The Importance of Solvent System Selection
The Liquid-Liquid Nature of CCC
For “traditional” solid-liquid chromatography much though goes into choosing the correct column packing (silica gel, reverse phase, ion exchange, etc…), solvent composition, and solvent gradient. On the other hand, the selection of CCC solvent systems is equivalent to simultaneously choosing both the column, solvent, and gradient.
The basic requirement for a CCC solvent system is that it consists of two immiscible phases. Many functional solvent systems have been proposed, studied and successfully employed over the years. One popular method of concocting a solvent system involves the mixing of a hydrocarbon solvent such as hexane with ethyl acetate, methanol and water (HEMWat). Another very familiar method of arriving at reasonable solvent system is mixing chloroform, methanol and water (ChMWat).
At the simplest level, only two different solvents are needed as long as they are mutually immiscible such as a hydrocarbon solvent (lipophilic) and water (hydrophilic). Chlorinated hydrocarbons are also usually immiscible with water. A solvent chart will sometimes indicate which solvent combinations are mutually immiscible. However, most functional CCC solvent systems contain three or more solvents. The “modifiers” will typically be soluble in one the phases more so than the other. Hence, you can think of them as organic modifiers or aqueous modifiers depending on if they favor the lipophilic or hydrophilic phase.
Solvent systems are usually identified by their sequence of solvents arranged from lipophlic to hydrophilic separated by a slash. Their relative proportions are described separately in the same order as the solvents separated by a colon or slash. Thus, hexane/ethyl acetate/methanol/water 6:4:5:5. Relative proportions represent volumes of pure solvents before mixing. The total volume of the mixture is always less than the combined solvent volumes due to solvent mixing effects.
Abbreviated names for solvent system families are used in our lab: Ch = chloroform, E = ethyl acetate, H = hexane, M = methanol, and Wat = water. As a result, cumbersome solvent system combinations can be written and even pronounced in a manageable fashion, such as HEMWat (pronounced “hemwat”) and ChMWat (pronounced “kemwat”). Three other solvents that have been used are: Ac = Acetonitrile, Bu = n-Butanol, and ter = t Butylmethylether.
The ratio of final volumes of upper and lower phases varied by solvent system formulation. Typically, it is desirable to have a formulation that gives you approximately equal volumes of upper and lower phases, but this is not absolutely necessary.
Once you have identified a 3 or more solvents that form a biphasic mixture, combinations that you can mix are practically limitless. CCC researchers have spent a lot of energy dreaming up solvent combinations with interesting performance characteristics. It is an advantage to have so many “columns” from which to chose. However, it can also cause a lot confusion.
Ternary diagrams have been used extensively to map every possible combination of 3 and 4 member solvent system families. They are very instructive, but organizing solvent combinations into solvent system families seems to be a more accessible approach to making CCC practical.
Most of the time, combinations solvents are mixed together in rationally varying proportions and called a solvent system family. This way of formulating solvent system families adds a measure of organization to the solvent systems selection process.
ChallengesAnother aspect that is illustrated by the ChMWat table
Another challenge to creating solvent system families is that certain combinations “misbehave” due to similar densities. Usually, biphasic solvent systems separate easily into an upper and lower phase with a clean interface between the two layers. However, if the density of the two layers is close, gravity may not be able to get them to separate cleanly. This has happened to me when I have mixed hexane/dichloromethane/methanol/water. Hexane and dichloromethane are both immiscible with water but hexane is less dense than water and dichloromethane is more dense.
One may also encounter problems with emulsions when mixing solvents. Basically, an emulsion is formed when two phases don’t separate completely. A layer of tiny “bubbles,” made up of both phases, is formed between the upper and lower phases. In my experience, this is usually not a problem when mixing pure solvents. Emulsions are a big problem, however, when introducing plant extracts into biphasic solvent systems.
The Sweet Spot
While CCC does not retain any compounds on the “column,” it may not separate many of them in any appreciable way unless the solvent system has been chosen very carefully. There is a “window of opportunity” present in CCC separations that is related to the KD value of a given compound in a particular solvent system. The distribution constant, KD, can be expressed as the concentration of the compound in the stationary phase divided by the concentration of the compound in the mobile phase.
