In the past decade, lanthanide ions (Ln3+) doped nanocrystals with fixed compositions capable of emitting tunable upconversion colors have attracted increasing attention, mainly due to their wide prospects in the application fields of 3D display devices, information storage, and advanced anti-counterfeiting1-3.
Previously, different upconversion luminescence colors were obtained using different samples, through strategies such as changing the dopant concentrations4, doping with other lanthanide ions5, introducing luminescent quenchers6, and adjusting the distance among the emitting ions located in different layers in the same particle7. However, by using these approaches, upconversion emissions with certain colors were obtained, which might be suitable exclusively for specific applications. For instance, near infrared (NIR) photons exhibit deep penetration in biological tissues8, red light falling in the spectral biological window I (BW-I) is promising for visible imaging9, green light is most sensitive to human eyes, and blue and UV light are expected to trigger easily photochemical reactions10. Despite of these potential applications, a luminescent material capable of emitting different colors, i.e. a color-tunable material, is highly desired11-14.
Controlling the upconversion luminescence color in nanocrystals with a fixed composition of active ions is relatively difficult since they are almost insensitive to the surrounding conditions. Thus, changes of the ambient temperature, surrounding moisture, electric field, and pressure, typically, only alter slightly the upconversion luminescence color.
To date, the external strategy that resulted more effective in controlling color tunability in a single particle has been the use of different excitation wavelengths to generate different emission colors, by exciting respectively the different lanthanide ions. This strategy demonstrated that it is possible to tune the emission color with high brightness and good color purity at a single nanocrystal level, as it is in principle a spatial combination of different nanocrystals doped with different lanthanide ions that emit different colors through core-shell structural engineering. In 2016, Li et al. published a pioneer work that generated two different colors in a single nanocrystal, blue and green by exciting at 980 nm and 796 nm, respectively, on the basis of the absorption filtration effect15. Later on, Wu et al. demonstrated tunable blue (Ex=980 nm) and red (Ex=800 nm) emissions based on a similar effect16. In 2020, Lei et al. developed tunable green and red upconversion nanocrystals by manipulating the energy migration processes through a dual-excitation strategy17. In the same context, our group reported tunable red/green upconversion emissions upon 980/1530 nm excitation18,19. The above described excitation wavelength-dependent color tuning strategies require a rational design of the compositions of active ions in every layer in a core-shell nanocrystal, where the different layers are responsible for the different emission colors, and the introduction of an inert shell is usually needed between two color layers to avoid the detrimental color interference. Therefore, this type of color tuning nanocrystals is inevitably cumbersome in structure, typically requiring the formation of 4 to 6 layers19. This high structural complexity results in long time highly difficult synthesis processes and high costs, which highly restrain their applicability in real devices. A further consequence of the complexity of these structures has been the difficulty in the obtaining of the three-primary-color (RGB) emission. This has not been obtained until very recently by Hong et al. using an even more complex structure, a 7-layer nanocrystal after excitation at three different excitation wavelengths: 800, 980, and 1530 nm20. Another important limitation of these strategies is the requirement of using multiple excitation sources, which would highly increase the cost and complexity of a real setup for implementing this luminescence color tunability.
From another side, by regulating the excitation in the time-domain allows achieving varied emission colors in a single nanocrystal, and thus, it also becomes a promising strategy towards the generation of high quality color tunability without using multiple excitation sources. Tunable green and red upconversion emissions have been demonstrated by using pulsed 980 nm excitations with different widths and frequencies21. Also, by increasing the pulse width, Kibrisli et al. demonstrated color tuning of orange → green → white emissions using a pulsed 980 nm laser source as well22. But clearly, time-domain modulation of the excitation source requires the introduction of additional devices to manipulate the pulse width or frequency, and also results in cumbersome excitation setups and complicated operations. Furthermore, to date, this strategy has only demonstrated the possibility of achieving two tunable components of the RGB emission.
