Journal of Chromatography A, 1545 (2018) 67–74

Contents lists available at ScienceDirect

Journal of Chromatography A
journal homepage: www.elsevier.com/locate/chroma

Evaluation of supercritical fluid chromatography coupled to tandem

mass spectrometry for pesticide residues in food
Víctor Cutillas, María Martínez Galera, Łukasz Rajski, Amadeo R. Fernández-Alba ∗
European Union Reference Laboratory for Pesticide Residues in Fruit & Vegetables, University of Almeria, Agrifood Campus of International Excellence
(ceiA3) Department of Hydrogeology and Analytical Chemistry, Ctra. Sacramento S/N◦ , La Cañada de San Urbano, 04120, Almería, Spain

a r t i c l e i n f o a b s t r a c t

Article history: Supercritical fluid chromatography coupled to triple quadrupole mass spectrometry has been evaluated
Received 22 December 2017 for pesticide residues in food. In order to check its advantages and limitations it was developed a method
Received in revised form 17 February 2018 to identify and quantify 164 pesticides in three different matrices (tomato, orange and leek). A carbon
Accepted 22 February 2018
dioxide gradient with methanol (containing 1 mM ammonium formate) was used allowing a flow rate
Available online 23 February 2018
of 1.5 mL/min that made the total run time of 12 min without any problem of overpressure. Addition
of a post column flow 150 ␮L/min of Methanol with ammonium formate/formic acid was necessary to
Keywords:
improve the ionization.
SFC-MS/MS
Pesticides
The matrix effect study revealed that the percentages of pesticides with irrelevant matrix effect
Matrix effect (suppression lower than 20%) was 99% in tomato, 87% in orange and 62% in leek, whereas significant
Supercritical fluid chromatography suppression (higher than 50%) was not found in tomato and only 1% of the compounds in orange and 3%
in leek.These results compare favorably with that typically obtained in LC–MS/MS. The absence of water
in the mobile phase, also provided some important advantages regarding LC–MS/MS as (i) higher reten-
tion of polar compounds in the column, which elute with high sensitivity and good peak shape and (ii) a
general increase of the sensitivity of the analysis, consequence of the high ionization and ion extraction
efficiency.
Pesticides evaluated were identified following the SANTE/11813/2017. At the spiking concentration
of 5 ␮g/kg, 98% of the pesticides were identified in tomato, 98% in orange and 94% in leek, whereas for
the concentration of 10 ␮g/kg all the compounds were identified in tomato and only spiromesifen was
not identified in orange and leek. At the concentration of 20 ␮g/kg, spiromesifen was also identified in
these two matrices. The linearity and reproducibility of the method were evaluated with results which
guarantee high quality in the analytical measurements. Even though only 2 ␮L of final extract were
injected, the sensitivity of the SFC method was enough to achieve stringent LOQs.Real samples, including
6 different fruits and vegetables, were analyzed by the SFC-MS/MS proposed method, the results being
similar to those obtained by LC–MS/MS. The method was also applied to a proficiency test of fipronil in
eggs with good results in all the cases.
Carbon dioxide as mobile phase with methanol as modifier can represent a good alternative to
LC–MS/MS with reduction of matrix effects and shorter run times.
© 2018 Elsevier B.V. All rights reserved.

1. Introduction total quantity of pesticide sales in the EU-28 amounted to close to

400000 tons. Spain (19.9%), France (19.0%), Italy (16.2%), Germany
Pesticides are chemicals with an extended range of polarity (11.6%) and Poland (5.9%) were the Member States in which the
and molecular weights which are used to control weeds and pests highest quantities of pesticides were sold and, together, they made
during the crop growth and post-harvest. The use of pesticides up 72.7% of the EU-28 s pesticide sales [1].
improves the productivity and profitability of agricultural prod- As the selection and use of pesticides depends on the target crop,
ucts. According to data from the European Union (EU), in 2014, the user and country, the development of analytical methods for multi-
residue analysis of pesticides in a variety of agricultural products,
which are simple, quick and accurate, results of crucial importance
∗ Corresponding author.
for food safety and international trade.
E-mail address: [email protected] (A.R. Fernández-Alba).

https://doi.org/10.1016/j.chroma.2018.02.048
0021-9673/© 2018 Elsevier B.V. All rights reserved.
68 V. Cutillas et al. / J. Chromatogr. A 1545 (2018) 67–74

