Food process development
Freeze Drying Cycle Development for Food: Pilot to Production
A good-looking sample leaves critical questions unanswered: did the piece dry thoroughly to its center, will that result repeat across the intended loading conditions, and does the evidence justify releasing a production recipe? Build development work around resolving those questions, rather than searching for an arbitrary drying time supposed to fit every batch.

Freeze drying cycle development means defining a food-specific sequence of freezing, primary drying and secondary drying, then testing it against agreed product and loading requirements. The result should be a traceable recipe revision, justified stage transitions, a demonstrated operating window and clear reasons to reassess the cycle when material, loading or equipment changes. One successful sample cannot establish all of these. [1] [3]
What must the finished food achieve?
Define clear acceptance criteria for the finished food before choosing machine settings. A crisp fruit piece, a meal component needing rehydration, and material destined for milling have distinct product-specific requirements. When a development brief merely specifies that food “looks dry,” the team cannot tell whether a change in a trial cycle improved the process or merely produced a different result.
Agree in advance on quality measures, test methods, and post-drying sample handling. Record the full product identity and intended loading right alongside those criteria to maintain a comparable basis across later trials.
- Product identity: formulation, raw-material lot, preparation, solids or initial moisture, piece dimensions or layer depth, and freezing history.
- Required result: the specified moisture test and reporting basis, center dryness where relevant, color, texture, rehydration and other product-specific quality measures.
- Handling and stability: sampling conditions, exposure before sealing, intended packaging and the storage verification needed for the product. Water activity may be specified separately; it is not another name for moisture content.
- Intended operating range: tray type, loading geometry, planned batch loading and the equipment on which the result must be repeated.
The stages do different jobs. Freezing establishes the frozen starting condition; primary drying removes ice by sublimation; secondary drying removes additional non-frozen water by desorption. The food’s composition, structure and freezing history affect the subsequent drying response. Set quality targets from the product specification and testing, not a generic final-moisture percentage. [3]
Evaluate moisture results using the same analytical method and reporting basis across all trial runs. Record sampling positions on the trays and product exposure before sealing; unrecorded differences in handling or test procedures can easily be mistaken for a process change or an apparent improvement. When water activity is specified, track and report it separately from moisture content.
Keep the questions separate. The temperature and pressure guide explains the control variables. The lab and pilot guide covers test equipment. This article addresses how trial evidence supports a development decision.
What should the next trial prove?
Start each trial plan with the specific operational question to answer and the possible explanation being tested. Identify the deliberate change or planned set of changes, keeping other processing conditions comparable across runs. When batches show damp centers, long drying times, or patchy quality, define beforehand what observations would support the tested explanation and what findings would weaken it. That distinction makes every trial an informative comparison, even when a batch fails quality checks.
Choose a documented reference cycle to establish your comparison baseline, retaining its existing records of raw-material lots, equipment setup, tray layout, recipe version, sampling steps, and test results. Meeting quality targets in a candidate or baseline sample does not approve the run as a production cycle.
Use the planning ledger below to connect each practical question to its target evidence. These hypothetical entries illustrate proposed comparisons, not actual measured results or promised universal settings.
| Question | Candidate change | Keep comparable | Evidence for the decision |
|---|---|---|---|
| Does layer geometry explain a persistent wet center? | A defined thickness or fill-depth comparison. | Product lot, preparation, freezing basis, tray arrangement and the other recipe settings. | Position-based moisture and center checks, product-temperature response, stage duration and quality. |
| Is primary drying unnecessarily conservative? | A justified heat-input or pressure comparison inside the product and equipment limits. | Geometry, freezing history, loading and the endpoint method. | Product-limit margin, pressure and condenser response, transition evidence and final quality. |
| Does the proposed secondary stage reach the target? | A planned temperature-profile or hold comparison within the product’s allowable exposure. | Confirmed primary-drying completion, sampling method and handling before testing. | Residual moisture, specified quality measures and location-to-location variation. |
| Does the recipe remain suitable at a different load? | A defined partial-load versus intended-load comparison. | Product form, per-tray geometry, freezing basis and documented equipment configuration. | Temperature distribution, transition behavior, pressure stability, condenser response and product acceptance. |
These are engineering planning examples. Actual settings, test levels, repetitions and acceptance limits must be defined for the product and dryer.
