*2.3. Water-Soluble Upconverted Nanomaterials*

Two mL of the cyclohexane solution of the upconverted nanomaterial prepared in the above experiment was prepared and ultrasonically dispersed for 5 min. Twenty mg of NaBF4 was dissolved in two mL of acetonitrile solution and was added to the fully dispersed cyclohexane solution of upconverted nanomaterials by stirring at 1000 rpm/min for 30 min to obtain a mixed solution of water and oil separation. Then, the water-soluble NaYF4:Yb, Tm upconverted nanomaterials were obtained by centrifuging at 8000 rpm/min for 15 min.

#### *2.4. Preparation of Carboxylated Proteins*

First, we dissolved 1 g of bovine serum albumin in 20 mL of ultrapure water, then excess oxalic acid was added, and we adjusted the pH to 7–8 with an aqueous sodium carbonate solution under magnetic stirring. Then, 5 mmol of EDC was added and the mixed solution was stirred overnight. A dialysis bag with a molecular-weight cut-off of 10k–30k Da was used for dialysis. The denatured proteins were put into the dialysis bag, clamped on both sides with dialysis clips to prevent leaking, and immersed into a sodium bicarbonate aqueous solution (1000 mL, 2 mmol) leaving it in a refrigerator at 4 ◦C. The sodium bicarbonate aqueous solution was replaced every 4–6 h to ensure that the protein was always in a slightly alkaline environment. The protein was dialyzed and purified under this condition for at least 72 h, and the carboxylated bovine serum albumin was obtained, which was divided into centrifuge tubes and stored in a −20 ◦C refrigerator.

#### *2.5. Carboxylated Protein-Modified Upconverted Nanoparticles*

Next, 0.5 mL of the prepared carboxylated protein, described above, was dissolved in 2 mL of DMF solution, then water-soluble upconverted nanomaterials were added to the DMF solution and the pH was adjusted to about 8 with aqueous sodium bicarbonate solution by stirring for 2 h. After the reaction was completed, the mixed solution was centrifuged at 10,000 rpm/min for 10 min. The precipitates were washed twice with DMF centrifugation, and finally dispersed with DMF for preservation. The above operations were all carried out at room temperature.

#### *2.6. DNA Probes Linked to Carboxylated Protein-Modified Upconverting Nanoparticles*

The pH of 1 mL of the carboxylated protein-modified upconverted nanomaterials obtained in the above steps was adjusted to about 6 with 20 μmol of hydrochloric acid aqueous solution, then 5 mg EDC and 5 mg NHS were added and stirred at room temperature for 25 min to let the carboxyl groups fully react with the nanoparticles. Next, the aminated DNA probe with FAM was added to a 40 μL PBS buffer to make a concentration of 100 μmol. An appropriate amount of the solution was immediately added to the carboxylated upconverted nanomaterial solution, and then put in a shaker to react for

30 min. The reaction was washed twice with PBS buffer at 10,000 rpm/min for 5 min each time, and the resulting precipitates were DNA probes with FAM linking to a carboxylated protein-modified upconverted nanomaterial fluorescent probe (DNA/dBSA /NaYF4:Yb, Tm).

