Synthesis and applications of porous non-silica metal oxide submicrospheres

Abstract

Nowadays the development of submicroscale products of specific size and morphology that feature a high surface area to volume ratio, well-developed and accessible porosity for adsorbates and reactants, and are non-toxic, biocompatible, thermally stable and suitable as synergetic supports for precious metal catalysts is of great importance for many advanced applications. Complex porous non-silica metal oxide submicrospheres constitute an important class of materials that fulfill all these qualities and in addition, they are relatively easy to synthesize. This review presents a comprehensive appraisal of the methods used for the synthesis of a wide range of porous non-silica metal oxide particles of spherical morphology such as porous solid spheres, core–shell and yolk–shell particles as well as single-shell and multi-shell particles. In particular, hydrothermal and low temperature solution precipitation methods, which both include various structure developing strategies such as hard templating, soft templating, hydrolysis, or those taking advantage of Ostwald ripening and the Kirkendall effect, are reviewed. In addition, a critical assessment of the effects of different experimental parameters such as reaction time, reaction temperature, calcination, pH and the type of reactants and solvents on the structure of the final products is presented. Finally, the practical usefulness of complex porous non-silica metal oxide submicrospheres in sensing, catalysis, biomedical, environmental and energy-related applications is presented.

---

Yash Boyjoo

Yash Boyjoo received his PhD in 2013 from Curtin University, Australia. After two years as a research associate at Curtin University working with Dr Jian Liu and Prof. Vishnu Pareek, he currently works as a postdoctoral research fellow at Université Lille 1 in France, where his studies deal with the synthesis of materials for VOC elimination from polluted air. His other research interests include photocatalytic wastewater treatment, synthesis of activated carbons for application as CO2 adsorbents and supercapacitors, and the design and modelling of photocatalytic reactors.

Meiwen Wang

Meiwen (Sharon) Wang received her bachelor's degree with first class honours in Chemical Engineering at Curtin University, where she is now pursuing PhD research under the supervision of Dr Jian Liu. She is interested in the design and fabrication of yolk–shell particles and their use as nanoreactors.

Vishnu K. Pareek

Vishnu Pareek is a Professor in the Department of Chemical Engineering at Curtin University, Australia. He received his PhD degree from the University of New South Wales, an MTech from IIT Delhi and a BE from the University of Rajasthan, Jaipur, all in Chemical Engineering and with high distinction. Prof. Pareek's expertise lies in the simulation and design of chemical processes with particular emphasis on the use of these tools in industrial-scale processes. He is an ardent cricket fan and an enthusiastic reader especially on subjects related to politics, environment and economy. In his spare time, he is a bird watcher with special interest in understanding the behaviour of birds of prey.

Jian Liu

Jian Liu obtained his PhD degree in Physical Chemistry from the Dalian Institute of Chemical Physics, Chinese Academy of Science, in 2008. Subsequently, he moved to Australia and worked as a Postdoctoral Research Fellow at AIBN, the University of Queensland. He started as a Lecturer at Curtin University in 2013 and was promoted to Senior Lecturer in 2014. Dr Liu has published more than 120 peer reviewed journal articles with over 6300 citations (H-index of 43). He has been honoured with a prestigious UQ Foundation Research Excellence Award and Australian Postdoctoral Fellowship. His current research interests are nanoreactor design, green chemical processes, and utilization of CO2.

Mietek Jaroniec

Mietek Jaroniec received his MS and PhD from M. Curie-Sklodowska University, Poland, in 1972 and 1976, respectively. Since 1991 he has been a Professor of Chemistry at Kent State University, Kent, Ohio (USA). Before joining Kent State he was a Professor of Chemistry at M. Curie-Sklodowska University, Poland. His research interests revolve primarily around inter-disciplinary topics of interfacial chemistry, and chemistry of materials, including physical adsorption at the gas/solid and liquid/solid interfaces, adsorbents, and catalysts. At Kent State he has established a vigorous research program in the area of ordered nanoporous materials such as ordered mesoporous silicas, organosilicas, inorganic oxides and carbons, focusing on their synthesis and environmental and energy-related applications.

---

1. Introduction

The design of functional nano- and microsized particles is a hot topic due to their application in a wide range of fields such as energy conversion and storage, catalysis, biomedicine, and environmental remediation. These tiny particles feature a high surface area to volume ratio, which is beneficial for diffusion and transport of reactants and products. Furthermore, depending on the material used, the particle size, crystal size, surface area, pore size distribution, and morphology can be tailored for specific applications by adjusting the chemical composition and synthesis conditions.

Metal oxides (MeO) exhibit several attractive features such as high mechanical strength, thermal stability, chemical inertness, non-toxicity, biocompatibility, oxygen vacancies, semiconductor properties and high isoelectric point, and can act as supports for noble metals such as Au, Pt and Pd, or rare earth metals to achieve synergistic catalytic activity toward specific chemical reactions. Furthermore, for nanosized crystals, quantum effects become important as reflected by significantly different optical and electronic properties from those observed for the bulk phases,¹ which can have favorable outcomes for their application in catalysis and photocatalysis; for example, new physicochemical phenomena such as ferromagnetism and paramagnetism can be achieved for otherwise antiferromagnetic systems, as in the case of NiO spheres.² As a result, countless studies have been performed to synthesize submicrosized porous non-silica metal oxide particles with different morphologies, phases, sizes, crystal sizes and pore size distributions. These particles are useful for a variety of industrial, biomedical, environmental and energy-related applications as illustrated in Fig. 1.

Fig. 1 Spherical metal oxide particles and their potential applications.

The porous non-silica metal oxide submicrospheres can be designed to have large surface areas and well-developed porosity to enhance interfacial interactions with reactants and facilitate the transport and diffusion of reactants and products. Also, the crystal size and growth directions need to be considered for specific catalytic and semiconductor applications. Furthermore, a proper balance between micropores, mesopores and macropores that respectively act as reaction sites, distribution/evacuation pathways and reservoirs is desired. The spherical morphology is the most stable shape that is achieved in nature. In the case of metal oxides, the advantages offered by spherical structures are high mechanical strength, short pathways for diffusion of species, dispersion enhancement due to the stabilization of electrostatic charges, high surface area to volume ratios, easy coating with other species or metal oxides, minimization of viscous effects and predictable hydrodynamics. Typical spherical morphologies of metal oxides discussed in this review can be classified into three groups: porous solid spheres (Fig. 2Aa), core–shell spheres (Fig. 2Ab) and hollow spheres (Fig. 2Ac). The core–shell spheres shown in Fig. 2Ab can be further extended into single-core particles with a multi-particle (raspberry-like) shell (Fig. 2Ab1), multi-core particles with a single shell (Fig. 2Ab2), single-core particles with a multi-shell (Fig. 2Ab3), or a combination of particles shown in Fig. 2Ab and b1 (see Fig. 2Ab4). Also, hollow spheres can be further classified into yolk–shell spheres (Fig. 2Ac1) and multi-shell hollow spheres (Fig. 2Ac2). In principle, each compartment of these spheres can be nonporous or porous with different pore sizes (micropores, mesopores, and macropores). More complex metal oxide spheres can be proposed and synthesized by combining and modifying the aforementioned nine patterns as shown in Fig. 2A.

Fig. 2 (A) Graphically illustrated classification of metal oxide spheres: (a) porous spheres; (b) core–shell spheres: (b1) raspberry-like core–shell spheres; (b2) multi-core spheres with a single shell; (b3) single-core spheres with multi-shells; (b4) raspberry-like single-core with multi-shells; and (c) hollow spheres: (c1) yolk–shell spheres; (c2) multi-shell hollow spheres. (B) SEM and TEM images of various metal oxide spherical particles: (a) dense spheres; (b) porous spheres; (c) core–shell spheres; (d and e) yolk–shell spheres; (f) hollow spheres. Panel B reproduced with permission from ref. 3. Copyright © 2007, American Chemical Society.

Several excellent reviews have been published previously focusing on the synthesis and applications of metal oxides. However, these reviews consider either a specific metal oxide such as TiO₂^4,5 and iron oxides,⁶ or a specific application such as gas sensors,⁷ energy storage and conversion applications^8,9 or a specific morphology such as hollow spheres¹⁰ and multi-shell structures.¹¹ Indeed, a concise and up-to-date review on the synthesis strategies of different types of porous non-silica metal oxide submicrospheres with various spherical morphologies as well as their emerging applications is timely as this research field continuous to rapidly grow. This review is focused on the synthesis and applications of non-silica-based metal oxide spheres based on the literature for the past five years. One of the main sections of this review is devoted to the major strategies frequently used for the preparation of MeO particles such as hydrothermal/solvothermal synthesis at elevated temperatures (>100 °C), solution precipitation synthesis at low temperatures (<100 °C), and the aerosol-type synthesis. Other methods are only briefly mentioned. The hydrothermal and low temperature precipitation syntheses are general methods that allow for the development of more sophisticated structures by hard templating, soft templating or controlled hydrolysis, or by taking advantage of Ostwald ripening and the Kirkendall effect.^11,12 The aforementioned section on the synthesis methods is supplemented by a critical appraisal of the effect of different experimental parameters such as reaction time, reaction temperature, calcination, pH and the type of reactants and solvents on the structure of the resulting metal oxide particles. Finally, the last section presents the major applications of metal oxide spheres in industrially relevant sensing and catalysis, in the development of biomedically relevant photoluminescent devices and drug delivery vehicles, in environmentally relevant photocatalysis and adsorption, and in energy-relevant applications such as lithium-ion batteries, supercapacitors and dye sensitized solar cells.

2. Synthesis methods

Metal oxide particles have been generated by using different synthetic strategies, for example, colloidal synthesis, sol–gel process, aerosol process, precipitation, hydrothermal/solvothermal synthesis, hot injection, and non-aqueous and non-hydrolytic chemistry method (e.g., solid-state reactions, solid–gas reactions). In this section we summarize the main methods for the preparation of MeO particles with spherical structures such as the hydrothermal synthesis at elevated temperatures (>100 °C), low temperature solution precipitation synthesis (<100 °C) and the spray method. The first two methods include structure defining approaches such as hard templating, soft templating, hydrolysis, Ostwald ripening and the Kirkendall effect. This section ends with a brief overview of other methods involving electrodeposition, laser irradiation or ultrasonic irradiation.

2.1 Hydrothermal and solution precipitation methods

The hydrothermal/solvothermal synthesis of metal oxide particles involves the use of a batch reactor at elevated temperatures (above 100 °C) and pressures. It offers a controlled environment for interactions of the salts used as metal oxide with other reactants in solution to form nanosized crystallites. The elevated pressure and temperature conditions facilitate dissolution of precursors and recrystallization of materials that can be insoluble under ambient conditions. As a result, high purity, homogeneous, metastable and often crystalline products are formed with unique properties and narrow particle size distribution.¹³ This synthesis method is flexible and well suited for the control of morphology and crystallinity of the porous metal oxide spheres by varying experimental parameters such as reaction temperature and time, type of reactants and solvents, and the chemical composition of the synthesis mixture. During the hydrothermal process, the nanosized crystallites self-assemble (with or without additives such as polymers and surfactants) into more complex architectures with optimal stability and lowest surface energy; hence, spherical shapes are favored. Also, metal oxide spherical structures can be synthesized via solution precipitation using milder conditions (i.e., temperature below 100 °C and atmospheric pressure) as reported elsewhere.^13–20

Tables 1 and 2 present the pertinent experimental conditions and characteristic features of solid spheres and shell-type particles obtained by hydrothermal synthesis together with relevant references.^21–119 As can be seen from these tables, the hydrothermal method is very popular for the synthesis of various categories of metal oxides such as alkaline earth metals (MgO), rare-earth metals (CeO₂, Y₂O₃), transition metals (TiO₂, V₂O₅, Cr₂O₃, MnO₂, Fe₂O₃, Co₃O₄, NiO, CuO, ZnO, ZrO₂, Nb₂O₅, MoO₂, Ta₂O₅ and WO₃) and post-transition metals (Al₂O₃, Ga₂O₃, In₂O₃, Bi₂O₃ and SnO₂). Composites, perovskites and doped metal oxides can also be created by controlling the ratio of the precursors.

Table 1 Porous solid spheres synthesized by the hydrothermal method

Ref.	Particle type	Reactants	Solvents	Hydrothermal conditions	Calcination conditions	Particle size	BET SA (m^2 g^−1)	Pore volume (cm^3 g^−1)	Pore size (nm)
a Microwave heating. b By varying the concentration of NbCl5 between 0.3 g and 0.5 g in 25 ml solution. c Prepared by an antisolvent method. d For nanowire-, flower- and urchin-like spheres respectively by increasing urea concentration. e For nanowire-, flower- and urchin-like spheres respectively. f For W(CO)6 concentration between 4.26 mM and 28.4 mM. g For WO3/TiO2 between 2% and 10%. h Supercritical temperature. NR: not reported.
22	Bi2O3 spheres	Bi(NO3)3·5H2O, PVP	EG	180 °C, 0.17 h^a	400 °C, 3 h in air	10 μm	NR	NR	NR
23	Bi2O3 spheres	Bi(NO3)3·5H2O, HNO3, urea	Water, EG	150 °C, 3 h	None	350 nm	8	0.018	NR
		Bi(NO3)3·5H2O, NaOH, HNO3, PVP				100 nm	23	0.15
24	Bi2WO6 spheres	Na2WO4·2H2O, Bi(NO3)3·5H2O, PVP K30	Water	180 °C, 12 h	None	4 μm	NR	NR	NR
25	Bi2WO6 perovskite spheres	Bi(NO3)3·5H2O, Na2WO4·2H2O, NaHCO3, citric acid	Water	200 °C, 18 h	None	2 μm	24	NR	NR
26	CeO2 spheres	Ce(NO3)3·6H2O	Water, C2H5COOH, EG	180 °C, 3.3 h	None	130 nm	216	NR	3.8
27	Co3O4 spheres	Co(CH3COO)2·4H2O, NH3	Water, EG	180 °C, 12 h	500 °C, 4 h in air	2–5 μm	13	NR	NR
28	Co3O4 spheres	Co(NO3)2·6H2O, urea	Water	160 °C, 6 h	300 °C, 2 h in air	8–20 μm	30	0.245	17
29	CoFe2O4 spheres	CoCl2·6H2O, FeCl3·6H2O, urea	Water, ethanol	170 °C, 0.42 h^a	500 °C in air	1 μm	25	0.18	25
30	Cr2O3 spheres	C15H21CrO6, NH4HCO3	Ethanol	250 °C, 2 h	500 °C, 4 h	1–1.2 μm	15	NR	20–80
31	Cr2O3 spheres	Cr(NO3)3·9H2O, H2C2O4, urea	Ethanol, PEG	180 °C, 5 h	500 °C, 2 h in air	2–3 μm	NR	NR	NR
32	CuO spheres	Cu(CH3COO)2	Water, EG	160 °C, 1 h	None	412 nm	168	NR	5
33	CuO spheres	Cu(CH3COO)2, NH3, sodium alginate	Water	160 °C, 2 h	None	500 nm	21	NR	NR
34	α-Fe2O3 spheres	FeCl3·6H2O, ascorbic acid, urea	Water	160 °C, 4 h	500 °C, 4 h in air	0.5–5 μm	20	0.11	2–50
35	α-Fe2O3 spheres	Fe(NO3)3·9H2O	Water, 2-butanone	140 °C, 12 h	None	100 nm	NR	NR	NR
36	α-Ga2O3 spheres	Ga(NO3)3, oxalic acid	Water	200 °C, 10 h	450 °C, 3 h	0.5–4 μm	62	0.193	12.3
37	Cubic-In2O3 spheres	InCl3·4H2O, citric acid	Water, ethylenediamine	180 °C, 7 h	400 °C, 0.17 h in air	150–200 nm	88	NR	NR
	Hexagonal-In2O3 spheres	InCl3·4H2O, tartaric acid	Water, ethylenediamine	180 °C, 7 h	400 °C, 0.17 h in air	150–200 nm	85	NR	NR
38	In2O3 spheres	InCl3·4H2O, urea, sodium citrate	Water, EG	200 °C, 16 h	400 °C, 2 h in air	600–700 nm	19	NR	NR
39	Nb2O5 spheres	NbCl5	Ethanol	200 °C, 24 h	550 °C, 2 h in air	200–900 nm^b	23–68^b	NR	NR
40	Nb2O5 spheres	Glycolated Nb2O5 spheres^c	Water	180 °C, 12 h	None	400–500 nm	312	0.567	2
41	NiO spheres	Ni(NO3)2·6H2O, Na2SO4, NaOH, glycine	Water	180 °C, 0.5 h^a	300 °C, 3 h in air	2 μm	202	NR	25
42	NiO spheres	Ni(NO3)2·6H2O, NaCl, sodium acetate	EG	190 °C, 8 h	300 °C, 2 h in air	600 nm	222	NR	4–10
2	NiO spheres	NiCl2, sodium acetate, polyethyleneimine	Water, triethanolamine	200 °C, 8 h	270 °C, 0.5 h in air	500 nm	60	NR	10–30
43	NiO spheres	NiCl2·H2O, urea	Water	100 °C, 20 h	300 °C, 2 h in air	3–4 μm	200–240^d	NR	3.2, 8.9, 4^e
44	La doped NiO spheres	Ni(NO3)2·6H2O, La(NO3)2·6H2O, NH3, glucose	Water	140 °C, 12 h	550 °C, 4 h in air	1–2 μm	278	0.79	2–50, >50
45	SnO2 spheres	Na2SnO3·3H2O, sodium alginate	EG, water	180 °C, 24 h	None	200–400 nm	29	NR	15
46	SnO2 spheres	SnCl2·2H2O, NaClO, HCl	Ethanol	180 °C, 12 h	None	150 nm	62	NR	4
47	SnO2 spheres	SnCl4·5H2O, PVP	Methanol	180 °C, 3 h	500 °C, 2 h in air	400–700 nm	78	NR	10
48	SnO2 spheres	SnCl4·5H2O, PVP	Methanol	180 °C, 3 h	500 °C, 2 h in air	500–700 nm	78	NR	10
49	SnO2@C spheres	K2SnO3·3H2O, glucose	Water	180 °C, 4 h	450 °C, 4 h in N2	100 nm	NR	NR	NR
50	C–V2O3 spheres	NH4VO3, citric acid	Ethanediol, water	180 °C, 24 h	600 °C, 3 h in N2	2 μm	45	NR	20
51	V2O5 spheres	VO(OiPr)3	Acetic acid	200 °C, 1.5 h	350 °C, 0.5 h in air	4–10 μm	42	NR	NR
52	WO2 spheres	W(CO)6	Ethanol	200 °C, 24 h	None	0.7–1.5 μm^f	78–114^f	NR	NR
53	WO3 spheres	WCl6, carbon microspheres	Dimethylformamide	120 °C, 4 h	420 °C, NR	150–220 nm	22	0.0447	NR
54	WO3/TiO2 spheres	(NH4)10H2(W2O7)6, TiOSO4, P123	Water, ethanol	140 °C, 16 h	500 °C, 6 h in air	NR	45–64^g	0.21–0.26^g	12.2–15.2^g
55	Eu^3+:Y2O3 spheres	Y(NO3)3·6H2O, Eu(NO3)3, KOH	Water, 1-propanol	400 °C^h, 0.17 h	1000 °C, 1 h in air	2–3 μm	NR	NR	NR
56	ZnO spheres	Zn(NO3)2·6H2O, L-asparagine, urea	Water	100 °C, 3 h	300 °C, 0.5 h in air	Several μm	194	NR	5
57	ZnO spheres	Zn(NO3)2, urea	Water	120 °C, 2 h	450 °C, 2 h in air	10 μm	38	NR	8.67
58	ZnO spheres	Zn(NO3)2·6H2O, trisodium citrate, urea	Water	120 °C, 6 h	300 °C, 2 h in air	4–6 μm	40	NR	20–60
59	ZnO spheres	Zn(CH3COOH)2·2H2O, MEA, urea	Water	120 °C, 12 h	450 °C, 2 h in air	1–2 μm	40	NR	5–50
60	ZnO spheres	Zn(CH3COO)2·2H2O, NaOH, citric acid	Water, ethanol	120 °C, 24 h	None	2–3 μm	42	NR	2–30
61	ZnO spheres	Zn(CH3COO)2, thiourea	Water	180 °C, 10 h	500 °C, 3 h in air	3–5 μm	21	NR	22.6
60	Ag loaded ZnO spheres	Zn(CH3COO)2·2H2O, AgNO3, NaOH, citric acid	Water, ethanol	120 °C, 24 h	None	2–3 μm	37	NR	NR
62	ZrO2 spheres	ZrOCl2·8H2O, HCl, urea	Ethanol, water	160 °C, 2 h	None	1–2 μm	102	0.09	2–102

