As a PSA Gas Generator supplier with over 15 years of hands-on experience troubleshooting, optimizing, and delivering these systems to industries ranging from food and beverage to pharmaceuticals and manufacturing, I’ve learned that the single most common mistake new operators make is treating their PSA unit as a “set-it-and-forget-it” device. Too often, a customer will reach out after six months of operation, frustrated because their nitrogen purity is 2% lower than promised, or their oxygen output is fluctuating wildly—only to find the root cause is not a faulty generator, but a mismatch between the inlet gas composition they feed and the unit’s design parameters. PSA (Pressure Swing Adsorption) technology relies entirely on the selective adsorption of target gases (or impurities) by molecular sieve materials, and every component of that process—from adsorbent lifespan to purity output to energy efficiency—hinges on how closely the inlet gas composition aligns with what the unit was engineered to handle. PSA Gas Generator

Let’s start with the basics, for those who might not be as familiar with PSA systems. Our PSA nitrogen generators, for example, work by forcing compressed air (the inlet feed) through two vessels filled with carbon molecular sieve (CMS). Oxygen and trace impurities adsorb onto the CMS at high pressure, leaving nitrogen to pass through as product. Once the adsorption vessel is saturated, we drop the pressure to release the adsorbed gases (a process called regeneration), and the cycle repeats between the two vessels. For PSA oxygen generators, it’s similar, but we use zeolite sieve to adsorb nitrogen from compressed air, leaving high-purity oxygen as output. The adsorbent is the heart of the unit, and its performance is not static—it shifts dramatically based on the composition of the gas entering the system.
First, let’s talk about the impact of variable nitrogen or oxygen content in the inlet gas. Most of our standard PSA generators are designed for a dry, clean inlet gas with a fixed composition of ~78% nitrogen, ~21% oxygen, and ~1% argon and other trace gases. If a customer feeds in air that has a higher nitrogen content—say, they’re recycling off-gas from a process that’s already 85% nitrogen, and they want to recover it with our PSA unit—this is actually a benefit. The higher nitrogen concentration means less CMS is needed to separate it from oxygen, so the unit can run more efficiently, with lower pressure drop and longer adsorbent lifespan. But here’s the catch: if the inlet gas has a lower nitrogen content (e.g., only 70% nitrogen, which might happen if you’re pulling air from a poorly ventilated confined space with higher CO2 levels), that’s a problem. The CMS is designed to handle a specific ratio of impurities to target gas. With more non-nitrogen gases (oxygen, CO2) in the feed, the adsorbent gets saturated faster. We’ve seen customers run their units at 90% capacity for a year without issues, then drop to 60% because their inlet composition shifted and they didn’t adjust the cycle time or maintenance schedule. In one case a few years back, a metal plating company was feeding their PSA nitrogen generator with vent gas from their electroplating tanks, which had only 65% nitrogen—they didn’t tell us upfront, so the adsorbent was fully saturated in just 18 months instead of the expected 5 years. We had to replace the sieve early, and that cost them $12,000 in downtime and parts. That’s the kind of avoidable hit we work to prevent by educating customers on their inlet gas composition before we even design a unit.
Next, impurities in the inlet gas—water vapor, oil, CO2, and hydrocarbons—are maybe the biggest challenge when it comes to gas composition. PSA adsorbents (both CMS for nitrogen and zeolite for oxygen) are incredibly porous, which is what makes them effective at separating gases, but that same porosity makes them prone to contamination. Let’s break this down: water vapor in the inlet gas is a silent killer. Even trace amounts of moisture will adsorb onto the sieve’s pores, blocking the sites that should be adsorbing oxygen (for nitrogen generators) or nitrogen (for oxygen generators). We once had a craft brewery customer who skipped the aftercooler and dryer on their compressed air line, feeding raw, moist air into their PSA nitrogen unit. Within three months, their oxygen adsorption rate dropped by 12%, so their nitrogen purity went from 99.9% to 97.2%—a huge issue for their beer packaging line, which required nitrogen to prevent oxidation. When we inspected the sieve, it was caked with a thin layer of water, rendering most of the surface area useless. We had to reactivate (reheat) the sieve, but even that only recovered 70% of its original performance; they still needed a full dryer upgrade to prevent future issues. Oil is even worse—compressor oil vapor in the inlet gas will coat the sieve like plastic, sealing the pores completely. That’s not something you can reactivate; you have to replace the sieve, which is a major expense. CO2 and hydrocarbons, meanwhile, are also adsorbed by the sieve, but they take up more space than oxygen or nitrogen. In a natural gas feed (for PSA hydrogen generators, another common use case), CO2 and methane in the inlet gas can compete with the target hydrogen for adsorption sites. If the inlet gas has a CO2 content of more than 0.5%, standard CMS will start adsorbing CO2 instead of the impurities we want, leading to lower hydrogen purity. We had a refinery customer who ran their PSA hydrogen generator for two years without a pre-treatment filter for CO2, and their product purity dropped from 99.99% to 98.5%—which was enough to ruin their catalytic reforming process, costing them $50,000 a day in lost production. That’s why we always test a sample of a customer’s inlet gas first, before finalizing a PSA unit design, to add pre-treatment steps (like dryers, filters, or pre-adsorbers) if needed to remove harmful impurities.
