Pasteurization is a log reduction, not a temperature
Heat does not sterilise food at some threshold temperature. It kills a fixed proportion of the microbial population per unit time, and that proportion depends on temperature. This makes pathogen death a first-order process, exactly like radioactive decay, and it means there is no temperature above which food is instantly safe and below which it is instantly dangerous.
The unit that expresses this is the D-value: the time at a given temperature to reduce the surviving population by 90%, one decimal place, one log. Hold for one D and 90% of the organisms are gone. Hold for two D and 99% are gone. Hold for 6.5 D — the common target for poultry — and you have removed 99.99997% of them, which turns a contamination of ten million cells into three.
The second unit is the z-value: the temperature rise that cuts the D-value by a factor of ten. For Salmonella in poultry z is around 6.5 °C, so 55 °C is roughly ten times slower than 61.5 °C and a hundred times slower than 68 °C. That exponential sensitivity is why a table of times looks so extreme: about 15 minutes at 55 °C, under three minutes at 60 °C, under half a minute at 65 °C for the same reduction.
This is the entire justification for sous vide chicken at 60 °C. The familiar 74 °C (165 °F) instruction is not a threshold either — it is a temperature at which the required hold is so short, well under a second, that no cook could fail to achieve it. Holding at 60 °C for a few minutes reaches the same log reduction with far less protein damage. The trade is temperature for time, and the arithmetic on this page is how you make that trade honestly.
The model, and where the numbers come from
The core equation is DT = Dref × 10^((Tref − T) / z). Put T at the reference temperature and the exponent is zero, so DT is Dref — the definition. Drop T by exactly z and the exponent is 1, so D is ten times larger. The required hold is then simply the target log reduction multiplied by D at your temperature.
The D and z values are the weak point, and you should treat them that way. Heat resistance varies with strain, fat content, salt, sugar, pH and water activity, and published figures for the same organism in similar foods differ by factors of two or three. The presets here are typical literature values: Salmonella in poultry near D₆₀ = 0.40 min with z ≈ 6.5 °C, higher in fattier red meat, and Listeria monocytogenes distinctly more resistant at D₆₀ ≈ 2 min with z ≈ 7.5 °C. Fat is protective, which is why the red-meat preset carries a larger D. If you are writing a HACCP plan rather than cooking dinner, you need heat-resistance data for your own product and a process authority to review it. That is what the custom option is for.
The come-up credit is the part tables omit. Food does not sit at 5 °C and then jump to 60 °C. It passes through 50, 55 and 58 °C on the way, and lethality accumulates the whole time. The calculator estimates the come-up from thickness and shape using the same conduction model as the sous vide timing calculator, tracks the coldest point through the heating curve, and integrates the instantaneous kill rate along it. Because it tracks the coldest point, the credit is conservative: every other part of the food is hotter and accumulates more.
Why 54.4 °C (130 °F) is a hard floor. Below it the D-values stop being reliably measurable, some strains show injury-and-recovery behaviour rather than clean first-order death, and several pathogens can still grow. The equation will happily return a number at 48 °C; that number is not a process.
Worked example: a 25 mm chicken breast at 60 °C
A boneless chicken breast 25 mm at its thickest, from a 5 °C fridge, into a 60 °C bath. You want a 6.5-log Salmonella reduction.
- D-value at 60 °C. The reference temperature is 60 °C, so the exponent is (60 − 60) ÷ 6.5 = 0 and D = 0.40 × 10⁰ = 0.40 minutes.
- Hold with no credit. 6.5 × 0.40 = 2.6 minutes at the core, once the core is at 60 °C.
- Come-up time. A 25 mm slab has a Biot number of 2.47, an eigenvalue of 1.1395 and a coefficient of 1.1959. With the gap now 55 °C, θ = 0.5 ÷ 55 = 0.00909, so Fo = ln(131.5) ÷ 1.2984 = 3.758 and the come-up is 3.758 × 1116 s = about 70 minutes.
- Credit for the heat-up. Over those 70 minutes the coldest point climbs through the lethal range, and the integral of the kill rate along that curve comes to several log reductions on its own. The calculator reports the figure; whenever it exceeds 6.5, no extra hold is needed at all.
- Total. Come-up plus whatever hold remains. For this breast it is essentially the come-up time: by the time the centre reaches 60 °C it has already spent long enough above 55 °C to be pasteurized.
Compare with 55 °C, where the same breast needs D = 0.40 × 10^(5 ÷ 6.5) = 0.40 × 5.878 = 2.35 minutes per log, so 6.5 logs is 15.3 minutes of hold. Five degrees of bath temperature multiply the required hold nearly six-fold, and that is the single most important intuition on this page.
How to use the three time figures
The hold time is the conservative answer. It assumes zero lethality before the core arrives, which is never true but is always safe. If you want a number you can defend without arguing about heating curves, use this one and start counting when your probe reads the target temperature.
