Metal-catalyzed copolymerizations of epoxides and carbon disulfide for high-refractive index low absorbance adhesives and plastics

High-refractive index plastics are useful materials due to their optical properties, ease of processing, and low-costs compared to their inorganic counterparts. Catalytic carbon disulfide (CS2) copolymerization with epoxides is one method for producing low-cost high refractive index polymers. The reaction is accompanied by an oxygen-sulfur exchange reaction which produces irregular microstructures in the repeating units. In this study, metal salen catalysts were investigated with different metal centers (Al, Cr, Co) and salen ligand electronics, sterics, backbones, and co-catalyst in the copolymerization of CS2 with propylene oxide (PO) and cyclohexene oxide (CHO). The results reveal the essential nature of Cr metal centers on reactivity and the backbone geometry on monomer selectivity. There were no significant impacts on the O-S exchange reaction when ligand design changed, however PO and CHO/CS2 copolymers yield different monothiocarbonate microstructures. Additionally, the effects of microstructure on optical and thermal properties were investigated using spectroscopic ellipsometry and calorimetry, respectively. The CHO system produced high T g plastics (93°C) with high refractive indexes (n up to 1.64), modest absorbance (κ < 0.020), and Abbe numbers of 32.2 while PO yielded low T g adhesives (T g = 9°C) with high refractive indexes (n up to 1.73), low absorbance (κ < 0.005), and low Abbe numbers (V D = 19.1).


Introduction
High-refractive index materials are essential components for imaging technologies and optics (Vollmer and Mollmann, 2017).Infrared (IR) imaging systems, for example, are commercially derived from inorganic composites such as metal fluorides, or metal chalcogenides (e.g., Ge, Si, Cd, Ga, Zn); however, advances in high refractive index polymers opens the possibility of flexible, processible, and cost-efficient alternatives (Kleine et al., 2020).In order to deploy these active components in devices, they must adhere to detectors or multilayers and also match refractive indices with other components.In addition to controlling refractive index (n), the extinction coefficient of absorption (κ) should be low across the optical window to lessen potential signal interference.
While episulfide/CS 2 copolymerization produces well-defined trithiocarbonate repeat units (Nakano et al., 2007), epoxides introduce oxygen-sulfur exchange reactions that afford multiple thiocarbonate microstructures (Figure 1).Zhang reported this O-S exchange in double metal cyanide (DMC) catalyzed reactions between propylene oxide (PO) and CS 2 (Zhang et al., 2008;Zhang et al., 2009), and Darensbourg observed the exchange in cyclohexene oxide (CHO) and CS 2 using a chromium (III) salen catalyst (Darensbourg et al., 2009).Mechanistically, O-S exchange occurs either in the monomer or at the active chain-end (Zhang et al., 2008;Darensbourg et al., 2009;Yang et al., 2020).If catalyst systems can be devised that are capable of tuning the degree of O-S exchange, both the optical and thermo-mechanical properties of epoxide/CS 2 copolymers could be tuned to access high T g plastics or low T g adhesives for optical devices.Since the catalyst is involved in both exchange and propagation pathways, we hypothesized that catalyst design may be a method for tuning the O-S exchange and therefore the optical properties of the produced polymers, while the monomer selection (i.e., CHO or PO) could afford access to high T g plastics or low T g adhesives.
Previous approaches to controlling O-S exchange include reaction temperature and [CS 2 ]: [epoxide] which also bias the yield of cyclic and polymeric products (Darensbourg et al., 2009).Through control over O-S exchange chemistry, different thermal properties are obtained due to the crystallinity differences between polycarbonate and polythiocarbonate repeat units.Zhang and coworkers accessed a range of epoxide/CS 2 thermal properties through the copolymerization of ethylene oxide with CS 2 (Yang et al., 2021).In their system, metal-free Lewis acid/base pairs were used to produce semicrystalline blocky materials with varying degrees of O-S exchange and T m s as high as 245 °C.The resulting semi-crystalline poly (ethylene trithiocarbonate) and poly (ethylene monothiocarbonate) linkages are segmented within low T g poly (ethylene carbonate) (T g = −35 to −18 °C) and afford controlled degrees of crystallinity (0%-78% crystallinity).Darensbourg and coworkers reported the use of cyclopentene oxide/CS 2 copolymerization using a Cr salen catalyst, which yielded O-S exchange microstructures and episulfide homopolymer linkages (i.e., polythioether) with low T g s (−17 °C-32 °C) (Darensbou et al., 2013).Werner and coworkers discovered propylene oxide, butylene oxide, and cyclohexene oxide copolymerize with CS 2 in the presence of LiOtBu initiators to produce regioirregular polythiocarbonates with high molecular weights (up to 109 kg/mol) (Diebler et al., 2016).Through NMR spectroscopic analysis of the polythiocarbonates, the regioirregular head-to-head arrangements could be detected in thiocarbonate and tail-to-tail arrangements in the trithiocarbonate linkages.
As part of our interests in catalyst design, epoxide copolymerization, and polymer properties we herein investigate the effects of metal salen active site, sterics, electronics, and geometries on the O-S exchange reaction, polymerization activity, and monomer selectivity.The catalyst design enables the synthesis of either high T g CHO/CS 2 copolymers (93 °C) or low T g (−23 °C) PO/CS 2 copolymers that exhibit low absorbance and high refractive indices across the visible spectrum.

