Calcium Determination by Complexometric Titration with Calcein Indicator Using Webcam for Endpoint Detection

Abstract

Precise and convenient analytical methods are needed for the quantitative determination of calcium in water and food. Complexometric titration remains a reliable technique to determine calcium in milligram amounts. The titrations have been performed automatically by detecting color transitions with a webcam. Classical complexometric indicator calcein provided a sharp color transition. In diffuse reflection mode, the color appearance parameter (Hue) provides better precision and is more resistant to ambient light fluctuations compared to RGB primaries. In fluorescence mode with LED illumination, the fluorescence brightness of calcein is independent of ambient light, and the primary green color provides the sharpest endpoints. The color change during titration is better in the upper part of the acquired images due to the internal filter effect in calcein solutions. The automatic titration with a digital burette provides a standard deviation as low as 0.1 μmol. An example of its application is in the determination of calcium in commercial mineral waters. Based on the AGREE and ComplexMoGAPI rating scales, the semi-automatic titration showed better environmental assessment compared to the standard ASA method.

1. Introduction

The determination of calcium in water is important for many reasons. In heat exchange installations, calcium and magnesium ions can result in the formation of scale on hot surfaces. Therefore, water used as a heating or cooling medium should be free of calcium ions. For technical reasons, water used for domestic purposes should also have a hardness of low to medium. One study on tap water samples collected in 33 countries showed significant variation in calcium and magnesium content [1]. It turned out that European waters are among the hardest, which is due to their geological origin. So, water softening is necessary, and its control requires determining the content of calcium and magnesium ions. The opposite situation occurs in natural seawater, where calcium plays an important role in the structure of many marine organisms, including corals, mussels, macroalgae, and bacterial biofilms. Therefore, an important task for owners of marine aquaria is the regular control of calcium concentration in order to maintain it at a suitable level. In the human body, calcium is 1 of 11 macroelements. Calcium forms the structure of bones and teeth. Calcium plays important roles in the homeostasis of blood plasma, and in the proper functioning of cells and mucous membranes. Calcium is involved in metabolism, muscle contraction, and blood coagulation. It has anti-inflammatory, anti-swelling, and anti-allergic properties [2]. Therefore, the body’s high demand for calcium and magnesium, and on the other hand, the low supply of these elements in the diet, are the cause of many diseases. These include cardiovascular diseases and an associated increased risk of mortality [3]. Sources of calcium in the diet are mainly dairy products, as well as fish and plant products [4,5]. Calcium can be supplemented with mineral water, too. Drinking water containing appropriate amounts of calcium ions (within the recommended range of 20–200 mg L⁻¹) can, among other things, prevent circulatory system diseases and resulting mortality [3]. However, highly mineralized water can only be a supplement to the diet. For example, the deteriorating quality of drinking water in Pakistan, which, due to anthropogenic causes, is manifested, for instance, as an increase in calcium content (often significantly above 300 mg L⁻¹), and is attributed to diseases such as urinary tract infections, stomach ulcers, and tooth decay [6]. Anthropogenically significant increases in calcium content in groundwater to levels exceeding 440 mg L⁻¹ are also a problem in approximately 43% of studied districts in India [7].

Precise and convenient analytical methods are needed for the routine quantitative determination of calcium in water and food. The most used methods are atomic absorption spectrometry, ICP-OES, ICP-MS, calcium-selective electrodes, and complexometric titration [8,9,10,11]. The method of complexometric titration with ethylene diamine tetraacetic acid (EDTA) has a long history and still remains popular in analytical practice [12,13,14,15,16,17,18,19,20]. The advantages of the titration method are its high accuracy and low hardware requirements. The disadvantages of manual titration are increased labor costs and subjective errors in color perception, which reduce the accuracy of endpoint detection. Automatic titration with a calcium ion-selective electrode provides better repeatability; however it is time-consuming because each titration step takes several minutes for the electrode potential to stabilize [21,22]. More convenient detectors include photometric immersion probes that provide immediate responses and have high sensitivity [23]. However, air bubbles in the optical gap may disturb the photometric signal [24]. Suspended particles and droplets also interfere with the optical immersion probes, leading to spikes and erroneous detection of the endpoint.

