Wenting Fan, Fang Zhao,*, Ming Chen, Jian Li, Xuhong Guo,2,*

1 State Key Laboratory of Chemical Engineering, School of Chemical Engineering, East China University of Science and Technology, Shanghai 200237, China

2 International Joint Research Center of Green Energy Chemical Engineering, Shanghai 200237, China

Keywords:

ABSTRACT

1. Introduction

Superparamagnetic Fe3O4nanoparticles (NPs), usually within the size range of 3–50 nm[1],have attracted tremendous research interest due to their applications including protein separation [2],environmental remediation [3,4], drug delivery [5] and catalysis[6]. Many methods have been developed to prepare Fe3O4NPs[7–9], among which coprecipitation of iron salts in the aqueous solution is probably the most popular for its low cost,environmental benignancy, and high level of convenience [10,11]. The traditional coprecipitation process, implemented in macroscale vessels in small batches,suffers from problems such as poor reproducibility from batch to batch,relatively low quality and difficulty in tuning the nanoparticle properties.

Microfluidics has become an enabling technology for the preparation of nanoparticles [12–14]. Compared with conventional batch vessels, continuous flow microreactors for NP synthesis can offer significant advantages such as enhanced mixing efficiency and precise control over the reaction conditions, thereby creating a homogeneous and controllable chemical environment for particle nucleation and growth and minimizing batch-to-batch variations[15–17].

The most widely used microreactors for the coprecipitation process to synthesize Fe3O4NPs at a laboratory scale is the capillary microreactor [18,19]. Though the capillary microreactor set-up is easily built in the lab, researchers need to take countermeasures against the relatively low mixing efficiency in the capillary compared with other types of microreactors.(1)Using a coiled capillary is almost always the first method, which however could only enhance the mixing in the capillary to a certain degree.(2)The second method is increasing the flow rate,which however will reduce the residence time correspondingly. This is not beneficial for the synthesis of Fe3O4NPs which requires enough residence time for the rate-limiting reaction step and the improvement in crystallinity. A longer capillary will be needed to maintain the same residence time under higher flow rates and then result in a cumbersome set-up and reduce the flexibility. (3) The third method is decreasing the inner diameter of the capillary, which is also inappropriate for the synthesis of Fe3O4NPs due to the increased risk of blockage. (4) Researchers also used the segmented flow where the circulation in the aqueous liquid slug could improve the mixing efficiency, which however obtained Fe3O4NPs with a low saturation magnetization (30 emu?g-1, 1 emu?g-1=1 A?m2?kg-1) [20]. Therefore, it is necessary to develop a new type of microreactor supplying both high mixing intensity and adequate residence time for the production of Fe3O4NPs.

In this work,a new type of microchannel with continuous serially connected micromixers (CSCM) was proposed. The fluid flow conditions and mixing performance of two CSCM microchannels(V-typed and U-typed) were investigated by computational fluid dynamics (CFD) simulation. Furthermore, a CSCM microreactor with good mixing efficiency even at relatively low flow rates was fabricated to conduct the coprecipitation process for Fe3O4NPs.The morphology, crystallinity, and magnetic properties of the Fe3O4NPs synthesized in the CSCM microreactor were compared to those obtained in the capillary microreactor.

2. Microchannel Design

The overall reaction that takes place during the coprecipitation method to form Fe3O4can be expressed as:

As described in the mechanism of the coprecipitation process by most studies[21–24],Fe2+and Fe3+ions immediately coprecipitate upon the addition of an alkali, forming iron hydroxides and oxyhydroxides which then transform slowly to form the phase of magnetite. Based on this mechanism, there are three essential requirements to design a microreactor for the coprecipitation process to synthesize Fe3O4NPs, which are listed as follows.

(1) The mixing efficiency should be high enough to match the rapid initial precipitation of iron ions and produce a high level of saturation instantly, forming a large number of nuclei in a short time.In addition,good mixing performance is also required for the nuclei to grow into Fe3O4NPs with uniform size distribution.

(2) The residence time in the microreactor should be long enough to match the aging time needed for complete coprecipitation and a high level of crystallinity in the final Fe3O4products.

