TOTAD interface: A review of its application for LVI and LC-GC

4.1 Adsorption temperature

Starting with the adsorption temperature, which is the temperature of the interface during the transfer from LC to GC or during LVI, there are two possibilities: to use high temperature or low temperature. With high temperature (60–150°C), solvent evaporation is favored. Most of the solvent evaporates, but a small quantity of solvent is eliminated as liquid, so elimination occurs in evaporative and nonevaporative modes. Although the temperature in the injector is programmed to a given value, the temperature at the point where evaporation occurs is lower, since evaporation of the solvent requires the input of heat (the heat of vaporization), which lowers the temperature at the point where evaporation occurs, that is, in the retention material inside the liner. The process is mainly evaporative, and so it is suitable for analytes of low volatility, since very volatile compounds would be removed along with the solvent. In 1999, a study was published that included, among other parameters, the influence of adsorption temperature on the sensitivity of an analysis of a standard solution of pesticides in acetonitrile by LVI using a PTV [37]. The optimum temperature for most of the pesticides tested was between 60°C and 120°C. These conditions were applied to the determination of pesticide residues in olive oil by online LC-GC using the TOTAD interface [51]. Figure 4 shows the LC and GC chromatograms obtained. The temperature of the interface body during the transfer was 100°C. High adsorption temperatures of 125°C and 150°C were also used to analyze sterols [48] and waxes [52] in edible oils, respectively. For the enantiodifferentiation of a mixture of γ-lactones (compounds of relatively low volatility), an adsorption temperature of 75°C was used [53].

Figure 4

Liquid and gas chromatogram obtained from the online LC-GC-FID analysis of (a) an olive oil sampled spiked with a mixture of pesticides at 1 mg L⁻¹ and (b) an unfortified olive oil sample. The thick line in the LC chromatogram indicates the LC fraction transferred from the liquid chromatograph to the gas chromatograph. Reprinted from ref. [51] with permission from Jennifer Diatz (Director of Publications, AOAC International).

Under high temperatures conditions, it is better to transfer or inject at a low flow rate, and in most applications, a flow rate of 0.1 mL min⁻¹ has been used [51,54]. The nature of the retention material inside the liner has little influence on the sensitivity.

However, below 60°C, the process is essentially one of absorption or adsorption depending on the packing material used inside the liner. In this case, the retention material inside the liner has a strong influence. As in the case of high temperatures, it is a partially evaporative process, although the percentage of solvent that evaporates during the transfer is relatively small. The process that takes place inside the liner is similar to a solid-phase extraction (SPE). The temperature inside the liner is lower than the set temperature because of the heat of vaporization of the solvent. These conditions are suitable for highly volatile analytes [18], which, at high temperatures, would evaporate together with the solvent. For instance, for the determination of methyl jasmonate, a compound of medium volatility, in commercial jasmine essence, the temperature of the interface was set at 50°C [55], while for the analysis of chiral volatile compounds in strawberries [56] and polychlorinated biphenyls (PCBs) in transformer oils [57], the temperature was set at 40°C. Methanol:water was used as a mobile phase in the LC step in all these determinations. Under these conditions, it is better to use higher transfer or injection flow rates. Flores et al. reported that in the analysis of methyl jasmonate, a transfer flow rate of 0.3 mL min⁻¹ provided double the peak areas of those obtained at 0.1 mL min⁻¹ [55].

