Microscale titration of acetic acid using digital colorimetry and paper-based analytical devices

Talanta 276 (2024) 126254 Available online 12 May 2024 0039-9140/© 2024 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/bync-nd/4.0/). Microscale titration of acetic acid using digital colorimetry and paper-based analytical devices Irati Berasarte a , b , * , Ane Bordagaray a , Rosa Garcia-Arrona a , Miren Ostra a , William Reis de Araujo b , Maider Vidal a a Department of Applied Chemistry, University of the Basque Country (UPV/EHU), 20018, Donostia/San Sebastian, Spain b Portable Chemical Sensors Lab, Department of Analytical Chemistry, Institute of Chemistry, State University of Campinas (UNICAMP), 13083-970, Campinas, S˜ ao Paulo, Brazil ARTICLE INFO Keywords: Paper-based device Micro-titration Paper microfluidics Colorimetry Digital image analysis Process automation ABSTRACT A quantitative method for acid-base titrations in paper-based devices (PADs) is described to analyze acetic acid in vinegar samples. In this work, two different types of PADs were developed: a device for individual spot testing and a microfluidic device. Digital colorimetry was used as the detection method, and the images were acquired using a smartphone and a homemade box with LED lights for controlled image acquisition. Titration curves were built with just eight points, using the R channel based on the gradual color transition from red to blue of litmus, a natural indicator. The endpoint was accurately determined by second derivative calculations. Both systems were applied to fifteen vinegar samples of different types, and good concentration results were obtained in comparison to the reference method. The proposed methodology is simple, fast, environmentally friendly, and surpasses the need for calibration curve construction. Moreover, the subjective endpoint identification is eliminated, and the method was automated to provide a high throughput workflow, suitable for quality control processes and realtime measurements. 1. Introduction One of the most popular traditional methodologies that enables quantitative analysis is volumetric analysis. As a result, the concentration of a particular analyte can be measured using a solution with a known concentration (usually used as titrant). There are many different titration types, and they can be categorized based on the measuring technique, the order of the titrants, or the type of response. The response can be based on an acid-base, redox, complexometric, or precipitation reaction [1]. Focusing on acid-base titrations, the most common methods employed are manual titration with visual indication of the endpoint (hereafter EP, set to correspond to equivalence point). Instrumental and automated lab-based methods using spectrophotometric detectors and the use of pH electrodes (potentiometry) provide accurate measures, exempting the analyst from interpreting the color change. However, over the last decades, significant technological advances have been made, enabling the implementation of digital colorimetric analysis to follow these titrations and to detect the endpoint at the point-of-need [2,3]. Digital colorimetry constitutes a powerful tool due to its simplicity, low cost, and the diversity of imaging options, ranging from desktop scanners and webcams to digital cameras and smartphones [4]. An image can be decomposed into different color models, such as RGB, HSV or CIELAB, and then, color intensity can be related to analyte concentration [5]. Many examples can be found where digital colorimetry has been applied. For instance, for the screening of disease markers in serum [6], the determination of sulfonamides in water [7], nickel in electroless coating baths [8], and total phenolic content in tea and infusion samples [9]. Regarding the titration-based determination of total acidity, it has been calculated using a microplate, either in wines with a flatbed scanner [10], or in citric fruits employing a smartphone [11], among others. In both cases, a vector is created with RGB values extracted from the image, which is plotted against the added titrant volume to obtain a titration curve. Then, EP is calculated with first or second derivatives. In fact, derivatives are the employed mathematical calculations to determine the inflection point of sinusoidal curves as the ones obtained in titrations. Sodium chloride has also been determined by titration and employing a microplate – smartphone system [12]. In this case, G * Corresponding author. Department of Applied Chemistry, University of the Basque Country (UPV/EHU), 20018, Donostia/San Sebastian, Spain. E-mail address: [email protected] (I. Berasarte). Contents lists available at ScienceDirect Talanta journal homepage: www.elsevier.com/locate/talanta https://doi.org/10.1016/j.talanta.2024.126254 Received 21 February 2024; Received in revised form 6 May 2024; Accepted 11 May 2024 Talanta 276 (2024) 126254 2 channel values were plotted against added titrant volume and EP was calculated as the intersection between two straight lines. Paper-based analytical devices (PADs) have also drawn attention as a promising tool in analytical chemistry in order to detect and quantify the analyte of interest. Paper is made of cellulose, which gives the ability to transport liquids without instrumentation, by capillarity [13]. There are many ways to manufacture these devices, including photolithography, plasma treatment, wax-printing, ink-printing and laser treatment [14]. The detection can also be done by several methods, such as fluorescence, Raman scattering, electrochemistry, or colorimetry. Because of its flexibility and malleability, many different designs can be created, including spot tests, microfluidic analysis (known as μ PADs), and distance-based measurements [15]. In general, PADs are used in many fields, including environmental monitoring, point-of-care diagnostics, forensics, clinical analysis, and food safety [16–19]. In addition, literature shows that they can be combined with titrimetric analyses: chelate titration for Ca 2+ and Mg 2+ determination [20], redox titration for alcohol content analysis [21], acid-base titration for formaldehyde quantification [22] and iodometric titration for ascorbic acid and dopamine determination [23], for example. Teprek et al. presented four different systems including alkalinity determination by acid-base titration, hardness quantification by complexometric titration, redox titration for thiosulfate quantification and chloride analysis by precipitation titration [24]. In these references, EP was identified visually on the PADs based on the color change of the indicator. Scaling from semi-quantitative to quantitative analysis can be achieved with digital colorimetry, as it can be employed to build a calibration