- Email: [email protected]
Heat penetration determinations and thermal process calculations
Heat penetration determinations and thermal process calculations
3
3.1 Introduction In 1917, the use of thermocouples in the United States was introduced by the National Canners Association to measure temperatures during heating and cooling of foods in sealed containers. This led to graphic mathematical and computerised procedures to estimate the minimum retorting conditions required to produce ‘commercially sterile’ foods without excessive damage to their eating quality or nutritional value. Since that time, progressive scientific, technological, and engineering improvements have been achieved in canning equipment and in packaging. The conventional steam-still retort has lost popularity in favour of continuous hydrostatic and reel-and-spiral cookers, overpressure and agitating retorts, and, albeit to a lesser extent, aseptic processing equipment. Packaging has developed to the position in which metal cans have evolved from cylinders into complex geometries, with considerable consumer preference for glass jars, plastic pouches, semi-rigid trays, and even cartons. These present unique challenges for heat penetration (HP) measurement. Food that is thermally processed in hermetically sealed containers is heated throughout by employing an external source of heat. If the food is a liquid (e.g. broth) or contains a liquid of low viscosity (e.g. peas in brine), the heat is transferred via convection. If it is a solid (e.g. meat, fish) or highly viscous (e.g. cream-style corn), it is heated primarily via conduction; as a result, a greater portion of the contents is severely overprocessed in order to sterilise the small volume occupying the geometric centre. Agitation during cooking (axial or end-over-end) may effectively reduce the times and temperatures required for safe processing of some foods as a result of the stirring action that eliminates conduction as the sole mode of heat transfer. In addition, processing at higher temperatures for shorter periods of time effectively reduces damage to product quality while achieving commercial sterility. Hence, increasing interest is devoted to continuous rotary cookers (axial rotation) and to batch agitating retorts (end-over-end rotation). Each of these technological and engineering improvements, in turn, opens up new possibilities in the variety of foods that can be preserved through heat processing. On the other hand, each imposes upon the processor the necessity for more stringent and sophisticated procedures to assure adequate control. Furthermore, the newer heat processing methods are employed for the production of high-quality formulated foods that were not possible to produce using the conventional still retort. Many of these formulated foods are designed to be consumed without further heating, which, in itself, imposes the need for the processor to exercise greater stringency in order to assure product sterility. These processes are discussed in more detail in the section on sterilization systems. A Complete Course in Canning and Related Processes. http://dx.doi.org/10.1016/B978-0-85709-678-4.00003-8 Copyright © 2015 Elsevier Ltd. All rights reserved.
44
A Complete Course in Canning and Related Processes
3.2 pH classification of canned foods Canned foods may be classified into two categories based upon pH and acidity and on the necessary thermal processing required to effect safety as well as microbiological stability: 1. Low acid: above 4.6 2. High acid: 4.6 or lower
The low-acid category is the only one of importance from the standpoint of the botulinum hazard. No Clostridium botulinum spore, regardless of type, has been detected that will germinate, grow, and produce toxin in a food having a pH of 4.7 or lower. Foods in the low-acid range must be fully retorted to assure safety; this is sometimes referred to as a botulinum cook, which requires a minimum equivalent process of 3 min at 250 °F (121.1 °C) – the F0 3 process. The US Food and Drug Administration (FDA) has promulgated regulations designed to control production of low-acid canned foods. The regulations, among other things, (1) provide for the registration of all firms producing low-acid canned foods; (2) provide for the filing by each firm of a detailed description of the heat processes used for each low-acid food in each container size (scheduled process); (3) define good manufacturing practices in the processing of low-acid canned foods; and (4) require establishments that fail to comply with the regulations to operate under emergency permit control of the FDA. These regulatory controls serve to further assure the safety of canned foods. Food processors in countries that export to the United States are required to follow these regulations in order that their products are accepted into the country. However, not all countries apply the same level of regulatory compliance and instead rely on national guidelines that are enforced locally, often by the retailer. Standards in the manufacture of thermally processed foods are very similar to those in the United States, and the safety record in most countries is every bit as high. Any low-acid food that achieves an F0 3 process is referred to as commercially sterile; however, both pasteurised and sterilised foods can be considered commercially sterile because of the caveat regarding the micro-organisms deemed ‘capable of growing in the food’. One of the definitions of commercial sterility is given by the UK Department of Health (DoH, 1994): “Commercial sterility is the condition achieved by the application of heat which renders food free from viable micro-organisms, including those of known public health significance, capable of growing in the food at temperatures at which the food is likely to be held during distribution and storage.” An F0 3 process will be safe with respect to C. botulinum spores, but it is important to remember that the product may contain a small number of surviving spores of more heat-resistant spoilage micro-organisms. This is because there are many more types of spoilage micro-organisms than those of concern to public health, and therefore it is not surprising that a greater range of heat resistance is found. Those foods in the high-acid range (pH 4.6 or lower) generally require a considerably lower level of heat processing to effect preservation because the processing schedule is primarily designed to kill vegetative cells. A few spore-bearing bacteria, particularly Clostridium butyricum, Bacillus coagulans, and related species, are able
Heat penetration determinations and thermal process calculations
45
to grow in foods having a pH value as low as 4.1. However, they have no known public health significance. Food processors have learnt to prevent spoilages of this nature by employing good food hygiene practices that reduce the initial spore load of the food combined with appropriate heat processing schedules. The efficacy of heat in destroying micro-organisms is influenced by various factors. It is for this reason that mild heat alone cannot be used to produce safe stable pasteurised foods. The most common factors used in conjunction with heat are pH, water activity, and low-temperature storage. Even if an appropriate heat treatment is applied at the desired level, if these other factors are not in control, there is the potential for micro-organisms to grow and spoil the food. The following statements are usually true: • Most spore-forming bacteria are inhibited by a combination of pH < 4.6 and aw < 0.90. • Acid tolerant spore-forming bacteria will survive and grow at pH > 3.8. This includes many of the butyric anaerobes, a group of bacteria that under anaerobic conditions can ferment sugars to butyric acid, including Bacillus macerans and Bacillus polymyxa. • Some xerophilic and osmophilic spore-forming yeasts and moulds can grow below aw 0.85. • Antimycotic agents (preservatives) have been successfully used in foods to prevent the outgrowth of yeasts and moulds. • C. botulinum does not usually germinate and grow in foods with a pH < 4.6 or an aw < 0.94, although some experimental conditions have shown growth and toxin production by C. botulinum at pH less than 4.5 (Raatjies & Smelt, 1979). • Chilling is only a short-term barrier to microbial growth.
It should be noted that the thermal destruction of some yeasts and moulds is more complex than simple logarithmic destruction. Ascospores can get activated at temperatures as high as 85 °C. Unusual heat resistance has been reported in pasteurised beverages and sauces that contained ascospores of Talaromyces trachyspermus and Neosartorya fisheri (Campden BRI, 2006). In a product with pH < 3.8, there is very low risk of germination and outgrowth of bacterial spores, but certain yeasts and moulds can thrive. The heat process for foods of pH < 3.8 only needs to destroy those yeasts and moulds. In products with pH > 3.8, other organisms are capable of surviving, so the heat process applied must be sufficient to ensure that their numbers are destroyed. As a generalization, within the acid foods group, products can be further divided into pH categories so that a minimal, but effective, heat treatment can be applied.
3.2.1 High acid: pH < 3.8 This group includes cherries, plums, apricots, berries, citrus, and sometimes pineapples. The spoilage organisms include yeasts, moulds, and lactic acid bacteria. Due to the high acidity of these fruits, they generally are very susceptible to softening during heat processing, and processors usually try to process them as close to the minimum requirement as possible. Process recommendations for high acid products sometimes target a can-centre temperature (CCT) over 70 °C, with a holding time of 2–3 min, but practically a CCT of 80–82 °C is usually targeted for most products in this group. Ascospores of Byssochlamys fulva or Byssochlamys nivea can survive this process,
46
A Complete Course in Canning and Related Processes
and good hygiene in production is essential to ensure that they are not present in the canned product. If they grow, they can produce enzymes that can soften the fruit.
3.2.2 Acid: pH 3.7–4.2 This group of fruits includes peaches, apples, and pears. General process recommendations are for the CCT to reach 85 °C for 5 min or 95 °C for 30 s. This is effectively a P-value of 5 min at 85 °C. The process is intended to target spoilage organisms and includes yeasts, moulds, lactic acid bacteria, and some acid-tolerant bacteria, such as B. coagulans. For all acid and high-acid fruit products, good hygiene during product peeling and sorting is critical because many of the organisms that can spoil canned fruit are sufficiently heat resistant that they cannot be processed away without damaging the texture and colour.
3.2.3 Acid: pH 4.2–4.5 This group includes tomatoes and some pears. The heat resistance of micro-organisms is greater at the slightly higher pH, so processing to a CCT of 100–104 °C is recommended. If the fruit texture is unable to withstand this high process, the product must be acidified and a lower process applied.
