Elsevier

Journal of Chromatography A

Volume 1356, 22 August 2014, Pages 180-187
Journal of Chromatography A

Assessment of the complementarity of temperature and flow-rate for response normalisation of aerosol-based detectors

https://doi.org/10.1016/j.chroma.2014.06.055Get rights and content

Highlights

  • Temperature and flow-rate gradients are used with an isocratic aqueous mobile phase.

  • Highly uniform response is obtained on a corona charged aerosol detector for a wide range of analytes.

  • Post-column addition of acetonitrile can be used to greatly enhance detector response.

  • The proposed approach is useful for impurity testing of pharmaceuticals using calibration based only on the major component.

Abstract

The solvent dependency of the detection response is a major limitation of corona-charged aerosol detection (C-CAD). The present study empirically investigates the utility of temperature and flow-rate gradients to overcome solvent gradient limitations of C-CAD. In preliminary flow-injection investigations, it is demonstrated that the response of C-CAD remains relatively unaltered with variations in flow-rate when used with water-rich eluents. Based on these findings two separation approaches were developed and their utility for C-CAD response normalisation was demonstrated using a mixture of eight analytes. In the first approach the use of a solvent gradient is replaced with a temperature gradient performed under isocratic mobile phase conditions. Detection response is further enhanced by mixing a secondary stream of pure acetonitrile with the column effluent, yielding a 3-fold increase in detection response. In the second approach, flow-rate programming is used to improve speed of isocratic-temperature gradient separation. The use of simultaneous variation in flow-rate and column temperature reduced the separation time by 30%, with relatively uniform analyte response. Lastly, an inverse-gradient solvent compensation approach was used to evaluate the response homogeneity and the applicability of the above approaches for quantitative analysis. Good peak area reproducibility (RSD% < 15%) and linearity (R2 > 0.994, on a log-scale) over the sample mass range of 0.1–10 μg was achieved. The response deviation across the mixture of eight compounds at seven concentration levels was 6–13% compared to 21–39% when a conventional solvent gradient was applied and this response deviation was comparable to that obtained in the inverse gradient solvent compensation approach. Finally, applicability of these approaches for typical pharmaceutical impurity profiling was demonstrated at a concentration of 5 μg/mL (0.1% of the principal compound).

Introduction

Aerosol-based liquid chromatography detectors, such as the evaporative light-scattering detector (ELSD) and the corona-charged aerosol detector (C-CAD), are regarded widely as quasi-universal detectors. These detectors have been proven to be useful for quantitative analysis of non-volatile analytes since they provide detection response irrespective of the optical absorption properties of the analyte and mobile phase. Nevertheless, non-linearity of response with sample concentration and solvent dependency of detection response are major concerns which limit the widespread acceptance of these detection techniques. Log-normalisation is generally used to relate detection response to sample concentration and it works well under isocratic conditions. However, with solvent gradient separation, response correction requires multiple calibration curves, since the response of aerosol detectors varies with the mobile phase composition. This poses practical difficulties in quantification, and it is highly desirable to have single calibrant quantification.

Different approaches to mitigate the solvent effect have evolved since the introduction of ELSD. Mathews et al. [1] have proposed a generic calibration method to account for the response non-linearity of the ELSD associated with sample concentration and solvent effects. A 3-D surface plot of detection response as a function of sample concentration and mobile phase composition, constructed by multiple injections of a single non-volatile and non-retained calibrant, was used to minimise the quantification errors. Similarly, a two-step response model was developed by Hutchinson et al. [2] to relate the C-CAD response to the variation in sample concentration and mobile phase composition. Another technique is the real time gas control feature of the ELSD, which allows control over the evaporation process and aerosol particle transport, providing more uniform detection response in solvent gradient separations [3]. In 2007, Gorecki et al. proposed a post-separation inverse gradient approach to compensate solvent gradient effects in C-CAD detection [4]. Subsequently, the utility of the same approach for the response normalisation in ELSD detection was demonstrated by de Villiers et al. [5]. In this approach a secondary stream of solvent is mixed with column effluent in a way that the mixture of aqueous–organic eluent being introduced in the detector inlet remains in equal proportions throughout the run.

All the above-mentioned approaches involve correcting the response variations resulting from the solvent gradient separation. Relative uniformity of response is an attractive goal, especially for impurity testing in the pharmaceutical industry. In this case, a calibration plot prepared for the active pharmaceutical ingredient could be used to determine the levels of impurities without the need for standards or calibration plots for these impurities. One potential solution to the non-uniformity of response of aerosol detectors would be to avoid the solvent gradient altogether and to optimise the separation under isocratic conditions. This might be achievable using alternate separation tools, such as temperature or flow-rate variation, which facilitate a comparable elution profile to that obtained using solvent gradient elution.

