Column chromatography separates the components of a mixture by passing a solvent (the mobile phase) through a tube packed with an adsorbent solid (the stationary phase, almost always silica gel). Each compound in the mixture partitions between the two phases according to its polarity: less polar compounds spend more time dissolved in the mobile phase and travel down the column faster, while more polar compounds adsorb more strongly to the silica and travel more slowly. The result, run correctly, is a series of bands that elute from the column and can be collected as separate fractions.
This is a preparative technique — the goal is to isolate milligram-to-gram quantities of a purified compound, not just to detect or quantify it. It is almost always run after thin-layer chromatography (TLC) has already been used to scout a workable solvent system and confirm the mixture is separable at all. If you have not yet run TLC on the mixture, do that first: the Rf values it gives you are the direct input to the solvent-selection step below.
Choosing the Stationary Phase and Column Format
Silica gel is the default stationary phase for the large majority of organic separations because it is cheap, chemically inert to most reagents, and available in a range of particle sizes. Alumina is used for a smaller set of applications (notably compounds that decompose or streak badly on the mildly acidic surface of silica), but the procedure below assumes silica.
The particle (mesh) size of the silica determines how the column is run and how it must be packed:
| Format | Typical mesh size | Approx. particle size | How solvent moves | Typical run time |
|---|---|---|---|---|
| Gravity column | 60–230 mesh | roughly 63–200 µm | Solvent drips through under gravity alone | Tens of minutes to a few hours |
| Flash column | 230–400 mesh | roughly 40–63 µm | Compressed air or nitrogen forces solvent through under low positive pressure | Minutes to tens of minutes |
Finer (flash-grade) silica gives sharper bands and better resolution for a given column length because the smaller, more uniform particles pack more densely and reduce the diffusion path between particles, but it is too fine to flow under gravity alone in a reasonable time — hence the need for applied pressure. Coarser silica is slower and gives somewhat broader bands but needs no pressure source and is the standard choice for a simple gravity separation. Exact mesh ranges are set by the manufacturer and vary by supplier, so treat the ranges above as a guide, not a specification, and check the label on the bottle you are using.
As a starting point for column and silica sizing, a widely used rule of thumb is 20–100 g of silica for every 1 g of crude sample, with the lower end of that range for a well-separated mixture (large Rf difference between the target compound and its nearest neighbour on TLC) and the higher end for a difficult separation (closely spaced spots). The column length-to-diameter ratio is typically kept around 8:1 to 10:1 for gravity columns; flash columns run shorter relative to their diameter because the applied pressure substitutes for some of the length otherwise needed to develop resolution.
Selecting the Solvent System
The solvent system is chosen from the TLC result, not guessed independently. The general target is a solvent (or solvent mixture) in which the compound of interest runs at an Rf of roughly 0.2–0.3 on TLC using the same silica. An Rf in that range on TLC translates to a compound that elutes reasonably promptly from a column without co-eluting with a more polar neighbour that has barely moved.
Most solvent systems are binary mixtures — a nonpolar solvent (commonly hexane or heptane) combined with a more polar one (commonly ethyl acetate) — because a mixture gives finer control over polarity than switching between pure solvents. The table below lists common column chromatography solvents in order of increasing polarity (the eluotropic series for a silica stationary phase); moving down the table increases how strongly a solvent displaces adsorbed compounds from the silica surface, which is why a gradient elution moves down this list over the course of a run.
| Solvent | Relative polarity on silica | Common role in a solvent system |
|---|---|---|
| Hexane / heptane | Lowest | Nonpolar carrier; used alone for very nonpolar mixtures (e.g. some hydrocarbons) |
| Toluene | Low | Slightly more eluting power than hexane while still weakly polar |
| Dichloromethane (DCM) | Low–moderate | Common alternative polar component; higher eluting strength than ethyl acetate mixtures at equivalent volume fraction |
| Ethyl acetate | Moderate | Most common polar component, paired with hexane; typical starting mixtures range from about 5% to 40% ethyl acetate in hexane |
| Acetone | Moderate–high | Alternative polar component when ethyl acetate does not resolve close-running spots |
| Ethanol / methanol | High | Used in small percentages (often 1–10%) added to DCM or ethyl acetate for polar compounds that otherwise streak or will not move |
Two practical points follow from this. First, if the initial mixture pushes the target compound’s Rf too high (it runs near the solvent front), reduce the proportion of the polar solvent; if the Rf is too low (the compound barely moves), increase it. Second, for mixtures containing compounds of noticeably different polarity, a single isocratic (constant-composition) solvent system often cannot resolve everything: increasing polarity in steps over the course of the run (a stepwise gradient) is the standard fix, and is discussed further below.
