TD-DFT Guided Advanced E-Eye Sensing Technique for On-site Quantification of Fe, Cr, F, and As in the Environmental, Biological, and Food Samples

Summary

This work has discussed the protocol for designing and developing an E-Eye-Enable POCT device to detect Fe, Cr, As, and F in environmental samples, biological samples, and food and beverages. Nearly 2000 samples have been tested, and the results accord with the gold-standard technique.

Abstract

This paper demonstrates the step-by-step protocol of designing and developing an E-Eye-enabled multiplexed point-of-care testing (POCT) kit. We use time-dependent density function theory (TD-DFT) to explore the basic working principle of the sensors before going into the details of the device's hardware. The TD-DFT analysis gives the λ_max of the targeted element, which helps find the most accurate route for detecting the pollutant in the analytes. The TD-DFT analysis is performed in Gaussian 09 and Gauss View 5.0 software. An optical sensor called the electronic eye (E-Eye) has been developed based on the λ_max value and the Light Emitting Diode/Light Dependent Resistor (LED/LDR) principle. The E-Eye device has been fabricated with an LED, an LDR placed opposite each other, and other hardware, including a liquid crystal display (LCD) interfaced with a microprocessor across a voltage divider, acting as a central microprocessor. Initially, all the samples (environmental, biological, and food and beverages) were pre-processed to extract all the targeted elements in the aqueous phase. The specific (lock and key) reaction has been carried out in the specified reactor, and readings observed in the display unit have been recorded. An indigenously multiplexed device has also been fabricated to detect Fe, Cr, As, and F simultaneously. The efficacy of the sensors has been tested against more than 2000 samples and compared with the gold standard methods. A good precision has been confirmed with the accuracy of 95.3%, 94.7% and 95.4% with respect to the samples of environmental, biological, and food and beverages. It is important to note that the performance of the arsenic sensor has been compared with Atomic Absorption Spectrometry (AAS) results, and all other results, i.e., sensors for iron, chromium, and fluoride, have been compared with the results obtained from UV-Vis analysis.

Introduction

Heavy metals play a significant role in maintaining the Earth's ecosystem, which is essential to human beings¹. On the contrary, the industrial growth and manhandling of natural resources in the recent era have led to heavy metals becoming a contaminant to the food chain and other essential commodities of living beings, which is a serious alarm to modern civilization. It is due to their bioaccumulation in living and non-living bodies and the nature of non-degradability². In this regard, the US Government has continuously evaluated the food grade, including baby foods, for a long time to monitor the food quality. Altogether, they analysed more than 22 elements and more than 3200 samples across their country³.

In many cases, the levels of heavy metals crossed the threshold limit recommended by the WHO and the US Food and Drug Administration, and the Total Diet Study Report (TDS), July 2022. A similar conclusion can be drawn on Europe's water bodies and soils^4,5. In addition to soil and water, the strong presence of heavy metals endangers life^6,7. It is important to note that India is a river-based country, and a recent survey shows that more than 43% of the rivers have an alarming level of many heavy metals and other pollutants^8,9. In this program⁸, it was accepted that Blood Lead Levels in more than 27.5 crore children are more than 5 µg/dL, which is considered unsafe.

The above discussions describe the need for continuous monitoring of heavy metals and other hazardous chemicals for the well-being of global society. There are various techniques for detection and concentration measurement for heavy metals, such as UV-Vis spectroscopy, AAS, Inductively Coupled Plasma-Atomic Emission Spectroscopy (ICP-AES), and Inductively Coupled Plasma-Mass Spectrometry (ICP-MS), etc. (US Food and Drug Administration, Total Diet Study Report (TDS), July 2022)³. The necessity of high-end equipment and a skilled workforce restricts these methods for quick and on-site detection of heavy metals. The situation discussed in the above paragraph demands a user-friendly, quick, and accurate on-site detection kit for heavy metals, which can help to monitor the quality of soil, air, and water.

