Agarose-Based Gel-Phase Microextraction Technique for Quick Sampling of Polar Analytes Adsorbed on Surfaces.

Sampling and extraction of chemical residues present on flat or curved surfaces as well as touch-sensitive objects are challenging. Hydrogels are characterized by high mechanical flexibility and water content. Thus, they are an ideal medium for transferring water-soluble analytes from a sampled surface to the next stage of an analytical workflow. Here, we demonstrate gel-phase microextraction (GPME), in which disks of blended hydrogels are utilized to lift traces of water-soluble substances adsorbed on surfaces. The protocol has been optimized in a series of tests involving fluorometric and mass spectrometric measurements. Compared with the pure agarose hydrogel, most of the tested blended hydrogels provide a higher efficiency for the sampling/extraction of a model analyte, fluorescein. The blended hydrogel disks are incorporated into three-dimensional (3D)-printed acrylonitrile-butadiene-styrene chips to create easy-to-use sampling probes. We exemplify the suitability of this improved GPME approach in sampling chemical residues present on the skin, glass, and daily use objects. In these tests, the extracts were analyzed on a triple quadrupole mass spectrometer fitted with an electrospray ion source operated in the positive- and negative-ion modes. The method enabled the detection of diclofenac on excised porcine skin fragments exposed to a topical nonsteroidal anti-inflammatory drug and sweat residues (lactic acid) left on surfaces touched by humans. The limits of detection for diclofenac and lactic acid in hydrogel extract were 6.4 × 10-6 and 2.1 × 10-5 M, respectively. In a model experiment, conducted using the presented approach, the amount of lactic acid on a glass slide with fingerprints was estimated to be ∼1.4 × 10-7 mol cm-2.

Introduction

The early steps of most analytical workflows entail sampling real matrices, capturing the target analytes, concentrating the analytes in small volumes, and transferring them to detection systems. Sampling is one of the key steps that affect analytical workflows, and, in some cases, it constitutes a major bottleneck in analytical methodology.[1,2] The sampled matrices can be present in different phases and have different shapes and aspect ratios. It is always necessary to obtain a representative portion of the sampled matrix with minimum perturbation to the sampled objects.[3,4] Perturbation can be limited by decreasing the size of the sampling tool as well as minimizing mechanical and chemical stresses imposed on the sampled object.

Two popular microsampling techniques are solid-phase microextraction[5] and single-drop microextraction.[6] They utilize thin fibers coated with polymers or hanging liquid droplets to capture analytes from the liquid or gas phase. It is yet challenging to collect trace amounts of analytes present on soft, flat, and curved objects. Hydrogel is a biocompatible, hydrophilic, and flexible material. For example, cross-linkable solutions of dextran-methacrylate were used to form hydrogels for sampling amino acids and deoxyribonucleic acid from fingerprints.[7] In previous studies, we implemented agarose hydrogel micropatches to collect low-molecular-weight compounds (metabolites, drugs) from the human skin surface.[8−11] The main component of sweat is water. Water-soluble compounds can readily diffuse to a water-rich hydrogel matrix. However, that procedure involving single-component hydrogel micropatches showed practical limitations related to sampling efficiency and long-term storage of hydrogel probes. The goals of the present study are to improve the sampling efficiency, to verify the possibility of implementing various hydrogels for sampling analytes adsorbed on various surfaces, to increase the maximum storage time of the probes, and to demonstrate alternative applications of the method. Mixing two kinds of polymers is a simple way to produce a blended material with the advantages of both constituents.[12−14] Therefore, in addition to other modifications to the analytical procedure, here, we have combined agarose with other water-soluble polymers. One of the proposed applications concerns about the profiling retention of a topical drug component (diclofenac[15,16]) on the skin. The short sampling time with hydrogel allows one to conduct repeated sampling of the same specimen and temporal profiling of dynamic processes such as absorption of a topical drug by the skin. Another application concerns about the detection of human activity by detecting skin metabolites on surfaces. Lactic acid is among the most abundant metabolites present in human sweat.[10,17,18] Thus, in another application, we aimed to detect lactic acid in fingerprints present on solid surfaces.

