Millifluidic Ion-Selective Optodes with Integrated Optical Detection for Colorimetric Quantification of Urinary Calcium

Abstract

Measurement of urinary calcium concentration is essential for the diagnosis and management of various disorders related to calcium metabolism, including kidney stones, hypoparathyroidism, and hyperparathyroidism. Current testing protocols require patients to visit clinical laboratories or hospitals, and no at-home devices are currently approved for quantitative self-testing of urinary calcium. In this work, we report a low-cost and user-friendly colorimetric sensor for urinary calcium measurement. The sensor consists of a segment of sensing oil contained within a millifluidic channel, such as a pipette tip. The oil incorporates hydrophobic sensing chemicals typically used in ion-selective optodes, including a chromoionophore, an ion exchanger, and an ionophore. It is specially formulated to fully extract calcium ions from urine under exhaustive conditions, providing a highly sensitive response. The sensor exhibits a linear response for 0 to 10 mM calcium ions, enabling accurate diagnosis of both hypercalciuria and hypocalciuria. A custom-built absorbance detector, consisting of a paired light-emitting diode and photodetector, is attached to the pipette tip to measure the absorbance of the sensing oil and quantify calcium ions in urine. The colorimetric sensor is applied to five urine samples before and after calcium spiking. The calcium concentrations of the unspiked samples determined by the current method overall match those obtained by the reference methods although there are biases. The recovery rate ranges from 94% to 103% for the spiked samples, suggesting high reliability of the biphasic colorimetric sensor for high concentrations of urinary calcium. Given the cost of < 20¢ for the chemicals and pipette tip for each test and the cost of < $50 for the custom-built optical detector, this sensing method holds promise for urinary calcium testing in resource-limited settings such as at patients’ homes and in clinics.

Graphical abstract

Introduction

Measurement of urinary calcium provides valuable clinical information for the diagnosis and management of disorders related to calcium metabolism. It reflects how much calcium is being excreted by the kidneys, which helps distinguish between different causes of abnormal blood calcium levels. Low urinary calcium suggests hypocalciuria, which may occur in hypoparathyroidism or vitamin D deficiency, whereas high urinary calcium (hypercalciuria) is commonly associated with hyperparathyroidism, excessive vitamin D intake, or kidney stone disease.[1] In patients with hypoparathyroidism, urinary calcium monitoring is particularly important during treatment with calcium and active vitamin D, as excessive supplementation can lead to hypercalciuria and subsequent renal complications such as nephrocalcinosis or kidney stones.[2] Regular measurement of urinary calcium therefore guides therapeutic adjustment, helps assess renal safety, and provides a noninvasive indicator of calcium balance and treatment efficacy.

Despite its clinical significance, urinary calcium testing is currently limited to centralized laboratories and requires patient visits to hospitals or diagnostic facilities. Conventional laboratory methods for urinary calcium measurement, such as colorimetric assays using o-cresolphthalein complexone or Arsenazo III, ion-selective electrode potentiometry, and inductively coupled plasma spectroscopy, offer high accuracy but require specialized instrumentation and trained personnel.[3] Therefore, there is a growing need for simple, rapid, and low-cost analytical platforms that enable point-of-care or at-home quantification of urinary calcium to improve disease management and patient outcomes. Semiquantitative specific test papers have been developed as the only commercially available at-home urine test.[4] However, these paper-based dry chemistry methods are susceptible to interference from urinary components and are therefore not approved by FDA to guide medical decisions.[5] For instance, it has been reported that urinary Mg²⁺ can falsely elevate the measured Ca²⁺ levels, making reliable quantification challenging.[6] There are recently published sensor technologies for urine calcium measurement based on advanced versions of ionophore-based ion-selective electrodes and optodes.[7–12] However, for the paper-based optical sensors, the calibration sensitivity based on the color or fluorescence signal is low for the clinically relevant concentration range.[7,8] Electrochemical sensors suffer from limitations such as long stabilization time and sensor-to-sensor variations.[9–12] Overall, questions remain about their reliability and usability in resource-limited settings, underscoring the need for new sensor technologies for urinary calcium measurement.

