Gas analysis solutions for your ethylene oxide process

Ethylene Oxide transcript

MH: Welcome everybody to another Servomex podcast, this time on the theme of ethylene oxide or EO, as you’ll hear us refer to it throughout. I’m joined today by Karen Gargallo, our Application Manager. Hi, Karen.

KG: Hi Matt

MH: And you all know me by now, Matt Halsey, Application Development Manager here at Servomex. So, Karen, we’re going to be talking about EO today. This is one of the most widely produced chemical precursors in the world. It’s used in loads of other stuff. About 75% of ethylene oxide that’s manufactured is used to produce different types of ethylene glycols, which go into a number of useful things that we will come across in everyday life, antifreeze being one of them, that’s a really big one, surface disinfectants and sterilization, things like that. Also the manufacture of PET which is used in plastic bottles, things like that. And other types of coolants and solvents.

It’s very widely produced, as I say, so there’s about 20 to 30 million tonnes produced every year, something like that. And, as it turns out, about 3 million tonnes, so 10% of the global production, is actually produced by one manufacturer in the US, so they have the largest stake. There’s somewhere between 250 and 300 plants worldwide, according to the research that we’ve done.

So, on paper, the reaction that’s required to produce ethylene oxide is actually very simple. It’s the oxidation of ethylene, which as I said, sounds very simple, but there’s a lot more to it. The process is both kind of simple at the same time as being quite complex, especially when it comes to the analysis requirements. There’s a lot of challenges that our customers face and obviously today we’re going to go through the needs for the Servomex analyzers and why the analyzer is so good at making these measurements.

MH: On paper, the ethylene oxide reaction is quite simple. It’s quite literally the oxidation of ethylene. It’s a very exothermic reaction, produces lots of energy and is therefore quite dangerous. So, Karen, could you give us a little lowdown of how the ethylene oxide process actually works?

KG: Thanks, Matt. So ethylene oxide exists as a colorless flammable gas at ambient temperature. If you look at the structure of ethylene oxide, it is a three-member ring structure and one of the atoms is oxygen. So it’s classified as the simplest epoxide. When ethylene oxide is formed, what happens is that single bonds from two adjacent carbons are connected to an oxygen atom and this is then what forms the three-member ring. However, such a three-member ring makes it quite an unstable molecule due to the geometry of the molecule.

However, this is also the same reason why ethylene oxide is very reactive. So it’s a very good component to make other important chemicals, as you mentioned, ethylene oxide, and PET. However, the instability of the molecule makes it highly flammable, toxic, and explosive. That’s why it’s important to make safety measurements across the whole ethylene oxide process.

MH: In a nutshell, what we’re doing in this reaction is we’re taking oxygen, pure oxygen, and we’ll talk about that as a measurement a little later, ethylene, again, a pure feed of ethylene, and we’re mixing them together in a predetermined ratio, feeding them into reactors and the reactors is really where the magic is happening. So, this is where the oxidation reaction is occurring over a very specific pressure, temperature and dwell time in the reactor. So the speed that that mixture passes from the inlet to the outlet of the reactor, and then what we get out is ethylene oxide, which is our product, but we also have some other by-products, which are effectively considered to be contaminants. So what contaminants are we are we getting out of this reaction?

KG: If we go back, one step, looking back again at the ethylene oxide reaction, sometimes called the ethylene epoxidation reaction, depending on the license that is being utilized by specific plants, the oxidation can normally be using either compressed air, or 100% oxygen. I think we can touch on, again, the positive and the negatives of using either air or pure oxygen to make sure that the oxidation occurs. In this case, the feed to the reaction is ethylene. And of course, oxygen is being used for the oxidation. What we have to make sure, though, is that the oxygen concentration is controlled at very specific concentrations, for then not to create an explosive reaction, You’ve already said, this is a highly exothermic process, we want to make sure, of course, that the temperature and the conditions within the reaction are controlled.

MH: That’s a very good point, Karen, mentioning the two types of feed. There’s lots of manufacturers, various manufacturers globally that produce the equipment for the ethylene oxide process, and a lot of them are quite region-specific or country-specific. But I think there’s primarily maybe two main licenses, that develops the methodology behind the ethylene oxide reaction, and we won’t name them in the podcast. But there’s two primary licenses to consider. As you say, they tend to differ in the way that they bring in the oxygen to the process, either through pure oxygen or through air.

