Recent Progresses in NIR-II Luminescent Bio/Chemo Sensors Based on Lanthanide Nanocrystals

Abstract

Fluorescent bio/chemosensors are widely used in the field of biological research and medical diagnosis, with the advantages of non-invasiveness, high sensitivity, and good selectivity. In particular, luminescent bio/chemosensors, based on lanthanide nanocrystals (LnNCs) with a second near-infrared (NIR-II) emission, have attracted much attention, owing to greater penetration depth, aside from the merits of narrow emission band, abundant emission lines, and long lifetimes. In this review, NIR-II LnNCs-based bio/chemo sensors are summarized from the perspectives of the mechanisms of NIR-II luminescence, synthesis method of LnNCs, strategy of luminescence enhancement, sensing mechanism, and targeted bio/chemo category. Finally, the problems that exist in present LnNCs-based bio/chemosensors are discussed, and the future development trend is prospected.

1. Introduction

Bio/chemical detection is of great significance in biological research and medical diagnosis [1,2]. So far, various detection methods have been developed, such as colorimetry [3], chromatography [4], electrochemical method [5], and fluorescence spectroscopy [6]. These methods have the disadvantages of complicated operation processes, low sensitivity, inability to perform real-time detection, and in vivo application compared with fluorescent spectroscopy. Fluorescent spectroscopy is a new type detection method, which has the advantages of high sensitivity, good selectivity, low detection limit, and high resolution [7,8]. Meanwhile, it may be applied to real-time monitoring of biological substances [9].

Fluorescent spectroscopy is divided into visible, NIR-I (700~900 nm), and NIR-II (1000~1700 nm), according to the different emission wavelengths [10,11,12,13]. Nevertheless, the autofluorescence generated (Figure 1a) [14], interference of scattering (Figure 1b) [14], and absorption (Figure 1c) [15] in the biological tissues severely limits the application of visible light and NIR-I in the biological body when the excitation and emission light pass through biological tissues [14,15]. In this wavelength range, especially, the scattering and autofluorescence of organisms are significant, forming a strong signal background and interference. To circumvent these restrictions, NIR-II has been emerging in recent years since it can dramatically reduce scattering lights (Figure 1d) [16] and increase penetration depth (Figure 1e) [17,18,19] in biological applications, compared with those emitting in the visible or the NIR-I. Therefore, the research on NIR-II luminescent materials has gradually increased in bioimaging and bio/chemical sensing due to the unique superiority of NIR-II luminescent materials (Figure 2).

The current study of NIR-II luminescent materials mainly includes the following: quantum dots (QDs) [20,21], single carbon nanotubes (SWCNTs) [22], organic fluorescent probe [23,24,25,26,27,28,29], and lanthanide nanocrystals (LnNCs) [30,31,32,33,34,35,36,37]. Most QDs contain heavy metal elements such as lead, mercury, or arsenic [20]. This potential biological toxicity severely limits the application of QDs in organisms. The low luminescence quantum yield (<0.1%) and poor biocompatibility of SWCNTs limit their application in NIR-II [22]. Organic probes generally face extremely low fluorescence quantum yield, poor water solubility, and are prone to photobleaching and fluorescence quenching, which greatly limit the application of organic fluorescent probes in vivo [23,24,25,26,27,28,29]. Meanwhile, the emission wavelength of organic probes is short, most of which are around 1000 nm, which severely limits its application in the second near-infrared region [27,29]. In contrast, lanthanide ions (Ln³⁺) have a unique 4f electronic structure [38,39,40]. The spectrum produced by the f-f transition has narrow-band emission characteristics and is not sensitive to the environment, and the emission only involves atomic transitions, so they have strong resistance to photobleaching [12,38,39,40]. Meanwhile, the Ln³⁺ can effectively protect the internal 4f orbitals from the interference of the external crystal field due to the 5s²5sp⁶ electron orbitals, make Ln³⁺ have a longer luminescence lifetime, and effectively avoid the interference of the autofluorescence of biological tissues [41,42]. Therefore, LnNCs, as a new generation of NIR-II probes, with low photobleaching, long luminescence lifetimes, low long-term cytotoxicity, narrow emission band widths, chemical stability, and large Stokes shifts, have gained more attention recently for bioimaging applications and bio/chemo sensors.

At present, the reviews on LnNCs mainly focus on discussing the application of LnNCs for in vivo imaging [12,38,39,40,41,42], and the discussion of LnNCs in bio/chemo sensing is only briefly mentioned as part of the imaging application. Therefore, a comprehensive and detailed review of NIR-II LnNCs in bio/chemo sensing is of great significance for researchers to thoroughly understand the research progress of LnNCs in bio/chemo sensing, and to provide ideas for the design of NIR-II LnNCs sensing probes. In this review, the latest research progress on NIR-II LnNCs bio/chemo sensors has been retrospected. Firstly, the design and synthesis of NIR-II LnNCs are discussed from a spectral point of view. Secondly, various application fields and sensing mechanisms of bio/chemo sensors are categorized and commented. Finally, the disadvantages of NIR-II LnNCs chemo/bio sensors and their prospects are given.

2. Design and Synthesis of NIR-II LnNCs

2.1. Mechanisms of NIR-II Luminescence

In LnNCs, three components are commonly included: a host matrix, a sensitizer, and an activator [36,37,38,39,40,41,42]. The host material should meet the requirements of optical transparency and low lattice phonon energy. At present, fluoride is mainly used as the host. The sensitizers and activators are generally Ln³⁺. Nd³⁺, Er³⁺, and Yb³⁺ are demonstrated to be efficient sensitizers due to their large absorption cross-sections in NIR-I or NIR-II regions (Figure 3a) [36]. The absorption peaks of Nd³⁺ are located at 730 nm and 808 nm; the absorption peaks of Er³⁺ are located at 808 nm, 980 nm, and 1550 nm; the absorption peak of Yb³⁺ is located at 980 nm.

