Sciencemadness Discussion Board

fractional distillation

spoofy - 21-6-2009 at 14:04


I'm trying to do a basic fractional distillation using water since it is my first time.

The problem is that the water is not coming over to the small flask. In fact it is just condensing in the fractionating column and dripping back down .. before it even seems to make it to my 3 way adapter. Why isn't the water condensing in the condenser instead?


bfesser - 21-6-2009 at 14:05

Insulate the column. Water is high boiling, and so will condense relatively easy with only air cooling.

BromicAcid - 21-6-2009 at 14:09

Follow above suggestion to help, or add more heat, it can take a lot of energy to force your distilate over through the fractioning collumn. Also, even though you are using a fractioning column, if you're not collecting fractions, is it really a fractional distillation?

spoofy - 21-6-2009 at 15:15

Quote: Originally posted by bfesser  
Insulate the column. Water is high boiling, and so will condense relatively easy with only air cooling.


insulation was enough to push it over. thanks

Paddywhacker - 21-6-2009 at 15:52

Insulating the column will allow the distillate to come over, but it is not an ideal solution. Insulating the column will decrease it's efficiency, and that is not good for real fractionating.

I think what you are using is too large a column for the amount of liquid that you are distilling. Or perhaps, not enough heat on the flask.

watson.fawkes - 21-6-2009 at 16:01

Quote: Originally posted by Paddywhacker  
Insulating the column will allow the distillate to come over, but it is not an ideal solution. Insulating the column will decrease it's efficiency, and that is not good for real fractionating.
Fractionating depends upon the establishment of a thermal gradient between the vaporization end and the condensation end. Insulation will shift the burden of cooling from the walls of the column onto the condenser. If the condenser can reject heat at an adequate rate, you haven't lost efficiency. If your condenser is underpowered (meaning, insufficient rate of heat energy removal), then you will lose efficiency.

Paddywhacker - 22-6-2009 at 14:21
Quote: Originally posted by watson.fawkes  
Fractionating depends upon the establishment of a thermal gradient between the vaporization end and the condensation end. Insulation will shift the burden of cooling from the walls of the column onto the condenser.
That's right.
Quote: Originally posted by watson.fawkes  
If the condenser can reject heat at an adequate rate, you haven't lost efficiency. If your condenser is underpowered (meaning, insufficient rate of heat energy removal), then you will lose efficiency.

Nope. Efficiency is a is not a measure of how well you condense the distillate, but of how well you separate the fractions, and that depends on the temperature gradient along the fractionating column, which you will lose by insulating, as you said in your first sentence.

1281371269 - 22-6-2009 at 14:25

I think I'm missing something here...what fractions are you collecting?

watson.fawkes - 22-6-2009 at 16:52

Quote: Originally posted by Paddywhacker  

Quote: Originally posted by watson.fawkes  
If the condenser can reject heat at an adequate rate, you haven't lost efficiency. If your condenser is underpowered (meaning, insufficient rate of heat energy removal), then you will lose efficiency.

Nope. Efficiency is a is not a measure of how well you condense the distillate, but of how well you separate the fractions, and that depends on the temperature gradient along the fractionating column, which you will lose by insulating, as you said in your first sentence.
I wasn't talking about efficiency of condensation at all. You'll lose efficiency if your condenser is underpowered because the temperature at the condenser rises above the condensation point, reducing the temperature difference between the boiling flask and the condenser, and thus lowering the temperature gradient. As long as your condenser is adequately powered, the temperature difference is governed by the temperature of the lowest boiling fraction.

As long as your condenser remains at a fixed temperature, you lose no gradient by insulating the column. You have one temperature at the condenser, another at the boiling flask. There's a continuous function of temperature connecting those two temperatures. The temperature gradient is the first derivative of that temperature function. If you're column is insulated (assume infinitely well), then the only heat exchange is up and down the column. Assuming a column of constant cross section, this yields a linear gradient.

If you have heat leaking out the sides, that's heat that's not available for vaporization, lowering the number of theoretical plates of the column.

Paddywhacker - 22-6-2009 at 17:26

I think you are mistaken. The fractionating efficiency ... the number of plates ... depends in part on the temperature gradient between the top and the bottom of the fractionating column.

At the top of the column the take-off vapour spills into the condenser.

The separation is done in the column, not in the condenser, and so depends on the gradient along the column. The condenser is irrelevant except that if its hold-up volume is large then small fractions will mix and smear as they pass through.

watson.fawkes - 22-6-2009 at 17:40

Quote: Originally posted by Paddywhacker  
I think you are mistaken.
Nope. I looked it up before replying, just to make sure.

Paddywhacker - 22-6-2009 at 19:39

You have got to be yanking my chain. You cannot believe that separation of the fractions takes place in the condenser.

The very best separation of fractions takes place when there is no take-off into the condenser and the lowest-boiling fraction is condensing at the top of the condenser and the lower-boiling fractions are separated out along the column in a refluxing equilibrium.

Then, to achieve practical separation the vapours at the top are bled into the condenser at such a slow rate that the equilibrium established on the column is maintained. Bleeding off at too high a rate will have vapour rushing up the column disturbing the temperature gradient and smearing the fractions.

Read a bit more. And think about it.

watson.fawkes - 22-6-2009 at 20:31

Quote: Originally posted by Paddywhacker  
You have got to be yanking my chain. You cannot believe that separation of the fractions takes place in the condenser.
I don't. And I have no idea how you came to believe that. Really, none.

