How to Choose a GC Column: Types, Phases and Dimensions Explained
Key Takeaways
Choose the phase first. Match the polarity of the phase to the polarity of your compounds.
The safe default is a 5% phenyl column - DB-5ms, HP-5ms, Rtx-5ms or ZB-5ms. Start here for unknown or mixed samples.
Narrow columns give sharper peaks. Wide columns take bigger samples. That is the main trade-off.
Thick films hold onto light, volatile compounds. Thin films get heavy compounds off the column faster.
Change length last. Doubling the length adds about 40% more separating power but doubles the run time.
If your method quotes a USP G-number, the phase is already chosen for you.
What is a column in GC?
A GC column is a length of tubing with a coating inside it. That coating is the stationary phase, and it is what does the separating.
Here is what happens. Your sample is vaporised and pushed through the tubing by an inert gas - helium, hydrogen or nitrogen - called the carrier gas. Compounds that stick to the coating more strongly move more slowly, so they come out later. Compounds that barely stick come out early. That difference in travel time is your separation. Everything else in the instrument exists to get the sample onto the column and to detect what comes off the other end.
Columns come in two physical shapes: thin hollow tubes (capillary) and wider tubes filled with packing material (packed). Capillary columns handle almost all modern work. Our capillary versus packed comparison covers when you still need the packed kind.
What are the two main types of GC columns?
The two main types are capillary columns and packed columns. Capillary columns separate far better and are the normal choice. Packed columns hold much bigger samples and are still named in some older methods and gas analyses.
Capillary columns come in three builds:
Capillary build | Where the coating sits | What it is for |
|---|---|---|
WCOT | Coated straight onto the tube wall | Almost everything - this is the standard column |
PLOT | A porous solid layer on the wall | Gases like methane and carbon dioxide |
SCOT | Coating on a support layer on the wall | Old methods only; rarely sold now |
WCOT is what you get unless the catalogue says otherwise.
PLOT columns exist for one reason. Very light compounds - methane, ethane, carbon dioxide - barely stick to a liquid coating at usable temperatures, so they all rush out together. A porous solid layer grips them properly. If you analyse permanent gases, you need PLOT.
Choosing the stationary phase
This is the decision that matters most. The table below covers the phases behind almost all routine GC work. Every manufacturer sells each one - they just use different names for the same chemistry.
Phase chemistry | Polarity | Sold as | Good for |
|---|---|---|---|
100% dimethylpolysiloxane | Non-polar | DB-1, HP-1, Rtx-1, ZB-1, BP-1, SPB-1 | Hydrocarbons, solvents, general screening |
5% phenyl / 95% dimethylpolysiloxane | Non-polar | DB-5ms, HP-5ms, Rtx-5ms, ZB-5ms, BPX5, SLB-5ms | The all-rounder: pesticides, drugs, GC-MS screening |
6% cyanopropylphenyl / 94% dimethylpolysiloxane | Mid-polar | DB-624, Rtx-624, ZB-624, BP-624, VOCOL | Residual solvents and volatiles - the USP G43 phase |
14% cyanopropylphenyl / 86% dimethylpolysiloxane | Mid-polar | DB-1701, Rtx-1701, BPX-10 | Pesticides, alcohols, drug impurities |
50% phenyl / 50% dimethylpolysiloxane | Mid-polar | DB-17, DB-17ms, Rtx-17, BPX-50 | PAHs, steroids, triglycerides |
Polyethylene glycol (PEG, also called Carbowax) | Polar | DB-WAX, HP-INNOWax, Stabilwax, ZB-WAX, Supelcowax-10 | Alcohols, fatty acids, flavours, fragrances |
Acid-modified PEG (FFAP) | Polar | DB-FFAP, Stabilwax-DA, Nukol | Acidic compounds that tail badly on normal PEG |
Highly cyanopropyl | Very polar | SP-2560, HP-88, Rt-2560, BPX70 | Fatty acid methyl esters, cis/trans isomers |
The rule of thumb is simple: like holds like. Non-polar coatings mostly separate by boiling point, so hydrocarbons come out in size order. Polar coatings add extra chemical attraction, which is how a PEG column pulls apart alcohols that a non-polar column would bunch together.
Two warnings save a lot of money.
PEG columns die easily. Their maximum temperature is low - usually around 250-260 C. Go above it, or let air in while the column is hot, and the coating is ruined for good. Treat them as shorter-lived than the silicone-based phases.
The "ms" suffix matters. DB-5ms and ZB-5ms are low-bleed versions made for mass spectrometry. Use them for GC-MS. A standard column sheds more coating, which lifts your baseline and dirties the ion source.
What is a DB-624 GC column?
A DB-624 is a mid-polar capillary column used for residual solvents and volatile compounds. Its coating is 6% cyanopropylphenyl / 94% dimethylpolysiloxane.