In our recent work, we reported an ionophore-based colorimetric sensor capable of measuring electrolytes including calcium from a small volume of blood.[13] The sensor consists of a pipette microtip preloaded with a segment of plasticizer (oil phase) containing sensing chemicals for ion-selective optodes (ISOs). A stepper motor such as that in an electronic pipette or a syringe pump can control the aspiration of an accurate volume of sample and facilitate the extraction-based response. This liquid ISO based on an oil segment, when appropriately formulated, permits highly effective ion extraction and highly sensitive optical response compared to traditional ISOs based on membranes or nanoparticles. The platform has improved sensor-to-sensor reproducibility due to the liquid nature of the sensor and its associated low manufacturing variation. More importantly, the configuration of two liquid segments within a liquid channel enables us to analyze the color of the sensing oil without any optical interference from the sample. Although urine is less complex than blood, it exhibits varying degrees of yellowish color due to different concentrations of urobilin and becomes turbid in the presence of crystals, cells, bacteria, mucus, or proteins. Therefore, our biphasic sensor design may be advantageous over traditional single-phase optical assays for urine analysis. In this work, we tailor the sensing oil based on urinary calcium levels and develop a point-of-care device for quantitative analysis of urine calcium.

Experimental Section

Reagents and Materials:

Calcium ionophore II (CaII, N,N,N′,N′-tetra[cyclohexyl]diglycolic acid diamide), sodium tetrakis[3,5-bis(trifluoromethyl)phenyl]borate (NaTFPB), chromoionophore I (ChI, ETH 5294, 3-octadecanoylimino-7-diethylamino-l,2-benzophenoxazine), hexadecane, dioctyl sebacate (DOS) as well as all salts and buffers were purchased from MilliporeSigma and directly used without further purification.

Preparation of oil-loaded pipette tips and aqueous samples:

Sensing oils are prepared by dissolving sensing chemicals in DOS via sonication. A single-channel Integra VIAFLO or a 12-channel Integra VOYAGER Electronic Pipette is used for liquid handling. Although a multi-channel pipet is used to test responses of multiple aqueous solutions in a high throughput, a real-world sensing system will only need a single-channel pipet or a similar fluid handing device controlled by a stepper motor. The reagent-loaded pipette tip is prepared by sequentially aspirating the sensing oil and hexadecane into the pipette tip, separated by an air gap. Another air segment is aspirated after hexadecane to allow extra space for reassembly of the pipette tip to the electronic pipette. The tip is sealed at the needle end with soft beeswax, and the opposite end is closed using a 9 mm diameter Lasenersm aluminum foil film sticker to prevent leakage. This design allows the pipette tips to be stored and transported. When performing a test, the single‑channel pipette punctures the aluminum sticker seal and removes the beeswax cap to expose the needle tip to the sample. The pipette then executes the programmed aspiration and mixing sequence to initiate the sensing process. Color development is monitored photographically using an iPhone 13 Pro illuminated by a white LED light panel. The hue value of each oil segment is quantified using the Adobe Capture app.The aqueous sample is always diluted by tenfold in this work by adding 10 μL of the sample into 90 μL of 0.1 M MES (4-Morpholineethanesulfonic acid) buffer at pH 6.0. The diluted sample is then aspirated into the pipette tip.

UV–Vis Absorption Spectroscopy:

The absorption spectrum of the sensing oil is collected using a Varioskan LUX multimode microplate reader equipped with a μDrop plate that requires a sample volume of 3 μL.

