Deuterated polystyrene (DPS) shells have been used to study the physics of highly compressed fuel plasmas in laser-driven inertial confined fusion (ICF), laboratory astrophysics, high field physics and so on.^1–3 Like other polymer shells needed in the above extreme conditions, there are stringent specifications on the quality of DPS shells such as size requirements, monodispersity and surface finish.⁴ With DPS synthesized by radical polymerization, DPS shells with ∼ 100 μm–1500 μm diameter and ∼3 μm–15 μm wall thickness have been successfully fabricated by the density-matched emulsion method.^5,6 However, the method shows unsatisfactory controllability on the quality of DPS shells. Therefore, it is necessary to introduce a new technique to fabricate DPS shells and investigate and reveal the factors which affect the quality of DPS shells.

With the development of the microfluidic technique, devices with complicated structures, such as T-shape, co-flowing and flow focusing, have been developed to prepare compound droplets.^7–12 It has been reported that the CV (the coefficient of variation: the ratio of the standard deviation of the size distribution to its arithmetic mean) value for the diameter of compound droplets prepared by the microfluidic technique is less than 2%, which indicates excellent size monodispersity.^13,14 It is promising, to improve the monodispersity of DPS shells by this method. The microfluidic technique has wide applications in various fields such as biology, drug delivery and the food industry. However, focus has been on the fabrication of compound droplets; their solidification has not been considered. Moreover, the compound droplets in these fields are at the nano-/micro-scale, and therefore much smaller than the corresponding droplets of DPS shells.

In this paper, the synthesis of DPS is optimized to control the molecular weight and the molecular weight distribution, while purification is improved to remove hydrophilic substances. The microfluidic technique is introduced to prepare DPS shells with good monodispersity. The surface finish of DPS shells and the defects appearing during the drying process are also investigated.

DPS-C was obtained from commercial products and synthesized by anionic polymerization, while DPS-S was synthesized by radical polymerization.¹⁵ Poly(acrylic acid) (PAA) and poly(vinyl alcohol) (PVA) were purchased from Polyscience and Sigma Aldrich, respectively. Anhydrous calcium chloride (CaCl₂) was purchased from the Chengdu Kelong Chemical Reagent Factory. Fluorobenzene (FB) was obtained from Shanghai Jingchun Reagent Ltd.

As shown in Fig. 1, the fabrication process of DPS shells is similar to that of PS shells as reported previously.^16,17 Pure water and aqueous solutions containing surfactants were used as the inner and outer phases (W1 and W2), respectively, while the DPS/FB solution was used as the oil phase (O). The flow-focusing microfluidic device was designed by our group and is composed mainly of two parts. One is the tubes while the other is the supporting body made of Teflon. The stainless steel tube, plastic tube and capillary are used as the inner, middle and outer tubes, respectively. Moreover, the tubes are fixed coaxially.

The molecular weight and molecular weight distribution of DPS were characterized by gel permeation chromatography using multiangle laser light-scattering (GPC-MALLS). The deuterated percentage of DPS was characterized by nuclear magnetic resonance (NMR). The residual lithium in DPS was measured by inductively coupled plasma mass spectrometry (ICP-MS) while the residual solvent was confirmed by pyrolysis gas chromatography mass spectrometry (PY-GC-MS) at 500 ^oC for 3s. The thermal stability of DPS was tested by thermogravimetric analysis (TGA) at 5 ^oC/min from 50 ^oC to 500 ^oC.

The morphology of the W1/O compound droplets and DPS shells was characterized using a digital microscope. Figure 1 shows that there are two interfaces between the inner and middle phases; therefore, it is difficult to determine where the real interface is. Therefore, the real inner diameter (t_i) of the W1/O compound droplets was calculated by Eq. (1), where t_o is the outer diameter and t_p is the diameter of the corresponding droplets by rupturing the compound droplet, shown in Fig. 2.¹⁸ The outer diameter and wall thickness of the DPS shells were obtained from their X-ray digital images.¹⁹ti=(to3−tp3)13.(1)

DPS-C is first considered for the preparation of DPS-shells. However, the stability of W1/O compound droplets prepared by DPS-C is poor. As shown in Fig. 3, when the mass fraction of PVA in the W2 phase is 2.0%, some W1/O compound droplets still break up to form O droplets within 10 min after their generation. This is probably due to the presence of residual lithium salts or hydrophilic solvents in the DPS-C, which originate from the initiator (butyl lithium (BuLi)) in the anionic polymerization of DPS-C or the purified process.²⁰ It is difficult to obtain DPS shells by DPS-C. Moreover, there are many vacuoles in the walls of the DPS shells, which will be discussed later.

