Figure 2. (a) Comparison of FTIR spectra of glycerol and glycerol/inhibitor mixture before (control) and after 5 min PTFE deposition. Here, to make the substrate, an equal amount (30 μl) of corresponding solution was spin-coated on plasma-cleaned Si substrates, with comparable areas. Then, PTFE deposition was performed in the iCVD reactor within 5 min at T_S=15 ^oC and P=300-1200 mTorr. (b) The reaction pathway for thermal hemolysis activation of the initiator in the gas phase followed by quenching the primary free radicals on the liquid surface.

As mentioned earlier, for the samples prewetted by the glycerol/HQ, both chemical and physical mechanisms are in action. The deposition was limited to the top surfaces and the liquid prevents the reactants from coating the inner walls of the porous support. At the same time, HQ quenches the free radicals absorbed into glycerol. Figure 2 (b) shows the pathway for thermal hemolysis (activation) of the initiator (perfluorobutane sulfonyl fluoride),³¹ and the proposed reaction for chemical inhibition. We propose that the primary radicals (), adsorbed on the surface of glycerol/inhibitor mixture, are quenched by HQ, through electron transfer.^32,33 The FTIR results of the glycerol/inhibitor mixture after primary radical adsorption also showed that HQ is reduced to quinone, confirming that HQ inhibits polymerization by quenching the radicals; see Section S3, Supporting Information.

To investigate the effectiveness of physicochemical inhibition, we evaluated the surface porosity of PVDF support by measuring nitrogen permeation at various transmembrane differential pressures. The PTFE deposition was performed at different pressures ranging from 300 to 1200 mTorr. Figure 3 (a) compares the nitrogen permeation for the pristine PVDF support with those of coated samples prepared at different iCVD pressures. As shown, we did not observe any significant reduction in nitrogen permeation after depositing of PTFE. This observation indicates the deposition was limited to the solid domains, and the surface porosity of the membranes was preserved.

To measure the wettability of the substrates, we performed the water contact angle (CA) measurement on the top and bottom surface of the processed membranes. Figure 3 (b) shows that the water CA on the top surface of all samples was above 140 degrees. In contrast, the water droplet placed on the back side of the membrane wicks through. We note that the contact angle of water on the top surface of the Janus membranes increases with increasing the pressure of iCVD processing. Section S4, Supporting Information, includes the SEM images showing the effect of varying the deposition pressure on the morphology of Janus membrane surfaces. Here we attribute the observed trend in the membranes’ CA to the induced roughness as a result of a transitionig the deposition process from the adsorption limited to a gas-phase limited one ³⁴ Figure 3 (c) also shows the optical image of water and ethylene glycol droplets placed on the top and the back side of the Janus membranes. As shown, both liquids formed contact angles above 140 degrees on the top surface of the Janus membrane. However, the contact angle of all liquids on the back side was zero. This observation shows that the back side of the membrane remained pristine. The X-ray photoelectron spectroscopy (XPS) C1s core electron spectra collected from the bottom surface of the coated membrane also matched that of a pristine PVDF; see Section S5, Supporting Information.

To examine the underwater oleophobic properties of the hydrophilic side of the Janus membranes, we measured the contact angle for oil droplets when the hydrophilic side is immersed in water. Figure 3 (d) displays the side-view optical image from a fabricated Janus membrane placed on the water surface with the hydrophilic side facing the water. As shown, the oil droplet has a contact angle of ~140 degrees with the membrane and does not wet the surface, confirming underwater oleophobicity of the untreated side. To illustrate this functionality more clearly, we recorded a video of the testing procedure of a Janus membrane, shown in Movie S2, Supporting Information. As shown, after suspending a Janus membrane on the top surface of a water bath, water penetrates the Janus membrane such that the membrane becomes semi-transparent. However, the thin PTFE layer (2-3 μm) hinders water from reaching the top surface. As a result, when water droplets impinged on the top surface of the Janus membrane, they bounced back off the surface due to the air gap formed within the PTFE domain; see Movie S2, Supporting Information.