A Cooperative Effect of Bifunctionalized Nanoparticles on Recognition

Tamkang University, Tamsui, Taiwan 25137. Reported here is a cooperative effect that the sensing efficiency of the active group on gold nanoparticles ...
0 downloads 0 Views 264KB Size
Download PDF
Anal. Chem. 2005, 77, 4821-4828

A Cooperative Effect of Bifunctionalized Nanoparticles on Recognition: Sensing Alkali Ions by Crown and Carboxylate Moieties in Aqueous Media Shu-Yi Lin,† Chun-hsien Chen,*,† Meng-Chieh Lin,‡ and Hsiu-Fu Hsu‡

Department of Chemistry, National Tsing Hua University, Hsinchu, Taiwan 30013, and Department of Chemistry, Tamkang University, Tamsui, Taiwan 25137

Reported here is a cooperative effect that the sensing efficiency of the active group on gold nanoparticles (GNPs) can be significantly influenced by another proximal functional group. We previously developed a visual sensing scheme for K+ by 15-crown-5-CH2O(CH2)12SH functionalized GNPs in aqueous matrix. Upon adding K+, the GNP solution changes from red to blue. Such a transform is triggered by a 2-to-1 sandwich complexation of crown to K+, resulting in the red shift of surface plasmon absorption due to GNP aggregation. Herein, we discover that introducing a second functionality, thioctic acid (TA), onto GNPs significantly affects the sensing efficiency of crown moieties (15-crown-5-CH2O(CH2)nSH and 12-crown-4CH2O(CH2)nSH, where n ) 4, 8, and 12). The rate constant of K+ recognition by TA- and 15-crown-5-CH2O(CH2)4S-bifunctionalized GNPs is more than 4 orders of magnitude faster than the others containing longer methylene chains. The same chain-length dependence is also found in the case of Na+ sensing by 12-crown-4 functionalized GNPs. The discrepancy in sensing performance is attributed to a cooperative effect that the negatively charged carboxylate of TA may preorganize the crown moiety for K+ recognition. This method is applied to measure K+ and Na+ in human urine by UV-visible spectrometry. By adjusting the concentrations of GNPs, the dynamic ranges tuned for K+ and Na+ are, respectively, 6.25 µM-1.12 mM and 0.156-4.00 mM, suitable for real samples pretreated simply by 10-fold dilution. The results ([K+] ) 20.3 mM, [Na+] ) 45.1 mM) agree with those obtained from ICP-AES ([K+] ) 19.8 mM, [Na+] ) 43.8 mM). Spherical gold nanoparticles (GNPs) with the size of 5-20 nm in diameter exhibit an intense red color due to surface plasmon (SP) absorption,1 a result of the collective oscillations of the GNP surface electrons upon interacting with visible light of suitable * To whom correspondence should be addressed: (phone) +886 3 573 7009; (fax) +886 3 571 1082; (e-mail) [email protected]. † National Tsing Hua University. ‡ Tamkang University. (1) Bohren, C. F.; Huffman, D. R. Absorption and Scattering of Light by Small Particles; John Wiley and Sons: New York, 1998. 10.1021/ac050443r CCC: $30.25 Published on Web 06/22/2005

© 2005 American Chemical Society

wavelength. The peak position of the SP band is sensitive to the interparticle distance, and thus, the color generally turns blue or purple when GNP aggregation is taking place. The extinction coefficients of GNPs are nominally in the range of 108-1010 M-1 cm-1,2 so high that it becomes an increasingly important colorimetric reporter3 to signify the events associated with transforming GNPs from dispersion to aggregation, such as recognition of DNA,4-8 proteins,9-11 metal ions,12-14 anions,15 and saccharides.16-18 Although extraordinary in their colorimetric absorptivity, for GNP applications to be practical, it requires properties such as satisfactory specificity and quick response toward recognizing targeting species. These recognition properties rely on the functional group modified on GNPs. A rudimentary approach to selecting the sensing functionality is by screening those learned from solutionphase host-guest chemistry. However, due to the essential difference in their molecular conformations at the solution-GNP interface from those in solution, the performance of functionalized GNPs as a sensor is sometimes not up to the expectation. To relax the steric constrain of the sensing moiety, a spacer is generally (2) Link, S.; El-Sayed, M. A. J. Phys. Chem. B 1999, 103, 8410-8426. (3) Drechsler, U.; Erdogan, B.; Rotello, V. M. Chem. Eur. J. 2004, 10, 55705579. (4) Mirkin, C. A.; Letsinger, R. L.; Mucic, R. C.; Storhoff, J. J. Nature 1996, 382, 607-609. (5) Elghanian, R.; Storhoff, J. J.; Mucic, R. C.; Letsinger, R. L.; Mirkin, C. A. Science 1997, 277, 1078-1081. (6) Storhoff, J. J.; Mirkin, C. A. Chem. Rev. 1999, 99, 1849-1862. (7) Mirkin, C. A. Inorg. Chem. 2000, 39, 2258-2272. (8) Sato, K.; Hosokawa, K.; Maeda, M. J. Am. Chem. Soc. 2003, 125, 81028103. (9) Mann, S.; Shenton, W.; Li, M.; Connolly, S.; Fitzmaurice, D. Adv. Mater. 2000, 12, 147-150. (10) Niemeyer, C. M. Angew. Chem., Int. Ed. 2001, 40, 4128-4158. (11) Nam, J.-M.; Park, S.-J.; Mirkin, C. A. J. Am. Chem. Soc. 2002, 124, 38203821. (12) Kim, Y.; Johnson, R. C.; Hupp, J. T. Nano Lett. 2001, 1, 165-167. (13) Lin, S.-Y.; Liu, S.-W.; Lin, C.-M.; Chen, C.-h. Anal. Chem. 2002, 74, 330335. (14) Obare, S. O.; Hollowell, R. E.; Murphy, C. J. Langmuir 2002, 18, 1040710410. (15) Itoh, H.; Naka, K.; Chujo, Y. J. Am. Chem. Soc. 2004, 126, 3026-3027. (16) Aslan, K.; Zhang, J.; Lakowicz, J. R.; Geddes, C. D. J. Fluoresc. 2004, 14, 391-400. (17) Aslan, K.; Lakowicz, J. R.; Geddes, C. D. Anal. Biochem. 2004, 330, 145155. (18) Aslan, K.; Lakowicz, J. R.; Geddes, C. D. Anal. Chem. 2005, 77, 20072014.

