Biosensors and Bioelectronics 63 (2015) 61–71

Contents lists available at ScienceDirect

Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios

Fluorescent carbon nanoparticles for the fluorescent detection of metal ions Yongming Guo n, Lianfeng Zhang, Shushen Zhang, Yan Yang, Xihan Chen, Mingchao Zhang College of Chemistry and Pharmaceutical Engineering, Nanyang Normal University, Nanyang 473061, China

art ic l e i nf o

a b s t r a c t

Article history: Received 13 May 2014 Received in revised form 4 July 2014 Accepted 4 July 2014 Available online 14 July 2014

Fluorescent carbon nanoparticles (F-CNPs) as a new kind of fluorescent nanoparticles, have recently attracted considerable research interest in a wide range of applications due to their low-cost and good biocompatibility. The fluorescent detection of metal ions is one of the most important applications. In this review, we first present the general detection mechanism of F-CNPs for the fluorescent detection of metal ions, including fluorescence turn-off, fluorescence turn-on, fluorescence resonance energy transfer (FRET) and ratiometric response. We then focus on the recent advances of F-CNPs in the fluorescent detection of metal ions, including Hg2 þ , Cu2 þ , Fe3 þ , and other metal ions. Further, we discuss the research trends and future prospects of F-CNPs. We envision that more novel F-CNPs-based nanosensors with more accuracy and robustness will be widely used to assay and remove various metal ions, and there will be more practical applications in coming years. & 2014 Elsevier B.V. All rights reserved.

Keywords: Fluorescent carbon nanoparticles Detection Metal ions Fluorescence

Contents 1. 2. 3.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Detection mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Assays of metal ions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Mercury ions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Copper ions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Iron ions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4. Other metal ions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Conclusion and future perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1. Introduction A variety of functional materials contained metal ions have been widely used all over the world and bring about great convenience to our life. However, metal ions that are released from these materials are detrimental to human health and the environment. Therefore, the detection of these metal ions is urgent for researchers and the government. Researchers have developed a wealth of methods for sensing metal ions, including optical methods (Nolan and Lippard, 2008), capillary electrophoresis (Ali and Aboul-Enein, 2002), n

Corresponding author. E-mail address: [email protected] (Y. Guo).

http://dx.doi.org/10.1016/j.bios.2014.07.018 0956-5663/& 2014 Elsevier B.V. All rights reserved.

61 62 62 62 65 68 69 70 70 70

electrochemical methods (Aragay and Merkoçi, 2012), atomic absorption spectrometry (Fang et al., 1984), inductively coupled plasma mass spectrometry (Townsend et al., 1998), and so on, which can provide satisfactory results. However, most of these methods show some limitations, such as complicated processing, high-cost instruments and time-consuming operations. Therefore, simple and lowcost methods for the detection of metal ions are highly desirable. As a new kind of carbon nanomaterials, fluorescent carbon nanoparticles (F-CNPs) have recently captivated the attention of scientists due to their unique properties, such as robust chemical inertness, low photobleaching, low toxicity, good biocompatibility, good water solubility, easy preparation, etc. (Qu et al., 2012b). These properties are, however, not found in organic dyes or semiconductor quantum

62

Y. Guo et al. / Biosensors and Bioelectronics 63 (2015) 61–71

Table 1 Comparison of the performance of various analytical methods for detecting Hg2 þ based on the LOD. Analytical methods

LODa

Reference

Optical methods (fluorescence) Capillary electrophoresis

10 μM 46.5 nM  3.86 nM  7.48 nM 5 nM 0.2 nM  2.19 nM 0.15 nM 10 nM 0.5 nM

Zhang et al., 2014b; Tian et al., 2014 Ge et al., 2014; Yang et al.,2014 Zhang et al., 2013a; Xuan et al., 2013 Türker et al., 2014; Giakisikli et al., 2013 Guo et al., 2013; Qin et al., 2013

Electrochemical methods Atomic absorption spectrometry F-CNPs-based methods

a

Limit of detection.

dots (QDs). F-CNPs are thus promising to replace these highly toxic semiconductor QDs. Fluorescent carbon nanomaterials were accidentally found by researchers in the procedure of purification of singlewalled carbon nanotubes (SWCNTs) fabricated by arc-discharge methods in 2004. They found that the short SWCNTs oxidized from long SWCNTs showed fluorescence, which is dependent on the size of SWCNTs (Xu et al., 2004). This accidental discovery has drawn great attention and provides new ideas for the fabrication of F-CNPs. Researchers have developed a handful of methods to prepare F-CNPs which exhibit desirable properties (Baker and Baker, 2010). F-CNPs have recently been widely applied in various fields, including bioimaging, sensors, photocatalysis, optoelectronics, and so on (Esteves da Silva and Gonçalves, 2011; Li et al., 2012; Shen et al., 2012). The fluorescent detection of metal ions is one of the most important applications due to their high quantum yield (QY), good photostability and low toxicity. The F-CNPs with high QY can still exhibit strong fluorescence intensity even at very low concentration. The good photostability of F-CNPs guarantees the stability of the fluorescence signal, ensuring the accuracy of the detection results. And there is no need to worry about their poisoning effect on human health and the environment. Therefore, it is facile to qualitatively determine metal ions without costly instruments or complicated operations, and quantification can also be simply achieved by monitoring the change of the fluorescence intensity. Moreover, F-CNPs are water soluble, thus the recontamination resulted from toxic organic solvent that is commonly required to disperse organic dyes can be avoided. Hence, F-CNPs exhibit great promise in the fluorescent detection of metal ions and many case studies have been explored and investigated. And we compare the performance of various analytical methods for detecting metal ions based on the limit of detection (LOD) taking Hg2 þ as an example. The performance of capillary electrophoresis, electrochemistry and atomic absorption spectrometry is superior to the optical methods (fluorescence) and F-CNPs-based methods. However, these methods including capillary electrophoresis, electrochemistry and atomic absorption spectrometry require expensive instruments and complicated operations. And the performance of the F-CNPs-based methods is better than the performance of the optical methods (fluorescence) (Table 1). Particularly, the F-CNPs-based methods are relatively inexpensive and very simple when compared to the performance of these instrumental methods and the optical methods (fluorescence). The F-CNPs can be easily synthesized from low-cost carbon sources via simple methods, such as hydrothermal methods and microwaveassisted methods (Guo et al., 2013; Barman and Sadhukhan, 2012). And the quantification of metal ions can be easily achieved based on the change of the fluorescence intensity. The F-CNPs-based methods are thus promising candidates for the fluorescent detection of metal ions in aqueous solution. Review on the recent

advances in F-CNPs for the fluorescent detection of metal ions is necessary. In the review, we mainly summarize the recent progress on the development of F-CNPs for the fluorescent detection of metal ions. It should be notified that F-CNPs in the review include all kinds of fluorescent carbon nanomaterials, including carbon dots, carbon nanodots, graphene quantum dots (GQDs), g-C3N4, and so forth.