The above drawbacks motivate analysts to seek a more convenient technique for detecting the endpoint. Currently, the most convenient and low-cost option for optical detectors is webcams. Commercially available CMOS sensors have high sensitivity and a wide dynamic range, which facilitates the automation of the titration procedure. Averaging signals from multiple pixels minimizes fluctuations caused by air bubbles and suspended particles. An additional advantage of this technique is the lack of a washing procedure for electrodes or immersion optical probes.

Digital cameras and scanners are already used in analytical chemistry, mainly for the direct measurement of transmittance and reflectance [25,26,27,28,29,30,31,32]. Several scientific reports have addressed the use of digital cameras for endpoint detection in titration techniques [33,34,35,36,37,38,39]. Different color models are used to detect the color transition [33,38,40]. The Red–Green–Blue color model is most commonly used. The Hue–Saturation–Intensity and Hue–Saturation–Value models may provide less fluctuation in the titration line due to the separation of color and luminance information [40]. The Hue parameter provides smooth titration graphs using an open mixing chamber without rigorous lighting control [38]. The Hue parameter and the inner product of RGB vectors have been compared using the same set of digital images [41]. Both parameters proved to be excellent in detecting endpoints.

So far, there have been few published scientific reports on testing complexometric indicators with RGB sensors for the titrimetric determination of calcium. Image analysis was successfully used to detect the endpoint in the complexometric determination of water hardness at pH = 10 [35]. The Eriochrome Black T indicator was examined with both the RGB and Hue–Saturation–Intensity models. The conclusion was that the most sensitive signal is the Red one [35]. The Eriochrome Black T indicator was also used within the framework of the Hue–Saturation–Value color model [37]. A digital camera was also used to document the course of the titrimetric determination of alkali metals and calcium with ionophore-based reagents [36,39]. Water hardness was determined by EDTA titration using a Calmagite indicator and a 6-channel spectral sensor board as an endpoint detector [42]. The six spectral signals were processed using the K-nearest neighbor classification algorithm, which allowed for reliable detection of color change.

This work describes the optimal lighting conditions and color parameters for recording sharp endpoints with a digital camera, using calcein as an indicator. Calcein is a typical calcium-sensitive indicator used in complexometric titrations [14,15,43,44,45]. It has been proven that the optimal selection of measurement conditions for a webcam allows for the precise determination of calcium in water solution.

2. Materials and Methods

2.1. Reagents, Instrumentation, and Lighting

The standard volumetric solutions of EDTA (0.025 M) and calcium chloride (0.01 M), as well as sodium hydroxide, for titrimetric determination were purchased from Chempur (Piekary Śląskie, Poland). The CRM was calcium carbonate, purchased from Johnson Matthey Chemicals Ltd. (London, UK), and the indicator was calcein (Figure 1a), purchased from Sigma-Aldrich (Saint Louis, MO, USA). The reagents used for the AAS method were 65% nitric acid suprapur (Merck, Darmstadt, Germany), a standard solution of calcium nitrate in 0.5 mol L⁻¹ nitric acid of Ca concentration 1000 mg L⁻¹ (Merck, Darmstadt, Germany), and lanthanum chloride (Sigma-Aldrich, Saint Louis, MO, USA). The commercial mineral waters used were of Polish origin. The other samples of water were taken from the municipal water supply system and from the Vistula River, in the city of Bydgoszcz, Poland.

The measuring set is shown in Figure 1b. The Titronic 300 burette and the Logitech C270 webcam were controlled by a computer with custom software ChemiON (v. 1.4.8.5 of the author’s program by Jan Lamkiewicz). ChemiON is a Java program used for controlling and collecting data from various input and output devices [46]. In diffuse reflection mode, the background was white, and the light source was an LED lamp (1260 lm, 4500 K) placed 80 cm above a milky polypropylene beaker. In fluorescence mode, the background was a sheet of black paper, and the light source was an LED spotlight (360 lm, 6000 K) placed 25 cm above the glass beaker. The automatic gain function of the webcam was turned off. The AAS measurements were performed using spectrophotometer iCE 3500 Series AA (Thermo Scientific, Waltham, MA, USA) at a wavelength of 422.7 nm.

