Published: September 07, 2011

r 2011 American Chemical Society 13242 dx.doi.org/10.1021/la202055f | Langmuir 2011, 27, 13242–13247

ARTICLE

pubs.acs.org/Langmuir

Synthesis of Titania�Silica Core�Shell Microspheres via a ControlledInterface Reaction in a Microfluidic DeviceWenjie Lan, Shaowei Li, Jianhong Xu,* and Guangsheng Luo*

The State Key Laboratory of Chemical Engineering, Department of Chemical Engineering, Tsinghua University, Beijing 100084, China

1. INTRODUCTION

In recent years, the preparation and study of core�shellmicrospheres with well-defined structures have attracted sub-stantial interest because of their potential applications in con-trolled drug delivery systems, lightweight fillers, catalysis, chro-matography, vessels for confined reactions, and photonic bandgap material.1�5 Among the various types of core�shell materi-als, titania-coated silica particles are of particular interest forcatalysis.6,7 Although there has been extensive work on thepreparation of core�shell microspheres, many of these studieshave involved batch, multistep techniques. For example, col-loid�particle templating methods8,9 require the preparation ofwell-dispersed silica colloidal suspensions and subsequent coat-ing of titania. Layer-by-layer deposition methods10,11 usuallyrequire the repeated adsorption of polyelectrolyte layers in atime-consuming, multistep procedure. Mechanical stirring usedin batch process would cause local fluctuation of temperature andconcentration, which may lead to uneven coating. Frequentbatch-to-batch process would adversely influence the reproduci-bility. Therefore, we thought it would be intriguing if hybridparticles could be synthesized continuously in a single step.

Indeed, although many studies have already demonstratedviable one-step fabrication methods for producing core�shellmicroparticles, many of these have been based on double-emulsion formation techniques.12�21 In this method, the middlephase is usually a precursor solution for the shell material. Themiddle phase is sheared by the continuous phase to formdroplets. The inner phase is encapsulated in the middle phase,and a double emulsion is formed. The middle phase is thensolidified to obtain core�shell particles. The solidifying methods

include photopolymerization,12�15 thermopolymerization,16 andcross-linking.17�21 However, the formation of stable doubleemulsions has many significant requirements regarding the sys-tem’s physical properties, especially the interfacial tension andviscosity, which can greatly limit system selection and applica-tion.22�24 In addition, liquid/liquid/liquid three-phase flow pat-terns are much more complicated than those of two-phase flows,thereby increasing operational complexity. Finally, the productionof double emulsions requires more complex devices compared tothe devices used to produce single emulsions.

Our research aims to present an alternative to the commonlyused preparationmethod for core�shell microspheres by use of asingle emulsion formed by liquid/liquid two-phase flow. In thismethod, uniform droplets are generated first, and then only theoutermost surface of the droplets is solidified to form core�shellmicroparticles, rather than solid ones. Several groups have fabri-cated core�shell microparticles using this method. Takeuchiet al.25 and Quevedo et al.26 fabricated polymer hollow micro-capsules via interfacial polymerization by adding monomers andcross-linkers to each phase. Steinbacher et al.27 and Eun et al.28

prepared hollow organosilicon microcapsules and hollow titaniamicrocapsules, separately. Both groups dissolved the precursor ofthe shell material in the dispersed phase and obtained hollowparticles via hydrolysis of the precursor at the outermost surfaceof the droplets.

Received: June 1, 2011Revised: August 31, 2011

ABSTRACT: In this work, we describe a novel, simple microfluidic methodfor fabricating titania�silica core�shell microspheres. Uniform droplets ofsilica sol were dispersed into an oil phase containing tetrabutyl titanate via acoaxial microfluidic device. The titanium alkoxide hydrolyzed at the water�oil interface after the formation of the aqueous droplets. A gel shellcontaining the titanium hydroxide formed around the droplets, and thetitania�silica core�shell microspheres were obtained after calcinations.The X-ray diffraction results show that titania coatings crystallized into apure anatase structure. The scanning electron microscopy and energy-dispersive spectrometry characterization shows that the microspheres aremonodispersed with uniform titania coating on the surface. The dispersityand size of the microspheres could easily be controlled by changing themicrofluidic flow parameters. The titania content on the surface could beadjusted in the large range of 1.0�98.0mol % by varying the continuous phase composition and the reaction time, and the structuresof the core�shell microshperes could also be controlled.

