Gas chromatography (GC) and gas-liquid chromatography (GLC) are the same technique described two different ways: in the vast majority of GC systems in routine use, the stationary phase is a liquid film coated on a solid support or on the inner wall of a capillary column, and the mobile phase is an inert gas — which is exactly what “gas-liquid chromatography” describes. The two terms are used interchangeably in practice, and this guide treats them as one topic. What actually separates one GC method from another, and what determines whether a method works at all, is a much narrower set of choices: which column, which carrier gas, and which detector. This guide covers all three, plus temperature programming and a troubleshooting reference for when a run stops behaving the way it should.
How a GC system works, in brief
A sample is injected (usually as a liquid, flash-vaporized in a heated inlet) into a stream of inert carrier gas, which sweeps it onto the column. The column sits inside a temperature-controlled oven; as the oven temperature rises through the run, compounds separate because they partition between the moving gas phase and the stationary liquid or solid phase differently depending on their volatility and chemical interaction with that phase. Compounds that interact more strongly with the stationary phase, or that are less volatile, take longer to elute. As each separated compound exits the column, it passes through a detector, which converts its presence into an electrical signal plotted against time as a chromatogram. Every design decision in a GC method — column, carrier gas, detector, and the oven’s temperature program — exists to control how well that separation and detection happen for the specific compounds in your sample.
Column selection: packed vs. capillary
The first fork in the road is column format. Packed columns are metal or glass tubes, typically 2-4 mm internal diameter, densely packed with a solid support material coated in the stationary phase. Capillary (open tubular) columns are long, narrow, flexible fused-silica tubes — typically 0.10-0.53 mm internal diameter and 15-105 m in length — with the stationary phase coated as a thin film (commonly 0.1-5 micron) directly on the inner wall, leaving the center of the tube open for gas flow. Capillary columns dominate modern GC because their open, unobstructed flow path and thin, uniform film produce far higher resolution and sharper peaks than a packed bed. Packed columns still see use for specific applications — permanent-gas analysis, some preparative-scale work, and situations needing high sample capacity or resistance to fouling — where their robustness outweighs the resolution penalty.
| Property | Packed column | Capillary column |
|---|---|---|
| Typical internal diameter | 2-4 mm | 0.10-0.53 mm |
| Typical length | 1-3 m | 15-105 m |
| Resolution | Lower | Much higher |
| Sample capacity | Higher | Lower (narrower bore, thinner film) |
| Typical use today | Permanent-gas analysis, some preparative work, rugged/dirty samples | The default choice for most analytical GC and GC-MS methods |
Stationary phase polarity
Within capillary columns, the stationary phase’s chemistry determines what property drives the separation. Nonpolar phases (100% dimethylpolysiloxane, sold under names such as DB-1, HP-1, or equivalent) separate compounds primarily by boiling point and are the standard general-purpose and “screening” choice. Mid-polarity phases (commonly 5% phenyl-substituted dimethylpolysiloxane, e.g. DB-5-type phases) are the single most widely used family, offering a workable balance of selectivity for a broad range of compound classes. Polar phases (polyethylene glycol-based, e.g. Carbowax/DB-WAX-type phases) separate compounds more by polarity and hydrogen-bonding character and are preferred for alcohols, free fatty acids, and other polar analytes that resolve poorly on nonpolar phases. Highly specific phases (cyanopropyl-substituted phases for fatty acid methyl esters, for example) exist for particular application classes and are generally chosen by matching the method to a validated application note for that compound class rather than by first principles alone.
Length, internal diameter, and film thickness: what each trades off
- Longer column: more theoretical plates and higher resolution, at the cost of longer analysis time and higher required inlet pressure (or lower flow) for a given carrier gas.
- Narrower internal diameter: higher efficiency and resolution per unit length, but lower sample capacity (easier to overload) and higher inlet pressure requirements.
- Thinner stationary-phase film: faster elution and less column bleed, but lower sample capacity and less retention for very volatile compounds, which can co-elute with the solvent front.