2. Detection mechanism F-CNPs have been widely used to detect various metal ions in aqueous solution. There are mainly four types of fluorescence response modes: fluorescence turn-off, fluorescence turn-on, fluorescence resonance energy transfer (FRET) and ratiometric response. Most of F-CNPs-based nanosensors are fluorescence turn-off and the real fluorescence quenching mechanism induced by metal ions has not been fully elucidated. However, lots of researchers have tried to explain the fluorescence quenching mechanism. Most of the fluorescence quenching mechanism is mainly attributed to the electron, charge or energy transfer resulted from the selective interaction between F-CNPs and metal ions. The reason is ascribed to the functional groups on the surface of F-CNPs, such as carboxyl groups, hydroxyl groups, amino groups, etc. These functional groups can selectively interact with the specific metal ions, resulting into the complex of F-CNPs with metal ions. The complex may change the electronic structure of F-CNPs and affect the distribution of excitons, which accelerates the non-radiative recombination of the excitons through effective electron, charge or energy transfer, further results in the fluorescence quenching (Guo et al., 2013; Zhou et al., 2012). Moreover, the inner filter effect is another reason for fluorescence quenching caused by metal ions. The inner filter effect is due to the absorption of the excitation and/or emitted light by absorbers in the detection system when the absorption spectrum of the absorbers overlap with the fluorescence excitation or emission spectra of F-CNPs. The inner filter effect can enhance the sensitivity compared to other mechanisms because the changes in the absorbance of sensors can transform exponentially into fluorescence intensity changes (Badarau and Dennison, 2011; Dong et al., 2012; Zheng et al., 2013). The above mentioned reasons can contribute the fluorescence quenching of F-CNPs in the presence of the specific metal ions. For the F-CNPsbased nanosensors with fluorescence turn-off, the fluorescence signal is easily influenced by the detection medium, thus the obtained results may be not accurate. The nanosensors with fluorescence turnon, FRET or ratiometric response may improve the accuracy of the detection results. However, there are only a few case studies. In the future, more efforts have to be devoted to the development of F-CNPs-based nanosensors with fluorescence turn-on, FRET or ratiometric response.

3. Assays of metal ions 3.1. Mercury ions Mercury, one of the most toxic metal ions, has received considerable attention due to its extremely harmful effects on human health and the environment (Nolan and Lippard, 2003, 2008). Mercury vapor is easily oxidized to water-soluble mercury ions and the subsequent biotransformation will produce more toxic organic mercury, which can accumulate environmentally through the food chain (Benoit et al., 1998; Harris et al., 2003; Ma et al., 2011; Renzoni et al., 1998). Particularly, mercury is a severe neurotoxin and long-term exposure to high levels of mercury can cause damage to the brain, nervous system and

Y. Guo et al. / Biosensors and Bioelectronics 63 (2015) 61–71

63

Fig. 1. (A) Fluorescence turn-on for the detection of Hg2 þ with CuDTC2 modified F-CNPs. Reprinted with permission of American Chemical Society from reference Yuan et al. (2014). (B) Ratiometric fluorescent detection of Hg2 þ based on the nanohybrid of F-CNPs and QDs. Reprinted with permission of Elsevier from reference Cao et al. (2013).

immune system (Clarkson et al., 2003; Harris et al., 2003; Mutter et al., 2005; Zheng et al., 2003). Therefore, sensitive and selective detection of Hg2 þ is very significant for environmental, analytical and biomedical applications. Up to now, numerous techniques and methods have been well developed to monitor the level of Hg2 þ in various real-world samples (Leopold et al., 2010). The fluorescent methods for the detection of Hg2 þ have attracted widespread interest due to their simplicity and ease of operation. As a kind of fluorescent nanomaterials, many F-CNPs have been used to detect Hg2 þ in aqueous solution. Ultraviolet (UV) pulsed laser is used to prepare carbon nanoparticles with the size of about 100 nm. After functionalization with NH2-polyethylene-glycol and N-acetyl-L-cysteine (NAC), the functionalized carbon nanoparticles show fluorescence with emission wavelength at 450 nm and can be used to detect Hg2 þ . The selectivity is attributed to the presence of sulfur atom from NAC. The interaction between Hg2 þ (soft donor) and sulfur atom (soft acceptor) is very strong (Gonçalves et al., 2010). The functionalized F-CNPs can also be immobilized on the optical fiber through layerby-layer assembly method to assay Hg2 þ . The LOD decreases with increasing of the number of layers of F-CNPs on the tip of the optical fiber. The LOD with six layers of F-CNPs is 0.01 μM, but the

LOD is 0.1 μM for F-CNPs in solution. The interesting phenomenon may result from the availableness of F-CNPs to interact with Hg2 þ when immobilized in discrete layers within such a thin film (Gonçalves et al., 2012). With the help of concentrated H2SO4, ethylene glycol can be carbonized into graphitized F-CNPs at 140 ° C with the QY up to 25%. The F-CNPs can be used to assay Hg2 þ with good selectivity and sensitivity because of the static quenching mechanism from F-CNPs-Hg2 þ complex structure. The LOD in the system reached as low as 35 nM (Liu et al., 2012c). Recently, Zhang's group reported a new fluorescence turn-on nanosensor for the selective detection of Hg2 þ with bis(dithiocarbamato) copper(II) (CuDTC2) functionalized F-CNPs. They synthesized amine-coated F-CNPs and conjugated CuDTC2 complex on their surface through the condensation between carbon disulfide and the nitrogen atoms in the surface amine groups. The conjugated CuDTC2 complex on the surface of F-CNPs can effectively quench the fluorescence of F-CNPs due to the combination of electron and energy transfer. The addition of Hg2 þ can cause the recovery of the fluorescence of F-CNPs because the conjugated Cu2 þ is replaced by Hg2 þ , cutting off the energy transfer pathway (Fig. 1A). The nanosensor can detect Hg2 þ with a LOD as low as 20 nM. Interestingly, the nanosensor can be fabricated onto paper

64

Y. Guo et al. / Biosensors and Bioelectronics 63 (2015) 61–71

Fig. 2. (A) Fluorescence turn-off for the detection of Hg2 þ with F-CNPs synthesized from PEG. Reprinted with permission of Elsevier from reference Liu et al. (2013). (B) The fluorescent detection of Hg2 þ and biothiols with F-CNPs. Reprinted with permission of The Royal Society of Chemistry from reference Zhou et al. (2012).

and acts as a portable Hg2 þ nanosensor. This fluorescence turn-on nanosensor can eliminate the disturbance from the detection medium and provides an idea for detecting other metal ions with F-CNPs (Yuan et al., 2014). Ratiometric fluorescent probes have recently attracted growing interest because of their high accuracy. Cao et al. have constructed a ratiometric fluorescent nanosensor for Hg2 þ through simple mixing of the blue-emission F-CNPs with red-emission carboxylmethyldithiocarbamate modified CdSe@ZnS QDs (GDTC-QDs). The hybrid nanosensor shows dual emissions at 436 nm and 629 nm at a single excitation wavelength (365 nm). Because of the strong chelating ability of GDTC to Hg2 þ , the fluorescence of GDTC-QDs in the nanohybrid nanosensor can be effectively quenched by Hg2 þ , but the fluorescence of F-CNPs remains constant, resulting in a continuous fluorescence color change from red to blue with increasing of the concentration of Hg2 þ (Fig. 1B). And the LOD of the nanosensor is 0.1 μM. This strategy can effectively eliminate the disturbance from the detection medium because the fluorescence of F-CNPs can be used as the standard signal. More importantly, the strategy provides an insight into the detection of other metal ions just by changing the ligands on the surface of QDs (Cao et al., 2013). The above mentioned F-CNPs can detect Hg2 þ in aqueous solution with proper surface modification. Lately, many F-CNPs without complicated surface functionalization can be directly used to selectively detect Hg2 þ in aqueous solution. Sodium hydroxide-assisted reflux with poly(ethylene glycol) (PEG) has been used to prepare F-CNPs with good watersolubility and high stability. The F-CNPs exhibit excellent upconversion fluorescence properties and can be used to detect Hg2 þ with a LOD of 1 fM (Fig. 2A). The fluorescence quenched by Hg2 þ is due to the electron or energy transfer, because the oxygen-containing groups on the surface of F-CNPs can preferentially interact with Hg2 þ . In addition, the feasibility of F-CNPs has been tested with river, lake, and tap water (Liu et al., 2013). F-CNPs with graphitic carbon nitride structure (CNx) can be efficiently synthesized from formamide through a simple microwave-assisted method. The fluorescence emission of F-CNPs is strongly dependent on solvent, pH and excitation wavelengths. The complex of Hg2 þ and CNx sheet results in the fluorescence quenching because the complex involves π delocalized electron moieties of the CNx. Accordingly, the F-CNPs can be used to detect Hg2 þ with a LOD of about 1 nM (Barman and Sadhukhan, 2012). The pyrolysis of ethylenediaminetetraacetic acid salts has been employed to prepare F-CNPs, which can be directly used to detect Hg2 þ with a LOD of 4.2 nM due to Hg2 þ -quenched fluorescence. The

