Conductivity Detection in Cation Chromatography–Pros and Cons of Suppression
Dr. Joachim Weiss, International Technical Director, Dionex Corporation
The necessity for suppression in cation-exchange chromatography has been discussed for approximately 20 years. Other vendors believe that nonsuppressed conductivity detection for cation analysis is more sensitive than suppressed conductivity detection. Here, we examine this subject in a scientific way. Let’s start with the general requirements for a detection system:
- High sensitivity and specificity
- Low matrix dependence
- High reproducibility
Nonsuppressed vs Suppressed Conductivity Detection
Conductivity detection is separated into two approaches: suppressed and nonsuppressed. For the latter, the term indirect conductivity detection is also used in literature because background conductivity resulting from the eluent electrolyte is high (typically between 600 and 800 µS/cm). In addition, we subdivide detection methods as direct and indirect methods. Conductivity detection belongs suppressed conductivity detection, and is a bulk-property detection method because it only measures conductivity including eluent, analytes, anions, or cations. In this way, conductivity detection in ion chromatography (IC) is analogous to refractive index detection in classical HPLC. By combining it with a suppressor system, the conductivity detector is converted into a solute-specific detector. Indirect detection methods are characterized by a limited working range, low sensitivity and specificity, and strong dependence on temperature. These methods are only employed today if there is no alternative or if different compound classes are to be analyzed in the same run that do not require sensitivity (e.g., the analysis of wine for determination of a combination of sugars, alcohols, and major organic acids).
In 1979, Prof. Jim Fritz (Iowa State University, Ames, USA) introduced his concept of indirect (nonsuppressed) conductivity detection. He used salts of organic acids such as potassium dihydrogenphthalate as an eluent to separate anions because the resulting background conductivity was lower than that of inorganic salts. For many years, analytical chemists dealt with having positive and negative system peaks in the chromatogram along with interfering analyte peaks. At that time it was difficult to understand why such a method was proposed, because system peaks do not occur when using a suppressor system. Fritz also used organic acids such as pyridine-2,6-dicarboxylic acids (dipicolinic acid) for indirect conductivity detection of cations, however, the resulting problems were similar. The system peak in this case consisted of all the unseparated anions, and appeared in the column void volume. Depending on its size, this peak can interfere with early eluting analyte cations. Even with organic acid eluents, background conductivity is approximately 600 µS/cm, so that low-conducting analyte cations appear as negative signals in the chromatogram. By reversing the polarity of the detector, the resulting chromatograms can be acquired with a conventional chromatography data system. Depending on the anion concentration, the system peak in the void volume is then either positive or negative. A major disadvantage of this method is the low cation-exchange capacity of the separator used, which is a requirement for working with low eluent concentrations. These in turn are required to assure low background conductivity is maintained. As a consequence, column overloading effects are often observed in this form of cation analysis, even at relatively low analyte concentrations.
Analytical chemists agree that suppressor systems are necessary for sensitive and specific anion-exchange chromatography; however similar agreement does not exist on whether suppressor systems for cation analysis are necessary.
The most common arguments against suppressed conductivity detection for cation analysis include the nonlinear calibration demonstrated for weakly dissociated cations such as ammonium and amines. While this issue can be solved by employing a quadratic curve fitting, suppressed conductivity for ammonium and amines is less sensitive due to the equilibrium between ammonia and the dissociated ammonium as the suppression product. Moreover, analytical chemists prefer a linear calibration behavior requiring fewer calibration levels.
Other vendors claim that nonsuppressed conductivity detection is more sensitive than suppressed conductivity, proving that the difference between response factor and sensitivity is not always clearly understood. Also, nonsuppressed conductivity detection allows simultaneous analysis of transition metals and alkali-/alkaline-earth metals, which may only be feasible in special cases.
The problem of nonlinear calibration for weakly dissociated cations can be solved by combining a Cation Self Regenerating Suppressor with an Anion Self Regenerating Suppressor in converter mode. In this form of suppression, all cations are first converted into methanesulfonic acid (MSA) salts, while in the second suppressor all cations are replaced by hydronium ions to form MSA. Because MSA is a strong organic acid and thus completely dissociated, higher sensitivity is observed for all cations, along with linear calibration graphs.
Nonsuppressed Conductivity Detection is Put to the Test
In his book Ion Chromatography (1989), Hamish Small discussed suppression in cation analysis using a computer simulation to compare the relative sensitivity of both application forms of conductivity detection for the separation of sodium and potassium. To achieve an unbiased comparison, he kept parameters such as the stationary phase volume, capacity of the separator column, injection volume, and the selectivity coefficients for sodium and potassium constant. The result of this computer simulation proved that in nonsuppressed conductivity detection, the response factors for cations are higher, but sensitivity is lower due to the significantly higher noise level caused by the high background conductivity.