curve. For example, by the use of the R channel intensity [25] or Euclidean norm [26], which can be related to concentration. Acetic acid, also known as ethanoic acid or methylcarboxylic acid, is a weak organic acid that can be used in the production of latex emulsion resins, synthesis of terephthalic acid, fruit vinegar production, dyes or pharmaceutical formulations, among others [27]. It is also the primary organic component of vinegar, responsible for its scent and bitter flavor. Vinegar is defined as a liquid product that is produced by alcoholic and acetic fermentation of agricultural-origin substances. It is mostly known for its applications in the food industry as salad seasoning, pickling, flavoring, or as a preservative agent in bottled sauces [28]. Table 1 shows some works related to acetic acid determination found in the literature. Regarding titration-based techniques, a pH ISFET cell for wine analysis has been reported [29], as well as visual detection of endpoint for vinegar samples [30,31], and digital colorimetry for acetic acid determination in water [32]. In relation to spectroscopic methods, UV–Vis [33] and NIR [34] have been used for vinegar analysis, and fluorescence-based determinations of bacteria [35] and Chinese wine [36]. Finally, HPLC for vinegar, probiotic and water kefir beverages [37–40], and 1 H-RMN for multivariate analysis of strawberries [41] have been reported. In the majority of the cases, analyte quantification was carried out using an analytical curve, either with univariate calibration using potential difference, absorbance or peak area, for example; or by multivariate analysis methods. Literature shows that acetic acid can also be determined using PADs. Acid-base titrations have been carried out by different methods such as spot tests [43] or μ PADs, either by star-like designs [44] or distance-based measurements [45]. However, the majority of the systems are based on qualitative or semi-quantitative EP determination, some works require building a calibration line previously, which do not employ the titration curve itself for acid determination. In other cases, images need to be taken after each titrant addition, which make the methods laborious and significantly increases experimental work and data analysis complexity. In this work, microscale acid-base titrations for acetic acid determination have been carried out in two types of paper-based devices, a spot testing device and a star-like design following a microfluidic approach. In the proposed systems, just an eight-point titration is necessary to obtain a quantitative and accurate determination of EP based on second derivative calculations, which is achieved using a single image. To our knowledge, it is the first μ PAD system where EP identification is completely objective and based on the titration curve, without the need for visual detection or previously building a calibration curve. A simple, low-cost, and environmentally friendly method is proposed, automated for fast sample analyses. 2. Experimental 2.1. Chemicals and materials All reagents were of analytical grade and used without additional purification. Litmus, acetic acid, potassium hydrogen phthalate and absolute ethanol were purchased from Panreac (Barcelona, Spain). Sodium hydroxide was supplied by Sigma-Aldrich (St. Louis, USA) and Labkem (Barcelona, Spain). Phenolphthalein was provided by Merck (Darmstadt, Germany). Deionized water (resistivity ≥18 MΩ cm at 25 ◦C) was obtained from a Milli-Q Advantage-0.10 purification system (Millipore, Germany). Solutions of litmus 0.1%, litmus 0.166% and phenolphthalein 0.1% were prepared by dissolving the appropriate weighted mass of the indicators in a small amount of absolute ethanol and diluting to mark with deionized water. Standard solution of acetic acid 4% was prepared from 96% glacial acetic acid and NaOH 0.1 mol L −1 was standardized with 0.056 mol L −1 potassium hydrogen phthalate. 2.2. Fabrication of paper-based devices Paper-based devices were designed using Microsoft Office PowerPoint 2016 software and were printed on qualitative filter paper using a wax printer (Xerox ColorQube 8570, CT, USA). The printed sheets were placed in a thermal press (Maquinatec, MG, Brazil) for 15 s at 100 ◦C. Upon heating, wax penetrates the paper and defines hydrophobic barriers around the pattern. The back of the printed device was covered with transparent tape to prevent solutions from leaking through the device. Two approaches were used in this work, spot test analysis and microfluidic analysis, so two PADs were designed and fabricated (Fig. 1). Spot tests were performed in 5 mm diameter circles, patterned in a 3 by 8 design in order to analyze eight titration points in triplicate (Fig. 1a). Microfluidic devices were fabricated modifying the star-like design reported by Karita and Kaneta [44], with a sample reservoir located at the center, eight branches with a reaction reservoir and a detection zone at the end of each branch (Fig. 1b). Table 1 Summary of analytical methods used for acetic acid determination found in the literature. Technique Sample Determination Ref. Titration Wine pH ISFET cell and analytical curve [29] Vinegar Visual endpoint detection [30,31] pH-meter and analytical curve [42] Water Digital colorimetry and titration curve derivatives [32] Spectroscopy Vinegar UV–Vis and analytical curve [33] NIR and multivariate regression [34] Bacteria Fluorescence biosensor and analytical curve [35] Chinese wine Fluorescent uranyl-organic framework and analytical curve [36] HPLC Vinegar, probiotic and water kefir beverages UV–Vis and analytical curve [37–40] 1 H-RMN Strawberry Multivariate analysis (PCA, HCA and PLS-DA) [41] I. Berasarte et al. Talanta 276 (2024) 126254 3 2.3. Colorimetric acid-base titration In both approaches, eight-point acid-base titrations were carried out. In order to find the optimal measurement conditions, some parameters were optimized. These parameters include device size, indicator, titrant and sample volumes, drying method and image acquisition time. Regarding spot tests, first, mixtures of litmus, NaOH and water were added to the spots, and then, sample was added. Immediately, PADs were left to dry and images were taken at different times to find the optimal conditions. For microfluidic analyses, final concentrations in spot tests were taken as reference. First, varying NaOH concentration solutions were added to all reaction reservoirs. Then, litmus was added to all detection zones. Finally, sample was placed in the center of the μ PAD. Devices were left to dry and images were taken at different times to achieve the optimal condition. In both cases, digital images of PADs were acquired with a smartphone OnePlus Nord CE (three cameras, 64 +2 +8 MP) in automatic mode. In order to maintain image acquisition conditions constant, a home-made light box was used. The portable wooden box (dimensions: 10 cm height, 12 cm width, 15 cm length) was fabricated to eliminate external light interferences, with 24 V white LED lights along the interior. Both PADs and the smartphone were always placed in fixed positions to avoid further variation between photographs. RGB values were extracted using the free image-editing software GIMP (version 2.10.34). Image analysis was also carried out using MATLAB R2020b version (The Mathworks Inc., Massachusetts, USA). 