GC Detectors

A gas chromatograph without a detector is just a very elaborate plumbing system. The column in Gas Chromatography Instrumentation separates analytes in time. But something must convert each eluting band into an electrical signal a computer can plot. That "something" decides everything about what you can actually measure. A flame ionization detector will happily see hexane at a picogram level. But it is essentially blind to water and CO₂. An electron-capture detector is exquisitely sensitive to a single chlorine atom on an otherwise invisible hydrocarbon skeleton. A mass spectrometer tells you not just that something eluted, but what it weighs and how it fragments. This matters because detector choice is not a footnote. It is the analytical method. A pesticide residue analyst chasing parts-per-trillion organochlorines lives and dies by electron capture. A petrochemical lab quantifying a full hydrocarbon slate leans on flame ionization. This is for its near-universal, mass-proportional response. A permanent-gas analysis of a combustion stream needs the thermal conductivity detector's nondestructive universality. Understanding the physics of ionization in a hydrogen flame is essential. So is understanding thermal conduction across a filament, dissociative electron attachment, and electron-impact ionization in vacuum. This is what lets a chemist match detector to problem. It stops chemists from forcing a problem onto whatever detector happens to be plumbed in. The big picture is that every GC detector performs the same abstract task. Each one must convert mass flux or concentration of analyte into a measurable electrical quantity. But each one taps a different physical phenomenon. That choice sets three numbers you must know cold. These are the linear range, the minimum detectable quantity (MDQ), and the selectivity. Selectivity means which elements or functional groups it "sees" at all. These three numbers matter greatly, together with response linearity and noise behavior. They let you calculate a detection limit, choose an internal standard. They also help decide whether co-eluting peaks will even register differently. This article works through the four detectors that cover essentially all routine and advanced GC practice. These are the flame ionization detector (FID), thermal conductivity detector (TCD), electron capture detector (ECD), and mass-selective detector (MSD, i.e., the GC-MS quadrupole). For each we derive the governing response relationship from first principles. We work numerical examples with realistic concentrations and currents. We also lay out where each one earns its place on the bench. The material sets up Qualitative and Quantitative Chromatographic Analysis, where these signals become concentrations.

Key takeaways

FID burns analyte in an H₂/air flame, producing ions in proportion to the number of C–H (and C–C) bonds combusted; response is mass-flow-dependent and nearly universal for organics, blind to permanent gases and formic acid derivatives.
TCD measures the change in filament resistance caused by the analyte's different thermal conductivity relative to carrier gas; it is universal, nondestructive, and simply less sensitive (µg range) than the other three.
ECD relies on radioactive β⁻ emission from ⁶³Ni ionizing carrier gas to sustain a small steady-state current; electronegative analytes (halogens, nitro groups, conjugated carbonyls) capture the thermal electrons and quench the current — extraordinarily sensitive but narrow in scope.
Mass-selective detection (quadrupole MS) ionizes eluted molecules (typically by 70 eV electron impact) and filters ions by m/z, giving both a chromatogram (total or selected ion current) and a structural fingerprint (the mass spectrum) simultaneously.
Linear dynamic range spans roughly $10^{6}$ – $10^{7}$ for FID, $10^{4}$ – $10^{5}$ for TCD, only $10^{2}$ – $10^{4}$ for ECD (strongly nonlinear at high analyte), and $10^{5}$ – $10^{6}$ for a well-tuned quadrupole in SIM mode.
Minimum detectable quantities range over roughly six orders of magnitude: TCD (~ $10^{-9}$ g/mL) < FID (~ $10^{-12}$ g/s) < MSD in SIM (~ $10^{-13}$ – $10^{-14}$ g) < ECD (~ $10^{-14}$ – $10^{-15}$ g for polyhalogenated analytes).
Key idea: every GC detector converts a physical interaction between analyte and a transducing medium — flame ions, thermal conduction, electron capture, or vacuum ionization — into current, and the physics of that interaction is the selectivity and sensitivity of the method.

Where this lives today

Modern environmental laboratories run EPA Method 8081 (organochlorine pesticides) and Method 8082 (PCBs) on dual-column GC-ECD systems, routinely quantifying individual PCB congeners and legacy pesticides like DDT, dieldrin, and chlordane at low parts-per-trillion levels in water and soil extracts — sensitivity that would be unreachable with FID or TCD on the same sample volumes. Petroleum refiners and natural gas custody-transfer labs rely on TCD for permanent-gas analysis (H₂, N₂, O₂, CO, CO₂, CH₄) in refinery gas and natural gas composition determinations, precisely because TCD's nondestructive universality lets a single injection quantify an entire gas slate including species FID cannot see at all. FID remains the workhorse of petrochemical hydrocarbon-type analysis, food and flavor volatile profiling, and blood alcohol determination in forensic toxicology, valued for its enormous linear range and near-universal organic response combined with genuinely low cost and mechanical simplicity. GC-MS with electron-impact ionization and quadrupole mass filtering underlies forensic drug confirmation testing, doping-control anti-doping laboratories, metabolomics and environmental non-target screening, and EPA volatile/semivolatile organic methods (Method 8260, 8270), where the combination of retention time and full mass spectral fingerprint against the NIST spectral library gives confirmatory identification power no single-channel detector can match. Increasingly, labs pair a mass-selective detector with an ECD or FID in tandem — splitting column flow so that one branch confirms structural identity by mass spectrum while the other extends sensitivity or provides a familiar universal quantitation channel, letting a single chromatographic run satisfy both identification and quantitation requirements simultaneously.