A solvent system, where the KD value of a particular compound is close to one, is considered to be the ideal system for separating the compound. According to the figure, small KD values result in a loss of peak resolution, while large KD values tend to produce excessive sample band broadening and long run times. In addition, the decision of which phase (upper or lower) will be the mobile phase is less important if KD = 1, since the retention volume of the target compound will be very similar in either mode.
Goals of CCC separations
Chromatographic separations usually address a particular goal in the process of starting from crude mixture of compounds to identification/characterization of pure compounds. The goal the researcher is trying to achieve will determine the methodology he or she adopts. Let us examine some possible goals.
Purification of a Single Compound
In some cases the purpose is to isolate a single compound. Chromatography is done to get rid of unwanted impurities and purify a single compound. This would be the case in purifying a synthetic product or isolating a single bioactive compound from an extract. Typically, more than one chromatographic step is needed, particularly in the case of isolating natural products.
Single compound purification can be accomplished in CCC by putting the target compound in the sweet spot. By definition, the sweet spot gives you the best purification zone for your target compound.
The distinct advantage of CCC in purifying single compounds is that separations can be routinely done in preparative scale. One other advantage worth mentioning is that CCC gives “clean cuts” between compounds. If anyone have ever done a preparative silica gel column, they know that compounds tend to streak, tail, and bleed into one another on even the best-packed column. If CCC is able to separate a compound, it will do so cleanly with no tailing.
Separation of Several Compounds
In other cases, the research scientist would like to purify more than one compound from mixture. This is often the case when the researcher is doing original research. In this case, the researcher adopts a strategy that gives her/him the best separation with a minimum of effort. The chromatography method selection is more complex in this case, and involves a certain amount of artistic flair by the research scientist.
Usually, the process begins with one or more preparative methods that can handle large amounts of crude mixture but don’t give very good resolution (separation between compounds). The conventional separatory funnel liquid-liquid extraction is an excellent example of a crude preparative method.
From preparatory methods, the process moves to more high-resolution methods. The trade-off is that high-resolution methods tend to be selective for certain types of compounds so you need to do more of them. At this point, the research scientist is working with much less material than they started out with so the high-resolution methods don’t have to be as high-capacity as the preparative methods.
CCC has the advantage of being both a high-capacity and a high-resolution technique. One can introduce any crude mixture on a CCC column as long as it is soluble in the solvent system being used. The fractions collected from the preparative CCC column can be re-chromatographed on a smaller capacity CCC column with a solvent system that compliments the selectively and resolution of the first one. The ability of CCC to make “clean cuts” mentioned before helps the research scientist who is going after minor compounds. In column chromatography minor compound tend to get overwhelmed by major compounds in the mixture and it is difficult to “draw them out” without repeated chromatographic steps that continually deplete the mass of minor compound in the test tubes at the end of the run. CCC methodology gives consistently better separation, with higher recovery.
Bioassay Guided Fractionation
Bioassay guided fractionation (BGF) represents a special case of chromatographic goals. In BGF a crude mixture is examined for a particular biological activity with the goal of isolating and identifying all the compounds in the mixture that have a desired biological activity. In beginning the research scientist does not know how many compounds, if any, exist in the mixture. In addition, the scientist does not know if these compounds are major, minor, or micro constituents of the original mixture. In theory, the biological activity of the active fractions will increase as the biologically active component becomes a larger and larger part of the fraction being tested. In this way, the researcher “follows the activity” by chromatographically separating a mixture into fractions and testing the activity of the fractions. Those fractions (each containing a number of different compounds) that exhibit the desired bioactivity are re-chromatographed and the process is repeated. In the end, the researcher will have fractions containing pure compounds that show potent biological activity.
In reality, the process does not always work that way. Sometimes the (initially promising) activity of the crude mixture diminishes as the mixture is fractionated. This may be due to loss of the active compound during some chromatographic step. It may also be due to synergistic effects of two or more compounds that have to be together in order for bioactivity to be observed.
