Comprehensive GCxGC System
Generally, comprehensive 2D GC (GCxGC) experiments are carried out on two columns connected in series (located in a single, or independently in two ovens), using a transfer system, defined as modulator. Separations on the second column should be very fast, typically in the 4-8 s range. For such a reason, normally a short segment of micro-bore column is used in the second dimension (1-2 m x 0.10 mm ID). The primary column is usually of conventional dimensions (30 m x 0.25 mm ID). Ideally, all analytes should elute from the second column before the end of the modulation period. If elution times exceed the modulation period, then a phenomenon defined as wrap-around occurs. Second-dimension peaks are typically very narrow (e.g., 200-300 ms) and so detectors must be characterized by high acquisition rates, negligible internal volumes, and rapid rise times. The main advantages of comprehensive 2D GC, over conventional 1D GC, can be summed up in five points: I) increased separation power; II) enhanced selectivity; III) higher sensitivity; IV) speed, considering the number of peaks resolved per unit of time; V) formation of highly-organized patterns of compounds with the same functional groups (e.g., alkanes, fatty acid methyl esters, etc.).
Instrumental set-up and method development
The function of the modulator is to cut, possibly re-concentrate, and re-inject portions of the primary column effluent, onto a short column, in a sequential and continuous manner, throughout the analysis. The time required to complete such a process is defined as the modulation period. The latter parameter is very important since it must be sufficiently low to maintain the resolution achieved on the primary column, but high enough to avoid a loss of sensitivity (through excessive peak sampling) and to enable the complete elution of the compounds from the second column before the following re-injection.
The different types of developed modulators can be divided in three main classes: I) heat-based modulators; II) cryogenic modulators; and III) flow modulators. Heat-based and cryogenic modulators are part of the class of thermal modulators.
The very first modulator, developed by Liu and Phillips (1991), was a heat-based one consisting of a 15-cm segment of thick-film capillary column, divided equally into two-stages, painted with an electrically-conductive film and looped outside the oven under room temperature conditions, and heated periodically by the application of an electrical pulse.
The first commercial modulator was a heat-based one known as "thermal sweeper". It consisted of a moving metallic slotted heater, with a gap through which the capillary column was passed. The thermal sweeper is now considered as an obsolete device.
In 1998, Kinghorn and Marriott proposed the first cryogenic modulator, called the longitudinally modulated cryogenic system (LMCS). It consisted of a small CO2 cryogenic trap, characterized by a hollow-sleeve configuration, which moved upwards and downwards across a segment of column located at the head of the second dimension. The longitudinal movement enabled efficient entrapment, re-concentration and re-injection of analyte bands from the first to the second column.
One of the most popular cryogenic modulators is the loop-type modulator, a device commercialized by Zoex Corporation. Dual-stage modulation is performed by looping a segment (1-1.5 m) of capillary column (modulator tube), through the pathway of a cold jet of N2 gas (Figure 1). Though the modulator loop can be created by using the last part of the first dimension or the initial segment of the second, such options are not advisable because breakages can occur when a capillary column is coiled tightly (columns are expensive!). A better choice is to use an uncoated column, or a segment of stationary-phase coated capillary. The focusing gas, which is cooled in a heat exchange coil located in a small liquid-N2 dewar, flows continuously throughout the GCxGC analysis. The cold jet is directed vertically downward onto the modulator tube, thus generating two cold spots; the cold jet is diverted from the cold spots by a hot jet of nitrogen gas, which is activated for a brief period (e.g., 300-375 ms), in a periodic manner (e.g., every 4-6 s, corresponding to the modulation period). The hot jet is located in a perpendicular manner with respect to the cold one, and rapidly heats the cold spots, remobilizing the entrapped analytes (Figure 1).
