Reverse Titration Using Tablets for Accurate Water Hardness Measurement with Improved Resistance to Interference

Abstract

We report a novel tablet-based reverse titration system for rapid, point-of-use measurement of water hardness, overcoming key limitations of conventional EDTA titration. Reagents are encapsulated in pullulan matrix giving two separate tablets. The first tablet contains the Eriochrome black T (EBT) and N-cyclohexyl-3-aminopropanesulfonic acid (CAPS) buffer, while the second encapsulates ethylenediaminetetraacetic acid (EDTA) disodium salt dihydrate. The system employs a trimodal detection strategy: qualitative screening via immediate color change with the EBT tablet, semi-quantitative estimation through combined tablet dissolution and adjusting the sample volume to a reference level, and quantitative determination using reverse titration, where water is gradually added until the red wine endpoint appears. This approach enhances interference tolerance from competing metal ions and improves accuracy over traditional methods. Testing with real water samples showed excellent agreement with standard titration. The tablets remain stable for over seven months, and the system eliminates the need for skilled personnel, laboratory equipment, or bulky instrumentation. This low-cost, user-friendly, and interference-tolerant platform enables rapid and accurate water hardness assessment at the point of use.

1. Introduction

Water, the universal solvent, can dissolve a broad range of solutes. When their concentrations exceed permissible limits, the overall quality deteriorates. Elevated levels of calcium and magnesium ions, for example, are the primary contributors to water hardness [1]. Hardness is typically expressed as the combined concentration of calcium and magnesium ions, reported as calcium carbonate equivalents [2]. According to the World Health Organization (WHO), hardness is classified into four categories based on this equivalent concentration: soft, moderately hard, hard, and very hard [3].

Water hardness is a significant concern in both domestic and industrial applications. In industrial systems, it facilitates scale deposition that obstructs pipelines, impairs heat exchanger performance, degrades reverse osmosis membranes, and reduces the operational efficiency of boilers and wet scrubbers, thereby increasing energy demand and overall operational costs [4,5]. In some homes, hard water contributes to the scaling of plumbing systems, degradation of appliances, and increased soap consumption during laundry activities [6]. In addition to its physical impacts, hard water has been associated with potential health effects; for instance, prolonged consumption has been linked to an increased incidence of atopic dermatitis in children [7], and soft tissue mineralization of the heart, kidney and renal tubule, especially in animals [8]. Despite its detrimental effects on infrastructure, the consumption of hard water has also been associated with potential health benefits. Studies have reported a reduced risk of cardiovascular disease and protection against atherosclerosis in children and adolescents. Hard water can also serve as a supplementary source of calcium and magnesium for individuals who do not meet the World Health Organization’s (WHO) recommended daily intakes of approximately 1000 mg of calcium and 200–400 mg of magnesium through diet alone [9,10]. Given the dual impact of water hardness on infrastructure and human health, there is a clear need for rapid, cost-effective, and accurate tools to detect and quantify hardness levels in water. Previous studies have explored several analytical platforms for water hardness detection and measurement, including paper-based microfluidic devices that employ potentiometric [11], spectrometric [12], fluorometric, and colorimetric detection methods [13]. However, these approaches face multiple challenges, such as non-uniform color distribution on paper substrates, reagent degradation due to air exposure, the need for costly packaging, and, in some cases, reliance on expensive and complex fabrication processes. In addition to these miniaturized approaches, the traditional ethylenediaminetetraacetic acid (EDTA) complexometric titration has long been considered a gold standard for water hardness measurement. However, this method is limited by several drawbacks, including the requirement for skilled personnel, bulky instrumentation, extensive chemical handling, and potential accuracy loss due to interference from other metal ions such as copper and iron when present in high concentrations. Recently, a new class of sensors known as tablet-based sensors has been developed for applications in both health and environmental monitoring. These sensors work by dissolving a pre-formulated tablet in the sample of interest, after which the resulting colorimetric response is measured and compared to a calibration curve to quantify the analyte.