#### **3. Experimental Results and Discussion**

#### *3.1. The Effect of Yb3+ Doping Concentration on the Luminescence of Upconverted Nanomaterials*

Figure 1 shows the energy structure and the transitions of Yb3+ and Tm3+ ions in the upconverting luminescence process. It can be seen that upconversion luminescence is a complex multi-photon energy transfer and conversion process. The energy level transition of 2F7/2 → 2F5/2 of the sensitizer Yb3+ ion matches the energy of the near-infrared photon at 980 nm, so it can continuously absorb the excitation energy and then transfer it to the adjacent luminescent center Tm3+. The 3H5, 3F2 (3F3), and 1G4 energy levels are from Tm3+ ions. Among them, there are three methods for upconversion luminescence: (1) the 3H6 energy level absorbs three photons continuously and transitions to the 1G4 energy level; (2) the 3H6 energy first absorbs two photons continuously and then transitions to the 3F2, then, through the cross-relaxation process 3F2,3 + 3H4 → 3H6 + 1D2 to the 1D2; (3) the non-radiation transitions from 1G4, 1D2 of Tm3+ to the lower energy levels 3F2,3,4, 3H4,5,6 to achieve upconversion luminescence. From the emission spectra of NaYF4:Yb,Tm and NaGdF4:Yb,Tm in Figure 3b,d, it can be seen that different host materials NaY(Gd)F4 will not affect the position of the emission peak, but the doping concentration of Yb3+ ions affects the intensity of the emission peak. However, the change is not a simple linear increase or decrease with the increase or decrease in the doping ratio of Yb3+.

**Figure 1.** Schematic diagram of energy level transition of Yb3+ and Tm3+ ions.

*3.2. The Effect of Rare-Earth Ion Doping Concentration on the Luminescence of Upconverted Nanomaterials*

3.2.1. Effect of Yb3+ Doping Concentration on Upconverted Nanomaterials

NaYF4:x%Yb3+, 0.5%Tm3+ and NaGdF4:x%Yb3+, 0.5%Tm3+ (x = 5, 10, 20, 50, 80) nanomaterials were synthesized by high-temperature thermal decomposition under the same experimental conditions by varying the doping molar fraction of Yb3+ ions with a fixed Tm3+ ions molar fraction of 0.5%. Figures 2 and 3 show the multi-directional characterization results of the prepared NaYF4:x%Yb3+, 0.5%Tm3+ and NaGdF4:x%Yb3+, 0.5%Tm3+. As shown in Figures 2a–j and 3a–j, all the nanomaterials exhibit the characteristics of high size dispersion and good crystallinity. It can be seen from Figures 2k and 3k that NaYF4: 20%Yb3+, 0.5%Tm3+ and NaGdF4:20%Yb3+, 0.5%Tm3+ are uniform in morphology and size, forming a complete and regular hexagonal phase. Obviously, the doping concentration of Yb3+ ions does not have much effect on the morphology of

the upconverted nanomaterials. The nanomaterials with different Yb3+ doping ratios can be synthesized with particle sizes between 25–38 nm, which lays a good foundation for the subsequent preparation of upconversion fluorescent probes. NaYF4:x%Yb3+, 0.5%Tm3+(x = 5, 10, 20, 50, 80) and NaGdF4:x%Yb3+, 0.5%Tm3+ (x = 5, 10, 20, 50, 80) were subjected to phase analysis, as shown in Figures 2l and 3l, the XRD patterns were compared with standard card No.16-0994 (NaF4), and the diffraction peaks obtained all corresponded to the standard card one by one, indicating that the samples obtained under this condition were all pure hexagonal NaYF4:Yb, Tm and NaGbF4:Yb, Tm.

**Figure 2.** SEM, particle size analysis, TEM and XRD images of NaYF4: x%Yb3+, 0.5%Tm3+. SEM: (**a**) x = 5, (**c**) x = 10, (**e**) x = 20, (**g**) x = 50, (**i**) x = 80; particle size analysis: (**b**) x = 5, (**d**) x = 10, (**f**) x = 20, (**h**) x = 50, (**j**) x = 80; (**k**) TEM image of NaYF4: 20%Yb3+, 0.5%Tm3+; (**l**) XRD of NaYF4:x%Yb3+, 0.5%Tm3+ (x = 5, 10, 20, 50, 80).

Figure 4 shows the fluorescence spectra of nanomaterials with different concentrations of Yb3+ doping. It can be seen from Figure 4a,c that the emission peak position is not affected by the host material NaY(Gd)F4 or the doping concentration of Yb3+ ions. The intensity of the emission peak changes as the doping ratio of Yb3+ changes, but the change is not a simple linear increase or decrease with the increase or decrease in the doping ratio of Yb3+.

