In a typical experiment, C@SiO₂ PC beads were fabricated by the evaporation of droplet templates containing monodisperse PDA@SiO₂ nanoparticles followed by pyrolysis. The detailed fabrication for C@SiO₂ PC beads is shown in Figure 1(a). Firstly, PDA@SiO₂ particles were obtained by polymerization dopamine forming a PDA shell on the uniform SiO₂ particles in 10 mM tris-buffer. Then, PDA@SiO₂ particles were assembled into spherical PC microspheres (PDA@SiO₂ PC beads) with a droplet-based microfluidic technique. Finally, C@SiO₂ PC beads were obtained after pyrolysis treatment of PDA@SiO₂ PC beads in the absence of oxygen. In the microfluidic system, the PDA@SiO₂ particles could directly assemble into the stable PC beads without any extra treatment due to the PDA binder properly around each building block. Figure 1(b) shows the transmission electron microscope (TEM) image of the prepared PDA@SiO₂ particles with 252 nm particle sizes. These building block-PDA@SiO₂ nanoparticles with thin-coated PDA shell presented an admirable spherical shape and good monodispersity, which showed a brilliant red color after centrifugation (Figure S1). The standard spherical PDA@SiO₂ PC beads displaying a homogeneous slightly yellowish structural color were obtained after drying the droplet containing PDA@SiO₂ nanoparticles (Figure S2). Followed by pyrolysis, the C@SiO₂ PC beads, as shown in Figure 1(c), were obtained and exhibited a more visible and luminous green color under natural light, superior to the SiO₂ PC beads (Figure S3). The nanoparticles of the barcode beads mainly formed spherical assemblies with a regular arrangement and ultimately performed as a close-packed colloidal crystal array structure in the sphere (Figure 1(d)). The barcode beads were derived from microfluidic droplets that contained monodispersed silica nanoparticles, and their initial size and subsequent variations after being PDA-coated and carbonized could be customized from several to hundreds of micrometers. The adjustable preparation strategies are performed not only by varying the flow rates of the water and oil phases but also by using different concentrations of the nanoparticles for droplet generation. Here, barcode particles with a diameter of about 300  μm were generated for the functional and morphological characterization as cell-culturing and capture carriers.

Based on the optical principle of PC, the central diffraction wavelength (structural color) of the C@SiO₂ PC beads is estimated by Bragg’s equation under normal incidence [18],where \lambda is the central diffraction wavelength, d is the center-to-center distance between two nanoparticles, and n_a is the average refractive index of C@SiO₂ beads. By changing the diameters of the C@SiO₂ beads, four classes of colored C@SiO₂ beads were obtained (Figures 2(a) and S4). These beads displayed visible colors under natural light that can be clearly distinguished by the naked eyes. The signal capture of the visible color is schemed in Figure 2(b), (i). Since the incident light may be acquired from various directions, the color signal range of the PC microspheres can be basically equal to their volume. Nevertheless, the spectrum signal is usually obtained by detecting the diffraction spectrum via injecting incident light perpendicular to the vertex tangent of the microspheres (Figure 2(b), (ii)). This means that the detected signal comes from only one point of the microspheres. The optical photographs in Figure 2(c) also demonstrated that the spectral signal is just one point located on the top of the spheres. By contrast, the signal area of the spectral mode only accounts for 0.08 parts of the structural color mode (Figure 2(d)). These results indicated that the structural color-encoded beads have a significant advantage on the signal acquisition than the spectrum encoded for PC barcodes.