FIGURE 1
Comparative study of PO and CHO copolymerization with CS 2 using variations on the salen ligand, metal center, and cocatalyst.

Materials
CHO and PO were purified over CaH 2 for 24 h, followed by vacuum transfer and degassing by three freeze-pump-thaw cycles.CH 2 Cl 2 was purified through a solvent purification column equipped with molecular sieves.All catalysts were prepared using literature procedures (Hosseini Nejad et al., 2012) which followed salicylaldehyde condensation onto diamine backbones using formic acid in methanol, and subsequent metalation with Et 2 AlCl, CoCl 2 , or CrCl 2 accordingly (Hosseini Nejad et al., 2012).Tethered binuclear catalyst 9 was synthesized according to the procedure of Lu and coworkers (Ren et al., 2009).Additional material details and the experimental spectra are available in the Supplementary Material.

Polymerization procedures
PO/CS 2 copolymerization: a 25 mL Schlenk flask was dried in the oven overnight, brought into glovebox, and catalyst (0.01 mmol, 1.0 equiv.)and bis (triphenylphosphine) iminium chloride (0.01 mmol, 1.0 equiv.)were added.The solids were dissolved in 2 mL of CH 2 Cl 2 and allowed to incubate for thirty minutes before evaporating under reduced pressure.Carbon disulfide (0.60 mL, 1,000 equiv.)and then propylene oxide (0.35 mL, 500 equiv.)were added and stirred at 23 °C for 5 h.An aliquot was removed for crude NMR analysis of conversion based on remaining propylene oxide and the remainder of the system was quenched and precipitated with 5% HCl/MeOH, to yield a red viscous polymer and red MeOH layer.The samples were centrifuged at four thousand rpm for 5 minutes to separate the viscous and liquid components.The viscous layer was redissolved in CH 2 Cl 2 and reprecipitated with acidic MeOH a total of three times until the MeOH layer was no longer colored yielded a yellow polymer precipitate.The tacky material was collected, dried under vacuum, and analyzed by GPC, DSC, and NMR and taken as the polymer mass, while the methanol layers were concentrated for analysis of the cyclic byproducts. 1 H NMR (500 MHz, CDCl 3 ) δ = 5.94 (m, CH dithiocarbonate C=S), 5.60 (br s, CH monothiocarbonate C=S), 5.25 (m, CH monothiocarbonate C=S), 5.01 (br s, CH carbonate), 3.84-3.35(m, CH 2 and CH trithiocarbonate), 1.45-1.40(m, CH 3 ) ppm. 13   ,965, 2,909, 1740, 1716, 1,647, 1,449, 1,377, 1,348, 1,259, 1,223, 1,013, and 797 cm −1 .
CHO/CS 2 copolymerization: a 25 mL Schlenk flask was dried in the oven overnight, brought into glovebox, and catalyst (0.01 mmol, 1.0 equiv.)and bis (triphenylphosphine) iminium chloride (0.01 mmol, 1.0 equiv.)were added.The solids were dissolved in 2 mL of CH 2 Cl 2 and allowed to incubate for thirty minutes before evaporating under reduced pressure.Carbon disulfide (2.42 mL, 4,000 equiv.)and then cyclohexene oxide (2.02 mL, 2000 equiv.)were added and heated at 50 °C for 4 h.An aliquot was removed for crude NMR analysis of conversion based on remaining CHO and the remainder of the system was quenched and precipitated with 5% HCl/MeOH, to yield a red viscous polymer and red MeOH layer.The samples were centrifuged at four thousand rpm for 5 minutes to separate the viscous and liquid components.The viscous layer was redissolved in CH 2 Cl 2 and reprecipitated with acidic MeOH a total of three times until the MeOH layer was no longer colored yielded a yellow polymer precipitate.The powdery solids were collected, dried under vacuum, and analyzed by GPC, DSC, and NMR and taken as the polymer mass, while the methanol layers were concentrated for analysis of the cyclic byproducts. 1

