Lanthanide-Activated Upconversion Luminescent Nanomaterials

Abstract

As a nonlinear optical process, photon upconversion, utilizing consecutive absorption of multiphotons within the long-lived ladder-like intermediate states of lanthanide ions embedded in inorganic matrix to yield high-energy anti-Stokes luminescence, is essentially critical for a variety of emerging applications including multiplexing sensing, anticounterfeiting, volumetric display, super-resolution imaging and optogenetics, and so on. To promote the practical applications, the fundamental understanding of lanthanide-doped upconversion nanomaterials remains a frontier area of research, including host crystal structure, size, dopants, and core–shell construction. Recent progresses also witnessed that the local micro-environment of the activator ions plays the most vital role in regulation of the radiative and non-radiative relaxation mechanisms. This chapter focuses on the recent development of the following aspects: energy transfer processes, lanthanide dopants, host lattices, and nonlinear quantum yield within the frame of inorganic upconversion nanosystems. The state-of-the-art approaches to enhance the upconversion luminescence are also reviewed. Finally, the challenges and future perspectives of lanthanide-activated upconversion luminescence nanomaterials are prospected.

1 Introduction

Among all upconversion materials, those containing lanthanide ions with trivalent state as luminescent centers, typically Nd³⁺, Er³⁺, Ho³⁺, or Tm³⁺ [1], are identified as lanthanide-activated upconversion systems. Since the upconversion phenomenon within the lanthanide ions was firstly discovered by Auzel [2] and independently found by Ovsyankin and Feofilov [3] in the mid-1960s, great efforts have been devoted to discover more and more lanthanide-doped upconversion luminescence systems. Due to the inner 4f ladder-like electronic energy levels of lanthanide ions, which unquestionably lead to abundant optical transitions, the relatively small collection of lanthanide ions dominates the vast majority of all the work in the field of upconversion luminescent materials. Moreover, the partially filled 4f electronic energy levels of lanthanide ions are very insensitive to the structural nature of the surrounding host lattice, such as exact crystal field or the local site symmetry, due to the shielding effect of 5s and 5p subshells. As a result, the intrinsic spectroscopic character of Ln³⁺ has no obvious changes in different chemical microsurroundings, being either doped into crystal lattice or in the form of free ions, which also results in the acute emission peaks and long-lived excited electronic states in host lattices for upconversion processes to intervene [4].

The research field of lanthanide-activated upconverters used to focus mainly on bulk inorganic materials in the solid state. It was not until the 1990s that the study of nanosciences and nanotechnology entered the period of rapid development. Early attempts were reported only by using ball milling method to crush bulk phosphors to prepare nanophosphors, which severely limited the minimum size of phosphors to a few sub-micrometers, preventing wider applications of these phosphors. The beginning of twenty-first century has seen the development of new synthetic methods, such as hydro-/solvo-thermal processing, coprecipitation, and thermal decomposition [5,6,7,8], to synthesize high-quality nanocrystals with controllable sizes and morphologies. Due to high surface-to-volume ratio of the nano-sized powders, the surface of nanoparticles is always full of surface defects or attached by ligands or solvent molecules that possess high phonon energies. Therefore, most of the energy transitions of lanthanide ions are severely quenched by surface defects on/around surface. It can be realized in two ways: (i) Energies are deactivated directly by neighbor surface quenching sites when dopants locate on/around nanocrystal surface. (ii) Energies of dopants locating in the inner nanoparticles travel a long distance to the surface traps or quenching sites. The quenching process results in remarkably low luminescent efficiency of lanthanide emission in nanosystems. Hence, how to enhance the luminescent efficiency of lanthanide-doped upconversion nanomaterials becomes the most challenging issue.

To produce highly efficient upconversion luminescence, there are several key technical specifications that need to be addressed: (i) The energy transfer mechanisms. (ii) Dopants. (iii) Host Matrix. (iv) Emission manipulation strategies. In this chapter, the fundamental principles of the upconversion process are briefly described. Subsequently, the selection of dopants and host matrix as well as the nonlinear nature of upconversion is emphasized. The strategies of boosting upconversion are then reviewed. Finally, several challenging issues and prospects are discussed, expecting to broaden the understanding of these special upconversion nanosystems.

2 Energy Transfer Processes of Lanthanide-Activated Upconverters

Lanthanide upconversion, defined as an anti-Stokes process, usually refers to the sequential absorption of two or more long-wavelength photons via low-lying intermediate long-lived energy levels followed by the emission of short-wavelength photons due to the unique ladder-like 4f electronic states of lanthanide ions. Hence, the upconversion properties of lanthanide-activated nanosystems are resulted from different energy transfer pathways owing to the abundant energy levels of different lanthanide ions. In order to modulate and optimize the optical properties of these nanomaterials, it is essential to study the related energy transfer process. Until now, there are four main classes of models, including excited state absorption, energy transfer upconversion, energy migration-mediated upconversion, and photon avalanche, reported documentarily for the UC process, as shown in Fig. 1. These four basic processes are discussed as follows.

Fig. 1

General schematic illustration of a excited state absorption, b energy transfer upconversion, c energy migration-mediated upconversion, and d photon avalanche processes. Red and green arrows represent the excitation and emission, respectively. Pink and green curved arrow represents energy transfer and migration, respectively

2.1 Excited State Absorption

Excited state absorption process includes the consecutive absorption of pumping photons within the ladder-like energy levels system of the same ion, usually occurs in singly doped upconversion materials. As shown in Fig. 1a, energy transfers from the ground state G to a metastable state E₁ after the absorption of one excited photon. Then, the electrons are excited and transferred from the intermediate state E₁ to a higher excited state E₂ after the resonant absorption of a second pumping photon. As a result, when the high-energy electrons falls from excited state E₂ down to ground state G, the energy will be consumed in the form of light, which lead to the generation of upconversion emission. In this case, the identical ions act as sensitizer as well as activator at the same time. To date, only a few lanthanide ions, such as Er³⁺, Tm³⁺, Ho³⁺, and Nd³⁺ [9], present efficient excited state absorption when they are singly doped into the host matrix and excited by commercially available 980 or 808 nm laser diodes. However, these ions require low doping concentration to avoid inevitable non-radiative processes that severely quench the upconversion luminescence.

2.2 Energy Transfer Upconversion

Energy transfer upconversion, which is also called the ATPE phenomenon (French acronym for Addition de Photons par Transfert d’Energie) firstly found by Auzel in Yb³⁺–Er³⁺ system [2], includes an efficient energy transfer between two neighboring ions that act as sensitizer and activator, respectively, compared to the ESA process. It describes primarily that the sensitizer ion absorbs lower-energy photons, which are successively transferred to luminescent centers or activators in an excited state and raised to higher-energy states via thermally assisted phonon-relaxation process (if there is energy difference between absorbing photon and the transition energy levels). In the case of energy transfer upconversion (Fig. 1b), the sensitizer absorbs a pump photon and then transfers it to the activator, leading to the depopulation of sensitizer state E₁’ and the population of the activator metastable state E₁. The sensitizer ion absorbs another pump photon with the same energy and subsequently transfers it to the upper emitting state E₂ of the activator, while the excited state E₁’ of the sensitizer relaxes again back to its ground state G’. Upconversion emission then occurs when electrons of emitting state E₂ fall back to ground state G of the activator. The efficient energy transfer between sensitizer and activator lies to the larger absorption cross-section of intermediate states of the sensitizer compare to that of the activator, as well as the spacing between these two ions, of which the latter one is determined by the dopant concentration of lanthanide ions. Energy transfer upconversion process is a crucial tool for lanthanide-doped upconversion nanomaterials, especially utilizing sensitizer/activator ion pairs of Yb³⁺/Er³⁺, Yb³⁺/Tm³⁺, and Yb³⁺/Ho³⁺ for enhanced excitation at about 980 nm [10,11,12,13]. Generally, energy transfer upconversion possesses much higher efficiency than excited state absorption process.