We were talking about insulation, right?

Ramiel - 23-6-2009 at 04:00

Woo, science battle. I love these

It's agreed, fractionation occurs in the column.

Gentlemen, removing egos from the debate for a minute, is it possible you're talking about two different definitions of "efficiency". I don't understand the fine point that the argument seems to be over.

I was taught (and I may be wrong) that insulating the column is bad, but you must guard against gusts of air. I had a rather large column (40L or so scale), and had to insulate the steel column with a loose sleeve of plastic, so that gusts of wind didn't disrupt the column.

Mossydie, I think spoofy was just doing a trial run with water to get the hang of fractioning. It is quite tricky, as you have to achieve an steady-state system in the column, with a little bit of your distillate coming over into the condenser.

[edit - thanks watson, I replaced the misleading "equilibrate" with "steady-state", I did not mean to imply equilibrium ^_^ ]

[Edited on 24-6-2009 by Ramiel]

Saerynide - 23-6-2009 at 04:13

Paddywhacker is correct. Separation occurs in the column. Too much heat, and the column will take off (and flood, if its packed), ruining the separation.

Insulation does affect the gradient in the column. Insulation keeps the heat in the column all the way until the top. If you wait for the vapor to reach the condenser without letting it lose heat on the way up there, it is too late, as the heavy keys will have already come over the top. Best separation occurs when equilibrium is established, with the tops at the light key bp and the bottoms at the heavy key bp.

[Edit]: I was also taught that columns rarely need to be insulated on the bench scale. For water, you can definately do without it.

I've only ever seen pilot scale and larger columns being insulated.

[Edited on 6/23/2009 by Saerynide]

Nicodem - 23-6-2009 at 05:03

This same question and the resulting nonsense about noninsulated columns being more efficient was already answered and discussed in another thread on the same topic.
Most commercial 29/32 wide distillation columns come with a vacuum jacket insulation. I wander why?

The temperature gradient along the column must depend on the vapour/liquid equilibration process if you want the column to work well. Loosing the heat along the column is just deteriorating this equilibrium, causing the vapours to condense at a faster rate as they can evaporate (thus a reflux which can cause column flooding).
The temperature at each point must correspond to the boiling point of the mixture with the composition at that same point. In other words, the temperature gradient must correspond to the composition gradient. And that's why you need good insulation for better efficiency.

watson.fawkes - 23-6-2009 at 05:51

Quote: Originally posted by Saerynide  
Insulation does affect the gradient in the column. Insulation keeps the heat in the column all the way until the top.
This is true only in an out-of-steady-state condition. It keeps heat from leaking out of the column, that true. But thermal loss from the walls of the column is not the only way that the column loses heat. The condenser itself is a heat extraction device. Heat goes in the bottom into the boiler and comes out the top from the condenser.

A column in a steady-state heat flow condition is not in a state of thermal equilibrium. Thermal equilibrium is the state where everything is the same temperature. I know it's ordinary usage to say that a column under total reflux is in equilibrium, but it's not true in the physics sense of the word equilibrium. Thermal equilibrium is impossible in a column because operating a column means putting a both a heat source and a cold sink in it. What we have here is modeled thermodynamically in what's called the "quasi-static approximation", where you take all the first derivatives of temperature as constant (and thus all the second derivatives as zero). This isn't the case when the column is taking off product, but the approximation also works when the second derivatives are small.

And there's no magic discontinuity in temperature at the condenser, either. Condensate at the temperature of the condenser drips onto the packing, cooling it. The top of the column, at steady-state, is close to the temperature of the condenser. The bottom of the column, at steady-state, is close to the temperature of the boiler.
Quote:
If you wait for the vapor to reach the condenser without letting it lose heat on the way up there, it is too late, as the heavy keys will have already come over the top. Best separation occurs when equilibrium is established, with the tops at the light key bp and the bottoms at the heavy key bp.
That's right. My point is that there are two sinks to which the vapor can reject heat: the condenser and the walls. You only need one of them. With a packed column, you can do without the heat loss through the walls. In certain cases you can do without the condenser and just use the walls. That's essentially what a Vigreux column does (*). The best fractionation, however, happens with insulated walls.
Quote:
I was also taught that columns rarely need to be insulated on the bench scale.
That's also right. When the bulk of your heat rejection is through the condenser, the heat loss through the walls makes little difference.

(*) Correction: I should have said "That's what a Vigreux column does when it's run without a condenser."

[Edited on 23-6-2009 by watson.fawkes]

entropy51 - 23-6-2009 at 05:53

I agree with Nicodem. (I may be crazy, but not dumb enough to disagree with him.)

When I worked in industry, all the columns were always insulated. Little columns, big columns, Vigreux columns, packed columns, bubble columns, spinning band columns. As Nicodem says, the best columns come with a vacuum jacket. Sometimes they were heated with heat tapes, for very high boilers.

If you don't believe us, just check Vogel, pg 95 of the 3rd edition:

"Even slight heat losses considerably disturb
the delicate equilibrium of an efficient column, and almost perfect thermal
insulation is required for the separation of components with boiling
points only a few degrees apart. Theoretically, the greatest efficiency
is obtained under adiabatic conditions."

Adiabatic means zero heat loss or gain.

This is not cutting edge science. It's pitiful how many questions that have known answers are debated endlessly back and forth when a quick check of an authoratative reference would settle the issue.