Analytical Chemistry, Vol. 77, No. 15, August 1, 2005 4821

introduced19,20 and the sensing molecules are often diluted within a matrix of short and innocent alkanethiols.21-27 Herein, we engineer an example demonstrating that the matrix can play an active role in shaping the conformation of the sensing moiety and in improving the recognition performance. The sensing group of the monolayers on GNPs is crucial to the recognition performance. Ligand place-exchange reactions are one of the key processes to introduce the tailored functionality onto monolayer-protected clusters (MPCs) in organic solvents. This method is carried out simply by mixing excess free thiols with the alkanethiol-protected MPCs. The ω-functionalized alkanethiols can thus be incorporated into the straight-chain alkanethiol monolayers.19-21,28-30 The exchange reactions involve a fast and a slow step. The fast step takes place at the edge and vertex sites where free thiols can access the gold substrate relatively easily. On the faceted area where the molecules are tightly packed and exhibit relatively strong intermolecular interactions, the rate of ligand exchange is so slow that the extent of functionalization is never completed.31-33 The results of incomplete exchange are similar for GNPs in aqueous solution.34 In our previous work of functionalizing water-soluble GNPs prepared by reducing HAuCl4 with trisodium citrate, we developed an approach that the physisorbed chloride and citrate are first displaced by thioctic acid (TA), which is then exchanged by thiols containing the desired functionality in the second step.34 Because TA adsorbates are never entirely displaced, there are practically two types of functionalized molecules on GNPs (or termed bifunctionalized GNPs). Other than ligand place-exchange reactions,30-32,34-36 nucleophilic substitution and coupling derivatization can prepare MPCs21,37-41 and GNPs42 with multiple functionalities. In literature examples, however, only one of the multiple ω-func(19) Hostetler, M. J.; Green, S. J.; Stokes, J. J.; Murray, R. W. J. Am. Chem. Soc. 1996, 118, 4212-4213. (20) Ingram, R. S.; Hostetler, M. J.; Murray, R. W. J. Am. Chem. Soc. 1997, 119, 9175-9178. (21) Templeton, A. C.; Wuelfing, W. P.; Murray, R. W. Acc. Chem. Res. 2000, 33, 27-36. (22) Fullam, S.; Rao, S. N.; Fitzmaurice, D. J. Phys. Chem. B 2000, 104, 61646173. (23) McIntosh, C. M.; Esposito, E. A., III.; Boal, A. K.; Simard, J. M.; Martin, C. T.; Rotello, V. M. J. Am. Chem. Soc. 2001, 123, 7626-7629. (24) Norsten, T. B.; Frankamp, B. L.; Rotello, V. M. Nano Lett. 2002, 2, 13451348. (25) Watanabe, S.; Sonobe, M.; Arai, M.; Tazume, Y.; Matsuo, T.; Nakamura, T.; Yoshida, K. Chem. Commun. 2002, 2866-2867. (26) Hong, R.; Fischer, N. O.; Verma, A.; Goodman, C. M.; Emrick, T.; Retello, V. M. J. Am. Chem. Soc. 2004, 126, 739-743. (27) Verma, A.; Rotello, V. M. Chem. Commun. 2005, 303-312. (28) Templeton, A. C.; Cliffel, D. E.; Murray, R. W. J. Am. Chem. Soc. 1999, 121, 7081-7089. (29) Chen, S.; Murray, R. W. Langmuir 1999, 15, 682-689. (30) Ionita, P.; Caragheorgheopol, A.; Gilbert, B. C.; Chechik, V. J. Am. Chem. Soc. 2002, 124, 9048-9049. (31) Shaffer, A. W.; Worden, J. G.; Huo, Q. Langmuir 2004, 20, 8343-8351. (32) Donkers, R. L.; Song, Y.; Murray, R. W. Langmuir 2004, 20, 4703-4707. (33) Gu, T.; Whitesell, J. K.; Fox, M. A. Chem. Mater. 2003, 15, 1358-1366. (34) Lin, S.-Y.; Tsai, Y.-T.; Chen, C.-C.; Lin, C.-M.; Chen, C.-h. J. Phys. Chem. B 2004, 108, 2134-2139. (35) Warner, M. G.; Reed, S. M.; Hutchison, J. E. Chem. Mater. 2000, 12, 33163320. (36) Woehrle, G. H.; Brown, L. O.; Hutchison, J. E. J. Am. Chem. Soc. 2005, 127, 2172-2183. (37) Templeton, A. C.; Hostetler, M. J.; Kraft, C. T.; Murray, R. W. J. Am. Chem. Soc. 1998, 120, 1906-1911. (38) Templeton, A. C.; Hostetler, M. J.; Warmoth, E. K.; Chen, S.; Hartshorn, C. M.; Krishnamurthy, V. M.; Forbes, M. D. E.; Murray, R. W. J. Am. Chem. Soc. 1998, 120, 4845-4849.

