## Featured Articles

Check out the latest featured articles.

## New Article

Product Viscosity vs. Shear

## Featured File

Vertical Tank Selection

## New Blog Entry

Scrubber Design for Desulfurization- posted in Ankur's blog

# Molecular Sieve Column Calculations

This topic has been archived. This means that you cannot reply to this topic.
5 replies to this topic
|

### #1 lalibo

lalibo

Brand New Member

• Members
• 4 posts

Posted 29 March 2011 - 10:33 PM

Hi all,

I am currently doing a design project at uni and am required to give a list of design criterias and size two adsorption column that uses molecular sieves.

The feed will be 1349mol/hr water, 7095mol/hr ethanol and 5.5mol/hr isopropanol (This is what I obtained in Aspen using one distillation column). A first adsorption column will be used to remove the water whilst the second column will be used to extract the ethanol. The only thing I know for now is that I will be using a type 3A sieve for the first column, and a type 4A sieve for the second.

Can anyone point me out to relevant equations to calculate the size, surface area required, work for pump and heating requirements for regeneration? Also, is there any recommendations for how long the adsorption column should be running and how long the regeneration time is to be (my group has not given me any specific info on this yet)?

Attached is a rough version of our PFD, distillate stream table (which gives the feed flow and composition for the adsorption column) - please let me know if anything there is not feasible or wrong. Also, what is an efficient method to extract glucose and xylose from water at approx 100oC? I was considering evaporative drying because of the already high temperature but at the same time it seems extremely energy inefficient because of the large volumne of water (107122mol/hr) to be evaporated. Any advice on this aspect?

Any kind of help would be greatly appreciated as I have never been taught anything on this topic before.

Cheers

### #2 riven

riven

Gold Member

• ChE Plus Subscriber
• 178 posts

Posted 30 March 2011 - 02:03 AM

On getting data for simulation, I cannot help you. Some general websites
http://www.molecular...cular_Sieve.htm

Regarding the use of 4A molecular sieve to remove ethanol, this is a practical solution if ethanol is a minor component. I would not use this as a bulk extraction technique as you will have to adsorp all the ethanol and then deadsopt it; high energy ecomsumption.
The amount of IPA leads me to believe that you can ignore it. Even if you lose 10% of your ethanol, all the water and keep all the IPA in the first stage , the ethanol composition will be around 99.9% which is good for most (if not all as far as I am aware) applications. Why do you need to get rid of the IPA?

### #3 lalibo

lalibo

Brand New Member

• Members
• 4 posts

Posted 30 March 2011 - 02:24 AM

We thought that IPA was to be removed to fit regulations as combustion fuel for vehicles. Say that the IPA was to be removed from the ethanol stream, what would be a recommended process to do so then?
I will go double-check on the ethanol purity required by the Australian govt.

Currently just using the information here (http://www.tower-pac...cular_sieve.htm) to calculate the bed length needed. I suppose I might eventually be able to scab equations here and there from articles to work everything else out..

Thanks for the help!

### #4 lalibo

lalibo

Brand New Member

• Members
• 4 posts

Posted 30 March 2011 - 10:43 AM

Okay I've worked out the bed length, need to find requirements for the feed pump and energy required for regeneration.

Can anyone review my work and advise me on relevant equations for pressure swing regeneration?

### #5 Art Montemayor

Art Montemayor

Gold Member

• 5,529 posts

Posted 30 March 2011 - 01:50 PM

I have been a practicing professional engineer for 51 years and have been in these Forums for quite some years now. I am still surprised by what I see and read in some of our threads.

Your query is one that requires me to ask you a series of questions in order to identify how best to help you:
• What university do you attend in Australia, and what year of Chemical Engineering are you in?
• How is it that you have been assigned this project although you "have never been taught anything on this topic before"?
• If you have no knowledge of adsorption, how is it that you have decided that you "will be using a type 3A sieve for the first column, and a type 4A sieve for the second"?
• How can it be that a Chemical Engineering student actually believe that an assigned project can be resolved using the logic "I suppose I might eventually be able to scab equations here and there from articles to work everything else out"?
I ask these questions for my own enlightenment because the frank reality of your situation is that unfortunately there are no "relevant" equations for pressure swing regeneration that we can advise you on and that will quickly, easily, and painlessly obtain for you the right answers and the grade you are looking for. Put succinctly in vernacular U.S. engineering language, "there are no free rides and no free lunches in engineering". You are going to have to face up to the hard, complex, and taxing job of confronting the adsorption Unit Operation and work out all the details. But that is what any engineer would expect.