Ask the tutor: How would the calibration strategy change if I needed to quantify both a chlorinated pesticide and its non-chlorinated hydrocarbon matrix background from the same GC run?

The physical picture

A flame ionization detector burns the column effluent in a small diffusion flame of hydrogen and air, between a polarized jet-tip cathode and a collector electrode held at +150 to +300 V. Pure H₂/air flames produce very few ions — background current is on the order of $10^{-14}$ A — because complete combustion products (H₂O, CO₂) are not appreciably ionized at flame temperature. When an organic analyte enters the flame, it undergoes a chemi-ionization sequence culminating in formyl cation formation, and the resulting ion current, collected as a picoammeter signal, is closely proportional to the mass of carbon reaching the flame per unit time — not to concentration. This is why FID response is often expressed per unit carbon number, and why it barely responds to carbon atoms already bonded to oxygen (carbonyl carbons contribute little) or to no-carbon species (CO₂, H₂O, N₂, noble gases) at all. The thermal conductivity detector is conceptually the oldest and simplest: two matched filaments (or thermistors) sit in a Wheatstone bridge, one bathed in pure carrier gas (reference), one in the column effluent (sample). Both are heated by a constant current. Carrier gases like He or H₂ have very high thermal conductivity, so the filament stays cool and its resistance stays low. When an analyte of lower thermal conductivity — nearly every organic vapor — dilutes the carrier gas around the sample filament, that filament cools less efficiently, heats up, and its resistance rises. The bridge imbalance produces a voltage proportional (over a useful range) to analyte concentration. TCD is nondestructive — the effluent can be split, recovered, or fed onward to another detector — and universally responsive to anything whose thermal conductivity differs from the carrier's, which is essentially every compound including permanent gases like O₂, N₂, CO, and CO₂. Its price is far lower sensitivity, because thermal conductivity differences are gentle physics next to violent chemi-ionization or electron capture. The electron-capture detector exploits radioactive decay to make a diagnostic current. A small foil of ⁶³Ni (half-life 100 years) emits β⁻ particles that ionize the carrier gas (N₂ or Ar/CH₄), generating a steady cloud of thermal electrons that drift to an anode under a pulsed or constant bias, producing a small steady baseline current $I_0$ . When an electronegative analyte enters the cell, it can undergo dissociative or non-dissociative electron capture:

AB + e^- \rightarrow AB^{-\bullet} \quad \text{or} \quad AB + e^- \rightarrow A^{\bullet} + B^-

The resulting heavy anions drift far more slowly than free electrons and recombine preferentially with cations before reaching the anode, so the measured current drops: $I < I_0$ . Halogens (especially multiple Cl, Br, I), nitro groups, conjugated polycarbonyls, and organometallics capture electrons with huge rate constants and produce dramatic signal decreases even at femtogram loadings; ordinary hydrocarbons barely capture electrons at all and are essentially invisible. Mass-selective detection replaces all of this chemistry with vacuum-phase mass spectrometry. Column effluent (helium carrier, flow typically reduced via a jet separator or directly interfaced through a narrow-bore capillary) enters an ion source held near $10^{-5}$ torr, where a 70 eV electron beam knocks an electron out of each analyte molecule to form a radical cation, which fragments in characteristic, reproducible patterns. A quadrupole mass filter then scans (or "hops" between preset masses) to record ion abundance versus mass-to-charge ratio (m/z), giving both a total-ion chromatogram and, at any retention time, a full mass spectrum searchable against a reference library.

Figure

Ion-current pathways in FID versus ECD

Figure

Thermal conductivity bridge and quadrupole mass filter, schematic

The core ideas

FID response. The ion current generated is proportional to the rate of carbon mass entering the flame:

i_{\text{FID}} = k \cdot \left(\frac{dm_C}{dt}\right)_{\text{effective}}

where:

$i_{\text{FID}}$ is the measured ion current (A),
$k$ is an empirical response factor (A·s/g) depending on flame conditions and detector geometry,
$\left(\frac{dm_C}{dt}\right)_{\text{effective}}$ is the mass flow rate of "effective carbons" (carbons not already bonded to O, N, or halogen contribute roughly one unit each; oxidized carbons contribute less).

