Industrial Vacuum Pumps Types Applications and Compressed Air Integration

Atlas Copco manufactures the GHS VSD+ dry screw vacuum pump and the ZR oil-free screw compressor on production lines that share tooling. The bearing specifications cross over between the two products. A maintenance technician trained on one can rebuild the other. The purchasing happens through separate channels because the vacuum pump goes on the process equipment budget and the compressor goes on the utilities budget, and by the time both are installed in the same plant, the mechanical commonality between them is invisible to everyone except the person ordering spare parts.

That purchasing separation has consequences for cooling systems, heat recovery, piping, leak management, and controls. Vacuum pumps and compressors do the same thermodynamic work. The vacuum pump usually does it at a harder compression ratio. At 100 mbar inlet pressure exhausting to atmosphere, the compression ratio is about 10:1. A plant air compressor at 7 bar gauge works at roughly 8:1. Below 10 mbar the vacuum side hits ratios above 100:1, and the energy cost per molecule removed starts climbing on a steep exponential.

01Cooling Water and Liquid Ring Pumps

Vapor Pressure Limits

A liquid ring pump's vacuum depth depends on the vapor pressure of whatever liquid forms the ring. With water at 15°C the limit sits around 17 mbar. At 30°C it jumps to 42 mbar. At 40°C, 74 mbar.

Those numbers by themselves do not convey what happens to the pumping speed curve. Near the vapor pressure limit, pumping speed does not taper off. It drops sharply because vapor recirculating inside the pump displaces gas that should be getting evacuated. A pump rated 500 m³/h at 100 mbar with 15°C sealing water can fall to 200 m³/h with 30°C water at the same suction pressure.

Nash built the largest installed base of liquid ring pumps in petroleum refinery vacuum distillation. Their application engineers ask about sealing water temperature first because that resolves most performance calls.

The cooling tower serving a liquid ring pump also serves compressed air aftercoolers. When summer conditions push cooling water return temperature up, the aftercoolers produce wetter compressed air with higher pressure dew points at the same time the liquid ring pump loses vacuum capacity. One shared thermal constraint degrades two utility systems simultaneously. Preventing it requires sizing both to summer design-day conditions during the original engineering, which means the process engineer specifying the vacuum pump and the utilities engineer specifying the compressed air cooling need to coordinate. In most projects they do not overlap.

Liquid ring pumps use a lot of electricity for the vacuum they produce. Maintaining the water ring against gravity and viscous drag uses power that does not compress gas. At moderate vacuum a liquid ring pump draws close to double what a dry screw draws for the same capacity. Insurance underwriters in chemical and petrochemical plants often mandate liquid ring for flammable solvent and explosive atmosphere service. Discharge temperature stays within about 12°C of the sealing liquid because of the isothermal compression, eliminating ignition risk from hot surfaces.

02Temperature Inside Dry Screw Vacuum Pumps

Dry screw pumps have taken more market share in the past twenty years than any other vacuum pump type, moving from semiconductor fabs into pharmaceutical, chemical, and food processing. What drives most of the application engineering difficulty is temperature.

Rotor tip clearances at cold assembly sit around 0.04 to 0.06 mm depending on manufacturer and pump size. Compression work concentrates at the discharge end because that is where gas reaches its highest pressure. The discharge-end rotor lobes run well over 100°C hotter than the inlet-end lobes during deep vacuum operation. Thermal expansion along the rotor is non-uniform. The clearances that were set cold must survive this gradient without closing to zero at operating temperature.

Rotor profiles are proprietary and computationally optimized per manufacturer. They cannot be sourced aftermarket. A replacement rotor set can cost a third of the pump's original price with several weeks of OEM lead time.

If the hot end of the rotor expands too far, the rotor contacts the housing and the pump seizes. New rotor set, possible housing rework, weeks of downtime. This failure mode constrains how hot the pump can run. The other constraint pushes in the opposite direction.

In chemical and pharmaceutical vacuum drying, the process gas contains air, water vapor, and solvent vapor. Every point along the compression path inside the pump has a local pressure somewhere between the inlet vacuum and atmospheric. If the gas temperature at any of those points drops below the dew point of any vapor component at that local pressure, liquid forms on the rotor surfaces.

What condenses determines the damage. Acidic condensate eats into the rotor coating. Monomers like styrene polymerize into solid residue that builds up over weeks until the rotors seize. Water accumulation can hydraulically lock the pump during restart.

