Copolymerization of a Bisphenol a Derivative and Elemental Sulfur by the RASP Process

: Fossil fuel reﬁning produces over 70 Mt of excess sulfur annually from for which there is currently no practical use. Recently, methods to convert waste sulfur to recyclable and biodegradable polymers have been delineated. In this report, a commercial bisphenol A (BPA) derivative, 2,2 (cid:48) ,5,5 (cid:48) -tetrabromo(bisphenol A) ( Br 4 BPA ), is explored as a potential organic monomer for copolymerization with elemental sulfur by RASP (radical-induced aryl halide-sulfur polymerization). Resultant copolymers, BAS x ( x = wt% sulfur in the monomer feed, screened for values of 80, 85, 90, and 95) were characterized by thermogravimetric analysis, di ﬀ erential scanning calorimetry, and dynamic mechanical analysis. Analysis of early stage reaction products and depolymerization products support proposed S–C aryl bond formation and regiochemistry, while fractionation of BAS x reveals a sulfur rank of 3–6. Copolymers having less organic cross-linker (5 or 10 wt%) in the monomer feed were thermoplastics, whereas thermosets were accomplished when 15 or 20 wt% of organic cross-linker was used. The ﬂexural strengths of the thermally processable samples ( > 3.4 MPa and > 4.7 for BAS 95 and BAS 90 , respectively) were quite high compared to those of familiar building materials such as portland cement (3.7 MPa). Furthermore, copolymer BAS 90 proved quite resistant to degradation by oxidizing organic acid, maintaining its full ﬂexural strength after soaking in 0.5 M H 2 SO 4 for 24 h. BAS 90 could also be remelted and recast into shapes over many cycles without any loss of mechanical strength. This study on the e ﬀ ect of monomer ratio on properties of materials prepared by RASP of small molecular aryl halides conﬁrms that highly cross-linked materials with varying physical and mechanical properties can be accessed by this protocol. This work is also an important step towards potentially upcycling BPA from plastic degradation and sulfur from fossil fuel reﬁning.

Radical-induced aryl halide-sulfur polymerization (RASP, Scheme 1B) has recently emerged as another way to produce HSMs [47][48][49]. RASP employs an aryl halide as the comonomer with elemental sulfur, thus expanding the scope of potential monomers for HSM preparation beyond the olefins required for inverse vulcanization. Although 1,4-diiodobenzene [47], chlorolignin [48], and 2,4-dimethyl-3,5-dichlorophenol have been used in the RASP process, the extent to which the ratio of organic small molecules to sulfur influences the properties of the resultant HSMs has yet to be fully elucidated.
Herein, RASP of sulfur and a commercial monomer, 2,2′,4,4′-tetrabromo(bisphenol A) (Br4BPA), is undertaken to produce BASx (where x = wt% sulfur in the monomer feed). BASx copolymers were prepared having four different monomer ratios, ranging from 80 to 95 wt% sulfur in the monomer feed. Characterizing these copolymers provides (1) insight into the influence of monomer ratio on properties in HSMs produced by RASP and (2) the first HSMs produced by RASP wherein there is no possibility for benzylic cross-linking to supplement cross-linking due to direct S-Caryl bondforming reactions. BASx materials range from thermosets to thermoplastics, and all of them form materials wherein polymeric sulfur is stabilized by its incorporation into a cross-linked network. The influence of monomer feed ratio on the processability, composition, thermal properties, and mechanical properties was evaluated, and the recyclability of remeltable materials was validated.

Synthesis and Structure
RASP proceeds by the Macallum mechanism whereby thermal reaction of aryl halides forms aryl radicals (Scheme 2A), generally at temperatures of ≥220 °C [48]. In the presence of excess sulfur, aryl radicals and halogen radicals rapidly react with elemental sulfur, leading to S-Caryl bond-

