Rapid Isolation of High Solute Amounts Using an Online Four

Apr 11, 2014 - Peter Q. Tranchida , Mariarosa Maimone , Flavio A. Franchina , Thiago Rodrigues Bjerk , Cláudia Alcaraz Zini , Giorgia Purcaro , Luigi ...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/ac

Rapid Isolation of High Solute Amounts Using an Online FourDimensional Preparative System: Normal Phase-Liquid Chromatography Coupled to Methyl Siloxane−Ionic Liquid−Wax Phase Gas Chromatography Danilo Sciarrone,† Sebastiano Pantò,† Peter Quinto Tranchida,† Paola Dugo,†,‡,§ and Luigi Mondello*,†,‡,§ †

Dipartimento di Scienze del Farmaco e dei Prodotti per la Salute, Università di Messina, Viale Annunziata, 98168 Messina, Italy Chromaleont s.r.l. A Start-Up of the University of Messina, c/o University of Messina, Via Á lvaro del Portillo, 21, 98168 Messina, Italy § Centro Integrato di Ricerca (CIR), Università Campus Bio-Medico, Via Á lvaro del Portillo, 21, 00128 Roma, Italy ‡

ABSTRACT: This study reports the recent evolution of a multidimensional GC-GC-GC preparative system, now combined with an online LC preseparation step, operated under normal phase conditions. It is herein shown that the four-dimensional instrument can collect sample components with a concentration lower than 10%, in a short time period, while maintaining a high level of analyte purity. The LC dimension allows (I) the injection of higher sample amounts, compared to “direct” GC injection; (II) a polarity-based preseparation, leading to the GC injection of simplified subsamples, and thus reducing the possibility of coelutions; (III) to eliminate the essential-oil “matrix”, replacing it with the LC mobile phase (the GC system is more protected from potential contamination); (IV) the LC mobile phase is of much lower viscosity with respect to a pure, or highly concentrated essential oil, avoiding difficulties in the syringe sample withdrawal process, prior to GC injection. System optimization was performed by using standard solutions; in addition, a very complex sample, namely, vetiver essential oil, was subjected to the preparative process, with the scope of isolating two low-amount constituents (namely, α-amorphene and β-vetivone). The latter two sesquiterpenoids, which accounted for 1.7 and 4.0% of the sample (considering the volatiles), respectively, were successfully collected at the milligram level, in a one-day work period, with a purity degree in excess of 90%.

T

preparative techniques, namely, the low analyte amounts collected per analysis and/or the poor level of solute purity due to peak overlap. The first drawback can be resolved, in part, by using mega-bore GC columns (0.5−0.7 mm ID), as such tools are characterized by an increased sample capacity. However, it is well-known that mega-bore columns possess a limited efficiency, and can fully separate only samples of low complexity. In this respect, the complexity of many natural mixtures greatly exceeds the peak capacity of a mega-bore GC capillary (e.g., 10 m × 0.53 mm ID). Heart-cutting multidimensional gas chromatography (MDGC) is a highly useful option, enabling a considerable increase in the peak capacity.14 MDGC systems equipped with Deans switches, have been successfully used in prep GC applications, by different research groups.4,6,7,9,12,13 Recently, a three-dimensional GC (GC-GC-GC) preparative system, equipped with three Deans switch devices and three different stationary-phase mega-bore columns, has been described.15,16 The system purified the fractions of interest in

he research of new molecules, to be used in various academic and industrial fields (e.g., pharmaceutical, flavors, fragrances, etc.), requires adequate chromatography− mass spectrometry techniques for reliable identification purposes. Often, such an objective is not reached by using GC-MS approaches, because of a nonsufficient GC separation and/or MS identification, involving complex samples. In such a respect, preparative GC (prep GC) systems can be employed for the aim of isolating compounds, then subjected to other analytical techniques, such as NMR, FTIR, etc.1 However, the collection of a specific volatile, from a complex matrix, is often an excessive challenge for a single GC column, especially if characterized by low efficiency. During the last decades, prep GC has been exploited for the isolation of isomers in complex mixtures,2,3 of pure substances in the flavor and fragrance4−6 and environmental fields,7,8 of low-concentrated essential oil components for the assessment of organoleptic properties,9 and for the collection of pheromones.10 In many cases, a prep GC analysis precedes an 1H NMR4,7,11,12 or compound-specific radiocarbon analysis.8,13 Besides the difficulties related to the collection of GC fractions, additional issues have greatly limited the use of © 2014 American Chemical Society