(3) The characteristic dimension of the microchannel reactor should not be too small, otherwise, the microchannel will be highly prone to be clogged by the nanoparticles generated during the coprecipitation process.In addition,it will lead to a high pressure drop when both narrow and long microchannel is used to ensure enough residence time.

To meet the above requirements, a microreactor with continuous serially connected micromixers (CSCM) was designed in this study for the coprecipitation process to synthesize Fe3O4NPs.The CSCM microchannel is characteristic of a significant number of cylindrical mixing chambers (diameter 2.0 mm, height 0.7 mm) connected in series by straight and tangential channels(width 0.7 mm,height 0.7 mm).Two types of CSCM microchannels were proposed in this work,the schematic representations and key geometry sizes of which are shown in Fig. 1 and Table 1 respectively. It can be seen that the main difference between these two types of CSCM microchannels is the angle between the two straight channels connecting with the same mixing chamber: 23° and 90°in Fig. 1, respectively. For purpose of convenience, the two types of CSCM microchannels in Fig. 1 are called V-typed and U-typed microchannels, respectively.

3. Methods

3.1. Mixing performance assessment

CFD simulation was carried out using ANSYS Fluent 14.5 to evaluate the mixing performances of the two types of CSCM microchannels presented in Fig. 1. For comparison, a Zigzag microchannel and a capillary microchannel were also simulated.The models of the four microchannels used for CFD simulation,all of which have a volume of 89 μl, are presented in Fig. S1 in the Supplementary Material, besides the meshing and modeling details.

To quantitatively evaluate the mixing efficiency, the mixing indices on the cross-sections of the microchannel were calculated from the CFD simulation results. The standard deviation of the mass fraction on any cross-section normal to the primary flow direction is given by the following formula:

Fig. 1. Schematic diagrams and local magnification of the (a) V-typed and (b) U-typed microchannels.

Table 1 Geometric parameters of the V-typed and U-typed microchannels

where n is the number of sample points on the specified crosssection(n>400 to ensure accuracy in this study),aiis the mass fraction value at the sampling point i,and b is the optimal mixing value and equals 0.5 for identical settings of the two inlet fluids except their compositions. The mixing efficiency of the two fluids can be then calculated using the following formula:

where MI is the mixing index, and σmaxis the maximum standard deviation and calculated to be 0.5 using Eq. (4).

The value of MI is from zero (total fluid segregation) to one(complete fluid mixing).

3.2. Chemicals and experiments

Magnetite nanoparticles were precipitated by the reaction between the precursor and base solutions. The precursor solution was prepared by dissolving 1.08 g ferric chloride hexahydrate(FeCl3?6H2O)and 0.34 g ferrous chloride tetrahydrate(FeCl2?4H2O)in 100 ml of degassed deionized(DI)water.The base solution was prepared by diluting 2.7 ml NH3?H2O (28% (mass)) to 100 ml ammonia solution using degassed DI water. All chemicals except DI water were purchased from Aladdin Reagent Co., Ltd. (China).

The precursor and base solutions were delivered to the two inlets of the microreactor, respectively, at an identical flow rate with a syringes pump (PHD ULTRA 70-3007, Harvard Apparatus,USA).The microreactor was immersed in a ～25°C ultrasonic water bath.The Fe3O4NPs were separated from the suspension by a magnet and washed three times with DI water. Finally, the dry Fe3O4NPs were obtained by a freeze dryer (SCIENTZ-10 N, SCIENTZ,China) and characterized.

4. Results and Discussion

4.1. Mixing performance in the CSCM microchannel

The CSCM microchannel proposed in this work, possessing millimeter-scale mixing chambers connected consecutively by submillimeter-scale straight channels and thus being able to afford sufficient volume for reaction, can directly satisfy the demands in residence time and channel dimension for the coprecipitation process to synthesize Fe3O4NPs. It still needs to be verified whether the CSCM microchannel can meet the demand in the mixing efficiency.

Fig. 2. Mass fraction contours on the central transverse planes of the (a) V-typed, (b) U-typed, (c) Zigzag and (d) capillary microchannels at a flow rate of 2 ml?min-1.

