Glass Bottle Sterilizer Tunnel: Placement and Heat Curve

A sterilizing stage on a glass line belongs in one of two places: ahead of the filler, where it treats the empty container, or behind the closing machine, where it treats the sealed package. Only the second can make a product shelf stable, and an open tunnel can only do that for acid or acidified products, because its water cannot go above boiling point. A low-acid product needs a pressurised process. The heat curve then has to be proven at the slowest-heating point inside the container, and the glass, closure and decoration have to be specified against that curve, cooling included.

Start from the product, then the machine, then the container

Sterility describes a process. It does not describe a jar. One and the same glass container can hold a pasteurized product, a commercially sterile one or an aseptically filled one. What differs between those routes is the heat the package must survive, the pressure across the closure, what the decoration is exposed to, and the records that must exist afterwards. Good behaviour on one route tells you nothing about another until the container has been tested on that route's real profile.

The decisions therefore run in a fixed order. The food technologist or process authority first classifies the product:

• whether it is acid, acidified or low-acid;
• whether particulates, pulp, starch or fat alter the way heat travels through it;
• whether it is carbonated or will receive a nitrogen dose before capping;
• the shelf life required and the temperature it will be stored at.

Those answers give the target organism, the reduction needed and so the intensity of the heat step. Equipment is chosen next. The container comes last, as a design problem against a profile that already exists.

On most projects the heat step is run by a filler, a co-packer or the brand's own plant, while the glass comes from a supplier who never sees the tunnel or its process record. Risk collects in that gap. The filler knows the worst case but seldom writes it in a form a glass supplier can test to, and the glass supplier offers a general heat-resistance statement that cannot be inspected. Before a specification is frozen, we ask the process side to put the worst case in writing, stops and restarts included, and ask the container side for evidence taken against that figure, not a nominal one.

Four machines sold under the word sterilizer

Catalogues use the term loosely. When a quotation says "sterilizer" with no qualifier, the first thing to establish is which of these is on offer:

• Empty-container sterilizer. Treats the bottle or jar before filling, normally with hot water, steam or a chemical rinse, then a rinse and a dry.
• In-package tunnel sterilizer or pasteurizer. Treats the closed package. Unless the tunnel is pressurised, the boiling point of water caps what it can do.
• Steam autoclave or retort. Treats a static or rotating load under pressure and reaches conditions no atmospheric tunnel can.
• UV or pulsed-light tunnel. Works on surfaces only and shares almost nothing with the other three.

Several neighbouring jobs are not sterilization at all. Washing removes soil, residue and scale; drying makes sure the container reaches the filler dry; aseptic filling builds a sterile space in which product and container meet, so that suitable products need no in-package heat step.

Before the filler or after the closer

Both positions are legitimate and they solve different problems. A container stage deals with what the glass carries into the filling room. A package stage deals with what the product and the headspace carry. The costly mistake is to sterilize empties, fill a product that was never stabilised, and assume the two operations add up to one assurance. They do not. Shelf stability comes only from the package stage, and only when that stage is sized against the organism in the product.

Placing a container stage

The usual position is after washing and drying and directly ahead of the filler. Every metre between sterilizer and filler hands the clean container back to the room air. A container that leaves the stage wet turns its own water into a carrier, which is why a wet container stage sits badly beside a dry filling area. Either the container is dried thoroughly after treatment, or the room is accepted as a wet area and designed that way from the start.

Placing a package stage

Its position is given: it follows the closing machine. What it does to the line around it is the real decision. Dwell is fixed by food safety and cannot be cut to win back lost time, so once installed the tunnel sets the pace. The filler has to be sized to the tunnel, not the other way round. Belt speed and pitch for a given container determine dwell, so a format change is a process change, never just a mechanical changeover.

A container cleared for a sterilizing route also has to run on the line that feeds the tunnel. The tolerances a line needs before it will handle a container at all are set out in our guide to the glass bottle filling line.

Pasteurization and commercial sterility aim at different organisms

Describing pasteurization as mild sterilization is misleading. The two routes differ in their target, not only in how hard they push.