The dew point calculation demands complete gas composition. Not approximate, not "major components only." A 2 percent trace solvent left off the process data sheet can shift the calculated dew point at discharge conditions by 20°C. Pharmaceutical drying applications with polar non-ideal mixtures need NRTL or UNIQUAC thermodynamic models. Peng-Robinson handles hydrocarbons. Using the wrong model or incomplete composition data produces a dew point answer that looks plausible. The pump runs for weeks or months before polymer or condensate accumulation causes a seizure that gets written up as a mechanical failure.

For condensable service the cooling water has to be restricted to keep pump internals above the dew point along the entire compression path. The machine runs deliberately hotter than the cooling system could make it. Maintenance crews dislike this because they are being asked to restrict cooling on a machine with clearances measured in hundredths of a millimeter, and they are right to be cautious. The temperature window between "too cold, condensation occurs" and "too hot, thermal expansion closes the clearances" can be narrow. Staying inside it requires temperature sensors at multiple housing locations, cooling water valves with PID control, and alarm setpoints on both limits. A manually set valve will drift with seasons and cooling water conditions.

Heat Recovery

Discharge gas temperature from dry screw pumps runs 150 to 200°C, depending on vacuum depth and gas composition. That is substantially hotter than the 80 to 90°C coming off a compressed air compressor's oil cooler. The temperature difference opens a specific application: desiccant dryer regeneration. A heatless desiccant dryer purges 15 to 20 percent of its rated airflow to strip moisture from the desiccant bed. That purge is compressed air produced at full cost and then thrown away. Routing vacuum pump discharge heat into the dryer regeneration circuit drops the purge requirement to 3 to 5 percent. Parker and Beko both build heated desiccant dryers that accept external heat sources. A heat exchanger, some piping, and a control valve connect the two systems. In new construction projects where both get specified together, this connection sometimes gets made. In retrofit situations the dryer and the vacuum pump came through separate capital projects, and connecting them means reopening a closed specification, which takes more organizational energy than the engineering warrants.

A different compressed air connection shows up in semiconductor fabrication. Load-lock chambers vent to atmosphere between wafer transfers using clean dry air. SEMI F6 specification calls for below 1 ppm moisture. If the CDA system delivers above that, water molecules deposit on chamber walls every vent cycle and have to be pumped out during the next evacuation. Over thousands of daily cycles the effect accumulates into measurably longer pump-down times and higher base pressures. The degradation looks like the vacuum system aging. Checking CDA moisture at the point of use rather than at the dryer outlet is the diagnostic step that usually resolves it. The CDA system is compressed air infrastructure. The symptom presents in the vacuum system. The two are connected through the load-lock vent valve, and that connection spans two utility budgets and two maintenance teams.

03Rotary Vane

Oil-sealed, eccentric rotor, sliding vanes. Atmospheric to 0.5 mbar single-stage.

The discharge oil separator uses a coalescent element, borosilicate microfiber, identical technology to what sits inside oil-injected screw compressors. Compressor rooms have differential pressure monitoring on the separator and replace on condition. Walk out to the production floor and look at the vacuum pump's separator. Usually no gauge. Identical element, identical degradation, and it gets changed when it starts smoking or when a calendar reminder fires, whichever comes first.

Backstreaming oil vapor toward the process chamber worsens as the pump oil thermally degrades over its service life. In clean-chamber applications, the resulting hydrocarbon contamination drifts upward slowly enough to be attributed to other process variables for months before someone analyzes the chamber wall residue and traces it back to the pump.

Losing share to claw pumps in packaging. Holding everywhere else.

04Roots Boosters

Roots boosters do not get specified often enough, given what they deliver per dollar and per kW.

A 100 m³/h backing pump delivering 40 m³/h effective at 10 mbar (internal leakage consumes displacement at low pressures) gets a 500 m³/h Roots booster upstream. The booster compresses from 10 mbar to about 50 mbar. The backing pump recovers to 80 m³/h at that pressure. System effective speed at the chamber reaches around 400 m³/h for an additional few kW on the booster motor. Hard to find a better ratio of performance gain to added cost anywhere in vacuum system design.

Roots machines have no internal compression, just displacement. Two figure-eight lobes trap gas and carry it from inlet to discharge. Compression happens by backflow from the discharge plenum when the trapped pocket opens to the discharge port. Loud, and not great as a standalone vacuum pump. Put one in front of a decent backing pump and the system performance changes by an order of magnitude.

Aerzen has been manufacturing Roots-type machines since 1868. They build vacuum boosters and positive-pressure blowers for wastewater aeration and pneumatic conveying on an identical mechanical platform. The rotors, casings, timing gears, and bearings interchange between vacuum and pressure versions. A plant that runs Aerzen Roots boosters on vacuum and Aerzen blowers on process air can maintain one spare parts inventory for both services.

