Elsevier

Journal of Chromatography A

Volume 1546, 20 April 2018, Pages 97-105
Journal of Chromatography A

Observation and explanation of two-dimensional interconversion of oximes with multiple heart-cutting using comprehensive multidimensional gas chromatography

https://doi.org/10.1016/j.chroma.2018.02.035Get rights and content

Highlights

  • Oxime compounds undergo thermal E/Z interconversion during MDGC analysis.

  • Here, 1D and 2D wax columns generated considerable interconversion of the isomers.

  • An ionic liquid 2D column produced a moderate interconversion behaviour.

  • Comprehensive MDGC mode with sequential sampling gave unusual 2D patterns.

  • Careful GC × GC column arrangements were also observed to give similar behaviour.

Abstract

Real-time interconversion processes produce unconventional peak broadening in gas chromatography (GC), and can be used to generate kinetic and thermodynamic data. In this study, an unusual separation situation in comprehensive two dimensional GC where two dimensional interconversion (i.e. a raised plateau in both first and second dimension, 1D and 2D) was observed in analysis of oxime isomers. This resulted in a characteristic and unusual rectangular peak shape in the two dimensional result. A related theoretical approach was introduced to explain the peak shape supported by simulation results which can be varied depending on concentration profiles and kinetics of the process. The simulated results were supported by experimental results obtained by a comprehensive heart-cut multidimensional GC (H/C MDGC) approach which was developed to clearly investigate isomerisation of E/Z oxime molecules in both 1D and 2D separations under different isothermal conditions. The carrier gas flow and oven temperature were selected according to initial results for 1D interconversion on a poly(ethyleneglycol) stationary phase, which was further used in both 1D and 2D separations to result in broad zones of oxime interconversion in both dimensions. The method involved repetitive injections of oxime sample, then sampling contiguous fractions of sample into a long 2D column which is intended to promote considerable interconversion. Comprehensiveness arises from the fact that the whole sample is sampled from the 1D to the 2D column, with the long 2D column replacing the short 2D column used in classical comprehensive two-dimensional gas chromatography, where the latter will not promote sufficient interconversion. Data processing and presentation permits a ‘rectangular’ distribution corresponding to the separated compounds, characteristic of this experiment.

Introduction

Separation in gas chromatography (GC) conventionally involves non-reactive and inert partition/adsorption processes, producing a single peak for each compound in the chromatogram. However, some configurationally labile molecules may undergo molecular transformation into different species on the separation time-scale, which can lead to unusual non-Gaussian peak shapes (strong peak broadening; overlapping or incompletely resolved peaks) [1,2]. This phenomenon may present difficulty in interpretation of GC data, for instance of some aldehydes or unsymmetric ketones which undergo interconversion within the separation timeframe. In general, interconversion is a process where two isomers undergo mutual conversion into each other, i.e. A ⇌ B (such as enantiomerisation or diastereomerisation processes) [3]. Interconversion in 1DGC has been studied in detail and is adequately understood. Model on-line reactions have been studied in GC, especially interconversion of E/Z oxime isomers. Proposed mechanisms for oxime isomerisation include 1) inversion, via sp-hybridisation of the nitrogen atom with the 180° C single bond N single bond O bond angle in the transition state [4,5] or 2) rotation around the C single bond N bond axis facilitated by polarisation of the C single bond N bond to result in E/Z isomerisation [4,5].

The overall separation/interconversion event defines the peak shape observed. Instead of two separate peaks of the E and Z isomers, on-line conversion of one isomer into the other causes observations somewhat like peak fronting or tailing in chromatograms. Interconversion on the separation timeframe may lead to a characteristic plateau between the two isomer peaks. The effect of temperature programming on interconversion has also been explored. Increasing the temperature of the system increases the magnitude of the observed plateau, due to increasing reaction rate at higher temperature [6]. The stationary phase used in the experiment also affects the extent of interconversion. Among different phases, it was found that oxime interconversion was more readily observed on a poly(ethyleneglycol) (PEG) phase [7].

Multidimensional GC (MDGC) conventionally employing two sequentially-arranged columns with different selectivity [8] has emerged as a high resolution technique, and proven useful for application to complex samples [9,10]. MDGC modes include single and multiple heart-cut (H/C) MDGC (GC−GC) [9,11]. A narrow H/C window will avoid sampling too many components into the 2D column. A long 2D column will improve 2D separation, at the expense of longer analysis time. As a result, normally only a few target regions are sampled in conventional GC−GC. The second mode, comprehensive two-dimensional GC (GC×GC), subjects the total sample to 2D separation, with transfer of 1D effluent zones usually less than the peak width of 1D peaks, for further separation on a short 2D column. A short narrow bore 2D column provides fast analysis time, and maintains high resolution [9] although this does reduce the 2D separation magnitude compared with the 2D column in H/C MDGC analysis.

MDGC has been applied to interconverting compounds, to study the molecular behaviour of these processes [12]. Interconversion in 1D of GC × GC has been studied where chiral phases were employed in both 1D and 2D [13], for a chiral oxime, to accomplish R/S separation and E/Z isomerisation. Although enantioseparation is difficult to achieve with a short column, the strategy employed a longer enantioselective column and allowed ‘wrap-around’ i.e. 2tR > PM, to provide sufficient resolution of the closely eluting enantiomers. In general, interconversion is not observed on short 2D columns, due to insufficient time to promote interconversion. Characteristic interconversion profiles on 2D in GC × GC were observed such as the small extent of interconversion of 2-phenylpropanaldehyde oxime on a 2D IL111 column at 140 °C [12]. The interconversion in both 1D and 2D led to an unconventional overall distribution defining rectangular peak shapes that have not been previously explained or investigated in detail. This could affect purity of target peaks in 2D, e.g. causing error in quantitative analysis. In order to investigate such phenomena and the resultant peak shape, an arrangement that has long columns in both 1D and 2D can be employed. Considering that a long 2D column with suitable phase (PEG) is required to promote interconversion, an option is to use a comprehensive H/C GC−GC (comprehensive MDGC) approach, with a sequential sampling strategy which progressively shifts the 1D H/C over the 1D elution profile with multiple injections applied to the whole sampled region. This allows both long 1D and long 2D columns to be investigated to deliver interconversion in both dimensions of the separations. A strategy of using long 1D and 2D columns has been previously used in GC–GC with multiple injections for high resolution alternative fuel characterisation of trace oxidation products [14]. The final data can be reconstructed into what mimics a GC × GC data presentation format.