Results and Discussion

Figure presents the workflow of this study on gel-phase microextraction (GPME), while method details are provided in the Experimental Section below. For this study, we chose several hydrogel-forming polymers that are relatively biocompatible. In fact, they are used in food-related, pharmaceutical, or cosmetic products. Fluorescein was used as a model analyte in the initial tests because it is soluble in water and has low molecular weight. Hence, it can enter the pores within the hydrogel matrix. It could also be quantified at low concentrations by fluorescence spectrometry (FS). Analyses involving other compounds were conducted by mass spectrometry (MS).

Figure 1

Gel-phase microextraction: (A) preparation of hydrogel disk; (B) tests with FS; (C) tests with MS; (D) sampling porcine skin surface; (E) GPME probes placed on solid surfaces.

Optimization of GPME Sampling

Initially, we performed a number of tests, in which hydrogels with different compositions were submerged in fluorescein solutions (loading) and the absorbed fluorescein was subsequently re-extracted to a solvent mixture (extraction). All of the results for fluorescein concentrations in the extract vs loading and extraction times show similar ascending trends (Figure ). The dependencies exhibit plateaus for the loading times exceeding 20 min and extraction times exceeding 30 min. At long loading times, the hydrogel matrix is saturated with fluorescein and the system is equilibratedAt long extraction times, an equilibrium is established between the hydrogel matrix and the extraction solventMoreover, various kinds of hydrogels exhibit different characteristics of loading analytes to and extraction of analytes from the hydrogel. Most of the blended hydrogels provide higher Cmax,L and Cmax,E than the single-component agarose hydrogel, although blended hydrogels generally provide lower kL and kE (Figure A,B and Table S1). These differences may be due to the different character of intermolecular interactions such as Coulombic repulsion, hydrogen bonding, and polar forces (cf. ref (19)), as well as pore structure. The optimal component ratio for poly(vinylpyrrolidone) (PVP)–agarose hydrogel was 1.5:1.0 and for poly(ethylene glycol) (PEG)–agarose was 3.0:1.0, while for poly(vinyl alcohol) (PVA)–agarose and sodium hyaluronate (HA)–agarose was 2.5:1.0 (Figure C).

Figure 2

Optimization of GPME using a fluorescent marker (fluorescein): (A) effect of extraction time; (B) effect of loading time; and (C) effect of component ratio in blended hydrogel (loading time, 20 min; extraction time, 30 min). Symbols in (A) and (B): (○) 1% agarose; (□) 5% agarose; (△) ratio of 1.0:1.0; (×) ratio of 1.0:1.5; (◇) ratio of 1.0:2.0; (◁) ratio of 1.0:2.5; (☆) ratio of 1.0:2.0. Data points in (A) and (B) are average values from three replicates. Data points in (C) are average values from six replicates. Standard deviations are shown. Note that the results for pure agarose (1 and 5%) in (A) and (B) are the same and are displayed in every graph for comparison.

By comparing the mass spectra of blank extracts from different hydrogel disks, we found out that complex spectral background related to the PVP–agarose hydrogel appears in the m/z range from 800 to 2000, while the spectral background related to the PEG–agarose and PVA–agarose hydrogels is in the m/z range from 500 to 2000 (Figure S1). Thus, using the PVP–agarose hydrogel, one can detect analytes in the m/z range from 20 to 800 with only minor spectral interference. Although the HA–agarose hydrogel exhibits a lower spectral background, the extracted fluorescein concentrations are much lower in this hydrogel than in the other blended hydrogels (Figure C). Therefore, we chose PVP–agarose hydrogel with the component ratio of 1.5:1.0 for further work. The extraction efficiency of fluorescein from the PVP–agarose hydrogel to the liquid phase was estimated to be ∼48%.

To ensure convenient sampling of solid surfaces, the hydrogel disks were incorporated into 3D-printed bases. Sealing the probes in plastic packs and storing them at 4 °C (Figure S2A,B) efficiently minimize water evaporation from the hydrogel for at least two weeks (Figure S2C).

