Dually functioned core-shell NaYF4:Er3+/Yb3+@NaYF4:Tm3+/Yb3+ nanoparticles as nano-calorifiers and nano-thermometers for advanced photothermal therapy

To realize photothermal therapy (PTT) of cancer/tumor both the photothermal conversion and temperature detection are required. Usually, the temperature detection in PTT needs complicated instruments, and the therapy process is out of temperature control in the present investigations. In this work, we attempt to develop a novel material for achieving both the photothermal conversion and temperature sensing and control at the same time. To this end, a core-shell structure with NaYF4:Er3+/Yb3+ core for temperature detection and NaYF4:Tm3+/Yb3+ shell for photothermal conversion was designed and prepared. The crystal structure and morphology of the samples were characterized by X-ray diffraction (XRD), scanning electron microscopy (SEM) and transmission electron microscopy (TEM). Furthermore, the temperature sensing properties for the NaYF4:Er3+/Yb3+ and core-shell NaYF4:Er3+/Yb3+@NaYF4:Tm3+/Yb3+ nanoparticles were studied. It was found that the temperature sensing performance of the core-shell nanoparticles did not become worse due to coating of NaYF4:Tm3+/Yb3+ shell. The photothermal conversion behaviors were examined in cyclohexane solution based on the temperature response, the NaYF4:Er3+/Yb3+@NaYF4:Tm3+/Yb3+ core-shell nanoparticles exhibited more effective photothermal conversion than that of NaYF4:Er3+/Yb3+ nanoparticles, and a net temperature increment of about 7 °C was achieved by using the core-shell nanoparticles.


Results and Discussion
Structure and morphology of NaYF 4 core and core-shell structure. The XRD patterns for the pulverized NaYF 4 : Er 3+ /Yb 3+ core and NaYF 4 : Er 3+ /Yb 3+ @NaYF 4 : Tm 3+ /Yb 3+ core-shell are shown in Fig. 1. All the diffraction peaks can be well indexed in accordance with diffraction pattern of hexagonal NaYF 4 powder (in JCPDS card no. , which indicates that the samples prepared are free of impurities, neither the introduction of RE 3+ into the host nor the shell coating has effect on the crystal phase. The diffraction peaks of NaYF 4 :Er 3+ / Yb 3+ @NaYF 4 :Tm 3+ /Yb 3+ core-shell nanoparticles are slightly narrower than that of the core. From the diffraction peaks at 2θ angle of 17°, the crystallite sizes of the samples are estimated based on the Debye-Scherrer formula ( = λ θ D K BCOS ) to be approximately 21.2 nm for the core and 26.2 nm for the core-shell nanoparticles, respectively. This means that the expected NaYF 4 :Er 3+ /Yb 3+ @NaYF 4 :Tm 3+ /Yb 3+ core-shell structure might be received via the synthesis approach stated in Experimental section.
To observe the microscopic morphology for the NaYF 4 :Er 3+ /Yb 3+ core and NaYF 4 :Er 3+ /Yb 3+ @NaYF 4 :Tm 3+ / Yb 3+ core-shell structure, SEM images were taken and are shown in Fig. 2a,b, respectively. It can be seen from the images that both core and core-shell samples are composed of uniform and regular sphere-like particles. For more clearly observing histological morphology of the samples, the TEM images of the synthesized core and core-shell samples were measured and are displayed in Fig. 2c,d, respectively. It is found that particles in the naked core sample in Fig. 2c are monodispersed and sphere-shaped particles, but fewer particles showed hexagonal shape. Figure 2d for the core-shell sample displays that the particles remain monodispersed, but the morphology of all particles is of hexagonal shape. To estimate the particle sizes of the core and core-shell samples, the sizes of 70 particles for each sample were measured from the TEM images by using noncommercial software. The particle size histograms of core and core-shell samples are shown in Fig. 2f,g, respectively. It is found from Fig. 2f,g that the particle size distribution is narrow, and the average diameter of core sample is determined to be 19.2 nm, and the mean particle size of core-shell sample increases up to 23.2 nm. The results are nearly consistent with the calculation results derived from XRD patterns. These statistical results for particle sizes imply that NaYF 4 :Er 3+ /Yb 3+ nanoparticles were successfully coated with about 2.0 nm thickness NaYF 4 :Tm 3+ /Yb 3+ shell. Figure 2e exhibits the HRTEM images of core-shell UCNPs, the two adjacent lattice fringes with a lattice spacing of 0.52 nm matches well with that of the (100) plane in hexagonal NaYF 4 . The obvious lattice fringes in the HRTEM images further confirm the core-shell structure possesses high crystallinity.