4822

Analytical Chemistry, Vol. 77, No. 15, August 1, 2005

Chart 1. Structures of Crown Thiols Employed in the Modification of GNPs

tionalized alkanethiols performs actively and the others silently take the passive role. Improving the performance of the active functionality toward sensing and recognition by the comodified molecules is an important and as yet unexplored subject. In this study, the classical crown-alkali complexation is utilized as a model system to demonstrate that the conformation of crown moiety at the bifunctionalized GNPs can be preorganized for recognizing alkali cations by proximal carboxylate thiols. In aqueous solution, crown molecules appear not a good host for alkali cations due to the highly flexible ethylene glycol moiety, whose oxygen atoms are driven outward by the surrounding polar water molecules. To complex with an alkali cation, it requires reorganizing the conformation such that the hydrophobic ethylene becomes the exterior against the hydrophilic environment.43 Therefore, the association constant for 15-crown-5 to K+ is only 5.5-5.7,44 and 12-crown-4 does not complex with Na+ in aqueous media.44 We previously found that 15-crown-5-CH2O(CH2)12SH (1c, see Chart 1) functionalized GNPs are a specific and efficient colorimetric sensor toward K+ in water, presumably due to a preorganized structure of 15-crown-5 at the water-alkyl interface.13 The preorganized structure expedites a 2-to-1 sandwich complexation of 15-crown-5 to K+,45-48 GNP aggregation, and a distinct change in solution color from red to blue. In the current study, TA and crown thiols (shown in Chart 1) are comodified on GNPs (Figure 1) via the aforementioned two-step method in aqueous solutions.34 The numbers of crown thiols to TA molecules reach a stable ratio, consistent with the result of ligand place-exchange reactions. Between the TA and crown domains, there are boundaries where interactions between the two functional groups take place. We will show that such interactions can be fine-tuned by adjusting the chain length of crown thiols (Figure 1) and exhibit a cooperative effect on improving the performance of GNPs in K+ and Na+ recognition. RESULTS Characterization of the Bifunctionalized GNPs. Figure 2 shows the transmittance FT-IR spectra of the bifunctionalized (39) Jordan, R.; West, N.; Ulman, A.; Chou, Y.-M.; Nuyken, O. Macromolecules 2001, 34, 1606-1611. (40) Patolsky, F.; Weizmann, Y.; Lioubashevski, O.; Willner, I. Angew. Chem., Int. Ed. 2002, 41, 2323-2327. (41) Sung, K.-M.; Mosley, D. W.; Peelle, B. R.; Zhang, S.; Jacobson, J. M. J. Am. Chem. Soc. 2004, 126, 5064-5065. (42) Grubisha, D. S.; Lipert, R. J.; Park, H.-Y.; Driskell, J.; Porter, M. D. Anal. Chem. 2003, 75, 5936-5943. (43) Steed, J. W.; Atwood, J. L. Supramolecular Chemistry; John Wiley & Sons: New York, 2000. (44) Izatt, R. M.; Pawlak, K.; Bradshaw, J. S. Chem. Rev. 1991, 91, 1721-2085. (45) Christensen, J. J.; Hill, J. O.; Izatt, R. M. Science 1971, 174, 459-467. (46) Xia, W.-S.; Schmehl, R. H.; Li, C.-J. J. Am. Chem. Soc. 1999, 121, 55995600. (47) Yamauchi, A.; Hayashita, T.; Nishizawa, S.; Watanabe, M.; Teramae, N. J. Am. Chem. Soc. 1999, 121, 2319-2320. (48) Kim, J.; McQuade, D. T.; McHugh, S. K.; Swager, T. M. Angew. Chem., Int. Ed. 2000, 39, 3868-3872.

Figure 1. Relative positions of the carboxylate and crown moiety for crown thiols with different chain lengths.

Figure 2. Transmission IR spectra of the TA and crown bifunctionalized GNPs in KBr pellets. The crown molecules are (A) 15-crown-5 and (B) 12-crown-4 thiols. The crown thiols employed in the functionalization of GNPs are specified by the notation (Chart 1 and Figure 1) denoted to the left of each spectrum.

GNPs dispersed in KBr. The spectra exhibit the characteristic features of crown moiety at 949, 1117, 1252, and 1300 cm-1; carboxylate at about 1409 (νsym(CO2-)) and 1559 cm-1 (νasy(CO2-)); and carboxylic acid at 1700 cm-1 (ν(CO2H)), confirming the presence of both crown thiols and thioctic acid on GNPs. Also apparent in Figure 2 is that the intensity ratios of νasy(CH2) at 2850-2930 cm-1 to those of crown moieties are larger for GNPs modified by longer crown thiols. The positions of νasy(CH2) are centered at 2928, 2924, and 2920 cm-1, respectively, for 1a-, 1b-, and 1c-GNPs with 4, 8, and 12 methylene units for the corresponding 15-crown-5 thiols. Similar peak positions of νasy(CH2) and the same trend of red-shift associated with increasing number of methylene units are observed for 2-GNPs (Figure 2B). The νasy(CH2) peak positions49,50 indicate that the monolayers are liquidlike for the n ) 4 crown thiols (1a- and 2a-GNPs), disorganized for the relatively short ones (1b- and 2b-GNPs), and highly crystalline with primarily trans-zigzag extended polymethylene for the n ) 12 ones (1c- and 2c-GNPs).51,52 The numbers of crown thiols per GNP are roughly estimated by the averaged particle diameter deduced from transmission (49) Porter, M. D.; Bright, T. B.; Allara, D. L.; Chidsey, C. E. D. J. Am. Chem. Soc. 1987, 109, 3559-3568. (50) Laibinis, P. E.; Whitesides, G. M.; Allara, D. L.; Tao, Y.-T.; Parikh, A. N.; Nuzzo, R. G. J. Am. Chem. Soc. 1991, 113, 7152-7167. (51) Bain, C. D.; Troughton, E. B.; Tao, Y.-T.; Evall, J.; Whitesides, G. M.; Nuzzo, R. G. J. Am. Chem. Soc. 1989, 111, 321-335. (52) Chidsey, C. E. D.; Bertozzi, C. R.; Putvinski, T. M.; Mujsce, A. M. J. Am. Chem. Soc. 1990, 112, 4301-4306.

electron microscopy (TEM) electromicrographs, the weight ratio of organic layers to gold cores of GNPs measured at elevating temperatures by thermogravimetric analysis, and the abundance ratio of oxygen to sulfur atoms by X-ray photoelectron spectroscopy (XPS). The abundance ratio is estimated by taking the peak areas and atomic sensitivity factors into consideration. From the abundance ratio and the numbers of sulfur and oxygen atoms per molecule (two sulfur and two oxygen atoms per TA; one sulfur and 6 oxygen atoms per 15-crown-5 thiol; one sulfur and five oxygen atoms per 12-crown-4 thiol), the percentages of the crown thiols to total number of thiol legs per GNP are roughly in the range of 50-70%. The numbers of crown thiols and TA per GNPs are around 9500 ( 2000 and 2300 ( 600, respectively. The detail numbers are listed in Table S1 (Supporting Information). Colorimetric Response of Recognizing Alkali Cations by the Bifunctionalized GNPs. The most intriguing finding in this study is that the response time and sensitivity of recognizing alkali by GNPs are significantly dependent on the chain length of the crown thiols. To illustrate this chain-length-dependent effect, Figure 3 shows a series of photographs taken 5 min after introducing the targeting alkali cations into the bifunctionalized GNP solutions. The GNPs are very soluble in aqueous solutions and infinitely stable at pH g7.34 Prior to adding the targeting alkali cations, the GNP solutions appear ruby red. For example, the solution shown in Figure 3A contains 2.4 mL of 17.2 nM 1a-GNPs and 2.5 mM NaCl. NaCl is introduced deliberately to serve as the Analytical Chemistry, Vol. 77, No. 15, August 1, 2005

