Functionalization of 3D-Printed Plastics for the Photocatalytic Removal of Organic Pollutants in Air

3.1. Materials and Coatings Morphology A microscopic inspection of the coatings over the different substrates was performed. Samples have been inspected and photographed by scanning electron microscopy (SEM) (Phenom-World Pro). The plastic samples (1 cm × 1 cm) were put inside the vacuum chamber and observed at different magnifications; the selected images are shown below in Figure 2. Figure 2 shows micrographies of the samples after 8 coating cycles, at 3000× magnification for each tested material. A complete coverage of all the polymeric substrates can be seen, with a reasonable uniformity distribution of the catalyst over the entire surface. In the PLA sample, small micro-cracks can be seen, probably due to the thermal expansion of the film, while in PETG, some regions with micro-clusters of TiO₂, from 0.3 to 0.47 μm in size, can be observed, which may be due to imperfections in the material surface. The ABS surface shows the most homogeneous distribution of the catalyst. 3.2. Catalyst Load During the evaluation process of the materials, one of the first parameters that were considered was the catalyst load on different samples. The weight of each sample was recorded after each immobilization cycle, which included dip-coating and drying. As shown in Figure 3, the amount of fixed catalyst increased proportionally with the number of coating cycles. All three materials incorporated catalysts in the same order of magnitude, with PLA having the largest load per unit area. PETG, on the other hand, had a lower capacity for adding catalysts. After 12 immobilization cycles, the superficial catalyst loads achieved for the different materials are approximately 2.75 mg·cm⁻² for PETG, 3.36 mg·cm⁻² for ABS, and 3.6 mg·cm⁻² for PLA. To provide a reference point regarding the capacity of these materials to support the catalyst, we can compare them to coatings on stainless steel meshes (25 mesh size). Following a procedure reported in the literature [24], after performing 4 coating cycles, a catalyst load of approximately 3.0 mg·cm⁻² is reached. 3.3. Catalyst Adherence Test The catalyst film is adhered to the substrate but may eventually be detached. In this sense, we selected an additional parameter for the substrate comparison, a test for catalyst adherence. To evaluate the resistance of the adhered films in high-stress conditions, the coated samples were tested in an aqueous medium using the ultrasonic method, as per the methodology described in the literature [25]. Although the coated materials will be used for gaseous streams in our case, a milder condition, we deemed this to be a valid comparative examination. The test consists of the measurement of the weight loss caused by the exposition of the sample to ultrasound waves. The coated samples were immersed in distilled water and then subjected to a 2.5-min ultrasonic bath using a Biobase^® Ultrasonic Cleaner at maximum power and room temperature. Following this, the samples were dried on a stove at 50 °C for 4 h. The weight loss was calculated by comparing the mass of the samples before and after the ultrasonic test. The results of the adherence test showed a lower adherence of the coating on PETG with a remaining adhesion (expressed as the percentage of remaining TiO₂ mass) of approximately 45%. On the other hand, ABS and PLA presented similar behavior, with a remaining adherence of between 67 and 71% for ABS and between 70 and 72% for PLA. The adhesive and cohesive strength of TiO₂ films could be of further investigation—but left out of our scope—considering factors such as thickness, application techniques, pretreatments of the substrates, and number of coating cycles. Nonetheless, it’s important to note that the adhesion testing method used here puts more stress on the materials than what is expected in gas phase operation. During the photocatalytic activity tests, there were no instances of catalyst detachment observed. 3.4. Optical Properties The absorbed fraction of a film supported on any material, defined as the ratio of absorbed radiation to incident radiation, is an indirect property that can be calculated from reflectance and transmittance measurements. Among the materials studied, PLA and ABS are opaque within the relevant wavelength range. In contrast, PETG is semi-transparent, allowing for the application of models to determine the radiation absorbed fraction of catalyst layers as a function of wavelength. In this discussion, we will propose a simple model to explain the chemical conversion of DCM based on the absorbed fraction. We will focus exclusively on PETG, as it has measurable, non-zero transmittance, while PLA and ABS are opaque, resulting in negligible transmittance for those materials. The reflectance and transmittance of bare support materials and coated samples were measured in the range of lamp emission power wavelengths (300–400 nm, 5 nm intervals) and depicted in Figure 4. Figure 4 illustrates that the spectral reflectance curves for the three uncoated plastics differ significantly, particularly the curve corresponding to PETG. In contrast, the total reflectance curves of all coated samples are nearly identical. This similarity can largely be attributed to the properties of the catalyst. Because the catalyst loads are high and comparable (equal to or greater than 3 mg·cm⁻²), the films exhibit very similar optical behaviour, regardless of the support material used. Consequently, the radiation model developed for PETG can also be considered valid for ABS and PLA. Based on the net-radiation method proposed by Howell et al. [26], it is possible to calculate the absorbed fraction (A) of the photocatalyst layer simply by: In Equation (1), T is for transmittance, and