DI-SDME

DI-SDME, in its simplest format, involves the immersion of a drop of solvent at the tip of a syringe, into a water sample, withdrawal into the syringe and analysis by GC, GC-MS atomic absorption spectroscopy (AAS), or HPLC. The theoretical concentrating, or EF can range from 10 to 1,000 SDME, like all LLE extractions, is an equilibrium process, with actual analyte extraction amounts ranging from 10% to 100%, though it is consistent when all extraction parameters are consistent. It is also possible, however, in some cases to achieve nearly 100% recovery, when a complexing, or derivatizing agent or acid/base conversion reaction is used to force the equilibrium towards the extraction solvent. Typical sample sizes range from 0.5–10 mL. Typical solvent volumes range from 0.5–10 µL. Extraction times vary from 5 min to 90 min, though typical extraction times are about 15 min. Reverse DI-SDME, where the sample is an oil and the extraction solvent is aqueous or very polar solvent (ionic DES, IL) is also possible. DI-SDME suffers from drop instability, as well solvent water solubility or evaporation, and the limited solvent volume possible. These problems have been addressed, in part, by using magnetic solvents and the simple directly suspended droplet microextraction (DSDME) and LLLME methods. Nevertheless, this is an important sample preparation technique and has been found useful for food analyses [105–110].

HS-SDME

HS-SDME involves exposing the solvent drop in the HS of a liquid (aqueous or oil) or solid. This technique is much less susceptible to drop instability and has been used successfully for analysis of volatiles in spices, wine and other food products [118–123]. Typically, sample temperatures range from 25–40℃, to avoid evaporation of the solvents, such as hexane. Less volatile solvents, such as DESs, NADESs, and ILs can be used at higher temperatures. Sample sizes range from 10 mg–10 g for solid samples and and 10 µL–10 mL for liquid samples. This technique can compete quite effectively with purge and trap for volatiles analysis, and the extraction can be fully automated as well and is comparatively economical [108].

In an example by Abreu et al. [118], HS-SDME analysis for 2-phenoxyethanol (an anesthetic) was performed on 2 g tilapia homogenate using 1.8 µL of octane for 30 min at 35℃. The octane extract (0.5 µL) was analyzed by GC-MS. The LOQ was 0.56 mg L^-1 of sample. The results were comparable to those obtained by SPME, but at requiring less extraction time and cost, with no carryover. In a second example, Triaux et al. [119], analyzed 6 spices for 29 terpenes. The powdered spices (50 mg were extracted for 90 min with 1.5 µL of DES solvent (tetrabutyl bromide:dodecanol, 1:2). The DES extract was analyzed by GC-MS, using a DB-WAX column. LOQs ranged from 0.47 µg/g (borneol) to 86 µg/g (α-farnecene), with most terpenes at less than 2 µg/g. In a third example by Abolghasemi et al. [120], HS-SDME was used to analyze the amounts of 7 triazole fungicides in fruit juices and vegetables. Prepared aqueous samles (10 mL containing 10% salt were extracted for 30 min at 85℃ with 2 µL of DES (choline chloride:4-chlorophenol, 1:2) and the extract analyzed by GC-FID. The LODs for the fungisides ranged from 0.82−1.0 µg L^-1.

HF-LPME applications

In HF-LPME extractions, the extraction solvent is contained within the lumen of a so-called porous hollow fiber, usually polypropylene, though other polymeric materials can be used. The fiber is prepared to length, as needed, and typically contains 5–50 µL of extraction solvent. This technique has been used primarily for extraction of polar or ionizable analytes from aqueous solutions. Two-phase HF-LPME is also an equilibrium process while three phase can result in nearly exhaustive extraction given long enough time, and the use of a complexing, derivatizing or pH sample, receptor solution differential to effectively make the extraction irreversible. The major deficiencies of the technique are long extraction times (45–99 minutes) and pore plugging of the fiber. The technique can also be used with a flat sheet of membrane, separating sample and receiving solution, discussed below [124–129].