Gas and liquid chromatography (GC and LC) linked to mass of molecular weights (112.1–888.6) and polarities (logPow = −4.6
spectrometry (MS) are the most widely used techniques for the to 7.05) within 11 min. Then, the same authors [7] developed a
comprehensive multiresidue analysis of non-polar and polar pes- high-throughput screening method with a 20 min run time, using
ticides, respectively, in fruits and vegetables [2]. SFC coupled to a high-performance benchtop quadrupole Orbi-
However, highly polar pesticides need special analytical meth- trap mass spectrometry (SFC/Q Exactive) with positive/negative
ods using hydrophilic interaction chromatography (HILIC) [3,4] or polarity switching and automated library-based detection, for 444
ion-pair LC [5]. pesticides covering a wide polarity ranges (log Pow = −4.2 to 7.7)
As, in addition, the above mentioned methods are time- and wide molecular range (99.0–872.5). In this work, 373 pesticides
consuming, an important goal is the development of multiresidue were confirmed in QuEChERs spinach extracts without purification,
analytical methods to analyze simultaneously pesticides with a spiked at 10 ␮g kg−1 levels. Recently, Y. Fujito et al. [8] analyzed
wide range of polarity in a short time interval. These requirements 441 pesticides with a wide polarity range (logPow from −4.21 to
are fulfilled by the Supercritical Fluid Chromatography (SFC) [6–8] 7.7) and pKa (acidic, neutral and basic) using LC–MS/MS and SFC-
even though other columns than reversed phase are sometimes MS/MS. The comparison of both methodologies revealed that in
necessary to separate highly polar compounds as in LC–MS/MS [9]. SFC-MS/MS a higher number of compounds can be detected with
This technique shows unique capabilities and advantages; namely, good repeatability in real sample analysis.
it is considered as a technique with “liquid-like solvation power Liquid chromatography coupled to tandem mass spectrome-
and gas-like viscosity‘ and “tunable solvent’ capability, also being a try with electrospray ionization (LC–ESI–MS/MS) is the analytical
‘green alternative’ to conventional chromatography using organic technique of choice for multiresidue pesticide analysis due to its
solvents [10]. undeniable precision, robustness and high sensitivity and selectiv-
Carbon dioxide (CO2 ) is the most frequently used mobile phase ity. In addition, operation in MRM mode permits the measurement
for SFC due to their favorable properties: (i) low critical point of multiple analytes (typically more than 150) in a single run [26].
(Tc = 304.13 K, Pc = 73.773 bar), (ii) low toxicity and flammability, However, despite these advantages, an important drawback is that
(iii) carbon-neutral and widely available, (iv) health and safety related with the high presence of endogenous matrix interferents
issues and concerns over the environment and hardware damages and low efficiency during the ESI ionization stage.
and (v) capability to mix with a wide range of liquid organic solvents The goal of the present work is to evaluate application of super-
[10]. The non-polar properties of CO2 make it unsuitable for work- critical CO2 as mobile phase combined with MeOH as modifier
ing with polar analytes. However, its high miscibility with polar increasing the ionization efficiency of the ESI source and shortening
organic solvents (or co-solvents) such as methanol, ethanol, etc. the run time. The method was applied for the multiresidue analysis
and the use of additives like water, citric acid, trifluoroacetic acid, of 164 pesticides with a wide polarity range (log Pow from −0.85
isoprpylamine, triethylamine, ammonioum acetate, etc. makes SFC to 6,9) in orange, tomato and leek, by SFC coupled to the highly
a technique with particular chromatographic potential, which is sensitive and robust QqQ detection (SFC-QqQMS).
versatile enough to analytes with a wide range of polarity [11–15].
The influence of co-solvents and the role of additives on SFC
2. Experimental
have been summarized by A. Tarafder [10]. Co-solvents (i) modify
the polarity, and hence the solvating power, of the mobile phase
2.1. Reagents and materials
in such a way that its effect on solvent elution is greater than
any other factors, (b) modify the density of the mobile phase, (c)
All high-purity pesticide standards were obtained from
block active sites on the stationary phase and inhibit adsorption,
LGC (Teddington, United Kingdom), Sigma-Aldrich (Steinheim,
(d) modify stationary phase characteristic through sorption, (e)
Germany) and Riedel-de-Haën (Seelze, Germany) and were stored
the sorbed co-solvent molecules can lead to increase the net vol-
at −30 ◦ C. Individual pesticide stock solutions (1000–2000 mg/L)
ume of the stationary phase, altering the phase ratio and (f) solvat
were prepared in acetonitrile and were stored in amber screw-
selectively polar compounds in the mobile phase forming clus-
capped glass vials in the dark at −20 ◦ C. Individual standard
ters with different distribution properties. Moreover, additives act
solutions, used for the optimization, along with standard-mix
by (a) enhancement of solvating power of mobile phase, (b) ion-
solutions, used for the calibration, were prepared from the stock
pairing with charged analytes and (c) modification of stationary
standards.
phase surface properties by covering active sites, changing polarity,
The carbon dioxide was obtained from Air Liquid (Madrid,
etc.
Spain) and methanol from Fluka Analytical (Steinheim, Germany).
In addition to the above mentioned advantages, the SFC mobile
Ammonium formate, formic acid, anhydrous magnesium sul-
phase can be used with both polar and non-polar stationary phases
phate, sodium chloride, sodium hydrogenocitrate sesquihydrate
and gradients of polar solvents can also extend the availability of
and sodium citrate tribasic dihydrate were purchased from
this technique.
Sigma–Aldrich (Steinheim, Germany). PSA was supplied by Supelco
SFC has been widely applied to both chiral and achiral sep-
(Bellefonte, PA, USA).
arations in different areas: (i) enantioselective separation and
determination [16], (ii) clinical and pharmaceutical analysis [17]
and (iii) food analysis, among others [18]. 2.2. Supercritical fluid chromatography-triple quadrupole-mass
To the best of our knowledge and regarding pesticide analysis, spectrometry analysis
SFC-MS/MS has been applied to the enantioseparation and deter-
mination of flutriafol in vegetables, fruits and soil [19], dinetofuran A Nexera UC (Shimadzu Corporation, Kyoto, Japan) was used for
in bee products [20], tebuconazole in water and zebrafish [21], the SFC analysis. In addition to two binary pumps, a CO2 pump, an
isofenfos in wheat, corn and peanut [22], pyrisoxazole in cucum- autosampler and a column oven, this instrument includes a back-
ber, tomato and soil [23], propiconazole in wheat, grapes and soil pressure regulator (BPR) connected in series with the MS detector.
[24] and fenbuconazole in fruits, vegetables, cereals and soil [25]. The BPR allows to transfer the total eluate from the analytical col-
As for multiresidue methods for pesticides, M. Ishibashia et al. umn to the MS analyzer, without diffusion of peaks. The CO2 is
[6] showed the potential of SFC coupled to triple quadrupole returning to gas state in the ion source, as pressure after the back-
mass spectrometry (QqQMS) under multiple reaction monitoring pressure regulator is coming back to atmospheric one and it will be
(MRM) for the determination of 17 pesticides with a wide range going through the rough pump of the MS system.
 V. Cutillas et al. / J. Chromatogr. A 1545 (2018) 67–74 69

