Video Upload Options

1. Introduction

The proposal that a solid could detect infrared light by converting two low-energy photons into one higher-energy photon was made by Bloembergen in 1959, as a scheme for an infrared quantum counter that would add visible-light sensitivity to an infrared wavelength.^[1] What he had in mind as a detector became, within a few years, a general phenomenon of trivalent lanthanides in solids: the sequential absorption of two or more near-infrared photons by a single ion, using the ladder of the 4f energy levels as a temporary store. Auzel's comprehensive review collected the early observations, named the process and set the framework that has guided the field since ^[1][2]. Figure 1 shows the canonical energy-level picture.

Figure 1. Energy-level schematic of the archetype Yb³⁺–Er³⁺ upconversion pair. A 980 nm photon is absorbed by Yb³⁺ (the sensitizer) and the excitation is transferred non-radiatively to Er³⁺ (the activator); a second such transfer reaches the higher excited state, from which green or red emission returns the ion to lower levels. The long-lived intermediate level is what lets the ion hold the first photon until the second arrives.

2. Historical Development

2.1. From Infrared Detection to a General Phenomenon

Bloembergen's quantum counter was a proposal for detection rather than for emission, but the underlying idea — that a lanthanide ion could hold an absorbed quantum long enough for another to arrive — was quickly demonstrated in bulk crystals. The surprising step was the discovery that the process could become self-sustaining: Chivian and co-workers described the photon avalanche, in which a vanishingly weak pump produces strong upconverted emission through a feedback loop between excited-state absorption and cross-relaxation ^[3], establishing that the phenomenon was richer than a simple two-photon process.

2.2. Mechanisms

The framework that organises all later discussion was set out by Pollnau and colleagues, who derived the power dependence of upconversion luminescence for each of the competing mechanisms and showed how the slope of emission intensity against pump power identifies which one dominates in a given material ^[4]. The three canonical processes are excited-state absorption, in which a single ion absorbs two photons consecutively; energy-transfer upconversion, in which two neighbouring ions that are each excited pool their energy onto one of them; and photon avalanche, in which a slow-seeded population multiplies through cross-relaxation. The reviews by Haase and Schäfer ^[5], and by Wang and Liu ^[6], lay out the nanocrystal-era restatement of these mechanisms and are the standard entry points to the field.

3. Lanthanide Host and Dopant Chemistry

3.1. Why NaYF₄

The host lattice determines the local environment of the dopant ions, and therefore the radiative rates and the non-radiative losses that compete with them. Fluoride hosts win on the non-radiative side because the lanthanide–fluoride phonon energies are low, which makes multi-phonon relaxation slow, and among fluorides the hexagonal phase of NaYF₄ has proven the most efficient host for green and blue upconversion. Doping is typically with Yb³⁺ as sensitizer, because its absorption at around 980 nm is strong and its single excited state transfers efficiently to Er³⁺, Tm³⁺ or Ho³⁺ as activators. The low-phonon host and the resonant Yb-to-activator transfer together give upconversion quantum yields that no other class of nanocrystal approaches.

3.2. Colloidal Synthesis

Two synthetic routes dominate. Hydrothermal synthesis from rare-earth fluorides gives nanocrystals of controllable size and phase from aqueous precursors at moderate temperatures ^[7], which makes the chemistry scalable but leaves the surfaces with hard-to-remove ligands. Thermal decomposition of lanthanide trifluoroacetates in high-boiling coordinating solvents gives monodisperse hexagonal NaYF₄:Yb,Er with luminescence an order of magnitude brighter than the cubic phase ^[8], and is the route that produced the nanocrystals on which most modern bioimaging work rests. Surface quality, phase purity and dopant concentration are the three levers that any synthesis must control, and the reviews collect the trade-offs between them.

4. Core-Shell Architectures and Luminescence Enhancement

A naked nanocrystal loses a large fraction of its excitation to surface defects: dopants within a few nanometres of the surface transfer their energy to quenching sites instead of emitting. Direct experimental evidence for the size-dependence of this surface quenching came from systematic measurements across a series of nanocrystal sizes ^[9], and the fix is architectural — grow an undoped shell of the same host lattice around the emitting core, which physically separates the dopants from the surface. The shell suppresses quenching and also allows dopant concentrations well above the bulk optimum ^[10], which raises the effective absorption cross-section without losing emission. Absolute quantum-yield measurements on shelled NaYF₄ established how much of the theoretical maximum a real nanocrystal reaches ^[11], and the general design space — inert shells, active shells, and multi-layer structures in which energy migrates across an interface before emitting — is reviewed for the whole family ^[12]. Reviews of the enhancement strategies treat the shell, the dopant distribution and the excitation geometry as the three independent levers ^[13]. Sensitisation by an antenna dye attached to the surface, absorbing in the near-infrared and transferring to Yb³⁺, extends the excitation beyond what the lanthanide itself can absorb ^[14]. Narrowing the emission to a single band, by combining the dopants with a transition-metal ion that suppresses all other channels, produces colours that are useful in multiplexed applications ^[15]. Figure 2 summarises the core-shell geometry.