2.2. Titration Procedure and Image Analysis

Test solutions for the determination of linearity range, effect of calcein dose, and sampling area were prepared by diluting exact volumes of standard 0.01 mol L⁻¹ calcium chloride solution with deionized water. A solution of CRM with a calcium concentration of 10 mg L⁻¹ was prepared by dissolving CRM using ultrapure water and HNO₃. The river water was filtered through a medium-hardness filter while tap water and mineral waters were not filtered. If necessary, mineral water samples were diluted to maintain the amount of calcium in the linearity range. Then, the pH was adjusted to 12.5 with 8 mol L⁻¹ NaOH solution, and the exact volume (of 0.1 to 1.2 mL) of 1 mmol L⁻¹ calcein indicator solution was added.

The titration procedure was performed by adding standard EDTA solution (0.025 mol L⁻¹) in increments of 0.01 or 0.05 mL. Images were recorded 10 s after each dose step. The images of the reaction mixture were analyzed with ChemiON software. The sampling area was set by the operator. Each of the RGB color components was averaged over the sampling area. The transformation of the RGB values (8-bit integer) to the Hue parameter (degrees) was carried out according to the HSI color model. The variance of the endpoint volume was calculated from the results of ten determinations.

2.3. AAS Measurements

Samples with high concentrations of calcium were diluted with deionized water to obtain calcium concentrations within the range of the standard curve, i.e., from 1 to 20 mg L⁻¹. In total, 100 mL of solution was acidified with 1 mL of nitric acid. In order to eliminate matrix interference effects, 10 mL of lanthanum chloride solution (20 g La L⁻¹) was added.

3. Results and Discussion

The calcein indicator has four carboxyl groups that coordinate calcium ions (Scheme 1). Since the carboxyl groups are weakly dissociated, it is assumed that only two of the four groups bind the Ca ion. The complex has a light green color and, under appropriate conditions, fluoresces. Figure 1b shows the fluorescence of calcein under LED light. During titration, calcium combines with EDTA to form a more stable complex and the free indicator is released (Scheme 2). So, the color of the solution changes to that characteristic of free calcein (desaturated dark orange). Free calcein does not fluoresce.

3.1. Detecting the Endpoint in Diffuse Reflection Mode

Figure 2 shows color changes in analyte solution during titration in diffuse reflection mode. The color transition is well recognizable using the Hue and green signals (Figure 3a,b). The primary blue color is less suitable due to a narrow range of changes in parameter values, and the red signal is unsuitable for detecting the endpoint (Figure 3c,d). The optimal concentration of the calcein indicator is in the range of 4 to 10 μmol L⁻¹, but in the case of the Hue signal, a concentration of 2 μmol L⁻¹ proved to be sufficient.

The repeatability of the measurements was determined using high and low concentrations of calcium (

Table 1. Standard deviation, relative standard deviation, and recovery towards standard Ca solutions determined using 6 μM calcein indicator in diffuse reflection mode.

Ca Amount, μmol	Signal	SD, μmol	RSD, %	Recovery, %
50	Green	0.375	0.72	104.8
50	Blue	0.830	1.60	103.5
50	Hue	0.375	0.73	102.8
5	Green	0.305	6.07	100.5
5	Blue	0.250	5.00	100.0
5	Hue	0.100	2.08	96.0

). In full agreement with the titration graph shapes, calcein indicator provides the most precise results using the Hue parameter compared to the primary RGB colors. The Hue parameter is characterized by particularly good performance at low calcium contents (Table 1). At higher Ca contents, the green signal also ensures a relatively low value of relative standard deviation. Additionally, each of the tested signals provides a satisfactory recovery value (Table 1), which indicates good accuracy in the determination. It can be concluded that both the Hue and Green parameters are quite suitable for automatic endpoint recognition in diffuse reflection mode.

