A heating-assisted liquid-liquid microextraction approach using menthol: Separation of benzoic acid in juice samples followed by HPLC-UV determination

Accepted Manuscript A heating-assisted liquid-liquid microextraction approach using menthol: Separation of benzoic acid in juice samples followed by HPLC-UV determination

Irina Timofeeva, Daria Kanashina, Dmitry Kirsanov, Andrey Bulatov PII: DOI: Reference:

S0167-7322(18)31044-4 doi:10.1016/j.molliq.2018.04.040 MOLLIQ 8943

To appear in:

Journal of Molecular Liquids

Received date: Revised date: Accepted date:

28 February 2018 5 April 2018 8 April 2018

Please cite this article as: Irina Timofeeva, Daria Kanashina, Dmitry Kirsanov, Andrey Bulatov , A heating-assisted liquid-liquid microextraction approach using menthol: Separation of benzoic acid in juice samples followed by HPLC-UV determination. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Molliq(2017), doi:10.1016/j.molliq.2018.04.040

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT A heating-assisted liquid-liquid microextraction approach using menthol: Separation of benzoic acid in juice samples followed by HPLC-UV determination Irina Timofeeva*, Daria Kanashina, Dmitry Kirsanov, Andrey Bulatov Department of Analytical Chemistry, Institute of Chemistry, Saint-Petersburg University, St. Petersburg State University, SPbSU, SPbU, 7/9 Universitetskaya nab., St. Petersburg, 199034

* Corresponding author. Tel.: +7 952 2021787.

RI

E-mail address: [email protected] (Irina Timofeeva)

PT

Russia

SC

Abstract

A heating-assisted liquid-liquid microextraction (HA-LLME) procedure was developed

NU

as a new approach for pretreatment of complex sample matrix. The procedure was applied for HPLC-UV determination of preservative (benzoic acid) in juices. Menthol was investigated as an extractant for the HA-LLME. The procedure consists in heating of an aqueous sample in a

MA

polymeric vial with solid-phase menthol located at the bottom of extraction vial. The heating promotes the melting of menthol and its dispersion in a sample phase. The extractant drops are moving from the bottom to the surface of the sample, forming a liquid organic phase containing

D

target analyte. Thus separation is taken place without centrifugation. It was found that menthol

PT E

provides for microextraction of benzoic acid from fruit and berry juice samples with recovery from 93 to 117 %. The conditions of the HA-LLME were optimized using factorial experimental designs. Under optimal experimental conditions the linear detection range was found to be 0.5 –

CE

50 mg L-1 with LOD at 0.15 mg L-1. The HA-LLME with menthol allowed for significant improvement of the selectivity of benzoic acid determination. The advantages of the HA-LLME

AC

are the simplicity, low cost, and using of environmentally friendly extractant.

Keywords:

heating-assisted

liquid-liquid

microextraction;

high

performance

liquid

chromatography with ultraviolet detection; benzoic acid; juice

Introduction Benzoic acid and its salts are widely used as food preservatives and known as E-numbers E210, E211, E212, and E213 [1-3]. The efficiency of benzoic acid and benzoates depends on the pH values of particular food samples. Using these preservatives in food products with a slightly

1

ACCEPTED MANUSCRIPT acidic (above 5) or neutral pH is ineffective. Such acidic beverage as fruit juices, sparkling drinks, soft drinks, and pickles are well preserved with benzoic acid and benzoates [4]. Benzoic acid and its salts are permitted food additives according to international laws, but their content must be declared on the package and should not exceed the established limits, because their excessive use can lead to metabolic acidosis, convulsions, and hyperpnoea in humans [5]. European Union Directive 95/2/CE has established the maximal content for benzoic acid at 150 mg L-1 (0.015 %). Therefore, monitoring of benzoic acid’s content in foods and

PT

drinks is an important task for analytical chemistry. Especially strict control should be performed for children's nutrition.

RI

А variety of analytical techniques (Table 1) based on high-performance liquid chromatography-diode array detection and ultraviolet (UV)-spectrophotometry [4, 6-8], gas-

SC

chromatography with flame ionization detector [9], supercritical fluid chromatography [10], spectrophotometry [11] and capillary electrophoresis with UV and conductivity detection [12,

NU

13] have been developed for the determination of benzoic acid in food samples. Taking into account the complexity of foods matrixes, the developed methods include sample preparation

MA

procedures based on various separation methods such as liquid-liquid extraction [8, 13], salt and air-assisted homogeneous liquid-liquid extraction [6], dispersive liquid-liquid microextraction (DLLME) [7], solid phase microextraction [14, 15], silica gel thin-film microextraction [16] and

D

micro dialysis [11] (Table 1).

