**3. Results**

### *3.1. Synthesis and Characterization of Particles*

Figure 1a–f present the TEM images of particles tested in this study. In general, these particles had spherical or pseudospherical morphologies. Pristine LDH particles showed the characteristic flower-like morphology with hexagonal shape of the flaky sheets of LDH crystallites (Figure 1a). Significant changes in morphological features were observed when pristine LDH was modified by intercalation of casein. The presence of casein in the CLDH composite gave rise to the ordered inorganic lamellae represented by the darker lines (Figure 1b), which was similar to the Ca-Al-casein LDH reported earlier [27]. The schematic diagram (Figure 1g) shows the possible mechanism of casein intercalation through anion exchange when the negatively charged groups (–COO−) of casein gets electrostatically adsorbed between the positively charged LDH layers. The TEM micrograph of CCLDH (Figure 1c) showed mesoporous plates and rod-like cores of CLDH rendered by ordered CMC. Figure 1h demonstrates the scheme of CMC layered onto the CLDH particles. RS showed amorphous morphology with larger discrete particles with a size of ~5 μm (Figure 1d). E171 showed more consistent shapes with an estimated particle size of 100–120 nm (Figure 1e). Size distribution analysis (Figure S1) showed that ~20% of particles in E171 had size less than 100 nm (to be categorized as nanoparticles). Similar to previous observation from our group, E551 (food grade silica) were heavily agglomerated where the size of individual agglomerate was ~30 nm [28].

**Figure 1.** TEM micrographs of layered double hydroxide (LDH) powders: (**a**) LDH, (**b**) CLDH, (**c**) CCLDH, (**d**) Rice Starch, (**e**) E171, (**f**) E551, (**g**) scheme showing anionic exchange facilitating the intercalation of casein molecules between the brucite layers, (**h**) schematic of the CCLDH composite.

These particles were further characterized for their agglomeration behavior, surface charge, surface chemistry, and crystallinity. Based on the DLS results (Figure 2a), poly-

dispersity of CLDH decreased significantly from pristine LDH but showed a significant increase after CMC adsorption. Since the LDH samples were poly-dispersed, it was hard to evaluate the mean size based on the TEM images; thus, the size distributions of the particles were measured in solution by DLS (Figure S2). The hydrodynamic diameter (Figure 2a) of pristine LDH was larger (>1 μm) in comparison to CLDH (<500 nm). The average diameter by volume (Figure 2a) of CCLDH had also increased compared to CLDH. The zeta potential (Figure 2b) of pristine LDH was +10.24 mV while that of casein intercalated (CLDH) was +6.03 mV. CCLDH showed a net negative zeta-potential of −11.96 mV. Zeta potential values of CLDH were close to those of pristine LDH than that of casein, suggesting that the LDH accommodated the casein molecules between the LDH brucite layers. CCLDH however showed a negative zeta potential value because of the surface adsorption of CMC onto CLDH.

**Figure 2.** (**a**) Hydrodynamic diameter (bar graph) and polydispersity index (line graph) of particles (50 μg/mL) suspended in water. \* Indicates statistical significance (*t*-test) in polydispersity (PdI) of CLDH in comparison to pristine LDH and CCLDH compared to CLDH, *p* ≤ 0.05, *N* = 2; (**b**) Zeta potential of LDH particles and its composites; (**c**) FTIR spectra of the particles; particles were placed directly on the ATR probe prior recording the spectra for a wavenumber range of 4000–400 cm<sup>−</sup><sup>1</sup> using ATR–FTIR; (**d**) XRD spectra of the tested particles. The spectra of dry powder of samples were captured using Bruker D8 Discovery X-ray Diffractometer.

ATR-FTIR was performed to assess the surface groups present in the LDH composites and the commercially used white food pigments, and the result is summarized in Figure 2c. The peak of 3692.09 cm<sup>−</sup><sup>1</sup> on the LDH graph features the vibration of −OH groups present in the LDH inorganic Mg (OH)2 layer [29]. The sharp peak at 1356.00 cm<sup>−</sup><sup>1</sup> indicates the adsorption of CO2 contamination from air during the collection process which is difficult to be excluded during LDH synthesis [30]. This contamination occupies some sites, which may lead to incomplete anionic exchange for casein intercalation [30].