3.2. Detecting the Endpoint in Fluorescence Mode

The calcein indicator is well suited for detecting the endpoint under intense artificial lighting. The LED spotlight induces bright fluorescence in the calcein indicator complexed with calcium ions (Figure 4, Scheme 1 and Scheme 2). The emitted light is well suited for recording on a black background (Figure 1b).

The initial color for moderate calcein concentration is lime green with RGB values around 90, 250, and 100 (Figure 4). After all the calcium ions are bonded with EDTA, the color changes to very dark green with RGB values around 50, 100, and 5 (Figure 4). The color changes are far more pronounced compared with those in the diffuse reflection mode (cf. Figure 2).

Another difference between the image recording modes is color uniformity. In diffuse reflection mode, the color of the solution is fairly uniform over the height of the titration beaker. In fluorescence mode, the color of the solution changes from top to bottom. The top layer, closest to the LED light, has the lightest color, with deeper layers becoming progressively less bright. Figure 5a–c show the tested positions of the color sampling area. The location of the sampling area has a significant impact on the signal drop at the endpoint (Figure 5d,e). The higher the sampling area location, the larger the signal drop. The obvious reason for this is the internal filter effect in calcein solutions. The intensity of excited light is lower in deeper layers of the calcein solution because of absorption in the higher layers. For this reason, the sampling area should be located at the upper part of the image.

The applied calculation procedure allows for a free choice of sampling area size (Section 2.2). Figure 6 shows the influence of the sampling area size on the sharpness of the endpoint detection. The increase in sampling area results in a slight decrease in the signal drop at the titration endpoint. In the case of the blue component, increased signal fluctuations are observed for small sampling areas (Figure 6b). Therefore, in the case of the Blue component, it is averaging the signal over a larger area is recommended.

Another parameter influencing the titration curve is the amount of calcein in the analyzed solution. Figure 7 shows the titration curves obtained in the presence of different concentrations of the calcein indicator.

At the end point of the titration, the green signal drops rapidly by 130 units (Figure 7a). The concentration of the calcein indicator has little effect, especially above 6 μM. The blue signal drops by 60 to 100 units (Figure 7b). The red and Hue signals have smaller ranges of change (Figure 7c,d). It can be concluded that in fluorescence mode the calcein indicator is applicable in the concentration range from 1 to 12 μmol L⁻¹.

The selected optimal parameters allow for a reliable determination of the calcium content in the solution. Figure 8a,b show titration graphs obtained for calcium amounts ranging from 5 to 250 μmol. In fluorescence mode, the emitted light does not depend on ambient lighting due to stable luminous flux from the spotlight used (Figure 1b). Additionally, the black background minimizes the reflectance of diffused ambient light. As a result, the titration graphs have uniform character throughout the measurement series (Figure 8a,b). As the calcium concentration increases, the endpoint drop in the green signal decreases only slightly (Figure 8a). Therefore, the upper range of determination is probably far higher than the tested value of 250 μmol. Figure 8c shows the excellent linearity of the method. The coefficients of determination are R² = 0.9998 and R² = 0.9995 for the green and blue color components, respectively. Both the linear equations have the same slope of 0.0398 mL μmol⁻¹. The graph shift values are irrelevant (0.1293 mL and 0. 0986 mL for the green and blue color components, respectively).

The variance in calcium quantitation in fluorescence mode was determined using high and low concentrations of calcium (

Table 2. Standard deviation, relative standard deviation, and recovery towards standard Ca solutions determined using the calcein indicator (6 μM) in fluorescence mode.

Ca Amount, μmol	Signal	SD, μmol	RSD, %	Recovery, %
50	Red	1.63	3.14	103.7
50	Green	0.375	0.71	105.3
50	Blue	0.575	1.11	103.3
50	Hue	1.09	2.07	105.5
5	Red	0.295	5.92	100.0
5	Green	0.175	3.52	99.5
5	Blue	0.315	6.57	96.5
5	Hue	0.305	4.68	130.0

). The calcein indicator provides quite precise results for all the color characteristics in question (Table 2). The green signal provides lesser values of relative standard deviation due to a larger fall at the endpoint. The other color parameters also provide acceptable precision (Table 2). The accuracy of the determination expressed as percentage recovery is satisfactory, with the exception of the Hue parameter applied to low calcium contents. Thus, the obtained results correspond to the shapes of the titration curves shown in Figure 7.