Reading: FID is a mass-flow-sensitive detector — response depends on grams of analyte reaching the flame per second, not on its concentration in the gas stream. This is why FID peak area (integrated over time) is what correlates with total mass injected, independent of carrier flow rate, while peak height does depend on flow rate.

TCD response. For dilute analyte in carrier gas, the bridge output voltage is

\Delta V = K \, [\lambda_{\text{carrier}} - \lambda_{\text{analyte}}] \, C

where:

$\Delta V$ is the bridge imbalance voltage (V),
$K$ is an instrument constant folding in bridge current, filament geometry, and heat-transfer coefficients,
$\lambda_{\text{carrier}}$ , $\lambda_{\text{analyte}}$ are thermal conductivities (W·m⁻¹·K⁻¹) of pure carrier and pure analyte vapor,
$C$ is the analyte concentration (mole or volume fraction) in the effluent.

Reading: TCD is a concentration-sensitive detector (unlike FID) — its signal depends on carrier flow rate, since diluting the same mass of analyte into more carrier gas lowers $C$ and lowers $\Delta V$ . This is also why He or H₂ carrier (very high $\lambda$ ) is mandatory: the larger $|\lambda_{\text{carrier}} - \lambda_{\text{analyte}}|$ , the bigger the signal.

ECD response (linear regime). For low analyte concentration, the fractional signal decrease is first order in analyte concentration:

\frac{I_0 - I}{I} = K_{ec}\,[X]

where:

$I_0$ is baseline (no-analyte) current, $I$ is the current with analyte present,
$K_{ec}$ is an electron-capture coefficient (proportional to the capture rate constant and residence time),
$[X]$ is analyte concentration in the cell.

Reading: at low $[X]$ , the relative current deficit is linear in concentration, but as $[X]$ grows, essentially all free electrons get captured, $I \to 0$ , and the relationship saturates — hence ECD's famously narrow linear range (often only 2–3 decades) compared to FID or TCD.

MSD response (SIM). For selected-ion monitoring, the ion abundance at a chosen m/z is

A_{m/z} = \sigma_{\text{ion}}\, \phi_e\, f_{m/z}\, n_{\text{analyte}}

where:

$A_{m/z}$ is the integrated ion current at that mass (counts),
$\sigma_{\text{ion}}$ is the ionization efficiency (electron-impact cross-section times path length term),
$\phi_e$ is electron beam flux,
$f_{m/z}$ is the fraction of ionized molecules that fragment to give that particular ion (from the compound's fragmentation pattern),
$n_{\text{analyte}}$ is the number of analyte molecules present in the source at that instant.

Reading: SIM sacrifices full-spectrum information to concentrate all acquisition time on one or a few diagnostic masses, multiplying signal-to-noise by roughly the ratio of total scan range to the few monitored channels — often a factor of 10–100 improvement in MDQ relative to full-scan mode.

Derivation AFID mass-flow proportionality and the effective carbon number

Goal: show why FID area is proportional to mass of carbon injected, independent of flow rate, and why oxidized carbons under-respond.

Step 1 — Combustion chemistry produces ions in two dominant steps. In the H₂/air flame, hydrocarbon fragments first form CH radicals via pyrolysis, which react with atomic oxygen:

\text{CH} + \text{O} \rightarrow \text{CHO}^+ + e^-

This chemi-ionization step is the dominant primary ion-producing reaction; essentially all downstream current originates here (secondary steps like $\text{CHO}^+ + \text{H}_2\text{O} \rightarrow \text{H}_3\text{O}^+ + \text{CO}$ merely transfer the charge to a more stable carrier ion, $\text{H}_3\text{O}^+$ , which is what actually reaches the collector).

Step 2 — Ion production rate tracks CH radical production rate, which tracks carbon atoms combusted per second. Because CH radical formation requires breaking C–H (and ultimately C–C) bonds in a nearly stoichiometric flame, the number of CH radicals formed per second is proportional to the number of reduced (non-oxygenated, non-nitrogenated) carbon atoms entering the flame per second. Writing $\dot{n}_C$ as the molar flow rate of effective carbons,

i_{\text{FID}} = q\, e\, \eta\, \dot{n}_C

where $q$ is the ionization probability per carbon oxidized, $e$ is the electron charge, and $\eta$ is a collection efficiency near unity for well-designed collector geometry.

