Sensors & Actuators: B. Chemical 368 (2022) 132180 Contents lists available at ScienceDirect Sensors and Actuators: B. Chemical journal homepage: www.elsevier.com/locate/snb Rapid warning microanalyzer for heavy metals monitoring in natural waters Alex Pascual-Esco , Julián Alonso-Chamarro , Mar Puyol * Group of Sensors and Biosensors, Department of Chemistry, Faculty of Sciences, Universitat Autònoma de Barcelona, Carrer dels Til⋅lers s/n, Bellaterra, 08193 Cerdanyola del Vallès, Spain A R T I C L E I N F O A B S T R A C T Keywords: A warning microanalyzer for the rapid monitoring of different heavy metals in water using Carbon Dots (CDs) as Carbon Dots selective optical reagents is presented. The synthesized CDs have different surface functionalization and exhibit Heavy metals selective fluorescence quenching by heavy metal ions, that combined with the use of microfluidics, provide Fluorescence quenching sensitivity, ease of automation and reproducibility to the method. Moreover, they present maximum excitation Microfluidics wavelengths around 350 nm, allowing multiparametric analysis with a single light source. Although quantum yields range from 16 % to 78 % depending on the type of CDs, enough sensitivity is achieved for each heavy metal using the same measurement conditions of the optical detection system (lock-in modulating frequency, signal amplitude and measurement frequency). The microanalyzer is composed of a Cyclic Olefin Copolymer (COC) analytical microsystem, a flow management system, and a miniaturized customized optical detection system. In this paper, we demonstrate that our proposed system can be used as a toxicity control system by selectively measuring five different heavy metal ions (Co2+, Cu2+, Hg2+, Ni2+, and Pb2+) with detection limits ranging from 2 to 12 ppb. Spiked tap water samples were analyzed, giving recoveries from 98 % to 134 %. Polluted samples containing four of the five heavy metal ions studied (Co2+, Cu2+, Ni2+, and Pb2+) were also analyzed with no significant differences observed between both methods, the proposed microanalyzer and the reference method (ICP-OES). 1. Introduction new instrumentation capable of procuring analytical information rapidly and with high sensitivity. The miniaturization of analytical Contamination of water by heavy metals, coming from human ac­ procedures brings some important advantages for environmental tivities like mining or industrial wastes, has been a critical environ­ continuous monitoring like enhanced portability and reduced reagents mental concern for decades. One of the main preoccupations is their consumption and waste generation [7,8]. In this sense, microfluidic toxicity caused by the non-biodegradability and potential of bio­ devices are broadly used because of the increase in surface area to vol­ accumulation in different tissues and organs. The commonly employed ume ratio that improves process-control, reproducibility, allows the techniques for heavy metals detection (mainly atomic absorption spec­ integration of different processes and yields faster reaction times [5, trometry (AAS) and inductively coupled plasma mass spectroscopy (ICP- 9–12]. In the area of environmental pollution control and early warning MS)) require complex and expensive instrumentation and are not suit­ systems, the combination of microelectronics and microfluidics with able for on-site analysis [1–5]. Currently, it is very difficult to analyze optical sensors is gaining importance, because of the simplicity of the simultaneously several heavy metal ions in water with low detection integration [12,13]. limits without using these techniques and it is still a challenge to develop Photoluminescent nanoparticles have attracted a lot of interest in a sensor for on-site and real-time monitoring of various heavy metal ions recent years due to their potential applications in optical sensing [14, with high selectivity and sensitivity [6]. There is consequently great 15], as well as in medical and biological fields [16,17]. The use of these demand for regular water quality monitoring to identify and assess nanoparticles as analytical reagents enhances sensitivity, improves heavy metal pollution in water [5]. detection limits, and improves the stability of optical reagents [18]. The tendency towards on-site analysis requires new approaches and Carbon Dots (CDs) show unique characteristics as optical reagents * Corresponding author. E-mail address: mariadelmar.puyol@uab.cat (M. Puyol). https://doi.org/10.1016/j.snb.2022.132180 Received 15 March 2022; Received in revised form 6 May 2022; Accepted 6 June 2022 Available online 8 June 2022 0925-4005/© 2022 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by- nc-nd/4.0/). A. Pascual-Esco et al. Sensors and Actuators: B. Chemical 368 (2022) 132180 because of their excellent optical properties (tuneable fluorescence oven for the microwave-assisted processes. emission properties, high quantum yields, and photochemical stability) Absorption spectra were acquired with a Shimadzu UV-3101PC UV- [19–21] and other outstanding properties like their solubility in water, Vis-NIR double beam spectrophotometer (Kyoto, Japan) and fluores­ chemical inertness [22], ease of functionalization, and good biocom­ cence spectra were registered with a Horiba Jobin Yvon Fluorolog patibility [23]. Additionally, CDs’ low toxicity highlights them as an FL3–11 spectrofluorometer (Longjumeau, France). Dynamic Light alternative to other luminescent nanoparticles as quantum dots [24,25]. Scattering (DLS) measurements were performed in a