Paddywhacker - 23-6-2009 at 11:37

Watson: you seem to be referring to having the condenser mounted vertically above the fractionating column. The usual arrangement for distillation is for there to be something like a 180 degree bend at the top of the column with the condenser mounted in parallel. It isn't possible for the condenser to affect the functioning of the column. We may indeed have been talking about different things.

The information from the other posters gives me pause for thought. It is at odds with my understanding of things, but I can feel a paradigm shift coming on.

For sure, I have insulated columns myself when I wanted just to get the distillation going.

And I have seen research setups with heated columns ... but my understanding was that heating or insulating the column to decrease its temperature differential was for fine-tuning the efficiency of the column.... something I cannot now think of a reason for doing.

watson.fawkes - 23-6-2009 at 14:00

Quote: Originally posted by Paddywhacker  
Watson: you seem to be referring to having the condenser mounted vertically above the fractionating column.
I was referring to the situation where there's a fractionating head in any configuration that takes part of the condensate and returns it to the column and takes the rest of it out as product. But my argument is true even if the condenser is only in vapor-phase thermal contact with the top of the column. The only difference is the thermal resistance between the condenser and the top of the column. As a rule, liquid thermal transport yields a much smaller temperature difference between the condenser and the top of the column than does vapor transport.
Quote:

And I have seen research setups with heated columns ... but my understanding was that heating or insulating the column to decrease its temperature differential was for fine-tuning the efficiency of the column.... something I cannot now think of a reason for doing.
Let me ask you a diagnostic question. When you say "temperature differential", what are you specifically referring to? It sounds like there are two temperatures that aren't the same. What two things are those?

Paddywhacker - 23-6-2009 at 16:29

temperature differential - the difference in temperature between the bottom of the fractionating column and the top. That which is diminished by insulating the column.

A fractionating column is like a whole series of distillations. If there is no differential then it is if they are all taking place at the same temperature. If there is a differential then it is like a series of distillations at steadily increasing or decreasing temperature. Both will give separation... but which is optimal?

entropy51 - 23-6-2009 at 16:43

If there's no temperature differential between the bottom of the column and the top of the column then there will be no separation. You will just have a conducting Pyrex pipe instead of a fractionating column.

You two could have figured all this out if you'd been reading a few books instead of slugging it out.

Both of you are half right, but the whole is less than the sum of its parts.:D

watson.fawkes - 23-6-2009 at 17:19

Quote: Originally posted by Paddywhacker  
temperature differential - the difference in temperature between the bottom of the fractionating column and the top. That which is diminished by insulating the column.
How is it diminished because the column is insulated? What is the temperature difference between the top of the column and the condenser?

S.C. Wack - 23-6-2009 at 18:46

I have to wonder how many CAREFUL fractionations that the people arguing against lagging have done. Higher temperature ones are difficult without actual heating of the column (due to imperfect insulation) in my hands.
Quote: Originally posted by Paddywhacker  
Insulating the column will decrease it's efficiency, and that is not good for real fractionating.


In volume 4 of Techniques of Organic Chemistry (the 835 page distillation volume), the author is so thoroughly full of shit that pages 345-353 are entirely dedicated to the subject of insulating columns for ordinary fractional distillation in the laboratory. From the first paragraph:

"Internal reflux resulting from excessive heat loss may under some conditions bring about a slight increase in fractionating efficiency, but such results are probably fortuitous to a large extent."

Paddywhacker - 23-6-2009 at 19:13

Quote: Originally posted by watson.fawkes  
Quote: Originally posted by Paddywhacker  
temperature differential - the difference in temperature between the bottom of the fractionating column and the top. That which is diminished by insulating the column.
How is it diminished because the column is insulated? What is the temperature difference between the top of the column and the condenser?


Watson, if the column is operating at maximum efficiency, with the lightest vapour condensing at the top and with no takeoff into the condenser then the condenser is irrelevent. Heating it a bit more to get takeoff isn't going to suddenly make it relevent.

A fractionating column is like a whole series of distillations. If there is no temperature gradient along its length from it being insulated then it is if they are all taking place at the same temperature, so you will get separation. If there is a temperature gradient then it is like a series of distillations at steadily increasing or decreasing temperature, again, giving separation... but which is optimal?

not_important - 23-6-2009 at 19:23

SCW is correct on the need for insulation with higher boiling substances, as well as for distillation under reduced pressure where the less number of molecules per cc means less heat to counteract losses from the column.


Consider that with a proper still head the top of the column delivers vapour to the head's condenser. The condensed liquid is split between being taken off as product (the lesser part) and returned to the column as reflux (the greater part). The condenser is usually a fair amount cooler than the column top, less so if it is very efficient and there is no path that the vapour can take to escape the condenser+column (ie - liquid takeoff of the product). If there is a vapour path out of the head, to a product condenser, the temperature will be determined by the reflux ratio being attempted and is difficult to predict.

This means that the column top must be slightly warmer than the BP of the lower boiling component. The pot temperature, and thus that of the bottom of the column, starts out slightly above the BP of the low boiler. The pot will become hotter as distillation proceeds, ending up as hot as the BP of the next higher boiling component in order to fully vapourise the last of the lower boiling component. The bottom of the column will be near the pot temperature, depending on heat losses between the liquid surface and the bottom of the column; note that this region can contribute one theoretical plate of separation if some condensation occurs within it.