When to Consider Nonsuppresed Conductivity Detection
Nevertheless, a suppressor system for cation analysis is not always required. Analyte determinations at high concentrations can also be achieved with nonsuppressed conductivity detection to avoid large dilutions or very small injection volumes. Supporters of nonsuppressed conductivity detection also uphold the advantages of polymer-coated silica columns for isocratic separation of mono- and divalent cations together with hydrophilic amines such as alkanolamines. Mono-, di-, and tri-ethanolamine are used individually or in combination in many chemical processes including the conditioning of feed water in the power-generating industry.
Polymer-Coated Silica Columns
Although all polymer-coated silica columns exhibit poor uniformity of the polymer coating resulting in insufficient lot-to-lot reproducibility, they can be used with 3 mmol/L MSA as the eluent. This analysis can be performed on the Dionex IonPac® SCS 1 column. Based on a 250 mm column length, the relatively long analysis time of more than 30 minutes is a general disadvantage of this method. Like other polymer-coated silica columns, the IonPac SCS 1 column has a very low cation-exchange capacity in order to utilize low eluent concentrations to keep the background conductivity below 1 mS/cm. Instead of MSA, organic dicarboxylic acids, along with their complexing properties, can also be used as an eluent if different selectivities are required.
Advantages can be achieved by performing this analysis on IonPac CS18 column with suppressed conductivity detection under isocratic conditions. Analysis time can be cut in half, and even low analyte concentrations can be detected due to excellent baseline stability.
Another disadvantage of polymer-coated silica columns based on poly(butadiene-maleic acid) is their relatively poor compatibility with acidified samples. With an acid content in the sample of only 20 mmol/L, most of the carboxyl groups on the stationary phase surface are protonated, which makes the separation of divalent cations almost impossible. Polymeric cation exchangers can tolerate approximately five times more acid in the sample, because the carboxyl groups in the monomer used for the functionalization have a lower pK value.
Analyzing Several Classes of Cationic Compounds Using Nonsuppressed Conductivity
Regarding the ability to simultaneously analyze alkali- and alkaline-earth metals, ammonium, amines, and transition metals with nonsuppressed conductivity detection, in principle, such separation can also be carried out on IonPac SCS 1 column, but again requires a long analysis time. Even though zinc, cobalt, and manganese can be separated from other inorganic cations under the respective chromatographic conditions, it is still an exception as other transition metals can interfere with the determination of standard cations.
Addressing the Critics
Other vendors state it is common for transition metals to precipitate in the suppressor system. It is true that transition metals are converted to their respective hydroxides in the suppressor system which, with very few exceptions, are not dissociated and thus not detected. The precipitation of transition metals in the suppressor system, however, typically does not occur because only very small amounts of analytes are injected onto the column, such that the solubility product is not exceeded.
High-Capacity Columns and Gradient Elution
In addition to the sensitivity and specificity issues discussed above, the use of suppressor systems is mandatory for employing high-capacity columns or gradient elution techniques. The low cation-exchange capacity of separator columns required for nonsuppressed conductivity detection can lead to column overloading effects with analyte concentrations even in the mg/L range. To avoid this and to deal with large concentration ratios between cations (especially between sodium and ammonium), a high-capacity cation exchanger is required, which in turn requires a higher eluent concentration.
Using the IonPac CS16 column as an example, the eluent concentration under standard conditions is 30 mmol/L MSA which is not compatible with nonsuppressed conductivity detection because the background conductivity would be in the upper mS/cm range. Using a concentration gradient in combination with conductivity detection, a similar situation occurs. Although the standard application in cation analysis―the separation of alkali- and alkaline-earth metals―does not require a gradient elution technique, isocratic approaches are not suitable for strongly retained cations, screening analyses, or for complex samples containing various amines.
The gradient elution technique in cation analysis is ideal for separation of biogenic amines and standard inorganic cations. Biogenic amines are more strongly retained than alkaline-earth metals, which can lead to long analysis times and band broadening under isocratic conditions. Using a concentration gradient, late-eluting peaks can be focussed and thus detected almost as sensitively as early eluting peaks. Today, modern cation-exchange columns (e.g., IonPac CS17, IonPac CS18) allow such separations to be carried out with purely aqueous mobile phases that are compatible with Reagent-Free™ Ion Chromatography (RFIC™).
In conclusion, cation suppressors are not mandatory for all applications, but the advantages of cation suppression provides justifies the investment. The problems with cation suppression identified in the past―low sensitivity and nonlinear calibration for weakly dissociated cations―are no longer an issue.