A disproportionately large part of the chemical literature concerns itself with theory or the interpretation of experimental data- that is talking about what was done. By comparison there is relatively little written concentrating detailing methods and techniques- that is how it was done. We have a misplaced confidence in the pedagogical talents of chemists in general. As I look back on my own career development I see that it can be divided into three phases:
- Trying to do experiments with outdated, unsophisticated methods, which I had learned in introductory laboratory courses.
- Using a smattering of more refined methods which picked up from whatever more experienced co-workers I happened to come into contact with.
- Gathering together the advanced laboratory literature and text-book descriptions of the most up-to-date and refined methodologies for achieving success in common laboratory situations and selecting from these the most appropriate and powerful for the particular situation.
It is my experience that to pass through these stages over many years is an exceedingly wasteful process of professional development. I believe that there is a need for up-to-date reviews comparing all the methods available for attaining general laboratory goals- such as high synthetic yields, simple separations and pure products. Making this review available through KiloMentor addresses this problem.
In organic chemistry, product isolation and purification is the most time consuming aspect of the art. It is the least intensively dealt with in the literature. Because preparative column chromatography is by far the most general and powerful of these methods, a thorough understanding of the available options is important. Over the years some highly refined and powerful methods have been designed for carrying out preparative scale separations; however, there has been very few comprehensive description, tabulation or critical comparison of these second and third generation forms of preparative chromatography. This is the present goal.
This article, being placed on-line now in 2020, is at least 10 years out of date since I am now retired. Bear this in mind but it is a start- my belated contribution. Separation techniques age better than the best reaction methods!
TRADITIONAL COLUMN CHROMATOGRAPHY
There is really no definition of traditional column chromatography but it is desirable to start with the description of some basic procedure and then describe the recent innovations and variations on this general practice. For lack of something better, the author will present as a composite norm the best advice he received as a graduate student in Professor R.B. Woodward’s Harvard laboratories about 1972.
The ideal length for a column is twelve times the diameter and the ratio of adsorbent weight to the weight of the mixture should be 100:1 until experience proves differently in the particular application. To choose the correct diameter for your column, the density of moist adsorbent must be known. An easy way to do this is to add the adsorbent (100 times the weight of the mixture to be separated) to the starting solvent mixture in a large graduated cylinder and swirl to remove the air bubbles and dispel the heat. Let the solids settle and note the volume, which the wet solid occupies. The correct column diameter will be
d= (V/3 * 3.14159 )1/3 V= volume in milliliters d= diameter in centimeters
If you already know the density of moist adsorbent you can immediately calculate the required diameter from the weight of mixture to be chromatographed.
d = (100 * g/3 * 3.14159 *D)1/3 d= diameter in centimeters
g = weight in grams
D = density in grams/ cubic centimeter
These are to be interpreted, as approximate but should allow the selection of appropriate column geometry. The constriction just above the stopcock should be filled with a small tightly wound glass wool plug. Using longer columns this is not easy to do. Frustration can be avoided by using a long glass tube in combination with a still longer solid glass rod, which can slide easily but without much wobble, inside the tube. Place the glass wool plug in one end of the tube with the rod positioned just behind it like the plunger in a syringe. Position the assembly inside the column and ‘inject’ the glass plug at the correct position at the outlet of the column. The glass wool plug should then be covered with just enough coarse sand to hide all the glass wool threads. This space, between the bottom of the chromatographic layer and the stopcock exit port, is the dead volume in your column. Keeping it small is critical to achieving sharp bands and maximizing resolution. The column should be clamped securely with special care that it is exactly vertical as viewed both from the front and the side.
If the column is to be eluted with a series of solvents, i.e. hexane and then ethyl acetate, the column should be packed in 1:1 hexane/ethyl acetate. This deactivates the most active sites on the adsorbent so that heating and channeling will not occur when the solvent composition changes rapidly. The column should be slurry packed with the tap slightly open to allow a gentle solvent flow. The column should be periodically tapped with a wooden dowel or agitated with a vibrator to settle the adsorbent. For best results, the column should be covered with a good head of solvent, stoppered at the top to avoid evaporation and left to stand overnight before use. This part of the advice is very easy to ignore or discount because without preplanning it can be inconvenient but experience demonstrates that such columns run faster and better. This is an empirical observation; perhaps the overnight pause allows any remaining bubbles to work their way to the top of the column; perhaps the solvent and adsorbent become more thoroughly equilibrated or perhaps it is just that the operator begins the operation fresh, with the entire day available, if necessary, to complete the job right without short-cuts.
Before starting to apply the sample the column should be flushed with five volumes of the less polar solvent of the solvent system mixture (hexanes in the example using hexanes-ethyl acetate). The top of the column should be protected with a layer of coarse washed sand. This protective layer, which prevents the column body from being disturbed by turbulence during the addition of the elution solvents, should not be put in place until just before the column is to be loaded.
The mixture to be separated should be dissolved in no more than
{column diameter * 3.14159/8 cc} of the starting solvent (this is a column of liquid no more than one-half column diameter high) The solution should be applied evenly to the sand and run onto the top of the column as quickly as possible. At no time should the top of the column below the sand become dry.
