Pasteurization (named after Louis Pasteur) is a heat-treatment process that kills yeast and at least some bacteria (e.g. Lactobacillus and Acetobacter) in a (usually liquid) food product. The goal of pasteurization is to kill enough of the microbes so that the product is stable for several months to a year or so. This contrasts with sterilization, where the goal is to completely kill all microbes. When pasterurizing, a liquid is heated to a specific temperature for a specific period of time to produce a certain reduction in the microbial population. While both pasteurization and sterilization are heat treatments to kill microbes, the lower temperatures of pasteurization are better at preserving flavors and nutrients.
The amount of pasteurization that is applied to a product is measured using Pasteurization Units (PUs). Pasteurization Units indicate how much thermal "force" is applied to a product to reduce the microbial population, but they don't tell us by how much the population is reduced. Some microbes are stronger than others and can resist more of this force. In order to compute the reduction in the population of microbes, one needs to know both the level of pasteurization (in PUs) and the decimal reduction time, called the D value, that is associated with those microbes in this liquid environment, as explained below. If you know the required number of PUs for a product, you can determine the time and temperature needed to reach that level of pasteurization, even if you don't know the D value.
In this tutorial, the mathematical concept of power (xy) is important, in particular the power of 10 (10y). The exponent in a power equation specifies how many times the base (10) is multiplied by itself; 102 is the same as 10×10 and the exponent is 2. When the exponent is 4, then 104 is the same as 10×10×10×10 or 10,000. Exponents don't need to be integers; for example 102.5 is 316.23. The other math used here is limited to addition, subtraction, multiplication, division, and summation. Summation is indicated by the ∑ symbol. Capitalization is important in the variable names: T (representing temperature) is different from t (representing time).
The D value is a numerical description of how heat-tolerant the microbes are; less heat-tolerant microbes will be killed off more quickly than more heat-tolerant ones. The D value is defined as the time it takes, in minutes, to reduce (or divide) a microbial population by a factor of 10 at a specific temperature. This value depends on the type of liquid being pasteurized and which organism is in the liquid. It is common to indicate the target temperature (in Celsius) for a D value using a subscript, such as D60 = 4.0.
The D value is determined experimentally by adding one microbe species (e.g. Lactobacillus brevis) to a target product (e.g. beer) and measuring the population change at a target temperature. If D has been determined to be 4.0 at 140°F (60°C) for one microbe species in a product, then (by definition) it takes 4 minutes to reduce the population of that microbe by a factor of 10 at that temperature. Another 4 minutes will reduce the remaining population by another factor of 10.
For example, if there are 100,000 live microbes floating around in your liquid, then holding that liquid at 140°F (60°C) for 4 minutes (when D60=4.0 for those microbes) will result in 10,000 live microbes. Holding for another 4 minutes will result in another factor of 10 and only 1,000 live microbes, or a total reduction by a factor of 100. Mathematically, we can express this relationship by saying
This is where things get more interesting. We can revise these equations in such a way that we can change the actual temperature that we hold the liquid at, and still determine the amount of pasteurization (the number of PUs) at this new temperature using information based on the original reference temperature. The revised definitions are:
It may be helpful to understand LT through some examples. First, let's look at the case where T equals Tref. In this case, the difference is zero, and so (T − Tref)/z is also zero. Then, because 100 equals 1, when the actual temperature equals the reference temperature, LT equals 1, and PU equals t. This shows that the extended formula is still consistent with the definition that 1 PU is equivalent to holding the liquid at the reference temperature for one minute. Next, let's consider the case where T is larger than the reference temperature by z degrees. Then, T − Tref = z and so (T − Tref)/z equals one. Then LT equals 101 or 10, and (t × 10) / D is the same as t / (D / 10), and so we have increased the temperature such that the D value (the time it takes to reduce the population by a factor of 10) has itself decreased by a factor of 10. This is also equivalent to reducing the time needed for a certain level of pasteurization by a factor of 10. If we get 24 PUs when t = 24.0 and LT = 1.0, then we also get 24 PUs when t = 2.4 and LT = 10.0. If we know the value of z and the reference temperature, we can easily determine any combination of time and temperature that will yield a certain level of pasteurization.
Looking at a few more examples may be helpful. Using a z value of 6.94, a reference temperature for D of 140°F (60°C), and holding a cider at 158°F (70°C) for one minute will produce 27.6 PUs, because 10((70-60)/6.94) = 27.6. Holding this cider at 131°F (55°C) for 60 minutes will produce 11.4 PUs (60 × 10((55-60)/6.94) = 11.4). If we want 50 PUs, we can get that using one minute of pasteurization at 161.2 °F (71.8°C) or four minutes at 153.7°F (67.6°C). (In this case, T = (log(PUtarget / t) × z) + Tref). If we want 50 PUs with a target temperature of 149°F (65.0°C), then we'll need 9.5 minutes (t = PUtarget / 10(T-Tref)/z).