Received: December 16, 2013 Accepted: April 11, 2014 Published: April 11, 2014 4295

dx.doi.org/10.1021/ac404078u | Anal. Chem. 2014, 86, 4295−4301

Analytical Chemistry

Article

under the following gradient conditions: flow rate was 1 mL/ min (reduced to 0.35 mL/min during the transfer step): from 0 to 6 min, the LC effluent was directed to waste; from 6 min to 10 min (1400 μL−100% hexane); and from 14 min to 18 min (1400 μL−100% MTBE) the LC effluent containing the hydrocarbon and oxygenated fractions, respectively, were directed to the first GC. Data were acquired by the LCsolution software (Shimadzu). 2.3. LC-GC Interface. The LC-GC transfer device consisted of a dual-side-port 25-μL syringe (CTC Analytics AG, Zwingen, Switzerland), controlled by means of a Shimadzu Model AOC5000 autosampler. Chromatography band transfer was achieved, in the stop-flow mode. The lower part of the syringe was connected, via two transfer lines, to the LC detector exit and to waste. A Teflon plug was located at the end of the syringe plunger; the latter was characterized by a lower OD, with respect to the barrel ID, thus enabling mobile phase flow inside the syringe. In the waste mode, the plug was located below both lines and the effluent was directed to waste. In the cut position, the plug was positioned between the upper and lower line, and the effluent flowed to the first GC. For more details on the syringe interface, the reader is directed to the literature.17 2.4. Multidimensional Prep GC Analysis. The preparative MDGC instrument consisted of three GC2010 systems (GC1, GC2, GC3), equipped with three Deans-switch transfer devices, namely TD1 (between the first and second columns), TD2 (between the second and third columns), and TD3 (between the third column and the collection station). The Deans switch elements, located inside the ovens, were connected to three advanced pressure control systems (APC1, APC2, and APC3), which supplied carrier gas (He) at constant pressure, also to the three FID systems. The configuration of the system has been previously described elsewhere.16 GC1 was equipped with an Optic 3 (ATAS GL International, Eindhoven, The Netherlands) large volume injector (LVI) and a flame ionization detector (FID1). The LVI temperature program and flow rate were optimized for each chemical class. LVI conditions for the hydrocarbon fraction: during the transfer step (4 min) and for the first 0.75 min of the analysis time, the split mode was used (total flow rate was 230 mL/min, at 45 °C), followed by a 1 min splitless period; afterward, the split mode was applied (126 mL/min), heating the injector to 300 °C at a rate of 15 °C/s. LVI conditions for the oxygenated fraction: during the transfer step (4 min) and for the first 0.50 min of the analysis time, the split mode was used (total flow rate was 332 mL/min at 35 °C), followed by a 1 min splitless period; afterward, the split mode was applied (126 mL/min), heating the injector to 300 °C at a rate of 15 °C/s. Column 1 was an Equity-5 [poly(5% diphenyl/95% dimethylsiloxane)] 30 m × 0.53 mm ID × 5.0 μm df (Supelco), preceded by a 1-m segment of uncoated precolumn, with the same ID. Helium was the carrier gas, having the following pressure conditions: 80 kPa for 0.75 and 0.50 min, for hydrocarbon and oxygenated compounds, respectively; then to 140 kPa at a rate of 400 kPa/min, with the pressure remaining constant afterward (initial gas linear velocity ≈ 22 cm/s). Oven temperature program: 45 °C for 1.75 min (35 °C for 1.50 min in the case of the oxygenated compounds), to 300 °C at a rate of 15 °C/min. APC1 pressure: 27.5 kPa for 0.75 min (0.50 min in the case of the oxygenated compounds); then to 125 kPa at a