**Figure 3.** SEM, particle size analysis, TEM and XRD images of NaGdF4: x%Yb3+, 0.5%Tm3+. SEM: (**a**) x = 5, (**c**) x = 10, (**e**) x = 20, (**g**) x = 50, (**i**) x = 80; particle size analysis: (**b**) x = 5, (**d**) x = 10, (**f**) x = 20, (**h**) x = 50, (**j**) x = 80; (**k**) TEM image of NaGdF4: 20%Yb3+, 0.5%Tm3+; (**l**) XRD of NaGdF4:x%Yb3+,0.5%Tm3+ (x = 5, 10, 20, 50, 80).

As shown in Figure 4, the luminescence intensity at 450, 477 and 646 nm increased gradually with the increase in Yb3+ concentration up to 20%, then the emission intensity is decreased with the increase in Yb3+ concentration (Figure 4b,d).

Upon Yb3+ doping, with the change in doping concentration, the number of photons absorbed at 980 nm increases, and the energy transferred to Tm3+ ions increases, so that its luminescence is enhanced. When the concentration of Yb3+ continues to increase, the photon energy absorbed by the Yb3+ ions will pass through the "bridge" between Yb3+ − Yb3+ and surface defects. Through energy resonance transfer, the energy will be transferred to the surface defects and organic vibration groups, through the free radiation process. According to the experimental results, when the optimal Yb3+ doping mole fraction is 20%, the luminescence reaches its peak.

**Figure 4.** Fluorescence spectra of NaYF4: Yb3+, Tm3+ and NaGdF4: Yb3+, Tm3+ with different Yb3+ doping ratios. (**a**) Fluorescence spectra of NaYF4:x%Yb3+, 0.5%Tm3+ (x = 5, 10, 20, 50, 80) at 980 nm excitation. (**b**) Schematic diagram of the relationship between the maximum intensity of the luminescence peaks corresponding to NaYF4: Yb3+ and Tm3+ with different doping ratios and the concentration change. (**c**) Fluorescence spectra of NaGdF4:x%Yb3+, 0.5%Tm3+ (x = 5, 10, 20, 50, 80) at 980 nm excitation. (**d**) Schematic diagram of the relationship between the maximum intensity of the luminescence peaks corresponding to NaGdF4: Yb3+ and Tm3+ with different doping ratios and the concentration changes.

#### 3.2.2. The Effect of Tm3+ Doping Concentration on Upconverted Nanomaterials

NaYF4:20%Yb3+, x%Tm3+ and NaGdF4:20%Yb3+, x%Tm3+ were synthesized by the same method and conditions by varying the concentration of Tm3+ (x = 0.2, 0.3, 0.5, 0.8, 1.0) while the concentration of Tb3+ was fixed at 20%. Figures 5 and 6 show the multi-directional characterization results of the prepared NaYF4:20%Yb3+, x%Tm3+ and NaGdF4:20%Yb3+, x%Tm3+. As can be seen in Figures 5a–j and 6a–j, all the nanomaterials have high dispersibility in size distribution and good crystallinity. It can be seen from Figures 5k and 6k that the upconverted nanomaterials have uniform morphology and size, forming a regular hexagonal phase. Similarly, by changing the doping concentration of Tm3+, the morphology of the upconverted nanomaterials does not change too much. The particle sizes of nanomaterials with different Tm3+ doping ratios can be between 21–43 nm. The XRD pattern of NaGdF4: x%Yb3+, 0.5%Tm3+ (x = 0.2, 0.3, 0.5, 0.8, 1) and NaGdF4: 20%Yb3+, x%Tm3+ (x = 0.2, 0.3, 0.5, 0.8, 1) nanomaterials was also compared with the standard card No.27-0699 (NaGdF4), NaGdF4: Yb, Tm is a pure hexagonal phase. As shown in Figures 5l and 6l, the diffraction peaks obtained all corresponded to the standard card, which means that the samples obtained under this condition were all pure hexagonal NaYF4:Yb, Tm and NaGbF4:Yb, Tm.