The oxygen feed measurement, it’s very important. As you mentioned, it’s there to validate that the mixing nozzles that are in place are actually doing their job and that they’re not providing a mixture that has too little oxygen, in which case you won’t get the throughput from the reactor, or the reaction at all. And also avoiding too much oxygen, which can lead to explosive and hazardous conditions in the reactor, which is the last place that you want lots of oxygen, because of the exothermic nature of the reaction.

So there’s a measurement point here that’s very critical, you have the option feed measurement to start with which, if using 100% oxygen, you’re looking for your oxygen purity, effectively. So ‘have I actually got 100% oxygen?’. So that measurement is relatively simple in terms of oxygen analysis, it’s quite traditional to use a paramagnetic-type analyzer and Servomex there would utilize the OxyExact 2200.

Now the reason that we would use the 2200, for those of you listening that are familiar with our product range, and in previous podcasts, we’ve spoken about the Oxy 1900 as well and they’re quite similar products in many ways, the same technology. One of the main differences between the two is that the 2200 is capable of measuring enriched oxygen samples. So up to 100% oxygen while still being used in a Zone 1 or Div 1 hazardous area location which these plants typically are, they are zoned obviously for safety. So, there aren’t many products on the market from Servomex and competitors that can make such a measurement enriched option in Zone 1 or Div 1, you’re a bit restricted on what you what you choose. But the Servomex 2200 is used because of its high level of reliability. It’s incredibly accurate, it has a quoted accuracy across the entire measurement range of less than 0.02% oxygen. So very, very accurate analyzer highly repeatable, easy to maintain, you know, Paramagnetic technology, the Servomex application requires no reference gas. So, it has very, very minimal ongoing cost of ownership.

So that’s why the 2200 is typically used in that application. You then have a second application of Paramagnetic a little bit further downstream, which is then monitoring the oxygen that’s in the mix. So 10% or thereabouts, I think, is that is the target. And at this point, you could bring in an Oxy 1900 as well. But it would be very common to utilize a 2200. You’ll also see them later on in the reactor stage, which we’ll talk about in a little while, it means that you can just have a single model of analyzer across the whole plant rather than mixing and matching. So it’s quite common to use the 2200 as well. Then of course, the other gas going into this mix is the ethylene.

KG: Yes. So, in the oxidation we do need to oxidize the ethylene. That’s how we make ethylene oxide. And ethylene, of course, can be measured by a photometric analyzer. Ethylene because, again, of its composition has absorption bands in the infrared region. So it’s measured by a an Infrared analyzer like the SERVOTOUGH SpectraExact 2500. Because it’s the feed to the reaction, the range is zero to 100% ethylene and the requirements to make the measurement for controlling the process, you need an analyzer as with the OxyExact 2200, high performance very stable and again, because of the nature of the process, it has to be certified to be able to measure flammable samples.

MH: Yeah, that’s a very good point Karen.

KG: The Servomex SERVOTOUGH SpectraExact 2500 is the flagship photometric, Infrared analyzer of Servomex, utilizing single-beam dual-wavelength technique. With this technique, because of the reference wavelength, we’re able to counter common effects such as drift, source ageing and contamination of the cell.

MH: So the 2500, like a lot of the other Servomex analyzers is, you know, a low cost of ownership unit. So it doesn’t need reference gas or anything like that. It has virtually no consumable parts.

KG: That’s correct. That gives, of course, the analyzer high measurement availability, which is extremely important for the ethylene oxide process.

MH: The next step, after we’ve mixed the oxygen and ethylene, we start getting towards the reactor portion of the process, which is, of course, where the oxidation reaction is happening. Karen, you’ve already spoken a little bit about the reaction itself. What are the challenges here, what are the complications with the ethylene oxide reaction at the reactor stage?