The strong peak at 1388.11 cm<sup>−</sup><sup>1</sup> [due to ν(CO3 <sup>2</sup>−)] and bands below 1000 cm<sup>−</sup><sup>1</sup> (due to M–O vibrations and M-O-H bending of LDH) are all characteristics of Mg-Al LDH [28,31–33]. The peak at 450.78 cm<sup>−</sup><sup>1</sup> in CLDH is a unique characteristic of Mg-Al LDH materials at 2:1 ratio [31]. This strong peak and a broad peak at 610.75 cm<sup>−</sup><sup>1</sup> are attributed to the lattice vibrations of Mg2Al-OH layer [34]. The band at 1443.18 cm<sup>−</sup><sup>1</sup> from CLDH is attributed to the stretching vibration of C = O symmetric stretching, verifying the presence of casein. The FT-IR spectrum of CCLDH is the combination of both the CMC and CLDH spectra, which is the good reason for the successful preparation of CCLDH. The casein peak of 1658.58 cm<sup>−</sup><sup>1</sup> had also broadened to 1617.08 cm<sup>−</sup>1.

The phase purity and crystallinity of the CLDH and CCLDH were analyzed and compared to the pristine LDH, alongside other industrial used white pigments by measuring XRD peaks (Figure 2d). In general, the LDH particles possessed defined basal diffraction peaks, suggesting their crystallinity. The varied phase composition of the synthesized LDH-composites indicated the presence of a variety of hydroxides and oxide hydroxides. The diffraction peaks at 15.3◦ for a rhombical symmetry is characteristic of hydrothermally synthesized LDH [19]. Strong diffraction peaks of the LDH due to a long aging time are in good agreemen<sup>t</sup> with those reported for Mg-Al-OH LDH [19]. The larger and more intense peak of LDH is attributed to its larger crystallite size compared to the smaller ones (less intense diffraction peaks), suggesting a correlation of size-crystallinity in LDHs [35]. Notably, the increased intensity of the rhombohedral symmetry peak at 23.6◦ in the composite material suggested successful intercalation of casein into the crystal structure. This finding concurs with the FTIR results as well as with the results from TEM images. LDH peaks disappeared in CCLDH suggesting the rendering of the negatively charged CMC over the positively charged surface of LDH [36]. All in all, physicochemical characterization of parental LDH and variants synthesized demonstrated a typical floral morphology of primary particle size from 100 to 1000 nm, well dispersed in aqueous solution where the original positive surface charge changed to negative when CLDH was surface functionalized with CMC, and the crystalline characteristics of the original LDH were reduced when it was intercalated with casein and subsequently functionalized with CMC. E551 particles were amorphous as evidenced by the broad peak in XRD spectra, while E171 were anatase [7]. RS, however, showed both crystalline and amorphous XRD patterns [36].

### *3.2. Reflectance, Stability of Particle Suspensions, and Masking Power*

Among the synthesized particles, CCLDH showed excellent reflectance at the visible range which was second only to E171 (Figure 3a). This result complies with the masking power test whereby CCLDH had the strongest masking property in comparison with other LDH composites (Figure 4a). As shown in Figure 3a, the reflectance increased from 90% to 100% when the CLDH was engineered with CMC.

Since food additives are commonly exposed to different temperatures and pH during processing or preparation, we investigated if the aqueous suspensions of particles tested were stable at elevated temperatures and in pH range [25,37]. We observed that suspensions of CLDH and CCLDH were more thermally stable compared to pristine LDH, possibly due to the free organic anion after casein intercalation. Further, adsorption of CMC onto the LDH surface could have prevented aggregation of LDH because of static and stearic hindrance. Both RS and E551 showed increased % transmittance as the temperature increased, which suggested that the opacity of currently used white pigment alternatives is not stable at elevated temperatures. In addition, the particle suspension seemed stable at alkaline and neutral pH, but an increase in % transmittance of the LDH composites was observed at pH 3 as LDH particles started to aggregate. Notably, RS was not stable at an

alkaline pH as evidenced by an increased % transmittance. E551, however, was stable at an alkaline pH similar to that of the LDH composites.

**Figure 3.** (**a**) Reflectance of powdered particles (both synthesized and commercially used white pigment) were measured in the range of 200–800 nm with 1 nm interval, demonstrating that all the synthesized LDH particles had a good reflectance in the visible range; (**b**,**<sup>c</sup>**) % transmittance (550 nm) at varied temperature and pH, respectively, suggested relatively high stability of LDHs at high temperatures, but not in acidic pH. \* Indicate statistical significance (*t*-test) in % transmittance of LDH composites in comparison to pristine LDH, *p* ≤ 0.05, *N* = 3.