**Figure 5.** SEM, particle size analysis, TEM, and XRD images of NaYF4: 20%Yb3+, x%Tm3+. (**a**) x = 0.2, (**c**) x = 0.3, (**e**) x = 0.5, (**g**) x = 0.8, (**i**) x = 1.0; particle size analysis: (**b**) x = 0.2, (**d**) x = 0.3, (**f**) x = 0.5, (**h**) x = 0.8, (**j**) x = 1.0; (**k**) TEM image of NaYF4: 20%Yb3+, 0.5%Tm3+; (**l**) NaYF4: 20%Yb3+, x%Tm3+ (X = 0.2, 0.3, 0.5, 0.8, 1.0).

**Figure 6.** SEM, particle size analysis, TEM, and XRD images of NaGdF4: 20%Yb3+, x%Tm3+. (**a**) x = 0.2, (**c**) x = 0.3, (**e**) x = 0.5, (**g**) x = 0.8, (**i**) x = 1.0; particle size analysis: (**b**) x = 0.2, (**d**) x = 0.3, (**f**) x = 0.5, (**h**) x = 0.8, (**j**) x = 1.0; (**k**) TEM image of NaGdF4: 20%Yb3+, 0.5%Tm3+; (**l**) NaGdF4: 20%Yb3+, x%Tm3+ (X = 0.2, 0.3, 0.5, 0.8, 1.0).

The luminescence spectra of materials with different Tm3+ concentrations are shown in Figure 7. As the mole fraction of Tm3+ increases from 0.2% to 0.5%, the luminescence peaks at 450, 475, and 646 nm also change. As the mole fraction of Tm3+ increases from 0.5% to 1%, the emission intensities of these three luminescence peaks gradually decrease. When the mole fraction of Tm3+ is 0.2%, the luminescence intensity of the nanomaterial is not strong because there are not enough excitable Tm3+ ions in the nanomaterials. As the concentration of Tm3+ increases gradually, the number of excitable Tm3+ ions increases accordingly, and the luminescence of nanomaterials becomes stronger accordingly. When the mole fraction of Tm3+ reaches 0.5%, the upconversion luminescence intensity reaches the maximum, and then gradually becomes weaker. This is because the increase in Tm3+ ion concentration reduces the interionic distance and strengthens the interaction. Finally, the concentration quenching effect and cross-relaxation effect are observed, resulting in a decrease in luminescence intensity. At the same time, when the sample is excited with the same power of the 980 nm laser, since the total energy is fixed, the energy that each Tm3+ can receive will decrease with the increase in Tm3+, leading to the weakening of the luminescence.

**Figure 7.** Fluorescence spectra of NaYF4: Yb3+, Tm3+ and NaGdF4: Yb3+, Tm3+ with different Tm3+ doping ratios. (**a**) Fluorescence spectra of NaYF4:20%Yb3+, x%Tm3+ (x = 0.2, 0.3, 0.5, 0.8, 1.0) at 980 nm excitation. (**b**) NaYF4: Yb3+ with different Tm3+ doping ratios, the relationship between the maximum intensity of the luminescence peak corresponding to Tm3+ and the concentration change. (**c**) Fluorescence spectra of NaGdF4:20%Yb3+, x%Tm3+ (x = 0.2, 0.3, 0.5, 0.8, 1.0 at 980 nm excitation). (**d**) The corresponding doping ratios of NaGdF4: Yb3+ and Tm3+ with different Tm3+ doping ratios.