Ellipsometry
Polymer samples were dissolved in CH 2 Cl 2 (1 g/mL) and spuncoated onto Piranha-cleaned silicon wafers (ca. 2 cm × 2 cm).The samples were annealed in a vacuum oven above the samples T g overnight.
Two steps were taken to implement a reliable analysis of the ellipsometer data.First, we directly measured thickness of the spuncoated films using atomic force microscopy (AFM; Bruker Dimension Icon) in the tapping-mode.The spun-coated films were scored with a sharp razor blade to create a distinct edge, that allowed procurement of thickness profiles.The directly measured thickness eliminated the term as a fitting parameter in modeling the ellipsometric parameters.
Second, depending on the optical characteristics of the sample, if the sample does not absorb energy (imaginary part of the complex refractive index, i.e., κ ≈ 0) then the Cauchy equation can be used to use to determine the real part of the complex refractive index (n) profile of the material (Equation 1) (Pavunny et al., 2012).
where n is the real part of complex refractive index, λ is wavelength in nm, and A,B,C . . ., are fitting coefficients.Additionally, if the sample does absorb light, the Cauchy-Urbach dispersion model can be used which incorporates an equation to calculated κ (Equation 2) (Pavunny et al., 2012).
where κ is imaginary part of the complex refractive index, α and β are constants, λ is wavelength in nm, and E b is the band gap energy in eV.Thus, choosing the critical Kramers-Kronig consistent dispersion model is essential in modeling the ellipsometric parameters to calculate the complex refractive index (n*) dispersion.
The ellipsometric parameters [delta (Δ) and psi (ψ)] of the thin films were collected at 23 °C via a Nulling Ellipsometry method at varying angles of incidences (50 °, 55 °, and 60 °) across a spectroscopic wavelength from 360 nm to 1700 nm using an Imaging Ellipsometer (Nanofilm EP4; Accurion GmbH, Germany).The parameters of the n* dispersion model used in the fitting of the ellipsometric parameters are presented in Supplementary Table S1.The Δ and Ψ curves, with the appropriate thickness values and n* dispersion model, were modeled in the commercial EP4Model data analysis software by Accurion GmbH, Germany.
When the catalyst backbone was changed to non-planar geometries, as in the trans-cyclohexadiamine (7) and ethylene diamine (8) salen ligands, the Cr catalysts showed polymerization activity in both PO and CHO/CS 2 copolymerizations.By GPC/SEC analysis, CHO/CS 2 polymers were consistently higher molar mass than PO/CS 2 products synthesized from the same catalyst (entries 7, 8 & entries 9, 10).We hypothesize this reactivity difference arises from the ability for PO to react through axial binding to the metal center, whereas CHO requires a cis-ligand geometry (e.g., pseudooctahedral) to undergo the O-S exchange reaction and/or propagation; similar to what has been observed in epoxide/CO 2 copolymerization by porphyrin catalysts (Chatterjee and Chisholm, 2013).This cis-ligand geometry is less accessible in the planar phenylene diamine backbone (1) and thus only PO catalysis was productive.Whereas both Co. and Cr catalysts are active in epoxide/ CO 2 copolymerization and Al catalysts are active in epoxide/ anhydride copolymerization (Longo et al., 2016), only Cr catalysts were active in epoxide/CS 2 copolymerization.It is expected that the combined Lewis acidity, oxophilicity, and thiophilicity of Cr are key thermodynamic considerations for promoting catalyst activity in both propagation and O-S exchange mechanisms (Kepp, 2016;Zhang et al., 2020).
The electronics and steric environments of Cr (salph) systems (4-6) were next investigated.Unsubstituted salicylaldehyde derived catalyst (4) exhibited no CS 2 copolymerization with PO or CHO, which we attribute to poor catalyst solubility.Electronic deficient para-fluoro (5) and electron rich para-methoxy (6) catalysts revealed that electron rich systems were more reactive and produced higher molar masses by GPC (2,400 and 1,500 g/mol, respectively).The electronics enhance nucleophilicity of the dithiocarbonate chainend (Darensbou et al., 2013), and accelerates the cyclization reaction Measured from the second heating cycle inflection point of DSC, at 10 °per minute.f Measured at 95 wt% remaining according to thermogravimetric analysis.g Determined from molar mass ratios of repeat units observed in integrated 13C NMR, spectra.
Frontiers in Chemistry frontiersin.org04 Schwarz et al. 10.3389/fchem.2023.1287528and increases the observed yields of cyclic over polymer (entries 5 and 6).Dinuclear tethered cocatalyst system 9 (Ren et al., 2009), in the absence or presence of 1 equiv. of PPNCl failed to afford any conversion, consistent with the other cobalt catalysts studied.The obtained O-S exchange microstructures in productive catalyst systems were investigated by 13 C NMR to ascertain the sulfur content (wt%) and ratio of trithiocarbonate, mono thiocarbonate (C=O, C=S), and carbonate linkages (Figure 2).While there are six possible combinations (Diebler et al., 2016)