Saerynide - 23-6-2009 at 23:36

Quote:
Condensate at the temperature of the condenser drips onto the packing, cooling it. The top of the column, at steady-state, is close to the temperature of the condenser. The bottom of the column, at steady-state, is close to the temperature of the boiler.
Quote:
If you wait for the vapor to reach the condenser without letting it lose heat on the way up there, it is too late, as the heavy keys will have already come over the top. Best separation occurs when equilibrium is established, with the tops at the light key bp and the bottoms at the heavy key bp.
That's right. My point is that there are two sinks to which the vapor can reject heat: the condenser and the walls. You only need one of them. With a packed column, you can do without the heat loss through the walls. In certain cases you can do without the condenser and just use the walls. That's essentially what a Vigreux column does (*). The best fractionation, however, happens with insulated walls.
(*) Correction: I should have said "That's what a Vigreux column does when it's run without a condenser."

Perhaps we may indeed be talking about different things?

In the industrial setting, the column is cooled by the feed and liquid returning as reflux. The tops are condensed, and split into product and reflux. And so yes, the condenser is a heat extraction device and removes heat from the column.
Quote: Originally posted by watson.fawkes  
This is true only in an out-of-steady-state condition. It keeps heat from leaking out of the column, that true. But thermal loss from the walls of the column is not the only way that the column loses heat. The condenser itself is a heat extraction device. Heat goes in the bottom into the boiler and comes out the top from the condenser.

But in laboratory batch distillation, I can't see how the condenser can participate in refluxing, since, as soon as the vapor enters the condenser, it drips away as product. The condenser could be absent, with the product kept in vapor phase, and not change what goes on in the column. The condenser in this case, extracts heat only from the product, no? Or is there another way to set up a condenser that I don't know about (I'm thinking about a setup like this http://en.wikipedia.org/wiki/File:Fractional_distillation_la...)? Being a batch distillation, is it also not at steady-state, as the tray temperatures change (and so do the compositions) as the distillation continues?
Quote:
A column in a steady-state heat flow condition is not in a state of thermal equilibrium. Thermal equilibrium is the state where everything is the same temperature. I know it's ordinary usage to say that a column under total reflux is in equilibrium, but it's not true in the physics sense of the word equilibrium. Thermal equilibrium is impossible in a column because operating a column means putting a both a heat source and a cold sink in it.

Hmm... I think we are also talking about a different equilibrium. By equilibrium, I mean the vapor and liquid phases at each stage are in equilibrium in the VLE sense.

If there was indeed thermal equilibrium in the column, we would achieve nothing :)

This thread has indeed got me thinking. I understand why insulating continuous distillations improves efficiency and same for high boiling liquids where it is hard enough to get the stuff to even boil, but I still can't seem to see how insulating a laboratory column with low boiling liquids is not detrimental to separation. Where else can the heat go, if not through the walls (a la vigreux column)?

Nicodem - 24-6-2009 at 00:14

Quote: Originally posted by Paddywhacker  

A fractionating column is like a whole series of distillations. If there is no temperature gradient along its length from it being insulated then it is if they are all taking place at the same temperature, so you will get separation. If there is a temperature gradient then it is like a series of distillations at steadily increasing or decreasing temperature, again, giving separation... but which is optimal?

Don't get offended, but this shows your ignorance about the subject. I suggest you to read about vapour pressures above binary mixtures. Their boiling point depends on the composition and this is the origin of the temperature gradient along the column. Lack of insulation only deteriorates such a gradient. Just think logically for a moment instead of refusing to think at all. You said that the separation in a column "is like a whole series of distillations". This is only its theoretical model, though in reality it is about a continuous liquid-vapour equilibration along the path (hence the fillers such as Raschig rings, etc.). Anyway, the composition of the vapours is different than the composition of the liquid (the vapour phase is richer in the lower boiling component and this is the basic phenomenon of distillation at non-azeotropic points). When these vapours condense they form a liquid phase with this same enriched composition. However this phase enriched with the lower boiling component has a lower boiling point. Do you now finally get it where the temperature gradient comes from?

Let's use a specific case. You are separating a 1:1 mixture of toluene (bp 111°C) and benzene (bp 80°C). You can not separate these two with a simple distillation, but a fractionation can separate them. So let's say your fractionation gives a distillate containing 98% benzene and a residue containing 98% toluene. The mixture of 98% toluene / 2% benzene boils at 110°C and the mixture of 98% benzene / 2% toluene boils at 81°C (arbitrary estimations). So, guess what is the temperature gradient along the column when you have 110°C at the start of the column and 81°C at the top? And this is only because you have a composition gradient along the column. Nothing to do with insulation.
Your homework is to evaluate the influence of heat due to lack of insulation and how the resulting non-equilibrium reflux inside the column causes it to loose efficiency. But I warn you that you will need to read about liquid-vapour phase equilibriums to understand this! If you don't learn the science behind it, you will never understand the technology of it.

Magpie - 24-6-2009 at 07:34

from Saerynide:
Quote:

This thread has indeed got me thinking. I understand why insulating continuous distillations improves efficiency and same for high boiling liquids where it is hard enough to get the stuff to even boil, but I still can't seem to see how insulating a laboratory column with low boiling liquids is not detrimental to separation. Where else can the heat go, if not through the walls (a la vigreux column)?


The heat has been taken up by transforming the light liquid from the pot into a vapor at the top of the column, ie, enthalpy of vaporization. This "heat" leaves the column with the light component vapor.