Mechanical strength and stability are the main factors to be considered in the application of barcode beads, especially in cell culture and capture. The highly color visible C@SiO₂ PC beads are converted by pyrolysis of coating PDA, which plays a role as an adhesive polymer and increases the PC beads’ binding strength. As expected, the mechanical properties, i.e., compression resistance and ultrasonic damage resistance, increased upon PDA to C conversion. The compression tests were implemented by NanoTest, which were to test the load stresses of beads to reach 25  μm depths (Figures S5-S8). Figure 3(a) shows the resulted curves of load vs. depth for SiO₂ PC beads, PDA@SiO₂ PC beads, sintered SiO₂ PC beads, and C@SiO₂ PC beads. The PDA@SiO₂ PC beads endured nearly 82\pm 5\text{mN} stresses before breaking, which were double of the SiO₂ PC beads. After pyrolysis, the endured pressure increased, with values of 149\pm 2\text{mN} for C@SiO₂ PC beads and 120\pm 15\text{mN} for sintered SiO₂ PC beads. Furthermore, the residual rate after suffering two minutes of ultrasonic treatment increased after pyrolysis. As shown in Figure 3(b), the average residual rate was calculated to be 0.12 for SiO₂ PC beads, 0.48 for PDA@SiO₂ PC beads, 0.59 for sintered SiO₂ PC beads, and 0.95 for C@SiO₂ PC beads. As a result, compared to the initial SiO₂ PC beads, the C@SiO₂ PC beads exhibited remarkably increased compression resistance (365%) and ultrasonic damage resistance (792%). In order to understand the reasons for the increase of mechanical strength for C@SiO₂ PC beads, the section-cross of the above beads were investigated by the field emission scanning electron microscopy (FESEM), as shown in Figures 3(c) and S9. Except for SiO₂ PC beads, many grains were observed between the particles of PDA@SiO₂ PC beads, sintered SiO₂ PC beads, and C@SiO₂ PC beads. Raman spectra of SiO₂ PC beads, PDA@SiO₂ PC beads, and C@SiO₂ PC beads confirmed that the grains of PDA@SiO₂ PC beads could be contributed to a self-adhesion reaction of PDA shells, the grains for sintered SiO₂ PC beads might be the fusion of SiO₂ particles by pyrolysis, and the grains for C@SiO₂ PC beads may be due to the pyrolysis of PDA shells (Figure S10). Regardless of the fact that grains existed in all these PC beads, C@SiO₂ PC beads own the strongest mechanical strength. This may be attributed to the graphitic structure which was formed by pyrolysis of PDA, which is more stable than that of the PDA supermolecule and even stronger than the silicon-oxygen bond between the sintered SiO₂ PC beads. XPS was further used to analyze the chemical composition and confirm this speculation. The peaks at 284.8, 401.1, and 532.1 eV correspond to C1s, N1s, and O1s in the PDA@SiO₂ beads and C@SiO₂ beads (Figure S11), respectively. After pyrolysis, a significant decrease in the N1s and O1s peaks was observed due to chemical reduction and graphitization [33]. High-resolution N1s and C1s spectra from the C@SiO₂ beads revealed significant differences in the carbon chemistry of the PDA. The high-resolution N1s spectra indicated the presence of graphitic N at 400.9 eV, pyrrolic N at 399.0 eV, and pyridinic N at 398.4 eV (Figures 3(d) and 3(e)). The major N1s peak observed in the PDA@SiO₂ beads corresponded to pyrrolic N (57.55%), whereas the N1s peak observed in the C@SiO₂ beads corresponded to graphitic N (82.77%). The high-resolution C1s spectra exhibited similar shifts after pyrolysis. The high-resolution C1s spectra indicated the presence of C-H bonding at 281.3 eV, C-C bonding at 284.5 eV, C-O bonding at 285.5 eV, C-N sp² bonding at 286.1 eV, C-O-C bonding at 286.8 eV, and C-N sp³ bonding at 288.2 eV (Figure S12). The C-H level decreased significantly after pyrolysis, from 15.58% (PDA@SiO₂) to 3.17% (C@SiO₂). Moreover, the level of C-O bonding increased from 2.26% (PDA@SiO₂) to 32.68% (C@SiO₂). As a result, the strong graphitic interaction and additional interaction area induced by template pyrolysis PDA endow the C@SiO₂ PC beads with a strong mechanical strength, which brings great convenience to the real application of PC barcode beads.

