Applicability of Ultra Performance Convergence Chromatography, a New Generation of Supercritical Fluid Chromatography, for the Analysis of Pesticide Residues

Monitoring and controlling wide variety of pesticide residues is a crucial challenge of food safety. In our study ultra-performance convergent chromatography (UPC2), as the new generation of supercritical fluid chromatography coupled with ESI-MS/MS system was applied to separate a set of pesticides to investigate their chromatographic behavior under various UPC2 conditions. 30 components were selected representing the GC and LC measurable components. Capacity factors obtained from LC and GC runs UPC2-PDA were compared. Based on our data UPC2 should be considered as an alternative chromatographic approach with separation mechanisms not yet fully characterized. Interestingly the type of mobile phase modifier influences the ionization in an ESI-MS system.


Introduction
From the middle of the 20 th century the increased industrialization of food processing and agriculture made pesticides more frequent in our life.Treating crops with pesticides is believed as a proper tool to maintain stable supply of agricultural products to serve the needs of the increasing world population.However by applying pesticides the number of food safety issues is also raised therefore the registration and the monitoring of these chemicals are inevitable.More than 300 pesticides are registered nowadays.[1] These days pesticides analysis is preferably carried out by using GC or LC combined with mass spectrometry, predominantly with MS/MS or TOFMS [2,3].In some cases detection is based on UV. [4] The selection of separation method depends on the volatility and the thermostability of the target components.If targeted analytes are sufficiently volatile, or can be converted to volatile derivate with chemical modification, without decomposition, GC is the preferred method for separation due to the higher achievable chromatographic resolution.If the analyte contains more polar and/or thermally unstable compounds, the application of LC separation is the appropriate choice [5].
Supercritical fluid chromatography (SFC) as an alternative chromatographic approach might also have feasibility in pesticide analyses.However the application of SFC in this field was highly ignored until these days and only limited literature is available [7].Probably this is mainly due to hardwarerelated technical limitations, among which the insufficient reproducibility of former SFC instrumentation was considered the most severe one [6].Recently appeared SFC hardware solutions from Agilent Technologies called ultraperformance supercritical fluid chromatography (UHSFC) and by Waters Company called ultra-performance convergent chromatography (UPC 2 ) are advertised attributed to be able to achieve fast and easily reproducible results.Additionally since CO 2 (as the main mobile phase) can easily and reproducibly be modified with numerous co-eluents, e.g., hexane, acetonitrile, methanol, ethyl acetate etc., it is claimed that either GC-like and/or LC-like conditions can be similarly set in this system [8].
In this study we challenged convergence chromatographic separation to a carefully selected set of analytes that similarly contained representatives of those compound-types, which are typically considered as GC or LC amenable ones [9].Our goal was to investigate some basic chromatographic properties (capacity factor and peak shape) of these selected pesticides and comparing the chromatographic results of LC and GC runs related to the same compound.
Stock solutions were prepared by dissolving an accurately weighted portion of the pesticides (approximately 10 mg powder or liquid) in 5 ml of an appropriate solvent, and they were stored at -18 °C.Working mixtures (10.0 μg mL -1 ) were obtained by further dilution with CH 3 CN and CH 3 OH individually to the UPC2 PDA detection.The selection of appropriate solvents based on the type of the applied co-solvent.Using the single compound working solutions, two 0.6 μg mL -1 multicompound mixtures were prepared.Analytes were split into two mixture solutions to exclude the simultaneous presence of isobaric compounds (e.g., fenhexamid (M+H+ :302.0) and flutriafol(M+H+ :302.1)) in the same mixture.Both mixtures contained also common compounds such as Boscalid and Piperonyl-butoxide.Two working mixtures were prepared twice either in CH 3 CN or CH 3 OH for different UPC 2 -MS experiments.In the LC-ESI-MS experiments, mixtures were dissolved in 80:20 v/v% water and the appropriate mobile phase solvent CH 3 CN and CH 3 OH were used.

UPC 2 -ESI-MS
QTRAP 3200 triple quadrupole-linear ion trap mass spectrometer (Applied Biosystems/Sciex, Foster City, CA, USA) was applied as detection system.The instrument was equipped with a Turbo V interface and Turbo Ion Spray probe (Applied Biosystems), operating in positive ion mode.The UPC 2 -ESI-MS was coupled with an Agilent 1100 HPLC system (Agilent Technologies, Waldbronn, Germany) UPC 2 flow rate 1.5 mL min -1 and the HPLC flow rate 0.2 mL min -1 passed thought the splitter and the mixture was introduced to the ESI.The HPLC pump carried 50:50 v/v% CH 3 CN : water with 0.1 v/v% HCOOH.Since the controlling of HPLC-ESI-MS and UPC 2 systems with one PC was unsolvable, the systems were driven by two separate computers.
The UPC 2 -ESI-MS coupled system was able to work only with manual injection.The gradient was the same described before.The back pressure of the system was 1500 psi.This value was set after the coupling and applied to the UPC 2 -PDA method because by MS coupling the back pressure regulator was unable to keep neither 3000 nor 2000 psi pressure previously tried to set. Figure 1 shows the UPC 2 -ESI-MS system.
Capacity factor "k" was calculated from the TR with the following formula: Where TR is the retention time, TD is dead time, except of the GC MS data obtained from literature (Waters Application) [10].The k values were used to compare the behavior of the target compounds under different chromatographic conditions.

