Inverse Vulcanization of Styrylethyltrimethoxysilane–Coated Surfaces, Particles, and Crosslinked Materials

Abstract Sulfur as a side product of natural gas and oil refining is an underused resource. Converting landfilled sulfur waste into materials merges the ecological imperative of resource efficiency with economic considerations. A strategy to convert sulfur into polymeric materials is the inverse vulcanization reaction of sulfur with alkenes. However, the materials formed are of limited applicability, because they need to be cured at high temperatures (>130 °C) for many hours. Herein, we report the reaction of elemental sulfur with styrylethyltrimethoxysilane. Marrying the inverse vulcanization and silane chemistry yielded high sulfur content polysilanes, which could be cured via room temperature polycondensation to obtain coated surfaces, particles, and crosslinked materials. The polycondensation was triggered by hydrolysis of poly(sulfur‐r‐styrylethyltrimethoxysilane) (poly(Sn‐r‐StyTMS) under mild conditions (HCl, pH 4). For the first time, an inverse vulcanization polymer could be conveniently coated and mildly cured via post‐polycondensation. Silica microparticles coated with the high sulfur content polymer could improve their Hg2+ ion remediation capability.


Experimental Section
Synthesis of poly(Sn-r-StyTMS) 500 mg of Sulfur powder (> 99%, Alfa Aesar, Ward Hill, USA) and 500 mg of Styrylethyltrimethoxysilane (Gelest, Morrisville, USA) were weighed into a vial with a stir bar and sealed with a cap. The vial was heated to 130 °C for 8 h under stirring at 400 rpm. The vial was extracted stepwise with 4, 2 and 2 mL of THF, respectively. The solution was cooled to -20 °C for 20 min to precipitate residual and excess sulfur, which was then removed by centrifugation. Evaporation of volatiles yielded a THF soluble, brown oil. For the experiments, poly(Sn-r-StyTMS) was not isolated but instead used directly in THF solution after isolation from sulfur.

Solution processing of poly(Sn-r-StyTMS)
Poly(Sn-r-StyTMS) was diluted with THF to reach a mass concentration of 40 mg mL -1 (rel. to the initial mass of styrylethyltrimethoxysilane). Poly(Sn-r-StyTMS) was hydrolyzed by stirring with 5 vol% of dilute HCl (pH 4) for 60 min at room temperature.

Spin coating of silicon surfaces for ellipsometry
Ca. 250 µL of hydrolyzed poly(Sn-r-StyTMS) were pipetted onto pieces (ca. 2 × 2 cm) cut from a silicon waver until it was fully covered with liquid. The waver pieces were then spun at 6000 rpm for 35 s.

Spin coating of surfaces for AFM scratch analysis
All substrates were spun covered in poly(Sn-r-StyTMS) solution at 1500 rpm for 15 s. The same volume was then added on the substrate within 10 s and it was spun at 1500 rpm for 15 s again. For silicon, pieces cut from a silicon waver (ca. 2 × 2 cm) 250 µL were used for both pipetting steps. For gold coated glass slides and glass slides (2.5 × 7.5 cm) ca. 800 µL of hydrolyzed poly(Sn-r-styrylsilane) were used for both pipetting steps. For consecutive spinning cycles the specimen were air dried for at least 20 min before the next spinning cycle. Glass slides were activated by immersion into fuming HCl:MeOH (1:1 V:V) for at least one day. The water contact angles before and after activation were 49.4° ± 0.2° and 31.3° ± 0.3°, respectively.
Synthesis of net-poly(Sn-r-StyTMS) coated silica particles 3.00 g of sulfur powder (> 99%, Alfa Aesar, Ward Hill, USA) and 3.00 g of styrylethyltrimethoxysilane (Gelest, Morrisville, USA) were weighed into a vial with a stir bar and sealed with a stopper. The vial was heated to 130 °C for 8 h under stirring at 400 rpm. The vial was extracted with THF and poured into a beaker. The solution was then decanted to remove residual sulfur and filled up to a total of 300 mL THF. 50.13 g Silica particles (40-63 µm) were charged into a 1 L round bottom flask and the poly(Sn-r-StyTMS) solution was poured in. 1.5 mL of H2O (pH 4) were added and the dispersion was stirred vigorously for 24 h. The mixure was then filtered through a buchner filter. The filter cake was briefly (10 min) dried in an oven at 80 °C. The particles were then washed with 80, 50, and 3 × 20 mL of THF and then with 3 × 20 mL of CS2, until the washings were colorless. They were then dried for 20 min at 110 °C.
A total of 51.86 g coated particles could be obtained.