2.4. Vinegar samples Both of the proposed methods were tested in fifteen vinegar samples. Five samples were purchased in local markets from Campinas (S˜ ao Paulo, Brazil), including lemon, rice and alcohol vinegar (Castelo Alimentos S.A.) and two different apple vinegar samples (Castelo Alimentos S.A. and Almaromi, both from S˜ ao Paulo, Brazil). The other 10 samples were part of an industrial process, provided by JR Sabater S.A. (Murcia, Spain), and included five samples derived from alcoholic fermentation and another five samples of white wine fermentation. Commercial samples were diluted 10 times, as indicated concentration on the label is 4%, whereas industrial samples, with a theoretical concentration of around 10%, were diluted 25 times. Acid-base potentiometric titration was used as the reference method. Acetic acid was determined using the 794 Basic Titrino (Metrohm, Switzerland) automatic titrator coupled to a pH electrode. Firstly, NaOH was standardized with potassium hydrogen phthalate and phenolphthalein in triplicate. Then, 0.25 mL or 0.5 mL of vinegar sample (depending on its theoretical concentration) and 2.0 mL of litmus 0.1% indicator were added to a beaker. Finally, solutions were titrated with standardized NaOH solution until the color changed from red to blue. EP volume was used for the calculation of acetic acid concentration. Measurements were performed in triplicate. 3. Results and discussion 3.1. Optimization of PAD-based sensing 3.1.1. Color channel selection In this work, litmus was used as acid-base indicator (pK a 6.5), which color change occurs between pH 4.5–8.3 from red to blue, with purple color at neutral values. Acetic acid is a weak acid for which titration EP lies between pH 6.0–9.0. For that reason, litmus was chosen over some common indicators such as phenolphthalein, as it allows us to observe a broader color change closer to the EP, providing better precision and sensitivity for this weak acid-strong base titration. A synthetic solution of acetic acid 0.4%, after a 1:10 dilution of the 4% stock solution, was used for procedure optimization, and firstly, some preliminary tests were performed in 96-well microplates to select the optimal color channel. Digital images were analyzed to extract red (R), green (G) and blue (B) values of each spot, which were plotted against the titrant volume added to achieve a classic acid-base titration curve (Fig. S1). The three channels had more or less constant values before EP. Then, R and G channels drastically decrease, while B channel slightly increases. After EP, R values become constant, while G and B values increase a bit. For this reason, R channel was selected as the color channel for monitoring the titrations, as it gives the most sigmoidal curve. 3.1.2. Optimization of spot tests procedure Fabrication of PADs was done by wax printing and thermal heating, and no related problem was identified during the experiments. First and foremost, the optimal indicator volume for the designed and waxprinted device was investigated. Two indicator volumes were tested, 5 μ L and 7 μ L. It was easily observed by naked eye that 7 μ L was too much, as drying time increased significantly and sometimes the drop exceeded the wax, so 5 μ L was chosen as optimal indicator volume. Then, 96-well microplates were used to estimate titrant volumes, which were scaled to PAD proportions. Once NaOH volumes were fixed as 1.0–6.0 μ L, titrations were carried out using three different concentrations of litmus. Color values with litmus 0.05% were too light, and the Fig. 1. Design of paper-based devices. (a) Wax-printed paper device with individual spots for an eight-point titration, with three rows for analysis in triplicate. (b) Wax-printed μ PAD for microfluidic analysis, with eight identical branches located equidistantly from the center (sampling zone). Lengths are indicated in millimeters. I. Berasarte et al. Talanta 276 (2024) 126254 4 calculation of the concentration was more subjected to error. In contrast, a high concentration of litmus as 0.2% compromised the titration, as EP was shifted to higher values, and thus, the calculation of acidity was affected. With a concentration of 0.1%, both titration curves and acetic acid concentration showed good results, so it was selected as the optimal concentration. Important differences were observed between adding the reagents individually and adding them together. When reagents were added individually, titration curves were not homogeneous and the method became unreliable. When both reagents were added together, repeatability was improved, and for that reason, litmus and NaOH were added together. The mixture is stable for a long time, it can be previously prepared, and the correct mixture of the reagents is ensured. Mixtures of 5 μ L litmus, 1–6 μ L NaOH, and 1–5 μ L water were prepared in vials, and 11 μ L was added to each of the eight titration spots. Then, 4 μ L of a standard of acetic acid 0.4% was added to start the reaction. Thus, final concentrations of each compound were: NaOH 0.007–0.04 mol L −1 , litmus 0.03%, and sample 0.1%. Two of the most important factors to consider are the drying method and the drying/reaction time. On the one hand, devices were dried at room temperature and in the oven (Fig. 2a). PADs left at room temperature took a lot of time to dry, up to 90 min or more. This was caused by moderate to high air humidity, which varies greatly between days and seasons, impeding the standardization of the protocol. In contrast, if the devices were dried in an oven, PADs were completely dry in 30 min maximum, so it was selected as drying method. A temperature of 50 ◦C was used to avoid the deformation of the wax-based barrier. On the other hand, regarding drying time (Fig. 2b), it could be visually detected that spots changed color in the first minutes after addition, so images acquired immediately and after 5 min did not give good results. After 10 min, good EP and concentration results were obtained. However, because the drops were not completely dry, pixel selection for data extraction was limited, and thus, the shape of the curves was not completely sigmoidal. At 20 min, some spots were dry and others were not, so titration curves showed variable R values with high error bars, impacting EP definition and providing inaccurate results. After 30 min in the oven, all spots were completely dry and image quality was good, which resulted in good titration curves and concentration results in all cases, so this time was set as optimal. Finally, it needs to be noted that the reduction of total reagents and sample volume of each spot from 15 μ L to 10 μ L or 8 μ L, for example, could be investigated. It will depend on the system under study and the necessary concentration of each reagent, but that way, the total drying time would be reduced, accelerating the process. 