4.2 Injection or transfer flow rate

Another variable to take into account is the injection (in LVI) or transfer (in online LC-GC) flow rate, which, in both online LC-GC and LVI, is controlled by means of the HPLC pump. The injection or transfer flow has a great influence on the sensitivity when working at high temperatures: the lower the injection or transfer flow is, the higher the sensitivity [58]. Solvent elimination in the evaporative mode is easier if the injection or transfer flow rate used is low as the requirements for solvent elimination are easily fulfilled if the speed of introduction is close to the solvent elimination rate [37,42,59]. Figure 5 shows the GC chromatograms obtained in the online LC-GC analysis of minor compounds (free sterols, tocopherols, squalene, erythrodiol, and uvaol) in virgin olive oil using different transfer flow rates: 0.1, 1, and 2 mL min⁻¹. As can be observed, the lowest transfer flow leads to the highest sensitivity. Of course, a lower transfer flow rate increases the transfer time and consequently the analysis time. In the LC chromatogram (Figure 5), the elution time of the minor compound fraction can be seen. Bearing in mind that the flow rate in the LC system was 2 mL min⁻¹, the volume of the fraction to be transferred was 1.6 mL, and so, at a transfer flow rate of 0.1 mL min⁻¹, it took 16 min to transfer the total volume. However, at the lower transfer rate, squalene, which occurs at high concentrations in olive oil, could not be quantified because the signal was saturated. However, since sensitivity decreased when the transfer flow was increased, squalene could be quantified by using a transfer flow of 2 mL min⁻¹ although the rest of the sterols did not appear in the GC chromatogram (Figure 5c) because of the decrease in sensitivity. The GC chromatograms of Figure 5 were obtained when 20 µL of olive oil diluted 1:50 in 2-propanol was injected. A similar GC chromatogram to that of Figure 5a was obtained when undiluted olive oil was injected and the transfer flow was increased to 2 mL min⁻¹ [58], and duration of the transfer flow rate was only 0.8 min. It is clear, then, that a higher transfer flow rate decreases the overall analysis time while maintaining the sensitivity of the analysis if the oil is not diluted. Therefore, the transfer flow rate can be used to adjust the sensitivity when the injection is carried out at high temperatures.

Figure 5

LC and GC chromatograms obtained in the online RPLC-GC-FID analysis of minor compounds (free sterols, tocopherols, squalene, and erythrodiol and uvaol) in a virgin olive oil using different transfer flows: (a) 0.1 mL min⁻¹, (b) 1 mL min⁻¹, and (c) 2 mL min⁻¹. The thick line in the LC chromatogram indicates the LC fraction transferred from the liquid chromatograph to the gas chromatograph. Reprinted with permission from ref. [58]. Copyright 2020 American Chemical Society.

4.3 Time and desorption temperature

The desorption temperature and time must be sufficient to achieve desorption and volatilization of the retained analytes. These values basically depend, therefore, on the volatility of the analytes in question. In most applications described to date, the TOTAD interface was heated rapidly and maintained at the final temperature for 5 min [54,57]. The desorption temperature varied from 200°C to 350°C depending on the volatility of the analytes. Martínez et al. applied various temperatures for different durations to accomplish the thermal desorption of the analytes (γ-lactones) retained in the interface and concluded that a temperature of 290°C held for 15 min was best for the experimental conditions [53]. The maximum temperature that the retention material can reach must always be taken into account, and it is preferable not to exceed that temperature. For instance, Tenax TA cannot be heated higher than 350°C.

4.4 Injected volume

The injected volume refers to the volume injected in the LC injector, which in the case of online LC-GC is obviously lower than the volume transferred from the LC to the GC, which is usually of hundreds of microliters or even more. The injected volume must be selected according to the required sensitivity because the sensitivity is directly proportional to this variable.

The TOTAD interface accepts a high volume of any solvent, both polar and nonpolar because the solvent is efficiently removed, as indicated by the very low amount of solvent seen in the GC chromatograms. In the case of LVI, either sample or extract can be injected into the GC, but the presence of interfering substances limits the volume that can be injected, since such interfering substances may accumulate just like the analytes at the TOTAD interface. Figure 6 shows the chromatograms obtained during the analysis of pesticide residues in a tomato sample that was spiked with pesticides and a tomato sample collected from an experimental plot. This analysis represents a case of the LVI of an extract. The high sensitivity of the technique can be appreciated. The use of a specific detector, in this case a nitrogen–phosphorus detector (NPD), provides a fairly clean chromatogram. The ability of the TOTAD interface to efficiently eliminate the solvent allows any volume to be injected, although, in the aforementioned case, 50 µL were injected because this volume provided sufficient sensitivity. The high sensitivity achieved made it possible to avoid a concentration step and therefore to reduce the sample size required. Thus, only 5 g of sample, 2 g of sodium sulfate, and 5 mL of ethyl acetate were used in the extraction step, and, after shaking, 50 µL of the extract was injected into the GC.