A modifier consisting of methanol with 1 mM Ammonium for- ments of DG SANTE/11813/2017 [27]. As the peak width of the
mate was mixed with the carbon dioxide. The Make-Up solvent compounds analyzed were similar to those obtained in reverse
used after the column was methanol containing 5 mM ammonium phase liquid chromatography (around 0.2 min), acquisition win-
formate and 0,1% formic acid. Separation was carried out on a dows of ± 0.3 min were considered enough to adequately detect
Shim-pack UC-X RP (polar functional groups between silica and C18 them, even when small variation in the pressures of the SFC occur
groups) column with length, diameter and particle size of 250 mm, (±3 bar).
2.1 mm and 3 ␮m, respectively. The column was thermostatted at
40 ◦ C. The Back Pressure module (BPR) was operated under 150 bar
2.4. Sample preparation
and 50 ◦ C. The total flow used was kept constant at 1.5 mL/min and
gradient was as follows: From 0–3 min an isocratic flow of 2% of
Tomato, orange and leek from ecological crops were bought
modifier was used, from 3 to 7 min it was increased to 24% and
in a local market in Almeria (Spain). These matrices are consid-
from 7 to 7,5 min to 40%. This 40% of modifier was maintained dur-
ered representative of the widest difficulty range expected in a
ing 2 min. Following this, the modifier percentage was changed to
laboratory focused on fruit and vegetable analysis, tomato being
2% and maintained over 3 min for a properly re-equilibration of
a high water content matrix, orange an acidic matrix and leek hav-
the system. The injection volume was 2 ␮L. The autosampler was
ing a high number of matrix components. These matrices were
thermostatted at 10 ◦ C.
extracted according to Citrate QuEChERS method and were ana-
A LCMS 8060 triple quadrupole mass spectrometer (Shimadzu
lyzed to ensure that they could be considered as blanks.
Corporation, Kyoto, Japan) equipped with an ESI source was used,
A 10 g portion of homogenized tomato, orange and leek was
employing nitrogen as the nebulizer gas. The ESI parameters
weighed in 50 mL PTFE centrifuge tube. Then, 10 mL of acetonitrile
were as follows: interface temperature, 300 ◦ C; desolvation line
were added and the samples were shaken in an automatic axial
(DL) temperature, 250 ◦ C; heat block temperature, 400 ◦ C; inter-
extractor for 4 min. Afterwards 4 g MgSO4 , 1 g NaCl, 1 g trisodium
face voltage, 4 kV; nebulizer gas flow, 3 L/min; heating gas flow,
citrate dihydrate and 0.5 g disodium hydrogen citrate sesquihy-
10 L/min; drying gas flow, 10 L/min.
drate were added and the samples were shaken again in the
automatic axial extractor for 4 min. Each extract was then cen-
2.3. Optimization of the SFC-MS/MS parameters trifuged (3700 rpm) for 5 min 5 mL volumes of the supernatants
were transferred to a 15 mL PTFE centrifuge tubes containing
The creation of the method was carried out using the Shimadzu 750 mg of MgSO4 and 125 mg of PSA and were shaken in a vortex
Extended MRM Library which allows to select transitions in a sim- for 30 s and centrifuged again (3700 rpm) for 5 min. The extracts
ple way and to import them to the method. Three transitions at were transferred to amber vials and they were acidified with 10 ␮L
the top of the sensitivity rank were selected for each pesticide. of formic acid 5% per mL of extract.
After the injection of the pesticides, the most intense transition was These extracts were spiked with 164 pesticides (all of them
selected as the quantifier transition (Target ion), while the second included in the European Union Monitoring Program) as follow:
most intense was chosen as the qualifier transition (Reference ion). 100 ␮L of each blank extract were evaporated under a gentle stream
Optimization of MRM events from precursor ion search was of nitrogen and reconstituted with 100 ␮L of acetonitrile containing
used for the optimization of the compounds not present at the the mixture of analyzed pesticides at 5,10,20,50,100 or 500 ␮g/L.
library such as internal standards. With this aim, individual solu- Then, 400 ␮L of ultrapure water were added to dilute the extract
tions of Dimethoate D6, Carbendazim D3, Malathion D10 and before injection. Thus, the injected samples contained 0.2 g of
Dichlorvos D6 at the concentration of 100 ␮g/L were injected. matrix per 1 mL of extract and their pesticide concentrations were
Retention times, MRM transitions and collision energies (CE) for the 1, 2, 4, 20 and 100 ␮g/L.
target compounds are compiled in supplementary material (Table Real samples from conventional agriculture were extracted as
S1). described above and 100 ␮L of the acidified extracts were diluted
Identification of pesticides was evaluated by retention time with with 400 ␮L of ultrapure water before injection. All the concentra-
a ± 0.1 min range and the presence of the two transitions that meet tion values which appear in this paper refer to the concentration in
the ion ratio criteria were used, following the identification require- the extracts before dilution.

