soluble fraction is lower. In fact, for the sample irradiated for 20 h, insoluble components were visible by turbidity and deposits the sides of the vials, suggesting high cross-linked polymers were excluded in the GPC samples.

Further, when comparing with thermal polymerization of cyclic disulfides, e.g., 1,2-dithiane, it does not proceed at monomer concentrations below 4.0 M [35]. Thus, the concentration of the monomer is an important factor not only for the thermal polymerization [36], but for the photopolymerization of cyclic disulfides, wherein specifically for LA, the amount of PALA decreased with a decrease of LA concentration. Even thought, concentrations less than 5 mM might work for producing some PALA, the yield is lower compared with higher LA concentrations.

A common way to stabilize gold nanoparticles (AuNP) involves the formation of the thiolate S–Au bond, by performing reaction of thiols with the gold surface [9], therefore, due to the inherently strong interactions between the AuNP and sulfur [37], LA interacts and caps the AuNP surface. AuNPs were synthesized using a reported method [21] via reduction with the photoenol from 3,3,6,8-tetramethyl-1-tetralone with a lifetime around ~ 3 μs, mediated by the carbonyl triplet state of the ketone (τ ~ 1.9 ns) as a precursor. In this particular case, the excited photoenol has biradical character and is useful for the fast synthesis of AuNP. The photolysis of the substituted tetralone was examined in acetonitrile solution using Laser Flash Photolysis (LFP) with a 355 nm laser excitation [20]. Briefly, 2 mL sample of 1 mM HAuCl4 in 1:1 water:acetonitrile containing 10 mM substituted tetralone under nitrogen, was irradiated with a LED source centered at 368 nm (Fig. S11, †ESI), leading to a characteristic plasmon band of AuNP (~ 550 nm) in less than one minute. A nice characteristic of photoenol reductions is that any intermediates that fail to reduce gold, regenerate the original tetralone precursor [21]. The TEM images show a nanoflower-like shape for the NPs with a plasmon peak that extends to longer wavelengths, which suggests aggregation and increased polydispersity (Fig. S7, †ESI).

In an attempt to control particle growth and aggregation, LA was used as in situ stabilizer to control the particle size. Specifically, to the tetralone system described above, 0.2 mM, 0.5 mM, 1 mM, 2 mM of LA were added to the system before irradiation. Figure 5 shows the spectra of the AuNPs when adding LA as a stabilizer. Surprisingly, when LA was present in concentrations of 1 mM, more irradiation time was needed to see the change in solution color from light-yellow to ruby red–purple and particularly, the width of the plasmon peak was reduced and shifted towards shorter wavelengths. Additionally, it is worth noting that when concentrations of 2 mM LA and higher were present as part of the reaction, no plasmon peak was observed, but a red shift of the absorbance took place, as shown by Fig. 5.

AuNP spectra obtained immediately after irradiation of 10 mM tetralone, 1 mM HAuCl4, 0.2 mM, 0.5 mM, 1.0 mM, 2.0 mM LA in water-acetonitrile with UVA light (LED at 368 nm) for 35 s and 3 min of irradiation

While concentrations of LA between 0.2 and 1 mM provide a well-defined plasmon peak between 530 and 575 nm, smaller particle sizes are obtained when increasing the concentration of LA in the system, as we can see in Fig. S8 (†ESI). It is interesting how higher concentrations than 2 mM LA inhibited the formation of the plasmon peak of the nanostructures in this system. According to Bucher et al. [19] LA can undergo photoexcitation when irradiated with UVA light, producing an excited triplet state with a lifetime of τ = 75 ns. In this study, LFP was also used to monitor the LA radical species formed after UVA excitation and the results are consistent with a short-lived species like that reported by Bücher et al. [19]. The formation of a dithiyl radical and other radical species were observed in the reaction. Based on this, a question has been raised as whether LA could reduce Au in order to form plasmonic nanostructures. As shown in Fig. 6, when a solution of 1 mM HAuCl4 and 0.5 mM LA was irradiated using a 368 nm LED for a period of 30 min, a sharp plasmon peak is obtained with a maximum at 555 nm, with AuNP having circular and triangular-like shapes and a size around ~ 85 nm, as shown by the inset. This confirms that LA can reduce Au(III). Furthermore, 0.2 mM and 1 mM LA were also tested along 1 mM HAuCl4 and, remarkably, when having a concentration of 1 mM of LA in the system, the same trends of shifted absorption and no plasmon peak were observed, which confirms that concentrations of 1 mM LA and higher, allow LA to be polymerized under UVA light, wherein the process of polymerization is favored over the reduction of the metal ion.