4823

Figure 3. Colorimetric response for the TA and crown bifunctionalized GNPs subjected to adding alkali cations. The respective crown thiols on GNPs and the composition in the solutions are (A) 1a, 2.5 mM Na+, (B) 1a, 2.5 mM Na+ + 25 µM K+, (C) 1b, 2.5 mM Na+ + 100 µM K+, (D) 1c, 2.5 mM Na+ + 100 µM K+, (E) 2a, 2.5 mM K+ + 5.0 mM Na+, (F) 2b, 2.5 mM K+ + 5.0 mM Na+, and (G) 2c, 2.5 mM K+ + 5.0 mM Na+. 2.5 mM Na+ (B-D) and 2.5 mM K+ (E-G) represent interferences. The maximum of the surface plasmon band for the GNP solutions is adjusted to 1.36 AU prior to the test. The photographs were taken 5 min after introducing the target cations. Noted that the concentration of targeting alkali metal in panel B is the lowest and yet the colorimetric response is the most significant.

interference and to mimic the extracellular matrix for K+ recognition.53,54 The absorbance of all crown thiol-modified GNP solutions is adjusted to the same level and the color is identical to that of Figure 3A (not shown). Upon exposure to 25 µM KCl, the solution of 1a-GNPs transforms to blue instantly (Figure 3B and Figure 4C,D, vide infra). Considering the prompt response, the distinct transform in color, the sensitivity, and the 100-fold concentration ratio of Na+ to K+, the sensing performance appears adequate for real sample analysis. Interestingly, the response for 1b- and 1c-GNPs subjected to the same concentration level of K+ is slow (Figure 4, vide infra). Panels C and D show, respectively, solutions of 1b-GNPs and 1c-GNPs responding to 100 µM K+, a higher concentration level utilized to manifest the superior sensing abilities of 1a-GNPs to the others. The color of Figure 3C (1bGNPs) essentially remains the same but becomes fainter than that prior to adding K+. To observe apparent change in color within minutes, it required at least 0.3 mM K+ present in the 1b-GNP solution. For 1c-GNPs, upon exposing to 0.10 mM K+, the color promptly turns purple (Figure 3D) which, however, is not comparable to that of Figure 3B. 2-GNPs exhibit the same trend of chain-length dependence as 1-GNPs. Panels E-G of Figure 3 show intense blue for 2aGNPs, almost unchanged for 2b-GNPs, and deep purple for 2cGNPs upon responding to 5.0 mM Na+. The solutions contain 2.5 mM K+ as the interfering agent. The visual detection limit reaches 0.6 mM Na+ in the presence of 2.5 mM K+ by dilution of the 2a-GNP solutions (Abs ) 0.238 AU at λ525 nm) to reduce the intense background of red.13 Although the sensitivity and specificity are sufficient in examining serum (∼150 mM Na+ with the presence of 5 mM K+)53,54 or human urine samples (40-220 mequiv/24 h Na+ and 30-90 mequiv/24 h K+),53 it is, however, far inferior to that of recognizing K+ by 1-GNPs. We believe that (53) Fiereck, E. A. In Fundamentals of Clinical Chemistry: Appendix; Tietz, N. W., Ed.; Saunders: Philadelphia, 1970; pp 634-657. (54) Masilamani, D.; Lucas, M. E. Flourescent Chemosensors for Ion and Molecule Recognition; Czarnik, A. W., Ed.; American Chemical Society: Washington, DC, 1992.

4824 Analytical Chemistry, Vol. 77, No. 15, August 1, 2005

Figure 4. Changes in SP absorbance versus time upon exposure to alkali cations for (A-D) 1-GNPs and (E, F) 2-GNPs. GNPs bifunctionalized by crown thiols with 4, 8, and 12 methylene units are shown in the blue, red, and purple traces, respectively. Concentrations of alkali cations in the cuvettes: (A, B) 2.5 mM Na+ + 100 µM K+; (C, D) 2.5 mM Na+ + 25 µM K+; (E, F) 2.5 mM K+ + 2.5 mM Na+; where 2.5 mM Na+ (A-D) and 2.5 mM K+ (E, F) represent interferences. For panels C and D, the absorbance changes for 1band 2b-GNPs are indiscernible and thus their traces are not shown for clarity.

the relatively poor sensitivity of 2a-GNPs is due to the stronger hydrating tendency of Na+ (-402 kJ/mol) than that of K+ (-314 kJ/mol)55 and the weaker affinity of 12-crown-4 to Na+ (no complexation was found in aqueous solution)44 than that of 15crown-5 to K+ (K ) 5.7 in aqueous solution).44 Taking together the cases of alkali recognition by 15-crown-5 and 12-crown-4 moieties, Figure 3 demonstrates that the performances are the best for the shortest crown thiols (1a- and 2a-GNPs, n ) 4), intermediate for the longest ones (1c- and 2c-GNPs, n ) 12), and poor for the those with eight methylene units (1b- and 2bGNPs). We examined the specificity of 1- and 2-GNPs and found that the GNPs do not respond to physiologically important cations such as Mg2+, Ca2+, Zn2+, Cu2+, and NH4+ (6-8.5 mequiv/24 h, 1.25-3.75 mequiv/24 h, 2.3-18.3 µequiv/24 h, 0.05-0.55 µequiv/ 24 h, and 36-85 mequiv/24 h, respectively)53 (see Supporting Information Figures S1 and S2) because they do not form stable 2-to-1 complexes. Kinetic Study and Reaction Order. Figure 4 correlates the response time with the chain length of crown thiols by monitoring the SP absorbance, which is preadjusted to 1.36 AU. Upon responding to 100 µM K+ for 1a-GNPs (blue curves in Figure 4A and B), the diminishing of A527 nm and the concomitant rising of (55) Zumdahl, S. S. Chemistry: Chemistry; D. C. Heath and Co.: Lexington, 1986; p 284.

