Surface enhanced Raman scattering fiber optic sensor as an ion selective optrode: the example of Cd²⁺ detection

Abstract

Here in this work, we report the fabrication of a (metal) ion selective surface enhanced Raman scattering (SERS) optrode, the counterpart of an ion selective electrode, for the detection of metal ions in solution. Following our previous work, a layer-by-layer self-assembly strategy was used to fabricate the SERS optrode, followed by modification with an ion chelating reagent, 4-(4-phenylmethanethiol)-2,2′:6′,2′′-terpyridine (PMTTP). The SERS spectrum change after binding with metal ions was used to identify and detect metal ions in solution. Cd²⁺ in aqueous solution was chosen as a sample analyte. Similar to standard pH measurement, through simple single point known standard solution calibration, a quick (semi-)quantitative analysis of Cd²⁺ was realized.

---

Introduction

An optrode is a type of spectroscopic sensor that involves fiber optics as the light source and signal transmitting media, where the signal transforming active layer is on the distal end.¹ The main advantage of an optrode is that it suits places where it is dangerous or space is hindered for access, such as in underground sensing, and nuclear waste management. It can also be extended to a situation where there is very little sample volume, as in a single live cell.² Since the first optrode was invented for the detection of ammonia,³ a number of investigations have been made in this field, including chemical sensors,⁴ biosensors,⁵ ion sensors⁶ and so on. Optrodes for heavy metals⁷ based on absorption,⁸ chemiluminescence,⁹ or fluorescence¹⁰ methods have been widely reported.

In recent years, surface enhanced Raman scattering (SERS), a technique that not only can achieve single-molecule sensitivity with the help of nanostructured noble metals, but also is used for direct identification of substance using its inherent finger-printed spectra of molecules, has gained wide attention.¹¹ In fact, SERS is considered to be one of the most powerful spectroscopic analytical methods.^11–15

Inspired by both the merits of fiber optic sensor and SERS, different groups, including ourselves, have put efforts in building SERS optrodes.^16–26 Currently, the focus is mostly on the fabricating of effective signal transmitting media, i.e., SERS substrate, where the protocols include layer-by-layer self-assembly of metallic NPs,^20–24 chemical vapor deposition modification,²⁵ in situ photo reduction modification,^16,26 templated^27,28 or non-templated²⁹ physical vapor deposition,^30,31 and nano-fabricated structures on physical deposited metal films^32,33 on one end of the optical fibers (bundles).

However, the analytes of the SERS optrode currently demonstrated are limited, most of them are “sample analytes”, such as dyes,^21,24,32,34 thiols,^{24,25,29,31,33} and other small molecules.^24,27 Though there are reports using SERS technique for sensing metal ions in water with organic chelating probes,^35–45 there is only one reported work about the development of SERS optical fiber sensor for detecting metal ions.⁴² However, it does not really fall into the definition of optrode since optical fiber only served as SERS substrate and signal transmitting media.^42,46 In other words, it would be difficult to be applied in remote sensing situation. Further, unlike pH electrode and contrary to what the word optrode may imply, a series of concentration gradients have to be tested to obtain a calibration curve before each measurement.⁴² We believe two possible reasons might have hindered the development of ion selective SERS optrode: (1) the lack of methods for fabricating SERS optrode with high sensitivity and stability; and (2) lack of good chelating reagents that can flag the existence of different metal ions with SERS spectra change. Due to stability issue, it is hard to reuse the SERS optrode. On the other hand, since there is huge variation, it is almost impossible for quantification analysis through multiple SERS optrodes. All these make the idea of using SERS optrode for the analysis of metal ions less attractive compared with its counterpart, ion selective electrode.

