Simultaneous Spin Coating and Ring-Opening Metathesis Polymerization for the Rapid Synthesis of Polymer Films

We report a highly controlled technique for the synthesis of polymer films atop a substrate by combining spin coating with ring-opening metathesis polymerization (ROMP), herein termed spin coating ROMP (scROMP). The scROMP approach combines polymer synthesis and deposition into one process, fabricating films of up to 36 cm2 in under 3 min with orders-of-magnitude reduction in solvent usage. This method can convert numerous norbornene-type molecules into homopolymers and random copolymers as uniform films on both porous and nonporous substrates. Film thickness can be varied from a few hundred nanometers to a few tens of micrometers based on spin speed and monomer concentration. The resulting polymers possess high MW (>100 kDa) and low polydispersity (PDI) (<1.2) values that are similar to ROMP polymers made in solution. We also devise a model to investigate the balance between convective monomer spin-off and polymer growth from the surface, which allows the determination of critical kinetic parameters for scROMP. Finally, translation of scROMP to porous supports enables the synthesis of thin film composite membranes that demonstrate the ability to dehydrate ethanol by pervaporation.


SI.1 Synthesis Routes for scROMP on Various Substrates
Synthesis of a pNBD film by scROMP atop a pristine Au-coated wafer.The left pathway yields a robustly attached polymer film whereas the right pathway yields a film that can be peeled away, intact, from the surface.

SI.2 Polymer Film Stability Testing
Polymer film stability on PAN supports was evaluated by flowing a 90/10 v/v% EtOH heated to 55 ºC across a 4 cm 2 polymer film for 6 h, similar to the environment for the ethanol dehydration studies.Prior to flow testing, a pNBF8 film was analyzed with ATR-FTIR to establish a baseline  The similarity in these spectra indicates that the polymer film is robustly adhered to the PAN support and is not desorbed or eroded during membrane flow testing.

SI.3 Profilometric Measurements of Polymer Film on Au Substrate
Stylus profilometry was performed to quantify film thickness and surface roughness.Profilometry measurements were conducted after performing scROMP using neat NBDAC.Prior to profilometry, a portion of the pNBDAC film was removed to expose the Au substrate so that film height could be measured relative to the substrate.Over the sampled distance, the film shows a uniform thickness of 20 µm and a comparably low surface roughness (Ra = 85 nm).

SI.4 Surface Roughness of PAN and Au-Coated Wafer
The surface roughness of the PAN support and Au-coated wafer were quantified by stylus profilometry.Line scans across the surface of both substrates are shown below in  Line scans are offset to provide better visualization of the roughness of the surfaces.
Profilometry of the pristine surfaces shows that the non-porous Au substrate has a very low surface roughness (Ra = ~3 nm) whereas the porous PAN surface has orders-of-magnitude higher surface roughness (Ra = ~3 µm).We attribute the higher dependence of w observed on PAN to the increased surface roughness and presence of porosity.

SI.5 Impact of Monomer Dispense Volume on Film Thickness
The amount of monomer dispensed on the surface during the polymerization step of scROMP was assumed to impact the resulting film thickness.To investigate this, polymer films were made atop a 4 cm 2 Au-coated wafer at varying polymerization spin speeds using 100 and 200 µL of NBDAC as shown in Figure S.5.This study revealed that there is no significant difference between the final film thickness when either 100 µL or 200 µL of the neat liquid monomer NBDAC is used.

SI.6 Calculation of Active Catalyst Concentration from GPC/SEC
To determine the amount of G3 that participates in the scROMP process, we assumed that each polymer chain contains only one catalyst molecule and that chain transfer is negligible 1 .The number of moles of catalyst used in the polymerization was then determined by where gpolymer is the mass of the polymer synthesized and MW,polymer is the polymer molecular weight.These calculations were carried out for two systems.The first system was generated by dispensing 100 µL of G3 onto a static Au-coated wafer at various concentrations of G3 in DCM (Table S.1).Following solvent evaporation, a 1M solution of NB was dispensed onto the surface at a polymerization spin speed of 3000 RPM to yield pNB.The second system was generated by varying the spin speed during polymerization of NBDAC (Table S.2).First, a 5 mM G3 solution in DCM was dispensed onto a Au-coated wafer by spinning at 1000 RPM for 30 s.Then, neat liquid NBDAC monomer was dispensed on the surface at 1000, 2000, and 4000 RPM and allowed to polymerize for 60 s.After 60 s had elapsed, the surface was flooded with ethyl vinyl ether to terminate the polymerization reaction.The generated pNBDAC film was subsequently immersed into a 90/10 v/v% solution of THF/EtOH.This converted the acyl chloride side chains present on pNBDAC into ethyl ester side chains and yielded a THF soluble film that was then used for GPC/SEC analysis.The results in Table S.1 indicate that the scROMP approach is capable of synthesizing polymer films with high MW and low PDI, which is attributed to the fast initiation and propagation rates of G3.The polymer synthesized in the presence of 500 nmol of G3 is an extreme case where all catalyst dispensed on the surface while spinning would be retained.This perfect retention of catalyst is an unrealistic scenario for the scROMP approach and shows lower MW and higher PDI due to the participation of excess catalyst in the polymerization.The polymers synthesized using 100 and 50 nmol G3 are more representative of the scROMP process.These polymers show higher molecular weights and lower PDI, indicating that the scROMP approach can be used to produce well-defined polymers in ambient conditions with minimal solvent usage.
The results shown in Table S.2 demonstrate the impact that polymerization spin speed has on the resulting polymer MW and amount of catalyst participating in the polymerization.As spin speed is increased, the amount of catalyst initiating and participating in propagation is decreased.We attribute the reduction in active G3 to higher shear stresses at the surface at increasing spin speeds.
Polymer MW increases with increasing spin speed, which we attribute to reduced competition between initiated catalysts to consume monomer for polymer chain propagation.