If the mixture to be chromatographed will not completely dissolve in the solvent volume specified it should be preadsorbed onto a small portion ( no more than d * 3.141549/8 in volume) of the adsorbent by low-temperature evaporation of a solution in the presence of adsorbent. This can most easily be done by placing the solution and adsorbent together in a round-bottomed flask and ‘rotovapping’ to dryness with minimal heat. Care must be taken here to avoid further activation of the adsorbent by removing residual water. The solid obtained should be spread evenly onto the top of the column. If this option is used this layer needs to be placed on the top of the column before the protective layer of sand!
During the elution of a well-packed column, the solvent flow should be kept less than the maximum possible. The bottom stopcock should be at least partly closed. Otherwise, the bands become noticeably distorted because the flow rate in the central core of the column differs too greatly from that at the walls.
Research Techniques in Organic Chemistry by Robert B. Bates and John P. Schaefer 1 and Guide for the Perplexed Organic Experimentalist by H.J.E. Loewenthal also describe how to run a preparative column chromatography.
Bates and Schaefer recommend a height to diameter ratio of 7.5 to 1 for alumina chromatography. This is not a large divergence from the average of 12 to 1 because alumina is denser than silica gel, the more commonly employed adsorbent. These authors further advise an adsorbent-substrate ratio of 30:1. They too recommend a glass wool plug at the base of the column. It is interesting that no comment is made for or against the very popular commercial columns fitted with sintered glass discs. Bates and Schaefer recommend slurry packing with the starting least polar solvent even if a solvent change is planned during the development. For sample application, they recommend the least polar solvent in which the entire sample will dissolve. No quantitative guideline for the minimum volume of solvent, which may be used, is set down. This is a serious omission because it is a very common error to use too much. Grade IIII alumina is suggested as a good trial adsorbent. Such an empirical choice of adsorbent material is I think unnecessarily trying on the patience of the researcher. TLC examination on activated and deactivated plates of both silica and alumina is probably a preferred course.
In a more sophisticated laboratory text, Guide to the Perplexed Experimentalist; Loewenthal gives no advice about sample loading and column size. He recommends wet packing in all cases but distinguishes between the preferred techniques for silica gel and Florisil and that for alumina. For the former pair, he recommends that the practitioner slurry the solid in a beaker to dissipate the heat and expel the air bubbles and then pour the homogenate into the column, augmented with whatever additional solvent is needed to transfer almost everything at once. For alumina, his recommendation is to add dry solid in a fine stream to the column with a solvent head of three milliliters per gram. It is not stated whether during the operation the stopcock is open or closed. Loewenthal does emphasize that the column should be perfectly vertical and must be vibrated to settle the solid. About sand, glass wool or sintered glass disks in the base of the column nothing is said although he does consider Teflon stopcock indispensable to eliminate both grease and leakage. He urges special attention for the top of the column. He believes its levelness and homogeneity in large part predetermine whether good separation will occur. In contradiction with many workers, he specifically advises against disturbing the top layer in any way either to apply a covering of sand or even to spread a mixture preadsorbed on a portion of the adsorbent. Loewenthal’s instruction is to add the mixture in solution carefully to a sufficient solvent head above the adsorbent. What this means in practice is not clear. He quite properly emphasizes the importance of preparing adsorbents of the correct known activity and he makes the practical observation that although many authors advise selecting elution solvents by TLC study in practice and most routine cases economy and convenience factors may necessarily dominate. By and large, he prefers solvent mixtures differing widely in the boiling points of constituents with the less polar as the higher boiling. This makes recovery and recycling possible. Favorite mixtures are dichloromethane-hexanes, ether-hexanes, and ether-benzene. Gradient elution is recommended in all cases.
Loewenthal’s book is one of the few I know, which recommends anything, which I would describe as advanced laboratory practice. He specifically recommends multi-bore columns and the use of water-jacketed columns to eliminate overheating and column cracking. It should be noted that his advice increases the likelihood of these effects because he does not equilibrate his adsorbent with all the solvents before beginning the chromatography.
Neither these sources or others which I have examined have any suggestion for overcoming the most widespread error in column chromatography- that is the accidental running dry of the column, The best idea that I have seen for dealing with this mini-disaster is to cover the top of the column not with a layer of sand but with a layer of anhydrous magnesium sulfate. It is reported that for some reason the meniscus of the solvent will not run below this layer but instead the column will stop running. I do not have personal experience to confirm this claim! Of course, the composition of any eluent, which contains molecular species complexed by magnesium sulfate, would be altered by this layer but usually, this simply provides improved in-column solvent drying!
ADVANCED LABORATORY PRACTICE
Advanced laboratory practice is described in the primary literature, in reviews and also in commercial brochures in cases where a company supplies special apparatus for the method. In general, the techniques have not filtered down into textbooks and laboratory descriptions.
Because there are no data testing these techniques on a standard separation mixture problem, comparisons of relative resolving power are anecdotal and might even reflect mental biases or differences in technical expertise in practicing different methods. Nevertheless, even in the absence of comparative data, there are operational differences that should direct a choice to one rather than another method according to specific circumstances. All the advanced methods described here have some clear advantages over the standard column chromatography. Moreover, the descriptions and comparisons point up combinations of advanced chromatographic techniques that may not have been tried but which might be capable of even better results. Even if the reader does not adopt any procedure in its entirety knowledge of them all will improve his/her own intuitive feel for what will and will not work.
