Catalytic inverse vulcanization

The discovery of inverse vulcanization has allowed stable polymers to be made from elemental sulfur, an unwanted by-product of the petrochemicals industry. However, further development of both the chemistry and applications is handicapped by the restricted choice of cross-linkers and the elevated temperatures required for polymerisation. Here we report the catalysis of inverse vulcanization reactions. This catalytic method is effective for a wide range of crosslinkers reduces the required reaction temperature and reaction time, prevents harmful H2S production, increases yield, improves properties, and allows crosslinkers that would be otherwise unreactive to be used. Thus, inverse vulcanization becomes more widely applicable, efficient, eco-friendly and productive than the previous routes, not only broadening the fundamental chemistry itself, but also opening the door for the industrialization and broad application of these fascinating materials.

were commercially available and used as received without any further purification. Iron diethyldithiocaebamate and cobalt diethyldithiocaebamate were both synthesized from sodium diethyldithiocaebamate following a method reported in the literature. 1

Instrumentation used for characterization:
Gel permeation chromatography (GPC): The molecular weight of the soluble fraction of the polymers was determined by gel permeation chromatography (GPC) using a Viscotek system comprising a GPCmax (degasser, eluent and sample delivery system), and a TDA302 detector array, using THF as eluent.
Powder X-ray Diffraction (PXRD): Data was measured using a PANalytical X'Pert PRO diffractometer with Cu-K α1+2 radiation, operating in transmission geometry.
Differential Scanning Calorimetry (DSC) were performed on a TA Instruments Q200 DSC, under nitrogen flow, and with heating and cooling rates of 5 °C/min. Thermogravimetric analysis (TGA) samples were heated under nitrogen to 800 °C at a heating rate of 20 °C min -1 using a TA Instruments Q500.
Fourier-transform infrared spectroscopy (FT-IR) was performed using a Thermo NICOLET IR200, between 400 cm -1 to 4000 cm -1 . Samples were loaded either neat, using an attenuated total reflectance accessory, or in transmission after pressing into a KBr pellet.
Solution NMR was recorded in deuterated chloroform using a Bruker Advance DRX (400 MHz) spectrometer. 13 C magic-angle spinning (MAS) NMR spectra were performed on a Bruker Avance III operating at a 1 H Larmor frequency of 700 MHz, using a Bruker 4mm HX probe.
Chemical shifts were referenced using the CH 3 resonance of solid alanine at 20.5 ppm ( 13 C). A chemical shielding reference of 189.7 ppm was used, determined from a separate calculation on an optimized tetramethylsilane molecule.

Experimental procedures:
A note on the reproducibility and sensitivity of inverse vulcanization reactions: We have noted that these reactions are particularly susceptible to changes in apparatus and conditions. This results from the nature of the reactions, being driven by radical initiation, and exothermic polymerization. Care must be taken not to let the temperature of the heating medium (we use metal heating blocks on hot plate stirrers) to 'overshoot' in temperature, as this will affect the reaction time. The sulfur polymerization mixture is also quite a poor thermal conductor, and if care is not taken, the exothermic reaction causes the internal temperature to increase above the intended temperature. The reaction will therefore proceed differently depending on the heat transfer away from the reaction. As a result, using different sizes and shapes of heating blocks, different glassware volumes, stirrer speeds, stirrer geometries etc. will all affect this process. For this reason, we took great care to keep all of these variables constant in these reactions.
General procedure for the catalysts discovery and screening: preparation of poly (sulfur-random-(ethylene glycol dimethacrylate)) (Poly(S-r-EGDMA)) To a 40 mL glass reaction vial equipped with a magnetic stir bar was added 5 g (19.5 mmol) of elemental sulfur, catalysts (masses detailed below) and heated until molten by placing the vial in a metal heating block set to 135 °C. The melting point of sulfur is ~120 °C. The reactions were stirred at 200 RPM using cross shaped magnetic stirrer bars. When the sulfur was molten, 5 g (25.2 mmol) of Ethylene glycol dimethacrylate (EGDMA) cross-linker was added. The stirring rate was then increased to 900 RPM, and the reaction continued for up to 10 hours. Samples that were observed to react to form a homogeneous molten state (does not separate if removed on a spatula and cooled to room temperature), were then removed from stirring and cured in an oven at 140 °C for 10 hours further. Samples that showed no sign of reaction, and that were still two phases after 10 hours were aborted.
Preparation of Poly(S-r-EGDMA) with 1 w% of ZnO as catalyst: The copolymerization was carried out by the following the general method mentioned above with ZnO (100 mg, 1 w% loading, 1.22 mmol) to afford two layers of mixture with yellow solid at bottom and a clear liquid on the top (yield: 9.7 g). Elemental Analysis for (