The mass fraction contours of the two types of CSCM microchannels (V-typed and U-typed), the Zigzag microchannel and the capillary microchannel at a total flow rate of 2 ml?min-1are presented in Fig. 2. The flow rates of fluid A and fluid B were the same and the flow rates mentioned in the following are all the total flow rates of the two fluids.The composition of both fluid A and fluid B were set as water, with the same physicochemical properties(density 998 kg?m-3;dynamic viscosity 8.9×10–4Pa?s;diffusion coefficient 1.0×10–9m2?s-1).It can be obviously seen in Fig. 2(d) that, in the capillary microchannel, the two fluids were stratified and the mass fraction contour was nearly unchanged along the flow direction, indicating a low degree of mixing. The successive bends in the Zigzag microchannel could introduce disturbance in the flow and promote the mixing efficiency (Fig. 2(c)). But still, the concentration did not reach homogeneity at the outlet of the Zigzag microchannel. By contrast, in the V-typed and U-typed microchannels,the concentration distribution rapidly varied along the microchannel and almost became uniform at the microchannel outlet.

Fig. 3 shows the variations of the mixing index along the microchannel volume under different flow rate conditions in the four kinds of microchannels. Four flow rates, i.e., 2, 4, 6 and 8 ml?min-1, were investigated, corresponding to the Reynolds number at the inlet of the microchannel 27,53,80 and 106,respectively. It can be seen that, at 2 ml?min-1, only the V-typed microchannel achieved complete mixing at the outlet of the microchannel.At 4 ml?min-1,the V-typed and U-typed microchannels accomplished the mixing of the fluids at locations of 50%microchannel volume and 75% microchannel volume respectively,while mixing in the Zigzag microchannel was still unfinished at its outlet. When the flow increased to 6 ml?min-1, the Zigzag microchannel realized complete mixing of the fluids at the location of 75%microchannel volume.Under all the four flow rates investigated,the capillary microchannel did not complete the blending of the two fluids therein. It can be seen that the two types of CSCM microchannels outperformed the Zigzag and capillary microchannels,especially under lower flow rates(<6 ml?min-1). This implies that the CSCM microchannels were able to offer higher mixing performance and longer residence time simultaneously at a given volume, which is an important advantage for the synthesis of Fe3O4NPs.

To elucidate the reason why the two types of CSCM microchannels were capable of enhancing the mixing performance markedly,the velocity vector plots for the V-typed and U-typed microchannels are given in Fig. 4.The cross-section perpendicular to the primary flow direction after the flow exited the second mixing chamber was chosen for both two microchannels. As seen in Fig.4,two symmetrical and counter-rotating vortices were formed on the selected cross-section in both the V-typed and U-typed microchannels, indicating the Dean flow was induced due to the centrifugal force that the flow experienced when it passed the cylindrical mixing chamber. The Dean vortices could largely enhance the chaotic advection [25] and were responsible for the high level of mixing in the two types of CSCM microchannels.Besides,the periodic constriction-expansion structure in the CSCM microchannel, could also intensify the advection in the flow persistently.

4.2. Synthesis of Fe3O4 NPs in the CSCM microreactor

Fig.3. Variation of the mixing index along with the microchannel volume in the four microchannels at four flow rates:(a)2 ml?min-1,(b)4 ml?min-1,(c)6 ml?min-1 and(d)8 ml?min-1.

The simulation result in Fig. 3 shows that the V-typed microchannel stood out from the other three at a low flow rate of 2 ml?min-1. A microreactor having the V-typed microchannel(denoted as the V-typed microreactor for short)was then designed and manufactured, as shown in Fig. 5. The V-typed microreactor,which was made of borosilicate glass, had a total channel volume of 2.46 ml, and an exterior size of 17 mm × 10 mm × 7 mm(length × width × height).