KG: At the reactor stage, we have to be very precise on controlling the oxygen measurement in the process, I think you’ve already touched on it. So the challenge in an ethylene oxide reaction, as you’ve already touched on it, Matt, is that we want to make sure that the oxygen concentration is controlled at a very precise level, we don’t want it to be too low that we do not get the yield that we want from the process. But we don’t want it to be too high to then cause an explosion in the process, or runaway oxidation. The reason that we also want to control the oxygen level very precisely is there is actually another unwanted reaction that takes place. In this case, you can produce CO₂ and water vapor in the process. And we know of course, that CO₂ emissions pose environmental concerns. We want to minimize as much as possible for this unwanted reaction to take place. We also want to prevent is any further oxidation of ethylene to again produce more CO₂ and more water vapor. Ethylene oxide can cause dangerous conditions because it can explode. And it’s both highly flammable, reactive and toxic.

MH: Yeah, there’s been lots of reports over the years, they’re quite well documented reactor failures, reactor explosions at different plants around the world. And, sadly, that’s a very dangerous occurrence. And people have been killed with reactor explosions in the past, and it affects nearby towns and settlements and things like that as well. So, monitoring this reaction is so, so critical.

The residence time of this mix of gases in the reactors is, we’re talking around a second, something like that. It’s an incredibly fast reaction, a runaway reaction, as you described, can happen very, very quickly. What we would do here is we would monitor the outlet of the reactor, we’d monitor the oxygen level post-reactor, and looking for this explosive mix of gases or the explosive level of oxygen.

So once again, you’re looking for an analyzer that is incredibly accurate and reliable, but also very fast. Going back to the 2200, we’re looking at response times, excluding a system, of a number of seconds, four seconds, something like that. The system adds a little bit to that. But obviously, the systems are designed to maximize response time or reduce response time as much as possible depending on which way you want to look at it. And this is where we were able to utilize another one of the 2200’s aces up its sleeve, which is the fact that we can link multiple transmitters to a single control unit. Now why would we want to do that?

There’s a rule on these plants that this oxygen measurement is so critical that there has to be redundancy. And quite often there is a voting system set up. For example, you’d have three 2200 transmitters, which are the units making the measurement with the paramagnetic technology inside them. And it depends on the plant, but typically two of three must agree with each other to tell the plant operators that everything is safe, everything’s fine. If there’s a disagreement in the readings, if it’s severe enough, the plant would actually shut down the reaction, because that’s the safest thing to do, is just shut the reaction down, investigate if there’s a problem, and then get it up and running again and restart.

So this is another really useful feature of the 2200, you can take three, four, up to six transmitters and link them up to a single control unit, which gives the operator a single point of reference where they can go and see all of the diagnostics for all six units, however many there are on one screen and control it via one keypad interface. It just makes life very, very simple for the operators.

The analyzer is also SIL-assessed, it has a functional safety approval or certification to SIL 2, which means you just get this added peace of mind that you’ve got this element of high reliability maximum uptime and of course, the analyzer is capable of monitoring its own diagnostics, and telling you if there is a problem that it knows when there’s something wrong with it, effectively. It displays various faults and alarms and maintenance requirements. So it just gives the operator that extra peace of mind.

When methane is involved in the pure oxygen feed, or when pure oxygen is used as the feed, methane is added as a constituent, the oxygen level that appears in the reactor is very, very low. It’s like sub-1%.

KG: Yes, sub-1% is correct.

MH: Which is actually a relatively safe place to be. There’s not enough oxygen to form an explosive mix. But when you’re pulling oxygen from air, and you’re actually then using the nitrogen as the balance gas, you’re measuring now, you’re seeing oxygen in the 5 to 8% range, which is less than…

KG: Yes, it has to be less than 8%

MH: Which is which is actually a far more, in theory, a more unsafe region of oxygen to be in, because you’re closer to the point where the oxygen could cause an issue. I think the emphasis really there, is just saying that, it’s really just re-emphasizing that the inlet oxygen measurement and the outlet oxygen measurement are just so critical and are measurements that you just can’t go without. In fact, I think you wouldn’t be allowed to go without them on a plant.

KG: So, we now have the ethylene oxide from our process, but it’s not pure ethylene oxide. The next step is then to recover and separate the ethylene oxide from the reactor effluent. The advantage of ethylene oxide, just because of the nature of the component, is that it dissolves in water. In addition with the water ethylene glycol is then added to the ethylene oxide absorbers to prevent any foaming to happen. So, it acts as an anti-foaming agent. So, the ethylene oxide along with the other gases that are present on the reactor effluent are then separated from the water and the ethylene glycol in what we call a stripper, a steam stripper. We’re using steam of course, because ethylene oxide dissolves in water. Then the next step is to distil water and ethylene oxide to separate the ethylene oxide from the steam or the water.