There have been many attempts to develop the aforementioned on-site detection kits^10,11 for heavy metals. Some of them^12,13,14,15 have been successful, and some¹³ have failed to fulfil the requirements of WHO recommendations. A few of them are digital sensors, which display the digital values of concentrations of the analytes^16,17,18. The rest are qualitative sensors like paper-based microfluidic sensors used for the pregnancy test, which can mark the presence and absence of the targeted pollutants in the sample¹⁹. There are several kits of different brands, makes, and models commercially available in the market, which could be routinely deployed to test the presence of heavy metals in water and other samples. These kits are generally graded based on key performance indicators such as detection range, cost per sample, sensitivity, operational reliability, and the technical skills required. For example, for the detection of chromium, most commercial kits offer detection ranges up to 1 mg/L, and a few up to 10 mg/L (Chemetrics²⁰), but their limits of detection (LODs) and least counts are typically above the World Health Organization (WHO) recommended threshold for safe drinking water. For instance, the HACH²¹ kit (5- 100) and (50-1000) mg/L, while covering a broad range, demonstrates poor sensitivity at low concentrations (least count ≥5 mg/L) and costs 125.95 per sample, requiring basic experimental skills and visual interpretation from color change. The Hanna Hi3846²² and Himedia²³ kits, though more affordable (51.00 and 375.00 per sample, respectively), also exhibit limited sensitivity, with the Hanna kit's least count at 0.2 mg/L. Merck²⁴ and Chemetrics²⁰ kits employ digital readouts and span intermediate to high detection ranges (2-10 mg/L). Still, their costs per sample are significantly higher (up to 823.20), and their sensitivity remains suboptimal for regulatory compliance. The BARC kit²⁵, while user-friendly and suitable for rapid field categorization of water safety, also suffers from a relatively high detection limit (≥0.05 mg/L) and qualitative, rather than quantitative, output.

Similarly, for fluoride detection, the commercially available kits such as Aquasol AE210²⁶ (0.1-2 mg/L and 120 mg/L ranges) and Himedia²⁷ (0-2.5 mg/L) kits are among the most affordable at 6 and 11.60 per sample, respectively, but rely on visual colorimetry with unspecified sensitivity limits and time taken to analyse the sample. Mid-range options include the Merck Colourimetric²⁸ (detection range 0-0.8 mg/L, 292.20 per sample) and Photometric²⁹ kits (detection range 0.1-2.5 mg/L, 456 per sample), with the latter offering digital readouts at higher costs. Little sophisticated systems like the Hanna Hi739³⁰ (detection range 20-20 mg/L, 362 per sample) and Prerana Laboratories³¹ (150 per sample) demonstrate trade-offs between detection time (≥1 minute for Hanna versus 10-30 min for Prerana) and analytical precision, though both lack explicit sensitivity specifications. Existing kits either prioritize affordability with compromised sensitivity or improved precision at elevated costs, with the detection limits compromised in the mid-range up to 20 mg/L.

Commercially available iron detection kits exhibit varied performance characteristics, with significant trade-offs between sensitivity, cost, and operational complexity. The AE303³² offers a 0-2 mg/L detection range at 14 per sample, utilizing a visual colorimetric technique. Similarly, another such kit for iron detection, the HI3834³³, covers 0-5 mg/L (52 per sample) with a least count of 1 mg/L and a detection time of ≥ 4 min, requiring the ability to perform basic experiments. The Insa tech³⁴ serves the application of high concentration detection in the range of 50-800 mg/L, though its high cost and sensitivity (≥50 mg/L) with a detection time of 75 min limit its application for general purposes. The Prerana Laboratories kit³⁵ (0-1 mg/L detection range, 28 per sample) improves low-concentration detection (0.1 mg/L sensitivity) but takes an analysis time of 30 min. Hach's test strips³⁶ (0-5 mg/L detection range, unspecified cost) achieve a 0.15 mg/L least count but rely on subjective color interpretation.

Commercial arsenic detection kits demonstrate significant variability in analytical performance, cost, and operational complexity. The BARC field kit³⁷ (0-0.01 mg/L, 5-10 per sample) employs a non-toxic reagent and 10 min colorimetric analysis, offering rural accessibility but relying on subjective visual comparison. The AE408³⁸ (0-0.1 mg/L/0-3 mg/L, 77.40 per sample) prioritizes portability and APHA compliance, though its sensitivity remains unspecified. Higher sensitivity is achieved by the Lovibond kit³⁹ (0-0.5 mg/L, 500 per sample, which detects arsenic at 5 µg/L^, meeting WHO guidelines, through a safe, sulfide-interference-resistant protocol. Premium systems like the Cole-Parmer ITS kit⁴⁰ (2 µg/L detection limit, 480 per sample) and Palintest Visual Kit⁴¹ (200 tests/£560, ~280 per sample) offer enhanced precision with EPA verification or triple-filter systems but incur steep costs.