SI.8 GPC/SEC Data of p(NB-co-NBF4) Polymers
Polymer chain lengths and molecular weight distributions for p(NB-co-NBF4) were determined using GPC.The degree of polymerization (DP) was determined by dividing the mass-average molecular weight (MW) by the molar mass of the repeat unit, resulting in the average number of repeat units per chain.For copolymers, DP was determined through the same approach using the NB:NBF4 NMR ratios presented in Table 3 of the main text.The equation for DP calculation of copolymers is written as: (&'(% +,).$ ( )* /0(&'(% +,12).$( )*+, / (S.2) Both number-and weight-average molecular weights increased with fluorocarbon concentration, while polydispersity and degree of polymerization decreased.Increases in average molecular weights were expected since the NBF4 repeat unit is ~3x the molecular weight of a NB repeat unit.
The degree of polymerization decreased with increasing NBF4 monomer concentration; however, NMR data indicate that the NBF4 monomer is more rapidly incorporated into the polymer film than the NB monomer is.The presence of the NBF4 must hinder the polymerization of NB enough to decrease the overall degree of polymerization in monomer systems with high NBF4 ratios.The fluorocarbon and hydrocarbon monomers are largely incompatible, so the NB may be repelled from the catalyst, decreasing the overall number of monomers polymerized.The polydispersity index is lower for systems with high amounts of NBF4 since the presence of the fluorocarbon chain may decrease the likelihood for chain terminations, such as backbiting by the catalyst, 3 by screening susceptible olefinic bonds.

Table S.3.
Polydispersity and molecular weights of p(NB-co-NBF4) films using GPC.Degree of polymerization was calculated from Mw and 1 H NMR data.Films synthesized using ratios less than 1:1 were not soluble enough in THF to produce meaningful signals.

SI.9 Membrane Pervaporation Testing Setup
Membranes were tested in a pervaporation setup.A schematic of the equipment train is shown below in Scheme S.1.Briefly, a solution of 90/10 v/v% EtOH/H2O was added to a feed vessel.

Figure S. 1 .
Figure S.1.Schematic depicting two pathways for synthesizing pNBD films.(Left) Synthesis of absorbance.Following flow testing, the same film was reanalyzed with ATR-FTIR to observe the change in absorbance as shown in Figure S.2.

Figure S. 3 shows
an example profilometric line scan of a polymer film on a Au-coated wafer.

Figure S. 3 .
Figure S.3.Profilometric line scan of pNBDAC on Au by scROMP.The pNBDAC film was made Figure S.4.

Figure S. 4 .
Figure S.4.Profilometric line scans of the surface of PAN (black) and a Au-coated wafer (red).

Figure S. 5
Figure S.5 Profilometric measurements of films made on Au using 100 and 200 µL of NBDAC.

1 H
NMR for pNB shown in Figure S.6 is consistent with literature spectra for pNB 2 and features a trans to cis ratio of 1:1.32.Referencing the notation used in Figure S.6., (CDCl3): d cis: (a) 5.21 (2H); (b) 2.78 (2H); (c, e) 1.83 (3H); (d) 1.35 (2H); (f) 1.03 (1H); trans: (g) 5.34 (2H); (h) 2.43 (2H); (i, k) 1.83 (3H); (j) 1.35 (2H); (l) 1.03 (1H).Moisture from the CDCl 3 appears at 1.55 ppm, and pentane absorbed in the film from dissolving the monomers appears at 1.30, 1.26, and 0.88 ppm.The 1 H NMR spectrum for pNBF4 is considerably less straightforward due to the varied placement of the perfluoro chain along the hydrocarbon backbone.For example, six different chemical shifts for just the olefinic protons could be observed from the conformations presented in Figure S.7.However, the peaks in Figure S.8. at 3.25 and 2.05 ppm only appear in the pNBF4 spectra and have been used to determine the relative composition of the monomer units within the film.We conclude that the 3.25 ppm peak represents the proton labeled x in the cis conformation of Fig. S.8., and the 2.05 ppm peak represents the c* and e* protons in the cis conformation and the i* and k* protons in the trans conformation.The integrations are referenced relative to the olefinic area between 5.5 and 5.0 ppm, with the ratio between the 3.25 ppm peak and the 5.5-5.0 ppm area in the NBF4 representing a 100% pNBF4 composition.Values for composition deviated by <5% when using either the 3.25 or the 2.05 ppm peaks.

Figure S. 7 .
Figure S.7.Different conformations and placements of the fluorocarbon chain in pNBF4 affect

Figure S. 9 .
Figure S.9.HSQC for the pNBF4 film synthesized using scROMP with axis F2 showing the 1 H Scheme S.1.Schematic of membrane pervaporation setup used for ethanol dehydration.

Table S . 1 .
Molecular weight, polydispersity, and calculation of mols of G3 utilized during scROMP of NB with different amounts of G3 on the surface.All films were made using a 1M solution of NB in DCM spun on at 3000 RPM.

Table S . 2
Calculation of mols of G3 utilized in the polymerization of NBDAC at various spin speeds.