of exchange repeat units-S (C=S)S, S (C=S)O, S (C=O)S, O (C=S)O, S (C=O)O, O (C=O)
O-four prominent microstructures were produced by Cr salen catalysts.The NMR of entry 1 (Figure 2; Supplementary Figures S1, S2) exhibited predominantly trithiocarbonate resonances (43 mol %) at chemical shifts of 222 ppm and monothiocarbonates (C=S, 40 mol%) at 193 ppm.The remaining linkages were non-sulfur containing carbonate linkages (9 mol%) and trace amounts of dithio-and monothiocarbonates (C=O).This indicates that under metal-catalyzed conditions, O-S exchange can produce polymeric materials with higher sulfur content than the unexchanged dithiocarbonate products.
Whereas catalyst electronics and ligand geometry dictated monomer selectivity and polymerization productivity, there were only minor differences (within 5 mol% deviations by 13 C NMR) observed in the microstructures within a given epoxide study (Supplemental Figures S1-S10).However, we did observe significant differences in microstructures between CHO and PO.Compared to the above PO/CS 2 polymer (Figure 2), CHO/CS 2 copolymerization afforded predominantly trithiocarbonate (52%) and carbonate linkages (31%) with a minor monothiocarbonate (C=O) component (16%).Interestingly, the differences in monothiocarbonate isomers (C=S with PO and C=O with CHO) points to a difference in O-S exchange mechanisms that is dependent on the monomer and catalyst system.In both systems the unexchanged dithiocarbonate (C=S) is observed in only slight amounts (<5 mol%) by 1 H NMR. Given the observations that propylene oxide/COS copolymers selectively yield monothiocarbonates with C=O moieties rather than C=S, we propose our observed products form from postpolymerization O-S exchange of a carbonate moiety rather than direct incorporation of in-situ generate COS molecules (Luo et al., 2016).Because CHO is more sterically hindered and conformationally restricted, the necessary transition state for this post-polymerization exchange is suppressed and the carbonate moieties are observed (Zhang et al., 2008;Darensbourg et al., 2009;Darensbou et al., 2013;Yang et al., 2020).The thermal properties of the CHO/CS 2 and PO/CS 2 copolymers were measured by differential scanning calorimetry (DSC) in order to ascertain their potential as optical plastics and adhesives, respectively (Figure 3; Supplemental Figure S11).The cyclohexyl structure restricts molecular mobility throughout the chain and rigidifies the polymers giving rise to high T g materials (T g = 93 °C).The higher T g is essential for plastic and rigid optical components.Alternatively, the flexible linear propylene linkage imparts rubbery properties (T g = 9 °C) after precipitation and purification from cyclic plasticizers.Compared to ethylene oxide/ CS 2 copolymers (Yang et al., 2021), the obtained polymers are entirely amorphous due to the monosubstituted and disubstituted epoxides used which disrupt chain packing in these atactic poly (thio) carbonates.These amorphous PO/CS 2 copolymer structures exhibit the low T g s needed for imparting flexibility and peel strengths properties.