Previously, we have demonstrated that through layer-by-layer assembly of Ag NPs on single mode²⁰ and multimode²¹ optical fiber, SERS detection with near single molecule sensitivity could be achieved. Meanwhile, by incubating the SERS optrode in N₂ atmosphere at 120 °C for 1 hour, the stability could be greatly improved. Both multiplexing and continuous monitoring of different sample analytes in a microfluidic chip were realized.²¹ In this report, we will examine the feasibility of developing SERS optrode for metal ion detection. We modified the distal end of the SERS optrode with a metal ion chelating reagent, 4-(4-phenylmethanethiol)-2,2′:6′,2′′-terpyridine (PMTTP, Fig. 2 inset). Upon binding with heavy metal ions, characteristic new feature(s) will appear. Cd²⁺ was chosen as the sample analyte. It was found that the binding of Cd²⁺ will introduce a new peak around 1020 cm⁻¹, which was also reversible by treating with EDTA. By carefully choosing an intact band of PMTTP at 1040 cm⁻¹ as internal reference, the normalized SERS intensity at 1020 cm⁻¹ can be used for the quantification of Cd²⁺. Most importantly, this quantification could be done through a single point calibration. Cadmium in aqueous solution can be detected semi-quantitatively. The analysis procedure of our SERS optrode for determination of Cd²⁺ exactly mimics an ion selective electrode, such as pH electrode. Binding constant between PMTTP and Cd²⁺ was extracted and found to be in agreement with literature.⁴⁷ We believe this is a very important step in building ion selective SERS optrode for metal ions analysis, which may open up great opportunity for (SERS) optrode sensor development.

Experimental part

Chemicals

All the reagents used were from Sigma without further purification. High-purity water (Nanopure, Thermo Scientific) was used throughout this study.

The optrode was fabricated following our previous report by layer-by-layer self-assembly of Ag NPs based on APTMS sol–gel on the fiber tip.^20,21 Five rounds of Ag NPs depositions were performed before it was cured at 120 °C under N₂ atmosphere for 1 h. The synthesis of PMTTP (Fig. 2A inset) can also be found in our previous work.⁴⁸

Instrumentation

SERS measurements were performed on a customized Raman microscope. The detector was a Pixis-100BR CCD (Princeton Instrument), the dispersion system was an Acton SP-2500i spectrograph, and the excitation source was a 20 mW He–Ne laser (632.8 nm). The accuracy of Raman shift was calibrated using α-pinene as external standard. Small shift of Raman frequency (<1 cm⁻¹) is expected from day to day. A 20× objective (N.A. = 0.45) were used throughout the work. All spectra were average of 8 measurements, recorded with 500 μW laser power and the accumulation time was 20 s. The SERS measurement configuration of the ion selective SERS optrode can be found in our previous work.²⁰

Results and discussion

Characterization of the optrode and the binding to heavy metal ions

The as prepared SERS optrode was characterized by SEM. Fig. 1 shows the morphology of Ag NPs modified tip of the optrode. It is clear that the entire surface is covered by Ag NPs of various sizes and shapes. Similar to our previous work,²⁰ there are many gaps among the particles where the hotspots can be localized, which guarantee the high sensitivity of the optrode.

Fig. 1 SEM image of Ag NPs modified tip of the SERS optrode.

Later, the as fabricated optrodes were immersed into 1 mM PMTTP ethanolic solution for 10 min, resulting the formation of PMTTP layer on the optrode. Then, the SERS spectrum of the PMTTP on the optrode was recorded, as shown in Fig. 2A. Next, 5 optrodes were dipped into different solutions for 10 min and the respective spectra were recorded after, as shown in Fig. 2B, where special attention was given to the region between 960 to 1060 cm⁻¹. It is clear that dipping into water does not change the SERS spectrum of PMTTP (Fig. 2B(a) and (b)), as expected, indicating there is probably no structural and/or orientational change. The difference caused by introducing of 4 different metal ions varied. First of all, lead ion seems do not interact with the PMTTP, at least not flagged by SERS. The chelating with Pb²⁺ added a tiny shoulder around 1016 cm⁻¹ (shown by arrow). Cadmium, on the other hand, brought a more dominant band around 1020 cm⁻¹. The largest change appears to be zinc, where the new band shifted to 1025 cm⁻¹. Note that except Cadmium, which is 0.2 mM in concentration, the other three ions are all at 1 mM. It is also interesting to note that the degree of perturbation of the PMTTP spectrum falls into the inverse order of the atomic numbers. That is, for Pb²⁺, there is no observable change of the original spectrum of PMTTP. However, for the element with the smallest atomic number tested, Zn²⁺, it not only introduces a new peak, but also has the largest shift (higher wavenumbers). We suspect that the size of the metal ion plays a role here,^49–51 since the valence of all the ions tested are 2⁺. We further tested 8 more metal ions and the results were listed in Fig. s1.† Though all the ions responded differently on SERS spectra, potential interference might exist when mixing different ions together. However, such kind of spectral interference can be resolved through method such as spectral deconvolution.⁵²