The activators with NIR-II emission mainly include Pr³⁺, Nd³⁺, Ho³⁺, Er³⁺, and Tm³⁺ (Figure 3a,b) [36,41]. To overcome the weak light absorption problem of the activator ion itself, sensitizers with higher absorption coefficients are co-doped into the host, serving to harvest the excitation photons and transfer the excitation energy to activators, thus populating the radiative transition of activators for NIR-II luminescence. Therefore, NIR-II luminescence is the mechanism of LnNCs: firstly, the excited states of the sensitizers and activators split into different energy levels, in a ladder-like manner, in the crystal field maintained by the host matrix. The ground state electrons of the sensitizer are excited to the excited state under the excitation light of the appropriate wavelength. Then, the energy is transferred to the activator, and the activator is excited to the excited state. Finally, the excited state electrons return to the ground state and emit NIR-II luminescence (Figure 3b). According to NIR-II emission of different activators, it can be roughly divided into three types of NIR-II probes [36,41,43,44]: (1) Er³⁺ based NIR-II probes. NIR-II emission at intense 1525~1550 nm can be generated through ⁴I_13/2 → ⁴I_15/2 radiative transition. (2) Nd³⁺ based NIR-II probes. The radiative energy transitions of ⁴F_3/2 → ⁴I_11/2 and ⁴F_3/2 → ⁴I_13/2 in Nd³⁺ enable the generation of intense NIR-II luminescence, peaked at 1060 and 1330~1340 nm, respectively. (3) NIR-II probes of other ions. In a similar radiative transition, NIR-II emission can also be realized from Ho³⁺ at 1155 nm (⁵I₆ → ⁵I₈), Pr³⁺ at 1289 nm (¹G₄ → ³H₅), and Tm³⁺ at 1475 nm (³H₄ → ³F₄).

Figure 3. (a) Absorption (solid line) and NIR−II emission (solid line with color shade) spectra of different Ln³⁺. Adapted, with permission, from Ref. [36]. Copyright 2019, Wiley−VCH. (b) Energy level diagram and energy transfer of Ln³⁺ with NIR−II characteristic emission. Adapted, with permission, from Ref. [41]. Copyright 2019, Springer Nature. (c) NIR-II spectra of LnNCs with different contents of Ce³⁺ concentrations. Adapted, with permission, from Ref. [45]. Copyright 2019, American Chemical Society. (d) The energy transfer between Yb³⁺, Er³⁺, and Ce³⁺. Adapted, with permission, from Ref. [46]. Copyright 2017, Springer Nature. (e) NIR−II spectra of LnNCs with different Zn²⁺ concentrations. Adapted, with permission, from Ref. [47]. Copyright 2019, Springer Nature.

Figure 3. (a) Absorption (solid line) and NIR−II emission (solid line with color shade) spectra of different Ln³⁺. Adapted, with permission, from Ref. [36]. Copyright 2019, Wiley−VCH. (b) Energy level diagram and energy transfer of Ln³⁺ with NIR−II characteristic emission. Adapted, with permission, from Ref. [41]. Copyright 2019, Springer Nature. (c) NIR-II spectra of LnNCs with different contents of Ce³⁺ concentrations. Adapted, with permission, from Ref. [45]. Copyright 2019, American Chemical Society. (d) The energy transfer between Yb³⁺, Er³⁺, and Ce³⁺. Adapted, with permission, from Ref. [46]. Copyright 2017, Springer Nature. (e) NIR−II spectra of LnNCs with different Zn²⁺ concentrations. Adapted, with permission, from Ref. [47]. Copyright 2019, Springer Nature.

2.2. Strategies of Optimizing NIR-II Luminescence

The LnNCs with good luminescent efficiency can be obtained by rationally selecting the matrix, sensitizing ion, and activating ion. However, the luminescent efficiency of LnNCs is still low due to the inherent problems of LnNCs, which severely limits its application in sensing [38,39,40,41,42]. It is possible to design and optimize the structure of LnNCs to increase the quantum yield and enhance its luminescent intensity.

2.2.1. Ion Doping

It is possible to increase the luminescent intensity of a specific wavelength by adjusting the energy distribution of Ln³⁺, in different luminescent bands, because of the multi-emission characteristics of Ln³⁺. The up-conversion luminescence of LnNCs can be converted into down-conversion luminescence by doping other ions into LnNCs, and the NIR-II luminescence can be enhanced [45,46,47]. Li et al. [45] synthesized NaLnF₄: 40%Gd, 20%Yb, 2%Er, 5%Ce³⁺ nanorods (Ln = Y, Yb, Lu). The 1525 nm luminescence of Er³⁺ increased by 2.2 times, and the quantum efficiency reached 3.6% by doping Ce³⁺ to inhibit the up-conversion energy transfer of NaLnF₄ (Figure 3c). The incorporation of Ce³⁺ can promote the relaxation process of ⁴I_11/2 and ⁴I_13/2 of Er³⁺, as well as shorten the life of the ⁴I_11/2 state, thereby inhibiting the up-conversion process and enhancing Er³⁺ luminescence at 1525 nm (Figure 3d) [46]. However, high Ce³⁺ concentration (10%) will also cause concentration quenching and reduce the luminescence of NIR-II (Figure 3c).

In addition, the luminescent intensity of LnNCs involves 4f-4f electronic transition forbidden effect [46,47]. This prohibition can be broken by doping with other ions, resulting in luminescent enhancement of LnNCs. Zhong et al. [47] designed and synthesized a Zn-doped α-NaYbF₄: 2%Er, 2%Ce, 10%Zn@NaYF₄ NIR-II luminescent probe. The doping of Zn weakened the 4f-4f electronic transition forbidden effect, resulting in enhancing the NIR-II luminescence. When the doping concentration of Zn was 10%, the luminescence of NIR-II was 1.5 times higher than that of undoped α-ErNPs (Figure 3e).

2.2.2. Core-Shell Structure

Due to the small size of LnNCs, the excited state energy of its luminescent ions is easily transferred to the surface of the nanocrystals, through inter-ion energy transfer, and quenched by various surface defects, ligands, or solvent molecules. Therefore, coating different shell layers on mononuclear LnNCs can inhibit its surface quenching and cross-relaxation, thereby enhancing NIR-II luminescence.