Portable Optical Detector Coupled to the Pipette Tip:

The optical detection platform is modified from our recently developed one for pipette tips.[14] The detector is constructed using an Adafruit AS7341 10-channel sensor connected to an Arduino Uno R4 WiFi board. A white LED from ams OSRAM serves as the illumination source. The sensor enclosure is fabricated with 8K standard photopolymer resin using an ELEGOO Saturn 4 LCD MSLA 3D printer, while additional components for mounting the Arduino board, LED, and pipette are printed from PLA filament using a Creality Ender 3 V3 Plus FDM 3D printer. Modular designs are printed to accommodate the 12.5 μL pipette tip from Integra Biosciences, ensuring secure fit and reproducible alignment. Data acquisition and real-time visualization are handled via LabVIEW 2025 (National Instruments, USA), which communicates with the Arduino Uno R4 board. Absorbance values are calculated based on the light intensity transmitted through a pure solvent and the sensing oil.

Ca²⁺ Measurements for Human Urine Samples:

Five random single-donor human urine samples were obtained from Innovative Research, Inc. and used without further processing. Pipette tips preloaded with 2 μL of sensing oil and 2 μL of hexadecane are used as the sensors. 1.0 μL of pH 6.0 MES-buffered urine is aspirated into the tip. After completing the pipette-controlled mixing protocol, the absorbance of the oil phase is analyzed to determine the Ca²⁺ concentration. For reference measurements, calcium concentrations are determined using inductively coupled plasma–optical emission spectrometry (ICP-OES) and the Arsenazo III reagent kit from HORIBA (Pointe Scientific). Agilent 5110 ICP-OES is used for filtered (0.45 μm PTFE syringe filter) and 50× diluted urine samples prepared in 3% HNO₃.

Results and Discussion

Working principle of liquid ISOs under the exhaustive response mode

The liquid ISO for Ca²⁺ is a water-immiscible oil segment containing a pH-sensitive chromoionophore (ChI), a cation exchanger (NaTFPB), and a calcium ionophore (CaII). A polypropylene pipette tip is used as a transparent liquid channel to accommodate the sensing oil and the aqueous sample, similar to our previous study.[13] Scheme 1 illustrates the working principle of the ISO and Figure 1 shows the colorimetric response of the ISO toward Ca²⁺ in the aqueous phase. The original sensing oil appears reddish due to the electrically neutral state of the dye. It undergoes a dramatic color change from red to blue upon equibration with the MES buffer at pH 6.0, as protons are extracted into the oil to fully protonate ChI. Na⁺ from the cation exchanger is expelled to the aqueous phase to maintain electroneutrality of the oil phase. When the buffer contains Ca²⁺, Ca²⁺ is extracted into the oil to bind to the ionophore, reducing proton extraction since the total amount of cation exchanger, TFPB⁻, is fixed in the oil phase. This results in a lower degree of protonation of ChI. Correspondingly, a shift in color from blue to red is observed as the concentration of Ca²⁺ increases. When the extracted Ca²⁺ occupies all the anionic sites (TFPB⁻), the degree of protonation of ChI approximates zero and the response is saturated.

Scheme 1.

Response principle of the Ca²⁺ ISO under the exhaustive response mode. A, B, C indicates the response toward low, medium, and high Ca²⁺ concentrations, respectively.

Figure 1.

(A) Photograph of pipet tips containing 2 μL DOS-based sensing oil with 2 μL hexadecane additive equilibrated with 1 μL of 0.1 M MES-buffered solution containing different levels of CaCl₂. The MES buffer is at pH 6.0, adjusted with NH₄OH as the base. All labelled concentrations are before the 10-fold dilution of the sample with the buffer. The sensing oil contains 2 nmol ChI, 2 nmol NaTFPB, and 5 nmol CaII. (B) Absorption spectra of the sensing oil after equilibration with the aqueous phase. (C) Calibration curve based on the hue of the oil segment. (D) Calibration curve based on the absorbance of the oil at 659 nm. Data points in the calibration curve are average ± SD for n=3 measurements.

