Assessing and Applying Thermal Characteristics of Ready-to-Eat Products during Heat Processing
By Gene W. Bartholomew, Ph.D.
Thermal processing of foods has been practiced for thousands of years to impart certain organoleptic properties to foods and also to help preserve foods. The mechanism behind food preservation was not well understood until fairly recently. Perhaps the first glimpse into the mechanism of food preservation was given by Spallanzani in 1768, who disproved spontaneous generation through the application of heat. Soon thereafter, Nicolas Appert applied heat to foodstuffs for Napoleon’s troops and started the beginnings of a canning business. It was not until later that Pasteur and his contemporaries and students understood that this heat step was destroying microorganisms in the food, thereby extending its shelf life. We use these principles to this day to produce canned items, which are termed shelf stable, and fully cooked or ready-to-eat products, which go through typically milder heat processes and must normally be refrigerated to prevent spoilage. In either case, application of heat is essential for eliminating bacteria that may cause illness if consumed.
A Little Background
Thermal bacteriology, which applies knowledge about heat’s effects on bacterial viability, is a relatively mature field and has many applications to food processing. Two articles[1,2] have covered the use of thermal bacteriology in food processing, and they provide good fundamental knowledge in what has been termed process lethality. The reader is advised to read these articles for background material. As illustrated in these articles, I will use a widely available tool, the American Meat Institute Foundation’s (AMIF) Process Lethality Spreadsheet to demonstrate the advantages of calculating cumulative lethalities of several thermal processes. But before we can fully exploit this tool, we will have to understand better how characteristics of ovens or other cooking devices affect heating of foods and possess a means of measuring product temperatures. I will also explore ways in which thermal profiling of products during cooking can be used to establish Hazard Analysis and Critical Control Points (HACCP) critical limits (CLs), HACCP monitoring efforts and some real-world applications that may help processors with difficult situations.
Thermal Food Processing
Heat transfer from a heat source to the food product occurs through conduction, convection or radiation (or a combination of these means). Conduction is the transfer of heat by the exchange of kinetic energy between two particles in contact with each other. Convection occurs when heat is carried to an object by the flow of a liquid or gas, typically water or air. Radiation transfers heat by the transmission of electromagnetic radiation through a medium and absorption by the food. In the meat processing industry, with which I am the most familiar, all three transfer mechanisms can be and are used. Conduction is the primary means of transferring heat at the surface of a food to the interior so the item is fully cooked throughout. Delivery of heat to the surface of a food is typically achieved through convection or radiation. Radiant heating is usually through either infrared or microwave radiation. Both of these systems are typically deployed in a continuous-cook mode, where the product passes through a radiant zone on a conveyor on a continuous basis. Convection heating dominates the meat processing industry, and heat is delivered via air or water. Traditional smokehouses rely on heating air in a separate zone and then delivering the heated air to a chamber that holds product, which can be either stationary (batch mode) or moving through the chamber (continuous cook). Water cooking is normally accomplished by immersing foods in a tank holding hot water for a time, although there are many examples of product heated by a shower of hot water, in either batch or continuous-cook mode. I will spend more time describing a smokehouse using air as the heat exchange medium, as it is the most complex mechanically and arguably requires the most management and monitoring.
Most smokehouses consist of two components: an air heating and distribution system, and a treatment area where product is contained. The most common arrangement has the heating equipment sitting on the top of the smokehouse and the air entering the main body of the smokehouse along the sides, with the air sweeping down the sides and across the floor. The air returns to the heating unit through the return ducts running down the center of the house. If the air moved only in this pattern, product adjacent to the walls and in the center of the house would receive much more heat than product one-third of the way in from each side of the smokehouse. Most manufacturers of smokehouses overcome this stratification of air by incorporating oscillating dampers on each side of the air train entering the house, resulting in air flow being largely blocked on one side of the house, while the air on the other side is largely unimpeded. These dampers move back and forth 90 degrees out of phase with each other, resulting in air patterns sweeping across all of the product in the house at regular intervals. See Figure 1 for a pictorial representation of the air movement in relation to the damper setting.