Preparation of Poly(S-r-EGDMA) with 1 w% of Zinc as catalyst:
The copolymerization was carried out by the following the general method mentioned above with Zinc (100 mg, 1w% loading, 1.53 mmol) to afford two layers of mixture with yellow solid at bottom and a clear liquid on the top (yield: 9.3 g). PXRD and DSC confirmed the presence of unreacted sulfur.

Preparation of Poly(S-r-EGDMA) with 1 w% of Zinc Chloride as catalyst:
The copolymerization was carried out by the following the general method mentioned above with ZnCl 2 (100 mg, 1w% loading, 0.736 mmol) to afford two layers with gray-brown solid at the bottom and a clear liquid above (yield: 9.5 g). PXRD and DSC confirmed the presence of unreacted sulfur.

Preparation of Poly(S-r-EGDMA) with 1 w% of iron Chloride as catalyst:
The copolymerization was carried out by the following the general method mentioned above with FeCl 2 (100 mg, 1w% loading, 0.787 mmol) to afford two layers with brown-red solid at the bottom and clear liquid on above (yield: 9.3 g). PXRD and DSC confirmed the presence of unreacted sulfur.

Preparation of Poly(S-r-EGDMA) with 1 w% of Copper oxide as catalyst:
The copolymerization was carried out by the following the general method mentioned above with CuO (100 mg, 1w% loading, 1.26 mmol) to afford two layers of mixture with brown solid at bottom and liquid on the top (yield: 9.1 g). PXRD and DSC confirmed the presence of unreacted sulfur.

Preparation of Poly(S-r-EGDMA) with 1 w% of Copper Chloride as catalyst:
The copolymerization was carried out by the following the general method mentioned above with CuCl 2 (100 mg, 1w% loading, 0.743 mmol) to afford two layers with a brown-green solid at the bottom and a clear liquid above (yield: 9.3 g). PXRD and DSC confirmed the presence of unreacted sulfur.

Preparation of Poly(S-r-EGDMA) with 1 w% of Zinc Stearate (Zn-STR) as catalyst:
The copolymerization was carried out by the following the general method mentioned above with Zn-STR (100 mg, 1w% loading, 0.743 mmol) to afford an orange -red solution that cooled to a solid (yield: 9.3 g). Elemental Analysis for However, all reactions were left in the heating blocks for at least 24 hours before being removed and allowed to cool. The reactions were monitored for the first hour, and then in half hour intervals for the first 12 hours, then checked again after 24 hours (hence why the reaction time is listed as 12-24 hours for some reactions). After cooling samples were recovered by breaking the vials. All of these reactions were performed in triplicate to ensure the timings were consistent, and allow DSC to be performed on three separate reactions. All crosslinkers were prepared according to the above method, except for limonene, which is known to produce low molecular weight byproducts including cymene, where the reaction was performed under vacuum distillation as reported by Chalker et al. 2 .

Synthesis of moulded objects: Moulded objects (such as the 'robots' in
Supplementary figure 2) were produced using a modified version of the synthesis above. The initial phase of the reaction was the same, but when the reaction mixture had formed a 'pre-polymer' state (increase in viscosity, darkened colour, single phase, no separation on cooling) it was transferred from the glass vial into a silicone mould and placed in an over at 140 °C for a further 12 hours to cure.

