-
- admin
- Site Admin
-
- Posts: 647
- Joined: Wed Aug 11, 2004 12:19 am
By Anonymous on Saturday, June 5, 2004 - 03:00 pm:
We've been having a debate at work and I am interested to get other opinions. My view is that in most cases (excepting ion chromatography) HPLC analysis can be done without buffers. The way I see it there are two reasons to use a buffer: to change the pH and to stabilize the pH. So let's take them in turn.
If you think about reasons to change the pH two things come to mind. One is lowering the pH to a range where a silica based column is more stable, the second is lowering pH to minimize tailing of basic compounds. But it seems to me that HPLC column technology has reached the point where neither of these issues is that critical. There are several columns on the market (that are silica based) that manufacturer's claim are stable over the full pH range. And with an unbuffered mobile phase the pH would be relatively neutral anyway. Virtually *all* manufacturer's now have columns that can handle basic compounds well without buffers (due to higher purity silica and - especially - polar embedded groups). So it seems there would not be that much need to change the pH.
With respect to keeping pH constant: one would think the pH shouldn't change drastically if the mobile phase is always prepared the same way. And, further, minor changes in pH shouldn't be that critical unless you are at the pKa of one of the analytes.
Anyway, those are my thoughts. I would be interested to see what others have to say.
-------------------------------------------------------------------------------------------------------
By Anonymous on Saturday, June 5, 2004 - 04:16 pm:
Why ask the question? Try it, see what's happening and tell us what you see!
-------------------------------------------------------------------------------------------------------
By Consumer Products Guy on Saturday, June 5, 2004 - 07:29 pm:
Mostly (95%) we use water-methanol or water-acetonitrile as HPLC mobile-phase (all HPLC-filtered grade); with acidic or phenolic analytes we add some phosphoric or acetic acid to the water. This gives us the advantages (1) we don't need to prefilter as would be required with adding solid buffers (2) we can shut off the pump through the software instead of the necessity of keeping some flow going, then flushing, as would be required with dissolved-solid buffers (to protect the equipment). When necessary (5%), we do use dissolved-solid buffers, and have to hassle with filtration, keeping flow going, flushing, etc., but that's time-consuming and I've got a lot to do !!!
-------------------------------------------------------------------------------------------------------
By bill on Sunday, June 6, 2004 - 05:13 am:
suggested experiemnts to test your hypothesis:
try terephthalic acid, methanol/water. you will need to adjust pH of sample so that terephthalic acid is partially neutralized to be able to get it into solution. this experiment will illustrate how adequate buffering(or not adequate buffering) of the eluant will compensate for pH conditions in a sample that are inappropriate for the chromatography. Repeat with phosphoric acid in the eluant and compare difference. Don't do therephathalic acid in your lab? Maybe not, but many labs receive samples that are inappropriate pH for the separation. Buffering of mobile phase is labor saving way to cure the problem.
Next try a mixture of simple aliphatic acids, say acetic, propionic, etc. in just methanol/ water with and with out phosphoric acid added to mobile phase. This experiment will illustrate that retention and separation of acids is greatly affected by mobile phase pH.
As you gain experience you will no doubt encounter other good examples, presuming you encounter the separation of acdic and basic compounds.
-------------------------------------------------------------------------------------------------------
By Anonymous on Sunday, June 6, 2004 - 10:10 am:
Consumer Products Guy
I found your response very interesting. I have often wondered why the large-molecule people can simply add 0.1% TFA whereas us small-molecule folks need to prepare buffers (I am familiar with the issue of TFA as an ion pair reagent - so no need to mention that). But it seems that it would be so much easier to - as you say - just add some acid. It's easier and there is less concern with filtration, flushing out the column, etc (all the points you made). I guess the counter argument is that by doing things this way we lower the pH but we don't do anything to stabalize it. But then if the method is, for example, for a pharmacuetical analysis, where the components are always the same, it would seem to me that if it works during validation you're home free.
Bill
Thanks for your response. One additional thought I had (I decided not to mention it in my original message as it was already too long) there are reversed phase columns now with polar groups included in the C-18 chains. I would imagine this may allow one to work without buffers - in some cases - even with acidic and basic compounds. But I certainly was not suggesting that buffers are no longer an important issue in HPLC (that would be way off base).
-------------------------------------------------------------------------------------------------------
By Anonymous on Sunday, June 6, 2004 - 05:44 pm:
I have experience with pH of mobile phase in vitamins analysis. In low pH, acidic compounds, such as ascorbate, will be eluted behind thiamine and nicotinamide, but in higher pH (6 - 7), it will be eluted faster than thiamine and nicotinamide.
So, I think pH still has an important role in separation technique by HPLC.
-------------------------------------------------------------------------------------------------------
By HW Mueller on Monday, June 7, 2004 - 12:36 am:
No flushing of column necessary after work with H3PO4, AcOH mobile phases??
Very often the quicky methods quickly lead to wrong or at least questionable results. Your nonbuffered solutions might work fine under somewhat idealized developmental conditions, then come real, maybe difficult samples, and your nice quicky fails without warning or any indication of failure.
The "simple" 0.1% TFA failed so often with macromolecules in this lab that we havn´t touched it for years. Just personal experience: The bigger the molecule the more likely the quackery.
-------------------------------------------------------------------------------------------------------
By MG on Monday, June 7, 2004 - 07:09 am:
Just to add my two cents: I do exclusively LC/MS, very seldom with concurrent UV detection. My compounds of interest usually have acidic or basic functionality. I have found empirically that it is not usually necessary to have a true buffer, but it *is* necessary to add "a little bit of acid" as others have posted, to control the pH. Without any sort of pH control, even on modern endcapped columns, I have found that acidic and basic compounds will have horrible peak shape.
-------------------------------------------------------------------------------------------------------
By Anonymous on Monday, June 7, 2004 - 05:14 pm:
If we add some acid, without a true buffer system, and bring the pH to say 3; I think in a sense we are still buffering. The presence of acid should cause a resistance to pH change from addition of base. It's true that there is nothing to resist the lowering of pH by addition of acid. But we are already at 3. Doesn't seem likely that stuff in the sample (usually at ppm levels) would push the pH down further than that.