A700 nm56 are so rapid that the recognition event is visually assured instantaneously. At 1700 s, the magnitude of ∆A527 nm decreases 1.5 AU, suggesting that 1a-GNPs are almost exhausted. Although the response of 1c-GNPs (purple curves) to K+ recognition is distinctive to naked eyes, the magnitude and rate of the absorbance changes are far inferior to those of 1a-GNPs. 1b-GNPs are not sensitive enough to respond discernibly (red curves). The enhanced sensitivity of 1a-GNPs over 1c-GNPs is emphasized by panels C and D of Figure 4, where a low concentration level of 25 µM K+ is employed in the presence of 2.5 mM Na+. The response of 2-GNPs to Na+ (Figure 4E and F) exhibits the same chain-length dependence as that of 1-GNPs. To explore the kinetic aspect, empirical equations for the recognition event and the corresponding reaction rate are, respectively, described by eqs 1 and 2. The coupled SP absorbance

Table 1. Summary of Initial Rates and Rate Constants for Recognizing K+ and Na+ by 1- and 2-GNPs CGNPs,t0 (nM)

CM+,t0 (µM)

initial rate Rt0 (M s-1)

rate constant k (M-2 s-1)

average k (M-2 s-1)

1a-GNPs

16.5 16.5 16.5 5.3 8.9

25 50 100 100 100

4.5 × 10-11 7.8 × 10-11 1.5 × 10-10 2.3 × 10-11 5.8 × 10-11

5.4 × 106 5.1 × 106 5.4 × 106 5.5 × 106 5.8 × 106

5.5 × 106

1c-GNPs

15.8 15.8 15.8 5.1 8.5

50 100 200 100 100

5.9 × 10-11 1.0 × 10-10 1.4 × 10-10 2.5 × 10-11 4.6 × 10-11

1.1 × 102 1.2 × 102 1.1 × 102 1.2 × 102 1.2 × 102

1.2 × 102

2a-GNPs

8.5 8.5 8.5 4.8 6.0

4000 5000 7500 5000 5000

6.3 × 10-12 8.2 × 10-12 1.2 × 10-11 2.9 × 10-12 4.6 × 10-12

5.2 × 105 5.3 × 105 5.2 × 105 5.3 × 105 5.6 × 105

5.3 × 105

2c-GNPs

8.4 8.4 8.4 4.0 5.0

5000 6250 7500 5000 5000

7.3 × 10-12 8.3 × 10-12 9.9 × 10-12 2.7 × 10-12 3.5 × 10-12

3.1 × 101 3.0 × 101 3.1 × 101 3.1 × 101 3.0 × 101

3.1 × 101

GNPs

pM+ + qGNPs f aggregates (aq) f aggregates (s)V (1) reaction rate ) -(1/q)(d[GNPs]/dt) ) k[M+]p[GNPs]q (2) at A700 nm and A680 nm (see Figure 4) generally soars up first and then decreases gradually. The decrease is attributed to the sedimentation of aggregated GNPs whose size grows to a certain extent that they can no longer disperse in the solution. Thus, the dispersive aggregates and sedimentary forms are denoted as aggregates (aq) and aggregates (s) in eq 1, respectively. Note that sandwich complexes may proceed by adjacent crowns in the same GNP.57,58 This reaction pathway is not discussed here because it does not induce GNP aggregation and thus is not probed by UV-visible spectra. Equation 2 shows that the reaction rate, R, can be obtained by the consumption of the reactant as a function of time. The apparent reaction rate is measured by monitoring the decrease in the SP absorbance at 527 nm for 1-GNPs or 525 nm for 2-GNPs. The initial-rate method59 is employed to determine the reaction orders for M+ and GNPs as well as the apparent rate constants. The necessary equations and mathematical details are provided in the Supporting Information. The values of p and q for 1a-, 1c-, 2a-, and 2c-GNPs are 0.87 and 1.67, 0.62 and 1.23, 1.03 and 1.79, and 0.75 and 1.35, respectively. The quotient of q to p ranges from 1.7 to 2.0, suggesting a 2-to-1 complexation of GNPs to K+ or Na+ at the initial stage. The results derived by the initial-rate method are summarized in Table 1 where the apparent rate constants for GNPs modified by short crown thiols (i.e., 1a- and 2a-GNPs) are more than 4 orders of magnitude faster than those modified by long ones. Quantitative Analysis of Human Urine Sample. The analytical application of 1a- and 2a-GNPs is demonstrated by the quantification of K+ and Na+ in human urine samples where the typical concentrations fall in 30-90 and 40-220 mmol/24 h, respectively.53 The calibration curves are obtained by correlating the concentration of analytes to the decrease in peak intensity of the dispersive SP band. To cope with real samples, the dynamic (56) Storhoff, J. J.; Lazarides, A. A.; Mucic, R. C.; Mirkin, C. A.; Letsinger, R. L.; Schatz, G. C. J. Am. Chem. Soc. 2000, 122, 4640-4650. (57) Flink, S.; Boukamp, B. A.; van den Berg, A.; van Veggel, F. C. J. M.; Reinhoudt, D. N. J. Am. Chem. Soc. 1998, 120, 4652-4657. (58) Flink, S.; van Veggel, F. C. J. M.; Reinhoudt, D. N. J. Phys. Chem. B 1999, 103, 6515-6520. (59) Birk, J. P. J. Chem. Educ. 1976, 53, 704-707.