Tablet-based sensors offer several advantages, including portability, ease of use, pre-measured and accurate dosing, cost-effectiveness, scalability, long shelf life, stability, minimal skill requirements, eco-friendliness, low waste generation, and excellent suitability for on-site testing. However, they are also associated with certain limitations, such as their single-use nature and the potential for subjective color interpretation. Several studies have focused on developing tablet-based sensors for a wide range of applications, often using different polysaccharides such as dextran, chitosan, and pullulan as encapsulation matrices due to their ability to stabilize and preserve reagents [14]. For example, dextran has been used to encapsulate gold nanoparticles for the detection of phosphates, hypochlorite, and hydrogen peroxide [15,16,17] in water. Similarly, pullulan has been used to encapsulate and stabilize gold nanoparticles for glucose detection in human saliva [18].

We developed a tablet-based sensor by encapsulating Eriochrome black T (EBT), N-cyclohexyl-3-aminopropanesulfonic acid (CAPS), and ethylenediaminetetraacetic acid disodium salt dihydrate (EDTA) within pullulan. Notably, this is the first report of pullulan being used to encapsulate EBT and EDTA for water hardness measurement. This work builds on our previous contribution to Chemosensors (2020), in which we introduced a dual-modal assay kit for the qualitative and quantitative determination of total water hardness [19]. The tablet-based system developed in this study offers several notable improvements as compared to the classical EDTA titration method. For example, we introduce a new solid-based methodology for water hardness measurement through a reversed titration approach, in which preloaded EDTA in the tablet is titrated with the water sample until the endpoint is reached. To our knowledge, this approach has not been previously reported and is particularly suitable for use in resource-limited settings. Second, this is the first time that pullulan, a polysaccharide, has been used as an effective matrix to encapsulate both EBT and EDTA for water hardness testing. While most water hardness measurement platforms rely on liquid or paper-based substrates, our work demonstrates that polysaccharides can serve as stable and effective matrices for encapsulating reagents used in water hardness analysis. Third, we show for the first time how to encapsulate methanol-containing solutions within a pullulan matrix by introducing the injection-drop-casting method. This innovation overcomes the precipitation problem typically encountered when methanol is combined with pullulan (video is provided in SI file). Furthermore, this study is the first to integrate all three modes of measurement qualitative, semi-quantitative, and quantitative into a single water hardness testing platform, providing multiple options for the end user (Figure 1). Finally, the tablet-based methodology reduces the need for specialized skills or laboratory equipment and, importantly, the semi-quantitative method demonstrates strong tolerance to common interfering ions.

2. Experimental

2.1. Materials and Chemicals

Magnesium sulfate, ethylenediaminetetraacetic acid (EDTA) disodium salt dihydrate, Eriochrome black T (EBT), sodium hydroxide, N-cyclohexyl-3-aminopropanesulfonic acid (CAPS), manganese acetate tetrahydrate, ferrous sulfate, ammonium fluoride, copper (II) nitrate trihydrate, and methanol were obtained from Sigma-Aldrich (Oakville, ON, Canada). Calcium chloride dihydrate and pullulan were purchased from Fisher Scientific (Edmonton, AB, Canada). Deionized water was sourced from the HU-265 Laboratory at Concordia University (Montréal, QC, Canada). The EBT stock solution (0.1% w/v) was prepared by dissolving 100 mg of EBT in 100 mL of methanol. The CAPS buffer (0.45 M) was prepared by dissolving 2 g of CAPS powder in 20 mL of deionized water, adjusting the pH to 10 using 0.1 M NaOH under gentle stirring. The pH was measured with an Accumet AB200 pH and conductivity meter (Fisher Scientific, Hampton, NH, USA) to ensure accuracy. All weighing was performed using a Sartorius Quintix124-1S analytical balance (Sartorius AG, Göttingen, Germany). Tablets were dried at room temperature in a fume hood. Colorimetric responses were recorded using a Samsung A04S phone, and color intensity analyzed with ImageJ version 1.54p software (National Institutes of Health, Bethesda, MD, USA).

