Where should F₀ be measured in retort processing?

Educational engineering overview only. It does not replace a qualified process authority, applicable regulatory requirements, or the validated scheduled process for a specific product, package and retort system.

Ask an experienced thermal-process or regulatory engineer where F₀ should be measured, and the instinctive answer is usually: “At the cold spot, of course”.

Correct. But deceptively incomplete.

Which cold spot? The cold region in the retort load? The slowest-heating container? Or the slowest-heating location inside that container? And how was that location established?

That distinction is not semantic hair-splitting. It separates two fundamentally different studies: heat distribution, which asks whether the retort delivers heat adequately throughout the load, and heat penetration, which asks how the food inside an individual package responds to that delivered heat. Codex, FDA-related guidance, Canadian Food Inspection Agency (CFIA) guidance and the Institute for Thermal Processing Specialists (IFTPS) all treat these as related but distinct parts of thermal-process establishment [1,2,3,4].

From cook charts to measured lethality

Commercial thermal processing began with Nicolas Appert’s early nineteenth-century preservation method. By the early twentieth century, Bigelow, Esty, and Ball shifted process design from empirical cook times to microbial thermal-resistance data, measured heat penetration, and mathematical process calculations. Modern retorting still rests on that same combination of microbial resistance, product heating behaviour, and controlled heat delivery by the retort [5,6].

The historical progression matters because retort temperature alone does not describe what the microorganism experiences. The retort may be at 121°C while the critical part of the product is still substantially cooler. Conversely, the product may continue accumulating lethality during early cooling. The engineering question therefore moved from “How hot was the retort?” to “What temperature–time history occurred at the location receiving the least adequate thermal treatment?” [7,5].

First, what does F₀ actually represent?

For the conventional Clostridium botulinum reference convention, F₀ expresses a variable-temperature exposure as the equivalent number of minutes at 121.1°C, using a z-value of 10°C:

Here, T(t) is the measured product temperature as a function of time. FDA terminology similarly treats an unsubscripted F-value as referenced to 250°F with a z-value of 18°F. The selected reference temperature, z-value, integration interval and target organism must therefore be stated rather than left implicit [7,5].

F₀ is an integrated lethality metric, not a sensor reading and not simply the programmed hold time. If you use full temperature-history integration, every recorded temperature contributes according to its lethal rate, including relevant portions of heating, holding, and cooling. You can also calculate a scheduled process using recognised formula methods rather than relying solely on direct numerical integration, so regulatory compliance should not be reduced to “Did the logger display the target F₀?” [7,8].

Most importantly, F₀ does not locate the cold spot. A perfectly calculated F₀ at the wrong physical location is still the wrong answer.

“ Cold spot” has more than one meaning.

The Institute for Thermal Processing Specialists explicitly distinguishes three related uses of the term:

Scale or study	Precise meaning	Recommended engineering wording
Retort temperature distribution	The retort area that is last to reach the required processing temperature during come-up	Slowest-to-come-up region in the retort/load
Heat-transfer distribution	The load position where transfer of heat into packages is slowest	Least-favoured or slowest-heating load position
Product heat penetration	The experimentally determined slowest-heating location inside a container; for non-homogeneous foods this may include the slowest-heating particle	Slowest-heating product location within the container

These definitions explain why saying only “cold spot” can be ambiguous. A thermocouple hanging in the retort’s coldest part measures the heating medium or retort environment. It does not measure the temperature history inside the food’s slowest-heating part [4,9].

For a technically literate audience, I would use “retort/load cold or slow zone” for the equipment-scale problem and “container slowest-heating location” for the product-scale problem. You can introduce “cold spot” in brackets after each definition, but do not use it unqualified.

The relationship is illustrated below. Heat distribution narrows the search at retort scale; heat penetration then establishes the relevant measurement location at product scale [4,3].

Figure 1: Where F₀ is evaluated is the end of a two-scale evidence chain: retort/load distribution identifies the least-favoured load region, then container heat-penetration work identifies the slowest-heating product location for the specific product–package–process combination.

Heat distribution asks: does the retort deliver heat adequately?

A temperature-distribution study places multiple calibrated temperature-measuring devices throughout the retort and load. It aims to establish venting, come-up requirements, temperature stability and uniformity, and the positions that take the longest to reach process temperature. The sensors measure the retort heating environment around the packages, not the internal product temperature [4,7].

Sensor placement covers the three-dimensional load, including suspected difficult regions, areas near the reference and control sensors, and multiple positions in each basket or rack. If an initially broad survey suggests one region is consistently slow, further runs may concentrate sensors there. The objective is evidence, not confirmation of an assumed location [9,4]

A useful nuance is that uniform retort temperature does not always guarantee uniform package heat transfer. This is particularly relevant in steam-air, water-spray, cascade, immersion and other overpressure systems where flow, fan performance, nozzles, racks, load density or entrained non-condensable gases may affect local surface heat transfer. IFTPS therefore identifies heat-transfer distribution as a separate study where appropriate. It compares standardised packages or heat-input units throughout the production retort to identify load positions with the slowest heat uptake [4,6].