There is another energy transfer upconversion model, namely cooperative upconversion, similar to that of the above upconversion process. The only difference lies in the presence of long-lived intermediate levels as well as the upper-lying emitting levels, which results in the cooperative sensitization upconversion and cooperative luminescence, respectively. The cooperative upconversion requires the very close spacing of dopants at their high concentration, usually takes place in Yb³⁺/Eu³⁺- or Yb³⁺/Tb³⁺-codoped systems [14, 15]. However, due to the absence of these intermediate excited levels, the luminescent efficiency of the cooperative upconversion process is several orders of magnitude lower than that of energy transfer upconversion, leading to weak upconversion emission.

2.3 Energy Migration-Mediated Upconversion

In 2011, a novel UC phenomenon, namely energy migration-mediated upconversion, was found by Chen’s group [16] and developed and proposed by Liu’s group [17] for efficient upconversion emissions through energy migration within Gd sublattice based on a large amount of lanthanide activators without long-lived intermediate energy states of NaGdF₄:Yb³⁺,Tm³⁺@NaGdF₄:A³⁺ (A = Sm, Eu, Tb, Dy) core–shell nano-blocks. As shown in Fig. 1c, the mechanism includes four types of ions: sensitizer, accumulator, migrator, and activator, which are rationally designed at precisely defined ion types and concentrations within a core–shell nanoarchitecture. The general process is described as follows: a sensitizer S absorbs pumping photons and enables a neighboring accumulator A’ to higher excited levels. Subsequently, the energy migration takes place between migrators M after energy transfer within the high-lying excited states from the accumulator to the migrator. Finally, the energy is captured by the activator A with the random hopping of energies through the migrator ions in the core–shell interface. To effectively eliminate the dispensable cross-relaxation, sensitizers, accumulators and activators are strictly separated in different layers of the core–shell structure. Under such circumstances, upconversion luminescence of the activator can be achieved without energy transfer through long-lived intermediate energy levels, which generates modulated upconversion emissions. Generally, the energy migration-mediated upconversion process is achieved in Gd-sublatticed core–shell nanoarchitecture. However, the shortcoming of such migration strategy depends strongly on the efficiency of population of ¹I₆ state of Tm³⁺ pumped by Yb³⁺, which is usually low because of the participation of five-photon process. Hence, developing high efficient upconversion in energy migration-mediated core–shell structure will be very important. Recently, Nd³⁺-sensitized upconversion has been developed due to its larger absorption cross-section at ~800 nm compared to that of Yb³⁺ at ~980 nm (about an order of magnitude), and water absorption at 800 nm weaker than at 980 nm (about 20 fold). However, the design only allows for Nd³⁺ doping at rather low concentrations, leading to weak absorption at 800 nm, which is attributed to the deleterious cross-relaxation and multiphonon relaxation processes due to the very close distance between Nd³⁺ ions and activator ions its high concentration. Therefore, the core–shell structure, which spatially separates Nd³⁺ and activator in the shell layer and core layer respectively, will be a promising strategy to avoid the unnecessary energy transfer processes that consume the near-infrared excitation energies. Liu et al. developed a core–shell strategy that was precisely controlled over the concentration of Nd³⁺ in the core and shell layers of a nanoparticle. The upconversion emission were remarkably enhanced through Nd³⁺-sensitization in the core–shell nanostructure [18]. Using efficient energy transfer between Nd³⁺ and Yb³⁺, the core–shell strategy based on Nd³⁺-sensitization and Yb³⁺-migration were then designed to greatly enhance and modulate upconversion emission [19,20,21]. The high efficient upconversion of above Nd³⁺/Yb³⁺ energy migration-mediated process is successful because they avoid the dispensable multiphoton upconversion of high-lying states of Tm³⁺ of Gd³⁺-based migration process, which provides broad applications for bioimaging and phototherapy without concern of overheating.

2.4 Photon Avalanche

Photon avalanche process, firstly, discovered in Pr³⁺-doped infrared quantum counters [22], is considered as looping process that involves excited state absorption and energy transfer upconversion processes within the same ion as illustrated in Fig. 1d. At the beginning, the weak excited state absorption process enables the population of E₁ state of activator ion. In this way, high-lying state E₂ of activator ion could be further populated by energy transfer upconversion, where the energy transfer occurs from the excited sensitizer ion to the activator ion, populating its the emitting state E₂. Subsequently, an efficient cross-relaxation process between sensitizer ion and activator ion occurs (E₂(A) + G’(S) → E₁(A) + E₁’(S)), leading to the population of the intermediate state for both ions. Sensitizer ion then transfers its energy to the activator ion to populate its intermediate state E₁, leading to the formation of the first energy recycling transfer. Alternately, the activator state E₂ is populated again by another energy transfer upconversion process to further stimulate cross-relaxation and promote the exponential population of E₂ by excited state absorption process, which finally induces the strong upconversion emission as an avalanche process. The phenomena were also observed in Nd³⁺- [23] and Tm³⁺-doped [24] luminescent materials. Consequently, the photon avalanche-induced upconversion process can be easily recognized due to its presence as an unusual power-dependent mechanism of multiphoton process, of which the value of slope of logarithmic intensity versus power intensity is always larger than 2. Besides, the photon avalanche-induced upconversion costs a long responsive time, usually up to several seconds, to achieve the avalanche emission.

3 Screening of Dopants and Host Matrix

Generally, both inorganic materials and lanthanide ions do not emit upconversion light under near-infrared excitation at room temperature in most cases. However, when lanthanide ions are introduced into an inorganic host, upconversion luminescence is then formed by the lanthanide dopants served as luminescent centers. Therefore, to achieve better upconversion luminescence, two vital factors should be considered: (i) suitable host materials with different crystal structures providing optimal atomic positions for luminescent centers, and (ii) suitable dopant ions as efficient luminescent centers. In the so-called sensitized luminescence scenario, two kinds of ions play important roles in the luminescence process: an ion would radiate at its higher-lying level after the excitation of absorbed energy obtained from another ion in the way of non-radiative energy transfer. The ion that accepts the energy and emits the luminescence radiatively is called an activator, while the other one that donates the energy is the sensitizer. The ion-to-ion spacing and the spatial arrangement play important roles in the sensitized luminescence process. The upconversion luminescence can be observed in many lanthanide activator-sensitizer combination systems, and most of crystalline materials can be developed as host materials. However, highly efficient upconversion only occurs by the rational design and screening of activator/sensitizer ions, dopant concentration, and host matrix.

3.1 Activators

Multiple intermediate energetic levels make lanthanide ions suitable for upconversion process. The lanthanide ions, which usually present as trivalent state and possess electronic configuration 5s²5p⁶4fⁿ (n = 0–14), range from the lanthanum ion (La³⁺) to the lutetium ion (Lu³⁺). The outer electronic shells 5s² and 5p⁶ orbitals can shield 4f electrons, which leads to weak electron–phonon interaction and hence the parity forbidden f–f transitions, inducing acute 4f transition bands as well as low transition rates and remarkable long-lived (magnitude of orders of about tenth ms) excited states. Other than those lanthanide ions such as La³⁺, Ce³⁺, Yb³⁺, and Lu³⁺, the rest have two or above excited 4f energy levels [25].

