27 replies to this topic

[pmoneysh]

Hello experts,

 

I am now involved in a project to design a distillation column for PF5 generator. This project is a scale-up of existing plant. Existing design of distillation column is:

 

Bottom part of the column is a reactor and there is a re-circulation pump. There is only 1 outlet from the top. However, there is a required spec of impurity at the top (less than 10 ppm- for the new project). Existing plant has an impurity of around 100 ppm. How effective would it be if we increase the column height? I am thinking of using an additional tail column but the operating cost is quite high. It is the 1st time I am seeing a distillation tower with only one outlet.

 

Operation philosophy of the existing plant: 

 

The reactor is initially charged (50% level) with 2 of the 3 reactants (Reaction does not proceed at this time). Then we have continuous feeding of reactants and simultaneously distillate the product.

 

I am just wondering how it has been possible to operate this tower with no major issues for years. I need to learn more about distillation:(

 

Impurity here refers to HF.

 

Any of your suggestions or experience would be appreciated

 

Thankyou

Attached Files

•  schematic diagram.jpg   172.59KB   6 downloads

[breizh]

Hi,

Share with us more information about your product (composition of the feeds, the distillate), the mass or moles balances, the operating conditions (Pressure, temperature), the Boiling point of the materials if not available in literature, the nature of the coolant, the heat source.

How do you discharge the heavies? 

Share with us the existing SOP for better understanding of the operation.

What is PF5?

As long you have a reflux in place, you could modify it and check for the results.

Is your operation under vacuum?

Do you have means to control the operation and where are they located? (Temperature sensors, pressure sensors, flow meters, analytical means)

How do you monitor the purity (analytical method) and what is the yield of your operation?

Do you have a model for this existing operation or it's purely empirical, trial and errors from pilot?

Going from 100 ppm to 10 ppm should be very challenging.

 

More details better answer!

Good luck

Breizh 

[shvet1]

What is PF5? 

Gas phosphorus pentafluoride

Electronics industry

Edited by shvet1, 05 July 2023 - 11:17 PM.

[pmoneysh]

Hi Breizh, 

 

Thankyou for the reply. We do not discharge heavies. There is only an emergency vent line at the bottom. PF5 means phosphorous pentafluoride. Condenser temperature is very low. Heat source is LP steam. Existing plant (currently in operation for a very long time- probably more than 10 years) impurity is 100 ppm. We do have the sensors and controllers. The yield is very good almost equal to theoretical yield. 

 

I have not been able to understand why there is no accumulation when there is reflux and only one outlet stream. Therefore, I wanted to know if any one in this group has ever desgined or operated such tower.

 

Thankyou

[shvet1]

  (currently in operation for a very long time- probably more than 10 years) 

 

50 years is a very long time in chemical industry

 

Seems it is a lab scale unit. Did you provide a test run to find out max capacity of packing? Do you know reflux ratio limit you are able to reach? Is it able to decrease tower capacity (evaporation rate)? Why you did not try to adjust reflux rate (it is a simple procedure)? Are you able to replace packing with more efficient one? Which is a bottlenecking element of this existing unit?

 

What do you mean "scale-up" actually - you want to convert lab scale unit to bench scale unit (bench to pilot, pilot to semi-industrial unit, semi-industrial to industrial - does not matter)? If you want to "scale-up" then why you did not design and test existing unit properly (much more cheaper and less risky option)? If the goal is to design a pilot scale unit then where is your process engineering team? What is current technology readiness level? Is this an emerging technology or a replica of existing one?

 

Note that as you have only one outlet stream impurities are not affected by reflux as this is a question of mass balance. You still have an option to obtain "contaminated" batch (let's say 1000 ppm) and "pure" batch (let's say 10 ppm) of gas but average contamination rate will be still 100 ppm (google how whiskey is distilled).

 

Looks like you are lack of basics in chemical engineering. Do you have a process engineer in your team? Not an assistant, technician or scientist - an engineering person. Who did provide a design of this tower? calculated packing, heat exchangers and pump? Where is this person now?

 

If this is a lab/bench scale unit why did you post in industrial professional forum? If this is an pilot/industrial scale unit where is a process engineering service now (internal or outsourced)?

Edited by shvet1, 06 July 2023 - 12:23 AM.

[breizh]

hi,
weird! what are the impurities?
Breizh.

[pmoneysh]

Thankyou Shvet for the response. 

 

I would not say it is a lab scale since the capacity is 127 kg/hr. We do have the documents and the design details. Also, we do have process engineers and I am one of them (though I have only 3 years of experience). I have designed 4 distillation columns- basic stage (and heat exchangers) to date but never got a chance to deal with this kind of. 

 

Senior engineers get involved during detailed engineering phase and of course, they will correct us. Here, my main concern is I am amazed to see the role of the  recirculation pump. Is it really because of this pump?

 

Thankyou

[shvet1]

hi,
weird! what are the impurities?
Breizh.

 

 

 Impurity here refers to HF. 

[pmoneysh]

hi,
weird! what are the impurities?
Breizh.

Impurity is Hydrogen fluoride. Relative volatility difference between the components is very high.

[shvet1]

@pmoneysh

 

You still have provided no info related. We keep know nothing other than oversimplified PFD.

 

If there is an engineering team why they did not provide calculations and a test run? This is a simple, cheap, fast operation. You mentioned detailed engineering, design and some documentation what means for me that you have all data required. If there has been a detailed engineering then a basic engineering also. A lot of efforts I should say - where is all this staff?

 

Honestly speaking 127 kg/h says nothing to me, capacity a day or a year is more informative. Still looking as lab or bench scale.

Edited by shvet1, 06 July 2023 - 12:41 AM.