Fig. 2 (A) SERS spectrum of PMTTP, the inset shows its molecular structure. (B) SERS spectrum of (a) PMTTP; (b) after immersing in water; (c) to (f) after immersing 10 min in 1 mM Pb²⁺, 1 mM Hg²⁺, 0.2 mM Cd²⁺, and 1 mM of Zn²⁺ solution.

(Semi-)quantitative application of SERS optrode: Cd²⁺ as example

In this section, we will show how the SERS optrode can work the same way as its counterpart ion selective electrode.

Since the reusability of SERS fiber optic substrate is one of the key factors in optrode development,²⁴ we first examined whether the binding of cadmium ion to PMTTP is reversible. As shown in Fig. 3, the PMTTP modified optrode was firstly immersed in 1 mM cadmium nitrate aqueous solution for 10 min, and then in 0.1 M EDTA solution for 30 min, then back into Cd²⁺ solution again. It is clear that due to the chelating with Cd²⁺, a new peak at 1020 cm⁻¹ appears (Fig. 3b). The washing step with 0.1 M EDTA aqueous solution made it back to PMTTP's original SERS spectrum (Fig. 3c). However, even after 5 cycles, with the introducing of Cd²⁺, the peak at 1020 cm⁻¹ appears again (Fig. 3d).

Fig. 3 Recovery by EDTA. (a) The initial SERS spectrum of the optrode; (b) after immersing into 1 mM of Cd²⁺ for 10 min; (c) after 5 cycles of Cd²⁺ dip-EDTA washing; (d) immersing into 1 mM of Cd²⁺ for another 10 min.

From Fig. 2B and 3 we show that upon binding of Cd²⁺, new characteristic peak at 1020 cm⁻¹ will appear. We will further show that quantification of Cd²⁺ in solution can be realized based on this new feature.

As shown in Fig. 4A, it is clear that the new band at 1020 cm⁻¹ increase with increasing cadmium concentration. However, there is almost no correlation (data not shown) between the absolute intensity and cadmium concentration. This is not surprising. In this type of indirect SERS analysis, usually a ratio of the affected and the intact bands were used, to minimize any environmental variations.⁵³ Thus, we decide to use the band at 1040 cm⁻¹ as the internal reference. In Fig. 4B, the SERS signal ratio I₁₀₂₀/I₁₀₄₀ (y-axis) was plotted against the concentration of cadmium ion solution (x-axis), and a non-linear response was found. To extract the binding constant between PMTTP and Cd²⁺, we further fitted the data with the well-known Langmuir isotherm model:⁵⁴

x/y = x + 1/K_ads, (1)

where y is I₁₀₂₀/I₁₀₄₀, x is the concentration of Cd²⁺, K_ads is the binding constant between the Cd²⁺ and PMTTP (Fig. 4B inset). It was found that there is very good correlation coefficient R (R² = 0.992), indicating the binding is well described by the Langmuir model. Furthermore, the K_ads was extracted and was found to be in perfect match with that between PMTTP's precursor 2,2′,2′′-terpyridine and Cd²⁺ (lg K of 5.0 versus 5.1).⁴⁷

Fig. 4 (A) SERS spectra of PMTTP at different Cd²⁺ concentrations (from bottom to top (arrow): 0, 10 μM, 20 μM, 80 μM and 200 μM) (B) calibration curve (I₁₀₂₀/I₁₀₄₀ (y) versus Cd²⁺ concentration (x)) built using data from 3 optrodes. Inset shows the Langmuir isotherm plot. See text for more detail. In all cases, optrodes were immersed in 0.1 M of EDTA for 30 min to remove bound Cd²⁺ before any new measurement.