MULTIBORE COLUMNS
In a 1964 article, George A. Fischer and J.J. Kabara investigated the effect of column shape on the resolution of lipids on Florisil. They concluded that using a multibore column has significant advantages over the conventional single diameter column. These are:
Almost twice as fast to run
Uses only one-half the solvent
Quantitative separation for multi-bore column versus a crude enrichment of the components in the fractions from an equivalent monobore column
Higher substrate capacity
It was found empirically that the best results were obtained when the radius ratio between adjoining sections was 1.4:1 and the sections were of equal length. This corresponds to a volume ratio between the connected sections of 2:1. The large cross-sectional area of the head of the column allows initial adsorption of the mixture on a very thin concentrated band and rapid initial ordering of the solutes.
In a regular column, the solute moves fastest at the walls. The effect of column narrowing is to retard the solvent at the walls and concentrate the solute band.
The consecutively smaller lower sections of the column have the advantage of being less sensitive both to the influence of channeling and to the degree of tilt of the column; furthermore, since these narrower column sections allow for a more efficient contact of the solute with adsorbent particles, the lower part of the column produces increasingly refined resolution.
Overall the decreased total volume of the column, compared with a conventional design of the same length and capacity, gives increased speed and decreased solvent usage.
Multibore columns are, most, often custom-made; however, the graduated column diameter has the result that in practice the same column can be used for the separation of a wide range of sample sizes simply by adjusting the amount of adsorbent.
The combination of multibore columns and medium pressure or vacuum preparative chromatography does not seem to have been investigated. Preparing a multicore column from a thin separatory funnel is another idea that may or may not have been tried.
CONTINUOUS COLUMN CHROMATOGRAPHY
Continuous chromatography is a specialized technique that has been used predominantly for the purification of very poorly soluble materials which have impurities that make them intractable to recrystallization. The method is in principle more general and should be applicable to the chromatography, using hydrocarbon and other non-polar solvents, of many difficultly separable mixtures wherein substances have very small but different Rfs .
Another use might be for reverse phase chromatography of entire reaction mixtures by refluxing a polar elution solvent either at atmospheric pressure or reduced pressure to elute first the inorganic reagents and byproducts and then the purified organic products. Higher than ambient temperature elution would provide higher Rfs with less polar solvents and the continual recycling system would dramatically lower the solvent requirement. The chromatography becomes a continuous extraction off of the solid adsorbent rather than a traditional chromatography.
Another potential use for the system would be the decolorization of material by chromatography through a column of charcoal or mixed Celite and charcoal.
All glass commercial apparatus are available commercially. The system is pictured in Figure 3. It is available in a variety of sizes. One disadvantage of this chromatography method is that an entirely separate assembly is required for each size operation.
The elution solvent is heated to reflux in the boiling flask and the vapors led up around the column to a condenser. The descending liquid runs onto the chromatographic solid and slowly percolates through the adsorbent moving the substrates short distances as it passes back to the boiling flask. Eventually, the fastest eluting substance in the mixture is transported into the boiling flask where it must be detected. Because the only way to collect fractions is to change the eluting solvent in the boiling flask it should be obvious that this method will only be convenient for mixtures with a very wide difference in Rfs and which are for some reason unsuitable to other methods. Of course, these are exactly the situations where one is looking for a special technique.
VACUUM LIQUID CHROMATOGRAPHY
Chromatographic resolution can be improved by using fine grades of adsorbent but if this is done some added impetus must be supplied to maintain a satisfactory flow rate and faster operation. O new answer to this difficulty is to maintain a pressure differential between the top of the column and the stopcock. Primitive efforts to use vacuum to enhance the solvent velocity resulted in bubble formation, extensive cracking and channeling when the vacuum was interrupted to change fractions. To prevent these problems in vacuum liquid chromatography:
The column is kept under vacuum continuously
The system uses a long narrower column
A preadsorption layer of Celite is used
The apparatus is illustrated in Figure 3. It should be noted that most of the components are readily available in the laboratory. E_F is a modified separatory funnel. H is a large cow more usually used in reduced pressure fractional distillation. If elegance is not vital concern the ground glass joint between the column and the fraction collector E can probably be replaced with a carefully bored vacuum-tight rubber stopper D with no loss of effectiveness. This would allow the upper part of the apparatus to be replaced by any available column. Because any size standard column can be substituted at will it would make the set-up readily adaptable to a variety of sample separations.
To begin the column is dry-packed with 10-40 μm sorbent using vacuum and a tamping tool to compact the layer. A Celite preadsorption layer of about 10% of the sorbent weight is packed in the same way. Although problems might be expected from solvent loss or bubble formation in the column these are not observed when operating at 2.5 mm of mercury the usual operating conditions. The cow, receivers, and eluent reservoir have sufficient ballast volume that the column can be run with valve G closed once a vacuum has been created. An occasional rotation of the tap is all that is reported to be needed to maintain it.