An example by Sun et al. [138] illustrates the applicability of HF-LPME for determination of organophosphorus pesticides in fish tissue. Samples were first extracted with acetone, and centrifuged. The acetone extracts were rotory evaporated, and redissolved in 10 mL 5% methanol/water. After filtration, a 2-phase HF-LPME extraction with 30 µL of o-xylene for 30 min was followed by GC-MS analysis. LODs ranged from 2.1–4.5 µg kg^-1 for 8 pesticides. A second example by Afshar Moggadam et al. [141] illustrates the ability of HF-LPME to concentrate trace impurities, using a solvent stir bar HF-LPME mode [referred to as solvent bar microextraction (SBME)], combined with simultaneous derivatization. The procedure involved using a fiber containing a DES extraction solvent (8-hydroxyquinoline:pivalic acid), chloroacetyl chloride derivatizing agent and an iron wire inserted into the fiber, with both ends heat sealed. The solvent bar was placed in the sample and stirred for 5 min. The derivatized extracts were analyzed by GC-MS and the procedure was validated using FDA recommendations for the determination of 12 phenols in fruit juices packaged in plastic. The technique had LOQs ranging from 29–76 ng L^-1 and EFs of more than 1,000. In a third example by Gonzalez-Dominguez et al. [143] a simultaneous derivatization-extraction of organotin compounds in a solution of 1 mL fruit juices, 4 mL of buffer solution and sodium tetraethyl borate, and a 10 min 2-phase extraction with hexane contained in the lumen of the fiber. The extract was analyzed by GC-MS, with LOQs of 0.8–1.8 µg L^-1. In a fourth example, Lorenzo-Parodi et al. [146] compared 3-phase HF-LPME and 3-phase PALME for the determination of aromatic amines in the urine of smokers. PALME was determined to be faster, less labor intensive, capable of automation and required much smaller volumes of sample. GC-MS analyses resulted in LODs ranging from 45–75 ng L^-1 for 13 aromatic amines.

EME applications

EME was originally conducted with a three phase HF-LPME system, with electrodes placed inside of the fiber lumen and the sample. A potential placed on the electrodes enabled the extraction of ionized drugs from biological fluids in much shorter timeframes than HF-LPME alone [130–137]. Both HF-LPME and EME have more recently have transposed to a 96 well system, using the PALME methodology, will be discussed in section CF, microfluidics and 96 well LPME.

Two examples illustrate the potential of EME for the analysis of foods, Rezaee et al. [149] extracted melamine adulterant in infant formulae using EME, with a fiber reinforced with nano graphine oxide (NGO). The NGO acted as an adsorbent interface, enhancing the electrical conductivity and migration of analyte into the acceptor solution, thus enhancing recoveries of polar analytes. The fiber solvent (SLM) used was 1-octanol, the sample solution was adjusted to pH 3 and the acceptor solution (20 µL) pH 2 with a 70 V potential for 10 min, followed by HPLC/UV. An LOQ of 0.1 µg kg^-1 for infant formula was obtained. In the second example by Nsubuga et al. [150] seafood samples were microwave digested and the solutions subjected to pEME. pEME was performed in polypropylene envelopes containing an electrode, 120 µL of 0.1 mol L^-1 NAOH and 1-hexanol as the SLM solvent. EME was conducted with 25 mL of sample solution (pH 6) at 12 V for 10 min. Following EME, the extract was analyzed with IC for an LOD of 40 µg kg^-1.

Solvent assisted DLLME applications

Solvent assisted (SA)-DLLME involves dissolving a water-insoluble extraction solvent (such as C₂Cl₄, 1-octanol, DES, MDES, NDES, IL, or MIL) in a water-soluble solvent (ACN, ethanol, methanol, or acetone) and injecting the solution rapidly into the aqueous sample, to obtain a dispersion [72, 156, 157]. The proper volumes of sample, extraction solvent and dispersion solvent must be obtained by experiment. Published data should be used as guides only. The dispersion is broken up by centrifugation, salt or dispersion solvent addition. Chlorinated solvents are sometimes still used, since their high density allows easy removal from the bottom of the centrifuge tube. Low density solvent may require using a thin neck tube or removing the water with a long needle syringe first [157, 165]. High melting liquids (such as undecanol, some DESs, and ILs) can be solidified and physically removed to a sample vial (DLLME-SFO) [157]. MDESs and MILs can be retrieved, sometimes without centrifugation, using a neodymium magnet, though this must be followed up by back extraction with 10–20 µL of solvent compatible with the analytical instrument. NADES and some DESs can be directly used in GC. Ionic DESs and ILs are normally chromatographed with HPLC [155–157].