PT E

Development of a faster, simpler, less expensive and environmentally-friendly sample preparation techniques is very important in modern analytical practice [17]. Recently menthol has been proposed as an alternative to conventional organic solvents in the DLLME due to its

CE

low melting point (40 ºC), low vapor pressure and its high viscosity, which allows to carry out the green and reproducible extracting process [18]. In accordance with the reported DLLME

AC

procedure menthol was dissolved in a disperser solvent (acetone) and injected into an aqueous sample solution. After phase separation menthol phase containing the analytes (phthalate esters) was solidified by contact with air and formed a single solid piece. The developed procedure is simple and convenient for sample preparation without centrifugation. However, the DLLME has some limitations. The disperser organic solvent used in the DLLME procedure can increase solubility of the hydrophobic analytes in an aqueous phase reducing the extraction efficiency [19]. Moreover, relatively high volume of acetone (1.25 mL) is required to dissolve menthol in disperser organic solvent resulting in a high waste generation. In this research a new approach for the pretreatment of complex sample matrix, a heatingassisted liquid-liquid microextraction (HA-LLME) procedure using menthol, is presented. To the best of our knowledge for the first time it was found that menthol provides for fast and efficient 2

ACCEPTED MANUSCRIPT extraction of benzoic acid from fruit and berry juice samples. The proposed HA-LLME procedure consists in the heating of an aqueous sample in a polymeric vial with solid-phase menthol located at the bottom of extraction vial. The heating leads to the melting of menthol and its subsequent dispersion in a sample phase. The extractant drops are moving from the bottom to the surface of the sample, forming a liquid organic phase containing target analyte. Thus separation is taken place without centrifugation. In order to make the microextraction process as rapid and easy as possible, the HA-LLME was carried out in one step. The HA-LLME procedure

PT

was successfully used for the HPLC-UV determination of preservative (benzoic acid) in juices.

RI

2. Experimental

SC

2.1. Reagents and solutions

Benzoic acid was purchased from Sigma-Aldrich (USA). Stock solution of 1 g L-1 of

NU

benzoic acid was prepared by dissolving an appropriate amount of substance in 0.5 mL methanol, and then it was diluted to 25 mL with water and stored in the dark at +4 °C. Working

MA

solutions were prepared right before the experiments by dilution of the stock solution with water and adjusting the pH value to the required level with sulfuric acid (Sigma-Aldrich, USA). Menthol was purchased from Vekton (Russia). Ultra pure water from Millipore Milli-Q RG

D

(Millipore, California, USA) was used for all the experiments. All chemicals were of analytical

PT E

reagent grade.

2.2. Eхtraction vials preparation

CE

Eхtraction vials with solid menthol were prepared by injecting 100 µL of molten menthol (45 ºС) onto the bottom of polypropylene vials (2.5 mL), which were then closed and stored at

AC

room temperature. 2.3. Samples

Apple, multifruit, orange, cherry and berry baby juices from different brands produced by domestic companies were purchased in local supermarkets (Saint Petersburg, Russia) and employed for further analysis. 2.4. HA-LLME procedure 2 mL of juice sample was placed into the extraction vial prepared as described in 2.2 and then the vial was placed into the thermostatic bath (LT-116a, LOIP, Russia) at 45 ºС for 3 min 3

ACCEPTED MANUSCRIPT for the extraction and phase separation (Fig. 1). After that 20 µL of the top extraction solvent phase (molten menthol) containing analyte were mixed with methanol (1:1) (to exclude the solidification of menthol in the measuring system) and then analyzed by HPLC-UV.

2.5. Reference extraction procedure The reference extraction method was performed according to [7] with minor modification. Briefly, an aliquot of 5 mL of juice sample (after 5 times dilution with water, adjusting the pH to

PT

3 and adding 1 g of NaCl) were placed into a 10 mL vial. The solution consisting of 500 µL of ethanol (disperser solvent) and 100 µL of 1-octanol (extraction solvent) was rapidly injected into

RI

the extraction vial. The mixture was thoroughly shaken using a flat shaker for 2 min, and then centrifuged for 10 min at 4000 rpm. After that, 20 µL of the organic phase was injected directly

SC

into the HPLC-UV system.