Synthesis of net-poly(Sn-r-StyTMS)
Hydrolyzed poly(Sn-r-StyTMS) was charged in a Teflon mold or round bottom flask. Volatiles were removed at room temperature, in an oven, or in a rotary evaporator. The yielded orange-red solid was washed thoroughly with THF and CS2. For DSC, TGA, solid state NMR and PXRD measurements, the orange-red solid was milled into a powder and was allowed to age for two months.

Mercury Removal Test
Silica particles coated with net-poly(Sn-r-StyTMS) (200 mg) were stirred in 10 mL of aqueous HgCl2 solution (15 mg L -1 ) for 1 h. The dispersion was then centrifuged, and ca. 3 mL of the solution were decanted and filtered. The filtrate was diluted with 1:30 v/v% water (HPLC grade) and 300 µL of concentrated HCl were added. The Hg 2+ concentration was determined with Hydride-AAS (atom absorbance spectroscopy). Pristine silica particles (40 -63 µm) were used as a reference. The pH of the filtrate after the remediation experiment was neutral.

Thin film interference color calculation
To calculate the thin film interference in dependence of viewing angle and thickness, a script created by Dr. Jens Raacke was used (http://www.raacke.de/index.html?airy.html).
Based on the Cauchy parameters n0, n1 and n2 and the absorption coefficients k0, k1 and k2 the functions n(λ) and k(λ) are calculated in steps of 5 nm from 380 nm to 780 nm. n and k are combined as complex refractive index N according to the Kramers-Kronig relation (I) for further calculations. 2) The refracted beams are a superposition of multiple beams. To calculate the total reflectivity, the sum of all refracted beams is formed based on series development.
3) The reflectivity spectrum is multiplied with the sensitivity spectra of all three human color receptors to obtain a XYZ color scheme.
4) The obtained XYZ colors are converted to RGB colors.

5) Light source correction from Illuminant E (emission intensity assumed constant for all wavelengths of
the visible light spectrum) to the selected light source (5300 K) and averaging over TE and TM polarization.

Methods
Specifications of analytical devices in an alphabetical order are listed below.

AFM
The surface topology of samples was investigated with a Dimension Icon with ScanAsyst from Bruker (Billerica, USA). Cantilevers with a resonance frequency of 325 kHz from Olympus (Shinjuku, Japan) were used.

Nanosized particles were measured on a Zetasizer Nano ZS from Malvern Instruments (Malvern, United
Kingdom) equipped with a 633 nm laser.

DSC
Differential Scanning Calorimetry was performed on a Discovery DSC from TA Instruments (Newcastle, USA). The heating rate was 10 K min -1 .

EDX/SEM
Specimen were analyzed with a LEO 1530 scanning electron microscope from Leica (Hillsboro, USA) with an accelerating voltage of 5-10 kV. For SEM analysis the specimen were sputtered with a thin layer of gold. For EDX a NORAN System SIX from Thermo Scientific (Waltham, USA) was used.

Ellipsometry
Spectroscopic Ellipsometry measurements were performed on a M2000 from Woollam (Lincoln, USA).
Samples were prepared via spin-coating on silicon wafers (ca. 2 × 2 cm 2 ). Measurements were taken at an angle of incidence of 45°, 55° and 65° in the spectral range of 300-1000 nm. To evaluate the experimental data, an optical box model was applied using the instrument software CompleteEase (V6.51). Silicon substrates were fitted with database values for Si and SiO2. [1] To determine the thickness and optical constants of the polymer layer, all spectra of samples of varying thickness were fit together in a multi-sample analysis. For this, the Cauchy function with Urbach extension terms was used in order to consider the absorption band of the polymer in the UV range. Common fit parameters were An, Bn, as well as Ak and Bk. The thickness was allowed to vary freely between samples. As the band edge, 400 nm was assumed.