Synthesis and Structure
RASP proceeds by the Macallum mechanism whereby thermal reaction of aryl halides forms aryl radicals (Scheme 2A), generally at temperatures of ≥220 • C [48]. In the presence of excess sulfur, aryl radicals and halogen radicals rapidly react with elemental sulfur, leading to S-C aryl bond-formation and loss of the halogen as S 2 X 2 (Scheme 2). The HSMs initially formed by this reaction contain oligomeric or polymeric sulfur chains that link aryl rings together. Such extended sulfur catenates are unstable and under ordinary conditions will revert back to an S 8 ring allotrope, so the end product of the Macallum polymerization is a poly(aryl sulfide) in which a single sulfur atom bridges aryl units. If a highly cross-linked network is present, however, the polymeric sulfur domains can be stabilized indefinitely. Stabilizing the polymeric sulfur domains is important for accessing materials with high mechanical strength, so establishing a highly cross-linked structure is a primary objective in selecting the organic monomer for RASP. formation and loss of the halogen as S2X2 (Scheme 2). The HSMs initially formed by this reaction contain oligomeric or polymeric sulfur chains that link aryl rings together. Such extended sulfur catenates are unstable and under ordinary conditions will revert back to an S8 ring allotrope, so the end product of the Macallum polymerization is a poly(aryl sulfide) in which a single sulfur atom bridges aryl units. If a highly cross-linked network is present, however, the polymeric sulfur domains can be stabilized indefinitely. Stabilizing the polymeric sulfur domains is important for accessing materials with high mechanical strength, so establishing a highly cross-linked structure is a primary objective in selecting the organic monomer for RASP.
The mechanism for S-Caryl bond formation in Macallum polymerization/RASP ((A), X = Cl, Br or I) relies on thermal decomposition of sulfur and aryl halides, eventually leading to the net reaction (B).
In some prior demonstrations of RASP, a pre-existing polymeric network was supplied in the form of lignin [48,49], or a network comprised by a combination of S-Cbenzylic and S-Caryl bonds was prepared. For the current work, an affordable aryl halide monomer that could form a highly crosslinked network comprised entirely from S-Caryl bond-forming reactions (lacking the possibility for S-Cbenzylic cross links) was desired. A commercially available bisphenol A (BPA) derivative, 2,2′,5,5′tetrabromo(bisphenol A) (Br4BPA), was selected for this purpose. Br4BPA should support a high cross-link density by merit of its ability to form up to four S-Caryl cross-links per molecule, whereas the organic BPA subunit is a well-established and thermally stable structural element of durable polymers. Furthermore, the thermal reactivity of Br4BPA under RASP conditions, a prerequisite for S-Caryl bond formation as delineated in Scheme 2, has already been established [50].
As indicated in Scheme 2, the halogens are initially removed as S2X2 (X = Br in the current case), which are toxic and corrosive compound, so caution should be exercised when performing the RASP reaction. After reaction in the pressure tube, the tube should be fully cooled to room temperature before opening it in a fume hood, where S2Br2 in the product can be readily hydrolyzed by controlled reaction with water or slow reaction with humidity in the air. These RASP reactions can also be performed in a Schlenk flask vented to a water bath into which S2Br2 can be distilled as it forms, where it will be hydrolyzed to HX. Both the Schlenk flask and pressure tube methods prove effective for preparing the HSMs. Should the reaction be scaled up in the laboratory, trapping S2Br2 by its hydrolysis in a water bath as it forms could be an appropriate measure to prevent its release into the atmosphere. In the current context, at least some of the S2Br2 generated during RASP could also react with phenolic sites as well [51].
Another goal of the current research was to study the extent to which the monomer ratio influences thermal and mechanical properties of the resulting HSMs, something that has not been studied for RASP of small molecular aryl halides. Elemental sulfur and Br4BPA were subjected to RASP by heating at 240 °C in a pressure tube. Four different sulfur monomer feeds, ranging from 80 to 95 wt%, were employed to yield four copolymers BASx (where x = wt% sulfur in the monomer feed, Scheme 3). All of the materials have a dark brown-orange color characteristic of polymeric sulfur ( Figure 1). Copolymers having ≥90 wt% sulfur in the monomer feed were remeltable, whereas BAS85 In some prior demonstrations of RASP, a pre-existing polymeric network was supplied in the form of lignin [48,49], or a network comprised by a combination of S-C benzylic and S-C aryl bonds was prepared. For the current work, an affordable aryl halide monomer that could form a highly cross-linked network comprised entirely from S-C aryl bond-forming reactions (lacking the possibility for S-C benzylic cross links) was desired. A commercially available bisphenol A (BPA) derivative, 2,2 ,5,5 -tetrabromo(bisphenol A) (Br 4 BPA), was selected for this purpose. Br 4 BPA should support a high cross-link density by merit of its ability to form up to four S-C aryl cross-links per molecule, whereas the organic BPA subunit is a well-established and thermally stable structural element of durable polymers. Furthermore, the thermal reactivity of Br 4 BPA under RASP conditions, a prerequisite for S-C aryl bond formation as delineated in Scheme 2, has already been established [50].
As indicated in Scheme 2, the halogens are initially removed as S 2 X 2 (X = Br in the current case), which are toxic and corrosive compound, so caution should be exercised when performing the RASP reaction. After reaction in the pressure tube, the tube should be fully cooled to room temperature before opening it in a fume hood, where S 2 Br 2 in the product can be readily hydrolyzed by controlled reaction with water or slow reaction with humidity in the air. These RASP reactions can also be performed in a Schlenk flask vented to a water bath into which S 2 Br 2 can be distilled as it forms, where it will be hydrolyzed to HX. Both the Schlenk flask and pressure tube methods prove effective for preparing the HSMs. Should the reaction be scaled up in the laboratory, trapping S 2 Br 2 by its hydrolysis in a water bath as it forms could be an appropriate measure to prevent its release into the atmosphere. In the current context, at least some of the S 2 Br 2 generated during RASP could also react with phenolic sites as well [51].
