Flexible Potentiometric Sensor System for Non-Invasive Determination of Antioxidant Activity of Human Skin: Application for Evaluating the Effectiveness of Phytocosmetic Products

: In contemporary bioanalysis, monitoring the antioxidant activity (AOA) of the human skin is used to assess stresses, nutrition, cosmetics, and certain skin diseases. Non-invasive methods for skin AOA monitoring have certain advantages over invasive methods, namely cost-effectiveness, lower labor intensity, reduced risk of infection, and obtaining results in the real-time mode. This study presents a new ﬂexible potentiometric sensor system (FPSS) for non-invasive determination of the human skin AOA, which is based on ﬂexible ﬁlm electrodes (FFEs) and membrane containing a mediator ([Fe(CN) 6 ] 3–/4– ). Low-cost available materials and scalable technologies were used for FFEs manufacturing. The indicator FFE was fabricated based on polyethylene terephthalate (PET) ﬁlm and carbon veil (CV) by single-sided hot lamination. The reference FFE was fabricated based on PET ﬁlm and silver paint by using screen printing, which was followed by the electrodeposition of precipitate containing a mixture of silver chloride and silver ferricyanide (SCSF). The three-electrode conﬁguration of the FPSS, including two indicator FFEs (CV/PET) and one reference FFE (SCSF/Ag/PET), has been successfully used for measuring the skin AOA and evaluating the impact of phytocosmetic products. FPSS provides reproducible (RSD ≤ 7%) and accurate (recovery of antioxidants is almost 100%) results, which allows forecasting its broad applicability in human skin AOA monitoring as well as for evaluating the effectiveness of topically and orally applied antioxidants.


Introduction
The skin is the largest organ of the human body and performs several vital functions, such as protective, sensory, thermoregulatory and others. Similar to any other organ, the skin is exposed to reactive species (RS), which may exist as radicals and nonradicals [1,2]. In physiological concentrations, RS impact cellular metabolism processes, such as protection against infectious pathogens, intracellular and intercellular signaling, redox regulation, and they are neutralized by antioxidants [1][2][3]. Various environmental factors (e.g., radiations, pollutants) and physiological factors, such as unhealthy lifestyle or unbalanced exercise, lead to RS overproduction and may change the redox homeostasis of the skin [1,2,4]. In pathological (elevated) concentrations, RS can cause irreversible changes in cellular compartments, inflammation, weakening of immune functions, and tissue degradation [1,2]. In addition, current research shows a link between high levels of RS, aging [1,2,5], and skin diseases [1,2,6]. The use of antioxidants in topical cosmeceuticals and oral nutraceuticals aims to achieve and maintain the redox balance of the skin [2,7]. For this reason, methods and sensors that measure the antioxidant status of the skin have It is known that the reproducibility of tactile skin sensors is related to the following factors: effective sampling, i.e., transport of the analyte to the sensor surface [33]; sensor contamination [33,34]; skin condition variability [34]; and poor contact resulting in an unreliable analytical signal [35]. This paper adopts three strategies aimed at improving the result reproducibility when determining the human skin AOA by the CHPM. Following the first strategy, the proposed sensor system should be disposable, which eliminates its contamination during repeated use and makes it safe for a respondent. The second strategy focuses on the need to create a three-electrode sensor system configuration, which enables receiving two results during one measurement, thus eliminating a possible impact of time-varying skin conditions. The third strategy implies the use of flexible film electrodes (FFEs) as part of the sensor system, which ensures better contact at the heterogeneous electrode/membrane with [Fe(CN) 6 ] 3-/4-/skin interface and contributes to the measurement of a stable analytical signal (potential) under load conditions (local pressure). Finally, a carbon veil (CV) was used for fabricating the indicator FFE. Since the CV displays such features as electrical conductivity, softness, and flexibility, it is these features that explain the CV popularity in applied electrochemistry. The CV can be used for manufacturing electrodes for batteries [36], capacitors [37], microbial fuel cells [38], and generator cells [39]; heating elements [40]; and sensors for the determination of ascorbic acid [41] and nitrite ions [42]. The analysis of Web of Science and Scopus publications has not revealed any case reporting the use of CV for monitoring the AOA of biological objects.