NU

2.6. Chromatographic conditions

Analytical separation was carried out using LC-20 Prominence HPLC system (Shimadzu,

MA

Japan) with Supelco C18 column (250 mm × 4.6 mm, 5 µm) at 43 °C. UV-vis spectrophotometer set at 230 nm was used as the detector, and the volume of the injected sample was 20 µL. The mobile phase was methanol – ammonium acetate buffer (0.05 M) (30:70, v/v); the flow rate was

PT E

D

0.5 mL min-1.

2.7. Optimization of HA-LLME conditions with fractional factorial design In order to optimize the conditions of microextraction experiment fractional factorial

CE

design was applied [20]. The concentration of benzoic acid in aqueous phase was 25 mg L-1 for all experiments. The following parameters were considered to be important for the extraction

AC

efficacy: temperature, pH, the volumes of extractant and sample, presence/absence of the saltingout reagent. With this comparatively large number of factors full factorial design would require rather big number of experiments (32), thus fractional factorial design (1/2) for five factors at two levels (25-1 design) was considered. The design matrix contained 16 rows corresponding to 16 different combinations of parameters, and 16 columns, containing one column for b0, five columns for individual factors and 10 columns for interaction terms. The complete design matrix with encoded factors can be found in Supplementary material (SM Table 1). Experimental levels for studied parameters are presented in Table 2. The lower level of temperature (45 0C) was chosen to ensure the melting of menthol, the upper one (70 0C) was limited by the capabilities of the thermostatic bath. The pH range was chosen to be in acidic 4

ACCEPTED MANUSCRIPT media, because in alkaline solution benzoic acid is ionized with formation of water-soluble form. The salting-out effect can increase extraction efficiency and it was studied at two categorical levels. The lower volume of the extractant was the minimal one suitable for convenient handling, the upper one – to achieve required sensitivity. The sample volume was chosen in this particular range to ensure reasonable extraction efficiency.

3. Results and discussion

PT

3.1. Preliminary studies

RI

Sample preparation based on HA-LLME prior to the HPLC-UV detection was chosen to remove the juice sample matrix in order to minimize the risk of blocking the column of the

SC

HPLC-UV system with matrix components and impairing the chromatographic performance. The developed HA-LLME procedure involves several processes: the thermostating of a sample

NU

with solid-phase extractant located at the bottom of the vial at the temperature higher than extractant melting point; the melting of the solid-phase and its dispersion in a sample; the

MA

movement of extractant drops from the bottom to the top of the sample vial; separation of the organic phase containing analyte. Hence, the extractant dispersion, analyte extraction and complete phases separation are carried out in one step by heating.

D

In our research menthol was studied as extractant due to the fact that it is cheap, and non-

PT E

toxic substance with melting point close to normal conditions (39 – 40 °C). The distribution coefficient (KD) was determined by the ratio of the molar concentrations of benzoic acid in the menthol phase and in the aqueous phase after the establishment of the phase equilibrium (1:1,

CE

v/v; concentration of benzoic acid – 25 mg L-1). The estimated KD was equal to 50±2 and its value was almost independent on the benzoic acid concentration in the whole studied range (0.5 – 50 mg L-1).

AC

The Supelco C18 column was used for the chromatographic separation of benzoic acid after the HA-LLME. Previously, chromatographic separation was carried out using acetonitrile– phosphoric acid (40:60) as a mobile phase and detection at 230 nm according to [11]. It was found that food preservative sorbic acid was also extracted into menthol, and benzoic and sorbic acids had the same retention time (RT=10.6 min) (SM_Fig.1a). The mobile phase based on methanol–ammonium acetate buffer (30:70, v/v) [8] provided for effective chromatographic separation of benzoic (RT=6.9 min) and sorbic (RT=7.9 min) acids after the HA-LLME (SM_Fig.1b). 3.2.