In an upconversion process, the sequential excitation to high-lying excited levels pumped with a single monochromatic light can be achieved due to the similar energy gap between ladder-like energy levels. Owning to the long-lived intermediate states, most lanthanide ions hold a great promise and theoretical expectation in upconversion luminescence. However, it is required that each energy gap between an excited state and its lower ladder-like states (including ground state) should be close enough in favor of absorbing pumping photons and energy transfer participating in upconversion processes for the formation of practically useful luminescence. Under such circumstances, not all lanthanide ions meet the criteria of this upconversion excitation due to the lack of suitable energy gap structures. Only Nd³⁺, Er³⁺, Ho³⁺, and Tm³⁺ are served as the most popular dopants (activators) for upconversion nanomaterials due to their adjacent ladder-like energy levels at the excitation of 980 and 800 nm [26] (Fig. 2).

Fig. 2

Schematic illustration of energy levels exhibiting typical excited state absorption processes of Er³⁺, Tm³⁺, Ho³⁺, and Nd³⁺. The pink and grey dashed arrows and full arrows represent excitation, multiphonon relaxation, and emission, respectively. The excitation occurs by the way of photon excitation or energy transfer. Since the energy difference between key pair of energy levels is a little inconsistent, upconversion processes happen with the assistance of phonon relaxation

Due to the similar energy gap of Er³⁺ between two pair of sequential energy states, such as the one between the ⁴I_11/2 and ⁴I_15/2 levels (~10,350 cm⁻¹) compared to that between the ⁴F_7/2 and ⁴I_11/2 levels (~10,370 cm⁻¹), the upconversion efficiency is hereby particularly high. Therefore, the two-photon excited state absorption process can be observed using those levels of ⁴I_15/2, ⁴I_11/2, and ⁴F_7/2 under 980 nm monochromatic excitation. In addition, the energy difference between ⁴F_9/2 and ⁴I_13/2 states is close to the energy gap of the above two pairs of energy levels, and hence, the ⁴F_9/2 state can also be populated by absorbing another 980 nm photon by ⁴I_13/2 state, followed by the multiphonon relaxation of electrons from ⁴I_11/2 state. In this case, the green and red emissions are generated after both the sequential absorption of two photons when the electrons transit from ²H_11/2/⁴S_3/2 and ⁴F_7/2 manifolds to the ground state, respectively. Situations are similar to Tm³⁺ and Ho³⁺ ions. The only difference is that the energy gap between sequential pair of levels are obvious, which differs from the energy of pump photons. Therefore, the absorption of pump photons of these states has to be completed by the phonon-assisted energy transfer, which leads to the lower efficiency of upconversion process in these ions. The upconversion emissions can also be observed in Nd³⁺ doped upconversion nanomaterials [27, 28], only excited with different excitation wavelength due to larger coupled pair of electronic states.

For luminescent materials at nanoscale as well as the Laporte-forbidden 4f–4f transitions, the ineffective trapping of the pumping light can be a critical problem for the lanthanide-doped upconversion luminescence. It is reported that increasing the lanthanide dopant concentration in nanomaterials could remarkably improve the absorption of pumping light. However, high doping concentration brings severe depletion of excitation energy with the emergency of non-radiative multiphonon relaxation and the cross-relaxation, which seriously quenches the upconversion luminescence and hence limits the range of useful dopant concentrations. As a result of electron–phonon interaction, an excited state would usually attenuate to the adjacent lower-lying states in the form of non-radiative transition, which is then converted into phonon energy. The population relaxation of the energy gap between two energy states determines the non-radiative multiphonon relaxation rate, which constitutes another vital element that determines the population of intermediate and emitting levels and straightly influences the efficiency of the upconversion process. By the assumption that phonons participate with equal energy, a commonly used formula for the multiphonon relaxation rate that is temperature-dependent is as follows [29]

$$W\left(T\right)=W\left(0\right){\left[\frac{\mathrm{exp}\left(\mathrm{\hslash }{\omega }_{\mathrm{m}}/kT\right)}{\mathrm{exp}\left(\mathrm{\hslash }{\omega }_{\mathrm{m}}/kT\right)-1}\right]}^{\frac{\Delta E}{\mathrm{\hslash }{\omega }_{\mathrm{m}}}}$$

W(0) represents the spontaneous transition rate with all the phonon modes initially in their ground state when T = 0. ħω_m stands for the cut-off phonon energy of the host lattice going with the sharp transition bands of rare-earth ions. ΔE is the energy gap of the adjacent energy levels (usually between the populated state and the next lower-lying state) of a lanthanide ion. At low temperatures where ħω_m ≫ kT, the multiphonon relaxation rate determined mainly by the constant W(0) for 4f levels of lanthanide ions is described as

$$W\left( 0 \right) = C\exp ( - \alpha \Delta E/\hbar \omega_{{\text{m}}} )$$

where, C and α are variables determined by the particular host. The above equation describes an exponential dependence relationship between the transition rate and the energy gap, suggesting the general exponential decay of multiphonon relaxation rates with increasing energy gap. In identity with the energy gap law, the most effective upconverters hitherto are realized with Er³⁺ and Tm³⁺ as the doping emitters.

3.2 Sensitizer

In the case of singly doped upconversion nanomaterials, the upconversion process is severely affected by two parameters: the interionic distance between two adjacent activator ions and the intrinsic absorption cross-section of the activator ions. As to the former one, the efficiency of energy transfer between two neighboring activators is positively correlative to the distance between them. However, the shorter distance also lead to the deleterious cross-relaxation that severely quenches the excitation and emission energy. Thus, the optimal doping content should be investigated for different types of activator ions and even host matrix for various phase structures. It is reported that singly doping concentration of lanthanide ions would always keep low and be precisely adjusted to avoid the quenching effect. As expected, the low doping levels will lead to inferior absorption of the pump light as well as low energy transfer efficacy, together resulting in low emission luminescence efficiency. Besides, the commonly used activator ions for upconversion exhibit quite low absorption cross-section, resulting in small pumping efficiency. The above reasons give rise to the inconsiderable total upconversion efficiency for singly doped nanomaterials.

To take advantage of the efficient energy transfer upconversion process, an ion with massive absorption cross-section, usually called sensitizer, of the near-infrared pump light should be introduced, which ensures sufficient energy transfer to the activator ion and hence enhance the upconversion luminescence efficiency [25]. It is reported that the most widely used sensitizer for activator Er³⁺ is Yb³⁺ ion due to two major aspects: On one hand, the trivalent Yb³⁺ possesses a simple energy level scheme with only two electronic levels: one ground state ²F_7/2 and one excited state ²F_5/2, where the absorption band owning to Yb^{3+ 2}F_7/2 → ²F_5/2 transition is located at around 980 nm matching well with the energy transitions of, for instance, Er^{3+ 4}I_11/2 → ⁴I_15/2 and ⁴F_7/2 → ⁴I_11/2. On the other hand, Yb³⁺ possesses relatively large absorption cross section compared with that of Er³⁺ (about one magnitude of order). Hence, it enables the efficient energy transfer upconversion process between sensitizers and activators. The similar codoping strategy can also be adopted for Yb³⁺/Tm³⁺ and Yb³⁺/Ho³⁺ combinations. These spectacular optical features make Yb³⁺ suitable for use as an upconversion sensitizer. According to previous reports, Yb³⁺ is codoped into the lattice in relatively high concentration range (18–40 mol%), which would inevitably cause concentration quenching effect due to the cascade energy transfer processes that hamper the upconversion luminescence. To overcome the problem, some special structural designs were adopted, such as by the construction of multilayer sandwich-like nanoarchitecture that spatially separates the dopant ions to enhance the quenching concentration [20, 30], or selecting proper host for intra-clustering energy transfer of dopant ions at sublattice level to suppress significant concentration quenching effect of sensitizer [31].