At present, the main thing is to wrap an inert shell on the surface of a single core [48,49,50,51]. However, it should be noted that the lattice matching, between the inert shell layer and the core, is the main factor to be considered in the preparation. If the lattice parameter gap is too large, it is difficult to form a complete core-shell coating structure. Therefore, considering the factors of lattice matching, common inert shell layers include NaYF₄, NaLuF₄, and CaF₂, etc. [49]. The luminescent intensity of the NaYF₄: 10%Yb³⁺, 30%Nd³⁺@CaF₂ core-shell structure at 1000 nm is 45 times higher than that of single-core NaYF₄: 10%Yb³⁺, 30%Nd³⁺ under the excitation of 808 nm (Figure 4a) [48]. Likewise, Li et al. [49] completed a detailed investigation of the relationship between shell thickness and NIR-II emission intensity in the NaErF₄@NaYF₄ core/shell. As the inert shell path increases, the luminescence of Er³⁺ in NIR-II gradually increases (Figure 4b).

However, when the inert shell exceeds a certain thickness, it will weaken the light intensity received by the inner core, resulting in the weakening of the luminescent intensity. From Figure 4c, the luminescent intensity of LnNCs at 1525 nm was significantly reduced after NaYF₄: Yb³⁺, Er³⁺ coated inert shell NaLuF₄ with a thickness of 13.2 nm [50]. Sun et al. [51] reported a highly efficient NIR-II luminescent NaErF₄@NaYF₄@NaYF₄: 10%Nd@NaYF₄ nanocrystal. The strong NIR-II luminescence was successfully obtained by adjusting the thickness of the NaYF₄ buffer layer, to limit the energy transfer between Er³⁺ and Nd³⁺, and optimizing the thickness of the outermost NaYF₄ to passivate surface defects. Therefore, reasonable adjustment of the coating thickness of the inert shell layer and the position of the inert shell side can achieve high-efficiency luminescence in NIR-II.

In addition to wrapping the inert shell to achieve the enhancement of NIR-II luminescence, it can also wrap the active shell to enhance the NIR-II luminescence. The active shell layer refers to the presence of an activator or sensitizer in the shell layer, which mediates energy transfer and enhances the luminescence in the NIR-II [52]. Wang et al. [52] prepared β-NaGdF₄@Na (Gd, Yb) F₄: Er@NaYF₄: Yb@NaNdF₄: Yb. The Nd³⁺ in the outermost layer can be excited by an 800 nm laser, and the energy was transferred to Er³⁺ in the inner layer through Yb³⁺, resulting in an enhancement of its 1525 nm luminescence emission. Its 1525 nm luminescence penetrates the pork tissue to a depth of 18 mm.

2.2.3. Dye-Sensitized Luminescence

The absorption coefficient of Ln³⁺, in some specific wavelength bands, is very low, which greatly affects its luminescent intensity. The absorption coefficient of π-π transitions of some organic dyes is very high, has strong absorption in the spectrum 700~860 nm, and it can be used as a ligand to bind with Ln³⁺ [53,54,55,56]. The excited state energy level of the organic dye matches the excited state energy level of the Ln³⁺, which can produce the energy transfer process, so that the LnNCs can carry out efficient NIR-II luminescence emission (Figure 5a) [54].

Wang et al. [55] used indocyanine green dye (ICG) as an organic sensitizer to transfer energy to NaYF₄: Er nanoparticles to enhance the NIR-II luminescence (Figure 5b,c). ICG, as a donor with high absorption cross-section at 808 nm, increased the excitation efficiency of Er³⁺ through the energy transfer mechanism (Figure 5d). However, the NIR-II luminescent intensity decreased when the concentration of ICG was higher, which was ascribed to the self-quenching effect between the ICG (Figure 5b). The luminescence at 1525 nm was enhanced by 10 times at the optimal ICG concentration. Similarly, Ren et al. [56] modified LnNCs with dye-brush polymer (Dye-BP), which led to an impressive 675 times enhancement of NIR-II luminescence in aqueous solution. When the concentration of Dye-BP exceeds a certain concentration, it will also reduce the luminescence of NIR-II. Therefore, it is the key to choosing suitable organic dyes as sensitizers, but it is necessary to find the optimal concentration of organic dyes in the regulation of enhancing NIR-II luminescence to achieve the strongest NIR-II luminescence.

In addition to organic sensitizers, inorganic nanoparticles are also used as sensitizers to enhance the NIR-II luminescence of LnNCs due to their wide absorption and narrow emission and excellent resistance to photobleaching. Zhang et al. [57] introduced an inorganic nanoparticle sensitized system, which uses near-infrared-emitting Ag₂S QDs, as a sensitizer with broadband photon absorption, to enhance NIR-II luminescence. The NIR-II luminescence of LnNCs@Ag₂S enhanced ~17-fold in intensity and ~10-fold in brightness over bare LnNCs because of increased absorptivity and overall broadening of the LnNCs@Ag₂S absorption spectrum.

Except for the main methods mentioned above, it is also possible to enhance NIR-II luminescence by controlling the size and crystal form of synthetic LnNCs [47]. However, due to the inherent properties of LnNCs and the complexity of the external environment (-OH) during application, it will severely quench the luminescence in the NIR-II. Therefore, enhancing the NIR-II luminescence is a difficult point in research.

2.3. Preparation Methods

The grain size, crystal phase, morphology and shell of LnNCs are all key parameters that directly affect their luminescence performance. Researchers have been committed to developing various methods to synthesize LnNCs with different properties, such as thermal decomposition, hydrothermal synthesis, co-precipitation, and sol-gel process [42]. The most commonly used methods are thermal decomposition [58,59,60,61,62,63,64,65], hydrothermal synthesis [66,67,68,69,70,71] and co-precipitation [72,73,74,75,76,77,78,79,80,81].