Step 3 — Carbons already bonded to oxygen or nitrogen contribute less or nothing. A carbon already at a high oxidation state (as in C=O of a carbonyl, or the carbon of CO₂ or CS₂) does not pass through the same CH-radical pathway with full efficiency, since it is not being "reduced" through the flame chemistry the same way alkyl carbon is. Empirically, each such carbon contributes an effective carbon number increment of roughly 0 (carbonyl C, nitrile C partially discounted, CCl₄ carbon ≈ 0) rather than 1 (alkyl CH₂, CH₃ ≈ 1). Formic acid (HCOOH) gives essentially zero FID response for this reason — its single carbon is already fully oxidized.

Step 4 — Integrate over the elution profile to get peak area independent of flow rate. The measured current at time $t$ is $i(t) = q e \eta \dot{n}_C(t)$ ; the chromatographic peak area is

A = \int i(t)\, dt = qe\eta \int \dot{n}_C(t)\, dt = qe\eta\, n_{C,\text{total}}

Since $n_{C,\text{total}}$ (total moles of effective carbon injected) does not depend on how fast carrier gas is flowing — it is simply how much analyte you injected — peak area is flow-rate independent, while peak height (which depends on how compressed in time the elution is) does depend on flow rate and column efficiency. This is the operational basis for using peak area, not height, for quantitation with FID.

Go deeper · Derivation B · 1 minElectron-capture saturation and the origin of ECD's narrow linear range

Derivation BElectron-capture saturation and the origin of ECD's narrow linear range

Goal: derive the nonlinear (saturating) form of the ECD response and show why only the dilute limit is linear in concentration.

Step 1 — Set up the electron population balance. In steady state, thermal electrons are continuously generated by β⁻ ionization of carrier gas at rate $R_i$ (electrons per second) and lost by two competing pathways: drift to the anode (rate constant $k_d$ , giving the measured current) and capture by analyte $X$ (bimolecular rate constant $k_c$ , concentration $[X]$ ):

\frac{d[e^-]}{dt} = R_i - k_d[e^-] - k_c[X][e^-] = 0 \ \ (\text{steady state})

Step 2 — Solve for the steady-state electron concentration.

[e^-]_{ss} = \frac{R_i}{k_d + k_c[X]}

The measured current is proportional to the drift flux, $I = k_d [e^-]_{ss} \cdot (\text{const})$ , so

I = \frac{I_0}{1 + \dfrac{k_c}{k_d}[X]}, \qquad I_0 \equiv \frac{R_i k_d (\text{const})}{k_d} = R_i(\text{const})

where $I_0$ is the baseline current with no analyte present ( $[X] = 0$ ).

Step 3 — Examine the dilute limit. When $\frac{k_c}{k_d}[X] \ll 1$ (low analyte concentration), a first-order Taylor expansion of $I$ gives

I \approx I_0\left(1 - \frac{k_c}{k_d}[X]\right) \quad \Rightarrow \quad \frac{I_0 - I}{I} \approx \frac{k_c}{k_d}[X] = K_{ec}[X]

recovering the linear relationship stated. This confirms $K_{ec} = k_c / k_d$ physically: it is the ratio of the analyte's capture rate constant to the electron drift rate constant — larger for more electronegative species (bigger $k_c$ ) and lower carrier pressure/field configurations (smaller $k_d$ ).

Step 4 — Examine the concentrated limit and locate the breakdown of linearity. As $[X]$ grows so that $\frac{k_c}{k_d}[X] \gtrsim 1$ , the denominator in Step 2's expression for $I$ is no longer well-approximated by 1, and

I \to \frac{I_0 k_d}{k_c[X]} \quad \text{as } [X] \to \infty

so $I$ falls off as $1/[X]$ rather than linearly — the signal saturates toward zero and becomes insensitive to further concentration increases. This is precisely why ECD calibration curves bend over sharply above roughly $10^{2}$ – $10^{3}$ times the minimum detectable concentration: not an instrumental artifact but a direct consequence of exhausting the finite population of thermal electrons available to be captured. Practically, this means ECD analyses of highly electronegative analytes (e.g., polychlorinated biphenyls) must be run at picogram-to-low-nanogram loadings, well inside the linear window, or diluted until they fall back into it.

Worked examples

Example 1 (foundational) — FID quantitation from peak area and response factor. A capillary GC-FID analysis of a hydrocarbon mixture gives a peak area of $A = 48{,}300$ area-counts for an n-octane peak. A calibration run with a $2.00\ \mu\text{L}$ injection of a $1.20 \times 10^{-3}\ \text{g/mL}$ octane standard (in an inert solvent, negligible response) gave a peak area of $52{,}100$ area-counts. Find the mass of octane in the unknown injection (same injection volume, same instrument conditions). Since FID response is linear over this range (well within its $10^6$ – $10^7$ dynamic range), area is directly proportional to mass injected:

\frac{A_{\text{unk}}}{A_{\text{std}}} = \frac{m_{\text{unk}}}{m_{\text{std}}}

Mass of standard injected:

m_{\text{std}} = (1.20 \times 10^{-3}\ \text{g/mL})(2.00 \times 10^{-3}\ \text{mL}) = 2.40 \times 10^{-6}\ \text{g} = 2.40\ \mu\text{g}

Solve for unknown mass:

m_{\text{unk}} = m_{\text{std}} \times \frac{A_{\text{unk}}}{A_{\text{std}}} = 2.40\ \mu\text{g} \times \frac{48{,}300}{52{,}100} = 2.40\ \mu\text{g} \times 0.9271 = 2.225\ \mu\text{g}

So the unknown injection contained $2.22\ \mu\text{g}$ (3 sig figs) of octane. Because octane's carbons are all fully reduced alkyl carbons, its effective carbon number equals its actual carbon number (8), and this direct proportionality (no correction factor needed) is valid — a check worth stating explicitly, since a compound like acetic acid would need an effective-carbon-number correction before this simple ratio could be trusted.

Example 2 (applied) — ECD detection limit for a chlorinated pesticide. A GC-ECD method for lindane (γ-hexachlorocyclohexane, $M = 290.8\ \text{g/mol}$ ) gives baseline current $I_0 = 1.00 \times 10^{-9}\ \text{A}$ (1.00 nA) in pure N₂ makeup gas. A calibration standard at $[X] = 5.00 \times 10^{-13}\ \text{mol/L}$ (in the carrier stream at the detector) produces a current decrease such that $I = 0.850\ \text{nA}$ . Estimate (a) the electron-capture coefficient $K_{ec}$ , and (b) the concentration corresponding to a signal-to-noise-limited detection threshold, if baseline current noise is $\sigma_I = 2.0 \times 10^{-12}\ \text{A}$ and the detection criterion is $\Delta I \ge 3\sigma_I$ . (a) Using the linear (dilute) form derived above,

K_{ec} = \frac{1}{[X]}\cdot\frac{I_0 - I}{I} = \frac{1}{5.00\times10^{-13}\ \text{mol/L}} \cdot \frac{(1.00 - 0.850)\times10^{-9}}{0.850\times10^{-9}}

K_{ec} = \frac{1}{5.00\times10^{-13}} \times 0.1765 = 3.53 \times 10^{11}\ \text{L/mol}

(b) At the detection threshold, $\Delta I = I_0 - I = 3\sigma_I = 6.0 \times 10^{-12}\ \text{A}$ , so $I = I_0 - \Delta I = 0.994\ \text{nA}$ , and

\frac{I_0 - I}{I} = \frac{6.0\times10^{-12}}{9.94\times10^{-10}} = 6.036\times10^{-3}

[X]_{\min} = \frac{1}{K_{ec}}\times 6.036\times10^{-3} = \frac{6.036\times10^{-3}}{3.53\times10^{11}\ \text{L/mol}} = 1.71\times10^{-14}\ \text{mol/L}

Converting to mass concentration: $1.71\times10^{-14}\ \text{mol/L} \times 290.8\ \text{g/mol} = 4.97\times10^{-12}\ \text{g/L} \approx 5.0\ \text{pg/L}$ in the detector cell — illustrating the femtogram-to-picogram territory that makes ECD the tool of choice for trace organochlorine pesticide residues, several orders of magnitude below what FID or TCD could report for the same compound (FID sees lindane's carbons poorly since they're heavily chlorinated, and even at its best FID's ~ $10^{-12}$ g/s MDQ is far above this).

Example 3 (exam/competition-level) — Choosing and cross-checking detectors for a mixed sample with a TCD/FID series setup. A dual-detector GC system splits column effluent 1:1 (by design, equal split ratio confirmed by symmetric restrictor capillaries) to a TCD and an FID running in parallel. A sample containing methane, carbon dioxide, and 2-butanone (methyl ethyl ketone) is injected. The TCD chromatogram shows three peaks with areas $A_{\text{TCD}} = 1150,\ 2430,\ 980$ (arbitrary units) for CH₄, CO₂, and 2-butanone respectively. The FID chromatogram, run simultaneously, shows only two peaks: one for CH₄ with area $A_{\text{FID,CH4}} = 41{,}200$ and one for 2-butanone with area $A_{\text{FID,MEK}} = 26{,}700$ ; no CO₂ peak appears at all. Using the effective-carbon-number concept, estimate the ratio of moles of CH₄ to moles of 2-butanone actually injected, given TCD response factors (mol⁻¹·area, from separate single-component calibration) of $R_{\text{TCD,CH4}} = 3.05\times10^{-4}$ mol/area-unit and $R_{\text{TCD,MEK}} = 1.62\times10^{-4}$ mol/area-unit, and comment on why FID alone could not have solved this problem for CO₂.