Yb³⁺ sensitized upconversion encounters a realistic challenge that its excitation wavelength (980 nm) can be absorbed by water and animal or human tissues, resulting in local overheating. For better in vivo bio-applications, an additional near-infrared light source should be introduced. Nd³⁺ sensitization at about 800 nm pump energy for Er³⁺ or Tm³⁺ upconversion emission is attracting more and more research attention. However, since there is no suitable resonant energy levels between Nd³⁺ and Er³⁺/Tm³⁺, Yb³⁺ ions are still used as energy reservoir state for bridging the sensitizer and the activators. The early attempt of 800-nm pumping Nd³⁺ sensitization for Yb³⁺-Tm³⁺ triply doped nanoparticles [32] and Yb³⁺–Tm³⁺–Ho³⁺ quadruply doped nanoparticles [33] revealed that the energy transfer pathways were followed as Nd³⁺ → Yb³⁺ → Tm³⁺/Ho³⁺, where Nd³⁺ absorbed 800-nm pump source and Yb³⁺ acted as bridging ions accepting energies from Nd³⁺ and then transferring the energies to Tm³⁺/Ho³⁺ activators. The multiply doped strategy leads to deleterious energy transfer processes between those lanthanide ions, resulting to the remarkably decrease of upconversion. Recent improved strategies were to construct NaYF₄:Yb,Tm,Nd@NaYF₄:Nd and NaGdF₄:Yb,Er@NaGdF₄:Nd,Yb active core and active shell nanoarchitecture [18, 19], of which the upconversion emission was obviously enhanced compared to the triply-doped non-spatially separating nanoparticles.

3.3 Host Matrix

The screening or selection of appropriate host materials is important for highly efficient upconversion emission due to its interaction with the dopant ions, namely the spatial distribution and chemical surroundings of dopant ions such as the spacing between the dopant ions, their relative spatial atomic positions, coordination environment including the type of ions and the number of ions around the dopant, has a strong impact on the upconversion process. An ideal host material should be spectrally transparent in the range of ultraviolet to infrared light, thermally and chemically stable, with high threshold for optical damage, and highly tolerable with small lattice mismatch for dopant ions.

As stated above, the phonon-assisted non-radiative process dominates the energy loss mechanism for upconversion process. According to the above-mentioned energy gap law, the non-radiative transition rate depends on the ratio between energy gap of activator ions and the cutoff phonon energy of the host lattice, which denotes the number of the phonons participating in the non-radiative process. Generally, the larger is the number of involved phonons is, the smaller is the non-radiative rate, resulting into the higher efficiency of upconversion luminescence. Hence, the host materials should possess low cutoff lattice frequency phonons for the enhancement of the emission efficiency by reducing the non-radiative rate. To date, lanthanide-doped upconversion investigation has been widely performed in various nano-scale host matrix, such as halides, fluorides, oxides, oxysulfides, phosphates, vanadates, etc. Heavy halides such as chlorides, bromides, and iodides generally possess low phonon energies (usually less than 200 cm⁻¹). However, they suffer from low chemical stability and cannot be widespreadly used. Metal oxides (including oxysalts) usually show high chemical and thermal stability, and yet they own relative high cutoff phonon energies larger than 500 cm⁻¹. Among all investigated hosts, fluoride nanomaterials possess competitive advantages (relatively low phonon energies and high chemical/physical stability) and thus are usually served as the host materials for upconversion.

To maximize the tolerance for lanthanide dopants, trivalent lanthanide-based inorganic compounds are chosen as ideal host matrix for upconversion nanomaterials due to the similar ionic radii and chemical properties of all trivalent lanthanide ions. In addition, host lattices based on cations such as alkali ions or alkaline-earth ions with ionic size in the proximity of those of the lanthanide ions effectively minimize the formation of crystal defects and lattice strain. Therefore, lithium/sodium/potassium ion- or calcium/magnesium/barium ion-based lanthanide tetrafluorides are now considered to be excellent host materials for upconversion luminescence. Among those reported fluoride hosts, the NaREF₄ series were proven to be the most excellent hosts due to their unique structural characteristics (cubic and hexagonal phases are widely investigated). For example, the upconversion efficiency of Yb³⁺ and Er³⁺ codoped NaYF₄ is over 20-fold and sixfold higher than that of Yb³⁺ and Er³⁺ codoped lanthanum oxide and lanthanum molybdate, respectively. Although there is a huge distance between present studies and actual applications for upconversion fluorides, NaYF₄:Yb³⁺/Er³⁺ upconversion nanoparticles are now commercially available and launched in several famous enterprises such as Sigma-Aldrich and Aladdin Corporation, etc.

It is well known that changing the nanoscopic environment electronically and structurally can effectively manipulate the excitation energy of a dopant ion, which significantly influences the optical properties of upconversion nanomaterials. So far, hexagonal-phase NaYF₄ (also refers to β-NaYF₄) is reported to be the most efficient host material for visible upconversion phosphors among the rare earth doped fluoride hosts, activated by Yb³⁺–Er³⁺, Yb³⁺–Tm³⁺, and Yb³⁺–Ho³⁺ ion pairs, respectively. In comparison, its cubic counterpart, namely, α-NaYF₄, also exhibits upconversion luminescence with the same Yb³⁺–Er³⁺, Yb³⁺–Tm³⁺, or Yb³⁺–Ho³⁺ doping, of which, however, the corresponding upconversion efficiency is approximately an order of magnitude lower than that in hexagonal phase counterparts [34]. The superior upconversion performance of hexagonal NaYF₄ is found to be originated from the specific distribution of dopant ions randomly located on two types of lattice sites [35], which was then proven to be true from the high-resolution spectra of Eu³⁺ dopants that the high spectroscopic site symmetries of Eu³⁺ descended to low site symmetries in both α- and β-NaYF₄ [36]. According to Judd–Ofelt theory, the parity forbidden 4f–4f electronic transitions are realized by coupling the 4f states with the unoccupied electronic configurations in higher energies of the rare earth ions or with a ligand-to-metal charge transfer configuration through crystal–field interaction. Hence, to develop high efficient upconversion materials, symmetry breakdown in host materials is an effective method, particularly for nanomaterials. Typically, low-symmetry hosts impose an extra crystal field containing more uneven components around the dopant ions compared with high-symmetry counterparts, which then strengthen the electronic coupling between 4f energy levels and higher electronic configurations and subsequently enhance f–f transition probabilities of the dopant ions. Rare earth ions doping is an effective strategy to generate more site symmetry breakdown by inducing lattice distortion of the crystalline lattice. It is confirmed that Eu³⁺ doping into α- and β-NaYF₄ crystals will lower the symmetry of lanthanide crystallographic sites without altering the crystalline structure of the doped crystals [36]. Doping with non-luminous lanthanide ions such as Y³⁺, Gd³⁺, and Lu³⁺, will not only induce the modification of local site symmetry, but also the particle size and phase transition and even the upconversion properties. Liu’s group studied NaYF₄:Yb,Er system just to find that through host ion replacement of Y³⁺ by Gd³⁺, the size (down to ten nanometers), phase (cubic to hexagonal) and upconversion emission intensity (optimal concentration) could be rationally tuned [37]. Xu et al. proposed the phase transition delay protocol in KLu₂F₇:Yb,Er upconversion nanoparticles by replacing Lu³⁺ with Y³⁺ or Gd³⁺, which exhibited enormous enhancement of upconversion luminescence at high doping concentration without changing the crystalline structure of the nanocrystals [38].