Pasteurization is sized against vegetative organisms that spoil a product or harm a consumer but form no spores: typically yeasts, moulds, lactic acid bacteria and non-spore-forming pathogens. It runs below the boiling point of water for minutes to tens of minutes, to a reduction target agreed for the product and its market. That suits beer, cider, juices, functional drinks and dairy drinks. On the container side the issues are pressure balance across the closure, water chemistry and decoration exposure, not extreme heat. The profiles, label exposure and audit questions of that route are covered in the article on the tunnel pasteurization line.

Sterilization of packaged food means commercial sterility: nothing in the product can grow under the intended storage conditions. The target is the heat-resistant spore formers that come through a pasteurization-grade treatment. Killing spores takes either a far higher temperature or a far longer dwell, and here the physics of an open tunnel intervenes.

What the boiling-point limit means for each pH class

For acid and acidified products the pH itself holds back the organism of concern, and the remaining risk is spoilage, not pathogen growth. A near-boiling hot fill or hot tunnel can be enough, documented as a scheduled process for an acidified food.

For low-acid products the temperature that would do the job in seconds is out of reach at atmospheric pressure. The recognised route is a retort, an autoclave or a pressure-balanced tunnel that keeps the package above atmospheric pressure while the contents pass the boiling point.

The public rules behind the split

In the United States, FDA 21 CFR 113 sets scheduled-process requirements for low-acid canned foods, 21 CFR 114 covers acidified foods, and 21 CFR 120 is the HACCP framework for juice. 21 CFR 1700.20 is something else entirely, a child-resistance protocol. In the European Union the starting references are Regulation 852/2004 on general hygiene and Regulation 1935/2004 on food contact materials, with LFGB as the German national reference.

A glass supplier cannot claim any of these. They frame the work of the filler's process authority. For a buyer the practical test is simple: a sterilization temperature quoted without the product's pH class and the target organism does not describe a process.

Four ways of delivering the heat

With intensity fixed, the next choice is the delivery method. Four inputs drive it: whether pressure is needed at all; how sensitive the product is to agitation and uneven temperature; the container's format, wall and shape; and the decoration, because a route that wets the whole container narrows what can be applied beforehand.

Two points deserve emphasis. In a spray tunnel, nozzle condition is a quality variable, so nozzle inspections should be filed with the process data. In a recirculating bath the water is shared between packages, and its microbiological quality matters for the same reason cooling water quality does.

Which stage owns which contamination risk

Risk, not machinery, decides which stages exist and what proof each one owes. The table gives directions and bands on purpose. The values for a real schedule come from your product, your container and your process authority.

The rows separate two complaints that tend to be argued as one. Spoilage may be a thermal shortfall, which is a process fault. It may equally be recontamination of a batch that was heated correctly, which is a packaging and hygiene fault. The organism and the timing of the complaint usually tell you which.

Two studies that turn a machine into a controlled process

Sign-off rests on a heat distribution study and a heat penetration study. They are often confused and they answer different questions. Both are run by the process side. The container supplier's part is to show evidence against the profile they produce.

Heat distribution

This study maps the equipment. It checks that every position a package can occupy gets the same treatment, across the tunnel section and along the belt, at the worst-case belt speed and not a comfortable one. Cold locations tend to turn up near a wall, at the rim of a nozzle pattern, low in a spray zone, in a lane with less flow, or in the zone furthest from the steam header. Any change to nozzles, water distribution or loading pattern calls for a repeat. A map drawn once and never revisited may describe a machine that no longer exists.

Heat penetration

This study tracks the product. It records the temperature at the slowest-heating spot inside the package over time and compares the lethality accumulated there with what the scheduled process demands. Come-up, hold and come-down all count, so it is not a hold-zone check. The cold point must be located, not presumed. It shifts with the heating mechanism, container geometry, fill weight, particulates and fill temperature. A viscous puree heating by conduction and a thin liquid heating by convection have different cold points, and one recipe in two jar sizes may too.

Changes that require a new study

Write the revalidation triggers into the same document. Each of the following can move the cold point or alter the delivered process, and the right response is a fresh study, not a log entry:

• recipe, viscosity, solids content, pH or water activity;
• fill weight, fill temperature or headspace;
• container size, shape or wall;
• closure or liner;
• belt speed, pitch or zone temperature;
• a replacement nozzle set;
• the procedure followed when the line stops.

Measuring the cold point inside a jar

Most validation exercises fail here, on what looks like a technician's routine. Four details decide whether the reading means anything.

