Progress 08/01/24 to 07/31/25

Outputs
Target Audience:For 2025: Higher education and Industry : year-long heat transfer and modeling project led by CO-PI Tom Diller. This featured laboratory instruction on vacuum steam pasteurization, thermal measurements and design of instrumentation. Knowledge was shared with industry vacuum steam stakeholders and industry representatives that manufacture thermal heat flux instrumentation. Higher education: Year-long project developing methods for inactivation of pathogens and surrogate bacteria on dried nuts and seeds. Laboratory instruction on development of novel methods to simulate conditions associated with contamination of pathogens in the environment. Industry: Consultations with Cosmed Group and PhytoVac. both companies process products using vacuum steam pasteurization to kill harmful pathogens on food or plant products. Changes/Problems:As reported in the year one report there was a delay in the start of the project so that will continue to impact the rate of expenditures. Although the steam vacuum pasteurization process appears to be simple in concept, it actually is a very complicated physical system. This has required a number of changes to our research approach. 1. The initial plan was to use commercially available heat flux sensors mounted to thin aluminum blocks to aid in modeling the heat transfer and temperature of the product at different positions in the package. This approach gave good results for individual product pieces, but did not work in packages of product, even though we worked with the instrument manufacturer to improve the water resistance of the heat flux sensors. Consequently, a better method was incorporated using instrumented aluminum balls and actual product. Using the measured rate of temperature change gave consistent and reliable values of the corresponding heat flux. 2. Following industry practice we assumed that the steam would penetrate quickly throughout the product package. Measurements showed that this was clearly not the case. Further investigation demonstrated the importance of the non-condensable gases present even after pulling a vacuum. This was confirmed by computational fluid dynamic (CFD) modeling. 3. Because of the poor steam penetration, the temperature distribution within a package was inconsistent and unpredictable. A tube with a small fan was implemented to provide a uniform flow throughout the package. This new approach to using steam vacuum greatly enhanced repeatability as well a much faster temperature rise of the product. One of our future directions is to consider how to transition this concept to industrial scale. Our protocols for working with the biohazards in this study were approved by the Virginia Tech Institutional Biosafety Committee in 2023 and reviewed in 2024. No additional changes were required. What opportunities for training and professional development has the project provided?Training activities: One-on one training for one M.S. and one Ph.D. students in thermal transfer measurements, preparation of thermal transfer measurement instrumentation and modeling. In addition, two additional Ph.D. students have assisted with the project, expanding their training in vacuum steam processing, thermal measurement and performing process validations, critical skills needed by food safety professionals. Professional development: Scientific communication development for one Ph.D. student on the project with activities including development of a short (3 minute) research presentation (2 hours) incorporating facilitated discussion and peer review before the student presented his talk. Additional scientific communication workshops covered writing abstracts (1 hr), and designing scientific posters (2 hr). How have the results been disseminated to communities of interest?Meeting with Dr. Mark White, CEO of PhytoVac in December 2024 to discuss the vacuum steam process. Emphasis was on the shared difficulty of moving the steam through the product in the chamber. A presentation of experimental and analytical results using the new fan system was given by one of the graduate students working on the project in March, 2025. Details of the air and steam interaction in the PhytoVac system and possible alternatives were discussed with Dr. White. Shared heat transfer model for macadamia nuts processed using vacuum steam with David Howe of Cosmed Group, one of the largest vacuum steam processors of LWAF foods. What do you plan to do during the next reporting period to accomplish the goals?One of the primary objectives of this proposal is to develop validated