We and others previously used hue, a color parameter, as the optical signal of ISOs when employing a smartphone as a convenient optical detector.[13,15–17] However, unlike absorbance, the color parameter is not linearly proportional to the concentration of the protonated or deprotonated form of the dye. This leads to a nonlinear calibration curve even if the deprotonation of the dye is proportional to the analyte concentration in exhaustive-mode ISOs as proposed by the Bakker group[18]. As shown in Figures 1A and 1C, for the dark blue color of the oil arising from 0–5 mM Ca²⁺, the change in hue is very small. When the color turns purplish, there is an abrupt increase in hue, resulting in a distorted calibration curve that is not ideal for quantification. In contrast, the use of absorbance of the oil phase allows for the construction of a truly linear response curve, as is shown in Figures 1D. The absorbance at 659 nm is from the protonated form of ChI. The concentration of the protonated ChI ($C_{\mathit{HInd}}$) after equilibration is related to the Ca²⁺ concentration in the aqueous phase ($C_{Ca}\left(aq\right)$) by a simple equation when the response is in its exhaustive mode:

$$C_{\mathit{HInd}}\times V_{\mathit{oil}}=n_{\mathit{HInd}}=n_{\mathit{HInd}}\left(0\right)-2\times n_{Ca}\left(\mathit{oil}\right)=n_{\mathit{Ind}}\left(\mathit{tot}\right)-2\times C_{Ca}\left(aq\right)\times V_{aq}$$

where $n_{\mathit{HInd}}$ is the amount of protonated ChI after extraction of Ca²⁺, $n_{\mathit{HInd}}\left(0\right)$ is the amount of protonated ChI when there is no Ca²⁺ in the buffer, $n_{\mathrm{C}\mathrm{a}(\mathit{oil})}$ is the amount of Ca²⁺ extracted into the oil phase, $n_{\mathit{Ind}}\left(tot\right)$ is the total amount of ChI in the oil phase, $V_{\mathit{oil}}$ and $V_{aq}$ are the volume of the oil and the aqueous phase, respectively. Since ChI is fully protonated when the aqueous phase is the buffer at pH 6.0 without Ca²⁺, $n_{\mathit{HInd}}\left(0\right)equals{n}_{\mathit{Ind}}(tot)$. Given the equal amount of ChI and TFPB in the oil phase, extraction of each Ca²⁺ reduces protonation of two ChI.

The response curve of the liquid ISO matches the theoretical linear relationship for the full dynamic range of 0–10 mM, confirming the unprecedentedly effective extraction of Ca²⁺ from the aqueous sample to the sensing oil. In contrast, the response of previously reported exhaustive-mode ISOs using DOS-based nanoparticles deviates from linearity as the concentration of Ca²⁺ increases.[18,19] In this work, one key innovation that ensures the full ion extraction is the inclusion of hexadecane as an ancillary oil. As shown in Figure 2, when the same sensing chemicals are dissolved in pure DOS at the same concentrations without using hexadecane, the response to higher concentrations of Ca²⁺ also becomes significantly smaller, deviating from the linear relationship. Hexadecane, an ultra-low polarity hydrocarbon with a dielectric constant of only 1.9, lowers the overall polarity of the sensing phase. The less-polar sensor matrix enhances the extraction of primary ions and improves the optical response presumably due to the increased ionophore-ion binding affinity.[20] Due to the high specificity of the calcium ionophore, the optode does not respond to other chemicals in urine such as K⁺, Mg²⁺, and creatinine (Figure S1).

Figure 2.

(A) Photograph of pipet tips containing 2 μL DOS-based sensing oil with 2 μL DOS additive equilibrated with 1 μL of 0.1 M MES-buffered solution containing different levels of CaCl₂. The MES buffer is at pH 6.0, adjusted with NH₄OH as the base. All labelled concentrations are before the 10-fold dilution of the sample with the buffer. The sensing oil contains 2 nmol ChI, 2 nmol NaTFPB, and 5 nmol CaII. (B) Absorption spectra of the sensing oil after equilibration with the aqueous phase. (C) Calibration curve based on the hue of the oil segment. (D) Calibration curve based on the absorbance of the oil at 659 nm. Data points in the calibration curve are average ± SD for n=3 measurements.