In a paper by Barzegar et al. [166] barbequed meat was analyzed for aromatic amines. One gram of homogenized sample was mixed with KOH, ACN and ethanol and microwave extracted. The resulting solution was centrifuged, the pH adjusted to 3 and Carrez I and II solutions (potassium ferricyanide, ZnSO₄) added to clarify the proteins and derivatize amines in the solution. After changing the pH to 11 and centrifugation, the upper liquid phase was used for DLLME. A solution of 100 µL of 1-octanol in 600 µL ethanol was injected into the prepared solution and, after centrifugation, 20 µL of the upper layer of 1octanol analyzed by HPLC/UV. LODs ranged between 0.25–0.71 µg kg^-1 for 4 aromatic amines. In a second example, Mardani et al. [167] analyzed cocoa powder for organochlorine residues using a DLLME-SFO preparation procedure. Cocoa powder (1 g) was mixed with 4 mL of 5% NaCl solution, and 80 µL 1-decanol in 1 mL ethanol injected after raising the temperature to 80℃. After refrigerated centrifugation, the solidified 1-decanol extract was analyzed with GC/ECD. LOQs ranged from 0.091–0.175 µg kg^-1 for 5 organochlorine pesticides (OCPs), including Aldrin.

VA-DLLME applications

VA-DLLME avoids the need for a dispersion solvent, which, when used in excess, can decrease the yield of extraction solvent recovery. Typically, the extraction solvent is rapidly injected by syringe into the sample, which is then immediately vortexed for 30–90 s. Vortexing does not always yield a true dispersion, which should not immediately separate into separate layers. However, vortexing still results in significant extractions. The sample is retrieved as above [157].

Total iron in water and food samples was determined Zhang et al. [171] using reductive-derivatization, followed by VA-DLLME-SFO. Solutions containing Fe (II) and Fe (III) were mixed with a DES (thymol:lauric acid) and a complexing agent (1,10-phenanthroline). The DES acted as both the extraction solvent and a reducing agent, converting Fe (III) to Fe (II), which then complexed with the 1,10-phenantroline. After VA-DLLME and cooling, the solidified DES extraction solvent was analyzed with VIS detection. The LOQ for total iron was 1.5 µg L^-1. Wang et al. [173] analyzed strobilurin fungicides in water, juice and vinegar using MIL-VA-DLLME. The MDES was prepared from methyltrioctylammonium chloride, ferric chloride and heptanoic acid. Addition of 200 µL of the MDES to 5 mL of sample and vortxing for 3 min. was followed by addition of 90 mg of carbonyl iron powder (CIP) to aid in separating the MDES. The MDS, was then filtered to remove the coagulated CIP-ferric chloride, and the DES extract analyzed by HPLC/UV. The resulting LODs for 3 fungicides ranged from 1–2 µg L^-1.

UA-DLLME applications

UA-DLLME is comparable to VA-DLLME, but usually results in a more stable dispersion system. Ultrasound should be limited to as short a time as possible, to prevent analyte degradation and possible loss of more volatile analytes. Sample retrieval is the same as above [157].

An example of DES-UA-DLLME-SFO for the determination of nickel and cobalt in broccoli, spinach and environmental water samples was demonstrated by Tavakoli et al. [176]. Food samples were digested with HNO₃/H₂O₂ and 150 µL DES (d,l-menthol:decanoic acid) added to 50 mL of prepared sample, 2 mL pH 6 buffer and 100 µL of complexing agent [2-(5-bromo-2-pyridoazol)-5-(diethylamino) phenol] dissolved in ethanol. After ultrasonic treatment for 2 min., the sample was centrifuged and cooled in an ice bath. The solidified DES was removed, allowed to melt and diluted to 1 mL with ethanol, and the extract analyzed by FAAS. LOQs were 1.1 µg L^-1and 1.3 µg L^-1, for Co and Ni, respectively and the EF was 50. A second example of UA-DES-DLLME-SFO was used by Elik and Altunay [178] for propineb fungicide in water and foods, including cereal based baby foods and tomato. The procedure involved the addition of 330 µL of DES (8-hydroxyquinoline:pivalic acid) to a 5 mL sample solution (pH buffered to 4.6), 1.5 min sonication, centrifugation, solidifying the DES in an ice bath, diluting the solidified DES in 500 µL of ethanol and UV detection. The LOD was 6.1 µg L^-1, and the EF was 93.