2.3.1. Thermal Decomposition

Thermal decomposition is the most popular method to synthesize high-quality monodisperse LnNCs [58,59,60,61,62,63,64,65]. Thermal decomposition is a synthesis process in which organic precursors of metal ions are dissolved in a high-boiling organic solvent and then decomposed at high temperatures to obtain corresponding metal fluoride products.

The LnNCs with different crystal phases and sizes can be synthesized by controlling different reaction parameters. Suter et al. [59] studied the crystal phase process of synthesizing NaYF₄ in oleic acid (OA) and octadecene by thermal decomposition method. There were three stages in this thermal decomposition synthesis process. The first stage was the nucleation and growth of small α-NaYF₄ (<300 °C); the next stage was the growth of α-NaYF₄ (300 °C, 60 min); the third stage was the phase transition from small α-NaYF₄ to larger β-NaYF₄ (300~310 °C, 30 min). Li et al. [60] also proved that α-NaYF₄ can be synthesized at low temperature, while β-NaYF₄ can be synthesized at high temperature, and the longer the time, the higher the crystallinity and the larger the particles. The content of OA in the reaction mainly affects the aspect ratio of the particles, and a high ratio of OA can synthesize a rod-like structure [60]. NH₄F provides F⁻ ions in the process, which affects the nucleation rate of LnNCs, thus affecting the size of LnNCs, and the high NH₄F content leads to the formation of small particles [58,60].

The LnNCs with different crystal types, sizes, and core-shell structures can be prepared using this method, especially for synthesizing ultra-small LnNCs. However, the requirements for the operation during the reaction and the amount of reactants added are particularly high when synthesizing small-sized materials, especially when the whole reaction is in an inert atmosphere. If oxygen enters, the OA oxidation solution will turn black and cause the reaction to fail. Wang et al. [58] studied, in detail, the problems and solutions that easily occur in synthesizing LnNCs via thermal decomposition. This has important guiding significance for the efficient synthesis of LnNCs.

2.3.2. Hydrothermal Synthesis

The hydrothermal synthesis usually occurs in a closed environment under high temperature and pressure. The high solubility and reactivity of the reactants in the sealed reaction vessel, under high temperature and high pressure, are used to grow nanocrystals in aqueous solutions (such as water or organic solvents) [42,66,67,68,69,70,71]. The reaction temperature is relatively low (generally less than 200 °C). It has become a simple and effective method to synthesize monodisperse LnNCs with adjustable morphology and structure [67,69,71].

Wang et al. [67] studied the effects of reactant concentration, temperature, and time on LnNCs. High concentration (Ln³⁺) would result in nanorods, while low concentration resulted in nanoparticles. Higher temperature will decrease the length-to-diameter ratio of the as-prepared nanorods. Too short of a reaction time resulted in only nanoparticles instead of nanorods. The morphology and size of LnNCs can also be controlled by controlling the amount of F⁻. Qu et al. [69] used NH₄F as the F⁻ source to prepare NaYF₄ samples and found that different F⁻/Y³⁺ ratios have a significant impact on the morphology, crystal phase, and size of the nanocrystals. The high molar ratio of F⁻/Y³⁺ would benefit the α to β phase transition and formation of hexagonal phase products. The influence of different pH on the synthesis of LnNCs was also studied, NaYF₄ presented an octahedral morphology and a smooth surface (pH 1.5). When the pH was 6, NaYF₄ microtubules were formed [66].

The hydrothermal method can control the synthesis of LnNCs of different types of morphology, size, and crystal form, but the reaction rate of the system is slow, and it generally takes 20 h of reaction time to obtain monodisperse LnNCs with uniform size and morphology. The size is generally larger (sub-micron level). Meanwhile, it is difficult to observe the growth process of LnNCs. In addition, the operability of the hydrothermal reaction is slightly poor, so it is difficult to effectively control the growth of the core-shell structure.

2.3.3. Co-Precipitation

The co-precipitation method refers to the preparation of LnNCs with uniform composition in a variety of cationic solutions through a precipitation reaction after adding a precipitant [72,73,74,75,76,77,78,79,80,81]. Generally, the growth of LnNCs is controlled by adding ligands, such as polyvinylpyridinone (PVP), polyethyleneimine (PEI), or ethylenediaminetetraacetic acid (EDTA), to the solvent.

The co-precipitation method may be one of the most effective methods to prepare ultra-small LnNCs, and the size of LnNCs prepared by this method is uniform and controllable. Compared with other methods, this method does not need expensive equipment, the reaction conditions are relatively mild, the experimental operation process is relatively simple, and it saves time [75,78,79,80,81]. Yi et al. [78] improved this method and used simple water-soluble inorganic compounds as precursors to synthesize LaF₃ nanocrystals, with uniform size distribution and size around 5 nm. Guan et al. [79] employed a facile co-precipitation method to synthesize the pure hexagonal NaYF₄ and NaYF₄: Yb, Er by controlling the molar ratio of anions to cations at room temperature and ambient pressure. They can realize the crystal phase transition of LnNCs by controlling the pH of the precursor liquid and obtaining the hexagonal phase in the pH range of 5.0~6.5. The whole process does not use any organic additives, nor high temperature and high pressure, and it is an environmentally friendly technology suitable for large-scale industrial production.

However, the morphology of LnNCs, obtained by the co-precipitation method, is relatively simple, the surface is rough, and the morphology is difficult to control. In addition, the luminescence of LnNCs is also relatively weak, and annealing is generally required to improve the crystallinity of the material and to improve the luminescent intensity. After the annealing treatment, the surfactant wrapped on the surface of the LnNCs will be carbonized, which will destroy the hydrophilicity of the LnNCs and affect its subsequent application.

3. Sensing Mechanisms of NIR-II LnNCs

The energy gap between some Stark sublevels of Ln³⁺ is very small, and the thermal coupling is very strong. Small temperature variations may result in remarkable changes in their emitting intensity [41,82,83,84], which are thus used for temperature sensing. Recently, Nexha et al. [85] gave a comprehensive review on LnNCs-based temperature sensing.