-------------------------------------------------------------------------------------------------------
By Uwe Neue on Monday, June 7, 2004 - 05:41 pm:
Control of the ionization of the analytes is important to get reproducible chromatography. If you can do this by just adding some mild acid (pH 3) without true buffering, this might be perfectly fine for most of the analytes. It may not be good enough for amino acids, though.
If you are in the intermediate pH range and want to get control of the ionization of the analytes, you will be forced in most cases to use a real buffer.
Remember, the retention time difference between the ionized (salt) form and the nuetral form of most compounds is about a factor of 30 in most cases. This is not a good thing, if this is changing - for whatever reason.
In addition, you need to remember that an ion needs a counterion to migrate reproducibly. I bet that you will get quite interesting results by mixing an acidic drug (in the salt form) with a basic drug (in the salt form) and chromatograph the two together without a buffer. If I had a student in my lab with nothing good to do, I would have him make this experiment...
-------------------------------------------------------------------------------------------------------
By HW Mueller on Tuesday, June 8, 2004 - 02:06 am:
Anon, June7, 5:14pm,
take a look at some buffer index (buffer capacity) curves. At pH = 3 (or especially at pH = 2), attained via a nonbuffering acid, you get very strong buffering on adding more acid, none on adding base. This is due to buffering by the water equilibrium, actually.
Theoretically, one can fathom that you could have a stable situation if your chosen pH essentially is shifting your system far out of equilibrium, possible??
-------------------------------------------------------------------------------------------------------
By Anonymous on Tuesday, June 8, 2004 - 04:13 am:
I think there is some nonsense in this thread about "true buffers" and the like. Look up the titration curve of any strong acid in a high school text book -HCl will do. What happens to 0.1M HCl when you add some acid-nothing much. What happens when you add base-even 0.1M NaOH? Nothing much either as long as you're not near the end point. So 0.1M HCl is a reasonable buffer. The same is true of TFA-it's a perfectly good buffer for HPLC at about 0.1% concentration (pH about 2.3). That's why "simply adding TFA" works - because it's a perfectly good buffer under these conditions.
-------------------------------------------------------------------------------------------------------
By HW Mueller on Tuesday, June 8, 2004 - 08:06 am:
That´s what buffers are for, to reduce the "nothing much" to considerable less.
We do not need a re-invention of buffering.
A practical definition used to be that one has buffering when, as an example, one adds a given amount of H+ to water, but gets a considerably lower concentration of H+ at equilibrium.
Incidentally, if you plot a titration curve as concentration vs concentration then you will see that the "nothing much" is actually quite a bit at most parts of the curve (exceptions are noted in connection with the buffer index, above).
-------------------------------------------------------------------------------------------------------
By A.Mouse on Tuesday, June 8, 2004 - 07:04 pm:
At pH 3, the pure buffer capacity of a strong acid such as TFA is close to nothing: the concentration is about 1 mM. The story is different at pH 2: there, a strong acid has a concentration of 10 mM, comparable to the buffer concetrations that we commonly use in LC/MS.
-------------------------------------------------------------------------------------------------------
By bill tindall on Tuesday, June 8, 2004 - 08:17 pm:
I. M. Kohthoff, who some call the father of modern analytical chemistry and Roger Bates of NIST(NBS), the most trusted name in pH, both pointed out in books (and in my case classes) that strong acids and strong bases are buffers at low and high pH respectively. Furthermore, at a given concentration, they are a better buffers in the concentration range where they are effective, than buffers prepared from weak acids and their salts.
It follows that dilute phosphoric is an excellent buffering agent for the reproducible separation of acidic compounds and I have done hundreds if not thousands of them. In the begining we prepared these eluants from mixtures of phosphoric and potassium phosphate but for the reasons eloquently stated in a previous post above it was much better to simply use phosphoric acid. If one does a Scifinder search of my publications one will find several papers on successful application of this approach to acid determination used daily in various industrial applications.
Therefore, from both a theoretical and practical consideration, strong acids and bases are good buffers that have advantages over acid/salt buffers in some applications. There are cases where you can save your company $$$ by using them, and thats the bottom line.
-------------------------------------------------------------------------------------------------------
By Anonymous on Tuesday, June 8, 2004 - 08:47 pm:
Seems like we are reaching a consensus. In some cases neither acid or buffer is needed. In other cases simply adding some acid will suffice (or base in less common applications). And in a few instances a true buffer is needed: mostly when it is necessary to stabilize the pH in a more intermediate range.
Sound like a fair summary?
-------------------------------------------------------------------------------------------------------
By bill tindall on Wednesday, June 9, 2004 - 04:46 am:
Anon,
I am not sure from your summary if you understand when buffering is essential becasue you didn't specifically mention acids and bases. This would be my summary.....
1. Something like acetate or benzoate has almost no retention in any sort of reversed phase separation, polar embedded column or not. At the sorts of concentrations that separations are made at, something like acetic acid or benzoic acid will be essentially fully ionized to acetate and benzoate in just water and hence not retained. To get retention one acidifies the eluant to supress dissociation. The buffer objective in this case is to get subtatntial to near complete supression of ionization. The undissociated acids are nicely retained on many reversed phase columns. The phosphoric acid that might be used in such a case to get pH of 2.2(for example) is as "true" a buffer as one prepared from a mixture of phosphoric acid and sodium phosphate to get a pH of 2.2.
Occasionally one encounteres an unresoved pair of components. If they have different pKa's or pKb's they can be separated by diddling with the pH of the eluant to make one more or less ionized relative to the other. In this case the buffering objective is to create differences in dissociation between two components of the separation.
3. For cases where the components to be separated are neither acids or bases a buffer will likely have no effect on the separation.
-------------------------------------------------------------------------------------------------------
By HW Mueller on Wednesday, June 9, 2004 - 07:21 am:
(Bill, you know of course, that to get a 2.2 buffer with H3PO4 one is limited to one conc. (unless you add other species), whereas if you make it with H3PO4 + phosphates you have a very wide choice of concentrations......but that is getting near hair splitting again).
