Heart-cutting multidimensional gas chromatography: A review of recent evolution, applications, and future prospects

Analytica Chimica Acta 716 (2012) 66–75

Contents lists available at SciVerse ScienceDirect

Analytica Chimica Acta journal homepage: www.elsevier.com/locate/aca

Review

Heart-cutting multidimensional gas chromatography: A review of recent evolution, applications, and future prospects Peter Q. Tranchida a , Danilo Sciarrone a,b , Paola Dugo a,c , Luigi Mondello a,c,∗ a b c

Dipartimento Farmaco-chimico, Facoltà di Farmacia, Università degli Studi di Messina, Viale Annunziata, 98168 Messina, Italy Chromaleont s.r.l. A spin-off of the University of Messina, Via Industriale, 143, 98123 Messina, Italy Università Campus-Biomedico, Via Álvaro del Portillo, 21, 00128 Roma, Italy

a r t i c l e

i n f o

Article history: Received 19 October 2011 Received in revised form 7 December 2011 Accepted 8 December 2011 Available online 19 December 2011 Keywords: Multidimensional gas chromatography Heart-cutting Deans switch Valve switching

a b s t r a c t The present contribution is focused on the main advances made in the field of heart-cutting multidimensional gas chromatography (MDGC), over approximately the last decade. Brief details on the history of classical MDGC are also given. A series of applications, carried out with modern-day commercially available instrumentation are shown, demonstrating the usefulness of the bidimensional methodology in specific analytical situations. Finally, the future prospects of MDGC are considered, within the shadow projected by a very powerful GC technique, namely comprehensive two-dimensional gas chromatography. © 2011 Elsevier B.V. All rights reserved.

Contents 1. 2. 3.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Heart-cutting MDGC over the past decade . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Future prospects of heart-cutting MDGC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Peter Q. Tranchida is Associate Professor in Food Chemistry at the School of Pharmacy in Messina (Italy). He has co-authored approximately 80 publications and presented work in over 120 oral/poster presentations in national and international meetings. His main research activities are focused on the analysis of volatiles and semi-volatiles contained in complex samples, by using advanced gas chromatographic methodologies. In particular, he has acquired considerable experience in the use of fast and veryfast gas chromatographic techniques. Furthermore, he is actively engaged in the development and application of classical multidimensional GC, comprehensive two-dimensional GC (GC × GC) and LC–GC (LC × GC) systems.

Danilo Sciarrone is Assistant Professor in Analytical Chemistry at the School of Pharmacy at the University of Messina (Italy). He has co-authored approximately 30 publications/book chapters and presented work in over 50 oral/poster presentations in national and international meetings. His research interests regard the application of chromatographic techniques coupled to mass spectrometry detection (GC, SPME-GC, chiral GC), heart-cut (MDGC) and comprehensive chromatography (LC × LC, GC × GC, LC × GC), as well as the development of fast chromatographic techniques.

∗ Corresponding author at: Dipartimento Farmaco-chimico, Facoltà di Farmacia, Università degli Studi di Messina, viale Annunziata, 98168 Messina, Italy. Tel.: +39 090 6766536; fax: +39 090 358220. E-mail addresses: [email protected] (P.Q. Tranchida), [email protected] (D. Sciarrone), [email protected] (P. Dugo), [email protected] (L. Mondello). 0003-2670/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.aca.2011.12.015

67 68 73 75

P.Q. Tranchida et al. / Analytica Chimica Acta 716 (2012) 66–75

Paola Dugo is Full Professor of Food Chemistry at the University of Messina (Italy) and at the “Campus Biomedico” in Rome. Her research interests include the development of comprehensive chromatography and high throughput separation methods, applied to the study of Citrus products (essential oils and juices), of compounds with a possible biological activity in natural samples (carotenoids, anthocyanins, coumarins), of plant essential oils, of the aromatic fraction of wine and other alcoholic beverages, and of triglycerides in food lipids. She is the author of approximately 130 scientific papers, and has been invited speaker in several national and international congresses.

1. Introduction Before going into the details on modern heart-cutting multidimensional gas chromatography (MDGC) one must ask oneself: what is the approach, and why/when is it needed? In responding to such questions, some historical milestones will also be mentioned. The answer to the first question is straightforward: a classical (or heart-cutting) MDGC set-up usually consists of two conventional capillary columns (e.g., 30 m × 0.25 mm ID × 0.25 ␮m df ), connected in series and characterized by a differing selectivity (e.g., apolarpolar, polar-chiral, etc.). Today, MDGC experiments are mainly carried out with capillary columns, for obvious reasons related to efficiency. Capillaries were employed for the first time, in the MDGC field, in 1964 [1]: two 150 ft columns were connected via a pneumatically operated diaphragm 6-port valve. In general, the columns can be located in a single or two separate ovens, with the latter solution being certainly the most flexible one. The first double-oven MDGC experiment was carried out nearly 40 years ago [2]. A transfer device is located between the two dimensions, and enables the passage of chromatography bands, from the primary to the secondary column. A cryotrap, situated somewhere between the two capillaries, is an option bringing the following advantages: (I) sensitivity enhancement through analyte re-concentration; (II) peak capacity increase, by substantial reduction of the width of primary-column chromatography bands; (III) the concentration of trace-amount compounds, co-eluting (or not) with other constituents, applying sequential analyses. It must be added that, for many compounds, focusing can be attained by keeping the second oven at a low temperature during heart-cutting. The transfer systems developed can be classified in three groups: (I) in-line valve, (II) out-line valve and (III) valveless systems. In the first group, a valve interfaces the two columns in a direct manner; out-line valves are employed to regulate the direction of gas flow towards the column interface, while valveless systems form a third minor MDGC group. When an MDGC instrument (in- or outline valve) is in the stand-by mode, a one-dimensional analysis is being developed; when the configuration is switched to the cutting mode, the primary-column effluent is directed towards the second capillary. The greater the number of transfers achieved, the higher the possibility of a mix-up of previously separated compounds. Although such an occurrence goes against a “golden” rule of multidimensional chromatography, namely that “all compounds resolved in the first dimension, must remain so in the second”, the event is acceptable if all target analytes remain separated. With regards to the second question, the reason why classical MDGC is needed appeared evident right from the first experiment, carried out over 50 years ago [3]: Simmons and Synder described