Just thinking about it in this way I can't see why you would not want any column, laboratory or industrial, to not be insulated. This would provide more stable and predictable control, regardless of weather. Also economics would be greatly improved as you would not be losing all that expensive heat to the environment. I'm guessing that for much laboratory work columns are not insulated just because it is not worth the effort.

But my answer doesn't really satisfy me. We all know that a good reflux ratio makes for an efficient separation. A good ratio might be 5:1, ie, there is 5 times as much liquid dropping back into the pot as is being taken off at the condenser. Now with an industrial column I understand that reflux is achieved by taking a large portion of the condensate and re-inserting it into the top of the column. But how is reflux achieved in a laboratory column where such plumbing does not normally exist? The only answer I can come up with is that there has to be heat loss in the column. Is this right?

[Edited on 24-6-2009 by Magpie]

[Edited on 24-6-2009 by Magpie]

[Edited on 24-6-2009 by Magpie]

entropy51 - 24-6-2009 at 09:48

Magpie, I'm not certain that I understand what you're asking, but I think you're asking about the energy lost from vapor that condenses to sustain the reflux. (The heat of evaporation that the vapor must lose to condense and form the liquid reflux flowing downward.) Is that what you meant?

If so, I think that this energy goes into evaporating liquid in the column into the vapor that is leaving to enter the condenser (assuming a downward condenser). That is to say that the liquid of the low boiling component is evaporated as it cools the vapor of the higher boilng component, which is condensed and flows downward as reflux.

All the books I've consulted say that an adiabatic column is best, so it appears that heat loss from the column is not necessary (or desirable) for the operation of the column.

Saerynide - 24-6-2009 at 10:14

Quote:
But my answer doesn't really satisfy me. We all know that a good reflux ratio makes for an efficient separation. A good ratio might be 5:1, ie, there is 5 times as much liquid dropping back into the pot as is being taken off at the condenser. Now with an industrial column I understand that reflux is achieved by taking a large portion of the condensate and re-inserting it into the top of the column. But how is reflux achieved in a laboratory column where such plumbing does not normally exist? The only answer I can come up with is that there has to be heat loss in the column. Is this right?


Magpie: I was trying to reconcile the exact same thing: how heat is lost to achieve reflux with no condenser and reflux drum ala industrial distillation collumns.

I was always under the impression that the temperature of the trays is a function of the rates and temperatures of the feed, reflux and reboiler. The bottoms is the hottest, with each tray getting colder due to colder feed and downcoming liquid as a result of reflux from the top.
Quote:
If so, I think that this energy goes into evaporating liquid in the column into the vapor that is leaving to enter the condenser (assuming a downward condenser). That is to say that the liquid of the low boiling component is evaporated as it cools the vapor of the higher boilng component, which is condensed and flows downward as reflux.


Hmm... Say you have a 50/50 mixture of benzene and toluene in a closed pot. You heat it to, say, 95C. About half by mol is liq with composition of xb ~ 0.4 and the other half exists as vapor with composition of yb ~ 0.65 (see http://upload.wikimedia.org/wikibooks/en/5/52/Benztol_txy_di...)

Now add a perfectly insulated column to the top of the pot, with no condenser. What's at the top of the column? The whole set up should still be at 95C with yb~0.65, all the way to the top of the column, or am I wrong? If I am, it sounds like I should heat a closed perfectly insulated tube containing B and T, to 95C and leave it. And when I come back, I should find one end colder than the other and they will be separated out? :o

According to the Txy digram, how do we go from tray 1 to 2, 3..etc? The vapor from tray 1 flows up to tray 2. True that the vapor has a lower bp, but if it is currently at 95C due to insulation and no condenser, I don't understand why the heavy keys would want to condense and evaporate the light keys?

I was under the impression that tray 2 is lower temp than tray 1 because of the liquid coming down from tray 3. Ultimately, the source of down coming liquid is from the feed and reflux streams. Vapor from tray 1 enters tray 2, where a new VLE is established, and so on.

In a batch chemlab setup, with no feed and the bottom temperature remaining relatively constant, it leaves the reflux to change the tray temperatures. Since in the lab, the vapor at the top is not passed through the condenser then split into reflux and distillate streams (like an industrial column), the condenser plays no role in determining the reflux ratio. Can there really be a temperature profile in the column and reflux occuring spontaneously with no heat lost through the walls? I'm really quite confused now :(
Quote:
All the books I've consulted say that an adiabatic column is best, so it appears that heat loss from the column is not necessary (or desirable) for the operation of the column.

Hmm.. All the books I have also say an adiabatic column is best, but then, they are all chemE books and assumes there is condenser and reflux drum, so that makes sense.

I'm not really familar with the math of bench scale distillations, so I've just been trying to reason it out with what I know from engineering.

[Edited on 6/24/2009 by Saerynide]

entropy51 - 24-6-2009 at 11:37

Since us engineers seem to be talking to each other, I've tried to derive a simple thermodynamic model for a simple column. Consider a distillation of a pure single component with no heat loss or gain from the column. Assume steady state. Let

mi = mass flow rate from boiler to column (boiloff)
mo = mass flow rate from column to condenser (takeoff)
mr = mass flow rate of reflux of saturated liquid from column to boiler

hv = saturated vapor enthalpy
hl = saturated liquid enthalpy

R = Reflux Ratio = mr/mo

by conservation of mass mi = mo + mr

by cons of energy (mi)hv = (mo)hv + (mr)hl

solving for reflux ratio R = ((mi/mo) - 1)(hv/hl)

which requires mi > mo, as does physical intuition. Also hv/hl > 1

This shows that if mass takeoff to condenser mo is held constant,
increasing heat to boiler increases mi which increases R.