Fig. 1. Chromatogram of the 164 pesticides studied in the method spiked at the concentration of 10 ␮g/Kg in orange matrix.
70 V. Cutillas et al. / J. Chromatogr. A 1545 (2018) 67–74

3. Results and discussion

3.1. Identification of compounds

The low viscosity and high diffusivity showed by the supercrit-

ical CO2 , allowed to reach the high flow rates which are customary
in SFC [6] without reaching the pressure limit of the pump. In this
case, a 1.5 mL/min flow rate was selected, which reduced the time
consuming per run, the last compound of the method (Ioxynil) elut-
ing at 7.3 min (Fig. 1). Despite this high flow rate, using a dwell time
of 6 msec for all the compounds, the highest cycle time occurred
in the retention time interval of 1.30–1.32 min and corresponded
to 0.570 s, providing more than 20 points per peak for the 76
transitions involved. Various post column flows were evaluated to
increase the sensitivity of the ionization (100,150 and 200 ␮L/min).
The best peak shapes and sensitivity was obtained with a flow of
150 ␮L/min.
This fast multiresidue analysis promoted the coelution between
matrix compounds and some analytes of interest, leading to iso-
baric interferences that modify their ion ratios showing differences
higher than 30% to the expected values (false negatives) [27]. This
situation is frequent working with matrices with high number of
interfering compounds
The overcoming of these isobaric interferences leading to “false
negatives” was accomplished taking advantage of the maximum Fig. 2. Compounds identified at the concentrations of 5 ␮g/Kg and 10 ␮g/Kg for the
cycle time used in the method (0.570 s), which provided a wide three matrices studied. The area of each bar surrounded by a dashed-line box refers
to those compounds that presented isobaric interferences and have been identified
margin to continue adding transitions without affecting the sensi-
after adding a third transition.
tivity. Thus, using the Extended MRM Library (see 2.3), additional
transitions were quickly and easily added when some target analyte
was not detected because it did not meet the ion ratio criteria. 2 pesticides with the same situation in orange (Isoprocarb and
Once an extra transition was added and considering the con- Lufenuron) meet the ion ratio and were identified. As for leek,
centration level of 5 ␮g/kg, the 2 pesticides affected by isobaric which is the matrix with the highest number of interferences
interferences in tomato (Isoprocarb and Fenhexamid) and the coeluting with the analytes [28], the number of compounds iden-

70%

62%

60%

50% 48%

40%
% of compounds

40%

35%

30%

20%

12%

10%

3%

0%
<20 20-50 >50
Signal suppression

SFC LC

Fig. 3. Comparison of the signal suppression for leek between the LC–MS/MS method published by S. Uclés et al. [28] and the SFC-MS/MS method developed in this article.
 V. Cutillas et al. / J. Chromatogr. A 1545 (2018) 67–74 71