Application of GPME in Kinetic Analysis of Exogenous Residues on the Skin

Characterizing retention of chemical compounds on the skin is important for the development of topical drug and cosmetic products (cf. refs (20−23)). To test retention kinetics of such products, we applied selected commercial products to excised porcine skin and analyzed their residues at different time points. GPME probes were used to lift the target analytes adsorbed on and absorbed by porcine skin periodically. The active ingredient of the tested nonsteroidal anti-inflammatory drug (NSAID) gel is diclofenac sodium. In the negative-ion mode electrospray ionization–mass spectrometry (ESI–MS), diclofenac gives high signals at the m/z 294 and 296 (in addition to others), revealing the characteristic isotope distribution of chlorine (Figure S3). Note that some of the signals are likely related to the excipients present in the NSAID pharmaceutical formulation. The intensity of diclofenac ion at the m/z 294 in the spectra obtained for GPME probes exposed to the porcine skin decreased with incubation time (Figures A and S3). All of the other diclofenac-related ions also decreased with time. However, they were still detectable after 24 h incubation. This long retention of diclofenac on excised porcine skin is in line with the literature discussing retention of diclofenac in humans.[15,16] Most detectable species from the tested cosmetic products were rapidly absorbed by the skin, in some cases, even within one hour (Figures B–D and S4; for blank datasets, see Figure S5). Absorption of specific ingredients is likely affected by the polymers added to the cosmetics.[24] Based on the decay rate constants (k), the analytes displayed dissimilar absorption kinetics in the porcine skin (Table S2). This can be attributed to different physicochemical properties of various compounds and the resulting different affinities to the skin matrix (cf. ref (25)).

Figure 3

Sampling porcine skin with hydrogel disks (two replicates). (A) NSAID gel; (B) moisturizing cream; (C) Nashi blossom moisturizing body lotion; and (D) aloe moisturizing gel. See Figure S4 for other results from this experiment. The extracts were analyzed by electrospray ionization–triple quadrupole mass spectrometry (ESI–QQQ-MS) operated in the positive-ion (+) and negative-ion (−) modes [Q3 scan, average from 1 min extracted ion current (EIC)].

To evaluate quantitative capabilities of the described method, we prepared a calibration plot for diclofenac (Table S3 and Figure S6A). Acetaminophen was used as an internal standard, and blank hydrogel extract was used as the solvent to simulate matrix effect. The limit of detection (LOD) and limit of quantification (LOQ) for diclofenac in the hydrogel extract were 6.41 × 10–6 and 2.14 × 10–5 M, respectively (3.62 × 10–10 and 1.21 × 10–9 mol, respectively). The amount of diclofenac on the porcine skin was estimated to be (2.26 ± 0.42) × 10–7 mole cm–2 at the time point “0 min” (assuming that the extraction efficiency from the hydrogel to the liquid phase was 48%, assuming that the extraction efficiency from the specimen surface to the hydrogel was 100%, and taking into account that the contact area between the hydrogel and skin was 0.28 cm2). The recoveries of diclofenac from the extract of individual hydrogel disks (exposed to the porcine skin smeared with the NSAID gel) were found to be in the range of 121–182%. The excipient ingredients in the NSAID gel are most likely responsible for the observed matrix effect.

In a recently disclosed sampling approach, a polymer-based blotting paper is used to lift topical residues from the skin prior to gas chromatographic analysis.[26] The advantage of using GPME and direct infusion MS for analysis of topical residues is that nonvolatile species can readily be detected.

Imaging Transfer of the Analyte from the Sampled Surface to the Hydrogel Disk

To further investigate the kinetics of GPME sampling, the NSAID gel was spiked with fluorescein (final concentration, 10–3 M). The mixture (0.1 g) was spread onto the porcine skin fragment (∼9 cm2). PVP–agarose (1:5:1.0) hydrogel disks (without the base) were affixed onto the porcine skin. The disks were removed periodically, placed under stereomicroscope, and cut in half using a sharp metal blade. During the exposure of the hydrogel disks to the porcine skin, fluorescein molecules diffused into the hydrogel matrix. A band of these fluorescent species is clearly visible near the edge of the hydrogel that contacted the skin surface (Figure A). The G values for pixels situated along six lines perpendicular to the disk edge were averaged, the obtained pixel intensity versus distance profiles were smoothed by the Savitzky–Golay algorithm (Figure B), and the areas under the curves (AUCs) were computed (Figure C). The relation between AUC and sampling time (t, in minutes) is described by the equationIt is striking that, after 50 min exposure of hydrogel disks to the porcine skin, the absorbed fluorescein was evenly distributed within the hydrogel matrix. However, it should be noted that this sampling/extraction time was longer than the loading time selected based on the previous experiments (Figure ; with hydrogel disks submerged in the solvent, in the presence of shaking). It is understandable that molecular transport between the solution within the hydrogel matrix and the external environment is faster if the hydrogel disk is surrounded by the solvent. Nevertheless, to expedite the sampling process, the loading/sampling time in the following experiments was set to 20 min.