Another goal of the current research was to study the extent to which the monomer ratio influences thermal and mechanical properties of the resulting HSMs, something that has not been studied for RASP of small molecular aryl halides. Elemental sulfur and Br 4 BPA were subjected to RASP by heating at 240 • C in a pressure tube. Four different sulfur monomer feeds, ranging from 80 to 95 wt%, were employed to yield four copolymers BAS x (where x = wt% sulfur in the monomer feed, Scheme 3). All of the materials have a dark brown-orange color characteristic of polymeric sulfur ( Figure 1). Copolymers having ≥90 wt% sulfur in the monomer feed were remeltable, whereas BAS 85 and BAS 80 were thermosets. Although highly cross-linked polymers with structures comprising primarily C-C bonds are not generally remeltable in this way, the thermal processability of BAS 90 and BAS 95 is possible by merit of the thermal reversibility of S-S bond formation.
Sus. Chem. 2020, 1, FOR PEER REVIEW 4 of 15 and BAS80 were thermosets. Although highly cross-linked polymers with structures comprising primarily C-C bonds are not generally remeltable in this way, the thermal processability of BAS90 and BAS95 is possible by merit of the thermal reversibility of S-S bond formation. The formulation of the BASx microstructure as being assembled via S-Caryl bond formation with the regiochemistry shown in Scheme 3 and in which the BPA backbone is intact is further supported by (1) analysis of depolymerization products and (2) reaction of Br4BPA with smaller quantities of sulfur to give soluble products that could be analyzed by 1 H NMR spectrometry.
Quantification of S-Caryl bond formation was sought by analysis of depolymerization products of BASx. Depolymerization was affected by the reaction of BASx with AlBr3 in toluene, a process that cleaves S-S bonds and S-Calkyl bonds while leaving S-Caryl bonds intact (Scheme 4A) [52,53]. After depolymerization and workup, one aryl thiol is produced for each C in a -C-Sn-C-unit for which n ≥ 2. Any carbon atoms bridged by a single S atom will not be decomposed to thiols by this process. By quantifying the number of resulting thiols, the extent to which oligomeric/polymeric sulfur chains bridge carbon sites in the composite can thus be quantified. Quantification of thiols following depolymerization was undertaken by their reaction with Ellman's reagent, which produces a selective colorimetric response to thiols that can be monitored by absorbance of the resultant thiolate at ~412 nm in the UV-VIS spectrum (Scheme 4B). The absorbance increase induced by exposure of Ellman's reagent to the depolymerization products allowed quantification of aryl thiols in the sample and BAS80 were thermosets. Although highly cross-linked polymers with structures comprising primarily C-C bonds are not generally remeltable in this way, the thermal processability of BAS90 and BAS95 is possible by merit of the thermal reversibility of S-S bond formation. The formulation of the BASx microstructure as being assembled via S-Caryl bond formation with the regiochemistry shown in Scheme 3 and in which the BPA backbone is intact is further supported by (1) analysis of depolymerization products and (2) reaction of Br4BPA with smaller quantities of sulfur to give soluble products that could be analyzed by 1 H NMR spectrometry.
Quantification of S-Caryl bond formation was sought by analysis of depolymerization products of BASx. Depolymerization was affected by the reaction of BASx with AlBr3 in toluene, a process that cleaves S-S bonds and S-Calkyl bonds while leaving S-Caryl bonds intact (Scheme 4A) [52,53]. After depolymerization and workup, one aryl thiol is produced for each C in a -C-Sn-C-unit for which n ≥ 2. Any carbon atoms bridged by a single S atom will not be decomposed to thiols by this process. By quantifying the number of resulting thiols, the extent to which oligomeric/polymeric sulfur chains bridge carbon sites in the composite can thus be quantified. Quantification of thiols following depolymerization was undertaken by their reaction with Ellman's reagent, which produces a selective colorimetric response to thiols that can be monitored by absorbance of the resultant thiolate at ~412 nm in the UV-VIS spectrum (Scheme 4B). The absorbance increase induced by exposure of Ellman's reagent to the depolymerization products allowed quantification of aryl thiols in the sample The formulation of the BAS x microstructure as being assembled via S-C aryl bond formation with the regiochemistry shown in Scheme 3 and in which the BPA backbone is intact is further supported by (1) analysis of depolymerization products and (2) reaction of Br 4 BPA with smaller quantities of sulfur to give soluble products that could be analyzed by 1 H NMR spectrometry.
Quantification of S-C aryl bond formation was sought by analysis of depolymerization products of BAS x . Depolymerization was affected by the reaction of BAS x with AlBr 3 in toluene, a process that cleaves S-S bonds and S-C alkyl bonds while leaving S-C aryl bonds intact (Scheme 4A) [52,53]. After depolymerization and workup, one aryl thiol is produced for each C in a -C-S n -C-unit for which n ≥ 2. Any carbon atoms bridged by a single S atom will not be decomposed to thiols by this process. By quantifying the number of resulting thiols, the extent to which oligomeric/polymeric sulfur chains bridge carbon sites in the composite can thus be quantified. Quantification of thiols following depolymerization was undertaken by their reaction with Ellman's reagent, which produces a selective colorimetric response to thiols that can be monitored by absorbance of the resultant thiolate at~412 nm in the UV-VIS spectrum (Scheme 4B). The absorbance increase induced by exposure of Ellman's reagent to the depolymerization products allowed quantification of aryl thiols in the sample corresponding to 90% ± 10% of the available bromide sites in the Br 4 BPA staring material. Elemental microanalysis for Br likewise revealed that 85-100% of the Br was removed following RASP.  