a primary 5% diphenyl, a secondary poly(ethylene glycol), and on a third medium polarity ionic liquid; finally, the fraction was directed to the collection station. The novel instrument was capable of collecting components present in a concentration ranging from 10% to 30%, at the mg level, and in a short time period. However, the isolation of chemicals with a sample concentration lower than 10%, in an acceptable time period, remains a complicated task. The present research describes an evolution of the GC-GCGC preparative instrument, inasmuch that an online prepurification normal-phase (NP) HPLC step was performed (LCGC-GC-GC). In fact, an NP HPLC process, prior to a GCbased one, can be very useful for the preseparation of a sample in chemical classes, on a polarity basis.17 LC was also used to eliminate the essential-oil “matrix”, and replace it with a pure organic solvent (the LC mobile phase), of low viscosity, reducing the possibility of contamination of the GC system and difficulties in the syringe sample withdrawal process relative to the injection of large volumes of a pure, or highly concentrated essential oil. Afterward, it is possible to transfer simplified subsamples to a GC-based system, using higher sample volumes, compared to direct GC analysis. Such a concept has been applied in the present investigation, enabling the collection of “95%) in a “working-day” period. The LC optimization step was straightforward, inasmuch that the separation involved hydrocarbons and oxygenates. The optimization of the LC-GC transfer parameters was performed with the objective of enabling solvent evaporation within the GC injector, during the transfer step, and to reduce possible sample losses (low temperature conditions were applied). Tuning of the three GC processes was a compromise between the total analysis time, and the final degree of analyte purity. No optimization of the collection conditions was necessary, because of the low volatility of the target compounds.



EXPERIMENTAL SECTION Samples and Sample Preparation. Vetiver essential oil (Haitian) [Vetiveria zizanioides (L.) Nash] was provided by Firmenich SA (Genève, Switzerland). n-Hexane (GC grade), tert-butyl methyl ether, n-hexane (LC grade), C14/C16 alkanes, caryophyllene, and caryophyllene oxide were kindly provided by Supelco/Sigma−Aldrich (Bellefonte, USA). Spirogalbanone was kindly provided by L’Oreal (Aulnay-sous-Bois, France). Two stock solutions, containing 300 mg of C14 and C16 alkane (solution A) and 300 mg of caryophyllene oxide and spirogalbanone (solution B), were prepared each in 5 mL of hexane. Vetiver essential oil was diluted 1:5 (v/v) in hexane prior to injection in the LC-GC-GC-GC system. 2.2. LC Preseparation. The LC preseparation of vetiver oil was performed by using an LC system (Shimadzu, Kyoto, Japan), equipped with a Model CBM-20A communication bus module, two Model LC-20AD dual-plunger parallel-flow pumps, a Model DGU-20A online degasser, a Model SPD20A UV detector, a Model CTO-20A column oven, and a Model SIL-20AC autosampler. Five microliters (5 μL) and 50 μL of a vetiver oil solution were injected into a 250 mm × 4.6 mm ID × 5 μm dp SUPELCOSIL LC-Si column (Supelco, Milan, Italy), operated 4296

dx.doi.org/10.1021/ac404078u | Anal. Chem. 2014, 86, 4295−4301

Analytical Chemistry

Article

Figure 1. Scheme of the LC-GC-GC-GC preparative system.