Microtrac Nanotrac Regarding the aim of this work, N-modified CDs have also been Flex nanoparticle size analyzer (Krefeld, Germany). High-Resolution demonstrated to provide sensitivity to heavy metals, and an easy syn­ Transmission Electron Microscopy (HR-TEM) images were taken on a thesis based on cheap precursors [26,27]. high-resolution JEOL JEM-2011 Microscope (Tokyo, Japan). In this sense, we propose an automated microanalyzer for the determination of heavy metal ions as an on-site warning system based on 2.3. Microfluidic platform and experimental setup the fluorescence quenching effect of these ions on the emission of CDs. For the multiparametric determination and as demonstrators of heavy The fabrication process of the microfluidic platform starts with the metal pollution, we selected five different types of CDs from the bibli­ design of the different layers with Computer-Aided Design (CAD) soft­ ography with proven selectivity to Co2+, Cu2+, Hg2+, Ni2+, and Pb2+. ware. Every COC layer substrate was machined on an LPKF Laser and Hydrothermal and microwave-assisted bottom-up methods [28–32] Electronics Protomat S63 Computer Numerical Control (CNC) micro­ were performed to synthesize five types of CDs as reagent demonstrators milling machine (Garbsen, Germany). The different layers were ther­ of heavy metals pollution. The microanalyzer is composed of a micro­ molaminated using a Talleres Francisco Camps hydraulic press fluidic platform and a customized miniaturized optical detection system. (Granollers, Spain). The microfluidic platform is formed by a structural The platform is fabricated in cyclic olefin copolymer (COC), which is layer of TOPAS 5013, which was previously laminated with two films of ideal for optical detection methods because of its high transparency in lower glass transition temperature (TOPAS 8007) acting as sealing near-ultraviolet and visible regions compared to other polymers [9,33] layers [40], and two outer layers of TOPAS 5013. Once overlapped and and suited to multi-layered microfabrication techniques. Signal acqui­ aligned, COC tapes were laminated to obtain the final device. As it can sition is computer controlled and the analytical method is based on a be seen in Fig. 1A, the microfluidic platform (30 mm wide, 50 mm continuous flow strategy, specifically reverse Flow Injection Analysis height, and 2 mm depth) has two inlets, a two-dimensional meander (rFIA). The sample is continuously flowing as a carrier, into which a micromixer (0.8 mm wide and 1 mm depth), an optical flow cell (4.5 mm specific volume of the different CDs is periodically injected [34]. In this diameter and 1 mm depth), and an outlet. way, a transient signal is obtained with maximum emission when the The experimental setup is based on a rFIA, where CDs are sequen­ sample does not contain heavy metals, and fluorescent quenched signal tially injected into a blank (buffer solution) and samples (standard so­ when heavy metals are present in the sample [35–38]. Fluorescence lutions, spiked tap water and soil extracts). The microfluidic set-up (measured as the peak height) is related to the analyte concentration allows the pre-buffering of the samples by an in-line T connector mixer following the Stern-Volmer equation. Finally, FIA systems procure (sample/buffer solution in 1:1 proportion). For continuous flow injec­ reproducibility and automatic baseline recovery [39]. tion, the microfluidic platform was connected with 0.8 mm internal Chemical and hydrodynamic parameters of the developed micro­ diameter Teflon tubing (Tecnyfluor, Barcelona, Spain), to a Gilson analyzer have been optimized to achieve the lowest limit of detection of Minipuls 2 peristaltic pump (Middleton, WI, USA) fitted with 1.14 mm all the heavy metals and the highest repeatability by analyzing standard internal diameter Tygon tubing (Ismatec, Wertheim, Germany). Con­ solutions. Finally, its applicability has been demonstrated by measuring nections were secured with O-rings FPM75 (Epidor, Barcelona, Spain). A heavy metals in spiked tap water and extracts of polluted soil samples Hamilton MVP TMI-6116 six-port injection valve (Bonaduz, from areas near a lead-zinc metal smelter, increasing confidence in the Switzerland) was used to inject the CDs. accuracy and validity of the results. The microfluidic platform is inserted in a customized miniaturized optical detection system (Fig. 1B) by using an Optical Lock and Key 2. Material and methods Reader approach [8]. It consists of a LED emitting at 365 nm (OSA Opto Light GmbH EOLD-365–525), a Thorlabs 460/60 25 mm band-pass filter 2.1. Chemicals and materials (Munich, Germany), and a PIN photodetector (Hamamatsu S1337–66BR) with 33 mm2 of active area mounted and integrated on a Acrylic acid, citric acid, ethylenediamine, glycerol, imidazole-2- printed circuit board (PCB). The optical detection system is modulated carboxylic acid, quinine sulphate, sodium citrate, urea, the metal salts mercury (II) nitrate, nickel (II) nitrate, cobalt (II) nitrate, lead (II) ni­ trate, copper nitrate, zinc nitrate, cadmium nitrate, sodium nitrate, calcium nitrate, magnesium nitrate, potassium chloride and potassium carbonate were purchased from Sigma-Aldrich (Barcelona, Spain). Pol­ yethyleneimine, branched, was purchased from Fisher Scientific (Madrid, Spain). All the CDs synthesized were purified by dialysis against MilliQ water using a Pur-A-Lyzer Mega 1000 dialysis kit with a cut-off of 1 kDa. Dispersions for the CDs syntheses and citric/citrate buffer were prepared in MilliQ water. All solutions for the microanalyzer charac­ terization and the metal ion sensing were buffered at pH 4 with a 0.1 M citric/citrate buffer. Cyclic Olefin Copolymer (COC) sheets were supplied from TOPAS Advanced Polymers GmbH (Florence, KY, USA). 