3.2.3. Comparison of Luminescence Properties of Two Upconverted Nanomaterials with the Best Doping Ratio

From the above results, the luminescence intensity of nanomaterials is the strongest when the mole fraction of Yb3+ is 20% and the mole fraction of Tm3+ is 0.5%. Therefore, the optimal doping concentration ratio (40:1) is selected to prepare these two kinds of upconverting nanomaterials. Figure 8 shows the luminescence intensity comparison of NaYF4:20%Yb3+,0.5%Tm3+ and NaGdF4:20%Yb3+, 0,5%Tm3+. At 345, 362, 450, 477, 646, 802 nm, the luminescence intensity of NaYF4:Yb,Tm is 4.4, 3.0, 4.2, 3.4, 2.3, and 2.7-times stronger than that of NaGdF4:Yb,Tm, respectively. The luminescence intensity of NaYF4:Yb, Tm is much higher than that of NaGdF4:Yb, Tm under the same power of 980 nm laser excitation. Therefore, we chose NaYF4:20%Yb3+,0.5%Tm3+ upconverted nanomaterials as the substrate to further investigate protein detections.

**Figure 8.** Comparison of fluorescence spectra of NaYF4: 20% Yb3+, 0.5% Tm3+ and NaGdF4: 20% Yb3+, 0.5% Tm3+ at 980 nm excitation.

#### *3.3. Analysis of Fluorescence Characteristics Based on NaYF4:Yb3+, Tm3+ Biological Probes*

After the carboxylated bovine serum albumin was added to the upconverted nanomaterial sample for reaction treatment, we carried out repeated centrifugal washing on the sample to remove the carboxylated bovine serum albumin, and then measured the infrared absorption spectrum of the remaining samples, as shown in Figure 9a (the red line). As shown in Figure 9a, the broad infrared (IR) absorption peak at 3295.86 cm<sup>−</sup><sup>1</sup> represents the stretching vibration peak of OH in the carboxyl functional group; the sharp IR absorption peak at 1650.82 cm<sup>−</sup><sup>1</sup> represents the C=O formed after the reaction between the carboxylated protein and NH2-PEG group stretching vibration peak; 1024.05 cm<sup>−</sup><sup>1</sup> represents the stretching vibration absorption peak of -O- in PEG; 700.06 cm<sup>−</sup><sup>1</sup> broad absorption peak represents the out-of-plane rocking vibration absorption peak of NH in bovine serum albumin and NH2-PEG. Repeated centrifugal washing can completely remove the free carboxylated bovine serum albumin in the solution, and the upconversion material is inorganic and will not absorb at these positions. Therefore, it is considered that the new characteristic peaks belonging to organic functional groups can only come from carboxylated bovine serum proteins that have been attached to the surface of upconverted nanomaterials, which cannot be washed away. The above results prove that the surface of NaYF4:Yb3+, Tm3+ nanomaterials contains many carboxyl functional groups, and the carboxylated bovine serum albumin has been successfully modified the surface of water-soluble nanomaterials. Figure 9b shows the excitation and emission spectra of DNA/dBSA/NaYF4:Yb,Tm excited at 480 nm. The excitation spectrum is from 450 nm to 490 nm, and the emission spectrum is from 510 nm to 530 nm, which is mainly the contribution of FAM. As described in

Section 2.6, we carried out repeated centrifugal washing to fully wash the excess DNA, and then obtained the fluorescence spectrum in Figure 9c. As shown in Figure 9c, when the DNA/dBSA/NaYF4:Yb,Tm fluorescent probes were excited at 480 nm, a strong emission peak is observed at 520 nm, which is consistent with the FAM fluorescence peak. It indicates that some DNA strands were not washed away due to their attachment to the carboxylated protein-modified upconversion material, namely DNA was attached successfully to the surfaces of UCNP via a protein and a FAM was added to the UCNP/DNA complex. Similarly, NaYF4:Yb,Tm and DNA/dBSA/NaYF4:Yb,Tm was excited at 980 nm, and the upconversion intensity of the luminescence is quite strong, as shown in Figure 9d. These nanocomposites have strong upconversion luminescence with good water solubility and biocompatibility, making them a new type of fluorescent probe.

**Figure 9.** (**a**) Infrared absorption spectra of NaYF4: Yb,Tm before and after modification with carboxylated bovine serum albumin. (**b**) Absorption and emission spectra of the FAM fluorophore. (**c**) Fluorescence spectra of NaYF4:Yb,Tm and novel fluorescent probes under excitation at 480 nm. (**d**) Fluorescence spectra of NaYF4:Yb,Tm and novel fluorescent probes under excitation at 980 nm.