3.1.3. Optimization of microfluidic analysis procedure Microfluidic paper-based devices ( μ PADs) were fabricated as mentioned in the Experimental part. Firstly, NaOH concentration and volume were optimized. Mixtures with different concentrations of titrant were prepared in small vials, considering the final titrant mmol added in spot tests. Reaction reservoirs could be filled with 1 μ L in the original design [44], but as our design consisted of 8 branches instead of 10, optimization was carried out with 0.50, 0.75, and 1.00 μ L of titrant. With the lowest volume, inaccurate results were obtained due to the incomplete filling of the whole reaction reservoir. In contrast, the highest volume spread through the channels of the device, and interfered with concentration calculations. It was determined that 0.75 μ L (NaOH standards between 0.02 and 0.12 mol L −1 ) was the optimal volume of titrant to be added to the reaction reservoirs, as it avoided both problems mentioned and offered good concentration results. Regarding litmus volume and concentration, it was visually detectable that 0.5 μ L was enough to fill the detection zone correctly. The concentration was recalculated according to the litmus-NaOH relation used in spot tests configuration, so a new stock solution of litmus 0.166% was prepared and used. In this μ PAD system, sample was placed in the center to spread through the channels. Preliminary tests were carried out using 28, 24, 20 and 16 μ L of sample. It was observed that the two smallest volumes were not enough to complete the titration. By naked eye, colors obtained with 28 μ L and 24 μ L were similar. However, R channel values and titration curves showed that the most accurate concentration values were obtained using 24 μ L of the sample. Once the optimal sample volume was set, reaction time was investigated. Images were acquired using the lighting box at 0, 2.5, 5 and 10 min. It was observed that 10 min was the necessary time to completely dry the device. Room temperature was selected as the drying method due to the size of the μ PAD and the small volumes used in the analysis. 3.1.4. Summary of colorimetric procedures Regarding spot tests, each of the titration points was analyzed in triplicate with the same device. Firstly, 11 μ L of a reagent mixture was added to all spots. The mixture contained 5.00 μ L of litmus 0.1%, varying volumes of NaOH 0.10 mol L −1 (1.00–6.00 μ L), and water up to 11.0 μ L. After, 4.00 μ L of sample 0.4% was added. Immediately, PADs were placed in the oven at 50 ◦C, and images were taken after 30 min. For μ PAD analysis, 0.75 μ L of varying NaOH concentration solutions (0.02–0.12 mol L −1 ) were added to all reaction reservoirs, and then, 0.50 μ L of litmus 0.166% was added to all detection zones. Next, 24 μ L of sample 0.4% were placed in the center of the μ PAD (sampling zone). The device was dried at room temperature and images were acquired after Fig. 2. (a) Titration curves obtained with two different drying methods. Drying at room temperature took up to 90 min, in contrast with the oven, which took only 30 min. (b) Titration curves obtained from PADs dried in the oven for different times. I. Berasarte et al. Talanta 276 (2024) 126254 5 10 min. 3.1.5. Method validation Once optimal analysis parameters were identified, a standard solution of acetic acid (0.4%) was used for method validation. In the case of spot tests, intra-day precision was calculated using three PADs measured on the same day (n =9), and inter-day precision with four devices from three different days (n =12). In the case of μ PADs, fifteen devices divided into 5 images (3 devices per image) were used for intra-day precision calculation (n =15), and inter-day precision was evaluated using nine μ PADs from 3 images taken in three different days (n =9). Tests performed with the standard solution showed good precision and accuracy results by both the proposed methods (Table 2), which come in agreement with the guideline of AOAC [46]. 3.2. Colorimetric titration and acetic acid calculation As previously stated, images of spots PADs were obtained after 30 min in the oven, while images of μ PADs were captured after 10 min at room temperature. In both cases, R channel values were used to build titration curves, obtaining sigmoidal shapes. After extracting R channel values, the first and second derivatives were calculated and plotted, as it is the correct way of mathematically calculating the inflection point of this type of curve. In this titration, the first derivative results in a minimum value at the endpoint, whereas it equals zero on the second derivative. Assuming that the reaction between the weak acid and the strong base is complete, and due to the 1:1 stoichiometry of the reaction, it can be assured that the NaOH mmol number equals acetic acid mmol exactly at EP. Using the R color values and the titrant volumes, experimental derivatives were calculated. Then, a regression line was built between the maximum and minimum points of the second derivative curve. Subsequently, EP volume or mmol number was calculated for y = 0 using the slope and intercept of the regression line. Finally, acetic acid concentration (%) was calculated. In the case of spot tests, sample volume is a known value (4 μ L), so R values were plotted against the volume of NaOH added, and acetic acid concentration was calculated directly using that value. In relation to microfluidic analyses, 0.75 μ L of different titrant solutions were added to the reaction reservoirs, and then, the sample was added to the center of the device, so it could spread to the channels of the μ PAD. To build titration curves, R channel values were plotted against the added NaOH mmol number. In this case, added NaOH volume was the limiting factor, so the amount of sample that took part in the titration is the amount of titrant that was fixed (0.75 μ L). A schematic illustration of the procedure for EP calculation is shown in Fig. 3. 