The z value, like the D value, depends on the type of liquid being pasteurized and which organism is being targeted. The z value is determined experimentally by adding one microbe species to the product, measuring the population change at several target temperatures to determine D at several temperatures, and finding the increase in temperature that reduces D by a factor of 10.
So far, we have considered holding a product at a single temperature for some period of time t. The same concepts can be applied to holding the product at different temperatures over a period of time, for example as the product heats up and cools down. Above we defined PU as t × LT which is the same as t × 10(T−Tref)/z. We can break the time period t down into one-minute increments, like this: t = 1 + 1 + ... + 1, where this addition happens t times, and represent that with summation: t = ∑t 1. (The superscript t above the ∑ indicates that the summation happens t times, and summing 1 t times yields the value t.) Then we can consider the temperature at each of these individual times, Tm, where m is a one-minute time interval within the total time t. In this case, the definition of LT undergoes a very minor change, where T becomes Tm to indicate that the temperature is now for a specific time (minute) m:
Again, some examples may help. In the simplest case, Tm is the same as Tref at all times, and according to the new formula, PU still equals t. In a more complicated example, let's say that we apply heat treatment to the product for 5 minutes. Let's also say that Tref = 60, D = 5.6 minutes, and z = 6.94. During the first minute, the heat is still increasing and T1 = 57°C. During the second and third minutes, the heat is held at 65°C, so T2 and T3 = 65°C. At the start of the fourth minute, the heat is turned off and the product starts to cool down. Then, T4 = 61°C and T5 = 59°C. With this information, we know that LT1 = 10−3.0/6.94 = 0.37, LT2 and LT3 = 105/6.94 = 5.25, LT4 = 101/6.94 = 1.39, and LT5 = 10−1/6.94 = 0.72. The total amount of pasteurization is the sum of all of these LT values, or 0.37 + 5.25 + 5.25 + 1.39 + 0.72 = 12.98 PUs. (A real-life calculation of PUs is much more easily done using a pasteurization units calculator.) With a D value of 5.6, we have reduced the microbial population by a factor of 1012.98/5.6 = 102.32 or 208. If there were originally 100,000 microbes, after this heat treatment there are only 480 microbes. This reduction by a factor of 208 or 2.32D is still far short of the typical goal of a 6D reduction (with 33.6 PUs for a factor of 1,000,000).
Increasing the temperature above the reference temperature will rapidly increase the destruction of microbes during each minute of heat treatment. Similarly, decreasing the temperature will cause fewer microbes to be killed each minute. How low can we make this temperature until the microbes are "uncomfortably warm" but not being killed by the heat treatment? In other words, what is the lower temperature limit for pasteurization? A minimum pasteurization temperature of 140°F (60°C) is sometimes recommended because it is around this temperature that proteins undergo irreversible denaturation. (This is the same temperature as the reference temperature in this tutorial's examples.) But is this the lowest temperature at which PU values accumulate?
Grzegorz (Greg) Rachon at Campden BRI (Rachon, 2019) used a peak temperature of 129°F (54°C) to achieve more than a 6D reduction in the microbial population by heat treatment with 1.6 PUs. The target microorganism was Lactobacillus brevis in beer, with a reference temperature of 140°F (60°C), a D value of 0.2 minutes, and a z value of 9.48°C. The peak temperature was held for one minute, producing 0.23 PUs, and so the other 1.37 PUs were achieved during the 45-minute heating-up and 45-minute cooling-down periods, with a minimum temperature of 95°F (35°C). In this case, the observed reduction in microbial population (as part of a validation step) was actually larger than the computed reduction, indicating that temperatures between 95°F (35°C) and 129°F (54°C) are sufficient to greatly reduce the microbial population, even for heat-resistant microorganisms.
Splittstoesser et. al (1975) looked at 29 yeasts and 5 lactic acid bacteria (LAB) in wine (12% ABV). Most of the yeasts had D values less than 0.6 minutes at 120°F (49°C), with the largest D value of 1.6 minutes at 120°F (49°C). Four of the five LAB were less heat-tolerant than the yeasts, with D values less than or equal to 0.35 minutes at 113°F (45°C). The bacterium Lactobacillus fructivorans behaved differently, however, with a D value of 1.7 minutes at 140°F (60°C) and a z value of 12°F (6.67°C). At 136°F (58°C), the D value was about 3.2 minutes. (If the changes in D values were linear, the two values would imply a z value of 20°C. Apparently, at low temperatures the D values do not change linearly for this microbe.) (The authors also found an inverse relationship between D values and ethanol concentration; higher-alcohol products result in smaller D values.) Wikipedia states that L. fructivorans can grow in temperatures as high as 53°C and is very tolerant of both heat and ethanol. By way of analogy, L. fructivorans is like a human who is most comfortable when relaxing in a hot tub with a bottle of whiskey in hand.