Alternatively, by adjusting the excitation power density might be the most convenient and lowest-cost way to tune the upconversion emission color, among external stimuli. Although some research publications have demonstrated color tuning of Er3+ emission through variation of the excitation power density, the obtained color purity remains not ideal (usually only from green to yellow or orange), mainly due to the lack of the pure red component23-25. In 2018, in a rationally selected doping composition, Meng et al. achieved high quality tunable green and red upconversion emissions by adjusting the 980 nm excitation pumping power density26. Unfortunately, the further combination of the blue component generated by Tm3+ in this green/red tunable emitting nanocrystal tends to generate white light due to color interference27.
From what described above, it can be concluded that using three different excitation sources can allow achieving tunable RGB emissions, but leads to cumbersome excitation setups, while using a single excitation source with a varied power or pulse widths and frequencies induces only two RGB components at most. Clearly, by combining the two above described strategies might generate the tunable RGB emission and meanwhile partially simplify the excitation setup. In 2015, Deng et al. combined the excitation regulations in wavelength and time-domain, and for the first time demonstrated tunable RGB emissions using only two excitation sources (800 and 980 nm)1.
Achieving tunable RGB upconversion emissions using a single excitation wavelength can largely reduce the operation difficulty and cumbersome of the excitation setup for real applications of these luminescent nanocrystals. However, up to now no successful case has been still reported. In this paper, for the first time, we developed relatively simple-structured upconversion nanocrystals capable of emitting tunable RGB colors by using only one excitation wavelength through what we call photon-order dependent upconversion luminescence color mechanism. Just by adjusting the excitation power density, the RGB color components appear successively in turn. The mechanisms governing the proposed RGB color tunability are investigated. This work presents a novel upconversion luminescent nanocrystal, simultaneously holding tunable RGB emissions, high color purity, relative compact structure, and good brightness, which can find applications in a wide range of fields, such as advanced anti-counterfeiting, multifunctional bioprobes, and controlled release of drugs.
In this work, hexagonal-phase of NaLnF4, the well-accepted highly efficient upconversion host materials, is chosen for the synthesis of luminescence color-tunable nanocrystals. Figure 1a shows the relatively simple structure of the core-shell-shell (C-S1-S2) nanocrystals constructed, with rational designed compositions of Er3+/Yb3+ and Tm3+/Yb3+ in the core and the outermost shell S2, respectively. In addition, an inert layer of pure NaLuF4 isolating the two luminescent layers (core and S2) is inserted as S1 to prevent the cross-relaxation effect between Er3+ and Tm3+, after absorbing the excitation energy. The details of the synthesis procedure can be found in the Supporting Information (SI). Yb3+ was used as an efficient sensitizer for harvesting NIR light excitation at 980 nm and transfers it effectively to Er3+ and Tm3+, that are responsible for the green/red and blue emissions, respectively.
Figure 1a also depicts a TEM image of the 1Er:NaYbF4@NaLuF4@2Tm/30Yb:NaLuF4 (in mol%) C-S1-S2 nanocrystals. It can be seen that the formed nanocrystals are well mono-dispersed and highly uniform in size and shape, indicating the lack of phase separation during the epitaxial growth. The 1Er:NaYbF4 core-only, 1Er:NaYbF4@NaLuF4 core-shell (C-S1), and C-S1-S2 nanocrystals are all nanodisks with mean diameters of 105, 144, and 193 nm, respectively (see Figure S1 in the Supporting Information for additional TEM images and their sizes distributions). The gradual increase in particle size evidences the successful growth of the encapsulation shells. The spatial distributions of the different Ln3+ in a single C-S1-S2 nanocrystal are shown in Figure 1b. The doughnut-like distribution of Lu3+ as well as the first decreased and then increased radial distribution of Yb3+ demonstrate the successful implementation of the structural design to form the lumienscence color-tunable nanocrystals. The hexagonal phase of the as-synthesized nanocrystals was confirmed by X-ray powder diffraction (XRD) (see Figure S2 in the Supporting Information).