Air-assisted DLLME applications

The term air-assisted (AA) is a misnomer, since dispersion is actually obtained by the shearing forces resulting from the rapid movement of sample and extraction solvent into and out of a syringe. The process works best with small volumes of sample, to ensure maximum sample-solvent interaction [157, 165].

Elik et al. [179] used a MDES-AA-DLLME procedure for the analysis of melamine in milk and milk products. Prepared solutions were extracted by drawing a solution of 260 µL magnetic DES [octanoic acid:Aliquat-336:Co (II), 3:2:1] and 125 µL acetone in and out of a the sample with a syringe 6 times. The MDES was separated from the sample solution with a neodymium magnet, the DES back extracted and the extract analyzed by UV-Vis. The LOQ was 0.9 µg L^-1, with an EF of 144.

Switchable solvent DLLME applications

Switchable solvents (SSs) are simply water insoluble acids or bases which become soluble or insoluble, depending on the pH of the solution. As an example, sodium hexanoate added to an aqueous sample will becomes an insoluble dispersion when acid is added. Recovery is as above [157, 162].

A DES-SS-DLLME extraction was developed by Mogaddam et al. [180] for the extraction of phenolic antioxidants from edible oil samples. The antioxidants were first extracted from the 5 mL of oil at 80℃ with 1.25 mL of 1.5 mol L^-1 NaOH for 5 min and centrifuged, The extracted antioxidants were concentrated for analysis by dissolving DES (tetrabutylammonium chloride:hydroquinone) in the alkaline solution, followed by dispersion formation of the DES upon acidification. After centrifugation, the DES extract was analyzed by GC-FID. The LOQs ranged from 0.36−1.41 µg L^-1 with EFs averaging 400. Wang et al. [182] developed a simple SS-DLLME for on-site sample preparation of benzoylurea insecticides in water and honey samples. The water and honey samples (5 mL honey diluted to 100 mL) were filtered to remove particulates, and 15 mL added to a plugged 20 mL plastic syringe, along with 1.1 mL (100 mg mL^-1) of salicylic acid solution. After addition of H₂SO₄ solution to acidify the sample, a salicylic acid dispersion formed. NaCl (0.5 mL, 20%) was added, the syringe plunger attached, the plug removed and replaced by a needle containing a porous fabric to retain the salicylic acid. The water was expelled from the syringe, which was then returned to the laboratory for analysis of the extract. After dissolving the salicylic acid in methanol, the solution was filtered and analyzed by HPLC/UV. The LODs for the insecticides were 1.5 µg L^-1 for the water samples and 30–90 µg kg^-1 for the honey samples.

In situ dispersion formation DLLME applications

In situ-DLLME, also referred to as homogeneous liquid-liquid microextraction (HLLME) most commonly involves formation of a dispersion by the in situ formation of an insoluble extractant solvent, or by temperature control to form a SUPRAS [157, 160, 163].

Three example are discussed below.

Antibacterial sulfonamides were determined in milk with MIL-in situ-DLLME sample preparation in a paper by Yao and Du [184]. The technique involved adding a water soluble MIL (containing an organic free radical, 4-OH-Tempo) to milk samples which had been denatured and filtered. KPF₆ was then added to the solution, resulting in the substitution of the chloride ion in the soluble IL with the PF₆^-1 ion and formation of an emulsion, extracting the contaminants. The MIL was separated from the solution using a neodymium magnet, back extracted into methanol and HPLC/UV used for final analysis for 5 sulfonamides. LOQs ranged from 1.8–3 µg L^-1, with EFs between 42–47. Timofeeva et al. [185] developed an automated system for the analysis of PAHs in tea infusion using SUPRAS-in situ-DLLME, based on a syringe pump system interfaced with HPLC/fluorescence detection (FLD). The procedure involved first: the formation of a SUPRAS emulsion by mixing hexanoic acid with NaOH in a 1:4 molar ratio. This was followed by addition of 1.2 mL of the emulsion to 3.8 mL of tea sample in the syringe, with stirring. Stirring was then stoped to allow separation of the phases. The upper phase (50 µL) was then transferred from the syinge directly to the HPLC for analysis. The syringe was automatically cleaned during the HPLC run for a another sample. The LODs for PAHs ranged from 0.02−0.04 µg L^-1, and EFs ranged from 38−46. As a final DLLME example, Piao et al. [187] developed a MDES-in situ-DLLME sample preparation for the analysis of triazine herbicides in rice. The procedure involved extraction of 1 g of powdered rice with 4 mL hexane and centrifugation. Addition of 250 µL DES (ethylene glycol: tetrabutylammonium chloride), 40 mg iron chloride, and 90 mg of CIP to the hexane was followed by removal of the generated MIL/CIP. Separation of the MIL/CIP with a neodymium magnet extracted the herbicides from the hexane. The MIL/CIP was then back extracted with 5 mL of diethyl ether, the ether evaporated, the residue dissolved in 100 µL of ACN, and 20 µL of the extract analyzed by HPLC/UV. LOQs ranged from 5–10 µg kg^-1.