Another possibility is that the active compound is present in so small an amount in the original mixture that it is only present in vanishingly small amounts as the fractionation process proceeds.
The great advantage of CCC in bioassay guided fractionation is its characteristic of complete recovery. If activity is lost during CCC fractionation the best explanation is synergy. It may even be possible to go back and recombine the compounds responsible synergistic interactions. In addition, CCC is easily scaled up in order to search for micro-constituents.
How to Select an Appropriate Solvent System for a CCC separation
Selecting an appropriate solvent system is one of the major undertakings in preparing a CCC separation. Unfortunately, this process is highly empirical and fraught with pitfalls. Its kind of like digging through a box of random nuts to find the appropriate nut for a bolt that you want to use.
Assuming that you have an idea of what compound (or at least what class of compound) that you are looking for the obvious approach is to look through the literature for someone who has already published a method that works with a compound or compounds similar to yours. This is commonly done for all sorts of chromatographic methods. You may not find any. On the other hand, you will most likely find many. The nagging question is: “Even if you find something that appears to work, how do you know it’s the best alternative for your particular need?”
Another strategy is to start with a simplistic approach and then fine-tune it. Every laboratory has a few “one-size-fits-all” chromatography methods. The problem with CCC is that it is so selective that this method may involve a lot of trial and error even if you have a reasonable strategy. In general, has already done a TLC of the mixture and therefore, has a good idea of the relative polarity of the compound of interest. The TLC information can be used to predict the best plan of action.
The ultimate strategy is good old “trial and error.” What are graduate students for anyway? Usually trial-and-error is combined with the two above approaches. The most popular trial-and-error process is called “partition studies.”
Over the years, several other methods of solvent system selection for CCC have been proposed, studied and utilized. An accepted method of predicting CCC behavior is to perform a partitioning study of a compound by measuring the relative concentrations of the compound in the upper and lower layers of a biphasic solvent system. The P value can be expressed as the concentration of the compound in the upper phase divided by the concentration of the compound in the lower phase. P values obtained by partitioning studies predict the retention time of a particular compound, e.g., in an HSCCC instrument, when the proper consideration is made for the mobile and stationary phase of the HSCCC run.
The most common form of partition study is descriptively called the “shake-flask” method. This method involves dissolving a small amount of a compound or mixture in a biphasic system, shaking them together, and allowing the system to equilibrate before measuring the concentration of the target compound(s) in each layer. The concentration in each layer can be measured by three principle methods:
'(i)' The two phases may be separated and the solvents evaporated in order to obtain the mass of the residues. This gravimetric method requires relatively large amounts of compound to get a reliable result. It is also not very useful for mixtures, which may contain large amounts of extraneous compounds.
'(ii)' The two phases may be separated and the solvents reduced in volume in order to spot the compounds on a TLC plate. The shake-flask/TLC method does not give a very accurate determination of the relative amounts of compounds because the researcher is only going by the density of the spots to make a judgment. I
'(iii)' The relative concentrations can be measured by measuring the UV-vis absorption of each layer. This spectroscopic method works well for targeting a particular chromophore by itself, or in a mixture of non-absorbing compounds. It can be done with small amounts of compounds. However, the spectroscopic method does not work for compounds that do not absorb in UV-vis and for mixtures where compounds’ absorptions interfere with each other. Also, since the compound is being measured in two different solvents, steps must be taken to minimize solvent interference with spectroscopic measurements.
'(iv)' In the case of mixtures, each phase can be analyzed by high pressure liquid chromatography (HPLC) or gas chromatography (GC) and the relative amounts compounds present in each layer can be determined. This chromatographic method requires the development of a reliable HPLC or GC protocol that gives a reasonable separation of the compounds of interest. The chromatographic method is relatively time consuming when several solvent systems must be tried. In addition, for many natural product samples the target analyte may not even be known, such as is always the case in bioassay-guided fractionation.