We measured also the upconversion luminescent spectra of the as-synthesized C-S1-S2 nanocrystals, dispersed in cyclohexane using a concentration of 0.1 mmol/mL, as shown in Figure 1c. The excitation power density is controlled by changing the distance from the fiber tip of the 980 nm laser to the sample, i.e. changing the excitation spot size. Upon weak excitation (far distance of the fiber laser tip and large excitation area on the sample), the nanocrystals generate dominant green emission bands centered at 520 and 540 nm, which can be attributed to the 2H11/2 and 4S3/2 to 4I5/2 electronic transitions of Er3+, respectively. Intriguingly, the dominant color changes gradually to red and then to blue by increasing the excitation power density (moving the fiber laser tip closer to the sample). The red emission band at 650 nm can be ascribed to the 4F9/2→4I15/2 transition of Er3+, while the blue emission bands at 450 and 475 nm are originated by the 1D2→3H5 and 1G4→3H6 transitions of Tm3+, respectively.
Remarkably, upon weak excitation, the green/(red+blue) emission intensity ratio is 3.3. The red/(green+blue) intensity ratio increases to 3.5 when using a moderate excitation power density. Finally at strong excitation power density, the blue/(green+red) intensity ratio reaches 1.5 (see Figure S3 in the Supporting Information). Figure 1d shows the luminescent images generated by the as-synthesized nanocrystals upon changing the excitation power density, exhibiting relatively good color purity for every component. The results presented here demonstrate that, by using this elaborately designed C-S1-S2 nanocrystal, we achieved the RGB color tuning using a single continuous-wave (CW) 980 nm excitation source.
Theoretically, due to the presence of activators Er3+ and Tm3+ and sensitizer Yb3+, the formed nanocrystals could in principle emit red (Er3+), green (Er3+), and blue (Tm3+) light upon 980 nm excitation. However, allowing everyone of the RGB components to appear upon different 980 nm excitation power density is the main challenge28.
Generally, high-order upconversion emissions require a bigger number of excitation photons to be involved. Thus, an emission generated from a high-lying electronic energy level requires a relatively high excitation threshold. From another side, an increase in the power density is likely to induce a larger rate of increase in the emission intensity of an emission generated from a high-lying energy level, than that occurring at an emission generated from a low-lying energy level29. Considering that the red and green emissions from Er3+ and the blue emission from Tm3+ originate from energy levels with gradually increased energy, it is possible to exploit this order-dependent upconversion nature to achieve the tunable RGB color using only a single CW excitation source.
The main obstacle towards the above scenario is that the Er3+ green and red emissions are usually both related to two-photon absorption processes, i.e. they are generated by the same photon-order. Thus, if they are generated by the same photon-order, it means it is unlikely to be able to extract the pure green and red components by changing the excitation power density of the 980 nm laser in the traditional Er3+-Yb3+ codoped materials used to generate efficient visible emissions by upconversion (see Figure S4 in the Supporting Information).