This is the USP G43 phase, which is why pharmacopoeial residual solvent methods name it so often. The same column from other suppliers is the Rtx-624, ZB-624, BP-624 or VOCOL. Same chemistry, different badge.
DB-624 columns usually come with a thick coating - 1.4 um or 1.8 um. The compounds being measured are very volatile, so they need the extra coating to be held back long enough to separate. If your method quotes a G-number rather than a product name, our USP cross-reference table resolves it.
Internal diameter, film thickness and length
Once you have the phase, these three settle the practical trade-offs.
Dimension | Smaller value gives | Larger value gives | Go larger when |
|---|---|---|---|
Internal diameter (0.10-0.53 mm) | Sharper peaks, faster runs | Bigger sample capacity, copes with dirty extracts | Samples are concentrated or messy |
Film thickness (0.10-5.0 um) | Heavy compounds elute sooner, less bleed | Holds onto volatiles, more capacity | You are measuring gases or early solvents |
Length (10-60 m) | Shorter runs, lower pressure | More separating power | Two peaks still overlap after everything else |
A 30 m x 0.25 mm x 0.25 um column is the standard starting point, and it is the right answer more often than not.
Move up to 0.32 mm when you need to inject more. Move to 0.53 mm when you are converting an old packed-column instrument. Drop to 0.18 mm or 0.10 mm only when run time genuinely matters.
On length, the maths is unkind. Separating power rises with the square root of length. Going from 30 m to 60 m gives you roughly 1.4 times the resolution for twice the run time and twice the price. That is why length is the last thing to change, not the first.
Is GC better than HPLC?
Neither is better. They handle different samples.
GC needs your compound to turn into a vapour and survive the heat without breaking down. In practice that means a boiling point below roughly 300 C, or a compound you can chemically modify to get it there. HPLC handles everything GC cannot: heavy, fragile, non-volatile and ionic compounds.
Question | Gas chromatography | Liquid chromatography |
|---|---|---|
What the compound must be | Vaporisable and heat-stable | No such requirement |
Typical column | 15-60 m of thin tubing | 5-25 cm packed steel tube |
What carries the sample | Inert gas | Solvent, used continuously |
Separating power per column | Very high | Lower, but made up for by fine packing |
Typically used for | Solvents, volatiles, fatty acids, environmental and forensic work | Drugs, proteins, sugars, non-volatile actives |
Most labs that run one eventually need the other. Where a compound could go either way, GC usually separates isomers better and costs less to run, because carrier gas is far cheaper than HPLC-grade solvent.
Guard columns, maintenance and dirty samples
A guard column is a short piece of empty, deactivated tubing fitted in front of your real column. It costs a fraction of the price and takes the damage that sample residue would otherwise do to the front of your coating. If you inject plant extracts, soil extracts or biological fluids, fit one. Then trim the guard instead of replacing the column.
When peaks start looking wrong, work through it in this order.
What you see | Usual cause | What to do first |
|---|---|---|
Tailing on some compounds only | Active sites in the inlet liner or column front | Change the liner and septum, then trim 0.5-1 m off the column inlet |
Baseline climbing at high oven temperature | Coating bleeding, or air damage | Check carrier gas purity and traps; confirm the column temperature limit |
Every peak coming out later | A leak, or carrier flow dropping | Leak-check the fittings and confirm flow before changing the method |
Early peaks split or shouldered | Injection band too wide, or wrong solvent | Inject less; check your solvent against the coating polarity |
Retention suddenly gone altogether | Coating stripped by air or overheating | The column is usually finished - replace it |
Two habits save more columns than any cleaning procedure. Never exceed the stated maximum temperature, and never let air into a hot column - leave the carrier gas on until the oven has cooled.
Columns for water analysis
Water reaches a GC by purge-and-trap, headspace or solvent extraction, and the route decides the column.
Volatile compounds purged out of water run on a 6% cyanopropylphenyl column - DB-624 and its equivalents, the same G43 chemistry used for residual solvents. Heavier compounds extracted into solvent run on a 5% phenyl low-bleed column for GC-MS.
One thing not to do: never inject liquid water onto a PEG column. Water attacks the coating and the column will not recover.
A note on availability
LS Scientific does not stock GC capillary columns. That is why this guide carries no prices and names no supplier. Where a column is named above, it is named because a method or a USP code requires that chemistry - not as a recommendation to buy it from anyone in particular.
LS Scientific does supply the instruments these columns run on, including the Thermo Scientific ISQ7610 single quadrupole GC-MS, and the derivatisation reagents used to bring stubborn compounds into GC range.
Where to go next
Deciding between column formats? Capillary vs packed GC columns sets out where each one still applies.
Method quotes a USP G-number? USP GC column designations and cross-brand equivalents turns the code into a real product from each manufacturer.
Choosing the instrument, not the column? Best GC-MS systems in 2026 compares the four platforms most UK and Nigerian labs look at.
Comments