3.3 Thermal death time Safety from botulism in canned low-acid foods stems from the pioneering research done by scientists of the National Food Processors Association in the early 1920s. They determined the thermal resistance of spores harvested from the most heat- resistant C. botulinum strain known to them. Their studies demonstrated that, by extrapolation from the exponential survival curve, it was necessary to heat a spore suspension in phosphate buffer for 2.78 min at 250 °F (121.1 °C) to reduce the survival population from just over 1011 spores/unit to less than one spore/unit; from this study came the 12D concept (12 decimal reductions in survival population). Later, a correction in come-up time resulted in the amendment of that heating time to 2.45 min to achieve the same lethal effect. Data on thermal death times combined with HP studies can be employed to calculate a safe heat process for any canned food. In conducting thermal death time studies on spore suspensions, the logarithmic survival curve permits determination of decimal reduction values (D values) – the time in minutes at constant temperature necessary to destroy 90% of the spores. By plotting determined D values on a logarithmic scale (ordinate axis) against temperature on a linear scale (abscissa axis), a so-called thermal death time curve can be constructed. From this plot, the z value can be obtained, which, in essence, is the negative inverse of the slope of the curve and represents the number of degrees Fahrenheit or Celsius required for the curve to traverse one logarithmic cycle. In other words, the z value denotes the degrees required to effect a tenfold change in time to achieve the same lethal effect.
Heat penetration determinations and thermal process calculations
47
The sterilizing value of a process is generally expressed as the F-value, which is equivalent to the number of minutes required to destroy a specified number of spores at 250 °F (121 °C) when z equals 18 °F (10 °C). The z value for C. botulinum spores is generally regarded to be 18 °F (10 °C), while the time required at 250 °F (121.1 °C) to reduce the survival population by a factor of 12 decimal values (12D) in a phosphate buffer is 2.45 min (F0 = 2.45 or F18250 = 2.45). The z value indicates, for example, that at 232 °F (250 − 18 °F), 24.5 min heating would be required to achieve the same lethal effect. Theoretically, spore concentration (numbers per unit volume) does not influence the amount of heat necessary to achieve an equivalent lethal effect on a given spore load. For example, whether 1012 spores are distributed among 1000 containers or one spore residues in each of 1012 containers, each container should have to receive the same thermal process to effect an equivalent reduction in the total initial spore population. While adherence to the 12D concept for the thermal processing of low-acid foods might appear to be excessively conservative, it has served the canning industry well in minimizing the incidence of botulism. In actual practice, most thermal processes are calculated to provide additional safety; few recommended processes for canned low-acid foods (other than cured meats or sweetened foods) provide for an F0 value of less than 3.0. Over and above safety considerations, several non-toxin-producing bacterial species are known to occur that produce spores with heat resistances significantly higher than C. botulinum spores. To cope with these potential spoilage micro-organisms (both mesophilic and thermophilic), the canner generally chooses to process many low-acid foods at values significantly higher than those required to achieve safety. Other foods may receive even higher processing levels to achieve a desirable texture. In such instances, the 12D concept is of no consequence, except in evaluating safety of a particular lot of such foods that has received an inadvertent underprocess. Most recommended thermal processes for foods subjected to one of the high- temperature/short-time (HTST) methods (e.g. in aseptic processing) provide for F values substantially higher than those used for in-can processes. Such high processing schedules are feasible because of the decreased impairment to quality of the foods subjected to high temperatures for a very short time. Because of the importance of a few seconds or minor temperature fluctuations at the high sterilizing temperatures, a reasonable margin above the minimum processing values is required to ensure a safe process.
3.4 HTST processing More thermally efficient methods for heat processing foods, particularly the HTST methods, impose even greater demands for strict controls. Furthermore, the range of new ingredients on the market for the formulation of new canned foods necessitates more rigid control during production. The canner who wants to take advantage of higher quality products or formulate new products to be cooked by HTST processes must understand the principles underlying these procedures and appreciate the significance of any change of ingredients or operation. The following are critical points that
48
A Complete Course in Canning and Related Processes
must be included in an effective quality control programme for food processed by one of the HTST methods: 1. The formula of the product and the character ingredients must be constant because the heat process is designed for a very specific set of conditions. If any of these ingredients are changed, the heating parameters of the product must be changed and a new time and temperature process is required for product safety. 2. The consistency of each batch of product must be measured, controlled, and recorded. If the consistency changes during heat processing, the change must be taken into consideration when establishing a safe process. A procedure to lower the consistency of a given batch as it is made up or provisions for correcting mistakes in formulation must be available. If variations in ingredients cause the batch to be too thick for appropriate agitation, adding water may be appropriate as a corrective measure. 3. An experimentally inoculated pack should be made before the initial startup of any processing line. Known numbers of bacterial spores, with a predetermined heat resistance in the specific product, are added to the formulated food, and the product is heat processed through the line. Attempts are made to select processing times and/or temperatures to assure some spoilage by the added spores at the lower processing levels. From the results obtained, combined with HP data (if it can be taken), a safe processing schedule can be estimated.
3.5 Inoculated pack studies The most common spores employed in inoculated pack studies for sterilised foods are those obtained from the putrefactive anaerobe, Clostridium sporogenes, NFPA strain 3679 (P.A. 3670). This micro-organism possesses characteristics quite similar to C. botulinum, except that it does not produce a toxin. Spores are harvested from a culture grown in a laboratory medium under conditions known to yield spores having a heat resistance slightly greater than the most heat resistant C. botulinum spore. To conduct a test on the inoculated pack, a formulated food is generally inoculated with enough of the spore suspension to yield at least 10,000 viable spores per container. The following considerations must be taken into account when carrying out these studies: 1. The inoculated pack studies should be conducted in the actual processing equipment used for the commercial food production. Data obtained solely from pilot line equipment are inadequate. 2. If the formulated food contains a dry ingredient (e.g. spaghetti, macaroni, noodles, tapioca), spores used for the inoculated pack should be incorporated into the dry ingredient as fabricated. A representative portion of the inoculated dry ingredient is then added to the formula to yield the desired initial spore count. Care should be taken to ensure that the inoculated dry ingredient has rehydration characteristics similar to the ingredient that will be used commercially and that the individual particle size is as large as, or larger than, any found in the commercially available material. 3. It is well known that spores are more resistant to dry than to wet heat. Therefore, it is necessary to heat process the product adequately after the dry ingredient has been rehydrated in order to ensure safety. Clumping of spores or product, and other physical barriers to
Heat penetration determinations and thermal process calculations
49
HP, must be avoided. Failure to adequately control the above factors may require a more severe process, thus defeating the main advantage of the HTST system. 4. The filler bowl temperature for the formulated product should be controlled. In addition, when the cook is started, the initial temperature (IT) of the product must be controlled and recorded. In the event of delays due to line breakdown or other causes resulting in low IT, the ensuing thermal process must be adjusted to compensate. 5. Container headspace must be positively controlled. In high-speed agitating cookers, dependence is placed upon the bubble from the headspace to stir the product, thus simulating a convection heat process. Therefore, constancy of bubble size must be maintained. Controlling headspace by determining container weight is not adequate in most instances because of variations in specific gravity of the product. Precautions must be taken to avoid incorporation of air into the product because this will affect the headspace needed to achieve the desired weight. To assure constancy of agitation, the ratio of solid to liquid must be controlled. 6. While the product is in an agitating cooker, the time and temperature of the cook, as well as the rotation speed of the cooker, must be controlled and recorded. Specific precautions must be followed to avoid inadvertent changes in time, temperature, or rotational speed of the cooker. 7. Any canner who installs a new high-speed system to produce new products, or to increase production of an existing product, must take into account the capacity of preparation facilities.
The development of new systems, techniques, and formulations for new canned products is inevitable. However, it is necessary that the food processor and the research, development, and quality assurance staff devise appropriate control and monitoring procedures. The following factors will need consideration: formula, consistency, IT, headspace, processing time, processing temperature, and rotation speed. Regardless of the canning technique employed to process low-acid canned foods, failure to achieve commercial sterility in any lot may result from failure to recognise certain product characteristics or operating conditions that affect lethality. Ideally, fail-safe procedures should be devised to assure that commercial sterility is achieved.
3.6 HP determinations The aim of a HP study is to determine the heating and cooling behaviour of a specific product in order to establish a safe thermal process regime and to provide the data to analyze process deviations. Design of the study must ensure that all of the critical factors are considered to deliver the thermal process to the slowest heating point in the product. The numbers of instrumented sample containers and replicate retort runs is subject to much discussion (Bee & Park, 1978; Campden BRI, 1977; IFTPS, 1995; NFPA, 1985), with the final decision linked to the measured variability between samples and between runs. Modern data logging systems can provide the facility for taking multiple temperature measurements; therefore, large quantities of data can be taken more easily than was the case when the Campden BRI guidelines were written in 1977. These
50
A Complete Course in Canning and Related Processes
recommendations were to use three samples in three replicate runs, providing a total of nine measurements. This 3 × 3 system has served the industry for many years. The more common situation now is to take up to 10 samples in two replicate runs, providing that the variability between runs is within acceptable limits. However, there can be limitations on the number of probes that can be inserted through a packaging gland or through the central shaft of a rotating system, and in these situations at least two replicate runs should be completed. The HP study should be carried out prior to commencing production of a new product. Changes to any of the criteria that change the time–temperature response at the product’s slowest heating point will require a new HP study to be conducted. This includes changes to the product formulation, preparation procedures, and the packaging or packing conditions (e.g. vacuum). The conditions determined in the study are referred to as the scheduled heat process and must be followed for every production batch, with appropriate records taken to confirm that this was followed. No further temperature measurement within containers is required in production, although some companies do measure temperatures in single containers at defined frequencies. However, the conditions used in single container tests do not represent the worst case, and it is expected that the instrumented container shows a process value higher than that measured from the HP study. Such data are intended to show due diligence and are at best a comfort factor. An HP test is usually subdivided into two further stages when conducting the tests: • Firstly, to locate the product cold point in the container. • Secondly, to establish the process conditions that will lead to the scheduled process.