The coprecipitation process of Fe2+and Fe3+in presence of NH3?H2O was conducted in the V-typed microreactor shown in Fig. 5(b), as well as in a capillary microreactor (inner diameter 0.8 mm) with the same internal volume. Fig. 6 shows the TEM images of Fe3O4NPs obtained in the two microreactors at a flow rate of 2 ml?min-1. The agglomeration of Fe3O4NPs synthesized in the capillary microreactor was clearly observed in Fig. 6(a),while the NPs obtained in the V-typed microreactor, as shown in Fig. 6(b), exhibited much better dispersity and uniformity. This should be because the greatly enhanced mixing condition in the V-typed microreactor created a more homogeneous environment for both the rapid nucleation of iron hydroxides/oxyhydroxides at the initial stage of the coprecipitation process and the subsequent (relatively) slow transition to magnetite, resulting in Fe3O4NPs with better uniformity. In addition, the chaotic advection caused by the Dean vortices in the V-typed microreactor could mitigate the aggregation of NPs, rendering Fe3O4NPs with improved dispersity. Based on size statistics of the TEM images(see Fig. S3 in the Supplementary Material), the average size of the Fe3O4NPs synthesized in the V-typed microreactor was 9.3 nm which is within the typical size range of superparamagnetic NPs.

Fig.4. Velocity vector plots on the cross-section of the straight channel at the outlet of the second micromixer at a flow rate of 2 ml?min-1 in the(a)V-typed and(b)U-typed microchannels.

Fig. 5. Microchannel reactor with continuous V-typed serially connected micromixers. (a) Schematic diagram of the internal microchannel structure; (b) photo of the microreactor.

The XRD patterns of Fe3O4NPs obtained from the V-typed microreactor and capillary microreactor are shown in Fig. 7(a),which were both in agreement with the reference pattern of Fe3O4in the Powder Diffraction File database (PDF card 88-0866). The XRD pattern of Fe3O4NPs produced by the V-typed microreactor exhibited evidently sharper peaks, indicating improved crystallinity.

Further, different Fe3O4samples were synthesized in the V-typed microreactor under different flow rates and their XRD graphs are shown in Fig. 7(b). As the flow rate elevated from 2 to 8 ml?min-1, the mixing efficiency was intensified (Fig. 3) but the residence time was reduced correspondingly. As a result, the NPs did not have sufficient time for thorough phase transformation to magnetite and the crystallinity of Fe3O4became lower with the increase in flow rate(Fig.7(b)).The result in Fig.7(b)verifies again the ability of the V-typed microreactor presented in this work to offer concurrently high mixing efficiency and sufficient residence time for the synthesis of high-quality Fe3O4NPs.

As shown in Fig. 8, the magnetic hysteresis loops of the Fe3O4NPs obtained from both capillary and V-typed microreactors are reversible S-shaped curves, confirming their superparamagnetic nature. The saturation magnetization of Fe3O4NPs synthesized in the V-typed microreactor was slightly higher than that obtained in the capillary microreactor probably due to the combined effect of the improved crystallinity and reduced NP size.

Fig. 6. The TEM images of the Fe3O4 NPs obtained from (a) capillary and (b) V-typed microreactor at a flow rate of 2 ml?min-1.

Fig. 8. The magnetization curves of the Fe3O4 NPs obtained from the capillary and V-typed microreactors at a flow rate of 2 ml?min-1 (1 emu?g-1 = 1 A?m2?kg-1, 1 Oe = 80 A?m-1).

5. Conclusions

In summary, a new type of microreactor with CSCM was customized for the continuous flow synthesis of Fe3O4NPs requiring both high mixing intensity for rapid nucleation and sufficient residence time for complete crystallization.The V-typed and U-typed CSCM microchannels displayed superior mixing performance than the common Zigzag microchannel and the capillary microchannel.The V-typed CSCM microchannel could achieve complete mixing in 2.7 s at a Reynolds number of 27. CFD simulation results demonstrated that the special structure of the CSCM microchannel could bring about Dean vortices,which greatly intensified the micromixing.Compared with NPs obtained in the capillary microreactor,the Fe3O4NPs synthesized in the V-typed CSCM microreactor exhibited a more uniform size distribution and higher crystallinity and saturation magnetization. For future work, control of the properties of Fe3O4NPs and synthesis of other metallic oxides (e.g., NiO) via coprecipitation method will be investigated in the new V-typed CSCM microreactor developed in this work.

Data Availability

Data will be made available on request.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

We gratefully thank the financial support from the National Natural Science Foundation of China (21808059) and the Fundamental Research Funds for the Central Universities(JKA01221712).

Supplementary Material

Supplementary material to this article can be found online at https://doi.org/10.1016/j.cjche.2022.12.008.