2.2. Apparatus Fig. 1. (A) Microfluidic platform fabricated for the heavy metal ions detection. A Labolan IDL.AI36 oven (Navarra, Spain) was used for the hydro­ (B) Customized miniaturized optical detection system with the microfluidic thermal synthesis processes and a Bifinett KH 1106 domestic microwave platform inserted. 2 A. Pascual-Esco et al. Sensors and Actuators: B. Chemical 368 (2022) 132180 with a digital lock-in amplifier to avoid light interferences and the signal 2.7. Analytical quality parameters of the microanalyzer is obtained with a NI USB-6211 Data Acquisition Card (National In­ struments, Austin, TX, USA). Selectivity was tested in the microfluidic system using several stan­ dard solutions with a 0.5 ppm concentration of the target heavy metal (Co2+, Cu2+, Hg2+, Ni2+, and Pb2+ in each case). Every standard solution 2.4. CDs synthesis and optical characterization also contained the rest of the heavy metal ions tested at a concentration of 10 ppm and the other potentially interfering ions tested at a con­ Five different CDs were synthesized following published synthesis centration of 1000 ppm. procedures: CDs synthesized with citric acid and ethylenediamine (ED Limit of detection was evaluated from the calibration curves ob­ CDs) [28], CDs synthesized with citric acid and polyethyleneimine (PEI tained by injecting per triplicate each type of CDs in standard solutions CDs) [29], CDs synthesized with glycerol and ethylenediamine (EA CDs) containing increasing concentrations of the five heavy metals studied [30], CDs synthesized with citric acid, urea, and imidazole (UREA CDs) (0.01, 0.05, 0.1, 0.5, and 1 ppm) and the other interfering ions. Fluo­ [31], CDs synthesized with acrylic acid and ethylenediamine (ACR CDs) rescence quenching of the CDs was studied through the Stern-Volmer [32]. All batches were dialyzed against MilliQ water using a dialysis equation [42]. membrane with a cut-off of 1 kDa for their purification and kept in a Repeatability of the analytical system was evaluated by performing refrigerator. Since the concentration of the CDs is unknown, dilution ten injections of each CD into a standard solution containing 0.5 ppm of factors in citric/citrate buffer are used as an approximate measurement. each target heavy metal and the other studied ions. Absorption spectra and photoluminescence spectra (excitation and emission) of CDs’ water dispersions were obtained. Quantum Yields (QYs) were calculated with the optically dilute measurement method 2.8. Samples analysis [41] using quinine sulphate in 0.1 M sulphuric acid as a standard. To assess the applicability of this method, spiked tap water samples and extracts of polluted soil samples were tested. Tap water samples 2.5. Optimization of the chemical and hydrodynamic parameters of the were spiked to obtain concentrations of 0.03, 0.06, 0.1, 0.4, and 0.8 ppm microanalyzer of the target heavy metals. They were also spiked with the rest of the heavy metal ions tested at a concentration of 10 ppm and the other Some chemical and hydrodynamic parameters were optimized using potentially interfering ions tested at a concentration of 1000 ppm. a univariate optimization procedure taking as compromise signal to Nitric acid extracts of polluted soil samples coming from the neigh­ noise ratio and sensitivity for all the five target heavy metals. bourhood of a smelter plant (located in Torreon, Mexico) and containing Citric/citrate buffer solutions prepared at four different pH values (3, high concentrations of four of the five heavy metals studied (Hg2+ is not 4, 5, and 6) and four different concentration levels (0.001, 0.01, 0.1 and present), were directly supplied, and analyzed by the reference method 1 M) were tested to obtain the maximum fluorescence signal and taking ICP-OES and the proposed microanalyzer, for comparison purposes. into account the solubility of the different heavy metal ions. 0.5 g of the samples were digested with 10 mL of nitric acid in a Mile­ Regarding hydrodynamic parameters, the flow rate of the sample stone Ultrawave microwave digestion system at 175ºC for 15 min. (carrier) was tested from 0.5 to 2 mL/min and CDs injection volume Samples were filtered with a 0.45 µm filter, diluted to 50 mL with MilliQ from 0.1 to 1 mL. Moreover, some operational parameters of the optical water, and finally analyzed with a Perkin-Elmer Optima 4300DV detection system like signal amplification and integration time were spectrometer. optimized. In this case, CDs were injected into buffer solution and Calibration curves were performed with standard solutions at higher 0.01 ppm heavy metal standard solutions to evaluate the signal to noise concentrations (different concentration ranges depending on the ana­ ratio and detection limit. lyte). As stated, to avoid pH changes of these acid extracts, samples were Finally, different dilution factors of the CDs dispersions were also automatically buffered in the system before the injection of the CDs. tested. For the five types of CDs, 10, 50, 100, 500 and 1000 dilutions in Moreover, an exhaustive cleaning of the microsystem consisting of buffer solution, were prepared. flowing buffer solution for 2 min was necessary between samples. 