67

Luigi Mondello is Full Professor of Analytical Chemistry at the Dipartimento Farmaco-chimico of the University of Messina, Italy and at the University “Campus Biomedico” in Rome. He is the author of approximately 200 scientific papers, 29 book chapters, and co-editor of a book on Multidimensional Chromatography (Wiley), and editor of a book on Comprehensive Chromatography in combination with Mass Spectrometry (Wiley). His research interests include one-dimensional chromatography techniques (HRGC, HPLC, GC–MS, HPLC–MS), combined particularly with mass spectrometry, as well as multidimensional chromatography methods, such as LC–GC–MS, GC–GC–MS, GC × GC, LC × LC, LC × GC, and their use in the study of complex samples.

a first-dimension boiling-point separation of C5 –C8 hydrocarbons, and a polarity-based analysis of each of the four hydrocarbon groups in the second dimension, in sequential applications. The sum of the four applications produced a comprehensive multidimensional one. The valve-based instrument produced quite remarkable results and the two affirmed that “separations can be obtained with this column arrangement which are not normally possible with previously described arrangements of single columns and multiple columns connected in series.” Classical MDGC is usually employed for the bidimensional analysis of a part of the sample, while the other fraction is of no analytical interest. In short, the primary-column separation can be considered as a prefractionation step, with no mess involved. Although, in principle, it is possible to carry out sequential cuts on the entire primarycolumn effluent, such an approach is hardly ever used if all the sample constituents require analysis in both dimensions. Other rather unpractical and elaborate alternatives for such an analytical scope have been developed, viz., the use of an array of seconddimension capillaries, one for each heart cut, or parallel traps located between the two dimensions, used for storage purposes [4]. Deans achieved a major breakthrough at the end of the 1960s, with the introduction of pressure switching [5], what is here defined as an out-line valve system. The invention of the Deans switch brought a series of unquestionable advantages, such as no temperature limitations, artefact formation, memory effects, or direct contact between valve mechanical parts and sample compounds; additionally, the contribution of such interfaces towards band broadening was very low. The Deans switch also enabled basic MDGC operations such as venting and backflushing. The research work carried out by Deans layed the basis for the development of the most popular, commercially available MDGC instrument used during the 1980s and 1990s, viz., the Siemens instrument [6]. The transfer device employed in the Siemens MDGC system was defined as a “live” switch, and was characterized by a Pt capillary which was introduced directly into the outlet and inlet of the first and second dimension. During the same time period other options became available, such as the SGE multidimensional conversion system (a Deans-switch interface) [7], the IBM (6-port valve) instrument [8], the Perkin-Elmer (Deans switch) instrument [9], the Analytical Controls [10], or the Gerstel [11] Deans switches. In a further research carried out by Deans (and Scott), a coupledcolumn system was described in which the pressure, at the junction point between the two capillaries (mid-point pressure), was controlled independently [12]. Though such a GC configuration was simple, it was also very interesting. As will be seen, such an approach is capable of MDGC analysis. In general, the work carried by Deans is today exploited in advanced MDGC systems [5,12].

68

P.Q. Tranchida et al. / Analytica Chimica Acta 716 (2012) 66–75

Vacuum conditions have also been exploited to capture specific chromatography bands, from a capillary column [13]. The rather elaborate system, which belonged to the out-line valve approaches, did not find wide use. Considerable technological improvements in valve design were made in the 1980s, with the introduction of micro-volume connections and use of thermally stable elasteromeric material. Satisfactory results were attained (using a Valco six-port valve) on a test solution and real-world samples (gasoline, lemon essential oil, urine extract), with no apparent problems related to valve activity observed [14]. During the 1980s and 1990s, lab-constructed in-line valve systems appeared [15]. When the use of MDGC is necessary, is a thing dependent on sample complexity, as well as on the judgement, expertise and habits of the analyst. In chromatography, it is always a nice challenge to solve things with a single column and a mass spectrometer (GC–MS). In many situations, mass spectrometry can circumvent the presence of overlapping compounds at the GC column outlet. For example, electron-impact “unit-mass resolution” mass spectrometry can help overcome a non-sufficient GC separation by using selected ion monitoring (SIM), extracted ions or simply by deriving mass spectra at different points across a peak. Chemicalionization “unit-mass resolution” mass spectrometry exalts the MS separation capability, but is characterized by a diminished potential for identification. Tandem MS systems are highly selective and sensitive, but the analyst needs to know what to look for, prior to the analysis. High resolution mass spectrometry is also characterized by high selectivity, but requires a substantial investment and in some instances cannot overcome an insufficient GC separation (e.g., enantiomers). If the isolation of a limited number of target analytes, from interfering compounds, proves to be an exceedingly difficult task, then the use of a second chromatography dimension is an excellent option. The total peak capacity of the twin-column system becomes that of the primary column (n1 ), plus that of the second capillary (n2 ), with the latter multiplied by the number of cuts. The exposition of a simplified sample (heart-cut), to the full separation power of a conventional column (100,000–130,000 N), usually does the trick. Classical MDGC can be considered a prime choice in the analysis of a limited number of target analytes in medium to highly complex samples. The word “medium”, in the present authors’ opinion, relates to samples with 100–200 volatile constituents. The hyphenation of a mass spectrometer to an MDGC system produces a very powerful system, with three analytical dimensions. A brief note on the main differences between classical and comprehensive multidimensional GC (GC × GC), a technique first reported in 1991 [16]. GC × GC experiments are performed using a short second column (1–2 m), because it must receive primarycolumn cuts in a continuous and sequential mode. The GC × GC transfer device, defined as modulator (usually cryogenic), enables the rapid accumulation and re-injection of chromatography bands from the first to the second column. The total peak capacity of a GC × GC twin-column system becomes roughly that of the primary column (n1 ), multiplied by that of the second capillary (n2 ). Comprehensive 2D GC can be considered as an excellent option in the analysis of many target analytes or of all the constituents of medium to highly complex samples. GC × GC (or GC × GC–MS) is a fantastic technique for the generation of a specific sample “fingerprint”. Though more details will be given later on GC × GC technology, the reader is directed to recent literature for thorough information [17].