- Thicker film: more retention and capacity for volatile analytes, at the cost of longer analysis times, more column bleed at high oven temperatures, and reduced maximum operating temperature.
In practice, a 30 m x 0.25 mm x 0.25 micron column is the closest thing GC has to a default starting configuration for a general-purpose method, with narrower/thinner variants used to speed up fast-GC methods and wider/thicker variants used for trace volatiles or dirty, high-capacity samples.
Carrier gas selection: helium, hydrogen, and nitrogen
The carrier gas is not inert to method performance even though it’s chemically inert to the sample. Its choice affects separation efficiency, analysis speed, safety requirements, and running cost, and — because global helium supply has been repeatedly volatile in price and availability in recent years — more labs are actively re-evaluating a choice that used to be made once and never revisited.
| Carrier gas | Separation efficiency | Safety | Cost & availability | Best for |
|---|---|---|---|---|
| Helium | High efficiency across a broad range of flow rates; the traditional default, and directly compatible with mass spectrometer vacuum systems without modification | Inert, non-flammable, no special hazard controls needed | Non-renewable, geopolitically concentrated supply; price and availability have both been volatile | Labs that need a drop-in-compatible gas for existing MS interfaces and don’t want to manage a flammable gas |
| Hydrogen | The most efficient carrier gas by separation theory: its optimal linear velocity is higher than helium’s or nitrogen’s, so it supports faster runs with less efficiency loss at high flow than either alternative | Flammable; requires a documented risk assessment, leak detection, adequate ventilation, and — for generator-supplied systems — routine leak-check procedures before use | Can be generated on-site from a hydrogen generator, decoupling the lab from cylinder supply and helium market volatility; generally the lowest ongoing cost of the three | Labs converting away from helium for cost or supply reasons; fast-GC methods; routine analyses where the safety infrastructure is in place. Some unsaturated or reactive analytes can be altered by hot metal surfaces in a hydrogen stream, and GC-MS interfaces may need re-optimization when converting from helium |
| Nitrogen | Lowest efficiency of the three: its Van Deemter curve has a narrower, more sharply-defined optimal flow window, and efficiency drops off faster than helium or hydrogen outside that window | Inert, non-flammable | Cheapest option and easy to generate on-site with a nitrogen generator | Routine, less demanding separations with detectors such as FID or ECD where maximum resolution isn’t the limiting factor; generally not the first choice for high-resolution capillary work or GC-MS, where vacuum-pumping capacity and efficiency requirements favor helium or hydrogen |
Switching an existing method between carrier gases is not a drop-in change: flow rates, and sometimes the temperature program, need to be re-optimized for the new gas’s optimal linear velocity, and any switch to a flammable gas requires updating the lab’s safety documentation and gas-handling infrastructure before the switch, not after.
Detector selection
The detector determines what you can see and how selectively you can see it. Universal detectors respond broadly to almost anything eluting from the column; selective detectors respond strongly to specific elements or functional groups and largely ignore everything else, which is often exactly what’s needed to pull a target analyte cleanly out of a complex matrix.