13243 dx.doi.org/10.1021/la202055f |Langmuir 2011, 27, 13242–13247

Langmuir ARTICLE

In spite of the research mentioned above, there is still littleresearch available focusing on the application of a single-emul-sion method rather than the double-emulsion method in thepreparation of core�shell microparticles. We feel that focusingon research involving a single-emulsion method is needed toadvance and widen the scope of methods currently used in thefield. Currently, the single-emulsion method is mostly used tofabricate hollow microparticles.25�28 Fabrication of compositecore�shell microparticles using this method is seldom reported.

In our work, a single-emulsion method was used to preparecomposite core�shell microparticles. This method presentsunique advantages over the more commonly used double-emul-sion method in that the equipment is simpler and flow patternsare much easier to control. Composite core�shell microparticleswere successfully fabricated by dissolving the shell materialprecursor in the continuous phase and the core material pre-cursor in the dispersed phase. The titania precursor in thecontinuous fluid hydrolyzed at the interface forming a shellaround the silica precursor, which is substantially different fromthemethods used for the preparation of hollowmicroparticles. Inaddition, as our study has shown, the single-emulsion methodcan be used to control the silica and titania composition (rangingfrom 2.0 to 99.0 mol % silica) of the shell layer by simplyadjusting the reaction rate and reaction time.

2. EXPERIMENTS

2.1. Materials.Methyl cellulose (MC) was purchased from Sigma-Aldrich Corp. (St. Louis, MO, USA). Other chemicals, includingtetraethoxysilane (TEOS), acetic acid, liquid paraffin, Span 80, oleicacid, tetrabutyl titanate (TBT), and n-octanol, were all purchased fromSinopharm Chemical Reagent Beijing Co. (Beijing, People's Republic ofChina).2.2. Experimental Microdevice. A flowchart of the experiment

and the microfluidic device are shown in Figure 1a,b. The device wasfabricated on a 40 mm � 20 mm � 3 mm poly(methyl methacrylate)

(PMMA) chip using micromachining technology. A Teflon tube with0.5 mm inner diameter was inserted as the main channel for the multi-phase flow, while a stainless steel microneedle was inserted into theTeflontube as the dispersed phase inlet. The outer and inner diameters of themicroneedle were 0.34 and 0.16mm, respectively. Themicrofluidic devicewas sealed using another 40 mm � 20 mm � 1 mm PMMA chip whichwas bonded to the first chip using the ultrasonic assisted sealingtechnique.29 Three microsyringe pumps and three gastight microsyringeswere used to pump the fluids into the microfluidic device. The dropletspassing through the Teflon tubing were collected in a round containerfilled with continuous phase fluid in which further hydrolysis occurs. Asshown in Figure 1a, a microscope coupled with a high-speed CCD videocamera was used to observe the flow pattern in the microfluidic device.2.3. Preparation of Titania�Silica Core�Shell Micro-

spheres. A silica sol was prepared by dissolving 2.0 wt % MC, 2.0 wt% acetic acid, and 5 wt % tetraethoxysilane (TEOS) in water and stirredfor 24 h before being used as the dispersed phase. As the dispersed phasewas pumped through the microneedle, monodispersed droplets wereproduced by the sheath focusing force of the continuous phase flow. Aliquid paraffin solution with 2 wt % Span 80 and 3 wt % TBT was used asthe continuous phase. Oleic acid was added to the continuous phase atconcentrations ranging between 0.1 and 1.0 wt % to prepare micro-spheres with different surface composites. As soon as the droplet wasformed, the hydrolysis of TBT occurred at the interface to form a thin gelshell around the droplets. The droplets flowed through the 2 m longTeflon tube and then into a solidification bath containing the continuousfluid for further hydrolysis. Finally, the microspheres were washed withn-octane and freeze-dried. The driedmicrospheres were calcined in air at500 �C for 4 h.2.4. Characterization. Droplet formation and the resulting hybrid

microspheres were both observed with an optical microscope(Olympus). More detailed structures were observed using scanningelectron microscopy (SEM, FEI XL30) operating at 1 kV. Energyspectrum analysis (EDS) was performed on the SEM samples usingthe EDS function of the SEM. The X-ray diffraction (XRD) pattern ofthe microsphere surface was obtained using an X-ray diffractometermodel (D/max-RB, Japan).

