Sub-ppm quantification of Hg(II) in aqueous media using both the naked eye and digital information from pictures of a colorimetric sensory polymer membrane taken with the digital camera of a conventional mobile phone

Abstract

We present colorimetric sensory membranes for detecting Hg(II) in aqueous media. The colour response of the sensory materials can be tuned for detection with the naked eye, such as the maximum contaminant level of Hg(II) that is set by the United States Environmental Protection Agency (EPA) for drinking water. Furthermore, the concentration of Hg(II) can be monitored using digital pictures of the membranes taken with conventional cameras. Thus, nanomolar concentration of Hg(II) could be detected by the naked eye due to colour changes of membranes, and the concentration of Hg(II) could be quantified, within the millimolar to nanomolar range, by means of analysing the digital information of pictures taken of the membranes after dipping them in water containing this environmentally poisonous cation.

---

Detecting and quantifying chemical species are some of the key challenges in chemical research, particularly in environmental control, where the detection of contaminants by easy, rapid, and inexpensive techniques has always been of utmost importance. Among the different methods developed for this purpose, mass spectrometry, inductively coupled plasma atomic emission spectroscopy, high-performance liquid chromatography, and electrochemistry have been widely utilized.¹ Within this framework, and mimicking nature, the recognition of target molecules by sensory receptor motifs of discrete molecules² and of polymers³ (i.e., the specific interaction of two chemicals) with a concomitant variation in a measurable macroscopic property has received significant attention. This is primarily because the sensing phenomena can be easily used to selectively quantify chemical species with non-sophisticated instruments. Among the target molecules, divalent mercury has been deeply studied because it is very toxic, and its bioaccumulation within the brain and kidneys causes serious problems to human health, ultimately leading to neurological diseases. The contamination of water sources comes from the erosion of natural deposits, discharge from refineries and factories, as well as runoff from landfills and croplands, to name but a few. Regarding this point, fluorogenic and chromogenic chemical probes are useful for the fast detection of mercury contamination. They operate by complexation of the guest mercury cation with the probe, or host, giving rise to a macroscopically detectable colour change or change in fluorescent behaviour.

Herein, we present an advancement in the application of supramolecular chemistry to the quantification of chemicals using pictures of colorimetric chemosensory materials taken with conventional digital cameras to quantify the concentration of Hg(II) in aqueous media.⁴ Moreover, the sensory membranes promote colour changes within the limit of Hg(II) in drinking water (2 ppb, or 1 × 10⁻⁸ M) that are established by the United States Environmental Protection Agency (EPA).⁵ Sensory polymer membranes⁶ and other support materials⁷ have been previously utilized as “dip-in” solid kits in the naked eye detection of chemicals with good results but, to the best of our knowledge, without a clear naked eye quantification of the target molecules.

Dense polymer membranes, with gel behaviour in aqueous media, were prepared as colorimetric and fluorogenic sensory materials. The copolymer structures were designed to have the correct balance of hydrophobic and hydrophilic character in order to have both good mechanical properties, thus making tractable materials, and suitable water swelling. Water uptake permits the diffusion of solvated chemicals into the membrane, where the sensing is achieved through interaction with the sensory motifs. The sensory substructures were provided by a water insoluble comonomer rhodamine derivative (2) shown in Fig. 1.⁸ The rhodamine sensory motif was chosen because rhodamine derivatives have been widely studied as fluorogenic and chromogenic chemosensors for the detection and quantification of mercury species.^9,10 This hydrophobic comonomer was introduced in mole percentages of 1.0, 2.5, and 5.0%, producing membranes Mem1%, Mem2.5%, and Mem5%.

Fig. 1 Chemical structures of the monomers and the copolymers. The copolymer structures and codes are shown over a digital picture of the sensory film Mem2.5%.

The water uptake and the mechanical properties were controlled also by use of a crosslinker. The water swelling percentages (WSP) of the materials were primarily controlled by the hydrophilic comonomer 2-hydroxyethyl acrylate (1) and by the nominal cross-linking ratio corresponding to 1% of the mole content of the crosslinker (3), and to a lesser extent by the mole content of (2). The WSP were 66% for Mem1% and 55% for both Mem2.5% and Mem5%. WSPs higher than 40% and lower than 100% are considered to be optimum for sensory materials for application in aqueous environments. The membranes were highly tractable, having rubber-like behaviour with a Young's modulus of 4 MPa and a tensile strength of 3 MPa.