The authors have stated, “ Extensive tests on identical columns have shown that separation of material was at least twice as efficient using the VPC methods as using slurry or dry-packed low-pressure columns”. The procedure has the additional advantage that solvent programming or gradient elution is applicable without difficulty because the top of the column is unencumbered and at atmospheric pressure. The most obvious disadvantage of the methods is that contamination of small poorly separated fractions would seem to be inevitable since each droplet of eluent must run through a long stretch of glass tubing before reaching the receiver. This will be compounded by the increased tendency for the solvent to evaporate while running through the intermediate collection bulb and the impossibility to wash this container between fractions.
DRY COLUMN CHROMATOGRAPHY
The advocacy of dry column chromatography is largely on the hypothesis that high resolution such as is obtained in thin-layer chromatography is primarily attributable to development by capillary action on a dry but deactivated adsorbent which need not be a fine TLC grade.
The developers of this technique, Bernard Love and Marjorie Goodman claim direct transferability of TLC conditions to the dry column but emphasize the importance of checking to match the activity of the plates used in developing the analytical separation with the adsorbent used in preparing the column.
In the most convenient embodiment of the dry column technique deactivated column grade adsorbent mixed with a florescent indicator is dry packed into a nylon tube with a constricted lower end. The substrate mixture may be run onto the top of the column in a minimum amount of the eluting solvent but the preferred loading technique is by preabsorbing the mixture onto a small (5 X the weight) amount of the adsorbent and distributing this solid evenly on the top of the column and overlaying with sand or glass beads. The column is developed by slowly adding solvent until gravity and capillary action has drawn the solvent down the full length of the column. Because the nylon tube is transparent to UV light the separation of the bands may be followed using a handheld UV lamp to see the selective quenching of the fluorescent dye incorporated in the adsorbent. Cutting the column in sections with a sharp knife and extracting the individual components isolate the bands.
Because no bands are eluted out of the column much less solvent is used than the wet column displacement methods. The choice of solvents is somewhat restricted in this method because benzene or toluene completely obscure the fluorescence required to see separation; also with silica gel, but not alumina, methylene chloride and chloroform cannot be used they soften the nylon. Special adsorbent conditioning id required with mixed solvents. Although some common solvents are ruled out by choosing this technique other expensive and unusual solvents become possible. Since the method uses very little solvent otherwise expensive solvents can become acceptable. Also much less volatile solvents can be used because the solvent is removed while it is still adsorbed on the solid phase and this removal can be done in a vacuum oven. The combination of using stronger vacuum and spreading the solvent on a solid surface to make it more available to the reduced pressure accelerates the rate of solvent removal.
In rough terms for an easy separation, a delta Rf of 0.4 about 70 gm of adsorbent is needed per gram of mixture, while for a difficult separation one needs 300 gm of adsorbent for one gram of mixture.
The preparation of the dry column as described by the authors is certainly simple and the development time is reported to be 15-30 minutes. To this must be added the time required to grade the adsorbent both on the TLC plates and the deactivated adsorbent to be used in the column, since close correspondence is critical to success. For very long columns applying reduced pressure to the column as seen below can further shorten the time. Even if UV cannot see the separation the column can be sampled by puncturing it with a small syringe. Dye mixtures co-chromatographed with the sample can also be used to indicate a satisfactory place to cut fractions.
SHORT COLUMN CHROMATOGRAPHY
Both short column chromatography, which will be described in this section, and TLC mesh column chromatography, which will be described in the subsequent section, use relatively short stubby columns to conserve solvent and to reduce the time because the flow rate becomes much slower using fine grades of adsorbent. To compensate for the reduced resolution both use TLC grades of absorbent to increase the contact between the mixture constituents and the adsorbent particles, which is a surface effect.
In short column developed by B.J. Hunt and W. Rigby the column length to diameter ratios are in the range 7/3 to 15/10. For optimal performance, the column containing the adsorbent must be perfectly cylindrical with a sintered glass disk covering the entire bottom of the column. Any tapering of the working section of the column apparently leads to markedly inferior results. The top of the column should accept a flat flanged joint or some form of constriction and closure, which allows a modest pressure to be applied during development without interfering with the careful construction of the column. Although such equipment is ideal for reducing the dead volume and allowing the even development of bands the authors claim that many successful separations have been done in plateless tubes 20 to 30 cm long or in ordinary cylindrical separatory funnels with reasonably wide mouths. (B29) If such a set-up without a sintered glass disk is to be employed the conical portion above the stopcock must be completely filled with washed sand, to reduce this dead volume as much as possible. This review will only consider the more ideal situation of having the truly cylindrical column because the author feels that the less optimized situation would fare poorly in any comparison of advanced methods as are being described here.
Sample loading on short columns must ultimately be determined by experimentation but as a rule of thumb, a ratio of 1/33 is recommended for easy separations as judged by TLC and 1/100 for rather difficult ones. Short columns work more effectively than preparative layer plates for reasonable separations but they cannot compete in the case of very fine separations in which there is hardly any empty space between the resolved components, In these latter instances, only zero dead volume of preparative layer plates is satisfactory. Sensitive compounds, which readily oxidize, or photolyze when exposed on dry adsorbent are of course better treated on short columns. Additional advantages vis-à-vis plates are the avoidance both of the messy scraping operation and the time and material losses during extraction.