To break through this obstacle and achieve the tunable green and red emissions, we choose a low-Er3+-high-Yb3+ doping composition of 1Er/99Yb (in mol%). As depicted in Figure 2a, successive energy transfer (ET) processes from Yb3+ to Er3+, along with the rapid nonradiative relaxation, enable the population of Er3+ 2H11/2/4S3/2 levels that generate the green emissions. This population pathway is similar to that of the 2Er/18Yb:NaYF4 nanocrystals, traditionally used to generate efficient green emission by upconversion, corresponding to a 2-photon upconversion process30. By further increasing the excitation power density, more electrons at 4S3/2 level are excited towards the higher-lying 2G7/2 level. After a rapid relaxation to the 4G11/2 level, the Er3+ 4F9/2 level, from which the red emission is generated, can be populated by a back-energy transfer (BET) process described by: 4G11/2(Er3+) + 2F7/2(Yb3+) → 4F9/2(Er3+) + 2F5/2(Yb3+), leading to a 3-photon upconversion pathway that generates the red emission (see Figure 2b). The ultrahigh doping level of Yb3+ in the as-prepared nanocrystals validates the efficient BET process mentioned above. This mechanism is equivalent to increase the photon-order necessary to generate the Er3+ red emission, and enables to discriminate the upconversion mechanism of the red emission from that of the green emission. Thus, in this way it is possible to tune the green and red emissions just by changing the excitation power density. As a consequence, the 3-photon red upconversion luminescence dominates the emission spectrum of the formed nanocrystals upon medium excitation power density. Zhang et al. populated the 4G11/2 level of Er3+ directly through an excitation at 378 nm and generated a bright red luminescence31, confirming the feasibility of the 3-photon upconversion pathways involved in the generation of the red emission. Another evidence supporting the proposed mechanism of the 3-photon red emission is the appearance of the emission bands at 407 and 557 nm, with a low intensity, in the spectrum, marked by asterisks in Figure 1c. The 407 and 557 nm emission bands are both generated from the 2H9/2 energy level (see Figure S5 in the Supporting Information), which is populated by three-photon processes upon Yb3+ sensitization32. Taking the 557 nm emission band for instance, the gradual increase of the luminescent intensity ratio between the intensity of the 557 and 540 nm emission bands when the excitation power density increases, confirms that the 3-photon process plays a more important role in the Er3+ upconversion luminescence when increasing the excitation power density, no matter if the core-only or the C-S1 nanocrystals are considered (see Figure S6 in the Supporting Information).
After achieving the tunable R/G emission, a relatively high doping composition of 2Tm/30Yb:NaLuF4 was chosen to build the outermost layer, after an inert shell of pure NaLuF4 was grown outside the 1Er:NaYbF4 core. Herein, the NaLuF4 instead of the most popular NaYF4 was chosen as the host material to build the S1 and S2 layers, due to the high similarity between Lu3+ and Yb3+ facilitating the construction of the core-shell-shell structure. The 475 and 450 nm emissions from Tm3+ 1G4 and 1D2 levels correspond to 3- and 4-photon upconversion processes, respectively (see Figure 2c). The faster increase in high-order upconversion intensity with the increase of excitation power density enables the fraction of 4-photon blue upconversion emission intensity to gradually increase, and finally it dominated the emission spectrum of the formed nanocrystals upon strong excitation power density. Similar to the analysis of the luminescence intensity ratio between the 557 and 540 nm emissions of Er3+, the appearance of the 507 nm emission (due to the 1D2→3H5 electronic transition of Tm3+, see Figure 1c and Figure S7 in the Supporting Information), even with a low intensity, indicated the occurrence of four-photon process.
In addition, the above-discussed excitation photon-order determined upconversion emission color can be directly confirmed by their corresponding power dependences. The obtained slopes of 1.7 for the Er3+ green emission, 3.1 for the Er3+ red emission, and 3.9 for the Tm3+ blue emission, respectively, clearly demonstrate the proposed 2-, 3-, and 4-photon processes (see Figures 2d-2f).
Besides the tunable green to red emission obtained via the low-Er3+-high-Yb3+ composition in the core of the nanocrystals, the key aspect towards the fully tunable RGB color is the generation of a “switchable” blue emission. The suppression of the blue emission (“off” state) was implemented to avoid the mixture of the blue component with the G/R emission upon weak/medium excitation power densities, while the boost of the blue emission (“on” state) was implemented to ensure that the appearance of the blue component could overwhelm the G/R emission upon strong excitation.