2. Heart-cutting MDGC over the past decade During the 1990–2000 period two reviews on classical MDGC appeared [4,18]. Additionally, a book, devoted to

“Multidimensional Chromatography”, contained a specific chapter based on classical MDGC [19]. Over the decade, the use of MDGC was rather limited, a reason being the great popularity of GC–MS. The number of GC–MS systems (the majority of systems employed were single quadrupoles) commercialized obscured that of MDGC instrumentation. In the later of the two reviews (1999), the author reported that ca. 2500 (two-thirds) of the current papers in the GC field were based on the employment of GC–MS, whereas only about 50 publications/year described an MDGC experiment. During the last decade, it could be thought that comprehensive 2D GC has outshadowed classical MDGC in a similar manner as GC–MS did during the 80s and 90s. Such an opinion is true to a certain extent. Although GC × GC is a great multidimensional approach and has gained a lot of popularity, it is also true that in much published research, classical MDGC would have probably provided a better analytical result. There has certainly been, at times, an over-emphasis of GC × GC, even though it is an outstanding method. The high resolution power of a conventional column (in the second dimension) is often the best choice in the analysis of target compounds. Even though cryogenic GC × GC offers unmatched sensitivity, the use of a 1–2 m micro-bore column segment as second dimension gives a reduced separation power. To have an approximate evaluation of the diffusion of heartcutting and comprehensive MDGC methods, over the past decade, the terms “comprehensive two-dimensional gas chromatography” and “multidimensional gas chromatography” were used in a search, over the period 2000–2010, on Journal of Chromatography A. The feed-back was: 240 GC × GC papers, 34 heart-cutting MDGC works and 8 contributions containing both technologies. Though it is almost certain that a series of papers were missed, the 7:1 proportion gives a good idea on the favour that GC × GC has gained over classical MDGC. Although the authors are not aware of the number of commercial systems sold worldwide, effective classical MDGC instruments are now available on the market. Such systems are characterized by accurate electronic pressure units, are not of such elaborate construction as what they were in the past, are supplied with simple-to-use software enabling complete automatic instrumental control, and give highly repeatable results, even under multiple-cut conditions. In terms of transfer device, the Deans switch has conquered the scene. If one desires to use in-line valve MDGC, then the required hardware (a switching valve, connections, an auxiliary pressure source, etc.) must be purchased and then installed in the GC. Before the descriptions of the commercial systems (and applications), attention will be devoted to the concept of in-series twin-capillary GC, the coupled-column approach proposed by Deans and Scott [12], and recent versions of that technology. Let us consider an extremely simple sample, composed of two constituents, defined as ␣ and ␤. If such a sample is analyzed on two conventional capillaries (e.g., 30 m × 0.25 mm ID) linked in series, with differing selectivities, and under ideal conditions of temperature (with a single oven) and flow, then the possible analytical outcomes are four: (I) ␣ and ␤ overlap to some degree on both capillaries; (II) ␣ and ␤ remain entirely resolved on both columns; (III) ␣ and ␤ are separated on the first column but overlap to some degree on the second; (IV) ␣ and ␤ co-elute to some degree on the first column, and are separated on the second. In cases I and II, the employment of the second stationary phase is useless, while it is entirely negative in the third situation. On the contrary, a clear advantage is attained in the fourth situation. If the aforementioned example is extended to a medium-complexity sample (e.g., 100 compounds), then the number of possible combinations becomes exceedingly high. However, the final number of compounds that one could expect to resolve would most probably be not that much

P.Q. Tranchida et al. / Analytica Chimica Acta 716 (2012) 66–75

69

Fig. 1. Scheme of a coupled column system, with stop-flow operation. Abbreviations: C1 = first dimension; C2 = second dimension; V = pneumatic valve; BC = ballast chamber; PC = pressure controller; I = injector; CG = carrier gas source. Reprinted with permission from T. Veriotti, R. Sacks, Anal. Chem. 73 (2001) 3045–3050. Copyright 2001 American Chemical Society.

higher than that attained using a 60 m × 0.25 mm ID column, with a single stationary phase. One way to improve the chromatography of a coupled-column system would be to operate each stationary phase using independent temperature programs [20]. A further way of improving the separation performance of a twin-capillary system can be achieved by manipulation of the gas flows in each column, by using an additional pressure source connected to the columns connection point [12]. A series of interesting papers, based on such an approach, have been described by Sacks and co-workers [21–23]. A scheme representing a “series-coupled column ensemble with stop-flow operation” is shown in Fig. 1. The system illustrated is one belonging to the out-line valve [a pneumatically activated one (V)] MDGC systems. The valve is connected on one side to the columns junction point, and on the other to an aluminium ballast chamber (BC). The pressure in BC is controlled by an electronic pressure controller (PC). An FID is also located between the two chromatography dimensions. If a pressure pulse, equal to that of the injector (I), is applied to the columns connection point, then a condition of stop-flow will arise in the first column, while elution continues in the second dimension. The stop-flow MDGC instrument was employed in combination with a third separative dimension, viz., a time-of-flight (ToF) mass spectrometer, in the high-speed analysis of essential oil compounds [23]. Such MS systems possess the unique capacity to unravel coeluting GC peaks through spectral deconvolution. Fig. 2 illustrates three chromatograms relative to the analysis of 9 compounds, using the tandem-column ToF MS system. About 10% of the first-dimension (7 m × 0.18 mm ID × 0.2 ␮m trifluoropropylmethyl polysiloxane) effluent was directed to an FID (Fig. 2a); the single-column GC-FID chromatogram shows that peaks 3–4 overlap completely, while compounds 5–6 coelute partially. Observing the middle total-ion-current (TIC) chromatogram, that is an application carried out with no mid-pressure regulation, it can be easily concluded that the chromatography situation has become worse: peaks 1–2, separated on the polar column, co-elute completely on the apolar second column (7 m × 0.18 mm ID × 0.2 ␮m 5% phenyl); though peaks 3 and 5 are now entirely separated from components 4 and 6, respectively, peaks 4 and 6 now overlap completely; compounds 7-8-9, previously resolved, are now mixed together after their passage on the second capillary. A 4-point stop-flow experiment, carried at times indicated by the arrows in Fig. 2a, enabled the chromatographic separation of all nine volatiles. The first stop-flow operation was made just after the