2.6. Characterization of CDs selectivity 3. Results and discussion Before performing photoluminescence studies of the synthesized 3.1. CDs characterization CDs, a dilution was done to obtain an absorption of 0.05 a.u. and avoid possible self-absorption effects. As reproducibility and characteristics of the synthesized CDs will Selectivity of the different CDs was firstly tested in batch against impact directly on the analytical results and reproducibility of the different heavy metal ion solutions (Cd2+, Co2+, Cu2+, Hg2+, Ni2+, Pb2+, analytical method, CDs were fully characterized. Regarding optical Zn2+), and other potentially interfering ions present in water (Na+, properties, they show similar absorbance and photoluminescence Mg2+, Ca2+, Cl-, CO2-3 ). Emission spectra of 0.5 mL of the corresponding spectra, with a wide absorption band around 350 nm and a wide emis­ CD dispersion at the optimized dilution factor showing an absorption of sion band around 450 nm, obtaining Stokes’ shifts of approximately 0.05 a.u. with the addition of 2.5 mL of citric/citrate buffer at pH 4 were 100 nm in all cases, as it can be seen in Fig. 2. Therefore, all of them compared with the same CDs dispersion with the addition of 2.5 mL of a could be used as optical reagents with the same miniaturized optical set- buffered solution containing a specific ion at a concentration of 10 ppm up. (in the case of the heavy metal ions) or at a concentration of 1000 ppm QYs were calculated, obtaining values of 78 % for ED CDs, 32 % for (for other potentially interfering ions) after the dilution in the cuvette. PEI CDs, 16 % for EA CDs, 44 % for UREA CDs, and 25 % for ACR CDs. 3 A. Pascual-Esco et al. Sensors and Actuators: B. Chemical 368 (2022) 132180 Fig. 2. : Excitation and emission spectra of the five synthesized CDs. Size and shape were obtained with HR-TEM images and DLS measure­ As an example, Fig. 3A shows optimization of the buffer pH. The ments. Results showed spherical shape nanoparticles with sizes ranging optical properties of CDs are sensitive to pH due to the ionizable func­ from 2 to 4 nm. Although high-quality contrast images are difficult to tional groups on the surface. Therefore, the analytical system must obtain with CDs with HR-TEM, we observed two populations of CDs, one operate at a constant pH. In general, the higher the pH, the higher the with a crystalline structure, where the diffraction planes could be fluorescence of the CDs. However, the selected pH must ensure to have distinguished and the other amorphous one. heavy metal species free in solution. As it can be seen, the highest fluorescence intensity is obtained at pH 4 for all the different CDs. On the 3.2. CDs selectivity tests in batch other hand, the effect of the flow rate on the measured signal can be seen in Fig. 3B with the injection of 0.5 mL of a dilution of 100 of ED CDs into First, the quenching effect of different heavy metals, and other a 0.1 M citric/citrate buffer at pH 4 as an example. A faster flow rate potentially interfering ions present in water was investigated in batch. A implies a more rapid analysis and therefore, higher sample throughput, great number of publications demonstrate that different types of N- which are desirable. However, slightly less intense peaks are obtained modified CDs present very strong interactions towards different heavy because the sample is less time in contact with the CDs. The lower is the metal ions through the surface functional groups (-COOH, -OH, -CONH-) flow rate, the wider and more intense are the obtained peaks until [28–32]. According to the bibliography and as shown in the supple­ achieving the steady-state signal. Therefore, a 1.5 mL/min sample flow mentary information (Fig. S2A-E), each CD dispersion was selective to a rate was chosen as the optimal one for subsequent experiments. At this specific heavy metal. ED CDs show good selectivity to Hg2+, PEI CDs to flow rate, response time is calculated in 50 s Cu2+, EA CDs to Pb2+, UREA CDs to Ni2+, and ACR CDs to Co2+. The quenching effect of all the other tested ions was not significant. In these 3.4. Analytical quality parameters of the microanalyzer experiments, it could also be noticed the different sensitivity of the synthesized CDs to the target heavy metal, which would affect the Analytical quality parameters were calculated from calibration optimization of the measurement conditions of the optical detection curves obtained performing the separate analysis of each heavy metal at system of the microanalyzer for multiparametric measurements. the optimized operational values. Fig. 4A show the transient response of the microsystem by injecting the five optimized CDs dispersions into 3.3. Optimization of the microanalyzer buffer solution and into 0.5 ppm standard solutions of the target heavy metal (Co2+, Cu2+, Hg2+, Ni2+, and Pb2+) which also contained the rest Different chemical, hydrodynamic and optical measuring variables of the heavy metal ions tested at a concentration of 10 ppm and the of the microanalyzer were optimized to make possible the multi­ other potentially interfering ions tested at a concentration of 1000 ppm. parametric detection of heavy metals in water, as a compromise between Results obtained confirm the selectivity of the different CDs tested in the best S/N, limits of detection, response time and reagents saving. batch. The fluorescence signal quickly returned to the baseline after Tested values intervals and the final optimal selected values are shown each signal peak and there was not any significant signal drift of the in Table 1. baseline. This demonstrates the robustness of the microanalyzer for the proposed application. Table 1 Calibration curves