ranges are adjusted to 6.25 µM-1.12 mM for K+ and 0.156-4.00 mM for Na+ by diluting the concentration of GNPs to 2.30 nM for 1a-GNPs and 0.58 nM for 2a-GNPs. The spectra and calibration curves were obtained 15-20 min after introducing the samples because the SP absorbance declines only 3-6% of the overall decrease within the time span. The continuous decay in absorbance does not cause serious problems in reproducibility if the timing for data acquisition is followed. The two calibration curves show good linearity with slopes of -2.57 × 10-2 and -1.34 × 10-3 AU/mM, intercepts of 0.116 and 0.527 AU, and correlation coefficients of 0.998 and 0.997 for determining K+ and Na+, respectively (see Figure S3 in Supporting Information). The urine samples were treated simply by 10-fold dilution to minimize the matrix effect. The concentrations of K+ and Na+ are found 20.3 ( 1.5 and 45.1 ( 2.0 mM, respectively (n ) 3 each). To ascertain the correctness of the results, the samples are spiked with K+ (2.0 mM) and Na+ (5.0 mM), and the respective recovery rates are 93 and 95%. Parallel comparison is made by ICP-AES measurements, which give 19.8 mM for K+ and 43.8 mM for Na+ in the urine sample (RSD < 2%). These results ensure the validity and selectivity of 1a- and 2a-GNPs in the application of real samples. DISCUSSION In aqueous solutions, crown ethers complex poorly with cations,44 orders of magnitude smaller in formation constants than those in organic solvents.60 Why do the recognition events shown here appear efficient and chain-length dependent? In organic media, the nonpolar ethylene backbones dominate the outer surface, resulting in a preorganized polar cavity ready for capturing alkali ions.43,61 Contrarily, in aqueous media, the interior and exterior of crown are inverted from what appeared in nonpolar media. It required a structural turnover for the crown to host a (60) Gokel, G. W.; Leevy, W. M.; Weber, M. E. Chem. Rev. 2004, 104, 27232750. (61) Cram, D. J. Angew. Chem., Int. Ed. Engl. 1986, 25, 1039-1134.

Analytical Chemistry, Vol. 77, No. 15, August 1, 2005

4825

Figure 5. Proposed structures of the crown moiety preorganized due to the neighboring molecules of (A) thioctic acid and (B) thioctic amine. The cuvettes contain 16.5 nM concentrations of the corresponding GNPs, 100 µM K+, and 2.5 mM Na+. The solutions are buffered to pH 5.85 by 10 mM (2-(N-morpholino)ethanesulfonic acid).

metal ion and thus results in a small formation constant in aqueous media. Given the prompt sensing of some GNPs studied here, preorganization of the crown moiety has to take place to certain degrees which lead to the chain-length-dependent performance. For GNPs functionalized by long-chain crown thiols (1c- and 2c-GNPs, n ) 12), the peak position of νasy(CH2) at 2920 cm-1 indicates crystalline packing of methylene chains. Note that on a planar surface self-assembled monolayers prepared by the same molecule, 1c, is essentially a liquidlike film whose νasy(CH2) is centered at 2930 cm-1.57,58 This is ascribed to the steric effect arising from the bulky terminal crown moiety, which interrupts the interchain van der Waal attractions. The difference between a planar surface and the spherical feature of GNPs generates the discrepancy in molecular packing. Namely, as away from the curved surface, the void space created due to reduced packing density (number of chains per unit area) plays an important role in relaxing the steric effect from the bulky crown moiety. The crown moieties poise at the interface of the aqueous and polymethylene phases, resembling a polar-to-nonpolar boundary facilitating preorganization of crowns whose oxygen atoms face the polar solution and the ethylene groups rest on the organic shell of GNPs. Therefore, the 2-to-1 sandwich complexation of 1c-GNPs to K+ and 2c-GNPs to Na+ becomes favorable. For the other GNPs, their νasy(CH2) modes appear at ∼2930 cm-1, indicative of disordered organic shells. It is peculiarly interesting to discover that the sandwiched complexation is efficient for 1a- and 2a-GNPs (n ) 4) but imperfect for 1b- and 2b-GNPs (n ) 8), given their similarity in constituents and liquidlike monolayers. Because the disparity in sensing performance arises from the chain length of crown thiols, the effects attributable to the relative positions between the carboxylate and the crown moiety are pondered thoroughly. A plausible explanation is depicted in Figure 5A that the electrostatic repulsion between the proximal groups shapes the crown of 1a (n ) 4) into a preorganized host. Such preorganized microstructures are likely chain-length dependent because electrostatic force is inversely proportional to the square of the distance. Therefore, such preorganization is less likely for GNPs modified by longer crown thiols (n ) 8 and 12). 4826 Analytical Chemistry, Vol. 77, No. 15, August 1, 2005

To prove the proposed hypothesis, GNPs are functionalized by thioctic amine, as opposed to thioctic acid, such that the preorganization of crown and sensing performance is influenced. The association constant for 15-crown-5 to ammonium is 51 in aqueous medium.44,62 Thioctic amine is synthesized, and 1a′-, 1b′-, and 1c′-GNPs denote GNPs functionalized by thioctic amine and 15-crown-5 thiols of n ) 4, 8, and 12, respectively. Figure 5B illustrates that the crown-to-amine affinity60,63 attracts the crown moiety and makes it difficult to preorganize for K+ recognition. The test results are consistent with our proposed mechanism. 1a′GNPs are insensitive to addition of K+ (e.g., the cuvette shown in Figure 5B), and the sensing performance is even inferior to that of 1b-GNPs (n ) 8). For 1b′- and 1c′-GNPs, the amino-tocrown spacing becomes longer and their performance is not very different from that of 1b- and 1c-GNPs, respectively. This chainlength-dependent behavior of 1′-GNPs supports the proposed electrostatic interactions sketched in Figure 5. Another possibility is the electrostatic interactions between the charged functional group and K+. This factor is concomitantly associated with using either thioctic acid or amine to prepare the bifunctionalized GNPs. The positively charged K+ is attracted by the carboxylate on 1a-GNPs or repelled by the amino group on 1a′-GNPs. The electrostatic attraction spaced by crown thiols is also chain-length dependent. However, electrostatics is not selective and attracts more interfering agents than the analyte. In the case of sensing K+ by 15-crown-5, the major interferance is Na+ whose size matches well with the crown cavity. Namely, electrostatics draws Na+ to block the active sites and probably impedes the subsequent recognition toward K+. Therefore, the electrostatic attraction is less likely a dominant factor for the enhancement in K+ recognition by TA. Because the factors of preorganization and electrostatics cannot be studied independently, the contribution of electrostatic attractions in K+ recognition cannot be ruled out. Nevertheless, the exceptional performance of 1a-GNPs represents (62) Izatt, R. M.; Terry, R. E.; Haymore, B. L.; Hansen, L. D.; Dalley, N. K.; Ayondet, A. G.; Christensen, J. J. J. Am. Chem. Soc. 1976, 98, 7620-7626. (63) Weber, E.; Toner, J. L.; Goldberg, I.; Vogtle, F.; Laidler, D. A.; Stoddart, J. F.; Bartsch, R. A.; Liotta, C. L. In Crown Ethers and Analogs; Patai, S., Ed.; John Wiley & Sons: Chichester, 1989; pp 339-340.