Next, the regiochemistry of S-Caryl bond formation was assessed by reacting Br4BPA with 0.75 equivalent to S8 under RASP conditions. If the RASP mechanism mirrors the Macallum mechanism in Scheme 2 initial formation of thiol 1 (Scheme 5A) would be expected after workup. Proton NMR spectrometric analysis of the product mixture (Scheme 5B) reveals that under these conditions, primarily one new aromatic resonance emerges at 7.74 ppm. Most of the starting material remains unreacted, with an aromatic resonance at 7.43 ppm. Compound 2 (Scheme 5C) appears to be the most similar model compound in the literature for which NMR data have been reported [54]. In this compound, a resonance at 7.66 ppm is attributable to proton Hc′, analogous to proton Hc in 1 (at 7.74 ppm) that gives rise to the most prominent new signal in Scheme 5B. The resonance for Ha is likely coincident with Hc and the aryl protons on unreacted Br4BPA because the influence of a thiol group and bromo group on a m-aryl proton are nearly identical. The resonance for the thiol proton, Hd, is observed at 5.64 ppm and its integration confirms its presence in the expected 1:1 ratio with Hc. Broadened resonances centered at around 7.2 and 8.6 ppm are attributed to the -OH protons He and Hf, respectively. Although some minor resonances in the aromatic region were observed at about 7.55 and 7.40 ppm (and possibly some overlapping the broad peak at 7.2 ppm), the predominant new species observed is consistent with the regiochemistry of S-Caryl bond formation shown for BASx in Scheme 3. Further evidence that phenol functionalities are intact in the copolymers was provided by IR spectra, which reveal characteristic peaks at 3400 cm −1 (O-H stretch) and 1153 cm −1 (C−O stretch). Next, the regiochemistry of S-C aryl bond formation was assessed by reacting Br 4 BPA with 0.75 equivalent to S 8 under RASP conditions. If the RASP mechanism mirrors the Macallum mechanism in Scheme 2 initial formation of thiol 1 (Scheme 5A) would be expected after workup. Proton NMR spectrometric analysis of the product mixture (Scheme 5B) reveals that under these conditions, primarily one new aromatic resonance emerges at 7.74 ppm. Most of the starting material remains unreacted, with an aromatic resonance at 7.43 ppm. Compound 2 (Scheme 5C) appears to be the most similar model compound in the literature for which NMR data have been reported [54]. In this compound, a resonance at 7.66 ppm is attributable to proton H c , analogous to proton H c in 1 (at 7.74 ppm) that gives rise to the most prominent new signal in Scheme 5B. The resonance for H a is likely coincident with H c and the aryl protons on unreacted Br 4 BPA because the influence of a thiol group and bromo group on a m-aryl proton are nearly identical. The resonance for the thiol proton, H d , is observed at 5.64 ppm and its integration confirms its presence in the expected 1:1 ratio with H c . Broadened resonances centered at around 7.2 and 8.6 ppm are attributed to the -OH protons H e and H f , respectively. Although some minor resonances in the aromatic region were observed at about 7.55 and 7.40 ppm (and possibly some overlapping the broad peak at 7.2 ppm), the predominant new species observed is consistent with the regiochemistry of S-C aryl bond formation shown for BAS x in Scheme 3. Further evidence that phenol functionalities are intact in the copolymers was provided by IR spectra, which reveal characteristic peaks at 3400 cm −1 (O-H stretch) and 1153 cm −1 (C−O stretch). The HSMs previously prepared by RASP contained some sulfur that was not covalently incorporated into the cross-linked network. Sulfur that is not covalently incorporated is readily extractable into CS2, whereas highly cross-linked polymers are generally insoluble [8,28,55]. This provides a convenient way to fractionate such materials. BASx samples were thus fractionated in CS2, and the relative masses of soluble and insoluble fractions were determined as shown in Table 1.
Elemental analysis of the CS2-soluble fractions of BASx confirms that they consist of >98.5% sulfur. With the aforementioned data in hand, the sulfur rank (average number of sulfur atoms per cross-linking chain) can be calculated as follows. The number of carbon atoms cross-linked by sulfur (Cxlink) is known by subtracting the remaining bromine (from elemental analysis, Brec) from the initial amount of bromine in the reaction feed (Bri). Fractionation data allowed for the calculation of the number of moles of sulfur atoms covalently incorporated (Scov) into the materials by subtracting the free sulfur removed in the CS2 wash (Sfr) from the total sulfur content (Sadd). From these data, the sulfur rank (average number of sulfur atoms in each cross-linking chain) was calculated using Equation (1) (the factor of 2 is necessary because each sulfur chain must link to two carbon atoms to form a cross-link).
The sulfur ranks in BASx (Table 1) are similar to those found in materials prepared by RASP utilizing a dichloroxylenol derivative (sulfur rank = 5) [49] but lower than in materials prepared by RASP of a lignin derivative (12-31 sulfur atoms). The lower sulfur rank in BASx is expected on the The HSMs previously prepared by RASP contained some sulfur that was not covalently incorporated into the cross-linked network. Sulfur that is not covalently incorporated is readily extractable into CS 2 , whereas highly cross-linked polymers are generally insoluble [8,28,55]. This provides a convenient way to fractionate such materials. BAS x samples were thus fractionated in CS 2 , and the relative masses of soluble and insoluble fractions were determined as shown in Table 1.
Elemental analysis of the CS 2 -soluble fractions of BAS x confirms that they consist of >98.5% sulfur. With the aforementioned data in hand, the sulfur rank (average number of sulfur atoms per cross-linking chain) can be calculated as follows. The number of carbon atoms cross-linked by sulfur (C xlink ) is known by subtracting the remaining bromine (from elemental analysis, Br ec ) from the initial amount of bromine in the reaction feed (Br i ). Fractionation data allowed for the calculation of the number of moles of sulfur atoms covalently incorporated (S cov ) into the materials by subtracting the free sulfur removed in the CS 2 wash (S fr ) from the total sulfur content (S add ). From these data, the sulfur rank (average number of sulfur atoms in each cross-linking chain) was calculated using Equation (1) (the factor of 2 is necessary because each sulfur chain must link to two carbon atoms to form a cross-link).