tified using this approach increased notably, in such a way that 6 5–500 ␮g/kg in the samples). In most cases the response was lin-
pesticides affected by isobaric interferences (Epoxiconazole, Fena- ear across the whole range and only for those pesticides which
rimol, Fenthion-sulfone, Flutriafol, Phoxim and Trichlorfon) were remained unidentified at 5 or 10 ␮g/kg, the linearity corresponded
identified at 5 ␮g/kg and the other 2 pesticides which remained to a narrower range.
unidentified (Imazalil and Fenoxycarb) could be identified at Coefficients of determination equal or higher than 0.99 were
10 ␮g/kg. achieved in the corresponding linear ranges and the back-
Focusing on the unidentified compounds at 5 ␮g/kg due to its calculated concentration deviations were lower than 20%.
low sensitivity, we could observe that 4 pesticides were not identi- The reproducibility of the method was calculated by mean of five
fied at this concentration in tomato (Prothiophos, Meptyldinocap, injections of each matrix sample extract with very similar results
Fenthion and Spiromesifen). Also, in orange, another 4 compounds for all of them. For the compounds identified at 5 ␮g/kg, the 99%
could not be identified at 5 ␮g/kg (Pendimethalin, Meptyldinocap, showed RSDs lower than 20%, the 80% showed RSDs lower than
Fenthion and Spiromesifen). As for leek, the most complex 10% and the 1% showed RSDs in the range 20–25%, the latest corre-
matrix, 8 compounds were not identified at 5 ␮g/kg (Acephate, sponding to those less sensitive compounds.
Fenpropathrin, Formetanate, Kresoxim-Methyl, Dichlorvos, Dini-
conazole, Pendimethalin and Spiromesifen). All the pesticides
exposed above were identified at the concentration of 10 ␮g/kg, 3.3. Matrix effect
except for Spiromesifen in orange and leek matrices. This com-
pound was also identified at 20 ␮g/kg in both matrices. Coextracted matrix components which chromatographically co-
Fig. 2 shows the total number of identified compounds and the elute with the target analytes can affect the ionization efficiency of
percentage of compounds that were identified once the interfer- the target compounds in the MS detector. This situation, referred as
ences were overcome by adding an extra transition. As it can be “matrix effect”, leads to suppression or enhancement of the signal
observed, the number of pesticides affected was low taking into in the ion source with an increase or a decrease in the detector
account the wide scope of the method. responses, the suppression being more usual in ESI sources [29,30].
This number of unidentified pesticides was similar than the one Suppression of the signal caused by matrix effect in ESI is consid-
obtained by S.Uclés et al. [28], which carried out the evaluation of ered as a consequence of the competition for the charges available
coextracted compounds as interfering components in the analy- between the analytes and the matrix compounds that coelute,
sis of 160 pesticides in twenty five different matrices (tomato and decreasing ionization efficiency in the interface [30,31]. High ion
leek, among them) using the citrate QuEChERS method and LC- suppression can result in alterations of the sensitivity, reproducibil-
ESI–MS/MS. In this method, 5 ␮L of final extract were injected in the ity and linearity of the affected compound.
chromatographic system, whereas similar results were obtained The matrices with a high number of interfering compounds hin-
injecting only 2 ␮L in the SFC-MS/MS with the same concentration der the ability of the ionized analytes to go out of the microdroplets
of matrix (0.2 g of sample per mL). Using this injection volume in and reach the gas phase. Therefore, the fact that a matrix can cause
SFC-MS/MS we obtain a good sensitivity even when the amount of suppression depends on the nature of the compounds present; of
sample introduced in the system is reduced. (1 mg vs 0.4 mg). Sub- course, vegetables matrices such as leek contain a large number
section 3.4 discusses in more detail the possible explanation of this of compounds in the matrix and are more likely to cause ion sup-
effect. pression. For the evaluation of the matrix effect, the slopes from
the calibration curves obtained in extracts of the different matri-
ces were compared with the ones obtained from calibration curves
3.2. Linearity and repeatability built in solvent, considered as no suppression reference [32].
I this way, it can be stated that zero matrix effect take place
The linearity of the method was evaluated employing matrix- when calibration graphs built in solvent and in matrix have the
matched standards in the 1–100 ␮g/L range (corresponding to same slope. Suppression of the signal between the range 0–20%

4
logPow

3 SFC
LC
2

0
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

-1
RT(%)

Fig. 4. Graphical representation of the logPow for the compounds of the LC–MS/MS method used by S. Uclés et al. [28] and the SFC-MS/MS developed method. The information
about LogPow was acquired from EURL Datapool and Pesticides Properties Database (PPDB). The retention time is expressed as a percentage considering 100% the retention
time of the last compound to be eluted in each method.
72 V. Cutillas et al. / J. Chromatogr. A 1545 (2018) 67–74