2.2. Fabrication of Reagent Tablets

The EBT/CAPS working solution was prepared by mixing 2 mL of 0.1% w/v EBT stock with 5 mL of 0.45 M CAPS buffer. Pullulan solution 7% (w/v) was prepared by dissolving 350 mg of pullulan in 5 mL of deionized water. Next, the injection-drop-casting technique was used to fabricate tablets. For this, 150 µL of pullulan solution was first drop-cast onto a sterile carbon–steel tray, followed by the careful injection of 200 µL of the EBT/CAPS solution into the pullulan matrix. The tray was then left to dry under ambient conditions in a fume hood for 24 h, resulting in well-formed EBT tablets, which were stored in airtight glass vials at room temperature. For the fabrication of EDTA tablets, we made a 4 mM EDTA stock solution by dissolving 149 mg of EDTA disodium salt dihydrate in 100 mL of deionized water. A 5 mL aliquot of this solution was combined with 350 mg of pullulan, and the mixture was gently stirred until fully homogeneous. A 300 µL portion was drop-cast onto a sterile carbon–steel tray and left to dry for 24 h under ambient conditions. Once dried, the EDTA tablets were stored in a separate glass vial.

2.3. Interference Studies

The fabricated tablet sensors were evaluated for selectivity against commonly occurring interfering ions in water, including fluoride (F⁻), copper (Cu²⁺), zinc (Zn²⁺), iron (Fe³⁺), and manganese (Mn²⁺). Test solutions were prepared at concentrations of 4 mg/L, 5 mg/L, 7 mg/L, 1.5 mg/L, and 1 mg/L, which are well above the typical regulatory limits for drinking water. Each interfering ion was tested individually in deionized water without Ca²⁺/Mg²⁺ ions, in a mixed ion solution without Ca²⁺/Mg²⁺ ions, and in a mixed ion solution containing 100 mg/L of Ca²⁺ and Mg²⁺ ions to evaluate potential interference under realistic conditions. To assess the effect of these ions on tablet sensor performance, images of the solutions were captured using a Samsung smartphone at the moment the solution turned blue following EDTA tablet dissolution, just before reverse titration. The captured images were analyzed using the histogram analysis function in ImageJ version 1.54p software. The software has been widely used in previous studies to quantify analyte concentrations by analyzing the color intensity of colored sample solutions [20].

2.4. Real-World Sample Analysis

Finally, the tablet system was validated using real-world water samples, including bottled mineral water, spring water (Maxi, Lachine, Quebec), and three residential tap water samples collected from Downtown Montreal, Lachine and Brossard, Quebec, Canada. Each sample was tested using qualitative, semi-quantitative, and quantitative modes as outlined in Section 3.4.

3. Results and Discussion

3.1. Optimization

Key formulation parameters influencing the performance of the tablet sensor were systematically investigated. The mixing ratio of EBT to CAPS buffer, CAPS concentration, EDTA concentration, and pullulan content were examined to ensure optimal color development, pH stability, and tablet functionality. For EBT–CAPS optimization, five ratios (1:1, 1:5, 2:5, 5:1, and 5:2) were prepared by mixing 0.1% w/v EBT solution with 0.45 M CAPS buffer and evaluated for their ability to produce a distinct red wine color and maintain the target pH 10. For each test, a 200 µL aliquot of the prepared mixture was added to 0.6 mL of a 200 mg/L hard water solution. The resulting mixtures were qualitatively evaluated for the development of a red wine color, indicating metal–indicator complex formation, and for pH stability. The pH of each solution was measured using an Accumet AB200 pH and conductivity meter by dipping the pH probe into the solution and reading the pH value. The reading was taken after the pH meter was calibrated using standard solution at pH 4, 7 and 10 and cleaning the probe with sample before measurement. All tested ratios produced a distinct red wine color upon reaction with hard water; however, the 5:2 ratio yielded the best visually distinct color and maintained a stable pH of 10, attributed to its higher CAPS content. Based on this, the 2:5 ratio was selected for subsequent experiments (Figure 2a).