Optical properties
Investigation of the optical parameters in the CS 2 /PO or CHO systems was carried out using variable-angle spectroscopic ellipsometry.This optical measurement (Figure 4) provides the n and κ across wavelengths (λ) in order to ascertain the utility of a material in high refractive index, low-absorbance, multilayer devices.In the PO/CS 2 sample (entry 1), the range of n was 1.73-1.60(λ = 360-1700 nm), which compares favorably to n = 1.78 of episulfide derived CS 2 poly (trithiocarbonates) (Nakano et al., 2007).Encouraging was the low absorbance extinction coefficient for PO/CS 2 copolymer (κ = 0.004-0.005)across the studied spectrum (λ = 360-1700 nm).
Another characteristic that can be derived from these optical constants is the Abbe number (V D ) which is a measure of dispersion, how the refractive index changes with wavelength, within a transparent material and is defined by equation 4. A higher V D corresponds to a low dispersion material, which improves that ability for the material to focus light.Entry 1 exhibits a very high dispersion with a V D = 19.1, while the CHO/CS 2 copolymer (entry 10) exhibits a more favorable Abbe number (V D = 32.2),which is comparable to polycarbonate glasses.While the CHO-based materials demonstrates improved dispersive properties, it is accompanied by lower sulfur content (28 wt%), a lower refractive index (n = 1.64-1.56),and higher absorbance (κ = 0.009-0.020).
Equation 3, where: n F is the refractive index at wavelength of the Fraunhofer F spectral line (486 nm), n d is the refractive index at wavelength of the Fraunhofer d spectral line (587 nm) and n C is the refractive index at wavelength of the Fraunhofer C spectral line (656 nm).

FIGURE 4
Refractive index and extinction coefficient values across 360-1,700 nm as measured by variable-angle spectroscopic ellipsometer for PO and CHO/CS 2 copolymers.

FIGURE 5
Comparison of refractive index and sulfur content in oxirane/CSX (X = S or O) copolymers (Nakano et al., 2007;Zhang et al., 2008;Darensbourg et al., 2009;Zhang et al., 2009;Yang et al., 2020).Black corresponds to low T g propylene materials, blue corresponds to high T g cyclohexyl materials, open points are literature values, and closed points are from this study.
The metal catalyst design in epoxide/CS 2 copolymerization was shown to dictate monomer selectivity and polymerization productivity.By matching the catalyst selection with epoxide choice (CHO or PO), high T g plastics (T g = 93 °C) with high refractive indexes (n up to 1.64), modest absorbance (κ < 0.020), and good Abbe numbers (V D = 32.2) could be synthesized along with low T g adhesives (T g = 9 °C) with high refractive indexes (n up to 1.73), very low absorbance (κ < 0.005), and low Abbe numbers (V D = 19.1).Within a given epoxide monomer the microstructures were consistent, PO/CS 2 copolymers all produce materials with 45 ± 1 sulfur wt% and CHO/CS 2 copolymers produce materials with 30 ± 2 sulfur wt%.However, CHO microstructures were dominated by trithiocarbonates and monothiocarbonates with thiocarbonyl moieties (C=S); whereas PO microstructures were dominated by trithiocarbonates and carbonate moieties.In the theoretical polymerization where no O-S exchange occurs PO yields polymers with 48 wt% and CHO 37 wt%, which indicates the metal-catalyzed exchanged decreases the optimal sulfur content for high refractive index.However, by suppressing the formation of carbonate linkages, CHO/CS 2 copolymers are relatively closer to their optimal sulfur wt%.From these findings, research may be pursued wherein O-S exchange favors trithiocarbonate formation and S content rivaling episulfide/CS 2 copolymerization (64 and 51 wt% for PO and CHO, respectively) (Nakano et al., 2007;Zhang et al., 2008;Darensbourg et al., 2009;Zhang et al., 2009;Yang et al., 2020).Context for the materials synthesized in this study and in the literature are presented in Figure 5, wherein various combinations of epoxides, episulfides, CS 2 , CO 2 , and COS enable synthetic control over the sulfur content which directly relates to the material refractive index.From this polymer toolset both the thermal and optical properties can be targeted for a given application.

FIGURE 3
FIGURE 3 DSC of adhesive PO/CS 2 (Entry 1) and plastic CHO/CS 2 (Entry 10) copolymers.Second heat shown at a rate of 10 °C min -1 and glass transition values taken at the inflection point.