3.6.1 Locating the product cold point Within each food container, there will be a point or region that heats up more slowly than the rest; this is referred to as the ‘slowest heating point’ or ‘thermal centre’. It should be located using thermocouples or some other sensing method positioned at different places in a food container. For foods that heat mainly by conduction, the slowest heating point is at the container geometric centre. Also, if the process utilises rotation or agitation, the slowest heating point is forced to the container geometric centre. For foods that permit movement and heat by convection, this point is between the geometric centre and approximately one-tenth up from the base (in a static process). However, during the retort process, the food viscosity decreases in response to increasing temperature; as a result, the slowest heating point moves downward from the container geometric centre. The critical time in the process is when the lethal effect on the target micro-organisms is at its most significant, which is toward the end of the hold phase.
3.6.2 Establishing the scheduled process time and temperature The thermal process is finally established by measuring temperatures at the container slowest heating point, for a number of replicates that are placed in the cold spot(s) of
Heat penetration determinations and thermal process calculations
51
the thermal processing system. The data obtained are usually referred to as HP data. A point open to discussion is the number of replicates required for confidence in the data. As described earlier in this chapter, this depends on the variability between data sets, with 3 × 3 and 2 × 10 being common approaches.
3.7 Process establishment methods The theory behind microbiological reductions by heat is that the kinetics is approximately first order. Thus, an F-value calculated from a spore reduction test should be the same as that calculated from the time-temperature integration (see Eqn (3.1)). t
F
³ 10 0
T T ref ]
dt
N D T log § 0 · © N ¹
(3.1)
where T is the product temperature (°C), Tref is the reference temperature for the DT value (°C), t is the process time (minutes), and z is the kinetic factor or the temperature change required to effect a ten-fold change in the DT value (°C). This integration is nowadays usually done within the data logger software to allow a thermal process to be operated until the target F- or P-value is reached. This calculation method is referred to as the General method. Other methods that use the measured times and temperatures within predictive models are also acceptable and used widely. Details of how each method works are given later in this chapter. For a process value to be calculated, it will need to be based on measurements of temperature distribution (TD) in the retort and HP in the product. TD is almost impossible to measure without temperature sensors; however, HP usually uses temperature sensors but can also use log reduction methods. The next sections provide outline information on temperature measurement systems and log reduction methods, with references for further information should this be required.
3.7.1 Temperature measurement systems for TD and HP testing Modern data loggers are typically multichannel systems with digital outputs, allowing data to be recorded directly to a laptop computer for display and to maintain permanent records. Thermocouples based on type T (copper/constantan) with polytetrafluoroethylene insulation are most common because they are inexpensive, accurate over the desired temperature range, and respond rapidly to changing temperature. Other types of temperature measurement are based on a change in electrical resistance with temperature, such as a thermistor, and platinum resistance thermometers (e.g. PT100). These are most commonly used in data loggers where the logging unit is a self-contained unit (e.g. Ball Datatrace, TMI Orion, or Ellab Tracksense). These are also referred to as resistance temperature detectors. Calibration of a temperature sensor against a traceable instrument is essential; otherwise, the user does not have confidence in the numbers obtained. This is usually done each time the instrument is used in a set of TD or HP trials. The method recommended
52
A Complete Course in Canning and Related Processes
by IFTPS (1995) and Bee and Park (1978) is to use the master temperature indicator (MTI) on the retort, which itself must be calibrated at no less than six-monthly intervals. It is worth noting that the new versions of Ellab Tracksense systems are claimed to hold the factory calibration for 12 months and are supplied with a calibration certificate. The use of time and temperature measurements to validate thermal processes is likely to remain the most widely used method. Advances in microchip technology are providing more computing power to analyze these results and increase the accuracy in defining the calculated process values. Historic canning processes were evaluated using lethal rates at time steps of 1 min because of the capabilities of data recorders at that time, but modern process values can be estimated from temperature measurements taken at much reduced frequencies (e.g. every second). The increased data storage capabilities also allow for more temperature probes to be used in each TD or HP test, with multiple loggers linked together in the software. With such systems, it is possible to exceed the number of suggested working probes to define an HP test (3 × 3 or 2 × 10). Several calculation methods are available for setting the scheduled process conditions. The most commonly used methods will be described in outline here, with the references provided for further information.
3.8 Process calculation methods 3.8.1 General method The reference calculation method is known as the General method (Bigelow et al., 1920), and it provides a process value (F-value) that is sometimes sufficient for validation purposes. It converts measured times and temperatures to F-values from accumulated lethal rates. A lethal rate is a relative term that compares the micro-organism killing effect at a measured temperature to 1 min at a reference temperature. Equation (3.1) presents the lethal rate equation as part of the integrated F-value calculation, but the lethal rate (L) itself is as given in Eqn (3.2).
L = 10
T − Tref z
(3.2)
Integration of lethal rates over the measured times and temperatures is usually done using the trapezoidal method, with the calculation routines embedded within a data logger software. The General method allows the user to follow the F-value calculation during the process, usually in real time (with probes); however, for loggers, this is done when they are removed from the processing system. Time to achieve a target F-value is determined from the data and the process conditions can be established. While this is a simple method, it does have some major limitations: • F-values are relevant to the TD and HP test conditions on the day of testing and cannot be easily transferred to another set of conditions. • Changes in critical variables cannot be evaluated without a new HP test, such as initial product temperature, retort temperature, come-up time, and cooling profile.
Heat penetration determinations and thermal process calculations
53
• Deviations to the established process conditions cannot be evaluated by calculation and need to be experimentally simulated.
3.8.2 Ball method For the above reasons, there are alternatives to the General method that are widely used in the industry. Apart from process calculations based on log reduction and General method calculations, most process calculation methods use heating (fh) and lag (j) factors. Heating factors originated in the canning industry (Ball, 1923, 1927) as a measure of the product heating rate used for calculating process times for canned foods. By definition, the heating factor is the time taken for the difference between environment and product temperature to reduce by 90%. They are important terms in thermal process data analysis because they provide information on the rate of heating for a container of food. Figure 3.1 shows an example of a logarithmic heating curve used to calculate fh and j factors. The Ball method is widely used within the United States and countries exporting thermally processed foods to the United States. Its use in Europe is less widespread. However, the Ball method offers options to the General method that enables some analysis of ‘what-if’ scenarios to be calculated. The Ball method has three parts to it and can be described in its most simple form as • An equation to the straight heating line in Figure 3.1, which is effectively the gradient at the end of heating. • This is connected by an experimentally determined complex hyperbolic function to the cooling line. • An equation to the straight line cooling line (not shown in Figure 3.1), which is effectively the gradient at the end of cooling.
±ORJ 7U
± ± ±
&DPSGHQ &KRUOH\ZRRG)RRG5HVHDUFK$VVRFLDWLRQ&7HPS3URJUDP
Figure 3.1 Example of a logarithmic heating curve for canned meat used to calculate fh of 50.4 min and j of 1.45.
54
A Complete Course in Canning and Related Processes
The above description does not do the Ball method justice but does describe the constituent parts. The method was derived for steam retorts because these were the main type used around the time Ball carried out his experiments (Ball, 1923, 1927). Hence, the complex connecting routine between heating and cooling is most accurate for steam retorts; adjustments to this are required for other retort types, such as water retorts. It is important to note that the Ball method calculates an F-value at the end of cooling. Routines can be used to back calculate an end of heating F-value but these begin to lose their accuracy. If an end of heating F-value is required, then it is easier to use only the first equation to the heating part of the logarithmic heating curve and express this as in Eqn (3.3).
RT – T –2.303t/fh (3.3) = j.e RT – I'T'
where j is the heating lag factor, defined as a measure of the thermal lag before the can temperature responds to the changing environment temperature; fh is the heating factor, defined as the temperature response parameter derived from the logarithmic heating curve (minutes); RT is the retort temperature during the hold period (°C); I′T′ is the product pseudo-IT at steam on (°C); and T is the product temperature after (t) minutes heating time (°C). The I′T′ is the temperature at which the straight line for heating (log-linear line) would have started had there not been a lag phase. Most Ball method calculations are now computerised versions of the tables and nomograms published by Ball (Ball & Olsen, 1957). Several attempts have been made since Ball to improve on the method, particularly for cooling lethality calculations.
3.8.3 Numerical methods The use of more advanced mathematical models to evaluate and predict process times and temperatures has increased as the computing power available to a thermal process authority has increased. Examples of numerical calculation software are CTemp (Campden BRI) and NumeriCAL (JBT Corporation). These models deal with the physics of heat transfer and use finite differences to solve Fourier’s partial differential equations. Therefore, they have a more scientific approach that allows variable retort profiles as the boundary conditions. This means that come-up and cool-down profiles can be defined, as well as deviations in processes caused by retort temperature fluctuations (Tucker, Noronha, & Heydon, 1996). These predictive modelling approaches not only help with deciding the fate for batches of product that have undergone a process deviation, but the task of process establishment is made more straightforward. For example, the models can be used to evaluate changes to initial product temperatures, shortened come-up times, or low retort temperatures. They offer considerable flexibility to make the job of process establishment easier.
