Intracellular Thermal Probing Using Aggregated Fluorescent Nanodiamonds

Abstract Intracellular thermometry provides important information about the physiological activity of single cells and has been implemented using diverse temperature‐sensitive materials as nanoprobes. However, measuring the temperature of specific organelles or subcellular structures is challenging because it requires precise positioning of the nanoprobes. Here, it is shown that dispersed fluorescent nanodiamonds (FNDs) endocytosed in living cells can be aggregated into microspheres using optical forces and used as intracellular temperature probes. The aggregation of the FNDs and electromagnetic resonance between individual nanodiamonds in the microspheres lead to a sevenfold intensity enhancement of 546‐nm laser excitation. With the assistance of a scanning optical tweezing system, the FND microspheres can be precisely patterned and positioned within the cells. By measuring the fluorescence spectra of the microspheres, the temperatures at different locations within the cells are detected. The method provides an approach to the constructing and positioning of nanoprobes in an intracellular manner, which has potential applications in high‐precision and flexible single‐cell analysis.


S2.1. FND Characterization
. a) Fourier-transform infrared spectra (FTIR) of the FNDs with diameter of 100 nm. b) Zeta potential of the FNDs in deionized water. c) Raman spectrum of the FNDs acquired with an excitation laser at 785 nm. d) Absorption and emission spectra of the FNDs. Figure S2. Optical trapping stiffness of the FNDs in aggregation (in x and y directions).

S2.4. Electrostatic adsorption in aggregated FNDs
As presented in Figure S1, there exist negatively charged −OH groups (−8 mV) on the surface of the FNDs in deionized water, which results in the electrostatic repulsion between the FNDs.
Theoretically, when the electrostatic forces between the particles is repulsive, an increase in ionic strength will rise the absorption capacity. [S4] The experiments on the forming and stabilization time of the aggregated FND microspheres in NaCl solution at different concentrations were carried out, as shown in Figure S4. The trapping laser power was set as 100 mW, and the FND concentration is 0.05 mg/ml. As a control group, the formation and stabilization times of the FND aggregates were 20 and 38 s with a diameter of 2.0 μm in deionized water, respectively. With the concentration of the NaCl solution increased to 0.01 (inset I of Figure S4) and 0.02 mol/L (inset II of Figure S4), the formation time was shortened to 4.2 and 4 s while the stabilization time was shortened to 12 and 5 s, respectively. Note that 5 as the concentration of the NaCl was further increased to 0.03 mol/L (inset III of Figure S4), the strong electrostatic adsorption between the FNDs caused an aggregation in large numbers and the spherical shape of the aggregate was no longer maintained.

S2.5. Roundness of aggregated FND microspheres
The roundness of the aggregated FND microspheres is evaluated by estimating the roughness acquired from the SEM images ( Figure S5). Here the roughness is defined as the average difference between the diameters of the FND microspheres and the diameter of an absolute sphere. The results show that the roundness gets smaller with the increase of the diameter of the aggregated FND microspheres ( Figure S5e).
6 Figure S5. a-d) The SEM images of aggregated FND microspheres with diameter of 2.5 (a), 2.0 (b), 0.6 (c) and 0.4 μm (d), respectively. e) Statistics of roughness and diameter of the aggregated FND microspheres.

S2.6. Minimum size of stable FND aggregates
The minimum size of the stable aggregates was 0.4 μm, which was formed by applying a trapping laser power of 50 mW within irradiation duration of 5 s, as shown in Figure S6.
Theoretically, smaller aggregates can be obtained with lower trapping power or shorter irradiation time. However, the roughness of the aggregates, as to be discussed previously ( Figure S5), will be higher than 0.08 μm, which cannot be considered as a spherical structure. Therefore, the smallest size of aggregate FND microsphere was 0.4 μm that contained ~800 individual FND particles to ensure a satisfactory roundness.

S2.7. Aggregation process at different FND concentrations
The FND concentration affects the aggregate diameter and the formation and stabilization time, as shown in Figure S7. At a fixed power of 150 mW and an irradiation time of 5 s, the diameter of the FND aggregates gradually increases with the concentration (Figure S7a), while the formation and stabilization time decreases ( Figure S7b).  Figure S8. Bright-field microscope and fluorescence images of aggregated Au (a1,2), BaTio 3 (b1,2) and Si nanospheres (c1,2) with diameters of 60, 100, and 80-200 nm, respectively.

S2.9. Aggregation of other temperature-sensitive nanoparticles
Although the size of aggregates can be controlled by the concentration, trapping laser power and irradiation duration, it is strongly related to the size of the individual nanoparticles that form the aggregates. As shown in Figure S9 (Table S1).