The development of PC barcodes should be devoted to gain a low ontology fluorescence to improve the detection signal-to-noise ratio in fluorescence-reading applications. To demonstrate the fluorescent property of C@SiO₂ PC beads, the sintered SiO₂ PC beads, PDA@SiO₂ PC beads, and C@SiO₂ PC beads were compared excited by different light (Figure 4(a)). As shown in Figure 4(b), the fluorescence intensity of C@SiO₂ PC beads is 0.36 of the sintered SiO₂ PC beads under ultraviolet (UV) light (EX: 361~385 nm), 0.25 under blue light (EX: 465~495 nm), and 0.20 under green light (EX: 540~580 nm). Adjusting the exciting light intensity to different degrees, the ultralow ontology fluorescence of C@SiO₂ PC beads remains unchanged. Obviously, the ultralow fluorescence intensity of C@SiO₂ PC beads is more remarkable when the exciting light tends to a short light (Figure 4(c) and Figure S13). Based on the above structure analysis on C@SiO₂ beads, the ultralow ontology fluorescence of C@SiO₂ PC beads may result from pyrolysis eliminating the organic small molecules which could produce fluorescence in the original PDA@SiO₂ beads, thereby reducing the ontology fluorescence of the C@SiO₂ beads [34, 35]. Furthermore, despite the fact that the addition of PDA can reduce the ontology fluorescence of PDA@SiO₂ PC beads, the extent of the reduction is very limited, only about 0.80 of the sintered SiO₂ PC beads. Thus, the advantage of ultralow ontology fluorescence may allow C@SiO₂ PC barcodes to improve the detection sensitivity through depressing the ontology signal interference.

Achieving both high mechanical stability and color visibility, it is important to ensure the biocompatibility of the developed C@SiO₂ PC barcodes for further biomedical applications [9, 13]. The cytocompatibility of C@SiO₂ PC barcodes was assessed by quantitative analysis of cell proliferation activity by MTT assay and CCK-8 tests and qualitative analysis of morphology from fluorescence micrographs by coculturing with human embryonic lung fibroblasts (MRC-5) cells and comparison with the control, sintered SiO₂ PC beads, and PDA@SiO₂ PC beads. The final OD values in direct proportion to the numbers of viable MRC-5 cells are given in Figure 5(a), which reveals the viability of MRC-5 cells cocultured with control, sintered SiO2 PC beads, PDA@SiO₂ PC beads, and C@SiO₂ PC barcodes. The OD value of MRC-5 cells cocultured with each kind of beads showed a significant increase from day 1 to day 7, indicating that cell proliferation activity increased sharply in a manner similar to growth on native extracellular matrix or tissue culture plates (TCPs). Compared with the cell viability on day 5, the OD value on day 7 did not have a significant increase, suggesting the saturation of cells on TCPs in the same culture medium. Furthermore, cell proliferation assays demonstrated that the contained SiO₂ nanoparticles, PDA, and carbonated PDA did not release any chemical pigment resulting in acute cytotoxicity and did not significantly influence cell viability. Before a typical cell experiment, CCK-8 tests have also demonstrated that the C@SiO₂ barcodes present excellent biocompatibility coculturing with MRC-5 (Figure S18).

The biocompatibility of the C@SiO₂ PC barcodes was also further confirmed by observing the morphology of cells cocultured with them for 5 and 7 days via an inverted fluorescence microscope (IFM) (Figures 5(b) and 5(c)). The viable cells can be visually observed after calcein AM staining with green fluorescence. MRC-5 cells have spindle morphology and polarity with long axis direction. There will be some vacancies among cells for contact inhibition when the long axis direction of cells has contact with each other after quick extension and proliferation as shown in Figure 5(b). Meanwhile, this morphology of MRC-5 cells proliferated coculture with sintered SiO₂ beads, PDA@SiO₂ beads, and C@SiO₂ beads similar to TCPs as shown in Figure 5(c), indicating that these barcodes have excellent cytocompatibility to promote cell proliferation. Furthermore, cell density showing a significant increase when cultured for 5 days further verified the accuracy of the MTT cell proliferation assay. Therefore, barcodes possessed excellent biocompatibility with improved proliferation viability and good phenotypic shape of MRC-5 cells, which should find various promising applications in the fields of fluorescence-reading-based multibioassays.