rate of 400 kPa/min. Transfer line between GC1 and GC2 was maintained at 280 °C. The FID1 (330 °C) was connected via a 0.25 m × 0.18 mm ID stainless steel uncoated column to the TD1. The second column was a custom-made ionic liquid one (SLB-IL59) of the following dimensions: 30 m × 0.53 mm ID × 0.85 μm df (Supelco). Oven temperature program for the hydrocarbon fraction: from 50 °C to 100 °C (20.21 min), at a rate of 5 °C/min, then to 240 °C, at a rate of 5 °C/min. Oven temperature program for the oxygenated fraction: from 50 °C to 150 °C (23 min), at a rate of 5 °C/min, then to 240 °C, at a rate of 5 °C/min. APC2 pressure: 7.8 kPa for 0.75 min (0.50 min in the case of the oxygenated compounds) to 95 kPa at 400 kPa/min. The transfer line between GC2 and GC3 was maintained at 240 °C. The uncoated column used to connect the TD2 to the FID2 was of the same dimensions as that reported for the FID1. The third column was a Supelcowax-10 (100% poly(ethylene glycol)) of the following dimensions: 30 m × 0.53 mm ID × 2.0 μm df (Supelco). Oven temperature program for the hydrocarbon fraction: from 50 °C to 110 °C (16.96 min) at a rate of 2 °C/min, and then to 240 °C at a rate of 5 °C/min. Oven temperature program for the oxygenated fraction: from 50 °C to 150 °C (11.65 min), at a rate of 2 °C/min, then to 240 °C, at a rate of 5 °C/min. APC3 was maintained off during the LC-GC transfer step, then to 35 kPa at 400 kPa/min. The connection of the FID3 with the TD3 was realized by means of a 1 m × 0.32 mm ID uncoated column. Detector gases for FID1, FID2, and FID3 (330 °C) were as follows: H2, 50.0 mL/min; air, 400 mL/min; makeup (N2), 40.0 mL/min (sampling rate = 5 Hz). Data were collected by the MDGCsolution software (Shimadzu). The collection system was a dedicated preparative Brechbühler Prep9000 station (Brechbühler AG, Switzerland), connected to the Deans switch system (TD3) by means of a flexible heated transfer line (280 °C) containing a 1.4 m branch of uncoated column. The system was equipped with a ten-position carousel with Carbotrap C (40 mesh) adsorption tubes. The collection system was linked to a vacuum circuit, isolated by a solenoid valve. During the collection process, the valve was opened, with the effects of the vacuum enabling a more rapid and effective analyte accumulation; additionally, the condensation of high boiling components, at the conjunction point between the transfer-line end and the adsorption tube, is avoided. During

normal operation (no collection), the solenoid valve was closed. A scheme of the LC-GC-GC-GC system is reported in Figure 1. 2.5. GC-FID. A Shimadzu Model GC2010 gas chromatograph, equipped with an AOC-20i series autoinjector, was used in all applications to evaluate recovery and the degree of purity. The column was an SLB-5ms [silphenylene polymer, virtually equivalent in polarity to poly(5% diphenyl/95% methylsiloxane)] of the following dimensions: 30 m × 0.25 mm ID × 0.25 μm df (Supelco). Temperature program: from 100 °C to 300 °C, at a rate of 5.0 °C min. Injection conditions: split/splitless injector (280 °C); injection mode, split (1:100 ratio); injection volume, 0.2 μL; inlet pressure, 110 kPa; carrier gas, He (constant gas linear velocity: 30.0 cm/s). FID (310 °C) gases: same as in subsection 2.4; sampling rate, 10 Hz. Data were handled through the use of the GCsolution software (Shimadzu). 2.6. GC-MS Analysis. Samples were analyzed through electron−ionization GC-MS, on a GCMS-QP2010 Ultra system (Shimadzu). GC conditions, in terms of column type, injection mode and oven temperature program were the same as for GC-FID analysis, apart from the inlet pressure that was 30.6 kPa. Single quadrupole MS conditions: ion source temperature, 200 °C; interface temperature, 250 °C; mass scan range, 40−400 m/z. Data were handled through the use of GCMSsolution software (Shimadzu); the FFNSC 2.0 mass spectral database was used for peak identification (Shimadzu). Procedures for Calculation of LC-GC Transfer Recoveries. Normalized peak areas were used for recovery calculations. The reference peak area was determined through the splitless injection of 1 μL of the standard solutions (A and B) into GC1. The amounts of components introduced onto the column were ∼60 μg for each solution. Calculation of the Collected Amounts. Two five-point calibration curves were constructed (n = 3), at the 10, 50, 100, 250, and 500 μg mL−1 levels, using caryophyllene as representative of the hydrocarbon sesquiterpene family and spirogalbanone for the sesquiterpene oxygenated family (regression coefficients were always higher than 0.9985). The concentration levels were selected in order to cover a wide concentration range for the isolated components, diluted in ∼1 mL of hexane when flushed from the collection tube. For more details on the collection stage, the reader is directed to the literature.16 4297

dx.doi.org/10.1021/ac404078u | Anal. Chem. 2014, 86, 4295−4301

Analytical Chemistry

Article

Figure 2. GC-MS chromatogram of vetiver essential oil (peak A, α-amorphene; peak B, β-vetivone), and GC-MS chromatograms of the preseparated LC hydrocarbon (middle trace) and oxygenated sesquiterpene fractions (lower trace) obtained on an SLB-5MS 30 m × 0.25 mm ID × 0.25 μm df microbore column.