1. Placement. The coldest spot is not the geometric centre by default. In a liquid heating by convection it sits low, since warm liquid rises and cooler liquid pools underneath. In a conduction-heating puree it lies nearer the centre, shifted by container shape and fill level. A shoulder or narrow neck changes the heat path again. The method is exploratory: several probes on a grid in representative packs, the slowest position identified for the real recipe and container, then that position fixed for routine monitoring.
2. Sensor behaviour. Type T thermocouples and wireless loggers are both in use. A needle probe touching the glass reads the wall, and during heating the wall runs ahead of the product, so the batch looks safer than it is. A probe in a gas bubble, or in a concentrated phase of a particulate product, reports a local condition. Wireless loggers intrude less but bring their own response time and placement limits.
3. The seal. A jar with a probe through the lid is no longer the production pack. Headspace and pressure behave differently, so the probe may see a different process. Seal the penetration so that measuring does not change what is measured, and state the remaining difference openly.
4. Calibration. Check probes before and after the run against a stated allowable drift. The check afterwards matters most: a probe that began correct can read low by the end of a long process.

The recurring failures are predictable. Calibration is done beforehand only. A single pack is instrumented when it should be the coldest position in the slowest lane. Air temperature stands in for product temperature. Readings from a nominal size are applied to another size. A belt average replaces the worst position. Each gives a record that looks finished and proves nothing.

What the heat load does to the product

Thermal load is time multiplied by temperature at the cold point, and for a set lethality one can be exchanged for the other. Hotter and shorter generally keeps heat-labile nutrients and volatile aromas better than cooler and longer. That reasoning lies behind flash treatments, high-temperature short-time processes and the aseptic route.

An atmospheric in-package stage has little room for that exchange, because boiling point caps its temperature. If lethality cannot be reached by going hotter, dwell is all that remains, and long dwell is what hurts the product. The route is dependable for food safety and costly for quality:

• cooked notes in dairy and tea;
• browning in high-sugar products;
• colour change in anthocyanin-rich drinks;
• losses of thiamine, folate and vitamin C;
• breakdown of the volatile compounds that give the product its identity.

Carotenoids and chlorophyll respond differently again, and citrus can turn bitter instead of merely losing freshness. Non-enzymatic browning and caramelisation mark an over-processed batch, and consumers tend to notice them first.

Headspace oxygen pushes the same way as heat. Residual oxygen speeds vitamin loss, oxidises aroma compounds and fades colour, so a thermally correct process can still give a short shelf life. Shelf-life reviews usually look at both together. How the headspace is handled before capping, and why some products get nitrogen in place of air, is explained in the piece on the nitrogen dosing station.

One consequence is that a recipe cannot simply change routes. A formula built for a near-boiling hot fill has to be re-evaluated by the food technologist before it goes to a pressurised high-temperature process, and likewise in reverse. The thermal window has a floor, set by the lethality the process authority requires, and a ceiling, where the product stops being itself. Write the container requirement against the full window. The ceiling sets the worst differential the glass will meet.

How much thermal shock the glass can take

Thermal shock is a temperature difference through the container wall imposed faster than the glass can even it out. Glass is brittle and cannot yield. Whichever face changes temperature first strains against the remainder of the wall, and the stress either dissipates or opens a crack. The curve decides the result more than the peak does: the rate at which the wall is driven, the time the difference is held and, most of all, the speed at which it is reversed.

Heating and cooling are not mirror images. While a package heats, its contents warm at a comparable rate and back up the inner face, and the outer face is in compression, which glass bears fairly well. While it cools, the outer surface is drawn into tension. That is the direction in which containers break, and it explains why a specification of "so many degrees" cannot be used. Without a direction and a rate, the figure can be neither inspected nor argued from.

Four properties of the container shift the real limit:

• Wall thickness distribution. It counts for more than nominal thickness, since heat crosses the thinnest section first and the gradient concentrates there.
• Residual stress. An uncontrolled annealing curve leaves stress that uses up part of the resistance before processing starts. Annealing records belong in the evidence pack.
• Shape. Sharp shoulders, square corners and sudden changes of section concentrate stress. A jar designed for the retail shelf is not thereby designed for a tunnel.
• Fill level and contents. A jar full of dense product heats and cools along a different path from one that is partly filled.