fluorescence quenching mechanism may be attributed to the non-radiative electron/hole recombination annihilation through an effective electron transfer process (Fig. 2B) (Zhou et al., 2012). To shorten the hydrothermal reaction time and enhance the QY of F-CNPs, we have developed a simple NH4HCO3-assisted hydrothermal method for synthesizing F-CNPs with high QY using sodium citrate as carbon source. The as-prepared F-CNPs have strong emission and excellent stability, and the fluorescence emission is independent on excitation wavelengths. The F-CNPs show excellent specificity for Hg2 þ and as low as 10 nM of Hg2 þ can be detected, which is attributed to the strong interaction between F-CNPs with Hg2 þ (Guo et al., 2013). Yan et al. synthesized two kinds of hetero atom-doped F-CNPs with citric acid as carbon source. The first kind of F-CNPs with a high QY of 65.5% is prepared through hydrothermal treatment of citric acid and ethylenediamine at 200 °C. The second kind of F-CNPs with a QY of 55.4% is synthesized through the pyrolysis of citric acid in the N-(β-aminoethyl)-γ-aminopropyl methyldimethoxy silane at 240 ° C for 1 min. The two types of F-CNPs exhibit high sensitivity and selectivity towards Hg2 þ with the LOD of 226 nM and 845 nM, respectively. The high selectivity is attributed to the fact that Hg2 þ has a stronger affinity toward the electron-rich surface of F-CNPs than other metal ions. In addition, the two F-CNPs can be used to detect and image intracellular Hg2 þ (Yan et al., 2014). Nitrogendoped F-CNPs with the size of 4.5 71.0 nm have been prepared via hydrothermal treatment of folic acid. The F-CNPs show excitation wavelength-dependent fluorescence with the maximum emission and excitation at 470 and 395 nm, respectively. It has been found that the F-CNPs can be used to detect Hg2 þ with a LOD of 0.23 μM because of the effective quenching effect of Hg2 þ . The fluorescence quenching mechanism may be attributed to the nonradiative electron-transfer from the excited states to the d orbital of Hg2 þ and the Hg2 þ -induced conversion of the functional groups (–CONH–) from spirolactam structure to an opened-ring amide (Zhang and Chen, 2014). These unmodified F-CNPs show excellent simplicity in the fluorescent detection of Hg2 þ with fluorescence quenching mechanism, which cannot rule out the interference from the detection medium. The nature can provide a variety of carbon sources that can be directly prepared into F-CNPs. Sun's group has recently proposed a microwave-assisted method for the rapid synthesis of F-CNPs using flour as carbon source. The resulting F-CNPs show good sensitivity towards Hg2 þ with the LOD as low as 0.5 nM. The excellent result is due to the strong interaction between Hg2 þ and the carboxyl groups on F-CNPs surface, which results in

Y. Guo et al. / Biosensors and Bioelectronics 63 (2015) 61–71

fluorescence quenching through electron or energy transfer (Qin et al., 2013). The rhizome of the plant Giant Knotweed Rhizome, a traditional Chinese medicine, is used as carbon source for synthesizing F-CNPs through hydrothermal method. The as-prepared F-CNPs show blue emission and the QY is about 11.5%. Hg2 þ can

65

effectively quench the fluorescence of F-CNPs and a broad linear range from 50 nM to 100 μM is obtained. The LOD is 8.2 nM. The Hg2 þ -induced fluorescence quenching is attributed to the energy transfer between Hg2 þ and F-CNPs via a chelation reaction mechanism (Wu et al., 2013). Huang et al. synthesized nitrogendoped F-CNPs with blue emission and a QY of 6.3% via the one-pot hydrothermal treatment of strawberry juice. The fluorescence of F-CNPs can be specifically quenched by Hg2 þ (Fig. 3A). Thus, the F-CNPs can be used as Hg2 þ sensor with a linear range from 10 nM to 50 μM and a LOD of 3 nM. The fluorescence turn-off mechanism is attributed to the fact that the F-CNPs-Hg2 þ complexes facilitate charge transfer and restrain the radiative recombination of excitons (Huang et al., 2013). Sweet potatoes can also be used as carbon source to prepare F-CNPs with size of 1–3 nm via a hydrothermal method. The as-synthesized F-CNPs can be used as an effective sensor for Hg2 þ because of the Hg2 þ -induced fluorescence quenching caused by the electron or energy transfer. And the LOD is 1 nM (Lu et al., 2013). Sun's group synthesized F-CNPs with a QY of 6.9% through hydrothermal treatment of pomelo peel. Hg2 þ can specifically quench the fluorescence of the F-CNPs due to the electron or energy transfer caused by the strong interaction between Hg2 þ and the carboxyl groups on F-CNPs surface. Thus, the F-CNPs can be used as a probe for selectively sensing Hg2 þ with a LOD of 0.23 nM (Fig. 3B). The feasibility of F-CNPs is also demonstrated with lake water samples (Lu et al., 2012). With the rapid development of research, we believe that more F-CNPs-based nanosensor for Hg2 þ will be synthesized and the performance of these nanosensors will be much improved. In order to compare the performance of those F-CNPs-based nanosensors for Hg2 þ , we summarized a variety of F-CNPs-based nanosensors for Hg2 þ (Table 2). 3.2. Copper ions

2þ

Fig. 3. (A) Synthesis of F-CNPs from strawberry juice and detection of Hg with F-CNPs. Reprinted with permission of The Royal Society of Chemistry from reference Huang et al. (2013). (B) The fluorescent detection of Hg2 þ with F-CNPs prepared from pomelo peel. Reprinted with permission of American Chemical Society from reference Lu et al. (2012).

Copper is an essential trace element in the human body and Cu2 þ is commonly found in natural water (Gaggelli et al., 2006). It is, however, toxic at high concentration and can cause liver or kidney damage after long-term exposure of high levels of copper

Table 2 List of F-CNPs for the fluorescent detection of Hg2 þ . Probes PEG200, NAC-F-CNPs PEG200, NAC-F-CNPs Graphitized F-CNPs CuDTC2-CDsf

e

F-CNPs, CdSe@ZnS QDsg F-CNPs Graphitic carbon nitride QDs F-CNPs F-CNPs F-CNPs Nitrogen-doped F-CNPs F-CNPs F-CNPs Nitrogen-doped F-CNPs F-CNPs F-CNPs

a

QYa

Read out

LODb

LRc

RSd

Reference

– – 25% –

Turn-off Turn-off Turn-off Turn-on

– 0.01 μM 69.9 nM 20 nM

0.1–2.69 μM – 0–1.00 μM –

Gonçalves et al., 2010 Gonçalves et al., 2012 Liu et al., 2012c Yuan et al., 2014

5% – 29%

Ratiometric 0.1 μM Turn-off 1 fM Turn-off 1 nM

– 1 fM–10 nM –

– – Tap water Tap water, lake water, human urine sample Tap water, lake water River water, lake water, tap water –