Next, we optimized the concentration of N-cyclohexyl-3-aminopropanesulfonic acid (CAPS) buffer to ensure pH stability during the reaction. CAPS was chosen because its pKa of 10.6 closely matches the required pH of 10 and has been successfully used in previous studies [21]. CAPS solutions at 0.1 M, 0.25 M, 0.45 M, and 0.55 M were prepared, each mixed with 0.1% w/v EBT in 2:5 mixing ratio. A 200 µL aliquot of each buffer was added to 0.6 mL of hard water (200 mg/L), followed by 300 µL of EDTA. After EDTA addition, the pH was measured using Acummet AB200 pH and conductivity meter. Solutions containing 0.1 M and 0.25 M CAPS failed to maintain pH 10, resulting in incomplete color transitions to pink. In contrast, 0.45 M and 0.55 M CAPS maintained a stable pH of 10 and gave the needed blue color upon EDTA addition. CAPS concentration of 0.45M was used for our tablet formulation (Figure S1b).

Also, we evaluated the optimal EDTA concentration needed to complex calcium and magnesium ions across a wide hardness range. Water samples of known hardness (18.75, 25, 37.5, 75, 100, 150, 200, 300, 400 and 500 mg/L CaCO₃) were tested using EDTA concentrations of 1, 2, 3, 4, 5, and 10 mM. In each case, 300 µL of EDTA was added to 0.6 mL water sample buffered with EBT/CAPS solution. We observed that EDTA at concentrations of 1, 2, and 3 mM was insufficient to produce a complete color transition. EDTA concentrations of 4.0 and 5.0 mM consistently produced a complete color shift from red wine to blue for hardness levels up to 400 mg/L In contrast, the 10 mM EDTA caused a substantial pH shift, preventing any color change. Therefore, 4 mM EDTA was selected for subsequent experiments.

After establishing optimal chemical conditions in solution, we transitioned to optimizing the pullulan concentration used for tablet fabrication. Pullulan solutions were prepared at 4%, 5%, 7%, 10%, and 13% (w/v) and cast into tablets incorporating the optimized EBT, CAPS, and EDTA formulations. Tablets were evaluated for dissolution time under mild agitation by recording the time taken to fully dissolve in the sample. Durability was assessed by applying slight pressure to each tablet and observing whether it disintegrated easily, indicating fragility. Tablets formed with 4% and 5% pullulan dissolved quickly but were too fragile. Higher concentrations (10 and 13%) resulted in more durable tablets but delayed dissolution. A concentration of 7% (w/v) pullulan was found to provide the best compromise, yielding tablets that were both sturdy and dissolved quickly in water under practical conditions (Figure 2).

3.2. Fabrication of EBT and EDTA Tablets

To effectively encapsulate EBT/CAPS buffer into the pullulan matrix, we introduced the injection-drop-casting method (Video S1). Here, the pullulan solution devoid of any reagents was drop-cast onto a sterile carbon steel surface and allowed to spread as a thin liquid film. Once the pullulan was evenly cast, we carefully injected the EBT/CAPS methanol-containing solution into the pullulan matrix. EBT is very soluble in methanol and pullulan forms precipitate with methanol; hence, when methanol solution containing EBT and CAPS buffer was mixed directly with pullulan, precipitation was observed. This precipitation of pullulan in methanol solutions has been reported in the literature [22]. However, using our injection-drop-casting approach, we overcame this challenge. Encapsulating EDTA within pullulan was much more straightforward because both EDTA and Pullulan are very soluble in water; we followed established procedures reported in previous studies on the encapsulation of gold nanoparticles within dextran matrices [16]. Here, the pullulan powder was simply dissolved directly into the aqueous EDTA solution, and the mixture was drop-cast onto a sterile carbon steel tray. After air drying under ambient conditions for 24 h, the EDTA tablets were fully formed and durable (Figure 3b).

The tablets were characterized using FTIR, and the interactions between pullulan and the encapsulated reagents (EBT and EDTA) were examined using ATR–FTIR spectroscopy (Figure 4c). Pullulan displays a broad O–H stretching band at 3292 cm⁻¹ [23] and a strong C–O–C glycosidic vibration at 1148 cm⁻¹ [24]. The abundant hydroxyl groups in pullulan act as both hydrogen-bond donors and acceptors, enabling potential non-covalent interactions with polar functionalities of the encapsulated molecules.