Heat penetration determinations and thermal process calculations
55
3.9 Some causes of unreliable heat penetration data Any of the process calculation methods, whether based on Ball or numerical methods, require validation against a set of conditions that are representative of the thermal processes. Using a model for prediction outside of these validation boundaries should be done with caution because there are no guarantees that the model will predict with sufficient accuracy. If the purpose is only to provide a rough estimate of process time for a new product, then this is acceptable. Any process times and temperatures will need to be confirmed by HP data so that the model can be used again for fine tuning the conditions. It would be unwise to apply a model for establishing safe process times and temperatures without a basis of a good set of HP data. It makes sense that any process calculation method can only be as accurate as the HP data upon which it is based. This is where the skills of a thermal processing specialist are needed. It is easy to set up experiments to measure HP during a thermal process, but it is much more difficult to do this with a high level of accuracy. There are many ways that the data can be gathered incorrectly. The following is a partial list of some factors that are known to be common causes of unreliable HP data: 1. Thermocouple (temperature) readings are not continued for a sufficient length of time to adequately define heating rate or rates. The greatest accuracy when calculating heating and lag factors is when the heating curve becomes asymptotic, which is when the CCT is within one degree of the retort. 2. HP tests are conducted in a retort load of commercial production and stopped at the end of the scheduled process for quality determinations, rather than continued long enough to obtain sufficient data. 3. Frequency of readings is not sufficient to obtain accurate heating rate or rates. This is critical for fast heating products (fh < 20 min), usually under agitation, in which the rate of change in temperature tends to be uneven as regions of colder and hotter liquids are mixed. 4. Erroneous temperatures are received as a result of inadequate electrical grounding of the data logger. This is a rare occurrence nowadays as more remote loggers are used. It used to happen regularly with water immersion retorts. 5. No note of the retort come-up time is taken, so an estimate must be made from the retort temperature readings. It is useful to know the difference between the retort MTI and the retort temperature sensor used for HP measurements. A difference is common because of the physical size of the MTI, which can cause a thermal lag. 6. The come-up time for HP testing is significantly different from that used in commercial practice. This places a reliance on the process calculation method being able to predict the effect of different come-up times with sufficient accuracy. 7. Multiple thermocouples are used in small cans of product. Heat conduction to the probe tips and heat capacity of the probes are issues that should be considered. 8. No cold spot study is carried out, or insufficient replicates are taken at the cold spot thermocouple location. In an agitation process, it is likely that the cold spot will be at the geometric centre. 9. No times are noted on the temperature recorder, such as the time ‘steam on’, which may not be obvious from the HP data or recorder trace. It is good practice to note as many events as possible. It is better to have more events noted than too few.
56
A Complete Course in Canning and Related Processes
10. N o reference readings are taken from the MTI during the HP test. These are essential at the start of the process while the retort is coming up to temperature. 11. Erratic retort temperature control occurs during the test. 12. Critical factors associated with the product and processing system are not recorded or controlled. It is essential to set up a HP test based on the worst heating conditions that a product might experience. This requires a precise thought process so that each factor can be set at the level that would lead to the product heating slowly. 13. Large temperature disagreement exists between the thermocouple attached to the MTI and the MTI itself. The MTI is always taken as the reference instrument. 14. ITs of the test cans are significantly different from those used in commercial production. This is likely to result from cans filled warm/hot, which then cool before the HP test is ready to start. This will artificially increase the lag factor but not the heating factor, making process calculations less accurate. 15. There is an excessive delay in running the HP test after the containers are sealed. This can create differences in IT but is more critical with products that rehydrate during the process and change their heating rate as a result. Examples are pasta, pulses, and rice products. 16. Product for tests not prepared according to procedures used commercially for raw product preparation or condition. This can affect the heating rates. 17. Differences in processing temperatures between HP tests and commercial practice occur, especially when pilot retorts are used. This is critical for process calculations made with the General method. In theory, heating and lag factors should not be changed, although it is good practice to minimise retort temperature differences.
3.10 HTST: a special type of heat penetration test HTST processing is limited to fluid or pumpable foods. Current thermal processing regulations assume that the entire sterilisation or pasteurisation is achieved in the holding tube and no credit is given to that during heating or cooling. Therefore, the holding time must be sufficiently long to achieve the desired process. The safe design criterion for calculating the length of a holding tube is based on the fluid elements with the shortest residence time, which travels along the centre line of the holding tube. As a safe criterion, the maximum velocity is assumed to be equal to twice the average velocity, as calculated from the volumetric flow rate and holding tube dimensions. This assumption is based on laminar flow in the holding tube. There are two types of flow that will be experienced by a flowing food – laminar and turbulent – but with a transitional region between the two types as the flow changes from one to the other. It is important to know which flow regime is present in all parts of a continuous process so the residence times can be calculated correctly. Laminar flow is assumed to occur up to Reynolds numbers of around 2100, whereas turbulent flow occurs at greater than 10,000 (see Eqn (3.4)). The region in between 2100 and 10,000 is referred to as the transitional region, because the flow regime is changing from laminar to turbulent. This is a region that equipment designers will try to avoid because of the uncertainties in flow behaviour and the key relationship between fastest and mean velocity. Should a heat exchanger be operated under transitional flow,
Heat penetration determinations and thermal process calculations
57
then it would be safe to assume that the fastest liquid along the pipe centre could be travelling twice as fast as the mean velocity. The US FDA takes a Reynolds number of 4000 as the division between laminar and turbulent flow and assumes there is no transitional region. This is a simplification of what happens in practice, but it does make process calculations much clearer. For viscous products, the flow conditions are nearly always laminar. A tomato paste steriliser, for instance, operates at Reynolds numbers (Re) around 1. However, liquid foods that are low in viscosity, such as milk and juice, are more likely to experience turbulent flow, so the assumption is made that the maximum velocity is 1.2 times the mean.
Re =
d h .ρ .v
µ
(3.4)
where dh is hydraulic diameter of the processing system (m); μ is dynamic viscosity of liquid (Pa s); ρ is density of liquid (kg/m3); and v is velocity in the processing system (m/s). Equation (3.4) is for Newtonian liquids that have a viscosity independent of shear rate. Most, if not all, formulated foods contain thickening agents in which the viscosity depends on the shear applied in the processing system. These non-Newtonian foods are certain to flow under laminar conditions because of their high viscosity values, and they also display velocity profiles that are still more complex. As the degree of non-Newtonian behaviour increases, the velocity profile increases in flatness across the pipe cross-section. This means, in practice, that the maximum velocity can be less than its Newtonian value of twice the mean velocity. However, there are few commercial operations that do not apply a factor of 2 when calculating holding tube length, irrespective of the measured flow behaviour index.
3.11 Summary 1. A process is the application of heat to food for a scientifically determined time and temperature, which is known as the scheduled process. 2. A process that has been scientifically determined is specific for the given product, its formulation, methods of preparation, container size, and type of retort system. 3. The determination of a process depends upon reliable heating information for the product and the heat resistance of micro-organisms in the product. 4. The heat resistance of micro-organisms depends upon the micro-organism used, the food in which it is heated, and the food in which the organisms grow. 5. The determination of heating data (time/temperature) should be conducted on a product that simulates commercial preparation or in the commercial process itself. 6. The data obtained from thermal resistance and HP tests are used to calculate a process target of equivalent time at a reference temperature. 7. It is sometimes desirable to check the calculated process by means of an inoculated test pack, particularly if the process does not lend itself to HP studies.
58
A Complete Course in Canning and Related Processes
References Ball, C. O. (1923). Thermal process time for canned food. Bulletin of the National Research Council, Washington, DC, 7 Part 1, Number 37. Ball, C. O. (1927). Theory and practice in processing. The Canner, 64(5), 27. Ball, C. O., & Olsen, F. C. W. (1957). Sterilization in food technology. Theory, practice and calculation. New York: McGraw-Hill Book Co. Bee, G. R., & Park, D. K. (1978). Heat-penetration measurement for thermal-process design. Food Technology, 32(6), 56–58. Bigelow, W. D., Bohart, G. S., Richardson, A. L., & Ball, C. O. (1920). Heat penetration in processing canned foods. National Canners Association Bulletin, 16-L, 128 Washington, DC, USA. Campden BRI. (1977). Guidelines to the establishment of scheduled heat processes for low-acid foods. Technical Manual No. 3. Chipping Campden, Glos., GL55 6LD: Campden BRI. Campden BRI. (2006). Pasteurisation: A food industry practical guide (2nd ed.). Chipping Campden, Glos., GL55 6LD: Guideline No. 51. Campden BRI. Department of Health. (1994). Guidelines for the safe production of heat preserved foods. London: HMSO (out of print). IFTPS. (1995). Protocol for carrying out heat penetration studies. PO Box 2764, Fairfax, VA, USA: IFTPS. NFPA. (1985). Guidelines for thermal process development for foods packaged in flexible containers. Washington, DC. Raatjies, G. J. M., & Smelt, J. P. (1979). Clostridium botulinum can grow and form toxin at pH value lower than 4.6. Nature, 281, 398–399. Tucker, G. S., Noronha, J. F., & Heydon, C. J. (1996). Experimental validation of mathematical procedures for the valuation of thermal processes and process deviations during the sterilization of canned foods. Transactions of the Institution of Chemical Engineers, Food & Bioproducts Processing, 74, 140–148 Part C.