showed a good linear correlation between 0.01 Optimization of chemical, operational and hydrodynamic parameters. and 1 ppm, as can be seen in Fig. 4B. And even though the Stern-Volmer Parameter Tested interval Optimal value model has some limitations that affect the accuracy, caused by de­ Buffer concentration (M) 0.001 – 1.000 0.100 viations at high concentrations [43,44], concentration ranges as high as Buffer pH 3–6 4 50–600 ppm were tested for real polluted samples, showing also good Injection volume (mL) 0.1 – 1.0 0.5 linear correlation but lower sensitivity. Carrier flow rate (mL/min) 0.5 – 2.0 1.5 Detection limits, LODs (Table 2) were calculated as 3 SD/KSV (where Signal amplification 1 – 100 10 SD is the standard deviation of the blank and KSV (Stern-Volmer con­ Integration time (s) 0.1 – 1.0 0.1 ED CDs dilution factor 10–1000 100 stant) is the slope of the linear fit near to unity). PEI CDs dilution factor 10–1000 50 Repeatability was calculated as relative standard deviation (RSD) EA CDs dilution factor 10–1000 10 obtained with 10 repeated injections of the same CDs into the 0.5 ppm UREA CDs dilution factor 10–1000 50 heavy metal standard solutions. Data is detailed in Table 2, showing ACR CDs dilution factor 10–1000 10 values under 1.2%. 4 A. Pascual-Esco et al. Sensors and Actuators: B. Chemical 368 (2022) 132180 Fig. 3. (A) Normalized fluorescence intensity obtained by injecting the five CDs at their optimal dilution factor into 0.1 M citric/citrate buffer solutions of different pH values. (B) Effect of flow rate into the obtained signal for an injection of 0.5 mL of ED CDs (100 dilution factor) into buffer solution. Fig. 4. (A) Transient response of the microsystem by injecting the CDs into buffer solution (as intensity value with no quencher) and to different standard solutions of heavy metals containing potential interferents (Zn2+, Cd2+, Na+, Mg2+, Ca2+, Cl-, CO2- 3 ). (B) Stern-Volmer representation of quenching effect of the target heavy metals on the fluorescence of the five types of CDs. industry were analyzed. Nitric acid extracts were prepared and analyzed Table 2 by a reference method (ICP-OES) and the proposed microanalyzer. Fig. 5 Results summarizing the analytical performance of the proposed microanalyzer shows, as an example, the calibration curve for Co2+ at a higher con­ for the monitoring of heavy metals in water samples. centration range (10–70 ppm) where it can be noticed from the slope of Type of CDs Selectivity KSV (ppm− 1) Detection Limit (ppb) RSD (%) the calibration curve a lower sensitivity. Fig. 6 shows the comparative ED CDs Hg 2+ 0.3800 5.9 ± 0.8 0.5 test for the measurement of Co2+ in 8 different samples. Regression PEI CDs Cu2+ 0.2782 11.7 ± 0.7 1.2 equations of the comparison test (intercept, slope, and correlation co­ EA CDs Pb2+ 0.3293 8.8 ± 0.8 1.1 efficient) for the five heavy metal ions as well as tcalc values obtained UREA CDs Ni2+ 0.2405 2.5 ± 0.3 0.4 ACR CDs Co2+ 0.2883 3.9 ± 0.5 0.9 with the paired t-test are shown in the Supplementary Information (Table S1). From the statistical data treatment, it can be concluded re­ sults obtained with the developed microanalyzer were not significantly 3.5. Samples analysis different from the ones measured with the reference method according to the paired t-test (tcalc < ttab) for the four heavy metals analyzed. Spiked tap water samples were analyzed at the optimized conditions Table 4 shows accuracy expressed as % error for the differences and after performing a calibration experiment for each heavy metal. between results obtained with both methods. As it can be seen, accuracy Table 3 shows recovery rates. In general, an overestimation is noticed is different for each heavy metal and, as it is more affected at lower for the less concentrated samples near the limit of detection. At this concentrations, the determination of Pb2+ is less accurate. However, a concentration level, results are also less precise because the error of the maximum value of 7.5% error is calculated. interpolated signal is higher at lower concentrations. However, as a warning system, results allow determining the estimated value of heavy 4. Conclusions metals as indicators of pollution. To validate the usability of the developed microanalyzer, real In this paper we have developed a new multiparametric micro­ polluted soil samples taken from areas surrounding a metallurgic analyzer as a warning system for heavy metals pollution monitoring. 5 A. Pascual-Esco et al. Sensors and Actuators: B. Chemical 368 (2022) 132180 Table 3 Determination of the heavy metals studied in spiked tap water samples. Analyte recovery rates Sample Concentration of the analyte in the sample (ppm) Hg2þ Cu2þ Pb2þ Ni2þ Co2þ 0.03 129 % 131 % 134 % 123 % 120 % Spiked 0.06 116 % 112 % 120 % 110 % 111 % tap 0.1 105 % 103 % 105 % 104 % 105 % water 0.4 101 % 100 % 102 % 100 % 100 % 0.8 99.1 % 99.0 % 101 % 98.8 % 98.2 % Thanks to the reduced dimensions of the fluidic system, it consumes the analytes to determine, as far as other selective CDs (with absorption small volumes of reagents and has a fast analysis time. Although sample bands around 350 nm) could be synthesized. In most of the cases, LODs treatment (buffering) must be performed, which for acid extracts was comply with the regulation [5] concerning heavy metal pollution in crucial, this is performed by the same flow management systems before water (between 2 and 12 ppb) but in the case of Hg2+, the established CDs are injected. Performance of the microanalyzer, tested with stan­ LOD is still lower. Some future work