The relatively higher doping concentrations of Tm3+/Yb3+ in the outermost layer is critical for the generation of this switchable blue emission. The concentration of Tm3+ in the S2 layer is 2 mol%, 4-10 times higher than optimal Tm3+ concentration (0.2-0.5 mol%) used traditionally in NaLnF4 to generate an efficient blue emission33. As illustrated in Figure 3a, under weak and medium excitation power densities, the concentration quenching effect caused by the higher Tm3+ doping content in the selected S2 layer quenched the blue emission. This “off” state of the blue emission avoided the mixture of blue component with the red and green ones, when nanocrystals are irradiated with a weak/medium excitation power density. In stark contrast, large amount of “dark” Tm3+, those not involved in the luminescence processes, are activated and followed by a steep increase in intensity upon strong excitation33, corresponding to the “on” state of the blue emission (see Figure 3b). Meanwhile, the green and red upconversion emission intensities tend to saturate within the strong excitation power densities region, as evidenced by the much lower slopes of ~1 in their power dependences (see Figure S8 in the Supporting Information). The faster increase of the blue emission intensity along with the gradually saturated green and red emission intensities contribute together to the high purity of the blue emission upon strong excitation.
To verify the critical role of the 2Tm/30Yb doping composition for the generation of the switchable blue emission, a series of control experiments were performed by changing the concentration of these two lanthanide ions. It was found that the C-S1-S2 nanocrystals with a too diluted Tm3+ concentration in S2 (0.5 mol%) leaded to the absence of the blue emission component, even upon strong excitation, corresponding to an persistent “off” state (see Figure 3b). It is noteworthy that Zhao et al. demonstrated that 8 mol% Tm3+ doping yielded the brightest upconversion blue emission in NaYF4 nanocrystals33 when the excitation power density reached a value as high as 2.5×106 W/cm2. Generally, higher doping levels of Tm3+ in this family of materials require higher excitation power density to fully release their most intense blue emission. Herein, the maximum power density used in this paper was 2.5×103 W/cm2 (the maximum output of our laser), which obliged us to use a much lower optimal Tm3+ concentration of only 2 mol%.
As for the Yb3+ doping concentration, the conventionally optimized doping level of 20 mol%34 resulted in the premature appearance of the blue emission upon weak and medium excitation power densities, and a limited luminescent potential of the Tm3+ blue emission upon strong excitation (see Figure 3c). From another side, an excess of Yb3+ in S2 suppressed the energy harvest of the R/G core, the lack of incident energy absorbed by the core leaded to a persistent “on” state where the blue emission is maintained all the time. As shown in Figure 3d, a high Yb3+ concentration (60 mol%) in the outermost layer S2 absorbed most of the incident photon energy, resulting in an insufficient energy reaching the core to activate the generation of the green emission component upon weak excitation power density.
Note that using different compositions of Er3+/Yb3+ and Tm3+/Yb3+ in different layers is also critical for obtaining the tunable RGB emission. Control experiments on nanocrystals without the core-shell structure (1Er/2Tm/97Yb in mol%) generated green light and the color of the emission could only be slightly changed with the excitation power density (see Figure S9 in the Supporting Information). This confirmed that the core-shell structure prevents color interference between the emissions of Er3+ and Tm3+. In addition, the arrangement of the G/R emitting layer and the blue emitting layer is crucial. It is obvious that the nanocrystals with a reversed spatial arrangement of a 2Tm/30Yb core and a 1Er/99Yb S2 layer would produce tunable green and red emissions, but without the blue component, due to the strong energy filtration produced by the ultrahigh doping Yb3+ level of 99 mol% in the outermost layer15.
By adjusting the excitation power density of the 980 nm laser, we could therefore achieve a dynamic control of the upconversion emission color of the nanocrystals. By increasing the excitation power density from 100 to 103 W/cm2, the dominant emission color gradually changed from green to red and to blue due to changing the fraction of each component in the RGB emission bands (see Figure 4a). The track of CIE coordinates corroborates the color change as the excitation power density increased (see Figure 4b). Videos showing the R/G tuning using the core-only nanocrystals and the RGB tuning using the C-S1-S2 nanocrystals can be found in Video S1 and S2 in the Supporting Information, respectively. A summary of the results published in the literature on dynamic upconversion emission color tuning by manipulation of the 980 nm excitation power density are presented in Table S1 in the Supporting Information. There it can be seen that we achieved tunable full RGB upconversion emission color using a single CW laser excitation for the first time.