In EBT-pullulan tablet, the O–H stretching region of pullulan overlaps with that of the phenolic –OH groups in EBT (3200–3300 cm⁻¹) [25], precluding direct separation of these bands. Evidence of interaction is instead observed in the fingerprint region: the azo (–N=N–) band of EBT, originally at 1506 cm⁻¹ [25], exhibits a slight red shift (3–7 cm⁻¹) and reduced intensity, while the sulfonate S=O stretching envelope (1192–1043 cm⁻¹) [25] appears broadened. These spectral changes are consistent with hydrogen bonding between pullulan hydroxyl groups and the electron-rich sulfonate oxygens or azo nitrogen of EBT, potentially stabilizing the encapsulated form and restricting vibrational freedom.

In EDTA–pullulan tablet, the asymmetric and symmetric COO⁻ stretching bands (1666–1607 and 1394 cm⁻¹, respectively) [26], display subtle shifts and broadening compared to free EDTA. The C–N stretch at 1474 cm⁻¹ [26], also becomes less defined in the composite. These modifications suggest hydrogen bonding between pullulan hydroxyl groups and EDTA’s carboxylate oxygens, along with possible weak interactions involving the amine functionalities. Such interactions likely alter the local polarity and vibrational coupling within the encapsulated structure. For both systems, the preservation of distinct, non-overlapping marker bands—EBT: 1506, 1334, 1192–1043 cm⁻¹; EDTA: 1666–1607, 1394, 1014 cm⁻¹—together with peak broadening and minor shifts confirms that encapsulation proceeds via non-covalent interactions, predominantly hydrogen bonding.

3.3. The Principle of Tablet-Based Inverted Microtitration

Our tablet system follows the same fundamental chemistry as the classical EDTA titration. However, in conventional titration, when competing ions are present, the endpoint is often less distinct because the titrant is introduced gradually, allowing interfering ions to compete with calcium and magnesium during the titration process. In our approach, EDTA is preloaded in excess within the tablet, which allows sufficient time for both fast- and slow-complexing ions to bind before the titration step begins. As a result, the color change observed at the endpoint primarily reflects the presence of calcium and magnesium ions, producing a clearer and more reliable signal. Furthermore, our semi-quantitative method uses pre-calibrated vials (A, B, C, and D) that correspond to specific water hardness ranges. This design ensures that small amounts of interfering ions do not cause the measurement to shift into a different hardness category, giving the end user confidence that the reported classification is robust. In our approach, the water samples (0.6 mL) containing Ca²⁺ and Mg²⁺ ions are initially combined with EBT tablet which simultaneously buffers the solution and introduces the metal ion indicator, followed by the addition of the EDTA tablet which turns the mixture blue, indicating that all calcium and magnesium ions have been complexed by the EDTA present in the tablet. Subsequently, water samples are added gradually to the system. As the concentration of free metal ions (Ca²⁺ and Mg²⁺ ions) increases and exceeds the complexing capacity of EDTA, a distinct red wine coloration appears, marking the endpoint. This color transition allows for the determination of water hardness either semi-quantitatively, by comparing the final solution volume to pre-marked reference levels on the vial, or quantitatively, using a calibration curve derived from known standards. The underlying reaction mechanism is illustrated below (Figure 5).

3.4. Determination of Water Hardness Using Tablet Sensor

3.4.1. Qualitative Detection Method

After completing the optimization studies, we used the finalized tablet sensor to screen water sample for hardness; this test was carried out by preparing water samples at total hardness concentrations of 0, 50 mg/L, 100 mg/L, and 250 mg/L. For each sample, a 0.6 mL aliquot was pipetted into four distinct clean vials, and one EBT tablet was added to each vial. The vials were gently agitated to ensure complete tablet dissolution. Once the tablets dissolved, the resulting solution was observed for color change. A red wine color indicated the presence of hardness ions (Ca²⁺ and/or Mg²⁺) while a blue color indicated the absence of hardness. This simple visual method allows for fast, on-site screening for the hardness of water without requiring specialized equipment or technical expertise (Figure S4a).

