Xylene isomerization and the recycle loop
Separation gets the attention, but the isomerization loop is where a para-xylene plant is won or lost. It is what turns a 24 % para-xylene equilibrium into near-complete conversion — and it is where the losses accumulate.
Updated
Why there is a loop at all
Take mixed xylenes at equilibrium and remove the para-xylene. What is left is a stream of meta- and ortho-xylene that has no large market of its own and is chemically almost identical to what you wanted. Discarding it would mean throwing away three quarters of the feed.
Instead it is passed over an isomerization catalyst, which re-scrambles the methyl group positions until the mixture returns toward its thermodynamic equilibrium — roughly 24 % para, 52 % meta, 24 % ortho. Fresh para-xylene has been created out of the depleted stream. Send it back to separation, take the para-xylene out again, and repeat.
Run this loop enough times and almost all of the xylenes fed to the complex eventually leave as para-xylene. The limit is not the chemistry — equilibrium always regenerates para-xylene — but the losses taken on each trip around.
Approach to equilibrium
A real reactor does not quite reach the equilibrium distribution. How close it gets is reported as the approach to equilibrium: the para-xylene actually produced, divided by the para-xylene that full equilibration would have produced, as a percentage.
A higher approach means more para-xylene per pass, which means fewer passes, which means less circulation through every column in the loop. Pushing it, though, generally means more severe conditions — and severity is what destroys rings. The approach to equilibrium is therefore never optimised alone; it is traded against xylene loss.
The ethylbenzene fork
Ethylbenzene arrives with the xylenes and cannot be left to accumulate in the loop. Distilling it out is possible but brutally expensive — it boils only 2.2 °C below para-xylene. So the isomerization catalyst is asked to deal with it, and there are two schools:
| EB isomerization | EB dealkylation | |
|---|---|---|
| What happens to EB | Converted into xylenes via naphthene intermediates | Stripped of its ethyl group → benzene + ethane |
| Xylene yield | Higher — the ring is kept | Lower — the EB ring leaves as benzene |
| By-product | Little | Benzene (a saleable product) |
| Per-pass EB conversion | Moderate | High |
| Effect on the loop | More circulation, EB converts slowly | Less circulation, cleaner recycle stream |
| Hydrogen | Required | Required |
Neither is simply better. Keeping the ring preserves feedstock you have already paid for; dealkylating it gives a cleaner, smaller recycle and a benzene credit. Which wins depends on the relative prices of para-xylene, benzene and naphtha — which is exactly the kind of question a process model is for.
Where the losses go
Every pass through the reactor destroys some aromatic rings. The usual routes out:
- To toluene — demethylation, taking one methyl group off a xylene. Recovered and sold, or recycled to transalkylation.
- To C9+ heavies — methyl transfer producing trimethylbenzenes and heavier. Partly recoverable by transalkylation, partly a bottoms product.
- To light ends — cracking to C1–C5 gas, essentially fuel value only.
- To naphthenes and paraffins — ring saturation. These non-aromatics accumulate in the loop and must be purged, carrying aromatics out with them.
Why a small per-pass loss is not a small loss
This is the point most often missed. A 2 % xylene loss per pass sounds negligible. But a molecule of meta-xylene does not go around once — it circulates until it happens to emerge as para-xylene, which takes several passes on average. It is exposed to that 2 % loss every single time.
The same multiplication works in reverse and produces genuinely counter-intuitive results. Increasing ethylbenzene conversion, for example, does not simply increase output: it reduces how many times each molecule circulates, which can *reduce* the total cumulative flow of by-products even though the conversion per pass went up. Effects like this are why the loop has to be solved numerically rather than reasoned about a stage at a time.
What the loop costs
Because everything in the loop is sized for the circulating rate rather than the fresh feed, the recycle ratio is the multiplier on most of the plant:
- Xylene column and raffinate column reboiler duty
- Separation unit size — crystallizer surface or adsorbent inventory
- Reactor volume and catalyst inventory
- Recycle compression and pumping
- Hydrogen consumption and purge losses
Anything that raises per-pass para-xylene recovery or approach to equilibrium shrinks all of these at once. That is why the separation and isomerization sections cannot be designed independently — they are two halves of one loop.
Frequently asked questions
What is the equilibrium composition of xylenes?
Approximately 24 % para-xylene, 52 % meta-xylene and 24 % ortho-xylene, varying somewhat with temperature. This distribution is the reason para-xylene plants recycle: no single pass can do better than about a quarter.
What does approach to equilibrium mean?
The para-xylene actually produced by the reactor as a percentage of what complete equilibration would produce. Higher is better for the loop, but it is usually bought with more severe conditions, which increases ring loss.
Should ethylbenzene be isomerized or dealkylated?
Isomerization keeps the ring and gives a higher xylene yield; dealkylation converts it to benzene, giving a cleaner recycle, less circulation and a benzene credit. The right choice depends on the relative values of para-xylene, benzene and feedstock.
How much of the feed is lost around the loop?
Per-pass xylene losses are typically 1–3 %, but they compound with recirculation. With two to five passes per molecule, the cumulative loss on fresh feed is several times the per-pass figure.
Why do non-aromatics have to be purged?
Naphthenes and paraffins formed by ring saturation are not converted back to aromatics and do not leave with the product, so they accumulate in the recycle. A purge is needed to hold them at a steady concentration, and that purge inevitably carries aromatics out with it.
Change the approach to equilibrium, the per-pass loss or the ethylbenzene conversion and see the whole loop respond — circulation, by-products, utilities and margin.
Related reading
How para-xylene is made from mixed xylenes: crystallization or adsorption, the isomerization recycle loop, yields, losses and the economics.
The two para-xylene separation routes compared: how each works, per-pass recovery, purity, energy, consumables — and which feed makes each cheaper.
Where the money goes in a para-xylene plant: the PX–naphtha spread, refrigeration and reboiler duty, the recycle multiplier, ring losses and by-product credits.