3
3.1 Introduction In 1917, the use of thermocouples in the United States was introduced by the National Canners Association to measure temperatures during heating and cooling of foods in sealed containers. This led to graphic mathematical and computerised procedures to estimate the minimum retorting conditions required to produce ‘commercially sterile’ foods without excessive damage to their eating quality or nutritional value. Since that time, progressive scientific, technological, and engineering improvements have been achieved in canning equipment and in packaging. The conventional steam-still retort has lost popularity in favour of continuous hydrostatic and reel-and-spiral cookers, overpressure and agitating retorts, and, albeit to a lesser extent, aseptic processing equipment. Packaging has developed to the position in which metal cans have evolved from cylinders into complex geometries, with considerable consumer preference for glass jars, plastic pouches, semi-rigid trays, and even cartons. These present unique challenges for heat penetration (HP) measurement. Food that is thermally processed in hermetically sealed containers is heated throughout by employing an external source of heat. If the food is a liquid (e.g. broth) or contains a liquid of low viscosity (e.g. peas in brine), the heat is transferred via convection. If it is a solid (e.g. meat, fish) or highly viscous (e.g. cream-style corn), it is heated primarily via conduction; as a result, a greater portion of the contents is severely overprocessed in order to sterilise the small volume occupying the geometric centre. Agitation during cooking (axial or end-over-end) may effectively reduce the times and temperatures required for safe processing of some foods as a result of the stirring action that eliminates conduction as the sole mode of heat transfer. In addition, processing at higher temperatures for shorter periods of time effectively reduces damage to product quality while achieving commercial sterility. Hence, increasing interest is devoted to continuous rotary cookers (axial rotation) and to batch agitating retorts (end-over-end rotation). Each of these technological and engineering improvements, in turn, opens up new possibilities in the variety of foods that can be preserved through heat processing. On the other hand, each imposes upon the processor the necessity for more stringent and sophisticated procedures to assure adequate control. Furthermore, the newer heat processing methods are employed for the production of high-quality formulated foods that were not possible to produce using the conventional still retort. Many of these formulated foods are designed to be consumed without further heating, which, in itself, imposes the need for the processor to exercise greater stringency in order to assure product sterility. These processes are discussed in more detail in the section on sterilization systems. A Complete Course in Canning and Related Processes. http://dx.doi.org/10.1016/B978-0-85709-678-4.00003-8 Copyright © 2015 Elsevier Ltd. All rights reserved.
44
A Complete Course in Canning and Related Processes
3.2 pH classification of canned foods Canned foods may be classified into two categories based upon pH and acidity and on the necessary thermal processing required to effect safety as well as microbiological stability: 1. Low acid: above 4.6 2. High acid: 4.6 or lower
The low-acid category is the only one of importance from the standpoint of the botulinum hazard. No Clostridium botulinum spore, regardless of type, has been detected that will germinate, grow, and produce toxin in a food having a pH of 4.7 or lower. Foods in the low-acid range must be fully retorted to assure safety; this is sometimes referred to as a botulinum cook, which requires a minimum equivalent process of 3 min at 250 °F (121.1 °C) – the F0 3 process. The US Food and Drug Administration (FDA) has promulgated regulations designed to control production of low-acid canned foods. The regulations, among other things, (1) provide for the registration of all firms producing low-acid canned foods; (2) provide for the filing by each firm of a detailed description of the heat processes used for each low-acid food in each container size (scheduled process); (3) define good manufacturing practices in the processing of low-acid canned foods; and (4) require establishments that fail to comply with the regulations to operate under emergency permit control of the FDA. These regulatory controls serve to further assure the safety of canned foods. Food processors in countries that export to the United States are required to follow these regulations in order that their products are accepted into the country. However, not all countries apply the same level of regulatory compliance and instead rely on national guidelines that are enforced locally, often by the retailer. Standards in the manufacture of thermally processed foods are very similar to those in the United States, and the safety record in most countries is every bit as high. Any low-acid food that achieves an F0 3 process is referred to as commercially sterile; however, both pasteurised and sterilised foods can be considered commercially sterile because of the caveat regarding the micro-organisms deemed ‘capable of growing in the food’. One of the definitions of commercial sterility is given by the UK Department of Health (DoH, 1994): “Commercial sterility is the condition achieved by the application of heat which renders food free from viable micro-organisms, including those of known public health significance, capable of growing in the food at temperatures at which the food is likely to be held during distribution and storage.” An F0 3 process will be safe with respect to C. botulinum spores, but it is important to remember that the product may contain a small number of surviving spores of more heat-resistant spoilage micro-organisms. This is because there are many more types of spoilage micro-organisms than those of concern to public health, and therefore it is not surprising that a greater range of heat resistance is found. Those foods in the high-acid range (pH 4.6 or lower) generally require a considerably lower level of heat processing to effect preservation because the processing schedule is primarily designed to kill vegetative cells. A few spore-bearing bacteria, particularly Clostridium butyricum, Bacillus coagulans, and related species, are able
Heat penetration determinations and thermal process calculations
45
to grow in foods having a pH value as low as 4.1. However, they have no known public health significance. Food processors have learnt to prevent spoilages of this nature by employing good food hygiene practices that reduce the initial spore load of the food combined with appropriate heat processing schedules. The efficacy of heat in destroying micro-organisms is influenced by various factors. It is for this reason that mild heat alone cannot be used to produce safe stable pasteurised foods. The most common factors used in conjunction with heat are pH, water activity, and low-temperature storage. Even if an appropriate heat treatment is applied at the desired level, if these other factors are not in control, there is the potential for micro-organisms to grow and spoil the food. The following statements are usually true: • Most spore-forming bacteria are inhibited by a combination of pH < 4.6 and aw < 0.90. • Acid tolerant spore-forming bacteria will survive and grow at pH > 3.8. This includes many of the butyric anaerobes, a group of bacteria that under anaerobic conditions can ferment sugars to butyric acid, including Bacillus macerans and Bacillus polymyxa. • Some xerophilic and osmophilic spore-forming yeasts and moulds can grow below aw 0.85. • Antimycotic agents (preservatives) have been successfully used in foods to prevent the outgrowth of yeasts and moulds. • C. botulinum does not usually germinate and grow in foods with a pH < 4.6 or an aw < 0.94, although some experimental conditions have shown growth and toxin production by C. botulinum at pH less than 4.5 (Raatjies & Smelt, 1979). • Chilling is only a short-term barrier to microbial growth.
It should be noted that the thermal destruction of some yeasts and moulds is more complex than simple logarithmic destruction. Ascospores can get activated at temperatures as high as 85 °C. Unusual heat resistance has been reported in pasteurised beverages and sauces that contained ascospores of Talaromyces trachyspermus and Neosartorya fisheri (Campden BRI, 2006). In a product with pH < 3.8, there is very low risk of germination and outgrowth of bacterial spores, but certain yeasts and moulds can thrive. The heat process for foods of pH < 3.8 only needs to destroy those yeasts and moulds. In products with pH > 3.8, other organisms are capable of surviving, so the heat process applied must be sufficient to ensure that their numbers are destroyed. As a generalization, within the acid foods group, products can be further divided into pH categories so that a minimal, but effective, heat treatment can be applied.
3.2.1 High acid: pH < 3.8 This group includes cherries, plums, apricots, berries, citrus, and sometimes pineapples. The spoilage organisms include yeasts, moulds, and lactic acid bacteria. Due to the high acidity of these fruits, they generally are very susceptible to softening during heat processing, and processors usually try to process them as close to the minimum requirement as possible. Process recommendations for high acid products sometimes target a can-centre temperature (CCT) over 70 °C, with a holding time of 2–3 min, but practically a CCT of 80–82 °C is usually targeted for most products in this group. Ascospores of Byssochlamys fulva or Byssochlamys nivea can survive this process,
46
A Complete Course in Canning and Related Processes
and good hygiene in production is essential to ensure that they are not present in the canned product. If they grow, they can produce enzymes that can soften the fruit.
3.2.2 Acid: pH 3.7–4.2 This group of fruits includes peaches, apples, and pears. General process recommendations are for the CCT to reach 85 °C for 5 min or 95 °C for 30 s. This is effectively a P-value of 5 min at 85 °C. The process is intended to target spoilage organisms and includes yeasts, moulds, lactic acid bacteria, and some acid-tolerant bacteria, such as B. coagulans. For all acid and high-acid fruit products, good hygiene during product peeling and sorting is critical because many of the organisms that can spoil canned fruit are sufficiently heat resistant that they cannot be processed away without damaging the texture and colour.
3.2.3 Acid: pH 4.2–4.5 This group includes tomatoes and some pears. The heat resistance of micro-organisms is greater at the slightly higher pH, so processing to a CCT of 100–104 °C is recommended. If the fruit texture is unable to withstand this high process, the product must be acidified and a lower process applied.