Fig. 4B demonstrates that there is a correlation between the ratio (I₁₀₂₀/I₁₀₄₀, y) and Cd²⁺ concentration (x) in the range of 10 μM to 200 μM. Since the binding of PMTTP and Cd²⁺ gives distinct vibrational spectrum change, it is possible to use the proposed SERS optrode to detect heavy metal ion Cd²⁺ in solution qualitatively and quantitatively. However, this procedure would still be tedious since a calibration curve needs to be built first. We then turn back and check the counterpart of ion selective SERS optrode. The simplest ion selective electrode is pH electrode. In field, people only carry one or two standard pH solution(s), and calibrate the pH meter right before any measurement. We feel that we could make our ion selective SERS optrode work the same way as a pH electrode. To do that, the Langmuir isotherm would be a better choice as the working calibration curve since it is linear form. For a specific optrode, the constant (1/K_ads) in eqn (1) will then have to be amended via single-point calibration before applying to unknown samples.

As a proof of concept, in the following part, we will demonstrate how to apply the optrode to quantify cadmium ion in solution with single point calibration. Firstly, a SERS optrode was immersed into a known Cd²⁺ standard solution (40 μM in this work) and the SERS spectrum was collected. Thus the relative SERS intensity ratio I₁₀₂₀/I₁₀₄₀ (y) for the known standard was generated using the spectrum collected in this first step. This y value and the concentration x (40 μM in this work) were then substituted into eqn (1), so that a new constant was produced (C, column 2 in Table 1). Then, eqn (1) turned into

x/y = x + C, or x = C/(1/y − 1) (2)

and this was the working equation for this optrode. Later, the optrode was immersed into 0.1 M EDTA solution for recovering, and then in samples to obtain the y_exp (i.e., I₁₀₂₀/I₁₀₄₀). By substituting the y_exp into eqn (2), concentration of the unknown sample could be calculated. In Table 1, a sample containing 20 μM Cd²⁺ was tested by 7 different optrodes, respectively. The calculated concentrations were listed in column 4 in Table 1. It is worth noting that the single-point calibration process was implemented for each optrode. It is found that the recovery rate for 20 μM of Cd²⁺ was in the range of 69.5–157% (last column in Table 1), which is acceptable for a fast onsite screening. We believe this successfully proves the concept of ion selective SERS optrode, and could be used for future onsite applications.

Table 1 Recovery rates for samples containing 20 μM Cd²⁺ measured from 7 different optrodes based on the proposed single point calibration method (known standard: 40 μM)

Optrode	Constant^a (C)	yexp	xcal [M]	R%
a Working equation: x = C/(1/y − 1) (eqn (2)).
1	1.70 × 10^−5	0.562	2.17 × 10^−5	109%
2	0.621 × 10^−5	0.694	1.41 × 10^−5	70.5%
3	2.09 × 10^−5	0.601	3.14 × 10^−5	157%
4	2.17 × 10^−5	0.572	2.90 × 10^−5	145%
5	1.77 × 10^−5	0.519	1.91 × 10^−5	95.5%
6	1.46 × 10^−5	0.487	1.39 × 10^−5	69.5%
7	1.98 × 10^−5	0.524	2.18 × 10^−5	109%

---

Conclusions

In this paper, an ion selective SERS optrode that mimics ion selective electrode was developed. PMTTP, a chemical that can chelate with different metal ions, was synthesized and immobilized on the SERS optrode. The presence of a few heavy metal ions: Hg²⁺, Cd²⁺, and Zn²⁺ was flagged through the changes of SERS spectrum in the 980–1080 cm⁻¹ range, respectively. Most importantly, we propose that a SERS optrode can work exactly the same way as ion selective electrode. That is, through a single known standard amending of pre-built calibration curve, metal ions can be (semi-)quantitatively determined. As an example, Cd²⁺ in water was investigated with the proposed method. The existence of Cd²⁺ was identified by a new peak at 1020 cm⁻¹ and further quantified by relative intensity ratio of I₁₀₂₀/I₁₀₄₀ in the concentration range from 10 μM to 200 μM. The recovery rate for 20 μM of Cd²⁺ was found to be in the range of 69.5–157%. We believe that such kind of ion selective SERS optrodes will have broad applications in many disciplines.

Acknowledgements

This research is financially supported by the National Natural Science Foundation of China (Grant no. 21277131, 21105092, 51103141) and the Fundamental Research Funds for the National Universities of China (2682014RC06).