To test the superiority of C@SiO₂ beads’ usability, sintered SiO₂ PC beads, PDA@SiO₂ PC beads, and C@SiO₂ PC beads were implemented to capture 5-8F cells, which were labeled by corresponding fluorescent probes that gave blue and green fluorescence to the nucleus and cell membrane, respectively. Figure 6(a) exhibited the fluorescent microscopy images of the staining cell nucleus (Hoechst 33342), staining cell membrane (DiO), and merged images of Hoechst 33342 and DiO of SiO₂ PC beads, PDA@SiO₂ PC beads, and C@SiO₂ PC beads. These images stated that C@SiO₂ carriers showed an ultralow ontology fluorescence, which could even be completely ignored, compared with SiO₂ PC beads. Figure 6(b) and Figures S14-S16 further revealed different exposure-time-obtained fluorescence signal-to-noise ratios between Hoechst 33342 staining cell nucleus (i), DiO staining cell membrane (ii), and merged images of Hoechst 33342 and DiO (iii) of the captured 5-8F cells with the PLL-coated sintered SiO₂ beads, PDA@SiO₂ beads, and C@SiO₂ beads, respectively. The results showed that C@SiO₂ beads are available to eliminate the interference of ontology fluorescence and improved the signal-to-noise ratio by an average of 2.5 times than the sintered SiO₂ PC beads (Figure S17). Notably, although PDA@SiO₂ also was capable of depressing ontology fluorescence and thus increased the signal-to-noise ratio to some extent, the increased capacity was not ineffective for the Hoechst 33342 staining cell nucleus. The discrepancy may result from the transformation of PDA to C-PDA. The phenolic hydroxyl group and amino group in PDA have a strong electron ability, which promotes the efficiency of energy transfer and/or electron transfer when PDA@SiO₂ beads contact with fluorescent molecules. But after pyrolysis, these electron-withdrawing groups are destroyed, and the ability to quench the fluorescence of C@SiO₂ beads disappears.

Cell analysis plays a key role in the biomedical diagnosis [36–38]. In the real case of biomedical diagnosis, analysis for only one kind of cells is not sufficient to diagnose a specific disease from the extraordinary complexity of biological specimens [39, 40]. According to the unique optical encoding properties, sintered SiO₂ PC beads have served as cell capture carriers for the capture of various circulating tumor cells. Moreover, the surface of the barcode beads with a spherical periodic array topography further realized cell analysis and release for multiple types of circulating tumor cells simultaneously [9]. C@SiO₂ PC barcodes employ the characteristic reflection peaks as their distinct encoded elements and their peak positions are based on the periodic structure or refractive index of the materials and remain constant during cell capture, adhesion, and culture on their surface; this indicates the high encoding accuracy of the microcarriers in multiplexing [6, 17]. C@SiO₂ PC barcodes provide a nanopatterned surface topography with ordered hexagonal symmetry of the nanoparticles and reduce steric hindrance and increased density of the biomaterial molecules, such as protein antibodies (e.g., antiepithelial cell-adhesion molecule) or nucleotide probes (e.g., DNA aptamers) [9, 10]. They are free to interact with some specific surface proteins of multiple cells and, thus, increase the efficiency of the cell adhesion and targeted capture. An important requirement for using barcode particles for multiple cell capture and detection is the accurate identification of their coded information during the entire process of cell capture. To demonstrate the reliability of the barcode particles in capturing and detecting multiple types of cells, three types of barcode particles that exhibit red (A), green (B), and blue (C) structural colors were modified with three types of targeted probes (A’, B’, C’), respectively. These targeted probes modified barcode particles which were then mixed and incubated in a multitarget solution of red fluorescence-stained cells (A”), green fluorescence-stained cells (B”), and blue fluorescence-stained cells (C”) (Figure 7(a)). Because of the specific binding between the targeted probes and their corresponding target cell types, we expected to observe the specific cells on the surface of the barcode particles when their corresponding targeted probes were present.