must be conducted in order to dard solutions, spiked tap water and real samples, combined with the finally validate the real applicability of the proposed methodology for stability of the luminescent reagents (N-modified CDs) demonstrates its heavy metals monitoring, such as long-term studies. Despite current potential application for long-term and continuous monitoring of heavy maintenance criteria of quality control stations is approximately every metals in water. Five different heavy metals (Co2+, Cu2+, Hg2+, Ni2+ and 15 days and no critical issues are expected regarding the stability of CDs, Pb2+) have been selectively detected. However, it is possible to expand this should be assessed. Taking into account results for repeatability, selectivity, recoveries of spiked tap water samples and accuracy of real polluted samples, we demonstrate its applicability as an on-site warning system of heavy metals contamination. Although the presented set-up requires bulky components for flow management in continuous mode, such as a peristaltic pump, the equipment can be further miniaturized, and it is easily automatable. We are currently working on the development of an automatic and computer-controlled microanalyzer based on the use of solenoid peri­ staltic micro-pumps and three-way solenoid valves, enabling automa­ tion of the whole analytical procedure. CRediT authorship contribution statement Alex Pascual-Esco: Methodology Development, Investigation, Data analysis, Statistical analysis, Validation and Writing – original draft. Julián Alonso-Chamarro: Resources, Project administration, Funding acquisition, and Writing – review & editing. Mar Puyol: Research Conceptualization, Resources, Supervision, Project administration, Fig. 5. : Stern-Volmer representation of quenching effect of Co2+ on the fluo­ Table 4 rescence of ACR CDs in a higher concentration range (10–70 ppm). Concentration of the heavy metals studied in polluted soil samples determined by the reference method (ICP-OES) and with the proposed method, and accuracy expressed as (% error). [Cu2þ] [Pb2þ] [Cu2þ] [Pb2þ] Sample proposed % proposed % ICP-OES ICP-OES number method error method error (ppm) (ppm) (ppm) (ppm) 1 576.2 587.0 1.9 11.5 12.0 4.8 2 255.7 251.1 1.8 4.8 4.9 2.1 3 511.7 507.3 0.9 5.3 5.2 1.7 4 119.7 116.2 2.9 3.7 3.9 7.4 5 231.9 236.5 2.0 4.0 4.0 1.0 6 458.3 477.9 4.3 7.1 7.0 1.3 7 90.5 92.2 1.9 0.8 0.8 4.0 8 95.7 92.4 3.4 2.2 2.3 7.5 [Ni2þ] [Co2þ] [Ni2þ] [Co2þ] Sample proposed % proposed % ICP-OES ICP-OES number method error method error (ppm) (ppm) (ppm) (ppm) 1 42.3 42.8 1.0 49.4 49.1 0.5 2 31.8 31.6 0.7 34.2 36.0 5.1 3 61.3 61.9 1.0 61.3 63.9 4.3 4 35.8 35.9 0.3 24.8 26.0 4.8 5 57.9 58.7 1.4 39.1 37.9 3.1 Fig. 6. : Comparative study between the results for the analysis of Co2+ in 6 33.4 32.9 1.4 43.3 41.4 4.4 7 25.7 26.4 3.1 13.6 14.3 4.5 polluted soil samples (n = 8) obtained by the developed microanalyzer and the 8 33.9 35.4 4.4 19.9 21.0 5.7 reference method (ICP-OES). 6 A. Pascual-Esco et al. Sensors and Actuators: B. Chemical 368 (2022) 132180 Funding acquisition, and Writing – original draft. carbon quantum dots, Sens. Actuators B Chem. 279 (2019) 305–312, https://doi. org/10.1016/j.snb.2018.10.026. [16] M.L.C. Passos, P.C.A.G. Pinto, J.L.M. Santos, M.L.M.F.S. Saraiva, A.R.T.S. Araujo, Declaration of Competing Interest Nanoparticle-based assays in automated flow systems: a review, Anal. Chim. Acta 889 (2015) 22–34, https://doi.org/10.1016/j.aca.2015.05.052. [17] Y. Liu, F. Seidi, C. Deng, R. Li, T. Xu, H. Xiao, Porphyrin derived dual-emissive The authors declare that they have no known competing financial carbon quantum dots: customizable synthesis and application for intracellular Cu2 interests or personal relationships that could have appeared to influence + quantification, Sensors Actuators B Chem. 343 (2021), 130072, https://doi.org/ the work reported in this paper. 10.1016/j.snb.2021.130072. [18] T. Chen, S. Yin, J. Wu, Nanomaterials meet microfluidics: improved analytical methods and high-throughput synthetic approaches, Trends Anal. Chem. 142 Acknowledgements (116309) (2021) 1–19, https://doi.org/10.1016/j.trac.2021.116309. [19] K.J. Mintz, Y. Zhou, R.M. Leblanc, Recent development of carbon quantum dots regarding their optical properties, photoluminescence mechanism, and core We gratefully thank the Instituto Tecnológico de la Laguna (Torreón, structure, Nanoscale 11 (2019) 4634–4652, https://doi.org/10.1039/c8nr10059d. México) for soil samples supply and Dr. Eva Arasa for the soil extracts [20] A. Simpson, R.R. Pandey, C.C. Chusuei, K. Ghosh, R. Patel, A.K. Wanekaya, preparation. This work has been financially supported by the Spanish Fabrication characterization and potential applications of carbon nanoparticles in the detection of heavy metal ions in aqueous media, Carbon 127 (2018) 122–130, Ministry of Science and Innovation (MICINN) through project CTQ2017- https://doi.org/10.1016/j.carbon.2017.10.086. 85011-R and project PID2020-117216RB-I00, co-funded by FEDER [21] R. Jelinek, Carbon Quantum Dots: Synthesis, Properties and Applications, (Fondo Europeo de Desarrollo Regional), and the Catalonia Government electronic ed, Springer, Cham (Switzerland), 2017. [22] Y. Wang, A. Hu, Carbon quantum dots: synthesis, properties and applications, through 2017SGR-220. The support of Met-Mex Peñoles for the devel­ J. Mater. Chem. C. 2 (2014) 6921–6939, https://doi.org/10.1039/c4tc00988f. opment of this work was also gratefully acknowledged. [23] S. Anwar, H. Ding, M. Xu, X. Hu, Z. Li, J. Wang, L. Liu, L. Jiang, D. Wang, C. Dong, M. Yan, Q. Wang, H. Bi, Recent advances in synthesis, optical properties, and biomedical applications of carbon dots, ACS Appl. Bio Mater. 