3.3 Thermal death time Safety from botulism in canned low-acid foods stems from the pioneering research done by scientists of the National Food Processors Association in the early 1920s. They determined the thermal resistance of spores harvested from the most heat- resistant C. botulinum strain known to them. Their studies demonstrated that, by extrapolation from the exponential survival curve, it was necessary to heat a spore suspension in phosphate buffer for 2.78 min at 250 °F (121.1 °C) to reduce the survival population from just over 1011 spores/unit to less than one spore/unit; from this study came the 12D concept (12 decimal reductions in survival population). Later, a correction in come-up time resulted in the amendment of that heating time to 2.45 min to achieve the same lethal effect. Data on thermal death times combined with HP studies can be employed to calculate a safe heat process for any canned food. In conducting thermal death time studies on spore suspensions, the logarithmic survival curve permits determination of decimal reduction values (D values) – the time in minutes at constant temperature necessary to destroy 90% of the spores. By plotting determined D values on a logarithmic scale (ordinate axis) against temperature on a linear scale (abscissa axis), a so-called thermal death time curve can be constructed. From this plot, the z value can be obtained, which, in essence, is the negative inverse of the slope of the curve and represents the number of degrees Fahrenheit or Celsius required for the curve to traverse one logarithmic cycle. In other words, the z value denotes the degrees required to effect a tenfold change in time to achieve the same lethal effect.
Heat penetration determinations and thermal process calculations
47
The sterilizing value of a process is generally expressed as the F-value, which is equivalent to the number of minutes required to destroy a specified number of spores at 250 °F (121 °C) when z equals 18 °F (10 °C). The z value for C. botulinum spores is generally regarded to be 18 °F (10 °C), while the time required at 250 °F (121.1 °C) to reduce the survival population by a factor of 12 decimal values (12D) in a phosphate buffer is 2.45 min (F0 = 2.45 or F18250 = 2.45). The z value indicates, for example, that at 232 °F (250 − 18 °F), 24.5 min heating would be required to achieve the same lethal effect. Theoretically, spore concentration (numbers per unit volume) does not influence the amount of heat necessary to achieve an equivalent lethal effect on a given spore load. For example, whether 1012 spores are distributed among 1000 containers or one spore residues in each of 1012 containers, each container should have to receive the same thermal process to effect an equivalent reduction in the total initial spore population. While adherence to the 12D concept for the thermal processing of low-acid foods might appear to be excessively conservative, it has served the canning industry well in minimizing the incidence of botulism. In actual practice, most thermal processes are calculated to provide additional safety; few recommended processes for canned low-acid foods (other than cured meats or sweetened foods) provide for an F0 value of less than 3.0. Over and above safety considerations, several non-toxin-producing bacterial species are known to occur that produce spores with heat resistances significantly higher than C. botulinum spores. To cope with these potential spoilage micro-organisms (both mesophilic and thermophilic), the canner generally chooses to process many low-acid foods at values significantly higher than those required to achieve safety. Other foods may receive even higher processing levels to achieve a desirable texture. In such instances, the 12D concept is of no consequence, except in evaluating safety of a particular lot of such foods that has received an inadvertent underprocess. Most recommended thermal processes for foods subjected to one of the high- temperature/short-time (HTST) methods (e.g. in aseptic processing) provide for F values substantially higher than those used for in-can processes. Such high processing schedules are feasible because of the decreased impairment to quality of the foods subjected to high temperatures for a very short time. Because of the importance of a few seconds or minor temperature fluctuations at the high sterilizing temperatures, a reasonable margin above the minimum processing values is required to ensure a safe process.
3.4 HTST processing More thermally efficient methods for heat processing foods, particularly the HTST methods, impose even greater demands for strict controls. Furthermore, the range of new ingredients on the market for the formulation of new canned foods necessitates more rigid control during production. The canner who wants to take advantage of higher quality products or formulate new products to be cooked by HTST processes must understand the principles underlying these procedures and appreciate the significance of any change of ingredients or operation. The following are critical points that
48
A Complete Course in Canning and Related Processes
must be included in an effective quality control programme for food processed by one of the HTST methods: 1. The formula of the product and the character ingredients must be constant because the heat process is designed for a very specific set of conditions. If any of these ingredients are changed, the heating parameters of the product must be changed and a new time and temperature process is required for product safety. 2. The consistency of each batch of product must be measured, controlled, and recorded. If the consistency changes during heat processing, the change must be taken into consideration when establishing a safe process. A procedure to lower the consistency of a given batch as it is made up or provisions for correcting mistakes in formulation must be available. If variations in ingredients cause the batch to be too thick for appropriate agitation, adding water may be appropriate as a corrective measure. 3. An experimentally inoculated pack should be made before the initial startup of any processing line. Known numbers of bacterial spores, with a predetermined heat resistance in the specific product, are added to the formulated food, and the product is heat processed through the line. Attempts are made to select processing times and/or temperatures to assure some spoilage by the added spores at the lower processing levels. From the results obtained, combined with HP data (if it can be taken), a safe processing schedule can be estimated.
3.5 Inoculated pack studies The most common spores employed in inoculated pack studies for sterilised foods are those obtained from the putrefactive anaerobe, Clostridium sporogenes, NFPA strain 3679 (P.A. 3670). This micro-organism possesses characteristics quite similar to C. botulinum, except that it does not produce a toxin. Spores are harvested from a culture grown in a laboratory medium under conditions known to yield spores having a heat resistance slightly greater than the most heat resistant C. botulinum spore. To conduct a test on the inoculated pack, a formulated food is generally inoculated with enough of the spore suspension to yield at least 10,000 viable spores per container. The following considerations must be taken into account when carrying out these studies: 1. The inoculated pack studies should be conducted in the actual processing equipment used for the commercial food production. Data obtained solely from pilot line equipment are inadequate. 2. If the formulated food contains a dry ingredient (e.g. spaghetti, macaroni, noodles, tapioca), spores used for the inoculated pack should be incorporated into the dry ingredient as fabricated. A representative portion of the inoculated dry ingredient is then added to the formula to yield the desired initial spore count. Care should be taken to ensure that the inoculated dry ingredient has rehydration characteristics similar to the ingredient that will be used commercially and that the individual particle size is as large as, or larger than, any found in the commercially available material. 3. It is well known that spores are more resistant to dry than to wet heat. Therefore, it is necessary to heat process the product adequately after the dry ingredient has been rehydrated in order to ensure safety. Clumping of spores or product, and other physical barriers to
Heat penetration determinations and thermal process calculations
49
HP, must be avoided. Failure to adequately control the above factors may require a more severe process, thus defeating the main advantage of the HTST system. 4. The filler bowl temperature for the formulated product should be controlled. In addition, when the cook is started, the initial temperature (IT) of the product must be controlled and recorded. In the event of delays due to line breakdown or other causes resulting in low IT, the ensuing thermal process must be adjusted to compensate. 5. Container headspace must be positively controlled. In high-speed agitating cookers, dependence is placed upon the bubble from the headspace to stir the product, thus simulating a convection heat process. Therefore, constancy of bubble size must be maintained. Controlling headspace by determining container weight is not adequate in most instances because of variations in specific gravity of the product. Precautions must be taken to avoid incorporation of air into the product because this will affect the headspace needed to achieve the desired weight. To assure constancy of agitation, the ratio of solid to liquid must be controlled. 6. While the product is in an agitating cooker, the time and temperature of the cook, as well as the rotation speed of the cooker, must be controlled and recorded. Specific precautions must be followed to avoid inadvertent changes in time, temperature, or rotational speed of the cooker. 7. Any canner who installs a new high-speed system to produce new products, or to increase production of an existing product, must take into account the capacity of preparation facilities.
The development of new systems, techniques, and formulations for new canned products is inevitable. However, it is necessary that the food processor and the research, development, and quality assurance staff devise appropriate control and monitoring procedures. The following factors will need consideration: formula, consistency, IT, headspace, processing time, processing temperature, and rotation speed. Regardless of the canning technique employed to process low-acid canned foods, failure to achieve commercial sterility in any lot may result from failure to recognise certain product characteristics or operating conditions that affect lethality. Ideally, fail-safe procedures should be devised to assure that commercial sterility is achieved.
3.6 HP determinations The aim of a HP study is to determine the heating and cooling behaviour of a specific product in order to establish a safe thermal process regime and to provide the data to analyze process deviations. Design of the study must ensure that all of the critical factors are considered to deliver the thermal process to the slowest heating point in the product. The numbers of instrumented sample containers and replicate retort runs is subject to much discussion (Bee & Park, 1978; Campden BRI, 1977; IFTPS, 1995; NFPA, 1985), with the final decision linked to the measured variability between samples and between runs. Modern data logging systems can provide the facility for taking multiple temperature measurements; therefore, large quantities of data can be taken more easily than was the case when the Campden BRI guidelines were written in 1977. These
50
A Complete Course in Canning and Related Processes
recommendations were to use three samples in three replicate runs, providing a total of nine measurements. This 3 × 3 system has served the industry for many years. The more common situation now is to take up to 10 samples in two replicate runs, providing that the variability between runs is within acceptable limits. However, there can be limitations on the number of probes that can be inserted through a packaging gland or through the central shaft of a rotating system, and in these situations at least two replicate runs should be completed. The HP study should be carried out prior to commencing production of a new product. Changes to any of the criteria that change the time–temperature response at the product’s slowest heating point will require a new HP study to be conducted. This includes changes to the product formulation, preparation procedures, and the packaging or packing conditions (e.g. vacuum). The conditions determined in the study are referred to as the scheduled heat process and must be followed for every production batch, with appropriate records taken to confirm that this was followed. No further temperature measurement within containers is required in production, although some companies do measure temperatures in single containers at defined frequencies. However, the conditions used in single container tests do not represent the worst case, and it is expected that the instrumented container shows a process value higher than that measured from the HP study. Such data are intended to show due diligence and are at best a comfort factor. An HP test is usually subdivided into two further stages when conducting the tests: • Firstly, to locate the product cold point in the container. • Secondly, to establish the process conditions that will lead to the scheduled process.