3.4.2. Semi-Quantitative Detection Method

After developing a method for qualitative screening of water samples, we developed a semi-quantitative method to estimate water hardness using the tablet sensor. To align with the World Health Organization (WHO) classification of water hardness, we prepared test solutions representing the lower and upper concentration thresholds for each category: 18–60 mg/L for soft water, 61–120 mg/L for moderately hard water, 121–180 mg/L for hard water, and 181–260 mg/L for very hard water. For each test, a 0.6 mL aliquot of the prepared sample was transferred into a 15 mL calibrated vial. The EBT tablet was added first, followed by the EDTA tablet. Initially, the solution appeared blue due to the presence of excess EDTA. Water was then gradually added, with gentle agitation, until a distinct red wine coloration appeared, indicating that the EDTA had been fully consumed, and free calcium and magnesium ions were now available to complex with EBT. Each test was performed in triplicate to ensure consistency and reproducibility of the results. At this color transition point, the total volume of liquid in the vial was measured and recorded. These measured volumes were then marked directly on the vial, to create a set of calibrated reference points for the semi-quantitative hardness measurement. It is worth to mentioning that our reference points on each vial were accurately marked by measuring the accurate volume for each water hardness category and carefully marking the lower meniscus of the liquid in vial. This allowed us to overcome manufacturing errors seen in some vial calibrations. The observed endpoint volumes were approximately 3.5 mL for 18 mg/L, 2.0 mL for 61 mg/L, 1.5 mL for 121 mg/L, and 1.1 mL for 181 mg/L, corresponding to soft, moderately hard, hard, and very hard water, respectively.

It is important to clarify that the lower concentration threshold within each WHO-defined hardness category represents the maximum volume of water required to reach the red wine endpoint for that range. Although samples at the upper end of a range may reach the endpoint with less added water, they still maintain the characteristic red wine coloration at the calibrated reference mark. By calibrating the system using the lower concentration threshold, we ensured that even harder water samples within the same category could still be reliably classified as the mark (Figure S4b).

3.4.3. Quantitative Detection Method

Building on the semi-quantitative findings, we developed a fully quantitative approach to precisely determine water hardness. Standard water solutions with total hardness values of 18, 37.5, 50, 75, 100, 150, 200, 300, and 400 mg/L, were prepared to cover a broad and practically relevant concentration range. For each standard, 0.6 mL of the solution was pipetted into a clean 15 mL calibrated vial. One Eriochrome Black T (EBT) tablet and one EDTA tablet were sequentially added, and the vial was gently agitated to ensure complete dissolution, forming a homogenous mixture. Using a 10 μL micropipette, incremental additions of the same standard solution were made, with careful observation for the appearance of a distinct red wine coloration at the endpoint, indicating complete EDTA consumption and subsequent complexation of free calcium and magnesium ions by the EBT indicator. Each measurement was performed in triplicate. Notably, trials revealed that the tablet system is ideal for quantitative hardness measurements, specifically within the 18–400 mg/L CaCO₃ range, as no clear or reproducible endpoint was observed below 18 mg/L, likely due to insufficient ion availability for consistent complex formation. The total volume of added hard water required to reach the red endpoint at each standard concentration was recorded and plotted to generate a calibration curve (Figure 6). This calibration curve provides a practical way for estimating unknown water hardness levels using the tablet system, making it highly applicable for routine field and laboratory use within the specified concentration range. Using a 10 μL micropipette may be less user-friendly for point-of-use detection. In our study, it was necessary to precisely deliver a controlled amount of water to the solution in order to generate the calibrations curve; however, the end user can use any water handler that can transfer a known volume of water to the system and then use the calibration curve to quantify the water hardness. (Figure 6).

3.5. Stability

The tablet sensor stability was evaluated at room temperature under ambient conditions using spiked hard water at 200 mg/L as CaCO₃. At monthly intervals we recorded the volume of water required to produce the first visible red wine color change and calculated retained activity as a percentage of day zero. The tablet sensor remained fully functional for the entire observation period, with consistent endpoints (Figure S5).

As a laboratory reference, we prepared an Eriochrome Black T indicator solution and stored it under the same conditions. During routine use, brief uncapping led to solvent loss and visible crystallization within 24 h, after which the solution was not usable (Figure S5). This finding aligns with literature reports recommending fresh preparation of EBT solutions due to their limited stability, especially when exposed to light [26]. We note that this in-use exposure is stringent and not directly comparable to commercial titration kits. Many commercial products supply liquid indicators that are formulated and packaged for long shelf life in sealed containers, with optimized solvents, preservatives, light protection, and minimal headspace.