A TLC Alternative to Shake-Flask Methods
A reliable method of solvent selection should be available that is accessible by both experienced and novice CCC users. An exemplary method of selecting an appropriate solvent system for a CCC separation would satisfactorily address the following criteria: *systematic in its approach,
- versatile for a wide range of natural products,
- supple enough to allow some margin of error in making a judgment,
- time efficient,
- adaptable to rational fine-tuning,
- and applicable to mixtures of unknown composition as well samples of known composition.
Since thin layer chromatography (TLC) has traditionally played the role as solvent system selection method in solid-support chromatography, a method that involves the estimation of CCC solvent system choice based on TLC behavior may meet the above criteria with some degree of satisfaction. A TLC-based method was developed for the Generally Useful Estimation of Solvent Systems in CCC, allowing a good first "G.U.E.S.S.", and being able to replace conventional procedures.
Without a doubt, TLC is a common denominator of all natural products separations. Samples ranging from crude extracts to purified compounds are subjected to TLC as a quick and easy way to assess their composition, identity and purity. Many useful TLC solvent systems are known and routinely used in laboratories all over the world. In fact, the G.U.E.S.S. method has been done in reverse for decades. It is customary to separate an extract or column fraction by CCC, and then perform TLC on the collected CCC fractions in order to ascertain their composition and purity. If TLC can be routinely used to analyze CCC fractions, then it should be possible to use TLC to predict CCC elution performance.
However, relating TLC and CCC is fundamentally challenging since their respective physicochemical means of separating compounds is quite different. At least one method of predicting droplet countercurrent chromatography (DCCC) behavior based on TLC observations has been proposed. In this method, silica gel TLC was done with the organic layer of a chloroform/methanol/water biphasic solvent system in order to predict the best mobile phase for optimal DCCC performance in that solvent system.
No matter how efficient or reliable the shake-flask method may be, the problem of “where to start” still needs to be addressed. The same bewildering choice of solvent systems is present when choosing the solvent system for a shake-flask partition study as it is for a CCC separation. Therefore, the TLC based G.U.E.S.S. system is at least complementary to the shake-flask method, and at best can replace the shake-flask and similar methods.
For more detailed information about the G.U.E.S.S. method please see the reference section at the bottom of this page.
A TLC-based CCC Solvent Selection Guide
The purpose of this page is to show how we use TLC to predict the best solvent system for a CCC separation.
How We Developed This Approach
The major drawback in the employment of CCC separations by both experienced and inexperienced natural products chemists seems to be a lack of clear guidelines for solvent system selection. The choice of solvent system for CCC separations is absolutely crucial. Compared to the far more popular solid-support chromatography, the selection of CCC solvent systems is equivalent to choosing both the column and the eluant at once.
The G.U.E.S.S. Mix (GUESSmix)
The GUESSmix is a mixture of reference compounds that can be used to link the TLC behavior of compounds with their CCC performance. Twenty-two GUESSmix compounds were chosen based on the following criteria:
- Natural Products
- Varying polarity
- Varying functional groups
- Varying size
- Commercially available chemicals
- Absorb in UV
- Identifiable with standard TLC visualization techniques
- Low toxicity
- Reasonably inexpensive
- Available in high purity formulations
Basically, it’s a versatile test mixture that can be used in several capacities in CCC. It is somewhat like the famous GROB mixture in GC chromatography.
The HEMWat Solvent System Family
We used the HEMWat solvent system family for our studies. The advantages of using the HEMWat solvent system family are numerous:
- It has a long history of use in CCC separations.
- It is an extensive, but not overly cumbersome, solvent system family.
- It covers a wide range of polarities
- It tends to include compounds of interesting (drug-like) polarity.
- It employs economical solvents
- It employs solvents of relatively low toxicity.
- The solvents are relative easy to remove (evaporate) from fractions.
- Tends not to bleed during CCC runs.
- Exhibits high stationary phase retention ratio volumes with CCC.
The G.U.E.S.S. Method
The key to the G.U.E.S.S. method involved establishing a link between shake-flask P values and TLC Rf values. An obvious approach to establishing this link is to compare the shake-flask P value of a compound in a particular HEMWat solvent system with the TLC Rf value of that compound developed in the organic phase of the HEMWat solvent system. This is an effective method, but it still involves making a series of biphasic solvent systems.