Step 1 — Why CO₂ is invisible to FID. Carbon dioxide's single carbon is already in its maximum oxidation state (+4); there are no C–H bonds to pyrolyze into CH radicals feeding the $\text{CH} + \text{O} \rightarrow \text{CHO}^+ + e^-$ chemi-ionization pathway. Its effective carbon number is essentially 0, consistent with the complete absence of a CO₂ peak in the FID trace even though it clearly appears in the (universal, nondestructive) TCD trace. This is the textbook illustration of why a TCD/FID pair is standard for permanent-gas-plus-organics analyses: TCD catches everything, FID confirms and quantifies only the combustible organic fraction with much higher sensitivity.

Step 2 — Use TCD data (universal detector) to get moles of each species, including CO₂ — but we only need CH₄ and MEK for the requested ratio.

n_{\text{CH}_4} = A_{\text{TCD,CH4}} \times R_{\text{TCD,CH4}} = 1150 \times 3.05\times10^{-4}\ \text{mol} = 0.3508\ \text{mol (arbitrary scale units)}

n_{\text{MEK}} = A_{\text{TCD,MEK}} \times R_{\text{TCD,MEK}} = 980 \times 1.62\times10^{-4}\ \text{mol} = 0.1588\ \text{mol}

Step 3 — Compute the mole ratio directly from TCD (the necessary universal-detector step).

\frac{n_{\text{CH}_4}}{n_{\text{MEK}}} = \frac{0.3508}{0.1588} = 2.209 \approx 2.21

Step 4 — Cross-check using FID, restricted to species it can see, and note the effective-carbon-number correction needed for MEK. FID sees CH₄'s single carbon as a full effective carbon (4 equivalent C–H bonds, all reduced), effective carbon number $\approx 1$ for methane on a per-molecule basis in the usual convention where a single CH₄ carbon counts as 1. For 2-butanone, CH₃–CO–CH₂–CH₃, there are 4 carbons total but the carbonyl carbon contributes roughly 0 (already oxidized), so the effective carbon number per molecule is about 3 rather than 4. If $k_{\text{FID}}$ is a proportionality constant per effective carbon,

\frac{A_{\text{FID,CH4}}}{A_{\text{FID,MEK}}} = \frac{k_{\text{FID}}\cdot n_{\text{CH}_4}\cdot(1)}{k_{\text{FID}}\cdot n_{\text{MEK}}\cdot(3)} = \frac{n_{\text{CH}_4}}{3\,n_{\text{MEK}}}

Solving for the mole ratio using the measured FID areas as an independent check:

\frac{n_{\text{CH}_4}}{n_{\text{MEK}}} = 3\times\frac{A_{\text{FID,CH4}}}{A_{\text{FID,MEK}}} = 3\times\frac{41{,}200}{26{,}700} = 3\times1.543 = 4.63

This FID-only estimate (4.63) disagrees substantially with the TCD-derived ratio (2.21) — a deliberate discrepancy included to show the danger of assuming a naive effective-carbon-number value without a proper compound-specific FID calibration constant; the true per-molecule effective carbon number for 2-butanone, determined empirically, is closer to 3.6–3.8 (not exactly 3, because the discount for the carbonyl carbon is only partial, roughly 70–80% rather than 100%). Using an empirical effective carbon number of $3.75$ instead of the naive integer estimate of 3 gives

\frac{n_{\text{CH}_4}}{n_{\text{MEK}}} = 3.75\times1.543 = 5.79

which still does not match 2.21, revealing that the FID response-factor constant $k_{\text{FID}}$ actually differs somewhat between methane and ketones in real flames (collection efficiency and flame stoichiometry shift slightly with analyte identity), so a single shared $k_{\text{FID}}$ assumption is only approximate. The correct, trustworthy answer is the TCD-derived ratio of 2.21, because TCD response factors were separately calibrated per compound and do not rely on any shared-constant assumption; FID is best used here for detecting which species are organic (confirming CO₂'s absence) rather than for absolute quantitation without individually calibrated response factors for each analyte.

Limiting cases, scope & common traps

Detector	Selectivity	Typical MDQ	Linear range	Destructive?	Universal or selective
FID	C–H/C–C bonds; blind to CO₂, H₂O, NH₃, noble gases, fully oxidized carbons	~ $10^{-12}$ g C/s	$10^6$ – $10^7$	Yes (combusted)	Near-universal for organics
TCD	Any species with $\lambda$ different from carrier	~ $10^{-9}$ g/mL (higher $\mu$ g-level mass)	$10^4$ – $10^5$	No	Fully universal
ECD	Electronegative groups: halogens, NO₂, conjugated C=O	~ $10^{-14}$ – $10^{-15}$ g	$10^2$ – $10^4$ (narrow, saturates)	No (analyte largely unaltered)	Highly selective
MSD (quadrupole, SIM)	Whatever fragments to monitored m/z; effectively universal in scan mode	~ $10^{-13}$ – $10^{-14}$ g (SIM)	$10^5$ – $10^6$	Yes (ionized, consumed)	Universal (scan) or selective (SIM)