3.3. Vinegar sample analysis Volumetric analysis performed with an automatic titrator and potentiometric detector was the reference method used for vinegar sample analysis. In addition, the proposed paper-based methods, both spot tests and μ PADs, were employed. All industrial and commercial samples were analyzed in triplicate in all cases. Some vinegar types may contain small amounts of tartaric and citric acid. However, due to the dilution applied to the samples (1:10 or 1:25), no interference should be expected from those compounds. Fig. 4 shows the titration curves obtained with both methods for a sample (C101) selected as example. Regarding spot tests, acetic acid concentrations were 4.4%–4.9% for commercial vinegars and 10.1%–17% for industrial samples. Microfluidic measurements gave concentrations between 4.2% and 4.8% for commercial samples and 10.2%–17% for industrial samples. Precision values (RSD%) ranged between 0.6% and 13% in spot testing, and between 0.9% and 9.9% for μ PAD analysis. It needs to be noted that, even if some RSD values were a bit high, they were below 10% for the great majority of samples. Considering the different measuring locations and that some samples were not measured as soon as they were opened or produced, especially in the case of the industrial samples, acceptable results were obtained with the proposed PADs. Besides, relative errors obtained by spot tests were all below 10.5%, with nine samples below 5%. Errors obtained by microfluidic measurements were similar, with 14% being the highest relative error and with ten samples below 5%. In conclusion, similar results were obtained with both methods in comparison to the reference method. Detailed values are summarized in Table 3. Results obtained by potentiometric titration and the proposed digital colorimetry titration on PADs were statistically evaluated. Two different tests were conducted. First, since sample concentrations vary in a wide range, a joint confidence ellipse test (EJCR) for slope and intercept was performed [47,48]. In the case of spot tests, the confidence interval for the slope was 1.00 ±0.09, and 0.2 ±0.8 for the intercept. For the μ PAD approach, the slope was 1.00 ±0.08, and intercept was −0.1 ±0.8 (Fig. S3). Considering these values, the ideal point (slope 1 and intercept 0) lies within these ranges in both PADs, so the proposed methods do not differ significantly from the reference method. Then, a paired t-test was performed with a 95% confidence level. No significant differences were found between reference and spot tests (t exp 1.80 <t crit 2.14), nor between reference and microfluidic measurements (t exp 1.20 <t crit 2.14), since the calculated t value was below the tabulated one. In conclusion, it can be assured that none of the proposed methods presents systematic errors and that they are reliable methods for acetic acid determination in vinegar samples. Collectively, the proposed microscale titration on PADs combined with digital colorimetry presents significant improvements from the Green Analytical Chemistry point of view. It reduces the amount of reagents consumed and waste generated due to its miniaturization, and provides automated analyses compared to classical titration, providing an AGREE score [49] of 0.71 for spot testing, 0.78 for the μ PAD approach and 0.49 for the classical titration method (Fig. S4). These values demonstrate that microscale titrations performed on PADs are more sustainable than the traditional potentiometry method. In addition, the use of digital colorimetry for titration improves the accuracy of the method because it exempts the analyst from visually identifying the color endpoint, which may be subjective. 3.4. Method automation Method automation in the food industry, specifically for analyte monitoring, is crucial for ensuring accuracy and efficiency in the testing process. Automation reduces human errors, enhances precision, and allows for a higher throughput of samples. In this regard, two image analysis software were used and compared in this work. On the one hand, the free software GIMP was employed. A Microsoft Excel (Microsoft, 2016 version) template was created to simplify and speed up data analysis, and just by inserting the R channel values extracted manually by GIMP, the template makes the necessary first and second derivative calculations, corresponding titration graphs, and acetic acid concentration calculations. Even if analysis time was reduced, the color extraction work becomes tedious when dealing with a large number of measurements. As an alternative, a user-friendly automated function was designed in MATLAB using the Image Processing Toolbox™. This way, manual extraction is avoided, a single software is used, and a stack of images Table 2 Method validation results for the two types of PADs developed in this work. PAD type Acidity (%) Precision (RSD, %) Accuracy (RE, %) Intra-day Inter-day Spots 0.427 ±0.007 0.4–1.9 4.8 2.7–9.1 μ PAD 0.39 ±0.02 1.0–6.4 5.4 1.4–3.8 I. Berasarte et al. Talanta 276 (2024) 126254 6 taken in the same position can be analyzed in a single run, providing a high throughput workflow. A schematic illustration of the procedure is shown in Fig. S2. A few inputs need to be specified to run the function: image name or format, dilution factor of the sample, and in the case of spot tests, NaOH concentration. Then, all necessary steps from dimension selection to acetic acid calculation are performed automatically. Finally, a double graph with all the results becomes visible (titration curve, 2 nd derivative graph, EP, and acetic acid concentration in the sample), as well as an output matrix in the workspace with the concentration results. The results obtained with both methods were almost identical, indicating the usefulness and effectiveness for the monitoring of acetic acid concentration either in academia or industrial processes. Moreover, the proposed PADs can be employed to perform other types of titrations, as the function is available under request and it is easily adjustable to other reactions just by modifying titrant volumes or concentrations. 