Adjust one variable at a time whenever a single-factor comparison can test the explanation. If factors may interact, use a planned multivariable comparison rather than drawing conclusions from unrelated runs. In both cases, log uncontrolled differences alongside intentional modifications, including incoming ingredient lot shifts or variations in frozen product temperature at loading.
What if the faster batch changed more than one thing?
Suppose a trial run completes more quickly after both the heating profile and the raw ingredient preparation are changed in a hypothetical trial. The comparison cannot isolate which simultaneous change caused the time reduction. Document both changes together with the resulting product quality, then plan an isolated trial with preparation kept comparable, or design a planned experiment to separate their effects.
Never treat a confounded trial as the reason for a recipe change, though an inconclusive result may still justify holding a candidate recipe, repeating the run, or investigating further. Conclude each test run with an unambiguous disposition: adopt, reject, repeat due to inconclusive evidence, or investigate another explanation. Identify the specific records supporting that finding along with the next engineering question. A faster cycle that misses finished product specifications is a rejected run, not an acceptable outcome.
Has primary drying actually finished?
Use only endpoint methods that the installed dryer and its approved procedures support. A pressure-rise test needs suitable chamber-isolation hardware and an approved operating procedure; do not improvise one by isolating an active chamber.
Before advancing to secondary drying, check what evidence shows that all ice has sublimated from this specific food and batch load. Seeing a controller advance simply confirms that it ran a programmed step or triggered an event; it does not prove that primary drying has finished across the chamber. Verifying that the transition rule works for this product requires supporting endpoint evidence rather than assuming the recipe or timer holds true.
Pisano, Barresi and Fissore studied pressure-rise monitoring in food trays using sugar solutions and frozen spinach. Near-endpoint gas behavior depended on cycle setup, loading and product type. They also showed why inserting a probe can change local heat transfer or vapor paths, so the probed location may dry differently from nearby product. Their experimental times and conditions are not commercial food recipes. [1]
Product-temperature response, pressure comparison and a suitable pressure-rise method can contribute evidence about sublimation completion. Interpret each within its instrument, equipment and test conditions. A temperature rise alone does not show that the last ice has disappeared throughout the load. [1] [3]
Document the transition logic so a qualified reviewer can examine it. State the primary signal or combination of indicators, sensor positions, calibration basis, tested load configurations, independent product checks, and the required response when signals conflict. Whenever equipment features allow, correlate sensor indications with independent product checks across representative tray positions.
Store this transition logic alongside the recipe revision and raw trend exports. Refer to the monitoring and batch-record guide for sensor selection and routine trend review. This article does not recommend universal pressure ratios, rigid pressure-rise thresholds, or extra hold durations.
What should happen when the signals disagree?
Retain the conflicting data alongside the specific recipe revision and its raw trend logs. Have a qualified reviewer check probe calibration, physical placement, data timestamps, sampling locations, handling practices, and the diagnostic limits of each test before assigning any cause to the discrepancy.
Select an approved discriminating check compatible with the installed equipment to separate competing explanations. If the discrepancy cannot be explained, log the transition evidence as inconclusive and hold the proposed transition change. Testing finished product moisture answers a final-condition question rather than ice sublimation timing, so it cannot resolve a primary endpoint dispute on its own.
Does the product meet its specification?
Confirming primary-drying completion and accepting the finished product are separate decisions. Secondary drying removes additional non-frozen water by desorption. Evaluate its temperature exposure and duration against the product’s moisture target and sensitivity to quality damage. Nowak and Jakubczyk’s food-process review explains the dependence of desorption and stability on the material and its water content; it gives no single endpoint suitable for every food. [3]
Evaluate secondary drying only after primary drying reaches a consistently justified endpoint. Otherwise, differences in residual ice or moisture entering the secondary phase can confound comparisons between runs. Standardize sampling positions across trays, and keep the time between chamber unloading, sealing, and testing comparable across trials.
- Record the actual product-temperature exposure as well as the programmed heating profile.
- Compare moisture results using the same specified method and reporting basis.
- Check the defined quality responses and variation among representative positions, rather than accepting only the batch average.
- Where specified, measure water activity under a defined test condition and retain it separately from moisture.
- Verify packaging and storage performance through the product’s own program. A drying endpoint does not by itself establish shelf life or a microbial kill step.
Approve the secondary drying endpoint only when test data confirms that agreed product requirements are satisfied within a justified operating window. Adding an unexamined long final hold may mask process uncertainty rather than resolving it. Conversely, shortening stage time without comparative moisture and quality testing leaves residual moisture unverified.