Figure 4

Diffusion of fluorescein from excised porcine skin into the hydrogel probe during sampling. (A) Red–green–blue (RGB) images of the hydrogel disk cross-sectional area viewed under a stereomicroscope. Scale bar: 1 mm. (B) Relation between average G value and pixel positions at different sampling times. The average G values were calculated based on the G values of pixels along the six yellow vertical lines in (A) (one replicate is shown). (C) Relation between the area under curve (AUC) in (B) and sampling time (n = 3; error bars represent standard deviation).

Application of GPME in the Analysis of Residues Present on Solid Surfaces

A potentially interesting application of GPME is tracing human skin metabolites present on surfaces with fingerprints. Every time we touch a surface, we leave small amounts of endogeneous and exogenous compounds.[27] We have applied the developed GPME probes to collect traces of molecules from the daily use objects such as computer keyboard, mouse, and laboratory bench. Subsequently, we implemented MS to analyze the resulting extracts from hydrogel disks. The results show that it is possible to detect various chemical species present on those target surfaces (Figure ). The blank result was prepared from a clean glass slide, sampled by the GPME probe, and displayed as sample number 1. In one experiment, three volunteers were asked to deposit fingerprints on clean glass slides for 5 min and the fingerprints were sampled by GPME probes (sample numbers 2, 3, and 4). For example, the MS signals at the m/z 89, 171, and 247 were higher in the control and real sample than in the blank. The signal at the m/z 89 was identified as lactic acid (for product ion scans, see Figure S7). The same signal appeared in the samples obtained from daily use items, such as bench and computer peripherals. While identification of other signals related to the species lifted from solid surfaces by GPME is out of scope of this study, in principle, such untargeted analysis could be carried out with the aid of a high-resolution mass spectrometer.

Figure 5

Sampling flat surfaces by GPME followed by MS analysis. Sampled surfaces: (1) clean glass slide; (2) fresh human fingerprint on the glass slide (first replicate); (3) fresh human fingerprint on the glass slide (second replicate); (4) fresh human fingerprint on the glass slide (third replicate); (5) backspace key on the computer keyboard; (6) E key; (7) enter key; (8) J key; (9) space bar; (10) computer mouse left button; (11) mouse right button; and (12) laboratory bench. Legend: black, m/z 89 (−); gray, m/z 171 (−); white, m/z 247 (−). The extracts were analyzed by ESI–QQQ-MS operated in the positive- and negative-ion modes (Q3 scan, average from 1 min EIC; n = 3; error bars represent standard deviation).

Because the extracts of hydrogel disks that were in contact with daily use objects do not contain a complex matrix (low matrix effect in ESI–MS), the recoveries for lactic acid were in an acceptable range (106–118%; Table S3 and Figure S6B). Thus, in some cases, the method could be applied for quantitative determinations. For that purpose, we prepared a calibration plot for lactic acid (Table S3 and Figure S6B). Threonine was used as an internal standard, and blank hydrogel extract was used as the solvent to simulate the matrix effect. The LOD and LOQ for lactic acid in the hydrogel extract were 2.10 × 10–5 and 7.00 × 10–5 M, respectively (1.19 × 10–9 and 3.96 × 10–9 mol, respectively). The amount of lactic acid on the glass slide with fingerprints was estimated to be (1.43 ± 0.08) × 10–7 mol cm–2 (assuming that the extraction efficiency from the hydrogel to the liquid phase was 48%, assuming that the extraction efficiency from the specimen surface to the hydrogel was 100%, and taking into account that the contact area between the hydrogel and skin was 0.28 cm2). Note that the properties of the support of the sampled analytes (surface hydrophilicity) should affect the sampling efficiency. However, the anticipated application of this method variant is in qualitative and semiquantitative analyses, for example, to confirm human activity in a certain area revealed by residues of skin excretions (here, lactic acid) left by humans on daily use objects.

Conclusions

This study presents the use of blended hydrogel as a medium to transfer the water-soluble analytes from the target surface to the detection step. The composition of the hydrogel matrix has been optimized in a series of experiments. The PVP–agarose hydrogel with the component ratio of 1.5:1.0 provided better loading efficiency than other tested hydrogels. By incorporating hydrogel disks into a 3D-printed base, one can easily produce a tool (GPME probe) for sampling water-soluble analytes from various surfaces. The GPME technique can be utilized to study the retention kinetics of drug/cosmetics on porcine skin. It can also be used to sample chemical residues left by humans on furniture and appliances. In perspective, this noninvasive sampling method may also be applied for collecting clinical and veterinary specimens containing skin excretions, topical drug residues, or toxic compounds adsorbed on foodstuffs. Taking into account the rapid progress in automation of MS,[28] and the requirement for fast sample processing and analysis, it is appealing to combine the proposed GPME technique with automated MS workflows.