Our conclusion is therefore limited to common laboratory practice with simple in-house indicator solutions. In that context, the pullulan tablet format removes the need for frequent solution preparation and maintained function for months at room temperature. Future work will include a head-to-head assessment under sealed storage, using matched temperature, light exposure, and opening cycles, to measure the relative shelf life of tablets versus commercial liquids [27].

3.6. Interfering Test

To assess whether common ions found in water interfere with the performance of our tablet system, aliquots were analyzed both individually (for each ion) and as a combined-ion mixture. Iron (Fe³⁺), copper (Cu²⁺), zinc (Zn²⁺), manganese (Mn²⁺), and fluoride (F⁻) were tested at concentrations of 1.5 mg/L, 5 mg/L, 7 mg/L, 1 mg/L and 4 mg/L, respectively. The selected concentration for each ion were substantially higher than the maximum allowable concentrations (MAC) in drinking water (Fe³⁺ ≤0.10 mg/L, Cu²⁺ 2.0 mg/L, Zn²⁺ ≤ 5.0 mg/L, Mn²⁺ 0.12 mg/L, and F⁻ 1.5 mg/L; (Table S2) [28,29,30]. All solutions were prepared in deionized water. In the single-ion tests, 0.6 mL of each solution was transferred to a micro test tube, while for the combined-ion condition, equal volumes of the individual ion solutions were mixed, and a 0.6 mL aliquot of this mixture transferred to a micro test tube. A hardness standard of 100 mg/L as CaCO₃ served as the positive control. The Eriochrome Black T tablet was added to each tube and allowed to dissolve completely, followed by the addition of an EDTA tablet. All samples were photographed using a Samsung smartphone mounted 20 cm from the tubes on a white background under constant lighting. Images were saved as JPEG files with no camera filters and analyzed in ImageJ software. A rectangular region of interest covering the liquid column, excluding the glass edges and the meniscus, was applied consistently to all tubes. The mean blue value (0–255) from the color histogram was recorded as the intensity metric, since the blue signal after EDTA tablet dissolution indicates readiness for the subsequent water titration step. Previous studies has also used imagej software to quantify analyte concentrations by analyzing the color intensity of colored sample solutions [20].

None of the interfering ions, either singly or in combination, produced a blue color close to the 100 mg/L hard water reference after the EDTA tablet was added, as shown in Figure 7a, confirming that the tablet system does not generate false positive signals in the presence of these common ions. Following this, we evaluated interference at concentrations reported in previous studies by spiking each interfering ion into 100 mg/L hard water and repeating the same protocol. As shown in Figure 7b, the blue color intensity closely matched that of the 100 mg/L control. Finally, hardness level of bottled mineral water was measured using both the EDTA titration and tablet system and found 103 mg/L and 101 mg/L, respectively. Next, all interfering ions at twice the concentrations used in the interference test were spiked into the bottled mineral water. Using the semi-quantitative procedure (Section 3.4.2), the sample remained within the moderately hard water range. In contrast, the quantitative measurement for the combined ion sample gave a hardness below 101 mg/L as CaCO₃ for the tablet system and a value higher than 103 mg/L as CaCO₃ for the classical EDTA titration. These findings indicate that combined ions can bias absolute quantification, whereas the semi-quantitative mode shows better tolerance to interference. Future studies will introduce masking agents into tablet formulation to improve quantitative accuracy [21].

3.7. Application of the Tablet System to Real Water Samples

The practical applicability of the tablet system was shown by testing real-world water samples obtained from bottled mineral and spring water (purchased from Maxi stores in Lachine, Quebec) and tap water from downtown-Montreal, Brossard, and Lachine, Quebec, respectively. The test was carried out in three modes qualitative, semi-quantitative, and quantitative. The first test was qualitative screening to check for the presence of water hardness. For each sample, a 0.6 mL aliquot was dispensed into a 15 cm calibrated vial, and one EBT tablet was added. The vial was gently agitated by hand until the tablet was fully dissolved. In all three replicates, a visible red wine coloration developed, confirming the presence of water hardness across all tested sources. This quick test provided a simple, yes-or-no visual confirmation of hardness.