2 (6) (2019) Appendix A. Supporting information 2317–2338, https://doi.org/10.1021/acsabm.9b00112. [24] M.L. Liu, B.B. Chen, C.M. Li, C.Z. Huang, Carbon dots: synthesis, formation Supplementary data associated with this article can be found in the mechanism, fluorescence origin and sensing applications, Green. Chem. 21 (2019) 449–471, https://doi.org/10.1039/c8gc02736f. online version at doi:10.1016/j.snb.2022.132180. [25] J.H. Liu, Y. Wang, G.H. Yan, F. Yang, H. Gao, Y. Huang, H. Wang, P. Wang, L. Yang, Y. Tang, L.R. Teisl, Y.P. Sun, Systematic toxicity evaluations of high-performance References carbon “Quantum” dots, J. Nanosci. Nanotechnol. 19 (4) (2019) 2130–2137, https://doi.org/10.1166/jnn.2019.15807. [26] X. Sun, Y. Lei, Fluorescent carbon dots and their sensing applications, Trends Anal. [1] Z. Wang, C. Xu, Y. Lu, X. Chen, H. Yuan, G. Wei, G. Ye, J. Chen, Fluorescence sensor Chem. 89 (2017) 163–180, https://doi.org/10.1016/j.trac.2017.02.001. array based on amino acid derived carbon dots for pattern-based detection of toxic [27] P. Li, S.F.Y. Li, Recent advances in fluorescence probes based on carbon dots for metal ions, Sens. Actuators B Chem. 241 (2017) 1324–1330, https://doi.org/ sensing and speciation of heavy metals, Nanophotonics 10 (2) (2020) 877–908, 10.1016/j.snb.2016.09.186. https://doi.org/10.1515/nanoph-2020-0507. [2] C.Y. Wang, C.C. Hsu, Online, continuous, and interference-free monitoring of trace [28] F. Du, F. Zeng, Y. Ming, S. Wu, Carbon dots-based fluorescent probes for sensitive heavy metals in water using plasma spectroscopy driven by actively modulated and selective detection of iodide, Microchim. Acta 180 (2013) 453–460, https:// pulsed power, Environ. Sci. Technol. 53 (2019) 10888–10896, https://doi.org/ doi.org/10.1007/s00604-013-0954-2. 10.1021/acs.est.9b02970. [29] A. Salinas-Castillo, M. Ariza-Avidad, C. Pritz, M. Camprubí-Robles, B. Fernández, [3] N. Ullah, M. Mansha, I. Khan, A. Qurashi, Nanomaterial-based optical chemical M.J. Ruedas-Rama, A. Megia-Fernández, A. Lapresta-Fernández, F. Santoyo- sensors for the detection of heavy metals in water: recent advances and challenges, Gonzalez, A. Schrott-Fischer, L.F. Capitan-Vallvey, Carbon dots for copper Trends Anal. Chem. 100 (2018) 155–166, https://doi.org/10.1016/j. detection with down and upconversion fluorescent properties as excitation sources, trac.2018.01.002. Chem. Commun. 49 (2013) 1103–1105, https://doi.org/10.1039/c2cc36450f. [4] J.P. Devadhasan, J. Kim, A chemically functionalized paper-based microfluidic [30] Y. Jiang, Y. Wang, F. Meng, B. Wang, Y. Cheng, C. Zhu, N-doped carbon dots platform for multiplex heavy metal detection, Sens. Actuators B Chem. 273 (2018) synthesized by rapid microwave irradiation as highly fluorescent probes for Pb2+ 18–24, https://doi.org/10.1016/j.snb.2018.06.005. detection, N. J. Chem. 39 (2015) 3357–3360, https://doi.org/10.1039/ [5] A. Lace, J. Cleary, A review of microfluidic detection strategies for heavy metals in c5nj00170f. water, Chemosensors 9 (60) (2021) 1–26, https://doi.org/10.3390/ [31] X. Wang, D. Wang, Y. Guo, C. Yang, A. Iqbal, W. Liu, W. Qin, D. Yan, H. Guo, chemosensors9040060. Imidazole derivative-functionalized carbon dots: using as a fluorescent probe for [6] X. Liu, Y. Wang, Y. Song, Visually multiplexed quantitation of heavy metal ions in detecting water and imaging of live cells, Dalton Trans. 44 (2015) 5547–5554, water using volumetric bar-chart chip, Biosens. Bioelectron. 117 (2018) 644–650, https://doi.org/10.1039/c5dt00128e. https://doi.org/10.1016/j.bios.2018.06.046. [32] N. Jing, M. Tian, Y. Wang, Y. Zhang, Nitrogen-doped carbon dots synthesized from [7] Ö. Biçen Ünlüer, F. Ghorbani-Bidkorbeh, R. Keçili, C. Mustansar Hussain, Future of acrylic acid and ethylenediamine for simple and selective determination of cobalt the modern age of analytical chemistry: Nanominiaturization, in: C. Mustansar ions in aqueous media, J. Lumin. 206 (2019) 169–175, https://doi.org/10.1016/j. Hussain (Ed.), Handbook on Miniaturization in Analytical Chemistry: Application jlumin.2018.10.059. of Nanotechnology, Elsevier, Amsterdam, 2020, pp. 277–296. [33] M. Denz, G. Brehm, C.Y.J. Hémonnot, H. Spears, A. Wittmeier, C. Cassini, [8] O. Ymbern, M. Berenguel-Alonso, A. Calvo-Ló pez, S. Gó mez-de Pedro, D. Izquierdo, O. Saldanha, E. Perego, A. Diaz, M. Burghammerd, S. Köster, Cyclic olefin J. Alonso-Chamarro, Versatile lock and key assembly for optical measurements copolymer as an X-ray compatible material for microfluidic devices, Lab Chip 18 with microfluidic platforms and cartridges, Anal. Chem. 87 (2015) 1503–1508, (2018) 171–178, https://doi.org/10.1039/c7lc00824d. https://doi.org/10.1021/ac504255t. [34] L. Chen, J. Xu, T. Wang, Y. Huang, D. Yuan, Z. Gong, Toward a versatile flow [9] R.K. Jena, C.Y. Yue, Y.C. Lam, Micro fabrication of cyclic olefin copolymer (COC) technique: Development and application of reverse flow dual-injection analysis based microfluidic devices, Microsyst. Technol. 18 (2012) 159–166, https://doi. (rFDIA) for determining dissolved iron redox species and soluble reactive org/10.1007/s00542-011-1366-z. phosphorus in seawater, Talanta 232 (122404) (2021) 1–9, https://doi.org/ [10] N. Yogarajaha, S.S.H. Tsai, Detection of trace arsenic in drinking water: challenges 10.1016/j.talanta.2021.122404. and opportunities for microfluidics, Environ. Sci.: Water Res. Technol. 