---

Table 2 Porous shells synthesized by the hydrothermal method

Ref.	Particle type	Reactants	Hollowing mechanism	Solvents	Hydrothermal conditions	Calcination conditions	Particle and shell dimensions	BET SA (m^2 g^−1)	Pore volume (cm^3 g^−1)	Pore size (nm)
a Microwave heating. b For F/Fe fractions between 0 and 1. c By varying the fraction of glycerol in water between 0.05 and 0.125. d (acac) = CH3COCH=C(O−)CH3. e Ratio SiO2 : Ta2O5 = 1 : 0.85. f Ratio SiO2 : Ta2O5 = 1 : 1.7. g Nanosheet shell. h Porous shell. NR: not reported.
63	γ-Al2O3 hollow spheres	KAl(SO4)2·12H2O, urea	Ostwald ripening	Water	170 °C, 3 h	600 °C, 2 h in air	4–6 μm, shell thickness 700–900 nm	149	0.45	12.3
64	Perovskite BaZrO3 hollow spheres	Ba(NO3)2, ZrOCl2·8H2O, KOH	Ostwald ripening	Water	200 °C, 24 h	None	160 nm, shell thickness 15 nm	NR	NR	NR
65	Bi2O3/Co3O4 hollow spheres	Bi(NO3)3·5H2O, Co(NO3)3·6H2O, PEG, NaAc	Ostwald ripening	EG	180 °C, 12 h	500 °C, 2 h in air	2–6 μm	46	0.16	NR
66	BiFeO3 hollow spheres	Bi(NO3)3·5H2O, Fe(NO3)3·9H2O, citric acid	Ostwald ripening	Glycerol, ethanol	160 °C, 24 h	500 °C, 24 h in air	1.5 μm, shell thickness 0.2 μm	15	NR	NR
67	CeO2 hollow spheres	Ce(NO3)3·6H2O, HCl, citric acid	Ostwald ripening	Water	160 °C, 24 h	365 °C, 1.5 h in air	2–4 μm	56	NR	NR
68	CeO2 hollow spheres	Ce(NO3)3·6H2O, PVP, H2O2, urea	Ostwald ripening	Water	180 °C, 24 h	None	126 nm	21	NR	4
69	CeO2 hollow spheres	Ce(NO3)3·6H2O, adipic acid	Ostwald ripening	Water, EG	180 °C, 5 h	None	135 nm	145	NR	4
70	CeO2 hollow spheres	CeCl3·7H2O, H2O2, urea	Ostwald ripening	Water	180 °C, 10 h	None	300 nm, shell thickness 50 nm	85	0.23	3–10
71	CeO2 hollow spheres	Ce(NO3)3·6H2O, PVP	Ostwald ripening	EG, ethanol, water	180 °C, 24 h	None	160 nm	66	0.181	3–30
72	CeO2 hollow spheres	CeCl3·7H2O, urea	Ostwald ripening	Water	180 °C, 4 h	None	300 nm, shell thickness 30 nm	37	NR	36
73	Yolk–shell CeO2	Ce(NO3)3·6H2O, PVP, NH4Ac·2H2O	Carbon spheres	Ethanol	180 °C, 12 h	600 °C, 3 h in air	180 nm	NR	NR	NR
74	Multi-yolk–shell Pd@CeO2 spheres	Pd@SiO2, Ce(NO3)3·9H2O	SiO2 etching	EG, CH3COOH, water	130 °C, 12 h	350 °C, 2 h in H2	150–200 nm	104	0.078	2–25
75	Co3O4 hollow spheres	Co(NO3)2·6H2O, sodium citrate, HMT, sucrose	In situ carbon from sucrose	Water	140 °C, 24 h	500 °C, 5 h in air	Shell thickness 130 nm	60	NR	7.8
76	Co3O4 hollow spheres	Co(NO3)2	Ostwald ripening	Glycerol, isopropanol	180 °C, 6 h	200 °C, 2 h in air	1 μm	180	NR	2–150
77	CoFe2O4 double shell spheres	CoSO4·7H2O, (NH4)2Fe(SO4)2·6H2O, sucrose	In situ carbon from sucrose	Water	180 °C, 24 h	600 °C, 2 h in air	0.5–1.5 μm, 200–500 nm hollow core	NR	NR	NR
	Same as above with half sucrose concentration						NR	38	NR	30
33	CuO hollow spheres	Cu(CH3COO)2, NH3, sodium alginate	Ostwald ripening	Water	160 °C, 6 h	None	500 nm	72	NR	NR
78	CuO hollow spheres	Cu(NO3)2, urea	Ostwald ripening	Water	180 °C, 18 h	400 °C, 2 h in air	4.5–6.5 μm	NR	NR	NR
79	CuO hollow spheres	Cu(CH3COO)2·H2O	Ostwald ripening	Water	120 °C, 24 h	None	3.5 μm, shell thickness 1.25 μm	NR	NR	NR
32	Cu2O hollow spheres	Cu(CH3COO)2, glucose	In situ carbon from glucose	Water, EG	160 °C, 1 h	None	1.5 μm, shell thickness 400 nm	37	NR	50
15	Cu/Cu2O hollow spheres	Cu(Oac)2·H2O, PVP	Ostwald ripening	Ascorbic acid	100 °C, 0.5 h^a	None	150–500 nm	19	0.118	2–100
80	CuO/Cu2O composite hollow spheres	Cu(NO3)2·3H2O, ethanolamine	Ostwald ripening	Water	180 °C, 12 h	None	1.5–3 μm	16	NR	NR
31	Cr2O3@C core shell spheres	Cr(NO3)3·9H2O, H2C2O4, urea	Controlled annealing	Ethanol, PEG	180 °C, 5 h	750 °C, 6 h in 5% Ar and 95% H2	2–3 μm	NR	NR	NR
81	α-Fe2O3 hollow spheres	FeCl3·6H2O	Ostwald ripening	Water, DMF, TFA	180 °C, 24 h	None	2 μm	4	NR	65.8
82	α-Fe2O3 hollow spheres	K3[Fe(C2O4)3]	In situ gas bubbles	Water, EG	150 °C, 48 h	450 °C, 3 h in air	190 nm	41	NR	4–12
83	α-Fe2O3 hollow spheres	FeSO4·7H2O	Quasi-emulsion droplets	Water, glycerol	145 °C, NR	None	1 μm, shell thickness 100–200 nm	103	NR	<30
84	Double-shelled α-Fe2O3 spheres	K3[Fe(CN)6], NH4H2PO4	Ostwald ripening	Water	200 °C, 30 h	None	350 nm, 200 nm core, 20 nm outer shell, 40 nm inner shell	98	NR	11.2
85	Fe3O4 hollow spheres	FeCl3·6H2O, NaOH, SDBS	Precursor templated	EG	200 °C, 1.5 h^a	300 °C, 1 h in N2	2–4 μm	62	0.131	10.2
86	Fe3O4@TiO2 double shelled yolk–shell spheres	Fe3O4@SiO2@TiO2, NaOH	Ostwald ripening + NaOH etching	Water	150 °C, 24 h	None	560 nm	150	0.27	7.5
85	γ-Fe2O3 hollow spheres	FeCl3·6H2O, SDBS, NaOH	Precursor templated	EG	200 °C, 1.5 h^a	300 °C, 1 h in air	2–4 μm	56	0.159	16.3
87	α-Fe2O3 four shelled hollow spheres	Fe(NO3)3·9H2O, L-histidine	Amino acid templated	Water	180 °C, 12 h	600 °C, 2 h in air	3 μm	14	0.07	NR
88	γ-Fe2O3 hollow spheres	FeCl3·6H2O, NH4F, ethylenediamine	Ostwald ripening	EG	200 °C, 20 h	250 °C, 5 h in air	250 nm, shell thickness 20–40 nm^b	9–19^b	NR	13.3–34.5^b
89	Fe3O4 hollow spheres	Fe(NO3)3·6H2O	Kirkendall mechanism	Glycerol, isopropanol, water	190 °C, 12 h	350 °C, 3 h in N2	900 nm, shell thickness 10 nm	89	NR	4, 5, 7
90	α-FeOOH hollow spheres	FeSO4·7H2O	Quasi-emulsion	Water, glycerol	120 °C, 24 h	None	1 μm, varied shell thickness^c	54–97^c	0.28–0.36^c	<20
91	Perovskite LaFeO3 hollow spheres	La(NO3)3·6H2O, Fe(NO3)3·9H2O, citric acid	Ostwald ripening	Water	180 °C, 24 h	800 °C, 2 h in air	2–5 μm, shell thickness 40–60 nm	49	NR	30–80 and 100–300
92	β-Ga2O3 hollow spheres	Metallic Ga, HCl, urea	In situ gas bubbles	Acetone	200 °C, 4 h	700–800 °C, 2 h in air	1–2 μm	22	NR	3
	γ-Ga2O3 hollow spheres					500–600 °C, 2 h in air	1–2 μm	31	NR	7
93	Er doped In2O3 hollow spheres	InCl3·4H2O, Er(NO3)·7H2O	Carbon spheres	Water	180 °C, 6 h	500 °C, 3 h in O2	300 nm, shell thickness 40 nm	NR	NR	NR
94	Rh-loaded In2O3 hollow spheres	In(NO3)3·xH2O, RhCl3·xH2O, D(+) glucose monohydrate	In situ carbon from glucose	Water	180 °C, 24 h	500 °C, 2 h in air	2.1 μm, shell thickness 180 nm	NR	NR	40
95	MgO hollow spheres	MgCl2·6H2O, urea	Ostwald ripening	Water, EG	120 °C, 10 h	450 °C, 1 h in air	3–4 μm	130	0.414	7
96	MgO hollow spheres	Mg(Oac)2·4H2O, PVP K-30, NH4OH	Ostwald ripening	EG	185 °C, 5 h	500 °C, 1 h in Ar + 1 h in air	1 μm	343	1.9	<30
97	MnO2 hollow spheres	KMnO4, SiO2 spheres, Pluronic F127	SiO2 etching	Water	150 °C, 48 h	None	210 nm	233	NR	NR
98	MnO2 hollow spheres	KMnO4	Hollow carbon spheres	Water	160 °C, 5 h	None	316 nm, shell thickness 69 nm	30	0.112	19.4
99	MnO2 hollow spheres	KMnO4, Ce(NO3)3·6H2O, HNO3	Ostwald ripening	Water	140 °C, 3 h	None	3–4 μm	29	0.3	2
100	C@MnO2 spheres	MnSO4·H2O, (NH4)2S2O8, glucose	Ostwald ripening	Water	180 °C, 3 h	None	1.5 μm	142	0.27	3–4
101	MoO2 hollow spheres	MoO3, diethylenetriamine	Ostwald ripening	Water	200 °C, 144 h	700 °C, 4 h in Ar	3–5 μm	NR	NR	NR
102	MoO2@MoO2 yolk–shell particles	MoO2(acac)2^d, HNO3	Ostwald ripening	Isopropanol, water	180 °C, 24 h	350 °C, 2 h in N2	1 μm, shell thickness 80 nm	31	NR	3–4
103	NiO hollow spheres	Ni(NO3)2·6H2O, NH3, L-cysteine	Ostwald ripening	Water	120 °C, 10 h	600 °C, 1 h in air	2–3 μm, shell thickness 400 nm	66	0.442	10–50
104	NiO multi-shelled spheres	Ni(NO3)3·6H2O, NH3, D-glucose	In situ carbon from glucose	Water	150 °C, 15 h	500 °C, 6 h in air	2–3.5 μm, shell thickness 50 nm	29	NR	NR
105	Core-in-double shell NiCo2O4 spheres	Ni–glycerate spheres prepared hydrothermally	Kirkendall mechanism	None	NA	350 °C, 2 h@1 °C min^−1	400 nm outer shell, 200 nm inner shell, 40 nm core, 70 nm and 40 nm outer and inner shell thickness	61	NR	<10
81	SnO2 hollow spheres	SnCl4·5H2O	Ostwald ripening	Water, DMF, TFA	180 °C, 48 h	None	2 μm	108	NR	6.04
106	SnO2 hollow spheres	SnSO4	Ostwald ripening	Water	120 °C, 48 h	None	100–200 nm	69	NR	4
107	SnO2 hollow spheres	SnF2, H2O2	Ostwald ripening	Water	180 °C, 12 h	None	100–200 nm, shell thickness 40–50 nm	156	NR	NR
108	SnO2 hollow spheres	K2SnO3·3H2O, urea	Ostwald ripening	Water, ethanol	150 °C, 24 h	None	150–400 nm	110	NR	4
109	SnO2 hollow spheres	SnCl2·2H2O, HCl, urea	Hollow polystyrene spheres	Mercaptoacetic acid	120 °C, 6 h	400 °C, 2 h in air	650 nm, shell thickness 100 nm	62	NR	3–8
73	Yolk–shell SnO2	SnCl2·2H2O, HCl	Carbon spheres	DMF, water	180 °C, 12 h	600 °C, 3 h in air	420 nm	43	0.073	6.8
110	SnO2 multishell spheres	SnCl4·5H2O, sucrose	In situ carbon from sucrose	Water	190 °C, 24 h	600 °C, 3 h in air	0.5–2 μm	36	0.197	2.50
111	SnO2/C hollow spheres	Sn spheres, glucose	Kirkendall mechanism	Water	180 °C, 3 h	500 °C, 3 h in N2	100 nm	NR	NR	NR
108	SnO2/C hollow spheres	SnO2 hollow spheres, glucose	In situ carbon from glucose	Water	180 °C, 3 h	550 °C, 3 h in N2	150–400 nm	NR	NR	NR
112	Perovskite SrTiO3 hollow spheres	Anatase TiO2, SrCl2·6H2O, NaOH	Kirkendall mechanism	Water	180 °C, 6 h	None	3–5 μm, shell thickness 700 nm	NR	NR	NR
113	SiO2–Ta2O5 hollow spheres^e	Tantalum isopropoxide, CTAB, TEOS, NH3	Ostwald ripening	Water, ethanol	120 °C, 48 h	550 °C, 5 h in air	100–250 nm, shell thickness 50 nm	249	0.48	14.8
	SiO2–Ta2O5 hollow spheres^f						200 nm, shell thickness 60 nm	225	0.26	13.5
73	Yolk–shell Tb4O7	Tb(NO3)3, NH4Ac·2H2O	Carbon spheres	Ethanol	180 °C, 12 h	600 °C, 3 h in air	200 nm	NR	NR	NR
114	V2O5 hollow spheres	NH4VO3	In situ gas bubbles	EG	180 °C, 24 h	500 °C, 2 h in air	3 μm, shell thickness 1.125 μm	22	NR	5–8
115	V2O5 hollow spheres	VO(C5H7O2)2, PVP	PVP micelles templated	EG	140 °C, 12 h	350 °C, 2 h in air	800 nm	NR	NR	NR
116	V2O5@V2O5 yolk–shell spheres^g	Vanadium oxytriisopropoxide	Carbon spheres	Isopropanol	200 °C, 12 h	350 °C, 2 h in air	1 μm, shell thickness 200 nm	NR	NR	NR
	V2O5@V2O5 yolk–shell spheres^h	Vanadium oxytriisopropoxide	Carbon spheres	Isopropanol, water	200 °C, 12 h	350 °C, 2 h in air	2 μm, shell thickness 100 nm	NR	NR	NR
117	V2O5 yolk–shell spheres	V2O5, oxalic acid	Ostwald ripening	Water, isopropanol	200 °C, 2.5 h	350 °C, 2 h in air	1 μm, shell thickness 100 nm	28	0.15	NR
118	ZnO hollow spheres	ZnCl2, glucose	In situ carbon from glucose	Water	180 °C, 24 h	500 °C, 4 h	0.8 μm	63	0.17	<5, 9–90
119	ZnO single shell hollow spheres	ZnSO4·7H2O, glucose	In situ carbon from glucose	Water	180 °C, 12 h	550 °C directly, 3 h in air	1 μm	10	0.04	NR
	ZnO double shell hollow spheres					550 °C, 5 °C min^−1, 3 h in air	1 μm	19	0.07	NR
	ZnO triple shell hollow spheres					550 °C, 2 °C min^−1, 3 h in air	1 μm	25	0.09	NR
62	ZrO2 hollow spheres	ZrOCl2·8H2O, HCl, urea	Ostwald ripening	Ethanol, water	160 °C, 24 h	None	1–2 μm	136	0.1	2–105
120	Yolk–shell ZnCo2O4	ZnAc2·2H2O, CoAc2·4H2O	Carbon spheres	EG	180 °C, 12 h	600 °C, 3 h in air	300–500 nm	16	0.063	20

---

Tables 3 and 4 provide a summary of metal oxide particles obtained by a solution precipitation method together with relevant references.^{13,16–20,120–159} However, the range of metal oxide particles that can be made by this method is not as extensive as in the case of hydrothermal synthesis, probably due to the lower solubility of metal oxide precursors at lower temperatures and the smaller flexibility at the temperatures used. Nevertheless, such mild conditions are attractive in terms of green technology and cost effectiveness.