What changes when the cycle moves to production?
Review the physical differences between pilot and production chambers before transferring recipe settings. In particular, distinguish adding more trays at comparable geometry from deepening the food layer on each tray, because these changes can affect local drying paths, vapor resistance, and total vapor demand in different ways. Their cycle effects cannot be assumed equivalent.
In a yogurt experiment, Sharma and Arora found that the relationship between thickness and production rate depended on the heating method. The useful lesson is to consider geometry and heat supply together. The study’s settings are not a scale-up rule for other foods. [2]
Do not treat matching shelf-temperature setpoints as proof that pieces experience the same product-temperature history. Instead, consult the transfer matrix below to identify the physical differences present on the receiving unit, and request the relevant verification evidence for those specific changes. Review whether each difference could alter product vapor resistance, conductive or radiant heat transfer, dryer vapor handling, or sensor interpretation. Evaluate each possible mechanism as a hypothesis rather than an established fact, and confirm the relevant receiving-system performance before accepting the transferred cycle.
| Change | Why comparability may change | Evidence to request |
|---|---|---|
| Formulation, solids, preparation or freezing history | The product’s structural limit and vapor-flow resistance may change. | Comparable material characterization, geometry and freezing records; renewed product checks where relevant. |
| Tray, contact, shelf arrangement or heating method | Contact and radiation can change the relationship between the heat-input setting and product response. | Representative temperature histories and product results across the receiving load. |
| Layer depth, piece distribution or total loading | Local drying paths, position effects and total vapor demand may change in different ways. | A documented loading map, stage behavior and location-based acceptance results at the intended conditions. |
| Vapor path, condenser or pressure-control behavior | The receiving system may handle the evolving water-vapor load differently. | Loaded pressure and condenser trends, alarms and evidence that the agreed process window is maintained. |
| Sensor system or stage-transition logic | The same displayed value or event name may represent a different measurement or rule. | Sensor identification, locations, control logic and rechecked transition evidence. |
This matrix identifies verification needs. It does not predict the receiving dryer’s cycle time or prescribe a transfer protocol.
Who closes each transfer difference?
Designate an assigned reviewer to record each system difference, consider its possible mechanisms, and evaluate the change against receiving data, product response, and final quality results under agreed limits. For example, when tray-contact mechanics differ between dryers, verify representative food temperatures and finished product attributes directly rather than relying on matching shelf-temperature curves. A receiving dryer may require an adapted recipe supported by traceable evidence; identical machine trends across different equipment are not required.
Record the explicit disposition and engineering rationale for every recipe modification to maintain traceability. If actual production conditions fall outside demonstrated receiving, product, or equipment scope, place the unresolved recipe revision on hold until renewed product- and system-specific confirmation trials are completed. Neither an arbitrary percentage adjustment nor a generic trial count can release an unverified recipe.
Verify frozen food handling, frozen-load release, and product loading criteria against the pre-freezing requirements guide. Once cycle recipe and operating load inputs are established, use measured cycle and load inputs to plan factory production. The cycle-time guide and batch-capacity guide support that production planning.
What evidence should move with the recipe?
Provide the receiving technical team with records establishing what was tested, under what operating conditions, and which boundaries remain unverified. The team should be able to trace each accepted recipe change directly back to the trial comparisons that justified it, including what was learned from rejected batches and inconclusive tests.
- Product and load specification: the tested formulation, preparation, geometry, tray arrangement and operating-load range.
- Recipe and rationale: the accepted version, product limits, programmed stages and reasons for important changes.
- Stage and product evidence: original trends, transition-method evaluation, sampling map, test methods and final results.
- Trial dispositions: accepted, rejected and inconclusive comparisons, including deviations and unresolved questions.
- Transfer and reassessment conditions: the receiving configuration, demonstrated range and changes that require a new review.
Cycle development does not replace equipment acceptance or a food-safety program. Contractual FAT/SAT and loaded-performance requirements belong in the food freeze dryer validation plan.
Agree the trial deliverables before comparing offers
Ask prospective development providers to itemize trial deliverables before comparing offers: clarify what is included, specify who supplies every measurement or record, and establish what happens when evidence is missing. Review these points when comparing a sample-only offer with a documented development package without assuming they are equivalent.