Materials
A polymer film based on polyethylene terephthalate (PET), 0.25 mm thick (Fellowes Inc., Itasca, IL, USA) was used as a substrate for FFEs fabrication. CV based on polyacrylonitrile (M-Carbo Ltd., Minsk, Belarus) was used in the preparation of the indicator FFE. CV specifications are given in Table S1 of the Supplementary Materials. Silver conductive paint Mechanic DJ912 (Shenzhen Welsolo Electronic Technology Co., Ltd., Guangdong, China) was used in the preparation of the reference FFE. Cementit Universal (Merz + Benteli AG, Niederwangen, Switzerland) was used as an insulator. Microporous film MFAS-OS-2 based on cellulose acetate (CJSC STC Vladipor, Vladimir, Russia) was used as a membrane for [Fe(CN) 6 ] 3-/4-. Five Russian-made phytocosmetic products were studied in the present work: cream-mousse, cream-gel, serum, day cream, and nourishing night cream. Their specifications are shown in Table S2 of the Supplementary Materials.

Apparatus
The LM-260iD laminator (Rayson Electrical MFG., Ltd., Guangdong, China) was used for single-sided hot lamination. The SPR-10 manual printer (DDM Novastar Inc., Ivyland, PA, USA) was used for screen printing. The IVA-5 inversion voltammetric analyzer (IVA Ltd., Yekaterinburg, Russia) was used for the potentiostatic electrodeposition of a precipitate containing a mixture of silver chloride and silver ferricyanide (SCSF). Two portable PA-S potentiometric analyzers (Ural State University of Economics, Yekaterinburg, Russia) were used for potentiometric measurements. A focused ion beam scanning electron microscope Auriga CrossBeam (Carl Zeiss NTS GmbH, Oberkochen, Germany) and Ultim Max detector (Oxford Instruments plc., Abingdon, UK) were used for scanning electron microscopy (SEM) and energy-dispersive X-ray spectroscopy (EDS), respectively. An ultrasonic liquid processor VCX 750 (Sonics & Materials Inc., Newtown, CT, USA) with a stepped microtip (Ø 2 mm) was used to produce emulsions. A digital multimeter Owon B41T+ (Fujian Lilliput Optoelectronics Technology Co., Ltd., Zhangzhou, China) with a thermocouple was used for the temperature control of emulsions. A pediatric sphygmomanometer LD-80 (Little Doctor Int. (S) Pte. Ltd., Yishun, Singapore) with 18-26 cm cuffs was used to fix the sensor system on the selected skin area. Installation Akvalab-UVOI-MF-1812 (JSC RPC Mediana Filter, Moscow, Russia) was used to obtain deionized water with a resistivity of 18 MΩ × cm.

FFEs Manufacturing
FFEs were prepared by applying scalable technologies [43]. The indicator FFE was prepared by applying the single-sided hot lamination technology described in [41,42] and in the Supplementary Materials. The manufactured electrode was marked CV/PET and was used as the indicator FFE of the potentiometric sensor system. The size of one CV/PET was 3 × 35 mm. The CV/PET was modified with gold nanoparticles (AuNPs) in order to improve its potential stability. The procedure of AuNPs synthesis and the technique of electrode modification are identical to those described in [30].
The reference FFE was manufactured on the basis of PET and silver paint using the screen-printing method, with the subsequent electrodeposition of a SCSF in the potentiostatic mode. The applied silver paint was oven-dried at 110 • C for 30 min. The silver electrodes were cooled to room temperature and then cut to the required dimension 3 × 35 mm. The middle part of the silver electrodes, which separated the working and contact zones, was covered with a mixture of insulator and acetone in a ratio of 1 to 5 by volume. The geometric area of the working area of one electrode was 6-9 mm 2 (3 × 2-3 mm). Modification of the silver electrode included the electrodeposition of a SCSF on the working surface, according to the technique described in [44]. For that purpose, the silver electrode was inserted in a continuously stirred buffer solution pH 5 (see Table S3 in Supplementary Materials) that contained 1 mM K 3 [Fe(CN) 6 ] and 0.05 mM K 4 [Fe(CN) 6 ], and then, it was polarized at a constant potential of 0.325 V (vs. Ag/AgCl/KCl, 3.5 M) for 120 s. The resulting electrode was labeled SCSF/Ag/PET and used as a reference FFE in a potentiometric sensor system.