Selection of an elution solvent mixture for short column chromatography can be guided by TLC Rfs but must be adjusted for mixtures of solvents. The rule of thumb here’s to decrease the percentage of the more polar component by half For example if a satisfactory TLC separation of a mixture was performed using 2% ethyl acetate in benzene, the column for short column chromatography should be made up and eluted with 1% ethyl acetate in benzene.
Although it is not specifically discussed, it would appear that Rigby-Hunt procedure is very flexible with regard to the choice of eluant system, because a wider range of silent mixtures is mentioned i.e. Benzene, ethanol-benzene, ether-benzene, ethyl acetate-chloroform, ethyl acetate-chloroform-1% formic acid, butanol-acetic acid-water (6:2:2) or ethanol-water (6:4). Indeed it would seem from the final few examples that partition chromatography has been successfully scaled up. The authors note that solvents do not have to be specially dry before use. The TLC grade adsorbent is slurry packed using a slightly positive pressure. Kieselgel G (Merck) is recommended as a general-purpose adsorbent. Apparently, TLC grade adsorbents vary widely in their suitability for short column work mainly because the rate of flow through them is highly variable. The calcium sulfate in Kieselgel G perhaps acts beneficially as a filter aid. The hydration of the calcium sulfate during development with undehydrated solvents does not give any reported problems.
To pack the column the adsorbent is shaken briskly in a stoppered flask with 3.5 ml of the starting solvent mixture for each gram of adsorbent. The sintered glass disk is covered with a sheet of open texture filter paper (Whatman 41) to prevent fines from blocking the sintered glass disk and impeding the flow. The tap s opened so that the filter paper will not be lifted when the adsorbent slurry is added. Particularly at first, the moderately thin slurry is added cautiously using a large controlled-addition pipette. Thereafter, when an initial layer covers the filter paper, thereafter the slurry can be carefully poured. Once all the slurry is in the column it can be capped and a pressure of 10-30 mm of mercury applied. During the bedding down process, which normally takes less than an hour, the column must be maintained exactly vertical and left undisturbed. The slow settling produces very uniform packing. The authors have selected the ratio of solvent to adsorbent such that the solvent front should just catch up with the adsorbent bed when it has completely compacted. Roughly 150 grams of Kieselgel G should settle to a 300 ml final volume. When the column hiss settled a second Whatman 41 filter paper is placed onto the top of the layer as an even 2-3 mm layer of washed sand is spread over the filter paper using as an applicator a large diameter glass tube which has been narrowed at one end and which can have this opening opened or closed by movement of a collinear control rod whose one end completely blocks the constriction. This heavy non-adsorbing sand layer is needed to allow the even distribution of the small volume of the material to be chromatographed and also to protect the surface of the column.
The solvent is drawn down so that it is below the sand level but still covering the adsorbent. The solution to be chromatographed with selected marker dyes added, is carefully and uniformly applied to the sand layer using a large controlled addition pipette. The solution should be applied uniformly and symmetrically. A final open texture filter paper is layered onto the top of the sand. Although the authors do not specify it in their article, presumably the sample solution is allowed to run onto the top of the adsorption layer at this point. Sufficient solvent to complete the entire chromatography is then added; the initial portion being run in by pipette right at the center of the filter paper. The top is placed on the column and clamped in place ready for pressurization. Usually a pressure of 10-70 torr. is sufficient to get a satisfactory flow rate but with longer columns and less satisfactory adsorbents pressures up to 300 torr. Have been used by the authors. Under such circumstances. however, it would appear that the advantage of convenience has been lost and regular medium pressure chromatography becomes more rapid, more efficient, more convenient and safer.
Even under these circumstances, short column chromatography would remain a method for consideration when using less common adsorbents such as Florisil magnesia, sucrose, and cellulose or modified silicas such as silver nitrate-silica gel.
Because it is important to retain good band shape and because it is difficult to otherwise follow the separation of colorless compounds, one or more marker dyes should routinely be added to the mixture. This has the added advantage that one can collect large fractions before and after the region of interest, thereby reducing the region where a fraction collector must be used and reducing the number of small fractions, which must be analyzed by TLC. The authors have recommended a series of marker dyes for use in benzene-ethanol mixtures up to 50% ethanol and another set for use in benzene-ether up to 10% ether. These are listed below:
Benzene-ethanol Benzene-ether
Phenolindo-2, 6-dichlorophenol Desaga test mixture red
Ethyl phenolindo-2,6-dichlorophenol
Methyl red
Neutral red
Rhodamine B
Only 0.002 to 0.02 milligrams of each dye are needed on a 6 to 10-centimeter diameter column. Even where the dye constituent overlaps with a constituent of interest it is usually exceedingly easy to remove the colored material after separation is complete.
In terms of time saved the technique is not competitive with flask chromatography or Taber’s TLC Mesh Chromatography but there are certain savings in solvent usage compared to standard chromatography because high Rfs (0.1-0.9) can be used. That is to say, a component of Rf 0.5 should elute in two-column volumes.
What makes this methodology most often impractical is that dye mixtures are not available readily at hand and the dye mixture may not be applicable to the solvent system, which you want to use. Since one of the most attractive aspects of this method is that one can immediately use the TLC system which one has validated for analytical use, one would like the dye system which matches that solvent system to be immediately at hand.
