Tuning the Color of Photonic Glass Pigments by Thermal Annealing

Thermal or solvent annealing is commonly employed to enhance phase separation and remove defects in block copolymer (BCP) films, leading to well‐resolved nanostructures. Annealing is of particular importance for photonic BCP materials, where large, well‐ordered lamellar domains are required to generate strong reflections at visible wavelengths. However, such strategies have not been considered for porous BCP systems, such as inverse photonic glasses, where the structure (and thus the optical response) is no longer defined solely by the chemical compatibility of the blocks, but by the size and arrangement of voids within the BCP matrix. In this study, a demonstration of how the concept of “thermal annealing” can be applied to bottlebrush block copolymer (BBCP) microparticles with a photonic glass architecture is presented, enabling their coloration to be tuned from blue to red. By comparing biocompatible BBCPs with similar composition, but different thermal behavior, it is shown that this process is driven by both a temperature‐induced softening of the BBCP matrix (i.e., polymer mobility) and the absence of microphase separation (enabling diffusion‐induced swelling of the pores). Last, this concept is applied toward the production of a thermochromic patterned hydrogel, exemplifying the potential of such responsive biocompatible photonic‐glass pigments toward smart labeling or anticounterfeiting applications.


Introduction
Block copolymers (BCPs) have been extensively explored for the production of structurally colored materials due to their ability to self-assemble into a myriad of nanostructures, ranging from simple micelles and lamellae, to more complex architectures, such as gyroids. [1][2][3][4] In such systems, the molecular properties of the BCP and optical properties of the microphase separation (to enable inward diffusion of water). Finally, we demonstrate how a post thermal-annealing strategy can be used to produce a thermochromic patterned hydrogel, exemplifying the potential of such responsive biocompatible photonic-glass pigments toward smart labeling or anticounterfeiting applications.
The ring-opening metathesis polymerization (ROMP) of both polyester macromonomers presented rapid polymerization kinetics. As evaluated by gel permeation chromatography with a multi-angle light scattering detector (GPC/MALS), the majority of the polymerization occurs over the first few minutes after the addition of third-generation Grubbs catalyst (G3), with only a small amount of conversion after this point ( Figures S4 and S5, Supporting Information). In contrast, due to the different side chain chemistry, [30,31] the polymerization of NB-PEG is much slower, needing at least 20 min to achieve a 95% conversion ( Figure S6, Supporting Information). As such, to ensure a well-defined BBCP architecture with a precise block ratio, the polyester macromonomer was polymerized by ROMP first, followed by the addition of NB-PEG. The two resultant biocompatible BBCPs were determined to be poly(NB-PDLLA) 75block-poly(NB-PEG) 67 and poly(NB-PCL) 71 -block-poly(NB-PEG) 66 (Figure 1a), as characterized by GPC/MALS and 1 H NMR spectroscopy (Figures S7-S10, Supporting Information), These BBCPs are respectively denoted as DLLA-EG and CL-EG in subsequent discussions and their properties are summarized in Table S1, Supporting Information. Importantly, their total backbone length (DP = 142 vs 137) and block ratio (DP 1 /DP 2 = 1.1 vs 1.1) are similar, which allows for a direct comparison of their self-assembly behavior.
Photonic pigments were produced through the confined self-assembly of the two BBCPs within emulsified toluene/ water droplets, as described in the Experimental Section. In short, solutions of DLLA-EG and CL-EG in toluene (3.3 wt%) were emulsified upon intersection with an aqueous polyvinyl alcohol solution (2.0 wt%) within a microfluidic flow-focusing device. The collected microdroplets were immediately placed in pre-heated water baths at different temperatures (40-80 °C) for 6 h, after which vibrant, structurally colored microparticles were obtained for both DLLA-EG and CL-EG. However, the two BBCPs presented a very different color response with processing temperature (Figure 1b, c and Figure S11, Supporting Information). While CL-EG photonic microparticles displayed a consistent blue reflection (Δλ < 10 nm across the series) during this processing time (6 h), DLLA-EG displayed a dramatic redshift with increasing temperature, allowing for the reflected color to be tuned across the visible spectrum ( Figure 1d). Furthermore, this temperature-induced redshift is both irreversible and independent of microparticle size, with larger photonic microparticles simply offering higher reflectance ( Figure S12, Supporting Information).