Design and operation of the liquid ISOs in pipette tips

The disposable sensor is prepared by preloading 2 μL of the sensing oil and 2 μL of hexadecane into a 12.5-μL pipette tip. The electronic pipette can accurately aspirate a defined volume of aqueous solution into the pipette tip. Then repeated aspiration and dispensing steps will be executed by the programmed pipette to move the liquids back and forth within the pipette tip for enhanced mass transfer. To study the response time of the colorimetric sensor, photos of the sensing oil are taken at different time points. The original oil is reddish before being exposed to aqueous samples due to the deprotonated state of ChI. As can be seen from Figure 3A, although there is still small color development from 10 mins to 20 mins, the color change becomes negligible after 20 mins for all examined concentrations of Ca²⁺, indicating a response time of less than 20 mins under these experimental conditions. We further used a custom-built absorbance detector integrated onto the pipette tip (see below details of the detector) to continuously track the absorbance of the oil phase within the pipette tip. As shown in Figures 3B, the absorbance reaches a plateau after 10 mins, consistent with the color change shown in photos. Figure 3C is an enlarged portion of the absorbance trace in Figure 3B. Notably, all tests in this work are conducted at 40 °C because the elevated temperature enhances mass exchange and reduces the response time from 30–40 minutes to less than 20 minutes. More importantly, temperature control eliminates potential errors caused by variations in ambient temperature across different users. Although we simply use an oven in this study, the temperature control can be readily implemented using a compact and cost-effective thermoelectric Peltier module.

Figure 3.

(A) Photograph of pipette tips containing the sensing oil and different concentrations of CaCl₂ after mixing for 10, 20, and 30 mins in a 40°C oven. (B) Real-time absorbance trace of the sensing oil in the pipette tip when the aqueous phase is 1 μL of 0.1 M MES buffer in a 40°C oven. The absorbance at the 620 nm channel is tracked by the optical detector. (C) The absorbance trace from 300–350 seconds.

For analytes like urinary calcium that vary over a broad physiological concentration range, effective detection requires both wide dynamic coverage and high resolution. The overall physiological range of urinary calcium excretion (100–300 mg/day) spans approximately 0.5 orders of magnitude. However, clinically meaningful thresholds, such as those for hypocalciuria and hypercalciuria, fall within much narrower subranges, particularly after incorporating age- and race-specific reference intervals. For example, the borderline hypercalciuria zone of 250–300 mg/day spans only ~0.08 orders of magnitude. This narrow decision window highlights the need for detection methods capable of resolving small concentration differences across a broad dynamic range.[21–24] As of the lower threshold, a 24-hour urine calcium excretion of less than 37 mg in older women and less than 23 mg in younger women is indicative of hypocalciuria.[23] Therefore, a capacity to show lower limits around 20–40 mg and upper limits covering 250–300 mg per 24 hours would be more recommended for all purpose clinical diagnosis of diseases associated with urinary calcium abnormalities. Given that normal adult urine output ranges from 800 to 2,000 mL per day, our sensor is designed to detect Ca²⁺ within the clinically relevant range of 0.5–10 mM, as is shown in Figure 1. By using absorbance of the oil phase, the sensor is also able to differentiate 0.1 to 0.5 mM Ca²⁺ with a resolution of 0.1 mM (Figure S2).