Generally, one gets a quantitative comparison of buffering capability by using the buffer index (capacity) equations, as mentioned above.
As to the situations where buffering is needed: In addition to the equilibria mentioned here, we are just having some cases with ions (not equilibrating with neutral species) which require pH control of the SiOH dissociation (discussed before?).
-------------------------------------------------------------------------------------------------------
By Anonymous on Wednesday, June 9, 2004 - 07:27 am:
Bill Tindall-many thanks for your reference to Kolthoff which points out that a strong acid is a good buffer at low pH.
A. Mouse- I wouldn't use 1mm TFA at pH 3 either. Most people use 0.1% TFA which has a pH about 2.3 and is a good buffer.
HWM- the buffer capacity of 0.1 % TFA is not far different from that of a mixture of phosphoric acid and KH2PO4 of similar molar concentration. I was being conservative by saying "not much" change. You could calculate this, or use one of the many programs available. Alternatively, you could try this with a burette and a pH meter.
-------------------------------------------------------------------------------------------------------
By Anonymous on Wednesday, June 9, 2004 - 12:40 pm:
Bill (or anyone else who feels qualified). I just want to followup on your comment that strong acids and bases are buffers. Please tell me where my logic is wrong (I wouldn't be at all surprised if it is).
I always thought that buffers were composed of weak acids or bases. And I always thought the reason was that with a strong acid or base, the conjugate is extremely weak. Let me illustrate with an example. If we think of the strong acid HCL. Its conjugate base is Cl-. But this is a *very* weak base. In fact most texts would call it a spectator ion cause it doesn't really do anything. So the question becomes: does a solution of HCl really have any ability to consume additional acid (hence act as a buffer). Or won't the Cl- ion just float around and do nothing.
Thanks. Once I get this straight I am set on this topic. Thanks to everyone for their valuable input.
-------------------------------------------------------------------------------------------------------
By bill tindall on Wednesday, June 9, 2004 - 02:43 pm:
Addressing the last Anon. question.........I wish you people had names. Cripes, call your self Joe and at least I could keep straight who was asking what. Pleased to discuss the topic...
The problem is that most people are only exposed to a limited definition of buffers, that which you stated. You need an accurate definition of a buffer and then all would be clear. "A buffer is something that resists changes in pH as a small increment of acid or base is added". It has a mathematical definition. It is the slope of the titration curve or d(Cb)/d(pH) where d(Cb) is a differential increment of base added. In any part of a titration curve that is flat the solution is highly buffered. This discussion can be found in any solution thermodynamics book or most advanced analytical chem books.
Let us take an extreme case to illustrate the point. Suppose we have a liter of water and a liter of 1 M HCl. Now we add 0.01 moles of sodium hydroxide to each. The water pH would zoom from 6 (CO2 present)to 12 or something like that. Clearly not buffered.
It is doubtful that one could even detect the pH change of the 1 M HCl solution after neutralizing 1 %. Hence, clearly this solution is highly buffered according to the general definition of buffers.
If one does the calculations it will be found that if buffers are needed at either the acid or base end of the pH range, a strong acid or a strong base is a better buffer than an acid/ salt pair of equivalent concentration. As a general guideline for pH 2 and less use a strong acid. Above 3 use an acid and its conjugate base. We can argue about what to use between 2 and 3 as it will depend on the objective of adding the buffer.
-------------------------------------------------------------------------------------------------------
By Anonymous on Wednesday, June 9, 2004 - 04:52 pm:
At the risk of being a pain, let me just "pick it apart" one last time.
(note: this question has to do with semantics of terminology and pH theory and has absolutely no practical significance. Anyone who's not interested in that - don't bother reading further)
Taking your example above: what if we add a small amount of acid to the 1M HCl solution. There's nothing to consume the additional acid. Having said that, I think it's also true that the additional acid will not cause much of a change in pH *but* it's because the H+ concentration is already high and a small amount of additional H+ doesn't consitute a significant change. It is not technically due to a buffering effect.
So - for someone like me who likes to be a huge nuisance about semantics - it would seem that a strong acid is only a unidirectional buffer. It resists change in pH when a base is added but not when more acid is added. But it experiences little change either way (and vica versa for a strong base).
My head hurts. I should have been a priest.
-------------------------------------------------------------------------------------------------------
By HW Mueller on Friday, June 11, 2004 - 12:06 am:
Last Anon:
Get a book on chemical equilibrium and look at pH index curves! A 1M HCl solution has a pH near 0. The buffering at such a low pH is enormous, it is real buffering and it´s done by the water equilibrium. If the water didn´t equilibrate you would have an increase in H+ CONCENTRATION corresponding directly to the amt of H+ added, conversly, you would have a H+ conc decrease corresponding directly to the amt of base added (by definition of strong acid or base).
Now what I said before about adding base refers to a pH of 2 and especially above. If you have only a completely dissociated acid present you will quickly raise the pH to regions where there is no buffering at all (adding base, NOT ACID, is dangerous in respect to buffering in such cases). Again, with "no buffering" I mean that you change the H+ Concentration by x moles if x moles of strong acid or strong base (completely dissociating) are added. Of course, if you add infinitally small amt of base you still have fairly good buffering even a bit above pH 2, but robustness is not about infinatly small amts.
-------------------------------------------------------------------------------------------------------
By bill on Friday, June 11, 2004 - 02:50 pm:
Last Anon: You want to think of buffering as a process- consuming the increments of acid or base added. This concept is useful in the mid ranges of pH but it does not explain buffering at the extremes of pH. Hence, it is not a useful concept to describe the general phenomena of buffering. And, buffers are not defined by a process, but by a result.
The general definition of buffering is simply resisting a change in pH and no mechanism of how that is accomplished is part of the definition.
Buffering = d(pH)/d(Cb).
