8 replies to this topic

[agporte]

I understand that there are many different technologies out there for the purification of ethanol/water mixtures from 95% to >99.5%.  I most recently worked with one of our clients on a molecular sieve skid and back in my college days did a project comparing extractive distillation to pervaporation.  I know people have used low pressure (vacuum) distillation for this process, but it is not commonly used due to energy requirements (to pull the vacuum and to sub-cool the condenser water).  That being said, even though it may not be the best technology, I am investigating the feasibility of using vacuum distillation to purify an ethanol/water mixture.

 

A little background: A 10wt% ethanol water mixture will be feed to an atmospheric column purifying the stream to ~95wt%.  This stream will enter the vacuum distillation column and be separated to 99+%.  I’m looking to do this on a relatively small scale continuous process (5 gpm feed rate).  Everything was going along smoothly until I got to designing the vacuum distillation column.  I was under the impression that the azeotrope shifts to the right on a VLE diagram (towards 100%) as the overall pressure decreases.  However, this was not the case for the many different sources/calculators/correlations I found on the internet.  Many of them showed a negative dip (below the equilibrium line), but the azeotrope stayed at ~95.5wt% in all case.

 

My main questions:  Am I just finding bad VLE data? Or does this separation require a “pressure swing” where the concentration would “jump” over the azeotrope and I would then separate the water from ethanol?  From my literature search it looks this purification has been done before with just standard vacuum distillation.

 

Any help or direction would be extremely appreciated.  And don’t hesitate to ask questions if I’ve left out any data.

[agporte]

These are the two sites I've used to find my VLE data:

 

http://www.ddbst.com...hanol;Water.php

 

http://vle-calc.com/...mberForVLE=0.06

 

Does anybody know of any other sites?

[PingPong]

Azeotrope should indeed move to the right in an Y-X diagram. In Perry's Chemical Engineers' Handbook 6th edition (1985) there is Figure 13-63 which shows that. At least down to 100 Torr (100 mmHg = 0.13 bar) as the graph does not go further.

 

If you do not have a professional process simulator at your disposal I suggest you use a liquid activity model (Wilson, Unifac, Uniquac, ...) to calculate it yourself. See for example Perry table 13-2 and 13-3 for Wilson.

 

I would never trust some calculator on the internet of which it is not clear what it does and what method/formulas it uses.

Edited by PingPong, 25 May 2013 - 07:22 AM.

[Dacs]

It's been years since I did this, but I do remember using NRTL model to obtain VLE data and it agrees with what Perry has given.

 

I attached a paper that I made when I was doing my undergraduate study, and maybe it can point you to the right direction on how to do this.

 

It includes how to solve it and there's a sample source code that you can use. It's been years, but I think this can easily done in MS Excel.

 

But if you have a commercial simulation software in hand (like Hysys), then it'll make your life much easier.

 

 

EDIT: Need to check the file, I just got it out of my archive.

Edited by Dacs, 26 May 2013 - 07:44 AM.

[thorium90]

That was the shortest dissertation i have ever read. Is it just an abstract? Perhaps it would help to post your entire paper if you are ok with it.

[Dacs]

Nope, it's just one of my requirements for one of my courses in my undergraduate degree.

 

It's more of a report than a paper in fact

[agporte]

Azeotrope should indeed move to the right in an Y-X diagram. In Perry's Chemical Engineers' Handbook 6th edition (1985) there is Figure 13-63 which shows that. At least down to 100 Torr (100 mmHg = 0.13 bar) as the graph does not go further.

 

If you do not have a professional process simulator at your disposal I suggest you use a liquid activity model (Wilson, Unifac, Uniquac, ...) to calculate it yourself. See for example Perry table 13-2 and 13-3 for Wilson.

 

I would never trust some calculator on the internet of which it is not clear what it does and what method/formulas it uses.

 

Pingpong - I have the 7th edition of Perry's and I've looked all up and down chapter 13 and can't find any data on water/ethanol at 100torr.  I only see the X-Y data for 101.3 kPa (I wish I had my old thermo book).  I'll try and use Wilson or NRTL. Thanks for your feedback on this.

 

 

It's been years since I did this, but I do remember using NRTL model to obtain VLE data and it agrees with what Perry has given.

 

I attached a paper that I made when I was doing my undergraduate study, and maybe it can point you to the right direction on how to do this.

 

It includes how to solve it and there's a sample source code that you can use. It's been years, but I think this can easily done in MS Excel.

 

But if you have a commercial simulation software in hand (like Hysys), then it'll make your life much easier.

 

 

EDIT: Need to check the file, I just got it out of my archive.

 

Dacs - I don't see the file attached.  How do I get to it?  I'm guessing that this will be exactly what I'll need.  Thanks for doing all the leg work on this  . If it's not in Excel yet, I'll make a document and post it back on here for all to use. Thanks!

[PingPong]

The figure 13-63 that I referred to is in Perry 6th edition. See attachment.

According to text in Perry there is no azeotrope below 70 mm Torr.

 

 

Table 1 in this document says basically the same, no azeotrope at 80 mm Torr or below:

http://profiles.uonb...c_published.pdf

Attached Files

•  Figure 13-63 Perry 6th edition.pdf   207.01KB   142 downloads

[agporte]

The figure 13-63 that I referred to is in Perry 6th edition. See attachment.

According to text in Perry there is no azeotrope below 70 mm Torr.

 

 

Table 1 in this document says basically the same, no azeotrope at 80 mm Torr or below:

http://profiles.uonb...c_published.pdf

 

Very strange.  It looks like Perry's took out some stuff on binary distillation and replaced it with enhanced and multi-component distillation.  One of my co-workers had a 6th edition and I was able to find the table that you posted as well.

 

I also came across that article when doing my initial internet search.  So it's confirmed. The azeotrope "disappears" after about 70-80 torr.  Now the search for the right VLE  correlation begins.