In addition to the dynamic emission color tuning ability, the formed nanocrystals also exhibit a relatively good luminescence intensity. The state-of-the-art 2Er/18Yb:NaYF4@NaYF4 (denoted as 2Er/18Yb@Y) nanocrystals used as efficient green upconversion emitters35 are chosen as a reference for the comparison of the luminescence intensity. The TEM images and luminescence images upon weak, medium, and strong power densities of 2Er/18Yb@Y nanocrystals can be found in Figure S10 in the Supporting Information. The integral intensity (400-700 nm) ratio between that generated by our emission color-tunable nanocrystals and that of the 2Er/18Yb@Y nanocrystals as a function of the excitation power density is depicted in Figure 4c. The integral intensity ratio slightly increased from 5.3 to 5.6, with increasing the excitation power density up to 103 W/cm2, during the 2-photon green and 3-photon red dominant luminescence processes. Notably, the 4-photon blue upconversion of our emission color-tunable nanocrystals enlarged this lead and achieved a 33-fold intensity while the excitation reaching 2.5×103 W/cm2. The ~5 times stronger upconversion emission of the C-S1-S2 nanocrystals upon weak excitation would be mainly attributed to the synergy of larger particle size (see Figures S1 and S10 in the Supporting Information) and stronger energy harvest of the 1Er/99Yb core, as compared to that of the 2Er/18Yb@Y nanocrystals. From another side, the gradually increased intensity ratio can mainly stem from the higher orders (3-photon and 4-photon) of the upconversion emissions with increasing the excitation power density, which exhibit larger rate of increases in the emission intensity as mentioned before.
Anti-counterfeiting demonstrations using upconversion emission color-tunable nanocrystals have been widely performed36,37. Herein, as a proof of concept, we decided to design a prototype of a portable laser power test strip, aiming to demonstrate a fast and convenient readout way of the emission power density of the 980 nm laser (see Figure 4d). For that, the RGB emission color-tunable nanocrystals were embedded into Polydimethylsiloxane (PDMS)38, forming a composite film exhibiting the advantages of good physical and optical stabilities, adjustable shape, and high flexibility and transparency (see the left inset in Figure 4d). The calibration of the laser power test strip presents different reference emission colors, corresponding to different excitation power densities (see the right inset in Figure 4d). After calibration, the excitation power density can be roughly estimated by comparing the emission color observed with the previously calibrated reference emission colors. Although similar to a pH test strip that can only provide an approximate range rather than an accurate value, the proposed portable, reusable, and power-free laser power test strip is promising for their use in certain application scenarios that require knowing the rough power density rapidly, for instance, pre-checking the excitation dosage to avoid damage before phototherapies or as a substrate in the in vitro applications that capable of indicating simultaneously the excitation power density.
In this work, we designed a compact core-shell-shell upconversion nanocrystal capable of emitting tunable RGB colors using a single CW excitation source at 980 nm. The 1Er/99Yb composition in the core is responsible for the generation of the dominant green and red emissions upon weak and medium excitation power densities, respectively. The 2Tm/30Yb doping composition, responsible for the generation of the switchable blue emission, is located in the outermost layer of these core-shell-shell nanocrystals. This switchable blue emission has a higher luminescence threshold and steeply increasing intensity with the increase of the excitation power density, thus it is possible to avoid its mixture of the G/R emissions, since it is only dominant at strong excitation power densities. The results reported here provide a clear evidence of the importance of the photon-order dependent upconversion processes in controlling the luminescence pathways, and thus, the emission colors of upconversion nanocrystals. We believe that this work will constitute a major step towards the dynamic control of the upconversion emission color using a simple excitation setup, and will offer new opportunities for applications in the fields of optical storage, advanced anti-counterfeiting, and potentially many others.