4 Upconversion Saturation and Upconversion Quantum Yield

It is well accepted that the long-lived intermediate levels (up to tens of millisecond) of lanthanide ions enable the sequential excitation of the high-lying states within one individual lanthanide ion and hence favor energy transfers between two or more lanthanide ions. These unique characteristics of lanthanide dopants dictate the basic upconversion processes. As mentioned in Sect. 2, the upconversion luminescence arises from four major energy transfer mechanisms, of which, however, only two excitation processes are most common available for lanthanide-activated upconversion luminescence that lead to emission from energy states higher than the terminating state of the first pump absorption step, namely energy transfer upconversion and excited state absorption. Generally, it is well known that the steady-state population density of an electronic state (thus its corresponding upconversion emission intensity) is non-linearly dependent on the pumping light power:

$$I_{{{\text{UC}}}} \sim P^{n}$$

where I_UC represents the upconversion luminescence intensity, P is the pumping power of excitation source, n is the number of the excitation photons required to excite the emitting state. As this power dependence relationship can be verified by carrying out the corresponding experiments, it is used as direct evidence to describe the number of excitation photons participating in the excitation of the upconversion process. Hence, n can be directly determined by the slope of the double-logarithmic relationship between the luminescence intensity and pump power at low excitation density. However, when the excitation sources become more powerful, it is noticed that the nonlinear process (or Pⁿ dependence) begins to reduce and finally disappears where increasing excitation intensity can no longer influence the upconversion emission at all. This phenomenon is known as the “luminescence saturation”, which originates from the dominant depletion competition between “the linear decay” and “upconversion” of the long-lived intermediate states along with variation of pump power [39,40,41].

As a result, the excitation power density should strongly determine the upconversion quantum yield (η_UC) or efficiency of upconversion luminescence due to the nonlinear nature of upconversion process. Specifically, η_UC is defined as the ratio between the total upconverting emission and the absorbed pumping energy (for a detailed demonstration, the total emitted photons against total absorbed photons):

$$\eta_{{{\text{UC}}}} = {\text{photons emitted}}/{\text{photons absorbed}}\sim I_{{{\text{UC}}}} /\alpha P$$

where α is the intrinsic absorption coefficient of the specific host under a particular excitation. The quantum yield is also referred to as the internal quantum yield or the absolute quantum yield, which can be experimentally measured by the combination of a commercially available fluorimeter and an integrated sphere [42]. The absolute quantum yield of several NaYF₄:20%Yb³⁺,2%Er³⁺ samples with different particle sizes were examined based on the above technique. It is found that the upconversion quantum yield of bulk sample is around 3.0%, while other samples whose particle sizes ranging from 10 to 100 nm possess quantum yield of 0.005–0.3%. As emphasized above, the falling quantum yield with decreasing particle size can be accounted for by the elevated surface-to-volume ratio of the smaller nanoparticles which own a large amount of the lanthanide ions in the few outermost shells near the surface of the nanoparticles. This contributes to the enhancement of non-radiative processes between emitting and intermediate levels enabled by solvent molecules as well as excitation energy migration quenching, and hence an overall decrease in the quantum yields. Using the above two equations, it can be easily deduced that:

$$\eta_{{{\text{UC}}}} \sim P^{n - 1}$$

According to this formula, it is very obvious that upconversion quantum yield has strong dependence on the excitation power density. Therefore, η_UC of a given upconversioin luminescence peak (a given state) can be easily determined by referring to a quantified upconversion quantum yield at one particular excitation density.

5 Enhancing Strategies of Lanthanide-Activated Upconversion Luminescence

Although lanthanide-doped upconversion nanomaterials have attracted a great deal of research attention due to their fascinating optical features, they still need to be improved for their low luminescent efficiency compared to that of the bulk materials, as well as strong luminescence quenching largely due to the comparatively low extinction coefficients of lanthanide dopants. This section focuses on the strategies for enhancing the upconversion efficiency of lanthanide-activated nanomaterials.

5.1 Crystal Lattice and Energy Transfer Modulation

As the emission of lanthanide ion is majorly derived from the energy transitions between the abundant energy levels within the parity-forbidden 4f configurations due to crystal field interaction and spin–orbit coupling, the mixing of opposite-parity configurations is able to cause the weakly allowed electric dipole transitions. That’s to say, a crystal field with more asymmetric portion is able to greatly increase the probability of the electric dipole transitions. Therefore, lowering the symmetry of the lanthanide sites by manipulating the crystal microstructure is expected to result in the enhancement of upconversion luminescence.

Doping, which enables various kinds of atoms or ions to be incorporated into a host material that modulates its crystallographic phase, morphology, size, and electronic configurations, become mostly important for inorganic compound materials to control optical, electrical, catalytic, and magnetic properties [43]. Ion doping can efficiently tailor the microstructures for the lattices of the nanoscale host matrix as different ions with various radii could remarkably affect the coordination condition and crystal field of lanthanides (see Fig. 3a), hence for better upconversion luminescence. By utilizing different types or concentration of the dopant ions, it can be easily achieved for enhanced upconversion emission of the doped nanomaterials. The typical codoped system NaYF₄:Yb,Er was previously investigated by its emission dependency on the ratio between Yb³⁺ and Er³⁺ ions, of which the best codopant combination was found to be 20%/2% for the highest upconversion intensity [44]. It was found that through varying Yb³⁺ concentrations to control energy transfer process between Er/Tm and Yb³⁺ ions, the multicolor fine-tuning of NaYF₄:Yb,Er and NaYF₄:Yb,Er,Tm upconversion nanoparticles could be easily achieved. Besides, it was also found that the minor variation in Er³⁺ content would lead to dramatic changes in the emission intensity ratio between Er³⁺ and Tm³⁺, resulting in multicolor output from blue to white of the Yb/Er/Tm tridoped system [11].

Fig. 3

Copyright 2011, 2014 WILEY–VCH Verlag GmbH & Co, KGaA, Weinheim

a Graphical sketching of lattice contraction (left) and expansion (right) due to the substitution of a host atom with a dopant ion of different sizes. b The setup used to measure the upconversion emission of BaTiO₃:Yb,Er thin film by applying an external electric field (left). Sinusoidal ac electric voltage applied to BaTiO₃:Yb,Er thin film and photoluminescence emission as a function of time while the sinusoidal ac electric field is applied to the thin film (right). Adapted with permission from Refs. [43, 61].

Li⁺ ion, possessing the smallest ionic radii in the periodic table of about 0.9 Å, is easily used to be incorporated into the host lattices such as oxides and fluorides. For instance, the upconversion violet and green emission of Y₂O₃:Yb,Er nanophosphors were enhanced up to 60- and 25-fold through Li⁺ ion doping [45], which was attributed to the lattice tailoring by lanthanide sites substitution or the interstitial lattice sites occupation with Li⁺ ions. The versatile Li⁺ doping approach to enhance the corresponding upconversion luminescence was later experimentally verified by several groups in other oxides such as ZrO₂ [46], BaTiO₃ [47], CaMoO₄ [48], GdVO₄ [49], Y₃Al₅O₁₂ [50], and TiO₂ [51], etc. Li⁺ ions can also be intercalated into fluoride hosts as they are excellent upconversion carrier for wider research interests. NaYF₄, one of the most efficient upconverted host materials, were inevitably investigated through Li⁺ doping, for its Tm³⁺ [52] and Er³⁺ [53] and Ho³⁺/Tm³⁺ [54] emissions. It was also found that the great improvement of upconversion emission could be easily achieved by the doping of transition-metal ions for the lanthanide ions in oxide and fluoride systems due to the strengthened electron–phonon coupling and enhanced crystal-field perturbation. [55,56,57].

