Enhancing the Scalability of Crystallization-Driven Self-Assembly Using Flow Reactors

Anisotropic materials have garnered significant attention due to their potential applications in cargo delivery, surface modification, and composite reinforcement. Crystallization-driven self-assembly (CDSA) is a practical way to access anisotropic structures, such as 2D platelets. Living CDSA, where platelets are formed by using seed particles, allows the platelet size to be well controlled. Nonetheless, the current method of platelet preparation is restricted to low concentrations and small scales, resulting in inefficient production, which hampers its potential for commercial applications. To address this limitation, continuous flow reactors were employed to improve the production efficiency. Flow platforms ensure consistent product quality by maintaining the same parameters throughout the process, circumventing batch-to-batch variations and discrepancies observed during scale-up. In this study, we present the first demonstration of living CDSA performed within flow reactors. A continuous flow system was established, and the epitaxial growth of platelets was initially conducted to study the influence of flow parameters such as temperature, residence time, and flow rate on the morphology of platelets. Comparison of different epitaxial growth manners of seeds and platelets was made when using seeds to perform living CDSA. Size-controllable platelets from seeds can be obtained from a series flow system by easily tuning flow rates. Additionally, uniform platelets were continuously collected, exhibiting improved size and dispersity compared to those obtained in batch reactions.


S3
captured on a Jeol 1400 Bio TEM microscope with an acceleration voltage of 80 kV.
Samples were prepared by drop cast 8 μL of solution on formvar-coated copper grids and then blotting away with filter paper.Another 8 μL of aqueous uranyl acetate solution (1 wt%) was dropped on the grid to stain the sample, and the excess was removed by filter paper.Samples were dried overnight in a desiccator box before testing.
The obtained images were analysed by ImageJ, and at least 100 particles were analysed to calculate the average size.The number average length (L n ) and weight average length (L w ) were calculated according to equations ( 1) and ( 2), and the distribution was the value of L w /L n . (1) (2) Similarly, the number average width (W n ), weight average width (W w ), number average area (A n ), and weight average area (A w ) were calculated according to equations (3) to (6), and their distributions were defined as W w /W n and A w /A n .
(     Flow reactors were set up by connecting several syringes with PTFE tubing (Figure 1a).
Seeds and unimer (PCL / PCL-PDMA=1:1, 10 mg•mL -1 in CHCl 3 ) were stored in different syringes, and the weight ratios of seeds and unimer were tuned by their flow rates.To make sure seeds and unimer were well mixed, a Y-mixer was equipped at their intersection, where living CDSA started.Before entering the Y-mixer, seed solution was pre-heated to the setting temperature in a tubing coil (inner diameter: 0.8 mm, residence: 2 minutes), and then a cycle of living CDSA was completed once the mixture flowed through another tubing coil (inner diameter: 0.3 mm, residence: 2 minutes) after the Y-mixer.Both the mixer and the tubing coil were buried in the pre-heated sand bath.
To conduct more living CDSA sessions, more mixers and tubing coils could be connected in series, as illustrated in Figure 4b.

Synthesis
Synthesis of dual-headed chain transfer agent (CTA)

Synthesis of bis-(ethylsulfanylthiocarbonyl) disulfide
Carbon disulfide (7.74 mL) was added to a solution of sodium ethanethiolate (10 g) in diethyl ether (500 mL), which is immersed in an ice bath with a constant stir.After stirring for another 2 hours, excess iodine was added to the flask until the colour of the reaction system became dark brown.The solution was kept stirred overnight, and then it was washed with a solution of sodium thiosulfate (1M, 3 × 100 mL) followed by brine (3 × 100 mL) after it was transferred into a separation funnel.The solvent was removed by rotary evaporation after the organic layer was dried with anhydrous magnesium sulfate.15.27 g (yield: 93.7%) of an orange liquid was obtained.