Optimization of the HA-LLME 5

ACCEPTED MANUSCRIPT Using the experimental design 25-1 described above (section 2.7) 16 experiments with predefined values of the factors were performed and the area under the peak of benzoic acid registered with HPLC-UV was employed as a target optimization parameter. The regression coefficients in the equation relating experimental parameters with the peak area were calculated using normal equation [20]. The observed values for each of the encoded factors are given in the equation (1):

PT

AUP = 3560.56 – 1116.06×t – 703.94×pH + 21.06×salt – 1020.56×Vex – 248.94×Vs + +244.94×t×pH – 238.06×t×salt + 254.31×t×Vex + 71.19× t×Vs – 396.44×pH×salt +

RI

122.19×pH×Vex+ + 408.31×pH×Vs – 91.06×salt×Vex – 2.44×salt×Vs – 449.94×Vex×Vs (1), where AUP is the area under the peak of benzoic acid, Vex– extractant volume, Vs– sample

SC

volume, t – temperature, salt – presence/absence of Na2SO4. The absolute values of the coefficients indicate the importance of corresponding factors and their interactions for the

NU

experiment outcome.

Examination of the values of resulted coefficients reveals the following:

MA

- the most important parameter is the temperature and since the corresponding coefficient has negative sign – the lower the temperature, the higher is the extraction efficiency. No further

D

lowering of temperature is possible due to the menthol melting point value;

PT E

- the lower volume of the extraction solvent also provides for better efficiency from the point of view of enrichment factor obtained. No further reducing of extraction solvent volume is possible due to the handling of solidified extract; - the same holds for pH – the lower it is, the better is the extraction due to the suppression of

CE

benzoic acid ionization;

- the next most important coefficient after pH is the one for the interaction term between

AC

sample and solvent volumes; - cooperative effects of pH and solvent volume, and of pH and salt are also important; - the rest of the coefficients are at least two times smaller and, thus, these parameters and interactions have very minor influence on the experiment outcome; - no effect of Na2SO4 addition (salting-out effect) was observed, thus it was not used in further experiments. Taking into account the values of interaction terms between pH, Vex and Vs another experimental design including these parameters was made. This time it was full factorial design 23 for three parameters at two levels. pH values were 1 and 2 in order to explore the effect of pH at lower region, solvent volume was 100 and 200 μL, sample volume was 1 and 2 mL. Design 6

ACCEPTED MANUSCRIPT matrix with encoded factors is given in Supplementary material (SM Table 2). Thus, another nine experiments were run at the fixed temperature of 45 0C and without salting-out agent (taking into account the results of the previous run). Based on the registered outputs of these experiments (benzoic acid peak area) corresponding regression coefficients were calculated. They are presented in the equation (2):

AUP = 6248.25 + 117.75×pH – 867.25×Vex + 1662.5×Vs + 7.75×pH×Vex – 253×pH×Vs – – 160×Vex×Vs – 46×pH×Vex×Vs

RI

The following conclusions can be made from these results:

PT

(2).

- menthol volume should be kept minimal;

SC

- sample volume of 2 mL provides better conditions for the HA-LLME;

- in this experimental layout pH changes did not have significant influence on the extraction

NU

efficiency;

- interaction terms were also not very significant.

MA

Based on the results of these two studies the following conditions were chosen for all further HA-LLME experiments: T= 45 0C, pH = 2, Vex = 100 μL, Vs = 2 mL. Extraction time is also an important parameter; however, it was not studied within the

D

factorial design due to the following considerations. It was found that the complete phase

PT E

separation at 45 0C was observed only after 3 min. Thermostating the system for less than 3 min led to higher RSD values in replicated measurements, while the times longer than 3 min did not

3.3.

CE

lead to any improvements. Thus, 3 min extraction time was chosen as optimal. Interference effect

AC

The effects of such compounds as ascorbic, lemon, oxalic, malic, succinic, and tartaric acids, rutine, inorganic salts (aluminum chloride, sodium chloride, sodium sulfate), glucose and sucrose, that can be found in juices, on the determination of benzoic acid by the developed method were investigated. The tolerable concentration of each interfering compound is considered to be less than 5 % of relative error in the signal. All these substances were found to be not interfering even at their 1000-fold excess. 3.4.