Sul f ur Rank
The sulfur ranks in BAS x (Table 1) are similar to those found in materials prepared by RASP utilizing a dichloroxylenol derivative (sulfur rank = 5) [49] but lower than in materials prepared by RASP of a lignin derivative (12-31 sulfur atoms). The lower sulfur rank in BAS x is expected on the basis of the high number of available cross-link sites. Similar sulfur ranks can also be obtained in HSMs prepared by inverse vulcanization. For example, inverse vulcanization HSMs wherein the olefin was a polystyrene derivative [8] or a tyrosine derivative [55] both have sulfur ranks of five. The number of available cross-link sites per sulfur in the copolymers appears to dictate the sulfur rank regardless of whether inverse vulcanization or RASP is used to prepare the material. Although the fractionation studies confirm the presence of sulfur that is not covalently bound to the cross-linked material, no phase separation was observed by scanning electron microscopy (SEM) with element mapping by energy-dispersive X-ray analysis (EDX), which confirmed uniform distribution of sulfur, carbon, and oxygen with almost no Br detectable ( Figures S1 and S2 in the Electronic Supplementary Information).

Thermal and Morphological Properties
HSMs can exhibit complex thermal behavior. Indeed, the S-S bonds in polymeric sulfur domains comprising important structural elements in HSMs undergo bond breakage at temperatures as low as 90 • C. This S-S metathesis behavior has been exploited in innovative approaches to lower temperature processing of HSMs [56][57][58][59]. The labile nature of the S-S bonds allows for a variety of interesting thermal characteristics. Differential scanning calorimetry (DSC, Table 2 and Figures S3-S6 in the Electronic Supplementary Information) analysis of BAS x revealed a number of notable morphological changes over the range of −60 to + 140 • C. The typical melt peak for sulfur domains of the materials was observed at 114-116 • C, slightly depressed from the melting point of pure orthorhombic sulfur. The presence of polymeric sulfur for all of the materials was also evident from the observation of the characteristic glass transition (T g ) at between −31 and −36 • C, consistent with the characteristic polymeric sulfur color of the materials (Figure 1). Determination of crystallization and melting enthalpies from the DSC data also allowed the quantification of the percent crystallinity of the composites. Crystallinity decreases predictably as more of the organic cross-linking unit is incorporated, to the extent that the crystallinity of BAS 95 is only 40% that of sulfur despite being composed of 95 wt% sulfur. A primarily amorphous polymer is attained with 20 wt% of organic cross-linker in the monomer feed, as reflected in the 8% crystallinity of BAS 80 . The chemical instability of the polymer covalent bonds at the polymerization temperature (for either inverse vulcanization or RASP) would seem to be in conflict with the fundamental idea that polymerization must be carried out at a temperature below the decomposition point of the product. It is the dynamic nature of the S-S bonds that allows the higher temperature inverse vulcanization and RASP processes to be exploited to prepare HSMs. Thermogravimetric analysis (TGA) of BAS x and of the BAS 90 fraction from which free sulfur had been removed provide some additional insight into the thermal behavior of the composites.
Thermogravimetric analysis of BAS x materials revealed onset of mass loss at between 222 and 240 • C (T d in Table 2, traces provided in Figure S7 of the Electronic Supplementary Information), which is below the RASP reaction temperature. Thermal mass loss was attributed primarily to sulfur sublimation as would be expected of materials composed of 80-95 wt% sulfur heated in an open TGA pan. When a bulk sample of BAS 90 is heated to 240 • C in a flask fitted with cold fingers, e.g., about 70 wt% of the material is collectable as sublimed sulfur. This is in contrast to heating during reaction in a sealed vessel from which escape of sulfur is not possible (this was confirmed by elemental analysis of the composites). Further evidence that thermal mass loss is primarily attributable to sublimation of sulfur was obtained by TGA analysis of the BAS 90 fraction after sulfur removal, revealing a T d value of 291 • C ( Figure S8A in the Electronic Supplementary Information), well above the reaction temperature. Additionally, isothermal TGA revealed that the BAS 90 sample from which free sulfur had been removed showed no major mass loss until after about 30 min of heating at 240 • C ( Figure S8B in the Electronic Supplementary Information). This time scale provides some insight into the induction period required for thermally induced rearrangement of polymeric sulfur domains in BAS 90 to form sublimable S 8 .