a result of cooperative effect of proximal crown and carboxylate functional groups. CONCLUDING REMARKS In summary, the colorimetric sensing of 1- and 2-GNPs exhibits an interesting chain-length dependence with the best performance for the shortest crown thiol (n ) 4), intermediate for the longest one (n ) 12), and poor for that with eight methylene units. Real sample analysis with 1a- and 2a-GNPs is demonstrated by spectrophotometric measurements of K+ and Na+ in human urine samples. The results match those obtained by ICP-AES. The reaction order derived from the kinetic study suggests the 2-to-1 sandwich complexation of 1-GNPs to K+ and 2-GNPs to Na+. The rate constants for 1a- and 2a-GNPs sandwiching alkali cations are more than 4 orders of magnitude faster than those of 1c- and 2c-GNPs. When the GNPs are bifunctionalized by thioctic amine and crown thiols, the trend is altered and the sensing performance for the shortest crown thiol (1a′-GNPs, n ) 4) becomes the worst. These findings manifest a cooperative effect that the comodified thiols may improve or undermine the sensing efficiency of the tailored sensing group on GNPs. This concept is important in designing and optimizing the performance of nanoparticle sensors. EXPERIMENTAL SECTION General Information. All chemicals were reagent grade and used as received. To avoid contamination due to releasing potassium and sodium ions from glassware, PP plastic assemblies (sodium-free containers) were used in synthetic procedures and quartz cuvettes were used for UV-visible measurements (Unicam UV 300 spectrophotometer). To adjust the concentrations of GNPs by regulating the SP absorbance of UV-visible spectra, the peak height was determined by using the local baseline. Surface functionalities were characterized by TEM, FT-IR (Perkin-Elmer, Spectrum 2000), and XPS (Physical Electronics, ESCA PHI 1600). Preparation of Crown Thiols. The synthetic procedures for crown-terminated alkanethiols were adapted from the van Veggel and Reinhoudt method.57 2-[(4-Mercaptobutyl)oxy]methyl-15-crown-5 (1a). A solution of 2-[(4-bromobutyl)oxy]methyl-15-crown-5 (0.80 g, 2.44 mmol) (this compound was synthesized from 2-hydroxymethyl15-crown-5 by alkylation with 1,4-dibromobutane) and thiourea (Showa, 0.93 g, 12.2 mmol) in ethanol (Tedia, 50 mL) was heated under reflux for 16 h. The solvent was evaporated under reduced pressure. Sodium hydroxide (0.63 g, 15.74 mmol, 6.45 equiv) and nitrogen-purged water were added to the residue. The mixture was heated under reflux for 2 h. After acidifying with 1 M HCl, CH2Cl2 (100 mL) was added and the organic layer was washed with water (3 × 100 mL). The solution was dried over MgSO4(s) and filtered. The solvent was evaporated, and the crude product was purified by column chromatography (SiO2, hexane/EtOAc 1:3) to yield 2a as a colorless oil (0.45 g, 54.4%): 1H NMR (500 MHz, Varian, Uniytinova-500, 11.8 T, CDCl3 (Merck), internal reference TMS) δ 1.25-1.35 (m, 2H), 1.57-1.68 (m, 2H), 2.51 (q, 2H), 3.33-3.50 (m, 4 H), 3.5-3.88 (m, 19 H). 2-[(8-Mercaptooctyl)oxy]methyl-15-crown-5 (1b). This compound was synthesized from 2-hydroxymethyl-15-crown-5 (0.99 g, 3.96 mmol) by alkylation with 1,8-dibromooctane (2.63 g, 9.64 mmol). The terminal bromide was converted into thiol,

analogous to the synthesis of 1a. The crude product was purified by column chromatography (SiO2, hexane/EtOAc 1:3) to give 1b as a colorless oil (0.50 g, 32.5%): 1H NMR (500 MHz, CDCl3) δ 1.25-1.35 (m, 6H), 1.57-1.68 (m, 6H), 2.51 (q, 2H), 3.33-3.50 (m, 4 H), 3.5-3.88 (m, 19 H). 2-[(12-Mercaptododecyl)oxy]methyl-15-crown-5 (1c). The synthetic procedures and characterization of 1c were described in detail elsewhere.13 2-[(4-Mercaptobuthyl)oxy]methyl-12-crown-4 (2a). NaH used in the van Veggel and Reinhoudt method57 was replaced by KH to avoid any ambiguity caused due to introducing Na+ in the synthetic procedures. To a 50-mL round flask of quartz was adding 20 mL of ethanol, 0.58 g of 2-[(4-bromobutyl)oxy]methyl-12-crown4 (1.27 mmol) (this compound was synthesized from 2-hydroxymethyl-12-crown-4 by alkylation with 1,4-dibromobutane), and 0.48 g of thiourea (6.54 mmol). The mixture was heated under reflux overnight. The solvent was then evaporated under reduced pressure. Lithium hydroxide (0.20 g, 8.19 mmol) and nitrogenpurged water were added to the residue. The mixture was heated under reflux for 2 h. After acidifying with 1 M HCl, CH2Cl2 (100 mL) was added and the organic layer was washed with water (3 × 100 mL). The solution was dried over MgSO4(s) and filtered, and the solvent was evaporated to give the crude compound, which was purified by column chromatography (SiO2, hexane/EtOAC ) 1:3) to yield 2a as a colorless oil (0.24 g, 47%): 1H NMR (500 MHz, CDCl3) δ 1.18-1.42 (m, 2H), 1.50-1.65 (m, 2H), 2.51 (q, 2H), 3.35-3.52 (m, 5H), 3.61-3.82 (m, 14H). 2-[(8-Mercaptooctyl)oxy]methyl-12-crown-4 (2b). This compound was synthesized from 2-hydroxymethyl-12-crown-4 (0.82 g, 3.88 mmol) by alkylation with 1,8-dibromooctane (3.72 g, 20.2 mmol). The terminal bromide was converted into thiol, analogous to the synthesis of 2a. The crude product was purified by column chromatography (SiO2, hexane/EtOAC ) 1:3) to give 2b as a colorless oil (0.099 g, 27%): 1H NMR (500 MHz, CDCl3) δ 1.22-1.42 (m, 10H), 1.45-1.65 (m, 2H), 2.49 (q, 2H), 3.353.49 (m, 5H), 3.50-3.85 (m, 14H). 2-[(12-Mercaptododecyl)oxy]methyl-12-crown-4 (2c). This compound was synthesized from 2-hydroxymethyl-15-crown-5 (0.77 g, 3.88 mmol) by alkylation with 1,12-dibromooctane (6.63 g, 20.2 mmol). The procedures of synthesis and purification were the same as that of 2a. The purified product, 2c, was a colorless oil (0.037 g, 24%): 1H NMR (500 MHz, CDCl3) δ 1.24-1.41 (m, 12H), 1.49-1.71 (m, 8H), 2.50 (q, 2H), 3.38-3.51 (m, 5H), 3.613.85 (m, 14H). 1,2-Dithia-3-(1-amino-n-pentyl)cyclopentane (Thioctic Amine) (3).64 DL-R-Lipoamide (thioctic amide, 1.0 g, 4.46 mmol) was dissolved in 100 mL of THF in a 250-mL round flask to which 50 mL of THF containing 826 mg of LiAlH4 (21.7 mmol, 4.87 equiv) was added. The solution was refluxed at 70 °C for 12 h. Distilled water (8 mL) was then added, and the solution was stirred at 0 °C for 30 min. After evaporation of solvent, methanol was added and the insoluble solid was removed. The supernatant was evaporated, and an aliquot of 100 mL of distilled water was added. The pH of the solution was adjusted to 6.5 with 1 N HCl, and the solution was stirred at room temperature for 24 h. The product was extracted with 1-butanol and washed with 1 N NaOH and 1 (64) Mortia, T.; Kimura, S.; Kobayashi, S. J. Am. Chem. Soc. 2000, 122, 28502859.