| Detector | What it detects | Relative sensitivity | Destructive? | Typical applications |
|---|---|---|---|---|
| FID (Flame Ionization Detector) | Almost universal for organic compounds containing carbon-hydrogen bonds; largely blind to permanent gases, water, and fully inorganic species | High, with a wide linear dynamic range | Yes — burns the sample in a hydrogen/air flame | Hydrocarbon and petrochemical analysis, environmental volatiles, food and flavor compounds — the default general-purpose GC detector |
| TCD (Thermal Conductivity Detector) | Truly universal — responds to any compound whose thermal conductivity differs from the carrier gas, including permanent gases (CO2, CO, N2, H2) that FID can’t see | Lower than FID | No — non-destructive, so it can be run in series with a second detector | Permanent-gas and fixed-gas analysis (natural gas composition, headspace gas analysis) where inorganic gases must be measured |
| ECD (Electron Capture Detector) | Highly selective for electronegative compounds — halogenated compounds and nitro compounds especially | Very high for its target compound classes (down to trace levels) | Not chemically destructive, but uses a sealed radioactive source (commonly nickel-63) that requires regulatory licensing and periodic wipe testing | Organochlorine pesticides, PCBs, and other halogenated environmental contaminants |
| MS (Mass Spectrometer, GC-MS) | Effectively universal, and uniquely provides structural identification via mass spectra that can be matched against a reference library, not just retention time | Very high, and confirmatory rather than just quantitative | Yes — ionizes and fragments the sample under vacuum | Compound identification and confirmation across forensics, environmental, pharmaceutical, and metabolomics work; the standard choice when you need to know what a peak is, not just that it’s there |
| NPD (Nitrogen-Phosphorus Detector) | Selective for nitrogen- and phosphorus-containing compounds, with greatly reduced response to plain hydrocarbons compared to FID | High for N/P compounds | Yes — flame-based, using a heated alkali-metal bead that degrades with use and needs periodic replacement | Organophosphate pesticide residues, drug and toxicology screening |
| FPD (Flame Photometric Detector) | Selective for sulfur- and phosphorus-containing compounds via characteristic flame chemiluminescence, measured through an optical filter | High for S/P compounds | Yes — flame-based | Sulfur speciation in petroleum products, organophosphate pesticides, environmental sulfur compounds |
Detectors can be combined — TCD’s non-destructive design in particular allows column effluent to be split to a second detector — and many labs run FID as a general-purpose workhorse while reserving GC-MS for samples that need positive identification, not just a retention-time match against a standard.
Temperature programming
Running a GC oven at a single fixed temperature (isothermal) works only for samples with a narrow volatility range. Most real samples contain compounds spanning a wide range of boiling points, and an isothermal run forces an unworkable compromise: a temperature low enough to resolve the earliest, most volatile peaks leaves the latest, least volatile compounds taking impractically long to elute (or not eluting at all), while a temperature high enough to elute the latest peaks in reasonable time crowds and co-elutes the earliest ones. Temperature programming solves this by ramping the oven temperature over the course of the run, so early peaks elute while the column is still cool (maximizing their resolution) and later, less volatile compounds are driven off as the oven heats, keeping total run time practical.
A generic, illustrative temperature program — the kind used as a teaching starting point, not a validated method for any specific analyte — looks like this:
- Initial oven temperature: 40°C, held for 2 minutes (allows the most volatile compounds to separate while the column is cool)
- Ramp 1: increase at 10°C/minute up to 250°C
- Final hold: 250°C for 5 minutes (elutes and clears the least volatile compounds, and bakes residual material off the column before the next injection)
More complex methods commonly use multiple ramp segments at different rates — for example, a slow initial ramp through a region where several target compounds are known to be closely spaced, followed by a faster ramp through a region with no analytes of interest, to shorten total run time without sacrificing resolution where it matters. The specific initial temperature, ramp rate(s), and final hold for any real method should come from a validated method or method-development process for the actual analytes and matrix involved, not from a generic template.