Table 3 Porous solid spheres synthesized by the solution precipitation method

Ref.	Particle type	Reactants	Solvents	Reaction conditions	Calcination conditions	Particle size	BET SA (m^2 g^−1)	Pore volume (cm^3 g^−1)	Pore size (nm)
a Sonicated. b When the ratio of Al2(SO4)3·16H2O : Al(NO)3·9H2O is between 0.33 and 0.167. c To obtain V2O3 phase. d To obtain V2O5 phase. e Depending on water or pyridine concentration. f Mol ratio of Na2WO4·2H2O : concentrated HCl = 1 : 50. g Removed by NaOH etching. h By varying the Ce/HMT ratio. i Assumed room temperature. NR: not reported.
19	Ag2O–MnO2 spheres	MnSO4, (NH4)2S2O8, Ag nanoparticles	Water	50 °C, 1 h^a	None	2.2 μm	NR	NR	NR
121	α-Al2O3 spheres	Al2(SO4)3·16H2O, Al(NO3)3·9H2O, urea	Water	98 °C, 1.5 h	1100 °C, 1 h in air	125–430 nm^b	76^b	NR	2–20
	γ-Al2O3 spheres	Al2(SO4)3·16H2O, Al(NO3)3·9H2O, urea	Water	98 °C, 1.5 h	900 °C, 1 h in air	NR	102^b	NR	2–20
18	CuO spheres	Cu powder, NaOH, (NH4)2S2O8	Water	25^i °C, 20 h	None	1–2 μm	8	NR	NR
122	CuO spheres	Cu(NO3)2·H2O, NH3, NaOH	Water, glycol	100 °C, 2 h	300 °C, 4 h in air	1–3 μm	88	NR	NR
17	MnO2 spheres	(CH3COO)2Mn·4H2O, AgNO3, H2SO4, oxone monopersulfate	Water	25^i °C, 36 h	None	1–3 μm	163	NR	65
21	MnO2 spheres	MnSO4, (NH4)2S2O8, FeSO4	Water	50 °C, 1.5 h^a	None	700 nm	NR	NR	NR
20	α-MnO2 spheres	MnSO4·H2O, K2S2O8, K2SO4, H2SO4, AgNO3	Water	40 °C, 12 h	None	2 μm	150	NR	2, 10–20
				60 °C, 12 h	None	2 μm	106	NR	2, 10–20
				80 °C, 12 h	None	2 μm	83	NR	2, 10–20
123	Nb2O5 spheres	NbCl5, HNO3, resol, PEO-b-PS diblock copolymer	THF	50 °C, 24 h + 100 °C, 24 h	350 °C, 3 h and 550 °C, 2 h in N2 + 400 °C, 3 h in air	0.2–1 μm	131	0.26	11.4
124	NiO spheres	Ni(NO3)2·6H2O, NH3	Water	97 °C, 1 h	300 °C, 2 h in air	5 μm	216	0.38	64.3
125	SnO2 spheres	Na2SnO3·3H2O, D-glucose monohydrate	Water	50 °C, 12 h	None	50 nm	160	0.196	2.55
					150 °C in air	50 nm	146	NR	NR
					300 °C in air	50 nm	103	NR	NR
					500 °C in air	50 nm	75	NR	NR
126	SnO2 spheres	SnSO4	Water, ethanol	25^i °C, 1 h	500 °C, 2.5 h in air	100–800 nm	29	NR	4
127	V2O5 spheres	Vanadium isopropoxide	Acetone, pyridine, water	25^i °C, 0.5 h	400 °C, 2 h in H2^c + 300 °C, 1 h in air^d	150–1000 nm^e	31	NR	<30
128	V2O5 spheres	NH4VO3, HCl, hydrazine	Water	25^i °C, 0.5 h	350 °C, 2 h in air	400 nm	12	NR	<50
129	WO3 spheres	Na2WO4·2H2O, HCl^f	Water, EG	75 °C, 12 h	450 °C, 2 h in air	3–5 μm	13	NR	3.3–5.4
130	WO3 spheres	Na2WO4, HCl, oxalic acid	Water	25^i °C, 1 h^a	500 °C in air	1–3 μm	13	NR	28.1
131	WO3·H2O spheres	Na2WO4·2H2O, HCl	Water	70 °C, 10 h	400 °C, 2 h in air	2–3 μm	11	NR	1.7–30
132	Y2O3:Er spheres	Mesoporous SiO2 spheres^g, Y(NO3)3, Er(NO3)3, urea	Water	90 °C, 2 h	700 °C, 3 h in air	560 nm	85	0.196	5.7
133	ZnCo2O4@CeO2 core–shell spheres	ZnCo2O4, Ce(NO3)3, hexamethylenetetramine	Water, ethanol	60 °C, 2 h	None	1.55–1.68 μm^h	34–57^h	NR	NR
134	ZnO spheres	Zn(CH3COO)2, TEA	Water	25^i °C, 2 h^a	None	520 nm	17	NR	25, 180
135	ZnO spheres	Zn(CH3COO)2·2H2O, hexamine, sodium citrate	Water	90 °C, 6 h	600 °C in air	2.5 μm	NR	NR	NR
136	ZrO2 spheres	ZrOCl2·8H2O, porous polymer spheres	Water, ethanol	25^i °C, 0.17 h^a	600 °C, 6 h in air	2.6 μm	22	0.17	31

---

Table 4 Porous shells synthesized by the solution precipitation method

Ref.	Particle type	Reactants (mol/mass ratios)	Hollowing mechanism	Solvents	Reaction conditions	Calcination conditions	Particle and shell dimensions	BET SA (m^2 g^−1)	Pore volume (cm^3 g^−1)	Pore size (nm)
a Hollow-microporous organic network (H-MON). b By varying the solvent ratio or the SiO2 amount. c For H-MON prepared with the ratio of toluene : triethylamine = 1 : 1. d Micrococcus lylae. e Sonicated. f Mol ratio of Na2WO4·2H2O : concentrated HCl = 1 : 15. g Polyvinylalcohol@glucose derived carbon rich polysaccharide spheres. h Poly(styrene-acrylic acid). i When the molar ratio of ZrOCl2·8H2O : ethanol is between 0.011 and 0.032. j Assumed room temperature. NR: not reported.
137	Al2O3 hollow spheres	Al2O3 spheres, PVP, NaOH	NaOH etching/Kirkendall	Water	25^j °C, few minutes	400 °C in air	190 nm, shell thickness 23–30 nm	292	0.442	6.3
138	Al2O3, ZrO2, ZnO shell	Metal salt	Solid core	Buffer solution	70 °C, 2 h	450 °C, 2 h in air	Shell thickness tuneable 1–20 nm	NR	NR	NR
139	CdO hollow spheres	Cd(CH3COO)2, NaOH	Yeast	Water	25^j °C, 12 h	500 °C, 4 h in air	2.3 μm, shell thickness 250–280 nm	5	0.009	3–30
140	CeO2 hollow spheres	Ce(NO3)3·6H2O, HMT	PS spheres	Water	75 °C, 2 h	600 °C, 2 h in air	190 nm, shell thickness 15 nm	66	0.19	NR
141	Co3O4 hollow spheres	Co(NO3)2	Untreated carbon spheres	Water	25^j °C, 1 h	450 °C, 2 h in Ar + 450 °C in air	240 nm, shell thickness 40 nm	223	0.29	15.3
		Co(NO3)2	Acid treated carbon spheres				240 nm, shell thickness 15 nm	301	0.36	9.9
		Co(NO3)2	Alkali treated carbon spheres				240 nm, shell thickness 70 nm	174	0.2	23
142	Co3O4 hollow spheres	Co2(CO)8	H-MON^a spheres	Toluene	100 °C, 12 h	500 °C, 5 h in air	500 nm, shell thickness 20–80 nm^b	64^c	0.32^c	NR
14	Co3O4 hollow spheres	CoCl2·6H2O, NaBH4	Bacterial suspension^d	Water	25^j °C, 12 h	None	1 μm	149	0.26	7.7
143	CuO hollow spheres	CuSO4, KOH, NH3	Ostwald ripening	Water	68 °C, 24 h	None	3–5 μm, shell thickness 500 nm	NR	NR	1–2.2, 5–30
144	CuO hollow spheres	Cu(CH3COO)2·H2O, urea	In situ gas bubbles	Water	80 °C, 2 h^e	None	400–500 nm, shell thickness 45 nm	60	0.104	3.6
145	Gd2O3 hollow spheres	Gd(NO3)3, urea	Carbon spheres	Water, ethanol	90 °C, 6 h	800 °C, 2 h in air	200–250 nm, shell thickness 20 nm	33	0.17	10.9
146	In2O3 hollow spheres	InCl3	Polymer spheres	C2Cl4	55 °C, 6 h	600 °C, in air	720 nm, shell thickness 110 nm	329	NR	3
					75 °C, 6 h	600 °C, in air	950 nm, shell thickness 140 nm	28	NR	3
					95 °C, 6 h	600 °C, in air	1180 nm, shell thickness 220 nm	27	NR	3
147	MnO2 hollow spheres	MnSO4	CH2Cl2/H2O interface	Water, CH2Cl2	25^j °C, 48 h	300 °C, 2 h in air	200–500 nm	219	0.451	5.9
148	NiO hollow spheres	Ni(NO3)2·6H2O, urea	Sulfonated polystyrene hollow spheres	Water, ethanol	80 °C, 12 h	450 °C, 2 h in air	500 nm, shell thickness 100 nm	62	NR	2–4
149	NiO hollow spheres	NiCl2·6H2O, (NH4)2C2O4	Calcination of organic species	Water	25^j °C, 0.67 h^e	500 °C, 1 h in air	1.7 μm	32	NR	3–20
150	SnO2 hollow spheres	SnCl2, HCl	Ostwald ripening	Water	90 °C, 12 h	None	100–300 nm, shell thickness 10 nm	89	NR	NR
151	SnO2 hollow spheres	Tin butoxide	Microemulsion template	CTAB, hexanol, n-dodecane, methanol, water	20 °C, 12 h	None	15–25 nm, shell thickness 3–5 nm	417	NR	NR
152	SnO2 hollow spheres	SnCl2·2H2O	Hollow SiO2 spheres	None	80 °C, 24 h	700 °C, 2 h in air	340 nm, shell thickness 50 nm	46	NR	2–5
129	WO3 hollow spheres	Na2WO4·2H2O, HCl^f	Ostwald ripening	Water, EG	75 °C, 12 h	450 °C, 2 h in air	3–4 μm	16	NR	5.4–89.6
153	WO3 hollow spheres with multiple shells	WCl6	PVA@GCP^g	Ethanol	0 °C, 12 h	450 °C, 1 h in air	500 nm	124	0.14	4.3
154	WO3/WO3·H2O hollow spheres	Na2WO4·2H2O, HCl, oxalic acid	Ostwald ripening	Water, isopropyl alcohol	80 °C, 12 h	200 °C, 2 h in air	2 μm, shell thickness 200 nm	22	0.079	14.2
155	Y2O3 hollow spheres	Y(NO3)3, urea	Melamine formaldehyde spheres	Water	85 °C, 3 h	800 °C, 2 h in air	1.8 μm, shell thickness 100 nm	NR	NR	NR
156	Y2O3:Ln^3+ hollow spheres	Y(NO3)3, Eu(NO3)3, urea	PS spheres	Water	90 °C, 4 h	800 °C, 2 h in air	2.1 μm, shell thickness 70 nm	62	0.313	20.7
157	Y2O3:Tb^3+ hollow spheres	Y(NO3)3, Tb(NO3)3, urea	PS spheres	Water, ethyl alcohol	85 °C, 3 h	800 °C, 2 h in air	1.3 μm, shell thickness 50 nm	NR	NR	NR
158	ZnO hollow spheres	Zn(NO3)2·6H2O, (CH2)6N4, sodium citrate	Ostwald ripening	Water	95 °C, 5 h	400 °C, 2 h in air	2 μm	42	NR	5–8
159	ZnO hollow spheres	Zn(CH3COO)2, HMT, sodium citrate	Calcination of organic species	Water	95 °C, 3 h	400 °C, 2 h in air	2–3 μm	138	NR	NR
160	ZrO2 hollow spheres	ZrOCl2·8H2O, NH3 vapour	PSA^h spheres	Ethanol	50 °C, NR	700 °C, 4 h in air	3.2–3.4 μm, shell thickness 80–200 nm^i	NR	NR	NR

---

In general, metal salts are used for the synthesis of metal oxide spheres, often supplemented by complexing/structure directing agents and basic reactants in the case of precipitation method. A reducing or oxidizing agent can also be used. For instance, NaBH₄ was used to reduce Co²⁺ into Co nanoparticles, which subsequently were spontaneously oxidized in air onto bacterial templates to form Co₃O₄,¹⁴ ascorbic acid was employed to reduce Cu(OH)₂ to form composite Cu/Cu₂O spheres under microwave hydrothermal conditions,¹⁵ and N₂H₄·H₂O was used to reduce Cu²⁺ in solution to Cu₂O in the reduction-induced precipitation method,¹⁶ while oxidizing agents such as oxone monopersulfate¹⁷ or (NH₄)₂S₂O₈^18,19 were employed to increase the valence states of metal cations. Moreover in some cases, metal ions were used as catalysts. For the synthesis of MnO₂ spheres, Ag⁺ was used as a catalyst to help the reaction proceed at low temperature,^17,20 while in another work, Fe²⁺ was used to oxidize Mn²⁺ to MnO₂ crystals.²¹

The hydrothermal and solution precipitation syntheses are general methods that include hard templating, soft templating, hydrolysis, Ostwald ripening and the Kirkendall effect, which are briefly discussed in the following subsections.

2.1.1 Hard templating.
2.1.1.1 Shell structures. Hard templating is the most common method for production of hollow spheres. In hard templating, nanoparticles are attached to the surface of a solid sphere and they aggregate to form a shell. A decade ago, the commonly used hard templates for the creation of spherical shells were carbon,^161–163 silica¹⁶⁴ or polymer spheres.¹⁶⁵ As can be seen from Tables 2 and 4 these templates are still very popular, while bio-organisms have also been used. The aforementioned templates have a large number of reactive oxygen functional groups that are electron donors and therefore attract positively charged metal cations, as depicted in Fig. 3. The size of the inner hollow space can be tailored by selecting solid templates with appropriate sizes, while the thickness of the shell can be adjusted by changing the concentration of the reactants¹⁶⁰ or by changing the hydrophilicity of the template's surface by alkaline or acid treatment.¹⁴¹ Just recently, hollow TiO₂ nanospheres with a thin single layer of TiO₂ nanoparticles were synthesized by using quasi-nanosized carbonaceous spheres as a template.¹⁶⁶ In addition, hollow mesoporous organic networks (H-MONs), a new class of functional materials prepared through various carbon–carbon coupling reactions between organic building blocks, were prepared with varying shell thickness and used as templates to generate Co₃O₄ shells of different thicknesses and having surface areas between 60 and 67 m² g⁻¹.¹⁴²

Fig. 3 Hard templating method for synthesis of hollow metal oxide spheres.

However, the use of templates does not restrict the final shape to hollow spheres. Lou and co-workers¹¹⁶ found that the deposition of vanadium species on carbon sphere (CS) templates via a non-hydrolytic hydrothermal reaction between vanadium oxytriisopropoxide and isopropanol produced core–shell CS@V particles with shells made from interconnected nanosheets. Interestingly, when a small amount of water was added to the process, yolk–shell structures with rough surfaces were obtained. The VO₂ species formed by sol–gel reaction on the carbon spheres underwent an Ostwald ripening process to form a well-defined gap between the core and the shell. Yolk–shell particles of V₂O₅@V₂O₅ were also created from C@V₂O₅ core–shell particles¹¹⁶ whereby the vanadium species bound to the carbon core surface would shrink into a core during annealing. Similarly, yolk–shell ZnCo₂O₄,¹²⁰ SnO₂, CeO₂ and Tb₄O₇⁷³ were generated by hydrothermal loading of the metal precursor into the carbon template pores followed by calcination in air. TGA studies revealed that the removal of the template occurred in two steps to give the shell first and then the core. Zeng et al.¹⁶⁷ reported the synthesis of multi-shell ZnO with single, double or triple shells by simply using carbon templates of different diameters for loading of the ZnO precursor.

Elsewhere, a penetration–solidification–annealing method was used to synthesize multi-shell spheres of Co_xMn_3−xO₄ by using carbon spheres as templates and controlling the molar ratios of Co and Mn oxide precursors.¹⁶⁸ Based on the anion-adsorption mechanism and usage of carbon spheres as templates, Wang and co-workers developed a sequential templating approach for production of multi-shell hollow spheres with different composition including ZnO, TiO₂, SnO₂, Co₃O₄, α-Fe₂O₃, Mn₂O₃, V₂O₅, etc.^{166,169–176} For example, multi-shell V₂O₅ hollow spheres were synthesized by a novel method involving competitive anion adsorption on carbon sphere templates followed by a Trojan catalytic combustion procedure (Fig. 4).¹⁷⁶ The carbon spheres were pre-treated to create a negative charge, which could facilitate adsorption of metal anions from a solution of ammonium salt. The NH₄⁺ cations also penetrated the templates, neutralizing the negative charges and stimulating further anion adsorption. As a result, single- or multi-shell structures were formed either by using different precursor concentrations or by performing multiple adsorption processes. The method was shown to be flexible and could be extended to the synthesis of MnO₂, MoO₃, Cr₂O₃ and WO₃ multi-shell hollow spheres.

Fig. 4 Multi-shell metal oxides prepared via an anion-adsorption mechanism: (A) schematic representation of two synthesis routes to obtain multi-shell hollow microspheres. (i) Cation-adsorption process. (ii) Anion-adsorption process. (B) Effects of synthesis conditions on the morphology of products. (C) Morphological and structural analysis of V₂O₅ spheres: (a–f) TEM images of the as-prepared samples. (g–i) SEM images of the as-prepared samples. Reproduced with permission from ref. 176. Copyright © 2016, Nature Publishing Group.