It is a fact that when a strong acid is titrated the pH doesn't change much until near the end point. Ditto for a strong base. So, there are at least two processes to accomplish buffering. Consuming the increments of added acid or base (mid range buffering) and overwhelming them (pH extremes process)
-------------------------------------------------------------------------------------------------------
By Chris Pohl on Saturday, June 12, 2004 - 04:32 pm:
One more perspective on this topic:
Contrary to what was suggested in the first sentence at the beginning of this thread, buffers (in the more classical sense of the definition, anyway) are actually relatively rarely used in ion chromatography. And even if they are, this aspect is for the most part largely irrelevant. The fact is that when the pH of the mobile phase is sufficient to render the analyte completely ionized (as is most commonly the case in ion chromatography) or completely neutral (as is often the case for acids retained by ion suppression in reverse phase chromatography), it really doesn't matter much what the pH of the mobile phase is, over a fairly broad range. In the case of ion chromatography, what generally matters is the electrolyte concentration. The associated pH is generally a largely irrelevant characteristic of the mobile phase, merely a consequence of the specific ionic strength required for the separation.
So, from my point of view, one can divide the answer to the question into three parts:
1). If all analytes are neutral over a pH range encompassed by a neutral mobile phase (actually, a fairly wide range of mobile phase pH values is possible in this case), no buffer or electrolyte of any sort is necessary.
2). If all analytes are being separated under conditions where changes in pH have essentially no effect on the ionization of the analytes, then buffer control is not necessary although ionic strength control may be necessary if analytes are fully ionized under separation conditions. In this case, it is simply necessary to have the right pH and a buffer is not necessary. If, for example, a set of cationic analytes are fully ionized at pH 4, pH 5 or even pH 6, then a mobile phase consisting of 0.1 mM HCl would work just as well as a true buffer chosen to buffer the system at pH 4 (hopefully, we can all agree that 0.1 mM HCl is not acting as a buffer at pH 4).
3). If analytes are being separated under conditions where at least one of the analytes is partially ionized, some sort of buffer control is required. Here I defer to the discussion above as to what comprises a buffer but my point is that only under these circumstances is buffering of the mobile phase advisable in order to control separation selectivity.
-------------------------------------------------------------------------------------------------------
By Anonymous on Sunday, June 13, 2004 - 03:38 pm:
Chris
Thanks for the clarification about ion chromatography. I see your point there. Regarding your summary, I like part 1 and 2. Part 3 seems a little contradictory with #2. 3 says 'if any analyte is partially ionized you need buffer'. But I don't think that's exactly true. I think you state it well in #2 that if the pH does not need to be changed or stabilized a buffer is not necessary. This could be the case with a partially ionized analyte.
-------------------------------------------------------------------------------------------------------
By HW Mueller on Monday, June 14, 2004 - 03:04 am:
Oops, Bill, you confused me for a second, as you have the inverse formular of that given in Butler, Ionic Equilibrium, Addison-Wesley, 1964:
Buffer index (buffer capacity) = dC/dpH
This doesn´t matter except if one gives some values like these for this equation:
0.01M HOAC// 0.01M TFA
pH1 0.230// 0.233
pH2 0.0231// 0.0234
pH3 0.00269// 0.00234
pH4 0.00315// 0.0023
pH5 0.00535// 0.000023
pH6 0.00118// 0.0000024
pH7 0.000131// 4.6x10E-7
a buffer index of ~0.0023 corresponds about to non-buffering here (in my definition above: One adds X moles of acid or base and gets X moles of H+ conc. change)
ARRGHH, tried to write a table, didn´t work, so I put // between value columns for the two acids,
-------------------------------------------------------------------------------------------------------
By Chris Pohl on Monday, June 14, 2004 - 10:50 am:
Anon (6-13-04),
Regarding your comments surrounding my point #3:
I guess, in retrospect, one could divide this case into two sub-cases:
3a. The analyte is partially ionized and the pKa or pKb is less than 4. In this case, buffering per se should be unnecessary. pH control using a defined concentration of hydroxide or hydronium ion would be sufficient.
3b. The analyte is partially ionized and the pKa or pKb is greater than 4. In this case, I really think you need to utilize a buffer since one cannot accurately produce a given pH between pH4 and 10 without the use of buffers. Assuming that analyte ionization affects retention (which is always true in my experience for reversed phase and ion exchange), good separation reproducibility necessitates reproducible pH control.
-------------------------------------------------------------------------------------------------------
By Anonymous on Monday, June 14, 2004 - 04:26 pm:
check
-------------------------------------------------------------------------------------------------------
By HW Mueller on Tuesday, June 15, 2004 - 12:01 am:
Darn, this doesn´t end, but Chris I have trouple understanding your detailing of 3.
An example of why: If you get a robust method of analyzing an acid in its undissociated form at lets say pH=3 (without buffering) then it should be more robust if the acid´s pKa is greater than 4.
-------------------------------------------------------------------------------------------------------
By Fred on Saturday, August 7, 2004 - 03:48 pm:
Anyone know of some good references discussing buffers. I have seen references that talk about various buffers and their pKa's but I've never seen one that also talks about preparation of buffers, and which buffers a good for use with mass spec or evaporative light scattering (i.e. volatile buffers). Basically "everything the liquid chromatography ever need to know about buffers"
Thanks
Fred
-------------------------------------------------------------------------------------------------------
By Kostas Petritis on Saturday, August 14, 2004 - 02:42 pm:
Fred,
There are several relevant references out there.
1) Wilson N.S. et al. "Buffers and Baselines" LC-GC June 2001.
2) Tindal G.W. "Mobile phase buffers. Part I-The interpretation of pH in partially aqueous mobile phases. LC GC North America, Novemember 2002
3)Tindal G.W. " Mobile-phase buffers, Part II- Buffer selection and capacity, LC GC North America, December 2002.
4) Petritis, K. et. al. Volatility evaluation of mobile phase-electrolyte additivites for mass spectrometry. LC GC Europe, February 2002.
You can find some more relevant articles in the above references. There are also some relevant books (general about buffers and pH), on-line literature (mostly commercial i.e. Radiometer Analytica) and some commercial software that can calculate buffer relative values (i.e. buffer capacity, ionic force etc.) for HPLC and CE -personally I like the software "pHoebus".