This gives the sequence:

1. Temperature distribution: Can the retort achieve the necessary temperature uniformity and operating profile?
2. Heat-transfer distribution, where needed: Does the heating medium transfer heat into packages uniformly across the load?
3. Heat penetration: How does the defined product/package combination heat and cool internally under adverse production conditions?
4. Process calculation: What scheduled process delivers the required lethality at the limiting product location?

CFIA states this relationship clearly: the retort’s temperature-distribution test identifies the slowest-heating retort region, which heat-penetration studies then use to establish the scheduled process [3,9].

Heat penetration asks: how does the product respond?

Heat-penetration studies measure temperature inside representative product containers. Codex requires the product temperature at the slowest-heating point in the container contents to be monitored under the most adverse conditions likely in production. It also requires enough studies to characterise variability and incorporates the resulting process and critical factors into the scheduled process [1,8].

The slowest-heating location cannot safely be inferred from package geometry alone because internal heat transfer depends on the entire product–package–process combination:

• Conduction-heated products: Highly viscous or solid products commonly heat slowly near the geometric centre.
• Natural-convection products: Mobile liquids often have a slow region below the geometric centre, but its exact position can vary.
• Mixed or broken heating: Products containing starch, protein networks, particles, or changing rheology may start with convection and transition to conduction.
• Particulate foods: The limiting measurement may be inside a particle located in a slow-heating region of the container.
• Agitated processes: Rotation, oscillation, or reciprocation change fluid motion, particle trajectories, headspace-bubble behaviour, and heat transfer.
• Package effects: Shape, thickness, orientation, fill mass, headspace, residual gas, nesting and rack restraint can alter both heating rate and the limiting location.

USDA training material presents the geometric centre and lower axial region as useful starting hypotheses for defined conduction and convection cases, respectively, but also states that the location depends on product, container, thermal-processing system and heat-transfer mechanism. IFTPS guidance further notes that rheology changes, particulates, residual gas, orientation and agitation can move or broaden the cold region [7,4].

This deserves particular attention for high-protein or starch-containing beverages. A product may thicken, aggregate or otherwise change flow behaviour during heating. The sensor location appropriate to the initial liquid cannot automatically be assumed to remain the location receiving minimum cumulative lethality throughout the cycle [4,6].

So where should the thermocouple actually go?

During exploratory cold-spot location work

Begin with a scientifically justified set of candidate positions informed by product rheology, heating mode, package geometry, orientation and agitation. Use multiple instrumented containers, or cautiously use multiple sensors in a container where their presence will not materially alter product movement. Comparing the temperature histories and calculated lethality identifies the limiting location experimentally [4,7].

The experiment should also address run-to-run and container-to-container variability. IFTPS recommends replication and notes that non-homogeneous or complex-heating products commonly show variation. Where results remain uncertain, the answer is more discriminating testing, not averaging the uncertainty away [4,1].

During validation or process-establishment work

Once the slowest-heating location is established, the sensing junction is fixed there. Instrumented test packages are then placed at the least-favoured load positions identified by retort distribution work. Testing challenges the adverse allowable combination of relevant critical factors, such as minimum initial temperature, maximum fill, viscosity, solids, package thickness, residual gas, orientation, headspace, rotation speed and retort operating conditions [4,8].

The probe installation itself must not create a heat-conduction path, stir the product, move during agitation, distort a pouch or compromise the package seal. Codex specifically warns that flexible and semi-rigid packages may require special fixtures to maintain the sensor at the predetermined location without changing heat-penetration characteristics [4,1].

When the location is variable

A process authority does not normally solve variability by choosing the average sensor result. Conservative approaches include:

• selecting the slowest credible temperature response or least accumulated lethality;
• testing the combinations of critical factors expected to produce worst-case heating;
• increasing replication where product or process variability is high;
• defining more than one critical location where the limiting position can shift;
• adding an appropriate process margin within the scheduled process;
• restricting the validated operating window if the uncertainty cannot be adequately bounded.

US regulations require addressing commercial variation and all relevant critical factors when establishing the scheduled process. Codex similarly requires testing under adverse production conditions and re-evaluation when product specifications change [8,1].

Practical decision logic

1. Define the system.

Specify the product formulation and physical state, container, fill, headspace, orientation, retort type, heating medium, load pattern and agitation condition.

2. Establish retort performance.

Use temperature distribution and heat-transfer distribution where appropriate, to demonstrate acceptable delivery and locate any least-favoured load regions [4,2].

3. Form a product-heating hypothesis.

Use conduction, convection, mixed-heating and agitation knowledge to nominate candidate internal sensor locations. Treat these as hypotheses, not conclusions [7,4].