Temperature sensing properties of NaYF 4 :Er 3+ /Yb 3+ and NaYF 4 :Er 3+ /Yb 3+ @NaYF 4 : Tm 3+ / Yb 3+ core-shell particles. Usually, the temperature at tissue surface is relatively easy to be detected with non-contact thermal camera, but the PTT process of tumor located in deep tissue would require real time temperature reading. Therefore, it is necessary to develop a non-contact thermometer. The present studied NaYF 4 :Er 3+ / Yb 3+ @NaYF 4 :Tm 3+ /Yb 3+ core-shell in which Er 3+ has two adjacent levels, namely 4 S 3/2 and 2 H 11/2 in thermal equilibrium, may be a good temperature sensing material. Because the emission intensity ratio of 2 H 11/2 → 4 I 15/2 to 4 S 3/2 → 4 I 15/2 is equal to the population ratio of 2 H 11/2 to 4 S 3/2 which is only dependent on the sample temperature for a certain system. These relations can be mathematically expressed as below.

H S
In above equation, R stands for the fluorescence intensity ratio; I H and I S represents the integrated emission intensities of 2 H 11/2 → 4 I 15/2 and 4 S 3/2 → 4 I 15/2 , respectively; C is a constant depending on the doped RE 3+ and host materials; ΔE is the energy gap between 2 H 11/2 and 4 S 3/2 states, k is the Boltzmann constant, and T is the absolute temperature. Therefore, the sample temperature can be readily derived by taking the measured fluorescence intensity ratio into equation (1) as long as ΔE and C are confirmed.
In order to obtain the parameters C and ΔE for the NaYF 4 :Er 3+ /Yb 3+ and NaYF 4 :Er 3+ /Yb 3+ @NaYF 4 :Tm 3+ / Yb 3+ core-shell particles, the temperature calibration experiments were carried out. In doing so, the cyclohexane solution containing OA-capped NaYF 4 :Er 3+ /Yb 3+ particles or OA-capped NaYF 4 :Er 3+ /Yb 3+ @NaYF 4 :Tm 3+ /Yb 3+ core-shell particles were prepared. The reason for using cyclohexane is that the OA-capped UCNPs could be well dispersed in cyclohexane to form stable colloidal solution directly without further surface modification, and the colloidal solution can maintain transparent and homogeneous for several days. The inset of Fig. 3 shows the images of the solution under irradiation of daylight and 980 nm fiber laser in dark background. The solution in both cases is homogeneous and ready for the spectral measurements. In the course of temperature calibration, Figure 1. XRD patterns of the NaYF 4 :Er 3+ /Yb 3+ core and NaYF 4 : Er 3+ /Yb 3+ @NaYF 4 : Tm 3+ /Yb 3+ core-shell UCNPs, the standard card of β-NaYF 4 (JCPDS: no. 28-1192) is shown as reference.
2 mL solution (4 Wt% of the nanoparticles) in cuvette was firstly heated via water bath to 373 K, and then moved to the sample chamber of the spectrometer for spectral measurement. The UC emission spectra at various temperatures were recorded when the solution was naturally cooled down until room temperature, and the solution temperatures were recoded via a K-type thermocouple connected to a proportional-integral-differential (PID) controller. It should be pointed out that when studying the temperature sensing property, the excitation power should be set as low as possible in order to avoid the laser-induced thermal effect 39,40,43,57 . To check this effect, the green UC emission spectra for the core-shell solution at room temperature were measured at different time under 980 nm fiber laser working at electric current of 1.0 A, and are shown in Fig. 3. It can be seen that all the spectral lines are accurately overlapped, thus implying that under the laser irradiation the solution temperature does not change obviously with time, that is to say, the thermal effect induced by constant laser irradiation can be omitted under these conditions. Therefore, these conditions except for the solution temperature were kept for the full processes of spectral measurements. It should be mentioned that the excitation power densities in this work were about 0.24 × 10 2 W/cm 2 and 1.29 × 10 2 W/cm 2 when the laser working currents were 1.0 and 2.0 A.