In exhaustive-mode ISOs, the molar quantities of sensing chemicals in the oil phase govern the extraction capacity for Ca²⁺ and determine the response range.[13,18,19,25] Extracting up to 10 mM Ca²⁺ in urine requires comparable concentrations of sensing chemicals in the oil assuming an equal volume of oil and sample. The color of the oil would be very dark with such a high concentration of dye, making both color analysis and absorbance measurement challenging unless an ultrathin channel is used to dramatically reduce the optical pathlength. At this stage, since we are using commercial pipette tips instead of customized channels, we dilute the aqueous sample by tenfold so much lower concentrations of sensing chemicals are needed and colorimetric analysis of the oil becomes feasible. To obtain an upper detection limit of 10 mM, 1 μL of 10-fold diluted sample contains 1 nmol Ca²⁺ and, therefore, requires 2 nm TFPB as the cation exchanger. Accordingly, we use 2 μL of DOS contains 2 nmol ChI, 2 nmol NaTFPB, and 5 nmol CaII, supplemented with 2 μL of hexadecane to enhance the extraction efficiency. The dilution using the MES buffer at pH 6.0 also ensures the same final pH of the aqueous phase, eliminating any pH interference. Even though the ISO in the exhaustive response mode is theoretically pH independent, pH still needs to be controlled especially when the sample has a large variation in pH (e.g., 4.5–8 for urine). This is because a low sample pH could make the ISO less exhaustive and a high pH would deprotonate the dye even without the analyte ions. Moreover, although ChIII with a much higher pK_a was used for blood calcium testing at pH 7.4 to ensure sufficient protonation at low calcium concentrations,[13,26] the pH of 6.0 in this work allows for the use of more affordable chromoionophores like chromoionophore I with lower pK_a values. Chromoionophore I is about 10 times cheaper than chromoionophore III. The total cost of the sensing oil for each test is less than 10¢ even using the price of Millipore-Sigma and the cost of each pipette tip is also below 10¢.

Calcium measurements in human urine samples using the liquid ISO coupled with a custom-built absorbance detector

Although the microplate reader allows for absorbance measurement of the oil phase transferred from the pipette tip to the μDrop plate, this process obviously is not suitable for point-of-care use. We recently developed an optical detection platform using open-source software and hardware components at a total cost below $50.[14] This self-contained optical detector enables in-situ measurement of the absorbance of the oil segment within the pipette tip. In this work, we modified the detector for the urinary calcium testing (Figure 4). The optical window is miniaturized and positioned to face the hexadecane region in the pipette tip, enabling acquisition of a clean blank signal from the transparent oil. A single-channel electronic pipette is employed to execute precise aspiration, dispensing, and multistep automated mixing protocols. The final oil phase is automatically moved into the optical window for measurement via a predefined pipette program. Since the platform measures transmitted light intensity, the data are converted into absorbance values based on the definition of absorbance. To account for slight manufacturing variability among pipette tips, an individual calibration function is included using the blank signal obtained from the preloaded hexadecane. Notably, the transmittance of hexadecane closely matches that of a 1:1 DOS/hexadecane mixture due to their nearly identical refractive indices (1.44–1.45), rationalizing the use of hexadecane as the blank. Figure 5A shows the absorbance measured by the optical detector consisting of a multi-spectral photodiode as the detector and a white LED as the light source. As can be seen, the overall absorption pattern matches the real absorption spectra obtained by the microplate reader. In Figure 5B, the linear response curve for 0–10 mM Ca²⁺ is constructed for the urine analysis for unknown samples.

Figure 4.

Schematic illustration (A) and photograph (B) of the optical detection system for the pipette tip.

Figure 5.

(A) Multispectral absorbance data acquired using the in-situ optical detector. (B) Calibration curve based on absorbance at the 620 nm channel of the sensing oil after equilibration with the aqueous phase. Data points in the calibration curve are average ± SD for n=3 measurements.