Analytical Chemistry, Vol. 77, No. 15, August 1, 2005

4827

N HCl. The solvent was evaporated to give the crude product, which was purified by column chromatography (SiO2, chloroform/ methanol/acetic acid 90:10:3 in volume ratio, Rf ) 0.18) to give 3 as a pale-yellow solid (0.29 g, 31%): 1H NMR (500 MHz, CDCl3) δ 1.47 (m, 4H), 1.65 (m, 4H), 1.89 (m, 2H), 2.47 (m, 2H), 2.76 (b, 2H), 3.15 (m, 2H), 3.56 (m, 1H). Preparation and Modification of GNPs. All glassware was thoroughly cleaned with aqua regia (3:1 HCl/HNO3) and rinsed with Millipore-Q water prior to use. To avoid the presence of the target cations in the preparation procedures, GNPs was prepared by reducing HAuCl4 with sodium citrate65 or potassium citrate for K+ and Na+ recognition, respectively. The average diameter measured by TEM (Hitachi, H-7500) was 18.0 ( 3.0 nm. The citrate-stabilized GNPs were bifunctionalized by the twostep method.34 For the first step, 2.3 mg of thioctic acid (TCI) was added to 100 mL of citrate-stabilized gold sols whose basicity was preadjusted to pH 11 by 1.5 mL of 0.5 M NaOH for 1-GNPs or KOH for 2-GNPs. The gold sols were stirred overnight (1624 h). The TA-functionalized nanoparticles were then centrifuged twice for 20 min (at 15 700g, 10 °C, Sorvall, Biofuge Stratos, Germany), followed by decantation of supernatants. The purified TA-stabilized GNPs were redissolved in 100 mL of Millipore-Q water (without electrolytes). For the second step, a 0.4-mL aliquot of the crown thiols (10 mM/ethanol) was mixed with 4.0 mL of TA-GNPs, and the solution was stirred for 2 days. Excess thiols were removed by 20-min centrifugation twice (at 15 700g, 10 °C), followed by decantation of supernatants. The purified bifunctionalized GNPs were redissolved in Millipore-Q water. The solution pH became 5.85. For GNPs bifunctionalized by thioctic amine and 15-crown-5 thiols, the negatively charged surface (due to adsorption of citrate) became positively charged, and therefore, the procedures were slightly different for step one. A 100-µL aliquot of thioctic amine (87 mM in methanol) was added to 1.0 mL of citrate-stabilized gold sols. The solution immediately became dark purple. Then (65) Grabar, K. C.; Freeman, R. G.; Hommer, M. B.; Natan, M. J. Anal. Chem. 1995, 67, 735-743.

4828

Analytical Chemistry, Vol. 77, No. 15, August 1, 2005

100 µL of HCl (1 M) was added, and the gold sol turned red immediately. After stirring for 1 day, the sols were centrifuged for 10 min (at 15 700g, 10 °C), followed by decantation of supernatants. The second step was exactly the same as the previously described. The purified GNPs was redissolved in 10 mM MES (2-(N-morpholino)ethanesulfonic acid, Sigma) which buffered the solution pH at 5.85. The extinction coefficients for 1a- and 2a-GNPs in the MES buffer were on the order of ∼108 cm-1 M-1. Preparation and Analysis of Human Urine Samples. The 24-h urine samples were collected from healthy volunteers. Immediately after each collection, the samples were stored in a 0 °C refrigerator and were defrosted to room temperature prior to analysis. The urine samples were 10-fold diluted by MES buffer (10 mM, pH 5.85), which was also used in preparing GNPs solutions. To cope with the amount of K+ and Na+ in real samples, the concentrations are 2.3 and 0.58 nM for 1a- and 2a-GNPs, respectively. The pH of MES buffer was adjusted with 0.5 N NaOH or 0.5 N KOH for analysis of potassium and KOH, respectively. UV-visible spectra were obtained 15-20 min after introducing a 0.70-mL aliquot of samples into 1.80 mL of GNPs. ACKNOWLEDGMENT The authors thank the National Science Council (R.O.C.), National Tsing Hua University, and the Department of Chemistry for generous financial and research support. Thanks also to Mr. Sung-Hsun Wu for his help in real sample analysis. SUPPORTING INFORMATION AVAILABLE Photographs and UV-visible spectra for studies of interference (Figures S1 and S2); calibration curves (Figure S3); details of initial rate method (Figure S4); number of crown thiols per GNPs (Table S1). This material is available free of charge via the Internet at http://pubs.acs.org. Received for review March 15, 2005. Accepted May 23, 2005. AC050443R