vacuum steam processes for the inactivation of pathogens and to identify surrogate non-pathogenic bacteria that behave similarly on LWAF. Therefore, in year 3 we will conduct vacuum steam processing on LWAF that are contaminated with three pathogenic bacterial species, Salmonella enterica, Escherichia coli and Listeria monocytogenes, that are most frequently associated with illnesses or prompt recalls of LWAF. To date our experiments have used Pediococcus acidilactici, which we have previously shown behaves similarly when exposed to heat on macadamia nuts. This was necessary because equipment and sensors used for heat transfer modeling could not be readily decontaminated. Since heat is necessary to kill the pathogen, we anticipate that the heat transfer models generated in year 2 will be able to predict the time necessary for the product to reach the target temperature. As the bacteria are exposed to heat their membranes are damaged allowing intracellular contents to leak out and proteins become damaged. Depending on the ability of bacteria to withstand or repair this damage the time of exposure to temperature must be increased. If the model does not reflect the behavior of the pathogens on a product that has similar heat transfer this may indicate that other factors are influencing bacterial survival. In year 2, we have shown that the dynamic model generated can predict the inactivation of another commonly used surrogate, Enterococcus faecium, when exposed to 70ºC vacuum steam pasteurization. We anticipate the models will be conservative and that the reductions of pathogens will be greater. However, some pathogens including Salmonella can be more thermal resistant when previously exposed to dry conditions resulting in low water activity stress. These bacteria will be inoculated by transfer from dry sand particles, more similar to what is likely happening in the environment. We will compare the suitability of currently used non-pathogenic surrogate bacteria to predict the behavior of the pathogens, necessary for commercial establishments that can not verify the effectiveness of their conditions by introducing a pathogen into the facility. Experiments are continuing in year 3 to develop dry transfer inoculation strategies that create stable inoculum for a period of 7-14 days for both pathogens and non-pathogenic surrogates. For commercial processors, it is necessary that the process achieves the expected inactivation of a surrogate bacteria. These products are inoculated off-site and shipped for processing, which may take between 7-14 days depending on workflow. The heat transfer models generated in year two will be incorporated into dynamic inactivation models to predict the inactivation at different locations within the package for different products. We have tested the model with two different surrogate bacteria and found that ~90% of the inactivation occurs within the first 200 sec after reaching the target temperature. However, the time needed to inactivate the remaining bacteria isn't linear. We plan to conduct a scaled-up package size test using a surrogate bacteria with one of our collaborating stakeholders in year 4. In year 3 we plan to publish 2 scientific papers. In the first we will describe how the density and composition of LWAF impacts the heat transfer to individual food pieces and considers how steam moves through a package. This paper will describe a different approach than relying solely on temperatures measured using thermocouples, and incorporates the need to consider heat flux from the steam to the surface of the product. This paper will also describe the limitations of currently available sensors that require uniform contact across a surface. For small food particles in a package this creates problems because the stacking of particles creates gaps where the sensor is not in contact with the food. We will share this information with manufacturers of heat flux sensors. In the second paper we will describe the dry transfer inoculation methods developed for macadamia nuts, pumpkin seeds, mustard seeds and mint. This paper will be noteworthy as it describes modifications needed for products considering the size, shape and fat composition. The final manuscript anticipated for this project describing the dynamic inactivation models of the pathogens and surrogate bacteria will be prepared in year 4. We anticipate the continued training of one PhD student performing the inactivation and modeling experiments, in addition one other PhD student and an undergraduate student will be assisting with the vacuum steam inactivation tests.