In the present study we have investigated the possibility of achieving relatively uniform detection response of the C-CAD for different analytes by employing isocratic separation with temperature and mobile phase-flow rate variation as optimisation tools. The capability of high temperature liquid chromatography (HTLC) to meet the ever-increasing requirements of separation throughput and control of separation selectivity has been investigated by many researchers. In recent times, several aspects of HTLC, including retention mechanism and principles, range of applications from analytical to capillary scale separations, challenges in broader acceptance and consequent developments in instrumentation have been published [6], [7], [8], [9]. It is well documented that with an increase in temperature, viscosity, surface tension and dielectric constant of water and aqueous-organic binary solvent mixtures drop significantly and approach the properties of the pure organic solvents [6], [7], [10]. It is noteworthy that an increase in temperature by 4–5 °C is required to achieve the effect comparable to 1% change in the amount of organic solvent (acetonitrile or methanol) present in the mobile phase. However, for chromatographic separations employing water-rich eluents, variation of temperature alone may not be sufficient to achieve elutropic strength comparable to a solvent gradient. Different strategies such as the use of isocratic binary solvent mixtures containing higher amount of organic solvents coupled with a high temperature ramp [9], [11], [12] or temperature pulsing [13] have been reported to improve the performance of HTLC separations. Thus, isocratic-temperature gradients can replace the use of solvent gradients in many instances and thereby complement aerosol detectors by minimising the contribution of solvent effects on non-linearity of response [6], [7], [14]. Nevertheless, the majority of the published work related to the use of temperature gradient separations has been limited to separations involving UV-detection. Even after quite intensive research in HTLC, the use of temperature gradients is not considered to be a mainstream approach for chromatographic optimisation. This is due primarily to potential practical limitations of temperature gradients, such as slow heat transfer in liquids, high thermal mass of analytical scale columns, and lack of a suitable heating device providing a sufficiently wide temperature range and a suitably steep temperature ramp. In view of these limitations, some combination of temperature gradients with additional optimisation parameters becomes inevitable.

Although the use of mobile phase flow-rate variation for improving the separation speed and peak efficiency has been investigated before [15], [16], [17], [18], [19], the risk of high column backpressure in a conventional high performance liquid chromatography set-up has deterred the widespread implementation of flow-rate programming. The dependence of column backpressure (ΔP) on viscosity (η) and linear velocity (u) of the fluid is commonly described by Darcy's equation:ΔP=β0×η×u×Lwhere β0 is the resistance factor and L is length of the column.

Column backpressure alters in direct proportion to viscosity and the linear velocity of the mobile phase. In HTLC separation, a temperature-induced decrease in solvent viscosity results in much lower column backpressure, permitting separations to be performed at high flow-rate. Variations in mobile phase flow-rate can be utilised as a means to improve efficiency and speed of HTLC separations. In HTLC separations, because of the reduced viscosity of the mobile phase, the diffusion coefficient alters significantly and thereby improves the inter-phase solute mass transfer, allowing separations to be performed at solvent velocities greater than the optimal Van Deemter value without significantly affecting efficiency. Therefore isothermal-HTLC separations are normally performed at higher flow-rates. However, in temperature gradient separations, the column backpressure varies across the length of the temperature ramp. Therefore, for optimal resolution and separation speed, a simultaneous alteration in linear velocity becomes necessary. Therefore, in order to fully realise the separation benefits of a temperature gradient, optimal mobile phase linear velocity needs to be maintained throughout the chromatographic run. Moreover, flow-rate programming can help to reduce the run time in a temperature gradient separation. Thus, a flow-rate gradient perfectly complements the use of a temperature gradient. Previous studies [15], [16], [17], [18], [19] involving simultaneous variation in column temperature and mobile phase flow-rate have been focused on separation throughput and were performed using a conventional UV-detector. In this study we have investigated the suitability of these separation strategies for response normalisation in aerosol detectors, which to the best of our knowledge has not yet been reported.

Section snippets

Reagents

HPLC gradient grade acetonitrile (ACN) and methanol used for the preparation of mobile phases was purchased from Scharlau Chemie (Sentmenat, Barcelona, Spain). High purity water (18.2  cm at 25 °C) was produced by a Millipore Milli-Q water purification system (Millipore, Molshiem, France). Analytical grade acetic acid, ammonium hydroxide and ammonium acetate were purchased from Sigma-Aldrich (Sydney, Australia). A mixture of sulfamethaxazole, sulfamethoxine, furosemide, prednisolone and

Effect of mobile phase composition and flow-rate on response of aerosol detectors (FIA)

To assess the practicality of using a flow-rate gradient, the effect of solvent composition and eluent flow-rate variation on the response of aerosol detectors was examined by FIA. As shown in Fig. 1(a), the response of both the detectors (ELSD and C-CAD) was enhanced with an increase in ACN up to 80%, which is in accordance with the commonly reported solvent-composition dependency of aerosol detection response. Effects of mobile phase flow-rate on detection response have been investigated