Results and discussion
The first purpose of our research was to choose the target compounds.The selection was based on former studies [4,10,11] and special attention was paid to pick only LC-measurable components, only GC-measureable pesticides and the group that can be determined both way.The selected compounds (depicted in Fig. 2) should have met the following criteria: (i) should be extractable with the citrate buffer of QuEChERS method, ii) should have different pKo/w values and (iii) their retention times should be distributed evenly over the entire chromatographic timescale.In the case of GC-measurable components the selections were made based on the date pool of EURL and the Waters application note.In the first experiment MS method was elaborated.All pesticides were diluted to 1μg mL -1 stock solution and injected individually to the QTRAP 3200 triple quadrupole-linear ion trap mass spectrometer by a syringe.The MS parameters were optimized to these compounds.

Results observed with UPC 2 -PDA coupled system
Originally the a Waters Acquity UPC 2 was coupled with a PDA detector that can only detect the target compounds if they are present in huge quantities 10.0 μg mL -1 , which has a high detection limit that makes this detector type unable to measure real-life samples that are containing the pesticides in lower amount.Therefore it was necessary to use the different systems that could detect different compounds in lower concentration.The UPC 2 -PDA system could detect all 30 compounds except spiroxamine, because at the pH of the sample this component did not show any UV activity [12].
Molecules present in the Waters GC application as well as measurable by LC-MS were selected.They were ranked based on their capacity factors (Table 1 and 2) .These rankings were compared among the different methods, and conclusions about the behavior of the pesticides in supercritical conditions were drawn.
The comparison of capacity factors of runs with methanol (LC) and methanol as co-eluent (UPC 2 ) showed that azinphosmethyl, diazinon, prochloraz and triadimenol behaved similarly under GC and UPC 2 conditions, while chloropham followed the rules of LC chromatography in this new system while the behavior of dichlorvos and piperonil-butoxid showed similarities both that of under LC and GC conditions.X mean no any similarity to between the delaminated sytem retention queue (Table 1).The capacity factors of runs with acetonitrile (LC) and acetonitrile as co-eluent (UPC 2 ) (Table 2) were different from those obtained with methanol (Table 1).Distinctively more components showed similar elution behaviour to GC conditions.Apart from dichlorvos and piperonil-butoxide only permethrin behaved as LC measurable.On the contrary the behaviour of dimethoate piromiphos-methyl, and azinphos-methyl showed no similarity neither GC nor LC-like (Table 2)

Results acquired with UPC 2 -ESI-MS and HPLCESI-MS coupled systems
Tests with the UPC 2 -HPLC-ESI-MS system showed that the number of detectable components depended on the quality of the eluent.Out of the 30 pesticides CH 3 OH allowed the detection of 25 compounds while using CH 3 CN as eluent only 19 pesticides could be detected.
By the LC-MS method all LC-compatible compounds were detected and surprisingly two other, only GC-compatible molecules (namely folpet and alpha-endosulphan).In this method CH 3 CN was the stronger eluent, but in the hyphenated UPC 2 system proved to be less effective because less component were measurable.In the Figure 3 the chromatograms of the LC -MS system is depicted.In the case of using LC-MS with CH 3 CN as eluent the peak shapes were Gaussian with less than 0.3 min baseline peak width.Figure 4 shows the comparison of UPC 2 -ESI-MS elution profile gained by the use of CH 3 CN and CH 3 OH.It is clearly perceptible that by using CH 3 OH containing eluent more components shoved regular peak shape.However it must be mentioned the baseline with this system is not so straight.The reason of this phenomenon is the assistant solvent from the HPLC pump, which was necessary for the ionization of the compounds.Firstly We coupled the UPC 2 directly to the ESIMS system but in this case there was no ionization observed.Therefore assistant pump was used to the coupling.(Fig. 1) However the assistant solvent increased the ionization of the compunds efficiently, it also elevated baseline.To handle this problem the UPC 2 system was coupled successfully either with single quadruple or with ESI TOF MS.The application of these stable coupled systems could result a scientific breakthrough in the field of pesticide analysis.
In general by increasing the temperature the density of the CO 2 eluent decreases and so does its solvating ability [13,14].Therefore, the higher the temperature the later the components would elute from the column.On the other hand no increase was found in the elution times of dimethoate, tebufenpyrade, prochloraz, piperonyl-butoxide with the addition of CH 3 CN to the eluent.It suggest in the Fig. 5.

Conclusion
All standards examined can be detected with either PDA or MS methods therefore UPC 2 has a potential new applicability in the field of pesticide analytics.The method needs further optimization before it can be applied in routine laboratories.The separation kinetics of the UPC 2 system does not resemble that either LC or GC therefore it has to be handled as a new, different method.GC is defined by using a gas as its mobile phase and LC is defined by using liquids as its mobile phase, however UPC 2 is using both gas and liquid.This convergence of mobile phases in combination with a far greater choice of stationary phases makes UPC 2 a powerful additional choice for laboratory scientists.

Fig. 3
Fig.3 Elution profiles obtained with the LC-MS system using methanol (left) or acetonitrile (right) as eluent in 40 °C

6 Fig. 4
Fig. 4 Elution profiles from the UPC2-ESI-MS system using methanol(left) or acetonitrile (right) as eluent in 40 o C

Fig. 5
Fig. 5 Elution behaviour in the case of using ancetonitrile coelunent by the examinated UPC 2 system

Table 1
Ranced and compared capacityfactors in the case of CH 3 OH

Table 2
Ranced and compared capacityfactors in the case of CH 3 CN