However, hard templating has a few disadvantages. An additional synthesis step is required to remove the templates, which is by either calcination in air for carbon-rich templates or alkaline or HF etching for SiO₂ templates. The use and subsequent removal of the solid templates represents a waste in resources, which goes against “green” processing. Furthermore, calcination can lead to the partial or full collapse of the shell architectures, while alkaline etching can implicate formation of unwanted crystalline phases, such as sodium titanate (by reacting with TiO₂) or impurities,^152,177 whereas HF is a very toxic chemical to deal with. Nevertheless, a recent study led to the successful synthesis of hollow spheres with sandwich-type heterostructured shells via SiO₂ templating, whereby hydrothermal treatment resulted in crystallization of metal oxide and simultaneous etching of SiO₂ in the superhot water.¹⁷⁸

Carbon templates can also be produced in situ during hydrothermal synthesis by adding glucose, sucrose or other organics to the metal oxide precursors during the one-pot hydrothermal process. For instance, this method afforded carbon-supported amorphous and crystalline V₂O₃ microspheres consisting of assembled ultrathin nanosheets of ca. 10 nm thickness by using NH₄VO₃ and citric acid in a mixed water and ethanediol solvent, followed by calcination in a N₂ atmosphere.⁵⁰ The carbon was generated by carbonization of citric acid and ethanediol, which resulted in its uniform distribution in the resulting composite having 10.6% and 8.1% of carbon in the amorphous and crystalline V₂O₃ respectively. Hollow spheres and multiple shells of α-Fe₂O₃, Cr₂O₃, Co₃O₄, NiO and ZnO were also successfully prepared by hydrothermal heating of metal chlorides in a solution of fructose at a moderate temperature of 135 °C for 6 h, followed by calcination in air.¹⁷⁹ The carbonaceous core of partially dehydrated fructose contains functional groups such as –C=O and –OH that can attract positive metal ions forming a metal oxide–carbonaceous composite. In another work,¹⁸⁰ composite carbonaceous and Y₂O₃ spheres produced under hydrothermal conditions were used to generate hollow Y₂O₃ spheres with 1 to 4 shells by controlled calcination at different heating rates. Other multi-shell structures were successfully produced via this method such as ZnO,¹¹⁹ NiO,¹⁰⁴ SnO₂¹¹⁰ and CoFe₂O₄.⁷⁷ However, the metal oxide shells produced this way sometimes tend to aggregate and may be non-uniform in size.^75,110,181

Nevertheless, hard templates could be environmentally friendly through the use of biotemplates such as yeast¹³⁹ and bacteria.¹⁴ The utilization of bacteria (Micrococcus lylae) as a template to synthesize flower-like hierarchical Co₃O₄ hollow spheres with a uniform size of 1 μm, a high surface area of 149 m² g⁻¹ and a pore volume of 0.26 cm³ g⁻¹ was achieved via a one-pot reduction/oxidation reaction at room temperature.¹⁴ The bacterial surface is naturally covered with carboxyl, ester, amine and hydroxyl groups, which could readily capture cations. The bio-templating method is very attractive as it is facile, scalable and cost effective and could be extended to fabricate other materials and composites.

The choice of the template is important in determining the quality of the produced metal oxide nanoparticles. This has been shown when TiO₂ shells were prepared by coating a core with a TiO₂ layer followed by adding an outer protective layer around the TiO₂.¹⁸² The core and the protective layer were made of either SiO₂ or a resorcinol–formaldehyde resin-derived carbon, which was then removed by either etching or calcination in air. It was found that better crystallization occurred with the use of resorcinol–formaldehyde resin as a template due to enhanced flexibility and volume shrinkage of the resin during carbonization, while the growth of TiO₂ crystals was inhibited by impregnated silicate species.

2.1.1.2 Solid spheres. The solid templates have been mainly used in the synthesis of metal oxide spheres for the creation of large mesopores. In this case, the metal species were infiltrated into mesoporous templates, which upon template removal created a porous spherical structure. Some reported examples are porous ZrO₂ spheres obtained by using the EDA-functionalized poly(GMA-co-EGDMA) microspherical templates, Y₂O₃:Er mesoporous spheres fabricated by using mesoporous SiO₂ spheres as templates, and porous WO₃ spheres created by employing carbon microsphere templates.^53,132,136 In another work, Chen et al.⁷⁴ prepared multi-core–shell Pd@SiO₂@mesoporous–CeO₂ by assembling CeO₂ nanoparticles around multiple Pd@SiO₂ cores via a hydrothermal method. Once the silica was removed with concentrated NaOH, a multi-yolk–shell Pd@mesoporous–CeO₂ structure was obtained with a size of 150–200 nm and a BET surface area of ∼103 m² g⁻¹.

2.1.2 Soft templating.
2.1.2.1 Shell structures. Soft templating is also a commonly used method for production of hollow spheres and yolk–shell spheres. In this method the gas–liquid, liquid–liquid or dissolved long chained organic micelles are used as soft templates for the formation and crystallization of metal oxide nanoparticles. Gas–liquid interfaces are in situ formed around gas bubbles that are generated during chemical reactions; for example, CO₂ bubbles are formed from urea, NH₄HCO₃¹⁸³ and ferric potassium oxalate⁸² during decomposition at high temperatures or N₂ bubbles are formed from the reaction of NH₄VO₃ with EG.¹¹⁴ Nevertheless, the hollowing of structures on gas bubbles is difficult to prove and it could be more likely that Ostwald ripening is responsible for the generation of hollow spheres.

A liquid–liquid interface occurs between immiscible liquids such as water and oil. Hierarchical mesoporous MnO₂ spheres were formed at the droplet interface between CH₂Cl₂ and H₂O.¹⁴⁷ Similarly, SnO₂ hollow spheres were prepared in a water-in-oil microemulsion of CTAB–hexanol-n-dodecane–methanol–water.¹⁵¹ The calcined SnO₂ particles had a surface area of 119 m² g⁻¹. In another work, a mixture of TBOT and PS in toluene was emulsified in formamide containing the triblock polymer Pluronic P123.¹⁸⁴ After removal of toluene and PS spheres via heat treatment, macroporous TiO₂ spheres of sizes between 500 nm and 2000 nm were obtained. Finally, a glycerol, water and Fe oxide precursor mixture was hydrothermally heated at 145 °C to obtain mesoporous α-Fe₂O₃via a quasi-emulsion templating mechanism. The product had a surface area of 103 m² g⁻¹ when using a glycerol/water volume ratio of 1 : 7;⁸³ however a different phase of α-FeOOH⁹⁰ was produced at 120 °C.

A long-chained polymer such as PVP K30 (M_w ∼ 40 000) can form micelles in solution with a hydrophobic core and a hydrophilic shell. These micellar structures were used as soft templates to attach vanadium species and upon calcination to form uniform V₂O₅ hollow spheres with sizes of ∼800 nm¹¹⁵ and with exposed [110] facets, as shown in Fig. 5. During the formation of hollow spheres, VO(acac)₂ accumulated on the hydrophilic PVP micelles in the EG solvent to form VO(acac)₂@PVP core–shell particles. The C₅H₈O₂ ligand subsequently was replaced by EG to form VEG, which oligomerized via the LaMer process into a tight layer on the surface of micelles. Upon calcination, V₂O₅ hollow spheres were formed. On the other hand, α-Fe₂O₃ with multi-shell morphology of ∼3 μm in size and with a surface area and pore volume of 14 m² g⁻¹ and 0.07 cm³ g⁻¹, respectively, was obtained with the L-histidine amino acid as a template.⁸⁷ The Fe(OH)₃–L-histidine–H⁺–NO₃– complex obtained under hydrothermal conditions underwent size shrinkage and phase separation during the calcination stage to form porous multi-shell particles. In another work,⁹⁷ a double templating method was employed by decorating soft micelles of the F127 copolymer onto SiO₂ spheres to obtain MnO₂ hollow particles following hydrothermal treatment of the template with KMnO₄. The morphology and surface area could be varied from urchin-like (233 m² g⁻¹) to flower-like (201 m² g⁻¹) and non-hierarchical (120 m² g⁻¹) by varying the F127/SiO₂ mass ratio to 0.2, 0.4 and 0.6 respectively.

Fig. 5 Hierarchical orthorhombic V₂O₅ hollow nanospheres prepared via soft-templating: (a) schematic illustration of the evolution of V₂O₅ hollow nanospheres. (b) FESEM and (c) TEM images of V₂O₅ nanospheres; (d) atomic resolution HRTEM image, from which the interlayer structure of V₂O₅ was directly observed. Reproduced with permission from ref. 115. Copyright © 2014, Royal Society of Chemistry.

2.1.2.2 Solid spheres. In the synthesis of metal oxide spheres, long chained organics are mostly used as soft templates for surface stabilization of building blocks. These organics accumulate on the nanocrystallites favoring their growth in certain directions or planes, which affects the spatial orientation of crystal nanoparticles and growth of hierarchical structures. Eventually, the removal of these organics by calcination frees hidden pores in the assembled metal oxide structures. Some examples of the organic compounds used as structure directing agents are amino acids such as L-asparagine⁵⁶ and glycine,⁴¹ sugars such as D-glucose monohydrate,¹²⁵ carbohydrates such as starch¹⁸⁵ and polymers such as sodium alginate,^33,45 polyethyleneimine² and also PVP, which has been commonly employed.^{22–24,47,48} To avoid unnecessary use of polymers and surfactants, alcohols and carboxylic acids have been proved to be very competent capping/structure directing agents by chelating with metal ions. Moreover when mixed together, they react at elevated temperatures to produce esters, which can further influence the growth rate and self-assembly process of nanocrystallites.^26,69

Some researchers have used micelles to create mesoporous structures. Luo et al.¹²³ examined a resol-assisted solvent evaporation method in the presence of THF, PEO-b-PS block copolymer and NbCl₅. Upon evaporation of THF, the block copolymer aggregated into cylindrical micelles covered by the resol/Nb⁵⁺ composite. Upon further evaporation, the micelles tended to bend and aggregate into spherical particles, while resol was acting as a binder. Subsequent pyrolysis and calcination produced Nb₂O₅ spheres of diameter 0.2–1 μm having uniform mesopores with an average size of 11.4 nm, a high surface area of 131 m² g⁻¹, and a pore volume of 0.26 cm³ g⁻¹. This solvent evaporation-driven self-assembly was also recently used to create mesoporous TiO₂ microspheres with [101] exposed facets from spherical composite micelles consisting of PEO–PPO–PEO and titania oligomers.¹⁸⁶ Similarly, surfactants such as Pluronic P123, due to its long hydrophobic chains, can be used to create large mesopores between 12 and 15 nm in the WO₃/TiO₂ composite spheres.⁵⁴ Elsewhere, Wang's group¹⁸⁷ used a water in oil system in the presence of acrylamide and azobisisobutyronitrile to synthesize hierarchically mesoporous hematite microspheres with high surface area and bimodal structure with mesopores of 2.5 nm and 9 nm.

2.1.3 Sol–gel and controlled hydrolysis. The sol–gel process is a viable method for the synthesis of metal oxides from alkoxides such as butoxides, ethoxides and isopropoxides. Alkoxides vigorously react with water molecules in the presence of an excess of alcohol and ammonia as catalyst to form an inorganic network or gel of a polymer-like metal oxide that condenses into particles shown in Fig. 6. Titanium alkoxides such as titanium butoxide and titanium isopropoxide and halides such as TiCl₄ are very well known and commonly used precursors for the synthesis of TiO₂ spheres. Because of the nature of these TiO₂ precursors, the reaction can be controlled by adding a small amount of water. Some of the techniques used to control the amount of water are discussed in the previous section.

Fig. 6 Schematic illustration of the formation of metal oxide spheres by a sol–gel process.

Controlled hydrolysis has been used to prepare TiO₂ spheres,^188–196 core–shell,¹⁹⁷ yolk–shell,^198,199 and hollow^{177,183,200–204} particles. A more detailed discussion on the synthesis of TiO₂ has been presented elsewhere; the reader is encouraged to refer to a comprehensive review on the synthesis of spherical TiO₂ nanostructures by Chen et al.⁵

Nevertheless, other types of metal alkoxides have also been used to produce the respective oxides, such as vanadium oxiisopropoxide,^51,116 vanadium(V) oxytriisopropoxide¹²⁷ and tin tert-butoxide.¹⁵¹ As shown in Fig. 7, V₂O₅ mesoporous spheres were synthesized at room temperature by reacting vanadium isopropoxide in a mixture of acetone, pyridine and water at a volume ratio of 983 : 500 : 1.¹²⁷ The average size of the particles was tuned between ∼1 μm and ∼150 nm by increasing the amount of water while maintaining the ratio of pyridine/acetone. The reduction of particle size with increasing water content was attributed to the increased number of sites for nucleation of particles.

Fig. 7 SEM image of the V₂O₅ porous microspheres; the inset shows the porous structure of a single sphere. Reproduced with permission from ref. 127. Copyright © 2011, Royal Society of Chemistry.

Silica supported Ta₂O₅ (SiO₂–Ta₂O₅) composite shells were produced by sol–gel synthesis using TEOS, tantalum isopropoxide, CTAB, H₂O, NH₃ and ethanol.¹¹³ NH₃ catalyzed the reaction but also assisted in the dissolution of cores at higher temperature. The diameter and shell thickness were tuned by changing the molar ratio of Si : Ta. The BET surface area of the calcined particles increased from 225 to 610 m² g⁻¹ with increasing Ta content.

The sol–gel method can be also used to manufacture templates for the synthesis of hollow and porous spheres. The one-pot sol–gel polymerization of formamide–resorcinol was employed to create vesicle templates for the synthesis of hollow In₂O₃ spheres,²⁰⁵ while the porosity in SnO₂ spheres was created upon removal of the carbon template from composite Sn–resorcinol–formaldehyde resin particles.²⁰⁶

2.1.4 Ostwald ripening. According to the IUPAC's terminology, the process of Ostwald ripening refers to “the growth of larger crystals from those of smaller size which have higher solubility than the larger ones”.²⁰⁷ As can be seen from Tables 2 and 4 the Ostwald ripening represents the major process of hollowing particles under hydrothermal conditions. Once a metal oxide amorphous precursor is formed, crystallization proceeds with increasing time; namely, smaller nanocrystals from the central core dissolve and recrystallize into loosely packed crystallites at the surface of spheres that act as seeds. The resulting void enlarges until a hollow structure is obtained, as shown in Fig. 8a. For example, flower-like hollow MgO spheres having a size of about 1 μm, cavities of ∼500 nm, a high surface area of 343 m² g⁻¹ and a pore volume of 1.9 cm³ g⁻¹ were fabricated by ripening magnesium glycolate under hydrothermal conditions.⁹⁶

Fig. 8 Ostwald ripening initiated by dissolution of the middle core (panel a), and localized Ostwald ripening (panel b).

In some cases, due to the localized Ostwald ripening around a dense core, an intermediate yolk–shell architecture is formed (Fig. 8b), as in the case of MoO₂@MoO₂,¹⁰² CeO₂@CeO₂,⁷⁰ SnO₂@SnO₂,¹⁰⁶ TiO₂@TiO₂²⁰⁸ and V₂O₅@V₂O₅,¹¹⁷ which subsequently is converted to a multi-shell structure having 2–3 shells after a prolonged hydrothermal process, and finally to hollow spheres. Similarly, double-shell CoO and Co₃O₄,²⁰⁹ double-shell α-Fe₂O₃,⁸⁴ as well as perovskite-type BaZrO₃@BaZrO₃⁶⁴ and LaFeO₃@LaFeO₃⁹¹ structures were obtained. In a different work, Li et al.⁸⁶ found that the two interfaces created by hydrothermal etching of silica from Fe₃O₄@SiO₂@TiO₂ in 1 M NaOH at 150 °C for 24 h allowed for the dissolution of TiO₂ crystals and their subsequent growth in an opposite direction to eventually form Fe₃O₄@TiO₂ double-shell spheres with flower-like morphology and with a uniform size of ∼560 nm and a high surface area of 150 m² g⁻¹. Ostwald ripening involving selective etching of crystals is favored in basic and acidic solutions and it is apparent from Tables 2 and 4 that bases such as urea, KOH, NaOH and amines or acids such as HCl, HNO₃ and other organic acids facilitate this process.

2.1.5 Kirkendall effect. The Kirkendall effect is an atomic diffusion phenomenon that occurs at elevated temperatures between two species separated by an interface (solid–solid, solid–liquid or solid–gas) and having different diffusion rates. The net directional flow of matter is balanced by an opposite flux of vacancies, which can result in the formation of voids, normally at the interface because of its high energy and defect density. Consequently, the Kirkendall effect has been utilized for the preparation of hollow nanocrystals,²¹⁰ yolk–shell particles²¹¹ as well as metal oxide hollow spheres.^{111,112,212,213}Fig. 9 shows the formation of a shell by the exchange of ions between the solid core and its surrounding.

Fig. 9 Hollowing by the Kirkendall effect.

The thermal treatment in air of a film decorated with the Cu(II) complex showed that the bulk diffusion of atoms/ions at the Cu–O interface gave Cu₂O-rich and CuO-poor spheres at 200 °C.²¹² As the temperature increases, the Cu from the core moves outwards through the oxide shell to react with oxygen, leaving a hollow space (since the outward diffusion of Cu ions is much faster than the inward diffusion of O ions) until a hollow sphere of pure CuO is formed at 400 °C. Core-in-double-shell hollow NiCo₂O₄ spheres were obtained by slow annealing of NiCo–glycerate spheres in air due to a combination of the Kirkendall effect and the contraction and adhesion forces during the oxidative degradation of organic species.¹⁰⁵ This method can also be extended to the synthesis of ZnCo₂O₄ and CoMn₂O₄ with complex interior structures.

The Kirkendall effect can also occur during the hydrothermal reaction stage. The reaction of TiO₂ microspheres with a solution of strontium chloride hexahydrate at 180 °C for 6 h generated perovskite-type SrTiO₃ hollow spheres having a size of 3–5 μm and a shell thickness of ∼700 nm.¹¹² With the assistance of NaOH, the Ti–O–Ti bonds can be broken to form Ti–O–Na on the surface of the sphere. Then, the Sr²⁺ ions can react with the sodium titanate to form a thin layer of SrTiO₃ shell, separating the inner TiO₃²⁻ ions from the Sr²⁺ ions in solution. Hence, the concentration gradient between these two types of ions permitted TiO₃²⁻ to diffuse out and the Sr²⁺ ions to diffuse in through the shell, resulting in hollow SrTiO₃ spheres. Additionally, composite SnO₂–C hollow spheres were prepared by Wu et al.¹¹¹ under hydrothermal conditions by reacting Sn spheres in a solution of glucose at 180 °C.