Aside from temperature, LnNCs have no direct relationship with any physiological and biochemical characteristics, due to their relatively stable structure and spectral characteristics. Therefore, in the construction of LnNCs-based bio/chemo sensors, LnNCs work either as energy donors or reference units, which further combine with an appropriate responsive unit to specifically respond to a target of interest. Some inorganic ions and organic dyes were used as the responsive units [44,51,86,87,88,89,90,91,92,93,94,95,96,97,98,99,100,101,102,103], and typical LnNCs-based bio/chemo sensors are summarized in

Table 1. Typical NIR-II LnNCs-based Bio/chemo sensors.

NIR-II LnNCs	Excitation (nm)	Emission (nm)	Response Unit	Influence Signal	Applications	Ref.
NaCeF4: Er, Yb	980	1530	Ce3+ → Ce4+	1530	H2O2	[86]
NaErF4: 2%Ho@NaYF4	1530	980/1180	IR1061	980	H2O2	[44]
NaErF4@NaLuF4	980	654/1550	Cypate	654	·OH	[88]
NaYF4: 18%Yb3+, 2%Er3+	808/980	1550	IR786s	808/1550	ROS	[89]
NaYF4: 20% Yb and 2%Er@NaYF4	808/980	1550	IR786s	808/1550	ROS	[90]
NaYF4: 50%Er@NaYF4	808/980	1550	Cy7.5	808/1550	HOCl	[91]
NaYbF4: 5%Er, 5%Ce@NaYF4: 20%Nd	808/980	1550	IR-783	808/1550	HOCl and ·OH	[92]
NaYbF4: Er@NaYF4: Yb@NaYF4: Nd	808	925/1525	Cy925	925	HOCl	[93]
NaYF4: 18%Yb3+, 2%Er3+	980	1150/1550	SeTT	1150	HOCl	[94]
NaErF4@NaYF4	808/980	1525	IR808	808/1525	HOCl	[95]
NaYF4: 20%Yb, 2%Er@NaYF4: 30%Nd	808/980	1550	4-nitrophenol-Cy7	808/1550	GSH	[96]
NaYF4: 20%Yb, 2%Er@NaYF4: 30%Nd	808	1530	HC-Ni	1530	GSH	[97]
NaYF4@NaYF4: 1%Nd3+	808	1064	MY-1057	1060	ONOO−	[98]
NaErF4@NaYF4@NaYF4: 10%Nd@NaYF4	808	1060/1525	A1094	1060	ONOO−	[51]
NaGdF4: 3%Nd@NaGdF4	808	1060	Compound 1	1060	ONOO−	[99]
NaYF4: 20%Yb, 2%Er@NaYF4	808/980	1050/1550	Ag+ → Ag2S	808/1050	H2S	[87]
NaGdF4: 2%Nd@NaGdF4	808	1060	Compound 1	1060	H2S	[100]
NaYF4: Gd, Yb, Er@NaYF4: Yb	808/980	1053/1525	Ag → Ag2S	808/1053	H2S	[101]
NaGdF4: 5%Nd@NaGdF4	730/808	1064	BSA-NPTAT	730/1064	BSA drug-release monitoring	[102]
NaGdF4:Nd@NaGdF4	808	1058	TA-Fe3+	1058	PPi	[103]

.

3.1. Inorganic-Ion-Based Recognition Mechanism

The LnNCs are internally doped or connected to inorganic ions on the surface, that are responsive to the external environment, to increase and decrease the luminescence of the NIR-II, thereby constructing a probe with a specific response ability to the external environment (Figure 6a). Inorganic ions must first participate in the energy transfer process of LnNCs NIR-II luminescence, and the increase and decrease in NIR-II luminescence can be controlled by the change of ion valence. Secondly, inorganic ions must respond to the external environment, and the change of the external environment will affect the change of the valence state of ions, as to realize the sensing function.

Lei et al. [86] designed and synthesized an H₂O₂ activated NaCeF₄: Er, Yb NIR-II luminescent probe and realized the detection of uric acid (Figure 7a). Among them, the doped Ce³⁺ effectively enhanced the cross-relaxation process of ²F_7/2 (Ce³⁺) + ⁴I_11/2 (Er³⁺) → ²F_5/2 (Ce³⁺) + ⁴I_13/2 (Er³⁺) and greatly increased the number of excited state electrons of ⁴I_13/2 (Er³⁺), thereby effectively inhibiting the up-conversion process and enhancing the down-conversion process, resulting in the nine-fold enhancement of the down conversion luminescence at 1530 nm. In this probe, Ce³⁺ existed as an external environmental responsive ion, which was used to construct a biosensing probe. Under the oxidation of H₂O₂, Ce³⁺ was oxidized to Ce⁴⁺, which weakened the cross-relaxation process and caused the luminescence quenching of the 1530 nm emission (Figure 7b). This responsive probe could be used to detect H₂O₂ or biomolecules that can generate H₂O₂. For example, the detection limit of uric acid could be reduced to 25.6 nmol·L⁻¹ by detecting H₂O₂ produced by the reaction of uric acid and uricase (Figure 7c), and its tissue depth could reach 10 mm in small animal imaging. Similarly, Wang et al. [87] realized the response to H₂S by attaching human serum albumin (HSA)-Ag⁺ (HSA-Ag⁺) to the surface of NaYF₄: 20%Yb, 2%Er@NaYF₄ (Figure 7d). In the presence of H₂S, Ag₂S QDs were formed in coated HSA through an H₂S-induced chemical reaction between H₂S and Ag⁺, which emitted luminescence at approximately 1050 nm (I_{808 Ex/1050}) on irradiation with an 808 nm (Figure 7e). However, the luminescence signal of the probe was stable at 1550 nm (I_{980 Ex/1550}) under 980 nm excitation (Figure 7f), generating a H₂S concentration-dependent ratiometric I_1050/1550 signal. Using this probe, the endogenous H₂S could be real-time monitored, in vivo, by the ratio variation of luminescence signals in two channels of 1050 and 1550 nm.

