Molecular sieves

Jan. 7, 2008
Knowing the plants of Saudi Aramco well and being in charge of product management of CECA SILIPORITE molecular sieves, I would like to comment on the article by Ahmed Ghazal (“Common errors can cause mol-sieve desiccant deterioration”; OGJ, Nov. 26, 2007, p. 62) to prevent misunderstanding.

Knowing the plants of Saudi Aramco well and being in charge of product management of CECA SILIPORITE molecular sieves, I would like to comment on the article by Ahmed Ghazal (“Common errors can cause mol-sieve desiccant deterioration”; OGJ, Nov. 26, 2007, p. 62) to prevent misunderstanding.

Molecular sieves remove dissolved water (speaking about liquids) but should not be in contact with entrained water, as the article relates. Entrained water will saturate them quite fast, given the likely huge amount of water entrained, may lead to local temperature peaks leading to faster aging, and may bring all kind of “aggressive” molecules (e.g., amines) with it, destroying either the binder or the zeolite (active material of the molecular sieve) or both.

In our more than 40 years’ experience, we cannot confirm the fact of ion leaching due to use of molecular sieves of different suppliers. Different suppliers would mean different types of zeolite and different binders. Indeed a single supplier may have to supply molecular sieves for natural sweetening (mercaptan removal), which would include different types (4A, 5A, 13X sieves) fabricated with different binders. No problems of ion leaching due to the composition would be noticed.

Moreover, the problems observed in YGP might be due to liquid carryover, causing chemical attack or simply hydrothermal damaging due to a poor regeneration procedure. Poor regeneration procedure means in this case heating too fast (no intermediate heating ramp) and causing water desorption from the layers at the inlet of heating gas and condensation of water at the still-cold layers at the opposite site of the bed.

This liquid water together with the high regeneration temperature will lead to boiling of the sieves in the liquid water, causing hydrothermal damaging. Hydrothermal damaging means attack of the binder, destruction of the crystal structure of the zeolite, or pore closure of the zeolite.

I agree with the author that it is better to use only fresh molecular sieve in order to get the process guarantee from the supplier, ensure the life time of the sieves (the costs of some days of production loss due to premature shutdown of the plant is huge compared with the price of a load of sieves), and start with low pressure drop as unloading and screening of the used sieves will decrease the average particle size leading to higher pressure drop.

It is strange, however, to see large beads (1⁄8-in. equivalent or 2.5-5 mm beads or 3.2 mm pellets) in a liquid dryer. Usually small particle size—1⁄16 in. equivalent—is recommended as the mass transfer is much shorter. For 1⁄8-in. particle, easily half of the vessel might be used for the mass-transfer zone. The mass-transfer zone is used to meet the specifications of the product—in this case 10 ppm (wt)—but does not adsorb a lot of water.

Using large particles in this case shows that the initial design is poor and short. This could be the case if fluidization should be prevented. This poor design might also be the origin of molecular sieve failure.

Peter Meyer
Business Manager Dynamic Applications
Molecular Sieves CECA SA

Ghazal replies

The occurrence of ion leaching at the mentioned site was verified by Aramco’s Research and Development Center through internal research and a series of lab tests.

Data obtained from subject study revealed a clear relationship between the instigation of ion leaching and presence of dissimilar desiccant materials, in the same vessel, that are supplied by different vendors and consist of different binder additives. Chemical attacks, poor regeneration procedures, and application of inaccurate loading configurations are all typical causes of desiccant failure and widely known among operating facilities.

The fact of the matter is that all troubleshooting attempts start by investigating the more obvious factors stated by Mr. Meyer and a rather vigorous study is initiated once the aforementioned bases are covered.

The highlighted desiccant particle sizes do not represent existing setup at Yanbu gas plant; the information was merely mentioned in my article to describe proper loading configuration for different layers within typical dehydration beds that serve sweet liquid hydrocarbon streams (propane and butane). I also might add that reflected desiccant size was obtained from engineering standards and manufacturer loading guidelines.

Although the mixture of different desiccant in the same dehydrator is no longer permitted at Saudi Aramco sites, we still allow use of different brands for adjacent vessels, meaning that Bed A could possibly be loaded with a different material than Bed B.

Ahmed S. Ghazal
Saudi Aramco
Dhahran