To understand the thermal dependence of these photonic microparticles, it was first necessary to establish the internal nanoarchitecture underlying the photonic response. As such, the photonic pigments were freeze-dried and fractured to allow their cross-section to be imaged using scanning electron microscopy (SEM, see Experimental Section). As reported in Figure 1e and Figure S13a, Supporting Information, the DLLA-EG microparticles contain a foam-like structure containing a dense arrangement of near-uniform pores, with the average pore size found to increase with the processing temperature. This short-range ordered structure is consistent with an inverse photonic glass, whereby coherent scattering from the correlated pores leads to an isotropic-colored reflection from the microparticles. [25] Segmenting the SEM images enabled a quantitative analysis of structural parameters, confirming that the increase in wavelength (λ = 445-656 nm) upon raising the processing temperature (T = 60-80 °C) correlates linearly with the correlation length (2ξ = 162-242 nm), following Bragg's law (Section S3 and Figures S11, S13 and S14, Supporting Information). Moreover, the increase in correlation length can be attributed to an expansion in pore radius (r = 68-115 nm), as opposed to changes in the wall thickness ( Figure 1e). Conversely, cross sections of CL-EG microparticles confirmed no significant change in either the correlation length (2ξ = 122 ± 4 nm) or pore radius (r = 52 ± 2 nm) over this temperature range ( Figure 1 and Figures S14 and S15, Supporting Information), resulting in a consistent peak reflection wavelength across the series (Figure 1d and Figure S11d, Supporting Information).
To explain the striking difference in behavior between DLLA-EG and CL-EG, we propose the final structure is formed in two phases, as illustrated in Figure 2a. First, during evaporation of the toluene droplet, the contained amphiphilic BBCP assembles into micelles that homogeneously swell with water (as evidenced by dynamic light scattering (DLS), see Figure 2b and Figures S16 and S17, Supporting Information), before densely packing upon final loss of solvent to form the inverse photonic glass architecture. Second, the desolvated microparticles, still dispersed in water, can undergo an additional thermal annealing step, whereby the high molecular mobility in the softened structures allows for further water uptake and thus additional swelling of the pores.
To disentangle these processes, first the desolvating time of the toluene droplet was estimated by evaporating droplets of macromonomers NB-PDLLA and NB-PCL at different temperatures, with microparticles found to form in approximately 15 min in all cases ( Figure S18, Supporting Information). As such, when microdroplets of BBCPs DLLA-EG and CL-EG were desolvated at elevated temperature (80 °C) for only 20 min (rather than approx. 6 h), both formed blue microparticles, with only a difference of 40 nm between their reflection wavelengths ( Figure S19, Supporting Information). This observation is in stark difference to the photonic responses of these BBCP systems when heated at this temperature for 6 h, where a further redshift of >230 nm for DLLA-EG was observed, but only <15 nm for CL-EG (see Figure 1). From this result we propose that the initial self-assembly process within the evaporating droplet is similar for both BBCPs, further evidenced by their similar chemical properties (total DP, DP ratio) and phase behavior (interfacial tension (IFT), DLS, see Figures S16 and S17, Supporting Information), however DLLA-EG is able to rapidly further evolve from blue to red color via thermal annealing. To test this hypothesis, we applied a discrete heating step to the blue microparticles prepared by desolvating for 20 min; while CL-EG maintained its blue color (Δλ < 10 nm), DLLA-EG underwent a small redshift at low temperatures (i.e., Δλ < 20 nm after 6 h at < 60 °C), which increased rapidly for microparticles treated at higher temperatures (e.g., Δλ > 200 nm after 6 h at 80 °C).
To understand why the nature of the polyester block gives rise to such different behavior upon annealing, their thermal properties were investigated by differential scanning calorimetry (DSC). This revealed that a homopolymer of P(NB-PDLLA) has a glass transition temperature (T g ) at 54 °C, while P(NB-PCL) is a crystalline polymer that melts at 48 °C (Figure 2c,d and Figure S20, Supporting Information). Notably, these are in a similar range to the melting temperature of NB-PEG of 52 °C ( Figure S21, Supporting Information).