[breizh]

BP Pf5 84.55 C

BP Hf 19.5 C

check with your process engineer.

[pmoneysh]

BP Pf5 84.55 C

BP Hf 19.5 C

check with your process engineer.

I think you missed -sign in BP of PF5. It is around -84oC

[shvet1]

 Relative volatility difference between the components is very high.

 

 

target 10 ppm is a rather low limit too, isn't it? 

Oversimplified approach is - You want to decrease an impurity contamination to 10 time (from 100 ppm to 10 ppm) what means you should increase some parameter to 10 times (let's say reflux ratio).

 

Ultrapure products are hard to produce as those require a lot of energy.

 

https://www.amazon.c...r/dp/0070349096

https://www.amazon.c...r/dp/007034910X

https://www.amazon.c...rnest E. Ludwig

Edited by shvet1, 06 July 2023 - 12:56 AM.

[shvet1]

check with your process engineer.

 

Topicstarter is that very process engineer

[pmoneysh]

@pmoneysh

 

You still have provided no info related. We keep know nothing other than oversimplified PFD.

 

If there is an engineering team why they did not provide calculations and a test run? This is a simple, cheap, fast operation. You mentioned detailed engineering, design and some documentation what means for me that you have all data required. If there has been a detailed engineering then a basic engineering also. A lot of efforts I should say - where is all this staff?

 

Honestly speaking 127 kg/h says nothing to me, capacity a day or a year is more informative. Still looking as lab or bench scale.

We cannot make a test run for the new project. As  I already mentioned, we have two options either to increase the tower height or to add an additional tower to match the purity. 100 ppm is the optimum impurity we can get in the existing plant. 

 

From what you said "Note that as you have only one outlet stream impurities are not affected by reflux as this is a question of mass balance".  In other words, increasing tower height is not an option, right?

 

127 kg/hr- i mean to say 127 kilogram/hour

Edited by pmoneysh, 06 July 2023 - 12:50 AM.

[breizh]

Yes you are right. - 84.5c
Work on the reflux and conceive to lose some material, purge.
Breizh.

[shvet1]

 This project is a scale-up of existing plant.  

 

 

 We cannot make a test run for the new project.  

 

1/ Make a test run of existing unit, not scaled-up one. This is cheaper and less risky.

 

2/ Increasing tower is an option. One should provide calculations to be able to compare all possible options and choose optimum one.

 

3/ 127 kg/h or kilogram/hour has no sense as time in operation may be, let's say 5%, of total. For batch operation capacity is measured a day, a month or better a year.

 

An engineer provides engineering, right? Engineering means knowledge+calculations, right? No calculations = no engineering = no feasible results = risk + time&money loss.

Edited by shvet1, 06 July 2023 - 01:16 AM.

[pmoneysh]

3/ 127 kg/h or kilogram/hour has no sense as time in operation may be, let's say 5%, of total. For batch operation capacity is measured a day, a month or better a year.

 

Understood. Thankyou.

 

2/ Increasing tower is an option. One should provide calculations to be able to compare all possible options and choose optimum one.

 

Currently, I am performing the simulation to tune with the real data but have problem on the recycle stream at the bottom (as expected) using aspen plus. If we decrease the tolerance of the recycle stream to 0.01, we can get a converged solution. Total imbalance is 18 kg/hr. 

 

1/ Make a test run of existing unit, not scaled-up one. This is cheaper and less risky.

 

We tried it. Changing any of the variables gave worse result.

 

Thankyou for the suggestions

[shvet1]

 Total imbalance is 18 kg/hr.  

 

18 kg has a potential to result in error 100'000 ppm HF. Tolerance is critical for traces of impurities.

 

 

We tried it. Changing any of the variables gave worse result. 

 

Tell us more. There is a chance we are able to improve results.

Edited by shvet1, 06 July 2023 - 01:50 AM.

[breizh]

why do you mention 3 feeds, are they different? feeding the still at the same time? recirculation should be there for better heat transfer and or homogeneity of the mixture, if some?
To me nothing to do with simulation, unless to drain some product from the bottom. Can you share some info about the operating procedure, start up? Breizh

[pmoneysh]

why do you mention 3 feeds, are they different? feeding the still at the same time? recirculation should be there for better heat transfer and or homogeneity of the mixture, if some?
To me nothing to do with simulation, unless to drain some product from the bottom. Can you share some info about the operating procedure, start up? Breizh

yes i agree with you. That is why I posted the problem in this forum. It is not possible to drain out the bottom product. Reactor size is about 1200 H* 900 ID mm and 50% is already charged with some of the reactants (two of the same reactant as of the continuous feed). There are 3 continuous feeds (reactants). The purpose of the pump is for mixing as there is no agitator in the reactor as you said. flowmeter in the pump shows upto 20,000 kg/hr recirculation rate. 

 

Could you think of any purging location since we cannot make any nozzle at reactor as it is already filled with reactants

 

Thankyou

Edited by pmoneysh, 06 July 2023 - 02:17 AM.

[breizh]

What is the height between structured packing and Pall, above the recirculation nozzle. You may be able to draw some material on the side of the column, if you can install a collecting pan to recover liquid (w part)and install a nozzle. Breizh

[pmoneysh]

Thankyou for the suggestions Breizh and Shvet. I will one more time look at the temperature profile and will recommend the nozzle location as Breizh suggested.  Shvet - your comments were really useful.

 

Thankyou

[latexman]

Take some of the 100 ppm product to a “distillation lab” and see what is needed to get to 10 ppm product. It seems that would be valuable information for you.  And, it's real data; not a simulation.

[breizh]

Hi,

I agree with Latexman but you need to be extremely cautious with these nasty materials (PF5 and HF).

Safety is key there.

To be handled by knowledgeable people, not students.

Breizh