was considered irrelevant matrix effect, between 20 and 50% low 3.4. Improvement performance by SFC
signal suppression and higher than 50% a significant suppression
of the signal. The high viscosity and surface tension of water can lead to poor
The percentage of compounds with irrelevant matrix effect ionization efficiency in ESI sources. These high values decreases the
(inside the range 0–20% of signal suppression) was 99% for tomato, solvent evaporation rate and leads to less analyte being available
87% for orange and 62% for leek, whereas the percentage of com- to reach the gas phase [33,34]. Working with SFC, methanol can be
pounds with signal suppression between 20 and 50% was 1% for the only solvent in the source. Then polarity and surface tension are
tomato, 12% for orange and 35% for leek. Significant suppression lower than those containing water, thus enhancing the ionization
was not found in tomato and only 1% of the compounds in orange efficiency. With the proposed SFC method, 56% of the compounds
and 3% in leek had an ion suppression higher than 50%. These results reach the ESI source with a solvent flow of 180 ␮L/min (including
compares favorably to the ones obtained in a previous study by 150 ␮L of the post column solvent).
S.Uclés et al. [28]. They evaluated the matrix effect in a reverse As can be seen in Fig. 4, pesticides in a reverse phase system elute
phase LC–MS/MS method for 160 pesticides in leek. The Fig. 3 shows in decreasing order of polarity but in SFC, the polarity does not play
the matrix evaluation for all the compound studied in both meth- an important role in their elution order. Typically, high polar com-
ods. pounds elute very early in reverse phase, in many cases providing
These improvements are due to the fact that the matrix com- a bad peak shape. On the contrary, in SFC they elute with retention
pounds perform differently in Supercritical fluid chromatography times above the average, without presenting these inconveniences.
with regard to LC reverse phase. The reduced initial charged- For example, Cyromazine, Pymetrozine and Methamidophos are
droplet size, which is achieved when using flow rates lower than compounds that elute at the beginning in reverse phase chromatog-
in conventional LC–MS/MS, increases the droplets surface minimiz- raphy; however, in SFC their retention times are 6.9, 5.8 and 3.5 min
ing the competition between surface-activated matrix components respectively.
and the analytes [33]. Also, the different surface tension properties Furthermore, SFC provides an increment of at least 80% in the
of methanol vs water improve the ionization efficiency (see below). areas of acidic compounds studied such as 2,4-D, Fluazifop, Haloxy-
Therefore, the analysis of matrices considered “difficult” can be fop or MCPA with regard to LC–MS/MS. The situation is showed in
carried out more efficiently in SFC-MS/MS than using LC–MS/MS as Fig. 5 that depicts the chromatographic peaks of the acidic pesti-
the suppression of the signal is lower for most compounds in the cides 2,4-D and Fluazifop at 100 ␮g/L, obtained by both techniques
matrices studied. using the same instrument which can work in both modes. These

Fig. 5. Chromatograms of 2,4-D and Fluazifop at the concentration of 100 ␮g/L in tomato extract, obtained by SFC-MS/MS and LC–MS/MS.
 V. Cutillas et al. / J. Chromatogr. A 1545 (2018) 67–74 73