Interestingly, a simple solution route by mixing hydrothermal carbon spheres in a solution of KMnO₄ at room temperature produced MnO₂ spheres of different morphologies.²¹⁴ Solid MnO₂ spheres were produced with 100 ml of 25 g L⁻¹ KMnO₄, C@MnO₂ yolk–shell spheres were obtained with 100 ml of 2.5 g L⁻¹ KMnO₄ and finally MnO₂ hollow shell spheres were obtained with 200 ml of 2.5 g L⁻¹ KMnO₄ solution. The formation of different morphologies was achieved due to different stages of the Kirkendall effect occurring through the soft surface of the hydrothermally synthesized carbon spheres by varying MnO₄⁻ concentration.

The Kirkendall mechanism provides a pathway for the selective etching of the surface-protected metal oxides to produce hollow structures. For instance, the PVP-protected TiO₂ solid spheres were selectively etched by fluoride ions to form hollow or yolk–shell TiO₂.²¹⁵ Similarly, it was also reported that NaOH and HCl were used to etch the PVP-protected colloidal Al₂O₃ and ZnO into hollow spheres, respectively.¹³⁷

It is apparent that during the Kirkendall mechanism, a solid core acts as a sacrificial template by reacting with its surrounding environment to form different hollow structures. This method could therefore be extremely useful to synthesize a variety of complex hollow compounds and composites from various solid templates.

2.2 Spray method

Spray methods include electrospray ionization^216–219 and gas phase processes such as aerosol^220–225 and flame or ultrasonic spray pyrolysis (USP).^226–229 These methods use high temperatures to evaporate the liquid from the colloidal precursor solution released by the spray nozzle to form solid or hollow spherical structures. Several types of metal oxides such as CeO₂ spheres,²¹⁶ TiO₂ spheres,²¹⁷ WO₃ spheres,²²⁸ ZnO spheres,^221,225 α-Fe₂O₃ microspheres,²³⁰ CuO hollow spheres,²²² Mn₃O₄ hollow spheres,²²⁴ WO₃ hollow spheres,²²⁶ hollow TiO₂ and ZrO₂ spheres,²²³ TiO₂ core–shell particles,²¹⁸ Bi₂WO₆ spheres,²²⁹ Fe₃O₄–carbon composite spheres,²²⁰ Li₂O–CuO–SnO₂ multi-deck cage-type spherical composites²¹⁹ and α-Fe₂O₃ multi-shell hollow spheres²³¹ have been produced via the spray method. Some morphologies including porous spheres,²²¹ hollow spheres,^222,223 yolk–shell spheres,²³² yolk–multi-shell spheres,²³³ and “ant-cave” spherical structure²³⁴ are shown in Fig. 10.

Fig. 10 Different types of metal oxide structures prepared by the aerosol method: (a) porous sphere (reproduced with permission from ref. 221; Copyright © 2014, Royal Chemical Society); (b) hollow sphere loaded with nanometals (reproduced with permission from ref. 223. Copyright © 2013, Wiley); (c) hollow shell (reproduced with permission from ref. 222. Copyright © 2013, Wiley); (d) yolk–shell sphere (reproduced with permission from ref. 232. Copyright © 2013, Wiley); (e) yolk–multi-shell sphere (reproduced with permission from ref. 233. Copyright © 2013, Wiley); and (f) “ant-cave” spherical structure (reproduced with permission from ref. 234. Copyright © 2013, American Chemical Society).

In the electrospray method, the liquid is evaporated via the potential difference between the nozzle and the metal receptor, while in the gas phase processes, the colloids pass through a flame or horizontal furnace (Fig. 11). Kang and co-workers have published numerous works on spray pyrolysis for the synthesis of hollow spheres, yolk–shell particles, multi-shell spheres and porous microspheres. Hollow WO₃ spheres with thin and porous shells were produced by USP using citric acid as the carbon source.²²⁶ Multi-shell structures were created from the precursor dissolved in sucrose solution. During decomposition, a dense carbon–metal oxide composite was formed, which upon further heating resulted in contraction and combustion of the carbon to form the multi-shell structures. This method was used to prepare yolk–shell TiO₂ and composite multi-component systems (composed of up to 5 components including TiO₂, Al₂O₃, ZrO₂, CeO₂ and Y₂O₃),²³² double-shell LiNi_0.5Mn_1.5O₄ particles,²³⁵ Pd loaded double-shell SnO₂ particles²³³ and double-shell SnO₂ spheres.²³⁴ Alternatively, the spray-pyrolysis method has been extended to the synthesis of yolk–shell structured metal oxide with 10 kinds of metal components in one step as shown in Fig. 11b.²³⁶ The method could also be modified to produce metal sulfide multi-shell spheres. SnO₂ yolk–double shell spheres were indeed treated with H₂S gas to produce SnS yolk–double shell spheres.²³⁷ Another work reported the synthesis of a new structured material named “ant-cave microball”, where polystyrene nanobeads were used as templates to create MoO₃–C composite spheres.²³⁴ The decomposition of these nanobeads resulted in unique morphology of porous composite spheres with nanochannels, effectively resembling an ant-cave.

Fig. 11 Spray pyrolysis for the formation of yolk–shell-structured LiNi_0.5Mn_1.5O₄ spheres (panel a; reproduced with permission from ref. 235. Copyright © 2013, Royal Chemical Society) and yolk–shell ten-component transition metal oxide powder (panel b; reproduced with permission from ref. 236. Copyright © 2014, Royal Chemical Society).

ZnO spheres were synthesized via an aerosol method using an organometallic precursor dissolved in toluene and Brij 58 as the structure directing agent.²²¹ The ZnO spheres had a BET surface area, crystal size and maximum pore volume of 61 m² g⁻¹, 8.6 nm and 13 nm, respectively. Al and S could be easily incorporated into the ZnO matrix by adding similar organometals into the precursor solution. Unfortunately, the dopants reduced the crystal size and hence the maximum pore size of the resulting ZnO spheres but this could be counteracted by using the triblock copolymer P123. Recently, very high surface area α-Fe₂O₃ microspheres with an average size of 560 nm, a BET surface area of 301 m² g⁻¹ and an average pore size of 2.1 nm were synthesized by USP using Fe(NO₃)₃ and Na₂CO₃ as precursors.²³⁰ The average particle size could be tuned by changing the concentration of the precursors.

Hollow or macroporous structures could be synthesized by using hard templates,^223,228 furnace synthesis at elevated temperatures,²²⁵in situ bubble reactions^222,224 or non-equilibrium air calcination.²³¹ Au nanorods, Pd nanocubes and Au core/Pd shell nanorods were successfully introduced into hollow TiO₂ and ZrO₂ spheres by initially embedding these nanometals in PS nanospheres. The PS spheres were dispersed in solution containing metal alkoxides and then sprayed by using N₂ through a tube furnace. Subsequently, calcination was performed to remove the PS template, leaving hollow TiO₂ and ZrO₂ spheres of average diameters of 0.8 μm and 0.6 μm, respectively, and containing nanometals inside the hollow space. In another work,²²⁵ ZnO spheres of various shapes were produced simply by changing the furnace temperature. Amorphous porous spheres were obtained between 40 °C and 100 °C, solid spheres at 400 °C, yolk–shell spheres at 600 °C and hollow spheres with different crystallite sizes between 700 °C and 1200 °C. The hollowing process was induced by the Kirkendall effect. An interesting in situ bubble hollowing method was devised by Jian et al.²²² to prepare hollow CuO spheres by adding sucrose and H₂O₂ to Cu(NO₃)₂ solution. The decomposition of the sucrose into CO₂ and H₂O (with H₂O₂ acting as a catalyst) within the aerosol at high temperature inflated the spheres like balloons to produce particles with an average size of ∼85 nm and very thin walls of 5–10 nm. The same strategy was used to prepare hollow Mn₃O₄ spheres.²²⁴ α-Fe₂O₃ multi-shell hollow spheres²³¹ were synthesized by spray drying a mixture of Fe(III) citrate and sucrose. The obtained Fe(III)–sucrose composite was then calcined in air to remove the carbon template. From the effect of non-equilibrium heating in air, the number of shells could be varied between 2 and 4 by simply changing the Fe(III) citrate/glucose ratio between 0.25 and 1.5.

The spray method is a simple and continuous process with a short residence time (a few seconds) of particles at a high temperature, which produces high purity products and can be easily implemented on an industrial scale. Moreover, other constituents can be included in the precursor solution allowing the preparation of composite or doped metal oxide particles. However, due to the low residence time of the particles at high temperature, further annealing may be required to improve the crystallinity of the products. Furthermore, the method has not yet been able to create hierarchical structures and the surface areas of the particles are often in the low to moderate range.

2.3 Other methods

Besides the methods presented above, there are a variety of other routes for the preparation of colloidal metal oxide structures such as template- and solvent-free methods, ultrasonic irradiation- and microwave-assisted syntheses, electrodeposition, direct printing, and methods involving lasers or taking advantage of gas–liquid diffusion.

A template- and solvent-free method was devised by Wang et al.²³⁸ for the synthesis of hierarchical metal oxide spheres (HMOS) of TiO₂, Fe₂O₃, ZrO₂ and their composites. This method involves grind milling of the metal oxides in the presence of PEG and some water to create a paste, which was spread into a film and annealed. The process generated microspheres via PEG modification and self-assembly and has great potential for large scale production of HMOS.

During ultrasonic irradiation, the formation and collapse of bubbles in the aqueous phase results in localized extremely high temperatures (>5000 K), high pressures (>20 MPa) and very high cooling rates (10¹⁰ K s⁻¹), which can supply enough energy to drive the formation of spherical metal oxide structures.¹⁴⁴ Various metal oxide spherical particles such as ZnO hollow nanospheres of size ∼80 nm,²³⁹ mesoporous NiO hollow spheres,¹⁴⁹ MnO₂ spheres made of interconnected nanoflakes,²¹ ZnO spheres with bimodal pores at 25 nm and 180 nm,¹³⁴ CuO hollow spheres,¹⁴⁴ composite Ag₂O–MnO₂ spheres with Ag₂O residing at the end of MnO₂ nanowires,¹⁹ and WO₃ spheres¹³⁰ were synthesized that way.

The electrodeposition method was used to prepare uniformly distributed 100–500 nm sized MnO₂ spheres with a very high surface area of 129 m² g⁻¹ and mesopores in the range of 5–12 nm, which were composed of randomly oriented nanorod-like structures.²⁴⁰ This synthesis is inexpensive, operates at room temperature and the deposition potential is an extra parameter that can be varied to achieve different morphologies. However, the synthesized oxides have low crystallinity and require further annealing. Other examples of particles prepared via electrodeposition are Y(OH)₃ and Y₂O₃ nanospheres,²⁴¹ Co₃O₄ hollow spheres deposited on PS spheres and organized into a close-packed monolayer array²⁴² and SnO₂ spheres.²⁴³

The microwave-assisted synthesis is analogous to the hydrothermal method but offers a much faster heating rate of the solution. An enormous advantage of this method is the very short time which is reduced to minutes instead of hours (as shown in Tables 1 and 2) as compared to the hydrothermal method. Some examples of metal oxides prepared by this method are Fe₃O₄ and γ-Fe₂O₃ hollow spheres,⁸⁵ NiO spheres⁴¹ and hollow spheres²⁴⁴ with very high surface areas reaching 200 m² g⁻¹, TiO₂ spheres^195,245 and Bi₂O₃ spheres.²²

Laser irradiation is another powerful and versatile way to obtain CuO²⁴⁶ and ZrO₂ spheres.²⁴⁷ This method was even used to make hollow spheres of metals and semiconductors such as Fe, Co, Ni, TiO₂, Co₃O₄, NiO, WO₃ and Fe₂O₃.²⁴⁸ The method offers control over the size of particles and high crystallinity, and therefore no annealing step is required. The high energy dispersed during laser heating was successful in producing single-crystalline rutile TiO₂ at room temperature with an average size of 540 nm²⁴⁸ from commercial anatase TiO₂ nanoparticles dispersed in acetone. The hollowing was attributed to the Kirkendall effect. The size of spheres could be tuned by controlling the laser beam and irradiation time; however the size of the void space could not be controlled. In another light-driven approach, UV irradiation was used to decompose titanium glycolate spheres into highly uniform mesoporous TiO₂ spheres with amorphous structure.²⁴⁹

Some other methods include thermal decomposition of various precursors such as Ni(CH₃COO)₂·4H₂O at 500 °C for 10 h to produce NiO mesoporous spheres,²⁵⁰ synthesis of hollow CuO spheres²¹² involving direct printing of metal–ion complex ink on a substrate followed by thermal heating, and a gas–liquid diffusion method to prepare Co₃O₄ hollow spheres by controlled precipitation from a solution of Co(NO₃)₂ in the presence of a vapor from crushed ammonium carbonate.²⁵¹

2.4 Methods for controlling the precipitation process

To prevent the uncontrolled precipitation of the metal oxide into large aggregates, the release of reactants in solution can be regulated. The commonly used procedures to control the precipitation process include metal complexation, pH control and controlled release of other reactants. Very low concentrations of metal precursors (usually in mmol amounts) are used as the first preventive measure.

2.4.1 Complexation. Metal ions present in solutions have a high positive charge to radius ratio, which makes them highly electronegative, i.e., they have a high affinity for electrons. As a result, the metal cations tend to attract polar molecules (or ligands) in solution by chelation to form stable dissolved metal complexes. Examples of ligands that form metal complexes are water, NH₃,¹²² simple organic compounds such as organic acids (e.g. citric acid,³⁷ ascorbic acid³⁴ and tartaric acid³⁷), alcohols (methanol⁴⁸ and EG^82,88), amines,^59,252 acetylacetonates,³⁰ acetates,^2,32,42 citrates^38,58 and oxalates.³¹ The metal complexes prevent the uncontrolled precipitation of the metal oxides into aggregates by slowly combining with hydroxide ions via multiple steps, as shown in eqn (1) and (2) for a hexa-aqua 2+ complex, until a neutral complex is precipitated out of the solution.

[M(H₂O)₆]²⁺ + OH⁻ → [M(H₂O)₅(OH)]⁺ + H₂O (1)

[M(H₂O)₅(OH)]⁺ + OH⁻ → [M(H₂O)₄(OH)₂] + H₂O (2)

Cations with high positive charge such as Ce³⁺,²⁶ Fe³⁺,³⁵ In³⁺,²⁵³ Sn^4+ 47,48 and W^6+ 52,53 can precipitate from the solution via forced hydrolysis at elevated temperatures without the need of a basic additive. The soluble hydroxyl complexes created by hydrolysis reaction (as per eqn (3)) form precursors for the nucleation of crystallites.

[M(H₂O)_n]^z+ ⇌ [M(H₂O)_n−y(OH)_y]^(z−y)+ + yH⁺ (3)

Ligands behave like surfactants or capping agents, since they adhere to preferred planes of the grains/crystallites, which then prevent their growth in specific directions. This gives rise to hierarchical spherical structures, built from differently shaped building blocks, such as nanocubes, nanosheets, nanopetals, nanowires, etc.

2.4.2 pH control. While the availability of metal ions is controlled via chelation, there are other methods to limit the availability of OH⁻ ions for precipitation. A slow release of OH⁻ ions for controlled precipitation of metal hydroxide complexes can be realized in solution via degradation of amides (such as urea) and amines (such as HMT, hydrazine and ethanolamine) at moderate temperatures (above 80 °C). Tables 1 and 4 show that this method is popular for the synthesis of solid and hollow spheres at moderate temperatures.

To successfully coat a homogeneous and very thin layer of metal oxide on a substrate can be a very challenging task. This subject has been studied by Zhang et al.,¹³⁸ who reported a simple but interesting method of controlling the amount of OH⁻ ions available for precipitation of metal hydroxides, which could be converted to metal oxides upon calcination. For controlled precipitation, the ionic product of the metal hydroxide precursor (K_mp) has to be equal to or slightly higher than its solubility constant (K_sp). For a metal hydroxide M(OH)_n, K_mp can be related to pH as expressed by eqn (4) and (5):

K_mp = [Mⁿ⁺_aq][OH_aq⁻]ⁿ (4)

K_mp = [Mⁿ⁺_aq][K_w × 10^pH]ⁿ (5)

Hence for precipitation,

K_mp ≥ K_sp (6)

where n is the valence number of the metal ion and K_w is the water dissociation constant, which has the value of 1.00 × 10⁻¹⁴ mol² L⁻². A buffer solution within the required pH range can then be used to provide a constant pH during reaction for controlled precipitation. Consequently, for Al₂(SO₄)₃ in a formic acid/ammonium formate buffer solution within the pH range of 3.8 and 4.6, a nanolayer of Al₂O₃ could be coated on different kinds of substrates such as Au, Si, SiO₂ and PS and different shapes could be created such as core–shell structures of multi-walled carbon nanotubes. The thickness of the shell could be controlled between 1 and 20 nm by adjusting the concentration of Al₂(SO₄)₃ salt. This method can be easily extended to other metal oxide coatings.

2.4.3 Controlled hydrolysis. The reaction of alkoxides with water during the sol–gel process is extremely fast. A few protocols are adopted to avoid uncontrolled precipitation of particles. Commonly, very small amounts of water are used in high purity alcoholic solvents (purity very close to 100%). Also, alkoxide or hydrated precursors are added very slowly, e.g., by vapor deposition¹⁹³ or in a dropwise fashion.^177,199,201 Gyger et al.,¹⁵¹ on the other hand, minimized the contact of Sn(Ot-Bu)₄ with water by slowly adding a mixture of Sn(Ot-Bu)₄ in dodecane to a water-in-oil microemulsion system. The hydrolysis of tin tert-butoxide occurred slowly on water droplets at the organic–water interface, eventually leading to the formation of SnO₂ hollow spheres.

The deposition of a very thin layer of TiO₂via the modified Stöber sol–gel method is normally difficult. However, it was successfully achieved by controlling the hydrolysis rate of the TiO₂ precursor by varying the amount of ammonia, which has a crucial role in controlling the reaction kinetics for the formation of TiO₂ shells.²⁰⁴ Accordingly, the thickness of the TiO₂ shell could be changed from 25 nm to 70 nm by varying the ammonia concentration from 0.25 vol% to 0.4 vol%.

Some researchers used the in situ generated water during chemical reactions for the synthesis of metal oxide particles. Liu et al.²⁵⁴ used the aldol condensation reaction between titanium isopropoxide and acetone to generate water molecules for the formation of TiO₂ spheres. Similarly, the water produced during the reaction of vanadium(V) oxytriisopropoxide with acetic acid catalyzed the hydrolysis–condensation process that led to the formation of vanadium oxide spheres.⁵¹ Guo et al. synthesized TiO₂ spheres by using water generated during esterification reaction of ethanol and acetic acid.²⁵⁵ In another example, TiO₂ shells were prepared on hydrated sulfate templates of ZnSO₄·7H₂O in ethanol.²⁵⁶ The spherical templates of ZnSO₄·7H₂O in ethanol that were formed acted as sites for deposition of titanium species. Moreover, water present in hydrated crystals acted as a supplier of water molecules for hydrolysis of titania precursors.

