A reactor engineering approach to describe bacterial inactivation during continuous UV-C light processing

Highlights

• Sucrose content, pH, and residence time (τ) affected the UV-C inactivation of Lactobacillus rhamnosus.

• Steady-state was reached faster and with a higher inactivation under the lowest τ.

• A variable order rate equation explains the unsteady-state inactivation of L. rhamnosus.

• Inactivation rate changes from first to near-zero order at about 10⁵ to 10^5.5 CFU/mL.

Abstract

A reactor engineering approach was used to mathematically describe microbial inactivation during continuous UV-C light processing of liquid foods. The method was followed to analyze the survival curves of Lactobacillus rhamnosus inoculated into sucrose model solutions prepared at different concentrations (8, 10 and 12 g sucrose per 100 g of solution) and pH values (pH 3, 4.5 and 6), and further processed at two different residence times (4.85 and 29.9 min). The inactivation process was considered as an irreversible elemental reaction of unknown order occurring in a continuous stirred tank reactor. The proposed model was expressed in terms of the logarithmic reduction in microbial population and its straight-line form allowed the easy estimation of the inactivation rate constant and reaction order. Results indicated that inactivation of L. rhamnosus followed a variable order kinetic, moving from a first-order rate during unsteady-state operation to a near zero-order inactivation when steady-state operation was reached. Steady-state was reached faster (0.48 ± 0.11 min⁻¹ vs. 0.37 ± 0.7 min⁻¹, p < 0.05) and with a higher steady-state log reduction (5.9 ± 0.2 vs. 5.4 ± 0.6 log CFU/mL, p < 0.05) in experiments conducted with the lowest residence time, UV-C dose and power density (4.85 min, 12.8 J/cm² and 6.6 J/cm³).

Introduction

Ultraviolet-C (UV-C) irradiation is a well-established technology used for the disinfection of drinking water and surfaces. This technology has been approved as a cold pasteurization alternative for liquid food products by the U.S. Food and Drug Administration; however, a minimum 5-log reduction in target pathogen numbers is required (FDA, 2001). In the last decade, UV-C light has been successfully applied to reduce the microbial load in several liquid foods, such as coconut water, grape must, juices (pure and blends), milk, among others (Antonio-Gutiérrez, López-Díaz, Palou, López-Malo, & Ramírez-Corona, 2019; Diesler et al., 2019; Fenoglio, Ferrario, Schenk, & Guerrero, 2019; Menezes, Tremarin, Junior, & de Aragão, 2019; Ochoa-Velasco et al., 2018).

In liquid foods, the effect of UV-C irradiation on microbial load reduction is studied in batch or continuous operation modes. Batch operation involves placing the liquid in a glass container (e.g., dish or beaker) within an UV-C irradiation chamber. The liquid is then irradiated during a predefined time corresponding to a known UV-C dose (Baysal, Molva, & Unluturk, 2013; Estilo & Gabriel, 2018). In some cases, the plots of log microbial counts versus UV-C processing time or irradiance deviate from the linear inactivation behavior, showing shoulders and/or tails (Gayán, Condón, & Álvarez, 2014a; Gabriel & Marquez, 2017; Gunter-Ward et al., 2017). Thus, log-linear plus tails, Weibull and other non-first-order models have been used to describe inactivation curves (Baysal et al., 2013; Menezes et al., 2019). Nonlinear inactivation behavior can arise under non-uniform dose distribution conditions and occurrence of microbial subpopulations (the test microbe exists in different physiological forms) (Fenoglio, Ferrario, García-Carrillo, Schenk, & Guerrero, 2020; Pendyala, Patras, Gopisetty, & Sasges, 2021). However, nonlinear inactivation behavior has been also documented in well-stirred systems with clear liquids. For example, Martin Jr. et al. (2016) reported a pronounced shoulder effect in the inactivation kinetics of Bacillus subtilis suspended in distilled, deionized water. On the other hand, Kaya and Unlurtuk (2016) reported a tail effect during inactivation of yeasts and lactic acid bacteria in pasteurized clear grape juice. Both studies were performed in well-stirred collimated-beam UV batch reactors.

A common practice in chemical reaction engineering is to use rate equations identified in batch experiments to predict the reactant conversion in continuous stirred-tank or plug-flow tubular reactors (as the reaction rate does not depend on the operation mode). However, the use of inactivation rate equations identified in batch experiments is not recommended for predicting the efficacy of continuous UV-C irradiation processes because of the existing differences in the way in which the incidence of UV-C light occurs (Gayán, Condón, & Álvarez, 2014b). Continuous UV-C inactivation is preferred from an industrial standpoint because it could offer some advantages over batch operation, for example, a higher productivity. This operation mode is conducted by pumping liquid foods into coiled tube or jacketed reactors, with or without recirculation. Despite of the evident differences between them, modeling of inactivation data in continuous systems is performed through the same models used for batch processes (Alberini, Simmons, Parker, & Koutchma, 2015; Antonio-Gutiérrez et al., 2019; Antonio-Gutiérrez, López-Malo, Ramírez-Corona, & Palou, 2017; Fenoglio et al., 2019; Fenoglio et al., 2020; Gouma, Gayán, Raso, Condón, & Álvarez, 2015; Müller, Stahl, Greiner, & Posten, 2014; Ochoa-Velasco et al., 2018; Unlurtuk & Atilgan, 2014). Consequently, design variables such as inlet and outlet streams, feeding cell concentration and reactor volume are left out from the model. These equations have time as independent variable and thus their parameters are evaluated under unsteady-state operation. The fitted equation reproduces the shape of the inactivation, but it lacks applicability for process scale up as the true inactivation rate equation is masked by the effects of the design variables listed above.

The efficacy of UV-C irradiation to achieve microbial inactivation depends on both food liquid characteristics and processing conditions. For example, the soluble and insoluble solids content is known to have a major effect on microorganism survival (Menezes et al., 2019). In continuous operation, the inactivation rate depends on the flow regimen and residence time (Alberini et al., 2015; Antonio-Gutiérrez et al., 2019).

This study aimed to develop a reactor engineering approach to describe microbial inactivation during continuous UV-C light processing. To fully achieve this purpose the following topics are covered: (i) the modeling of the continuous UV-C irradiation reactor and the development of a simple straight-line fit approach to identify the inactivation order, (ii) the identification of a non-elementary rate equation for simulating the dynamic and steady-state inactivation periods, (iii) the experimental validation of the proposed theory during processing of model solutions inoculated with Lactobacillus rhamnosus and (iv) the evaluation of effect of soluble solids content, residence time and pH on the performance of UV-C irradiation.