In most hot air systems and some water-cook systems, heat is lost to the first product upon which the stream impinges, and thus the last product to come into contact with the medium before it is reheated will be the coldest product in the chamber. This uneven heating of product is important for the producer to understand and quantify. For one thing, one cannot apply typical statistical analyses to product temperatures in the house, because they do not exhibit a normal distribution. And the cook schedule must be designed to take into account that there are colder spots in the house. Likewise, any monitoring for the CCP must include product in these cold spots, as they will have the lowest pathogen reductions.
As you can see from this discussion, with the possible exception of some water-cook systems, uneven heat distribution in a cook device is a fact of life, and the processor needs to know the basic pattern of heating to ensure minimum bacterial reductions. Before we discuss how a processor can model his cook systems, we need to introduce temperature monitoring for an entire cook cycle. When done properly, this information can define temperature distribution throughout the cook system (cold and hot spots) and allow one to calculate the cumulative process lethality by plugging the temperatures recorded during the cook into AMIF’s Process Lethality Spreadsheet.
Collection of Temperature Data
The best way to understand product temperatures during a cook is to place recording thermocouples or resistance temperature detectors in product throughout the chamber before the start of the cook. There are basically two ways of doing this: wired thermocouples leading to a recording module and wireless data loggers. I am familiar with models from many manufacturers of each type, and with few exceptions, prefer wireless data loggers by a wide margin. They are much simpler to place in the cook system, can be retrieved without fanfare after the cook has concluded, are easily identified when placed in the house and are usually the only way to monitor continuous-cook systems where product moves through a long heat tunnel. Most of the wireless data loggers have an upper temperature limit of 250 °F (121 °C), whereas the wired data recorders are much less heat resistant unless put in a thermal barrier. Whichever probe type you select, in most cases the probe tip needs to placed in the center of the mass of the item to be cooked prior to the start of the cycle. This location will be the slowest to reach the desired temperature.
To effectively determine heat distribution within a cooking chamber, you first have to understand the expected differences throughout the chamber, and the size and volume of product loaded in the chamber. Many manufacturers of cook systems have a good understanding of heat differences within their devices and can provide guidance to the user. That is a good place to start, but recognize that product type and loading also affect heat distribution, not to mention cook system maintenance practices and effectiveness. For small cooking devices or devices that are inherently uniform in heat distribution, I recommend a minimum of six product probes placed throughout the oven during cooking. They should be distributed equally throughout the device, unless you have prior knowledge of cold spots, which are of utmost importance to the operator. With larger systems, again try to get an even distribution of sensors throughout the cook device, thinking about natural cold spots and product loading. In the case of smokehouses in meat processing, and keeping in mind how hot air enters and exits the chamber, I try to use a minimum of 12 probes and focus on known colder areas such as the top of the rack of product and by the doors, as they tend to be the farthest removed from the heat source (Figure 2). If you measure product temperatures to establish compliance with a CL in your HACCP plan, place at least two probes in areas where your HACCP monitors take their temperatures to make sure these areas are actually the worst-case heating areas. Demonstration of consistent cold spots through this type of profiling can lead to a reduction in CCP monitoring sites while maintaining confidence that the process operated as designed.
Uses of Temperature Profiling
Speaking of CLs for pathogen reduction in foods, many processors in the United States use guidelines put forth by the U.S. Department of Agriculture (USDA) that establish safe harbors for setting time and temperature requirements.4 These tend to be very conservative guidelines that may lead to overcooking of some products. They also do not incorporate the differences that product size makes on the length of many cook schedules, plus they have hold times at the colder temperatures, which can be cumbersome to follow when monitoring the CCP. But if you have an understanding of the temperature profile of your product through continuous monitoring by data loggers, you can calculate the cumulative lethality during the entire cook and derive a CL based on this information. Figure 3 illustrates this principle with a steam house filled with roast beef of various weight ranges that were monitored throughout the cook. Because this schedule is rather lengthy compared with cook schedules for smaller items, by the time product reached 145 °F, it had already exceeded sufficient lethality to be considered fully cooked. Even if we could somehow immediately chill product at this stage below a temperature that is lethal to foodborne pathogens of concern with this product, we would have still exceeded the required pathogen reduction. Therefore, there is no reason to have any hold period as part of the CL, which simplifies CCP monitoring.