In order to simplify the formulation of the TLC solvent system, the HEMWat organic phase was replaced with a solvent system made by simply mixing hexane and ethyl acetate in the same ratio as the HEMWat solvent system. The following table illustrates how this is done. The problem here is that there are 10 SSE (solvent systems based on ethyl acetate) for 16 HEMWat solvent systems. This can be solved by just considering the ‘whole-step” HEMWat solvent systems: (-8, -7, -6, -5, -3, 0, +3, +5, +6, +7, and +8). The remaining solvent systems can be considered to be “half-steps” that can be exploited in fine-tuning the optimal solvent system.
In order to consistently compare Rf and P values, the partition results were expressed in terms of Pf, the concentration in organic phase divided by the sum of the concentrations in both phases. Most of the compounds in the GUESSmix exhibited a rather close correlation between SSE TLC Rf, and HEMWat Pf values.
Therefore a compound that has an Rf value close to 0.5 in a SSE TLC solvent system will very likely elute in the sweet spot of the corresponding HEMWat solvent system.
Of course there is a catch to this. It involves understanding a few things about the CCC technique.
Normal Phase of Reverse Phase
For a particular HEMWat solvent system, the choice of mobile phase is an important consideration. By comparing normal phase and reverse phase it can be shown that, as expected, compounds with a shake-flask P value close to 1 will be present in the HSCCC sweet spot in both normal and reverse phase. Therefore, for those compounds it does not matter much which phase is chosen to be mobile.
However, there is a significant difference between normal phase (organic phase mobile) and reverse phase (aqueous phase mobile) as to which compounds in a mixture are actually separated. As shown in the normal-phase/reverse-phase figure, the normal phase gathers less polar compounds, while reverse phase gathers more polar compounds.
Therefore, the choice of mobile phase must be taken into consideration, when the HSCCC run for a particular target compound or cluster of compounds is being planned. First of all, the selection of a solvent system that hits close to P = 1 for the target compound(s) (the ideal HEMWat number) should be attempted. It can then be decided whether to gather compounds with less polar behavior by doing a normal phase run or to gather compounds with more polar behavior by doing a reverse phase run.
The decision to go normal phase or reverse phase may be a way of “hedging your bets” that the right solvent system was chosen or it may be based on whatever other compounds are known or believed to be present in the mixture. The choice of normal phase or reverse phase is illustrated in the normal-phase/reverse-phase figure by comparing runs of standard compounds under similar conditions in both normal and reverse phase. There are three compounds that hit the sweet spot in both normal and reverse phase. The normal phase run also separates three compounds of higher P values that elute in front of the sweet spot. The reverse phase runs separates ferulic acid, which eluted too slowly in the normal phase run, as well as more polar compounds.
The value-added aspect of the G.U.E.S.S. is that, not only do you have the P vs. Rf connection, you can also directly compare GUESSmix compounds of known HEMWat behavior to a compounds of uncertain HEMWat behavior with TLC. You simply run the GUESSmix compounds right beside your mixture on TLC. The following figure is a summary of GUESSmix compounds an their optimal HEMWat solvent system as predicted by shake-flask P values.
Caveats and Responses
1) This method assumes that one of the HEMWat solvent systems is the optimal solvent system for your compound(s) of interest. The HEMWat solvent system family has some great credentials.
2) This method assumes that your compound(s) of interest will be soluble enough in hexane/ethyl acetate to give you a decent TLC. You can still do the side-by-side comparison with another TLC solvent system.
3) You probably won’t hit an optimal CCC solvent system dead-on your first try. It may be good-enough or you will probably have a head-start in optimization.
4) Some compounds don’t work in this method. Even if you have to go back to a shake-flask search you know where not to look.
HEMWat: The Quintessential Solvent System Family
What is a HEMWat?