Hope the above helps,
Kostas
We've been having a debate at work and I am interested to get other opinions. My view is that in most cases (excepting ion chromatography) HPLC analysis can be done without buffers. The way I see it there are two reasons to use a buffer: to change the pH and to stabilize the pH. So let's take them in turn.
If you think about reasons to change the pH two things come to mind. One is lowering the pH to a range where a silica based column is more stable, the second is lowering pH to minimize tailing of basic compounds. But it seems to me that HPLC column technology has reached the point where neither of these issues is that critical. There are several columns on the market (that are silica based) that manufacturer's claim are stable over the full pH range. And with an unbuffered mobile phase the pH would be relatively neutral anyway. Virtually *all* manufacturer's now have columns that can handle basic compounds well without buffers (due to higher purity silica and - especially - polar embedded groups). So it seems there would not be that much need to change the pH.
With respect to keeping pH constant: one would think the pH shouldn't change drastically if the mobile phase is always prepared the same way. And, further, minor changes in pH shouldn't be that critical unless you are at the pKa of one of the analytes.
Anyway, those are my thoughts. I would be interested to see what others have to say.
-------------------------------------------------------------------------------------------------------
By Anonymous on Saturday, June 5, 2004 - 04:16 pm:
Why ask the question? Try it, see what's happening and tell us what you see!
-------------------------------------------------------------------------------------------------------
By Consumer Products Guy on Saturday, June 5, 2004 - 07:29 pm:
Mostly (95%) we use water-methanol or water-acetonitrile as HPLC mobile-phase (all HPLC-filtered grade); with acidic or phenolic analytes we add some phosphoric or acetic acid to the water. This gives us the advantages (1) we don't need to prefilter as would be required with adding solid buffers (2) we can shut off the pump through the software instead of the necessity of keeping some flow going, then flushing, as would be required with dissolved-solid buffers (to protect the equipment). When necessary (5%), we do use dissolved-solid buffers, and have to hassle with filtration, keeping flow going, flushing, etc., but that's time-consuming and I've got a lot to do !!!
-------------------------------------------------------------------------------------------------------
By bill on Sunday, June 6, 2004 - 05:13 am:
suggested experiemnts to test your hypothesis:
try terephthalic acid, methanol/water. you will need to adjust pH of sample so that terephthalic acid is partially neutralized to be able to get it into solution. this experiment will illustrate how adequate buffering(or not adequate buffering) of the eluant will compensate for pH conditions in a sample that are inappropriate for the chromatography. Repeat with phosphoric acid in the eluant and compare difference. Don't do therephathalic acid in your lab? Maybe not, but many labs receive samples that are inappropriate pH for the separation. Buffering of mobile phase is labor saving way to cure the problem.
Next try a mixture of simple aliphatic acids, say acetic, propionic, etc. in just methanol/ water with and with out phosphoric acid added to mobile phase. This experiment will illustrate that retention and separation of acids is greatly affected by mobile phase pH.
As you gain experience you will no doubt encounter other good examples, presuming you encounter the separation of acdic and basic compounds.
-------------------------------------------------------------------------------------------------------
By Anonymous on Sunday, June 6, 2004 - 10:10 am:
Consumer Products Guy
I found your response very interesting. I have often wondered why the large-molecule people can simply add 0.1% TFA whereas us small-molecule folks need to prepare buffers (I am familiar with the issue of TFA as an ion pair reagent - so no need to mention that). But it seems that it would be so much easier to - as you say - just add some acid. It's easier and there is less concern with filtration, flushing out the column, etc (all the points you made). I guess the counter argument is that by doing things this way we lower the pH but we don't do anything to stabalize it. But then if the method is, for example, for a pharmacuetical analysis, where the components are always the same, it would seem to me that if it works during validation you're home free.
Bill
Thanks for your response. One additional thought I had (I decided not to mention it in my original message as it was already too long) there are reversed phase columns now with polar groups included in the C-18 chains. I would imagine this may allow one to work without buffers - in some cases - even with acidic and basic compounds. But I certainly was not suggesting that buffers are no longer an important issue in HPLC (that would be way off base).
-------------------------------------------------------------------------------------------------------
By Anonymous on Sunday, June 6, 2004 - 05:44 pm:
I have experience with pH of mobile phase in vitamins analysis. In low pH, acidic compounds, such as ascorbate, will be eluted behind thiamine and nicotinamide, but in higher pH (6 - 7), it will be eluted faster than thiamine and nicotinamide.
So, I think pH still has an important role in separation technique by HPLC.
-------------------------------------------------------------------------------------------------------
By HW Mueller on Monday, June 7, 2004 - 12:36 am:
No flushing of column necessary after work with H3PO4, AcOH mobile phases??
Very often the quicky methods quickly lead to wrong or at least questionable results. Your nonbuffered solutions might work fine under somewhat idealized developmental conditions, then come real, maybe difficult samples, and your nice quicky fails without warning or any indication of failure.
The "simple" 0.1% TFA failed so often with macromolecules in this lab that we havn´t touched it for years. Just personal experience: The bigger the molecule the more likely the quackery.
-------------------------------------------------------------------------------------------------------
By MG on Monday, June 7, 2004 - 07:09 am:
Just to add my two cents: I do exclusively LC/MS, very seldom with concurrent UV detection. My compounds of interest usually have acidic or basic functionality. I have found empirically that it is not usually necessary to have a true buffer, but it *is* necessary to add "a little bit of acid" as others have posted, to control the pH. Without any sort of pH control, even on modern endcapped columns, I have found that acidic and basic compounds will have horrible peak shape.
-------------------------------------------------------------------------------------------------------
By Anonymous on Monday, June 7, 2004 - 05:14 pm:
If we add some acid, without a true buffer system, and bring the pH to say 3; I think in a sense we are still buffering. The presence of acid should cause a resistance to pH change from addition of base. It's true that there is nothing to resist the lowering of pH by addition of acid. But we are already at 3. Doesn't seem likely that stuff in the sample (usually at ppm levels) would push the pH down further than that.