11.0% 68.22% 65.5%, 55.4% 15.7% 5.4% 11.5% 6.3% 2.8% 6.9%

Turn-off Turn-off Turn-off Turn-off Turn-off Turn-off Turn-off Turn-off Turn-off

0–3 μM 0–5 μM 1–12 μM, 1–15 μM 0–25 μM 0.5–10 nM 50 nM–100.0 μM 10 nM–50 μM 0.01–0.12 μM 0.5–10 nM, 500– 40000 nM

4.2 nM 10 nM 226 nM, 845 nM 0.23 μM 0.5 nM 8.2 nM 3 nM 1 nM 0.23 nM

QY: Quantum yield. LOD: Limit of detection. LR: Linear range. d RS: Real sample. e PEG200: Polyethylene glycol 200; NAC: N-acetyl-L-cysteine. f CuDTC2-CDs: Bis(dithiocarbamato)copper(II) complex functionalized carbon nanodots. g QDs: Quantum dots. b c

– Tap water, mineral water River water, mineral water Tap water, lake water Lake water River water Lake water Lake water Lake water

Cao et al., 2013 Liu et al., 2013 Barman and Sadhukhan, 2012 Zhou et al., 2012 Guo et al., 2013 Yan et al., 2014 Zhang and Chen, 2014 Qin et al., 2013 Wu et al., 2013 Huang et al., 2013 Lu et al., 2013 Lu et al., 2012

66

Y. Guo et al. / Biosensors and Bioelectronics 63 (2015) 61–71

(Georgopoulos et al., 2001). Moreover, many brain diseases, such as Alzheimer's, Parkinson's disease, are related to copper (Gaggelli et al., 2006; Millhauser, 2004). Therefore, detection of copper is of considerable importance for human health. Among many methods for the detection of Cu2 þ , fluorescent technique is popular because

Fig. 4. (A) The construction of a ratiometric fluorescent nanosensor for Cu2 þ with F-CNPs and CdSe/ZnS QDs. Reprinted with permission of John Wiley & Sons from reference Zhu et al. (2012). (B) Synthesis of F-CNPs from GO and the detection of Cu2 þ with F-CNPs. Reprinted with permission of John Wiley & Sons from reference Sun et al. (2013).

of its simplicity and high sensitivity. The past few years has witnessed a rapid development of F-CNPs-based nanosensors for Cu2 þ . Many F-CNPs with surface modification have been applied to detect Cu2 þ . Researchers have found that lysine and bovine serum albumin modified F-CNPs can assay Cu2 þ with the linear range of 0.002–1.5 nM and the LOD of 0.58 pM. The outstanding result is attributed to the strong coordination of Cu2 þ with lysine. The practicality of the method is also tested with human hair and tap water (Liu et al., 2012a). To further enhance the selectivity to Cu2 þ , N-(2-aminoethyl)-N,N,N′-tris(pyridin-2-ylmethyl) ethane-1,2-diamine (AE-TPEA) is chemically functionalized on the surface of F-CNPs. It is found that Cu2 þ can selectively quench the fluorescence and the LOD is 10 nM, because TPEA can specifically bind with Cu2 þ . Simultaneously, the modified F-CNPs can also be applied for intracellular sensing and imaging of Cu2 þ due to the low cytotoxicity and good cell-permeability (Qu et al., 2012a). To avoid the interference from the detection medium, a ratiometric fluorescent nanosensor for Cu2 þ has been constructed with the hybrid of CdSe/ZnS QDs and F-CNPs. CdSe/ZnS QDs are embedded in silica shells and their red emission serves as reference signals because the embedded CdSe/ZnS QDs are inert to Cu2 þ . The blueemitting F-CNPs are assembled on the surface of silica nanoparticles and then AE-TPEA is further chemically functionalized onto the surface of the nanohybrid. The system emits two bands centered at 485 and 644 nm, respectively, under a single excitation wavelength of 400 nm. It is found that Cu2 þ can selectively quench the fluorescence of F-CNPs (λem ¼485 nm) due to the strong interaction between Cu2 þ and TPEA, whereas the fluorescence of red CdSe@SiO2 (λem ¼ 644 nm) remains constant (Fig. 4A).

Fig. 5. (A) The detection mechanism of the g-C3N4 nanosheets-based sensor for Cu2 þ . Reprinted with permission of American Chemical Society from reference Tian et al. (2013). (B) The concept and process for detecting Cu2 þ based on F-CNPs and MOFs. Reprinted with permission of American Chemical Society from reference Lin et al. (2014).

Y. Guo et al. / Biosensors and Bioelectronics 63 (2015) 61–71

The intensity ratio of the two fluorescence peaks (I485/I644) can be used to assay Cu2 þ and the LOD is about 1 μM. Moreover, the ratiometric fluorescent nanosensor is successfully applied to image and sense Cu2 þ in living cells (Zhu et al., 2012). Again, the disturbance from the detection medium can be effectively eliminated, and thus the accuracy of the results is further improved. But the sensitivity needs to be further improved for practical applications. Amino-functionalized GQDs with a QY of 16.4% is prepared through the hydrothermal treatment of the greenish-yellow fluorescent GQDs with low QY in ammonia. The as-made F-CNPs can detect Cu2 þ due to the fact that Cu2 þ has a higher binding affinity and faster chelating kinetics with N and O on the surface of F-CNPs than other metal ions (Fig. 4B). A linear range from 0 to 100 nM and a LOD is 6.9 nM is obtained (Sun et al., 2013). Recently, many as-prepared F-CNPs without tedious surface modification have been directly used to detect Cu2 þ . Fluorescent cubic mesoporous graphitic carbon (IV) nitride (c-mpg-C3N4) can be prepared from a nanocasting approach using the ordered mesoporous silica KIT-6 as a hard template. It is found that Cu2 þ can efficiently quench the fluorescence of c-mpg-C3N4. The fluorescence turn-off mechanism is due to the photoinduced electron transfer from Cu2 þ to c-mpg-C3N4. And the LOD is as low as 12.336 nM (Lee et al., 2010). Tian et al. have developed an efficient fluorescent sensor for Cu2 þ with ultrathin g-C3N4 nanosheets comprised of only about three C–N layers, which is fabricated by ultrasonication-assisted liquid exfoliation of bulk C3N4. They found that the fluorescence of the ultrathin g-C3N4 nanosheets could be selectively quenched by Cu2 þ (Fig. 5A). The fluorescence quenched by Cu2 þ may be attributed to the close proximity of g-C3N4 nanosheets caused by the chelation of Cu2 þ with N of the g-C3N4 nanosheets. And the photoinduced electron transfer from the conduction band (CB) to the complexed Cu2 þ may also contribute to the fluorescence quenching because of the redox potential of Cu2 þ /Cu þ lying between the CB and valence band of g-C3N4. Thus the as-prepared F-CNPs can be used as a sensor for Cu2 þ and the LOD is as low as 0.5 nM. More interestingly, the use of test paper enables the naked-eye detection of Cu2 þ with a LOD