The loop-type modulator works
essentially in the same way as the quad-jet system, using only two jets: analytes entrapped in the upstream cold spot are injected in the modulator tube by activation of the hot jet. Before the analytes reach the downstream spot, the hot jet is disactivated, a new cold spot is created, and the previously remobilized analytes are subjected to a further re-concentration step. The following activation of the hot jet will inject the entrapped chromatography band onto the second dimension. Recently, Zoex Corporation has commercialized a loop-type modulator with no liquid N2 requirements (ZX-2): cooling of the N2 gas is achieved by using a refrigeration unit, with a reported minimum temperature of -90°C, sufficient for the entrapment of C7 alkane. Such a MW limit is sufficient for almost all applications in the GC field.
A suitable detector for GCxGC must be characterized by low internal volumes to avoid band broadening of the narrow second-dimension peaks; additionally, rapid rise times and fast acquisition rates are necessary for reliable peak reconstruction. It is generally accepted that at least 10 data points per peak are required for reliable quantification.
The flame ionization detector (FID) is a "perfect" device for GCxGC analyses, if the fact that it gives no type of structural information is considered. The FID is a universal detector, with a response proportional to the compound carbon content, thus suitable for most quantitative applications. During the period 1991-2000 the FID was by far the most common GCxGC detector, used in particular in petrochemical applications.
The use of selective detectors has also been reported, in particular in trace analysis. Detectors such as the electron capture (ECD), the nitrogen-phosphorous (NPD), the sulphur chemiluminescence (SCD) and the atomic emission (AED) have proven their usefulness in a variety of experiments (e.g., pesticides in food products), even though a variety of different problems have been observed (i.e., tailing, band broadening, low acquisition frequency, etc.).
A separate discussion has to be made on mass spectrometry (MS), the most informative detection system, considered as an additional third dimension by GCxGC users. Up until 2010, the low-resolution (LR) time-of-flight (ToF) MS was considered as the most suitable device for GCxGC qualitative and quantitative analyses. The reason was related to the capability of ToF MS systems to generate full-spectrum data in a very rapid manner (up to 500 spectra/s). Furthermore, ToF MS spectral ion profiles are very consistent enabling the deconvolution of overlapping peaks which present a minimum degree of chromatography resolution. The main ToF MS drawbacks are a decrease in sensitivity, with an increase of the spectral production rate, and the production of huge data files. Therefore, in GCxGC-ToF MS experiments an acquisition rate of 50 Hz is a frequent compromise. Another disadvantage is that practically all MS spectral databases have been constructed by using single quadrupole MS (quadMS) devices, and hence often ToF MS database matching produces low-quality results.
Such substiantial drawbacks have stimulated instrumental manufacturers to increase the spectral production frequencies of single quadrupoles. In 2010, a very fast quadMS was introduced (GCMS-QP2010 Ultra) and was used for the first experiment involving quadMS quantification in the GCxGC field (Purcaro et al., 2010). The GCMS-QP2010 Ultra is characterized by a scan speed of 20,000 amu/s and a very low inter-scan delay: an acquisition rate of 50 Hz was reported using a mass range of m/z 40-340. The introduction of the GCMS-QP2010 Ultra has contributed considerably to the diffusion of GCxGC combined with mass spectrometry.
Triple quadrupole (QqQ) MS systems are highly selective and sensitive instruments, very often used in combination with a GC separation step, in pre-targeted applications. The most classical tandem MS (MS/MS) mode corresponds to multiple reaction monitoring (MRM): the first (Q1) and third (Q3) quadrupoles are both operated in the selected-ion-monitoring (SIM) mode, while collision-induced dissociation (CID) reactions occur in a collision cell located between Q1 and Q3. Specifically, Q1 isolates a specific ion (defined as precursor), from the bunch of ions produced through EI of a target analyte, in the MS source. The precursor ion is trasmitted to the cell, which is practically always defined as (second) quadrupole (q), even though nowadays it is commonly a hexa- or octapole, and in all cases with no function as mass analyzer. The precursor ion collides with an inert gas (i.e., Ar) inside the cell, undergoes further fragmentation, generating a series of product ions. Q3 isolates one of the product ions, which is then directed to the detector. Usually, two product ions are used, one as quantifier and the other as qualifier, with these deriving from the same or a different precursor ion. The MRM mode enables a drastic reduction of background noise and matrix interferences. Other MSMS modes are product ion scan (Q1 SIM – Q3 scan), neutral loss scan (Q1 scan – Q3 scan), precursor ion scan (Q1 scan – Q3 SIM), as well as classical SIM and full scan.