The C@SiO₂ PC beads were placed in a system of cultured human nasopharyngeal carcinoma (5-8F) cells. Because of the small number of chemical groups on the C@SiO₂ surface that enable cell adhesion, the practical capture density of cells on the barcode beads is limited (Figure S19). To solve this problem, polylysine (PLL), widely used to promote cell adhesion, was modified on the surface of the beads [41, 42]. The process of 5-8F cells captured on the surface of C@SiO₂ PC beads is illustrated in Figure 7(b). The C@SiO₂ barcode beads were decorated by PLL molecular, and thus, they not only provided a large surface area and close-packed hemispheric array surface topography for cell capture but also provided a specific procell adhesion platform to increase the efficiency of cell capture. As shown in Figure 7(c), (i), the surface of the C@SiO₂ beads after being modified by PLL was still arranged with a hexagonal close packing. Taking into account this spherical array topography, 5-8F cells could successfully adhere to as well as extend to the whole surface of the sphere (Figure 7(c), (ii, iii)). To confirm the cell-captured capability, 5-8F cells were cultured for 6 days in the cultured medium mixed with PLL decorated C@SiO₂ carriers. The everyday condition of cells on the surface of beads was observed by a light microscope. As shown in Figure S20, with the growth of cultured time, the 5-8F cells gradually increased on the beads and even stacked with each other to cover the whole sphere to form a film with a thickness of up to 32  μm. More importantly, the shape of the C@SiO₂ beads remained unchanged for six days in the complex cell culture medium, indicating that the high stability of the beads is enough for a wider range of applications, superior to PDA@SiO₂ beads (Figure S21). The optical microphotograph in Figure 7(d) revealed that C@SiO₂ beads covered with thick 5-8F cells still maintained their initial brilliant yellow color. The corresponding fluorescent microscopy images of the cell-captured beads with a staining cell nucleus (Hoechst 33342), staining cell membrane (DiO), and merged images of Hoechst 33342 and DiO were shown subsequently. These fluorescent microscopy images exhibited that the ontology fluorescence of the C@SiO₂ carriers was basically negligible, so that the fluorescent information of the surface cells can be read out without consideration of background interference (Figure S22). Therefore, the C@SiO₂ carriers simplified the extraction of cellular fluorescence signal by eliminating the interference of carriers’ fluorescence during the analysis of cellular information. All these results demonstrated that the C@SiO₂ PC beads not only could maintain well both high mechanical stability and color visibility during cell capture and culture but also performed an excellent ability on fluorescent-reading after staining by fluorescence makers.

3-Hydroxytyramine hydrochloride (DA·HCl), tris(hydroxymethyl)aminomethane (Tris), and hexadecane were purchased from Aladdin (Shanghai, China). SiO₂ particles with different diameters were purchased from Nanjing Nanorainbow Biotechnology Co., Ltd. (China). Hoechst 33342 and 3,3-dioctadecyloxacarbocyanine perchlorate (DiO) were purchased from Shanghai Beyotime Biotechnology Co. Ltd. (Shanghai, China). The human NPC 5-8F cell line was provided by the Research Center of Clinical Oncology of the Affiliated Jiangsu Cancer Hospital (Nanjing Medical University, Nanjing, China). Polylysine (PLL) and fetal bovine serum without mycoplasma were purchased from Tianhang Biological Technology Co., Ltd., Zhejiang. PBS (\text{pH}=7.4) was laboratory homemade. Dimethyl sulfoxide (DMSO) was purchased from Sigma (USA). Deionized water (18.0 MΩ cm, Milli-Q Gradient System, Millipore) was used in all experiments. All chemical reagents were used without further purification.

PDA@SiO₂ was prepared by dispersing 0.1 g SiO₂ particles into 10 mL DA and Tris solution (\text{pH}=8.5) for stirring for 18 h. Then, the obtained particles were cleaned by centrifugalization for three times and further purification with a filter membrane.

Firstly, the particles of PDA@SiO₂ were homogeneously dispersed in water, with the particle quality fraction of 2%. Subsequently, the particle solution was used as the dispersed phase, and hexadecane with 1% Hypermer 2296 was used as the continuous phase. Then, when the hexadecane and aqueous suspension were simultaneously injected into the PTFE tube, the aqueous suspension was broken into droplets by the hexadecane flows at the needle tip. The hexadecane took the suspension droplets into the collection container which was filled with hexadecane with 2% Hypermer 2296 and maintained heated at 75°C, and then, solid PDA@SiO₂ beads were derived and then were thoroughly washed with hexane to remove the hexadecane. At last, the PDA@SiO₂ beads were pyrolyzed at 400°C for 8 hours with nitrogen protection.

The human MRC-5 cells were cultured in DMEM (Corning, Manassas, VA, USA) supplemented with 5% FBS (Gibco, Grand Island, USA) in the presence of 5% CO₂ at 37°C. The assay was carried out in a 96-well plate. The barcodes were sterilized with 75% alcohol and washed twice with PBS and then incubated in medium overnight. The number of cells in each well was 1\times {10}^4. After MRC-5 cell seeding, the 96-well plate was placed in the incubator for 24 h at 37°C and 5% CO₂. Then, we transferred the substrates to a new 96-well plate to get rid of the cells adhering on the plate, and new medium and 20  μL of MTT (98%) solution in PBS (5 mg mL^-1) were added into each well and incubated at 37°C for 4 h. The culture medium in each well was then removed, and 200  μL of dimethyl sulfoxide (DMSO, Sigma, USA) was added to completely dissolve the formazan crystals formed in the cells on the substrates. The absorbance of each well at 490 nm was measured by a microplate reader (Synergy HT, BioTek, USA).