Fig. 2. Chromatograms relative to the FID primary-column analysis (a), to the totalion-current MS result with no stop-flow operation (b), and to the total-ion-current MS analysis with stop-flow operation (c). Peak identification for the essential oil compounds: (1) camphene; (2) furfural; (3) eucalyptol; (4) terpinolene; (5) benzaldehyde; (6) octanal; (7) ␤-caryophyllene; (8) geranyl acetate; (9) eugenol. Reprinted with permission from T. Veriotti, R. Sacks, Anal. Chem. 75 (2003) 4211–4216. Copyright 2003 American Chemical Society.

first-dimension elution of peak 1: compound 2 was halted for 5 s in the final primary-column segment. Elution in the first dimension was interrupted a further 3 times: once to separate compounds 4 and 6, and the remaining two times to separate compounds 7-8-9. Mass spectral deconvolution was necessary for the chromatogram illustrated in Fig. 2b, but was clearly not needed in the stop-flow analysis. The stop-flow approach is both simple and attractive, and could be employed as an alternative MDGC method, though currently it is hard to encounter going through the literature. An MDGC instrument, characterized by a microfluidic transfer device (Agilent Technologies), has been recently described [24]. The interface is a thermally stable, leak free, chemically inert, low dead volume, Deans switch, manufactured through capillary flow technology (CFT): through holes and flow channels are etched into stainless steel plate halves, which are folded, heated to a very high temperature (>1000 ◦ C); then, high pressure is used to produce a diffusion-bonded metal sandwich. The internal channels, created in a similar mode to the production of integrated circuits, are deactivated with a coating layer. Leak-free connections are made by using metal ferrules. Optimum conditions for stand-by, cutting and backflushing processes are created by using electronic pressure control. The interface contains five ports: two are connected to the primary and secondary columns, while another is linked to a restrictor, with the same flow resistance as the second dimension. The latter requisite is important, because the pressure drop across the primary column must remain constant during the two operational modes. If

70

P.Q. Tranchida et al. / Analytica Chimica Acta 716 (2012) 66–75

Fig. 4. Capillary flow technology LTM MDGC: second-dimension separation of oxygenated compounds in fuel (a); first-dimension chromatogram of the hydrocarbons, after cutting (b); backflush operation, after cutting (c). Reprinted with permission from J. Sep. Sci. 31 (2008) 3385–3394. Copyright 2008 John Wiley and Sons.

Fig. 3. Scheme of the “Agilent” Deans switch in the by pass (stand-by) and inject (cut) modes. Reprinted with permission from Anal. Chem. 79 (2007) 1840–1847. Copyright 2007 American Chemical Society.

pressure-drop differences do occur, then there will be a mismatch between the programmed chromatography-band transfers, which are set on the basis of a preliminary stand-by separation, with those which occur during an MDGC analysis. The restrictor is usually connected to a detector, to monitor the first-dimension separation. The remaining two entrances are fixed, being linked to a two-way solenoid valve. Fig. 3 shows how the device achieves the bypass (stand-by) and inject (cut) states. The primary flow, which is always lower than the auxiliary flow, enters the interface through the central port. During the standby mode, the solenoid valve directs the auxiliary flow to the top left part of the Deans switch, which is connected, via an internal channel, to the port linked to the secondary capillary. Once inside the interface, the additional gas flow is divided in two parts, one directed to the second column and the other crosses the internal vertical channel ending up in the restrictor. Prior entry to the restrictor, the auxiliary flow is mixed with the first-dimension flow. When the solenoid valve is switched to the cutting state, the auxiliary flow is directed to the bottom left part of the transfer device, which is connected, via an internal channel, to the port linked to the flow restrictor. Once inside the interface, the auxiliary gas flow is split between the restrictor and the second column. As in the standby mode, the additional flow is mixed with the primary-column flow. In recent work, the “Agilent” instrument was combined with low thermal mass (LTM) GC [25]: the microfuidic interface was employed with success in multi-cut and backflushing operations, while LTM technology was used independently in both dimensions, for fast column heating and cooling. An application devoted to the analysis of light oxygenated compounds in fuel, is shown in Fig. 4. The target analytes were nicely isolated in the second dimension (10 m × 0.25 mm ID × 1.2 ␮m 100% polyethylene glycol) using the following temperature program: 45 ◦ C (3 min) to 150 ◦ C (3 min) at 30 ◦ C min−1 (upper chromatogram). A different

temperature program was applied for the characterization of the hydrocarbons in the first dimension (30 m × 0.25 mm ID × 1.0 ␮m 100% dimethylpolysiloxane): 150 ◦ C (4 min) to 300 ◦ C (3 min) at 30 ◦ C min−1 (middle chromatogram). During the stand-by mode, analytes exited the transfer device onto a 2.7 m × 0.18 mm ID fused silica restrictor. A backflush operation was carried out on the hydrocarbon fraction to reduce the analysis time and to prevent column contamination (lower chromatogram). Although the “Agilent” MDGC system is most suited for heartcutting operations, it has also been employed in GC × GC work [26]. Prior to the description of the specific experiment, further details will be given on the characteristics of a comprehensive 2D GC instrument: in the first dimension, a boiling-point separation (non-polar stationary phase) is the most common choice; then, single analytes or packets of isovolatile compounds are analyzed on a medium or high-polarity stationary phase. The linkage of a conventional capillary to a short micro-bore column segment guarantees slow and fast peak generation in the first and second dimension, respectively. The transfer device, defined as modulator, works continuously and sequentially throughout the analysis. Any modulator must enable the collection of peak sections (ideally, 3–4 per peak), from the primary column, and inject them onto the secondary one. Second-dimension analyses are very fast, normally completed within 4–8 s. The time between consecutive second-dimensions injections is defined as “modulation period”, and is equal to the secondary-column analysis timeframe. Each second-dimension separation is characterized by analytes with the same first-dimension retention time (expressed in min) and different second-dimension elution times (expressed in s). If a 4000 s GC × GC application with an 8-s modulation period is considered, then five hundred 8-s seconddimension chromatograms, positioned side-by-side, will form a (monodimensional) comprehensive 2D GC trace (only one detector is used). It is clear that the evaluation of a “raw” GC × GC chromatogram, as such, is entirely unpractical; hence, dedicated software is necessary to generate a two-dimensional separation space: each rapid chromatogram is positioned orthogonally to an x-axis, while the compounds separated in the second dimension are aligned along a y-axis, and are characterized by an oval shape. With regards to peak quantitation, it is necessary to sum the peak areas relative to the same compound in each fast seconddimension chromatogram; to do such an operation manually is