3.6.1 Locating the product cold point Within each food container, there will be a point or region that heats up more slowly than the rest; this is referred to as the ‘slowest heating point’ or ‘thermal centre’. It should be located using thermocouples or some other sensing method positioned at different places in a food container. For foods that heat mainly by conduction, the slowest heating point is at the container geometric centre. Also, if the process utilises rotation or agitation, the slowest heating point is forced to the container geometric centre. For foods that permit movement and heat by convection, this point is between the geometric centre and approximately one-tenth up from the base (in a static process). However, during the retort process, the food viscosity decreases in response to increasing temperature; as a result, the slowest heating point moves downward from the container geometric centre. The critical time in the process is when the lethal effect on the target micro-organisms is at its most significant, which is toward the end of the hold phase.
3.6.2 Establishing the scheduled process time and temperature The thermal process is finally established by measuring temperatures at the container slowest heating point, for a number of replicates that are placed in the cold spot(s) of
Heat penetration determinations and thermal process calculations
51
the thermal processing system. The data obtained are usually referred to as HP data. A point open to discussion is the number of replicates required for confidence in the data. As described earlier in this chapter, this depends on the variability between data sets, with 3 × 3 and 2 × 10 being common approaches.
3.7 Process establishment methods The theory behind microbiological reductions by heat is that the kinetics is approximately first order. Thus, an F-value calculated from a spore reduction test should be the same as that calculated from the time-temperature integration (see Eqn (3.1)). t
F
³ 10 0
T T ref ]
dt
N D T log § 0 · © N ¹
(3.1)
where T is the product temperature (°C), Tref is the reference temperature for the DT value (°C), t is the process time (minutes), and z is the kinetic factor or the temperature change required to effect a ten-fold change in the DT value (°C). This integration is nowadays usually done within the data logger software to allow a thermal process to be operated until the target F- or P-value is reached. This calculation method is referred to as the General method. Other methods that use the measured times and temperatures within predictive models are also acceptable and used widely. Details of how each method works are given later in this chapter. For a process value to be calculated, it will need to be based on measurements of temperature distribution (TD) in the retort and HP in the product. TD is almost impossible to measure without temperature sensors; however, HP usually uses temperature sensors but can also use log reduction methods. The next sections provide outline information on temperature measurement systems and log reduction methods, with references for further information should this be required.
3.7.1 Temperature measurement systems for TD and HP testing Modern data loggers are typically multichannel systems with digital outputs, allowing data to be recorded directly to a laptop computer for display and to maintain permanent records. Thermocouples based on type T (copper/constantan) with polytetrafluoroethylene insulation are most common because they are inexpensive, accurate over the desired temperature range, and respond rapidly to changing temperature. Other types of temperature measurement are based on a change in electrical resistance with temperature, such as a thermistor, and platinum resistance thermometers (e.g. PT100). These are most commonly used in data loggers where the logging unit is a self-contained unit (e.g. Ball Datatrace, TMI Orion, or Ellab Tracksense). These are also referred to as resistance temperature detectors. Calibration of a temperature sensor against a traceable instrument is essential; otherwise, the user does not have confidence in the numbers obtained. This is usually done each time the instrument is used in a set of TD or HP trials. The method recommended
52
A Complete Course in Canning and Related Processes
by IFTPS (1995) and Bee and Park (1978) is to use the master temperature indicator (MTI) on the retort, which itself must be calibrated at no less than six-monthly intervals. It is worth noting that the new versions of Ellab Tracksense systems are claimed to hold the factory calibration for 12 months and are supplied with a calibration certificate. The use of time and temperature measurements to validate thermal processes is likely to remain the most widely used method. Advances in microchip technology are providing more computing power to analyze these results and increase the accuracy in defining the calculated process values. Historic canning processes were evaluated using lethal rates at time steps of 1 min because of the capabilities of data recorders at that time, but modern process values can be estimated from temperature measurements taken at much reduced frequencies (e.g. every second). The increased data storage capabilities also allow for more temperature probes to be used in each TD or HP test, with multiple loggers linked together in the software. With such systems, it is possible to exceed the number of suggested working probes to define an HP test (3 × 3 or 2 × 10). Several calculation methods are available for setting the scheduled process conditions. The most commonly used methods will be described in outline here, with the references provided for further information.
3.8 Process calculation methods 3.8.1 General method The reference calculation method is known as the General method (Bigelow et al., 1920), and it provides a process value (F-value) that is sometimes sufficient for validation purposes. It converts measured times and temperatures to F-values from accumulated lethal rates. A lethal rate is a relative term that compares the micro-organism killing effect at a measured temperature to 1 min at a reference temperature. Equation (3.1) presents the lethal rate equation as part of the integrated F-value calculation, but the lethal rate (L) itself is as given in Eqn (3.2).
L = 10
T − Tref z
(3.2)
Integration of lethal rates over the measured times and temperatures is usually done using the trapezoidal method, with the calculation routines embedded within a data logger software. The General method allows the user to follow the F-value calculation during the process, usually in real time (with probes); however, for loggers, this is done when they are removed from the processing system. Time to achieve a target F-value is determined from the data and the process conditions can be established. While this is a simple method, it does have some major limitations: • F-values are relevant to the TD and HP test conditions on the day of testing and cannot be easily transferred to another set of conditions. • Changes in critical variables cannot be evaluated without a new HP test, such as initial product temperature, retort temperature, come-up time, and cooling profile.
Heat penetration determinations and thermal process calculations
53
• Deviations to the established process conditions cannot be evaluated by calculation and need to be experimentally simulated.
3.8.2 Ball method For the above reasons, there are alternatives to the General method that are widely used in the industry. Apart from process calculations based on log reduction and General method calculations, most process calculation methods use heating (fh) and lag (j) factors. Heating factors originated in the canning industry (Ball, 1923, 1927) as a measure of the product heating rate used for calculating process times for canned foods. By definition, the heating factor is the time taken for the difference between environment and product temperature to reduce by 90%. They are important terms in thermal process data analysis because they provide information on the rate of heating for a container of food. Figure 3.1 shows an example of a logarithmic heating curve used to calculate fh and j factors. The Ball method is widely used within the United States and countries exporting thermally processed foods to the United States. Its use in Europe is less widespread. However, the Ball method offers options to the General method that enables some analysis of ‘what-if’ scenarios to be calculated. The Ball method has three parts to it and can be described in its most simple form as • An equation to the straight heating line in Figure 3.1, which is effectively the gradient at the end of heating. • This is connected by an experimentally determined complex hyperbolic function to the cooling line. • An equation to the straight line cooling line (not shown in Figure 3.1), which is effectively the gradient at the end of cooling.
±ORJ 7U
± ± ±
&DPSGHQ &KRUOH\ZRRG)RRG5HVHDUFK$VVRFLDWLRQ&7HPS3URJUDP
Figure 3.1 Example of a logarithmic heating curve for canned meat used to calculate fh of 50.4 min and j of 1.45.
54
A Complete Course in Canning and Related Processes
The above description does not do the Ball method justice but does describe the constituent parts. The method was derived for steam retorts because these were the main type used around the time Ball carried out his experiments (Ball, 1923, 1927). Hence, the complex connecting routine between heating and cooling is most accurate for steam retorts; adjustments to this are required for other retort types, such as water retorts. It is important to note that the Ball method calculates an F-value at the end of cooling. Routines can be used to back calculate an end of heating F-value but these begin to lose their accuracy. If an end of heating F-value is required, then it is easier to use only the first equation to the heating part of the logarithmic heating curve and express this as in Eqn (3.3).
RT – T –2.303t/fh (3.3) = j.e RT – I'T'
where j is the heating lag factor, defined as a measure of the thermal lag before the can temperature responds to the changing environment temperature; fh is the heating factor, defined as the temperature response parameter derived from the logarithmic heating curve (minutes); RT is the retort temperature during the hold period (°C); I′T′ is the product pseudo-IT at steam on (°C); and T is the product temperature after (t) minutes heating time (°C). The I′T′ is the temperature at which the straight line for heating (log-linear line) would have started had there not been a lag phase. Most Ball method calculations are now computerised versions of the tables and nomograms published by Ball (Ball & Olsen, 1957). Several attempts have been made since Ball to improve on the method, particularly for cooling lethality calculations.
3.8.3 Numerical methods The use of more advanced mathematical models to evaluate and predict process times and temperatures has increased as the computing power available to a thermal process authority has increased. Examples of numerical calculation software are CTemp (Campden BRI) and NumeriCAL (JBT Corporation). These models deal with the physics of heat transfer and use finite differences to solve Fourier’s partial differential equations. Therefore, they have a more scientific approach that allows variable retort profiles as the boundary conditions. This means that come-up and cool-down profiles can be defined, as well as deviations in processes caused by retort temperature fluctuations (Tucker, Noronha, & Heydon, 1996). These predictive modelling approaches not only help with deciding the fate for batches of product that have undergone a process deviation, but the task of process establishment is made more straightforward. For example, the models can be used to evaluate changes to initial product temperatures, shortened come-up times, or low retort temperatures. They offer considerable flexibility to make the job of process establishment easier.