Commercially available detection kits for chromium, fluoride, iron, and arsenic exhibit critical sensitivity, cost, and operational reliability limitations that hinder their utility in field settings. For chromium, commercial kits like HACH (least count ≥5 mg/L) and Hanna Hi3846 (≥0.2 mg/L) operate above the WHO limit (0.05 mg/L), with costs ranging from 51-823 per test and reliance on error-prone visual interpretation. Similarly, fluoride kits such as AE210 (0.1 mg/L resolution) and Photometric (456 per sample) prioritize affordability or precision but lack rapid, sub-WHO-compliant detection. Iron detection suffers from high least counts (HI3834: 1 mg/L; strips: 0.15 mg/L) and prolonged analysis times (Prerana: 30 min), while arsenic kits like BARC (10 µg/L LOD) and Lovibond (5 µg/L) face reliability gaps below 70 µg/L and steep costs (up to 500/test). The proposed multi-analyte device addresses these lacunae through significant transformative advancements in ultra-sensitive detection, cost-effectiveness, user-friendliness, and rapid and on-site analysis. By bridging the sensitivity-cost trade-off and providing robust performance in complex matrices, this device resolves longstanding literature gaps and lacunae in field-deployable water quality monitoring, particularly in low-resource regions where affordability, precision, and ease of use are paramount. Its ability to detect compliance grade with digital point-of-care functionality represents a paradigm shift in environmental and healthcare analytics, overcoming the limitations of fragmented, analyte-specific commercial solutions.

In the present scenario, society demands a sensor to broadcast the actual concentration of the various pollutants. The existing work and point-of-care testing (POCT) kits fulfil this requirement. So, there is a need for a robust POCT kit with performance comparable to that of high-end equipment like UV-Vis, AAS, or ICP-MS, etc., capable of detecting multiple pollutants. The present work aims to develop a robust method and POCT kits to serve the purpose. Here, a method and a POCT kit for detecting different heavy metals in environmental, biological, and food and beverage samples have been developed. The samples' pre-processing technique was adopted and implemented using the indigenously developed optical sensor to quantify Fe, As, Cr, and F. The optical sensor was developed based on the working principle of LED-LDR; that is why it is called the E-Eyes-enabled POCT kit. The developed multiplexed POCT device reported in this work demonstrates strong practical applicability for on-site water quality monitoring, particularly in resource-limited settings. Each detection slot has a specific analyte using selective reagents and optimized conditions. The Fe²⁺ detection operates effectively within the concentration range of 0.01-5.0 mg/L, with a detection limit (LOD) of 0.017 mg/L. The Cr channel, employing diphenyl carbazide chemistry, covers a dynamic range of 10-500 µg/L and achieves an LOD of 10.78 µg/L. Fluoride sensing through the Fe-SCN complex decoloration method enables reliable detection from 0.5-47.5 mg/L, with an LOD of 0.46 mg/L. Based on the molybdenum blue reaction, arsenic detection is effective over a range of 5-100 µg/L with an LOD of 8 µg/L. All reactions are carried out at room temperature (25 ± 2 °C), and the entire assay is completed in less than 5 min without requiring sophisticated instrumentation or skilled personnel. These features affirm the method's applicability and compliance with WHO water quality standards. Before performing the reaction in the laboratory, the time-dependent density function theory (TD-DFT) was adopted to establish the specific chemical reactions (lock and key) for every analyte. This TD-DFT analysis also predicts the respective probable λ_max of the reagent and the main product of the reaction, which is the fundamental backbone of the sensor and ensures the probable efficacy of the sensor. A prototype has been developed for the E-Eyes-enabled POCT kit, and the prototype has been translated into a device. The whole workflow has been shown in Figure 1. The device has been tested against more than 2000 samples, and its performance was compared with the results obtained using UV-Vis, ICP-MS, and AAS. In all the cases, the device performances are highly satisfactory with minimum and maximum deviations of 1.25% and 6.25%, respectively.