P.Q. Tranchida et al. / Analytica Chimica Acta 716 (2012) 66–75

71

Fig. 5. Schemes of the Gerstel “selectable 1D/2D GC–MS” system. The one-dimensional GC–MS configuration is reported in (a); the transfer configuration is illustrated in (b); the two-dimensional GC–MS/backflush configuration is shown in (c). Abbreviations are defined in the text. Reprinted with permission from J. Chromatogr. A 1217 (2010) 2903–2910. Copyright 2010 Elsevier.

an arduous task and, again, the use of dedicated software is mandatory. In 2007, Seeley et al. extended the use of the “Agilent” MDGC system to GC × GC analysis [26]. A fast GC × GC analysis was carried out on gasoline, using a100% dimethylpolysiloxane 15 m × 0.25 mm ID × 0.50 ␮m first and a “wax” 2.5 m × 0.25 mm ID × 0.10 ␮m second dimension (a restrictor of the same dimension was used). The flow in the primary column corresponded to 1 mL min−1 , while that of the auxiliary flow was rather high, namely 9 mL min−1 . As aforementioned, the auxiliary flow was split between the secondary column and the restrictor. The primary-column flow was mixed with that of the secondary column, during the injection stage. High gas flows were necessary for rapid separations on the second column. A particularity of the experiment was that related to the modulation period applied, which was 1 s: the Deans switch was held in the bypass position for 0.93 s and in the cutting mode for 0.07 s. Consequently, during each modulation period, only a 7% fraction of the primary column flow was directed to the second column, while the remaining part was directed to waste. Such a low duty cycle can be considered as a main disadvantage. However, the experiment remained of great interest because it demonstrated, for the first time, that a Deans switch was capable of producing one-dimensional GC, heart-cutting MDGC and GC × GC data. In other terms, a unified instrument.

An integrated “selectable 1D/2D GC–MS” system, currently commercialized by Gerstel, is characterized by an Agilent capillary flow technology interface and low thermal mass GC modules [27]. The unified system is equipped with dedicated software to deal both with GC–MS and MDGC–MS applications. The main novelty is that the same mass spectrometer is employed in both application types, namely for stand-by and cutting analysis, meaning that peaks subjected to one- and two-dimensional analysis appear in the same GC–MS chromatogram. Apart from the MS system, other detectors can be used for one- and two-dimensional GC applications. Fig. 5 shows schemes of the unified instrument: an Agilent GC was equipped with a thermal desorption inlet, two independent LTM units, a CFT Deans switch and a CFT cross-union, to connect the restrictor (0.54 m × 0.10 mm ID), the second (5% phenyl) column (10 m × 0.18 mm ID × 0.40 ␮m), the MS transfer line, and the transfer line to other detectors. The Agilent Deans switch connections have been described previously. Apart from the PCM (pressure control module) connection (168.4 kPa) to the transfer device, a further line of the auxiliary pressure source (21.0 kPa) supplied make-up gas to the cross-union. A stand-by GC–MS application is carried out by diverting the “wax” primarycolumn (10 m × 0.18 mm ID × 0.30 ␮m) flow towards the restrictor (Fig. 5a), while a second-dimension GC–MS analysis is performed by activation of the solenoid valve (Fig. 5b); if required, the transferred chromatography band can be re-concentrated at the head of

Fig. 6. Scheme of the “Shimadzu” Deans switch in the stand-by (a) and cut (b) configurations. Abbreviation definitions are reported in the text.

72

P.Q. Tranchida et al. / Analytica Chimica Acta 716 (2012) 66–75

Fig. 7. Chromatogram relative to a first-dimension perfume analysis, after heart-cutting and with the position of each cut indicated by a Roman number. Reprinted with permission from LC GC Eur. 21 (2008) 130–137. Copyright 2008 Advanstar Communications.

the secondary column by using a cryotrap. At the end of the transfer period, the remaining part of the sample can be backflushed by reducing the head pressure to 10 kPa (initially 212.7 kPa), and the second-dimension LTM heating can begin (Fig. 5c). Though the “selectable 1D/2D GC–MS” system is certainly interesting it also appears to be characterized by a drawback, mainly the difficulty to perform multiple heart-cuts. The latter operation could be achieved by accumulating more than one heart-cut in the cryo-trap (or even at the ambient GC temperature if not excessively volatile) or by performing a single application for each heart-cut. The successful outcome of the first option would depend on the capability of the second column to separate all the entrapped compounds, while high time costs could characterize the second route. Gerstel also commercializes a more classical heart-cutting MDGC system (MCS: “multidimensional column switching” system), with a Deans switch, and a primary (monitor) and secondary (main) column detector (http://www.gerstel.com/pdf/p-gc-an-2005-02.pdf). A further effective Deans-switch MDGC system has been developed and introduced by Shimadzu Corporation. The MDGC instrument is commercialized in the double-oven configuration, and is equipped with a quadrupole mass spectrometer. The first and second dimension capillaries are linked by using a low

