Dehydration of ethanol using PSA 1

[unclewil]

I've got a feed of 93% by weight of ethanol with 7% water, and I've got an output of 99.5% by weight of ethanol. I've also been told that I must use PSA with 3A zeolite with regeneration by using some of the dry vapour at low pressures.
I have never been assigned to design such a project in my career so far and though I know the basics, I have yet to find any useful design information, so I was wondering if people could give some tips on operating pressures for dehydration and regeneration, whether liquid or vapour phase operation is best, and the fraction of the dry vapour to use for the regeneration bed?

 

[StoneCold]

Unclewil
If it is very big then UOP could do it for you, or others I am sure. If it is small snuggle up to Alcoa they can help you with the process stuff.

Goodluck
StoneCold

 

[Montemayor]

unclewil:

I’m going to assume you have a mandated scope of work and, as such, I will not make any recommendations or suggestions about the desirability of your selected process except to state that Mol Sieves will yield an almost “neat” alcohol and Alcoa’s activated alumina will do as well – although it seems you’ve already opted for Mol Sieves. There are various ramifications involving the adsorption unit process you’ve chosen, and I’ll recap some highlights to make sure we all understand what you’ve “bought” into:

1. Adsorption is a semi-continuous process (almost batch-type) and will require some surge rate capacity if you are trying to maintain or fit in with steady-state, continuous facilities down stream.

2. You don’t require above-normal process pressures for adsorption, although some would help;

3. Adsorption inherently will involve a “TEMA cycle” – which is composed of: (a) a normal “drying” time where the feed is passing through one Mol Sieve bed; (b) a regeneration time that is composed of draining a spent bed of Mol Sieve, passing a hot, regen gas through the bed, subsequently cooling the bed, refilling the bed with pure liquid product, and putting the regenerated bed into operation while taking out the other bed for regeneration. These sequenced and timed events must coincide with your process and regeneration needs.

4. The adsorbent Mol Sieve beds are usually in two (or more) vertical vessels and there are critical, block valves that are operated to switch the beds from process drying to regeneration. This can be done manual or automatically, using PLCs or other automatic instrumentation.

5. Mol Sieves require much higher regeneration temperatures than other adsorbents such as Activated Alumina. I’m talking about the range of 400 to 700 oF. Be sure you are aware of the regeneration requirements and it’s power consumption.

6. Your regen vapor exhaust stream is going to be a waste problem; I assume you will have to employ an inert gas such as Nitrogen and subsequently incinerate the waste.

7. You will probably regenerate at essentially atmospheric pressure – but with very high temperatures. What you are doing during regeneration is that you are stripping out the captured water molecules that are held on the adsorbent’s surface under van der Waal forces. So you have to sustain a high regen temperature for a stable period in order to ensure that you’ve stripped out all the captured impurities (you may be also adsorbing other goodies besides water).

8. The initial adsorbent charge will last indefinitely under a theoretical basis. However, in real life, we know that you will be pumping in some contingent impurities and other contaminants. Adsorption beds have to be replaced in due time – depending on the cleanliness and characteristics of the service. Mol Sieves are the most expensive adsorbent – in case you haven’t found this out already. Any adsorbent will plug, get “poisoned”, or simply start to lose sorptivity over time if it is subjected to tars, solids, salts, or any other contaminants that affect either its surface area or its sorptivity potential. The effectiveness and thoroughness of the regeneration is also a major factor in maintaining the adsorption beds at peak performance.

Your best method of reviewing and analyzing the process and its characteristics specific to your application is to discuss this with the company that will design and construct the unit. There are many engineering companies that specialize in this and I’m sure they’ll be very glad to go into details about your scope of work and basic data – especially that which regards your requirements for product purity and other process specifics.

My first, knee-jerk reaction is that 7 %wt of water is too large a load for an adsorption unit to be able to economically vie with other processes. But I’m not an expert in the design and fabrication field and know even less your scope of work. In the past, I've used adsorption to "polish" the purity of my liquid products - not as the main, sole process.

You’ve said that you’re instructed to employ “PSA”. Well, PSA means Pressure-Swing-Adsorption, and it a process applied to gaseous streams – not liquid streams. The reason it is used in gas applications is that a lower pressure is used as an aid to regenerate the adsorbent beds. This does not work with liquid streams. I’m very concerned about your use of terms like PSA and size 3A Molecular Sieves when you are merely at the evaluation stage and haven’t yet conceptually (nor economically) evaluated the type of process. You (the user) are not the one who will determine or pick the appropriate adsorbent – or its specific size or type. The designer / fabricator supplier of the unit will do that – after reviewing and studying your scope of work and specific needs. That’s why I strongly recommend you get together with a serious supplier and discuss your feasible, potential application.

I hope this experience helps you out.

 

[ibmem]

Unclewil,

The MolSieves will do the job, but an EtOH dehydration from 7% to 0.5% is an ideal application for a pervaporation or a vapour permeation (membrane recovery):
- less energy (50% steam consumption)
- no consumption of dry EtOH (required for the regeneration of Mol.Sieves)
- no contamination

hope it helps

ibmem

http://www.ibmem.com

 

[Montemayor]

unclewil:

This is not to deter from your main topic, but ibmem has mentioned a specific option to adsorption - as I indicated in my post. Pervaporation would seem to be a strong option, depending on your scope and existing situation. For a great overview on the subject, go to:

http://www.cheresources.com/pervaporation.shtml

Max Planck Street. What a great name for a street location of an engineering firm! And here in the USA we put up with Elm, Maple, Main, and Broadway streets....

 

[love4u]

13X molecular sieve is more suitable for ethanol dehydratio than 5A molecular sieve. This information can be found in the manufacures' website.

 

[bexcelant]

Ok I feel that after reading all the posts here I must interject. Having worked for both Alcoa and Union Carbide in molecular sieves and alumina. 3A molecular sieve is absolutely the only material you should consider. The amorphous nature of alumina will adsorb and hold onto excess ethanol while 3A will exclude all the ethanol and hold onto just the water. PSA technology is common in ethanol dehydration and so is thermal swing. The ethanol is deyhdrated at elevated temperatures in the vapor phase. This means that adsorption capacity is lowered due to the higher temperatures.

Typically first stage distillation is removes most of the water from the ethanol so we are at 6 percent water in ethanol. The remainder is removed by adsorption through the molecular sieve.