Protocol

1. Computing and validation of the reaction's sensing mechanism

1. Select the lock and key reaction based on the basic chemistry principle of the targeted analyte.
2. Use the Gaussian 09 program to perform all the calculations using the Becke, 3-parameter, Lee-Yang-Parr (B3LYP) hybrid method, and the Los Alamos National Laboratory 2 Double ζ (LanL2DZ) basis set¹⁴ to obtain more accurate results in electronic computation involving the targeted analyte. Here, the protocol is discussed for iron (Fe) only, which applies to all other analytes.
3. Introduce the self-consistent reaction field (SCRF) approach with the polarizable continuum (PCM) model to account for the solvent effect. Water is assumed to be a solvent in this protocol.
4. Optimize the molecular geometry using the geom=connectivity function. It helps to retain explicit bonding information in the process of optimization.
5. Calculate the ground state energy required to compute the optimized structure of the molecule, using the steps mentioned in 1.2 and 1.3.
6. Obtain the excited state electronic transition by the Time-Dependent Self-Consistent Field (TD-SCF) method (ns=60)¹⁴. Use the same method to obtain the UV absorption spectra.
7. Optimize the molecule's geometry by ground-state DFT calculations using the B3LYP functional and LanL2DZ basis set.
8. Get the optimized structure of 1,10-phenanthroline and Fe²⁺ as shown in Figure 2A,B.
9. Similarly, simulate the molecule structure of hexafluoroferrate (FeF₆), which has an octahedral structure with Fe³⁺ as the central atom surrounded by 6 F^- atoms as presented in Figure 2C.
   NOTE: The reaction, i.e., formation of ferroin, is specific for Fe²⁺. That is why, initially, Fe³⁺ should be reduced to Fe²⁺ according to the reaction below.
   
   The Fe²⁺ reacts quickly with 1,10-phenanthroline and forms ferroin, an orange-colored complex visible to the naked eye (see the reaction below). This reaction is the heart of the sensor, as shown in Figure 2.

2. Development of a colorimetric chemical sensor

1. Procure the analytical-grade chemicals required for the above reaction.
2. Transfer equal volumes of trisodium citrate dihydrate, ascorbic acid, and 1,10-phenanthroline into a transparent glass culture tube.
   1. In the experiment, 1 mL of each reactant was taken into a 5 mL transparent glass culture tube. After adding 1 mL of each reagent (1 mL of ascorbic acid, 1 mL of trisodium citrate, and 1 mL of phenanthroline), add 2 mL of stock. If 5 mg/L of stock was added, then the final concentration is 2 mg/L (C1V1 = C2V2), and the final volume is 5 mL.
3. Optimize the reagents' ratio to achieve the pH of the reaction medium using ascorbic acid, which also acts as a masking agent. Optimize the reagents to achieve a pH of 2 at maximum absorbance at a wavelength of 510 nm.
4. Use a masking agent to eliminate interference from copper, as 1,10-phenanthroline also forms a yellow-colored complex with copper in an alkaline environment.
   NOTE: Copper may interfere with the reaction as 1,10-phenanthroline also forms a yellow-colored complex with copper in an alkaline environment. So, maintaining the pH of the medium is very crucial. Ascorbic acid serves the purpose. However, the pH may change with time, and nitrogen of 1,10-phenanthroline may get protonated at this low pH. To overcome these difficulties, trisodium citrate buffer is used, which helps to maintain the pH and to prevent the protonation of 1,10-phenanthroline.
5. Prepare the solutions of Fe²⁺ with different concentrations ranging from 0.05 mg/L to 4.0 mg/L from FeSO₄.7H₂O.
6. Add 1 mL of this solution to the reaction mixture depicted in Figure 3A, which gives orange-colored ferroin as an indicator of the Fe²⁺ (Figure 3B).
   NOTE: Here, adding the stock solution to the reagent mixture is the visual checkpoint, which produces the orange color. Particular attention must be given to reducing agents, as 1,10-phenanthroline is selective to Fe²⁺ and does not form a complex with Fe³⁺.
7. Collect the raw samples and filter using a Whatman pore-size paper, 20-25 µm filter paper. Do the post-filtration dilution if necessary¹⁴.
   1. Collect water samples directly from the environmental sources (e.g., Brahmaputra River) in pre-cleaned, acid-washed high-density polyethylene bottles. Cap the bottles immediately and filter (within 2 h) using a Whatman filter paper of pore size 20-25 µm. Perform a pretreatment to remove the sediments and color from the samples other than spiked DI water (primarily environmental samples; see Supplementary File 1).
8. Measure the amount of iron in the samples using UV-Vis spectroscopy in a wavelength range of 250-750 nm, which is considered the gold standard technique in the present measurement.
   1. Place samples in quartz cuvettes with a 1 cm path length. Take the absorbance measurements at specific wavelengths corresponding to each analyte's peak absorbance in triplicate. Perform baseline correction using a blank containing the same water. Conduct all measurements at room temperature under standard laboratory conditions (scan speed: medium, measurement type: absorbance, slit width: 5.0 L, detectors: direct) to ensure consistency and reliability of results.
      NOTE: Quratz cuvettes used for UV-vis spectroscopy must be optically the same and clear to ensure accurate absorbance measurements. Any visible stain, residue, or scratches may introduce significant errors and warrant immediate replacement.
9. Spike the samples with Fe²⁺ in different concentrations ranging from 0.05 to 4 mg/L, if Fe²⁺ is not found in step 2.8. Reduce Fe³⁺ using ascorbic acid before step 2.8 to estimate total iron in the samples.
   NOTE: All the reactions occur at room temperature. The ferroin complex is formed instantly after adding Fe²⁺. Therefore, the sensor works fine at room temperature, and its response time is very short, within a few seconds.