Another often-overlooked factor is argon content in the inlet gas. For most standard PSA nitrogen generators, argon is a trace impurity (about 0.93% in dry air) that doesn’t affect performance much, because the CMS we use adsorbs oxygen far faster than argon. But if a customer is feeding a gas with higher argon content—like a glass manufacturing plant that uses argon in their production and recycles the off-gas—this becomes a problem. Argon has a similar molecular size to nitrogen, so the CMS can’t separate it effectively. That means the product nitrogen will end up with a higher argon concentration, which can be an issue for applications like semiconductor manufacturing, where even 0.1% argon can contaminate wafers. For that scenario, we design custom PSA units with a two-stage process: first, separate oxygen and CO2, then use a second stage with a different sieve (or adjust the cycle pressure) to reduce argon levels. But if we didn’t test the inlet gas for argon upfront, we’d deliver a standard unit that would fail the customer’s purity requirements within weeks. We saw this with a solar panel manufacturer a few years ago; they didn’t mention their argon recycling stream, so our standard unit delivered 99.5% nitrogen, but they needed 99.999% (5-nines) for their cell production. We had to retrofit their unit with a two-stage system, which saved their production but cost them more than a custom unit would have if we’d known about the argon upfront.
Now, let’s talk about how gas composition impacts energy efficiency. PSA units run on compressed air, which is one of their biggest operating costs. The more we have to adjust pressure or cycle time to compensate for an off-spec inlet gas, the more energy the unit uses. For example, if a customer’s inlet gas has a higher oxygen content than designed, the adsorption vessel has to run at a higher pressure to ensure all the oxygen is removed—this can increase energy use by 10-15% annually. If the inlet has water vapor, the unit has to work harder to regenerate the sieve, using more of the product gas to purge the vessels, which reduces the overall gas yield. We calculated for a 500 CFM PSA nitrogen generator, running 24/7, a 10% increase in energy use adds about $12,000 a year in electricity costs. That’s a number that doesn’t show up in the initial quote, but it’s a huge long-term expense if you don’t account for inlet gas composition. On the flip side, when the inlet gas is within design specs, our PSA units have an average energy efficiency of 0.3 kWh per cubic meter of nitrogen—something we can’t guarantee if the feed is variable or contaminated.
I want to be clear here: this isn’t to say that your inlet gas has to be perfectly pure or exactly 78% nitrogen. We’ve worked with customers on custom solutions for almost every scenario: off-gas recovery, variable composition feeds, even industrial exhaust gases that need to be purified. The key is transparency. When a customer comes to us and says, “My inlet gas is 75% nitrogen, 22% oxygen, 2% CO2, and it has 0.1% oil vapor,” we don’t just quote a standard unit. We run a series of tests on our in-house lab sieve samples, design a pre-treatment system to remove the CO2 and oil, adjust the cycle time to match the lower nitrogen content, and factor in a slight reduction in lifespan to give them an accurate cost breakdown. That’s the value we bring as a supplier—not just a unit, but a partner who understands how the details of your feed gas impact your operation.
Let’s also address a common myth: some operators think that they can “fix” a bad inlet gas composition by adjusting the PSA unit’s settings on their own. In my experience, that almost never works. We had a food processing customer who tried to lower the adsorption pressure to reduce energy costs, not realizing their inlet gas had a higher moisture content. They ended up with a unit that wouldn’t reach the desired nitrogen purity, because the lower pressure meant the sieve couldn’t effectively adsorb the oxygen, and the moisture already present was blocking the pores. It took our technician three days to recalibrate the unit, replace the desiccant in the dryer, and reset the cycle times. So many of these issues are avoidable with proper understanding of the connection between feed gas and PSA performance.
Looking ahead, as more industries move towards circular economy models—recovering and reusing off-gases instead of venting them—this conversation about gas composition will only become more important. Customers who want to use PSA units to recycle process off-gases can save thousands in raw material costs, but they need to work with a supplier who can test their specific feed and design a system that works with that variable composition. For example, a chemical plant that used to vent 10,000 CFM of off-gas (which was 60% hydrogen, 35% methane, and 5% CO2) now uses our custom PSA hydrogen generator to recover 99.9% pure hydrogen, cutting their raw hydrogen purchases by 70%. But that unit only works because we tested the off-gas composition, added a pre-treatment to remove methane, and adjusted the PSA’s cycle to handle the non-standard CO2 level. Without that upfront analysis, that project would have failed.

At the end of the day, a PSA Gas Generator is only as good as the gas you put into it. We’ve seen great units ruined by a lack of attention to feed composition, and good units turned into high-performing, cost-saving assets when the feed is accounted for correctly. If you’re considering a PSA Gas Generator, or if your current unit is underperforming and you’re not sure why, I’d encourage you to reach out to our team. We can run a free gas analysis test, walk through your current feed parameters, and help you design a system that meets your purity, efficiency, and budget needs. Don’t let a small detail like gas composition derail your process—let’s make sure your PSA unit works for you, not against you.
Nitrogen Generator References:
- Ruthven, D. M. (1997). Principles of Adsorption and Adsorption Processes. John Wiley & Sons.
- Yang, R. T. (1987). Gas Separation by Adsorption Processes. Butterworth Publishers.
- Jabbour, K., et al. (2018). “Pressure Swing Adsorption for Gas Separation: A Review.” Separation and Purification Technology, Vol. 194, pp. 186-203.
- International Society of Pharmaceutical Engineers (ISPE). (2020). “PSA Gas Generators in Pharmaceutical Manufacturing: Feed Gas Considerations.” Technical Report.
- Compressed Air and Gas Institute (CAGI). (2019). “Guidelines for PSA Generator Feed Purity and Preparation.” CAGI Publication 100-19.
Zhejiang Yuanda Air Separation Equipment Co., Ltd.
Zhejiang Yuanda Air Separation Equipment Co., Ltd. is one of the top level psa gas generator manufacturers and suppliers in China. If you are planning to buy psa gas generator from professional factory and seller, please feel free to contact us.
Address: No.300 Gushan Ave, Chun’an County, Hangzhou,Zhejiang, China.
E-mail: sales@ydget.com
WebSite: https://www.ydget.com/