Assembly of the Potentiometric Sensor System
The standard two-electrode configuration of the potentiometric sensor system includes one indicator electrode and one reference electrode, which are in contact with a membrane containing [Fe(CN) 6 ] 3-/4- [29]. In this study, prior to measurements, the membrane was kept in a buffer solution pH 5 (see Table S3 in Supplementary Materials) containing 1 mM K 3 [Fe(CN) 6 ] and 0.05 mM K 4 [Fe(CN) 6 ] for 3-5 min. The choice of the solution composition is due to the fact that it has sufficient electrical conductivity and its pH corresponds to the physiological pH value of healthy human skin [45]. An increase in the number of indicator electrodes in the potentiometric sensor system led to a change in its electrode configuration. The three-electrode configuration of the potentiometric sensor system was first proposed to determine the AOA of human skin in [30]. Figure 1a illustrates the relevant process. This three-electrode configuration where E 1 = E 3 = indicator electrodes (in our work, it is CV/PET) and E 2 = reference electrode (in our work, it is SCSF/Ag/PET), allows us to obtain two results simultaneously and correctly evaluate the reproducibility of the measurement. This approach is important, since the human skin is a complex analytical object and its parameters, including AOA, may change over time affected by various external and internal factors. To minimize the impact of a variable skin condition on the results, the time interval between repeated measurements should tend to zero. The use of a three-electrode configuration of a potentiometric sensor system meets this condition, since two results can be obtained during one measurement. In the present study, all components of the sensor system were flexible ( Figure 1b); hence, it was called the "flexible potentiometric sensor system" (FPSS).
Chemosensors 2021, 9, x FOR PEER REVIEW 5 of 14 a complex analytical object and its parameters, including AOA, may change over time affected by various external and internal factors. To minimize the impact of a variable skin condition on the results, the time interval between repeated measurements should tend to zero. The use of a three-electrode configuration of a potentiometric sensor system meets this condition, since two results can be obtained during one measurement. In the present study, all components of the sensor system were flexible ( Figure 1b); hence, it was called the "flexible potentiometric sensor system" (FPSS).

Model Conditions
In the model conditions, the membrane containing [Fe(CN)6] 3-/4-was placed on the table surface atop an inert (fluoroplastic) film. FFEs were applied to the membrane. Then, a steel 0.5 kg weight (base diameter = 36 mm) was set on top. The pressure exerted on the FPSS was approximately 4.82 kPa or 36.1 mm Hg.

Determining the AOA of Volunteers' Skin
Six young non-smoking women aged 24-28, with skin phototype II-III (hereinafter referred to as volunteers) were involved in the present study. The skin phototype of each volunteer was determined using the Fitzpatrick scale as described in [46,47]. Five volunteers participated in the skin AOA analysis in order to determine analytical characteristics of the developed FPSS. One volunteer was involved in evaluating the effectiveness of commercially available phytocosmetic products. The exclusion criteria were malnutrition, the intake of dietary supplements 7 days before the start of the measurements, and the use of cosmetic products other than those provided for the present experiment on the day of the analysis.
The three-electrode FPSS configuration was attached to the selected area of the skin (inside hand from the palm to the elbow) in such a way that the impregnated [Fe(CN)6] 3-/4-membrane was in contact with the skin surface, and FFEs were in contact with the membrane. The FPSS was fixed on the selected area of the arm with the help of a sphygmomanometer cuff, which was inflated to low pressure (35-40 mm Hg). This pressure was maintained during the entire measurement (10 min). Skin antioxidants diffused to the membrane, where they interacted with the oxidized form of the mediator by reaction (1):

Model Conditions
In the model conditions, the membrane containing [Fe(CN) 6 ] 3-/4was placed on the table surface atop an inert (fluoroplastic) film. FFEs were applied to the membrane. Then, a steel 0.5 kg weight (base diameter = 36 mm) was set on top. The pressure exerted on the FPSS was approximately 4.82 kPa or 36.1 mm Hg.