67

of 0.1 nmol under UV light (Tian et al., 2013). Polyamine-functionalized F-CNPs are synthesized through a simple one-pot method. Because branched poly(ethylenimine) (PEI) on the surface of F-CNPs can specifically recognize Cu2 þ , resulting in the effective fluorescence quenching. The F-CNPs can be used to assay Cu2 þ with the LOD as low as 6 nM and a linear range from 10 to 1100 nM (Dong et al., 2012). The polyamine-functionalized F-CNPs can also be encapsulated in metal-organic frameworks (MOFs) to form fluorescent MOFs, which can sense Cu2 þ through fluorescence turn-off mechanism and selectively accumulate Cu2 þ due to the adsorption of MOFs. The selective accumulation effect of MOFs can greatly amplify the sensing signal and selectivity of the fluorescent probe (Fig. 5B). The obtained F-CNPs/MOFs composites have been used to detect Cu2 þ with ultra-high sensitivity and high selectivity, and the LOD is as low as 80 pM. Most importantly, the composites have been successfully applied to assay Cu2 þ in environmental water samples (Lin et al., 2014). The combination of F-CNPs and MOFs greatly improves the specificity and sensitivity, which opens new avenues for F-CNPs in the fluorescent detection of other targets. The up-conversion fluorescent materials have plenty of advantages in biological applications, such as noninvasiveness, improved tissue penetration depth under NIR radiation and absence of autofluorescence in biological tissues (Li and Zhang, 2008; Li et al., 2008). Salinas-Castillo et al. have recently discovered that F-CNPs with down- and up-conversion fluorescent properties. They synthesized water soluble F-CNPs with downand up-conversion fluorescence via microwave-assisted pyrolysis of citric acid in the presence of PEI. The F-CNPs can be used to detect Cu2 þ through fluorescence turn-off mechanism caused by the inner filter effect. The LOD is 0.09 and 0.12 μM for UV and NIR excitation wavelength, respectively. The F-CNPs have been successfully used for sensing and imaging Cu2 þ in living cells (Salinas-Castillo et al., 2013). But the sensitivity needs to be further improved. F-CNPs can be prepared using mesoporous silica spheres as nanoreactors and citric acid as carbon source at high temperature. The as-synthesized F-CNPs can detect Cu2 þ due to the fact that Cu2 þ can efficiently quench the fluorescence of

Table 3 List of F-CNPs for the fluorescent detection of Cu2 þ . Probes e

Lys-BSA-F-CNPs TPEA-F-CNPsf F-CNPs, CdSe/ZnS QDsg Amino-GQDsh c-mpg-C3N4i g-C3N4 BPEI-CQDsj BPEI-CQDs, MOFsk PEI-Cdotsl F-CNPs GQDs F-CNPs F-CNPs F-CNPs F-CNPs a

QYa

Read out

LODb

LRc

RSd

References

– 10% 10% 16.4% – – 42.5% 42.5% 30% 7% – – 19.76% 5.1% 3.2%

Turn-off Turn-off Ratiometric Turn-off Turn-off Turn-off Turn-off Turn-off Turn-off Turn-off Turn-off Turn-off Turn-off Turn-off Turn-off

0.58 pM 10 nM 1 μM 6.9 nM 12.336 nM 0.5 nM 6 nM 80 pM 0.09 μM, 0.12 μM 23 nM 0.226 μM – 5 nM 1 nM 0.01 μM

0.002–1.5 nM 1–100 μM 1–100 μM 0–100 nM 10–100 nM – 10–1100 nM 2–1000 nM 0.3–1.6 μM 0–10 μM 0–15 μM 10–90 μM – – –

Human hair, tap water Cells Cells Cells – Lake water River water River water Cells – Synthetic water samples – – Lake water –

Liu et al., 2012a Qu et al., 2012a Zhu et al., 2012 Sun et al., 2013 Lee et al., 2010 Tian et al., 2013 Dong et al., 2012 Lin et al., 2014 Salinas-Castillo et al., 2013 Zong et al., 2014 Wang et al., 2014 Zhang et al., 2014a Zhao et al., 2014 Liu et al., 2012b Sha et al., 2013

QY: Quantum yield. LOD: Limit of detection. c LR: Linear range. d RS: Real sample. e Lys: Lysine/BSA: Bovine serum albumin. f TPEA: N-(2-aminoethyl)-N,N,N′-tris(pyridin-2-ylmethyl) ethane-1,2-diamine. g QDs: Quantum dots. h GQDs: Graphene quantum dots. i c-mpg-C3N4: Mesoporous graphitic carbon(IV) nitride. j BPEI-CQDs: Branched poly(ethylenimine)-functionalized carbon quantum dots. k MOFs: Metal-organic frameworks. l PEI-F-CNPs: Poly(ethylenimine)-functionalized carbon quantum dots. b

68

Y. Guo et al. / Biosensors and Bioelectronics 63 (2015) 61–71

F-CNPs via the charge transfer process, and the LOD is as low as 23 nM (Zong et al., 2014). Recently, reoxidized graphene oxide (GO) (Wang et al., 2014), activated carbon (Zhang et al., 2014a), ionic liquids (Zhao et al., 2014) have been used as carbon source to synthesize F-CNPs for the fluorescent detection of Cu2 þ . Biomass has been widely used as carbon source for the synthesis of F-CNPs to assay Cu2 þ . Hydrothermal treatment of grass has been used to synthesize nitrogen-doped F-CNPs for the fluorescent detection of Cu2 þ through fluorescence turn-off mechanism caused by the electron or energy transfer. The LOD is 1 nM (Liu et al., 2012b). However, much work is needed to improve the selectivity because Fe3 þ is a strong competitive cation to Cu2 þ . Pipe tobacco has also been proved to be an effective carbon source for nitrogen-containing F-CNPs prepared via the hydrothermal method. The resulting F-CNPs exhibit blue emission and can be utilized to sense Cu2 þ based on the Cu2 þ -induced fluorescence quenching caused by the electron or energy transfer. The LOD is 0.01 μM (Sha et al., 2013). The above mentioned F-CNPs show good response to Cu2 þ and have great potential applications in monitoring the level of Cu2 þ . However, most of F-CNPs for the fluorescent detection of Cu2 þ are based on fluorescence turn-off mechanism, which is easily affected by the detection medium. The detailed comparison of those F-CNPs-based probes for Cu2 þ is listed in Table 3. It is thus urgent to develop novel F-CNPs-based nanosensors for Cu2 þ . 3.3. Iron ions Fe3 þ is a biologically important metal ion and plays significant roles in oxygen uptake, oxygen metabolism, and electron transfer (Lynch, 1997; Touati, 2000). Fe3 þ is indispensable for most organisms, and both its deficiency and overload can induce various biological disorders (Andrews, 1999; Eisenstein, 2000). For example, Fe3 þ deficiency leads to anemia and excess iron in the body causes liver and kidney damage (Qin et al., 2008). Thus, sensitive detection of Fe3 þ is very important for human health. Currently, many F-CNPs have been used to assay Fe3 þ via fluorescence turnoff mechanism. A kind of F-CNPs has been prepared through hydrothermal treatment of thick syrup from citric acid and OHCH2 CH2OCH2CH2NH2 at 250 °C. It is found that Fe3 þ can specifically quench the fluorescence of the as-synthesized F-CNPs because of the ion selective chemical structure of F-CNPs and the inner filter effect. Thus, the F-CNPs can be used to detect Fe3 þ with a LOD of 11.2 μM (Sun et al., 2010). F-CNPs can also be prepared through one-pot oxidation of GO with periodic acid. The fluorescence of F-CNPs can be selectively quenched by Fe3 þ due to the strong interaction between Fe3 þ and phenolic hydroxyl groups. Thus, the