Experimental Section

Materials

Acetaminophen, fluorescein, l-threonine, and poly(vinylpyrrolidone) (PVP) (average molecular weight 40 000 u) were from Sigma-Aldrich (St. Louis, MO). Ammonium acetate and poly(vinyl alcohol) (PVA) (98%, hydrolyzed, molecular weight 31 000–50 000 u) were from Acros Organics (Geel, Belgium). Poly(ethylene glycol) (PEG) (average molecular weight 6000 u) was from Alfa Aesar (Haverhill, MA). Sodium hyaluronate (HA) (eye-drop grade) was from Seedchem (Miaoli, Taiwan). Agarose was from UniRegion Bio-Tech (New Taipei City, Taiwan). Ethanol was from Echo Chemical (Miaoli, Taiwan). Water (liquid chromatography–mass spectrometry grade) was from Merck (Darmstadt, Germany). Diclofenac sodium salt was from TCI (Tokyo, Japan). dl-lactic acid lithium salt was from MP Biomedicals (Santa Ana, CA).

A number of pharmaceutical and cosmetic products applied topically were purchased from a local pharmacy shop (Hsinchu, Taiwan). They include (i) a nonsteroidal anti-inflammatory drug (NSAID), gel-like, for treatment, control, and prevention of inflammation (ingredients: diclofenac sodium, propylene glycol, l-menthol, triethanolamine, isopropyl alcohol, carbopol, and purified water); (ii) a moisturizing cream (ingredients: water, dicaprylyl ether, ethylhexyl palmitate, glycerin, urea, cetearyl alcohol, glyceryl stearate, Prunus amygdalus dulcis (sweet almond) oil, sucrose polystearate, hydrogenated polyisobutene, biosaccharide gum-1, panthenol, Butyrospermum parkii (shea) butter, propylene glycol, Chamomilla recutita (matricaria) flower extract, Aloe barbadensis leaf juice, Calendula officinalis flower extract, tocopheryl acetate, citric acid, phenoxyethanol, caprylyl glycol, sodium stearoyl glutamate, sodium polyacrylate, triacetin, bisabolol, lecithin, tocopherol, ascorbyl palmitate, hydrogenated palm glycerides citrate, parfum (fragrance), and xanthan gum); (iii) Nashi blossom moisturizing body lotion (ingredients: water, caprylic/capric triglyceride, isononyl isononanoate, glycerin, propanediol, sodium acrylate/sodium acryloyldimethyl taurate copolymer, mineral oil, methylparaben, PEG-40 hydrogenated castor oil, phenoxyethanol, glycosyl trehalose, panthenol, hydrogenated starch hydrolysate, trideceth-6, allantoin, sodium polyacrylate, hydroxypropyl guar, disodium ethylenediaminetetraacetic acid (EDTA), betaine, tocopheryl acetate, xanthan gum, ethylhexylglycerin, squalane, and fragrance); (iv) red rose moisturizing body lotion (ingredients: water, caprylic/capric triglyceride, isononyl isononanoate, glycerin, propanediol, sodium acrylate/sodium acryloyldimethyl taurate copolymer, mineral oil, methylparaben, PEG-40 hydrogenated castor oil, phenoxyethanol, glycosyl trehalose, panthenol, hydrogenated starch hydrolysate, trideceth-6, allantoin, sodium polyacrylate, hydroxypropyl guar, disodium EDTA, betaine, tocopheryl acetate, xanthan gum, ethylhexylglycerin, squalane, and fragrance); and (v) aloe moisturizing gel (ingredients: water, 1,3-propanediol, phenoxyethanol/caprylyl glycol, acrylates/c10-30 alkyl acrylate crosspolymer, PEG-40 hydrogenated castor oil, urea, fragrance, methyl gluceth-20, glycerin/Prunus serrulata flower extract, A. barbadensis leaf juice, ethylenediaminetetraacetic acid disodium salt, allantoin, sodium hyaluronate, and sodium hydroxide).