dead-volume, thermally stable and chemically inert stainless steel interface. The latter is housed in the first oven, is characterized by very small dimensions (ca. 3 cm long), is connected to an auxiliary pressure source (2 ports) and to a stand-by detector. Furthermore, a fused-silica restrictor (R1 ) is fixed inside, and crosses the interface. Fig. 6 reports two schemes of the entire “Shimadzu” transfer system in the “stand-by” (Fig. 6a) and “cut” positions (Fig. 6b). Though the five-port metallic interface is located in the first GC, defined as GC1 (Fig. 6a), it is obvious that a web of external connections is necessary to create the required MDGC conditions. In both operational modes, an advanced pressure control unit (APC) supplies a gas flow at constant pressure to an external (with respect to the GC oven) fused-silica restrictor (R3 ) and to a two-way solenoid valve (V). The latter is connected to two metal branches, one with another fused-silica restrictor (R2 ) and one without: R2 produces a pressure drop, slightly higher than that generated by R3 (P2 > P3 ). In the stand-by mode (Fig. 6a), the ACP pressure is reduced on the side of the first dimension (e.g., 100 kPa – P3 ), while it reaches the second dimension branch, passing through the solenoid valve, unaltered. It is clear that through such a configuration, analytes eluting from the first (apolar) column are directed to FID1. Once the solenoid valve is activated, the transfer device passes

Fig. 8. TIC chromatogram relative to a second-dimension perfume analysis. For peak identification see Ref. [28]. Reprinted with permission from LC GC Eur. 21 (2008) 130–137. Copyright 2008 Advanstar Communications.

P.Q. Tranchida et al. / Analytica Chimica Acta 716 (2012) 66–75

to the cutting mode (Fig. 6b): the pressure on the first-dimension side of the interface remains unaltered, while the pressure on the second-dimension side becomes 100 kPa – P2 (a pressure lower than 100 kPa – P3 ). It is clear that, under such conditions, the primary-column eluate is free to reach the second (polar) capillary. The instrument is automatically controlled by using a dedicated software; the latter also enables the calculation of fundamental GC parameters, such as gas flows, linear velocities, and analyte recovery. With respect to the “Agilent” system, the main differences are that: the capillary linked to the stand-by detector does not need to be characterized by the same flow resistance as that of the secondary column (meaning that if one wants to change the second column, then one does not need to replace the restrictor also); the external design of the (3-restrictor) transfer system is a little more elaborate. However, both the commercial instruments work in an effective manner. Among a series of MDGC experiments, a 14-cut application on perfume allergens will be herein described [28]. In recent years, the relation between a series of perfumery ingredients and contact allergy, has been the subject of wide scientific discussion [29]. On the basis of European legislation (7th Amendment of the Cosmetic Directive), the 26 most frequently-recognized skin allergens must be reported on the final cosmetic product if specific concentrations are reached: 10 and 100 mg L−1 in leave-on and rinse-off products, respectively. Twenty-four compounds, out of the twenty-six, are amenable to GC analysis. Prior to an MDGC experiment it is of common use to inject a standard solution of target analytes to define the heart-cut windows. In the case of solutions containing perfume allergens, it is well known that they are characterized by a short life-time due to unstability. To circumvent such a problem a valid alternative was found: an alkane mixture was subjected to stand-by MDGC–MS analysis, prior to the injection of the allergen standard solution. Experimental linear retention indices (LRI) were derived for each compound on the primary (5% diphenyl) column; LRI cut windows were established by adding and subtracting 10 LRI units, with respect to the experimental value. Once an LRI cut window was established, the definition of a time window for each target analyte is straightforward. A total of fourteen cuts were extrapolated for 24 compounds, because the retention time difference between several allergens was only small. Through such an approach, the injection of an allergen solution prior to each MDGC application would not be necessary: it would be only necessary to derive the LRI values through preliminary hydrocarbon analysis. It is obvious that such an approach can be extended to any sample-type. A chromatogram relative to the first-dimension perfume analysis, after heart-cutting (the position of each cut is shown), is illustrated in Fig. 7. No retention-time shifts occurred during heart-cutting. The second dimension TIC MDGC-MS result is illustrated in Fig. 8. As it can be seen, 12 allergens were nicely separated from other matrix interferences. The general mass spectral purity was very good, with MS database similarities always over 90%. Each allergen was subjected to quantification; the data attained were in good agreement with the allergens reported on the perfume container. 3. Future prospects of heart-cutting MDGC It is rather curious to talk about the future of a technique, introduced over 55 years ago. However, such an evaluation is justified by the historically unstable popularity of the method. The future of heart-cutting MDGC, in the present authors’ opinion, is certainly positive. The main reasons for such a prospect, highlighted in the second section, are based on the availability of good instrumentation, on the increasing awareness of its utility and on

73

Fig. 9. Scheme of a unified MDGC instrument. Abbreviations: CT: cryotrap; V: 10port valve; 1 D and 2 D are the first and second dimension columns; UT: uncoated transfer line. Reproduced with permission from Anal. Chem. 75 (2003) 5532–5538. Copyright 2003 American Chemical Society.

the effectiveness/simplicity of the approach in specific analytical applications. In a way, the awareness of GC × GC shortcomings, for specific analysis-types (highlighted in the previous section), has contributed towards a renewed interest in classical MDGC. In terms of instrumental evolution, high levels of hardware and software quality have been achieved over the last decade, and so there does not appear to be room for excessive improvement. However, it could be imagined that attempts will be made to produce a unified GC instrument, namely one capable of achieving one-dimensional, as well as heart-cutting and comprehensive twodimensional GC analysis. As aforementioned, in 2007 the “Agilent” MDGC system was employed for both heart-cutting operations, and GC × GC work [26]. However, in that case, over 90% of the primary-column effluent was directed to waste. The construction of lab-made unified systems has been attempted and some examples will be herein discussed. The longitudinally modulated cryogenic system (LMCS) was first described as a device capable of increasing GC signal-to-noise ratios [30]. To demonstrate such a positive effect, the end part of a 30 m × 0.22 mm ID apolar column was thread through an LMCS device, that is, a moving cryotrap. The latter entrapped and re-concentrated chromatography bands through intense cooling, generated by an internal CO2 flow. After the entrapment period, the cold region of the column was exposed to the GC ambient temperature through the longitudinal movement of the trap, and the