To avoid the initiator taking hydrophilic substances in the DPS, DPS-S is synthesized by radical polymerization. As shown in Table I, the deuterated percentage of DPS-S synthesized by bulk polymerization is higher than that of DPS-S synthesized by solution polymerization, but still lower than that of DPS-C, since the deuterated percentage of the styrene monomer in our experiment is about 98.0%. Therefore, to ensure a high degree of deuterated DPS, bulk polymerization is preferred. Moreover, the molecular weight and molecular weight distribution is controlled by adjusting the reaction temperature, the time of adding the initiator, and the amount of initiator. The molecular weight and molecular weight distribution measured by GPC-MALLS are about 350 kg.mol⁻¹ and less than 2.0, respectively.

The purification and drying process of DPS is also improved to remove hydrophilic substances in DPS. The effects of purified methods on the quality of DPS has been investigated. DPS was purified by different methods including the traditional method, modified method A, and modified method B. In the traditional method, DPS was dissolved in toluene, precipitated from ethanol, and then dried in a vacuum oven. In modified method A, DPS was dissolved in toluene, washed with water and filtered. Then, DPS was precipitated from ethanol, and dried in a vacuum oven. In modified method B, DPS was first dissolved in toluene precipitated from ethanol, and dried in a vacuum oven. Then, DPS was dissolved in toluene again, washed with water and filtered. The DPS solution was firstly dried in a rotating cylindrical flask, and transferred to a vacuum oven to dry. DPS-S1, DPS-S2 and DPS-S3 are the corresponding products made from the traditional method, modified method A, and modified method B, respectively. Figure 4 shows that the weight loss percentage of DPS-S1 and DPS-S2 is about 3% and 1%, respectively, while DPS-C and DPS-S3 shows almost no weight loss below 200 ^oC, indicating that there are some residual solvents in DPS-S1 and DPS-S2.

For DPS-S1 and DPS-S2, the PY-GC-MS results show that there are two peaks, appearing at about 4.5 min and 5.8 min (Fig. 5(a)). As shown in Fig. 5(c), the corresponding MS spectrum of the product eluting at about 4.5 min presents high ion signals at the mass-to-charge ratio (m/z) of 91 and 92, which are the typical ion peaks of toluene (residual solvent in the DPS). The corresponding MS spectrum of the product eluting at about 5.8 min shows high ion signals at a m/z of 112, which are the typical ion peaks of deuterated styrene (Fig. 5(d)). The peak of toluene is much higher than that of deuterated styrene in Fig. 5(a). For DPS-1, the pyrolysis time is so short that most of DPS does not pyrolyse and the toluene quickly evaporates at 500 ^oC, so the toluene is more than the degradation product of DPS in the pyrolysis. For DPS-S3 and DPS-C, a single peak appears at about 5.8 min (Fig. 5(b)), indicating that there is almost no residual solvent. From the results of TG and PY-GC-MS, the modified method B is an effective method to remove residual solvent. Moreover, the ICP-MS results show that the lithium ion concentration of DPS-S3 is below the detection limit, since no substance containing lithium is used in the synthesis of DPS in our experiments.

To our knowledge, DPS shells were first fabricated by Takagi et al., where W1/O compound droplets with DPS in the O phase were formed by mechanical agitation and then solidified to remove the solvent of the O phase. We also used this traditional method to prepare DPS shells for the target capsule in the ICF experiments. It is difficult to find DPS shells meeting the size requirements, due to the wide distribution of diameter and wall thickness. Due to good monodispersity, the microfluidic technique was introduced to prepare DPS shells and a co-flowing microfluidic device was designed and constructed.

As shown in Fig. 6, the diameter of compound droplets prepared by mechanical agitation is from tens of micrometers to more than one thousand micrometers, while those prepared by the co-flowing microfluidic device exhibit good monodispersity of diameter and wall thickness. The size of compound droplets can be further tuned by optimizing the components of W1, O and W2 phases and their flow rates, leading to an improvement in monodispersity of DPS shells. Taking the size requirement (310 μm ± 10 μm diameter and 10 μm ± 1 μm wall thickness, for example, the size distribution of DPS shells prepared by our microfluidic technique is shown in Fig. 7. The CV values for the diameter and wall thickness of the DPS shells are 1.8% and 10.8%, respectively. The monodispersity of the diameter is much better than that of the wall thickness. The yield of the DPS shells meeting the size requirement is about 65%, much higher than that of DPS shells prepared by the traditional method (less than 1%). To improve the yield, it is necessary to improve the monodispersity of the wall thickness.