The conventional behaviour of Rhodamine B (Rho B) derivatives with the same sensory core as (2) in solution, in terms of its fluorescence and UV/Vis spectra, can be described as a pH-dependent equilibrium between two limiting chemical structures: the colourless and non-fluorescent spirolactam (neutral and basic media) and the fluorescent and pink amide (acidic media).^9a In neutral and basic media, the presence of Hg(II) causes the development of a pink colour and fluorescence due to the opening of the lactam to give the amide (see ESI, Scheme S2†). This behaviour was not observed in the solid state in our systems (Mem1%, Mem2.5%, and Mem5%), where the material is coloured and fluorescent both at high and low pH. The difference in colour and fluorescence intensity indicated that this equilibrium was displaced by the acidity of the medium to one of the limit structures, but never completely (Fig. 2). However, contrary to the behaviour of conventional and discrete Rho B sensory derivatives in organic solution, once the materials were swelled both in basic, or neutral, and acidic aqueous media, they showed colour and fluorescence development upon addition of Hg(II) to the medium (See Fig. 3, and ESI – Sections S7–S9†).

Fig. 2 Left: digital pictures of two pieces of Mem2.5% upon swelling in water at pH = 2 (left) and pH = 14 (right). Right: digital picture taken under UV irradiation (365 nm).

Fig. 3 Fluorescence (left, λ_exc = 361 nm) and UV/Vis (right) spectra corresponding to the titration of Mem5% with aqueous Hg(II) from 3.18 × 10⁻³ to 1.58 × 10⁻⁹ M under acidic (top, pH = 2, KCl–HCl) and neutral (bottom, pH = 7, TRIS–HCl) conditions.

The titration patterns of the UV/Vis and fluorescence spectroscopies may be tuned by modifying the content of the sensory motif within the membrane structure (2), as seen in Fig. 4. The increase in the Rho B membrane content, from Mem1% to Mem5%, prompted increased membrane absorbance, but to a lower relative increment upon adding increasing concentrations of aqueous Hg(II). Similarly, a higher sensory motif content did not prompt a notable increase in the relative fluorescence intensity at a given Hg(II) concentration. The response time of the sensory material was similar for all of the samples and was lower than 25 min for the colorimetric and fluorescence titration of a water sample containing a Hg(II) concentration of 1.8 × 10⁻⁵ M (Fig. 4).

Fig. 4 Left: titration of Hg(II) in water at pH = 2 (KCl–HCl) with the three membranes (red: UV/Vis; blue: fluorescence) (▲ = Mem1%; ● = Mem2.5%; ■ = Mem5%). The continuous lines are the linear fitting of the linear range. Right: UV/Vis (red) and fluorescence (blue) response of the membrane Mem1% in water (pH = 2, KCl–HCl) vs. time upon adding Hg(II) (1.8 × 10⁻⁵ M). The limit of detection (LOD) and the limit of quantification (LOQ) obtained with Mem1% using UV/Vis spectroscopy were 7.7 × 10⁻⁸ M and 2.3 × 10⁻⁷ M, respectively.¹¹

The colorimetric response of the sensory membranes to the presence of Hg(II) in aqueous media permitted the detection of this cation with the naked eye in addition to its quantification. Thus, considering that the maximum contaminant level (MCL) set by the EPA for drinking water is near 1 × 10⁻⁸ M (2 ppb),⁵ the membrane strips were dipped in acidic and neutral water (pH = 2 and 7, respectively) containing Hg(II) concentrations ranging from 1 × 10⁻¹⁰ to 1 × 10⁻³ M. Fig. 5 presents a sample of the responses of Mem1% and Mem5% strips (the behaviour of Mem2.5% was fairly similar to Mem5%). It is clear that the concentration of Hg(II) could be correlated with the naked eye to the colour of the membrane. Moreover, the colour response could be tuned with the acidity of the medium and/or with the sensory moiety content of the membrane. For example, Mem1% showed a significant colour change under neutral conditions at a Hg(II) concentration of 1 × 10⁻⁹ M, below the MCL, whereas this change was observed at about 1 × 10⁻⁷ M under acidic conditions. However, increased sensory concentration (i.e., Mem2.5%) prompted colour development that was more difficult to see with the naked eye. A similar behaviour could be observed upon irradiation of the strips with UV radiation (see ESI, Fig. S1†), where the fluorescence of the films permitted the semi-quantification of the concentration of Hg(II).

Fig. 5 Digital pictures of strips of Mem1% and Mem5% after dipping for 25 min into water containing different concentrations of Hg(II). The molar concentration of Hg(II) is written on each membrane strip. The digital pictures were taken with a mobile phone (Samsung Galaxy).