If boiloff mi is held constant, then increasing takeoff mo will decrease the reflux ratio R, as we know it must.

This simple model shows that energy is conserved without any heat loss term from the column, and the behavior is qualitatively correct.


[Edit] It appears that hv/hl is about 1.6 for benzene at the boiling point

This equation predicts that if you take off about 25% of the entering benzene vapor to the condenser, you will have the often cited reflux ratio of 5 - without any heat loss.


[Edited on 24-6-2009 by entropy51]



[Edited on 24-6-2009 by entropy51]

[Edited on 24-6-2009 by entropy51]

[Edited on 24-6-2009 by entropy51]

watson.fawkes - 24-6-2009 at 11:39

Quote: Originally posted by Nicodem  
I suggest you to read about vapour pressures above binary mixtures. Their boiling point depends on the composition and this is the origin of the temperature gradient along the column.
@Nicodem. Thank you for the clear exposition. I knew there had to be some fundamental misunderstanding at work; I just didn't know what it was. I'm responding to elaborate on the only point you made that might be unclear: this use of the word "origin". It's not wrong, as such, but it's also got another interpretation.

The way that it's right (and I'm assuming that's the way it's meant) is that if you observe that a column containing a mixture is in a condition of reflux then you must conclude that there's a temperature gradient inside the column. If there's no reflux somewhere in the column, that section of the column might be at constant temperature and filled only with vapor. If there's no mixture in the column, then you might have reflux at the boiling point of the compound. Because of the relationships of boiling points over mixtures you've explained, if you're seeing reflux, then the low boiling fraction must be concentrated at the top and the low boiling fraction concentrated at the bottom (near the boiler).

The way this could be misinterpreted is that putting a mixture into a fractionating column and heating it necessarily leads to a temperature gradient. In other words, the mixture in the column is not the sole origin of the temperature gradient. Consider an absurd example, but one that points out another necessary element. Put a mixture into a fully sealed column at atmospheric pressure and heat it to a constant temperature. Assuming there's reflux in the column (not true if the entire charge vaporizes), that reflux will be a constant percentage mixture at a constant temperature and pressure. Thus what's also necessary is that you not be in thermal equilibrium.

Theoretically speaking, you can do without contact cooling at all and still fractionate. I'll explain. As Magpie pointed out, the heat balance includes enthalpy of vaporization. Suppose you have a perfectly operating thermostat system in the foregoing. Now heat up the column as Magpie mentioned. At exactly the point of first reflux at the top of the column, start the take off. Extract, in vapor form, product from the top of the column and introduce replacement gas (air, say) at the same temperature. Let your thermostat exactly balance the heat input at the boiler and the amount of vapor enthalpy you're taking out at the top. This system fractionates without contact cooling. It's still cooling, from a thermodynamic point of view, because you the heat of vaporization can be extracted from the product. So you can't do without cooling of some form.

Now this system isn't very practical (although it could be constructed), and it's not even as good at fractionating as slower product take off under a condition of reflux return. But as soon as you're returning reflux, you're extracting enthalpy of vaporization at the condenser by heating the coolant in the condenser. This is the active cooling I've been speaking of all along.

Hence my focus this whole time on how the thermodynamic gradients are established with external heating and cooling systems. Such gradients are necessary for good fractionating efficiency.

S.C. Wack - 24-6-2009 at 12:05

elements_of_fractional_distillation_by_robinson_1950.djvu
http://ifile.it/93jyo6p

handbook_of_laboratory_distillation_by_krell_1982.pdf
http://ifile.it/7zrhe6n

laboratory_distillation_practice_by_coulson_1958.pdf
http://ifile.it/1ki8dhm

laboratory_fractional_distillation_by_carney_1949.pdf
http://ifile.it/rwhgc0b

Magpie - 24-6-2009 at 12:21

from entropy51:
Quote:

by cons of energy (mi)hv = (mo)hv + (mr)hl


Shouldn't there be a "+ heat loss" term on the right side? Otherwise there would be no condesation taking place, right?

Here's some relevant information from "Laboratory Text in Organic Chemistry," 1950, by Cason and Rapoport:

p. 234: "Since the fractionating zone in a column should be as nearly adiabatic as possible, the column must either be heated or insulated or both."

Then they go on to describe various "reflux heads." One provides partial reflux, another mentions a "total reflux head." These heads are uninsulated glass tube extentions of the column top and have a thermometer installed vertically from the top.

On p. 244 is shown a reflux head with a cold finger well that is cooled by a small condenser. This is all before the vapor take-off to the main condenser.

So, it looks like the more sophisticated lab fractionation columns can indeed have reflux plumbing.

[Edited on 24-6-2009 by Magpie]

[Edited on 25-6-2009 by Magpie]

entropy51 - 24-6-2009 at 12:27

S C Wack - many thanks for the excellent references! Krell in particular quantitatively considers the deleterious effects of heat loss on column performance. They all seem to say that heat loss is neither necessary nor desirable for maintaining equilibrium in columns.

Magpie, my intent was to show that a physically plausible heat and mass balance results without the heat loss term.

That's my point, that reflux occurs without heat removal - the vapor leaving the top is carrying off heat of evaporation and this results in condensation and reflux.