3. RESULTS AND DISCUSSION

3.1. Titania�Silica Core�Shell Microsphere Formation.The hydrolysis reaction of TBT took place vigorously and easily

Figure 1. Experimental setup.

Figure 2. (a) Micrograph of the drop flow in the coaxial microfluidicdevice. (b) Micrograph of the aqueous droplets with a gelike shelldispersed in the continuous phase. The scale bars are 500 μm. (c) Themechanism of the core�shell microsphere generation process.

13244 dx.doi.org/10.1021/la202055f |Langmuir 2011, 27, 13242–13247

Langmuir ARTICLE

broke the droplets. Thus, it is necessary to reduce the reactionrate for controllable preparation of hollow microspheres inthe microfluidic device. Carboxylic acids of longer chain lengthcontribute to the reduced rate of hydrolysis.30 Thus, oleicacid was added into the oil phase for chemical modification ofmetal alkoxides to reduce the reaction rate and extent in thiswork.Figure 2c shows the mechanism for generating titania�silica

core�shell microspheres in the microchannel. At the outletof the needle, the silica sol is sheared by the oil phase to formmonodispersed droplets (Figure 2a,b). The hydrolysis of TBToccurs at the interface as soon as the aqueous phase is in contactwith the oil phase. As the droplets flow through the Teflontube, the metal hydroxide undergoes further hydrolysis and

condensation to precipitate from the liquid phase. A thin gellikeshell containing metal hydroxide is then formed around thedroplet. The metal hydroxide is hydrophilic, and the water coulddiffuse through the shell toward the oil phase. Water transportedto the interface between the shell and the oil phase reacts withTBT, leading to the growth of the shell. A similar diffusion-basedprocess has been reported by Steinbacher et al.27 Besides, silicacolloidal particles in the silica sol are transported together withthe water. Thus, the shell of the microspheres contains bothtitania and silica.3.2. Characterization. By adjusting the oleic acid concentra-

tion (wOA) in the continuous phase and the residence time (t),microsphere samples with different titania molar fraction on thesurface (xtitania) were obtained. Figure 3a shows that the calcined

Figure 3. (a)Micrograph of calcined hybridmicrospheres. (b, e, h, k) SEM images of the hybridmicrosphere surface and its magnification with differenttitania loading content (xtitania). The residence time (t) and oleic acid concentration (wOA) used for the sample preparation are as follows: (b) t = 50min,wOA = 1%; (e) t = 0 min, wOA = 0.6%; (h) t = 0 min, wOA = 0.1%; (k) t = 50 min, wOA = 0.1%. (c, f, i, l) SEM images of the cross-section of the hybridmicrosphere in b, e, h, and k and its magnification. (d, g, j, m) EDS spectra of the hybrid microsphere at surface area 1 in b, 2 in e, 3 in h, and 4 in k. (n)EDS spectrum of the hybrid microsphere at cross-section area in i. Gold is present because the samples are gold-coated to prevent charging in the SEM.