GPC
Samples were dissolved in THF (ca. 2 mg mL -1 ), filtered through 0.43 µm PTFE filters and injected in a Tosoh EcoSEC GPC system from Tosoh (Tokio, Japan) equipped with a SDV 5 μm bead size guard column (50 × 8 mm) followed by three SDV 5 μm columns (300 × 7.5 mm, subsequently 100, 1000, and 105 Å pore size, PSS), a differential refractive index (DRI) detector, and a UV-Vis detector set to 254 nm. THF was used as eluent at 35 °C with a flow rate of 1.0 mL•min −1 . The SEC system was calibrated by using linear polystyrene standards ranging from 800 to 1.82 × 10 6 g mol -1 .

Solid-state NMR
29 Si high-power decoupled (HPDEC) and cross-polarization (CP) magic-angle spinning (MAS) NMR spectra were recorded using a Bruker Avance III 400 WB spectrometer from Bruker (Rheinstetten, Germany) equipped with a 4-mm double resonance MAS probe at a read-out temperature of 300 K. The spectra were acquired at a spinning speed of 12 kHz using TMS at 0 ppm as external reference.
Optimized 1 H and 29 Si 90° pulse lengths were 2.5 and 5.97 µs, respectively. The 29 Si CP spectrum was acquired with 1024 scans and a recycle delay of 5 s. 1 H to 29 Si magnetization transfer was achieved by using linear 70-100% 1 H-ramped CP with a contact time of 8 ms to fulfill the Hartmann-Hahn condition. [3] Heteronuclear decoupling during acquisition was achieved with swept-frequency two-pulse phase modulation (SWf-TPPM). [4,5] For HPDEC experiments, 664 and 736 scans were collected for powdered net-poly(Sn-r-StyTMS) and the background, respectively, with a recycle delay of 90 s. SWf-TPPM heteronuclear decoupling was used during the acquisition. Processed data were analyzed using PXRD Powder X-ray diffraction was measured using a PANanlytical X´Pert PRO diffractometer from Malvern Panalytical (Malvern, UK), operating in transmission geometry. All data were measured over the range 5-50° 2θ.

Raman
Raman spectra were recorded on a MultiRAM from Bruker Optik (Ettlingen, Germany) equipped with a Nd:YAG (1064 nm) laser.

UV/VIS
Spectra were recorded with a PerkinElmer Lambda 35 UV/VIS spectrometer from PerkinElmer Ink.

TGA-MS
Peaks corresponding to isomers of styrylethyltrimethoxysilane were assigned as follows: 1    To demonstrate the stability of net-poly(Sn-r-StyTMS) against depolymerization into elemental sulfur, a sample was milled into a powder and washed with THF and CS2. PXRD (powder x-ray diffractometry) and DSC (differential scanning calorimetry) measurements of powders aged for two months confirmed that no crystalline sulfur had formed ( Figure S4). This could be concluded from the absence of a sulfur melting peak (DSC) or sulfur crystal reflexes (PXRD). Crosslinked net-poly(Sn-r-StyTMS) was stable until ca. 220 °C in the presence of oxygen, where it began to decompose (Figure S4 C). Decomposition products above 220 °C were identified to be H2O, CO2, and SO2 by a TGA coupled mass spectrometer.
Non-destructive DSC measurements below the degradation zone led to the same result, i.e. no sulfur melting peak was observed ( Figure S4 D).       higher than S4were detected, while for elemental sulfur fragments up to S8could be observed.
Elemental sulfur was not distributed homogenously on the surface due to its fast crystallization upon evaporation of CS2. Instead, small crystallites formed. Figure S13. ToF-SIMS peak area integration for Snfragments for n = 1 -8 for different samples: Elemental sulfur spin-coated from CS2 on glass (violet) and on silica nanoparticles (blue) as well as netpoly(Sn-r-StyTMS) coated on glass (green) and on silica nanoparticles (orange). For the comparison in Figure 3D the silica particle free substrates were used. For net-poly(Sn-r-StyTMS) no significant amount of fragments higher than S4were detected, while for elemental sulfur fragments up to S8were observed.  Figure S14. 29 Si HPDEC-MAS NMR spectrum of powdered net-poly(Sn-r-StyTMS) and a background spectrum (rotor). There is an unknown Q species, which is also present in the background control spectrum. Thus, in the sample spectrum only the T1-3 species belong to the sample. Via line fit integration of T1, T2, and T3 an integral ratio of 4.5, 10.6, and 1.0, respectively, was obtained.