LPME using 96 well sampler application

Morelli et al. [196] developed a method for the extraction and analysis of 11 contaminants of concern in environmental water, including pesticides and hormones. The sample preparation involved extraction with DES impregnated in the pores of 1 cm long hollow fibers, placed over the pins of a 96 well sampler. The authors refer to this as HF-microporous membrane liquid-liquid microextraction (HF-MMLLME). A less cumbersome and more descriptive name which will be used here is DES-SLMME (DES-SLM microextraction). In effect, this technique is actually a variation of SDME, with the solvent held in the pores of the fiber polymer, rather than at the tip of a syringe. The procedure involved immersion of the fiber in 400 µL of DES (Thymol:Camphor, 1:1) for 10 min., immersion of the SLM into 1.5 mL of water sample for 50 min, and desorption with 300 µL of acetone:methanol (3:1) and analysis of 20 µL of the extract with HPLV/UV. The LODs ranged from 0.3–6 µg L^-1. The maximum EFs for this technique can be no higher than 5, due to the final ratio of sample and final desorption volumes, but the procedure allows automation of at least 30 samples simultaneously.

CF microfluidic EME application

An EME sample preparation for 5 biogenic amines in foods, using a CF microfluidic device, was developed by Zarhgampour et al. [201]. Samples (1 g), including sausage, were treated with trifluoroacetic acid, vortexed, filtered and the resulting solution diluted to 5 mL. The SLM for the microfluidic was a 3 mm × 4 mm porous polypropylene sheet, which was impregnated with 2-nitrophenyl octyl ether containing 10% 2-(2-ethylhexyl) phosphate. The receiving solution was 50 µL of 0.09 mol L^-1 HCl. A total volume of 2 mL of the sample was passed through the microfluidic device, and the receiving solution then treated with dabsyl chloride derivatizing agent, to allow analysis and detection by HPLC/UV. LODs ranged from 0.3–8 µg L^-1 for this semiautomated CF microfluidic system.

µ-QuEChERS development and applications

While the original QuEChERS method developers foresaw the possibility of using smaller amounts of sample, solvent and reagents, the approved methods used 10–15 g of sample, to provide the concentrations of pesticides in the final extracts to allow LOQ in the low ppb or high ppt ranges [204, 205]. An attempt to reduce sample requirements for QuEChERS was published in 2015 by Porto-Figueira et al. [11]. Their approach for monitoring zearalenone, a fungus aflatoxin in cereal grains, used 0.3 g of cereal sample (maize, corn) 0.7 mL ACN and 0.2 g of QuEChERS salt for the first step, and 75 mg MgSO₄, 12.5 mg of C₁₈ 12.5 mg of PSA for the cleanup. The final extracts were filtered, evaporated to near dryness and diluted with HPLC mobile phase. The samples were then analyzed using UHPLC-FLD.

Seven applications using the µ-QuEChERS approach are included in Table 9 [6–9, 11, 202–211]. The methodology has been used for pesticides, aflatoxins, alkaloids and polyphenols analysis in a variety of foods, food products and water runoff from washed foods. The reduced sample sizes, however, continue to require the use of highly sensitive chromatographic and detection instrumentation in the final analyses, though hazardous waste, analysis time and costs are reduced, while potentially increasing sample throughput and improving the green nature of the total methods.