Figure 6

Gas chromatograms obtained in the LVI-GC-NPD analysis of pesticide residues in (a) a tomato sample fortified at 0.05 mg kg⁻¹ and (b) a real tomato sample. Reprinted with permission from ref. [64]. Copyright 2020 American Chemical Society.

In the case of the online LC-GC, the volume injected in the loop of the LC is limited by the loading capacity of the LC column. Loops of different volumes (20, 100, and 250 µL) have been used [60]. The 100 and 250 µL volume loops enabled the authors to detect more radiolytic markers due to the evident increase in sensitivity resulting from the use of higher volume loops for introducing the sample into the HPLC system. Obviously, the volume of the LC fraction transferred to the GC is much higher and depends on the “window” of the LC fraction in which the analytes elute and the LC flow rate used in the preseparation step, so the transferred volume is the volume of the LC fraction. For instance, in the analysis of pesticides in olive oil, the volume of oil injected into the liquid chromatograph was 20 µL, and the volume of the LC fraction transferred to the GC was 3.6 mL [51].

4.5 Nature and length of the packing material inside the liner

The packing material placed inside the liner is another variable that must be considered. The correct choice of this material is of great importance because it has to retain the target analytes while allowing the solvent to be eliminated. The packing material should be inert and thermostable. In the case of the LVI of samples containing water and in online RPLC-GC, a water-resistant material, such as Tenax and Carbofrit, is required [36]. Glass wool has long been used as a packing material inside the PTV liner although it is difficult to deactivate and pack the liner in a reproducible way [32]. Mol et al. evaluated several packing materials to pack liners in PTV injectors and concluded that Tenax is especially suitable for volatile analytes [33]. Any adsorbent or absorbent material can be used with the TOTAD interface. In the case of absorbent materials, the analytes are retained by dissolution, while temporary bonds are formed on the surface of an adsorbent packing material. Whatever the case, the material must be chosen bearing in mind the target analytes and the sample matrix. Tenax and polydimethylsiloxane (PDMS) and poly (50% phenyl and 50% methylsiloxane) (both absorbents) were evaluated for determining pesticide residues in olive oil [61] and terpenes in orange juice [62]. These authors recommended the use of the PDMS since it is more thermally stable. Especially at low temperatures, the nature of the retention material strongly influences sensitivity and selectivity as, in such conditions, the adsorption or absorption process predominates, while at high temperatures the packing material has little influence, since the process is, as mentioned earlier, fundamentally evaporative.

The length of the retention material inside the liner is another factor to consider. When working at high temperatures, the retention material is best if short (≤1 cm), since the decrease in pressure that occurs when the liquid and the gas flow through it is more pronounced, which favors evaporation of the solvent. Flores et al. evaluated the effect of the amount of absorbent material (OV-17 and PDMS) for pesticide residue analysis, using 80°C and 110°C during the transfer step, and concluded that better sensitivity was attained with a 1 cm length than with 2 cm of either material [61]. However, at low temperatures, it is better if the retention material is long (>1 cm); since the retention in the material is mainly an absorption or an adsorption process similar to solid-phase extraction, the greater the length of the material, the greater break-through volume.

4.6 Carrier gas flow rate

Another variable is the flow rate of helium since there are two ways for the helium to enter the interface, one whereby the gas crosses the liner and impels the liquid through the retention material inside the liner (helium enters through B) and the other whereby the helium enters through A (Figure 1), the purpose of which is to prevent solvent accumulating in the chamber located at the end of the liner. Both helium flow rates have to be considered. However, in this case, the flow rate through B has little influence on the sensitivity and the flow rate through A has no influence.