Synthesis of 2-cyano-5-hydroxypentan-2-yl-ethyl carbonotrithioate (CHPET)
flask, and then the reaction apparatus was kept in the nitrogen flow and cooled to -78 °C with a dry ice -acetone bath.Dry tetrahydrofuran was used as the solvent and added to the flask with a syringe, and borane tetrahydrofuran (12 mL, 0.12 mmol) was dropped into the solution subsequently.After an hour of reaction at -78 °C, the solvent was left at room temperature and kept stirred overnight.2-propanol was added to react with the extra borane.After the reaction, the solvent was removed by rotary evaporation, and diethyl ether (200 mL) was added to dissolve the crude product.The solution was washed with saturated NaHCO 3 (3 × 200 mL) and then with brine (1 × 200 mL).The organic layer was concentrated by rotary evaporation after it was dried by anhydrous MgSO 4 .Silica gel column chromatography (n-hexane / EtOAc = 1:1) was used to purify the product further, and 2.13 g (yield: 74.9%) of orange oil was finally obtained.The synthesis route to access CTA was illustrated in Scheme S1.
1 H NMR (400 MHz, Chloroform-d, δ (ppm)): 3.72 (t, J = 6.1 Hz, 1H, CH 2 OH), 3.35 and then transferred into an ampoule.The solution was freeze-pump-thawed three times before the ampoule was immersed in an oil bath set at 70 °C.After 2 hours, polymerization was quenched by immersing the ampoule in the liquid N 2 .The crude product was precipitated into the cold diethyl ether once it reached room temperature and then collected by centrifugation, and this process was repeated three times.After drying in a vacuum oven for 3 days, 458.5 mg (monomer conversion: 80%) of a solid product was obtained.added into a vial and stirred at room temperature for 2 days.After the reaction, the undissolved substance was removed by filtration, and the filtrate was precipitated into Seeds (or original platelets) and unimer were loaded into syringes, and then the mixer and coil were buried in the sand bath at a pre-set temperature and equilibrated for 30 minutes before the flow started.The flow rates of seeds and unimer were set up according to Table 1-5.Every time flow rates, seeds, or unimer were changed, samples were collected after passing 3 reactor volumes to make sure they reached the steady state.TEM samples were prepared instantly once platelets were collected.The detailed flow setups for different experiments were listed in Table S1-5.
Scheme S1.The synthetic route of CHPET.
Table S1.Setup of flow reactor for epitaxial growth of platelets at various temperatures.Table S7.Batch platelets prepared using various amount of unimer.
Area (um 2 ) Length (um) Width (um) Unimer (eqv. of seed) and R u are abbreviations of the flow rate of original platelets and unimer, and b T is the time solution flows through the coil.The concentrations of original and unimer are fixed at 0.04 and 10 mg•mL -1 respectively in this work.

Figure
Figure S7.(a) Scheme to prepare seeds from PCL-PDMA.TEM images of (b) polydisperse cylinders and (c) seeds.Uranyl acetate aqueous solution (1%) was used for stain.(d) Length distribution of seeds.At least 100 particles were analysed to obtain statistical results.

Figure
Figure S8.(a) Scheme of living CDSA.TEM images of platelets prepared at seed/unimer ratio of (b) 1:10, (c) 1:20, and (d) 1:30.(e) Linear fit of area upon different unimer/seed ratios.Error bars represent the standard deviation of the area distribution.

Figure S9 .
Figure S9.Size comparison of platelets from 1-mL and 10-mL scales: (a) TEM image of platelets prepared from the 10 mL scale, (b) area, (c) length, and (d) width.

Figure
Figure S10.TEM image of platelets prepared at 1 mg mL -1 in 1 mL batch scale.

Figure S11 .
Figure S11.Linear fitting of area and unimer/seed ratio.The error bar represents the standard deviation.

Figure
Figure S12.(a) TEM images of extended flow platelets obtained from wider tubing, (b) flow platelets from seeds at the unimer/seed ratio of 10, (c) fragments of platelets, and (d) flow platelets from platelet fragments at the unimer/seed ratio of 10.

Figure S14 .
Figure S14.Size comparison of platelets from 1-mL and 10-mL batch scales and 20-mL flow: (a) length and (b) width.

Figure
Figure S15.TEM image of platelets prepared at 1 mg mL -1 in flow reactors.

Table S2 .
Setup of flow reactor for epitaxial growth of platelets at various flow rates.

Table S3 .
Setup of flow reactor for epitaxial growth of platelets with various amounts of unimer.

Table S4 .
Setup of flow reactor for epitaxial growth of seeds with various amounts of unimer.R s is the abbreviation of the flow rate of seeds, and " * " indicates the seeds were prepared by sonicating platelets. a

Table S6 .
SEC results of polymers detected by RI.

Table S8 .
Extended flow platelets prepared using original platelets as seeds at various temperature.

Table S9 .
Extended flow platelets prepared using original platelets as seeds at various flow rates.

Table S10 .
Extended platelets prepared using original platelets as seeds with various amount of unimer both in flow and batch reactors.

Table S11 .
Summary table of all condition studied for flow living CDSA and the size of the corresponding platelets.

Table S12 .
Flow platelets from seeds with various amount of unimer.