Thin Composite Carbon Molecular Sieve Membranes from a Polymer of Intrinsic Microporosity Precursor

Abstract

Ultra-thin composite carbon molecular sieve (CMS) membranes were fabricated on well-defined inorganic alumina substrates using a polymer of intrinsic microporosity (PIM) as a precursor. Details of the pyrolysis-related structural development were elucidated using focused-beam, interference-enhanced spectroscopic ellipsometry (both in the UV–vis and IR range), which allowed accurate determination of the film thickness, optical properties as well as following the chemical transformations. The pyrolysis-induced collapse of thin and bulk PIM-derived CMS membranes was compared with CMS made from a well-known non-PIM precursor 6FDA–DABA. Significant differences between the PIM and non-PIM precursors were discovered and explained by a much larger possible volume contraction in the PIM. In spite of the differences, surprisingly, the gas separation properties did not fundamentally differ. The high-temperature collapse of the initially amorphous and isotropic precursor structure was accompanied by a significant molecular orientation within the formed turbostratic carbon network guided by the laterally constraining presence of the substrate. This manifested itself in the development of uniaxial optical anisotropy, which was shown to correlate with increases in gas separation selectivity for multiple technologically important gas pairs. Reduction of CMS skin thickness significantly below ∼1 μm induced large losses in permeability coefficients with only small to moderate effects on selectivity. Remarkably, skin thickness reduction and physical aging seemed to superimpose onto the same trend, which explains and strengthens some of the earlier fundamental insights.

Nano-Confinement Effects on Structural Development and Organic Solvent-Induced Swelling of Ultrathin Carbon Molecular Sieve Films

“…We note here that in spectroscopic ellipsometry modeling thickness and refractive index data are to a large extent independent from each other and thus represent independent pieces of information about the properties of the pyrolyzed films. From Figure 3 a it is clear that progressively more volumetric collapse occurred going from the self-supported, ∼70 μm 15 to ∼300 nm and ∼30 nm films. At 700 °C, the ∼30 nm confined films collapsed approximately two times more (∼43%) as compared with the ∼300 nm films (∼22%).…”

“…The polymer combined a PIM character (high internal surface area and a rigid backbone) with a very high aromatic carbon content (84 wt %) which in our earlier work was proven to result in high-performance carbon molecular sieve membranes following pyrolysis. 15 …”

“…The presence of the ∼500 nm thick silicon oxide underneath each polymer/carbon film amplified the accuracy of the analysis versus the more broadly used native oxide wafers, as thoroughly investigated before. 15 , 32 , 33 An example of the optical model is presented in the Supporting Information (Figure S2). Error bars were estimated based on the fit parameter uniqueness analysis assuming as threshold a 25% reduction of the mean squared error (MSE).…”

“… 12 , 13 However, similarly to other high free volume amorphous membrane materials, CMS membranes suffer from a progressive structural densification (physical aging) leading to significant reductions of permeance. 14 , 15 Moreover, physical aging, and to some degree dilation (or swelling) induced by condensable gases like CO 2 or organic vapors, are well-known to amplify in thinner films, particularly in the submicron region. 16 − 19 …”

Nano Sn2S3 Embedded in Nitrogenous‐Carbon Compounds for Long‐Life and High‐Rate Cycling Sodium‐Ion Batteries

“…In addition, no diffraction peaks of carbon are detected by XRD because the carbon synthesized by the pyrolysis of high-molecular polymer below 700°C is amorphous. [23,24] According to the above discussion, it can be deduced that the Sn/CN compounds could be prepared successfully by using SnO 2 or SnCl 4 as the Sn-sources. Furthermore, the XRD patterns of heated Sn + S powder (defined as S-0) and Sn/CN + S powder were compared.…”