Considering first DLLA-EG microparticles, at lower temperatures (T ≲ 55 °C) it is in a glassy state and as such is unable to respond significantly to any applied heating (with the slight redshift observed in Figure 1d attributed to differing desolvating kinetics within the evaporating microdroplets). [25]  . Note only the position of the peak at a higher wavelength is reported for spectra where two peaks are visible, and all spectra were measured at room temperature. e) Pore radius r against correlation length ξ obtained from the analysis of the segmented SEM images. The inset shows a representative cross-sectional SEM microscopy image of photonic pigments prepared with DLLA-EG at 70 °C, revealing the internal porous nanoarchitecture.
However, at higher processing temperatures (i.e., 60-80 °C), both blocks exist in a viscous phase, where the polymer chains possess increasing mobility to rearrange in response to an external stimulus. During thermal annealing, the influx of water into the microparticles can continue, with it ultimately accommodated within the internal pores. As such the kinetics of pore swelling (and thus the degree of redshift of the reflected color) is dependent on the speed at which the BBCP can rearrange within the microparticle to stabilize the increasing internal surface area. Higher temperatures allow for higher chain mobility, leading to a faster growth of pore size and a larger redshift of the reflection peak (Figure 1b,d), but also can lead to uncontrolled fusion of the pores ( Figure S22, Supporting Information), which is exacerbated by the depletion of any remaining free BBCP to assemble at the rapidly expanding internal interfaces. Once the microparticles are again cooled below the T g of P(NB-DLLA), the well-ordered porous structures of the glassy photonic pigments are effectively frozen, with the color found to be stable over several months.
To understand why CL-EG does not follow a similar behavior, even though the processing conditions are typically above the melting temperature of both blocks (see DSC above), it is necessary to consider the phase separation between the different polymer block pairs (i.e., PDLLA vs PEG and PCL vs PEG). As reported previously for polymer blends, PDLLA is miscible with PEG, while PCL and PEG tend to form discrete crystal phases. [32][33][34][35][36][37][38] These trends are further supported by smaller Hansen solubility parameter distances for PDLLA/PEG than that for PCL/PEG (see Section S4, Supporting Information). DSC measurements of BBCPs DLLA-EG and CL-EG are consistent with these predictions: for DLLA-EG, a completely amorphous structure is obtained since no transition is observed between 10-150 °C (Figure 2f and Figure S23a, Supporting Information), while for CL-EG, two melting peaks between 45 to 50 °C were observed and attributed to the presence of discrete PCL and PEG crystalline domains (Figure 2g and Figure  S23b, Supporting Information). In addition, water contact angle measurements are consistent with this hypothesis. As expected, the homopolymer P(NB-PCL) has a high contact angle of 84°, which is more hydrophobic than the equivalent P(NB-PDLLA), with a contact angle of 73° ( Figure S24, Supporting Information). When considering the corresponding block copolymers, CL-EG has almost the same contact angle value (87°) as P(NB-PCL), while DLLA-EG reports a much lower value of 36°, which is the same value as measured for the hydrophilic macromonomer NB-PEG ( Figure S24, Supporting Information). In the context of CL-EG, these results suggest that once the solvent is lost the microparticle becomes hydrophobic, due to the isolation of hydrophilic PEG domains within PCL domains with low water permeability. In contrast, the permeability of DLLA-EG remains much higher due to the presence of a disordered phase for any remaining mobile BBCP, such that during thermal annealing water can diffuse through the structure (see Section S5, Supporting Information).
It is important to note that water diffusion into the CL-EG microparticles is not fully inhibited. For example, by increasing mobility through heating at an elevated temperature for a prolonged time (e.g., 12 h at 90 °C) a further modest redshift can be achieved, but with significantly reduced color purity ( Figure S25, Supporting Information). To explore this effect, a series of microparticles was prepared with CL-EG with different DPs. It was found that increasing the DP resulted in a much larger redshift upon thermal annealing ( Figure S26, Supporting Information), with the DP needing to be increased by 2-3 times to achieve a similar result to DLLA-EG (Figure 1d). This is attributed to reduced polymer chain mobility with increasing molecular weight, resulting in less phase separation and thus a more amorphous, hydrophilic structure within the microparticle. Furthermore, while specific colors can be achieved via multiple different combinations of DP, temperature, or time, the intensity of the reflection and the color purity are not equivalent for all pigments of the same color ( Figures S27-S29, Supporting Information). From the comparative analysis of the spectra, it is clear that when targeting a specific reflection wavelength, it is preferable to: i) minimize the duration of thermal annealing, as the occurrence of structural defects increases over time and ii) raise the temperature rather than raise the DP, as this leads to less coherent structures.