R is for reflectance. Notice that these are spectral properties; in the following equations, the subscript $$\lambda$$ will be dropped for the sake of clearness. The values of T and R of the catalyst layer are derived from the measured values of diffuse reflectance and diffuse transmittance of both the bare and the TiO₂-coated materials. Given that the plastic samples are coated on both sides, the equations relating to the optical properties are [27]: where the subscript c refers to the catalyst layer, s to the support material, and g to the global optical property of the composite medium: catalyst-support-catalyst. a radiation balance for the catalyst (Equation (1)), and the obtained values are plotted in Figure 5. The parametric curves show an increasing absorbed fraction with the number of catalyst immobilization cycles. The normalized radiative output power of the lamps employed can weigh the curves plotted in Figure 5. Calculating the incident radiation flux over the samples would yield the local superficial rate of photon absorption for each sample. The y-axis in Figure 5 is a normalized scale; it represents the spectral absorbed fraction of radiation (parametric curves) and the spectral emission of the lamp, relative to its peak. It is worth noting the steep decline in the absorbed fraction to reach almost null absorption for wavelengths larger than 380 nm; this result is consistent with the bandgap absorption of the P25 catalyst. 3.5. Photocatalytic Activity Tests To investigate the degradation of DCM in the reaction device, several blank tests were conducted. Firstly, the air supplied by the purification system was analyzed and no peaks were detected in the chromatogram, indicating no unwanted substances were present. Secondly, a blank test for photolysis was carried out by circulating the DCM mixture in the reactor with an uncoated plate (with no catalyst) and exposing it to maximum power. Finally, an additional test was done using the TiO₂-coated samples inside the reactor and lamps powered off. There was no significant difference between the concentrations of DCM at the inlet and outlet of the tests, suggesting steady-state operation without photolysis or dynamic adsorption phenomena. Experimental runs were conducted for each material sample, with identical dimensions, by varying catalyst coating cycles to measure photocatalytic activity. To ensure comparability of results, all operating variables were controlled, including inlet DCM concentration, total flow rate, radiation flux, and relative humidity. The details concerning operating conditions are summarized in Table 2. Each experimental run in the continuous reactor requires measurement of the inlet (C_in) and outlet (C_out) concentrations of DCM. At the beginning of the air circulation, before the lamps are turned on, samples are taken until C_in stabilizes at a constant value. Once this is achieved, the lamps are turned on; during the illuminated period samples are taken until C_out no longer changes over time. Finally, the lamps are turned off to check that the values of C_in (which is monitored during the whole reaction) and C_out match in the absence of radiation. The steady-state concentrations are then used to calculate the DCM conversions for each sample by: Table 3 summarizes the experimental results of DCM degradation for different coating cycles for the three plastic samples. All runs were done with a 50 ppm inlet concentration and maximum radiation. The variability indicated in the conversion column corresponds to one standard deviation of the mean value. Throughout the use of the coated plates, there were no signs of aging or catalyst poisoning; different runs were repeated, and no significant changes were observed as a function of the time of use. From the results in Table 3, it is evident that, although the conversions get higher when increasing the number of coating cycles, the intrinsic performance does not follow such a trend. The last column in Table 3 represents the DCM destruction efficiency per mass of the catalyst, which indicates the degree of utilization of each immobilized TiO₂ particle. As more layers of catalyst particles are added, it is natural that the first ones are not exposed to radiation or come in contact with the pollutant. Consequently, the ratio of pollutant moles reduced per mass of catalyst has an inverse relation with the number of coating cycles. The results in the last column of Table 3 show a decrease in the DCM concentration per gram of fixed catalyst (ppm/g) of 45.0 (ABS), 38.6 (PETG), and 26.8 (PLA). These values may be contrasted with a previously published study [24]. Using a stainless steel mesh as support, with four coating cycles, 45.9 ppm of DCM were converted per gram of fixed TiO₂, under similar conditions. The result is very similar to that obtained using ABS in the present study. It is worth noting that the comparison is performed against the number of coating cycles because of the different catalyst loadings achieved in the three plastics. By observing the results of the absorbed fraction by the catalyst layers, a saturation phenomenon could be seen. This allows us to propose a simple model of the absorbed fraction as a function of the superficial catalyst load (L_c) in the following terms: where A_c,∞ represents the absorbed fraction of an infinitely thick layer of catalyst, and k_c is a fitting parameter. Figure 6 shows the experimental points and the fitted model. The obtained values of the parameters are A_c,∞ = 0.8439 ± 0.0201 and k_c = 0.8446 ± 0.0624 mg·cm⁻²; with R-squared = 0.99955. According to the obtained parameters for Equation (5), it would be enough to reach approximately 3 mg/cm² of TiO₂ since for this value the absorbance no longer increases because there is a saturation phenomenon. An