Digital pictures taken of the sensory membranes after dipping in an aqueous solution of Hg(II) permitted quantification of the concentration of Hg(II) using the RGB colour model. The RGB colour model is used for the display of images in electronic systems and is related to the human colour perception. The acronym RGB stands for the three additive primary colours, red, green and blue. The RGB value defining each pixel in a picture is device dependent, and an RGB value does not define the same colour across devices. The opposite phenomenon is also true, with the same colour reproduced with different RGB values by different devices due to variation in the image sensors and colour management. Considering these phenomena, pictures of sensory membranes Mem1%, Mem2.5% and Mem5% dipped for 25 min into both neutral (pH = 7) and acidic (pH = 2) water with different concentrations of Hg(II) were taken with three digital cameras: an old model (Kodak EasyShare DX4330), a digital camera (Sony DSC-W100), and a smart phone camera (Samsung Galaxy). Then, the RGB values of each film strip, ranging from 0 to 255, were measured from the picture taken, averaging 5 pixels from each strip. The RGB parameters of a picture can be easily obtained from commercial or free graphic editing programs. The GNU Image Manipulation Program (GIMP), a free and open source software image retouching and editing tool, was used in this work. A distinct correlation was found between different values of the R, G and B parameters and the Hg(II) concentration, and these values were also correlated with naked eye observation.

To determine the overall information on the three RGB values and their correlation with the Hg(II) concentration, these parameters were subjected to principal components analysis (CPA). A relationship between the parameters was found that permitted the reduction of the three variables to one component (PC1). This component (PC1) could be correlated with the Hg(II) concentration (See Fig. 6, and ESI – Sections S10 and S11†), and this relationship was consistent with the naked eye response pattern described above. Thus, Mem1% under acidic conditions allowed the quantification of 1 × 10⁻⁹ to 1 × 10⁻⁷ M Hg(II) in water using the digital information from pictures of membrane strips. Therefore, these strips essentially provide a yes/no answer to the question of whether a sample exceeds the MCL (i.e., if it is drinkable). Furthermore, pictures taken of Mem1% under neutral conditions enabled the continuous quantification of Hg(II) content from 1 × 10⁻¹⁰ to 1 × 10⁻⁶ M. Increased concentrations of the sensory motifs within the membrane allowed tuning of the response behaviour. Thus, Mem2.5% and Mem5% showed continuous variation of the PC1 parameter over the range of Hg(II) concentrations measured (from 1 × 10⁻¹⁰ to 1 × 10⁻³ M, see Fig. 6 and ESI – Section S5†). The results obtained from analysing the pictures taken with completely different devices were comparable (See ESI, Fig. S5†), thus validating the reported methodology. At this point, it is worth noting that the digital pictures (Fig. 5 and 6) could yield better LODs than a commercial UV instrument (Fig. 4).

Fig. 6 First component PC1 [principal component analysis of RGB parameters from digital pictures of sensory membranes Mem1% and Mem2.5% dipped in water containing different concentrations of Hg(II)] vs. Hg(II) concentration. The PC1 data correspond to the mean of the PC1 values obtained with three digital pictures (taken with the cameras of: Samsung Galaxy smart phone, Kodak EasyShare DX4330, and Sony DSC-W100). The data corresponding to each individual device are shown in ESI – Fig. S5.† The depicted LOD and LOQ were calculated using the results obtained by PC1 components obtained with the latter device (similar results were obtained using the other devices).¹¹ The error bars are the standard deviation of the PC1 values corresponding to the RGB parameters of the pictures taken with the three devices.

The accuracy of the mercury concentration determination using the RGB parameters was verified by measuring the cation concentration of a sample prepared with tap water from our laboratory, thus emulating a real sample. The Hg(II) concentrations added to two samples of tap water were 1.00 × 10⁻⁸ M and 1.00 × 10⁻⁷ M and the concentrations determined were 2.07 × 10⁻⁸ M and 1.02 × 10⁻⁷ M, respectively, in good accordance with the real concentration of the sample (the concentration of Hg(II) in the tap water of our laboratory was found to be 1.75 × 10⁻⁹ M, see ESI – Section S4†). The data for the concentration evaluation, and for the titration curve, were obtained from the RGB data of each membrane obtained from a single digital picture taken with the Samsung Galaxy mobile phone containing the membranes dipped in tap water with known concentration of Hg(II), buffered at pH = 2, and the reference, i.e., a set of membranes previously dipped in pure water with known concentrations of Hg(II). The reference, once prepared, can be dried to be used in different Hg(II) measurements.