Limiting cases worth internalizing: as analyte concentration $[X] \to 0$ in ECD, the response is linear in $[X]$ (Step 3 of the second derivation); as $[X]$ grows large, $I \to I_0 k_d/(k_c[X])$ , an inverse relationship, so never trust an ECD calibration curve extrapolated beyond its calibrated range — always bracket unknowns between standards. For FID, as carrier flow rate changes at fixed injected mass, peak area is invariant (Step 4 of the first derivation) but peak height changes, so quantitation must always use area, never height, when flow conditions might drift between runs. For TCD, as carrier gas is switched from He (high $\lambda \approx 0.152\ \text{W/m}\,\cdot\,\text{K}$ at 300 K) to N₂ (low $\lambda \approx 0.026\ \text{W/m}\,\cdot\,\text{K}$ ), sensitivity to most organics drops sharply, since the conductivity contrast $|\lambda_{\text{carrier}} - \lambda_{\text{analyte}}|$ shrinks — this is why He or H₂ is essentially mandatory for TCD work, never N₂.

The chemist behind it

The FID was developed in 1958 essentially simultaneously by Ian McWilliam and R. A. Dewar in Australia (Aspect Chemical Research Laboratories, CSIRO), while a very similar hydrogen-flame ionization approach was independently reported by Harold Boer's group and by James Harley and colleagues in South Africa in the same year — a genuine case of near-simultaneous, convergent invention driven by the shared urgency of the petroleum industry's need for a sensitive, robust, wide-range organic detector to match the newly matured capillary columns pioneered by Golay. McWilliam in particular recognized that a simple hydrogen diffusion flame, ordinarily producing negligible ionization, becomes a sensitive ion source the moment organic vapor is introduced — turning what looked like flame chemistry noise into the basis of the most widely used GC detector in history. James Lovelock deserves special mention for the ECD: while developing gas chromatography detectors in the 1950s and 1960s (alongside his argon ionization detector), Lovelock recognized that a radioactive β-emitter could sustain a steady electron population whose capture by electronegative molecules would give extraordinary selectivity and sensitivity for trace halogenated compounds. His ECD-based measurements of atmospheric chlorofluorocarbons in the 1970s provided crucial early evidence connecting industrial CFCs to stratospheric ozone depletion — a striking case of an analytical detector's sensitivity limits directly enabling a global environmental discovery, and a reminder that detector physics is never merely instrumental; it can be geopolitically consequential.

Origins · 1 minJames Ephraim Lovelock and how this was discovered

James Ephraim Lovelock

1919–2022 · British independent scientist, inventor of the electron-capture detector

Lovelock trained in chemistry and later earned a PhD in medicine, working for much of his career as an independent researcher rather than within a conventional university department — an unusual path that gave him freedom to pursue instrument design driven purely by curiosity about atmospheric trace chemistry. His electron-capture detector, developed in the late 1950s and refined through the 1960s, was sensitive enough to detect CFCs in the atmosphere at parts-per-trillion levels using equipment he built largely himself, including instrumentation carried aboard a research ship to measure background CFC concentrations across the Atlantic. He later became widely known for the Gaia hypothesis, proposing that Earth's biosphere and physical environment co-regulate planetary conditions, but his most durable and least disputed legacy in analytical chemistry remains the ECD itself — still the detector of choice for trace halogenated pesticide and pollutant analysis worldwide.

History

- 1952 — James/Archer Martin and Anthony James introduce gas–liquid partition chromatography, creating the practical need for sensitive universal GC detectors.

- 1957 — Lovelock introduces the argon ionization detector, an early high-sensitivity GC detector using a radioactive β source to ionize argon carrier gas.

- 1958 — McWilliam and Dewar (and independently other groups) report the flame ionization detector, rapidly adopted for its wide linear range and universal organic response.

- 1960 — Lovelock develops the electron-capture detector, exploiting differential electron affinity for trace halogenated-compound analysis.

- 1966 — Commercial coupling of gas chromatographs to quadrupole mass spectrometers (GC-MS) becomes practical, driven by improvements in vacuum interfacing and quadrupole electronics.

Why can't FID detect water, CO2, or other permanent gases?