F-CNPs can be used as a fluorescent probe for Fe3 þ with a LOD of 17.5 μM (Wang et al., 2012). The electrochemical ablation of graphite electrodes has been utilized to make F-CNPs with graphitic crystallinity and oxygen-containing groups on the surface. The presence of oxygen-containing functional groups (e.g., hydroxyl, carboxyl) enables the F-CNPs to selectively respond to Fe3 þ via fluorescence turn-off mechanism resulted from the charge transfer and restrained exciton recombination. The LOD is as low as 2 nM due to the formation of complexes between Fe3 þ and the phenolic hydroxyls of F-CNPs (Zhang et al., 2013b). A ratiometric fluorescent nanosensor for Fe3 þ has been constructed based on water soluble F-CNPs prepared from the solution of citric acid and urea treated by the microwave method. The as-synthesized F-CNPs have two well-resolved fluorescence peaks at 455 and 520 nm under the excitation wavelength of 380 nm. It has been found that Fe3 þ can selectively quench the fluorescence at 455 nm, whereas the fluorescence at 520 nm remains constant. The Fe3 þ -induced fluorescence quenching may be ascribed to the electron or energy transfer process occurring between the excited F-CNPs and Fe3 þ . Based on the variation of the ratios from the fluorescent intensity centered at 455 and 520 nm, more than 0.04 μM Fe3 þ can be successfully detected (Qu et al., 2013b). Recently, Zhang's group has found that butylamine-modified GO (GO-C3Me) exhibits efficient fluorescence which can be specifically quenched by Fe3 þ through an electron transfer mechanism. Thus, the GO-C3Me can be utilized as a fluorescent probe for Fe3 þ with the LOD of 0.64 μM. Interestingly, Fe2 þ exhibits significant fluorescence quenching in the presence of H2O2 due to the dual quenching by the resulting Fe3 þ and hydroxyl radical evolved in the Fenton reaction of Fe2 þ with H2O2. According to these results, a three-input combinational logic gate is designed. The designed logic gate can be used to identify Fe3 þ , Fe2 þ , and their mixture in living cells (Fig. 6A) (Mei et al., 2012). Dopamine can be used as carbon source for the synthesis of F-CNPs with size about 3.8 nm through a hydrothermal method. The fluorescence of the asprepared F-CNPs can be selectively quenched by Fe3 þ due to the specific interaction between catechol groups and Fe3 þ (Fig. 6B). Thus, the F-CNPs can be used to detect Fe3 þ with a LOD of 0.32 μM (Qu et al., 2013a). Ionic liquids (1-butyl-3-methyl imidazole tetrafluoroborate) has also been used as carbon source to prepare F-CNPs for sensing Fe3 þ with the LOD of 20 nM through fluorescence turn-off mechanism caused by electron or energy transfer (Zhao et al., 2014). Pyrolysis of various plant leaves can yield bright F-CNPs with blue emission and the fluorescence intensity can be further improved through plasma and microwave-assisted techniques. The obtained F-CNPs have been applied to detect Fe3 þ with good sensitivity and selectivity based on fluorescence turn-off mechanism. There is a good linear correlation over a range

Fig. 6. (A) Schematic representation of F-CNPs logic gates for the discrimination of Fe3 þ and Fe2 þ in living cells. Reprinted with permission of The Royal Society of Chemistry from reference Mei et al. (2012). (B) Schematic illustration of F-CNPs for sensing Fe3 þ and dopamine. Reprinted with permission of John Wiley & Sons from reference Qu et al. (2013a).

Y. Guo et al. / Biosensors and Bioelectronics 63 (2015) 61–71

69

Table 4 List of F-CNPs for the fluorescent detection of Fe3 þ . Probes

QYa

Read out

LODb

LRc

RSd

References

GO nanosheetse GCQDsf F-CNPs GO-C3Meg F-CNPs F-CNPs F-CNPs F-CNPs

0.7% – –

Turn-off Turn-off Turn-off Turn-off Turn-off Turn-off Turn-off Turn-off

17.5 μM 2 nM 0.04 μM 0.64 μM 0.32 μM 20 nM – 0.025 μM

14.3–143.2 μM 0–1 μM – 5–50 μM 0–20 μM – 0–100 μM 1.0–5.0 μM, 5.0–50 μM

– – Lake water Cells Lake water, tap water – – Lake Water

Wang et al., 2012 Zhang et al., 2013b Qu et al., 2013b Mei et al., 2012 Qu et al., 2013a Zhao et al., 2014 Zhu et al., 2013 Xu et al., 2014

6.4% 25.80% 16.4% 15%

a

QY: Quantum yield. LOD: Limit of detection. LR: Linear range. d RS: Real sample. e GO: Graphene oxide. f GCQDs: Graphitic carbon quantum dots. g GO-C3Me: Butylamine-modified GO. b c

of 0–100 μM of Fe3 þ (Zhu et al., 2013). Potato has been hydrothermally treated to synthesize F-CNPs with a QY of 15%. The fluorescence of F-CNPs can be strongly quenched by Fe3 þ due to the strong affinity of Fe3 þ toward F-CNPs. Therefore, the F-CNPs can be utilized to sense Fe3 þ and it is found that the LOD is 0.025 μM and the linear ranges are both 1.0–5.0 μM and 5.0–50 μM (Xu et al., 2014). These F-CNPs can be utilized to assay Fe3 þ in aqueous solution, but much work is needed to improve the sensitivity of these F-CNPs-based nanosensors. Table 4 gives a brief comparison of these F-CNPs-based nanosensors for Fe3 þ . 3.4. Other metal ions F-CNPs have also been widely used to detect other metal ions, such as Pb2 þ (Wee et al., 2013; Qi et al., 2013), Ag þ (Ran et al., 2013), Cr6 þ (Zheng et al., 2013), Mn2 þ (Wang and Zhang, 2014),

K þ (Wei et al., 2012) and Sn2 þ (Yazid et al., 2013). The direct acid hydrolysis of bovine serum albumin has been utilized to prepare F-CNPs for sensing Pb2 þ via fluorescence turn-off mechanism, which is caused by the recombinant of excited electrons in the conduction band to the hole in the valence band. The sensing probe has a dynamic range up to 6.0 mM and the LOD is 5.05 μM (Wee et al., 2013). The detection of cerebral Pb2 þ can be realized using 3.9-dithia-6-monoazaundecane functionalized GQDs (DMAGQDs) and tryptophan. Pb2 þ acting as a cross-linker promotes the formation of a rigid structure between tryptophan and DMAGQDs, which can result in the fluorescence enhancement of the DMA-GQDs-tryptophan system because of the strong energytransfer interactions between tryptophan and DMA-GQDs. The system exhibits good selectivity to Pb2 þ and a linear range of 0.01–1 nM is obtained. The LOD is 9 pM and it is successfully used to monitor Pb2 þ in the striatum of rat (Qi et al., 2013). Ran et al. have described a novel method for the detection of Ag þ and

Fig. 7. (A) Illustration of fluorescent detection of Ag þ and biothiols using F-CNPs. Reprinted with permission of The Royal Society of Chemistry from reference Ran et al. (2013). (B) Schematic representation of the FRET model based on the F-CNPs-graphene hybrids and the mechanism of sensing K þ . Reprinted with permission of The Royal Society of Chemistry from reference Wei et al. (2012).

70

Y. Guo et al. / Biosensors and Bioelectronics 63 (2015) 61–71

Table 5 List of F-CNPs for the fluorescent detection of other metal ions. Ions

Probes

QYa

Read out

LODb

LRc

RSd

References

Pb2 þ Pb2 þ Ag þ CrO42  Mn2 þ Kþ Sn2 þ

F-CNPs DMA-GQDe, tryptophan GQDs F-CNPs F-CNPs Am-CDsf, 18C6E-rGOg F-CNPs

– – – 88.6% 5% – –

Turn-off Turn-on Turn-off Turn-off Turn-off turn-on Turn-off

5.05 μM 9 pM 3.5 nM – – 10 μM 0.36 μM

0–6.0 mM 0.01–1 nM 0–100 nM 0.01–50 μM 2–10 μM 0.05–10 mM 0–4 mM

– Rat brain microdialysates – – – Extracellular medium mimics –

Wee et al., 2013 Qi et al., 2013 Ran et al., 2013 Zheng et al., 2013 Wang and Zhang, 2014 Wei et al., 2012 Yazid et al., 2013