74

P.Q. Tranchida et al. / Analytica Chimica Acta 716 (2012) 66–75

entrapped chromatography band was injected onto the final column segment. In a later experiment, a 30 m × 0.25 mm ID apolar capillary was linked to a polar 0.6 m × 0.10 ID one, with the first 5-cm of the second column thread through an LMCS [31]; the movement of the trap was activated every 7.5 s, thus enabling cryogenic comprehensive 2D GC analysis. The same research group used an LMCS device, along with a Valco 10-port microswitching valve (with 4 ports blocked), to construct a unified heart-cutting and comprehensive two-dimensional system (illustrated in Fig. 9) [32]. The valve employed was characterized by high thermal stability (max. temp.: 350 ◦ C). Four operational modes were possible using such a system, namely, one-dimensional GC, cryogenically modulated GC × GC analysis (on the entire initial sample or on specific chromatogram regions), heart-cutting MDGC, with and without cryogenic entrapment. When the valve is in the stand-by configuration (one-dimensional analysis), analytes from the primary apolar column (30 m × 0.25 mm ID) are directed to FID1 through a 1 m × 0.10 mm ID uncoated capillary segment, while a gas flow deriving from the second injector reaches the second dimension (which is thread through the LMCS), and then FID2. In the cutting configuration, analytes from the first dimension are directed to the second analytical capillary (1 m × 0.10 mm ID), while the FID1 flame is maintained by the first injector. During GC × GC analysis, the microvalve is always in the cutting position; obviously, during heart-cutting MDGC the valve undergoes switching at precise times. Furthermore, during heart-cutting the cryotrap could enable a considerable sensitivity enhancement if desired. The different analytical modes, apart from the GC × GC one, are shown in Fig. 10: a monodimensional GC analysis of a spiked lime oil sample is illustrated in A (stand-by mode), with the symbols indicating the cut positions; a monodimensional GC analysis of a pesticide standard solution is illustrated in B (stand-by mode); the second-dimension chromatogram for the 4 heart-cuts applied to the spiked sample standard solution, with use of cryotrap, is shown in the lower righthand chromatogram; expanded insets relative to the spiked sample

and to the standard solution are shown in C (the cryotrap was not activated). Although the unified in-line valve instrument proposed was a rather elaborate one, requiring a high level of expertise for satisfactory operation, the research described was of interest. It must be added, however, that a 1-m long secondary column is rather short, for heart-cutting requirements. An evolution of the previously described unified MDGC instrument has been recently reported, and defined as a “switchable multidimensional/comprehensive two-dimensional gas chromatographic analytical system” [33]. The instrument, which conserved the possibility of the four operational modes, was constructed by using an Agilent GC, equipped with a microfluidic Deans switch and an LMCS transfer device (Fig. 11). The Deans switch was connected to a first-dimension 30 m × 0.25 mm ID apolar column, to a second-dimension 30 m × 0.25 mm ID polar capillary (for heart-cutting MDGC) and to a 0.786 m × 0.10 mm ID polar column, the latter replacing the restrictor used in conventional (Agilent) MDGC analysis. It is clear that the short column segment was exploited as second dimension for GC × GC analysis. Both columns were thread through the modulator, and then each connected to an FID. The constant head pressure applied was 114.1 kPa, while that generated by the auxiliary pressure source was 103.4 kPa. Now, considering an initial analysis temperature of 50 ◦ C, the hydrogen pressure conditions would generate a first-dimension flow of nearly 0.4 mL min−1 (ca. 7 cm s−1 ), and a second-dimension flow of approximately 2.7 mL min−1 . The latter value will produce an initial linear velocity of about 63 cm s−1 and 383 cm s−1 , in the secondary conventional and micro-bore column, respectively. Such operational conditions are far from ideal for either heart-cutting or comprehensive MDGC analysis. A drawback, if it can be defined as such, is the necessity of nimble hands to pass two columns through the modulator. However, the design appeared to be more attractive compared to the former microvalveequipped prototype. Importantly, a longer secondary column was employed for MDGC analysis.

Fig. 10. A one-dimensional GC chromatogram of a spiked lime oil sample (stand-by mode), with the symbols indicating the cut positions (A); a one-dimensional GC chromatogram of a pesticide standard solution (stand-by mode) (B); the second-dimension chromatogram for the 4 heart-cuts applied to the spiked sample, with use of the cryotrap (lower right-hand chromatogram); the expanded insets relate to the main second-dimension chromatogram for the spiked sample and for the standard solution (C). For peak identification see Ref. [32]. Reproduced with permission from Anal. Chem. 75 (2003) 5532–5538. Copyright 2003 American Chemical Society.

P.Q. Tranchida et al. / Analytica Chimica Acta 716 (2012) 66–75

Fig. 11. Scheme of a unified MDGC instrument. Abbreviations: CT: cryotrap; DS: Deans switch; 1 D, 2 DL and 2 DS , are the first and second dimension (short and long) columns. Reproduced with permission from J. Chromatogr. A 1217 (2010) 1522–1529. Copyright 2010 Elsevier.

75

modulation, to connect an uncoated fused-silica segment to a monitor detector [35]. The authors defined the LMCS-based transfer system as “double cool-strand interface” and used two conventional columns in most of the described experiments. A segment of uncoated column was looped, passed through the modulator, and was linked to the first and second column. Using such a configuration, chromatography bands were entrapped twice before passage onto the second dimension. The main difference, with respect to the approach proposed by Marriott et al. [34], was the presence of a monitor detector, and thus making the definition of cutting time-windows easier. The main advantage of the valveless approach is instrumental simplicity. In a way, such systems are unified ones because they are capable of GC, classical MDGC and GC × GC analysis. As shown, heart-cutting applications can be achieved, but one must consider that there is no vent column and, hence, everything will pass from the primary to the secondary column. Such a feature increases the possibility of second-dimension coelutions and decreases the possible number of cutting events. A further disadvantage is represented by the (present-day) use of cryogenic fluids, increasing the cost-per-analysis. Although the employment of valveless systems, in current classical MDGC analysis, has very few followers, it is a route that conserves its merits. References

