Ion Chromatography and Suppression

Suppressed conductivity IC is the default method (EPA Method 300.0/300.1 and ISO 10304 series) for anions in drinking water, wastewater, and environmental monitoring — chloride, nitrate, nitrite, sulfate, fluoride, phosphate, and bromate are routinely quantified in a single 10–20 minute run at low-ppb detection limits. Semiconductor fabs use suppressed IC to monitor ultrapure water for trace anionic contaminants (down to sub-ppb) that could poison wafer processing. Pharmaceutical quality control uses cation-suppressed IC to quantify counter-ions (Na⁺, K⁺, Ca²⁺) in drug formulations and to check for residual catalysts or excipient contaminants. Clinical and biomedical laboratories use suppressed IC for anion panels in biological fluids, and industrial hygiene labs use it to quantify oxyhalides (bromate, chlorite, chlorate) in disinfected water, a use case that depends critically on the low, stable suppressed background achievable only because the suppressor continuously converts a strongly conductive hydroxide or carbonate eluent into water or a very weak acid, preserving the sensitivity that makes trace-level oxyhalide monitoring possible at all. The same core hyphenation strategy — using an electrochemical or conductivity detector downstream of an Ion-Exchange Chromatography separation — underlies related techniques like capillary electrophoresis with conductivity detection, extending the suppression logic beyond pressure-driven liquid chromatography entirely.

Ask the tutor: How would the background conductivity and suppression gain change if the eluent were switched from carbonate/bicarbonate to a dilute KOH gradient generated electrolytically from water?

A conductivity detector measures the ability of the mobile phase to carry electric current between two electrodes, expressed as conductance $G$ (siemens, S) or as specific conductance $\kappa$ (S/cm, more often µS/cm for aqueous work). Every ion in solution — eluent counter-ion, eluent co-ion, and analyte alike — contributes to $\kappa$ according to the Kohlrausch law of independent migration: each ionic species contributes a term proportional to its concentration, its charge, and its own limiting equivalent conductance $\lambda^{\circ}$, essentially independent of what other ions are present (at the low ionic strengths typical of IC eluents, before ion-pairing or activity corrections become significant). The trouble is immediate: a carbonate/bicarbonate eluent used to elute anions like Cl⁻, NO₃⁻, and SO₄²⁻ from a high-capacity anion exchanger might run at 4.5 mM Na₂CO₃ / 1.4 mM NaHCO₃, giving a background conductivity in the hundreds of µS/cm — while the analyte anions of interest sit at 0.05–5 mg/L, contributing only fractions of a µS/cm on top of that background. Detecting a small signal riding on a huge, noisy baseline is the classic small-signal-on-large-background problem, and simply subtracting a baseline does not recover the lost signal-to-noise ratio, because baseline noise (pump pulsation, temperature drift) scales with the baseline itself. Suppression solves this by chemical conversion immediately downstream of the column, before the detector cell. For anion analysis, the suppressor is a strong-acid cation exchanger in the H⁺ form. As the eluent stream passes through, Na⁺ (from Na₂CO₃/NaHCO₃) is exchanged for H⁺:

Carbonic acid is a weak acid ($K_{a1} = 4.3\times10^{-7}$) and barely dissociates at the resulting concentrations, so the eluent's contribution to conductivity collapses. Simultaneously, every analyte anion — say Cl⁻ paired with Na⁺ in solution — is converted to its strong acid form:

HCl is a strong acid, fully dissociated, and H⁺ has an enormous limiting equivalent conductance ($\lambda^{\circ}_{\text{H}^+} = 349.8\ \text{S}\,\cdot\,\text{cm}^2/\text{mol}$) compared to Na⁺ ($\lambda^{\circ}_{\text{Na}^+} = 50.1\ \text{S}\,\cdot\,\text{cm}^2/\text{mol}$). The analyte signal is amplified roughly 7-fold in the cation contribution alone, while the background is suppressed 100- to 1000-fold. For cation analysis (using an acidic eluent like methanesulfonic acid, MSA), the suppressor is a strong-base anion exchanger in the OH⁻ form, converting MSA to water and analyte cations (e.g., Na⁺ paired with MSA⁻) to strong bases (NaOH).

Modern suppressors are almost universally membrane-based and electrolytically regenerated (e.g., Dionex's Self-Regenerating Suppressor, or Thermo Fisher's/Metrohm's equivalents), meaning the H⁺ (or OH⁻) needed for exchange is generated in situ by electrolysis of water, using an external regenerant reservoir of only deionized water and a small applied current — so the suppressor never needs manual chemical recharging and can run continuously for months.