3.2. Organic-Dye-Based Recognition Mechanism

The LnNCs are hardly individually used for bio/chemo sensing, due to their non-responsive property. However, detection pathways involving chemical reactions are the most common methods due to their high selectivity and sensitivity. To realize the NIR-II sensing of LnNCs, organic dyes can be introduced to the surface of the LnNCs nanoparticles.

The LnNCs themselves do not have the ability to respond to the external environment, and the ability to respond to the external environment is mainly accomplished by organic dyes. Organic dyes must affect the luminescence of LnNCs to change the responsive signal. On the one hand, they directly act on the luminescence of LnNCs to affect the luminescence of LnNCs. That is, organic dyes and LnNCs undergo energy transfer or luminous absorption (Figure 6b). On the other hand, organic dyes affect the luminescence of LnNCs by affecting the intensity of the excitation light source (Figure 6c).

4. Category of Sensing

4.1. Biomedical Related Species

The pathological microenvironment of diseased tissues is obviously different from the physiological environment of normal tissues, such as abnormal redox environment (ROS, HClO, ONOO⁻, GSH, and H₂S, etc.) [44,51,88,89,90,91,92,93,94,95,96,97,98,99,100,101]. These pathological parameters are exactly various diseases (such as cancer, inflammation, and cardiovascular disease) and essential biomarkers. According to the different response environments of LnNCs NIR-II probes, we have classified them.

4.1.1. Reactive Oxygen Species (ROS)

ROS including ·OH and H₂O₂, are widely present in organisms. It plays a vital role in physiological functions, which can regulate proteins, produce hormones, regulate cell signals, mediate inflammation, and eliminate pathogens [44,86,88,89,90].

Jia et al. [88] designed a cypate-modified NaErF₄@NaLuF₄ nanoprobe for detecting ·OH, on the basis of a typical reaction between cypate and·OH. Due to the presence of Er³⁺, NaErF₄@NaLuF₄ could produce 1550 nm NIR-II luminescence and 654 nm luminescence under excitation at 980 nm. The cypate strongly absorbed the luminescence at 654 nm, resulting in luminescence quenching at 654 nm, and it did not affect NIR-II emission at 1550 nm. When cypate encountered ·OH, the structure was destroyed, resulting in increased luminescence at 654 nm. Based on this, a proportional luminescent probe (I₆₅₄/I₁₁₅₀) responding to ·OH was designed. The detection limit was 4.20 μM. In addition, the NaErF₄@NaLuF₄ nanoprobe was successfully used, in mice, for the diagnosis of arthritis in vivo. Similarly, different types of ratiometric ROS sensors have been constructed through different LnNCs and surface modifications of ROS-responsive organic dyes [44,89,90]. Among them, Liu et al. [44] used 1530 nm as the excitation light source to design an H₂O₂ sensor using up-conversion (I₉₈₀/I₁₁₈₀), and Liao et al. [89,90] constructed the ROS sensor with dual excitation light sources (I_{980 Ex}/I_{808 Ex}). For the LnNCs NIR-II sensing of ROS, organic dyes are mainly used as response units, and the sensing is realized by affecting the NIR-II luminescence through organic dyes.

4.1.2. HClO

HClO is a weakly acidic reactive oxygen species, which has the physiological defense functions of antibacterial and anti-inflammatory. However, with the increase in HClO concentration, tissue damage and some diseases may occur, such as neuronal degeneration, arthritis, and so on [91,92,93,94,95].

Wang et al. [91] constructed a proportional luminescent probe, based on the Er³⁺ emission of 1550 nm, to detect the inflammatory site HOCl in vivo (Figure 8a). Here, the Cy7.5 fluorophore was chosen as the HOCl-responsive molecule. The Cy7.5 influenced the luminescence at 1550 nm through the selective absorption of different excitation light sources (I_{980 Ex}/I_{808 Ex}). Under 808 nm excitation, Cy7.5 had a stronger absorption at 808 nm, which greatly weakened the absorption of 808 nm by probe, resulting in weaker luminescence of Er³⁺ at 1530 nm. When HOCl molecules existed, Cy7.5 was degraded to weaken its absorption of 808 nm excitation. On the contrary, it enhanced the intensity of Er³⁺ NIR-II emission. Under 980 nm excitation, the luminescent intensity of Er³⁺ in 1550 nm remained almost unchanged, due to the weak absorption of Cy7.5 at 980 nm (Figure 8b). Therefore, under dual excitation wavelength irradiation, the ratio of the 1550 nm luminescent intensity of the two was closely related to the concentration of HOCl. The HOCl detection limit of the probe was 500 nmol/L, the resolution was 477 μm, and the detection depth was 3.5 mm in the living lymphoid inflammation model. Similarly, Zhang et al. [92] also designed a highly reactive oxygen species (HROS) sensor using dual excitation light sources (I_{980 Ex}/I_{808 Ex}). The detection limits for HOCl and ·OH were calculated as 0.3 and 1.6 μM, respectively. However, Cao et al. [93] reported a NIR-II ratio luminescent probe (I_{925 nm}/I_{1525 nm}) with a single excitation light source (I_{808 Ex}) to detect HOCl by attaching organic dye Cy925 to the surface of NaYbF₄: Er@NaYF₄: Yb@NaYF₄: Nd (Figure 8c). The response unit of HOCl is Cy925, which, itself, emits light at 925 nm. When Cy925 encounters HOCl, the structure of Cy925 will be destroyed, resulting in weakening of the fluorescence at 925 nm. The luminescence of Er³⁺, itself, at 1525 nm is not affected by HOCl. The detection range of the ratio probe for HOCl was 1~24 μM. The sensing of LnNCs NIR-II in HClO is similar to ROS, which is mainly realized by different organic dyes responding to HClO.

4.1.3. Glutathione (GSH)

GSH is an important endogenous antioxidant in the human body, which plays a vital role in the process of cell defense against toxins and free radical production. GSH has also been proven to be an important serum biochemical marker for the diagnosis of many diseases [96,97].