It can be concluded that the work carried out by the Seeley and Marriott groups provide a good basis for future research in this specific research field. Apart from the hardware, software aspects must also be taken in the highest consideration for each specific approach, for both real-time analysis and post-run data processing. A final note is devoted to valveless systems, that is the lessknown approaches towards heart-cutting MDGC analysis. In 2000, Marriott et al. used an LMCS device to achieve such an objective [34]. The authors reported the use of a primary 30 m × 0.25 mm ID apolar column, followed by either a 2 or 5 m × 0.10 mm ID mediumpolarity one, linked to the single detector (FID). The trap was cooled continuously throughout the entire analysis, but movement was activated at specific times during the analysis. A proof-of-principle experiment was carried out on a rather simple mixture of standard compounds (20 constituents); the analysis lasted just over 30 min and the trap was activated (to the release position) eight times. Groups of primary-column peaks were entrapped in the modulator and then injected onto the second dimension. It is obvious that, in such a manner, the first-dimension separation is destroyed, to a certain extent. The concept was essentially to subject simplified sub-samples, to a second-dimension analysis. The overall result reported was improved, compared to an analysis with the transfer device unactivated, considering both resolution and sensitivity. The valveless method proposed by Marriott et al. was simple and interesting, and could be used for the bidimensional analysis of target analytes in complex samples (e.g., enantiomers contained in essential oils). It is clear that a certain degree of preliminary work is necessary to determine the time windows for cutting. Such an operation would probably need to be carried out with a cold modulator, but with the column set positioned outside the trap. The reason is related to the fact that first-dimension retention times differ between applications carried with a cold and ambient-temperature modulator (the GC oven tends to contrast the cooling). After, the delay time due to retention in the second dimension should also be considered. In a system, similar to that of Marriott et al., Begnaud and Chaintreau employed a T-union (between the two dimensions) prior to

[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35]

D.J. McEwen, Anal. Chem. 36 (1964) 279–282. D.C. Fenimore, R.R. Freeman, P.R. Loy, Anal. Chem. 45 (1973) 2331–2335. M.C. Simmons, L.R. Snyder, Anal. Chem. 30 (1958) 32–35. W. Bertsch, J. High Resolut. Chromatogr. 22 (1999) 647–665. D.R. Deans, Chromatographia 1 (1968) 18–22. G. Schomburg, F. Weeke, F. Müller, M. Oreans, Chromatographia 16 (1982) 87–91. D.W. Wright, K.O. Mahler, L.B. Ballard, E. Dawes, J. Chromatogr. Sci. 24 (1986) 13–17. S.M. Sonchik, J. Chromatogr. Sci. 24 (1986) 22–25. G. Johnson, C. Marmonier, G. Garrabe, Spectra 2000 16 (1988) 23–35. L.L. Lamparski, T.J. Nestrick, D. Janson, G. Wilson, Chemosphere 20 (1990) 635–645. A. Hoffmann, R. Bremer, J.A. Rijks, Proceedings of the 15th International Symposium on Capillary Chromatography, Riva del Garda, 1993, pp. 830–836. D.R. Deans, I. Scott, Anal. Chem. 45 (1973) 1137–1141. W. Bertsch, F. Hsu, A. Zlatkis, Anal. Chem. 48 (1976) 928–931. W. Jennings, J. Chromatogr. Sci. 22 (1984) 129–135. L. Mondello, M. Catalfamo, P. Dugo, G. Dugo, J. Microcolumn Sep. 10 (1998) 203–213. Z. Liu, J.B. Phillips, J. Chromatogr. Sci. 29 (1991) 227–231. M. Adahchour, J. Beens, U.A.Th. Brinkman, J. Chromatogr. A 1186 (2008) 67–108. G. Schomburg, J. Chromatogr. A 703 (1995) 309–325. W. Bertsch, in: H.J. Cortes (Ed.), Multidimensional Chromatography Techniques and Applications, Marcel Dekker, Inc., New York, 1990, pp. 74–144. R.E. Kaiser, L. Leming, L. Blomberg, R.I. Rieder, HRC & CC 8 (1985) 92–94. T. Veriotti, M. McGuigan, R. Sacks, Anal. Chem. 73 (2001) 279–285. T. Veriotti, R. Sacks, Anal. Chem. 73 (2001) 3045–3050. T. Veriotti, R. Sacks, Anal. Chem. 75 (2003) 4211–4216. B. Quimby, J. McCurry, W. Norman, LC GC The Peak (April) (2007) 7–15. J. Luong, R. Gras, G. Yang, H. Cortes, R. Mustacich, J. Sep. Sci. 31 (2008) 3385–3394. J.V. Seeley, N.J. Micyus, S.V. Bandurski, S.K. Seeley, J.D. McCurry, Anal. Chem. 79 (2007) 1840–1847. K. Sasamoto, N. Ochiai, J. Chromatogr. A 1217 (2010) 2903–2910. L. Mondello, A. Casilli, P.Q. Tranchida, D. Sciarrone, P. Dugo, G. Dugo, LC GC Eur. 21 (2008) 130–137. S.C. Rastogi, T. Menné, J. duus Johansen, Contact Dermatitis 48 (2003) 130–132. R.M. Kinghorn, P.J. Marriott, J. High Resolut. Chromatogr. 21 (1998) 32–38. R.M. Kinghorn, P.J. Marriott, J. High Resolut. Chromatogr. 21 (1998) 620–622. P. Marriott, M. Dunn, R. Shellie, P. Morrison, Anal. Chem. 75 (2003) 5532–5538. B. Maikhunthod, P.D. Morrison, D.M. Small, P.J. Marriott, J. Chromatogr. A 1217 (2010) 1522–1529. P.J. Marriott, R.C.Y. Ong, R.M. Kinghorn, P.D. Morrison, J. Chromatogr. A 892 (2000) 15–28. F. Begnaud, A. Chaintreau, J. Chromatogr. A 1071 (2005) 13–20.