3. Effect of synthesis parameters

In this section, we examine the effects of the different experimental parameters such as reaction time, reaction temperature, calcination, pH and reactant and solvent types on the structure and phase of metal oxide spherical particles.

3.1 Reaction time

During initial stages of the reaction, nanocrystallites are formed and start aggregating into spheres to reduce their surface energy. The spheres are amorphous in nature but their size, long range ordering and the degree of crystallinity improve with increasing synthesis time.

Solution phase reactions at high temperatures favor hollowing via Ostwald ripening as time increases. The hollowing can be initiated at the central core whereby the relatively small crystallites dissolve and migrate to recrystallize on larger crystals on the surface of spheres. This mechanism has been used to obtain hollow spheres of different metal oxides, such as CeO₂,^67,69,72 CuO,^33,78,79,143 CuO/CuO₂,⁸⁰ ZnO,¹⁵⁸ MgO,⁹⁵ ZrO₂,⁶² SnO₂,¹⁵⁰ NiO,²⁴⁴ TiO₂²⁵⁷ and In₂O₃.⁹⁴ Some examples are presented in Tables 2 and 4. The shell thickness⁶⁷ or hollow core size⁹⁴ could be increased by increasing the reaction time. However, the synthesis time should be carefully adjusted because an excessive reaction time can corrode and collapse the formed hollow structures.^33,158

Another hollowing Ostwald ripening mechanism occurs at localized spots within the spheres, at fracture points or low crystal densities. Such cases can result in the formation of yolk–shell particles such as MoO₂@MoO₂,¹⁰² TiO₂@TiO₂,^81,203 CeO₂@CeO₂,⁷⁰ γFe₂O₃@γFe₂O₃,⁸⁸ BaZrO₃@BaZrO₃⁶⁴ and LnFeO₃@LnFeO₃ (Ln = La, Pr–Tb),⁹¹ multi-shell particles such as Fe₃O₄@TiO₂ double-shell structure⁸⁶ and V₂O₅ with up to three shells¹¹⁷ or porous In₂O₃ particles.³⁷ However these structures are, in general, eventually transformed into hollow shells due to further dissolution of the core if reaction time is further increased.

3.2 Reaction temperature

The reaction temperature is an important parameter, which affects the rate of crystal nucleation and therefore the size, morphology and polymorph of metal oxide particles.

Temperature affects the crystallization rate, which limits the quantity of crystals formed. This in turn affects the shape^32,36,64,79 and size^21,32,146 of the synthesized particles. As the nanocrystals continuously form, they aggregate into hierarchical structures. During hydrothermal growth of CuO spheres, the particle shape evolved from irregular CuO nanoparticles at 80 °C to uniform spindle-shaped CuO nanorods with sharp ends at 100 °C, small amounts of nanospheres at 140 °C and large scale monodisperse nanospheres with a wave-like surface at 160 °C.³² On the other hand, hollow flower-like spheres of α-GaOOH changed into microspheres and finally into rods as the temperature increased from 175 °C to 225 °C.³⁶ Therefore, the optimum temperature for achieving a maximum amount of the required structures needs to be experimentally determined. The diameters of solid spheres²¹ or shells^32,146 can also increase with temperature. The ultrasound-assisted solution precipitation of MnO₂ produced spherical particles with an average size that could be tuned between 0.4 and 1.28 μm by varying the temperature from 30 °C to 70 °C.²¹ High temperature was shown to favor the hollowing process of solid spheres via Ostwald ripening.^64,79,244 Conversely, the synthesis of hollow structures via bubble⁸² and emulsion templating⁸³ is not recommended at high temperatures due to reduced stability of the liquid medium.

Changes in temperature can also affect the growth patterns of the crystals, giving rise to different crystal phases. As an example, the Cu₂O content in CuO spheres increased with temperature^79,80 while α-Fe₂O₃ was favored at elevated temperatures instead of FeOOH.^35,90

3.3 Calcination

Calcination improves the crystallinity of metal oxides by densifying the crystallites and also removes organic impurities bound to the precursor during the hydrothermal process. As a result, the grain size of the crystals increases and the removal of water and CO₂ creates additional pores and interspaces. The sum of these effects generally decreases the surface area of particles (although an increase in the surface area has been observed in some cases⁵⁶) and increases the pore volume.

Calcination in an oxygen-containing atmosphere is required to remove carbon templates in the synthesis of porous or hollow spheres. Sometimes, multi-step calcination involving a pyrolysis step followed by oxidative calcination is required. During the synthesis of mesoporous Nb₂O₅,¹²³ the PEO-b-PS copolymer was used as a pore forming template and resol as a “glue” for the niobia composite following high temperature polymerization. The composite was pyrolyzed first at 350 °C to selectively decompose PEO-b-PS, then at 550 °C to crystallize Nb₂O₅ crystals without collapsing the polymer skeleton. Finally, the carbon skeleton was removed by calcination in air at 400 °C to obtain mesoporous crystalline Nb₂O₅ particles.

Some researchers reported the formation of yolk–shell structures of ZnCo₂O₄,¹²⁰ SnO₂, CeO₂ and Tb₄O₇⁷³via simple calcination of metal oxide precursors at 600 °C (using a rate of 5 °C min⁻¹) that were hydrothermally loaded into the pores of carbon spheres. The template removal occurred in two steps with the burning off of the first layer closest to the surface, separating the oxide shell and the composite core. Further calcination removed the carbon template from the core to produce the yolk–shell metal oxide structures.

In some circumstances, the heating rate was found to have some interesting effects on the final structure of calcined particles. A zinc oxide precursor–carbon composite was calcined in air via three methods.¹¹⁹ The first sample was calcined for 3 h in a preheated furnace at 550 °C, the second sample was calcined at a heating rate of 5 °C min⁻¹ to 550 °C and maintained for 3 h and the third sample was calcined at a heating rate of 2 °C min⁻¹ to 550 °C and maintained for 3 h. Single-shell hollow spheres were obtained from the first sample, while double- and triple-shell spheres were formed from the second and third samples, respectively. The formation of multi-shell particles was explained by the occurrence of a temperature gradient between the exterior and interior of the precursors, giving rise to an inside out Ostwald ripening process. Dong et al.¹⁷⁰ went a step further by proposing an easy way to control the number of shells and inter-shell spacing in hollow ZnO microspheres by controlling the heating processes (heating rate and final temperature) and precursor concentration on the carbon sphere templates. In another work, a gradual removal of the carbon core from C@V composite microspheres was successfully used to obtain V₂O₅@V₂O₅ yolk–shell particles.¹¹⁶ The core was much larger when the C@V sample was annealed at 350 °C at a rate of 3 °C min⁻¹ as compared to a rate of 1 °C min⁻¹, due to the relatively fast removal of the carbon core at the highest rate, leaving behind a V₂O₅ core formed from the vanadium species that were initially bound to the carbon core. However, annealing at 400 °C at a rate of 1 °C min⁻¹ gave porous single-shell spheres. In a similar fashion, the control of heating rate could generate core-in-double-shell NiCo₂O₄ spheres¹⁰⁵ and multi-shell Y₂O₃ spheres.¹⁸⁰

The calcination temperature can also govern the metal oxide polymorphism. For example, anatase TiO₂ forms between 300 °C and 600 °C and the more crystalline rutile phase starts to form at temperatures >600 °C,^182,191,258 the perovskite LaFeO₃ would not form below an annealing temperature of 750 °C,⁹¹ β-Bi₂O₃ was obtained at 350 °C while α-Bi₂O₃ was formed at 450 °C,²⁵⁹ the change of phase from γ-Ga₂O₃ to β-Ga₂O₃ occurred at T > 700 °C⁹² and γ-Al₂O₃ and α-Al₂O₃ were formed at 900 °C and 1100 °C, respectively.¹²¹

3.4 pH

We have already mentioned how pH alters the degree of precipitation and induces Ostwald ripening in the particles. Nevertheless, the solution pH can affect particle growth in various ways, for instance, by controlling adsorption of the precursor on carbon templates,¹⁷⁵ changing the degree of ionization of the ligands,⁶⁷ modifying the electric charge of the surface of nanoparticles²⁵³ and altering the degree of complexation of the metal ions in solution.¹⁰⁴

In the synthesis of Mn₂O₃, multi-shell structures were formed with the help of carbon sphere templates; the amount of metal cations that can be adsorbed on the carbon spheres was varied by using different solution pH (since high precursor concentration results in its accumulation on the template surface rather than in its infiltration into template pores).¹⁷⁵ This strategy was shown to be well suited to alter the zeta potential of carbon spheres and hence their electrostatic interaction with metal ions. As a result, depending on pH, single-, double- or triple-shell Mn₂O₃ structures were formed upon calcination.

CeO₂ particles of different shapes were respectively obtained at pH of 1, 2 and 3.5 in the presence of citric acid used as a ligand, namely, solid spheres, hollow spheres and microplates.⁶⁷ Due to the three levels of ionization of citric acid (pK_a = 3.13, 4.76 and 6.4), different ligands were formed at different pH. At pH below 1, most of the citric acid was in solution as H₃Cit without ionization, which coordinated with Ce³⁺ to give solid microspheres. At pH = 3.5, H₂Cit⁻ concentration was higher than H₃Cit concentration, giving rise to microplates. At pH = 2, the ligands were a mix of H₃Cit and H₂Cit⁻, which resulted in the formation of coordination polymers (metal organic frameworks) in a metastable stage between solid spheres and microplates, compromising the shape into hollow spheres.

The pH of a microemulsion system was found to alter the surface energy and the electric charge distribution on the surface of In(OH)₃ nanoparticles as well as the amount of adsorbed CTAB on the different crystal faces.²⁵³ As a result, the particles aggregated and grew in different directions to form bundles of nanorods at pH = 5 and spheres at pH = 3.

The hydrothermal synthesis of NiO in the presence of D-glucose was performed at different ammonia concentrations¹⁰⁴ to investigate its effect on the shape of final particles. A high concentration of ammonia decreased the concentration of free Ni²⁺ by forming stable [Ni(NH₃)_x]²⁺ complexes in solution and thereby controlling the rate of precipitation of Ni(OH)₂. Therefore, aggregated porous structures were obtained at pH < 10.5, and at pH 10.5 a controlled precipitation of the Ni species on the in situ formed carbon spheres followed by calcination afforded hollow microspheres. A further increase of pH to 10.9 reduced the rate of precipitation of nickel hydroxide, allowing formation of carbon layers on the surface of Ni(OH)₂ particles, which upon calcination were converted to multi-shell NiO microspheres. Therefore, pH is an important parameter to consider for controlling the morphology of particles.

3.5 Solvents and reactants

The type and composition of solvents is also an important factor that can be used to manipulate the morphology and surface area of metal oxide particles.^{55,83,90,95,122,251,260}

A mixture of water and ethanol was used to synthesize multi-shell Co₃O₄ when using carbon spheres as templates as opposed to single shells when only water was used.¹⁷¹ The effect of ethanol was used to decrease the number of aqua groups coordinated to Co ions, hence to decrease the size of the hydrated Co ions. This strategy was used to control the diffusion of ions through the carbon template, allowing the production of spherical structures, which upon calcination gave multi-shell structures. The same principle was applied to synthesize α-Fe₂O₃ multi-shell hollow microspheres.¹⁷²

The quasi-emulsion template was used for the formation of α-Fe₂O₃ hollow spheres;⁸³ the concentration of the soft template was varied by changing the glycerol/water ratio at 145 °C. As a result, the morphology was evolved from solid flower-like particles consisting of densely packed needle-like subunits to hollow spheres and finally solid spheres made up of nanosheets at the solvent ratios of 1 : 19, 1 : 7 and 1 : 4, respectively. The effect of the EG/water ratio was investigated during hydrothermal synthesis of MgO microspheres.⁹⁵ It was found that at a high EG volume, spheres were obtained, while at a high water volume, nanoplates were formed.

In another work, the glycol concentration had no effect on the morphology but rather affected the surface area of the synthesized particles.¹²² The BET surface area of CuO spheres (before calcination) decreased from 157 m² g⁻¹ to 104 m² g⁻¹ as the glycol content in water increased.

The solvent ratio can also influence the particle size. The solution phase precipitation at room temperature of V₂O₅ spheres in a mixed solvent of acetone, pyridine and water gave particles of different sizes (between ∼150 nm and ∼1000 nm) simply by varying the water or pyridine concentration.¹²⁷

An increase in the ratio of the structure directing agent (SDA) to the metal precursor improved the sphericity of particles such as ZnO,^56,185 SnO₂,¹²⁵ and Bi₂WO₆ spheres²⁴ as well as α-Fe₂O₃ multi-shell hollow spheres⁸⁷ because SDA acted as a capping agent and prevented uncontrolled aggregation, although an excess of SDA can be detrimental to the crystallinity¹¹⁸ or development of the structures.²⁴ Moreover, this ratio has also been used to control the size and morphology of particles such as NiO,⁴³ Y₂O₃,²⁶¹ ZnO,¹³⁴ ZnCo₂O₄@CeO₂,¹³³ Bi₂O₃,²³ SnO₂⁴⁸ and Co₃O₄.²⁸ Conversely, the effect of the change in the precursor concentration while keeping the overall ratio constant has been examined in the low temperature precipitation of ZnO hollow spheres using HMT at a 1 : 1 ratio.¹⁵⁹ Hollow fluffy-like spheres having a size of 2–3 μm and a surface area of 138 m² g⁻¹ were produced; however a decrease in the concentration of the reactants (but keeping the same ratio) gave solid spheres with a net-like surface and a much higher surface area of ∼368 m² g⁻¹.

The type of reactant can also impact the shape or phase of the final products. CuO/Cu₂O hollow spheres were obtained via a hydrothermal method using ethanolamine but CuO flower-like spheres were obtained when ammonia was used.⁸⁰ In the hydrothermal synthesis of NiO spheres, the β-Ni(OH)₂ phase was formed when NaOH or ethanolamine was used while α-Ni(OH)₂ was formed with urea instead.⁴³ This study indicates that the phase and structure of the final products depend on the special properties and structure of the reactants used.

4. Applications of spherical metal oxide particles

Metal oxides are used in diverse and multidisciplinary areas ranging from industrially relevant sensing and catalytic uses to biomedical, environmental and energy-related applications. This section shows how the characteristics of different metal oxides affect their respective applications.

4.1 Sensing and catalytic applications

4.1.1 Sensors. Sensors are used to measure and detect slight changes of a chemical compound or molecule in their surroundings, resulting in important applications in fields such as environmental monitoring, biomedical detection, oil and gas industry, metallurgy, food industry, etc. Some typical chemicals that require sensing are flammables such as ethanol,^{27,45,56,57,59,93,251,262–264} acetone^131,262 and hydrogen,^47,48,125 strong oxidizing agents such as H₂O₂,¹²³ health hazards such as 2-chloroethanol,⁴⁶ ammonia,¹⁵⁸ H₂S,¹⁴³ methanol,¹⁸³ toluene,^30,131 NO₂,¹³⁰ trimethylamine,²²⁶ formaldehyde^82,205 and CO^47,48,151 as well as glucose^239,240 to monitor its level in blood. While the most common materials that have been used are SnO₂ spheres and hollow spheres^{45,47,48,125,151,262,264} and ZnO spheres and hollow spheres^{56,57,59,158,239} due to their low cost and availability, other metal oxides have been tested too, such as Nb₂O₅ spheres,¹²³ In₂O₃ hollow spheres,^93,94,205 MnO₂ spheres,²⁴⁰ Co₃O₄ spheres,²⁷ Co₃O₄ hollow spheres,²⁵¹ α-Fe₂O₃ hollow spheres,⁸² yolk–shell α-Fe₂O₃ spheres,⁸⁴ CuO hollow spheres,¹⁴³ WO₃·H₂O spheres,¹³¹ WO₃ spheres,¹³⁰ WO₃ hollow spheres,²²⁶ Cr₂O₃ spheres³⁰ and TiO₂ hollow spheres.¹⁸³

Basically sensors operate in two ways. Biosensors make use of immobilized enzymes that can break down a biomolecule to give an electrochemical signal. On the other hand, n-type oxide semiconductors such as SnO₂, ZnO, In₂O₃, TiO₂ and WO₃ operate by the chemoresistive detection of reducing gases at their surface, involving the diffusion of the analyte gas towards the sensor surface and its electrochemical oxidation with a negatively charged adsorbed oxygen.⁹⁴ Sensors should have a high sensitivity (at the ppb or ppm level), very fast response (a few seconds), good selectivity towards similar substances, stability, a wide response range, a linear dependency with respect to gas concentration, good repeatability and reusability. These qualities are imparted by the proper choice of sensor material with a microstructure that has been tailored to have specific properties such as a high surface area, porosity, thermal and chemical stability, a short diffusion length for efficient mass transfer, non-toxicity and biocompatibility (for biosensors). From morphology viewpoint, porous or hollow particles have been found to be well suited as sensing materials since their pores allow for rapid mass diffusion and can accommodate numerous active sites, while hollow particles offer short diffusion lengths and an empty structure that can act as a reservoir to continuously replenish active sites with the monitored substance. To improve their performance, some sensors have been loaded with precious metals such as Au,^47,56,262 Er⁹³ and Rh,⁹⁴ made as composites, for example α-Fe₂O₃/In₂O₃ hollow spheres²⁶³ or as hierarchical structures with high surface area such as nanosheet-assembled WO₃ microspheres¹³⁰ or Co₃O₄ microspheres composed of large and thin nanoplatelets.²⁷

Kim et al.⁹⁴ observed that 1.67 atomic% Rh loaded into In₂O₃ hollow spheres showed an excellent response to 2–100 ppm ethanol, which was up to 180 times higher than in the case of unloaded In₂O₃ hollow spheres. Also, Rh additionally decreased the optimum operating temperature from 475 °C to 371 °C and enhanced the selectivity to ethanol 15–25 times. The sensing time was very short (0.4 s) but recovery time was relatively large (∼200 s) due to thermal promotion of the surface reactions. A higher Rh loading could decrease the recovery time at the expense of a lower response, which was attributed to the formation of unwanted rhodium and indium phases.

Hierarchical MnO₂ spheres consisting of nanorod subunits were synthesized via electrodeposition and loaded with the glucose oxidase enzyme for the mediator-less detection of glucose.²⁴⁰ The particles exhibited a high surface area of 129 m² g⁻¹ and pore size between 5 and 12 nm, which could easily accommodate the enzyme molecules. The sensor had a high sensitivity of 31.6 μA mM cm⁻², a large linear range up to 3.15 mM and a low detection limit of 0.35 μM. Moreover, no interference was observed with species that coexist with glucose in blood, such as ascorbic acid, uric acid and acetaminophen, indicating good selectivity of the sensor.