Determining the AOA of Volunteers' Skin
Six young non-smoking women aged 24-28, with skin phototype II-III (hereinafter referred to as volunteers) were involved in the present study. The skin phototype of each volunteer was determined using the Fitzpatrick scale as described in [46,47]. Five volunteers participated in the skin AOA analysis in order to determine analytical characteristics of the developed FPSS. One volunteer was involved in evaluating the effectiveness of commercially available phytocosmetic products. The exclusion criteria were malnutrition, the intake of dietary supplements 7 days before the start of the measurements, and the use of cosmetic products other than those provided for the present experiment on the day of the analysis.
The three-electrode FPSS configuration was attached to the selected area of the skin (inside hand from the palm to the elbow) in such a way that the impregnated [Fe(CN) 6 ] 3-/4membrane was in contact with the skin surface, and FFEs were in contact with the membrane. The FPSS was fixed on the selected area of the arm with the help of a sphygmomanometer cuff, which was inflated to low pressure (35-40 mm Hg). This pressure was maintained during the entire measurement (10 min). Skin antioxidants diffused to the membrane, where they interacted with the oxidized form of the mediator by reaction (1): indicator electrode and served as an analytical signal. The skin AOA (in mol-eq L -1 = M-eq) was calculated by Formula (2): where C Ox and C Red (in mol L -1 = M)-respective concentrations of K 3 [Fe(CN) 6 ] and K 4 [Fe(CN) 6 ] in the membrane; ∆E = E 2 -E 1 -the change in the potential of the indicator electrode for 10 min from the initial (E 1 , V) to the final (E 2 , V) value; F = 96485.332 C mol -1 -the Faraday constant; R = 8.314 J K -1 mol -1 -absolute gas constant; T (in K) = T (in • C) + 273.15-absolute temperature.

Analysis of Phytocosmetic Products
The effectiveness of phytocosmetic products (see Table S2 in Supplementary Materials) was measured in vitro and in vivo assays, applying the hybrid potentiometric method and CHPM respectively (see Table S6 in Supplementary Materials).
In vitro, AOA of phytocosmetic products was determined by the method described in [27] with some modifications. A sample of the phytocosmetic product was mixed with a pH 5 buffer solution containing 1 mM K 3 [Fe(CN) 6 ] and 0.01 mM K 4 [Fe(CN) 6 ] by ultrasound (frequency 20 kHz; amplitude 250 µmat; efficiency 20%) for 2 min. As a result, emulsions were obtained. The formation of emulsions by ultrasound was accompanied by their heating. The heated emulsions were cooled in a water bath to 25 ± 1 • C. The temperature was monitored by using a thermocouple of a multimeter lowered in the tube containing the emulsion. The emulsion AOA was expressed per 1 g of the initial phytocosmetic product according to Formula (3): where V (in L)-the volume of the obtained emulsion; m (in g)-the weight of the analyzed sample of a phytocosmetic product. In vivo, the skin AOA was measured without applying (control) and after applying the phytocosmetic product. A three-electrode configuration of FPSS was used for measurements, and the computation of the skin AOA was completed applying Formula (2). The effectiveness of the analyzed phytocosmetics product was estimated as the deviation of the skin AOA on the area with the applied phytocosmetic product relative to the control area by Formula (4): where AOA (skin+ phytocosmetics) -AOA of volunteer's skin with an applied phytocosmetic product; AOA skin -AOA of volunteer's skin without a phytocosmetic product (control).

Statistical Analysis
The measurements were performed in 2-6-fold replication (4-6 times in the model conditions and 2 times on the human skin). Statistical analysis was performed in Microsoft Excel 2010 with an accepted significance level of 0.05. The data are presented as X ± ∆X, where X is the mean value and ∆X is standard deviation. FPSS validation was performed by the spike recovery test with the use of non-enzymatic antioxidants present in the human skin. The recovery of non-enzymatic antioxidants was determined in accordance with IUPAC recommendations [48]. The relative standard deviation was used to determine the reproducibility of the measurement results. The correlation analysis was performed by calculating the Pearson correlation coefficient.