HEMWat is an acrynomn for a certain mixture of Hexane, Ethyl acetate, Methanol,and Water. The HEMWat solvent system has the general organization of: Organic/Organic Modifier/Aqueous Modifer/ Water. This is borne out by the observation that the upper phase is a mixture of hexane and ethyl acetate while the lower phase contains primarily methanol and water. However, the miscibility relationships are rather complex. Hexane is immiscible with water. Hexane and methanol form two phases when they are mixed together but they are still slightly miscible with each other. The same goes for ethyl acetate and water.
The following table describes some of the vital statistics of the four solvents. While their relative miscibility/immiscibility determines whether or not they will form 2 distinct phases when mixed together, several other parameters are also important. For example, relative density of the solvents determines which phase will be the lower phase and which phase will be the upper phase. It is interesting to note that in most categories Hexane and Water are at the opposite ends with ethyl acetate and methanol are somewhere in-between.
Solvent tables can be found at various addresses on the internet. Here is a small sample of what is available. Evidently, there are several "polarity indexes" that have been proposed. they all give slightly different values. http://organicdivision.org/organic_solvents.html http://macro.lsu.edu/HowTo/solvents.htm http://home.planet.nl/~skok/techniques/hplc/eluotropic_series_extended.html
Here is the definitive HEMWat table showing 17 formulations of the 4 solvents:
Why 17 formulations?
Solvent systems usually organized in families where the component solvents are combined in varying proportions. The HEMWat particular family will generally be useful for separating compounds of low to medium polarity. In general, CCC practitioners seek to find a solvent system where the distribution constant of their target compound is near 1. That means, at equilibrium, an equal concentration of the compound will be found in both upper and lower phases.
The 17 formulations are somewhat arbitrary in the sense that thousands of combinations are possible candidates for HEMWat. The HEMWat tabel shows a rational progression from hexane/methanol to ethyl acetate/water in meaningful steps. Too many steps adds unnecessary confusion and complexity to the solvent system selection process. Too few steps would leave gaps where useful solvent systems would not be indicated.
The amount of hexane + ethyl acetate is equal to the volumes of methanol + water in order for the resulting two phases of solvent system to be available in approximately equal amounts. As you can see from the table the lower phase tends to be have more volume than the upper. This is most likely due to the fact that ethyl acetate is tends to be found in both upper and lower phases. In the HEMWat solvent system family, the organic phase is mainly composed of hexane and ethyl acetate in the upper phase of biphasic mixture, while the aqueous phase is mainly composed of methanol and water in the lower phase of biphasic mixture. The solvent volumes found in both phases have been determined for five HEMWat solvent systems. [A. Berthod, M. Hassouon, M.J. Ruiz-Angel. "Alkane Effect in the Arizona Liquid Systems used in Countercurrent Chromatography, Anal. Bioanal. Chem. (2005) 383: 327-340][Li Z.C.; Zhou Y.J.; Chen F.M.; Zhang L.; Yang Y. "Property Calculation and Prediction for Selecting Solvent Systems in CCC" J. Liq. Chrom. & Rel. Technol. 2003, 26(9-10), 1397-1415.]
The HEMWat solvent system was inspired by two solvent system families that were published before the advent of HEMWat:
[A. Berthod, M. Hassouon, M.J. Ruiz-Angel. "Alkane Effect in the Arizona Liquid Systems used in Countercurrent Chromatography, Anal. Bioanal. Chem. (2005) 383: 327-340]
[Oka F.; Oka H.; Ito Y. Systematic Search for Suitable 2-Phase Solvent Systems for High-Speed Countercurrent Chromatography. J. Chromatogr. A 1991, 538(1), 99-108]
What makes it so good?
As part of a method development project in our lab we determined the partition coefficient of every GUESSmix compound in every HEMWat solvent system. The partition coefficient is obtained by dividing the concentration of a compound in the upper phase by its concentration in the lower phase to obtain a value between zero and infinity. This was a huge undertaking: 21 compounds x 17 solvent systems x 2 phases x 3 trials = 2,142 measurements! The simple “shake-flask” process was used to distribute the commercially available compounds between the two phases of a HEMWat solvent system. The relative concentrations of the compound in each phase was determined by UV absorption.