-------------------------------------------------------------------------------------------------------
By Uwe Neue on Monday, June 7, 2004 - 05:41 pm:
Control of the ionization of the analytes is important to get reproducible chromatography. If you can do this by just adding some mild acid (pH 3) without true buffering, this might be perfectly fine for most of the analytes. It may not be good enough for amino acids, though.
If you are in the intermediate pH range and want to get control of the ionization of the analytes, you will be forced in most cases to use a real buffer.
Remember, the retention time difference between the ionized (salt) form and the nuetral form of most compounds is about a factor of 30 in most cases. This is not a good thing, if this is changing - for whatever reason.
In addition, you need to remember that an ion needs a counterion to migrate reproducibly. I bet that you will get quite interesting results by mixing an acidic drug (in the salt form) with a basic drug (in the salt form) and chromatograph the two together without a buffer. If I had a student in my lab with nothing good to do, I would have him make this experiment...
-------------------------------------------------------------------------------------------------------
By HW Mueller on Tuesday, June 8, 2004 - 02:06 am:
Anon, June7, 5:14pm,
take a look at some buffer index (buffer capacity) curves. At pH = 3 (or especially at pH = 2), attained via a nonbuffering acid, you get very strong buffering on adding more acid, none on adding base. This is due to buffering by the water equilibrium, actually.
Theoretically, one can fathom that you could have a stable situation if your chosen pH essentially is shifting your system far out of equilibrium, possible??
-------------------------------------------------------------------------------------------------------
By Anonymous on Tuesday, June 8, 2004 - 04:13 am:
I think there is some nonsense in this thread about "true buffers" and the like. Look up the titration curve of any strong acid in a high school text book -HCl will do. What happens to 0.1M HCl when you add some acid-nothing much. What happens when you add base-even 0.1M NaOH? Nothing much either as long as you're not near the end point. So 0.1M HCl is a reasonable buffer. The same is true of TFA-it's a perfectly good buffer for HPLC at about 0.1% concentration (pH about 2.3). That's why "simply adding TFA" works - because it's a perfectly good buffer under these conditions.
-------------------------------------------------------------------------------------------------------
By HW Mueller on Tuesday, June 8, 2004 - 08:06 am:
That´s what buffers are for, to reduce the "nothing much" to considerable less.
We do not need a re-invention of buffering.
A practical definition used to be that one has buffering when, as an example, one adds a given amount of H+ to water, but gets a considerably lower concentration of H+ at equilibrium.
Incidentally, if you plot a titration curve as concentration vs concentration then you will see that the "nothing much" is actually quite a bit at most parts of the curve (exceptions are noted in connection with the buffer index, above).
-------------------------------------------------------------------------------------------------------
By A.Mouse on Tuesday, June 8, 2004 - 07:04 pm:
At pH 3, the pure buffer capacity of a strong acid such as TFA is close to nothing: the concentration is about 1 mM. The story is different at pH 2: there, a strong acid has a concentration of 10 mM, comparable to the buffer concetrations that we commonly use in LC/MS.
-------------------------------------------------------------------------------------------------------
By bill tindall on Tuesday, June 8, 2004 - 08:17 pm:
I. M. Kohthoff, who some call the father of modern analytical chemistry and Roger Bates of NIST(NBS), the most trusted name in pH, both pointed out in books (and in my case classes) that strong acids and strong bases are buffers at low and high pH respectively. Furthermore, at a given concentration, they are a better buffers in the concentration range where they are effective, than buffers prepared from weak acids and their salts.
It follows that dilute phosphoric is an excellent buffering agent for the reproducible separation of acidic compounds and I have done hundreds if not thousands of them. In the begining we prepared these eluants from mixtures of phosphoric and potassium phosphate but for the reasons eloquently stated in a previous post above it was much better to simply use phosphoric acid. If one does a Scifinder search of my publications one will find several papers on successful application of this approach to acid determination used daily in various industrial applications.
Therefore, from both a theoretical and practical consideration, strong acids and bases are good buffers that have advantages over acid/salt buffers in some applications. There are cases where you can save your company $$$ by using them, and thats the bottom line.
-------------------------------------------------------------------------------------------------------
By Anonymous on Tuesday, June 8, 2004 - 08:47 pm:
Seems like we are reaching a consensus. In some cases neither acid or buffer is needed. In other cases simply adding some acid will suffice (or base in less common applications). And in a few instances a true buffer is needed: mostly when it is necessary to stabilize the pH in a more intermediate range.
Sound like a fair summary?
-------------------------------------------------------------------------------------------------------
By bill tindall on Wednesday, June 9, 2004 - 04:46 am:
Anon,
I am not sure from your summary if you understand when buffering is essential becasue you didn't specifically mention acids and bases. This would be my summary.....
1. Something like acetate or benzoate has almost no retention in any sort of reversed phase separation, polar embedded column or not. At the sorts of concentrations that separations are made at, something like acetic acid or benzoic acid will be essentially fully ionized to acetate and benzoate in just water and hence not retained. To get retention one acidifies the eluant to supress dissociation. The buffer objective in this case is to get subtatntial to near complete supression of ionization. The undissociated acids are nicely retained on many reversed phase columns. The phosphoric acid that might be used in such a case to get pH of 2.2(for example) is as "true" a buffer as one prepared from a mixture of phosphoric acid and sodium phosphate to get a pH of 2.2.
Occasionally one encounteres an unresoved pair of components. If they have different pKa's or pKb's they can be separated by diddling with the pH of the eluant to make one more or less ionized relative to the other. In this case the buffering objective is to create differences in dissociation between two components of the separation.
3. For cases where the components to be separated are neither acids or bases a buffer will likely have no effect on the separation.
-------------------------------------------------------------------------------------------------------
By HW Mueller on Wednesday, June 9, 2004 - 07:21 am:
(Bill, you know of course, that to get a 2.2 buffer with H3PO4 one is limited to one conc. (unless you add other species), whereas if you make it with H3PO4 + phosphates you have a very wide choice of concentrations......but that is getting near hair splitting again).