Table 3. Comparison of results obtained for CRM with two analytical methods.

Titration			AAS
Signal	Ca Amount, mg L−1	Recovery, %	Ca Amount, mg L−1	Recovery, %
Red	10.52	105.2	9.72	97.2
Green	10.82	108.2
Blue	10.32	103.2
Hue	10.72	107.2

shows a comparison of the titration method and AAS used for the Certified Reference Material. The recovery of the titration method is slightly higher than that of the AAS method (Table 3).

The repeatability of the titration method and AAS was also compared by analyzing natural waters with different mineralization (

Table 4. The results of calcium determination using the calcein indicator (6 μM) in fluorescence mode as well as using the AAS method for natural water samples.

Sample	Titration				AAS
	Signal	Mean Ca Concentration, mg L−1	Intra-Day RSD, %	Inter-Day RSD, %	Mean Ca Concentration, mg L−1	Intra-Day RSD, %
A2	Red	2.89	7.44	9.35	2.70	1.98
	Green	2.81	5.05	4.83
	Blue	2.85	5.52	5.97
	Hue	2.97	6.55	4.36
C3	Red	302.2	3.05	1.15	301.1	0.95
	Green	306.6	1.17	1.00
	Blue	307.8	1.94	2.09
	Hue	310.2	1.22	1.58
Tap water	Red	74.93	2.16	7.73	70.03	1.60
	Green	75.45	1.04	0.96
	Blue	73.95	1.90	2.62
	Hue	76.25	1.37	1.00
River water	Red	79.53	1.99	4.38	74.38	0.64
	Green	79.91	0.83	0.94
	Blue	80.03	2.02	1.97
	Hue	80.79	0.60	1.16

). Again, the best titration repeatability was obtained using the Green signal. The largest values of RSD were observed for the titration method in the case of the low-mineralized water A2 (Table 4). The obvious reason is that the amount of analyte is close to the lower limit of quantification (Figure 8a). The inter-day and intra-day values of RSD are similar (Table 4), indicating a good stability of the titrant solution. Calcium concentrations determined by the AAS method confirm the correctness of the results obtained by the titrimetric method. The results shown in Table 3 and Table 4 suggest that the calcium concentrations determined by the AAS method are slightly underestimated, and the concentrations obtained by the titrimetric method are slightly overestimated.

Table 5. Reproducibility of calcium determination using the calcein indicator (6 μM) in fluorescence mode.

Sample	Signal	Operator 1		Operator 2
		Mean Ca Concentration, mg L−1	RSD, %	Mean Ca Concentration, mg L−1	RSD, %
C3	Red	302.2	3.05	308.6	2.22
	Green	306.6	1.17	312.0	1.60
	Blue	307.8	1.94	312.0	2.31
	Hue	310.2	1.22	316.6	1.21
River water	Red	79.53	1.99	87.37	17.22
	Green	79.91	0.83	79.92	0.49
	Blue	80.03	2.02	78.96	1.87
	Hue	80.79	0.60	80.96	0.93

presents the reproducibility of the measurements using the titrimetric method determined for mineral and river waters. The results obtained by two operators did not differ significantly, especially when the Green signal was used to determine the titration endpoint. Discrepancies in test results were probably due to human errors during the solution preparation.

Based on the measurement results shown in Table 4 and Table 5, it can also be seen that the repeatability of measurements for mineral water may be worse than that for natural water. The bottled water C3 is slightly carbonated. For this reason, despite degassing the sample before testing, the scatter of results may be slightly greater than in the case of non-carbonated waters (i.e., river and tap waters).