AuNP spectra obtained immediately after irradiation of sample containing 1 mM HAuCl4, 0.5 mM of LA with a 368 nm LED., Inset: TEM image of Au NPs using 0.5 mM LA, 100 nm size bar

Based on previous results, when the tetralone with triplet excited state of τ ~ 1.9 ns [21] and the excited species of LA (2 mM or higher concentrations) are generated in the same system, both seem to compete to reduce Au, but due to the polymerization of LA into PALA, neither the tetralone nor the LA can play that role. Of course, higher LA concentrations will also compete with the tetralone for absorption of UVA photons. To the best of our knowledge, this is the first time that PALA oligomers of at least 5–6 monomer units has been observed under direct UVA irradiation. Nonetheless, the assumption that the polymerization of LA is taking place at concentrations higher than 2 mM in a system containing tetralone, might involve the formation of self-assembled structures of PALA encapsulating gold and inhibiting further reduction of the metal ions; this may explain why we do not observe a plasmon peak under these conditions. Therefore, in this particular case, LA acts as an in situ stabilizer when being in concentrations no higher than 1 mM. However, when going above this concentration of LA, plasmon generation fails, at least in part because tetralone fails to absorb enough light given the competition by LA.

Considering the photochemical properties of LA, and a possible ring-opening polymerization upon UVA exposure, PALA was synthesized after irradiating 1 mM and 5.6 mM solutions of LA in CHCl3 for 20 h using UVA light. The solvent was evaporated for one day in order to add 1:1 water:acetonitrile as this was a suitable media for the synthesis of AuNPs in the presence of the substituted tetralone. In a typical experiment under nitrogen, a 3 mL sample containing 10 mM tetralone (1 mL), 1 mM HAuCl4 (1 mL) and PALA (1 mL), synthesized from 1 mM and 5.6 mM LA, was irradiated with a LED source centered at 368 nm. The PALA in this reaction included the high molecular weight fraction, in spite of its limited solubility. Figure 7 shows the spectra obtained after 3 min of UVA exposure. With this approach, i.e., by using PALA as an in situ stabilizer in one-pot reaction, the onset of formation of nanoparticles was confirmed from the change in solution color from light-yellow to purple and dark reddish (inset, Fig. 7), since the substituted tetralone only competes for absorption of UVA photons when LA is present, but when PALA is the stabilizing agent, the substituted tetralone is the only light absorber.

AuNP spectra obtained after irradiation of sample containing 10 mM tetralone, 1 mM HAuCl4, with PALA (1 mL) synthesized upon UVA exposure of 1 mM and 5.6 mM LA for 20 h. Inset: solutions of tetralone, HAuCl4 and PALA synthesized after UVA irradiation of LA 5.6 mM (left) and 1 mM (right)

We observe that when PALA was synthesized after irradiating 5.6 mM LA with UVA light for 20 h, the population of smaller AuNP of ~ 5–10 nm is higher, but structures of ~ 40 nm can also be formed (Fig. S10, †ESI). On the other hand, when using PALA obtained upon exposure of LA 1 mM, the shape is also better well defined and particles of around ~ 20–40 nm are obtained (Fig. S10, †ESI). This suggests that PALA can act as a stabilizing agent, capable of controlling size and agglomeration, passing from a nanoflower-like shape (when only the substituted tetralone and HAuCl4 are part of the reaction) to a plasmonic nanostructures with a more regular shape as we can see in Fig. S9 (†ESI). Likewise, amounts lower than 1 mL of 1 mM and 5.6 mM PALA solutions were used and amounts of 200 μL were tested, and as expected, a reduction in the agglomeration of the nanostructures was also observed with a maximum of 550 nm in the plasmon peak, as shown by Fig. 8 and different sizes of the nanostructures can also be obtained.

AuNP spectra obtained after 2–3 min irradiation of 1 mL of 10 mM tetralone, 1 mL of 1 mM HAuCl4 and PALA (200 μL) synthesized upon UVA exposure. Light scattering contributes to a vertical shift of the spectra. Inset: A TEM image from reaction containing PALA derived from 20 h irradiation of 1 mM LA solution and B 5.6 mM LA

Thus, amounts of about 200 μL–1 mL from a solution of PALA derived upon exposure of LA 1 mM and 5.6 mM for 20 h into the system that contains 10 mM substituted tetralone (1 mL) and 1 mM HAuCl3 (1 mL) in 1:1 water:acetonitrile, work to reduce the size of Au nanostructures, if the polymer is used in situ, as in this case only the substituted tetralone will play the role of the photo-initiator. Additionally, polymers such as PALA, can reduce the agglomeration of particles, especially when synthesized after irradiating LA in concentrations higher than 1 mM. This relies on the fact that when increasing the steric hindrance of capping agents, the size of the nanoparticles becomes smaller. Moreover, enhancement in polymer to metal molar ratio from 1:1 to 1:5 results in smaller nanostructures.