I said above "Consider a distillation of a pure single component with no heat loss or gain from the column".

Youre right about using a cold finger above the column for reflux, it's often used in industrial laboratories.

[Edited on 24-6-2009 by entropy51]

[Edited on 24-6-2009 by entropy51]

Paddywhacker - 24-6-2009 at 16:08

So, what is the consensus?

That best laboratory practice is to insulate the fractionating column, provide some sort of headspace to keep the column wet, and minimise the takeoff rate?

Saerynide - 24-6-2009 at 23:32

Quote: Originally posted by watson.fawkes  

The way that it's right (and I'm assuming that's the way it's meant) is that if you observe that a column containing a mixture is in a condition of reflux then you must conclude that there's a temperature gradient inside the column. If there's no reflux somewhere in the column, that section of the column might be at constant temperature and filled only with vapor. .... Because of the relationships of boiling points over mixtures you've explained, if you're seeing reflux, then the low boiling fraction must be concentrated at the top and the low boiling fraction concentrated at the bottom (near the boiler).

The way this could be misinterpreted is that putting a mixture into a fractionating column and heating it necessarily leads to a temperature gradient. In other words, the mixture in the column is not the sole origin of the temperature gradient. .... Thus what's also necessary is that you not be in thermal equilibrium.

Thank you for clearing up the semantics. That really had me scratching my head, thinking that either I or he must've been crazy :D

Entropy51: That is a surprising result! :o

Magpie: I couldn't find the book, but I think I found a pic of what you may be talking about :
http://www.labcommerce.com/prod_imagefiles/prod1955/chemglas...

[Edited on 6/25/2009 by Saerynide]

not_important - 24-6-2009 at 23:59

That is a total reflux head, no vapour is removed from the system. A collection flask is attacked at the male joint in the upper-right of the photo, the column connects with the lower joint. The stopcock closest to the column is closed at first, the still pot heated, all vapour is condensed by the cold finger and runs back into the column. Once a steady state condition is reach, with the vapour temperature being what is expected for the pure lower boiling component, that stopcock is opened slightly to allow slow drip of the condensed liquid into the collection flask. The pot temperature will need to be slowly raised as the lower boiling substance is removed; the take off stopcock would be closed if the pot is overheated (vapour temperature goes up) until the proper temperature is regained and the column has re-equilibrated.

Batch fractionation is a slow business, you just can't rush it and expect to get decent separation.

watson.fawkes - 25-6-2009 at 05:32

I'm addressing miscellaneous individual points here. They're all related by thermodynamics and heat transfer. Some points have been addressed (at least partly) and some posters have come to understand. Just consider, then, that I'm posting again for clarity.
Quote: Originally posted by Magpie  

But how is reflux achieved in a laboratory column where such plumbing does not normally exist? The only answer I can come up with is that there has to be heat loss in the column. Is this right?
("such plumbing" is a reflux head.) There's heat loss, but not energy loss, in the column because of vaporization. In other words, you can think of this as (internal) evaporative cooling. Don't take this analogy too far, because it's not exact, but there is an essential similarity. To sustain this cooling, you have to take off product. If you don't, then the pressure rises or you have vapor release, which means you're taking off product without recovering it (=foolish). This effect is the origin of the advice not to overheat the boiler when doing a careful fractionation.
Quote: Originally posted by Saerynide  
Can there really be a temperature profile in the column and reflux occuring spontaneously with no heat lost through the walls?
Yes. You get reflux even in systems in thermal equilibrium. Take a jar of water half full and seal it. You'll eventually see droplets of water on the sides of the jar. That's reflux condensation. Reflux simply means that there's both vaporization and condensation occurring simultaneously. It does not, however, require thermal equilibrium. In a column in a fractionating condition, the steady-state condition includes not only vaporization and condensation, but also net vapor flow up and net condensate flow down. When you model this with a set of a continuity equations. The net result, from top to bottom, implies a temperature difference. So you don't need to lose heat through the walls, but you do have to lose it somewhere: enthalpy of vaporization and (often) cooling at the condenser.
Quote: Originally posted by entropy51  
by cons of energy (mi)hv = (mo)hv + (mr)hl
To be precise, there should also be temperature difference terms. But the model does capture the essentials. If ΔT is small (say, a separation of close-boiling fractions), then the enthalpies of vaporization are much larger and can be disregarded for the purposes of understanding the basics.
Quote: Originally posted by Magpie  
Shouldn't there be a "+ heat loss" term on the right side? Otherwise there would be no condesation taking place, right?
You can't account for condensation without simultaneously accounting for vaporization. This is the countering energy term you're seeking. In adiabatic conditions, the heat lost to vaporization is the same as the heat gained from condensation. The point Nicodem was making was that the instantaneous heat equation is "heat of condensation = heat of vaporization + heat loss in walls", and that highest fractionating efficiency happens when the heat loss in walls is zero.

Magpie - 25-6-2009 at 07:46

Both the heavy and the light components leave the pot as vapor. So how can we have liquid formation in the column without heat loss?

watson.fawkes - 25-6-2009 at 12:09

Quote: Originally posted by Magpie  
Both the heavy and the light components leave the pot as vapor. So how can we have liquid formation in the column without heat loss?
I'm not sure I understand the question. So I'll address two things that may or may not be relevant to what you want. But I should ask you this: Are you thinking about net heat loss (over the column or apparatus) or instantaneous heat loss (at some point in the apparatus).