Analytical performance

Under the optimal conditions, the analytical performance of the proposed extraction procedure coupled with HPLC-UV was evaluated. The linear calibration range of 0.5 – 50 mg L7

ACCEPTED MANUSCRIPT 1

for benzoic acid was obtained with a correlation coefficient of 0.996. The LOD, calculated

from a blank test, based on 3σ, was found to be 0.15 mg L-1. The LOQ, equal to 0.45 mg L-1, was calculated based on the analyte peak with 10 times signal-to-noise ratio. The within-day precision of the method was investigated in standard solutions at the concentration levels of 1 and 25 mg L-1 of benzoic acid, by performing three replicates daily. The same solutions were analyzed three times each day, for a period of five days, for the day-today precision evaluation. The obtained relative standard deviation (RSD) values were in the

PT

range from 3 to 6 % and from 4 to 8 %, respectively. There was no significant influence of the stability of the preservative on the experimental results during sample preparation and HPLC-

RI

UV analysis.

The time required for the HA-LLME procedure is 3 min. Duration of the HPLC-UV analysis

Analysis of the real samples

NU

3.5.

SC

is 8 min.

The developed procedure was applied for the determination of benzoic acid in five juice

MA

samples intended for children consumption (apple, multifruit, orange, cherry, and berry). Samples were used without any pre-treatment and some of them contained pulp. According to juice producers all samples were manufactured without preservatives, as it was indicated on the

D

packages. The concentrations of the target analyte in real samples obtained by the standard

PT E

addition method are shown in Table 3. There was only one sample (berry juice) with benzoic acid detected (13.5 mg L-1). It can be explained by the fact that some berries naturally contain benzoic acid, for example, cloudberries.

CE

Recovery values varied from 93 to 117 %. The results demonstrated that the precision and accuracy of the present method are acceptable. The values of recovery were calculated using

AC

the following equation [21]: Recovery (%) =

Qst s  Qs  100 Qst

(3),

where Qst is the quantity of the analyte added (spike value), Qst+s – the quantity of the analyte recovered from the spiked sample and Qs – the quantity of analyte in the original sample. Accuracy and reliability of the resulting information were further studied by analyzing the mentioned samples using referent extraction procedure described in 2.5. As it can be seen from Table 3, the analytical results agreed well with the results obtained with reference method [7]. F-values ≤ 19.00 indicate insignificant difference in precision between both methods at the

8

ACCEPTED MANUSCRIPT 95 % confidence level. t-values ≤ 2.78 indicate insignificant difference between the results obtained using these methods (n = 3). 3.6. Comparison of the developed method with other methods Table 1 shows figures of merit for the proposed method and those of other methods reported for the determination of benzoic acid in various food samples. As it can be seen, LOD and linear range of the proposed method are comparable with those reported for the other

PT

methods. Unlike reported sample preparation procedures for the determination of benzoic acid, the described procedure is performed in one step and it can be considered as a simple, fast and

RI

inexpensive technique. Moreover, the proposed procedure is characterized by a low consumption of the extraction organic solvent, which is also environment-friendly, and therefore offers low

SC

waste production. In comparison with chromatograms of juice samples obtained in [4, 14], where several unknown peaks from complex matrixes were presented, the chromatograms obtained

NU

after HA-LLME included only one peak related to benzoic acid (Fig. 2). From this point of view the proposed method can be considered as selective.

MA

As for the using of methanol as extractant, microextraction process and complete phase separation can be achieved only by heating for 3 min instead 7 min for the DLLME in [18]. In the developed HA-LLME procedure a disperser solvent is not required. Due to availability, low

D

price and simplicity of the manufacture of extraction vials, the HA-LLME can be considered as

PT E

potential on-site sample pretreatment for a large number of samples. The single use of extraction vial eliminates the risk of cross-contamination.

CE

4. Conclusions

In this paper a simple, fast, selective, low-cost and environmentally friendly method for

AC

the determination of benzoic acid in juice samples was proposed. The method is based on the new microextraction procedure, HA-LLME, coupled with the HPLC-UV detection. The HALLME assumes the heating of aqueous sample in an extraction vial with solid-phase menthol located at the bottom. The heating leads to the menthol melting resulting in the extractant dispersion without the use of any disperser organic solvents, fast analyte extraction and phase separation. It is known, that disperser organic solvents used in the DLLME can increase solubility of the hydrophobic analytes in an aqueous phase reducing the extraction efficiency. Extraction vials can be preliminary prepared and used for on-site preconcentration. The single use of extraction vial eliminates the risk of cross-contamination. Capabilities of the HA-LLME were successfully demonstrated in separation and preconcentration of benzoic acid from juice samples. It was found that menthol provides for effective microextraction of benzoic acid from 9