4. Map the container experimentally.

Run exploratory heat-penetration tests across candidate positions and relevant critical factors.

5. Validate at the limiting combination.

Fix the sensor at the established slowest-heating product location and place test containers in the relevant least-favoured load region.

6. Evaluate F₀ or establish the process using the accepted calculation method.

Use the qualified temperature–time data, correct reference temperature and z-value, and the calculation procedure recognised by the process authority [5,8].

7. Control the scheduled process.

Routine production demonstrates compliance with the scheduled process and its critical factors. It does not rediscover the cold spot on every batch.

Common misconceptions

“The cold spot is always the geometric centre.”

Only a reasonable starting assumption for certain conduction-heated systems.

“The coldest position in the retort is where F₀ is measured.”

That position indicates where instrumented containers should go. F₀ uses the internal product temperature history.

“One thermocouple proves the cold-spot location.”

One thermocouple can characterise a known location. It cannot establish that no slower location exists.

“If the retort reached 121.1°C, the product achieved the same exposure.”

The product lags the heating medium; penetration behaviour must be measured or otherwise established scientifically.

“Achieving a target F₀ alone proves commercial sterility.”

Commercial sterility depends on the scheduled process, control of critical factors, retort delivery, package integrity, records and deviation management, not an isolated logger number [8,3].

Regulatory perspective

Codex: CXC 23-1979 is international guidance rather than automatically enforceable law. It calls for competent experts, temperature-distribution studies, adverse-condition heat-penetration testing and monitoring at the slowest-heating product point [1,10].

United States: 21 CFR Part 113 is legally enforceable for applicable thermally processed low-acid foods in hermetically sealed containers. It requires qualified persons to establish scheduled processes using recognised scientific methods that address critical factors and production variability. The regulation does not prescribe one universal thermocouple coordinate or one universal F₀ target for every food [8,2]

New Zealand: The 2023 MPI Food Notice requires applicable low-acid canned-food operations under a registered food control plan to follow the principles of Codex CXC 23-1979 and US 21 CFR Parts 113 and 114, with specified supervisory and process-development competency requirements [11,1].

Industry documents such as the IFTPS guidelines provide detailed study methodology. Still, they remain voluntary technical guidance unless incorporated into legislation, an approved food-control system or a company’s validated procedures [4,8].

Final takeaway

“Measure F₀ at the cold spot” is correct only after the difficult work has been done. A better engineering statement is:

Evaluate F₀ from the product temperature-time history at the experimentally established slowest-heating location within the relevant container, tested under the limiting product, package, load and retort conditions.

Heat distribution establishes where in the retort load to challenge the process. Heat penetration establishes where inside the product to evaluate the thermal response. F₀ is the mathematical result; it is not the method used to determine either location.

And no blog, guideline or generic sensor-position diagram can replace the qualified process authority responsible for the validated scheduled process.

Selected authoritative and foundational references

1. Codex Alimentarius Commission. Code of Hygienic Practice for Low and Acidified Low-Acid Canned Foods, CXC 23-1979, revised 1989 and 1993, editorial corrections 2011. (1)
2. US FDA. 21 CFR Part 113: Thermally Processed Low-Acid Foods Packaged in Hermetically Sealed Containers. eCFR current to 1 October 2026 when retrieved. (8)
3. US FDA. Guide to Inspections of Low Acid Canned Food, Chapters 10 and 12. (2,12)
4. US FDA. Sterilising Symbols D, Z and F, Inspection Technical Guide No. 7. (5)
5. Institute for Thermal Processing Specialists.Guidelines for Conducting Thermal Processing Studies. Approved March 2014. (4)
6. Canadian Food Inspection Agency.Commercial Sterilisation of Low-Acid Food in Hermetically Sealed Containers: Retort Systems, modified 20 June 2023. (3)
7. Canadian Food Inspection Agency.Temperature Distribution Test for Steam-Still Retorts. (9)
8. USDA FSIS.Principles of Thermal Processing, Thermal Processing for Meat and Poultry Products Training, Module 6. (7)
9. New Zealand Ministry for Primary Industries.Food Notice: Requirements for Food Control Plans and National Programmes, issued 24 April 2023. (11)
10. Jimenez PS, Bangar SP, Suffern M, Whiteside WS. Understanding retort processing: a review. Food Science & Nutrition. 2024;12:1545–1563. doi:10.1002/fsn3.3912. (6)
11. Bigelow WD, Esty JR. The thermal death point in relation to time of typical thermophilic organisms. Journal of Infectious Diseases. 1920;27:602–617. (6)
12. Ball CO. Thermal Process Time for Canned Food. National Research Council Bulletin 37. 1923; and Theory and Practice in Processing. The Canner. 1927;64(5):27. (4,6)