3.3. The Environmental Friendliness of the Titration Procedure

The developed procedure for semi-automatic calcium titration in fluorescence mode was compared with the standard AAS measurement method in terms of environmental friendliness. The tools used for this purpose were AGREE and ComplexMoGAPI tests.

The AGREE (Analytical GREEnness Metric Approach) test is based on the 12 principles of Green Analytical Chemistry [47]. Within each criterion, the calculator generates partial values, and the resulting final score ranges from 0 to 1, with higher values indicating better environmental friendliness. The ratings are assigned a color, which smoothly changes from Red (least environmentally friendly) through Yellow to Green (most environmentally friendly). The AGREE test has been frequently used to evaluate analytical methods used in pharmacy [48,49]. The GAPI (Green Analytical Procedure Index) tool [49] is a modification of the AGREE test. It assesses the greenness of the entire analytical methodology, including the sampling step, which is not included in AGREE. It assigns three values to individual criteria: Green, Yellow, or Red, understood in the same way as in the AGREE method. It was used, among others, to evaluate analytical procedures used in the determination of biogenic amines in wine samples and the determination of polycyclic aromatic hydrocarbons (PAHs) [50], as well as for determining pesticides in grapes [51]. The GAPI test is considered more informative compared to other tests. The procedure can be extended with additional criteria, which enables the assessment of the greenness of the synthesis/production of organic compounds or solvents, nanomaterials, or stationary phases (ComplexGAPI [52]). The ComplexMoGAPI test provides an overall assessment [53]. The assessment criteria according to the AGREE and ComplexMoGAPI tests are listed in

Table 6. Results of analytic method comparison using the AGREE test. The parameter weight is given in brackets.

No	Principle	ASA	Semi-Automatic Titration
1	Direct Analytical Techniques Should Be Applied to Avoid Sample Treatment	Off-line analysis (2)	Off-line analysis (2)
2	Minimal Sample Size and Minimal Number of Samples Are Goals	4 mL (1)	25 mL (1)
3	In Situ Measurements Should Be Performed	Off-line (2)	Off-line (2)
4	Integration of Analytical Processes and Operations Saves Energy and Reduces the Use of Reagents	4 steps (2)	3 steps (2)
5	Automated and Miniaturized Methods Should Be Selected	Semi-automatic, not miniaturized (2)	Semi-automatic, not miniaturized (2)
6	Derivatization Should Be Avoided	No derivatization (2)	No derivatization (2)
7	Generation of a Large Volume of Analytical Waste Should Be Avoided and Proper Management of Analytical Waste Should Be Provided	100 mL (2)	100 mL (2)
8	Multianalyte or Multiparameter Methods Are Preferred versus Methods Using One Analyte at a Time	1 analyte/run, 60 samples/h (2)	1 analyte/run, 4 sample/h (2)
9	The Use of Energy Should Be Minimized	Flame atomic absorption spectrometry (2)	Titration (2)
10	Reagents Obtained from Renewable Source Should Be Preferred	None of the reagents are from bio-based sources (2)	None of the reagents are from bio-based sources (2)
11	Toxic Reagents Should Be Eliminated or Replaced	Toxic reagents: 0.60 g (2)	Toxic reagents: 0.36 g (2)
12	The Safety of the Operator Should Be Increased	Toxic to aquatic life; highly flammable; explosive (2)	Toxic to aquatic life, corrosive (2)

and

Table 7. Results of analytic method comparison via ComplexMoGAPI test.