ACCEPTED MANUSCRIPT juice samples and the chromatograms obtained after HA-LLME have only one single peak related to benzoic acid. Unlike the DLLME procedure with menthol [18], the HA-LLME is twice as fast and does not require additional disperser organic solvent, which negatively affects the extraction efficiency. However, this method has such limitation as clogging a HPLC-UV system with menthol phase, thus, its dissolution in organic solvent or in mobile phase before assay is required. Comparative studies have indicated that the HA-LLME is a reliable analytical protocol

PT

with wide potential applications in food analysis. The HA-LLME has the opportunity to be coupled with other instrumental methods (gas-chromatography, mass spectrometry, etc). The

AC

CE

PT E

D

MA

NU

SC

RI

application of this method for other analytes is under investigation.

10

ACCEPTED MANUSCRIPT Acknowledgements The reported study was funded by Russian Foundation for Basic Research (project no. 16-33-60126 mol_а_dk). Some experiments were performed at the Center for Chemical Analysis

AC

CE

PT E

D

MA

NU

SC

RI

PT

and Materials Research of Research park of St. Petersburg State University.

11

ACCEPTED MANUSCRIPT Figure captions Fig. 1. Schematic representation of the HA-LLME procedure for the determination of benzoic acid in juice.

AC

CE

PT E

D

MA

NU

SC

RI

PT

Fig. 2. Chromatogram of real sample (berry juice) after the HA-LLME.

12

ACCEPTED MANUSCRIPT References

AC

CE

PT E

D

MA

NU

SC

RI

PT

[1] C.Z. Dong, W.F. Wang, Headspace solid-phase microextraction applied to the simultaneous determination of sorbic and benzoic acids in beverages, Anal. Chim. Acta, 562 (2006) 23-29. [2] C.M. Lino, A. Pena, Occurrence of caffeine, saccharin, benzoic acid and sorbic acid in soft drinks and nectars in Portugal and subsequent exposure assessment, Food Chem., 121 (2010) 503-508. [3] H.M. Pylypiw, M.T. Grether, Rapid high-performance liquid chromatography method for the analysis of sodium benzoate and potassium sorbate in foods, J. Chromatogr. A, 883 (2000) 299304. [4] C.P. Ene, E. Diacu, High-performance liquid chromatography method for the determination of benzoic acid in beverages U.P.B. Sci. Bull Series B, 71 (2009) 81-88. [5] S.A.V. Tfouni, M.C.F. Toledo, Determination of benzoic and sorbic acids in Brazilian food, Food Control, 13 (2002) 117-123. [6] R. Heydari, M. Mousavi, Simultaneous determination of saccharine, caffeine, salicylic acid and benzoic acid in different matrixes by salt and air-assisted homogeneous liquid-liquid extraction and high-performance liquid chromatography, J. Chil. Chem. Soc., 61 (2016) 30903094. [7] M. Kamankesh, A. Mohammadi, Z.M. Tehrani, R. Ferdowsi, H. Hosseini, Dispersive liquidliquid microextraction followed by high-performance liquid chromatography for determination of benzoate and sorbate in yogurt drinks and method optimization by central composite design, Talanta, 109 (2013) 46-51. [8] M.Z. Ding, J. Peng, S.L. Ma, Y.C. Zhang, An environment-friendly procedure for the high performance liquid chromatography determination of benzoic acid and sorbic acid in soy sauce, Food Chem., 183 (2015) 26-29. [9] S. Jawaid, F.N. Talpur, S.M. Nizamani, N.N. Memon, H.I. Afridi, A.A. Khaskheli, Rapid in situ esterification method for the determination of benzoic acid in dairy milk by GC-FID, Food Anal. Meth., 8 (2015) 1477-1483. [10] T.A. Berger, B.K. Berger, Rapid, Direct quantitation of the preservatives benzoic and sorbic acid (and salts) plus caffeine in foods and aqueous beverages using supercritical fluid chromatography, Chromatographia, 76 (2013) 393-399. [11] T. Fujiyoshi, T. Ikami, K. Kikukawa, M. Kobayashi, R. Takai, D. Kozaki, A. Yamamoto, Direct quantitation of the preservatives benzoic and sorbic acid in processed foods using derivative spectrophotometry combined with micro dialysis, Food Chem., 240 (2018) 386-390. [12] Y.J. Tang, M.J. Wu, The simultaneous separation and determination of five organic acids in food by capillary electrophoresis, Food Chem., 103 (2007) 243-248. [13] R.X. Wei, W.H. Li, L.R. Yang, Y.X. Jiang, T.Y. Xie, Online preconcentration in capillary electrophoresis with contactless conductivity detection for sensitive determination of sorbic and benzoic acids in soy sauce, Talanta, 83 (2011) 1487-1490. [14] A. Aresta, C. Zambonin, Simultaneous determination of salicylic, 3-methyl salicylic, 4methyl salicylic, acetylsalicylic and benzoic acids in fruit, vegetables and derived beverages by SPME-LC-UV/DAD, J. Pharm. Biomed. Anal., 121 (2016) 63-68. [15] Z.H. Wang, J.F. Xia, F.Y. Zhao, Q. Han, X.M. Guo, H. Wang, M.Y. Ding, Determination of benzoic acid in milk by solid-phase extraction and ion chromatography with conductivity detection, Chin. Chem. Lett., 24 (2013) 243-245. [16] L. Caia, J. Donga, Y. Wanga, X. Chena, Thin-film microextraction coupled to surface enhanced Raman scattering for the rapid detection of benzoic acid in carbonated beverages, Talanta, 178 (2018) 268-273. [17] A. Sarafraz-Yazdi, A. Amiri, Liquid-phase microextraction, Trends in Analytical Chemistry, 29 (2010) 1-14.