13245 dx.doi.org/10.1021/la202055f |Langmuir 2011, 27, 13242–13247

Langmuir ARTICLE

microspheres are highly uniform in size. More detailed structuresof the hybrid microcapsules were observed using SEM. EDSanalysis was performed on the microsphere surface, as shown inFigure 3d,g,j,m. High levels of silicon and oxygen and a low levelof titanium were detected in areas 1 and 2 of Figure 3d,g, whileareas 3 and 4 of Figure 3j,m show dramatically decreased levels ofsilicon and an increased level of titanium, indicating that xtitaniacan be adjusted in a wide range by controlling wOA and t. Theinfluence of the two parameters on the titania loading contentwill be investigated in detail in section 3.4. The same content oftitanium could be detected at different areas of the microspheresurface, indicating that titania has a uniform distribution on themicrosphere surface. Microspheres with different titania loadingcontent have different surface morphologies. The microspherewith low titania loading content has a smooth surface (Figure 3b).When the loading content increases, the surface becomes rough-er (Figure 3e,h,k). The cross-section images (Figure 3c,f,i,l) showthat the hybrid microspheres are composed of a compact thinsurface layer around the porous inner part. When the titaniacontent is low, the surface layer is in the nanoscale. With theincrease of the titania content, the surface layer becomes thickerand surface layers in the microscale are formed. This is mostlikely due to titania hydroxide being first formed by a fasthydrolysis reaction between TBT and a large amount of waterat the oil�water interface. As a result of fast reaction, highsupersaturation is reached and a very fast nucleation occurs,which leads to the formation of a large number of tiny primaryparticles. The primary particles assemble on the oil�waterinterface to form a membrane. If the reaction is ceased at thisstage, microspheres with smooth surface and low content oftitania will be obtained. Otherwise, water in the droplet willtransfer through the membrane and react with TBT to generaterelatively high supersaturation on the outer surface of themembrane. In this situation, heterogeneous nucleation will occuron the outer surface of the membrane to form many granules,leading to a rough and thick surface as well as a high titaniacontent on the surface. The EDS analysis result (Figure 3n)shows that the porous core is composed of pure silica. Further-more, XRD analyses were carried out on the microspheres withdifferent titania loading content. The results (Figure 4) show thattitania crystallized into a pure anatase structure. When the titaniacontent was extremely low (xtitania = 1.4%), the characteristicpeak of anatase titania at 2θ = 25.34� disappeared and a broadmaximum between 2θ values of 15 and 35� was observed,indicating amorphous silica.The above results verify that titania�silica core�shell micro-

spheres were successfully prepared using the novel microfluidic

method in a single step. Furthermore, the microspheres aremonodispersed, and titania are uniformly distributed on thesurface.3.3. Control of Microsphere Size. In our experiments, the

size of themicrospheres before calcination was nearly the same asthe droplet size. Previous studies using flow focusing methods incoaxial microfluidic devices have shown that the droplet forma-tion is determined by the balance between the shear andinterfacial forces.31,32 For the experimental system used, therewas very little variation in the liquid-phase viscosities andliquid�liquid interfacial tension. Besides, the influence of thedispersed phase flow rate on droplet size is slight.33,34 Thus,droplet size is mainly influenced by the continuous phase flowrate. The detailed investigation of the influence of flow rate ondroplet size and microsphere size was carried out and shown inFigure 5. The droplet size decreases with the increase of thecontinuous phase flow rate. After calcinations, the microsphereshrank to about half of the initial diameter. Therefore, we caneasily control microsphere size by changing the continuous phaseflow rate. The calcinated microsphere average diameter rangedfrom 100 to 300 μm in our experiments.3.4. Control of Titania Content on the Surface. Titania

loading capacity is an important hybrid microsphere property toconsider when determining potential applications. In this study,we investigated wOA in the continuous phase and t on xtitania. To

Figure 4. XRD patterns of microspheres with different titania contenton the surface.

Figure 5. Effects of continuous phase flow rate Qc on the dropletdiameter ddroplet and the calcined hybrid microsphere diameterdmicrosphere.

Figure 6. Aqueous drops were hanged in liquid paraffin solutions with3% TBT for 40 min. Oleic acid were added to the oil phase atconcentrations of (a) 0.1, (b) 0.3, and (c) 0.7%.

13246 dx.doi.org/10.1021/la202055f |Langmuir 2011, 27, 13242–13247

Langmuir ARTICLE

better understand the effect of oleic acid, we hanged one aqueousdrop with a microneedle in the oil phases with different oleic acidconcentrations for 40 min. Figure 6 shows the images capturedwith a CCD camera at different times. The droplets graduallyturn from transparent to opaque as time goes by. Droplet colorreflects the degree of reaction. Droplet immersed in the liquidparaffin solution with 0.1% oleic acid turns completely opaque in20 min. In contrast, the other two droplets immersed in liquidparaffin solutions with higher oleic acid concentrations are stillsemitransparent when immersed for 40 min.When immersed forthe same time, the lower the concentration of oleic acid, thedeeper the droplet color is. The results demonstrate that the oleicacid concentration significantly affects the reaction rate.EDS analysis was used to determine the molar ratio of the two