Packing the Column
The two standard packing methods are wet (slurry) packing and dry packing. Wet packing is more common in practice because it is easier to pack the silica bubble-free.
- Select and prepare the column. Choose a column with the length-to-diameter ratio noted above for the amount of silica calculated. Clamp it vertically. Add a small plug of glass wool or a fritted disc at the bottom to retain the silica, followed by a thin layer of sand if using glass wool, to give an even, flat base.
- Prepare the silica slurry. Slurry the calculated mass of silica gel in the least-polar solvent of your chosen system (or pure hexane/heptane) in a beaker, stirring to break up clumps and release trapped air.
- Pour the slurry. With the stopcock open and a small amount of solvent already in the column, pour the slurry down the column in a steady stream, rinsing the beaker with additional solvent to transfer all the silica. Keep solvent draining from the stopcock throughout so the bed settles under a continuous flow rather than a static one.
- Consolidate the bed. Tap the column gently and continuously (by hand or with a wooden rod, never metal on glass) as the silica settles, to dislodge air bubbles and encourage even packing. Air pockets and channels formed at this stage are the single most common cause of poor separation later, so this step is worth doing carefully rather than quickly.
- Add a top layer. Once the silica bed has settled and the solvent level has dropped to just above the silica surface, add a thin layer of sand on top. This protects the flat bed surface from being disturbed when solvent or sample is added.
- Never let the column run dry. At every stage from this point forward, keep the solvent level above the top of the silica (or sand) bed. A silica bed that is allowed to dry out cracks and channels, and the separation on a cracked column cannot be trusted even if you re-wet it.
Loading the Sample and Running the Column
- Load the sample. Once the solvent level sits just at the top of the sand layer, carefully add the sample — either as a concentrated solution in the starting solvent (for compounds that dissolve well), or pre-adsorbed onto a small amount of silica and applied as a dry powder (dry loading, generally preferred for compounds with limited solubility in nonpolar solvent, since it avoids diluting the top of the band before the run even starts).
- Add solvent and begin elution. Once the sample has been drawn onto the top of the bed, carefully add solvent down the side of the column (or through a solvent reservoir on top, for flash columns) without disturbing the sand layer, and open the stopcock to begin collecting eluent.
- Maintain a steady flow rate. For a gravity column, a flow rate of roughly one drop per second is a common practical target — fast enough to finish in a reasonable time, slow enough that the compounds have time to re-equilibrate between the two phases at each point down the column, which is what actually produces resolution. For a flash column, the applied pressure is adjusted to give a comparably steady, controlled flow rather than a rapid flood.
- Switch solvents for a gradient, if needed. If the solvent system was chosen as a stepwise gradient, increase the proportion of the polar solvent in planned increments (for example, 10% steps) once the least-polar target compound has eluted, rather than all at once — an abrupt jump in polarity can cause a slug of trapped, poorly resolved material to elute all at once.
Monitoring the Separation and Collecting Fractions
Fractions are collected in a rack of test tubes or vials, typically sized at somewhere between one and a few column volumes each, and the correct set of tubes to combine is determined by monitoring, not by guesswork:
- TLC co-spotting. The standard method: spot a sample of each collected fraction (or every second or third fraction, if there are many) onto a TLC plate alongside the original starting material, develop it in the same or a similar solvent system used on the column, and visualize under UV light and/or with a chemical stain. Fractions whose spots match in Rf and appear clean (a single spot, or the target spot free of others) are combined; fractions containing a mixture of spots are set aside or re-chromatographed.
- UV monitoring, where available. For compounds with a UV chromophore, an in-line UV detector or periodic checks with a UV-Vis spectrophotometer can flag which fractions contain absorbing material, which is a useful first pass before committing to TLC on every tube.
- Visual banding. For colored compounds, the bands are visible directly on the column and fraction collection can be timed to the band boundaries, though TLC confirmation of purity is still standard practice even when the band is visible.
Once fractions are combined by TLC match, the pooled solvent is removed — almost always by rotary evaporation — to recover the purified compound, which is then typically weighed and re-checked by TLC (and by whichever downstream method, such as NMR or mass spectrometry, is used to confirm identity and purity) before use.