The selectivity toward the analyte over other competitive species is a key parameter to evaluate the performance of a chemosensor. The possible interferences by other metal ions were assessed by means of dipping membrane strips in water containing Hg(II), Fe(III), Cr(VI), Cr(III), Cu(II), Ni(II), Fe(II), and Zn(II). The good selectivity toward Hg(II) could be observed qualitatively with the naked eye and quantitatively by means of applying a PCA to the RGB parameters of each piece of membrane, as shown in Fig. 7.

Fig. 7 Picture taken with a Sony DSC-W100 digital camera of strips of Mem1% after dipping for 25 min in water at pH = 2 in the presence of metal ions with a concentration of 1 × 10⁻⁵ M (left); and PCA (right) of RGB parameters of the pictures of the membranes using the standardization parameters and the PCA weights calculated for the titration of Hg(II) (Fig. 6, ESI – Section S12†).

In conclusion, we have prepared a colorimetric sensory membrane for detecting Hg(II) with the naked eye. The color changes were correlated with the Hg(II) concentrations, and the membrane composition was tuned to fit special detection requirements (e.g., determining whether water samples had Hg(II) concentrations that exceeded the legal limits). Furthermore, pictures taken with conventional cameras of these membranes after exposure to aqueous solutions of Hg(II) were used to quantify the Hg(II) from the digital parameters within the pictures. Thus, pictures taken of sensory membranes after dipping for 25 min in the aqueous solutions with the camera of a conventional smart phone over white paper gave excellent results.

This research was supported by the Spanish Ministerio de Economía y Competitividad-Feder (MAT2011-22544) and by the Consejería de Educación-Junta de Castilla y León (BU001A10-2).

