Green chemistry and engineering are creating a culture change for the future direction of industry1,2,3,4. Continuous flow technology has been identified as instrumental for its environmental and economic advantages leveraging superior mixing, heat transfer, and cost savings through the "scaling out" strategy as opposed to the traditional "scaling up"5,6,7,8,9,10.
Although the industries producing high-value products like the pharmaceutical industry have long favored batch processing, the advantages of flow technology have become attractive due to mounting economic competition and commercial production benefits11. For example, when scaling up batch processes, pilot scale units must be built and operated to ascertain accurate heat and mass transfer mechanisms. This is hardly sustainable and subtracts substantially from the marketable patent life of the product. In contrast, continuous flow processing allows for the advantages of scale out, eliminating the pilot-plant phase and engineering associated with production scale-a significant financial incentive. Beyond the economic impact, continuous technology also enables atomic and energy efficient processes. For instance, enhanced mixing improves mass transfer for biphasic systems, leading to improved yields, catalyst recovery strategies, and subsequent recycling schemes. Additionally, the ability to accurately manage the reaction temperature leads to precise control of reaction kinetics and product distribution12. The enhanced process control, quality of product (product selectivity) and reproducibility are impactful both from environmental and financial standpoints.
Flow reactors are available commercially with a wide variety of sizes and designs. In addition, customization of reactors to meet process needs can easily be achieved. Herein, we report experiments conducted in a glass continuous flow reactor (Figure 1). The assembly of microstructures (161 mm x 131 mm x 8 mm) made of glass is compatible with a wide range of chemicals and solvents and is corrosion-resistant over a wide range of temperatures (-25–200 °C) and pressures (up to 18 bar). The microstructures and their arrangement were designed for multi-injection, high-performance mixing, flexible residence time, and precise heat transfer. All of the microstructures are equipped with two fluidic layers (-25–200 °C, up to 3 bar) for heat exchange on either side of the reaction layer. Heat transfer rates are proportional to the heat transfer surface area and inversely proportional to its volume. Thus, these microstructures facilitate an optimum surface-to-volume ratio for improved heat transfer. There are two types of microstructures (i.e. modules): "mixing" modules and "linear" modules (Figure 2). The heart-shaped "mixing" modules are designed to induce turbulence and maximize mixing. In contrast, the linear modules provide additional residence time.
As proof of concept, we selected the well-described reaction of diphenyldiazomethane with carboxylic acids13,14,15,16,17. The reaction scheme is shown in Figure 3. The initial transfer of the proton from the carboxylic acid to the diphenyldiazomethane is slow and is the rate-determining step. The second step is rapid and yields the reaction product and nitrogen. The reaction was initially investigated to compare relative acidity of organic carboxylic acids in organic solvent (aprotic and protic). The reaction is first-order in the diphenyldiazomethane and first-order in carboxylic acids.
Experimentally, the reaction was conducted in presence of large excess of carboxylic acid (10 molar equivalents). As a consequence, the rate was pseudo first order with respect to the diphenyldiazomethane. The second order rate constant can then be obtained by dividing the experimentally obtained pseudo first order rate constant by the initial concentration of the carboxylic acid. Initially, the reaction of diphenyldiazomethane with benzoic acid (pKa = 4.2) was investigated. In batch, the reaction appeared to be relatively slow, reaching about 90% conversion in 96 minutes. As the reaction rate is directly proportional to the acidity of the carboxylic acid, we chose as a reaction partner the more acidic carboxylic acid, p-nitrobenzoic acid (pKa =3.4) to shorten the reaction time. The reaction of p-nitrobenzoic acid with diphenyldiazomethane in anhydrous ethanol was thus investigated in batch and flow (Figure 4). The results are provided in detail in the following section.
When the reaction is carried out in ethanol, three products can be formed: (i) benzhydryl-4-nitrobenzoate, which results from the reaction of p-nitrobenzoic acid with the diphenylmethane diazonium intermediate; (ii) benzhydryl ethyl ether that is obtained from reaction of the solvent, ethanol, with the diphenylmethane diazonium; and (iii) nitrogen. The product distribution was not studied as it is well documented in literature; rather we focused our attention to the technology transfer of the batch reaction to continuous flow13,14,15. Experimentally the disappearance of the diphenyldiazomethane was monitored. The reaction proceeds with a vivid color change, which can be visually observed by UV-Vis spectroscopy. This results from the fact that the diphenyldiazomethane is a strongly purple compound whereas all other products from the reaction are colorless. Therefore, the reaction can be visually monitored on a qualitative basis and quantitatively followed by UV spectroscopy (i.e. disappearance of the diphenyl diazomethane absorption at 525 nm). Herein, we first report the reaction of diphenyldiazomethane and p-nitrobenzoic acid in ethanol in batch as a function of time. Secondly, the reaction was successfully transferred and carried out into the glass flow reactor. The progress of the reaction was ascertained by monitoring the disappearance of diphenyldiazomethane using UV-spectroscopy (in batch and flow modes).