Notes and references

1. P. R. Stoddart and D. J. White, Anal. Bioanal. Chem., 2009, 394, 1761–1774.
2. J. P. Scaffidi, M. K. Gregas, V. Seewaldt and T. Vo-Dinh, Anal. Bioanal. Chem., 2009, 393, 1135–1141.
3. D. J. David, M. C. Willson and O. S. Ruffin, Anal. Lett., 1976, 9, 389–404.
4. V. G. Andreou and Y. D. Clonis, Anal. Chim. Acta, 2002, 460, 151–161.
5. J. I. R. De Corcuera, R. P. Cavalieri and J. R. Powers, Food Sci. Technol. Int., 2003, 9, 5–9.
6. C. E. Reese and S. A. Asher, Anal. Chem., 2003, 75, 3915–3918.
7. J. Janata, Anal. Chem., 1992, 64, 921A–927A.
8. S. O'Neill, S. Conway, J. Twellmeyer, O. Egan, K. Nolan and D. Diamond, Anal. Chim. Acta, 1999, 398, 1–11.
9. G. A. Crespo and E. Bakker, Analyst, 2012, 137, 4988–4994.
10. H. R. He, H. Li, G. Mohr, B. Kovacs, T. Werner and O. S. Wolfbeis, Anal. Chem., 1993, 65, 123–127.
11. M. K. Fan, G. F. S. Andrade and A. G. Brolo, Anal. Chim. Acta, 2011, 693, 7–25.
12. Y. K. Kim, S. W. Han and D. H. Min, ACS Appl. Mater. Interfaces, 2012, 4, 6545–6551.
13. X. M. Lin, Y. Cui, Y. H. Xu, B. Ren and Z. Q. Tian, Anal. Bioanal. Chem., 2009, 394, 1729–1745.
14. R. Aroca, Surface enhanced vibrational spectroscopy, Wiley, Hoboken, NJ, 2006.
15. E. C. Le Ru and P. G. Etchegoin, Principles of surface-enhanced Raman spectroscopy: and related plasmonic effects, Elsevier, Amsterdam, Boston, 1st edn, 2009.
16. Y. Kitahama, T. Itoh, J. Aoyama, K. Nishikata and Y. Ozaki, Chem. Commun., 2009, 6563–6565,  10.1039/b909603e.
17. J. W. Lim and I. J. Kang, Bull. Korean Chem. Soc., 2014, 35, 25–29.
18. L. X. Li, S. P. Liu, Z. Y. Chen, Z. M. Dai, N. Chen, F. F. Pang, Y. N. Shang, B. Lu and T. Y. Wang, Opt. Spectrosc., 2014, 116, 575–578.
19. Z. Y. Chen, Z. M. Dai, N. Chen, S. P. Liu, F. F. Pang, B. Lu and T. Y. Wang, IEEE Photonics Technol. Lett., 2014, 26, 777–780.
20. G. F. S. Andrade, M. K. Fan and A. G. Brolo, Biosens. Bioelectron., 2010, 25, 2270–2275.
21. M. K. Fan, P. H. Wang, C. Escobedo, D. Sinton and A. G. Brolo, Lab Chip, 2012, 12, 1554–1560.
22. A. Lucotti, A. Pesapane and G. Zerbi, Appl. Spectrosc., 2007, 61, 260–268.
23. E. Polwart, R. L. Keir, C. M. Davidson, W. E. Smith and D. A. Sadler, Appl. Spectrosc., 2000, 54, 522–527.
24. A. Pesapane, A. Lucotti and G. Zerbi, J. Raman Spectrosc., 2010, 41, 256–267.
25. A. Amezcua-Correa, J. Yang, C. E. Finlayson, A. C. Peacock, J. R. Hayes, P. J. A. Sazio, J. J. Baumberg and S. M. Howdle, Adv. Funct. Mater., 2007, 17, 2024–2030.
26. X. L. Zheng, D. W. Guo, Y. L. Shao, S. J. Jia, S. P. Xu, B. Zhao, W. Q. Xu, C. Corredor and J. R. Lombardi, Langmuir, 2008, 24, 4394–4398.
27. D. L. Stokes and T. Vo-Dinh, Sens. Actuators, B, 2000, 69, 28–36.