Most probably using this method may be almost completely limited to elutions where partitioning is utilized is wet and highly polar systems.
TLC MESH COLUMN CHROMATOGRAPHY
Douglas F. Taber has described a technique called TLC mesh chromatography which has many things in common with sort column. Both use a fine grade of adsorbent, for Taber, EM 7747 silica gel, 10-15 μm from Scientific products, and short, wide columns of a 3:2 length to diameter ratio. There are however interesting differences. The adsorbent bed is first allowed to settle by gravity and then compacted by applying air pressure. The sample is applied by a pre-adsorption method using a coarse silica gel, 60-200 mesh, as the carrier. The author points out that this assures an even application of the mixture to the top of the column guarantees a very narrow initial band and avoids the problem of mixture constituents, which are sparingly soluble in the non-polar column solvent. Using this method also avoids the need for a preliminary clean-up column as is often required in flask chromatography. I agree that the adsorption method is probably the major contributor to the enhanced separation using this technique but it should not be overlooked that the even spreading of the preadsorbed later over the surface of the column proper is made easy when the top of the column is most accessible as in a wide stubby column.
While the procedure of Hunt and Rigby seems to aim at exactly reproducing TLC behavior, both for an adsorption and partition modes, Taber seems more concerned with challenging the speed of flash chromatography using adsorption chromatography only. He collects fractions equal in volume to the weight of the slid adsorbent i.e. For a one gram column he collects 1 ml fractions; for a 500 gm column he collects 500 ml fractions. The pressure or vacuum is adjusted to give one fraction per minute.
The polarity of the eluent is selected so that the components of interest and/or the components, which present the most serious separation problem, have Rfs of about 0.4 and come out near fraction 10. The author states that twenty fractions should be sufficient to elute all the mixture's components. Because of this speed, an automatic fraction collector is unnecessary. This rate does, however, require that only low viscosity solvents be used. Taber recommends as his first choice for all applications mixtures of ethyl acetate and petroleum ether, then when the proper polarity mixture contains less than 5% ethyl acetate a switch should be made to methylene chloride-petroleum ether mixtures or when the correct polarity contains more than 40% ethyl acetate a switch to ethyl acetate-methylene chloride should be made. As can be seen, Taber does not give any special consideration to special solute-solvent interactions in attaining separation but concentrates on generalizing polarity considerations.
Clearly, if preliminary TLC study has demonstrated specific eluent-substrate interactions that markedly improve the resolution, this method should not be used.
The paper’s author claims that the entire procedure including column preparation should take less than an hour; moreover, he claims substantially better resolution than flash chromatography while using a less costly grade of silica gel.
On the largest column sizes, the vacuum is used to enhance the rate of elution and the method becomes equivalent to vacuum preparative chromatography except that the slurry column packing method is used.
FLASH CHROMATOGRAPHY
(MEDIUM PRESSURE DRY COLUMN CHROMATOGRAPHY)
Flash Chromatography has become the most commonly used of all the advanced methods of chromatography. In fact, it probably now represents the standard chromatography.
Flask chromatography is designed to separate substances having Rf differences s > 0.15 in 10-15 minutes. This time includes column preparation, sample application, and complete elution. The authors describe it as a refinement of short column chromatography using a specific particle size adsorbent, a rapid optimized flow rate, and specific low viscosity solvent combinations, with a standard height column of 5-6”. Actually, flash chromatography is more closely akin to dry-column chromatography (as indicated in the subtitle). One might think that because in dry-column chromatography the eluent is being drawn down by capillary action that might make a significant difference in chromatographic behavior. Love and Goodman, however, have already noted that in using their technique “addition of further solvent to the column after the front has reached the bottom does to correspond to developing a liquid-filled column. First passing solvent through a dry column, putting the mixture to be separated on the column and then developing the column with the same, proved this solvent. The separation was the same as was obtained in a regular “dry column” and far superior to that obtained in a “liquid filled” column. This demonstrated the advantage of dry-column chromatography would still be obtained when the flash method for preparing the column is used.
In flash chromatography, the solvents used must be dried and distilled prior to use. A solvent system is selected which gives an Rf of 0.35 on analytical TLC plates (E. Merck No. 5705) for the midpoint between the two most difficult to separate constituents or if the separation is easy an Rf of 0.35 for the slowest moving constituent of interest. If the solvent mixture in the TLC has the more polar component <2% by volume, then its ratio must be cut by a factor of 2 in the preparative run; i.e. 1% ethyl acetate/ petroleum ether would become 0.55 ethyl acetate/petroleum ether. The authors provide a table for choosing the correct column diameter for different mixture sizes.