3. Design and fabrication of the E-Eye-Enable POCT kit's prototype

1. Procure LDR, a white LED, an LCD, a microprocessor microcontroller, and resistors.
2. Arrange an LED and an LDR opposite each other across a 3-D printed box. The box holds a cuvette in which ferroin is formed through the lock and key reaction, as indicated in Figure 4A.
   NOTE: The LED used here emits white light; its diameter is about 5 mm, and its reverse voltage is 5 V. The LDR used in the prototype has a sensing ability of 5 Ω to 10 kΩ. Moreover, this LDR effectively senses the wavelength ranging from 400 to 700 nm emitted by the LED used here.
3. Complete the circuit with the necessary electronic hardware components: resistors (3.3 kΩ), connecting wires, and an LCD (SKU: 11497) unit.
4. Interface all the above components with a microprocessor across a voltage divider circuit portrayed in Figure 4B. Perform programming using the Arduino IDE v1.8.19. Read each channel's LED-LDR-based analogue signal using analog pins (A0-A3).
   NOTE: The microprocessor acts as the central processing unit of the circuit. The optoelectronic data acquisition system was built using an Arduino Uno (Rev3) microcontroller.
5. House the advanced and sophisticated electronic filters (resistance 10 kΩ, capacitance 1 µF) to improve the sensor's efficiency by suppressing the noise and boosting the device's instrumental vigor.
6. Print the printed circuit board (PCB) if the design is complete.
7. Supply the power to the circuit by turning on the electronic switch and recording the sensor's response in terms of resistance.
   1. After switching on the switch, the LED emits light, which passes through the reaction chamber. The LDR senses the intensity of the transmitted light and gives a resistance. Correlate this resistance with the concentration of Fe²⁺ present in the sample. This correlation of the known samples provides a calibration for the plot. Incorporate the equation obtained from this calibration plot within the derived algorithm.
8. The entire circuit diagram is depicted in Figure 4C. Print the device after completing all the steps mentioned above in this section. The actual image of the fabricated device is shown in Figure 4D.
   NOTE: Printing of the circuit board improves reliability and guarantees a robust integration of the electronic hardware components.

4. Development of a multiplexed device

NOTE: Multiple channels are required to detect multiple analytes in a single shot. For this, four channels are installed and integrated with the prototype.