Preparation of Blended Hydrogel Disks

The blended hydrogel disks were prepared with five kinds of polymers: PVP, PEG, PVA, HA, and pure agarose. The binary mixture solution was prepared by dissolving agarose and another polymer in water. The weight percentage of agarose in the hydrogel was in the range from 1.2 to 2.4 wt %, and the concentration of another added water-soluble polymer was from 2.4 to 3.6 wt %. The total polymer percentage in the aqueous solution was fixed at 4.8 wt % (following ref (13)). The mixture was gradually heated to ∼90 °C in a water bath to facilitate dissolution of the polymers. During this process, the appearance of the mixture changed from turbid to transparent. Subsequently, 4 mL of the solution was pipetted into a glass Petri dish (⌀ = 5 cm) and it was let to cool down to room temperature (∼22 ± 3 °C). After ∼10 min, the sol phase underwent transition to the gel phase. A glass tube (I.D., 0.6 cm) was used to cut disks out of the hydrogel mass (⌀ = 0.6 cm, h = 0.2 cm) (Figure A). In the application-oriented experiments, the hydrogel disks (2 or 3) were embedded in a miniature holder (⌀ = 0.6 cm, h = 0.2 cm), fabricated by a 3D printer (UP Plus 2; Tiertime Technology, Beijing, China) with an acrylonitrile–butadiene–styrene polymer filament, to produce GPME probes for sampling solid matrices (Figure S8). The blended hydrogels were inserted into the probe base. The probes were covered with glass slides (18 × 18 mm2), packed in low-density poly(ethylene) bags with the aid of a vacuum sealer (FM2000; Foodsaver, Oklahoma City, OK), and refrigerated at 4 °C (Figure S2A).

Extraction Protocol

In the early experiments, individual hydrogel disks were incubated in 10–5 M fluorescein solution in 10 vol % ethanol (Figure B). In the later experiments, involving mass spectrometric detection (Figure C), the following solid objects were sampled: porcine skin, glass slides with fresh fingerprints, laboratory bench surface, computer keyboard, and mouse (Figure D,E). An exposed hydrogel disk was transferred into a 1.5 mL microcentrifuge tube by tweezers. An aliquot of the extraction solvent was pipetted into the tube to re-extract the hydrogel disk. The extraction solvent was 20 mM ammonium acetate in 10 vol % ethanol in the case of FS and 20 mM ammonium acetate in 50 vol % ethanol in the case of MS. The volumes of the extraction solvent were 1 mL for FS and 0.5 mL for MS. To improve the re-extraction efficiency, the tube was placed in a thermoshaker (ThermoMixer C; Eppendorf, Hamburg, Germany) set to 37 °C and 300 rpm. Moreover, in the case of FS, the extracts were diluted 3× with 20 mM ammonium acetate (in 10 vol % ethanol) to fill up the standard quartz cuvette (optical path length, 10 mm).

Freshly excised porcine skin was obtained from a local morning market (acquired from a slaughterhouse on the same day; Hsinchu, Taiwan). It was rinsed with tap water for 5 min, then deionized water for 5 min, and subsequently cut into smaller fragments of ∼9 cm2 using scissors. The porcine skin fragments were stored at −80 °C and used for experiments during 7 days. Immediately before every experiment, the skin fragments were cleaned again in deionized water and wiped with cellulose tissue soaked in 70 vol % ethanol. The fragments were then placed on the top of a 0.5 wt % agarose hydrogel layer in a glass Petri dish and incubated at 30 °C for 1 h. The purpose of the agarose hydrogel layer under the skin was to assure skin hydration during the experiment (Figure D). The temperature of 30 °C was chosen because it is close to the average human skin temperature.[29] We applied 0.1 g of a selected pharmaceutical or cosmetic product onto the surface of porcine skin and spread it uniformly with a plastic spoon. The porcine skin was repeatedly sampled with GPME probes for 20 min at different time points following application of the product. Hydrogel disks were transferred from the probes to 1.5 mL microcentrifuge tubes by tweezers, and 0.5 mL of extraction solvent was pipetted into the tubes. The tubes were then placed in a thermoshaker (ThermoMixer C) set to 37 °C and 300 rpm for 30 min to re-extract analytes from the hydrogel disks. The consecutive samplings were performed on different (previously unsampled) skin areas to avoid the influence of analyte depletion on the result.