TVTCSi/DCPD) with 1 w% of Zinc diethyldithiocarbamate as catalyst:
The The reactions were stirred vigorously using magnetic stirrer bars for each crosslinkers until each reaction produced no more gas, typically under an hour. Another experiment was performed in the same manner with limonene, at 1 wt.% ZnD 2 loading, but at 180 °C, the same temperature as the uncatalysed reaction.
Observations: The reaction of sulfur and limonene without catalyst generated 63 mL H 2 S gas, while with 1 wt.% catalyst, the same reaction only generated 10 mL of the H 2 S gas. The former reaction was carried out at 180 °C for 15 min and the latter at 135 °C for 50 min. Similar observations were reached for DCPD, for which the reaction was carried out at 170 °C in the absence of catalysts for 27 min, the reaction produced 26.5 mL of gas. The same reaction with 1 wt.% catalyst was carried out at 135 °C for 45 min, the amount of gas generated was only 3.5 mL. The generation of H 2 S gas is dependent on the presence of α-proton of allyl groups and related to the reaction temperature. Therefore, with those cross-linkers without α-proton of allyl groups (e.g. DVB, DIB), lower amounts of gas were released, and only at the beginning of the reaction, there was not any gas generated after the first 3-5 min, especially for the reactions with catalysts. ENB is more reactive than DCPD, and as a result requires lower temperatures, producing less H 2 S. In the cases only small amounts of gas were produced (few mL from 10 g reactions), it is possible this recorded volume is at least partly the result of the desorption of gases and moisture dissolved in the reactants, because of heating. In the samples were larger amounts of gas were produced (DCPD and limonene without catalyst), the production of H 2 S was confirmed by exposing a H 2 S detector to the gas produced. Kinetic and capacity studies using mercury solution: Before conducting capacity tests for the sorbent, a kinetics study was preformed to determine the optimal time to leave the samples to fully adsorb the mercury from an aqueous solution. The kinetics study was performed by placing 240 mg of 5% catalysed S-Lim coated silica in to vials with 12 mL of 1000 ppm mercury chloride aqueous solution. Once capped the vials were placed on a tube roller for agitation before being removed at regular intervals.

Heavy Metal Remediation
Once a sample was removed, a small aliquot was removed, diluted by a factor 1:100 and analysed by ICP-OES. Mercury uptake was calculated by difference when compared to a control sample which did not contain the 5% S-Lim coated silica. A capacity study was conducted by sampling different mass of 5% S-Lim coated silica (120, or 240 mg) in different mercury chloride solution strengths (125 -2000 ppm). A peak capacity for the material was calculated at 65.250 mg of mercury per gram of sorbent used. However the majority of the sorbent mass is the fumed silica support, suggesting that the capacity of the polymer is likely to be x10 greater than the capacity for the whole polymer coated silica. To ensure that the silica itself played no part in the mercury removal process a control of uncoated fumed silica was added to 12 mL of 1000 ppm mercury solution and agitated for 16 hours on a tube roller as well.
Negligible mercury capture was noted (<3%).   There is then a slight, though reduced increase when the catalyst loading is increased to 5 %. There is then a slight, though reduced increase when the catalyst loading is increased to 5 %. ratio of sulfur to crosslinker, without catalyst, at 135 °C, and in triplicate. Standard deviation is given for the average of three parallel reactions. There is a slight increase in glass transition temperature for the 1 % catalyzed sample in comparison to the uncatalyzed sample. There is then a significant reduction when the catalyst loading is increased to 5 %, presumably as this is then effectively an excess loading of catalyst, beyond which the catalyst is not providing additional benefit to the reaction, but is plasticising the material, as a small molecular additive.

Affinity for other metals
It should be noted that the T g s for the 0 wt.% loaded samples were very faint, however the 1 and 5 wt.% catalyst loaded samples were more pronounced.
Supplementary figure 32. Glass transition temperatures (from DSC) as a function of catalyst loading for the polymerization of sulfur with GBDA. The reaction was carried out at a 1:1 mass ratio of sulfur to crosslinker, without catalyst, at 135 °C, and in triplicate. Standard deviation is given for the average of three parallel reactions. Only very faint transitions were detected, which is likely the cause of the poor accuracy in measurement. There appears to be a slight decrease in glass transition temperature moving from the zero to 1% loaded sample, and then an increase in the 5% loaded sample, however, in view of the poor reproducibility, this may not be significant.

Supplementary figure 33.
Glass transition temperatures (from DSC) as a function of catalyst loading for the polymerization of sulfur with limonene. The reaction was carried out at a 1:1 mass ratio of sulfur to crosslinker, according to the reported procedure. There appears to be a slight decrease in glass transition temperature moving from the zero to 1% loaded sample, and then an increase in the 5% loaded sample.