Deep vacuum occurs when pressure inside a sealed system drops to one millionth of atmospheric pressure or lower, creating conditions where normal physics breaks down

A deep vacuum is not just "very empty air." Once you drop below about one millionth of atmospheric pressure — roughly 0.001 pascals — the behaviour of gases, heat, and materials changes fundamentally. Molecules stop bumping into each other often enough to behave as a fluid. Instead, each one travels in a straight line until it hits a wall. This threshold marks the boundary between high vacuum and deep vacuum, and crossing it means the equipment, materials, and measurement methods you use have to change completely.

Deep vacuum matters because it is the only environment where certain manufacturing and research processes work. Semiconductor fabrication, particle physics experiments, and space simulation all require it. But reaching and maintaining deep vacuum demands equipment that is expensive, finicky, and unforgiving. A single leak the size of a pinhole can ruin months of work.

Key Takeaways

  • Deep vacuum begins below one millionth of atmospheric pressure, where gas molecules no longer collide with each other and instead travel in straight lines until hitting a surface.
  • Outgassing — the release of trapped gases from materials inside the chamber — becomes the dominant source of pressure and can take weeks or months to control.
  • Pumping systems must be staged: rough pumps bring pressure down to a manageable level, then turbomolecular or diffusion pumps take over for the final drop into deep vacuum.
  • Leak detection and prevention are critical because even microscopic holes will slowly destroy the vacuum and contaminate experiments or products.
  • Materials, seals, and measurement instruments must be chosen specifically for deep vacuum conditions, as many common substances outgas or fail under these pressures.

How pressure regimes change what happens inside the chamber

At atmospheric pressure, air molecules collide with each other roughly a billion times per second. These collisions create the pressure you feel and allow heat to move through convection. Below one millionth of atmospheric pressure, a molecule travels an average of 100 metres before hitting anything. This distance — called the mean free path — is now larger than most laboratory equipment.

This shift breaks the rules that governed the system at higher pressures. Heat no longer travels by convection; it can only move by radiation or conduction through solid contact. Gases no longer behave as fluids that flow smoothly; instead, individual molecules bounce around like billiard balls. Pressure itself becomes harder to measure because the instruments that work at higher vacuums — like Pirani gauges — stop functioning. You need ionization gauges or other specialized tools that count individual molecules instead of measuring bulk pressure.

Materials also behave differently. Metals that seemed inert at atmospheric pressure will outgas steadily in deep vacuum. Plastics and elastomers release trapped solvents and monomers. Even stainless steel, the workhorse of vacuum systems, sheds molecules. This outgassing is often the limiting factor: you can pump away the air, but the materials themselves become the source of pressure.

Outgassing and why it dominates deep vacuum systems

Outgassing is the release of gases trapped inside or on the surface of solid materials. At atmospheric pressure, outgassing happens but is invisible because the air around the material masks it. In deep vacuum, there is nothing to mask it. Every molecule released from the walls, the pump, the seals, or the sample itself contributes directly to the pressure inside the chamber.

The process happens in stages. First, loosely bound surface molecules leave quickly — this takes hours to days. Then, gases trapped deeper in the material begin to diffuse outward — this can take weeks. Finally, very deep pockets of gas release slowly over months. A stainless steel chamber that has been exposed to air will outgas for a long time after you seal it and start pumping.

To manage outgassing, vacuum systems use baking: heating the entire chamber to 150–250 degrees Celsius while pumping. Heat accelerates the release of trapped gases, so they leave the system instead of trickling out during your experiment. A bake-out cycle can take days or weeks, but it reduces outgassing by orders of magnitude. Without baking, reaching true deep vacuum is nearly impossible.

Pump stages and why one pump cannot do the job

No single pump can pull a system from atmospheric pressure down to deep vacuum. The process requires at least two stages, often three.

Rough pumps — usually rotary vane or rotary screw pumps — bring pressure down from atmospheric to roughly 0.1 pascals. They work by trapping gas in a shrinking volume and expelling it. They are fast and can handle the large amount of gas present at the start, but they cannot go lower because the gas they are trying to expel leaks back past the seals.

Turbomolecular pumps take over from there. They use a series of spinning rotors with angled blades that strike gas molecules and push them toward the exhaust. A turbomolecular pump can reach pressures below one millionth of a pascal, but it cannot start from atmospheric pressure — the high pressure would damage the rotor. It also cannot pump water vapour or other condensables, which would freeze on the blades.

Diffusion pumps are an older alternative that uses heated oil vapour to capture and carry gas molecules toward the exhaust. They are cheaper than turbomolecular pumps and can reach very low pressures, but they are slower, require careful temperature control, and can contaminate the chamber if the oil backstreams.

A typical deep vacuum system chains these together: a rough pump pulls the chamber down to a few pascals, then a turbomolecular pump takes over and reaches the target pressure. The rough pump stays running to remove the gas that the turbomolecular pump expels.

Leak detection and the cost of small holes

A leak the size of a human hair — roughly 100 micrometres across — will allow enough air to enter a deep vacuum system to ruin it in hours. Smaller leaks take longer but are equally fatal to long-term experiments. Finding and fixing leaks is often the hardest part of maintaining deep vacuum.