In 2013, a novel QqQ MS instrument (GCMS-TQ8030) was evaluated under challenging GC conditions, specifically those generated by flow-modulation (FM) GCxGC (Tranchida et al., 2013). The QqQ MS system was capable of operation under high speed conditions, in both full-scan (maximum scan speed: 20,000 amu/s) and MRM modes (up to 600 transitions/s). Apart from the speed, the QqQ MS instrument possessed a further unique property, i.e., the capability to generate full scan/MRM data, simultaneously, and also in a very fast manner. An FM GCxGC-MSMS method was used for the simultaneous untargeted analysis of essential oil compounds, and the MRM determination of targeted ones, specifically three preservatives (o-phenylphenol, butylated hydroxytoluene, butylated hydroxyanisole). The QqQ MS system generated a sufficient number of data points per peak, for both qualitative and quantitative purposes. The degree of sensitivity, reached through the MRM analysis, widely exceeded that necessary from a regulation viewpoint. The QqQ MS instrument employed is a very powerful unified instrument, inasmuch that it can perform both qualitative and quantitative analyses in a satisfactory manner. The combination of the GCxGC with the QqQ MS instrument has opened new prospects in the field of separation science.
GCxGC method optimization
As briefly described, hardware configuration is quite straightforward, while method optimization can be a cumbersome issue, which may have been limited the diffusion of GCxGC. Solid knowledge of chromatographic basic theory and experience in different branches of the GC field, such as conventional, classical multidimensional, fast micro-bore column, and high-speed mega-bore column low pressure GC, are of great support for rapid and effective GCxGC optimization. The main parameters to be considered can be very different if heat-based, cryogenic or flow modulator are considered. For reasons of space, method optimization will be herein considered for the one of the most common modulators, namely the loop-type one. In general, the main GCxGC optimization parameters are: modulation (temperatures, entrapping and re-injection period), stationary phase chemistries, capillary column dimensions, gas flow, temperature program(s), outlet pressure conditions, and the detector settings.
It is generally accepted that a temperature of circa -100°C lower than that of the GC oven is sufficient for effective entrapment of most compounds; lower temperature are needed for very volatile analytes, while the opposite is valid for components of lower volatility. Using the loop-type modulator, a trapping temperature of maximum 120-140°C lower than the first-dimension elution temperature is optimal for most applications; the use of an excessively low temperature can cause peak tailing due to incomplete re-mobilization. The hot-jet pulse should be at least 40°C higher than the elution temperatures, and the pulse duration should be a compromise between the reduction of analyte k values below a specific level and the necessity to avoid breakthrough. A pulse duration of about 300 msec is optimal for almost all applications.
The length and the diameter of the delay loop (normally 1-1.5 m) is a further important issue. If the loop is too short then breakthrough can occur during hot-jet operation, while, on the other hand, if the loop is too long, then re-focusing can be non-satisfactory or be missed completely (at the downstream cooling point). Finally, it is preferable to use a delay loop with a narrow ID (i.e., 0.10 mm) for more efficient re-injection in the second column.
The number of modulation per first-dimension peak is another important optimization, inasmuch that if one desires to preserve the resolution attained in the first column, then at least 3-4 modulations per peak must be applied. For example, if peak widths at the base are on average 20 s, then a modulation period of 5-6 s should be applied. A shorter modulation period, on the other hand, would cause excessive peak sampling and a general reduction in sensitivity.
Stationary-phase optimization is often obtained through trial-and-error testing with the main aim to maximize the amount of exploitable separation space, which is mainly achieved by choosing a proper stationary-phase combination. The separation mechanisms of the two columns must be different to obtain a so-called "orthogonal" separation. However, entirely dissimilar separation mechanisms do not exist, since analyte vapour pressures play a major role in all GC processes.