No	Criterion	Score
		AAS	Semi-Automatic Titration
Sample Preparation
1	Collection	Off-line	Off-line
2	Preservation	None	None
3	Transport	Required	Required
4	Storage	Under normal conditions	Under normal conditions
5	Type of method: direct or indirect	No sample preparation	No sample preparation
6	Scale of extraction	Not applicable	Not applicable
7	Solvents/reagents used	Non-green solvents/reagents used	Non-green solvents/reagents used
8	Additional treatments	None	None
Reagents and Solvents
9	Amount	10–100 (10–100 g)	10–100 mL (10–100 g)
10	Health hazard	Moderately toxic; could cause temporary incapacitation; NFPA = 2 or 3	Moderately toxic; could cause temporary incapacitation; NFPA = 2 or 3
11	Safety hazard	Highest NFPA flammability, instability score is 4	Highest NFPA flammability, instability score of 0 or 1. No special hazards
Instrumentation
12	Energy	≤1.5 kW h per sample	≤0.1 kW h per sample
13	Occupational hazard	Emission of vapors into the atmosphere	Hermetic sealing of analytical process
14	Waste	>10 mL (>10 g)	>10 mL (>10 g)
15	Waste treatment	No treatment	No treatment
16	Quantification	Procedure for qualification and quantification	Procedure for qualification and quantification
Yield and Conditions
I	Yield	>89%	>89%
II	Temperature/time	Room temperature, >1 h, Heating, <1 h	Room temperature, <1 h
Relation to Green Economy
III	Number of rules met	1	3
Reagents and Solvents
IVa	Health hazard	Moderately toxic; could cause temporary incapacitation; NFPA = 2 or 3	Moderately toxic; could cause temporary incapacitation; NFPA = 2 or 3
IVb	Safety hazard	Highest NFPA flammability, instability score is 4	Highest NFPA flammability, instability score of 0 or 1 No special hazards
Instrumentation
Va	Technical setup	Additional setups/semi-advanced instruments used	Common setup
Vb	Energy	≤1.5 kW h per sample	≤0.1 kW h per sample
Vc	Occupational hazard	Emission of vapors into the atmosphere	Hermetic sealing of analytical process
Workup and Purification
VIa	Workup and purification of the end product	Not applicable	Not applicable
VIb	Purity	Not applicable	Not applicable
	E-Factor	24	4

. Figure 9 shows the final result of the comparison.

In terms of environmental friendliness, the semi-automatic titration procedure and the AAS method have their own advantages and disadvantages. According to the AGREE test, semi-automatic titration requires the use of larger analyte portions and is less efficient (criteria 2 and 8 in Table 6 and Figure 9) compared to the AAS method. On the other hand, the semi-automatic titration is a more economical method than AAS in terms of the organization of the analytical procedure and energy consumption (criteria 4 and 9 in Table 6 and Figure 9). Moreover, titration is preferred due to the lower consumption of toxic reagents and their lower toxicity (criteria 11, 12 and 7, respectively, in Table 6 and in Figure 9). As a result, the total score of semi-automatic titration according to the AGREE test is comparable to the AAS method (Figure 9). The advantages of the semi-automated titration method are low energy consumption and better operator safety.

The ComplexMoGAPI test showed that the considered analytical methods differ in terms of energy consumption (parameters 12, Vb, and II in Table 7 and Figure 9), the degree of complexity of the procedure (parameter Va in Table 7 and Figure 9), and operator safety related to exposure to toxic reagents (parameters 11 and IVb, as well as 13 and Vc in Table 7 and Figure 9). The semi-automated titration method proved to be more environmentally friendly. Additionally, this method satisfies more green chemistry rules (parameter III in Table 7 and Figure 9). The overall rating of semi-automatic titration according to the ComplexMoGAPI test is much better than that of AAS method (Figure 9).

3.4. Determination of Calcium in Mineral Water Samples

Table 8. Calcium content in mineral water samples.

Brand	Carbonation	Ca Content, mg L−1
		Determined	Declared
A1	non-carbonated	156	162
A2	non-carbonated	2.81	3.27
AK	non-carbonated	50.7	40.3
C1	non-carbonated	111	130
K1	non-carbonated	56.3	68.1
KB1	non-carbonated	45.3	41.0
M1	non-carbonated	34.7	37.4
NZ	non-carbonated	101	96.2
O1	non-carbonated	144	123
OM	non-carbonated	144	148
P1	non-carbonated	51.3	48.1
S1	non-carbonated	152	124
SE	non-carbonated	128	107
Z1	non-carbonated	41.3	41.7
A2	low-carbonated	224	198
C2	carbonated	128	130.3
C3	low-carbonated	306.6	325.8
K2	carbonated	75.3	65
KB2	carbonated	45.3	41
M2	carbonated	42.7	37.4
MU1	low-carbonated	212	208
MU2	low-carbonated	131	200
O2	carbonated	142.7	123
P2	carbonated	54.7	48.1
S2	low-carbonated	111	155
Z2	highly carbonated	60.0	59.3