a

QY: Quantum yield. LOD: Limit of detection. LR: Linear range. d RS: Real sample. e DMA-GQDs: 3,9-dithia-6-monoazaundecane functionalized grapheme quantum dots. f Am-CDs: Aminated carbon dots. g 18C6E-rGO: 18-crown-6 ether-reduced graphene oxide. b c

biothiols by utilization of silver nanoparticles (AgNPs)-decorated F-CNPs prepared from GO. The addition of Ag þ can lead to the quench of the fluorescence due to the formation of AgNPs/F-CNPs hybrids. The LOD of Ag þ is 3.5 nM (Fig. 7A) (Ran et al., 2013). A fluorescence turn-off probe for Cr6 þ based on the inner filter effect has been reported by using the overlap of the absorption bands of Cr6 þ with the emission and excitation bands of F-CNPs. The linear range is 0.01–50 μM (Zheng et al., 2013). F-CNPs prepared from the oxidation of diesel soot have been utilized as a fluorescence turnoff probe for Mn2 þ with linear range of 2–10 μM (Wang and Zhang, 2014). The hybrids of crown ether modified GO and aminated F-CNPs have been constructed a fluorescence turn-on probe for K þ based on the FRET from F-CNPs to GO. In the absence of K þ , the fluorescence of F-CNPs is quenched by GO due to FRET from F-CNPs to GO. Whereas, the addition of K þ leads to the fluorescence recovery of F-CNPs because of the interaction of K þ with crown ether on the surface of GO and the release of F-CNPs from the surface of GO (Fig. 7B). The hybrids show a linear range from 0.05 to 10.0 mM and the LOD is about 10 μM (Wei et al., 2012). The study opens new opportunities for the construction of other F-CNPs-based nanosensors with FRET response. The carbonization and surface oxidation of preformed sago starch nanoparticles have been used to prepare F-CNPs for sensing Sn2 þ via fluorescence turn-off mechanism caused by the electron transfer. And the LOD is 0.36 μM (Yazid et al., 2013). In order to compare the performance of those F-CNPs for the fluorescent detection of other metal ions, we present a brief comparison between them in Table 5.

4. Conclusion and future perspectives In this review, we attempt to give a brief overview of recent advances of F-CNPs in the fluorescent detection of metal ions, including Hg2 þ , Cu2 þ , Fe3 þ and other metal ions. The studies described in the current review demonstrate that F-CNPs show great promise towards the detection of metal ions, but there remain several challenges to overcome before F-CNPs can be widely used in real samples. First, the fluorescence of F-CNPs is easily influenced by the detection medium, because the presence of minerals, organics and bacteria can easily cause the fluctuation of the fluorescence intensity. Many efforts are still required to improve the robustness of F-CNPs. Second, more F-CNPs-based nanosensors with fluorescence turn-on, FRET, ratiometric response and other novel response modes is much needed, because most of F-CNPs-based probes for metal ions are fluorescence turn-off, which is easily influenced by the detection environment. New detection strategies, such as ratiometric, turn-on and FRET, are still

needed because they can enhance the accuracy of the detection results. Third, the practical applications of F-CNPs in other fields needs further to be investigated because most of F-CNPs are only employed to assay metal ions in water samples. Last, but not least, most of F-CNPs are only used to assay metal ions. They cannot remove the detected toxic metal ions. The design of F-CNPs-based nanosensors with dual functions of detection and removal of metal ions is of great significance to the environment and human health. With the rapid development of F-CNPs, we really hope that more robust F-CNPs will be widely used in real samples for the fluorescent detection of metal ions for years to come.

Acknowledgments This work was supported by the Education Department of Henan Province (14B150022) and funded from Nanyang Normal University. The authors also thank Sha He for his valuable suggestions.

References Ali, I., Aboul-Enein, H.Y., 2002. Anal. Lett. 35, 2053–2076. Andrews, N.C., 1999. N. Engl. J. Med. 341, 1986–1995. Aragay, G., Merkoçi, A., 2012. Electrochim. Acta 84, 49–61. Badarau, A., Dennison, C., 2011. J. Am. Chem. Soc. 133, 2983–2988. Baker, S.N., Baker, G.A., 2010. Angew. Chem. Int. Ed. 49, 6726–6744. Barman, S., Sadhukhan, M., 2012. J. Mater. Chem. 22, 21832–21837. Benoit, J., Fitzgerald, W., Damman, A., 1998. Environ. Res. 78, 118–133. Cao, B., Yuan, C., Liu, B., Jiang, C., Guan, G., Han, M.-Y., 2013. Anal. Chim. Acta 786, 146–152. Clarkson, T.W., Magos, L., Myers, G.J., 2003. N. Engl. J. Med. 349, 1731–1737. Dong, Y., Wang, R., Li, G., Chen, C., Chi, Y., Chen, G., 2012. Anal. Chem. 84, 6220– 6224. Eisenstein, R.S., 2000. Annu. Rev. Nutr. 20, 627–662. Esteves da Silva, J.C., Gonçalves, H.M., 2011. TRAC, Trends Anal. Chem. 30, 1327– 1336. Fang, Z., Růžička, J., Hansen, E., 1984. Anal. Chim. Acta 164, 23–39. Gaggelli, E., Kozlowski, H., Valensin, D., Valensin, G., 2006. Chem. Rev. 106, 1995– 2044. Ge, Y., Guo, Y., Qin, W., 2014. Talanta 121, 50–55. Georgopoulos, P.G., Roy, A., Yonone-Lioy, M.J., Opiekun, R.E., Lioy, P.J., 2001. J. Toxicol. Environ. Health B: Crit. Rev. 4, 341–394. Giakisikli, G., Miró, M., Anthemidis, A., 2013. Anal. Chem. 85, 8968–8972. Gonçalves, H., Jorge, P.A., Fernandes, J., Esteves da Silva, J.C., 2010. Sens. Actuators B 145, 702–707. Gonçalves, H.M., Duarte, A.J., Davis, F., Higson, S.P., Esteves da Silva, J.C., 2012. Anal. Chim. Acta 735, 90–95. Guo, Y., Wang, Z., Shao, H., Jiang, X., 2013. Carbon 52, 583–589. Harris, H.H., Pickering, I.J., George, G.N., 2003. Science 301, 1203. Huang, H., Lv, J.-J., Zhou, D.-L., Bao, N., Xu, Y., Wang, A.-J., Feng, J.-J., 2013. RSC Adv. 3, 21691–21696. Lee, E.Z., Jun, Y.S., Hong, W.H., Thomas, A., Jin, M.M., 2010. Angew. Chem. Int. Ed. 49, 9706–9710.