S3.2. Comparision of aggregated and dispersed nanoparticles
For the dispersed single probes, such as FNDs (100 nm), UCNPs (35 nm) and QDs (5 nm), the fluorescence intensity (I 0 ) is usually weak so that the fluorescence imaging and detection are relatively difficult ( Figure S12a). By forming the dispersed probes into the aggregates, the fluorescence intensity (I 1 ) becomes much stronger at the same excitation power ( Figure S12b).
Statistically, the enhancement factor f of the fluorescence intensity, defined as f = I 1 /I 0 , is at least 4.3 ( Figure S12c). This means although the single probes dispersed in the cells can be

11
The power reduction σ is defined as σ = (P 0 −P 1 )/P 0 , where P 0 is the excitation power for the dispersed nanoparticles and P 1 is the excitation power for the aggregates. targeted to specific locations, they require a relatively high excitation power to detect multiple regions, which may cause certain damage to the cell. With the method of aggregating the FNDs, the excitation power can be reduced by up to 90% to obtain the same fluorescence intensity ( Figure S12d-f).

S4.2. FND aggregation in different cell types
To investigate the effect of different cell types on the FND aggregate formation at the same concentration of FND solution ( Figure S14a). It can be seen that attributed to the stronger endocytosis ability of the Ana-1 cells (Phagocyte), [S5] more FNDs were endocytosed and aggregated in the Ana-1 cells than those in other cells ( Figure S14b). Note that the aggregates in the HeLa cell in Figure S14b2 was smaller than that in Figure 3c3 in the manuscript because the laser irradiation duration was 1 min less. At the same FND concentration, the amount of the FNDs endocytosed by the cells was much less than that dispersed in deionized water, so the time required to stabilize the FND aggregates became at least 5 times longer in the cells ( Figure S14c). Because the endocytosed FNDs can be transported around the nucleus by the myosin so that it took less time to assemble the aggregates around the nuclei of the cells (red pillars in Figure S14c). Due to the limited endocytosis capacity of the cells, the diameters of the FND aggregates were smaller than that formed in deionized water under the same trapping laser power and duration of irradiation ( Figure S14d). In addition, the larger size of the FND aggregates in water resulted a stronger final fluorescence intensity than those in the cells ( Figure S14e).

S4.4. Cell viability
To investigate the possible cytotoxic effects caused by the trapping laser or the FND aggregates, an experiment was carried out to indicate the cell viability after trapping the FNDs by the optical tweezers in living cells ( Figure S16). In each group of the experiments, the effect was investigated for different trapping powers ( Figure S16a), irradiation durations ( Figure S16b) and amounts of FND aggerates ( Figure S16c). The cell viability was indicated by staining the cells with the cell permeable nucleus counterstain (Hoechst 33342) after aggregating the FNDs in the cells. The results show that for the cells treated with an irradiation of 3 min, the survival rates were higher than 80% at trapping powers lower than 150 mW ( Figure S16d). With a power fixed at 150 mW, the survival rates reached above 80% within irradiation duration of 5 min ( Figure S16e). In addition, the endocytosis of the FNDs has little effect on the cell survival because of the good biocompatibility ( Figure S16f).

S4.5. Microsphere precision in the cell
The precisions of the microsphere in the x and y directions can be experimentally measured, as shown in Figure S17. With a trapping laser power of 60 mW (the minimum power required to manipulate the aggregates in the cell), the vibration of an aggregated FND microsphere (2 μm) in the x and y directions was recorded ( Figure S17a), indicating that the precisions of the x and y directions were 60 and 70 nm, respectively. In general, the precision in the z direction 15 is twice of that in the x or y direction. By increasing the trapping laser power, the precisions in three dimensions can be improved ( Figure S17b). For the case in Figure 5, the trapping laser power was 150 mW, and the precisions of the x, y, and z were 27, 28, 56 nm, respectively.
The trapping laser was applied near the outline of the cell membrane determined by the bright-field observation. After 3 minutes of laser irradiation, the FND aggregates were formed near the membrane and excited with 546-nm laser in dark field.

S4.6. ZPL measurement at 40 °C
At each temperature between 20 and 60 °C, four aggregated FND microspheres particles were used for the ZPL measurement. For example, the ZPLs of the four aggregates were measured at the temperature of 40°C ( Figure S18), and the wavelength error was 0.06 nm. Figure S18. Fluorescence spectra of four aggregated FND microspheres in water at 40 °C.