After the qualitative test, we performed a semi-quantitative assessment to estimate the hardness range of each water sample. Four pre-calibrated vials were prepared and marked at 1.5 mL, 2.0 mL, 2.5 mL and 3.5 mL. These markings correspond to very hard, hard, moderately hard, and soft levels, respectively. For each replicate, 0.6 mL of the test sample was dispensed into the vial, followed by incremental additions of the water until the liquid level reached the distinct pre-calibrated mark. The calibrated volume at which the characteristic red wine color change occurred was then used to categorize the water sample into its corresponding hardness range. Results across all triplicates were consistent, confirming the reproducibility and reliability of the semi-quantitative method (

Table 1. Quantitative and semi-quantitative measurement of water hardness using the tablet-based sensor. Water samples were analyzed in triplicate using both the tablet sensor (reported as hardness ranges) and the classical EDTA titration method (mean ± SD). A two-tailed Student’s t-test yielded a p-value greater than 0.05, indicating no statistically significant difference between the tablet sensor and the classical EDTA titration. Hardness levels were classified according to standard categories.

Water Source	EDTA-Titration (Quantitative Analysis) (mg/L ± SD)	Tablet Sensor (mg/L ± SD)	Tablet Sensor (Semi-Quantitative Analysis) (Degree of Water Hardness) (mg/L)	Recovery (%)
Spring water	260 ± 1.2	255 ± 1.0	Very hard (≥181)	98.1
Tap water 1	123 ± 0.8	120 ± 0.8	hard (121–180)	97.6
Mineral water	103 ± 0.9	101 ± 0.6	moderately hard (61–120)	98.0
Tap water 2	76 ± 0.9	73 ± 0.8	Moderately hard (61–120)	96.0
Tap water 3	53 ± 0.6	51 ± 0.4	Soft (≤60)	96.2

).

For precise quantification, four 15 mL vials were each filled with 0.6 mL aliquots of the test sample. One EBT tablet and one EDTA tablet were added to each vial and allowed to dissolve fully. The respective water sample was then incrementally added with a 10 µL pipette until a red wine color endpoint was observed. The volume required to reach this endpoint was recorded and compared against a previously established calibration curve to determine total hardness in mg/L CaCO₃. The results obtained using the tablet-based system were in strong agreement with those from conventional EDTA titration, validating the sensor’s accuracy, reliability, and suitability for both field-based and laboratory application (Figure 8).

Future work will focus on applying a design of experiments (DoE) approach to further refine the optimization process and incorporate masking agents into the tablet formulation to enable accurate hardness measurement using the quantitative method for water samples with significant interference, such as wastewater. In addition, a comprehensive stability study comparing the tablet assay with commercially available water hardness test kits will be conducted to assess its long-term performance and reliability in future studies

4. Conclusions

The development of a portable, low-cost tablet-based detection system is crucial for effective water quality monitoring. Our proposed assay followed a trimodal approach to detect and quantify water hardness along with additional advantage of increased tolerance towards interfering ions in water. In this study, we successfully fabricated a dual-tablet system by encapsulating EBT and EDTA within a pullulan matrix. This system enables rapid qualitative screening within seconds and offers both semi-quantitative and quantitative measurement of water hardness. To address the challenge of pullulan precipitation in methanol, we introduced for the first time an injection-drop-casting method that enabled stable encapsulation of EBT. Furthermore, our inverted microtitration technique produced a clearer and more distinct endpoint color, while reducing the effect of interfering ions, particularly in the semi-quantitative measurement of water hardness. Although the quantitative method still exhibited some susceptibility to interference, similar to classical EDTA titration, future studies will focus on incorporating masking agents into the tablet formulation to enable reliable measurement of water hardness even in the presence of high levels of interfering ions. The tablet-based system was validated using water samples from three sources: bottled mineral, spring, and tap water, with results showing strong agreement with traditional EDTA titration methods. Overall, our sensor offers a convenient, rapid, equipment-free, and highly stable alternative for water hardness analysis, making it ideal for point-of-use applications in both domestic and field settings.