- Measurements and samples
- Name the tests, methods, reporting basis, sample positions and handling conditions. Identify results supplied by the buyer or an outside laboratory.
- Records, including unsuccessful trials
- Agree access to original trend exports, recipe revisions, and accepted, rejected or inconclusive trial records with their decisions.
- Transfer and remaining work
- Identify the receiving dryer, loading assumptions, unresolved differences and who performs the next verification. Mark missing evidence UNKNOWN.
This checklist outlines trial and transfer records to request. It does not imply that every test or transfer activity is included in a provider’s proposal.
Use a trial decision record
A trial record is useful when it states which explanation is being tested and how the result changes the next decision. The example below is hypothetical and contains no production recipe or customer result.
- Question
- Observed product problem and the explanation being tested
- Planned comparison
- Deliberate change; product, load and process factors to keep comparable
- Evidence
- Sampling locations, methods, trends and agreed product checks
- Disposition
- Adopt, reject, repeat as inconclusive or investigate another explanation; record the basis
- Next action
- Responsible person, next question, missing evidence and recipe revision if justified
Teaching example
- Question
- Does fill-depth variation help explain inconsistent centre dryness?
- Comparison
- Compare defined loading arrangements using the same formulation and freezing basis; agree the other controlled factors in the trial plan.
- Teaching observation
- Suppose sampling does not represent the difficult tray positions. The available results cannot resolve the original question.
- Disposition and next step
- Mark the trial inconclusive for that question; revise the sampling plan before attributing the result to fill depth.
This is an example of decision discipline, not an experimental protocol or a claim that a particular depth or cycle will succeed.
Copy the completed notes into the message field on the project enquiry page. Information entered here is not submitted or stored by this worksheet.
Prepare a focused cycle-development discussion
Send the food and formulation, solids or initial moisture, piece dimensions or layer depth, tray and batch loading, current recipe and available trend data. Include the required final quality and the specific uncertainty: wet centers, an unclear transition, excessive duration or transfer to a different load or dryer.
These inputs support a preliminary engineering discussion of the next useful test and the equipment evidence needed. Final settings and production performance require product-specific verification.
Food freeze-drying cycle development FAQ
Can a development cycle have a fixed total time?
A commercial production recipe may rely on fixed stage timers once those durations have been justified for the product, loading conditions, and endpoint verification method. Throughout development, however, check programmed timer settings against process data. Running a preset timer cannot by itself prove that ice sublimation has finished or that the recipe is ready for transfer.
Can laboratory shelf settings be copied to a production dryer?
Treat pilot settings as a documented starting reference condition, keeping them within the receiving system’s operating limits. Verify the actual product response, geometry, heat transfer characteristics, vapor handling, transition data, and finished food quality on the production unit. Equal controller setpoints do not establish equivalent drying conditions.
How many successful batches establish a cycle?
Establish a confirmation plan reflecting product variability, process changes under review, analytical measurement capability, and the intended operating range. No universal batch count applies across all food products. A successful run with a lightly loaded chamber leaves cycle repeatability across production loading unproven.
Should the shortest acceptable trial become the production recipe?
Before releasing a fast cycle, verify its operating margin and repeatability across agreed material and loading variations. A fast trial run is a promising candidate, but it becomes a supported recipe only after supporting confirmation covers its intended use.
Sources and evidence boundaries
- Pisano R, Barresi AA, Fissore D. Innovation in Monitoring Food Freeze Drying. Drying Technology. 2011;29(16):1920–1931. doi:10.1080/07373937.2011.596299. Food-tray monitoring experiments; no universal endpoint threshold or commercial recipe is adopted.
- Sharma NK, Arora CP. Influence of product thickness, chamber pressure and heating conditions on production rate of freeze-dried yoghurt. International Journal of Refrigeration. 1995;18(5):297–307. doi:10.1016/0140-7007(95)00025-7. Qualitative thickness/heating comparison for yogurt; numerical operating results are not generalized.
- Nowak D, Jakubczyk E. The Freeze-Drying of Foods—The Characteristic of the Process Course and the Effect of Its Parameters on the Physical Properties of Food Materials. Foods. 2020;9(10):1488. doi:10.3390/foods9101488. Review evidence for stages, desorption, endpoints and material-dependent behavior; individual study settings require their own context.
The development ledger and transfer matrix are proposed engineering worksheets. They contain no customer measurements and are not reproduced experimental protocols.