RESULTS AND DISCUSSION As previously mentioned, the scope of the present investigation was to construct a four-dimensional LC-GC-GC-GC instrument capable of collecting sample components with a concentration lower than 10%, rapidly and with a high level of purity. The LC step was operated with the aim to separate the sample in chemical classes, then injected into the GC preconcentrated. Furthermore, the introduction of semi/ nonvolatile components into GC1, due to the injection of high neat sample amounts, was reduced. With regard to the three GC stationary phases, these were selected on the basis of their different selectivities. In fact, a 5% diphenyl-based column was used in GC1, followed by two medium-polarity columns (namely, IL59 and Supelcowax-10), characterized by different selectivities.21 Initially, the LC-GC transfer and solvent evaporation steps were optimized using solutions A and B. One microliter (1 μL) of each of the two solutions were injected in the GC1 system (n = 5), in the splitless mode, to use the peak area values as a reference for recovery evaluation during the LC-GC transfer step: RSD% values were 3.3 (C14) and 6.8 (C16) for solution A (representative of sesquiterpene hydrocarbons) and 1.9 (spirogalbanone) and 5.4 (caryophyllene oxide) for solution B (representative of sesquiterpene oxygenated compounds). With regards to the optimization of the transfer/evaporation conditions, 5 μL of each solution were directed from the LC to GC1, under different conditions of LC flow, initial injector pressure/temperature, split flow and vent time, along with the injector pressure and temperature gradient. With respect to the recovery obtained during the transfer step, injector pressures of 70−140 kPa and temperatures of 35−50 °C were tested with the aim to evaporate the solvent at the lowest temperature condition in order to avoid sample loss. LC transfer flows from 1 mL/min to 0.25 mL/min, as well as different GC split flows and vent times were also tested. It was found that the most appropriate conditions for the hydrocarbon sesquiterpene components were as follows: initial injector temperature and pressure, 45 °C at 80 kPa; LC transfer flow, 350 μL/min with a split flow at 230 mL/min. Once the GC analysis was started, the split mode was used for 0.75 min,

Figure 3. Stand-by (upper traces) and heart-cut (lower traces) chromatograms, relative to the separation of the hydrocarbon sesquiterpene fraction in the first (A), second (B), and third (C) GC dimension.

4298

dx.doi.org/10.1021/ac404078u | Anal. Chem. 2014, 86, 4295−4301

Analytical Chemistry

Article

Figure 4. GC-MS chromatogram of the final α-amorphene fraction.

Figure 6. GC-MS chromatogram of the final β-vetivone fraction.

followed by the splitless mode for 1 min. During the splitless mode, the injector pressure and temperature were increased to 140 kPa at a rate of 400 kPa/min, and to 300 °C at a rate of 15 °C/s. After this step, a split flow of 126 mL/min was again used until the end of the GC analysis. With regard to the conditions for the oxygenated compounds, similar parameters were chosen after the optimization step, apart from injector temperature and pressure during the transfer step: 35 °C, split flow, 332 mL/ min. Once the GC analysis was started, the split mode was used for 0.5 min, followed by the splitless mode for 1 min. After this step, the same conditions of the sesquiterpene method were used. The initial low injector pressure conditions accelerated solvent evaporation and did not cause analyte losses;18 obviously, the reduced pressure conditions had to be extended to all three APC units, to avoid backflushing effects. Average recovery values were 81% ± 6% (C14) and 93% ± 3% (C16) for solution A, and 95% ± 5% (spirogalbanone), and 106% ± 3% (caryophyllene oxide) for solution B (n = 5). After optimization of the transfer conditions, a sample of vetiver essential oil was subjected to the preparative process. Vetiver oil is a complex mixture, composed mainly of sesquiterpenoid constituents, and it is used as an important raw material in perfumery, both as a fixative and in its own right as a fragrance ingredient.19,20 Two vetiver oil sesquiterpenes, namely, α-amorphene (hydrocarbon) and β-vetivone (oxygenated compound), were selected as target analytes, both being unavailable as standards and accounting for ∼1.7% and

Figure 5. Stand-by (upper traces) and heart-cut (lower traces) chromatograms, relative to the separation of the oxygenated sesquiterpene fraction in the first (A), second (B), and third (C) GC dimension.