All of the compounds tested showed similar trends in HEMWat systems. Generally, as the lower phase becomes more aqueous (the HEMWat number becomes more positive) organic compounds tend to flee towards the upper phase and P increases. An exponential increase in P as the HEMWat system becomes more positive can be observed by graphing the log10 of P versus the HEMWat number.
This was a rather exciting and intriguing result. There is no apparent reason why the LogP vs. HEMWat number plot should be linear. As I said, the choice of HEMWat proportions are somewhat arbitrary so there is not real reason why they should represent equal polarity steps. Of course, the P value is not strictly a measure of relative polarity, but rather the result of a complex interaction of a particular compound’s relative solubilities in 4 different solvents. Interestingly, as can be seen in, most of the compounds tested showed nearly linear LogP behavior with similar slopes. All compounds do not behave as linearly as Umbelliferone, but overall the correlations were surprisingly consistent.
How might this be useful? First of all, once the P value is determined for a compound in one HEMWat solvent system, its P value in the other 16 can be extrapolated. Second of all, this shows that the steps of the HEMWat solvent system family tend to progressively target more and more polar compounds.
The proposed HEMWat series of CCC solvent systems demonstrated its versatility by showing that many compounds, such as those shown in, have a P value equal to 1 in the range of HEMWat solvent systems. This means that many natural products are likely to be satisfactorily separated in one HEMWat solvent system or another. In addition, some compounds have LogP values within the sweet spot (-0.4 < LogP < 0.4 which is the same as 0.4 < P < 2.5) in one or more HEMWat solvent systems, even though they do not have an ideal P=1 value in any HEMWat solvent system. The currently proposed HEMWat method is a versatile and useful method for the separation of a variety of natural products, with polarities ranging from medium lipophilic to slightly polar (non-glycosidic).
How does the compound decide where to go?
Every organic compound, except the most lipophilic hydrocarbon-dominated molecules tend prefer methanol to hexane. On the other hand, organic compounds, in general, are not freely soluble in water. Therefore, the hexane/water dichotomy constrains compounds to be uncertain where to go. In effect they try to chose "the lesser of two evils" and end up chosing some of each. As the amount of hexane decreases (that’s good) the amount of water increases (that’s bad) so most molecules tend to waiver in both phases. This situation is beneficial to countercurrent chromatography since we are trying to find a solvent system where the target molecule is in nearly equal concentrations in each phase.
Interestingly, pure ethyl acetate is not a particularly good solvent for organic compounds, yet, it really works well along with the other three solvents. Methanol, on the other hand, tends to be a very good organic solvent, it has hydrogen bonding (both H donating and H accepting) properties but it is not as polar (aqueous?) as water.
Superficially, we could say that we are separating compounds based on their relative polarity. In reality, its a lot more complex than that. Many natural products possess complex structures sporting a variety of functional groups. These functional groups often are attracted to different kinds of solvents so the stage is set for this fickle behavior when it comes to relative solubility. The beauty of this is that even if two different solvent systems target compounds of similar polarity they may separate them differently. For example, the order of elution of the GUESSmix compounds is about the same if we replace ethyl acetate with t-butylmethyl ether, but there are some important differences.
Using HEMWat as a base to build solvent system families:
The redeeming qualities of HEMWat lead us to believe that we can use the HEMWat template to build solvent system families that compliment HEMWat in polarity window and separation selectivity. In first generation HEMWat variants we replace one of the solvents with a solvent of similar polarity characteristics. If we assume that the hexane/water brackets are performing an essential service we can focus on a group of first generation interior switches.
Already, we observe that some switches work better than others. For example hexane/dichloromethane/methanol/water 5:5:5:5 forms multiple phases because the density difference between the phases in not big enough. What other problems could occur?
Well, you may get a single phase. Many solvent system families are limited because in certain combinations of solvents will give a single phase: t-butylmethyl ether /ethyl acetate/methanol/water 5:5:5:5 is a good exmple.
On the other hand you may get more than 2 phases. This is interesting but requires adopting a different strategy to deal with them: hexane/t-butylmethyl ether/acetonitrile/water 5:5:5:5 gives three phases.