Impacts
What was accomplished under these goals? The overall goal of this project is to improve the safety of dry, ready-to-eat products such as nuts, seeds and spices (LWAF). These products may be contaminated with harmful pathogenic bacteria that may cause illness. One strategy for eliminating these bacteria is pasteurization, a process that applies heat energy, to kill bacteria. Vacuum-assisted steam processing delivers more heat in a short period of time compared to other thermal processes such as roasting. This may result in safer products with better sensory quality. In the second year of this project we advanced progress for objectives 1-3. Bacterial contamination on LWAF likely occurs from contact between a contaminated dry particle such as soil or sand that transfers bacteria to nuts and seeds in the environment. In year 2 we have also been developing inoculation strategies for nuts, mustard seeds and mint that mimics how these products might become contaminated in the environment. In our preliminary studies, we have achieved a stable inoculum of target human pathogens of 7 log CFU/g that can be used to demonstrate the desired 5 log reduction of pathogens after vacuum steam processing needed for objective 1. It has been necessary to customize the protocol for products based on the size (surface area) and fat composition. This project will use inoculation protocols that more closely resemble the stresses experienced by bacteria in the food chain. In order to predict the time needed to kill a pathogen on LWAF, it's necessary to understand how the heat is transferred from steam to the product, a process called convective heating. It's commonly believed that steam can travel through small crevices and gaps in the package of product, leading to uniform heating. However, our testing showed that this assumption is usually not valid. Even in a small package (~450g) of whole macadamia nuts we found that the steam only slowly penetrated through the package, which severely limits the temperature rise of the product. We found that the major reason for this poor performance was the remaining air (non-condensable gas, NCG) even after a pulling a vacuum. The air is trapped in the gaps and acts as a boundary that the steam has to penetrate through before it can condense, drastically limiting the efficiency. In order to improve the performance, a small fan was added inside the chamber to force the steam to pass through the package and flush out the air. This dramatically improved the time it took for all the product to reach the desired temperature. It's also necessary to predict how heat is transferred from one piece to another, a process called conduction. We created a conduction heat transfer model to predict the surface temperature of three products, whole macadamia nuts, macadamia nut pieces, and pumpkin seeds, products that differ in density and the nutritional composition (fat, protein, carbohydrate, ash, and moisture). The denser whole nuts required more energy to increase the temperature followed by the macadamia nut pieces, and fastest in pumpkin seeds. In general, as the size of the product increases so does the thermal mass. A higher thermal mass means more energy is required to increase the temperature. The heat transfer coefficient to the three products were found to be 170 W/m2*K for whole nuts, 200 W/m2*K for pieces, and 520 W/m2*K for pumpkin seeds. The higher coefficient for pumpkin seeds also contributes to the quicker temperature rise. By understanding the heat transfer from steam to the nuts (convection) and between nuts (conduction) we can predict the temperature within a package. This project is implementing new experimental and modeling technologies, not currently used in the food industry, to more accurately describe how steam moves through a product to raise its temperature and inactivate bacteria. To understand this complex heat transfer process a detailed computational fluid dynamic (CFD) model was created to predict the temperature at any location and point in time within a package. Based on the product size, packing density, and inlet velocity we successfully modeled the steam condensation process for five different products (whole macadamia nuts and pieces, pumpkin seeds, mustard seeds, and Brazil nuts). To determine the product surface temperature as a function of time we added a conduction heat transfer model within the nuts. The resulting calculations were consistent with the surface temperature history of the instrumented product as measured with tiny thermocouples embedded in the individual nuts. The model results were validated at multiple locations in the package and at three different process temperatures. This is a major advance in understanding the steam vacuum process. A temperature gradient was observed in the package with the fastest increases in temperature occurring at the top package. However, the time delay for the nuts to reach the same temperature was different at the top, middle and bottom of the package for each of the tested products. Whole nuts had very little delay (~2-5 s), nut pieces had a little more delay (~5-10s), and pumpkin seeds had the most (~30-40s). This is due to the differences in how steam penetrates the package impacted by the porosity, how the product stacks on top of each other, and steam flow resistance associated with the different products. The whole nuts have the highest porosity (.508) and lowest flow resistance, whereas the pumpkin seeds have the lowest porosity (.425) and the highest flow resistance. This causes the velocity through the whole nut package to be significantly quicker than the pumpkin seeds. The higher velocity leads to more uniform heating in the package with little time delay between the different positions, due to the steam penetrating through the whole package very quickly. By understanding how heat moves through the package we can build models that better predict the time it is needed to kill bacteria on LWAF. In year 2, we have also tested the ability of steam to kill two common surrogate bacteria, Enterococcus faecium and Pediococcus acidilactii on macadamia nuts in different package configurations and on pumpkin seeds.The reductions in bacteria will be used in creating a dynamic model that can be applied to different types of LWAF products. We built this model with Enterococcus faecium, a non-pathogenic bacteria (surrogate) that behaves similarly to the pathogen Salmonella, inoculated onto whole macadamia nuts and validated it using macadamia nut pieces, and pumpkin seeds. For each product, vacuum steam treatment (70C) was applied for nine different time intervals and the number of surviving bacteria determined. Inactivation occurs quickly within the first 200 s but then slows down requiring over 40 minutes to achieve the target inactivation. Overall, the inactivationates of E. faecium and Pediococcus were comparable. Based on these initial tests the inactivation rates were slightly faster for pumpkin seeds but comparable for whole nuts and pieces. These products differ in their densityand porosity which effects the heat transfer.Currently models don't consider the heat transfer and cannot be applied to different products. We have developed a dynamic model created using the densest product, whole macadamia nuts, that can predict the inactivation seen on different products demonstrated with pieces of macadamia nuts and pumpkin seeds at different time intervals. To quantify the accuracy of this model, the root mean square error (RMSE) was calculated. For whole macadamia nuts the RMSE was 0.12 log10(CFU/g), 0.39 log10(CFU/g) for macadamia nut pieces, and for pumpkin seeds it was 0.41 log10(CFU/g). This model will allow us to predict the inactivation of bacteria with a margin of safety for a variety of products because it considers how the heat is applied to the product.