Heat penetration determinations and thermal process calculations
55
3.9 Some causes of unreliable heat penetration data Any of the process calculation methods, whether based on Ball or numerical methods, require validation against a set of conditions that are representative of the thermal processes. Using a model for prediction outside of these validation boundaries should be done with caution because there are no guarantees that the model will predict with sufficient accuracy. If the purpose is only to provide a rough estimate of process time for a new product, then this is acceptable. Any process times and temperatures will need to be confirmed by HP data so that the model can be used again for fine tuning the conditions. It would be unwise to apply a model for establishing safe process times and temperatures without a basis of a good set of HP data. It makes sense that any process calculation method can only be as accurate as the HP data upon which it is based. This is where the skills of a thermal processing specialist are needed. It is easy to set up experiments to measure HP during a thermal process, but it is much more difficult to do this with a high level of accuracy. There are many ways that the data can be gathered incorrectly. The following is a partial list of some factors that are known to be common causes of unreliable HP data: 1. Thermocouple (temperature) readings are not continued for a sufficient length of time to adequately define heating rate or rates. The greatest accuracy when calculating heating and lag factors is when the heating curve becomes asymptotic, which is when the CCT is within one degree of the retort. 2. HP tests are conducted in a retort load of commercial production and stopped at the end of the scheduled process for quality determinations, rather than continued long enough to obtain sufficient data. 3. Frequency of readings is not sufficient to obtain accurate heating rate or rates. This is critical for fast heating products (fh < 20 min), usually under agitation, in which the rate of change in temperature tends to be uneven as regions of colder and hotter liquids are mixed. 4. Erroneous temperatures are received as a result of inadequate electrical grounding of the data logger. This is a rare occurrence nowadays as more remote loggers are used. It used to happen regularly with water immersion retorts. 5. No note of the retort come-up time is taken, so an estimate must be made from the retort temperature readings. It is useful to know the difference between the retort MTI and the retort temperature sensor used for HP measurements. A difference is common because of the physical size of the MTI, which can cause a thermal lag. 6. The come-up time for HP testing is significantly different from that used in commercial practice. This places a reliance on the process calculation method being able to predict the effect of different come-up times with sufficient accuracy. 7. Multiple thermocouples are used in small cans of product. Heat conduction to the probe tips and heat capacity of the probes are issues that should be considered. 8. No cold spot study is carried out, or insufficient replicates are taken at the cold spot thermocouple location. In an agitation process, it is likely that the cold spot will be at the geometric centre. 9. No times are noted on the temperature recorder, such as the time ‘steam on’, which may not be obvious from the HP data or recorder trace. It is good practice to note as many events as possible. It is better to have more events noted than too few.
56
A Complete Course in Canning and Related Processes
10. N o reference readings are taken from the MTI during the HP test. These are essential at the start of the process while the retort is coming up to temperature. 11. Erratic retort temperature control occurs during the test. 12. Critical factors associated with the product and processing system are not recorded or controlled. It is essential to set up a HP test based on the worst heating conditions that a product might experience. This requires a precise thought process so that each factor can be set at the level that would lead to the product heating slowly. 13. Large temperature disagreement exists between the thermocouple attached to the MTI and the MTI itself. The MTI is always taken as the reference instrument. 14. ITs of the test cans are significantly different from those used in commercial production. This is likely to result from cans filled warm/hot, which then cool before the HP test is ready to start. This will artificially increase the lag factor but not the heating factor, making process calculations less accurate. 15. There is an excessive delay in running the HP test after the containers are sealed. This can create differences in IT but is more critical with products that rehydrate during the process and change their heating rate as a result. Examples are pasta, pulses, and rice products. 16. Product for tests not prepared according to procedures used commercially for raw product preparation or condition. This can affect the heating rates. 17. Differences in processing temperatures between HP tests and commercial practice occur, especially when pilot retorts are used. This is critical for process calculations made with the General method. In theory, heating and lag factors should not be changed, although it is good practice to minimise retort temperature differences.
3.10 HTST: a special type of heat penetration test HTST processing is limited to fluid or pumpable foods. Current thermal processing regulations assume that the entire sterilisation or pasteurisation is achieved in the holding tube and no credit is given to that during heating or cooling. Therefore, the holding time must be sufficiently long to achieve the desired process. The safe design criterion for calculating the length of a holding tube is based on the fluid elements with the shortest residence time, which travels along the centre line of the holding tube. As a safe criterion, the maximum velocity is assumed to be equal to twice the average velocity, as calculated from the volumetric flow rate and holding tube dimensions. This assumption is based on laminar flow in the holding tube. There are two types of flow that will be experienced by a flowing food – laminar and turbulent – but with a transitional region between the two types as the flow changes from one to the other. It is important to know which flow regime is present in all parts of a continuous process so the residence times can be calculated correctly. Laminar flow is assumed to occur up to Reynolds numbers of around 2100, whereas turbulent flow occurs at greater than 10,000 (see Eqn (3.4)). The region in between 2100 and 10,000 is referred to as the transitional region, because the flow regime is changing from laminar to turbulent. This is a region that equipment designers will try to avoid because of the uncertainties in flow behaviour and the key relationship between fastest and mean velocity. Should a heat exchanger be operated under transitional flow,
Heat penetration determinations and thermal process calculations
57
then it would be safe to assume that the fastest liquid along the pipe centre could be travelling twice as fast as the mean velocity. The US FDA takes a Reynolds number of 4000 as the division between laminar and turbulent flow and assumes there is no transitional region. This is a simplification of what happens in practice, but it does make process calculations much clearer. For viscous products, the flow conditions are nearly always laminar. A tomato paste steriliser, for instance, operates at Reynolds numbers (Re) around 1. However, liquid foods that are low in viscosity, such as milk and juice, are more likely to experience turbulent flow, so the assumption is made that the maximum velocity is 1.2 times the mean.
Re =
d h .ρ .v
µ
(3.4)
where dh is hydraulic diameter of the processing system (m); μ is dynamic viscosity of liquid (Pa s); ρ is density of liquid (kg/m3); and v is velocity in the processing system (m/s). Equation (3.4) is for Newtonian liquids that have a viscosity independent of shear rate. Most, if not all, formulated foods contain thickening agents in which the viscosity depends on the shear applied in the processing system. These non-Newtonian foods are certain to flow under laminar conditions because of their high viscosity values, and they also display velocity profiles that are still more complex. As the degree of non-Newtonian behaviour increases, the velocity profile increases in flatness across the pipe cross-section. This means, in practice, that the maximum velocity can be less than its Newtonian value of twice the mean velocity. However, there are few commercial operations that do not apply a factor of 2 when calculating holding tube length, irrespective of the measured flow behaviour index.
3.11 Summary 1. A process is the application of heat to food for a scientifically determined time and temperature, which is known as the scheduled process. 2. A process that has been scientifically determined is specific for the given product, its formulation, methods of preparation, container size, and type of retort system. 3. The determination of a process depends upon reliable heating information for the product and the heat resistance of micro-organisms in the product. 4. The heat resistance of micro-organisms depends upon the micro-organism used, the food in which it is heated, and the food in which the organisms grow. 5. The determination of heating data (time/temperature) should be conducted on a product that simulates commercial preparation or in the commercial process itself. 6. The data obtained from thermal resistance and HP tests are used to calculate a process target of equivalent time at a reference temperature. 7. It is sometimes desirable to check the calculated process by means of an inoculated test pack, particularly if the process does not lend itself to HP studies.
58
A Complete Course in Canning and Related Processes
References Ball, C. O. (1923). Thermal process time for canned food. Bulletin of the National Research Council, Washington, DC, 7 Part 1, Number 37. Ball, C. O. (1927). Theory and practice in processing. The Canner, 64(5), 27. Ball, C. O., & Olsen, F. C. W. (1957). Sterilization in food technology. Theory, practice and calculation. New York: McGraw-Hill Book Co. Bee, G. R., & Park, D. K. (1978). Heat-penetration measurement for thermal-process design. Food Technology, 32(6), 56–58. Bigelow, W. D., Bohart, G. S., Richardson, A. L., & Ball, C. O. (1920). Heat penetration in processing canned foods. National Canners Association Bulletin, 16-L, 128 Washington, DC, USA. Campden BRI. (1977). Guidelines to the establishment of scheduled heat processes for low-acid foods. Technical Manual No. 3. Chipping Campden, Glos., GL55 6LD: Campden BRI. Campden BRI. (2006). Pasteurisation: A food industry practical guide (2nd ed.). Chipping Campden, Glos., GL55 6LD: Guideline No. 51. Campden BRI. Department of Health. (1994). Guidelines for the safe production of heat preserved foods. London: HMSO (out of print). IFTPS. (1995). Protocol for carrying out heat penetration studies. PO Box 2764, Fairfax, VA, USA: IFTPS. NFPA. (1985). Guidelines for thermal process development for foods packaged in flexible containers. Washington, DC. Raatjies, G. J. M., & Smelt, J. P. (1979). Clostridium botulinum can grow and form toxin at pH value lower than 4.6. Nature, 281, 398–399. Tucker, G. S., Noronha, J. F., & Heydon, C. J. (1996). Experimental validation of mathematical procedures for the valuation of thermal processes and process deviations during the sterilization of canned foods. Transactions of the Institution of Chemical Engineers, Food & Bioproducts Processing, 74, 140–148 Part C.