Mechanical Properties
The remeltable samples, BAS 95 and BAS 90 , can be fabricated into any number of shapes (Figure 1) by simply pouring the molten materials into a silicone mold and allowing the samples to harden upon cooling to room temperature. This allowed the two materials to be conveniently fabricated into rectangular prismatic beams appropriate for flexural strength analysis in a dynamic mechanical analyzer ( Figures S9 and S10 in the Electronic Supplementary Information). The flexural strengths (Table 3) of both BAS 90 (>3.4 MPa) and BAS 95 (>4.7 MPa) are comparable to that of portland cement (3)(4)(5). It is notable that the flexural strength values reported in Table 3 represent lower limits to the flexural strengths of the BAS x materials because the instrument was unable to apply enough stress to break the samples. In contrast, the portland cement sample of the same dimensions was broken by the instrument at a force corresponding to a flexural strength of 3.7 MPa. Portland cement is an important target for replacement by more sustainably sourced and recyclable materials because the production of portland cement is energy intensive and is responsible for 8-12% of anthropogenic CO 2 emissions [60,61]. In addition to their potential for high strength, sulfur cements, including BAS x , exhibit very little water uptake (<3 mass %) compared to portland cement (28 mass %), making them more resistant to fissure formation resulting from repetitive freeze-thaw cycles. The density of BAS x (1.8 g/cm 3 ) is also similar to that of portland cement (1.5 g/cm 3 ). One particularly attractive application space for sulfur cements like BAS x is in low-pH environments, because HSMs have proven highly resistant to mechanical deterioration by acid [28,29,62,63]. Recent studies have shown, e.g., that cellulose-sulfur composites are impervious to degradation by sulfuric acid under conditions that completely dissolve a portland cement block of the same dimensions. BAS x materials were likewise evaluated for their ability to withstand acid challenge by submerging samples in 0.5 M H 2 SO 4 for 24 h. Both BAS 90 and BAS 95 maintained their dimensional integrity and flexural strength under these conditions, whereas the portland cement sample was completely decomposed (Table 3).
Having shown that BAS 90 outperforms portland cement in flexural strength and acid resistance, the next test was to evaluate whether BAS 90 material could be recycled. For this test, the flexural strength of a sample of BAS 90 was determined after recursive cycles of pulverization, remelting, and curing in a silicone mold. The results of these measurements for seven cycles are shown in Figure 2. BAS 90 maintains of its mechanical strength, as measured from its storage modulus at 25 • C, within measurement error after this process, although some of the recycling steps produced material with somewhat lower (90-80%) strength. This observation is consistent with some previous reports on the recyclability of HSMs, in which full strength was maintained even after over a dozen cycles [8,28]. The recyclability of HSMs by repetitive melt processing represent a significant improvement over some widely used petrochemical polymers like polyethylene terephthalate (PETE) [64]. The recyclability for PETE is dramatically diminished after only a few thermal processing cycles, to the extent that the strain-at-break falls from 42% for freshly manufactured samples to only 0.7% after only the fifth cycle [65]. Even if BAS x can only be recycled four or five times while maintaining its mechanical durability, this would also be a marked improvement over portland cement, the vast majority of which is land filled at its end of life.