Other than metal ion doping, the recent advances have seen the successful anion doping strategy to efficiently enhance the upconversion luminescence of lanthanide-doped oxide materials for the manipulation of coordination environment of lanthanide sites. Zhang et al. found the efficient upconversion luminescence enhancement in Yb³⁺/Er³⁺ codoped Gd₂O₃ [58] and NaGd(MoO₄)₂ [59] through anion F⁻ doping. Ye et al. also investigated the role of F⁻ ions in upconversion luminescence enhancement and temperature sensing behavior of Ba₃Lu₄O₉:Yb³⁺,Er³⁺ [60].

Differ from the frequently-used chemical approaches, the physical methods can also tune the crystal lattice efficiently for real-time and in situ modulation of upconversion luminescence. Recently, Hao et al. investigated the lattice distortion driven by electric field of a ferroelectric host material BaTiO₃:Yb³⁺,Er³⁺ thin film, which then exhibited enhanced upconversion emission of the compacted device applied with bias ac voltage [61]. The modulation of upconversion luminescence was precisely synchronized with the sinusoidal ac electric current for the same frequency, suggesting the potential device applications such as electric-regulated upconvertors (see Fig. 3b).

In a typical upconversion process, the energy transfer between activator and sensitizer can be boosted with the decreasing activator-sensitizer spacing when the doping levels of activator ions are high enough (such as Er³⁺, Tm³⁺ or Ho³⁺). However, it will lead to the severe quenching of upconversion emission due to cross-relaxation processes between the ladder-like electronic energy levels of these activator ions. To overcome the problem, some researchers have discovered ingenious approaches to make use of high doping. Jin’s group discovered that by applying high-power excitation density on the upconversion nanoparticles with high Tm³⁺ concentration up to 8% under low-power excitation within an intriguing design (a microstructured optical fiber was inserted with a suspended core where the laser excitations was confined in this micro-meter-sized area, of which the corresponding excitation density reached up to 2.5 × 10⁶ W·cm⁻²), Tm³⁺ upconversion emission achieved a 70-fold enhancement [62]. Another similar report, involving a high doping concentration of Yb³⁺, also depicted the enhanced upconversion emission in NaLuF₄:Yb,Tm ultrasmall nanocrystals, ascribing the upconversion enhancement to the improved absorption of excitation and accelerated energy transfer from Yb³⁺ to Tm³⁺ in spite of the harmful effects such as surface and concentration quenching [63]. High doping of sensitizer Yb³⁺ usually causes enormous quenching of luminescence owing to an increased migration probability of excitation energy from sensitizers to the surface defects of host lattices. A new class of orthorhombic KYb₂F₇:2%Er upconversion nanocrystals was developed lately for its arranged Yb³⁺ tetrad clusters at sublattice level to effectively minimize the excitation energy migration, resulting in a peculiar four-photon-enabled violet upconversion emission intensity with more than eightfold higher than the previously reported one [31].

5.2 Core–Shell Structure Construction

Compared to the bulk counterparts, lanthanide-doped upconversion nanomaterials often possess strong surface quenching effects due to the high volume-to-surface area, which is quite common due to the surface-related effects dominating the energy-loss mechanism. The concentration of surface defects increases with decreasing particle size, resulting in obvious decrease of upconversion luminescence. Similar observations were previously reported in NaREF₄:Yb/Ln (RE = Gd, Y; Ln = Tm, Er) nanoparticles [64, 65]. Most dopant ions are always confined in the outermost few atomic layers of the nanoparticles in lanthanide doping process. Under such circumstances, high-energy oscillators, which originate from ligands, surface impurities, surface defects as well as solvent molecules through multiphonon relaxation processes, inevitably quench the luminescence of surface dopants. Besides, the excitation energies are likely to migrate from interior ions to the quenching sites of particle surface through energy migration-mediated process between sensitizer ions, leading to the dominant non-radiative relaxation and hence the decrease of upconversion luminescence. Therefore, by coating an epitaxial structure onto the surface of a core nanoparticle can effectively minimize the surface quenching effect and enable the efficient energy transfer between the dopants. More importantly, the core–shell design favors the spatially confinement of dopant ions for mitigating deteriorative quenching effects including cross-relaxation, multiphonon relaxation, and quenching of excitation energy. There are three typical coating strategies: amorphous shell coating, inert shell coating, and active shell coating (see Fig. 4a).

Fig. 4

Copyright 2014 WILEY–VCH Verlag GmbH & Co, KGaA, Weinheim. Adapted with permission from Ref. [70]. Copyright 2012 American Chemical Society

a Schematic illustrations of different types of core–shell structure for enhanced upconversion luminescence, including core nanoparticle, amorphous shell coating, inert crystalline shell coating, and active-shell design with sensitizers or activators doped into the shell layer. b Emission spectra of as-prepared NaGdF₄:Yb/Tm@NaGdF₄:A and NaGdF₄:Yb/Tm@NaGdF₄:A@NaYF₄ (A = Dy, Sm, Tb and Eu) nanoparticles. c Luminescence photographs of representative samples in cyclohexane solution under 980 nm irradiation laser. Adapted with permission from Ref. [43].

Amorphous shell coating is a simple and convenient method to increase the upconversion efficiency, which has been revealed by several studies. For instance, silica-/titania-coated Y₂O₃:Yb,Tm nanoparticles with different coating thickness were investigated and the improved upconversion luminescence was found [66]. In another demonstration, the carbon shell coating also triggered the enhancing upconversion luminescence of Yb/Er codoped fluorides [67], which exhibited much stronger emission intensities compared with that of silica-coated counterpart.

An epitaxial inert shell with similar optically inactive composition as the host lattice should minimize the energy loss paths from an activator to quenching sites, which significantly increase the upconversion luminescence intensity due to the provided strong crystal field as well as the effective preservation of excitation energy. Yi and Chow found a nearly 30-fold enhancement of upconversion luminescence in the coating core–shell structure, namely sub 10 nm NaYF₄:Yb/Tm nanocrystals coated with 1.5 nm NaYF₄ shell [68]. Liu’s group achieved the enhanced upconversion intensity of more than 450-fold for 10 nm NaGdF₄:Yb/Tm nanoparticles coated with 2.5 nm NaGdF₄ shell compared to that of the core [64].

An active shell with designed activator ions not only overcomes the problem of surface-quenching effect, but also holds the capacity for high-level doping. Capobianco et al. developed the NaGdF₄:Yb,Er@NaGdF₄:Yb active-core/active-shell structure and exhibited up to 20-fold enhancement of the red upconversion compared to that of the core-only nanoparticles [69]. By coating an inert NaYF₄ shell on the surface of NaGdF₄:Yb/Tm@NaGdF₄:A (A = Tb, Eu, Dy, and Sm) core/shell/shell nanoparticles [70], the upconversion emission intensity of the codoped A ions was drastically boosted (Fig. 4b) and hence lead to the variation of emitting colors (Fig. 4c). Recent progresses also verified the core/multishells strategies in greatly enhancing and manipulating upconversion luminescence of the lanthanide doped nanoarchitetures.