Pressure adds a second axis. In an open tunnel the surroundings stay at roughly atmospheric pressure while internal pressure climbs with temperature, so the closure carries the load. In a pressurised process an external pressure is applied to balance the internal one, and the glass is loaded by pressure as well as heat. The differential across the wall then depends on pressure control as much as on temperature, and losing that control is a container event, not just a process event.

For an aggressive process, we write the requirement as a temperature difference in a named direction, at the rate the line delivers, covering the worst case at a stop and restart, with a margin agreed between filler and glass supplier. The test method and an auditable pass criterion are a matter for thermal shock testing. A number quoted without its method is not evidence.

Cooling, water quality and the walk to the labeller

On a sterilizing route, cooling decides two things beyond the curve: whether the package stays sealed, and whether the water touching it becomes a route for contamination.

Stepped cooling from an intermediate temperature keeps the wall differential inside what the glass tolerates and lets internal pressure fall steadily instead of collapsing. Going from the hold zone directly into cold water breaks more containers and distorts more liners than any other arrangement. Local effects matter too. Cold spray striking one face of a hot jar sets up a gradient much steeper than the wall average, another reason nozzle condition and coverage are process variables.

Late in cooling, internal pressure drops below atmospheric and the pack is briefly under vacuum. A liner whose seal is already weakened will pull cooling water into the headspace. That is a microbiological failure, not a packaging defect, and it surfaces as spoilage days afterwards on a batch with a complete record. Treat the water reaching the finish at that moment as a product-contact medium of defined quality, and take the last rinse from a softened or treated supply, not from a recirculated bath in unknown condition.

Hard water causes a quieter fault. Dissolved solids dry onto the body, where they only show, and onto the sealing surface, where they alter the friction and surface the closure seats on. The results appear later as erratic capping torque, leakage in storage or complaints about marks. Blow-off after the final stage, planned draining and cleaning, filtration and control of biological growth in warm recirculated water are all routines for the process record.

Condensation and decoration after the tunnel

A bottle reaching the labeller warm and wet will not hold a paper label. A shrink sleeve traps water against the glass and shows wet marks, and print applied offline may fail to adhere. Bottle surface temperature at labelling is a specification in itself, because adhesive choice depends on it. The distance and accumulation between tunnel exit and labeller govern how much heat and water the bottle has shed by then.

The usual remedies are blow-off, a short warm hold or a dedicated drying station; the equipment and its checks are described under glass bottle dryers. On the container side the question is the decoration. Fired-on ceramic decoration is part of the glass and survives the full load. A pressure-sensitive label on a wash-resistant adhesive does better than wet-glue paper. Anything applied before the tunnel should be tested against the real water temperature and detergent, not a generic claim.

Diagnosing an under-processed batch

The reflex after spoilage or a failed release is more temperature or more dwell. It is expensive and often wrong, since the shortfall frequently lies outside the process. Work from the cheapest, likeliest cause to the costliest.

1. Measurement chain. Were probes calibrated after the run too? Was the probe in product and clear of the wall? Was the cold point actually located for this recipe and container? Was the record built from the slowest lane and not an average? Many findings disappear at this step.
2. Product. Reformulation, a new ingredient supplier, different solids, higher viscosity, larger particulates or a changed initial temperature all move the cold point and raise the process needed. So does a change in fill weight or fill temperature, the alteration most easily made without telling the process engineer.
3. Container. A different jar size alters the heat path and a different wall alters the rate. With the same recipe, a heavier container means the product heats more slowly. If the failed batch is the only one in a new size, suspect the container before the machine.
4. Line. Because speed and pitch fix dwell, a speed increase made for output quietly shortens the process. In a spray tunnel, look for blocked or worn nozzles, check flow and pattern, and confirm the distribution study still matches the machine. Compare zone temperatures with set points and inspect the supply side: steam pressure, water flow, blocked strainers. A zone can read correctly at its sensor and still be starved upstream.
5. Stoppage record. A pack halted in a hold zone gets more process than planned, and one halted in a cooling zone is affected in the opposite direction. Both are deviations needing a recorded decision. A plant that restarts the belt and says nothing cannot know which packs were involved.
6. A cause that is not thermal. When the record is complete and the batch still spoils, the organism usually identifies the stage. Flat-sour spoilage indicates spores and therefore lethality. Gas production often indicates a sealing or vacuum fault. Mould or yeast turning up later usually means recontamination after processing, most often via cooling water, a weakened seal, or handling between tunnel exit and the next dry station.