13

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

MA

NU

SC

RI

PT

[18] M.A. Farajzadeh, L. Goushjuii, Study of menthol as a green extractant in dispersive liquid– liquid microextraction; application in extraction of phthalate esters from pharmaceutical products, Anal. Methods, 5 (2013) 1975-1982. [19] J. Regueiro, M. Llompart, C. Garcia-Jares, J.C. Garcia-Monteagudo, R. Cela, Ultrasoundassisted emulsification-microextraction of emergent contaminants and pesticides in environmental waters, J. Chromatogr. A, 1190 (2008) 27-38. [20] R.G. Brereton, Chemometrics: data analysis for the laboratory and chemical plant, John Wiley & Sons, 2003. [21] D.T. Burns, K. Danzer, A. Townshend, Use of the terms "recovery" and "apparent recovery" in analytical procedures - (IUPAC recommendations 2002), Pure Appl. Chem., 74 (2002) 2201-2205. [22] J. Wang, X. Guo, L. Jia, A simple method for the determination of benzoic acid based on room temperature phosphorescence of 1-bromopyrene/γ-cyclodextrin complex in water, Talanta 162 (2017) 423–427.

14

ACCEPTED MANUSCRIPT Table 1. Comparison of methods for benzoic acid determination. Sample

Solvent consumption

Method of detection

LOD, mg L-1

Linear range, mg L-1

Recovery, %

Ref.

Filtration, sonication

-

HPLCDAD

7

20-180

85-93

[4]

Soy sauce

1g

LLE

5mL (diethyl ether)

HPLC-UV

0.2

20-100

96.1104.3

[8]

Juices

10 -100 g

SPME

-

LCUV/DAD

0.003

0.001-3

-

[14]

Milk

5 mL

SPME

5 mL( acetonitrile), 4 mL (methanol)

IC-SCD

0.001

0.1-20

88-93

[15]

Beverages

-

Derivatization

-

RTP

0.08

0-85.4

90-114

[22]

Carbonated beverages

10 mL

TFME

-

SERS

3.6

25-500

85-103

[16]

SP

-

up to 10

90-100

[11]

Soysauce, vinegar

2 mL

20 mL (acetonitrile)

MA

5 mL

CE-C D

0.01

0.04-2.44

95-99

[13]

1 mL (CCl4)

CE-UV

2.19

2.9373.27

98.2199.24

[12]

450 µL (ethanol), 60 µL (1-octanol)

HPLC-UV

0.06×10-3

0.001-10

91.25

[7]

Esterification, sonication

3 mL (methanol), 5 mL (hexane)