In short, the minimum temperature for accumulating PUs depends, like the D and z values, on the specific microbe and its environment. A temperature limit of 140°F (60°C) seems too high for the vast majority of microbes; lower temperatures can often still cause a reduction in micorobial population by a factor of 10 in just a few minutes. However, for the general case, a lower limit of 131°F (55°C) seems to be a conservative value that should be effective even for microbes such as L. fructivorans.
When computing PUs, we want to use the D and z values (and the minimum temperature) for the most heat-tolerant microbes that may be found in the product. Microbes that are not as heat tolerant will usually end up with lower population levels and will have less impact on the product shelf life. The D and z values are dependent not only on the type of microbe, but also the environment that this microbe is in (the product). The initial population count commonly found in a product will also depend on many factors. For example, raw unfiltered apple juice will have a different initial population count than apple juice that has been treated with potassium metabisulfite. The shelf life will also depend on various factors, such as the product's pH, the ethanol concentration, and if sugar has been added to the product.
The recommended number of pasteurization units therefore depends on the product and the specifics of how it was produced. For example, it has been recommended that pilsner and lager beer should have 15 to 25 PUs, ales and stouts should have 20 to 35 PUs, lemonade should have 300 to 500 PUs, and fruit juices should have 3000 to 5000 PUs. These recommended target PUs make assumptions about how the product has been made and treated, and what microbes are in the product. The target of 5000 PUs mentioned for fruit juices implies the presence of very heat-tolerant microbes with a very large D value and/or a 6D reduction being insufficient due to a very high initial concentration of microbes.
If we know the D value and accept a 6D reduction as a reasonable goal, then the target PU level is easily computed as PUtarget = 6 × D. The difficulty is in selecting the appropriate D value. For beer, a D value of 5.6 minutes is associated with a commonly-used z value of 6.94 that was determined from a study by Del Vecchio, Dayharsh, and Baselt back in 1951. Therefore, for beer we would want to target 5.6 × 6 or 33.6 PUs. More recent studies (e.g. Rachon, 2019) have demonstrated that this D value is probably much larger than needed for even heat-tolerant microbes in beer, and that as little as 1.2 PUs (with D=0.2 minutes) is sufficient for a 6D reduction in a beer's microbial population. To further muddy the waters, a target of more than 10.2 PUs may be required for a 6D reduction in L. fructivorans, with the exact target depending on the product's pH and ethanol concentration. There are sometimes no official recommendations on a target PU level for a specific product; lacking such recommendations you need to understand and weigh the risks and benefits yourself.
Botulism is an illness caused by ingestion of the toxin that is produced when the bacterium Clostridium botulinum grows from a spore to a full-fledged bacterium. Clostridium botulinum is very hardy and commonly found in the soil and on many food surfaces. A temperature of 176°F (80°C) for 30 minutes is required to destroy the toxin, and a temperature of 250°F (121°C) for 3 minutes is required to kill the spores. Pasteurization is therefore ineffective at preventing botulism spores from producing the toxin. However, Clostridium botulinum doesn't grow, and the toxin is not formed, when the pH is below 4.6. Therefore, pasteurization is considered sufficient heat treatment for anything with a pH less than 4.6. For products with higher pH levels, pressure cooking (which can reach temperatures above 212°F (100°C)) is required.
Estimating the final microbial population count that results from pasteurization requires knowing the initial microbial population, the level of pasteurization (in PU), and the D value. We usually don't know the initial population, but the goal is usually to reduce this population by a factor of 1,000,000 or 6D. Sometimes it's unclear what D value to use, as this value depends on not only the characteristics of the product (pH, ethanol concentration, sugar concentration) but also the microbes that may be found in the product.
If the target PU level is known, then pasteurization can sucessfully reach this level knowing only the reference temperature and the z value that is associated with this product and its potential microbes; the initial population count and the D value are not needed. A commonly-used reference temperature is 140°F (60°C) and z values often range from about 5 to about 10, although a z value of 15.8 has been estimated for alcohol-free beer.
The material above was based on various sources, especially:
Copyright © 2025 John-Paul Hosom, all rights reserved. Please feel free to contact me with comments, questions, suggestions, and/or bug reports.