Identifying the organism before touching the process stops a plant from re-cooking a product that was never under-cooked. The same logic applies upstream: soil and residue left on the glass before filling are a cleaning problem, dealt with in our guide to the bottle washing machine.

What to send when briefing a sterilizing route

Three sets of information, sent together, settle most of the open questions:

• The product. pH class, carbonation or nitrogen dosing, particulates or pulp, viscosity and fill weight.
• The shelf-life target. Required shelf life and storage condition, plus any target organism or process value a process authority has already fixed.
• The line. Intended container format, available belt speed range and pitch, atmospheric or pressurised tunnel, and the cooling arrangement.

From these it follows whether the stage goes ahead of the filler or behind the closer, whether the package has to act as a pressure vessel, and what differential the glass, closure and decoration must accept. We reply with a suggested route and the temperature and time window it implies, stated as the range in which the cold point must be reached, along with the stage position, the cooling sequence and a short list of evidence that should travel with the batch. If the product class calls for a pressurised process or an aseptic route, we say so plainly.

Every figure is a reference approach. It has to be confirmed against your own product, process authority and line, and we do not promise a process result or certification ahead of that work.

Questions buyers ask about sterilization tunnels for glass bottles

How do sterilization and pasteurization differ on a glass line?

By target and by physical reach. Pasteurization inactivates vegetative organisms such as yeasts, moulds and non-spore-forming bacteria, below boiling point, over minutes to tens of minutes. Sterilization means commercial sterility and goes after heat-resistant spore formers, which demands much more temperature or much more dwell. Since an open tunnel's water cannot pass boiling point, a low-acid product has to go through a pressurised process.

Should the glass be sterilized empty or after sealing?

Neither replaces the other. Treating empties controls what the container brings to the filler. Treating the sealed pack deals with product and headspace and is the only one of the two that delivers shelf stability. Let the product's pH and target organism decide which stage is needed, not layout convenience.

How is heat penetration validated in a sealed glass container?

A calibrated sensor is fixed at the slowest-heating position, product temperature is logged from come-up through hold to come-down, and the lethality accumulated there is set against the scheduled-process value. The position is found by exploratory runs on the real recipe and container, and the probe entry is sealed so the pack's pressure behaviour is unchanged. Probes are checked at both ends of the run. The study is redone whenever recipe, fill weight, container, closure or belt speed changes.

Where is the cold point in a filled jar?

Often not in the middle. Thin liquids that heat by convection are coldest low down. Purees that heat by conduction are coldest nearer the centre, offset by shape and fill level. Find it with a grid of probes in representative packs, once per size and per product. Borrowing a cold point from another recipe or jar size is a frequent reason a validation passes on paper and fails in the field.

Can a tunnel sterilizer handle low-acid products?

Only when it holds the package above atmospheric pressure. An open tunnel stops at the boiling point of water for the plant's altitude, which falls short of spore-level treatment. The accepted options are a retort, an autoclave or a pressure-balanced tunnel, with the scheduled process established by a qualified process authority. Settle the product's pH class before discussing equipment.

How much temperature shock will a bottle survive in a sterilizing tunnel?

No single figure is useful. The limit varies with the direction and rate of change, wall thickness and how it is distributed, residual stress from annealing, and container shape. Heating is the more forgiving direction, so most breakage occurs in cooling. Specify a difference in a stated direction at the line's real rate, stop-and-restart worst case included, and verify it by a defined method on production samples from the moulds concerned.

Why do jars break during cooling and not in the hold zone?

Cooling puts the outer surface in tension, whereas heating puts it in compression with the contents supporting the wall from inside. A cold spray on one side of a hot jar makes things worse locally. Stepping the temperature down from an intermediate stage, instead of applying one cold shock, is the standard control.

Where should troubleshooting of under-processing begin?

With the measurement chain: calibration drift, a probe against the wall, or a cold point inherited from another product. After that, check product and fill, then container, then line speed, pitch and nozzles, then the stoppage record. If spoilage occurs despite a complete record, classify the organism first, because recontamination after processing resembles under-processing and needs a very different remedy.