additional result is the trend of the DCM conversions versus the TiO₂ loading for all three materials studied, which is shown in Figure 7. It is evident that the support material loses relevance and the conversion is merely due to the attached photocatalyst. Also noticeable is a nonlinear trend for conversion, which can also be thought of as a saturation phenomenon. A Langmuir-type function was proposed to fit the experimental points: The parameter estimation yielded the following values: X_max = 24.548 ± 1.276 (%) and k_x = 0.6946 ± 0.1197 (mg·cm⁻²); with R-squared = 0.9711. According to this model, the maximum achievable conversion, for L_c tending to infinity, is 24.55%. Based on the model, adding more catalyst layers will increase pollutant conversion, but the efficiency gain will decrease. This is because the photons and pollutant molecules can’t reach the deepest layers of the catalyst. This raises the question of how many cycles of catalyst immobilization should be done. As the number of coating cycles increases, pollutant conversion increases but with diminishing returns. In this sense, it would be reasonable to think that there is an economic optimum for the catalyt’s loading. Additionally, there could be issues with adherence as the catalyst layer becomes thicker. Here it is very important to clarify how the variables used affect the decomposition reaction rate of the pollutant (r_DCM). This rate, averaged over the reactor volume, can be obtained from a global mass balance for DCM, resulting in: Given fixed values of the flow rate (Q), the catalytic area (A_cat) and the reactor inlet concentration (C_in) in Equation (7), as in our case, the reaction rate is directly proportional to the conversion. Therefore, whether we talk about conversion or reaction rate makes no difference. The catalyst load has been used as a reference for analysis as a fundamental variable. However, since the relationship between the load and the absorbed fraction (as illustrated in Figure 6) and the relationship between load and conversion or reaction rate (shown in Figure 7) demonstrate nearly identical behavior, it is evident that the key variable influencing the reaction rate or conversion is the absorbed fraction. Therefore, in photocatalytic reactors, the primary factor is the absorbed fraction of radiation, which directly reflects the catalyst’s efficiency in absorbing radiation for the reaction. This efficiency should be incorporated as a factor in a kinetic expression, similar to how catalyst efficiency is represented in conventional (thermal) fixed-bed catalytic reactors. The radiation absorption efficiency implicitly contains the catalyst loading via Equation (5). As a novelty, we can demonstrate this by plotting the conversion, or reaction rate (using Equation (7)) as a function of the absorbed fraction, as shown in Figure 8. The linear adjustment yielded the following parameters: slope = 30.96 ± 0.66 with an R-square: 0.9955. Finally, to make a well-founded selection of the most suitable plastic for its functionalization, a multi-parametric comparison is made. The comparison criteria in Table 4 include intrinsic material properties as well as photocatalytic performance-related results. Table 4 summarizes three major useful items for comparison: (i) material properties, (ii) manufacturing process, and (iii) performance. Recalling the information in Table 1 we can explain the rating in the material properties variables: PLA, a biopolymer derived from natural sources such as corn starch or sugar cane, is biodegradable under certain conditions; PLA polymers have been identified as a potentially viable alternative to petroleum-based products. Bio-based degradable polymers have been demonstrated to have numerous applications in the biomedical field and are used as disposable packaging materials [28]. PETG is not directly biodegradable and it is persistent in the environment, yet it can be recycled. ABS, on the other hand, is derived from petroleum and is not biodegradable, which poses an environmental issue. According to Table 1, PLA is the most durable material tested when it comes to aging under UV exposure. Concerning the manufacturing process of the samples, and based on our own experience, PETG is the most challenging material in terms of printability. When melted, PETG has a high viscosity, making it more susceptible to stringing, which can result in surface imperfections. Additionally, the material can adhere to the nozzle during printing, potentially causing clogs and loss of precision. To mitigate these issues, a higher temperature and proper retraction control settings are necessary. On the other hand, ABS is prone to warping, particularly in large or thin-walled parts, due to uneven cooling that creates internal stresses. This often leads to lifted corners or cracks between layers. These problems can largely be resolved by using higher temperatures during printing. In contrast, PLA is much easier to print and does not produce toxic fumes, making it a more user-friendly option. The low energy consumption parameter is related to the bed and extrusion temperatures in the printing process of the plastic support; PLA also has the advantage in this case. PETG is, out of the three plastics studied, the most difficult to manipulate when printing; although ABS is relatively easy to handle, it requires the highest temperatures. Finally, we analyzed performance aspects, based particularly on the results shown in Table 3. Although all three polymers provide similar DCM conversions, PLA seems the best choice for the functionalization of plastics to control airborne organic pollution. However, it must be noted that PETG would be the only viable option if the reactor’s internal configuration requires a semi-transparent support material.