Publications

Progress 08/01/23 to 07/31/24

Outputs
Target Audience:Higeher education classroom and a year-long student design project led by CO-Pi Tom Diller to refine the vacuum steam apparatus and to develop priliminary heat flux and heat transfer models for macadamia nuts Industry. consultations with Cosmed Group, a vaccum steam processor on challenges associated with steam movement industrially Changes/Problems:In year, we encountered a delay in the manufacturer of the necessary sensors to measure heat flux and improved thermocouples for recording temperatures in product. In our testing of the existing apparatus we identified that the previous methods to measure temperature needed to be improved and automation was added to increase repeatability. We had originally planned to complete the pathogen testing and thermal modeling by the end of year two, allowing a focus on dynamic modelling in year 3. We anticipate that our timeline for expenditures may be shifted by one year due to the inital delays. We will have 2 graduate students (one PhD, Food Science and Technology and one MS, Mechanical Engineering) working in tandem in years 2 and 3 on the inactivation work, thermal modeling and creation of dynamic models which consider the phyical characterisitics of the product and heat transfer in the system. What opportunities for training and professional development has the project provided?This project provided a year-long training opportunites for students in mechanical engineering to demonstrate their skills in project design and instrumentation for thermal processing , while developing new skills in design of sensors and dynamic models. Five students, mentored by CO_PI Tom Diller and PI Monica Ponder, enrolled in a senior design project were engaged in the project and recieved one on one and group instruction. Dr. Ponder introduced concepts of food safety anddiscussed different types of pasteurization used in the food industry. Together with Dr. Diller, the students measured the impact of product composition, size and density on the heat transfer. These are concepts introduced to ME students chiefly focused on metals and materials used in industry. This challenged students to apply this knowledge to another foods. Students developed presentations and papers that were course products at 5 different times through the year. Ultimately the final product was new instrumentation and a model. Graduate students in the summer of 2024, began to work with the project developing new strategies for inoculation and enumerations of bacteria from the food products. How have the results been disseminated to communities of interest?Food safety engineering and processing are not typical career paths for students majoring in mechanical engineering. Students in this major possess transferable knowledge and skills that can be applied to processing of safe and nutritious foods. In this project, a team of mechanical engineering students worked with developing new instrumentation and automation of thermal processing equipment, with the ultimate goal of improving food safety. The students gave 5 presentations to other mechanical engineering undergraduate students and presented in a University wide Engineering Research Day, where many faculty and students from different areas were exposed to the applications of engineering to addressing food safety and food quality. A news article, featured on the web pages of the College of Engineering and Colleges of Agriculture and Life Sciences, https://news.vt.edu/articles/2024/02/cals-research-coe-spices.html was also released to highlight the opportunities for careers and interdisciplinary strategies needed to tackle challenges in assuring safe an nutritious foods. What do you plan to do during the next reporting period to accomplish the goals?In year 2, the focus will split between objectives 2 and three. We will continue to refine the model by testing other LWAF that have similar properties compared to macadamia nuts (pumpkin seeds) and different properties (mustard seeds and mint leaves). This will allow us to determine suitability of one model to products that differ in their chemical and physical properties. We anticipate that intrinsic properties of the LWAF including density and porosity will alter the heat transfer. Commodities with similar compositions may have different heat flux if the products differ in density and porosity. The heat flux can be directly measured with a sensor in both laboratory and commercial scale systems. The predicted heat transfer from the mathematical model will be comparable to the measured heat flux using the miniatured sensors Due to safety concerns associated with the introduction of pathogens we will conduct sensory testing (objective 4) and heat transfer modelling for the other products in different configurations (objective 2) first, then move into refinement of dynamic models (objectives 1 and objective 3) using pathogens. We anticipate that pathogen testing may bein at the end of year 2 or beginning of year 3.

Impacts
What was accomplished under these goals? The overall goal of this project is to improve the safety of dry, ready to eat products such as nuts, dried fruits and spices (LWAF). These products may be contaminated with harmful bacteria that may cause illness. One strategy for eliminating these bacteria is pasteurization, a process that applies heat to kill bacteria. Vacuum-assisted steam processing can be used to deliver more heat in a short period of time compared to other thermal processes such as roasting. This may result in safer products with better sensory quality. In the first year of this project we focused on objective 2 of the project: developing models to describe heat transfer in one model LWAF, macadamia nuts. This project is implementing new technologies, not currently used in the food industry, to more accurately describe how heat moves through a product, called the heat flux. This year we designed and manufactured heat flux sensors and thermocouples needed to model both the static and dynamic heat transfer in our model system of macadamia nuts. Data from preliminary experiments determined that the temperature on the surface of the product varied more than 2% compared to the ambient temperature inside the chamber when processing via a common industry method. Our newly designed thermocouples and heat flux sensors and added instrumentation have allowed us to to automate the release of steam based on the temperature of the product itself at any position within a package, increasing repeatability of tests and the safety of the product. We have created a new mathematical model of the heat transfer in macadamia nuts that considers the temperature and the heat fluxto predict the temperature at any point within a package of macadamia nuts. We have also tested the ability of steam to kill a common surrogate bacteria, Enterococcus faecium, on macadamia nuts in different package configurations. The reductions in bacteria will be used in creating a dynamic model that can be applied to different types of LWAF products. In preparation for these next tests we are developing inoculation strategies for mint that mimics how it might become contaminated in the environment. The appearance and proximate conditions) properties of macadamia nuts (protein content, fat content) were not altered by the steam processing; in future years we will determine if humans can tell the difference between treated and untreated products using a human similarity test.

Publications