The detector's raw signal is specific conductance, related to measured conductance by the cell constant:

where:

• $\kappa$ = specific conductance (S/cm), usually reported in µS/cm
• $G$ = measured conductance (S), the reciprocal of measured resistance $R$
• $L/A$ = the cell constant (cm⁻¹), a fixed geometric property of the electrode pair (typical IC cells: $L/A \approx 1$–$10\ \text{cm}^{-1}$, calibrated with a standard KCl solution)

Reading: conductance is what the electronics actually measure; specific conductance is the intrinsic solution property, recovered by multiplying by the (fixed, pre-calibrated) cell constant. Everything else in this article is really about predicting $\kappa$ from solution composition. Kohlrausch's law of independent ionic migration gives $\kappa$ as a sum over all ionic species:

where:

• $\lambda^{\circ}_i$ = limiting (infinite-dilution) equivalent (molar, per unit charge) conductance of ion $i$, in S·cm²/mol (values in Table 1)
• $z_i$ = magnitude of the ionic charge (so the sum is really over equivalents, not moles, for multiply-charged ions)
• $C_i$ = molar concentration of ion $i$ (mol/L)
• the $10^{-3}$ factor converts mol/L·(S·cm²/mol) into S/cm units consistently (since 1 L = 1000 cm³)

Reading: every ion adds its own independent term; more mobile ions (larger $\lambda^{\circ}$) and higher charge contribute disproportionately. This is exactly why swapping Na⁺ (50.1) for H⁺ (349.8) on the analyte, while swapping Na⁺ for H⁺ on a weak-acid anion (which then barely ionizes) on the eluent side, produces such a large net gain in signal-to-background.

Table 1. Limiting equivalent ionic conductances at 25 °C (standard reference values, e.g. from Robinson & Stokes/CRC compilations). Divalent ions listed per equivalent so the Kohlrausch sum uses $z_iC_i$ directly.

Example 1 (Foundational): Converting a suppressed-mode chromatogram peak area to concentration via linear calibration. A lab runs external standards of NO₃⁻ at 0.500, 2.00, and 5.00 mg/L under suppressed conductivity detection, obtaining peak areas (µS·min) of 0.842, 3.35, and 8.41 respectively. A sample gives a peak area of 2.10 µS·min. Find the NO₃⁻ concentration, with an honest uncertainty estimate. Step 1 — linear regression. Fit peak area $y$ (µS·min) vs. concentration $x$ (mg/L):

Deviations: $x$: $-2.000, -0.500, 2.500$; $y$: $-3.359, -0.851, 4.209$.

The intercept is essentially zero (as expected for a well-suppressed baseline), giving confidence the calibration is not distorted by a residual background offset. Step 2 — invert for the sample.

Step 3 — report with appropriate significant figures. Given three calibration points (only 1 degree of freedom for residuals), a full confidence interval requires the standard IC/HPLC calibration-uncertainty propagation, but to the precision supportable by three points, report:

Limiting-case check: if the unknown's peak area exactly equaled a standard's (say 3.35 µS·min), the back-calculated concentration should recover 2.00 mg/L exactly: $(3.35-(-0.007))/1.683 = 1.994 \approx 2.00$ mg/L ✓, confirming the regression is self-consistent.

Example 2 (Applied): Estimating suppressor capacity lifetime and diagnosing a "breakthrough" peak. A packed-bed (or membrane) cation suppressor used for anion IC has an effective exchange capacity equivalent to continuously neutralizing 50.0 µequiv/min of incoming Na⁺ at the eluent flow used. The lab runs a Na₂CO₃/NaHCO₃ eluent identical to the one above (10.4 mM total Na⁺) at a flow rate of 1.00 mL/min for an unusually long unattended overnight sequence. Estimate how long until the suppressor's capacity is (in principle) exceeded, assuming no electrolytic regeneration (i.e., treating it as a finite, non-regenerated bed, as older "hand-regenerated" fiber suppressors required). Step 1 — compute the molar flux of Na⁺ arriving at the suppressor.

Since Na⁺ is monovalent, this equals 10.4 µequiv/min. Step 2 — compare to stated suppressor throughput capacity. The suppressor is rated at 50.0 µequiv/min of sustained exchange capacity (this is a steady-state regeneration rate for an electrolytic suppressor, or a total finite capacity divided by an assumed usable lifetime for a non-regenerated one — here we treat the number as a finite total capacity in µequiv for the non-regenerated case, e.g., total capacity budget $Q_{\text{total}} = 3.00\times10^{4}\ \mu\text{equiv}$, a realistic order of magnitude for a small fiber suppressor). Step 3 — estimate time to exhaustion.