(Instantaneous) In steady state reflux, the heat lost to the vapor (through condensation into liquid) is the same as heat gained by the liquid (through evaporation into vapor).

(Net) The continuity equation I was talking about was for the interior of the column. You get a different continuity equation at either end. The boiler end adds heat; the condenser end takes it away.

Magpie - 25-6-2009 at 14:34

Edit: I think I have seen the light.

As the fractionating column is being brought into equilibrium a temperature gradient is slowly established between the bottom and top of the column. And here is the key: each successively higher position in the column is a little bit cooler than that just below it. Therefore when the vapor reaches it a little of the mixture condenses, giving its heat to a little of the condensate coming down from the position just above it. The reason for this gradient is that the equilibrium condition at each position gives liquid and vapor slightly richer in the light component as you go up the column.

With this in mind, heat loss from the column seems undesireable.


[Edited on 25-6-2009 by Magpie]

[Edited on 25-6-2009 by Magpie]

entropy51 - 25-6-2009 at 16:46

Magpie, not only have you seen the light, but you managed to explain it much more clearly than I was able to.

I think that it's hard to discuss because it is not only complicated, but that it really requires a mathematical explanation to make it explicitly clear. If you look at the Krell book that SC Wack posted, it becomes clear how complex it is and how many physical phenomena are interacting in a simple column!

Paddywhacker - 25-6-2009 at 22:01

Congratulations gentlemen.

Having changed sides on this issue several times in my life, I now think I have reached the promised land.

Mind you, I'm a sucker for a well-reasoned argument, so the door on my mind isn't completely closed yet.

Paddywhacker - 27-6-2009 at 00:05

There is a patent here
http://www.patentstorm.us/patents/5453167/description.html
that talks about improving the performance of a fractionating column by introducing a higher-boiling mutually-soluble liquid instead of just the condensate.

The example of separating xylenes would be a useful thing to do. Something I would like to do.

But the discussion doesn't run to practicalities. I would think that the distillation flask would fill with this material and bring a halt to the process. Unless you can pump the flask contents back up to the top of the column....

not_important - 27-6-2009 at 04:01

The patent does address practicalities, it states "a continuous column" in a number of places. This means that the high boiling fractions are continuously removed from the reboiler while the low boilers are taken off the top, and feed is constantly introduced - generally within the column. Common industrial scale stuff.

In this case the high boilers would likely be feed to a second column to strip the xylenes from the additive, which would then be feed back to the 1st column.



JohnWW - 27-6-2009 at 05:15

Here are the URLs of some books on distillation that I downloaded some months ago, but I cannot guarantee that the links are still working:

Distillation_Operation.pdf - H.Z. Kister, 31.29 Mb: http://mihd.net/n8rdp2

http://mihd.net/5gkwab - distillation_design_and_control_using_aspen_simulation__-_william_l._luyben.pdf - 103.44 MB

http://ifile.it/h67vulw or http://dl3.s3.ifile.it/r9bs287c/lab_distillation_books.rar 16.98 Mb
- Laboratory Distillation Practice by Coulson E. A. & Herington F. G.(Newnes-1958), & Laboratory Fractional Distillation by C Thomas (Macmillan-1949)

P.S. Some additional ones I have downloaded in the past year or so:

Distillation_Operation.pdf - H.Z. Kister: http://mihd.net/n8rdp2 , 31.29 Mb

http://rs129cg2.rapidshare.com/files/84816426/Kister_-_What_... - In The Last 50 Years - 1,746 Kb - H.Z. Kister

http://rs258cg.rapidshare.com/files/129717333/Kizter_Practic... 23.48 Mb - H Z Kister

http://rs171tl.rapidshare.com/files/150265744/Distillation_D... 14.5 Mb

http://www.4shared.com/file/63740897/f4d7845/Special_Distill... 24.7 Mb


[Edited on 27-6-09 by JohnWW]

User - 27-6-2009 at 06:12

Very nice, its often in the little things :)
The last link is dead , maybe you could upload it again ?
Google didnt return any results when searched for.

JohnWW - 27-6-2009 at 06:21

There is one important point which seems to have been lost sight of here: the difficulty of separation of two liquids by fractional distillation is determined by the shapes of the boundaries between phases in the gas-liquid phase diagram of the mixture over the range of compositions. Such binary phase diagrams, and in some cases ternary phase diagrams, are given especially in the International Critical Tables and in chapter 13 of Perrys Chemical Engineers Handbook (which also covers multicomponent and extractive distillation), links for which are in the References section. Those in which constant-boiling azeotropes are formed, beyond which further purification by ordinary distillation of the mixture is impossible, are listed therein.

To use typical such phase diagrams, alternating vertical and horizontal lines are drawn on the phase diagrams between the two curved boundary lines, starting from the original composition, to determine the number of theoretical plates in a distillation column required for the desired purity of separation. (McCabe-Thiele graphical method). For simple laboratory-scale distillations or preparative-scale chromatographic separations, a further calculation based on diffusion coefficients of the required height of each single theoretical plate for such a distillation or packed column can be made, to arrive at the total required column height.

JohnWW - 27-6-2009 at 07:38

Quote: Originally posted by User  
Very nice, its often in the little things. The last link is dead , maybe you could upload it again? Google didnt return any results when searched for.
S.C. Wack has reposted those two titles on laboratory distillation, above, which were in the dead link.