This study reports observation of 2D interconversion in practical analysis of enantio- and stereo-isomers of oximes with GC × GC. The related theoretical approaches explain characteristic peak shapes in 2D interconversion. A comprehensive H/C GC–GC technique was further developed to investigate interconversion in both 1D and 2D separations employing PEG columns, with multiple repetitive heart-cuts, each offset by the sampling period, and a Deans switch (DS). Data analysis and the presentation approach are illustrated, and investigation of effects of temperature programs on separation results support the simulation results.

Section snippets

Theoretical

The characteristic shapes of the 1D and 2D interconversion peaks arise from: 1) interconversion of the isomers on the 1D column generating a broad zone of the isomers; 2) sampling of several regions along the interconversion zone resulting in injections of different initial concentration ratios of the two isomers (A0 to B0 ratio) to the 2D column; 3) combination of all the 2D profiles (injected with different isomer ratios) into an overall comprehensive 2D plot. Note that simulation of the 1D

Sample preparation

For 1DGC experiments, acetaldehyde oxime (acetaldoxime; 97%, Sigma-Aldrich, MO) was prepared in 1-hexanol with ethanol used as an internal standard. For GC–GC experiments, acetaldoxime was prepared in acetone. For GC × GC experiments, 2-phenylpropanaldehyde oxime (synthesis and characterisation procedures described elsewhere [13]) was prepared in HPLC grade n-hexane (Merck, Darmstadt, Germany).

GC–flame ionisation detection (FID)

In all experiments, the compounds were injected (1 μL, split ratio 10:1) into GC with an inlet

Results and discussion

In a recent report, [12] individual components of E/Z and R/S isomer forms of oximes were studied by use of a hybrid GC × GC–MDGC approach to allow isolation of the individual compounds and study their interconversion kinetics. Interconversion in dynamic GC × GC has been previously observed by this group where a classic GC × GC long 1D and short 2D column arrangement was employed in order to perform 2D analysis in a single run. However, this configuration only effectively allows interconversion

Conclusion

Real-time interconversion in GC × GC has formerly been observed where long 1D and short 2D columns were used for 2D analysis in a single run. However, interconversion in the 2D column has not been clearly observed due to the use of a short 2D column. Experimental situations where interconversion occurs in both 1D and 2D separation have been demonstrated here with a ‘rectangular peak shape’ characteristic of the separated isomers, plus a zone corresponding to interconversion constituting

Acknowledgements

The authors acknowledge funding from the ARC Discovery and Linkage program grants DP130100217 and LP130100048. YFW gratefully acknowledges the provision of a Monash Postgraduate Publication Award.

References (20)

There are more references available in the full text version of this article.

Cited by (10)

  • Multidimensional gas chromatography platforms for the analysis of flavours and odorants

    2022, Comprehensive Analytical Chemistry
    Citation Excerpt :

    For instance, a 30-min analysis with tsamp of 0.5 min will result in 60 repeat injections to complete a full C2DGC analysis. As a result, the latter method (multiple injections, each with multiple H/Cs) is more commonly employed, as reported in the analysis of volatile compounds [26,27], thermally oxidized biofuel [28], triacylglycerols (TAGs) [29], and on-column oxime isomerization (interconversion) [30,31]. If a classical C2DGC is on one end of the spectrum, where 2D separation power is maximized over the total analysis time, then GC×GC can be considered to be at the other end, where total separation time is greatly reduced with less 2D resolving power, but with substantial saving in total time.

  • Concepts, selectivity options and experimental design approaches in multidimensional and comprehensive two-dimensional gas chromatography

    2020, TrAC - Trends in Analytical Chemistry
    Citation Excerpt :

    The CH/C MDGC result for the comprehensive analysis is shown in Fig. 2A. A further example applied CH/C MDGC to study on-column molecular interconversion that generated intriguing observed peak shapes in MDGC, where interconversion occurred for both 1D and 2D separations [21], with the result shown in Fig. 2B. The CH/C MDGC analysis was also performed at different isothermal temperature using different 2tR,max and tsamp, with 2tR,max/tsamp values ranging from 10 to 20.

  • Multidimensional gas chromatography investigation of concentration and temperature effects of oxime interconversion on ionic liquid and poly(ethylene glycol) stationary phases

    2019, Analytica Chimica Acta
    Citation Excerpt :

    In order to promote interconversion in both dimensions, an option is to apply a long 2D column, but this will result in oversampling in GC × GC, with loss of detail and interconversion information for the 1D separation. Thus, multiple H/C 2DGC is preferred since this allows the use of both long 1D and 2D columns with minimum resolution loss [18,22]. This study develops comprehensive H/C MDGC for column selection, investigation of concentration effects on oxime interconversion and simultaneous investigation of interconversion in both 1D and 2D separations employing ionic liquid (IL) and PEG columns, respectively.

View all citing articles on Scopus
View full text