Step 4 — interpret the diagnostic symptom. As the suppressor approaches exhaustion, an increasing fraction of Na⁺ is no longer exchanged for H⁺; the background baseline conductivity rises progressively (since unconverted Na₂CO₃/NaHCO₃ contributes far more than H₂CO₃), and analyte peaks can show reduced amplification and peak-shape distortion (fronting/tailing) as the local suppressor chemistry becomes spatially non-uniform along the bed — an easily recognized symptom in an overnight batch sequence where late-run samples show systematically low recoveries and elevated baseline noise.

Limiting-case check: a modern electrolytically self-regenerating suppressor has, in effect, $Q_{\text{total}}\to\infty$ because H⁺ (or OH⁻) is continuously generated from water by the applied current rather than drawn from a finite chemical reservoir; setting $Q_{\text{total}}\to\infty$ in Step 3 correctly gives $t_{\text{exhaust}}\to\infty$, matching the real-world advantage of continuous electrolytic regeneration over finite packed-bed exchange capacity.

Example 3 (Exam/competition level): Resolving co-elution using selectivity and computing minimum resolvable concentration ratio from suppressed-mode noise. A suppressed anion-exchange separation resolves Cl⁻ and NO₂⁻ with resolution $R_s = 1.35$ under standard conditions (peaks nearly Gaussian, comparable widths). The baseline noise (peak-to-peak) in suppressed conductivity mode is $N_{pp} = 0.012\ \mu\text{S}$, and the detector's linear response constant (from calibration, analogous to Example 1's slope but per unit height rather than area) is $k = 3.40\ \mu\text{S per mg/L}$ for NO₂⁻ peak height. A regulatory method requires a limit of detection (LOD) defined as signal-to-noise = 3, and a limit of quantitation (LOQ) at S/N = 10. Additionally, determine the maximum concentration ratio $C_{\text{Cl}^-}/C_{\text{NO}_2^-}$ that still allows NO₂⁻ to be quantified without its peak being obscured, given that at $R_s = 1.35$ the peak-valley overlap contributes an interference equivalent to 4.0% of the larger peak's height superimposed on the smaller peak. Step 1 — LOD and LOQ concentrations from S/N criteria.

Step 2 — express the resolution-driven interference constraint. If Cl⁻ is present at concentration $C_{\text{Cl}}$ giving peak height $h_{\text{Cl}} = k_{\text{Cl}}C_{\text{Cl}}$ (with $k_{\text{Cl}} \approx 3.86\ \mu\text{S per mg/L}$, scaled from $\lambda^\circ_{\text{Cl}^-}/\lambda^\circ_{\text{NO}_2^-}$ relative to the NO₂⁻ constant — here we take it as given/measured directly at $k_{\text{Cl}} = 3.86\ \mu\text{S per mg/L}$), the interference riding under the NO₂⁻ peak is:

For NO₂⁻ to be quantifiable at its own LOQ, we require the true NO₂⁻ signal to exceed the interference by at least the same S/N=10 margin used for LOQ — i.e., the interference itself must not exceed the noise-equivalent contribution already budgeted, so a conservative working criterion is that interference stays below the LOD signal (0.036 µS), ensuring it does not itself introduce a spurious quantifiable "peak" mistaken for signal:

Step 3 — express as a maximum tolerable ratio at the NO₂⁻ LOQ. At $C_{\text{NO}_2^-} = C_{\text{LOQ}} = 0.0353\ \text{mg/L}$:

So Cl⁻ can outnumber NO₂⁻ by up to roughly 6.6-fold (by mass concentration) at the NO₂⁻ LOQ before peak overlap at $R_s=1.35$ compromises the quantitation; above this ratio, either better resolution (longer column, shallower gradient) or a different selectivity (different eluent concentration/column chemistry) is needed.

Limiting-case check: if resolution improved to baseline separation ($R_s \ge 1.5$, interference fraction → 0), the constraint in Step 2 relaxes to $C_{\text{Cl}} \le \infty$ in this idealized model — i.e., no ratio limit from peak overlap remains, matching the physical expectation that fully resolved peaks impose no such mutual constraint (real chromatograms would still have some finite tailing contribution, but the model's leading term correctly vanishes).

Suppressed ion chromatography traces to work at Dow Chemical in the mid-1970s by Hamish Small, Timothy Stevens, and William Bauman, who first demonstrated that a second ion-exchange column (or later, a membrane device) placed after the separator column could chemically neutralize eluent conductivity while amplifying analyte signal — turning conductivity detection, previously considered too nonspecific for trace ion analysis, into a viable quantitative tool. Their 1975 paper in Analytical Chemistry is generally credited as the founding publication of modern ion chromatography, and the technique was commercialized rapidly (Dionex Corporation, founded soon after specifically to exploit this discovery), transforming anion analysis from slow, wet-chemical, or specific-ion-electrode methods into a fast, multiplexed instrumental technique.