FFEs Study
Measuring circuit in the Figure 1a, where E 1 = E 2 = E 3 = CV/PET or AuNPs/CV/PET or SCSF/Ag/PET, was used to study the stability of the FFEs under the model conditions. The results are presented in Table 1. The time of establishing the potentials of all studied FFEs is less than the duration of measuring skin AOA (10 min or 600 s). However, the CV/PET has a potential of better reproducibility and a shorter period of its stabilization as compared with the electrode modified with gold nanoparticles (AuNPs/CV/PET). Unlike the carbon screen-printed electrode with a flat surface, whose modification with gold nanoparticles led to its better stability [30], the CV/PET surface is three-dimensional and consists of randomly coupled carbon fibers with a diameter of 5-10 µm and an unevenly distributed polyester binder (see Figure S1 in Supplementary Materials and [41,42]). The three-dimensional structure of the CV/PET excludes the formation of a surface flat layer containing AuNPs, since AuNPs penetrate through pores into the inner layers of the CV. This results in a mixed potential detecting AuNPs and CV simultaneously. Due to the registration of a mixed potential, the AuNPs/CV/PET demonstrates a worse potential stability than the CV/PET. The characteristics of the electrode that differ from the SCSF/Ag/PET by the non-conductive substrate (alumina ceramics), by SEM, were described in [44]. It was shown that small Ag 3 [Fe(CN) 6 ] crystals fill the pores between large AgCl crystals, which leads to a greater stability of the potential of the electrode with a mixed precipitate (Ag 3 [Fe(CN) 6 ] + AgCl) as compared with a silver chloride electrode (AgCl only) in a medium (solution, membrane) containing [Fe(CN) 6 ] 3-/4-. Since the non-electrically conductive substrate does not affect the potentiostatic electrodeposition of the precipitate, the results presented in [44] are fully valid for the SCSF/Ag/PET. In further studies, the CV/PET and SCSF/Ag/PET were used as the indicator and the reference electrodes of the FPSS, respectively.

FPSS Testing in Model Conditions
The three-electrode configuration of the FPSS (E 1 = E 3 = CV/PET and E 2 = SCSF/Ag/PET in Figure 1a) was tested under the model conditions to confirm its analytical effectiveness. Figure 2 illustrates the dependence of the FPSS potential on the logarithm of the concentration ratio of the oxidized and reduced forms of the mediator in the membrane, which is linear with an approximation coefficient R 2 > 0.99. The slope of this dependence is approximately 55 mV, which is in good agreement with the pre-logarithmic coefficient in the Nernst equation for a single-electron process, which is equal to 59 mV at 25 • C [49]. The content of the non-enzymatic antioxidants in the human skin decrease in lowing sequence: ascorbic acid > uric acid > reduced glutathione > α-tocopherol uinol 10 [11]. Due to normal aging and photoaging, the content of ascorbic acid, r glutathione, and a-tocopherol decreases, while the level of uric acid is relativel [12]. The non-enzymatic antioxidants prevailing in the human skin (ascorbic ac acid, and reduced glutathione) were used to validate the measurement procedu results are presented in Tables 2 and S4 of the Supplementary Materials. The res tained for ascorbic acid are characterized by good analytical characteristics: repro (RSD) does not exceed 5.5%, and the recovery is close to 100% in the range of AOA µM-eq (see Table S4 of Supplementary Materials).  There are no statistically significant differences between the expected and m AOA values for individual antioxidants and the binary mixture of ascorbic acid duced glutathione ( Table 2). The model solutions, including both ascorbic and ur showed a slight, but still statistically significant, decrease in the measured AOA pared to the expected value. Apparently, this is due to the insignificant (typical The content of the non-enzymatic antioxidants in the human skin decrease in the following sequence: ascorbic acid > uric acid > reduced glutathione > α-tocopherol > ubiquinol 10 [11]. Due to normal aging and photoaging, the content of ascorbic acid, reduced glutathione, and a-tocopherol decreases, while the level of uric acid is relatively stable [12]. The non-enzymatic antioxidants prevailing in the human skin (ascorbic acid, uric acid, and reduced glutathione) were used to validate the measurement procedure. The results are presented in Table 2 and Table S4 of the Supplementary Materials. The results obtained for ascorbic acid are characterized by good analytical characteristics: reproducible (RSD) does not exceed 5.5%, and the recovery is close to 100% in the range of AOA 15-990 µM-eq (see Table S4 of Supplementary Materials). There are no statistically significant differences between the expected and measured AOA values for individual antioxidants and the binary mixture of ascorbic acid and reduced glutathione ( Table 2). The model solutions, including both ascorbic and uric acids showed a slight, but still statistically significant, decrease in the measured AOA as compared to the expected value. Apparently, this is due to the insignificant (typical for this degree of concentration) antagonism of these antioxidants. An antagonistic effect of mixtures of ascorbic and uric acids was also recorded for conventional spectrophotometric ORAC and ABTS assays and was related to the formation of adducts between these antioxidants [50]. In general, biological objects are characterized by additive, synergistic, and antagonistic antioxidant effects [51]; that is why when characterizing the antioxidant status of a biological sample, preference should be given to the integral value due to its greater informativeness compared with the additive value. Table 3 presents the results of determining the AOA of five volunteers' skin using a three-electrode FPSS configuration (E 1 = E 3 = CV/PET and E 2 = SCSF/Ag/PET in Figure 1a). RSD does not exceed 7.5%, and the values of the recovery of the model antioxidant (L-ascorbic acid), introduced into the membrane, indicate the absence of significant matrix effects.