GC-FID

0.02

0.04-2.5

94-99

[9]

-

HPLC-UV

0.0052×10-

0.005-0.1

88-94

[6]

0.5-50

93-117

This work

Lipids LLE and dilution

Raw and processed milk

CE

2 mL 3

AC

mL

4

5 mL(anhydrous diethyl ether)

LLE

DLLME

Yogurt drinks

NU

2-5 g

D

Soy sauce

Micro dialysis

PT E

Processed meat, tsukudani, soft drink, dairy product and pickles

PT

-

SC

Juices

RI

Sample weight /volume

Sample pretreatment

Toothpaste, softdrink, tea

1 g, 5 mL

SAAHLLE

Juices

2 mL

HA-LLME

3

100 µL (menthol)

HPLC-UV

0.15

CE-UV сapillary electrophoresis with ultraviolet detection 4

CE-C D- capillary electrophoresis with capacitively coupled contactless conductivity detection DLLME - dispersive liquid–liquid microextraction GC-FID - gas-chromatography with flame ionization detector HA-LLME - heating-assisted liquid-liquid microextraction HPLC-DAD - high-performance liquid chromatography-diode array detection

15

ACCEPTED MANUSCRIPT HPLC-UV - high-performance liquid chromatography with ultraviolet detection IC-SCD - ion chromatography with a suppressed conductivity detector LC-UV/DAD - liquid chromatography with ultraviolet detection/ diode array detection LLE - liquid-liquid extraction RTP - Room temperature phosphorescence SAAHLLE - salt and air-assisted homogeneous liquid-liquid extraction SERS - Surface enhanced Raman scattering SP – Spectrophotometry

PT

SPME - solid phase microextraction

AC

CE

PT E

D

MA

NU

SC

RI

TFME - silica gel thin-film microextraction

16

ACCEPTED MANUSCRIPT Table 2. Experimental levels for studied parameters. Vex– extractant volume, Vs– sample volume, t – temperature, salt – presence/absence of Na2SO4. Level

Level

-1

+1

t, 0C

45

70

pH

2

7

Na2SO4

yes

no

Vex, μL

100

300

Vs, mL

1

PT

Factors

AC

CE

PT E

D

MA

NU

SC

RI

`

10

17

ACCEPTED MANUSCRIPT Table 3. The results of determination of benzoic acid in baby juices (n=3, P=0.95; Fk = 19.00; tk = 2.78).

Cherry juice



1

1.17±0.26

1.27±0.34

0.44

1.47

117

10

10.40±0.57

9.67±0.22

2.77

6.86

104

25

24.73±0.46

24.87±0.65

0.42

1.97

99

0



1

1.07±0.22

1.03±0.36

0.20

2.71

107

10

10.00±0.66

10.77±0.65

2.07

1.03

100

25

24.73±0.30

24.67±0.22

0.45

1.86

99

0



1

1.08±0.30

1.13±0.36

0.67

1.95

108

10

10.60±0.57

10.87±0.82

0.66

2.02

106

25

24.87±0.79

25.07±0.30

0.59

7.00

99

0


1

0.93±0.30

0.87±0.08

0.53

13.00

93

10.00±0.66

10.13±0.50

0.40

1.70

100

25.00±0.57

24.67±0.22

1.35

6.86

100

SC

RI

PT

0


0

13.50±0.38

13.27±0.58

0.84

2.33

1

14.66±0.36

14.63±0.48

0.73

2.58

116

10

23.80±0.66

23.87±0.60

0.19

1.21

103

25

38.63±0.60

38.00±0.52

1.99

1.33

101

10

AC

CE

25 Berry juice

Recovery, %

[7]

NU

Orange juice

F-value

This work

MA

Multifruit juice

t-value

Added

D

Аpple juice

Concentration of benzoic acid, mg L-1

PT E

Sample

18

ACCEPTED MANUSCRIPT Highlights • Heating-assisted liquid-liquid microextraction (HA-LLME) • Fast analyte extraction and complete phases separation in one step • Extraction vials for on-site sample preparation • HA-LLME for the HPLC-UV determination of benzoic acid in juices

AC

CE

PT E

D

MA

NU

SC

RI

PT

• Menthol as extractant for the HA-LLME

19

Graphics Abstract

Figure 1

Figure 2