Wang et al. [96] reported a NIR-II nanoprobe consisting of 4-nitrophenol-Cy7 (NPh) conjugated LnNCs (NaYF₄: 20%Yb, 2%Er@NaYF₄: 30%Nd) for in vivo ratiometric sensing of GSH (Figure 8d). The GSH-responsive dye, NPh, has a strong absorption peak at 786 nm. Meanwhile, NPh has no fluorescence due to the presence of intramolecular photoinduced electron transfer (PET). In the presence of GSH, NPh reacted with GSH to generate Cy7-SG, the PET disappeared, and fluorescence recovered. Under 808 nm, Cy7-SG absorbed the energy of 808 nm, and transferred the excitation energy to LnNCs through energy transfer process, thereby achieving sensitization to LnNCs and enhanced luminescence at 1550 nm (Figure 8e). In addition, the 980 nm laser could not excite Cy7-SG, the luminescence at 1550 nm remained unchanged (Figure 8f), resulting in a ratio luminescence signal (I_{808 Ex}/I_{980 Ex}). The ratiometric I_{808 Ex}/I_{980 Ex} value exhibited a linear ship with GSH concentration, ranging from 0~24 mM, with a detection limit of 0.3 mM. This dual excitation light source can effectively realize the sensing of GSH. However, the different absorption of different excitation light sources in vivo may lead to different detection depths and limits in different tissues.

4.1.4. ONOO⁻

ONOO⁻ is a very strong oxidant and plays an important role as an inflammatory mediator in physiology and pathology. ONOO⁻ and its secondary reactants can oxidize and destroy various biological molecules, such as proteins, lipids, and nucleic acids. This process has been implicated in cardiovascular disease, neurodegenerative diseases, host defense, and antitumor immune responses [51,98,99,104].

Zhao et al. [98] designed and synthesized the ONOO⁻ response NaYF₄@NaYF₄: 1%Nd³⁺@MY-1057 probe (Figure 9a). NaYF₄@NaYF₄: 1%Nd³⁺ absorbed 808 nm excitation energy and produced NIR-II luminescence at 1060 nm. MY-1057 molecules existed as response units that respond to ONOO⁻ and had a strong absorption at 1060 nm, thereby reducing the luminescent lifetime and intensity at 1060 nm. When ONOO⁻ existed, MY-1057 dye was degraded and the luminescent lifetime and intensity were restored (Figure 9b,c). The probe had no reference signal and was especially affected by tissue depth when using luminescent intensity for ONOO⁻ quantification. The slopes of intensity of the ONOO⁻ functions were 900, 432, and 60 a.u. µm⁻¹ under 0, 2, and 5 mm penetration depth, respectively (Figure 9d). This illustrates the unreliable quantitative detection of ONOO⁻ using intensity imaging based on signal attenuation due to homogeneous scattering and absorption. In contrast, consistent lifetime response was obtained regardless of penetration depth. Under phantom tissue with various penetration depths, lifetime values exhibited linear correspondence to ONOO⁻ concentration, and the lifetime ONOO⁻ function slopes were almost identical (Figure 9e). When no reference signal was present, the luminescent lifetime was taken as the signal with better stability and penetration depth than the luminescent intensity as the signal. Therefore, using the luminescence lifetime of LnNCs NIR-II as a signal to realize sensing may be a better research direction.

Sun et al. [51] tactfully designed a ratio type NIR-II probe (Figure 9f) that was constructed by modifying organic dyes A1094, responsive to ONOO⁻, on the surface of LnNCs (NaErF₄@NaYF₄@NaYF₄: 10%Nd@NaYF₄). The LnNCs exhibited two NIR-II luminescence emissions of 1525 nm from Er³⁺ and 1060 nm from Nd³⁺. Due to the substantial overlap between the A1094 absorbance at 1094 nm and the LnNCs@A1094 emission at 1060 nm, the 1060 nm emission was quenched by A1094 via the efficient Förster resonance energy transfer. The ONOO⁻ is a highly specific oxidant of A1094, which can destroy its absorbance at 1094 nm. Therefore, the emission at 1060 nm of LnNCs@A1094 could be activated in the presence of ONOO⁻ (Figure 9g). Meanwhile, A1094 had no effect on the luminescence at 1525 nm of LnNCs@A1094 (Figure 9h). The ratiometric luminescent intensity (I₁₀₆₀/I₁₅₂₅) of LnNCs@A1094 increased, linearly, with the augment of ONOO⁻ concentration in the range of 0~6 μM, and the detection limit was 0.8 μM (Figure 9i).

4.1.5. H₂S

H₂S is the third gas signal molecule discovered after NO and CO. It plays an important role in the pathogenesis of neurodegenerative diseases, diabetes, heart failure, inflammation, cancer, and other diseases [87,100,101].

Liu et al. [100] developed a NIR-II luminescent LnNCs (NaGdF₄:2%Nd@NaGdF₄) probe for sensitive and selective sensing of H₂S. In this probe, the H₂S response unit is compound 1, which has high absorption at 808 nm (Figure 10a). Due to the weakening of the luminescent intensity of 808 nm excitation, the NIR-II emission at 1060 nm (Nd³⁺) was weakened (Figure 10b). After the reaction with H₂S, the nucleophilic addition reaction between H₂S and the benzpyrole group in compound 1 could bleach the absorption of compound 1 at 808 nm and recover the 1060 nm emission of LnNCs (Figure 10c). The luminescent intensity at 1060 nm exhibited a good linear relationship in the range from 5 to 750 μM, and the detection limit was 17 nM.