results show that SFC is a good alternative for an efficiently analysis other aqueous normal-phase approaches for the liquid chromatography/mass
of some polar/acidic compounds. spectrometry-based determination of challenging polar pesticides, Anal.
Bioanal. Chem. 408 (2016) 4857–4869.
[5] J.V. Sancho, M. Ibáñez, S. Grimalt, Ó.J. Pozo, F. Hernández, Residue
3.5. Real samples determination of cyromazine and its metabolite melamine in chard samples
by ion-pair liquid chromatography coupled to electrospray tandem mass
spectrometry, Anal. Chim. Acta 530 (2005) 237–243.
12 real samples (4 watermelon, 3 tomato, 2 melon, 1 green pep- [6] M. Ishibashi, T. Ando, M. Sakai, A. Matsubara, T. Uchikata, E. Fukusaki, T.
per, 1 zuchini and 1 cucumber), provided by a food safety authority Bamba, High-throughput simultaneous analysis of pesticides by supercritical
of Andalusia, were analyzed by SFC-MS/MS and LC–MS/MS to eval- fluid chromatography/tandem mass spectrometry, J. Chromatogr. A 1266
(2012) 143–148.
uate the results obtained. An external calibration curve in blank [7] M. Ishibashi, Y. Izumi, M. Sakai, T. Ando, E. Fukusaki, T. Bamba,
tomato matrix was used to quantify the compounds detected. The High-throughput simultaneous analysis of pesticides by supercritical fluid
total number of detected pesticides was 13; pesticides were identi- chromatography coupled with high-resolution mass spectrometry, J. Agric.
Food Chem. 63 (2015) 4457–4463.
fied in 58% of the samples analyzed and 33% of the samples contain [8] Y. Fujito, Y. Hayakawa, Y. Izumi, T. Bamba, Importance of optimizing
at least 2 pesticides. The results in LC–MS/MS were similar; the chromatographic conditions and mass spectrometric parameters for
same pesticides were detected in both systems with a difference in supercritical fluid chromatography/mass spectrometry, J. Chromatogr. A 1508
(2017) 138–147.
the concentrations lower than 30% for all the cases.
[9] P. Kebarle, U.H. Verkerk, Electrospray: from ions in solution to ions in the gas
This method was also applied to a proficiency test of fipronil in phase what we know now, Mass Spectrom. Rev. 28 (2009) 898–917.
eggs [35] and the Z-scores calculated for the results were 0.2, 0.2 [10] A. Tarafder, Metamorphosis of supercritical fluid chromatography to SFC: an
Overview, TrAC Trends Anal. Chem. 81 (2016) 3–10.
and 0.3 for fipronil, fipronil sulfone and fipronil sum, respectively.
[11] L.T. Taylor, Packed column supercritical fluid chromatography of hydrophilic
Z-scores in the range of −2 to 2 are considered good results [36]. analytes via water-rich modifiers, J. Chromatogr. A 1250 (2012) 196–204.
[12] M.K. Parr, B. Wuest, E. Naegele, J.F. Joseph, M. Wenzel, A.H. Schmidt, M. Stanic,
X. de la Torre, F. Botrè, SFC-MS/MS as an orthogonal technique for improved
4. Conclusions screening of polar analytes in anti-doping control, Anal. Bioanal. Chem. 408
(2016) 6789–6797.
Low ion suppression was observed for most compounds in the [13] L. Nováková, V. Desfontaine, F. Ponzetto, R. Nicoli, M. Saugy, J.-L. Veuthey, D.
Guillarme, Fast and sensitive supercritical fluid chromatography –tandem
matrices studied including orange and leek, which are considered
mass spectrometry multi-class screening method for the determination of
complex matrices due to the large number of coextracted com- doping agents in urine, Anal. Chim. Acta 915 (2016) 102–110.
pounds that can coelute with the analytes. The low solvent flow [14] M. Ashraf-Khorassani, J. Yang, P. Rainville, M.D. Jones, K.J. Fountain, G. Isaac,
L.T. Taylor, Ultrahigh performance supercritical fluid chromatography of
applied and the 100% of organic solvent used in the SFC mobile
lipophilic compounds with application to synthetic and commercial biodiesel,
phase also improve the sensitivity of the analysis, in particular some J. Chromatogr. B 983 (984) (2015) 94–100.
polar/acidic compounds. [15] N. Riddell, B. van Bavel, I. Ericson Jogsten, R. McCrindle, A. McAlees, B. Chittim,
Supercritical fluid chromatography facilitates the high flow Coupling supercritical fluid chromatography to positive ion atmospheric
pressure ionization mass spectrometry: ionization optimization of
rates providing short analysis times (less than 8 min) without halogenated environmental contaminants, Int. J. Mass Spectrom. 421 (2017)
reaching the pressure limit of the pump. A small number of 156–163.
compounds presented isobaric interferences because the matrix [16] K. Kalíková, T. Šlechtová, J. Vozka, E. Tesařová, Supercritical fluid
chromatography as a tool for enantioselective separation: a review, Anal.
background increases when run times are reduced. Such effect, Chim. Acta 821 (2014) 1–33.
which can represent some difficulty on routine application, can be [17] E. Abbott, T.D. Veenstra, H.J. Issaq, Clinical and pharmaceutical applications of
overcome by selecting an extra transition for the affected analytes. packed-column supercritical fluid chromatography, J. Sep. Sci. 31 (2008)
1223–1230.
The developed SFC-MS/MS method allowed the identification of [18] J.L. Bernal, M.T. Martín, L. Toribio, Supercritical fluid chromatography in food
the majority of 164 target pesticides at the concentration of 5 ␮g/kg analysis, J. Chromatogr. A 1313 (2013) 24–36.
in tomato, orange and leek and for all the compounds at 10 ␮g/kg [19] Y. Tao, F. Dong, J. Xu, X. Liu, Y. Cheng, N. Liu, Z. Chen, Y. Zheng, Green and
sensitive supercritical fluid chromatographic–tandem mass spectrometric
except of Spiromesifen in the last two matrices where 20 ug/kg was
method for the separation and determination of flutriafol enantiomers in
the LOQ. vegetables, fruits, and soil, J. Agric. Food Chem. 62 (2014) 11457–11464.
[20] Z. Chen, F. Dong, S. Li, Z. Zheng, Y. Xu, J. Xu, X. Liu, Y. Zheng, Response surface
methodology for the enantioseparation of dinotefuran and its chiral
Acknowledgment metabolite in bee products and environmental samples by supercritical fluid
chromatography/tandem mass spectrometry, J. Chromatogr. A 1410 (2015)
The authors acknowledge funding support from The European 181–189.
[21] N. Liu, F. Dong, J. Xu, X. Liu, Z. Chen, Y. Tao, X. Pan, X. Chen, Y. Zheng,
Commission, DG SANTE (Document N◦ SANTE/11813/2017). The Stereoselective determination of tebuconazole in water and zebrafish by
authors also would like to thank Shimadzu Corporation for provid- supercritical fluid chromatography tandem mass spectrometry, J. Agric. Food
ing the equipment and constant support of the work. Chem. 63 (2015) 6297–6303.
[22] X. Chen, F. Dong, J. Xu, X. Liu, Z. Chen, N. Liu, Y. Zheng, Enantioseparation and
determination of isofenphos-methyl enantiomers in wheat corn, peanut and
Appendix A. Supplementary data soil with Supercritical fluid chromatography/tandem mass spectrometric
method, J. Chromatogr. B 1015–1016 (2016) 13–21.
[23] X. Pan, F. Dong, J. Xu, X. Liu, Z. Chen, Y. Zheng, Stereoselective analysis of novel
Supplementary data associated with this article can be found, chiral fungicide pyrisoxazole in cucumber, tomato and soil under different
in the online version, at https://doi.org/10.1016/j.chroma.2018.02. application methods with supercritical fluid chromatography/tandem mass
048. spectrometry, J. Hazard. Mater. 311 (2016) 115–124.
[24] Y. Cheng, Y. Zheng, F. Dong, J. Li, Y. Zhang, S. Sun, N. Li, X. Cui, Y. Wang, X. Pan,
W. Zhang, Stereoselective analysis and dissipation of propiconazole in wheat,
References grapes, and soil by supercritical fluid chromatography–tandem mass
spectrometry, J. Agric. Food Chem. 65 (2017) 234–243.
[1] e.P.s, Statistics, 2018 http://ec.europa.eu/eurostat/statistics-explained/index. [25] Y. Tao, Z. Zheng, Y. Yu, J. Xu, X. Liu, X. Wu, F. Dong, Y. Zheng, Supercritical fluid
php/Pesticide sales statistics, in. chromatography–tandem mass spectrometry-assisted methodology for rapid
[2] A. Stachniuk, E. Fornal, Liquid chromatography-mass spectrometry in the enantiomeric analysis of fenbuconazole and its chiral metabolites in fruits,
analysis of pesticide residues in food, Food Anal. Methods 9 (2016) vegetables, cereals, and soil, Food Chem. 241 (2018) 32–39.
1654–1665. [26] H.G.J. Mol, P. Zomer, M. García López, R.J. Fussell, J. Scholten, A. de Kok, A.
[3] G.P. Danezis, C.J. Anagnostopoulos, K. Liapis, M.A. Koupparis, Multi-residue Wolheim, M. Anastassiades, A. Lozano, A. Fernandez Alba, Identification in
analysis of pesticides plant hormones, veterinary drugs and mycotoxins using residue analysis based on liquid chromatography with tandem mass
HILIC chromatography –MS/MS in various food matrices, Anal. Chim. Acta 942 spectrometry: experimental evidence to update performance criteria, Anal.
(2016) 121–138. Chim. Acta 873 (2015) 1–13.
[4] A. Vass, J. Robles-Molina, P. Pérez-Ortega, B. Gilbert-López, M. Dernovics, A. [27] D. SANTE/11813/2017, http://www.eurl-pesticides.eu/docs/public/tmplt
Molina-Díaz, J.F. García-Reyes, Study of different HILIC, mixed-mode, and article.asp?CntID=727&LabID=100&Lang=EN,in.
74 V. Cutillas et al. / J. Chromatogr. A 1545 (2018) 67–74