Helium leak detection is the standard method. You flood the outside of the chamber with helium gas, then use a mass spectrometer connected to the pump line to sniff for helium. If helium appears in the pump exhaust, you have found a leak. The rate at which helium enters tells you how large the leak is. This method can find leaks as small as one billionth of a pascal-litre per second — roughly equivalent to a hole that would take a year to let in one molecule.

Leaks usually occur at seals, welds, or feedthroughs where wires or tubes pass through the chamber wall. Copper gaskets, indium seals, and welded joints are the most reliable, but they are also the most expensive and difficult to work with. Elastomer seals like Viton are cheaper but outgas and eventually harden, so they must be replaced regularly.

Materials and components rated for deep vacuum

Not every material that works in air works in deep vacuum. Outgassing rates vary wildly, and some substances will contaminate your experiment or damage your equipment.

Metals are generally safe: stainless steel, titanium, and aluminium all outgas at acceptable rates once baked. Copper and brass outgas more and are avoided in sensitive applications. Ceramics like alumina are excellent — they outgas very little and are chemically inert. Glasses work but must be borosilicate, not soda-lime, because soda-lime leaches sodium into the vacuum.

Elastomers and plastics are the problem children. Viton seals outgas significantly and must be baked and replaced regularly. Neoprene is worse. Teflon (PTFE) is acceptable for static seals but should not be used where it will be flexed. Polyimide (Kapton) is one of the few plastics that performs well in deep vacuum. Most other plastics — PVC, polycarbonate, acrylic — outgas so heavily that they are banned from vacuum systems.

Lubricants and adhesives are almost always a problem. Standard machine oil will evaporate completely. Even vacuum-rated oils outgas. Adhesives like epoxy will release solvents for months. If you must use adhesive, use a vacuum-rated type and bake it thoroughly before sealing the chamber.

Measurement and monitoring in deep vacuum

Measuring pressure in deep vacuum requires instruments that count individual molecules rather than measuring bulk pressure. The most common is the ionization gauge, which ionizes gas molecules with an electron beam and measures the resulting current. The current is proportional to the number of molecules present, so it indicates pressure.

Ionization gauges are sensitive and can measure down to 10 to the minus 12th pascal, but they require calibration and can be damaged by exposure to atmospheric pressure or by certain gases. They also cannot measure the composition of the gas — only the total pressure. If you need to know what gases are present, you need a residual gas analyser, which is essentially a small mass spectrometer that identifies molecules by their mass.

Pressure gauges must be chosen for the pressure range you are working in. A Pirani gauge works down to about 0.1 pascals. A Penning gauge works from 0.01 to 10 pascals. An ionization gauge starts around 0.1 pascals and goes down to the limits of detection. Using the wrong gauge for the pressure range will give you useless readings.

Common problems and why deep vacuum systems fail

The most common failure is slow pressure rise — the vacuum gets worse over time instead of staying stable. This is almost always outgassing from materials or a slow leak. To diagnose it, you bake the chamber again. If pressure improves, the problem was outgassing. If it does not, you have a leak and need to run a helium leak test.

Pump failure is the second most common problem. Turbomolecular pumps are precision instruments with bearings that spin at 60,000 rpm or higher. If they vibrate, overheat, or ingest liquid, they fail catastrophically. Backing pumps can fail from oil degradation or seal wear. Regular maintenance — oil changes, bearing inspection, and vibration monitoring — extends pump life.

Contamination is a third category. If you open the chamber to air and do not bake it afterward, outgassing will prevent you from reaching deep vacuum again. If you pump a substance that condenses at the pump inlet — like water vapour — it will freeze and block the pump. If you use the wrong lubricant or adhesive, it will outgas and contaminate your experiment.

Frequently Asked Questions

How long does it take to reach deep vacuum?

Reaching the threshold of deep vacuum — one millionth of atmospheric pressure — usually takes 24 to 48 hours with a properly sized pump and a clean chamber. Reaching lower pressures and stabilizing the system can take weeks, especially if baking is required. A new or recently opened chamber may take months to stabilize because outgassing is slow.

Can you use a regular air pump to start the vacuum process?

Yes. Any pump that can move air — a rotary vane pump, a scroll pump, or even a diaphragm pump — can serve as the rough pump to bring pressure down from atmospheric to a few pascals. After that, you need a specialized pump like a turbomolecular pump to reach deep vacuum. Using the wrong pump for the wrong stage will either damage the pump or fail to reach the target pressure.

What happens if the chamber leaks while you are running an experiment?

Pressure will rise slowly or quickly depending on the leak size. If you are monitoring pressure continuously, you will see it. If the leak is small, you may have hours to shut down safely. If it is large, pressure will rise to atmospheric in minutes. Either way, air will enter and contaminate whatever is inside the chamber. Preventing leaks through careful assembly and regular testing is far easier than dealing with a leak during an experiment.

Do you have to bake every time you open the chamber?

Yes, if you want to reach deep vacuum again. Opening the chamber to air allows moisture and gases to be absorbed by the walls and materials inside. Baking drives these out. If you do not bake, outgassing will prevent you from reaching the same pressure you had before, and the system may not stabilize for weeks.

Can deep vacuum be maintained indefinitely?

Not without a pump running. The moment you turn off the pump, outgassing and any small leaks will begin raising pressure. A well-sealed, well-baked chamber might hold deep vacuum for hours or days, but eventually pressure will rise. To maintain deep vacuum continuously, the pump must stay on.