Analysis of Phytocosmetic Products
The results of determining AOA of a volunteer's skin (a young woman) in the areas without application (control) and after application of the analyzed phytocosmetic product are given in Figures 3 and 4. The time of data collection equal to the duration of exposure of phytocosmetic products on the volunteer's skin was selected as a length of an average working day. All the analyzed phytocosmetic products demonstrated their effectiveness both in vitro (see Table S5 in Supplementary Materials) and in vivo assays (Figures 3 and 4). Figure 4 presents the ∆AOA = f(t) dependences which show how, after the application of phytocosmetic products, AOA of the volunteer's skin was changing for 8 h. The evidence of the antioxidant effect of the phytocosmetic product (t) was mounting in the following order: cream-mousse ≈ cream-gel ≈ serum (t ≤ 5 h) < day cream (t ≤ 8 h) < nourishing night cream (t > 8 h). We have identified two types of ∆AOA = f(t) dependencies. Creammousse, cream-gel, and serum demonstrated the first type of ∆AOA = f(t) dependence (Figure 4a), while day cream and nourishing night cream demonstrated the second type of ∆AOA = f(t) dependence (Figure 4b). We believe that the phytocosmetic products of the first type can be attributed to the express care category, and phytocosmetic products of the second type can be attributed to the cosmetics with the prolonged action. In this paper, we will limit ourselves with just this assumption and leave the further interpretation of the results to cosmetic chemists. without application (control) and after application of the analyzed phytocosmetic product are given in Figures 3 and 4. The time of data collection equal to the duration of exposure of phytocosmetic products on the volunteer's skin was selected as a length of an average working day. All the analyzed phytocosmetic products demonstrated their effectiveness both in vitro (see Table S5 in Supplementary Materials) and in vivo assays (Figures 3 and  4).    Figure 4 presents the ΔAOA = f(t) dependences which show how, after the application of phytocosmetic products, AOA of the volunteer's skin was changing for 8 h. The evidence of the antioxidant effect of the phytocosmetic product (t) was mounting in the following order: cream-mousse ≈ cream-gel ≈ serum (t ≤ 5 h) < day cream (t ≤ 8 h) < nourishing night cream (t > 8 h). We have identified two types of ΔAOA = f(t) dependencies. Cream-mousse, cream-gel, and serum demonstrated the first type of ΔAOA = f(t) dependence (Figure 4a), while day cream and nourishing night cream demonstrated the second type of ΔAOA = f(t) dependence (Figure 4b). We believe that the phytocosmetic products of the first type can be attributed to the express care category, and phytocosmetic products of the second type can be attributed to the cosmetics with the prolonged action. In this   Figure 4 presents the ΔAOA = f(t) dependences which show how, after the application of phytocosmetic products, AOA of the volunteer's skin was changing for 8 h. The evidence of the antioxidant effect of the phytocosmetic product (t) was mounting in the following order: cream-mousse ≈ cream-gel ≈ serum (t ≤ 5 h) < day cream (t ≤ 8 h) < nourishing night cream (t > 8 h). We have identified two types of ΔAOA = f(t) dependencies. Cream-mousse, cream-gel, and serum demonstrated the first type of ΔAOA = f(t) dependence (Figure 4a), while day cream and nourishing night cream demonstrated the second type of ΔAOA = f(t) dependence (Figure 4b). We believe that the phytocosmetic products of the first type can be attributed to the express care category, and phytocosmetic products The AOA of the phytocosmetic product in vitro was correlated with its in vivo effectiveness. The Pearson correlation coefficient between the AOA of the phytocosmetic product obtained in an in vitro assay and the maximum effectiveness of the phytocosmetic product in vivo is equal to r = 0.92 (p < 0.05). These data are presented in Figure 5 and indicate that phytocosmetic products with higher AOA are able to ensure a higher antioxidant status of the human skin. It is worth noting that the use of in vivo assay is preferable, since it most closely resembles the operating conditions of phytocosmetic products and therefore has a better reliability of the obtained results. In addition, in vivo assay enables evaluation of how long the effectiveness of the phytocosmetic product on the human skin can last (the duration of the antioxidant effect).
Chemosensors 2021, 9, x FOR PEER REVIEW 11 of 14 oxidant status of the human skin. It is worth noting that the use of in vivo assay is preferable, since it most closely resembles the operating conditions of phytocosmetic products and therefore has a better reliability of the obtained results. In addition, in vivo assay en ables evaluation of how long the effectiveness of the phytocosmetic product on the human skin can last (the duration of the antioxidant effect).