Figure 2. Cross-section of a core-shell upconversion nanocrystal. The NaYF₄ core carries the Yb³⁺ sensitizers and the Er³⁺ or Tm³⁺ activators; the undoped NaYF₄ shell of a few nanometres keeps the excited dopants away from surface quenching sites and raises the quantum yield. An outer functional layer (silica, amphiphilic polymer or a targeting ligand) makes the particle water-dispersible and gives it a biological handle.

5. Surface Chemistry and Biocompatibility

As-synthesised nanocrystals are hydrophobic, and converting them into a water-dispersible, biocompatible form without destroying the luminescence is a subject in its own right. Silica coating, first shown to leave the upconversion emission intact while giving a chemically versatile surface ^[16], remains the most widely used route. Ligand exchange with bifunctional molecules such as PEG-phosphate gives water dispersibility directly on the nanocrystal surface and a handle for further bioconjugation ^[17]. General reviews of surface modification and characterisation organise the options by what they preserve and what they sacrifice ^[18][19], and comparisons of the same nanocrystal under different surface treatments show that the luminescence intensity and the colloidal stability are coupled — improving one often costs the other ^[20].

6. Bioimaging

Upconversion nanoparticles entered biology as imaging agents. Two properties make them attractive in vivo: near-infrared excitation around 980 nm penetrates tissue far better than visible light, and the anti-Stokes emission produces essentially no autofluorescence background, because the tissue itself does not upconvert. Early demonstrations established in vitro and in vivo imaging with lanthanide-doped NaYF₄ at high contrast ^[21][22], and reviews of the imaging applications have traced the field from proof-of-principle toward quantitative and multiplexed use ^[23][24]. The practical limits are particle brightness, tissue absorption at the excitation wavelength and the choice of biological targeting ligand.

7. Therapy and Sensing

The same deep-tissue excitation is what makes upconversion nanoparticles useful in therapy. In photodynamic therapy the nanocrystal acts as a transducer: near-infrared light that would otherwise not reach the photosensitiser is converted locally to visible light that activates it. Reviews of the therapeutic applications catalogue the configurations and their clinical constraints ^[25], and early demonstrations of nanofabricated upconversion particles paired with photosensitisers showed the principle at work in cells ^[26]. The sharp, narrow emission bands also suit biosensing, in which the upconversion nanoparticle is the donor and a dye the acceptor: optimisation of the donor-acceptor geometry has been shown to reach limits useful for detecting single-molecule events ^[27].

8. Other Applications

Outside biology the same physics supports several technologies. In solar energy the interest is in harvesting sub-band-gap photons that silicon cannot absorb, and reviews of upconversion in solar cells lay out what is required in terms of efficiency and spectral coverage for a net gain ^[28]. Anti-counterfeiting exploits the multi-band, pump-dependent emission of lanthanide-doped nanocrystals as a security feature that is hard to counterfeit without the correct excitation ^[29], and inks based on them are now in commercial use. Super-resolution microscopy is the newest application: because the upconversion process is nonlinear in pump intensity, stimulated emission can be used to de-excite the nanocrystal selectively, and amplified stimulated emission from single upconversion nanoparticles has been used to break the diffraction limit in a way that ordinary fluorophores cannot ^[30], with recent reviews of the biological super-resolution variants reporting the resolutions now reachable ^[31].

9. Perspectives on Further Improvements

The field's open questions are those of a technology approaching its physical limits. The accessible quantum yield of a shelled NaYF₄ nanocrystal is bounded by the intrinsic branching ratios of the dopant, and further gains now come from architecture rather than from composition alone: reviews written for a general audience summarise the structural levers and the trade-offs between them ^[32], and the integration of upconversion with magnetic, radio-sensitising or therapeutic functions on a single core-shell particle is the subject of dedicated studies ^[33]. Designing the energy-transfer network explicitly — rather than accepting the spontaneous lanthanide coupling — is the direction the field has taken ^[34], and the loss mechanisms in the shell have been quantified to show where the remaining efficiency is going ^[35]. On the applications side, extending upconversion into organic-inorganic hybrids for dye-sensitised solar cells has been reviewed as a route that couples the nanoparticle to a broader spectral range than the lanthanide alone can absorb ^[36].

10. Current Status and Outlook

Upconversion nanoparticles occupy an unusual position in the family of luminescent materials: their quantum yield is far below that of a molecular fluorophore, but their excitation and emission wavelengths, their photostability and their absence of background make them the only option for certain imaging tasks. The synthetic chemistry is mature, the surface chemistry is standardised around a few well-understood routes, and the enhancement strategies are understood well enough to be specified in advance rather than discovered empirically. The open problems are those of a technology at scale — reproducibility from batch to batch, biocompatibility for clinical use and absolute efficiency at low pump power — and each is being addressed along the lines set out above.