Specifying a gas chromatograph starts with the analyte list, not the brand sheet: each compound class pushes the decision toward a specific detector, column phase, and oven program, and getting that pairing wrong is the single largest source of field rework.
Lab-grade instruments and online process GCs share the same physical separation principle but diverge sharply on enclosure rating, sample-conditioning budget, and detector choice; a wrong call on any of those gates costs more than the instrument itself once installation, calibration gas, and shelter work are added.
Analyte list and detection limit as the first gate
Component-level detection limits in the parts-per-billion to low-ppm range are achievable with bench-top GC-FID and GC-µECD configurations; FID is the default for hydrocarbon speciation, TCD for permanent gases, and ECD or mass-selective detection where halogenated or sub-ppb species are on the scope [S1].
The Welthagen et al. method published on ScienceDirect used direct thermal desorption coupled to GC×GC–time-of-flight MS to resolve hundreds of organic compounds in airborne particulate, demonstrating that resolution and not just sensitivity is the gating parameter when the analyte list exceeds roughly 50 named species [S1].
Yabumoto, Jennings, and Yamaguchi showed on wall-coated open-tubular glass capillary columns that Kováts retention indices measured on two columns of different polarity are realistic identification criteria, which is why dual-column confirmation remains a standard lab practice for unknown-peak work [S3].
Column and carrier-gas decision
Wall-coated open-tubular (WCOT) capillary columns, packed columns, and micropacked transitions cover the bulk of process and lab work; selection is driven by inner diameter, film thickness, and the polarity mismatch with the target analyte list [S3].
Helium remains the default carrier for lab work; hydrogen is increasingly specified for high-efficiency short columns on the basis of faster analysis time, and nitrogen is reserved for TCD work on permanent gases where sensitivity is acceptable for the application.
Erika Cremer and Fritz Prior built the first documented gas chromatographic system in Innsbruck during 1945–47, a fact now exhibited at the Deutsches Museum Bonn branch, and the principle they demonstrated — separation on a packed column with a thermal-conductivity detector — still anchors the modern packed-column method used in many process GC installations [S4].
Detector stack and dynamic range
Three detector families cover the bulk of analyzer specifications: flame ionization detection (FID) for hydrocarbons, thermal conductivity (TCD) for permanent gases, and electron capture / photoionization for halogenated or low-ppb work; a dual-detector setup is common when the analyte list spans organics and permanent gases on one stream. [S1]
FID linearity typically exceeds six orders of magnitude, which is why hydrocarbon speciation from ppm to percent level is straightforward on a single FID channel; TCD linearity is narrower, and method development must guard against over-ranging on high-concentration permanent gases.
Where the lab method needs to resolve hundreds of organic species in a single injection, comprehensive two-dimensional GC coupled to time-of-flight MS is the working technique in published methods for airborne particulate analysis, with the trade-off being capital cost, data-system complexity, and analyst time [S1].
Process versus lab enclosure and hazardous-area rating
Process GC analyzers mounted near the sample tap must meet the area classification of the plant — ATEX category for Zone 1 sites in the EU, IECEx for international projects, and Class/Division marking under NEC for North America — and the shelter or analyzer-house HVAC budget must hold the enclosure inside its certified ambient range. [S2]
Lab instruments sit in a controlled environment and typically carry only a general-purpose electrical rating; pushing a lab GC into a plant without recertifying the enclosure, purge, and cable glands is one of the most common audit findings in refinery and petrochemical deployments.
Sample-conditioning is the unglamorous gating step: the analyzer cannot outperform its sample line, so a process GC selection must be paired with a filter, coalescer, and temperature-regulated heat-traced sample train sized to the dew point and particulate load of the actual stream.
Comparing the main options against decision criteria
For a typical analyzer selection in a refinery or natural-gas plant, the three options most often shortlisted are bench-top GC-FID for lab speciation, process GC with dual FID/TCD for online hydrocarbon and permanent-gas analysis, and portable micro-GC for field screening or pipeline-quality spot checks. [S3]
Bench-top GC-FID wins on detection limit and resolution for C1–C6 speciation and detailed retention-index work, but loses on enclosure rating, sample-conditioning, and unattended duty cycle; a process GC reverses that trade, giving 24/7 unattended operation inside a Zone 1 enclosure at the cost of detection limit and method-development depth [S3].