The columns, which are used, are short and wide. The taper at the bottom leasing to the Teflon stopcock is abrupt and as flat as possible. To prepare a column the short section of tube leading to the Teflon stopcock is first blocked with a small plug of glass wool and this covered with a 1/8-inch layer of 50-100 mesh sand. The adsorbent layer is put in place simply by dumping the appropriate amount of 230-400 mesh (40-63 μm) silica gel 60(E. Merck No. 9385) into the column so that after tapping it vertically on the bench the layer is between 5,5 and 6 inches high. Finally, a 1/8” level of sand is placed flat on top of the adsorbent bed and the entire assembly is clamped carefully in a vertical position for pressure packing. The solvent for the elution is then run carefully onto the top of the column. Enough is used to completely cover the adsorbent after the air has been displaced. The flow controller is placed on the column in the open position and the main compressed air turned on slightly. The pressure is quickly raised, compressing the adsorbent and forcing the solvent down through the bed. The pressure is maintained without interruption until the air is completely expelled from the column, including the stopcock, and until the initial exothermic reaction with the dry adsorbent is over. Otherwise, the column will overheat, particularly in the lower portion and crack.
The sample is applied as a 20-25% solution in the eluent. It is forced onto the top of the column with pressure. The pressure is then broken and the walls of the column above the layer of sand are carefully washed down with pipetted solvent. This liquid is also forced onto the column with pressure. Now all the solvent that was used to pack the column is carefully run back into the column as a hydrostatic head of eluent. At all times during these manipulations, the column must be filled to above the adsorption layer. The silica gel must not be allowed to become exposed. With all the solvent required for the chromatography
Already contained in the column the pressure head is attached and elution begun at a rate of 2” per minute by adjustment of the flow controller. Appropriate sized fractions are collected in tared test tubes and assayed by TLC (or another method). Columns can be regenerated particularly if the color has been removed in a preliminary treatment.
The authors emphasize that this technique is for rapid rough separations. Removal of tars or difficultly soluble portions of the mixture should be done ahead of time in clean-up columns. The technique does not have the same resolving power as other methods. Another disadvantage is that the grade of silica gel recommended is more expensive than either regular column grade or TLC grade. Finally, gradient elution is not possible because access to the top of the column is not available during the separation. The flow controller blocks it. Nevertheless this mode of chromatography
Has gained the widest acceptance of any because it quickly simplifies any separation problems. Commercial columns with flow controllers are available from glassware and chemical equipment suppliers.
CENTRIFUGAL PREPARATIVE LIQUID CHROMATOGRAPHY
Centrifugal force can be used to increase the rate at which elution passes through an adsorbent. The preparative methods based on this concept are essentially circular analytical TLC modified by increasing the thickness of the adsorbent layer and increasing the rapidity of the solvent movement by spinning the adsorbent layer around the central point of solvent application. Commercial apparatus is available. The system can accept up to 200 gm of silica gel and presumably even more of the denser alumina. No special adsorbent is needed and the centrifugal flurry packing is said to produce homogeneous layers even from material of different particle sizes with packing times of less than five minutes. The major shortcoming is that the sample size is somewhat limited compared to other methods. Difficult separations with up to 2 grams of a mixture are possible while for less complex more easily resolved materials 10 gm is the limiting range. Unlike other forms of chromatography as separation occurs the bands are continuously expanding into progressively narrower concentric circles. As the solvent front moves further from the center the centrifugal force becomes greater and the solvent feed becomes faster which further compresses each band until the entire adsorbent has been wet with elution solvent. This concentration of the bands significantly reduces tailing and increases separation efficiency and resolution. As each concentric ring of pure material elutes out of the adsorbent at the perimeter of the Hitachi apparatus it is collected in a funnel-shaped concentrator and run past a detector.
Hard undeformable adsorbents are preferred to porous organic polymers in this method. The solvent can be programmed during elution and solvent mixtures are not reported to produce trouble although one might expect that solvents of markedly different densities might show some tendency to migrate through the layer at different rates.
Centrifugal preparative chromatography could be used for some of the separations done by high-pressure preparative chromatography so long as peak shaving and recycling are not involved. Since the former allows the packing of one's own column it would seem to be much cheaper. The commercial instrument comes with several modes of automatic collection including using UV detection to control fraction changes. It would also seem that special adsorbents such as silver nitrate-silica or boric acid-silica gel could be used with this technology.
MEDIUM & HIGH-PRESSURE LIQUID CHROMATOGRAPHY
Medium pressure, which we will now discuss, and high-pressure liquid chromatography, which will be presented thereafter both utilize mechanical pumping to get a reasonable flow through columns that finely divided adsorbents. Both methods have been widely developed commercially and complete systems are available. The suppliers of these systems provide extensive technical support for users. For this reason, these subjects require no additional description in this place.
Medium pressure systems can be made of glass and interconnections can be made from Teflon. Because the columns can be transparent it is much easier to see whether high molecular weight colored materials have been left behind on the column. High-pressure chromatographic pumps, injectors, feed lines, and columns are metal because of the need for high pressures. The adsorbent is purchased in a plastic cartridge, which fits inside the metal chromatography
Apparatus. For this reason, it is not possible to prepare your own columns. With medium and low pressure using commercial columns, it is possible to pack your own columns but they are usually purchased flushed thoroughly after use and reused. Both medium and high-pressure chromatography use more expensive hardware than other methods already described and although they are fast and have high resolving power they are not conservative with solvent. High-pressure chromatography has the advantage of a recycling capability, which allows peak shaving and results in the almost total separation of a mixture with no unresolved intermediate fractions. Peak shaving is the selective collection of only the front or back portion of a band.