In the first analytical method developed using the TOTAD interface, a high flow rate of 1,500 mL min⁻¹ was used for both inputs to ensure the total elimination of the solvent so as to avoid flooding of the GC system. This flow rate was maintained for 0.25 min before desorption of the analytes to eliminate the remaining solvent from the glass liner as well as from the CT tube [63]. Later it was confirmed that 500 mL min⁻¹ was sufficient for solvent elimination, and the time to eliminate the remaining solvent was increased to 2 min [64]; however, Flores et al. indicated that analyte absolute areas were practically the same using 500 or 400 mL min⁻¹ [55]. In the latest analytical methods developed, a flow rate of 200 mL min⁻¹ has been found to be sufficient for the effective removal of the solvent [60]. One aspect to consider here is that lower flow rates decrease the cost of analysis. Martínez et al. applied two helium flow rates (200 and 1,000 mL min⁻¹) to evaluate the effect on the orthogonality of the system [65]. The authors found that, for the enantiodifferentiation of γ-lactones, the mismatch between the plug of liquid transporting the selected fraction and the helium flow rates applied can lead to differences in viscosity great enough to cause “viscous fingering” and strangely distorted peaks. However, by increasing the helium flow rate from 200 to 500 mL min⁻¹, well-shaped peaks were obtained. No other authors have described the occurrence of viscous fingering in online LC-GC. Reducing the helium used when a large number of samples are to be analyzed is important due to the low resources of helium and its high cost. In this respect, a modification of the TOTAD interface has been proposed that consists of coupling a vacuum system to reduce the consumption of the helium needed to remove the eluent [66]. Another alternative is to use nitrogen, which is cheaper and more abundant than helium, to remove the solvent (no differences were observed in the performance of the system using helium or nitrogen) and only use helium in the GC analysis step [67].

5.1 Nature of the LC eluent

With regard to the polarity of the mobile and stationary phases, it must be first decided whether to use NPLC or RPLC.

As indicated previously, NPLC-GC coupling is simpler, since the LC eluents used have both lower boiling points and produce lower volume of vapor per unit of liquid than the LC eluents used in RPLC, which may form a solvent film inside the wall of a capillary and are also nonsuitable for GC. Hence, most of the applications that have been developed use the normal phase [12,42,68]. Nevertheless, at least some 70–80% of the LC separations are performed in the reversed phase, and in the case of aqueous samples, for instance, the use of reversed phase is necessary. However, the transfer of polar solvent to GC is difficult because eluents have high boiling points and produce a very high volume of vapor per unit of liquid. RPLC-GC coupling is much more difficult than NPLC-GC coupling. However, the TOTAD interface allows the LC-GC coupling to be carried out in both normal and reversed phases as the solvent is completely eliminated.

Therefore, the choice of RPLC-GC or NPLC-GC coupling will depend on the analytes and the nature of the matrix. For instance, RPLC-GC coupling was used for the analysis of free sterols in edible oils using methanol:water (95:5) as eluent in LC. The free sterols, which are more polar than triglycerides, were eluted first, and the sterol fraction was automatically transferred to the gas chromatograph, thus avoiding the transfer of triglycerides, which would damage the GC systems [58]. Grob et al. have previously proposed a method to analyze free and esterified sterols by online NPLC-GC using an on-column interface, which presented this problem. To prevent triglycerides from entering into the GC column, backflush is performed after each analysis [69,70]. However, waxes, compounds of long-chain fatty acids with long-chain alcohols, are less polar than triglycerides, and, when RPLC is used, they eluted after triglycerides and the long tail of the triglyceride peaks overlap the wax ester fraction and it is difficult to transfer the wax ester fraction without a small quantity of triglycerides. In this case, it is recommended to use the normal phase in LC, so that the wax ester fraction elutes before the triglycerides and does not overlap the triglyceride peak, thus avoiding this particular problem [47]. Grob’s group reported an NPLC-GC-MS method using two different interfaces an on-column interface with concurrent solvent evaporation [71] and a loop interface [65]. Nevertheless, these methods present the problem of the variability in the GC retention time, which is not a problem with the TOTAD interface because the analytes are refocused in the packed material inside the liner [72]. Indeed, relative standard deviations of the retention time lower than 0.2 have been calculated [51,64,73].