28. G. Kostovski, D. J. White, A. Mitchell, M. W. Austin and P. R. Stoddart, Biosens. Bioelectron., 2009, 24, 1531–1535.
29. C. Viets and W. Hill, Sens. Actuators, B, 1998, 51, 92–99.
30. D. J. White and P. R. Stoddart, Opt. Lett., 2005, 30, 598–600.
31. V. Guieu, F. Lagugne-Labarthet, L. Servant, D. Talaga and N. Sojic, Small, 2008, 4, 96–99.
32. S. Q. Chen, L. Han, A. Schulzgen, H. B. Li, L. Li, J. V. Moloney and N. Peyghambarian, Opt. Express, 2008, 16, 13016–13023.
33. E. J. Smythe, M. D. Dickey, J. M. Bao, G. M. Whitesides and F. Capasso, Nano Lett., 2009, 9, 1132–1138.
34. A. Lucotti and G. Zerbi, Sens. Actuators, B, 2007, 121, 356–364.
35. J. Duan, M. Yang, Y. Lai, J. Yuan and J. Zhan, Anal. Chim. Acta, 2012, 723, 88–93.
36. G. Wang, C. Lim, L. Chen, H. Chon, J. Choo, J. Hong and A. J. deMello, Anal. Bioanal. Chem., 2009, 394, 1827–1832.
37. D. Han, S. Y. Lim, B. J. Kim, L. Piao and T. D. Chung, Chem. Commun., 2010, 46, 5587–5589.
38. T. Senapati, D. Senapati, A. K. Singh, Z. Fan, R. Kanchanapally and P. C. Ray, Chem. Commun., 2011, 47, 10326–10328.
39. V. M. Zamarion, R. A. Timm, K. Araki and H. E. Toma, Inorg. Chem., 2008, 47, 2934–2936.
40. G. Aragay, H. Monton, J. Pons, M. Font-Bardia and A. Merkoci, J. Mater. Chem., 2012, 22, 5978–5983.
41. S. J. Lee and M. Moskovits, Nano Lett., 2011, 11, 145–150.
42. L. G. Crane, D. X. Wang, L. M. Sears, B. Heyns and K. Carron, Anal. Chem., 1995, 67, 360–364.
43. H. Qu, Y. Lai, D. Niu and S. Sun, Anal. Chim. Acta, 2013, 763, 38–42.
44. Y. Lai, J. Wang, T. He and S. Sun, Anal. Lett., 2013, 47, 833–842.
45. F. Li, J. Wang, Y. Lai, C. Wu, S. Sun, Y. He and H. Ma, Biosens. Bioelectron., 2013, 39, 82–87.
46. K. I. Mullen, D. Wang, L. G. Crane and K. T. Carron, Anal. Chem., 1992, 64, 930–936.
47. Chromogenic Reagents for Inorganic Ions, ed. Y. Zeng, H. Zhang and Z. Chen, Chemical Industry Press, Beijing, 1989.
48. X. Zhang, D. Li and X.-P. Zhou, New J. Chem., 2006, 30, 706–711.
49. K. Carron, K. Mullen, M. Lanouette and H. Angersbach, Appl. Spectrosc., 1991, 45, 420–423.
50. D. Buist, N. J. Williams, J. H. Reibenspies and R. D. Hancock, Inorg. Chem., 2010, 49, 5033–5039.
51. N. J. Williams, W. Gan, J. H. Reibenspies and R. D. Hancock, Inorg. Chem., 2009, 48, 1407–1415.
52. J. Y. Baron, J. Mirza, E. A. Nicol, S. R. Smith, J. J. Leitch, Y. Choi and J. Lipkowski, Electrochim. Acta, 2013, 111, 390–399.
53. O. Peron, E. Rinnert, T. Toury, M. L. de la Chapelle and C. Compere, Analyst, 2011, 136, 1018–1022.
54. M. De Keersmaecker, D. Depla, K. Verbeken and A. Adriaens, J. Electrochem. Soc., 2014, 161, C126–C137.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c4ra11260a

---