The insert in Fig. 4a shows the UC spectra for solution with core-shell sample at various temperatures ranged from 300 to 336 K. It can be seen that the emission intensities of 4 S 3/2 → 4 I 15/2 and 2 H 11/2 → 4 I 15/2 both decrease with increasing solution temperature, which is caused by the higher nonradiative transition rate of 4 S 3/2 level 58 . The integrated intensities at different temperatures for 4 S 3/2 → 4 I 15/2 and 2 H 11/2 → 4 I 15/2 transitions were calculated, and then the fluorescence intensity ratios were derived. The solid circles in Fig. 4a represent the dependence of integrated fluorescence intensity ratio on the solution temperature, and the solid line shows the fitting curve by using equation (1). The free parameters C and ΔE/k were confirmed to be 6.53 and 936.55 K from the fitting processes. In view of the above, the temperature response curve for the core-shell nanoparticles can be received by taking the parameters C and ΔE/k into equation (1). When the temperature response curve is known, the sensitivity (S) for temperature detection can be defined as derivative of fluorescence intensity ratio (R) with respect to temperature (T), and written as follows, On the basis of equation (2), the temperature sensing sensitivity curve (solid triangle dots) was obtained and is shown in Fig. 4a. It is confirmed that the sensitivity ranges from 0.0030 to 0.0033 K −1 in the temperature range of 300-336 K.
It should be mentioned that in above-studied core-shell structure the temperature sensing component, namely the NaYF 4 :Er 3+ /Yb 3+ core, is covered by NaYF 4 :Tm 3+ /Yb 3+ shell which is expected to play the part of calorifiers. Therefore, it is necessary to clarify if the temperature sensing performance of the core-shell particles becomes worse. For this purpose, the temperature sensing properties for the NaYF 4 :Er 3+ /Yb 3+ particles were studied in a analogical way as done for the core-shell sample. The obtained results are shown in Fig. 4b where the insert depicts the UC spectra measured at various temperatures, and the solid circles present the dependence of fluorescence intensity ratio on the sample temperature, the solid up-triangles gives the temperature-dependent sensitivity. The parameters C and ΔE/k in equation (1) were confirmed to be 7.73 and 991.68 K, and the sensitivity increases from 0.003 to 0.0035 K −1 as temperature increases from 301 to 337 K. In comparison, the temperature sensing properties of the naked core and the core-shell are almost the same, thus indicating that the temperature sensing performance of NaYF 4 :Er 3+ /Yb 3+ component in core-shell structure does not abate after coating with NaYF 4 :Tm 3+ /Yb 3+ shell. It should be mentioned that the UC luminescence intensity of NaYF 4 :Er 3+ /Yb 3+ @ NaYF 4 :Tm 3+ /Yb 3+ core-shell is 1.6 times higher than that of the naked NaYF 4 :Er 3+ /Yb 3+ core under the same experimental conditions. The mechanism of luminescence enhancement in the core-shell structure can not be easily deduced from the present spectral data, nonetheless, the intense UC emissions would be beneficial to the temperature measurement in the PTT application, since intense enough fluorescence signal would lower the requirement for high performance sensor in the spectral measurements, and meanwhile make the extraction of UC emissions from tissue easier. Here, we should emphasize that both the green emissions of Er 3+ are in the outside of the biological windows, therefore the core-shell nanoparticles can not be used for the tumor treatment in deep tissue, but they may be useful for treatment of superficial tissue.