4299

dx.doi.org/10.1021/ac404078u | Anal. Chem. 2014, 86, 4295−4301

Analytical Chemistry

Article

Table 1. Temperature, Pressure, and Heart-Cut Window for Each LC and GC Transfer α-amorphene

LC heart-cut, 6−10 min

1D heart-cut, 28.9−30.2 min

2D heart-cut, 46.5−46.9 min

3D heart-cut, 61.8−62.5 min

LC

LC Det l mobile phase

210 nm 100% hexane

GC1

injector temperature injector pressure transfer line oven temperature TD1 pressure

45 °C 80 kPa 280 °C 45 °C 27.5 kPa

300 140 280 300 125

GC2

transfer line oven temperature TD2 pressure

240 °C 50 °C 7.8 kPa

240 °C 100 °C (isoT) 95 kPa

GC3

transfer line oven temperature TD3 pressure β-vetivone

LC

LC Det l mobile phase

210 nm 100% MTBE

GC1

injector temperature injector pressure transfer line oven temperature TD1 pressure

35 °C 80 kPa 280 °C 35 °C 27.5 kPa

300 140 280 300 125

GC2

transfer line oven temperature TD2 pressure

240 °C 50 °C 7.8 kPa

240 °C 150 °C (isoT) 95 kPa

240 °C 240 °C (isoT) 95 kPa

240 °C 240 °C (isoT) 95 kPa

GC3

transfer line oven temperature TD3 pressure

280 °C 50 °C off

280 °C 136 °C 35 kPa

280 °C 150 °C (isoT) 35 kPa

280 °C 240 °C (isoT) 35 kPa

280 °C 50 °C off LC heart-cut 14−18 min

°C kPa °C °C kPa

300 140 280 300 125

280 °C 110 °C 35 kPa 1D heart-cut, 41.4−43.0 min

°C kPa °C °C kPa

°C kPa °C °C kPa

240 °C 184 °C 95 kPa 280 °C 110 °C (isoT) 35 kPa 2D heart-cut, 60.5−61.6 min

300 140 280 300 125

°C kPa °C °C kPa

300 140 280 300 125

°C kPa °C °C kPa

240 °C 240 °C (isoT) 95 kPa 280 °C 188 °C 35 kPa 3D heart-cut, 81.2−82.3 min

300 140 280 300 125

°C kPa °C °C kPa

of extensive coelution is evident, with the percentage of αamorphene in the hydrocarbon group reaching ∼27%. Such a percentage value does not consider possible coelutions (the same concept is valid for further values reported herein). The “heart-cut” LC-GC1-FID chromatogram (heart-cut window: 28.93−30.21 min), involving the transfer of α-amorphene to the second dimension, is also shown in Figure 3A (lower trace). The stand-by LC-GC2-FID trace (“ionic liquid” mega-bore column) of the hydrocarbon compounds is shown in Figure 3B (upper chromatogram). In such a fraction, α-amorphene is present in an amount of ∼24%. The α-amorphene peak was then transferred to the third dimension, through the application of a heart-cut window between 46.47 min and 46.96 min (lower trace in Figure 3B). As can be observed in Figure 3C, the separation on the “wax” mega-bore column enabled the separation of further interferences (upper trace). In the fraction separated on the third column, α-amorphene is present in a percentage of ∼73%. A third heart-cut window was finally applied from 61.77 min to 62.48 min (lower chromatogram in Figure 3C), to divert the α-amorphene peak to the collection station, while simultaneously activating the vacuum valve, allowing complete transfer to the adsorption tube. After collection, the tube was removed and flushed with 1 mL of hexane and analyzed by means of GC-MS and GC-FID, for qualitative and quantitative purposes, respectively.