You may get two phases but with a great inequality in phase volumes: hexane/dichloromethane/acetonitrile/water gives a 72/28 ratio. This is not an insurmountable problem but it does require some adjusting to the basic system.
Another problem that occurs is that the separation time may be slow and one phase remains cloudy for a period of time. Cloudy is not good. Cloudiness usually means the two phases are having difficulty separating cleanly.
Hexane/ethyl acetate/methanol/water solvent systems have widely been used to separate a variety natural products such as theaflavins, catechins, flavonoids, polyphenols, diterpenes, flavonoid glycosides, invermectins and macrolide antibiotics. The number of publications reporting on separations employing hexane/ethyl acetate/methanol/water solvent systems is around 50.
As far as I can tell, the first HEMWat separation was published in 1982 by Yoichiro Ito and Jesse Sandlin. They used a type J multi-layer coil planet centrifuge to separate a mixture of plant hormones with 3:7:5:5 solvent system. The aqueous phase was mobile. The flow rate was 4 mL/min. The speed of revoltion was 800 rpm. They observed 51% retention of stationary phase. The total coil capacity was 285 mL. They separated a mixture of plant hormones in order of reverse phase elution: indole-3-acetamide, indole-3-acetic acid, indole-3-butyric acid, and indole-3-acetonitrile. [Y. Ito & J. Sandlin. "High-Speed Preparative Counter-Current Chromatography with a Coil Planet Centrifuge" J. Chromaogr. 244 (1982)247-258.]
This experiment was reported again in 1988 with a different instrument. [y. Ito & F.E. Chou. "New High-Speed Counter-Current Chromatograph Equipped with a Pair of Separation Columns Connected in Series" J. Chromatogr. 454 (1988) 382-386] The same compounds were separated on HEMWat 5:5:5:5 in 1990.[Y. Ito, H. Oka, Y.W. Lee. "Improved High-Speed Counter-Current Chromatograph with Three Multilayer Coils Connected in Series. II Separation of Various Biological Samples with a Semi-Preparative Column" J. Chromatogr. 498 (1990) 169-178][H. Oka, Y. Ikai, N. Kawamaura, M. Yamada, K-I Harada, M. Suzuki, F.E. Chou, Y-W Lee, Y. Ito. "Evaluation of Analytical Countercurrent Chromatographs: High-Speed Countercurrent Chromatograph-4000 VS. Analytical Toroidal Coil Centrifuge" J. Liq. Chromatogr. 13 (1990) 2309-2328]
The heptane/ethyl acetate/methanol/water combination made it’s debut in 2000. The heptane/ethyl acetate/methanol/water system was reported to have a more-or-less constant relative viscosity (ηupper/ηlower = 3.21-3.86) as the formulations changed. [I.A. Sutherland, J. Muytjens,M. Prins, P. Wood "A New Hypothesis on Phase Distribution in Countercurrent Chromatography" J. Liq. Chrom. & Rel. Technol. 23 (2000) 2259-2276]
HSCCC stationary phase retention data, for HEMWat –6, –3, 0, +3, +6 and +7 at different flow rates, has been previously published.
[Du Q.Z.; Wu C.J.; Qian G.J.; Wu P.D.; Ito Y. Relationship between the flow-rate of the mobile phase and retention of the stationary phase in counter-current chromatography. J. Chromatogr. A 1999, 835(1-2), 231-235.]
Other Solvent Systems
Similar to HEMWat above the HEMBWat system incorporates Heptane instead of Hexane, and provides more polar options with the addition of Butanol.
Error creating thumbnail: /home/content/w/r/i/wrightwoodtech/html/theliquidphase/bin/ulimit.sh: line 3: ulimit: cpu time: cannot modify limit: Operation not permitted /usr/bin/convert: Unable to open file (/home/content/w/r/i/wrightwoodtech/html/theliquidphase/images/7/7e/HEMBWat.jpg) [No such file or directory].
- J.B. Friesen and G.F. Pauli “G.U.E.S.S. to make Generally Useful Estimations of Solvent Systems in CCC” Journal of Liquid Chromatography and Related Techniques 28, 2777-2806 (2005).