Micellization combined with thermal annealing provides a straightforward route to precisely control the structural colors of the pigments. For example, at a fixed temperature of 80 °C, the linear relationship between the annealing time and the peak reflected wavelength allows for tuning from blue (t = 1 h: λ = 437 nm) to red (t = 6 h: λ = 656 nm) while maintaining good color purity ( Figure S30, Supporting Information). Furthermore, it was found that thermal annealing can be applied to fully desolvated photonic pigments as a discrete postprocessing step. For example, the reflected color of DLLA-EG pigments stored at room temperature in water after an initial low-temperature drying step (60 °C for 6 h) was found to be unchanged for at least two weeks, however subsequent thermal annealing of the microparticles at a temperature above the T g of PDLLA (i.e., 60+ °C) resulted in their reflection peak to strongly redshift ( Figure S31, Supporting Information). This post-processing step is also size-independent ( Figure S32, Supporting Information). Indeed, the color of these post-thermally annealed microparticles approximated that of freshly prepared samples heated for the same cumulative time (cf. Figure S30, Supporting Information). In contrast, applying a post-thermal annealing step to CL-EG pigments resulted in only a broadening of the reflection peak, resulting in an increased contribution at blue wavelengths ( Figure S33, Supporting Information). However, post-annealing at high temperatures can cause unexpected deformation or expansion of the photonic pigments, possibly because of uncontrolled pore fusion during heating ( Figure S31d, Figures S33d and S34c, Supporting Information).
Finally, building on the concept of tuning the structural color of the pigment in a post-processing step, we explored its application to the selective patterning of a structurally colored coating. Photonic pigments with comparable optical response (CL-EG: 60 °C for 3 h; DLLA-EG: 60 °C for 1 h) were embedded separately into a polyacrylamide hydrogel (12 w/v %), cut into cubes, and then assembled into an alternating checkered pattern (see Experimental Section for more details). Initially, this composite gel displayed a near-uniform blue/violet color, however, upon applying a 70 °C post-thermal annealing treatment the DLLA-EG pigments were found to strongly redshift over time, revealing the underlying pattern (Figure 3).

Conclusions
We have demonstrated that inverse photonic glass structures can be thermally annealed, allowing for the reflected color to be tuned across the visible spectrum. By comparing two BBCPs with similar self-assembly behavior, we were able to determine that the thermally-induced swelling of the pores requires both a high molecular mobility of the polymer matrix and sufficient water permeability. While harsher treatments were found to lead to the formation of irregular pores and even distortion of the overall microparticle, it is worth noting that in all cases the swelling of the pores was sufficiently coherent throughout the microparticle to maintain a distinct color in reflection. Finally, we demonstrated that thermal annealing can be applied as a discrete step, allowing for thermo-responsive coatings or smart labeling from biocompatible photonic pigments.

Experimental Section
Preparation of BBCP Microparticles via a Microfluidic Device: Monodisperse microdroplets were generated within a hydrophilic, etched-glass microfluidic device (Dolomite #3 000 158, Droplet Junction Chip with 100 µm etch depth). The discontinuous phase was prepared by dissolving the block copolymers (DLLA-EG, CL-EG) in toluene (30 mg mL −1 ). The continuous phase contained polyvinyl alcohol (PVA, 200 mg) as a stabilizer, dissolved in Milli-Q water (10 mL). To form a toluene/water microemulsion, the aqueous PVA solution (2.0 w/v%) and the BBCP solution in toluene were injected into the microfluidic device using two syringe pumps with flow rates of 1000 and 200 µL h −1 , respectively. At the intersection of the two flows, shear forces resulted in the formation of toluene microdroplets with diameter Ø ≈ 160 µm. For each experiment, an aliquot of microdroplets (ca. 13 µL) was collected into a vial (7 mL) filled with PVA solution (ca. 2 mL) and stored in a preheated water bath at a defined temperature (60, 65, 70, 75, and 80 °C) and for different times. Note to produce microparticles with different sizes, the ratio of the flow rates within the microfluidic chip was altered to tune the microdroplet volume.