The total time is the realistic answer. It credits the kill accumulated during heat-up at the coldest point. For thick pieces at moderate temperatures the credit is large enough that no additional hold is required, which is why published sous vide tables for thick cuts often look like pure heating tables. Because the credit can only add lethality, the total is always less than or equal to come-up plus full hold — it can never exceed it.
The D-value tells you how much margin you have. If D is 0.4 minutes, an extra five minutes in the bath is twelve more log reductions and the process is effectively bulletproof. If D is 13 minutes, as it is at 50 °C, then five extra minutes buys you a third of a log and the process is fragile to any error in temperature. A process whose D-value is large relative to your control precision is a process you should not be running.
Choose the log target deliberately. Poultry processes commonly target 6.5 or 7 log for Salmonella. Intact whole-muscle beef is often treated at 5 log or less, on the reasoning that the interior of an unbroken muscle was never contaminated and the surface receives far more heat than the core. That reasoning collapses the moment the meat has been ground, needled, injected, mechanically tenderised or rolled, all of which carry surface organisms inward — and those products should be treated like the poultry case.
Hold time for a 6.5-log Salmonella reduction in poultry
| Core temperature | D-value | Hold for 6.5 log |
|---|---|---|
| 54.4 °C (130 °F) | 2.908 min | 18.9 min |
| 55 °C (131 °F) | 2.351 min | 15.3 min |
| 56 °C (132.8 °F) | 1.650 min | 10.7 min |
| 57 °C (134.6 °F) | 1.158 min | 7.5 min |
| 58 °C (136.4 °F) | 0.812 min | 5.3 min |
| 60 °C (140 °F) | 0.400 min | 2.6 min |
| 62 °C (143.6 °F) | 0.197 min | 1.3 min |
| 63 °C (145.4 °F) | 0.138 min | 0.90 min |
| 65 °C (149 °F) | 0.068 min | 0.44 min |
| 70 °C (158 °F) | 0.012 min | 4.5 s |
Read down the column: every 6.5 °C divides the time by ten. That is the definition of the z-value, and it is why the choice between z = 6 and z = 7.5 changes an answer more than most people expect.
This is a teaching tool, not a validated process
The D and z values here are typical published figures for broad food categories. Real heat resistance depends on the strain, the fat and salt content, the water activity, the pH and the recovery medium used in the original study, and values in the literature for the same organism differ by factors of two or three. A commercial process must be built on heat-resistance data for your specific product, validated by measurement in your specific equipment, and reviewed by a process authority. Nothing on this page substitutes for USDA FSIS lethality guidance, your regulator's requirements, or a HACCP plan. If you are cooking at home, the practical version of the same advice is simpler: use a calibrated probe, measure the coldest point, and give yourself margin.
Mistakes that make a pasteurization calculation meaningless
- Measuring the bath instead of the food. The bath is at temperature from the start; the coldest point in the food may be twenty degrees behind for an hour. Lethality happens at the coldest point.
- Starting the clock when the food goes in. Unless you are using the heat-up credit deliberately, the hold begins when the core reaches temperature, not when the bag hits the water.
- Applying whole-muscle logic to ground, rolled or injected meat. Any process that carries surface organisms into the interior removes the argument for a reduced log target.
- Using a D-value from the wrong matrix. Fat and low water activity protect bacteria substantially. A D-value measured in broth will understate the time needed in a fatty sausage.
- Forgetting what happens after the cook. Pasteurized is not sterile. Spores survive, and a pasteurized bag chilled slowly or held warm can support the growth of what remains. Chill rapidly or serve immediately.
- Treating the FDA's four-hour rule as optional. Below 54.4 °C, total time out of temperature control is the controlling limit and no log-reduction arithmetic overrides it.
Where this fits with the rest of the cook
Pasteurization arithmetic answers only one question: is it safe? It says nothing about whether the food is good. Texture is set by protein denaturation and collagen conversion, which follow their own time-and-temperature curves, and the reason 60 °C chicken is worth the trouble is that myosin has set while much less water has been squeezed out than at 74 °C. Use this calculator to establish the safety floor and the timing calculator to work out when the core actually arrives.
The same reasoning governs conventional cooking, just invisibly. USDA FSIS publishes time-temperature tables — the Appendix A lethality framework — that let a processor hold beef at 130 °F for a specified time instead of reaching 145 °F instantaneously, and those tables are built from exactly this D and z model. It is also why the guidance for a rib roast cooked to 130 °F depends on the cut being intact, and why a brisket in a low pit comes with a caution about getting the surface past 140 °F within four hours.
It also explains the thawing rules. A turkey thawed on a counter spends hours with its outer layers in the growth range, and growth is the mirror image of death: it too is exponential in time and steeply dependent on temperature. Every safe-handling rule you have been given is a statement about one of those two exponentials.
Finally, remember what pasteurization does not do. It reduces vegetative pathogens by the number of logs you specify. It does not eliminate spores, it does not destroy heat-stable toxins already formed, and it does not make food shelf-stable. Cook-chill operations pasteurize and then chill fast for exactly that reason, and any bag you are not eating immediately should go into iced water rather than a warm resting spot.