With the advent of more and more chemicals that are classified as toxic and the implementation of more stringent laws for their detection, the development of more versatile and sensitive sensors will be required in the future and metal oxides can effectively offer a solution.

4.1.2 Catalysts. The attributes that make metal oxide nanoparticles attractive as catalysts are their high surface area, easy accessibility of pores (for reactants as well as for loading noble metals), thermal resistance (since many catalytic reactions are performed at elevated temperatures), toughness (to avoid damage during multiple use), chemical stability, environmental friendliness, special inherent properties that are useful for specific applications and ability to interact synergistically with loaded nanometals for superior catalytic performances. Fig. 12A presents a simplified illustration of the operating principle of a supported catalyst.

Fig. 12 (A) Operating principle of a metal oxide support loaded with nanoparticles of a metal catalyst. (B) Catalytic reduction of 4-NP to 4-AP. (a) Reaction equation. (b) Time-dependent absorption spectra of the reaction solution in the presence of the Pd nanocube-embedded hollow mesoporous TiO₂ microspheres. (c) Plot of ln[C(t)/C(0)] against the reaction time. The R² = 0.9956 is the coefficient of determination obtained from the linear fitting. (d) Plot of C(t)/C(0) against the reaction time for five successive cycles of the reduction reaction catalyzed by Pd nanocube-embedded hollow mesoporous ZrO₂ microspheres. (e) TEM image of the ZrO₂ microspheres after five cycles of the reduction reaction (panel B reproduced with permission from ref. 223. Copyright © 2013, Wiley-VCH).

The environmentally friendly room temperature degradation of HCHO to CO₂ and H₂O was achieved by using Pt loaded γ-Al₂O₃ hollow spheres.⁶³ As shown in Fig. 12B, the hierarchical macro–mesoporous structure allowed for high dispersion of Pt and accessible pores facilitated diffusion of reactive molecules and products to/from reaction sites, respectively. In another work,²²³ the embedded Pd nanoparticles inside the hollow space of mesoporous TiO₂ and ZrO₂ spheres synthesized via the PS nanospherical template-assisted aerosol method prevented the aggregation of metal nanoparticles and reduced the loss of catalyst during recycling, enhancing the reduction reaction of 4-nitrophenol. Similarly, high CO oxidation at low temperatures was possible via the use of nanosized Au in Au@TiO₂ yolk–shell particles due to the unique synergy between Au and TiO₂ and the protection of nanometal by the shell.¹⁹⁹

Porous ZrO₂ microspheres have strong Lewis acidity and poor Brønsted sites but their treatment with phosphoric acid could improve the acidity of the Brønsted sites.¹³⁶ The synthesized catalyst was tested for the catalytic Friedel–Crafts alkylation of indoles with chalcones, giving 98% yield in 6 h and showed a negligible loss in its activity even after 22 reuses. Similarly, Friedel–Crafts reactions were carried out with mesoporous Nb₂O₅ spheres⁴⁰ and Fe₃O₄@TiO₂ double-shell yolk–shell particles⁸⁶ once the Lewis and Brønsted sites were improved by acidification with sulfuric acid and hydrothermal treatment in basic solution, respectively.

CeO₂ on the other hand shows strong oxygen storage and release capacity via facile conversion between Ce³⁺ and Ce⁴⁺. This makes ceria a model catalyst for oxidation reactions. Hollow CeO₂ spheres converted 43% CO at 295 °C,⁷⁰ while 10% Cu doped CeO₂ spheres showed >98% CO conversion at 210 °C⁶⁸ and a complete oxidation was achieved on the core–shell ZnCo₂O₄@CeO₂ particles at less than 200 °C.¹³³ However, loading CeO₂ with noble nanometals such as Pd,⁷⁴ Ag²⁶ and Au⁶⁹ drastically improved its performance at much lower temperatures. For instance, Pd@CeO₂ multi-yolk–shell particles achieved 100% CO oxidation at 110 °C.⁷⁴ The catalyst also performed excellently in the case of aerobic oxidation of cinnamyl alcohol into cinnamaldehyde with >99.9 conversion after 1.5 h. In another case, CeO₂ doped into MnO₂ hollow spheres could catalytically oxidize benzene with 90% oxidation at 252 °C and completely oxidize it at 340 °C.⁹⁹

Mesoporous NiO spheres and flowers having a size of 500 nm and mesopores in the range of 10–30 nm showed high catalytic activity for the transformation of toxic phenolic pollutants.² Furthermore, the NiO nanostructures with large scale nanocrystal domains and well-shaped morphologies imparted magnetic properties to an otherwise antiferromagnetic system due to the quantum confinement effect, allowing easy magnet-assisted separation.

Other types of metal oxides and their combinations have been investigated for several other reactions. Perovskite LnFeO₃ (Ln = La, Pr–Tb) hollow spheres showed an excellent catalytic performance for NO + CO reaction at high temperatures between 200 °C and 500 °C due to their outstanding thermal and chemical stability.⁹¹ MgO spheres and flowers performed much better than bulk MgO in the Claisen condensation of benzaldehyde and acetophenone under solvent-free conditions.²⁶⁵ Co₃O₄ hollow spheres were used for methane conversion¹⁴¹ and H₂O₂ oxidation.¹⁴² Fe₃O₄@SiO₂@void@TiO₂ particles were employed as a catalyst for epoxidation of styrene,¹⁹⁸ giving a high conversion and selectivity of 90.2% and 88.5% respectively. In the latter work, the Fe₃O₄ core was well protected by the SiO₂ layer and the void space was successfully loaded with Au nanometals, which respectively provided magnetic separability and high reactivity and reusability.

4.2 Biomedical applications

4.2.1 Photoluminescence. Rare earth doped materials have been widely used as high performance upconversion or downconversion luminescent devices. Upconversion is an optical process in which the interaction of two or more photons leads to the emission of light at shorter wavelength than the excitation wavelength such that energy is conserved, while in downconversion, one high energy (short wavelength) photon is converted to two or more, low energy (longer wavelength) photons. This behavior arises due to their unique electronic and optical properties resulting from their 4f electrons. The host matrix for the rare earth elements needs to have good chemical durability, thermal stability and low phonon energy. Lanthanide oxide spheres such as Y₂O₃ solid and hollow spheres^{55,132,155–157,261} or Lu₂O₃ hollow spheres²⁶⁶ have been used in upconversion devices. For example, under UV or low voltage electron beam excitation, the Ln³⁺ (Ln³⁺ = Eu³⁺ and Tb³⁺) ion doped Y₂O₃ samples showed strong red and green emission, respectively,¹⁵⁶ whereas multi-color upconversion emission was realized under 980 nm excitement when using Er doped Y₂O₃ spheres.¹³² Furthermore, multi-shell structures revealed that the upconversion luminous intensity increased as the number of shells increased due to multiple reflections and the enhanced light harvesting.¹⁸⁰ Luminescent materials could find application in the field of lighting, optoelectronic devices, biological labeling and drug delivery. Recently, hollow structured Y₂O₃:Yb/Er–Cu_xS nanospheres with multifunctionalities were created.²⁶⁷ Aside from upconversion luminescence properties, the material showed good biocompatibility and exhibited high ablation efficiency with regard to tumor cells because of the synergetic therapeutic effect between chemotherapy and photothermal therapy arising from loaded DOX and attached Cu_xS nanoparticles.

Some semiconductors were shown to have photoluminescence capabilities. This arises from oxygen vacancy defects that can accommodate photogenerated electrons, which release photons upon recombination with free photogenerated holes. Some examples of semiconductors that have been found to possess photoluminescence properties are perovskite BaZrO₃ hollow spheres,⁶⁴ In₂O₃ spheres^37,253 and γ-Ga₂O₃ hollow nanoflowers,⁹² with the latter also showing excellent solar blind detection performance.

4.2.2 Drug delivery and antimicrobial properties. Metal oxide spheres are also good candidates as drug delivery vehicles. TiZr oxide spheres showed high loading of drugs such as IBU, DEX and ERY as well as hydrolytic stability and sustained release profiles in simulated body fluids.²⁶⁸ Gd₂O₃ nanoparticles have a combination of magnetic and optical properties, which can be advantageous in the field of biomedicine. Indeed, Gd₂O₃:Yb/Er hollow spheres could simultaneously serve as a dual-imaging agent for optical/MR imaging and as a vehicle for drug storage/release.¹⁴⁵ Magnetic-based drug distribution was also a possibility with PEG-coated Fe₃O₄ and γ-Fe₂O₃ hollow spheres.⁸⁵ Both structures had high ibuprofen storage capacities of 297 mg g⁻¹ and 237 mg g⁻¹, respectively. Furthermore, a sustained drug release could be observed with most of the loaded drug released to the simulated body fluid in 136 h; the process of loading and releasing drugs is illustrated in Fig. 13A.

Fig. 13 (A) Schematic illustration of loading and unloading a drug to/from the PEG-coated hollow metal oxide particles. (B) (a) HRSEM image, (b) TEM image of hematite HNS, (c) HRSEM image, (d) TEM image of magnetite HNS, (e) HRTEM images of hematite HNS showing lattice fringes of the (006) plane, and (f) magnetite HNS showing lattice fringes of the (311) plane. (C) (a) Dynamics of extracellular dissolution of total Fe throughout the course of microbial growth in the absence and presence of bacteria and (b) intracellular content of Fe in the bacteria after 4 h incubation with nanoparticles of 500 μg ml⁻¹ concentration in LB broth (the inset illustrates the enhancement of the antimicrobial process). (Panels B and C reproduced with permission from ref. 269. Copyright © 2016, Wiley-VCH.)

Nor et al. reported hematite hollow nanospheres (HNS) synthesized by the hard templating method; reduction of hematite HNS by H₂ led to magnetite HNS (Fig. 13B). It was reported that magnetite HNS (hematite HNS and C-magnetite) shows superior antibacterial performance towards both E. coli and S. epidermidis (Fig. 13C). In comparison to hematite HNS, magnetite HNS allows for a multiple-fold increase in the generated soluble iron ions, showing that the control over both the composition and nanostructure is crucial for tuning the antimicrobial activity of iron oxides.

4.3 Environmental applications

4.3.1 Photocatalysis. Semiconductors can be photoexcited by light of suitable energy to generate electrons and holes, which can be used as mediators for the photocatalytic removal of pollutants (commonly dyes) and even for bacterial deactivation.^23,107 The pertinent features of a semiconductor nanomaterial that make it an attractive photocatalyst are its morphology, phase, size of crystals, chemical and physical stability and reduced amount of defects in the crystal lattice. These allow for increased lifetime of photogenerated electrons and holes and their fast migration towards the particle's surface for redox reactions, while reducing the probability of their recombination. Additionally, a proper engineering of semiconductor nanocrystallites can effectively reduce the band gap energy (hence the required energy of incident photons) with respect to the bulk sample, due to the quantum confinement phenomena, making them responsive to visible light. Numerous materials have been employed as photocatalysts using UV light such as SnO₂ spheres,^107,125 ZnO spheres and hollow spheres,^{58,60,118,270,271} ZnO multi-shell hollow spheres,¹¹⁹ γ-Fe₂O₃ hollow spheres,⁸⁸ In₂O₃ spheres,³⁸ α-Ga₂O₃ spheres,³⁶ CuO/Cu₂O composite hollow spheres,⁸⁰ SiO₂ supported Ta₂O₅ hollow spheres,¹¹³ perovskite type SrTiO₃ hollow spheres¹¹² and Fe₃O₄@TiO₂ double-shell yolk–shell spheres.⁸⁶ Catalysts that performed well under visible light due to their narrow band gap energy were Bi₂O₃ spheres,^23,259 WO₃ spheres and hollow spheres,^129,153 α-Fe₂O₃ spheres³⁴ and CeO₂ hollow spheres,⁷¹ perovskite types such as BiFeO₃ spheres⁶⁶ and Bi₂WO₆ spheres^25,229 and composites such as CeO₂/Bi₂O₃ nanospheres,²⁷² Bi₂O₃/Co₃O₄ spheres,⁶⁵ WO₃/TiO₂ spheres⁵⁴ and WO₃/WO₃·H₂O hollow spheres.¹⁵⁴

The morphology of the material is of utmost importance for designing a photocatalyst. A large surface area affords numerous reaction sites and a high crystallinity favors photocatalytic activity due to effective charge migration. Indeed, the annealing process improves the crystallinity of the material but at the expense of the total surface area. Therefore, a compromise needs to be reached for the best crystal size and the largest possible surface area for optimum photocatalytic performance. SnO₂ spheres calcined at 150 °C had the best activity under UV light for the degradation of methyl orange as compared to uncalcined samples and samples calcined at 200 °C, 300 °C and 500 °C respectively.¹²⁵ Similarly, Bi₂O₃ spheres calcined at 350 °C degraded 99% of methyl orange under visible light after 3.5 hours as compared to only 25% degradation for samples calcined at 500 °C. It was also shown that hollow structures performed better than solid ones due to multiple light reflection and absorption within the structure cavities (Fig. 14A) and stability of the photogenerated charge carriers.

Fig. 14 (A) Schematic illustration of multiple reflection and absorption of UV light in a hollow shell photocatalyst. (B) SEM image and schematic illustration of photoexcitation electrons and holes migrating to the opposite sides of the interface in hollow ZnO spheres (panel B reproduced with permission from ref. 225. Copyright © 2014, American Chemical Society).

Dilger et al.²²⁵ studied the effect of nanoarchitecture on the photoconductivity of ZnO particles (Fig. 14B). They found that the time needed for the current from photogenerated charge carriers to drop down to 10% of its maximum value (T_0.1) was <60 s for bulk and porous ZnO, whereas for yolk–shell and hollow spheres, T_0.1 was substantially higher reaching 521 s and 1150 s, respectively. This structural effect is duly reflected by photocatalytic experiments using various materials. Porous CeO₂ hollow spheres degraded 92% of gas phase acetaldehyde within 24 h, four times better than in the case of CeO₂ nanoparticles, despite the fact that the latter had a smaller band gap (2.88 eV as compared to 3.01 eV for hollow CeO₂ spheres) and better crystallinity.⁷¹ WO₃ hollow spheres had better performance than flower-like spheres, with first order kinetics of 0.056 min⁻¹ as compared to 0.0151 min⁻¹ for the removal of rhodamine B under UV light.¹²⁹ While both structures possessed similar surface areas (13 m² g⁻¹ and 16 m² g⁻¹ for the microspheres and hollow spheres, respectively), the improvement in the catalytic activity of the hollow spheres was attributed to the large textural porosity between 3.3 nm and 89.6 nm. Increasing the number of shells also resulted in better photoactivity. For example, ZnO hollow spheres with 1, 2 and 3 shells degraded 84.1%, 88.3% and 99% of rhodamine B respectively,¹¹⁹ while the required time for complete degradation of the same pollutant under visible light was 1.5 h for WO₃ with multiple shells as compared to 2.5 h for single shell particles.¹⁵³

The crystal phase, due to its geometric structure, is another important aspect that is related to photoactivity. Hou et al.²⁷³ showed that the β-phase of Ga₂O₃ exhibited superior photocatalytic activity towards the gas phase degradation of aromatic compounds as compared to α- and γ-Ga₂O₃. Also, studies of Fe₂O₃ as catalyst showed that α-Fe₂O₃ spheres could completely degrade rhodamine 6G after 1 h under visible light irradiation,³⁴ while F doped γ-Fe₂O₃ needed UV light to degrade rhodamine B.⁸⁸

TiO₂ has been the most extensively studied semiconductor for photocatalytic purposes due to its abundance, low cost, non-toxicity and high photoactivity accompanied by a high oxidation and reduction potential. TiO₂ can oxidize a large variety of organic compounds in water or air^{182,189,193,200,203,254–257,274–278} and has also been successful for reducing metal ions such as Cr^6+ 193 and for H₂ production.²⁵⁴ The major factors affecting the degree of photoactivity are surface area, porosity, crystallinity and morphology. A high crystallinity is desired rather than a large surface area to increase the separation between the photogenerated electrons and holes, and reduce their instantaneous recombination. However, the rutile phase despite its higher crystallinity and lower band gap energy has a lower activity than the anatase phase because of the lower surface area of the rutile particles. As a result, the surface area and the crystal size of particles have to be finely tuned due to their inverse role in photocatalysis, which is the reason why a wide discrepancy has been observed for the optimum calcination temperatures of the synthesized TiO₂ particles (between 450 °C and 800 °C for grain sizes varying between ca. 12 and 30 nm), depending on their sizes, morphologies and treatment methods.^{177,182,200,278} The activity of TiO₂ under visible light has been enhanced by doping with Pt,²⁵⁴ CO₃²⁻,²⁵⁴ N,^193,279 metal oxides,²⁵⁶ CdS¹⁸⁴ and even trace organics from the solvent or organic precursor,²⁵⁵ since the inclusion of a dopant within the TiO₂ matrix provides a sink for photogenerated electrons, thereby reducing the degree of electron–hole recombination. Similar to other semiconductors discussed in this section, the hollow shell morphology TiO₂ performed better than its other counterparts²⁰³ because of multi-reflection of light.

4.3.2 Adsorption. Due to their isoelectric point and porosity, semiconductors have been useful as adsorbents for removal of pollutants from wastewater. MgO spheres had excellent adsorption properties towards fluoride and arsenic in water, with 120 mg g^−1 95 and 569.7 mg g⁻¹ respectively.⁹⁶ ZrO₂ solid and hollow spheres,⁶² CuO spheres and Cu₂O hollow spheres,³² NiO spheres,^42,43 perovskite BaZrO₃ hollow spheres⁶⁴ and CeO₂ hollow spheres⁷⁰ were used for the removal of anionic dyes, while urchin-like α-FeOOH hollow spheres⁹⁰ had high removal capacities towards both organic dyes and heavy metal ions. MnO₂ spheres with a hierarchical dandelion-like surface structure, a high surface area of 163 m² g⁻¹ and main mesopores of about 65 nm showed both oxidation and adsorption properties over lethal As(III) species.¹⁷ The As(III) species could be effectively oxidized by MnO₂ particles to the less toxic As(V) species, followed by their adsorption.