1 (2015) [35] C. Hu, T.J. Lin, Y.C. Huang, Y.Y. Chen, K.H. Wang, K.Y.A. Lin, Photoluminescence 426–447, https://doi.org/10.1039/c5ew00099h. quenching of thermally treated waste-derived carbon dots for selective metal ion [11] L. Lin, Y. Yin, S.A. Starostin, H. Xu, C. Li, K. Wu, C. He, V. Hessel, Microfluidic sensing, Environ. Res. 197 (111008) (2021) 1–10, https://doi.org/10.1016/j. fabrication of fluorescent nanomaterials: a review, Chem. Eng. J. 425 (131511) envres.2021.111008. (2021) 1–15, https://doi.org/10.1016/j.cej.2021.131511. [36] F. Zu, F. Yan, Z. Bai, J. Xu, Y. Wang, Y. Huang, X. Zhou, The quenching of the [12] H. Gai, Y. Li, E.S. Yeung, Optical detection systems on microfluidic chips, in: B. Lin fluorescence of carbon dots: a review on mechanisms and applications, Microchim. (Ed.), Microfluidics: Technologies and Applications, Springer, Berlin, 2011, Acta 184 (2017) 1899–1914, https://doi.org/10.1007/s00604-017-2318-9. pp. 171–201. [37] H. Lee, Y.C. Su, H.H. Tang, Y.S. Lee, J.Y. Lee, C.C. Hu, T.C. Chiu, One-Pot [13] K.A. Lukyanenko, I.A. Denisov, V.V. Sorokin, A.S. Yakimov, E.N. Esimbekova, P. hydrothermal synthesis of carbon dots as fluorescent probes for the determination I. Belobrov, Handheld enzymatic luminescent biosensor for rapid detection of of mercuric and hypochlorite ions, Nanomaterials 11 (1831) (2021) 1–11, https:// heavy metals in water samples, Chemosensors 7 (2019) 1–10, https://doi.org/ doi.org/10.3390/nano11071831. 10.3390/chemosensors7010016. [38] C. Zhang, S. Wu, Y. Yu, F. Chen, Determination of thiourea based on the reversion [14] M. Berenguel-Alonso, I. Ortiz-Gómez, B. Fernández, P. Couceiro, J. Alonso- of fluorescence quenching of nitrogen doped carbon dots by Hg2+, Spectrochim, Chamarro, L.F. Capitán-Vallvey, A. Salinas-Castillo, M. Puyol, An LTCC monolithic Acta A 227 (117666) (2020) 1–6, https://doi.org/10.1016/j.saa.2019.117666. microreactor for the synthesis of carbon dots with photoluminescence imaging of [39] S. Karada, E.M. Görüsük, E. Çetinkaya, S. Deveci, K.B. Dönmez, E. Uncuoglu, the reaction progress, Sens. Actuators B Chem. 296 (126613) (2019) 1–8, https:// M. Dogu, Development of an automated flow injection analysis system for doi.org/10.1016/j.snb.2019.05.090. determination of phosphate in nutrient solutions, J. Sci. Food Agric. 98 (2018) [15] F. Copur, N. Bekar, E. Zor, S. Alpaydin, H. Bingol, Nanopaper-based 3926–3934, https://doi.org/10.1002/jsfa.8911. photoluminescent enantioselective sensing of L-Lysine by L-Cysteine modified 7 A. Pascual-Esco et al. Sensors and Actuators: B. Chemical 368 (2022) 132180 [40] TOPAS Advanced Polymers. http://www.topas.com, 2021 (Accessed October 28, Julián Alonso-Chamarro received his Ph.D. in Analytical Chemistry in 1987 at the 2021). Autonomous University of Barcelona. He is Professor of Analytical Chemistry at the same [41] G.A. Crosby, J.N. Demas, The measurement of photoluminescence quantum yields. university. He has co-authored more than 140 research papers on (Bio)chemical sensors A review, J. Phys. Chem. 75 (1971) 991–1024, https://doi.org/10.1021/ and automated flow systems and is author of different patents about sensors and analytical j100678a001. instrumentation for environmental monitoring. His research interests include the design [42] Y. Song, S. Zhu, S. Xiang, X. Zhao, J. Zhang, H. Zhang, Y. Fu, B. Yang, Investigation and fabrication of microsystems for chemical analysis (μTAS) based on LTCC, polymer and into the fluorescence quenching behaviours and applications of carbon dots, IC technologies, microfluidic platforms for the synthesis of nanomaterials and optical and Nanoscale 6 (2014) 4676–4682, https://doi.org/10.1039/c4nr00029c. electrochemical microsensors. Most of these developments were applied in the environ­ [43] A. de la Torre, S. Medina-Rodríguez, J.C. Segura, J.F. Fernández-Sánchez, mental and industrial control fields. A polynomial-exponent model for calibrating the frequency response of photoluminescence-based sensors, Sensors 20 (2020) 4635, https://doi.org/ Mar Puyol completed her Ph.D. in Analytical Chemistry in 2002 from the Autonomous 10.3390/s20164635. University of Barcelona. She was working as a researcher for two years in the Applied [44] T. Htun, A negative deviation from stern–volmer equation in fluorescence Physics Department of the Universidad de Zaragoza until 2006. She is author of more than quenching, J. Fluoresc. 14 (2004) 217–222, https://doi.org/10.1023/B: 40 papers related on the field. Her research interests include the study and characteriza­ JOFL.0000016294.96775.fd. tion of new chromo(fluoro)ionophores and its integration as recognition elements in miniaturized optochemical sensors, the development of microsystems for chemical anal­ Alex Pascual-Esco received a degree in Chemistry from Universitat Autònoma de Bar­ ysis (μTAS) based on LTCC, polymer and IC technologies and its application in environ­ celona (UAB) in 2017. In 2018 obtained a Master’s degree in Analytical Chemistry form mental, agri-food, space and biomedical fields, and the design and fabrication of Universitat de Barcelona (UB). Since October 2018, he is a Ph.D. candidate at the Sensors microfluidic platforms for the synthesis of nanomaterials and for the development of (bio) and Biosensors Group under the supervision of Dr. Mar Puyol. His research interests chemical miniaturized analyzers based on the use of nanostructured materials. include the design and fabrication of microsystems for water quality monitoring (μTAS) based on LTCC and polymer technologies, microfluidic platforms for the synthesis of nanomaterials and optical sensors. 8