3.5. Comparison with other works As was mentioned in the introduction, determination of acetic acid using PADs has been previously reported in the literature. Table 4 shows the works where spot testing and μ PADs have been used for that purpose. In relation to spot measurements, the works found were majorly related to teaching and adapting the paper-based method for experimental laboratory courses. The most used titrant was NaOH, and the Fig. 3. Schematic illustration of EP calculation procedure. Mean R values were extracted and plotted against added NaOH ( μ L or mmol, the latter in this example figure). Second derivatives were calculated for the curves. A regression line was constructed between the previous and the following point to the curve crossing at y =0, and from it, EP ( μ L or mmol) was calculated. Fig. 4. Titration curves obtained for C101 sample selected as example. (a) Spot testing, triplicate analysis in the same device, and (b) microfluidic measurements performed three times. I. Berasarte et al. Talanta 276 (2024) 126254 7 indicators varied from the classic phenolphthalein to others that offer gradual color changes such as jaboticaba peel extract or litmus, as proposed in this work. In general, the variable parameter was the volume of titrant, but the change in acetic acid concentration has also been reported. PAD analysis has been carried out either by naked eye, digital camera or smartphone, and EP has been determined based on visual determination, biggest jumps, or the intersection between two lines. One work proposed a methodology based on the 1 st derivative, similar to the one proposed. However, these methods are either subjective, semiquantitative or laborious, especially when they are based on images taken after each titrant addition, as proposed by Nogueira and coworkers [50], which significantly increases experimental work and data analysis complexity. The main advantage of the proposed spot-testing method is that the need of capturing various images is eliminated. Instead, multiple spots are used simultaneously, expediting the entire titration process with a unique image. Regarding μ PADs, only two works were found where acetic acid determination was carried out. In the work of Karita and Kaneta [44], a star-like design with 10 branches was used, titrant was placed in the reaction reservoirs, and indicator into the detection zones, as performed in this work. In that study, a semi-quantitative analysis was performed, and titration endpoint was determined visually based on the colorless to pink color change of phenolphthalein indicator. In this work, the design was modified and reduced to lower lobes in order to investigate if titrations can be carried out quantitatively with just two points around EP. As explained in previous sections, good results were obtained with an 8-point titration, demonstrating the precise applicability of the method. In the case of Dias et al. [45], acetic acid standards were used, placed on the μ PAD, and traveled distance was measured with a digital caliber. Distance of the colored channel of the μ PAD was related to the analyte concentration to build a calibration curve. Then, samples were measured following the same principle. Although these approaches are based on acid-base titration, the titration curve itself is not used for acid determination, the color change of phenolphthalein may lead to inaccurate EP determinations, and the need to build a calibration line increases data analysis time. As an alternative, the method proposed in this work enables the visualization of a gradual color change, which can be detected by smartphone imaging, quantitative determination of EP can be carried out, and thus, acetic acid concentration can be determined accurately based on an eight-point titration and just one image. As mentioned for spot testing, fast and easy-to-use methodologies that uplift previously reported works were developed. 4. Conclusions It was demonstrated that the proposed methods for microscale acidbase titrations using paper-based devices are reliable for the quantification of acetic acid in vinegar samples. Two types of devices – based on spot tests and μ PADs – were developed and optimized for analysis, eliminating the subjective detection of reaction endpoint and the need to build a calibration curve. Advances from other works were obtained: titrations were carried out using just eight analysis points and one image obtained with a smartphone, a completely quantitative μ PAD was Table 3 Acetic acid concentration values given as acidity percentage and obtained by the reference method, spot-based PADs, and microfluidic devices. Relative errors in comparison to potentiometry were calculated in each case. Sample Reference method Spot tests Microfluidic analysis C acetic (%) C acetic (%) RE (%) Cacetic (%) RE (%) A102 10.5 ±0.2 11.3 ± 0.5 7.6 10.2 ± 0.7 −2.8 E102 10.3 ±0.3 11.2 ± 0.3 8.5 10.3 ± 1.1 0.5 A106 12.1 ±0.1 13.1 ± 0.4 8.4 10.4 ± 0.8 −13.9 D106 11.0 ±0.1 10.1 ± 1.5 −8.1 10.4 ± 0.5 −5.1 B105 9.9 ±0.1 10.1 ± 1.3 2.6 10.3 ± 0.9 5.6 E105 11.0 ±0.2 10.8 ± 0.8 −1.8 11.0 ± 0.1 0.4 F105 11.15 ±0.03 12.2 ± 0.4 9.4 10.5 ± 0.6 −6.2 B101 12.95 ±0.04 12.8 ± 0.5 −1.4 12.9 ± 0.3 −0.4 C101 14.4 ±0.1 15.0 ± 0.1 2.8 14.8 ± 1.3 2.7 F101 17.0 ±0.3 16.8 ± 1.2 −1.2 17.3 ± 0.4 1.6 Lemon vinegar 4.30 ±0.01 4.4 ±0.3 3.3 4.2 ±0.3 −2.8 Rice vinegar 4.6 ±0.2 4.53 ± 0.05 −2.0 4.3 ±0.1 −6.7 Apple vinegar (1) 4.8 ±0.1 4.9 ±0.2 3.2 4.8 ±0.1 1.1 Apple vinegar (2) 4.35 ±0.04 4.8 ±0.2 10.4 4.39 ± 0.06 0.9 Alcohol vinegar 4.52 ±0.05 4.8 ±0.2 4.6 4.5 ±0.2 −1.8 Table 4 Summary of works reported in the literature for acetic acid determination using PADs. PAD Reagents Color change Procedure Detection EP V total Samples Precision and accuracy Ref. Spots NaOH and jaboticaba peel extract Magenta to green Varying V titrant Image after each addition Smartphone 1 st derivative 20 μ L Vinegar (3) RSD 2.8–4.8% RE 4.6–7.5% [50] Na 2 CO 3 and phenolphthalein Colorless to pink Varying C acetic One image Digital camera and lighting box Intersection of two lines a – – [51] NaOH and phenolphthalein Colorless to pink Varying V acetic Naked-eye Naked-eye 90 μ L Foods (15) RSD 0–12.6% RE 0.5–12.8% [43] NaOH and universal indicator a Varying V titrant Image after each addition Smartphone Biggest pH jump 100 mL Foods and others RSD 7.1–28% RE 5.8–40% [52] NaOH and litmus Red to blue Varying V titrant One image Smartphone and lighting box 2 nd derivative 110 μ L Vinegar (15) RSD 0.6–13% RE 1.2–10.4% This work μ PAD Na 2 CO 3 and phenolphthalein Colorless to pink Varying C titrant Naked-eye Naked-eye 45 μ L – – [44] NaOH and phenolphthalein Colorless to pink Calibration line Distance (digital caliber) – 160 μ L Vinegar (3) RSD 4.3–4.7% RE 1.9–2.4% [45] NaOH and litmus Red to blue Varying C titrant One image Smartphone and lighting box 2 nd derivative 34 μ L Vinegar (15) RSD 0.9–9.9% RE 0.4–14% This work a Not mentioned. I. Berasarte et al. Talanta 276 (2024) 126254 8 proposed for the first time, the use of litmus and second derivatives enabled a highly accurate endpoint calculation, and the method was automated either for free image-processing software or for MATLAB. This way, analysis time and complexity, and reagent and sample volumes were considerably reduced, minimizing waste generation and enabling a more environmentally-friendly method. Both methods are precise, accurate, easy to use and fast, making them suitable for quality control processes and real-time monitoring. CRediT authorship contribution statement Irati Berasarte: Writing – review & editing, Writing – original draft, Software, Methodology, Investigation, Conceptualization. Ane Bordagaray: Writing – review & editing. Rosa Garcia-Arrona: Writing – review & editing. Miren Ostra: Writing – review & editing, Supervision, Resources, Project administration, Conceptualization. William Reis de Araujo: Writing – review & editing, Resources, Conceptualization. Maider Vidal: Writing – review & editing, Supervision, Resources, Conceptualization. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Data availability Data will be made available on request. Acknowledgements This work was supported by the Basque Government (Research Group IT1662/22), the University of the Basque Country (pre-doctoral scholarship PIF 19/131), S˜ ao Paulo Research Foundation (FAPESP) [Grant numbers: 2018/08782–1, and 2022/03250-7], and National Council for Scientific and Technological Development – CNPq (Grant number: 310282/2022-5). Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.talanta.2024.126254. References [1] D.C. Harris, Quantitative Chemical Analysis, seventh ed., W. H. Freeman and Company, 2007. [2] E. da N. Gaiao, V.L. Martins, W. da S. Lyra, L.F. de Almeida, E.C. da Silva, M.C. U. Araújo, Digital image-based titrations, Anal. Chim. Acta 570 (2006) 283–290, https://doi.org/10.1016/j.aca.2006.04.048. [3] D.S.A. Beeckman, G. Meesen, P. Van Oostveldt, D. Vanrompay, Digital titration: automated image acquisition and analysis of load and growth of chlamydophila psittaci, Microsc. Res. Tech. 72 (2009) 398–402, https://doi.org/10.1002/ jemt.20694. [4] S. Soares, G.M. Fernandes, F.R.P. Rocha, Smartphone-based digital images in analytical chemistry: Why, when, and how to use, Trends Anal. Chem. 168 (2023) 117284, https://doi.org/10.1016/j.trac.2023.117284. [5] S.E. Umbaugh, Computer Imaging: Digital Image Analysis and Processing, CRC Press, 2005. [6] R. Deng, X. Chao, H. Li, X. Li, Z. Yang, H.Z. Yu, Smartphone-based microplate reader for high-throughput quantitation of disease markers in serum, Analyst 148 (2023) 735–741, https://doi.org/10.1039/d2an01571d. [7] D. Barzallo, L. Ferrer, E. Palacio, Eco-friendly screening method for sulfonamides using a 3D handheld smartphone-based fluorescence detection device and graphene nanoplatelet-packed pipette tip microextraction, J. Environ. Chem. Eng. 12 (2024) 111888, https://doi.org/10.1016/j.jece.2024.111888. [8] G. Albizu, M. Vidal, A. Bordagaray, R. García-Arrona, M. Ostra, Semi-automatic digital image control of nickel concentration in a Ni-P electroless acid bath using the hexaamminenickel blue complex, Microchem. J. 199 (2024) 110073, https:// doi.org/10.1016/j.microc.2024.110073. [9] M. Zugazua-Ganado, A. Bordagaray, J. Ezenarro, R. Garcia-Arrona, M. Ostra, Adaptation of the Folin-Ciocalteu and Fast Blue BB spectrophotometric methods to digital image analysis for the determination of total phenolic content : reduction of reaction time , interferences and sample analysis, LWT–Food Sci. Technol. 193 (2024) 115756, https://doi.org/10.1016/j.lwt.2024.115756. [10] D.A. Volmer, L. Curbani, T.A. Parker, J. Garcia, L.D. Schultz, E.M. Borges, Determination of titratable acidity in wine using potentiometric, Conductometric, and Photometric methods, J. Chem. Educ. 94 (2017) 1296–1302, https://doi.org/ 10.1021/acs.jchemed.6b00891. [11] L.M.A. de Oliveira, V.B. dos Santos, E.K.N. da Silva, A.S. Lopes, H.A. Dantas-Filho, An environment-friendly spot test method with digital imaging for the microtitration of citric fruits, Talanta 206 (2020) 120219, https://doi.org/10.1016/j. talanta.2019.120219. [12] K. Ruttanakorn, N. Phadungcharoen, W. Laiwattanapaisal, A. Chinsriwongkul, T. Rojanarata, Smartphone-based technique for the determination of a titration equivalence point from an RGB linear-segment curve with an example application to miniaturized titration of sodium chloride injections, Talanta 233 (2021) 122602, https://doi.org/10.1016/j.talanta.2021.122602. [13] L.A. Pradela-Filho, W.B. Veloso, I.V.S. Arantes, J.L.M. Gongoni, D.M. de Farias, D. A.G. Araujo, T.R.L.C. Paix˜ ao, Paper-based analytical devices for point-of-need applications, Microchim. Acta 190 (2023) 179, https://doi.org/10.1007/s00604023-05764-5. [14] E. Noviana, T. Ozer, C.S. Carrell, J.S. Link, C. McMahon, I. Jang, C.S. Henry, Microfluidic paper-based analytical devices: from design to applications, Chem. Rev. 121 (2021) 11835–11885, https://doi.org/10.1021/acs.chemrev.0c01335. [15] T. Ozer, C. McMahon, C.S. Henry, Advances in paper-based analytical devices, Annu. Rev. Anal. Chem. 13 (2020) 85–109, https://doi.org/10.1146/annurevanchem-061318-114845. [16] L.F. de Lima, A.L. Ferreira, I. Ranjan, R.G. Collman, W.R. de Araujo, C. de la Fuente-Nunez, A bacterial cellulose-based and low-cost electrochemical biosensor for ultrasensitive detection of SARS-CoV-2, Cell Reports Phys. Sci. 4 (2023) 101476, https://doi.org/10.1016/j.xcrp.2023.101476. [17] M. Dong, J. Wu, Z. Ma, H. Peretz-Soroka, M. Zhang, P. Komenda, N. Tangri, Y. Liu, C. Rigatto, F. Lin, Rapid and low-cost CRP measurement by integrating a paperbased microfluidic immunoassay with smartphone (CRP-Chip), Sensors 17 (2017) 684, https://doi.org/10.3390/s17040684. [18] M.S. Kavai, L.F. de Lima, W.R. de Araujo, A disposable and low-cost laser-scribed graphene electrochemical sensor for simultaneous detection of hydroquinone, paracetamol and methylparaben, Mater. Lett. 330 (2023) 133211, https://doi.org/ 10.1016/j.matlet.2022.133211. [19] M. Srisa-Art, K.E. Boehle, B.J. Geiss, C.S. Henry, Highly sensitive detection of Salmonella typhimurium using a colorimetric paper-based analytical device coupled with Immunomagnetic separation, Anal. Chem. 90 (2018) 1035–1043, https://doi.org/10.1021/acs.analchem.7b04628. [20] S. Karita, T. Kaneta, Chelate titrations of Ca2+and Mg2+using microfluidic paperbased analytical devices, Anal. Chim. Acta 924 (2016) 60–67, https://doi.org/ 10.1016/j.aca.2016.04.019. [21] S.A. Nogueira, A.D. Lemes, A.C. Chagas, M.L. Vieira, M. Talhavini, P.A.O. Morais, W.K.T. Coltro, Redox titration on foldable paper-based analytical devices for the visual determination of alcohol content in whiskey samples, Talanta 194 (2019) 363–369, https://doi.org/10.1016/j.talanta.2018.10.036. [22] N. Taprab, Y. Sameenoi, Rapid screening of formaldehyde in food using paperbased titration, Anal. Chim. Acta 1069 (2019) 66–72, https://doi.org/10.1016/j. aca.2019.03.063. [23] O. Heidary, M. Akhond, B. Hemmateenejad, A microfluidic paper-based analytical device for iodometric titration of ascorbic acid and dopamine, Microchem. J. 182 (2022) 107886, https://doi.org/10.1016/j.microc.2022.107886. [24] A. Teprek, V. Poetri Artono, W. Waiyawat, A. Limsakul, J. Shiowatana, A. Siripinyanond, Semi-quantitative analysis by spot counting on origami paperbased device for endpoint detection in titrimetric analysis, Microchem. J. 158 (2020) 105284, https://doi.org/10.1016/j.microc.2020.105284. [25] O.D. Oyewunmi, S.H. Safiabadi-Tali, S. Jahanshahi-Anbuhi, Dual-modal