Compared with NIR-II probes that use organic dyes to achieve H₂S response, Deng et al. [101] achieved H₂S responsive NIR-II ratio probes by coating LnNCs (NaYF₄: Gd, Yb, Er@NaYF₄: Yb) with SiO₂ and Ag nanodots (Figure 10d). Here, Ag nanodots were used as the signal unit of H₂S response. After sulfidation with S²⁻, LnNCs@SiO₂@Ag was quickly converted to LnNCs@SiO₂@Ag₂S. Under 980 nm laser, the luminescence of LnNCs@SiO₂@Ag and LnNCs@SiO₂@Ag₂S at 1525 nm were basically unchanged. It did not change with the increase in the amount of H₂S, which could be used as a reference signal (Figure 10e,f). Under 808 nm excitation, LnNCs@SiO₂@Ag has no luminescence emission at 1053 nm. However, when Ag₂S was formed, LnNCs@SiO₂@Ag₂S had a significant luminescence emission at 1053 nm under excitation at 808 nm (Figure 10e,g). Using the detection (I_{808 Ex/1053 Em}) and reference (I_{980 Ex/1525 Em}) signals, the NIR-II ratiometric luminescence signal, presented linearly, increases via improving the content of Na₂S from 0 to 0.08 μM, and the detection limit was 0.7 nM. More importantly, the in situ highly specific ratiometric imaging of the metformin-induced hepatotoxicity was successfully achieved by using the activatable orthogonal NIR-II emitting probe. Compared with organic dyes as the response unit, Ag, as the response unit, has better responsiveness and detection limit for H₂S. The main reason for this may be that the Ag has better stability and is less prone to falling off.

4.2. Chemical Species

In vivo tracking of drug carriers and monitoring of drug release processes are of great significance for evaluating drug delivery efficiency and pharmacokinetics. Wang et al. [102] proposed a competitive absorption strategy based on the multi-wavelength excitation characteristics of LnNCs (NaGdF₄: 5%Nd@NaGdF₄), and designed a micron-sized oral drug carrier that could monitor the release process of protein drugs in the intestine (Figure 11a). They coated LnNCs in a shell of silica. Then, its surface was wrapped a layer of mesoporous silica to form SiO₂-Nd@SiO₂@mSiO₂-NH₂ (mSiO₂-Nd) (Figure 11b). Protein drugs (BSA) and tetrasulfonated phthalocyanine sodium salt (NPTAT) will form BSA-NPTAT complexes due to electrostatic interactions. The high absorption coefficient of the NPTAT, at 730 nm, could dramatically quench the 1060 nm luminescence of mSiO₂-Nd when excited by 730 nm (Figure 11c,d). While under 808 nm excitation, the mSiO₂-Nd 1060 nm emission was barely influenced by the NPTAT due to the weak absorbance of NPTAT at 808 nm. When BSA-NPTAT was loaded into mSiO₂-Nd, the release process of the drug was monitored and quantified by the emission at 1060 nm under the dual excitation light source (I₇₃₀/I₈₀₈). With the gradual release of BSA-NPTAT, the luminescence of Nd³⁺, under excitation at 730 nm, also gradually recovered. The ratio of I₇₃₀/I₈₀₈ reflected the degree of drug release (Figure 11e).

The LnNCs NIR-II luminescence applications in chemical sensing are seldom compared to the broad applications in biosensing. The main reason is that the biggest bright spot of the current NIR-II luminescence, compared with the luminescence of other regions, is the deeper tissue penetration and lower biological background interference. Meanwhile, NIR-II luminescence requires a specific and expensive NIR-II luminous receiver. Therefore, the application of NIR-II chemical sensing can only be based on the distribution of pharmaceutical chemicals or the concentration of chemical substances in organisms. Recently, Su et al. [103] developed an LnNCs NIR-II probe to detect inorganic pyrophosphate (PPi) in vivo for the first time. However, there are very few reports on the NIR-II sensing of LnNCs for chemical substances in vivo, which is also a direction for future sensing research.

5. Conclusions and Outlook

In this paper, NIR-II luminescent LnNCs-based bio/chemo sensors are reviewed. Firstly, the luminescence mechanism of LnNCs is briefly presented, followed by the enhancement strategies of their NIR-II luminescence (ion doping, core-shell structure, and dye sensitization) and preparing methods (thermal decomposition, hydrothermal synthesis, and co-deposition method). We then comprehensively review the sensing mechanism of the LnNCs-based bio/chemo sensors, as well as their applications in detection of various chemical and biomedical-related species. Obviously, NIR-II LnNCs-based bio/chemo sensors show a great prospect in biomedical sensing and imaging. However, they also confront some practical problems when working under physiological environments.

One is the luminescence quenching of NIR-II LnNCs by ambient hydroxyl groups in aqueous medium [2,37]. Although core-shell structure has been proved to be an effective strategy to isolate Ln³⁺ ions from external environment, the non-radiative decay by hydroxyl group is very efficient. This put forwards strict requirements for shell coating, which should not only be thick enough to shield the vibrational decaying of nearby hydroxyl groups but should also be compact enough to prevent the infiltrating of H₂O molecules. On the other hand, the quantum yield of LnNCs is not high (<30%) due to its structure and lattice defects [12,36]. Higher luminescence can be achieved by structural design, doping ions to fill emission levels, or building shell structures in nanoparticles, but this also creates high requirements for synthesis conditions, which is not conducive to large-scale production and subsequent applications. The LnNCs-based bio/chemo sensors show rather weak luminescent intensity because of the inherent small absorption cross-section of Ln³⁺ ions.

In addition, organic dyes are commonly used as fluorescent probes in LnNCs-based bio/chemo sensors [36]. The coupling between organic dyes and organic LnNCs needs to be considered in case of dye leakage, which may reduce the sensing accuracy. It is well known that most of these emissions lie in the visible range, which would degrade the performance of NIR-II sensors with shallow penetration depth. Hence, it is urgent to develop more fluorescent organic probes with NIR emission to match the strong penetrability of NIR-II LnNCs. Meanwhile, current research shows that most LnNPs are harmless luminescent probes suitable for living cells and mice [38,39,105]. However, the application of LnNPs in the body system is still limited due to the complex environment and metabolic processes in the body. There is an urgent need for a systematic and comprehensive assessment of their toxicity and stability in cells and animals.

Last but not least, NIR-II luminescent LnNCs-based bio/chemo sensors exhibit high sensitivity by virtue of deep penetration and background-free detection, but this is at the cost of expensive detectors, such as InGaAs [41,42,105]. This severely hampers their widespread applications in the biomedical realm, particularly in the case of lifetime-based detection. NIR-II LnNCs-based bio/chemo sensors may witness broader applications with the further development of NIR photodetectors.