We used fluorescence staining to detect the cell viability of MRC-5 cells. Calcein AM (molecular prober, USA) is a kind of dye, which can penetrate cells and be used for determining the vitality of most eukaryotic cells. In living cells, a nonfluorescent calcein AM will convert to green fluorescent calcein after the intracellular esterase hydrolyzes acetoxymethylester. We incubated the MRC-5 cells cocultured with barcodes in 10  μM calcein AM solution for 20 min in an incubator. Finally, the stained MRC-5 cells were rinsed in PBS and observed by fluorescence microscope (Olympus IX71, Olympus, Japan).

Polylysine with concentration of 0.01% was prepared in a PBS solution. C@SiO₂ beads were firstly washed with ethanol (once) and PBS (three times). Then, the washed beads were immersed in the 0.01% PLL solution at room temperature. After 24 hours, the excess liquid was wiped off, and C@SiO₂ beads were washed repeatedly by PBS solution for three times with a sterile operation.

The human NPC cell line (5-8F) was cultured in RPMI-1640 (Corning, Manassas, VA, USA) supplemented with 5% FBS (Gibco, Grand Island, USA) in the presence of 5% CO₂ at 37°C. Before the experiment, SiO₂, PDA@SiO₂, and C@SiO₂ beads needed to be rinsed once with absolute ethanol and three times with PBS and soaked in 0.1 mg/mL polylysine (PLL) overnight. Then, digested cells in a logarithmic growth phase with trypsin were resuspended slightly, and the resuspended cells (3 mL, 1\times {10}^5 cells per mL) were added to new dishes which contained PLL-coated SiO₂, PDA@SiO₂, or C@SiO₂ beads. In the following 6 days, photos were taken with the light microscope to evaluate the mechanical strength. Meanwhile, in order to compare the ontology fluorescence of the beads and label these captured cells, a part of the beads with cells cultured to day 5 was selected, and the cell membrane and nucleus were stained with DiO (Beyotime, China) and Hoechst 33342 (Beyotime, China), respectively, and fluorescence photos were then obtained by a fluorescence microscope.

Ultrasonic damage test was conducted in the bottom of the ultrasonic instrument with thirty SiO₂ beads, PDA@SiO₂ beads, sintered SiO₂ beads, and C@SiO₂ beads soaked in a tube with 2 mL water. After two minutes of ultrasonic treatment, the final residual bead numbers (the integrated spheroidal was regarded as one bead number) were counted by the metalloscope (Olympus BX51). The compression test was performed by the Nantes with set-up parameters (maximum load was 150 mM, loading rate was 0.3 mN/s, and unloading rate is 5 mN/s). The beads lined up under the crosshair of the multiple-objective microscope and then were compressed by a 100  μm radius spherical diamond probe (Figure S4).

The SiO₂@PDA images were obtained by Transmission Electron Microscopy (TEM, JEM2100F). The images of appearance, surface, and cross-section of C@SiO₂ beads and the cells on C@SiO₂ beads were obtained by a Field Emission Scanning Electron Microscope (FESEM, Zeiss Ultra Plus). The surface element analysis of beads was implemented by X-ray photoelectron spectroscopy (XPS, GB-T19500-2004). The Raman spectra were measured using an inVia Renishaw Raman microscope system (Renishaw, New Mills, UK). The microphotographs of the beads were taken by the metalloscope (Olympus BX51) with a CCD camera (Media Cybernetics Evolution MP 5.0). The absolute reflectance of PC beads was measured by a spectrometer (QE65000, Ocean Optics) with light source (DH-2000UV-VIS-NIR, Mikropack) and optical fiber (QR200-7-UV-BX, Ocean Optics) using a diffuse reflection standard plate (WR-D97-30, Oceanhood) as a completely diffuse reflector. 1931 CIE coordinates were obtained from absolute reflectance calculated by a specific software. Fluorescence images of DiO, Hoechst 33342, and merged DiO and Hoechst 33342 were measured by an inverted fluorescence microscope (Olympus, MVC10). The cells on the beads for FESEM were pretreated with a graded series of concentrations of ethanol (0%, 25%, 50%, 75%, and 100%).