components. Figure 7a shows that higher oleic acid concentra-tion in the continuous phase leads to a lower loading capacity.When the oleic acid concentration was higher, a lower hydrolysisrate resulted in a slower growth of titania hydroxide on thedroplet surface. Figure 7b shows that the titania loading contentincreases with the increase of residence time. Longer residencetimes allow for extended titania loading Besides, the titanialoading rate also depends on the concentration of oleic acidand decreases with increasing oleic acid concentration. Thedecreased titania loading rate is due to the reduction in thehydrolysis reaction rate by oleic acid. The result shows thatthe addition of oleic acid could reduce the sensitive degree of theloading capacity to the operation parameters, which is beneficialto the accurate control of the loading content. In our experi-ments, the molar ratio of titania on the surface could be adjustedfrom 1.0 to 98.0 mol % by changing the concentration of oleicacid in the continuous phase (0.1�1.0 wt %) and the time thatthe hydrolysis reaction lasts (5�50 min).

4. CONCLUSIONS

In this work, we describe a novel and simple microfluidicmethod to prepare monodispersed titania�silica core�shellmicrospheres in one step. Titanium alkoxide precursor TBTwas dissolved in liquid paraffin containing oleic acid as a chemicalmodifier. Silica sol was emulsified in the oil phase via a coaxialmicrofluidic device. When the droplets were moving in thechannel, TBT hydrolyzed on the droplet surface with theassistant of water in silica sol. The titanium hydroxide formedon the surface, and the silica microspheres coated by titania weresuccessfully prepared after calcinations. Titania is uniformlydistributed on the sphere surface, and its loading capacity canbe adjusted from almost 1.0 to 98.0 mol % by changing the oleicacid concentration in the continuous phase and the residence

time. Furthermore, the microsphere average diameter can beeasily controlled from 100 to 300 μmby changing the continuousphase flow rate. The microfluidic approach has the advantages ofsimplicity and good reproducibility of a one-step process, uni-formity of the coating, good sphericity of the spheres, and easycontrol of sphere size and surface composition. Besides, themicroscale sphere coated with catalysts would protect catalystsfrom agglomeration and make them easily recoverable, both ofwhich are hard to achieve with the use of nanoscale particles.

’AUTHOR INFORMATION

Corresponding Author*E-mail: gsluo@tsinghua.edu.cn (G.L.); xujianhong@tsinghua.edu.cn (J.X.).

’ACKNOWLEDGMENT

We gratefully acknowledge the support of the NationalNatural Science Foundation of China (Grants 20876084 and21036002) and A Foundation for the Author of NationalExcellent Doctoral Dissertation of People's Republic of China(FANEDD: 201053) for this work.

’REFERENCES

(1) Yang, M.; Ma, J.; Zhang, C. L.; Yang, Z. Z.; Lu, Y. F. Angew.Chem., Int. Ed. 2005, 44, 6727.

(2) Shchukin, D. G.; Ustinovich, E. A.; Sukhorukov, G. B.; Mohwald,H.; Sviridov, D. V. Adv. Mater. 2005, 17, 468.

(3) Ren, M. M.; Ravikrishna, R.; Valsaraj, K. T. Environ. Sci. Technol.2006, 40, 7029.

(4) Jeong, U.; Wang, Y. L.; Ibisate, M.; Xia, Y. N. Adv. Funct. Mater.2005, 15, 1907.

(5) Li, Y.; Shi, J.; Hua, Z.; Chen, H.; Ruan, M.; Yan, D. Nano Lett.2003, 3, 609.

(6) Srinivasan, S.; Datye, A. K.; Hampdensmith, M.; Wachs, I. E.;Deo, G.; Jehng, J. M.A.; Turek, M.; Peden, C. H. F. J. Catal. 1991, 131,260.

(7) Khan, S. A.; Jensen, K. F. Adv. Mater. 2007, 19, 2556.(8) Caruso, R. A.; Susha, A.; Caruso, F. Chem. Mater. 2001, 13, 400.(9) Kamata, K.; Lu, Y.; Xia, Y. J. Am. Chem. Soc. 2003, 125, 2348.(10) Wang, D. Y.; Caruso, F. Chem. Mater. 2002, 14, 1909.(11) Nakamura, H.; Ishii, M.; Tsukigase, A.; Harada, M.; Nakano, H.