Conclusions

The present study has investigated the compatibility and complementarities of separation modes involving simultaneous variation in eluent temperature and flow rate, for use with C-CAD. The response of the C-CAD was found to remain relatively unaffected by the temperature and flow-rate variations, which reaffirms the ability of the C-CAD to provide truly mass dependent detection response. Both temperature and flow-rate gradients have their own limitations when applied separately, but in

Acknowledgements

This work was financially supported by the Australian Research Council’s Linkage Funding Scheme through the Grant no. LP0884030. Instrument support from Thermo Fisher Scientific, Australia is also gratefully acknowledged.

References (31)

  • B.T. Mathews et al.

    Improving quantitative measurements for the evaporative light scattering detector

    Chromatographia

    (2004)
  • S. Ball, I. Agilent Technologies, Universal quantification using the Agilent 385-ELSD Evaporative Light Scattering...
  • T. Gorecki et al.

    Universal response in liquid chromatography using charged aerosol detection

    Anal. Chem.

    (2006)
  • A.d. Villiers et al.

    Improving the universal response of evaporative light scattering detection by mobile phase compensation

    J. Chromatogr. A

    (2007)
  • C. McNeff et al.

    Practice and theory of high temperature liquid chromatography

    J. Sep. Sci.

    (2007)
  • Cited by (14)

    • Nebulizer-based detectors for liquid chromatography

      2023, Analytical Nebulizers: Fundamentals and Applications
    • Recent applications of the Charged Aerosol Detector for liquid chromatography in drug quality control

      2020, Journal of Chromatography A
      Citation Excerpt :

      Accordingly, in modern applications, the CAD is also described to perform with better sensitivity in HILIC methods, which utilize higher percentages of organic modifier in comparison to RPLC. This knowledge led to the technique of post-column addition of organic modifier in order to improve signals and signal-to-noise ratios as performed by several authors [8,43,95,97–99]. Since the signal intensities are significantly increased as presented in Fig. 5, it can be a helpful tool for improving the LOQs and the LODs of a method.

    • Simultaneous fingerprint, quantitative analysis and anti-oxidative based screening of components in Rhizoma Smilacis Glabrae using liquid chromatography coupled with Charged Aerosol and Coulometric array Detection

      2017, Journal of Chromatography B: Analytical Technologies in the Biomedical and Life Sciences
      Citation Excerpt :

      For those compounds whose reference substances were not commercially available, the quantification of those compounds is still a challenge. Charged aerosol detector (CAD) has been proven to be a useful for quantitative analysis of non-volatile analytes since it provides detection response irrespective of the optical absorption properties of the analyte and mobile phase [29]. The uniformity of responses between different analytes on CAD enabled the quantification of multiple components using a limited number of reference substances [30].

    • Chiral analysis of poor UV absorbing pharmaceuticals by supercritical fluid chromatography-charged aerosol detection

      2016, Journal of Supercritical Fluids
      Citation Excerpt :

      These detectors have often been referred to as “universal” detectors since they provide a near uniform response dependent upon the mass concentrations of analytes irrespective of their molecular structures. Among them, ELSD and CAD are probably the most commonly used due to their sensitivity, broad dynamic range and ease of use [38–47]. These detectors provide a nearly universal response at a constant mobile-phase composition for nonvolatile analytes, but the response factors vary with the gradient composition.

    • Systematic development of a group quantification method using evaporative light scattering detector for relative quantification of ginsenosides in ginseng products

      2016, Journal of Pharmaceutical and Biomedical Analysis
      Citation Excerpt :

      An essential prerequisite of utilizing these methods would be to equalize elution bandwidths of analytes in the original gradient since the compensation could hardly change the determined elution bandwidths. Recently, a complementary combination of temperature and flow-rate gradient instead of the solvent gradient for the response normalization in C-CAD was developed [19]. However, these successful developments still have not been utilized for ginseng products in practice due to the complicated requirements such as the surface model with a large number of injections, the equalized elution bandwidths of analytes, or the considerable changes in the liquid chromatographic system.

    • Performance of charged aerosol detection with hydrophilic interaction chromatography

      2015, Journal of Chromatography A
      Citation Excerpt :

      It is possible that highly-aqueous mobile phases produce excessively large droplets by condensation, which are removed in the nebuliser, resulting in reduced CAD response. Smaller droplets in greater numbers are formed with the lower viscosity, density and surface tension of highly organic eluents [16], which perhaps explains the better transport efficiency in these conditions. Table 4 shows that with AF pH 3 as buffer the uniformity of response was slightly improved in HILIC conditions compared with RPLC.

    View all citing articles on Scopus
    View full text