Fluorescence Spectrometry

The extracts containing fluorescein were analyzed by means of the LS55 fluorescence spectrometer from PerkinElmer (Waltham, MA) using FL WinLab (version 4.00, PerkinElmer, Waltham, MA). The excitation wavelength was 405 nm (slit, 10 nm). The emitted light was scanned from 470 to 650 nm (slit, 10 nm). The scan speed was 180 nm min–1. The final values were average emission intensities for the wavelengths from 506 to 516 nm. Fluorescein concentrations in the extracts were determined by building calibration plots for six concentration levels, always on the same day. The coefficients of determination (R2) for these calibration plots ranged from 0.992 to 0.999.

Mass Spectrometry

The extracts of complex matrices were analyzed by means of an LCMS-8030 triple quadrupole mass spectrometer (Shimadzu, Kyoto, Japan) using LabSolutions software (version 5.82). The electrospray ionization (ESI) interface was operated with a syringe pump (Legato Nano; KD Scientific, Holliston, MA) infusing the sample at a flow rate of 30 μL min–1. The nebulizing gas (nitrogen) flow rate was 2 L min–1, while the drying gas (nitrogen) flow rate was 15 L min–1. The desolvation line temperature was 250 °C, while the heating block temperature was 400 °C. The voltages of the ESI needle were +4.5 and −3.5 kV in the positive- and negative-ion modes, respectively. Depending on the experiment, the instrument was operated in Q3 scan (m/z range, 20–2000), product ion scan, or multiple reaction monitoring (MRM) mode. In the case of product ion scan and MRM modes, different collision energies were tested: −10, −20, and −30 V for positive ions and 10, 20, and 30 V for negative ions. MRM transitions: 294.00 → 250.00 for diclofenac, 152.00 → 110.00 for acetaminophen, 89.00 → 43.00 for lactic acid, 120.00 → 74.00 for threonine. Instrumental blank signals were recorded for ∼2 min, followed by recording sample signals for ∼2 min. The extracted ion currents were obtained for 1 unit intervals in the m/z scale. The final values correspond to the average of extracted ion currents from 1 min recording.

Imaging Diffusion of Fluorescein in Hydrogel Disks

RGB images of the hydrogel disk cross-sectional area were viewed under a stereomicroscope (SMZ745T; Nikon, Tokyo, Japan) with a 2× objective and a 10× eyepiece. A blue light-emitting diode (wavelength range: 410–500 nm), positioned at an angle of 60° and at a height of 1.5 cm, was used for excitation. The microscope objective was covered with three layers of a dark yellow-green filter (Lee Filters lighting gel sheet 090, wavelength range 460–590 nm; Knight Sound & Lighting, Cuyahoga Falls, OH). The exposure time and ISO value set in the digital camera (EM-1, Olympus, Tokyo, Japan) were 2.5 s and 100, respectively.

Collection of Fingerprint Residues

Three volunteers were asked to deposit their fingerprints by pressing fingertips on clean glass slides for 5 min. Note that the surface of this glass is hydrophilic (contact angle for water, 76°). In other tests, residues from various solid surfaces (bench, computer peripherals) were sampled. GPME probes were placed on the sampled surface for 20 min. Then, hydrogel disks were transferred from the probes to 1.5 mL microcentrifuge tubes by tweezers and 0.5 mL of extraction solvent was pipetted into the tubes. The tubes were placed in a thermoshaker (ThermoMixer C) set to 37 °C and 300 rpm for 30 min to re-extract analytes from the hydrogel disks.

Data Treatment

The experimental datasets were processed using OriginPro 8 software (OriginLab, Northampton, MA). The FS datasets were fitted with exponential functions to determine rate constants (min–1) of loading (kL) and extraction (kE)where CL and CE are the experimentally determined concentrations of fluorescein in the final extracts (moles per liter), Cmax,L and Cmax,E are the maximum concentrations of fluorescein for loading and extraction datasets (moles per liter), respectively, and t is the time (min).

In an experiment designed to study retention kinetics of cosmetic and pharmaceutical products on porcine skin, the MS datasets were fitted with the equationwhere I is the ion intensity recorded by MS (Q3 scan, average of extracted ion current (EIC) from 1 min), I0 is the ion intensity extrapolated to t = 0 h, k is the decay rate constant (h–1), t is the time (h), and b is the analysis baseline.

For further experimental details (materials, FS, MS, imaging diffusion of fluorescein, and collection of fingerprint residues), please, see the Supporting Information.