5.3 Surface-Plasmonic Enhancement

Another attractive strategy to enhance the upconversion luminescence is to couple the nanoparticles with noble metal in nanoscale. It is found that the intrinsic localized surface plasmon resonance around the nanoparticles can be generated from the interaction of incident light with noble nanoparticles of size much smaller than the light wavelength (<20 nm). Due to their fascinating properties, noble metal nanocrystals have been widely studied by manipulating the dimensions, morphologies, and spatial configuration. Surface plasmons are able to propagate along the metallic surface and thus produce intensive electromagnetic field (Fig. 5a). Consequently, such noble metal nanoparticles are considered as effective light-trapping agent that can boost the luminescence efficiency of the coupled nanoparticles, of which the intensity I is considered to be proportional with the photon flux of the incident light Φ. According to the literatures, the surface plasmon plays important role in enabling the active absorption of the sentitizer through electric-field interaction, as well as raising the radiative decay rate of the activator and increasing energy transfer probabilities from the sensitizer to the activator. Besides, the surface plasmon resonance induced strong local electric field E may enhance the incident light flux Φ quadratically. In this way, the localized surface plasmon resonance can enhance the overall luminescence intensity.

Fig. 5

Copyright 2014, 2015 WILEY–VCH Verlag GmbH & Co, KGaA, Weinheim. Adapted with permission from Ref. [71]. Copyright 2009 The Royal Society of Chemistry

a Schematic illustration exhibiting the plausible mechanism that governs the plasmonic enhancement of upconversion luminescence. b Optical microscopy images of upconversion nanomaterials without (i) and with (ii) 980 nm irradiation, and the corresponding emission spectra at points A (with Ag nanowires, iii) and B (without Ag nanowires, iv). c Large upconversion enhancement factor with two distinct surface plasmon resonance peaks perfectly matching both the excitation and emission wavelength of the upconversion nanoparticles. Adapted with permission from Refs. [43, 79].

Noble nanoparticles are the best candidate as upconverter booster for its tunable plasmon band from visible to near-infrared scope in favor of the efficient light harvesting. Yan et al. [71] has pioneered in the surface plasmon resonance enhanced upconversion luminescence by coupling Ag nanowires with NaYF₄:Yb,Er nanocrystals (Fig. 5b). Since then, much attention has been paid to the plasmonic-enhanced upconversion strategies. For instance, there are some reports concentrating on how the coupling structure tunes upconversion luminescence at the single particle level and what the related mechanisms are [72, 73]. The geometry of the noble metals has obvious impacts on the localized surface plasmon resonance properties. Researchers have developed noble nanosystems with different geometries, such as gold nanotip [74], gold/silver arrays and thin films [75], gold nanohole array [76], gold pyramid array [77], and gold nanocavity array [78], and so on. However, the upconversion enhancement factors in most of these reports were no more than 20 times. The large localized surface plasmon resonance-induced enhancement is with anticipation when the resonance peaks match the excitation and emission bands of the upconversion nanoparticles. Recently, an intriguing strategy was proposed by utilizing gold nanorods with two distinct surface plasmon peaks in order to match both the excitation and emission wavelength of ZrO₂ upconversion nanoparticles (see Fig. 5c) and an enhancement factor up to 35,000-fold was achieved when the upconversion nanoparticles were coupled with such gold nanostructures [79].

5.4 Broadband Sensitization Strategy

The above strategies have greatly proven the effectiveness in enhancing upconversion luminescence of the lanthanide-doped host materials. However, the inherent limitation to conventional single near-infrared excitation source, the extremely low absorption cross section and narrow absorption band in the near-infrared region greatly limit the practical applications of the upconversion nanomaterials, especially for applications in photovoltaics. Therefore, considerable efforts have recently been devoted to developing novel upconvertion nanosystems featured as broadband near-infrared absorption.

Multisensitizers doping is a simple method to expand the excitation region of the upconversion nanomaterials that meet the criteria of an ideal broadband absorption sensitizer for upconversion process. Yb³⁺, Nd³⁺, Er³⁺, and Ho³⁺ are typical spectral sensitizers used for upconversion for their main absorption wavelength 980, 808, 1480–1600, 1140–1250 nm attributed to ²F_7/2 → ²F_5/2, ⁴I_9/2 → ⁴F_5/2, ⁴I_15/2 → ⁴I_13/2, and ⁵I₈ → ⁵I₆ transitions, respectively. The fascinating work was conducted by Wang et al. [80], in which the NaGdF₄:Er@NaGdF₄:Ho@NaGdF₄ core–shell–shell nanostructures were prepared and effectively minimized the dispensable energy transfer between two sensitizers while maximized the light absorption in the near-infrared region. In the aforementioned Nd³⁺-sensitized upconversion, the triply doped Nd³⁺ with Yb³⁺ and Ln³⁺ also favors the efficient upconversion luminescence. However, the doping concentration of Nd³⁺ keeps very low to minimize unnecessary relaxation processes. The core–shell strategy is introduced to enable high doping level of Nd³⁺ and remarkably enhance upconversion luminescence under 800 nm excitation [18, 19], which shows comparable efficiency for in vivo bioimaging to Yb³⁺-based nanoparticles under the irradiation of 980 nm.

Drawing inspiration from the lanthanide coordination complexes that sensitize lanthanide luminescence by the coordinating organic compounds, organic dyes can be considered as wide-band sensitizers for boosted upconversion luminescence by improving the light-harvesting ability owing to their tunable optical properties, large absorption cross sections, developed preparation techniques and prevailing methods for functionalization. It is well-established that non-radiative energy transfer from dyes to lanthanide ions occurs via Förster and/or Dexter mechanisms originating from either electrostatic or exchange interactions. Zou et al. has pioneered the work by using an organic near-infrared dye IR806 as antenna to sensitize NaYF₄:Yb,Er nanoparticles [81], in which the dye antenna absorbed a broad range of wavelengths and then transferred to Yb³⁺ ions through Förster resonance energy transfer with the resultant overall upconversion luminescence enhanced up to 3300 times (Fig. 6a, b). Ever since, researchers developed several hybrid nanocomposites using near-infrared dyes such as IR780, IR806, IR808, ICG to sensitize the core- and core/shell-structured upconversion nanoparticles [82,83,84,85], in which the dye antenna boosted the upconversion brightness of the core/shell nanoparticles by 20–100 times and up to 33,000 times for core-only nanoparticles.

Fig. 6

Copyright 2012, 2018 Nature Publishing Group

a Principal concept of the dye-sensitized nanoparticle. Antenna dyes absorb near-infrared light and transfer it in the form of Förster resonance to the nanoparticle core to generate upconversion. b Steady-state upconversion emission spectra of β-NaYF₄:Yb,Er nanoparticles, IR-806, β-NaYF₄:Yb,Er nanoparticles/IR-780, and β-NaYF₄:Yb,Er nanoparticles/IR-806 in CHCl₃ excited by 800 nm continue-wave laser. c Mechanism of energy transfer within the dye-functionalized upconversion nanoparticles, depicting the antenna-like nature of the organic dye in sensitizing upconversion nanoparticle that conveys the much larger absorption cross-section of IR806 as well as the internal intersystem cross enhancement by Ln³⁺. d, e Shows the changes of quantum yields of dye-sensitized upconversion nanoparticles along with Gd³⁺ content, and emission enhancement from directly excited 30% Gd³⁺ core, directly excited 20% Gd³⁺ cores coated with a 1.2 nm NaYF₄ shell, the same core–shell nanosystem with dye sensitization and dye-sensitized 30% Gd³⁺ cores. Adapted with permission from Refs. [81, 83].