#### *3.4. Fluorescent Probes for the Detection of Different Proteins*

To test the detection of these upconverted nanomaterials for protein detection, 100 pM solutions of miRNA-155, single-base mismatch of miRNA-155, double-base mismatch of miRNA-155, complete-base mismatch of miRNA-155 and miRNA-150 solution with the same concentration of 100 pM were prepared, respectively. The prepared nucleotide sequences of miRNAs and fluorescent probes are shown in Table 1. The prepared fluorescent probes of upconverted nanomaterials were tested with different miRNAs and mismatched miRNAs, and the samples were excited by a fiber laser with a wavelength of 980 nm and their fluorescence spectra were measured.


**Table 1.** Nucleotide sequences of miRNAs and fluorescent probes.

As shown in Figure 10, the fluorescence spectrum of the fluorescent probe after connecting different miRNA-155 changed significantly. In general, the fluorescence spectra of the four groups were very similar, and the fluorescence intensity decreased significantly (compared with Figure 9d). The sample added with miRNA-155 had the strongest fluorescence intensity. The more mismatched bases, the more obvious fluorescence quenching and the smaller spectral intensity. It is worth noting that the fluorescence quenching at 345 nm, 362 nm, 450 nm, 477 nm and 646 nm is more obvious than that at 802 nm.

**Figure 10.** Fluorescence spectra of fluorescent probes connected to miRNA-155, miRNA-155 singlebase mismatch, miRNA-155 double-base mismatch, and miRNA-155 complete-base mismatch. (**a**) Fluorescence spectra of mismatched sequences relative to miRNA-155 at 980 nm excitation. (**b**) MiRNA-155 with different sequences, the relationship between the maximum intensity of the luminescence peak corresponding to sequences change.

We divided the peak intensity at 802 nm (*I*802) by the peak intensity at 345 nm (*I*345), 362 nm (*I*362), 450 nm (*I*450), 477 nm (*I*477) and 646 nm (*I*646) to calculate a group of fluorescence peak ratios for further analysis of the differences in fluorescence spectra of samples with different miRNA-155s. The results are shown in Table 2. As can be seen in Table 2, the five peak ratios of the fluorescent probes are very close to those of the upconverted nanomaterials; then, the value of the completely mismatched miRNA155 is relatively close to that of the upconverted nanomaterials, and intact miRNA-155 had the greatest effect on all five peak ratios. It seems that the completely mismatched miRNA-155 has little effect on the peak ratios, and the intact miRNA-155 has the greatest effect on the peak ratios. This result may be that miRNAs with different sequences have different effects on different molecular bonds of upconverted nanomaterials. It is believed that these peak ratios can be used for specific recognition of miRNA-155. For fluorescent probes with multiple emission peaks, in addition to identifying the target substance by simply comparing the changes in peak intensity, the ratios between peaks can also be used for substance-specific identification. The fluorescence probe with multiple emission peaks provides more abundant optical

information for the study of the characteristic changes of the detected object and has grea<sup>t</sup> application potential.

**Table 2.** The ratio of fluorescence peaks at 802 nm and 450 nm.


Noting: FP + CmiRNA-155 represents fluorescent probe with completely mismatched miRNA155; FP+ M2miRNA-155 represents fluorescent probe with miRNA-155 double-base mismatch: FP+ M2miRNA-155 represents fluorescent probe with miRNA-155 single-base mismatch; FP + miRNA-155 represents fluorescent probe with miRNA-155.

As shown in Figure 11, when the fluorescent probes were connected to miRNA-155 and miRNA-150, respectively, the fluorescence intensity of miRNA-155 was higher than that of miRNA-150. The experimental results show that the fluorescent probe can effectively distinguish different types of miRNAs.

**Figure 11.** Fluorescence spectra of fluorescent probes connected to miRNA-155 and miRNA-150 under excitation at 980 nm.