Two recent examples illustrate the scope of the µ-QuEChERS sample preparation technique. González-Gómez et al. [208] developed a µ-QuEChERS procedure for determination of tropane alkaloids in leafy vegetables, with GC-MS/MS analysis. The procedure reduced reagent and solvent requirements 10 fold, compared to QuEChERS. The leafy vegetables were lyophilized to remove water, and 0.5 mL of water and 1 mL of ACN added to 0.1 g of the powder for the initial extraction. After vortexing, 0.65 g of partitioning salts and citrate buffer were added with vortex and ultrasound mixing. After centrifugation, the upper layer was treated with MgSO₄ (150 mg) and PSA (25 mg), vortexed and centrifuged. Internal standards were added, and the supernatant blown dry, redissolved in ACN:water, (1:1), filtered and analyzed by UHPLC-MS/MS. The LOQs for the method were < 2 µg kg^-1. The presence of alkaloids like scopolamine in samples illustrates the co-dependence and connection between environmental and food analyses, since these alkaloids could have arisen either through co-harvesting weeds containing the alkaloids, or by uptake from the soil containing residual alkaloids. The second procedure by Câmara et al. [209] involved the analysis of patulin in apple juices using µ-QuEChERS, followed by UHPLC-MS/MS. The µ-QuEChERS extraction involved the addition of 1 mL of ACN/1% acetic acid to 100 mg of sample, salting out with 0.65 g partitioning salts, ultrasound mixing and centrifugation. The supernatant was filtered and subjected to HPLC-MS/MS analysis. LOQ for patulin was 1.2 µg kg^-1. These results should be compared to the similar results achieved by Li et al. [113] (section DSDME) using LLLME.

DLLME coupled with other sample preparation techniques applications

Three following examples illustrate the utility of combining LPME with other sample preparation techniques, for the further purification and concentration of samples in food analyses. Zhou et al. [212] developed a method involving SPE, followed by DLLME for the HPLC/UV analysis of carbamates of fruit and vegetables. Homogenized samples were treated with NaCl and MgSO₄, and ACN. After sonication and filtration, the liquids were rotary evaporated, and reconstituted in water and filtered. Portions of the filtrates were subjected to SPE, using C₁₈ packing, washing with water to remove polar matrix interferences and eluted with 1 mL ACN. For DLLME, 35 µL of chloroform was added to the ACN extract and rapidly injected into 5 mL of water. After centrifugation, the CHCl₃ was removed with a syringe and evaporated. The residue was dissolved in 25 µL of methanol and 5 µL analyzed by HPLC/UV. The LODs ranged from 5–60 pg kg^-1 for 3 carbamates. A second example by El-Deen and Shimizu [214] coupled µ-QuEChEERS to an AA-DLLME sample preparation for the analysis of PAHs in coffee, tea and environmental waters. Water samples were prepared by direct AA-DLLME for GC-MS analysis. The coffee and tea samples were powdered and 200 mg samples were extracted with 1 mL each of water and ACN. After vortexing, 200 mg of salts were added, and the sample vortexed and centrifuged. A mixture of MgSO₄, PSA and C₁₈ (200 mg) was added to 800 µL of the supernatant, the mixture vortexed, centrifuged and 600 µL of the extract evaporated to dryness and redissolved in 5 mL of water. AA-DLLME was performed by adding 500 µL of diethyl carbonate and drawing the mixture in and out of a syringe 6 times. The sample was then centrifuged, the diethyl carbonate (300 µL separated, and 1 µL analyzed by GC-MS. In addition to further concentrating the analytes, the solvent was changed to one compatible with GC chromatography and caffein interference was significantly reduced. LODs for the procedure were 0.01–2.1 µg kg^-1 for 15 PAHs. In the last example, Yu et al. [228] used a modified QuEChERS procedure for the extraction of OCPs from green leafy vegetables, followed by a further cleanup/concentration of the extract with DLLME. The QuEChERS procedure used fairly standard methodology, reduced in scale, for 2 g of the homogenized samples, which were first extracted with 4 mL of ACN, followed by the addition of 0.8 g of NaCl, but with no added MgSO₄. After shaking and centrifugation, 2 mL of the ACN extract was treated with 10 mg of PSA, 20 mg of carbon black (rather than GCB), and 20 mg of Fe₂O₃ magnetic nanoparticles (MNPs). The MNPs were used as an additional cleanup aid, and to facilitate separating the cleanup material from the extract using a magnet. For the DLLME procedure, 1 mL of the ACN extract was added to 4 mL of water and the mixture added to 40 µL of chloroform, the mixture briefly subjected ultrasound mixing, centrifuged, and the CHCl₃ dried with a small amount of MgSO₄. One µL of the CHCl₃ was used for GC-MS analysis of 8 OCPs. The LOQs ranged from 0.15–0.32 µg kg^-1, and EFs ranged from 2–3.