Generally, one gets a quantitative comparison of buffering capability by using the buffer index (capacity) equations, as mentioned above.
As to the situations where buffering is needed: In addition to the equilibria mentioned here, we are just having some cases with ions (not equilibrating with neutral species) which require pH control of the SiOH dissociation (discussed before?).
-------------------------------------------------------------------------------------------------------
By Anonymous on Wednesday, June 9, 2004 - 07:27 am:
Bill Tindall-many thanks for your reference to Kolthoff which points out that a strong acid is a good buffer at low pH.
A. Mouse- I wouldn't use 1mm TFA at pH 3 either. Most people use 0.1% TFA which has a pH about 2.3 and is a good buffer.
HWM- the buffer capacity of 0.1 % TFA is not far different from that of a mixture of phosphoric acid and KH2PO4 of similar molar concentration. I was being conservative by saying "not much" change. You could calculate this, or use one of the many programs available. Alternatively, you could try this with a burette and a pH meter.
-------------------------------------------------------------------------------------------------------
By Anonymous on Wednesday, June 9, 2004 - 12:40 pm:
Bill (or anyone else who feels qualified). I just want to followup on your comment that strong acids and bases are buffers. Please tell me where my logic is wrong (I wouldn't be at all surprised if it is).
I always thought that buffers were composed of weak acids or bases. And I always thought the reason was that with a strong acid or base, the conjugate is extremely weak. Let me illustrate with an example. If we think of the strong acid HCL. Its conjugate base is Cl-. But this is a *very* weak base. In fact most texts would call it a spectator ion cause it doesn't really do anything. So the question becomes: does a solution of HCl really have any ability to consume additional acid (hence act as a buffer). Or won't the Cl- ion just float around and do nothing.
Thanks. Once I get this straight I am set on this topic. Thanks to everyone for their valuable input.
-------------------------------------------------------------------------------------------------------
By bill tindall on Wednesday, June 9, 2004 - 02:43 pm:
Addressing the last Anon. question.........I wish you people had names. Cripes, call your self Joe and at least I could keep straight who was asking what. Pleased to discuss the topic...
The problem is that most people are only exposed to a limited definition of buffers, that which you stated. You need an accurate definition of a buffer and then all would be clear. "A buffer is something that resists changes in pH as a small increment of acid or base is added". It has a mathematical definition. It is the slope of the titration curve or d(Cb)/d(pH) where d(Cb) is a differential increment of base added. In any part of a titration curve that is flat the solution is highly buffered. This discussion can be found in any solution thermodynamics book or most advanced analytical chem books.
Let us take an extreme case to illustrate the point. Suppose we have a liter of water and a liter of 1 M HCl. Now we add 0.01 moles of sodium hydroxide to each. The water pH would zoom from 6 (CO2 present)to 12 or something like that. Clearly not buffered.
It is doubtful that one could even detect the pH change of the 1 M HCl solution after neutralizing 1 %. Hence, clearly this solution is highly buffered according to the general definition of buffers.
If one does the calculations it will be found that if buffers are needed at either the acid or base end of the pH range, a strong acid or a strong base is a better buffer than an acid/ salt pair of equivalent concentration. As a general guideline for pH 2 and less use a strong acid. Above 3 use an acid and its conjugate base. We can argue about what to use between 2 and 3 as it will depend on the objective of adding the buffer.
-------------------------------------------------------------------------------------------------------
By Anonymous on Wednesday, June 9, 2004 - 04:52 pm:
At the risk of being a pain, let me just "pick it apart" one last time.
(note: this question has to do with semantics of terminology and pH theory and has absolutely no practical significance. Anyone who's not interested in that - don't bother reading further)
Taking your example above: what if we add a small amount of acid to the 1M HCl solution. There's nothing to consume the additional acid. Having said that, I think it's also true that the additional acid will not cause much of a change in pH *but* it's because the H+ concentration is already high and a small amount of additional H+ doesn't consitute a significant change. It is not technically due to a buffering effect.
So - for someone like me who likes to be a huge nuisance about semantics - it would seem that a strong acid is only a unidirectional buffer. It resists change in pH when a base is added but not when more acid is added. But it experiences little change either way (and vica versa for a strong base).
My head hurts. I should have been a priest.
-------------------------------------------------------------------------------------------------------
By HW Mueller on Friday, June 11, 2004 - 12:06 am:
Last Anon:
Get a book on chemical equilibrium and look at pH index curves! A 1M HCl solution has a pH near 0. The buffering at such a low pH is enormous, it is real buffering and it´s done by the water equilibrium. If the water didn´t equilibrate you would have an increase in H+ CONCENTRATION corresponding directly to the amt of H+ added, conversly, you would have a H+ conc decrease corresponding directly to the amt of base added (by definition of strong acid or base).
Now what I said before about adding base refers to a pH of 2 and especially above. If you have only a completely dissociated acid present you will quickly raise the pH to regions where there is no buffering at all (adding base, NOT ACID, is dangerous in respect to buffering in such cases). Again, with "no buffering" I mean that you change the H+ Concentration by x moles if x moles of strong acid or strong base (completely dissociating) are added. Of course, if you add infinitally small amt of base you still have fairly good buffering even a bit above pH 2, but robustness is not about infinatly small amts.
-------------------------------------------------------------------------------------------------------
By bill on Friday, June 11, 2004 - 02:50 pm:
Last Anon: You want to think of buffering as a process- consuming the increments of acid or base added. This concept is useful in the mid ranges of pH but it does not explain buffering at the extremes of pH. Hence, it is not a useful concept to describe the general phenomena of buffering. And, buffers are not defined by a process, but by a result.
The general definition of buffering is simply resisting a change in pH and no mechanism of how that is accomplished is part of the definition.
Buffering = d(pH)/d(Cb).