If the most popular and orthogonal combination is used, that is a non-polar column (e.g., 100% dimethylpolysiloxane, 5% diphenyl + 95% dimethylpolysiloxane) as first dimension and a polar (e.g., 100% polyethylene glycol, 50% diphenyl + 50% dimethyl polysiloxane) one in the second, the non-polar compounds are normally located in the lower parts of the 2D plot, while the more polar compounds are more retained in the second dimension. The use of a so called "reversed" set (polar x non-polar) can give interesting resulted in some specific applications. For instance, in diesel samples the different classes (alkanes, mono-aromatics, di-aromatics, etc.) group tightly together, giving an advantage when group-type determination is required. The non-orthogonal approach improves also the peak shape of polar compounds, such as aliphatic acids and alcohols, because of reduced interaction with the stationary phase in the second dimension. Therefore, the conventional and reversed set must both be considered for the determination of target and unknown compounds in complex samples.
Carrier gas linear velocities and temperature programs
When using the most popular column configuration (D1: 0.25 mm ID; D2: 0.10 mm ID), the full potential of both GC columns is hardly ever expressed, mainly due to the generation of non-ideal gas linear velocities in both dimensions. In cryogenically-modulated systems, the carrier gas velocities, in both dimensions, are dependent on the primary-column inlet pressure. The main consequence is that most GCxGC applications occur at gas velocities near-to-ideal in the first dimension (usually slightly slower), and far higher in the second dimension. A possible solution would be to reduce the inlet pressure, decreasing linear velocities both in the first and second column; such a choice would have a negative impact on first-dimension resolution, and would cause higher elution temperatures, thus reducing the benefits of the decreased linear velocity in the second column. A further option would be to use a wider-bore ID second column (0.15-0.18 mm ID), thus generating lower second-dimension linear velocities; however, the main disadvantage of such a configuration would be an overall lower separation power. A final option could be the employment of a longer secondary column to reduce linear velocities; even so, there would be an increase in the occurrence of wrap-around, which would require extended modulation periods, thus causing potential losses in first-dimension resolution.
In truth, there is no ideal solution for the optimization of gas linear velocities, as well as the temperature program(s). The following advice may be useful: when using a column configuration of dimensions 30 m x 0.25 mm ID + 1-1.5 m x 0.10 mm ID, generate a first-dimension gas velocity of approx. 20 cm/s through the application of an adequate inlet pressure. With regards to the temperature program, start with a classical gradient of 3°C/min in the single or two GC ovens (if you possess a dual oven instrument). If retention is excessive in the second dimension, causing the generation of wrap-around, then increase the temperature gradient. Such a modification will lead to higher first-dimension elution temperatures, thus decreasing retention in the second column. Alternatively, if one wishes to maintain resolution in the first dimension and possesses a dual oven system, then use a positive temperature offset in the second dimension.
On the other hand, if there is limited retention in the second dimension, accompanied by general insufficient resolution, then decrease the temperature gradient. Such an option will lead to lower primary-column elution temperatures, thus increasing retention in the second column. In alternative, if one wishes to maintain resolution in the first dimension and possesses a dual oven system, then use a negative temperature offset in the second dimension.
- Kinghorn R.M., Marriott P.J., J. High Resol. Chromatogr. 21 (1998) 32.
- Liu Z., Phillips J.B., J. Chromatogr. Sci. 29 (1991) 227.
- Purcaro G., Tranchida P.Q., Ragonese C., Dugo P., Conte L., Dugo P., Dugo G., Mondello L., Anal. Chem. 82 (2010) 8583.
- Tranchida P.Q., Franchina F.A., Zoccali M., Pantò S., Sciarrone D., Dugo P., Mondello L., J. Chromatogr. A 1278 (2013) 153.
- "Method and apparatus for measuring velocity of a chromatographic pulse."
United States Patent Number 20050139076 A1, Edward B. Ledford, Jr., Inventor