As mentioned by Hyötyläinen and Riekkola [18], the use of gradient elution in LC is quite complicated, and it has not been used in any method developed using the TOTAD interface. However, it could be used without problem since the TOTAD interface has been demonstrated to totally eliminate any kind of solvent, polar or nonpolar, regardless of the composition of the eluent.

5.2 Characteristics of the LC column

Three parameters must be considered as regards the LC column: length, stationary phase, and diameter. The length and the stationary phases are related and basically depend on the application to be developed. It must be borne in mind that in the LC-GC coupling, the LC step is not itself an analytical process but an extraction, clean-up and concentration, and/or fractionation process, so the LC column does not have to be highly efficient. If a group of compounds is to be determined, short columns with stationary phases of low retention power should be used, so that the fraction of interest containing all the compounds of interest elutes in a small volume. For example, a C4 column of 5 cm length was used to analyze unsaponifiable compounds (free sterols, tocopherols, squalene, and erythrodiol and uvaol) in edible oils [58]. The same column was used in the multiresidue analysis of pesticides in olive oil [51], and in the LC conditions used, the volume of the pesticide fraction to be transferred was 3.6 mL.

However, when it is intended to analyze several analytes in very complex matrices, high retention power stationary phases and long columns must be used to provide the necessary high separation capacity. For instance, to determine individual pesticide residues in water, a C18 column 25 cm in length was used [72].

In the analysis of endogenous steroids in urine for the analytical control of doping, it is possible to discriminate between steroids of natural or synthetic origin based on the ¹³C/¹²C ratio by GC-IRMS (gas chromatography-isotope ratio mass spectrometry). Hence, it is essential to obtain very clean peaks in GC with no impurities that would reach the IRMS detector and modify the δ¹³C values. In addition, it is necessary to confirm the nature of the peak by GC-MS (gas chromatography–mass spectrometry). Toledano et al. developed an LC-GC-MS analysis method by using the TOTAD interface, which allowed the unequivocal identification of steroids [47]. The different testosterone metabolites were separated into different LC fractions, which were stored in a fraction collector and subsequently sent to the GC for analysis. The separation in LC of the steroids into different fractions was carried out using a 25 cm C18 column, which provided five fractions containing the nine steroids analyzed according to their concentration in the sample. This same LC column was used to discriminate between the endogenous or exogenous origin of Boldenone (Bo) and its main metabolite (BoM) in urine by LC-GC-IRMS. Bo and BoM were separated into two different fractions in LC and subsequently analyzed by GC-IRMS [74].

Another aspect of the LC column to bear in mind is the diameter. It should be remembered that in online LC-GC, the LC step acts not as an analytical separation process but as a sample preparation step, and any increase in the internal diameter of the LC column will increase sensitivity. The larger the diameter, the greater the load capacity and, therefore, the greater the sensitivity. The choice of the LC column will be a compromise between both aspects. But it must also be borne in mind that the greater the load capacity, the larger the fraction volume that must be transferred from LC to GC. In this case, since the flow rate for transferring analytes of low volatility must be low, the transfer time could be excessively long. For instance, in the analysis of pesticides in olive oil, the transferred volume was 3.6 mL; therefore, 36 min was needed for a transfer flow rate of 0.1 mL min⁻¹ [58]. In most of the analytical methods developed to date, a column of 4.6 mm internal diameter has been used for analyzing compounds occurring at trace levels in complex matrices [52,57,58].