In fact, the temperature sensing of lanthanide ions has been widely studied 45,[59][60][61] . To further evaluate the temperature sensing performance of the designed core-shell structure, the relative sensitivities for the naked core, core-shell structure and the bulk NaYF 4 :Er 3+ /Yb 3+ were calculated in the temperature region from 300 to 340 K by using the data obtained in this work and our previous work 60 . The relative sensitivity is defined as where R and S are determined by Eqs (1) and (2). The calculated results are shown in Fig. 4c. From Fig. 4c it can be seen that all the relative sensitivities for these three particles decrease with increasing temperature, and the relative sensitivity for NaYF 4 :Er 3+ /Yb 3+ @ NaYF 4 :Tm 3+ /Yb 3+ core-shell structure does not change obviously in comparison with the naked NaYF 4 :Er 3+ /Yb 3+ cores and the bulk NaYF 4 :Er 3+ /Yb 3+ produced by our group. To compare the relative sensitivity of our designed core-shell structure with those in references, the ΔE/k for NaYF 4 :Er 3+ /Yb 3+ particles under different excitation wavelengths and in different forms including core-shell structures and different crystal phases are collected and listed in Table 1 [62][63][64][65][66][67][68][69][70][71] . The 1 st column of Table 1 presents the size, morphology or structure; the phase and the excitation wavelength for each NaYF 4 :Er 3+ /Yb 3+ particles are listed in 2 nd and 3 rd columns; the 4 th column contains the ΔE/k values for all the NaYF 4 :Er 3+ /Yb 3+ samples; the last column gives references. From Table 1 it can be found that the parameter ΔE/k determining relative sensitivity of NaYF 4 :Er 3+ /Yb 3+ particles changes greatly from sample to sample, and our samples exhibit moderate ΔE/k values amongst all the results from different research groups. In this study, the temperature sensing properties of the naked core and core-shell structure were derived under the condition that the nanoparticles were immersed in cyclohexane liquid. Here we should point out that the most recent investigation proves that the temperature sensing properties of rare earth doped nanoparticles depend also on the immersion liquid environment 72 , thus the obtained results can be different when they are used in in-vivo systems. Photothermal conversion effect of NaYF 4 core-shell structure as nano-calorifier. In above section, the NaYF 4 :Er 3+ /Yb 3+ @NaYF 4 :Tm 3+ /Yb 3+ core-shell particles were qualified for temperature sensing as good as NaYF 4 :Er 3+ /Yb 3+ particles. The aim of this work is to develop bifunctional nanomaterials having both temperature sensing and photothermal conversion for PTT, therefore, in this section we will discuss about the photothermal conversion properties of the NaYF 4 :Er 3+ /Yb 3+ @NaYF 4 :Tm 3+ /Yb 3+ core-shell particles.
To examine the photothermal conversion effect of the core-shell nanoparticles, the cyclohexane solutions with NaYF 4 :Er 3+ /Yb 3+ @NaYF 4 :Tm 3+ /Yb 3+ core-shell particles and NaYF 4 :Er 3+ /Yb 3+ nanoparticles were again used, and 980 nm fiber laser was adopted as excitation source. The procedure for photothermal conversion measurements is designed as follows. The solution containing nanoparticles is continuously irradiated by 980 nm laser, and then the UC spectra are measured at different time when the laser works at the current of 2.0 A. For the sake of studying the influence of core-shell particles' content in solution on the heat generation, three solutions containing 5 Wt% naked NaYF 4 :Er 3+ /Yb 3+ nanoparticles (named as solution I), 6 Wt% (solution II) and 13 Wt% (solution III) NaYF 4 :Er 3+ /Yb 3+ @NaYF 4 :Tm 3+ /Yb 3+ core-shell particles in 7 mL cyclohxane were prepared, and 2 mL of each solution was used for the spectral measurements.
To confirm that the heat generation of the system was caused by the nanoparticles but not cyclohexane via absorbing 980 nm irradiation, the absorption spectra of the pure cyclohexane and solution III were measured and are shown in Fig. 5. The absorption peak at about 925 nm can be seen in both solutions, which originates from the solvent cyclohexane. It should be noted that the absorption of pure cyclohexane at 980 nm is very weak, and the absorbed light may convert to other lights with different wavelengths or transform to heat energy. To examine the emissions of pure cyclohexane under 980 nm excitation, its emission spectra in the ranges from 200-900 nm on F-4600 (Hitachi) and from 900-2250 nm on NIRQuest-256 (Ocean Optics) were measured, and no any emissions were observed, thus implying that the absorbed light energy by cyclohexane can completely converted into heat energy. This means that the heat generation by cyclohexane can not be neglected. The absorption peak at about 976 nm corresponding to the 2 F 7/2 → 2 F 5/2 transition of Yb 3+ is observed in solution III. The intense absorption centered at 976 nm in solution III containing core-shell nanoparticles is prerequisite for effective photothermal conversion, but this does not mean that the photothermal conversion can be achieved. Therefore, the photothermal conversion behavior should be further experimentally checked.