To help you out I am attaching a "How to" adsorption design booklet that should keep you very busy and active in discovering how an adsorption process is designed. You need to understand what this paper addresses in order to qualify to design your own adsorption process.

I am also submitting my Rev 1 of your workbook with some comments. Basically, you are taking this all too lightly. You fail in making a logical path to your solution by failing to specifically define WHAT PHASE of fluid that you are planning to adsorb. You state that you are adsorbing WATER. "Water" is a liquid; steam is a vapor. In your calculation you use NIST to find the specific volume of STEAM – not water; and yet, you get the data wrong. Both your logic and your basic data are wrong. For example:

You "seem" to be feeding Liquid Water (?) at 100 oC to the adsorber. But yet you make calculations that use the water figure and multiply it by the specific volume of steam! What is what? If you indeed are instead feeding a vapor stream (STEAM) to the adsorber, how do you intend to keep it in the vapor phase? Your bed initially will have to be higher than 100 oC in order to keep the stream in a vapor phase. What is your sorptivity for ethanol at 100 oC?

What is your logic for designing the diameter and length of your adsorption bed? You don't even address the need for a maximum superficial velocity through the bed. Show your algorithm and logic in arriving at your results. Otherwise no one in his right mind is going to believe your figures – even if you used some super-duper, extra-terrestrial equation from some website.

You say you are designing a Pressure Swing Adsorption (PSA) system – but yet you fail to show how you are going to establish a pressure and how it is that you can expect a regeneration simply by dropping the pressure.

This problem is more complex than what you would prefer to believe. And what is worse is that the most complex part is in defining specifically what it is that you plan to do and how you propose to go about it – an algorithm for the solution.

### #6 lalibo

lalibo

Brand New Member

• Members
• 4 posts

Posted 30 March 2011 - 08:46 PM

Thanks Art Montemayor,

I won't reveal what uni I attend as not to bring shame to it, but currently I am in my fourth year. We were only given one lecture on designing an adsorption column last year, and basically all I got out of it was "Solute in feed = capacity of adsorbent", ie cF*Q = qF*density of bed*A*LES , for the sizing of a fixed bed.

The decision to use 3A and 4A molecular sieves was made after searching through a series of company websites to look for suitable candidates, and all sites told me that 3A is suitable for adsorbing water from ethanol, and 4A adsorbs ethanol from IPA. However I have decided to scrap the second adsorber column because I found that there was actually no fuel standard for ethanol in fuel in Australia yet. I read this (http://www.environme...-fuel-grade.pdf) and decided not to use the second sieve. The exit molar composition should be 99.92% ethanol and 0.08% IPA, it should be okay for use.

The reason I have had to "scab equations" was because I spent hours and hours going through webofscience and my uni's library website trying to find anything that could be of assistance to me but without the basic theoretical knowledge on this topic, I really could not fathom how the column is supposed to be designed.

In the end I eventually gave up and wrote up a very rough version of what I was going to do to see if I were heading in the right direction (forgot to label the document as draft). I thought about the steam at 100oC, 1atm being liquid but haven't been able to find a source that gave me recommendations for temperatures to pump the feed at, so I just left it off at how it exited the distillation column for the time being, since it can be easily changed later. As for the bed area, I took the ratio 3:1 for bed length: bed diameter because that was just the dimensions from a tutorial question that I assumed to be about an average adsorption column. I think the "solid loading" part on my lecture note meant the amount of fluid that can be loaded onto the molecular sieve, kg sorbate/kg sieve. I had no idea where to find experimental data to obtain maximum superficial velocity (my lecture notes just gave a few questions with it being calculable from experimental data) so I just assumed a typical wavefront profile from articles I've read and estimated the length of unused bed from there.

I just started reading the guide on adsorber design. It should probably allow me to design something much more concise as it gives me what I need to start my work. I will keep you updated once I draft a version of the column.

Cheers!

Edited by lalibo, 30 March 2011 - 10:28 PM.