Figure 8 shows typical microphotographs of DPS shells prepared by the traditional method and our microfluidic technique. Generally, the diameter and wall thickness for most of DPS shells prepared by the traditional method are less than 1000 μm and 20 μm, respectively (Fig. 8(a)), which is unfavorable if preparing DPS shells with a larger diameter/thicker wall. In the microfluidic technique, DPS shells with a larger diameter/thicker wall are obtained using a microfluidic device with a larger tube, increasing the concentration of the O phase and adjusting the flowing rates (Fig. 8(b)–(d)). DPS shells with ∼300 μm – 2500 μm in diameter and ∼10 μm – 100 μm wall thickness were successfully prepared. The larger DPS shells have been used in experiments with the laser-driven spherically convergent plasma fusion scheme.³

It has been reported that numerous factors, such as hydrophilic residues in the polymer, the component of the W2 phase and the solidifying rate, affect the surface finish of polymer shells.^20–22 In Fig. 9, the experimental results show that the shells prepared from DPS-C without purification are opaque, indicating that there are many vacuoles in the shell. When DPS-C is purified to remove hydrophilic residues, the shells are transparent. Obviously, the residue from the butyl lithium initiator leads to water entering into the O phase, forming vacuoles in the resulting shells. Therefore, the reduction of hydrophilic residues in DPS to a level as low as possible is an effective method to eliminate vacuoles and improve the surface finish.

The effects of the component of the W2 phase on surface finish are also investigated. It has been reported that the addition of salts to the W2 phase suppresses vacuole formation.²¹ W2 phases with different salt concentrations were used to prepare DPS-3 shells. With 0.5% CaCl₂, there are still some vacuoles on the surface of DPS-3 shells, but no vacuoles can be observed when the concentration is increased to 2.0% (Fig. 10). Compared with polystyrene (PS) and Poly(α-methyl styrene) (PαMS), it seems more difficult to improve the surface finish of DPS shells. The improvement of the purification method is ongoing and the effects of other surfactants and salts on the surface finish are also under evaluation.

To obtain DPS shells, it is necessary to remove the internal water contained in the DPS shells. As shown in Fig. 11, different defects can appear during the drying process of DPS shells, including deflation, crazes, cracks, as well as crazes and cracks, similar to those found in PS shells.²³ The method to reduce defects for PS shells, including heat treatment, ethanol exchange, and lowered drying temperature are applied in the fabrication of DPS shells. The yield of DPS shells without these defects increases, but the improved effect is not as good as that in PS shells. Takagi et al reported that the shells made using lower molecular weights (30 kg.mol⁻¹-100 kg.mol⁻¹) were fragile and many cracks appeared. The defects that appear in the drying process of DPS shells may not, however, be linked to molecular weight, since the molecular weight of DPS used in our experiments is higher. The relevant experiments are carried out to make the mechanism clear and to improve the yield of DPS shells without defects.

DPS, deuterated to a high degree, is synthesized by bulk polymerization and purified to remove residual solvents. DPS shells with ∼ 300 μm – 2500 μm diameter and ∼ 10 μm – 100 μm wall thickness are successfully prepared by a microfluidic technique and used in ICF experiments. The monodispersity of the diameter is much better than that of the wall thickness in a batch of DPS shells. The yield of the DPS shells meeting the size requirement is about 65%. The vacuoles in DPS shells can be suppressed by both removing hydrophilic residues from DPS and introducing salts into the W2 phase. Defects of DPS shells that appear in the drying process are similar to those in PS shells, but their incidence is more difficult to reduce. In conclusion, the quality of DPS shells has been improved in recent years, but there are still some challenges: (Ι) Trace hydrophilic substances remain in DPS, but it is important to measure their amount and remove them completely. (Π) To meet the size requirements, not only does monodispersity of the wall thickness need to be improved further, but the size formation mechanism of DPS shells should also be addressed. (Ш) Since it is difficult to improve the quality of DPS shells, the effect of replacing hydrogen with deuterium on the different properties of PS and DPS, and the formation mechanism of vacuoles and defects in DPS shells should be investigated.