Another use of internal temperature profiling during a heat step is setting an HACCP CL and monitoring as a function of the product and oven operational characteristics, particularly when it is difficult or impossible to take product temperatures reliably at the HACCP monitoring point. If the product does not lend itself to temperature monitoring (in a can, for instance), is physically difficult to access in the heating device or rapidly loses heat as it exits the heat chamber, knowledge of how oven characteristics influence product temperatures can take the place of observing actual product temperatures at the conclusion of the heat step. But to link the two, product heating profiles tied to oven characteristics are needed. Probing as described earlier can be an effective way to demonstrate and document the relationship. To illustrate this point, consider heating neck bones so they can be considered fully cooked. Pork neck bones have very little meat on them, and operators in traditional meat processing cannot take CCP meat temperatures readily at the end of the cook step. To link the cook cycle to internal temperatures in the neck bones, we drilled holes in the bones so the cavity ended in the thickest part of the bone, inserted wireless data loggers in the holes and placed these samples in the smokehouse with product loaded in the usual fashion. We then ran an extended cook cycle (a dry bulb temperature held constant at 170 °F). This process was repeated four more times to control for normal size variation in the neck bones and to check a variety of placements throughout the house. One such trial is demonstrated in Figure 4. Based on these five probed trials where the oven temperature was held constant at 170 °F for the entire cook, we concluded that if the house were maintained at or above 170 °F for at least 4 hours, all product inside would reach the appropriate pathogen reduction. This became our CCP’s CL, and monitoring was recording house temperature versus time.
Another similar situation may occur in many facilities where the processor wants to deliver a heat treatment sufficient to eliminate pathogens on the surface of product (before or after packaging) without substantially warming the product core. This is extremely difficult to monitor with hand-held thermometers but can be modeled with data loggers. Many wired data loggers can use thermocouples with extremely small tips that can be placed just under the surface of food items and thus record surface temperatures during heating. In this way, one can define operating parameters of heat chamber temperature and hold time to determine lethality. An example is the case of some cooked deli slicing logs that were potentially contaminated on the surface by falling condensation. We took several logs out of a large group inside a smokehouse and placed small thermocouples just under the skin of each log, as well as a probe in the airstream. These wires ran out of the oven to a recording data logger, and the smokehouse was set to operate at a set temperature for 70 minutes. We collected temperature data from each sample for the duration of the reheating and calculated cumulative lethality values to determine the necessary reheat time. Figure 5 shows that after 30 minutes at 170 °F at high humidity, surfaces achieved a minimum of a 7-log reduction of Salmonella, while the core temperature of the logs rose only a few degrees (data not shown).
Other Applications of Temperature Data
So far, the applications I have discussed have focused on the use of heating profiles to allow us to calculate expected bacterial reductions in food items. But there are many other uses of temperature data that are unrelated to pathogen reduction. In the case of fermented foods, we use mild heat to promote growth of fermentative organisms, usually added as a starter culture, to change the organoleptic properties and spoilage characteristics of many different feedstocks. Just as we try to achieve uniform heating to give consistent lethality in a fully cooked food, we also want our fermentation chamber to provide optimum fermentation (growth) temperatures consistently to all product housed therein. Failure to do so may result in some areas of the chamber being colder than others, leading to incomplete fermentation and subsequent quality or safety issues. Likewise, high temperatures may injure the fermentation microorganisms, so again fermentation is not completed as desired. The best way to assess uniformity of temperature distribution throughout the chamber is the same as our cook process monitoring. Figure 6 shows internal temperature profiles of 11 Genoa salami sticks spread throughout a large chamber during both the fermentation step and a subsequent heat treatment for pathogen reduction. In this case, we used a starter culture as the fermenter and it functions well between 95 and 110 °F. Temperature distribution is extremely narrow in this chamber and leads to very consistent product.