It is a fact that when a strong acid is titrated the pH doesn't change much until near the end point. Ditto for a strong base. So, there are at least two processes to accomplish buffering. Consuming the increments of added acid or base (mid range buffering) and overwhelming them (pH extremes process)
-------------------------------------------------------------------------------------------------------
By Chris Pohl on Saturday, June 12, 2004 - 04:32 pm:
One more perspective on this topic:
Contrary to what was suggested in the first sentence at the beginning of this thread, buffers (in the more classical sense of the definition, anyway) are actually relatively rarely used in ion chromatography. And even if they are, this aspect is for the most part largely irrelevant. The fact is that when the pH of the mobile phase is sufficient to render the analyte completely ionized (as is most commonly the case in ion chromatography) or completely neutral (as is often the case for acids retained by ion suppression in reverse phase chromatography), it really doesn't matter much what the pH of the mobile phase is, over a fairly broad range. In the case of ion chromatography, what generally matters is the electrolyte concentration. The associated pH is generally a largely irrelevant characteristic of the mobile phase, merely a consequence of the specific ionic strength required for the separation.
So, from my point of view, one can divide the answer to the question into three parts:
1). If all analytes are neutral over a pH range encompassed by a neutral mobile phase (actually, a fairly wide range of mobile phase pH values is possible in this case), no buffer or electrolyte of any sort is necessary.
2). If all analytes are being separated under conditions where changes in pH have essentially no effect on the ionization of the analytes, then buffer control is not necessary although ionic strength control may be necessary if analytes are fully ionized under separation conditions. In this case, it is simply necessary to have the right pH and a buffer is not necessary. If, for example, a set of cationic analytes are fully ionized at pH 4, pH 5 or even pH 6, then a mobile phase consisting of 0.1 mM HCl would work just as well as a true buffer chosen to buffer the system at pH 4 (hopefully, we can all agree that 0.1 mM HCl is not acting as a buffer at pH 4).
3). If analytes are being separated under conditions where at least one of the analytes is partially ionized, some sort of buffer control is required. Here I defer to the discussion above as to what comprises a buffer but my point is that only under these circumstances is buffering of the mobile phase advisable in order to control separation selectivity.
-------------------------------------------------------------------------------------------------------
By Anonymous on Sunday, June 13, 2004 - 03:38 pm:
Chris
Thanks for the clarification about ion chromatography. I see your point there. Regarding your summary, I like part 1 and 2. Part 3 seems a little contradictory with #2. 3 says 'if any analyte is partially ionized you need buffer'. But I don't think that's exactly true. I think you state it well in #2 that if the pH does not need to be changed or stabilized a buffer is not necessary. This could be the case with a partially ionized analyte.
-------------------------------------------------------------------------------------------------------
By HW Mueller on Monday, June 14, 2004 - 03:04 am:
Oops, Bill, you confused me for a second, as you have the inverse formular of that given in Butler, Ionic Equilibrium, Addison-Wesley, 1964:
Buffer index (buffer capacity) = dC/dpH
This doesn´t matter except if one gives some values like these for this equation:
0.01M HOAC// 0.01M TFA
pH1 0.230// 0.233
pH2 0.0231// 0.0234
pH3 0.00269// 0.00234
pH4 0.00315// 0.0023
pH5 0.00535// 0.000023
pH6 0.00118// 0.0000024
pH7 0.000131// 4.6x10E-7
a buffer index of ~0.0023 corresponds about to non-buffering here (in my definition above: One adds X moles of acid or base and gets X moles of H+ conc. change)
ARRGHH, tried to write a table, didn´t work, so I put // between value columns for the two acids,
-------------------------------------------------------------------------------------------------------
By Chris Pohl on Monday, June 14, 2004 - 10:50 am:
Anon (6-13-04),
Regarding your comments surrounding my point #3:
I guess, in retrospect, one could divide this case into two sub-cases:
3a. The analyte is partially ionized and the pKa or pKb is less than 4. In this case, buffering per se should be unnecessary. pH control using a defined concentration of hydroxide or hydronium ion would be sufficient.
3b. The analyte is partially ionized and the pKa or pKb is greater than 4. In this case, I really think you need to utilize a buffer since one cannot accurately produce a given pH between pH4 and 10 without the use of buffers. Assuming that analyte ionization affects retention (which is always true in my experience for reversed phase and ion exchange), good separation reproducibility necessitates reproducible pH control.
-------------------------------------------------------------------------------------------------------
By Anonymous on Monday, June 14, 2004 - 04:26 pm:
check
-------------------------------------------------------------------------------------------------------
By HW Mueller on Tuesday, June 15, 2004 - 12:01 am:
Darn, this doesn´t end, but Chris I have trouple understanding your detailing of 3.
An example of why: If you get a robust method of analyzing an acid in its undissociated form at lets say pH=3 (without buffering) then it should be more robust if the acid´s pKa is greater than 4.
-------------------------------------------------------------------------------------------------------
By Fred on Saturday, August 7, 2004 - 03:48 pm:
Anyone know of some good references discussing buffers. I have seen references that talk about various buffers and their pKa's but I've never seen one that also talks about preparation of buffers, and which buffers a good for use with mass spec or evaporative light scattering (i.e. volatile buffers). Basically "everything the liquid chromatography ever need to know about buffers"
Thanks
Fred
-------------------------------------------------------------------------------------------------------
By Kostas Petritis on Saturday, August 14, 2004 - 02:42 pm:
Fred,
There are several relevant references out there.
1) Wilson N.S. et al. "Buffers and Baselines" LC-GC June 2001.
2) Tindal G.W. "Mobile phase buffers. Part I-The interpretation of pH in partially aqueous mobile phases. LC GC North America, Novemember 2002
3)Tindal G.W. " Mobile-phase buffers, Part II- Buffer selection and capacity, LC GC North America, December 2002.
4) Petritis, K. et. al. Volatility evaluation of mobile phase-electrolyte additivites for mass spectrometry. LC GC Europe, February 2002.
You can find some more relevant articles in the above references. There are also some relevant books (general about buffers and pH), on-line literature (mostly commercial i.e. Radiometer Analytica) and some commercial software that can calculate buffer relative values (i.e. buffer capacity, ionic force etc.) for HPLC and CE -personally I like the software "pHoebus".
Hope the above helps,
Kostas