Langmuir 2005, 21, 8918.(12) Wan, J.; Bick, A.; Sullivan, M.; Stone, H. A. Adv. Mater. 2008,

20, 3314.(13) Kim, J. W.; Utada, A. S.; Fernandez- Nieves, A.; Hu, Z. B.;

Weitz, D. A. Angew. Chem., Int. Ed. 2007, 46, 1819.(14) Wei, J.; Ju, X. J.; Xie, R.; Mou, C. L.; Lin, X.; Chu, L. Y. J. Colloid

Interface Sci. 2001, 357, 101.

Figure 7. Influence of (a) oleic acid concentration and (b) residence time on the titania loading content.

13247 dx.doi.org/10.1021/la202055f |Langmuir 2011, 27, 13242–13247

Langmuir ARTICLE

(15) Oh, H. J.; Kim, S. H.; Baek, J. Y.; Seong, G. H.; Lee, S. H.J. Micromech. Microeng. 2006, 16, 285.(16) Peng, S. L.; Zhang, M. Y.; Niu, X. Z.; Wen, W. J.; Sheng, P.; Liu,

Z. Y.; Shi, J. Appl. Phys. Lett. 2008, 92, 012108.(17) Kim, C.; Chung, S.; Kim, Y. E.; Lee, K. S.; Lee, S. H.; Oh, K. W.;

Kang, J. Y. Lab Chip 2011, 11, 246.(18) Seiffert, S.; Thiele, J.; Abate, A. R.;Weitz, D. A. J. Am. Chem. Soc.

2010, 132, 6606.(19) Gong, X. Q.; Peng, S. L.; Wen, W. J.; Sheng, P.; Li, W. H. Adv.

Funct. Mater. 2009, 19, 292.(20) Ren, P. W.; Ju, X. J.; Xie, R.; Chu, L. Y. J. Colloid Interface Sci.

2010, 343, 392.(21) Liu, L.; Yang, J. P.; Ju, X. J.; Xie, R.; Liu, Y. M.; Wang, W.;

Zhang, J. J.; Niu, C. H.; Chu, L. Y. Soft Matter 2011, 7, 4821.(22) Torza, S.; Mason, S. G. Science 1969, 163, 813.(23) Nisisako, T.; Hatsuzawa, T.Microfluid. Nanofluid. 2010, 9, 427.(24) Choi, C. H.; Yi, H.; Hwang, S.; Weitz, D. A.; Lee, C. S. Lab Chip

2001, 11, 1477.(25) Takeuchi, S. J.; Garstecki, P.; Weibel, D. B.; Whitesides, G. M.

Adv. Mater. 2005, 17, 1067.(26) Quevedo, E.; Steinbacher, J.; McQuade, D. T. J. Am. Chem. Soc.

2005, 127, 10498.(27) Steinbacher, J. L.; Moy, R. W. Y.; Price, K. E.; Cumings, M. A.;

Roychowdhury, C.; Buffy, J. J.; Olbricht, W. L.; Haaf, M.; McQuade,D. T. J. Am. Chem. Soc. 2006, 128, 9442.(28) Eun, T. H.; Kim, S. H.; Jeong,W. J.; Jeon, S. J.; Kim, S. H.; Yang,

S. M. Chem. Mater. 2009, 21, 201.(29) Li, S. W.; Xu, J. H.; Wang, Y. J.; Lu, Y. C.; Luo, G. S.

J. Micromech. Microeng. 2009, 19, 015035.(30) Dunuwila, D. D.; Gagliardi, C. D.; Berglund, K. A. Chem. Mater.

1994, 6, 1556.(31) Utada, A. S.; Fernandez-Nieves, A.; Stone, H. A.; Weitz, D. A.

Phys. Rev. Lett. 2007, 99, No. 094502/1-4.(32) Cramer, C.; Fischer, P.; Windhab, E. J. Chem. Eng. Sci. 2004,

59, 3045.(33) Cristini, V.; Tan, Y. C. Lab Chip 2004, 4, 257.(34) Xu, J. H.; Li, S.W.; Tan, J.;Wan, Y. J.; Luo, G. S. Langmuir 2006,

22, 7943.