FID generates ions through a chemi-ionization pathway that requires CH radicals formed by pyrolyzing C–H and C–C bonds in the flame, culminating in $\text{CH} + \text{O} \rightarrow \text{CHO}^+ + e^-$ . Water has no carbon at all, so there is no pathway to CHO⁺ formation. CO₂'s single carbon is already fully oxidized (oxidation state +4) with no C–H bonds to pyrolyze, so it does not feed the same radical chemistry either. Because complete combustion products like H₂O and CO₂ are the end state of the flame reaction rather than intermediates capable of generating CH radicals, they contribute negligible ion current. This blindness is a feature, not a bug: it lets FID selectively quantify organic analytes in aqueous or CO₂-rich matrices without background interference from the solvent or matrix itself.

Is ECD response linear over its whole working range?

No, and this is one of the most important practical facts about the detector. ECD is linear only in the dilute limit where the analyte captures a small fraction of the available thermal electrons; as concentration rises, the signal saturates because a shrinking pool of free electrons is being progressively exhausted by capture, following an inverse (1/concentration) trend at high loading rather than a linear one. Typical usable linear range is only about two to four orders of magnitude, versus six or seven for FID. Quantitative ECD work therefore requires standards that closely bracket the expected unknown concentration, with regular verification that the calibration curve has not been extrapolated beyond its validated linear segment.

What does "mass-selective detector" mean, and how is it different from a full research-grade mass spectrometer?

"Mass-selective detector" (MSD) is the common instrument-vendor term for the relatively compact, quadrupole-based mass spectrometers routinely coupled to GC systems for benchtop use. They typically use electron-impact ionization at a fixed 70 eV (standardized so spectra are library-searchable and reproducible across instruments and labs), and a single quadrupole mass filter scanning up to a few thousand amu per second. They are less capable than high-resolution or tandem (MS/MS) research instruments — lower mass accuracy, no MS/MS fragmentation stage — but are far more affordable, robust, and simple to operate, which is exactly why they dominate routine confirmatory GC-MS work in forensic, environmental, and clinical laboratories.

Why is helium the near-universal carrier gas choice, and does detector choice affect that decision?

Helium offers a good balance of low viscosity (permitting long capillary columns at reasonable pressure drop), chemical inertness, and — critically for TCD work — very high thermal conductivity that maximizes the conductivity contrast against nearly every organic analyte. For FID and ECD, carrier gas choice matters less directly to the detection mechanism itself (FID burns whatever reaches the flame regardless of carrier identity; ECD uses N₂ or Ar/CH₄ makeup gas specifically optimized for electron thermalization, often added downstream of a He-carrier column). Helium's supply-cost volatility has pushed some laboratories toward hydrogen carrier gas for FID work, since hydrogen is cheaper and gives comparable or better efficiency, though safety considerations (flammability) require additional engineering controls.

Can a single sample be analyzed by more than one detector at once?

Yes, and this is common practice. Column effluent can be split (using a Y-splitter or specially designed flow-splitting union) to feed two or more detectors simultaneously — for example, TCD plus FID for a sample containing both permanent gases and organics, or FID plus ECD to give both a universal organic quantitation channel and an ultratrace halogen-selective channel from a single injection. Dual-column, dual-detector configurations are also used for confirmatory identification, running the same sample on two columns of different selectivity to a common detector type, reducing the risk of misidentifying a co-eluting interferent as the analyte of interest.

Why does 2-butanone (or any ketone) give less FID signal per carbon than an alkane of the same carbon number?

The carbonyl carbon in a ketone is already substantially oxidized (bonded to oxygen via a C=O double bond), so it does not participate as efficiently in the CH-radical-forming pyrolysis chemistry that ultimately produces the CHO⁺ ion responsible for FID signal. This "effective carbon number" concept assigns roughly a full unit of response to ordinary alkyl carbons (CH₃, CH₂, CH bonded only to C and H) but a fraction — often 0.6 to 0.8, rather than zero, since the discounting is only partial — to carbonyl carbons, and often close to zero for a fully oxidized carbon like that in CO₂ or formic acid. Quantitative FID work with oxygenated compounds should always use compound-specific calibration rather than assuming a naive count of total carbon atoms.

How does mass-selective detection provide compound identification, not just detection?

When electron-impact ionization strikes an analyte molecule at 70 eV, it produces a molecular radical cation that typically fragments in a reproducible, compound-characteristic pattern determined by bond strengths and the stability of resulting fragment ions. The resulting mass spectrum — abundance versus m/z — acts as a structural fingerprint that can be compared against extensive reference libraries (such as the NIST/EPA/NIH mass spectral library) for identification, in addition to the retention time information every GC detector provides. This dual capability — quantitative response (as in any detector) plus qualitative structural confirmation from fragmentation pattern — is what distinguishes mass-selective detection from every purely electrical-response detector like FID, TCD, or ECD, and is why GC-MS is the standard confirmatory technique in regulatory and forensic contexts.

GC Detectors

Where this lives today

The physical picture

The core ideas

Worked examples

Limiting cases, scope & common traps

The chemist behind it

Before this

Where this leads