5.3 Fraction to be transferred from LC to GC

It is of great importance to establish accurately the “window” of the LC fraction containing the analytes of interest. Although the TOTAD interface effectively removes the solvent, whether nonpolar or polar, and so there is no problem with the volume of the fraction to be transferred, the unnecessary enlargement of the window may result in introducing undesired material into the GC, thus increasing the risk of peaks overlapping and unreliable determinations [75]. The precise determination of LC fraction to be transferred into the gas chromatograph is important to obtain relatively clear GC chromatograms, which will facilitate the identification and quantification of GC peaks. Conversely, as previously mentioned, the higher the volume of LC fraction, the longer the analysis time.

The LC fraction must be determined previously by injecting high concentrations of a solution of the target analytes into the LC so that analytes can be detected using an LC detector. For example, for pesticide residue analysis in water, a standard solution of the nine target pesticides, each at a concentration of 100 mg L⁻¹, was used to determine the LC retention time and peak widths. Methanol:water (70:30) at 1 mL min⁻¹ was used as mobile phase and an UV detector was used. The beginning and end of the LC peak determined the width of the LC fraction to be transferred to the GC. The volume of the fraction to transfer depends on the composition of the eluent, and the lower the eluent strength is, the larger the volume of the LC fraction. A compromise between a satisfactory degree of separation and the volume of the fraction obtained must be considered. However, the volume to transfer from LC to GC is not a problem using the TOTAD interface, but the lower the fraction transferred, the shorter the analysis time. For the analysis of PCBs in transformer oil by online RPLC-GC, methanol:water was used as an eluent and different proportions were tested (70:30, 80:20, 90:10, and 95:5). The ratio chosen was 90:10 since it provided a satisfactory degree of separation and a low LC fraction volume. The PCBs eluted between 0.85 and 2.35 min, resulting in an LC fraction volume of 1.5 mL when the flow rate was 1 mL min⁻¹. The flow rate during the transfer step was changed to 0.2 mL min⁻¹, and so the transfer time needed was 7.5 min [57].

In an analysis of unsaponifiable compounds in edible oils, using methanol:water (70:30) at 2 mL min⁻¹ as an eluent, the volume of the LC fraction to be transferred was 18 mL, which would take at least 1.5 h to transfer into the GC at a flow rate of 0.2 mL min⁻¹ during the transfer step. To decrease the LC fraction volume, the eluent composition was changed to 95:5 and the LC flow rate to 2 mL min⁻¹, and in these conditions, the LC fraction was fixed between 0.6 and 1.4 min, which resulted in a volume of 1.6 mL. Free sterols, tocopherols, squalene, and triterpene dialcohols can be determined in only one run (Figure 5). As can be seen, the squalene peaks overloaded in the experimental conditions were used. As the LC fraction can be modified to analyze compounds covering a wide range of concentrations [47,72], the LC fraction was split into two: one fraction from 0.6 to 1.0 min, corresponding to sterols and tocopherols, and the other from 1 to 1.4 min, corresponding to squalene and triterpene dialcohols (Figure 7). The two fractions were analyzed in two separate runs. To analyze both fractions, olive oil was diluted 1:50 in propanol, but under these conditions, the squalene peaks could not be quantified correctly, so the olive oil was then diluted 1:200 and the second fraction was analyzed [58].

Figure 7

Gas chromatograms obtained from the direct RPLC-GC-FID analysis of the two fractions indicated in the LC chromatogram. Fraction (a) corresponds to sterols and tocopherols, and fraction (b) corresponds to squalene and triterpene dialcohols. Fraction (b) was analyzed twice: (b1) olive oil diluted 1:50 in 2-propanol and (b2) olive oil diluted 1:200 in 2-propanol. Reprinted with permission from ref. [58]. Copyright 2020 American Chemical Society.