assay kit for the qualitative and quantitative determination of the total water hardness using a permanent marker fabricated microfluidic paper-based analytical device, Chemosensors 8 (2020) 1–19, https://doi.org/10.3390/chemosensors8040097. [26] E. Ghohestani, J. Tashkhourian, B. Hemmateenejad, Colorimetric determination of peroxide value in vegetable oils using a paper based analytical device, Food Chem. 403 (2023) 134345, https://doi.org/10.1016/j.foodchem.2022.134345. [27] F.S. Wagner, U.B. Staff, Acetic acid, Kirk-Othmer Encycl, Chem. Technol. (2014) 1–21. [28] C.W. Ho, A.M. Lazim, S. Fazry, U.K.H.H. Zaki, S.J. Lim, Varieties, production, composition and health benefits of vinegars: a review, Food Chem. 221 (2017) 1621–1630, https://doi.org/10.1016/j.foodchem.2016.10.128. [29] P. Gim´ enez-G´ omez, M. Guti´ errez-Capit´ an, J.M. Ríos, F. Capdevila, A. Puig-Pujol, C. Jim´ enez-Jorquera, Microanalytical flow system for the simultaneous determination of acetic acid and free sulfur dioxide in wines, Food Chem. 346 (2021) 128891, https://doi.org/10.1016/j.foodchem.2020.128891. [30] F. Benmeziane Derradji, N. Ghanemi, L. Djermoune Arkoub, Comparing physicochemical, nutritional and biochemical properties of three traditional fig vinegars produced in Algeria, Carpathian J. Food Sci. Technol. 15 (2023) 147–158, https://doi.org/10.34302/crpjfst/2023.15.2.13. [31] K.M.J.P. Bouatenin, K.A. Kouam´ e, M.E. Gueu-Kehi, N.T. Dj´ eni, K.M. Dj` e, Organic production of vinegar from mango and papaya, Food Sci. Nutr. 9 (2021) 190–196, https://doi.org/10.1002/fsn3.1981. [32] R.M. Caraballo, L.M. Saleh Medina, S.G.J. Gomez, P. Vensaus, M. Hamer, Turmeric and RGB analysis: a low-cost experiment for teaching acid-base equilibria at home, I. Berasarte et al. Talanta 276 (2024) 126254 9 J. Chem. Educ. 98 (2021) 958–965, https://doi.org/10.1021/acs. jchemed.0c01165. [33] S. ang Supharoek, K. Ponhong, W. Siriangkhawut, K. Grudpan, Employing natural reagents from turmeric and lime for acetic acid determination in vinegar sample, J. Food Drug Anal. 26 (2018) 583–590, https://doi.org/10.1016/j. jfda.2017.06.007. [34] S. Kasemsumran, A. Boondaeng, S. Jungtheerapanich, K. Ngowsuwan, W. Apiwatanapiwat, P. Janchai, P. Vaithanomsat, Assessing fermentation broth quality of pineapple vinegar production with a near-infrared fiber-optic probe coupled with stability competitive adaptive reweighted sampling, Molecules 28 (2023) 6239, https://doi.org/10.3390/molecules28176239. [35] M. Mormino, V. Siewers, Y. Nygård, Development of an Haa1-based biosensor for acetic acid sensing in Saccharomyces cerevisiae, FEMS Yeast Res. 21 (2021) foab049, https://doi.org/10.1093/femsyr/foab049. [36] G. Che, D. Gu, W. Yang, M. Li, Q. Pan, Turn-on fluorescence detection of acetic acid in wine using a uranyl-organic framework, Cryst. Growth Des. 22 (2022) 1984–1990, https://doi.org/10.1021/acs.cgd.2c00016. [37] S.T. Alemneh, S.A. Emire, M. Jekle, B. Hitzmann, Effect of refrigerated storage on some physicochemical characteristics of a teff-based fermented beverage and the viability of the fermenting Lactiplantibacillus plantarum and Lacticaseibacillus rhamnosus used, J. Food Process. Preserv. 46 (2022) e17034, https://doi.org/ 10.1111/jfpp.17034. [38] Z.B. Güzel-Seydim, G. S ¸atır, Ç. G¨ okırmaklı, Use of Mandarin and persimmon fruits in water kefir fermentation, Food Sci. Nutr. 11 (2023) 5890–5897, https://doi.org/ 10.1002/fsn3.3561. [39] A. Liu, W. Pan, S. Li, J. Li, Q. Li, L. He, S. Chen, K. Hu, X. Hu, G. Han, S. Li, J. Zhou, F. Chen, S. Liu, Seasonal dynamics of microbiota and physicochemical indices in the industrial-scale fermentation of Sichuan Baoning vinegar, Food Chem. X. 16 (2022) 100452, https://doi.org/10.1016/j.fochx.2022.100452. [40] S.H. Mohamed, A.I. Salim, Y.M. Issa, A.E. Ali, Detection and identification of adulteration in vinegar samples based on reversed-phase high-performance liquid chromatographic (RP-HPLC) strategies, ACS Food Sci. Technol. 2 (2022) 21–30, https://doi.org/10.1021/acsfoodscitech.1c00169. [41] S. Huang, S. Ying Lim, H. Lau, W. Ni, S. Fong Yau Li, Effect of glycinebetaine on metabolite profiles of cold-stored strawberry revealed by 1H NMR-based metabolomics, Food Chem. 393 (2022) 133452, https://doi.org/10.1016/j. foodchem.2022.133452. [42] N.S. Oliveira, A.F. Oliveira, A.A. Neves, M.E.L.R. Queiroz, Development of a nontitration potentiometric method for the determination of acidity in vinegar, Talanta 256 (2023) 124261, https://doi.org/10.1016/j.talanta.2023.124261. [43] J. Huang, L. Cai, A comprehensive experiment for training undergraduates majoring in nutrition education using a paper-based micro acid-base titration, J. Chem. Educ. 99 (2022) 3607–3612, https://doi.org/10.1021/acs. jchemed.2c00211. [44] S. Karita, T. Kaneta, Acid −base titrations using micro fl uidic paper-based analytical devices, Anal. Chem. 86 (2014) 12108–12114, 10.1021/ac5039384. [45] L.G.A. Dias, L.C. Duarte, K.M.P. Pinheiro, N.S. Moreira, W.K.T. Coltro, Distancebased paper microfluidics as environmentally friendly platforms for monitoring acid-base titrations, Talanta Open 7 (2023) 100216, https://doi.org/10.1016/j. talo.2023.100216. [46] Guidelines for Standard Method Performance Requirements, 2016, https://doi. org/10.1093/9780197610145.005.006. [47] H.C. Goicoechea, A.C. Olivieri, Sustained prediction ability of net analyte preprocessing methods using reduced calibration sets. Theoretical and experimental study involving the spectrophotometric analysis of multicomponent mixtures, Analyst 126 (2001) 1105–1112, https://doi.org/10.1039/b100422k. [48] A.G. Gonz´ alez, M.A. Herrador, A.G. Asuero, Intra-laboratory testing of method accuracy from recovery assays, Talanta 48 (1999) 729–736, https://doi.org/ 10.1016/S0039-9140(98)00271-9. [49] F. Pena-Pereira, W. Wojnowski, M. Tobiszewski, Agree - analytical GREEnness metric approach and software, Anal. Chem. 92 (2020) 10076–10082, https://doi. org/10.1021/acs.analchem.0c01887. [50] S.A. Nogueira, L.R. Sousa, N.K.L. Silva, P.H.F. Rodrigues, W.K.T. Coltro, Monitoring acid-base titrations on wax printed paper microzones using a smartphone, Micromachines 8 (2017) 1–10, https://doi.org/10.3390/mi8050139. [51] P. Namwong, P. Jarujamrus, M. Amatatongchai, S. Chairam, Fabricating simple wax screen-printing paper-based analytical devices to demonstrate the concept of limiting reagent in acid-base reactions, J. Chem. Educ. 95 (2018) 305–309, https:// doi.org/10.1021/acs.jchemed.7b00410. [52] J. Li, M.L. O’Neill, C. Pattison, J.H.W. Zhou, J.M. Ito, C.S.T. Wong, H.Z. Yu, N. Merbouh, Mobile app to quantify pH strips and monitor titrations: smartphoneaided chemical education and classroom demonstrations, J. Chem. Educ. 100 (2023) 3634–3640, https://doi.org/10.1021/acs.jchemed.3c00227. I. Berasarte et al.