Conclusions
In conclusion, the first step to evaluating the influence of monomer feed on the properties of copolymers produced by RASP reveals that both thermosets and thermoplastics can be accomplished. RASP also proves effective at stabilizing polymeric sulfur within a network made exclusively by S-Caryl bond-forming reactions, in a manner analogous to the stabilization of polymeric sulfur afforded in HSMs prepared by inverse vulcanization. The resultant copolymers have flexural strengths comparable to familiar commercial building materials like portland cement but have improved capabilities in terms of acid resistance and recyclability.

Chemicals and Materials
Tetrabromobisphenol A (>98%) was purchased from TCI America. Sulfur powder (99.5%) was purchased from Alfa Aesar. These chemicals were used without further purification. All processes were carried out under ambient conditions unless otherwise specified.

General Considerations
Thermogravimetric analysis (TGA) data were recorded on a Mettler Toledo TGA 2STARe instrument over the range of 25-800 °C, with a heating rate of 5 °C min -1 under a flow of N2 (20 mL min -1 ). Differential scanning calorimetry (DSC) data was obtained using Mettler Toledo DSC 3 STAR e System, over the range of −60 to 140 °C, with a heating rate of 5 °C min −1 under a flow of N2 (200 mL min -1 ). All the reported data were taken from the third heat/cool cycles.
Dynamic mechanical analysis (DMA) data was acquired using Mettler Toledo DMA 1 STAR e System in single cantilever mode. DMA samples were cast from silicone resin molds (Smooth-On OOMOO® 30 tin cure). The samples were cured for 120 h prior to stress-strain analysis at room temperature. The sample dimensions were approximately 40 mm × 11 mm × 2 mm and were clamped with 1 cN·m force. The samples cured for 120 h prior to stress-strain analysis at room temperature. The samples were clamped under single cantilever mode. The force was varied from 0 to 10 N with a ramp rate of 0.1 N·min -1 . Percent crystallinity of the composites was determined by using Equation (2):

Chemicals and Materials
Tetrabromobisphenol A (>98%) was purchased from TCI America. Sulfur powder (99.5%) was purchased from Alfa Aesar. These chemicals were used without further purification. All processes were carried out under ambient conditions unless otherwise specified.

General Considerations
Thermogravimetric analysis (TGA) data were recorded on a Mettler Toledo TGA 2STARe instrument over the range of 25-800 • C, with a heating rate of 5 • C min −1 under a flow of N 2 (20 mL min −1 ). Differential scanning calorimetry (DSC) data was obtained using Mettler Toledo DSC 3 STAR e System, over the range of −60 to 140 • C, with a heating rate of 5 • C min −1 under a flow of N 2 (200 mL min −1 ). All the reported data were taken from the third heat/cool cycles.
Dynamic mechanical analysis (DMA) data was acquired using Mettler Toledo DMA 1 STAR e System in single cantilever mode. DMA samples were cast from silicone resin molds (Smooth-On OOMOO®30 tin cure). The samples were cured for 120 h prior to stress-strain analysis at room temperature. The sample dimensions were approximately 40 mm × 11 mm × 2 mm and were clamped with 1 cN·m force. The samples cured for 120 h prior to stress-strain analysis at room temperature. The samples were clamped under single cantilever mode. The force was varied from 0 to 10 N with a ramp rate of 0.1 N·min −1 .
Percent crystallinity of the composites was determined by using Equation (2): where χ c is the percentage crystallinity with respect to sulfur, ∆H m is the melting enthalpy for the indicated material (sulfur or BAS x ), and ∆H cc is the cold crystallization enthalpy of the indicated material.