This case study also points out another parameter of thermal processing that I have not mentioned before—uniformity of product temperatures at the start of heating. Upon close examination of the starting product temperatures in Figure 6, you will see a large difference among the 11 samples probed. The reason for this difference is that this chamber, because of its size, takes several hours to load before it is started. So product that is stuffed at cold temperatures and loaded first into the nonrefrigerated chamber warms up considerably before the last product is placed and the cycle started. If not for the long period of the fermentation portion of this process, initial temperature differences of this magnitude would likely result in significant temperature differences through the heating portion, thus compromising uniform lethality. Another potential cause of uneven product temperatures at the start of a cook is the use of previously frozen materials. Unless they are carefully thawed, some product cores could still be frozen when placed in the oven and result in dramatic differences in final cook temperatures.
The final application for temperature profiling I want to discuss is monitoring of chilling after a thermal lethality process. Typically, unless a food processor is preparing canned (aseptic) foods, the heat process is not sufficient to destroy bacterial spores that may be present. One pathogenic spore former in particular causes issues in the food processing industry, and perhaps more so in the foodservice business. Clostridium perfringens, if present in spore form in a food item that receives heat processing, will survive the cook step and may proliferate rapidly after the cook if the product is held under warm conditions (100–120 °F). For this reason, USDA has issued performance standards that require prevention of growth of spore-forming bacteria in certain foods [9 C.F.R. Sections 318.17(a)(2), 318.23(d)(1) and 381.150(a)(2)]. Quickly chilling product to below 55 °F after cooking is effective at controlling grow-out of C. perfringens,5 and many manufacturers routinely monitor chilling of products after cooking in a manner similar to monitoring the cook step itself. Use of internal temperature data recorders can precisely show typical cooling curves of these products, which can be used predictively to fit against cooler and product characteristics as with neck bones or to identify slowest-cooling products in a lot and guide the operator to the worst-case items to further monitor. In Figure 7, we left the temperature data loggers in the samples of roast beef after the cook finished to monitor the chilling profiles of representative pieces. The main grouping of profiles shows that with product of similar weights, there was little difference in the rate of cooling based on placement within a given rack or among several racks spread throughout the blast chiller. The size of roast beef did affect the rate of cooling measurably, as the pieces from Racks 3A and 3B, which were roughly 4 pounds heavier than all of the other samples, required over 1 hour more than the smaller samples to reach 55 °F. If this profiling was repeated several times and similar results were observed, the data could be used to support monitoring only the larger sizes routinely or controlling chiller and product parameters so constant monitoring of product temperatures is not necessary.
Temperature profiling of processed foods during and after cooking provides the processor initially with critical information to confirm process suitability and periodically thereafter to support HACCP monitoring plans, to demonstrate equipment design issues and maintenance, and raw material uniformity and to allow one to make process improvements. While this report has focused on heating as a means of achieving appropriate pathogen reduction, the knowledge that temperature profiling gives us is useful in developing quality characteristics as well. It does require some capital investment and training, but I have found the return on investment has been well worth the effort.
Gene W. Bartholomew, Ph.D., is corporate director of food safety/HACCP affairs at the John Morrell Food Group. He provides technical expertise and consultation to the company’s 13 meat processing plants and one spice plant in the areas of HACCP plan development and validation, pathogen reduction efforts in sanitation, equipment design, product cooking and chilling and product shelf life extension. He has helped many of these plants achieve and maintain GFSI certification. He is also technical director of the corporate microbiology laboratory and consults with the plant labs. Gene helped develop the American Meat Institute Foundation’s Listeria Control Working Group and has presented to several industry groups on this subject, in addition to other presentations at the Food Safety Summit and The International Association of Food Protection. Prior to joining John Morrell Food Group, he developed new food packaging materials and designs and an aseptic packaging system for International Paper. He holds a B.Sc. in biology from Bucknell University, and M.Sc. and Ph.D. degrees in microbiology from Cornell University.
1. Sindelar, J.J. and A. King. 2013. Thermal processing with food safety in mind. Food Safety Magazine 19(5):58–64.
2. Scott, J. and L. Weddig. 1998. Principles of integrated time-temperature processing. Presented at the 1998 Meat Industry Conference, Philadelphia.
5. Taormina, P.J., G.W. Bartholomew and W.J. Dorsa. 2003. Incidence of Clostridium perfringens in commercially produced cured raw meat product mixtures and behavior in cooked products during chilling and refrigerated storage. J Food Prot 66:72–81.