05Claw Pumps in Packaging

Busch Mink dominates European packaging vacuum. Packaging machines cycle between vacuum and vent thousands of times per shift. Rotary vane pumps in this service lose pumping speed progressively as shock loading at each pressure transition wears the vane tips. Lines compensate by extending vacuum dwell time, trading throughput. The loss accumulates slowly enough to be blamed on product variability before anyone pulls the pump apart.

Claw rotors never contact the housing. Five-year performance holds close to commissioning curves. The capital premium over rotary vane, something like 40 to 60 percent depending on size, pays back through avoided vane replacement and recovered throughput. For plants running 16 or more hours a day, claw is the default choice in new packaging installations.

06Venturi Generators and Compressed Air Demand

Piab COAX multi-stage ejectors deliver about 30 percent better efficiency than single-stage Venturi generators. Still only 5 to 15 percent of the compressed air energy converts to useful vacuum work.

System Impact

One ejector on a robot gripper is negligible. Fifty on a packaging line can demand in the neighborhood of 200 kW worth of compressor capacity. If the compressed air system was sized without this load, header pressure sags during production and the compressor room raises the setpoint. Each bar of setpoint increase costs about 7 percent more energy across the entire system. Fast-cycling pick-and-place machines also create pressure pulsations in the compressed air header at 20 to 60 cycles per minute. Without a local receiver to buffer these pulses, the pressure oscillation degrades Venturi performance and affects other pneumatic consumers on the same header branch.

Below 20 percent duty cycle, Venturi wins on millisecond response and zero standby draw. Above 50 percent sustained duty, an electric pump is far cheaper. Fifty ejectors running at high duty cost roughly six times what a 7.5 kW claw pump costs in electricity for comparable vacuum output.

07Piping, Leaks, Central Vacuum, Heat Recovery

Gas density changes by orders of magnitude across the vacuum range. A pipe adequate at 500 mbar becomes undersized at 50 mbar, where volumetric flow for the same mass flow is ten times larger. The Hagen-Poiseuille conductance equation contains a mean-pressure term that makes conductance pressure-dependent in a way that has no parallel in compressed air pipe sizing. Vacuum pipe sizing requires iterative calculation at the operating pressure. Most undersized vacuum piping got sized from the pump's catalogue m³/h treated as a constant-density number. At 100 mbar the gas volume is ten times atmospheric. At 10 mbar, a hundred times. The undersizing shows up as elevated chamber pressure and gets misread as a pump or leak problem.

Conductance through a long tube in viscous flow goes as the fourth power of diameter. Dropping from 50 mm ID to 40 mm ID, which looks like a minor change on a drawing, cuts conductance by more than 59 percent. Going up one standard pipe size during installation adds negligible cost. Correcting the diameter later means draining the vacuum system, cutting and re-welding, and losing production time. The cost asymmetry between getting it right during construction and fixing it afterward is larger for vacuum piping than for most other utility piping, because vacuum lines tend to run through congested areas near process equipment where access is poor and the pipe itself may need to thread between existing services.

Ultrasonic leak detectors work on vacuum and compressed air lines interchangeably. A combined survey covering both saves mobilization cost. Vacuum leaks draw atmospheric air into the process space, carrying moisture, oxygen, and particulates. In food packaging and semiconductor fabrication, a vacuum leak is a contamination event whose product quality cost can dwarf the energy cost.

Compressed air centralized between the 1960s and 1990s. Vacuum still runs decentralized in most plants, with point-of-use pumps sized for individual machine peak demands and diversity factors of 0.5 to 0.7 across a multi-machine floor. Central stations capture that diversity at roughly a third less total motor power. Co-locating with the compressor room shares switchgear, cooling, and maintenance crews. For facilities using screw-type machines on both sides, the bearing and seal commonality between vacuum pumps and compressors becomes a concrete spare parts reduction.

Scroll pumps are confined to laboratory and analytical work by their displacement ceiling of about 35 to 50 m³/h. Edwards nXDS is the standard product. No meaningful compressed air interface.

Combined Heat Recovery

Combining compressor and vacuum pump waste heat changes recovery project economics. A 75 kW compressor rejects about 70 kW at 80 to 90°C. A 37 kW dry screw vacuum pump rejects about 35 kW at 150 to 200°C. Over 8000 hours that totals 840,000 kWh. Each stream alone may not clear a capital approval threshold for a heat exchanger system, but the combined total usually does. The vacuum pump heat, being hotter, serves desiccant dryer regeneration, process water heating above 90°C, and low-pressure steam generation.