Notes and references

1. (a) O. T. Butler, J. M. Cook, C. F. Harrington, S. J. Hill, J. Rieuwerts and D. L. Miles, J. Anal. At. Spectrom., 2006, 21, 217; (b) Y. Li, C. Chen, B. Li, J. Sun, J. Wang, Y. Gao, Y. Zhao and Z. Chai, J. Anal. At. Spectrom., 2006, 21, 94; (c) M. Leermakers, W. Baeyens, P. Quevauviller and M. Horvat, TrAC, Trends Anal. Chem., 2005, 24, 383.
2. (a) J. Du, M. Hu, J. Fan and X. Peng, Chem. Soc. Rev., 2012, 41, 4511; (b) R. Martínez-Máñez and F. Sancenón, Chem. Rev., 2003, 103, 4419; (c) R. Martıínez-Máñez and F. Sancenón, J. Fluoresc., 2005, 15, 267; (d) E. M. Nolan and S. J. Lippard, Chem. Rev., 2008, 108, 3443; (e) A. P. de Silva, H. Q. N. Gunaratne, T. Gunnlaugsson, A. J. M. Huxley, C. P. McCoy, J. T. Rademacher and T. E. Rice, Chem. Rev., 1997, 97, 1515.
3. (a) J. M. García, F. G. García, F. Serna and J. L. de la Peña, Polym. Rev., 2011, 51, 341; (b) H. N. Kim, Z. Guo, W. Zhu, J. Yoon and H. Tian, Chem. Soc. Rev., 2011, 40, 79.
4. (a) L. Shen, J. A. Hagen and I. Papautsky, Lab Chip, 2012, 12, 4240; (b) J. P. Lafleur, S. Senkbeil, T. G. Jensen and J. P. Kutter, Lab Chip, 2012, 12, 4651.
5. National Primary Drinking Water Regulations, United States Environmental Protection Agency (EPA), EPA 816-F-09–0004, May 2009, http://water.epa.gov/drink/contaminants/index.cfm.
6. (a) L. Ling, Y. Zhao, J. Du and D. Xiao, Talanta, 2012, 91, 65; (b) V. A. Lozano, G. M. Escandar, M. C. Mahedero and A. Muñoz de la Peña, Anal. Methods, 2012, 4, 2002; (c) F. García, J. M. García, B. García-Acosta, R. Martínez-Máñez, F. Sancenón and J. Soto, Chem. Commun., 2005, 2790; (d) S. Vallejos, P. Estevez, F. C. García, F. Serna, J. L. de la Peña and J. M. García, Chem. Commun., 2010, 46, 7951; (e) A. Gómez-Valdemoro, R. Martínez-Máñez, F. Sancenón, F. C. García and J. M. García, Macromolecules, 2010, 43, 7111; (f) S. Vallejos, H. El Kaoutit, P. Estévez, F. C. García, J. L. de la Peña, F. Serna and J. M. García, Polym. Chem., 2011, 2, 1129; (g) A. Gómez-Valdemoro, M. Trigo, S. Ibeas, F. C. García, F. Serna and J. M. García, J. Polym. Sci., Part A: Polym. Chem., 2011, 49, 3817.
7. (a) Z. Gu, M. Zhao, Y. Sheng, L. A. Bentolila and Y. Tang, Anal. Chem., 2011, 83, 2324; (b) N. Dave, M. Y. Chan, P. J. J. Huang, B. D. Smith and J. Liu, J. Am. Chem. Soc., 2010, 132, 12668; (c) P. Das, A. Ghosh, H. Bhatt and A. Das, RSC Adv., 2012, 2, 3714; (d) Z. Guo, J. Duan, F. Yang, M. Li, T. Hao, S. Wang and D. Wei, Talanta, 2012, 93, 49; (e) Y. Helwa, N. Dave, R. Froidevaux, A. Samadi and J. Liu, ACS Appl. Mater. Interfaces, 2012, 4, 2228; (f) N. Zhang, G. Li, Z. Cheng and X. Zuo, J. Hazard. Mater., 2012, 93, 49.
8. H. El Kaoutit, P. Estévez, S. Ibeas, F. C. García, F. Serna, F. B. Benabdelouahab and J. M. García, Dyes Pigm., 2013, 96, 414.
9. (a) M. Beija, C. A. M. Afonso and J. M. G. Martinho, Synthesis and applications of rhodamine derivatives as fluorescent probes, Chem. Soc. Rev., 2009, 38, 2410; (b) H. N. Kim, M. H. Lee, H. J. Kim, J. S. Kim and J. Yoon, Chem. Soc. Rev., 2008, 37, 1465.
10. (a) X. Chen, T. Pradhan, F. Wang, J. S. Kim and J. Yoon, Chem. Rev., 2012, 112, 1910; (b) Y. K. Yang, K. J. Yook and J. Tae, J. Am. Chem. Soc., 2005, 127, 16760; (c) L. Wang, B. Zheng, Y. Zhao, J. Du and D. Xiao, Anal. Methods, 2012, 4, 2369; (d) F. Yan, D. Cao, N. Yang, Q. Yu, M. Wang and L. Chen, Sens. Actuators, B, 2012, 162, 313; (e) W. Liu, J. Chen, L. Xu, J. Wu, H. Xu, H. Zhang and P. Wang, Spectrochim. Acta, Part A, 2012, 85, 38; (f) L. Wang, J. Yan, W. Qin, W. Liu and R. Wang, Dyes Pigm., 2012, 92, 1083; (g) B. N. Ahamed and P. Ghosh, Inorg. Chim. Acta, 2011, 372, 100; (h) H. Liu, Y. Zhou and C. Yao, Inorg. Chem. Commun., 2011, 14, 1798; (i) M. Kumar, N. Kumar, V. Bhalla, H. Singh, P. R. Sharma and T. Kaur, Org. Lett., 2011, 13, 1422; (j) S. Saha, P. Mahato, U. Reddy, E. Suresh, A. Chakrabarty, M. Baidya, S. K. Ghosh and A. Das, Inorg. Chem., 2012, 51, 336; (k) N. Wanichacheva, K. Setthakarn, N. Prapawattanapol, O. Hanmeng, V. S. Lee and K. Grudpan, J. Lumin., 2012, 132, 35; (l) H. H. Wang, L. Xue, C. L. Yu, Y. Y. Qian and H. Jiang, Dyes Pigm., 2011, 91, 350.
11. The limit of detection (LOD) and the limit of quantification (LOQ) were estimated by the following equation: LOD = 3.3 × SD/s and LOQ = 10 × SD/s, where SD is the standard deviation of blank sample and s is the slope of the calibration curve in the region of low Hg(II) content (below 3.5 × 10⁻⁶ M and 1 × 10⁻⁷ M for UV/Vis and RGB analysis, respectively).

---

Footnote

† Electronic supplementary information (ESI) available: Synthesis of a sensory monomer (Rho B derivative), membrane preparation, fluorescence and UV/Vis titration curves of Hg(II) with the sensory membranes, RGB data of pictures of membranes upon dipping in water containing different concentrations of Hg(II), and PCA analysis of RGB data. See DOI: 10.1039/c2ay26307f

---