Preparation of Patterned Hydrogels Containing Photonic BBCP Pigments: DLLA-EG and CL-EG photonic pigments (approx. 1.5 mg) were first respectively prepared at 60 °C for 1 and 3 h, such that the "dry" microparticle dispersion in water had a comparable violet coloration. These microparticles were washed with water at least 3 times to remove PVA and suspended in water (0.5 mL). Acrylamide aqueous solution (30.0 w/v%, 0.5 mL) was added in the microparticle suspension together with N,N' methylenebis(acrylamide) (2.0 w/v%, 200 µL) as a crosslinking agent and ammonium persulfate (6.6 w/v%, 50 µL) as an initiator. Tetramethylethylenediamine (20 µL) was added as an accelerator for hydrogel crosslinking. The hydrogels containing dispersed DLLA-EG and CL-EG photonic pigments were cut into cubes (4 × 4 × 2 mm) and assembled into a checkered pattern. The same hydrogel formulation was then used to bond the cubes together. To thermally anneal the hydrogel, it was heated at a fixed temperature and immersed in water to minimize dehydration.
Optical Microscopy and Micro-Spectroscopy: Optical microscopy and micro-spectroscopy were performed on a customized Zeiss Axio Scope A1 microscope fitted with a CMOS camera (Eye IDS, UI-3580LE-C-HQ, calibrated with a white diffuser) using a halogen lamp (Zeiss HAL100) as a light source. To perform micro-spectroscopy, the microscope was coupled to a spectrometer (Avantes, AvaSpecHS2048) using an optical fiber (Avantes, FC-UVIR200-2, 200 µm core size). The reflectance spectra were normalized against a white diffuser (Labsphere USRS-99-010). The BBCP microparticles dispersed in water were analyzed using a water immersion objective (Zeiss, W N-Achroplan 40x/0.75 M27 (FWD = 2.1 mm)). Note that optical measurements on thermally annealed microparticles were carried out once the microparticles had returned to room temperature. The collected spectra were analyzed based on the methodology presented in Section S3, Supporting Information.
Scanning Electron Microscopy: SEM was performed using a Mira3 system (TESCAN), operated at 3 kV and a working distance of 4-5 mm. The samples were mounted on aluminum stubs using conductive carbon tape and coated with Pt (10 nm) using a sputter coater (Quorum Q150T ES). The samples were obtained by lyophilizing the microparticle suspensions. The samples were fractured mechanically using a microspatula to expose the internal cross-section.
Differential Scanning Calorimetry: DSC analysis was carried out using a TA DSC Q20 instrument under the protection of argon gas (sample purge flow at 50 mL min −1 ). The samples were weighed out into an aluminum pan (Tzero, TA) using a Sartorius Micro Balance (MSE2.7S-000-DF, Sartorius Weighing Technology GmbH, Goettingen, Germany). Then an aluminum lid (Tzero, TA) was pressed on top of the pan using a Tzero sample encapsulation press. The experiments were started at 0 °C and were heated up to 160 °C at a ramping rate of 10 °C min −1 . After two minutes of isotherm, the samples were cooled down to 0 °C at the same ramping rate using a refrigerated cooling system (TA Instruments RCS40 Refrigerated Cooling System 40, RCS42-3110), followed by another two minutes of isotherm. The entire heating-cooling process was repeated twice. Note that only the second heat cycle was considered as the first cycle retained thermal history from the polymer preparation process (e.g., rapid precipitation), comparative traces of the two heat cycles are reported in Figures S20, S21, and S23, Supporting Information.
Dynamic Light Scattering: DLS was performed with a glass cuvette with round aperture on a ZETASIZER NANO-ZS with 633 nm He-Ne LASER (Malvern Panalytical). The BBCP-toluene solutions (1 mL, 5 mg mL −1 ) were added to the cuvette and equilibrated for 5 min at a specific temperature (30 °C) and water (200 µL) at the same temperature was added to initiate the measurements. Before each measurement, the mixture was gently shaken to obtain a uniform micelle suspension. Measurement angle, position, and attenuator were fixed at 173°, 6.0 mm, and 11, respectively. Constant refractive indices for the material (PDLLA: 1.460, PCL:1.470) and dispersant (1.496) were used throughout the measurements. As defined in the Zetasizer software, the viscosity of toluene at 30 °C was taken as 0.5259 cP.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.