Troubleshooting reference
| Symptom | Likely cause | What to check |
|---|---|---|
| Peak tailing | Active sites in the flow path, a contaminated or poorly-deactivated inlet liner, or column degradation | Replace or clean the inlet liner, check for exposed active sites at the column’s front end (trim a short section if needed), confirm the column hasn’t exceeded its maximum temperature rating |
| Ghost peaks or rising/drifting baseline | Column bleed (accelerated by exceeding the column’s rated maximum temperature), septum bleed, or system contamination | Confirm the oven program stays within the column’s rated temperature limits, replace the septum, run a bake-out, check for contaminated carrier gas or gas lines |
| Loss of resolution over time | Column degradation, a damaged or contaminated inlet liner, or carrier gas flow drifting from its set point | Verify actual carrier gas flow/pressure against the method’s setpoint, inspect and replace the liner, consider trimming or replacing the column |
| Retention time shifts between runs | Carrier gas leaks, gas purity issues, or oven temperature calibration drift | Leak-check fittings and connections, verify gas purity/supply, confirm oven calibration against a reference |
| Split or “shouldering” peaks | Injector discrimination, a poorly-deactivated liner, or mismatched injection technique for the sample solvent | Check liner deactivation and inertness, review injection volume/speed and inlet temperature relative to the solvent and analytes |
| Weak or absent signal | Detector fault, interrupted or depleted gas supply, or a cold spot in the flow path (transfer line, detector base) | Confirm all gas supplies (carrier, and detector-specific gases such as hydrogen/air for FID) are flowing and at correct pressure, verify detector temperature settings, check for a leak upstream of the detector |
| Elevated baseline noise | Detector contamination, carrier or detector gas purity, or an electrical grounding issue | Clean or replace detector components per the manufacturer’s maintenance schedule, verify gas purity grade matches the detector’s requirements, check grounding |
Frequently asked questions
Is gas chromatography the same as gas-liquid chromatography?
In routine practice, yes. Almost all GC methods in use today employ a liquid stationary phase (coated on a solid support or a capillary wall) with a gas mobile phase, which is precisely what “gas-liquid chromatography” describes. The terms are used interchangeably, and “GC” is now the more common shorthand.
Which carrier gas is best for GC-MS?
Helium has historically been the default because it’s directly compatible with mass spectrometer vacuum-pumping systems without modification. Hydrogen is increasingly used as a substitute, particularly where helium cost or supply is a concern, but converting an existing GC-MS method to hydrogen typically requires re-optimizing flow rates and, in some cases, the MS interface, and should be planned as a deliberate method change rather than a simple gas-cylinder swap. Nitrogen is generally not favored for GC-MS because its lower separation efficiency and the vacuum system’s pumping capacity make it a poor match for most mass spectrometer interfaces.
Why is my column bleeding?
Column bleed — a rising baseline, especially at higher oven temperatures — is normal at some level as the stationary phase slowly degrades with heat and use, but it accelerates sharply if the column is run above its rated maximum temperature, if it’s been exposed to reactive contaminants, or as it nears the end of its useful life. Persistent excessive bleed that doesn’t resolve with a bake-out or liner replacement usually means the column needs replacing.
Can nitrogen be used as a GC carrier gas with any detector?
Nitrogen works with FID and ECD for routine, less demanding separations, but its lower separation efficiency makes it a poor choice where maximum resolution is required, and it’s generally unsuitable for GC-MS given typical vacuum-system pumping capacity.
How do I choose between FID and TCD?
Choose based on what you need to see. FID is far more sensitive for organic, carbon-containing compounds but is essentially blind to permanent gases (CO2, CO, N2, H2, water). TCD is far less sensitive but is truly universal, including for the permanent gases FID can’t detect. Labs analyzing gas composition (natural gas, headspace gases) generally need TCD; labs analyzing organic volatiles generally default to FID.
Related reading
For the equipment upstream and adjacent to a GC workflow, see UV-Vis Spectrophotometer Basics for a comparison against a different analytical detection approach, and Pipette Calibration: How and When to Calibrate Lab Pipettes and pH Meter Calibration: Buffer Selection and Best Practices for the broader instrument-calibration discipline this guide’s troubleshooting section draws on. For lab accreditation requirements that govern how GC methods are validated and documented, see ISO/IEC 17025: What It Actually Accredits and How It Differs from ISO 9001. For informatics systems that track GC method data and instrument records, see What Is a LIMS System? and the related LIMS software comparison. If hydrogen carrier gas is part of your safety planning, see Chemical Spill Kits: What to Stock and How to Respond to a Lab Spill and PPE Selection for Chemical Handling in the Lab for adjacent hazard-response planning.
