Abstract
Metal-organic frameworks (MOFs), with their well-defined and highly flexible nanoporous architectures, provide a material platform ideal for fabricating sensors. We demonstrate that the efficacy and specificity of detecting and differentiating volatile organic compounds (VOCs) can be significantly enhanced using a range of slightly varied MOFs. These variations are obtained via postsynthetic modification (PSM) of a primary framework. We alter the original MOF's guest adsorption affinities by incorporating functional groups into the MOF linkers, which yields subtle changes in responses. These responses are subsequently evaluated by using machine learning (ML) techniques. Under severe conditions, such as high humidity and acidic environments, sensor stability and lifespan are of utmost importance. The UiO-66-X MOFs demonstrate the necessary durability in acidic, neutral, and basic environments with pH values ranging from 2 to 11, thus surpassing most other similar materials. The UiO-66-NH(2) thin films were deposited on quartz-crystal microbalance (QCM) sensors in a high-temperature QCM liquid cell using a layer-by-layer pump method. Three different, highly stable surface-anchored MOFs (SURMOFs) of UiO-66-X obtained via the PSM approach (X: NH(2), Cl, and N(3)) were employed to fabricate arrays suitable for electronic nose applications. These fabricated sensors were tested for their capability to distinguish between eight VOCs. Data from the sensor array were processed using three distinct ML techniques: linear discriminant (LDA), nearest neighbor (k-NN), and neural network analysis methods. The discrimination accuracies achieved were nearly 100% at high concentrations and over 95% at lower concentrations (50-100 ppm).