Figure 5. Correlation between AOA of phytocosmetic product (in vitro analysis) and
ΔAOAmax value of the phytocosmetic product (in vivo analysis). R 2 : approximation reliability value; r: Pearson's correlation coefficient; p: significance level.

Comparison of the Obtained Findings with Earlier Studies
Brainina et al. [29] first proposed a membrane based on cellulose acetate as a media tor carrier [Fe(CN)6] 3-/4-, which, complete with commercially available electrodes, was used to determine the AOA of human skin. A shortcoming of this technique was a multi ple use of the indicator platinum screen-printed electrode and the need for its sterilization (when the test volunteer changed) and regeneration (after a series of measurements). In their later work, Brainina et al. [30] proposed a disposable sensor system. Its effectiveness was exemplified with the biologically active food supplement Askorutin. However, the obtained findings showed a decrease in reproducibility as compared with the results re ceived in [29]. In the present work, a disposable FPSS is proposed, and its effectiveness is exemplified with commercially available phytocosmetic products. The comparative re sults of these three potentiometric sensor systems are presented in Table 4, and it is ap parent that the proposed FPSS enables obtaining the most reproducible results. commercial ECG-type electrode; SCSF/Ag/FG: silver screen-printed electrode on fiberglass modified with silver chloride and silver ferricyanide; SCSF/Ag/PET: silver screen-printed electrode on polyethylene terephthalate modified with silver chloride and silver ferricyanide. 3 The range of determined values of skin AOA can be broadened if concentration of the mediator changes (see Table S6 in Supplementary Materials). Figure 5. Correlation between AOA of phytocosmetic product (in vitro analysis) and ∆AOA max value of the phytocosmetic product (in vivo analysis). R 2 : approximation reliability value; r: Pearson's correlation coefficient; p: significance level.

Comparison of the Obtained Findings with Earlier Studies
Brainina et al. [29] first proposed a membrane based on cellulose acetate as a mediator carrier [Fe(CN) 6 ] 3-/4-, which, complete with commercially available electrodes, was used to determine the AOA of human skin. A shortcoming of this technique was a multiple use of the indicator platinum screen-printed electrode and the need for its sterilization (when the test volunteer changed) and regeneration (after a series of measurements). In their later work, Brainina et al. [30] proposed a disposable sensor system. Its effectiveness was exemplified with the biologically active food supplement Askorutin. However, the obtained findings showed a decrease in reproducibility as compared with the results received in [29]. In the present work, a disposable FPSS is proposed, and its effectiveness is exemplified with commercially available phytocosmetic products. The comparative results of these three potentiometric sensor systems are presented in Table 4, and it is apparent that the proposed FPSS enables obtaining the most reproducible results.

Conclusions
The use of flexible, soft, and stretchable materials is the basis for creating tactile, including wearable, sensors that are advantageous in terms stability, reliability, and convenience as compared with conventional "rigid" electronic devices. The application of scalable technologies and the use of available materials contribute to the rapid prototyping of sensor devices. In this study, we have proposed a new disposable FPSS for non-invasive monitoring of the human skin AOA, which has been developed on the basis of low-cost materials with the use of scalable technologies. The FPSS has been tested in the model conditions and applied for determining the skin AOA of volunteers and for evaluating the impact of phytocosmetic products. Validation of the assay procedure has been performed using non-enzymatic antioxidants present in the human skin. The obtained results have shown good accuracy and reproducibility, which makes it possible to predict a broad applicability of the FPSS in real-time monitoring of the skin AOA as well as for evaluating the effectiveness of topically and orally applied antioxidants.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/chemosensors9040076/s1, Table S1: Specification of the carbon veil used in the study; Figure  S1: SEM-images and EDS spectrum of CV/PET; Table S2: Characteristics of commercially available phytocosmetic products used in the study; Table S3: Composition of the buffer solution pH 5 used in the study; Table S4: Results of L-ascorbic acid determination using FPSS under model conditions (n = 4); Table S5: Effectiveness of commercially available phytocosmetic products;   Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.