[28] S. Uclés, A. Lozano, A. Sosa, P. Parrilla Vázquez, A. Valverde, A.R. [32] M.d.M. Gomez-Ramos, L. Rajski, A. Lozano, A.R. Fernandez-Alba, The
Fernández-Alba, Matrix interference evaluation employing GC and LC coupled evaluation of matrix effects in pesticide multi-residue methods via matrix
to triple quadrupole tandem mass spectrometry, Talanta 174 (2017) 72–81. fingerprinting using liquid chromatography electrospray high-resolution
[29] D. Remane, M.R. Meyer, D.K. Wissenbach, H.H. Maurer, Ion suppression and mass spectrometry, Anal. Methods 8 (2016) 4664–4673.
enhancement effects of co-eluting analytes in multi-analyte approaches: [33] A. Uclés Moreno, S. Herrera López, B. Reichert, A. Lozano Fernández, M.D.
systematic investigation using ultra-high-performance liquid Hernando Guil, A.R. Fernández-Alba, Microflow liquid chromatography
chromatography/mass spectrometry with atmospheric-pressure chemical coupled to mass Spectrometry—An approach to significantly increase
ionization or electrospray ionization, Rapid Commun. Mass Spectrom. 24 sensitivity, decrease matrix effects, and reduce organic solvent usage in
(2010) 3103–3108. pesticide residue analysis, Anal. Chem. 87 (2015) 1018–1025.
[30] H. Stahnke, L. Alder, Matrix effects in liquid Chromatography–Electrospray [34] R. Kostiainen, T.J. Kauppila, Effect of eluent on the ionization process in liquid
Ionization–Mass spectrometry, in: Mass Spectrometry for the Analysis of chromatography–mass spectrometry, J. Chromatogr. A 1216 (2009) 685–699.
Pesticide Residues and Their Metabolites, John Wiley & Sons, Inc, 2015, pp. [35] JRC-GEEL, https://ec.europa.eu/jrc/en/about/jrc-site/geel,in.
161–186. [36] C. Ferrer, A. Lozano, S. Uclés, A. Valverde, A.R. Fernández-Alba, European
[31] H. Trufelli, P. Palma, G. Famiglini, A. Cappiello, An overview of matrix effects Union proficiency tests for pesticide residues in fruit and vegetables from
in liquid chromatography–mass spectrometry, Mass Spectrom. Rev. 30 (2011) 2009 to 2016: overview of the results and main achievements, Food Control
491–509. 82 (2017) 101–113.