Portable micro-GC scores on footprint, carrier-gas economy, and rapid site deployment, but is constrained on column thermal program length and on detector count; if the application is continuous custody transfer or online control, the portable form factor is a screening tool rather than a primary measurement, while for sub-ppb environmental work a bench-top GC-µECD or GC-MS remains the working choice [S1].
Cross-reference against flow and level instrumentation is useful when the analyzer sits on a custody-transfer skid: a comparison of variable area flowmeter vs Coriolis selection in 2026 covers the same duty-cycle versus accuracy trade-off that applies to the GC sample stream itself.
Use cases and where each option fits
Natural-gas and LNG plants typically specify a process GC with dual FID/TCD channels for mole-percent C1 through C6+ speciation plus H2S and CO2 where the contract calls for it, with sample conditioning sized to a glycol- or amine-free dry sales-gas stream. [S4]
Refinery applications — reformer off-gas, ethylene cracker feed, catalytic cracker overhead — push the specification toward a multi-valve multi-column process GC with methanator/FID for trace CO and CO2, and the analyzer shelter must be rated for the surrounding hazardous area while keeping the sample train above its hydrocarbon dew point.
Environmental and atmospheric research laboratories work with bench-top GC coupled to FID, µECD, or mass-selective detection; published methods for airborne particulate show that two-dimensional separation is needed when the target list exceeds the resolving power of a single capillary column [S1].
For shop-floor or pilot-plant work where a turbidity-style single-parameter analyzer is being considered against a full GC, the turbidity meter vs gas chromatograph decision guide is a useful framing — turbidity is a bulk property, GC is speciation, and the wrong call is a recurring maintenance cost.
Limitations, failure modes, and standards
Common failure modes in process GC service are carrier-gas contamination, sample-line cold spots dropping below the hydrocarbon dew point, and detector flame-out on loss of fuel-gas or air supply; each one is mitigated by hardware interlocks and by a documented startup checklist on the analyzer itself. [S5]
Lab instruments face different failure modes — column bleed raising the baseline, septum leaks degrading retention-time reproducibility, and syringe-needle discrimination on high-boiling compounds — and the standard countermeasure is a scheduled consumables-replacement interval tied to the actual run count, not the calendar.
For hazardous-area installations the applicable regimes are ATEX 2014/34/EU for the European Union, IECEx for international projects, and the NEC Class/Division system in North America; the standard for analyzer shelters and purge is the IEC 60079-x family, and the analyzer selected must carry a certificate that covers the actual zone and gas group of the installed location.
Selection checklist before quotation
Before the request-for-quotation is released, the analyzer datasheet should carry: complete analyte list with expected concentration range, required detection limit, matrix composition including water and particulates, sample pressure and temperature at the tap, hazardous-area classification, ambient temperature range at the analyzer location, available carrier gas and detector gases, required cycle time, and the output protocol expected by the control system. [S1]
Any vendor quotation that does not address detection limit per compound, sample-conditioning scope, and enclosure certification as line items is a procurement risk; a process analyzer selection should also name the sample-conditioning vendor and the calibration-gas supplier at the same time, because lead times on those items often exceed the analyzer delivery time.
Documented chromatographic identification by retention index on a single column is acceptable for routine-quality work, while unknown-peak confirmation routinely uses a second column of different polarity under the same oven program; the published basis for that practice goes back to the wall-coated open-tubular work on Kováts indices in the 1970s [S3].
For analyzer shelters where the GC must coexist with temperature transmitter wiring and diagnostics, the same EMC and cable-segregation rules apply, and the shelter layout should keep the analyzer signal cables physically separated from power and from variable-frequency-drive runs.
Trackable signals to watch over the next procurement cycle: detector-gas price and supply security, because helium supply remains a recurring budget item on lab installations, and analyzer-shelter certification, because several vendors have shifted ATEX/IECEx certification scope on their process GC line in the past year; both items should be re-verified at quotation release.
For component-level specifications, see gas analyzer, and gas detector.