4% of the volatile fraction, respectively. Figure 2 illustrates a comparison between the GC-MS chromatogram of vetiver essential oil (peak A, α-amorphene; peak B, β-vetivone), and the GC-MS chromatograms of the preseparated LC hydrocarbon (middle trace) and oxygenated sesquiterpene fractions (lower trace). The fractions were collected and injected off-line in the GC-MS instrument. The two chemical families were collected from 6 to 10 min (hydrocarbons, 100% hexane) and from 14 to 18 min (oxygenated compounds, 100% MTBE); the LC flow was 1 mL/min when directed to waste. As can be observed, the NP LC step performed a clear separation between the two chemical classes, thus gaining the following advantages: (I) the complexity of the sample subjected to GC separation and, hence, the possibility of coelution, are both reduced; (II) the LC preseparation was useful to limit the general overloading of the GC column, because of the direct injection of the neat sample; (III) the introduction of excessive amounts of high-boiling-point components into the GC system is avoided. Fifty microliters (50 μL) of the vetiver oil 1:5 (v/v) solution were injected into the LC system, five times more, with respect to the amount of neat oil injected in the previous GC-GC-GC applications.15,16 The stand-by LC-GC1-FID chromatogram (“5% diphenyl” mega-bore column) of the hydrocarbon fraction is shown in Figure 3A (upper trace). The presence 4300

dx.doi.org/10.1021/ac404078u | Anal. Chem. 2014, 86, 4295−4301

Analytical Chemistry

Article

The GC-FID chromatogram, after a single collection process, reported a purity degree of 90% (a GC-MS chromatogram is shown in Figure 4). After seven LC-GC-GC-GC runs, ∼1 mg of α-amorphene was obtained, with an average amount per run of ∼160 μg. Such amounts were calculated using a calibration curve constructed for the hydrocarbon sesquiterpene family (see “Calculation of the Collected Amounts” in the Experimental Section). The time required for a single application was ∼80 min: 10 min for the LC run and 70 min for the GC-GC-GC one. The same analytical approach was used for the collection of β-vetivone. The stand-by LC-GC1-FID chromatogram of the hydrocarbon fraction is shown in Figure 5A (upper trace). The presence of extensive peak overlapping is again evident, with the percentage of β-vetivone in the oxygenated compound group corresponding to ∼5%. The “heart-cut” LC-GC1-FID chromatogram (heart-cut window: 41.40−42.97 min), related to the transfer of β-vetivone to the second column, is also illustrated in Figure 5A (lower trace). The stand-by LC-GC2FID trace of the oxygenated constituents is shown in Figure 5B (upper chromatogram). After the GC2 separation, analyte purity was increased to ∼64%. The β-vetivone peak was then transferred to the third column, through a heart-cut window between 60.55 min and 61.65 min (lower chromatogram in Figure 5B). As observed in the previous application, the “wax” column enabled the separation of further interferences (upper trace in Figure 5C). In the fraction separated on the third dimension, β-vetivone is present in a percentage of ∼94%. A final heart-cut window was applied from 81.22 min to 82.25 min (lower chromatogram in Figure 5C), to divert the βvetivone peak to the collection station. After collection, the adsorption tube was again removed, flushed with hexane, and analyzed: a purity degree of 95% was determined through GC-FID analysis (a GC-MS chromatogram is shown in Figure 6). Because of the higher amounts of the target analyte in the oil, after only two LC-GC-GC-GC runs, ∼1 mg of β-vetivone was obtained. The time required for a single application was ∼108 min: 18 min for the LC run and 90 min for the GC-GC-GC one. Table 1 reports the temperatures, heart-cut windows, and pressure conditions during each transfer step.

analytes in various sample types (i.e., food, fragrance, biological, etc.)



AUTHOR INFORMATION

Corresponding Author

*Tel.: +39-090-6766536. Fax: +39-090-358220. E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The Project was funded by the “Italian Ministry for the University and Research (MIUR)” within the National Operative Project “Hi-Life Health Products from the Industry of Foods” (Project No. PON01_01499).