As an example, Fig. 6 shows the UC spectra for solution III measured at different moments within 70 min under irradiation of 980 nm laser working at current of 2.0 A. To find the change of solution temperature with time, the UC emission intensity ratios R (I H /I S ) were calculated, and then the solution temperatures at different moments were derived based on the obtained temperature response curve (Fig. 4a). Hereby, the relations between the temperature of all the solutions and laser irradiation time are shown in Fig. 7. For comparison, the time-dependent temperature for the pure cyclohexane was also given in Fig. 7. The temperature for the pure cyclohexane was measured with a thermocouple under the same experimental condition. From Fig. 7 it can be seen that the temperature of pure cyclohexane grows steadily with the increase of irradiation time within 30 min, and then keeps unchanged after reaching around 304 K, thus indicating that a thermal equilibrium is achieved in the studied system. This means that though the absorption of pure cyclohexane is weak at 980 nm, the photothermal conversion can not be ignored. This observed result for pure cyclohexane coincides with the deduction we made based on Fig. 5. In fact, the absorption of body fluid is also existent in the practical applications of PTT. Therefore, the investigation for photothermal conversion in solution environment may have more practical significance than that in solid powders. It can be found from Fig. 7 that the other solutions with nanoparticles display a similar variation trend of temperature toward irradiation time as the pure cyclohexane does, but they are really different from each others. The temperature increasing rate for solution I at initial moments is larger than that of pure cyclohexane, and the temperature at the final thermal equilibrium is higher than that of cyclohexane.
These results indicate that NaYF 4 :Er 3+ /Yb 3+ nanoparticles have contribution to the photothermal conversion. In comparison with solution I, the solution II and III exhibit more effective photothermal conversion since the final thermal equilibrium temperatures are higher, and the temperature increasing rates in initial stage of laser irradiation are larger. It should also be noted that solution III, which contains more amounts of core-shell particles than that of solution II, displays more effective photothermal conversion behavior. Additionally, though solution   I and solution II contain approximately equal amount of nanoparticles (naked NaYF 4 :Er 3+ /Yb 3+ core in solution I; NaYF 4 :Er 3+ /Yb 3+ @NaYF 4 :Tm 3+ /Yb 3+ core-shell in solution II), the final equilibrium temperature of solution II is much higher than that of solution I, thus indicating that the NaYF 4 :Tm 3+ /Yb 3+ shell plays important role in the photothermal conversion. It should be emphasized that a net temperature incensement of about 15 K from room temperature 295 K to final equilibrium temperature around 310 K can be obtained in solution III. The net temperature increment is around 7 K (see Fig. 7) when the contribution of heat generation by the solvent (cyclohexane) was deducted. It should be mentioned that less than one fifth of volume for the solution but not the full volume were irradiated by the laser beam. Meanwhile, it should also be noted that this 7 K increment is accomplished in the solution with low amount of the studied core-shell nanoparticles (photothermal agent), and the solution in a thermally open environment, thus it can be expected that a 6 K temperature increment from 37 (normal body temperature) to 43 °C (effective tumor treatment temperature) can be obtained by increasing amount of the core-shell nanoparticles deposited on tumor tissues which are surrounded by other tissues other than the open environment, since the temperature increment is proportional to the absorbed heat quality.