Y. Guo et al. / Biosensors and Bioelectronics 63 (2015) 61–71

Leopold, K., Foulkes, M., Worsfold, P., 2010. Anal. Chim. Acta 663, 127–138. Li, H., Kang, Z., Liu, Y., Lee, S.-T., 2012. J. Mater. Chem. 22, 24230–24253. Li, Z., Zhang, Y., 2008. Nanotechnology 19, 345606. Li, Z., Zhang, Y., Jiang, S., 2008. Adv. Mater. 20, 4765–4769. Lin, X., Gao, G., Zheng, L., Chi, Y., Chen, G., 2014. Anal. Chem. 86, 1223–1228. Liu, J.-M., Lin, L.-P., Wang, X.-X., Lin, S.-Q., Cai, W.-L., Zhang, L.-H., Zheng, Z.-Y., 2012a. Analyst 137, 2637–2642. Liu, R., Li, H., Kong, W., Liu, J., Liu, Y., Tong, C., Zhang, X., Kang, Z., 2013. Mater. Res. Bull. 48, 2529–2534. Liu, S., Tian, J., Wang, L., Zhang, Y., Qin, X., Luo, Y., Asiri, A.M., Al-Youbi, A.O., Sun, X., 2012b. Adv. Mater. 24, 2037–2041. Liu, Y., Liu, C.-Y., Zhang, Z.-Y., 2012c. Appl. Surf. Sci. 263, 481–485. Lu, W., Qin, X., Asiri, A.M., Al-Youbi, A.O., Sun, X., 2013. J. Nanopart. Res. 15, 1–7. Lu, W., Qin, X., Liu, S., Chang, G., Zhang, Y., Luo, Y., Asiri, A.M., Al-Youbi, A.O., Sun, X., 2012. Anal. Chem. 84, 5351–5357. Lynch, S.R., 1997. Nutr. Rev. 55, 102–110. Ma, B., Xu, M., Zeng, F., Huang, L., Wu, S., 2011. Nanotechnology 22, 065501. Mei, Q., Jiang, C., Guan, G., Zhang, K., Liu, B., Liu, R., Zhang, Z., 2012. Chem. Commun. 48, 7468–7470. Millhauser, G.L., 2004. Acc. Chem. Res. 37, 79–85. Mutter, J., Naumann, J., Schneider, R., Walach, H., Haley, B., 2005. Neuroendocrinol. Lett. 26, 439–446. Nolan, E.M., Lippard, S.J., 2003. J. Am. Chem. Soc. 125, 14270–14271. Nolan, E.M., Lippard, S.J., 2008. Chem. Rev. 108, 3443–3480. Qi, Y.X., Zhang, M., Fu, Q.Q., Liu, R., Shi, G.Y., 2013. Chem. Commun. 49, 10599– 10601. Qin, C., Cheng, Y., Wang, L., Jing, X., Wang, F., 2008. Macromolecules 41, 7798–7804. Qin, X., Lu, W., Asiri, A.M., Al-Youbi, A.O., Sun, X., 2013. Sens. Actuators B 184, 156– 162. Qu, K., Wang, J., Ren, J., Qu, X., 2013a. Chem. Eur. J 19, 7243–7249. Qu, Q., Zhu, A., Shao, X., Shi, G., Tian, Y., 2012a. Chem. Commun. 48, 5473–5475. Qu, S., Chen, H., Zheng, X., Cao, J., Liu, X., 2013b. Nanoscale 5, 5514–5518. Qu, S., Wang, X., Lu, Q., Liu, X., Wang, L., 2012b. Angew. Chem. Int. Ed. 51, 12215– 12218. Ran, X., Sun, H., Pu, F., Ren, J., Qu, X., 2013. Chem. Commun. 49, 1079–1081. Renzoni, A., Zino, F., Franchi, E., 1998. Environ. Res. 77, 68–72. Salinas-Castillo, A., Ariza-Avidad, M., Pritz, C., Camprubí-Robles, M., Fernández, B., Ruedas-Rama, M.J., Megia-Fernández, A., Lapresta-Fernández, A., SantoyoGonzalez, F., Schrott-Fischer, A., 2013. Chem. Commun. 49, 1103–1105. Sha, Y., Lou, J., Bai, S., Wu, D., Liu, B., Ling, Y., 2013. Mater. Res. Bull. 48, 1728–1731. Shen, J., Zhu, Y., Yang, X., Li, C., 2012. Chem. Commun. 48, 3686–3699. Sun, H., Gao, N., Wu, L., Ren, J., Wei, W., Qu, X., 2013. Chem. Eur. J 19, 13362–13368.

71

Sun, W., Du, Y., Wang, Y., 2010. J. Lumin. 130, 1463–1469. Tian, J., Liu, Q., Asiri, A.M., Al-Youbi, A.O., Sun, X., 2013. Anal. Chem. 85, 5595–5599. Tian, M., Liu, L., Li, Y., Hu, R., Liu, T., Liu, H., Wang, S., Li, Y., 2014. Chem. Commun. 50, 2055–2057. Touati, D., 2000. Arch. Biochem. Biophys. 373, 1–6. Townsend, A.T., Miller, K.A., McLean, S., Aldous, S., 1998. J. Anal. At. Spectrom. 13, 1213–1219. Türker, A.R., Çabuk, D., Yalçınkaya, Ö., 2014. Anal. Lett. 46, 1155–1170. Wang, D., Wang, L., Dong, X., Shi, Z., Jin, J., 2012. Carbon 50, 2147–2154. Wang, F., Gu, Z., Lei, W., Wang, W., Xia, X., Hao, Q., 2014. Sens. Actuators B 190, 516– 522. Wang, Q., Zhang, S., 2014. J. Lumin. 146, 37–41. Wee, S.S., Ng, Y.H., Ng, S.M., 2013. Talanta 116, 71–76. Wei, W., Xu, C., Ren, J., Xu, B., Qu, X., 2012. Chem. Commun. 48, 1284–1286. Wu, D., Huang, X., Deng, X., Wang, K., Liu, Q., 2013. Anal. Methods 5, 3023–3027. Xu, J., Zhou, Y., Liu, S., Dong, M., Huang, C., 2014. Anal. Methods 6, 2086–2090. Xu, X., Ray, R., Gu, Y., Ploehn, H.J., Gearheart, L., Raker, K., Scrivens, W.A., 2004. J. Am. Chem. Soc. 126, 12736–12737. Xuan, F., Luo, X., Hsing, I.-M., 2013. Anal. Chem. 85, 4586–4593. Yan, F., Zou, Y., Wang, M., Mu, X., Yang, N., Chen, L., 2014. Sens. Actuators B 192, 488–495. Yang, F., Li, J., Lu, W., Wen, Y., Cai, X., You, J., Ma, J., Ding, Y., Chen, L., 2014. Electrophoresis 35, 474–481. Yazid, S.N.A.M., Chin, S.F., Pang, S.C., Ng, S.M., 2013. Microchim. Acta 180, 137–143. Yuan, C., Liu, B., Liu, F., Han, M.-Y., Zhang, Z., 2014. Anal. Chem. 86, 1123–1130. Zhang, R., Chen, W., 2014. Biosens. Bioelectron. 55, 83–90. Zhang, S., Wang, Q., Tian, G., Ge, H., 2014a. Mater. Lett. 115, 233–236. Zhang, T., Chan, C., Lan, R., Wong, W., Wong, K., 2014b. Chem. Eur. J 20, 970–973. Zhang, Y., Zhao, H., Wu, Z., Xue, Y., Zhang, X., He, Y., Li, X., Yuan, Z., 2013a. Biosens. Bioelectron. 48, 180–187. Zhang, Y.-L., Wang, L., Zhang, H.-C., Liu, Y., Wang, H.-Y., Kang, Z.-H., Lee, S.-T., 2013b. RSC Adv 3, 3733–3738. Zhao, A., Zhao, C., Li, M., Ren, J., Qu, X., 2014. Anal. Chim. Acta 809, 128–133. Zheng, M., Xie, Z., Qu, D., Li, D., Du, P., Jing, X., Sun, Z., 2013. ACS Appl. Mater. Interfaces 5, 13242–13247. Zheng, W., Aschner, M., Ghersi-Egea, J.-F., 2003. Toxicol. Appl. Pharmacol. 192, 1–11. Zhou, L., Lin, Y., Huang, Z., Ren, J., Qu, X., 2012. Chem. Commun. 48, 1147–1149. Zhu, A., Qu, Q., Shao, X., Kong, B., Tian, Y., 2012. Angew. Chem. Int. Ed. 51, 7185– 7189. Zhu, L., Yin, Y., Wang, C.-F., Chen, S., 2013. J. Mater. Chem. C 1, 4925–4932. Zong, J., Yang, X., Trinchi, A., Hardin, S., Cole, I., Zhu, Y., Li, C., Muster, T., Wei, G., 2014. Biosens. Bioelectron. 51, 330–335.