REFERENCES

(1) Scott, R. P. W. Preparative Gas Chromatography; Chrom-Ed Book Series; Library4Science, LLC, 2000. (2) Schurig, V.; Juza, M.; Green, B. S.; Horakh, J.; Simon, A. Angew. Chem., Int. Ed. Engl. 1996, 35, 1680−1682. (3) Ruhle, P. G.; Niere, J.; Morrison, P. D.; Jones, R.; CaradocDavies, T.; Canty, A. J.; Gardiner, M. G.; Tolhurst, V.-A.; Marriott, P. J. Anal. Chem. 2010, 82, 4501−4509. (4) Eyres, G. T.; Urban, S.; Morrison, P. D.; Dufour, J.-P.; Marriott, P. J. Anal. Chem. 2008, 80, 6293−6299. (5) Ledauphin, J.; Saint-Clair, J. F.; Lablanquie, O.; Guichard, H.; Founier, N.; Guichard, E.; Barillier, D. J. Agric. Food Chem. 2004, 52, 5124−5134. (6) Eyres, G. T.; Urban, S.; Morrison, P. D.; Marriott, P. J. J. Chromatogr. A 2008, 1215, 168−176. (7) Ruhle, C.; Eyres, G. T.; Urban, S.; Dufour, J.-P.; Morrison, P. D.; Marriott, P. J. J. Chromatogr. A 2009, 1216, 5740−5747. (8) Feng, X.; Benitez-Nelson, B. C.; Montluçon, D. B.; Prahl, F. G.; McNichol, A. P.; Xu, L.; Repeta, D. J.; Eglinton, T. I. Geochim. Cosmochim. Acta 2013, 105, 14−30. (9) Ochiai, N.; Sasamoto, K. J. Chromatogr. A 2011, 1218, 3180− 3185. (10) Nojima, S.; Kiemle, D. J.; Webster, F. X.; Apperson, C. S.; Schal, C. PLoS One 2011, 6, 1−7. (11) Kim, L.; Mitrevski, B.; Tuck, K. L.; Marriott, P. J. J. Sep. Sci. 2013, 36, 1774−1780. (12) Park, H. E.; Yang, S.-O.; Hyun, S.-H.; Park, S. J.; Choi, H.-K.; Marriott, P. J. J. Sep. Sci. 2012, 35, 416−423. (13) Ball, G. I.; Xu, L.; McNichol, A. P.; Aluwihare, L. I. J. Chromatogr. A 2012, 1220, 122−131. (14) Tranchida, P. Q.; Sciarrone, D.; Dugo, P.; Mondello, L. Anal. Chim. Acta 2012, 716, 66−75. (15) Sciarrone, D.; Pantò, S.; Rotondo, A.; Tedone, L.; Tranchida, P. Q.; Dugo, P.; Mondello, L. Anal. Chim. Acta 2013, 785, 119−125. (16) Sciarrone, D.; Pantò, S.; Ragonese, C.; Tranchida, P. Q.; Dugo, P.; Mondello, L. Anal. Chem. 2012, 84, 7092−7098. (17) Purcaro, G.; Moret, S.; Conte, L. J. Chromatogr. A 2012, 1255, 100−111. (18) Staniewski, J.; Rijks, J. A. J. Chromatogr. 1992, 623, 105−113. (19) Hyun, J. K.; Feng, C.; Xi, W.; Hau, Y. C.; Zhengyu, J. J. Agric. Food Chem. 2005, 53, 7691−7695. (20) Martinez, J.; Rosa, P. T.; Menut, C.; Leydet, A.; Brat, P.; Pallet, D.; Meireles, M. A. J. Agric. Food Chem. 2004, 52, 6578−6584. (21) Ragonese, C.; Sciarrone, D.; Tranchida, P. Q.; Dugo, P.; Dugo, G.; Mondello, L. Anal. Chem. 2011, 83, 7947−7954.



CONCLUSIONS The online four-dimensional preparative system herein proposed has shown an improved performance, compared to the three-dimensional GC-GC-GC system previously described.14,15 In fact, the instrument has proven to be an effective tool to collect milligram amounts of target essential oil compounds, contained in the sample at levels lower than 5%, in a short period of time. Specifically, for a specific constituent contained at the 4% level, only two LC-GC-GC-GC runs (108 min per collection) were necessary to isolate 1 mg (∼2.2 mg in a work day). Furthermore, the LC step is a very effective way to eliminate the essential-oil “matrix” and replace it with a pure organic solvent (the LC mobile phase) of low viscosity. In fact, the injection of large volumes of a pure, or highly concentrated, essential oil can lead to rapid contamination of the GC system and difficulties in the syringe sample withdrawal process. Even though the complexity and cost of the instrumentation are certainly high, these characteristics are justified by the analytical potential of the setup. Future research will be devoted to further instrumental improvement and to the isolation of target 4301

dx.doi.org/10.1021/ac404078u | Anal. Chem. 2014, 86, 4295−4301