**Figure 4.** (**a**) Masking ability of the synthesized LDH and its composites was compared with that of commercially used E171 and its alternatives by suspending them into a mixture to form a thick paste to be painted on a glass plate and visually compared against a worded background. From left: E171, E551, RS, CCLDH, CLDH, and LDH. The LDH composites showed better masking. (**b**) Suspension stability of particles (from left: LDH, CLDH (1 mg/mL), CLDH (10 mg/mL), CCLDH (1 mg/mL), CCLDH CLDH (10 mg/mL), and E171) dispersed in DI water were visually compared from time-lapse images taken at 30 s intervals up to 180 s, followed by 60 s intervals from 3 min to 6 min.

The masking power of LDH composites was compared with E171 and other TiO2 alternatives by painting slurries containing particles on to a clean glass sheet and visually compared over a worded background. Even though the masking power was not as good as E171 (Figure 4a) LDH composites were superior to current alternatives used in the industry. Notably, pure LDH was not suspended well in the slurry and showed granular appearance while composites showed smooth smearing (Figure 4a). The aim of the research was to

develop a white and opaque food pigment. Therefore, balancing the pigment structure and its stability in aqueous suspension is crucial. As seen in the time lapse images of particle suspension, casein intercalated LDH (CLDH) and modified with CMC (CCLDH) showed a more stable particle suspension.

### *3.3. Antigenicity and Cytotoxicity Assessment of LDH Composites*

Since LDH particles are efficient delivery carriers of proteins, we also explored the possibility of using LDH as a modulator to prevent the binding of immunoglobulin E (IgE)-mediated hypersensitivity to casein which is often associated with milk allergy symptoms. As shown in Figure 5a, the antigen response of caseins intercalated in the CLDH reduced significantly which confirms its ability to suppress casein-specific IgE binding capacity, therefore decreasing the casein antigenicity. However the same response was not maintained in CCLDH, although the effect was still significantly lesser than pure casein.

**Figure 5.** (**a**) IgE binding capacity was determined by ELISA, and its response was expressed as the casein antigen response (ng/mL) of casein samples under different LDH compounds. \* Indicates statistically significant (*t*-test) difference in casein antigenic response of LDH composites with casein in comparison to raw untreated casein molecules, *p* ≤ 0.05, *N* = 3. (**b**) Particle concentration dependent Caco-2 cell viability changes post exposure to particles tested. The result suggested low/no cytotoxicity of LDH particles. \* Indicates the significant differeces of the tested particles from the control (CdCl2) as verified by statistical *t*-test (*p* ≤ 0.05, *N* = 3).

The cytotoxicity response was evaluated based on the resazurin assay after the Caco-2 cells were treated with different concentations (0–25 μg/mL) of the synthesized pigments, its alternatives and E171 with cadmium chloride (CdCl2) as the positive control. The in vitro data clearly depicited that all of the synthesized LDH composites did not elicit any obvious cytotoxicity to Caco-2 with a viability above 80%.

### *3.4. Comparison of Particles for Efficacy, Cost, and Safety*

To demonstrate the applicability of the LDH composite as white pigments, we compared the tested materials in their performance, cost, and safety using MATLAB® analysis (Figure 6). The reliability of E171 was scrutinized against its current alternatives and CCLDH based on three factors: efficacy, cost, and safety. Each factor was scored from 0 to 3 as detailed in Supplementary Materials (Table S3). The scores were used to develop a

three-dimensional matrix using MATLAB® (Figure 6). As seen in the heatmap, CCLDH was better in comparison to other white pigments in this multi-parametric comparison. E551 with its poor performance and cost had the lowest score, followed by E171 and rice starch.

**Figure 6.** A three-dimensional matrix was developed by MATLAB® for comparing the efficacy (% reflectance), cost (USD/g), and safety (NOAEL) of CCLDH to TiO2 alternatives and E171. The application desirability of the four particles in food are expressed in the phase diagram developed by MATLAB®, with red representing likely low success, blue likely high success, and green having considerable success as white food pigments. Position of each particle in the matrix was determined from individual scores for the three parameters assessed.