1. Create four slots in the 3D box to hold the four cuvettes in their respective positions. Isolate each detection slot optically and electronically to minimize interference, physically by opaque barriers between channels, and electronically by using separate LED-LDR pairs and individual voltage dividers to prevent signal overlap.
   1. Assign one slot to a particular analyte; do not interchange the slots. The slot assigned to a particular analyte must hold the cuvette that only has reagents specific to that analyte.
2. Repeat the protocols mentioned in steps 3.3 to 3.7. All four pairs of LED-LDR are connected in parallel to the circuit board. One representative circuit diagram and image of the prototype for six such slots (6 slots) are shown in Figure 5.
3. Turn on the switch to supply power to the circuit. All the LEDs illuminate; consequently, the LDRs sense all four analytes simultaneously.
   NOTE: It works on the same principle as discussed in the note under step 3.7. It is essential to note the blank resistance; any change subject to external factors warrants recalibration. The device has been fabricated for UV-vis spectroscopy. However, the optical source (LED) used in this work encompasses a 400-750 nm wavelength. Therefore, any complex UV or near-infrared absorption may not give accurate results. The device can operate in its linear range only. Hence, any analyte contaminated in excess or below this range cannot provide a reliable output.

Results

This manuscript discusses the protocol for designing and developing an optical multiplexed POCT kit, which detects Fe, Cr, As, and F in various samples. Figure 1 shows the algorithms of the whole procedure. The colorimetric chemical sensor is developed based on the outcomes of the TD-DFT simulation. This chemical sensor is integrated with an indigenously developed optical sensor. The final device is an E-Eye-Enabled POCT digital sensor. Numerical optimization of the colorimetric chemical sen...

Discussion

The validation of the E-Eye-enabled POCT kit against the gold standard (UV-vis) is shown in Figure 5. For this purpose, spiked samples of Fe²⁺ were prepared. Its concentration was measured in triplicate with the developed E-Eye POCT kit and the UV-vis spectrophotometer, and the average value has been reported. Both readings were compared, and it was observed that the device accords with the gold standard. The results prove that the indigenously developed E-Eye-Enable POCT kit meas...

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
1,10-phenanthroline monohydrate (C12H8N2.H2O)	5144-89-8	Sigma Aldrich, India
1,5-diphenylcarbazide ( C13H14N4O)	140-22-7	Sigma Aldrich, India
Acetone (C3H6O)	67-64-1	Sigma Aldrich, India
Ammonium molybdate tetrahydrate ((NH?)?Mo?O??·4H?O)	12054-85-2	Sigma Aldrich, India
Arduino UNO development board	SKU A000066	Robu.in
Ascorbic acid (C6H8O6)	50-81-7	Himedia, India
Breadboard	SKU: 24441	Robu.in
Ferric chloride (FeCL3)	7705-08-0	Himedia, India
Ferrous sulphate heptahydrate (FeSO4.7H2O)	7782-63-0	Himedia, India
Hydrochloric acid (HCl)	7647-01-0	Finar, India
Hydrogen peroxide (H2O2)	7722-84-1	Finar, India
Jumper wires	SKU: 44281	Robu.in
Light emitting diode (LED)	SKU: R110553	Robu.in
Light-dependent resistor (LDR)	SKU: 1496181	Robu.in
Liquid crystal display (LCD)	SKU: 11497	Robu.in
Milli-Q water (18.2  MΩ)
Multimeter	SKU: 1720374	Robu.in
Nitric acid (HNO3)	7697-37-2	Finar, India
Phosphate Buffer Saline (PBS)	SKU: M1866	Himedia, India
Potassium dichromate (K2Cr2O7)	7778-50-9	Himedia, India
Potentiometer	SKU: 790452	Robu.in
Resistors	SKU: R225238	Robu.in
Sodium arsenate heptahydrate (Na2HAsO4·7H2O)	10048-95-0	Himedia, India
Sodium dihydrogen phosphate (NaH2PO4)	13472-35-0	Himedia, India
Sodium fluoride (NaF)	7681-49-4	Himedia, India
Sodium thiocyanate (NaSCN)	540-72-7	Himedia, India
Sulfuric acid (H2SO4)	7664-93-9	Finar, India
Trisodium citrate dihydrate (Na3C6H5O7.2H2O)	03-04-6132	Himedia, India