MH: That gives us our pure product.

KG: Yes. So, then we recovered the ethylene oxide and then of course purified and stored it, and that gives us our ethylene oxide to then be used as a precursor to the manufacturing of other important chemicals.

MH: So, on our process diagram, we monitor ethylene at the scrubber, so that’s the bit you were describing, right?

KG: Yes, because ethylene is also produced, is part of the reactor effluent, because we know we don’t produce just ethylene oxide or other gases. So, one of the components that we need to analyze to then determine and control the efficiency of the scrubber is to measure how much ethylene, again, is being scrubbed or removed from the reactor effluent.

MH: So, this is definitely that in theory could have been reacted in the reactors, so therefore shows the operator the efficiency of their reactors.

KG: The typical measurement range for measuring ethylene to look at the efficiency of the scrubber is typically zero to 5% ethylene. In this application, we would again use the SpectraExact 2500. It again provides high performance to control the process efficiently and it gives the fine measurement availability to the customer to ensure that they get an analysis throughout the scrubbing process. CO₂ is produced as part of the reaction and we also need to capture the CO₂ from the reactor effluent because what we want to do is to separate the CO₂ to then be able to recover the ethylene as a recycle loop directly back to the reactor. So we’re adding more ethylene as, you know, of course, ethylene is used as part of the reaction, so what we’re trying to do actually is to have additional ethylene as a recycle to our ethylene oxide reactor.

MH: So, once again here, this recycled ethylene that we’re pulling from, this is ethylene that hasn’t reacted, it’s been scrubbed back out, we can reuse it, it’s a useful product. This increases overall process efficiency, of course, if we’re reusing waste, and once again, here we’re seeing that the 2500 is used to monitor this recycled ethylene.

KG: That’s correct. And I’ve already mentioned that CO₂ has an adverse effect to the reaction, what we want to do is to make sure that the CO₂ is not at a high concentration, that it can then have a negative impact on the activity and the selectivity of the catalysts. So there’s an adverse effect of that, if you have high CO₂, it can actually damage the catalyst. As we know, as with any catalytic reaction, the catalyst plays a very important role in controlling the temperature, the residence time pressure, all those conditions in the reaction.

MH: And, of course, is incredibly expensive to replace, if it breaks.

KG: Yes, and how we then recover CO₂ is to use aqueous form of potassium carbonate because again it absorbs with CO₂, the reactions form then some potassium bicarbonate and this is then the way we ensure that the CO₂ concentration is not going to have an adverse effect on the catalyst activity.

MH: All these analyzers, of course, typically require a dry measurement to be made, or at least a non-condensing measurement, to be more precise. We have to utilize our systems expertise with a lot of these measurements. The paramagnetic certainly would require some kind of sampling system.

KG: Sample system again is key to make sure that the pressure and temperature are controlled.

MH: Yes.

KG: The flow is important, too.

MH: All our analyzers will require typically some kind of sample system, even in a very simplistic form, to control pressure and flow into the analyzer; most of these analyzers are flow-driven. So we would call on our Servomex systems expertise and Servomex have been designing and manufacturing ethylene oxide systems for decades. Servomex are, in many ethylene oxide applications, the preferred supplier, if not the sole supplier in some cases, due to our experience in these applications. So, yeah, we would call upon our systems expertise, and we can design bespoke systems for our customers, to fit in between the process and our analysis equipment, that meets any local or any particular requirements that our customers may have.

As uptime is so important with these analyzers, we would strongly recommend that customers make use of the Servomex service network for regular health checks and maintenance visits, picking up any issues with analysis equipment before something actually was to go wrong and fail, which in many of these measurements would ultimately lead to a shutdown of the process for a length of time, which is obviously not something that anybody wants.

MH: Thank you once again, everybody, for listening to another Servomex podcast, and thank you Karen, today, for your contributions.

KG: Thank you, Matt, for inviting me to join you in this podcast.

MH: Don’t forget we’ve got a load of podcasts now that you can go and listen to. Please do visit servomex.com, to find out more about ethylene oxide solutions. On the website, you’ll find videos, you’ll find links to literature. So, please do take a look. Thank you once again and see you next time.

KG: Thank you.