VACUUM DRY COLUMN CHROMATOGRAPHY
If large amounts of mixture are to be separated dry-column chromatography itself may not be sufficiently rapid. A further speed-up of the development can be obtained by applying a slight vacuum from the bottom of the column. A water aspirator has proven to be adequate. The pressure reduction inside the packed nylon tube may cause the flexible plastic wall to mold itself more tightly to the packing and reduce the void volume along the edge of the tube. If this is so, this method of chromatography will have some similarity with the patented radial compression used in Waters's high-pressure preparative chromatographic system. There is however no mention of this theoretical possibility in the publication describing this technique. The author recommends the method for special situations where speed is important even for small amounts of material such as the preparation of radioactive substances of extremely short half-lives and for materials that are easily decomposed on the adsorbent or which are very susceptible to air oxidation. No advantage in resolution is claimed over a regular dry column. The method uses inexpensive apparatus but some glass blowing is needed to modify standard equipment. The bottom of the column is made from a medium frit Buchner funnel of a size, which will just fit snuggly inside the nylon tube used in the column. Below the frit, the author has modified the exit tube by inserting a three-way stopcock. The three-way stopcock allows regulation of the vacuum and can be used to admit an inert gas for the protection of air sensitive compounds. This arrangement, however, requires a modified Buchner for each nylon tube diameter. Perhaps a better arrangement would be to use an unmodified funnel set into a plastic unbreakable filter bottle such as is now available commercially. The nylon tubing could be attached tightly to the funnel commercially available plastic straps.
A length of nylon tubing is selected to contain the required adsorbent with four to five inches left unpacked and the tubing is attached to the appropriate diameter Buchner funnel. Because the application of vacuum will cause the column to develop faster the length: diameter ratios can be higher than are usually used With a particularly difficult separation such a longer narrower path should be chosen along with an adsorbent: substrate ratio of >50:1. Vacuum fry-column chromatography makes use of deactivated adsorbent as in regular dry-column work and is simplest to conduct in single-solvent elution systems. If the best separations call for the use of mixed systems it is necessary to condition the water-deactivated adsorbent with 10% w. /v of the solvent mixture The original; author reports that there is a correspondence between Rfs on the column and on TLC plates except that the substances having TLC Rfs greater than 0.6 generally move to higher Rf on the column while conversely substances which move with Rfs below 0.4 in TLC are held back more tenaciously on the column. Between these values, the Rfs are closely comparable. Higher deactivation of the adsorbent is carried out as in regular dry column work. The Solent conditioning can be done simultaneously with the deactivation.
Because the nylon tube cannot now be packed by dropping it gently on a horizontal surface adding the adsorbent one-quarter at a time and gently tapping it with the hand or applying a vibrator packs it. A small amount of sand is applied to the top after the entire packing is firm. The mixture to be chromatographed is then packed as a 20-25% solid mixture with Celite above the sand.
The column is carefully and securely clamped in a vertical position. The large separatory funnel which is to dispense the eluent is firmly positioned above the column with the stem of the stopcock protruding down into the four or five inches of unfilled nylon tubing which had been left at the top of the column. This tubing is gathered tightly around the stem and fastened firmly there using plastic ties. At this point, if the compound is air-sensitive the full l column should be alternatively evacuated and refilled with inert gas several times. The solvent flow is then started with the column at atmospheric pressure. Only when the solvent has completely moistened the Celite and the sand layers and has reached the top of the adsorbent layer should the vacuum be applied. This procedure increases the likelihood of an even solvent front and is critical to avoid channeling in the column. In some cases, an uneven solvent front does develop but this does not seem to seriously affect the separation. Solvent addition to the top of the column is controlled by stoppering the solvent reservoir and opening the stopcock fully as in the usual procedure. Usually, a vacuum of about 100 mm of mercury is perfectly adequate to maintain a rapid flow rate. Typically a 150-gram silica column 21” long and 3” in diameter will be developed with methylene chloride in 15 minutes. For larger or very long columns >84” water vacuum aspirator still should be adequate but the vacuum should be somewhat better.
When the column development is complete the vacuum is broken by gradually admitting air or inert gas into the bottom of the column. The flow of solvent is interrupted leaving several inches of solvent at the top of the column. If the literature apparatus is being used the stopcock is closed. If a vacuum bottle is being used at the bottom of the column it must be quickly removed and the Buchner funnel spigot closed with a small septum, which should be wired securely. The separatory funnel is removed and the free nylon tubing folded over several times or wound onto a wooden dowel in order to press the remaining solvent into the column and to correct any slight cracking of the column which may have occurred when the vacuum was released. The top of the column is clamped to prevent the loss of solvent and the entire assemblage turned on its side. Using a scale and felt pen approximately appropriate Rf values for cutting the fractions are marked on the nylon-based on the TLC Rfs then a more exact determination is performed by taking small samples for either TLC of GC, using a small syringe. For fine particle size adsorbents, precautions must be taken to prevent blocking of the syringe. Either a bent needle or a special commercial needle can be used. If sampling proves to be difficult because of solvent loss additional liquid can be added either from the top, the bottom, or even by syringe directly into the middle of the column so that sampling will be facilitated. This difficulty is apparently particularly acute with very low boiling solvents such as ether. For this reason, if a choice is available for the eluent, preference should be given to those with boiling points above room temperature.