A few final comments need to be presented concerning this last example, which may also pertain to questions concerning other similar published procedures. The authors used granular carbon black, rather than the much more expensive commercial GCB to decolorize the extract, with no observable reduction in extract yield. This may be a useful approach for reduced sized QuEChERS methods, but needs further investigation. The authors also use MNPs as a cleanup aid, and to facilitate removal of PSA and carbon black from the extract, rather than centrifugation. It is unknown whether this method would work on a standard QuEChERS sized extraction, but is an interesting application and should also be further investigated. Most unusual, and completely unexplained, however, is the DLLME procedure. Rather than using standard SA-DLLME conditions (addition of CHCl₃ to ACN and then dispersion formation by addition of the ACN to water), the authors added the ACN directly to water and then added the mixture to the CHCl₃, and then used UA-DLLME to form a dispersion. This unexplained anomaly can also be seen in some of the other examples in Table 10. This unusual modification of DLLME should be compared to SA-DLLME conditions to determine whether it is an acceptable methodology, or not, before further use of this technique for sample preparations. As a final comment, several of the combined sample preparation procedures use chlorinated solvents for extractions. For situations involving a few sample preparations, these solvents may be acceptable, but for large scale operations, even the small amounts of chloroform or other chlorinated solvents should be replaced with greener solvents, since all aqueous and organic solutions, and solids in contact with chlorinated solvents should be treated as hazardous waste, requiring expensive remediation [40].

Choosing an LPME mode for plant or animal-based food sample analysis preparations

There are many factors which must be considered before choosing a food sample preparation method involving liquid extraction—with or without using LPME. In general, these include, but are not limited to: the sample characteristics, the extraction solvent, green requirements, available instrumentation, and whether the analyses are undertaken in the field or in a laboratory. There may also be more than one appropriate sample preparation choices for a specific sample. A simplified example of method choice is illustrated in Figure 3.

Display full size

Figure 3. 

Choosing an LPME mode for plant and animal-based sample preparations. QuEChERS: quick, easy, cheap, effective, rugged, and safe; µ-QuEChERS: micro-QuEChERS; SPE: solid-phase extraction; GC-MS/MS: gas chromatography-tandem mass spectrometry; UHPLC: ultra-high performance liquid chromatography; DLLME: dispersive liquid-liquid microextraction; HS-SDME: headspace-single drop microextraction; DI-SDME: direct immersion SDME; HF-LPME: hollow fiber liquid phase microextraction; PALME: parallel artificial liquid membrane extraction; EME: electromembrane extraction; pEME: parallel electromembrane extraction

Nearly all food analyses require some sample pretreatment, ranging from filtration or centrifugation, to denaturization, lyophilization, or acid degradation. Same analytes, such as elements, aldehydes or acids may also need to be derivatized. Some samples, such as commercial fruit juices or tea effusions, can be analyzed directly by techniques such as SDME or DLLME. This depends on the complexity or the sample matrix, including the presence of fats, proteins, sugars, pigments and other components which might be co-extracted and potentially interfere with the analysis or damage the analytical instrumentation. After a pretreatment step, the number of potential analytes and their physical and chemical characteristics need to be considered, including: are the analytes volatile, or non-volatile, polar or non-polar, ionizable or ionic, organic or inorganic, liquid, oil or solid. The QuEChERS, procedure has the advantages of being suitable for a very wide variety of food samples, providing sample cleanup of fats, proteins and sugars. The downside of the methodology is that little or no analyte concentration enrichment is obtained, necessitating, in most cases, very sensitivive, expensive to purchase, and very expensive to maintain analytical instrumentation. SPE and other traditional sample preparation procedures may overcome some of these deficiencies, but generally are more time and labor intensive. These factors, in addition to the large volumes of hazardous waste potentially resulting from these methodologies, make these techniques most suitable for large volume commercial or governmental analytical laboratories. LPME methods, on the other hand, are simple to use, requiring only standard laboratory supplies, and result in a 10 to more than 100 fold increase of the concentration of target analytes, in addition to sample cleanup, and often allow the use of less sensitive and much less expensive analytical instrumentation, such as GC, GC-MS, HPLC, IC or AAS. The scheme presented in Figure 3 is very general, and omits some protentional sample preparation choices. More detailed and specific selection guides for solvents, sample types and analyte families, as well as the metrics for calculating the relative greenness of an analytical method, are presented in the tables included in this review.