The accuracy of the liquid ISO method, coupled with an integrated optical detector for urine calcium analysis, is assessed using Inductively Coupled Plasma Optical Emission Spectroscopy (ICP-OES) as the reference method. As summarized in Table 1, the absolute difference between the two methods for samples 1, 2, and 5 was below 0.25 mM, satisfying the Total Allowable Error for urine calcium measurements defined by the Clinical Laboratory Improvement Amendments (CLIA) guidelines. However, the absolute difference exceeded this threshold for samples 3 and 4. While the two methods overall produce comparable trends, they do not consistently agree according to the Student’s t-test at the 95% confidence level (t_crit=2.776). The t_cal for samples 1 and 4 is smaller than the t_crit, suggesting no statistically significant difference for those samples. Conversely, the t_cal for samples 2, 3, and 5 exceeds the t_crit, indicating the presence of systematic biases. Notably, the Ca²⁺ concentration determined by the ISO method is higher than the ICP reference for 4 out of 5 samples. To further evaluate the ISO method, future studies will incorporate a larger cohort of urine samples. Moreover, refinements to the ISO chemistry and hardware setup may enhance the accuracy and precision of the technology. For example, a possible source of variation is the imperfect alignment between the pipette tip and the optical detector, likely caused by tip-to-tip geometric variations and minor displacement during testing. Future iterations will utilize dedicated pipette tips or millifluidic chips designed for a tighter, fixed fit into the optical detector.

Table 1.

Comparison of urinary calcium levels determined by the proposed ISO method and ICP-OES as a reference technique; recovery rates for spiked urine samples determined by the ISO method.

Urine sample	ICP / mM	ISO / mM	Difference / mM	tcal	Recovery rate
#1	0.41 ± 0.03	0.39 ± 0.13	0.02	0.260
#1 + 5 mM		5.35			99%
#2	0.16 ± 0.01	0.33 ± 0.05	0.17	−5.775
#2 + 5 mM		5.18			97%
#3	1.10 ± 0.13	1.46 ± 0.10	0.36	−3.802
#3 + 5 mM		6.62			103%
#4	3.03 ± 0.38	3.67 ± 0.25	0.64	−2.437
#4 + 5 mM		8.37			94%
#5	0.52 ± 0.05	0.71 ± 0.05	0.19	−4.654
#5 + 5 mM		5.80			102%

The calcium concentration for each unspiked sample is expressed as average ± SD for n=3 measurements. The critical t (t_crit) is 2.776 when the degree of freedom is 4 and the confidence level is 95% in the Student’s t test.

After spiking each urine sample with 5 mM Ca²⁺, recovery rates ranging from 94% to 103% are obtained, indicating better accuracy for higher concentrations of Ca²⁺. We also used the microplate reader to measure the absorbance of the oil phase and employed another reference method based on the commercial Arsenazo III colorimetric assay kit for more comprehensive comparison of various methods (Table S1).

Conclusions

In this study, we present a robust, low-cost, and user-friendly biphasic sensing system for quantitative detection of calcium ions in urine. The sensor utilizes hexadecane as an oil additive to enhance calcium ion extraction efficiency while minimizing the use of costly ionophores and incorporating a more economical chromoionophore compared to previous systems designed for blood calcium analysis. Engineered for complete ion extraction, the sensor exhibits high sensitivity across the clinically relevant urinary calcium range. Using a programmed multistep aspiration and mixing protocol controlled by an electronic pipette, the system achieves a response time of less than 20 minutes. An open-source Arduino-based optical detection platform, assembled for under $50, enables both spectral analysis and remote data acquisition. This technology provides a novel approach for at-home urinary calcium testing and also represents a flexible platform for future expansion into other ion-sensing applications. One limitation is that urine dilution is currently required to ensure complete calcium extraction and pH control of the sample phase. Future improvements, such as a custom-designed millifluidic channel with a shorter optical pathlength and dried buffer coatings, may eliminate the need for dilution. On the other hand, it is feasible to design simple devices to enable patients to dilute urine using provided buffer.

Highlights.

• A colorimetric sensor based on a liquid-type ion-selective optode is developed for calcium ion detection.

• The sensor operates under the exhaustive response mode and shows high sensitivity and linearity.

• A compact optical detector is used to detect the absorbance of the liquid optode in resource-limited settings.

• The sensor is applied for calcium ion measurements in five human urine samples.

Data availability

The data supporting this study are available within the article and its Supplementary Information File. Other data are available from the corresponding author upon reasonable request.