shows the results of calcium determination in the commercial mineral waters. Determined calcium content ranges from 2.81 to 306.6 mg L⁻¹. The measured values are usually in line with the values declared by the manufacturers. Discrepancies greater than 20% occur only in cases AK, S1, SE, MU2, and S2. This is not only related to extremely low or high calcium contents or carbon dioxide contents.

Discrepancies between the values determined and declared by the manufacturer (Table 8) may result from the fact that the composition of water changes over time. To study this question, water samples of the brand O1 were tested for a long time (Figure 10). This local mineral water source has been operating for 130 years and supplies water with a calcium–hydrogen carbonate character. The spring is fed by rainwater infiltrating through the ground into the deposit. It is the quaternary sediments that are the source of calcium, which is gradually washed away by precipitation. Such springs are typical in the area of north-western Poland [54].

It turned out that the concentration of calcium in O1 brand samples varies within a fairly wide range (Figure 10). The obtained results suggest that calcium concentration is subject to seasonal trends. Higher calcium concentrations occur in summer, while lower concentrations occur in spring and autumn. It has been reported that seasonal changes in groundwater calcium concentration are caused by changes in precipitation [55]. Probably, the observed changes in calcium concentration (Figure 10) are also related to the amount of precipitation. However, the effect of precipitation is not clearly pronounced because the infiltration process occurs rather slowly and is limited by absorption in soil as well as by the evaporation of rainwater in the summer. The concentration of calcium in groundwater may also slowly change as a result of climate change. Summers are becoming warmer and drier, with occasional heavy rains. As a result, an increase in calcium content in groundwater has been observed in many regions of Europe. For example, calcium–hydrogen carbonate groundwater from Debrecen (Hungary) was analyzed [56]. These waters are supplied with calcium by the slow dissolution of minerals: calcite, dolomite, and aragonite. Over a 5-year period, an increased dissolution of minerals was observed, leading to an increase in Ca ion concentration in groundwater. Similar observations have also been made in the Middle East. In Iran, a 10-year study showed an increase in calcium content in groundwater, which, however, is attributed to excessive groundwater extraction and drought, while insufficient aquifer recharge is maintained [57]. The decline in groundwater levels in this country has been confirmed by independent research [58]. Furthermore, it was shown that the titration method with semi-automatic end-point detection can also be used for other natural waters (Table 4). The result of the tap water test corresponds to the calcium content given in the current report of the local control station [59], with a difference of 8.5%.

4. Conclusions

Inexpensive webcams work well for automated endpoint detection in EDTA calcium titrations using the calcein indicator. The sharpness of the color transition depends on both the concentration of the calcein indicator and the lighting conditions during titration. In diffuse reflection mode, the Hue parameter and the Green component of the color provide greater precision than other primary colors. The Hue parameter is also more resistant to fluctuations in ambient light intensity. In fluorescence mode, the Green primary color provides very sharp endpoints. In this measurement mode, other color components and the Hue parameter can also be used, provided that the indicator concentration is properly selected. The advantage of the fluorescence mode is the minimal impact of the ambient light.

The automatic detection of the endpoint provides good precision. Taking into ac-count the minimum value of the registered standard deviation (0.1 μmol) and the sample volume (0.1 L), the Limit of Detection is estimated to be 2 μmol L⁻¹ and the Limit of Quantitation to be 6 μmol L⁻¹. It has been found that the calcium content in mineral waters given by the manufacturer may differ from the actual content. Fluctuations in calcium concentration may be seasonal and related to variable rainfall.

The proposed endpoint detection method is a convenient, inexpensive, and environmentally friendly solution for routine measurements of calcium concentration in aqueous solution. The automatic burette with a webcam detector can also be used to determine other metal ions in water samples.