Conclusions
The core-shell structured NaYF 4 :Er 3+ /Yb 3+ @NaYF 4 :Tm 3+ /Yb 3+ nanoparticles were successfully synthesized via a thermolysis reaction, and characterized by means of XRD, SEM, TEM/HRTEM, from which it was found that the average core diameter was 19.2 nm and shell thickness was about 2.0 nm. In assistance of water bath heating the optical temperature sensing for the NaYF 4 :Er 3+ /Yb 3+ @NaYF 4 :Tm 3+ /Yb 3+ core-shell nanoparticles was studied and compared with naked NaYF 4 :Er 3+ /Yb 3+ core nanoparticles. It was found that the coating of photothermal converting NaYF 4 :Tm 3+ /Yb 3+ shell did not obviously injure the temperature sensing performance of the core-shell particles, but improved the luminescence intensities of the thermally coupled 2 H 11/2 and 4 S 3/2 levels, that further made the optical signal measurements easy. The photothermal conversion behavior for the core-shell nanoparticles were studied, and it was found that the cyclohexane solution with NaYF 4 :Er 3+ /Yb 3+ @NaYF 4 :Tm 3+ / Yb 3+ core-shell nanoparticles exhibited more effective photothermal conversion than that of the NaYF 4 :Er 3+ / Yb 3+ nanoparticles, and a net temperature increment of 7 K was achieved.

Experimental section
Materials. Yttrium  Synthesis of NaYF 4 : Er 3+ /Yb 3+ core UCNPs. NaYF 4 : Er 3+ /Yb 3+ core particles were synthesized by a thermal decomposition route. As an example, a typical procedure is presented below. A total amount of 2 mmol RECl 3 ·6H 2 O (RE:78 mol% Y 3+ , 20 mol% Yb 3+ and 2 mol% Er 3+ ), 12 mL OA and 30 ml ODE were added into a 100 ml three-necked flask which was then vacuumized for 30 min to remove oxygen under magnetic stirring at room temperature. Next, the flask containing the mixture was heated to 150 °C to dissolve lanthanide chloride, and a yellow transparent solution under vacuum with magnetic stirring was formed. When the solution cooled down to room temperature naturally, 10 mL methanol solution with 5 mmol NaOH and 8 mmol NH 4 F was dropped into the flask slowly. After stirring and reacting at room temperature, the mixed solution was then heated to 100 °C in N 2 atmosphere and kept for 1 h to remove residual methanol, oxygen and water. Subsequently, the solution was heated to 310 °C, vigorously stirred and hold at this temperature for 1 h in N 2 atmosphere. After cooling down to room temperature, the final products were collected with absolute ethanol and centrifugated at 9000 rpm for 10 min. The white precipitate was washed three times with absolute ethanol/cyclohexane (3:1 v/v). Finally, the NPs were dispersed in 5 mL of cyclohexane for further characterizations.
Synthesis of NaYF 4 : Er/Yb@NaYF 4 : Tm/Yb core-shell structure. Following the same synthesis route above, the shell precursor solution containing 2 mmol RE 2 O 3 ·6H 2 O (RE:79.5% mol Y 3+ , 20%mol Yb 3+ , 0.5%mol Tm 3+ ), 12 mL OA and 30 ml ODE were prepared in the reaction vessel, after cooling down to room temperature, 2 mmol core product was added into the reaction vessel before adding methanol solution. After that the temperature was again increased up to 100 °C in N 2 atmosphere and kept for 1 h to remove cyclohexane. After removing cyclohexane, the following synthesis procedure is the same as that of core particles presented above.
Characterization. The crystalline structure of the samples was characterized by XRD (X-ray diffractometer using Cu-Kα1 radiation source, λ = 0.15406 nm, SHIMADZU, Japan). The XRD data in 2θ ranging from 10° to 80° were collected with a scanning step size of 0.02°. Morphologies and microscopic structures were observed by a field emission SEM (FE-SEM, SUPRA 55 SAPPHIRE, RIGAKU, Janpan) and a high resolution TEM (HR-TEM, JEM-2100F, JOEL, Japan). The UC emission spectra were recorded by using F-4600 spectrophotometer (HITACHI, Japan) under excitation of an externally introduced 980 nm fiber laser. It should be mentioned that if without specific statement, the excitation power density was around 0.24 × 10 2 and 1.29 × 10 2 W/cm 2 when the working currents of 980 nm laser were 1.0 and 2.0 A, respectively. The reason why the low power was used for the temperature calibration experiments is to avoid the laser-irradiation-induced heating effect. The temperature controlling experiments were carried out with quartz cuvette (10 m / m ) heated by water bath, and a thermocouple was used to monitor the solution temperature.