Troubleshooting
| Problem | Likely cause | Fix |
|---|---|---|
| Bands are diffuse or “streaky” rather than tight | Column overloaded with sample relative to silica mass; air bubbles or channels in the bed; compound is somewhat unstable or reactive on silica | Reduce sample load or repack with more silica; repack the column, tapping thoroughly to eliminate air pockets; consider adding a small percentage of a base (e.g. triethylamine) to the solvent if the compound is base-sensitive on silica, or switch to alumina |
| Solvent runs through unevenly on one side, or a visible crack/channel appears in the bed | Air pocket introduced during packing, or the column was allowed to run dry at some point | Repack the column from scratch; never let the solvent level drop below the top of the silica bed during a run |
| Two compounds that separated cleanly on TLC still co-elute from the column | Sample overloading (too much material for the column diameter/length); flow rate too fast for the compounds to re-equilibrate; solvent system’s eluting power too high for the resolution needed | Use a longer or larger-diameter column with proportionally more silica; slow the flow rate; reduce the proportion of polar solvent and consider a stepwise gradient instead of a single isocratic mixture |
| Nothing elutes even after several column volumes | Compound is more polar than expected and is stuck near the top of the bed; solvent system chosen was too weak | Increase the polar-solvent proportion in controlled steps until the compound begins to move, guided by TLC of the top of the bed if accessible |
| Compound elutes in the very first fractions, mixed with fast-running impurities | Solvent system too strong (too polar) for this compound | Re-run with a lower proportion of polar solvent, re-targeting an Rf of roughly 0.2–0.3 on TLC |
| Poor recovery of expected mass after evaporation | Compound is more polar than the solvent system used and was left on the column; product decomposed on silica; mechanical loss during transfer | Flush the column with a much more polar solvent (or pure methanol) at the end of the run and check by TLC/mass for remaining product; consider alumina or a shorter contact time if decomposition is suspected |
Frequently Asked Questions
How is the solvent system for column chromatography chosen?
From a prior TLC run on the same mixture and the same (or very similar) silica. The target is a solvent or solvent mixture that gives the compound of interest an Rf of roughly 0.2–0.3 on TLC; that solvent system is then used, or used as the endpoint of a stepwise gradient, on the column. See the solvent-selection table above for the relative polarity of common solvents.
What is the difference between flash and gravity column chromatography?
Gravity columns use coarser silica (roughly 60–230 mesh) and rely on gravity alone to move solvent through the bed, which is slow but requires no special equipment. Flash columns use finer silica (roughly 230–400 mesh) and use applied air or nitrogen pressure to force solvent through at a useful rate, which the fine particle size would otherwise make impractically slow under gravity. Flash columns generally give sharper bands and run faster; gravity columns are simpler to set up.
How much silica gel do I need for a column?
A commonly used starting ratio is roughly 20 to 100 grams of silica per gram of crude sample, with the lower end appropriate for a mixture that separates easily (large Rf difference between spots on TLC) and the higher end for a difficult, closely-spaced separation. This is a starting point to be adjusted based on how the separation actually performs, not a fixed rule.
Why did my column chromatography separation fail to resolve two compounds?
The most common causes are overloading the column with too much sample for its size, running the solvent too fast for the compounds to re-equilibrate between the stationary and mobile phases, or using a solvent system with too much eluting power for the resolution needed. See the troubleshooting table above for the corresponding fixes.
Can column chromatography be run without TLC monitoring?
It is possible for strongly colored compounds where bands are visible directly on the column, but TLC (or an equivalent method such as in-line UV or mass detection) is standard practice because a visible band does not guarantee the fraction is a single pure compound. For colorless compounds, monitoring by TLC or another detection method is effectively required to know which fractions to combine.
Related Guides
- Thin-Layer Chromatography: How to Run a Plate and Calculate Rf — the scouting step that determines the column’s solvent system and target Rf.
- HPLC: Columns, Mobile Phases, and a Peak-Problem Troubleshooting Table — the analytical, pressurized counterpart to open-column chromatography.
- Gas Chromatography: Columns, Carrier Gases and Detectors Explained
- LC-MS Explained: How Liquid Chromatography and Mass Spectrometry Are Coupled
- UV-Vis Spectrophotometer Basics: How It Works and Measures Concentration — useful for monitoring fractions from UV-active compounds.
- Analytical Balance Calibration: Weighing and Technique — for accurately weighing silica and sample before a run.
- How to Read a Safety Data Sheet (SDS): The 16-Section GHS Guide — column solvents such as hexane, DCM and methanol carry flammability and toxicity hazards worth checking before use.
- How to Write a Lab SOP: Step-by-Step Template and Guide