Synthesis of BAS 95
Elemental sulfur (9.5 g) and Br 4 BPA (0.5 g) was weighed directly into a pressure tube under inert environment. The tube was heated to 240 • C and heating was continued for 24 h with continuous stirring with a magnetic stir bar. The whole procedure was done under N 2 gas, giving quantitative yield of the material. Elemental combustion microanalysis calculated: C, 1

Preparation of Early Reaction Stage Products from Br4BPA
Elemental sulfur (0.088 g, 3.4 × 10 −4 mol of S 8 ) and Br 4 BPA (0.25 g, 4.6 × 10 −4 mol) was weighed directly into a pressure tube under an inert atmosphere. The tube was heated to 240 • C and heating was continued for 24 h with continuous stirring with a magnetic stir bar. The whole procedure was done under N 2 gas. After 24 h, the pressure tube was taken off the heat and allowed to cool to room temperature. An aliquot of LiAlH 4 (2.5 equivalent with respect to sulfur) suspended in toluene (10 mL) was then added to the pressure tube under an inert atmosphere to facilitate transformation of any S-S linkages to S-H units. The mixture was quenched with 0.5 M HCl solution followed by extraction with dichloromethane (3 × 30 mL). The organic layer was dried over Na 2 SO 4 and the solvent was evaporated under reduced pressure to yield a brown sticky solid. Efforts to separate the constituent compounds by preparative TLC were unsuccessful.

1.
CAUTION: Heating elemental sulfur with organics can result in the formation of H 2 S gas. H 2 S is toxic, foul smelling, and corrosive. This reaction may also produce S 2 Br 2 , a toxic and reactive material.

2.
CAUTION: Heating material in a sealed tube can generate high pressures. Use caution, and consult the manufacturer of the pressure apparatus used for safety guidance.

Depolymerization with AlBr3
The BAS x sample (30 mg) was powdered and mixed with 60 mg of AlBr 3 in a glovebox under an atmosphere of dry N 2 . The mixture of solid was suspended in 6 mL of anhydrous toluene for 30 min. At the end of the reaction time, the solvent was filtered and 5% (v/v) HCl:ethanol (5 mL) was added. Following three consecutive washes, the organic layer in toluene was separated out.

Procedure for the Reaction of Depolymerization Products with Ellman's Reagent
The procedure was carried out as previously reported. The stock solution of DTNB (5,5 -Dithio-bis-(2-nitrobenzoic acid)) was prepared by dissolving 40 mg of DTNB in 10 mL of DMSO. The solution was 100-fold diluted into 0.1 mM DTNB solution using Tris·HCl (Tris(hydroxymethyl)aminomethane hydrochloride) buffer solution at a pH of 7.4. A 1 mL aliquot of the solution containing depolymerized products and 19 mL of 0.1 mM DTNB were mixed together and then allowed to stand at room temperature for 2 min at room temperature. The samples were then analyzed using a Cary 50 UV-VIS spectrophotometer. The same procedure was carried out without adding AlBr 3 to confirm the absence of thiol groups in the initial sample as well as on a sample of BAS x that had not been depolymerized to rule out the presence of surface thiols prior to depolymerization.

Conclusions
In conclusion, the first step to evaluating the influence of monomer feed on the properties of copolymers produced by RASP reveals that both thermosets and thermoplastics can be accomplished. RASP also proves effective at stabilizing polymeric sulfur within a network made exclusively by S-C aryl bond-forming reactions, in a manner analogous to the stabilization of polymeric sulfur afforded in HSMs prepared by inverse vulcanization. The resultant copolymers have flexural strengths comparable to familiar commercial building materials like portland cement but have improved capabilities in terms of acid resistance and recyclability.
Supplementary Materials: The following are available online at http://www.mdpi.com/2673-4079/1/2/13/s1, Figure S1: Surface analysis of BAS90 by scanning electron microscopy (SEM) revealed a smooth surface consistent with those observed in high sulfur-content materials prepared by inverse vulcanization, Figure S2. Surface analysis of BAS90 by energy-dispersive X-ray (EDX) analysis revealed even distribution of sulfur (A), carbon (B), oxygen (C) and bromine (D) content on the polymer surface, Figure S3. Differential scanning calorimetry of BAS95, Figure  S4. Differential scanning calorimetry of BAS90, Figure S5. Differential scanning calorimetry of BAS85, Figure  S6. Differential scanning calorimetry of BAS80, Figure S7. TGA traces showing thermally-induced mass loss for monomer Br4BPA (red), BAS95 (blue), BAS90 (black), BAS85 (violet) and BAS80 (green) under nitrogen, Figure S8. TGA traces for the fraction of BAS90 from which free sulfur has been removed. In one experiment mass loss was monitored as the sample was heated from room temperature to 600 • C (A), and in another case mass loss was monitored as the temperature was held at 240 • C for 2.5 h (B), Figure S9