Compared with dye molecules that are chemically instable and photodegradable, another effective method for broadband sensitization is to exploit the energy transfer between electronic configuration of transition metal ions to 4f levels of lanthanide emitters because of the broad absorption spectra extended to the near-infrared range of transition metal ions, which is possible due to the sufficient overlap between transition metal ion emission bands and lanthanide ion absorption bands. Ni²⁺ and Cr³⁺ have been used as broadband sensitizers in some upconversion phosphors [86,87,88,89] for their broadband absorption range of 1100–1450 and 600–800 nm, respectively. It is found that the host materials using Ni²⁺ as dopants should be selectively conforming to the following criteria to realize broadband-sensitized upconversion: (i) The host lattice should possess octahedraon centers for cation sites rather than tetrahedron or other symmetric sites occupied by Ni²⁺ ions, which refers to the ABO₃ type perovskite structures. (ii) The ionic radius of cations should be comparable with that of Ni²⁺ ions to enable the high Ni²⁺ dopant concentration, which otherwise would oxidize to Ni³⁺ ions. (iii) The host lattice should be tolerable for bigger ions for sufficient amount of Ln³⁺ to be solubilized.

5.5 Photonic Crystals Engineering and Lensing Effect

Photonic crystals possess highly ordered structures with periodically varied refractive index for modulating localized electric fields. This periodicity leads to the formation of a photonic band gap, of which a band of frequencies for light propagation is forbidden in the photonic crystal. The photonic band gap can always be tuned by assembling the opal PMMA beads or polystyrene microspheres of different diameters. Therefore, the interaction between upconversion dopants and their local electromagnetic states influences the spontaneous and stimulated emission processes, leading to amplified upconversion emission at the resonant frequency. Song’s group presented the effective strategy to improve upconversion emission of β-NaYF₄:Yb,Tm nanoparticles by coupling with three-dimensional PMMA opal photonic crystals [90]. This novel-designed system yielded a maximum intensity up to 32 fold higher than the naked nanoparticles, in which the photonic band gap played an important role to the degree of the enhancement factor of the upconverison luminescence (Fig. 7a–c). Simultaneously, Yang et al. demonstrated the polystyrene-based opal photonic crystals could effectively enhance upconversion intensity of the Yb³⁺/Er³⁺ or Yb³⁺/Tm³⁺ codoped NaYF₄ nanocrystals [91]. Niu et al. developed the hybrid systems comprising a well-organized polystyrene frame filled with a periodic arrangement of air pores, within which the nanoparticles were intercalated [92]. A tens fold of enhancement was observed within the hybrid materials compared with that of the nanoparticles alone. Recently, Song et al. integrated the gold nanorods and PMMA opal photonic crystals with NaYF₄:Yb,Er nanoparticles to investigate the synergistic modulation of the upconversion luminescence [93], surprisingly to find an enhancement up to 1200 folds of the overall upconversion intensity and the reduction of detectable excitation threshold by three orders of magnitude (Fig. 7d–f).

Fig. 7

Copyright 2013 The Royal Society of Chemistry. Adapted with permission from Ref. [93]. Copyright 2016 WILEY–VCH Verlag GmbH & Co, KGaA, Weinheim. Adapted with permission from Ref. [94]. Copyright 2017 American Chemical Society. Adapted with permission from Ref. [95]. Copyright 2019 Nature Publishing Group

a Schematic illustration of the formation of PMMA OPCs/NaYF₄:Yb,Tm upconversion nanocrystals. b Upconversion emission spectra of NaYF₄:Yb,Tm and PMMA OPCs/NaYF₄:Yb,Tm. c Dependence of the upconversion enhancement factors as a function of photonic stop band of PMMA OPCs. d Schematic illustration of NaYF₄:Yb,Er/Au/OPCs upconversion nanohyrbids. e Enhancement factors of nanohybrids. f Upconversion detectable excitation threshold values of different kinds of systems. g Optical microscope images of upconversion luminescence of E. coli singly trapped, S. aureus singly trapped, and E. coli/S. aureus co-trapped by a fiber probe without and with a biomicrolen. h Schematic illustration of the experimental setup designed for luminescence amplification investigation. i Photographic images of a PDMS composite sheet comprising 50 μm BaTiO₃ microbeads. j Illustration of documentary security application using the composite sheet. The cross-section SEM image of composite sheet and experimental design for upconversion-based encrypted barcoding were performed. Photographic images of an encrypted quick-response code with and without the composite sheet were taken upon an incoherent light irradiation. Adapted with permission from Ref. [90].

Microlens have been widely used for increasing light-focusing efficiency for solar cells or light extraction efficiency for light-emitting diodes. Based on their lensing effect, the light field confinement by microlens provides the remarkable possibility for luminescence enhancement. Lately, Li et al. reported the spherical or discal-shaped biocells, such as yeast cells or human cells, were able to focus excitation light into a confined subwavelength region due to the photonic nanojet effect [94], which acted as microlens to boost the upconversion luminescence of NaYF₄:Yb,Tm nanoparticles up to 100-fold enhancement (Fig. 7g). Coincidentally, Liu et al. developed an approach that exploited bidirectional light confinement through the use of transparent dielectric microbeads [95], which could be operated as a superlens to restrict an incident light beam into a subwavelength high local intensive optical spot, to stunningly amplify the upconversion emission of NaYF₄:Yb/Ln@NaYF₄ core–shell nanoparticles up to five orders of magnitude while the luminescent decay lifetimes almost had no changes under the impact of the dielectric microbeads (Fig. 7h–j).

6 Conclusions

This chapter introduces the energy transfer mechanisms, dopants, host matrix, and the optimization strategies within the frame of lanthanide-activated upconversion nanomaterials. The past dozen decades have witnessed significant progresses in the efforts to enhance upconversion luminescence of the lanthanide-doped nanosystems, which have led to the emergence of many novel nanomaterials with exceptional optical properties that can expand the use in a broad range of scientific and technological fields.

Despite the inspiring prospects of lanthanide-doped upconversion nanomaterials, there are still some challenges to be addressed. For instance, the detailed energy transfer mechanisms should be further explored by experimental and theoretical works in the nano-particle level. Though many researches involve nonlinear energy transfer processes and physical mechanisms, there are still some arguments related to energy transfer between dopants. It is of most importance to proceed the investigation of novel host materials including unconventional architectures and corresponding properties, which can be manifested by nanoscale engineering and hybridization that can be controlled by various means of chemical methods and those beyond chemical methods such as light, sound, electricity, magnetism, mechanical force and biological incentive, etc. Moreover, the excitation threshold of existing nanoparticles is still quite large that it remains the challenge to reduce the power density of excitation source down to a few mW·cm⁻². At last but not least, the low quantum yield of current lanthanide-doped upconverters is still the major issue that limits the practical use in real life, though the recent doping strategies, plasmonic resonance enhancement, or core–shell building blocks were developed to remarkably enhance the upconversion luminescence.

Therefore, the investigation of upconversion optimization is an ongoing task. The vital answer to the above questions relies on a better comprehension of the spatial distribution of lanthanide ions within a nanoparticle, which requires additional high-end characterization and analytical techniques to unveil the accurate concentration and atomic positions at single particle level. Besides, high standards and requirement for facilities and protocols are demanded to allow quantitative measurement of upconversion properties. Nonetheless, it is believed that lanthanide-doped upconversion nanomaterials will have bright future in basic research and many breakthrough technological applications such as immunotherapy, photodetectors, super-resolution imaging, optogenetics, and many others.