4.4 Energy storage and conversion applications

4.4.1 Lithium-ion batteries. Transition metal oxides have been actively investigated as electrodes for lithium-ion batteries because they can reach high theoretical reversible capacities of more than 700 mA h g⁻¹,²⁸⁰ while the graphite used in commercial batteries has a relatively low Li storage capacity of 370 mA h g⁻¹.¹⁸ The energy density and life-cycle of the existing Li-ion batteries are unsatisfactory for many applications involving large devices such as electric vehicles and large scale energy grid storage,¹⁰² and better alternatives are being explored. The two major advantages offered by metal oxide colloidal structures that make them attractive as electrodes in lithium-ion batteries are high surface area for better contact of the active material with the electrolyte and porosity, which accommodates the volume change associated with the insertion and extraction of Li ions during charging and discharging processes, as shown in Fig. 15 for yolk–shell type particles.

Fig. 15 (A) Charging and discharging of Li ions in a yolk–shell type Li-ion battery; (B) schematic illustration of the formation process of yolk–shell MoO₂ microspheres; (C) (a) discharge–charge voltage profiles of the MoO₂ electrode at different current densities of 50, 500 and 2000 mA g⁻¹; and (b) cycling performance of the as-synthesized MoO₂ microspheres at 50, 500 and 2000 mA g⁻¹. (Panels B and C reproduced with permission from ref. 102. Copyright © 2013, Royal Chemical Society.)

V₂O₅ microspheres synthesized by an additive-free hydrothermal method involving a hierarchical assembly of nanoporous fibers possessed a moderate surface area of 42 m² g⁻¹.⁵¹ The structure endowed the material with shorter diffusion pathways for easier Li and electron transport hence enhancing electrochemical performance. The microspheres displayed a very stable capacity retention of 130 mA h g⁻¹ over 100 cycles at a current rate of 0.5C and showed an excellent rate capability with a capacity of 105 mA h g⁻¹, even at a high rate of 30C. Also, TiO₂ anatase spheres consisting of ultrathin nanosheets formed via crystals growing in the [001] direction¹⁸⁸ performed well at low and high rates. In another work,²⁵⁰ NiO spheres with loosely connected crystals, a surface area of 30 m² g⁻¹ and pore size distribution in the range of 10–60 nm provided suitable pathways for efficient transport of electrolyte ions, large surface to volume ratios, and good structural stability to deliver a reversible capacity of 800 mA h g⁻¹ after 100 cycles at a current density of 500 mA g⁻¹. The material was also tolerant to various charge and discharge currents, indicating high rate performance for high power applications. Aside from the morphology, crystal polymorphs can also be a deciding factor in the fabrication of metal oxide-based Li-ion batteries. It was shown that the anatase phase exhibited higher capacity than either brookite or rutile due to the more favorable Li ion insertion mechanism.^{188,191,281,282}

Hollow structures have also attracted great attention due to the increased contact area between the electrode material and electrolyte as well as their mechanical flexibility. Hierarchical orthorhombic V₂O₅ hollow spheres showed good performance in Na-ion batteries (an alternative to Li-ion batteries) due to the predominantly exposed [110] crystal planes, which provided channels for easy Na⁺ insertion and extraction as well as high tolerance to the deformation imparted by voids in the shells of hollow spheres.¹¹⁵

Mn₃O₄ hollow spheres synthesized by an aerosol method featured a high surface area of 96 m² g⁻¹ and thin shells of 5–10 nm. These hollow spheres exhibited good stability with a high capacity retention of 980 mA h g⁻¹ over 140 cycles and exceptional rate capability by retaining a capacity of 300 mA h g⁻¹ at an ultra-high current density of 10 000 mA g⁻¹.²²⁴ These values represent the best electrochemical performance for Mn₃O₄ anode materials to date and were attributed to the unique thin wall hollow structure, which provided considerably reduced diffusion paths for electrons and Li ions. On the other hand, multi-shell hollow α-Fe₂O₃ showed a superior capability of 1203 mA h g⁻¹ at a current density of 100 mA g⁻¹ after 128 cycles and an excellent cycling stability of 870 mA h g⁻¹ at a current density of 400 mA g⁻¹ after 300 cycles due to the material's hierarchical porosity and structure, which assured shorter diffusion pathways for efficient transport of electrons and Li ions.⁸⁷ An excellent performance of multi-shell hollow α-Fe₂O₃ spheres was indeed recently confirmed by another work,¹⁷² reporting up to 1702 mA h g⁻¹ at a current density of 50 mA g⁻¹, which was due to the enhanced volumetric capacity of the structure that allows for maximum lithium storage. Recently, multi-shell V₂O₅ hollow microspheres, synthesized by a simple method involving adsorption of anions on carbon templates, were found to exhibit an exceptionally high specific capacity of 447.9 mA h g⁻¹ (at a high current density of 1000 mA g⁻¹), exceptional rate capability and cycling stability, due to the ample charge storage sites, short transport paths and good structural stability of the material.¹⁷⁶ These attributes also belong to yolk–shell structures such as MoO₂@MoO₂¹⁰² and V₂O₅@V₂O₅.^116,117 The superiority of the yolk–shell structure over solid spheres was demonstrated for the V₂O₅@V₂O₅ yolk–shell particles retaining 89% of its specific discharge capacity after 50 cycles as compared to around 60% only for solid V₂O₅ spheres.¹¹⁷

Composites or multi-component structures are attractive for studies due to the synergetic effect between the different components. Carbon-supported amorphous and crystalline V₂O₃ microspheres both showed excellent high rate and electrochemical performance due to the uniform distribution of partly graphitized carbon within the framework and also due to the stable structure of spheres, which assured a low charge resistance, fast electronic transport, a large surface area and excellent stability.⁵⁰ The amorphous spheres showed 95% retention in the discharge capacity after 7000 cycles at a high current density of 2 A g⁻¹, while the crystalline ones retained 98% after 9000 cycles. Ant-cave structured MoO₃–C composite microspheres²³⁴ synthesized by USP exhibited high initial discharge and charge capacities of 1109 and 724 mA h g⁻¹ and still delivered a discharge capacity of 733 mA h g⁻¹ after 300 cycles. The high performance was attributed to the combination of ant-cave channel structure and conductivity of the carbon in the composite. A yolk–shell structure of CuO@NiO spheres exhibited much higher capacity than the theoretical value of 1061 mA h g⁻¹ after 200 cycles due to the unique multilayer hollow structure, which provided a large electrochemically active surface, more active sites for Li ion storage and facilitated Li diffusion.²⁸³ In another example, the carbon coated triple-shell hollow spheres of CoMn₂O₄ possessed a high specific capacity of 726.7 mA h g⁻¹ and nearly 100% capacity retention after 200 cycles.¹⁶⁸

Numerous other metal oxides and composites of various shapes and accommodating space for lithium have been successfully used in Li-ion batteries. The list of metal oxides used in Li-ion batteries is quite long and includes the following: CuO spheres and hollow spheres,^18,78,79,122 Fe₂O₃ hollow spheres,^83,88 α-Fe₂O₃ multi-shell hollow spheres,^172,231 Co₃O₄ spheres,²⁸ Co₃O₄ multiple shells,¹⁷¹ Gd₂O₃ hollow spheres,¹⁴⁵ NiO spheres and hollow spheres,^{41,103,149,284} TiO₂ hollow spheres,²⁸⁵ TiO₂ yolk–shell spheres,²⁸¹ multi-shell TiO₂ hollow microspheres,¹⁷⁴ Fe₂O₃@TiO₂ core–shell spheres,²⁰⁴ V₂O₅ spheres and hollow spheres,^114,127,128 SnO₂ spheres and hollow spheres,^{106,109,206,264} SnO₂ multi-shell spheres,¹¹⁰ Li₂O–CuO–SnO₂ multi-deck cage spheres,²¹⁹ CoFe₂O₄ and other metal ferrite spheres,^29,77 core-in-double-shell NiCo₂O₄ particles,¹⁰⁵ multi-shell LiMn₂O₄ hollow microspheres,²⁸⁶ Cr₂O₃–C core–shell spheres,³¹ SnO₂@C spheres,⁴⁹ SnO₂/C composite hollow spheres^108,111 and Fe₃O₄–C composite spheres.²²⁰

4.4.2 Supercapacitors. Supercapacitors are energy storage devices that provide a large specific energy density, a high specific power density, fast charging capability and a long lifetime due to the excellent cycling stability. Metal oxides are promising materials because of their high theoretical capacitance, low cost and environmental friendliness. To enhance the rate of ion and electron transport in the electrodes and at the electrode/electrolyte interface, the structural aspects such as short diffusion path lengths, a high surface area and porosity to accommodate the strain from the high rate of insertion and extraction of ions are required.

The surface area of MnO₂ hollow spheres prepared by a double templating method was varied by changing the ratio of the F127 surfactant and SiO₂ spheres used as soft and hard templates, respectively.⁹⁷ The urchin-like hollow spheres showed the best capacitance of 266.6 F g⁻¹ within the potential range of 1 V at a current density of 0.1 A g⁻¹ due to their high surface area of 233 m² g⁻¹ relative to the other hollow structures. Co₃O₄ hollow spheres with thin shells of 130 nm, a moderate surface area of 60 m² g⁻¹ and mesopores centered at 7.8 nm, prepared hydrothermally using sucrose as a precursor for carbon sphere templates, performed well giving a specific capacitance of 470 F g⁻¹ at 1 A g⁻¹ with no obvious capacitance decrease observed over 1000 cycles.⁷⁵ On the other hand, bacteria-templated Co₃O₄ hollow spheres prepared by a one-pot mineralization method at room temperature featured a similar average pore size of 7.7 nm but a larger surface area of 149 m² g⁻¹ due to their fluffy-like surface.¹⁴ These particles showed a high capacitance of 214 F g⁻¹ at 2 A g⁻¹, a Coulombic efficiency averaging over 95% and excellent cycling stability that showed a capacitance retention of about 95% after 4000 cycles. Mn₂O₃ triple-shell hollow microspheres with thin porous shells and a large pore volume of 0.52 cm³ g⁻¹ but a moderate surface area of 37 m² g⁻¹ showed a record high specific capacitance up to 1651 F g⁻¹ at 0.5 A g⁻¹, an excellent rate capability of 1422 F g⁻¹ at 10 A g⁻¹ and a cycling stability retention of 92% after 2000 cycles.¹⁷⁵ These findings show the importance of porosity, surface texture and high surface area for the performance of metal oxide particles as supercapacitors.

However, the particle morphology may surpass the effect of surface area in some cases, as it was shown in the case of NiO spheres prepared by a hydrothermal method.⁴³ It was found that NiO spheres made of ultrathin nanowires performed much better than urchin-like spheres with nearly 100% capacity retention after 200 cycles at a current density of 10 A g⁻¹, despite the higher surface area of the latter (243 m² g⁻¹vs. 215 m² g⁻¹). The high capacity was attributed to the bimodal pore size distribution of the ultrathin nanowire-assembled spheres at 3.2 nm and 8.9 nm as opposed to the monomodal 4 nm pores in the urchin-like spheres as well as to their network-like surface texture, which assured efficient diffusion paths for OH⁻ ions and greatly enhanced intercalation of the electrolyte ions, ensuring sufficient Faradic reactions.

Doped or composite structures can also perform well as supercapacitors as it was shown for Fe₃O₄ doped MnO₂ microspheres,²¹ Ag₂O–MnO₂ composite spheres¹⁹ and MnO₂/C composite spheres.¹⁰⁰ Core–shell C@MnO₂ had a high specific capacitance of 583 F g⁻¹ at a current density of 1 A g⁻¹ in 0.1 M Na₂SO₄ electrolyte,²¹⁴ which was attributed to the inner graphitized carbon core coupled with porous interconnected MnO₂ nanorods for enhanced electrolyte accessibility, short ion diffusion length and charge transfer pathways. Ternary metal oxides with two different metal cations also exhibited high electrochemical performance as it was demonstrated for core-in-double-shell hollow NiCo₂O₄ spheres¹⁰⁵ with a complex interior, porous shells and consisting of small nanocrystalline particles. This material delivered high pseudocapacitance values of 1141, 1048, 965, 862 and 784 F g⁻¹ at current densities of 1, 2, 5, 10 and 15 A g⁻¹. Furthermore, after 4000 cycles at 5 A g⁻¹, 94.7% of the specific capacitance was retained, proving the high stability of the spheres.

Other examples of metal oxides that have been used as supercapacitors are NiO spheres and hollow spheres,^{41,124,148,244,287} La doped NiO spheres,⁴⁴ Co₃O₄ hollow spheres²⁴² and MnO₂ hollow spheres.⁹⁸

4.4.3 Dye sensitized solar cells (DSSCs). In a DSSC, light is absorbed by dye molecules attached to the surface of a wide band gap metal oxide (usually TiO₂) to generate free electrons, which are then injected into the conduction band of the metal oxide to be transported to an external circuit, thereby generating a current. Generally, the photoanodes in DSSCs are in the form of a film consisting of anatase TiO₂ nanocrystals of size ca. 20 nm.¹⁹² These films have high surface area, allowing for high dye adsorption and high energy conversion efficiency. However, these films have poor separability of electrons and holes due to various defects, grain boundaries and non-homogeneity. Moreover, the small particle size means that most of the visible light of longer wavelengths passes through without being absorbed.

To address this issue, an additional layer of metal oxide particles has been deposited to enhance the light to current capability of DSSCs. The engineered particles have increased dye loading, interconnected crystals to improve the current transport, improved crystallization to intensify electron–hole separation and more importantly, enhanced back-scattering of light of longer wavelength (visible and near infrared region) onto the primary layer of DSSCs, which augments absorption of light, as shown in Fig. 16. There are numerous reports on the use of TiO₂ particles of varying architecture in DSSCs such as spheres,^{184,190,192,195,217,245,288–290} hollow spheres,^227,248,291 and yolk–shell structures.²⁰⁸

Fig. 16 Scattering of long wavelength light (red light) due to the layer of larger particles on the top of TiO₂ film in a DSSC.

Macroporous TiO₂ spheres were prepared by an emulsion templating method using PS particles as templates for the macropores,¹⁸⁴ and were formed via calcination of the PS–TiO₂ composite spheres. These spheres had size in the range of 500–2000 nm, pore size in the range of 200–300 nm, a surface area of 76 m² g⁻¹ and were composed of crystals with sizes of 6–8 nm. The high surface area assured a high amount of dye loading, which was about 1.5 times higher than that adsorbed on a commercial nano-TiO₂, and the proper size of spheres improved the scattering of light in the wavelength range of 500–800 nm. Hence, the geometry of particles assured a significant improvement in light scattering, and increased dye loading as well as interconnected pores (fewer grain boundaries) enhanced light to current efficiency as compared to conventional TiO₂ materials.

TiO₂ spheres of size 2.1 μm consisting of nanorods were prepared by a hydrothermal method.²⁸⁹ The large surface area (64 m² g⁻¹) and size of the particles resulted in a high power conversion efficiency of 10.34%. The high performance was also due to the crystal size and structure, which increased the electron transport rate and slowed down the recombination rate of electrons and holes. On the other hand, TiO₂ mesoporous microspheres of size ca. 800 nm, created by an evaporation-driven self-assembly method,¹⁸⁶ possessed a large surface area of 112 m² g⁻¹, a large pore volume of 0.164 cm³ g⁻¹, and highly crystalline walls with [101] exposed facets. These attributes resulted in DSSCs with high energy conversion efficiency of up to 12.1%.

The wavelength of the scattered light is closely related to the size of particles. Xu et al.²⁰² showed that TiO₂ particles with a size of 380 nm exhibited strong resonance with light of wavelength 366 nm. However, Yu et al.²⁸⁸ studied the effect of light scattering by the particles with sizes in the range of 260–800 nm and found that particles with a size of 450 nm showed the highest scattering of light in the range 600–750 nm and therefore the best photon to current efficiency; this was attributed to the size uniformity and long range ordering when the particles were applied as a light scattering layer in DSSCs.

Other semiconductors such as SnO₂,^126,243 quintuple shell SnO₂¹⁷³ and Nb₂O₅³⁹ have also been used in DSSCs. However, the power conversion efficiencies did not reach those of TiO₂-based electrodes.

5. Conclusions and perspectives

We have reviewed the synthesis methods and potential applications of metal oxide spherical nanoparticles. These particles were engineered to have a high surface area to volume ratio, a high pore volume, and high mechanical strength, which endorse them for a variety of applications in the catalytic, environmental, biomedical and energy fields. They can show synergistic interactions with noble metals when used as supports, making them potentially attractive catalysts. Furthermore, for specific crystal size, the quantum effects can impart semiconducting or magnetic properties to the particles, which can be beneficial in the field of photocatalysis and separation. The most common methods currently used for the synthesis of metal oxide spheres are the high temperature hydrothermal and low temperature solution precipitation methods. However, these methods are time consuming and suitable for batch production only. Thus, the emerging methods such as aerosol and spray drying or spray pyrolysis methods are well suited for continuous large scale production, and can be used to synthesize porous spheres and shells, allowing high loading of noble metals. However, further studies are required toward controlling the uniformity and size of spheres and the thickness of shells. For multi-shell hollow spheres, the finely controlled synthesis of concentric hollow spheres is still a challenge and more research in this direction is needed. Further studies on the design and synthesis of metal oxide structures, especially core–shell and yolk–shell particles with cores and shells of tailored surface properties, porosity and chemical composition for advanced catalytic, biomedical and energy-related applications, are highly desirable. Since chemistry of metal oxide nanoparticles is a rapidly growing research area, new discoveries in the synthesis and applications of these materials are anticipated.

In perspective, synergetic interactions of metal oxide composites need to be further investigated to possibly avoid the use of precious metals as active agents in catalysis, and enhance the energy storage capacities of the materials used in the energy field. Multiple functionalization within the same structure could generate catalysts that can cope with tandem reactions, hence avoiding the use of two or more catalysts. Furthermore, more work on the synthesis of hollow metal oxide spheres in the nanometer regime needs to be carried out for use as nanoreactors in colloidal solutions. The new generation of porous non-silica metal oxide submicrospheres with precisely controlled functionality can be envisioned to be created by taking advantage of prosperous nanochemistry synthesis methods under mild conditions. A further development of the computationally assisted design and fabrication of complex porous non-silica metal oxide particles is needed for structure prediction and reduction of the number of synthesized samples. In the past decade, there has been a remarkable evolution in porous non-silica metal oxide submicrospheres with unique properties as platforms for emerging applications, which envisages a bright and exciting future for these unique materials.

Abbreviations

BET	Brunauer–Emmett–Teller
CS	Carbon spheres
CTAB	Cetyltrimethylammonium bromide
DEX	Dexamethasone
DMF	N,N-Dimethylformamide
EDA	Ethylenediamine
EG	Ethylene glycol
ERY	Erythromycin
GMA-co-EGDMA	Glycidyl methacrylate-co-ethylene glycol dimethacrylate
HMT	Hexamethylenetetramine
IBU	Ibuprofen
IUPAC	International Union of Pure and Applied Chemistry
MEA	Monoethanolamine
PEG	Poly-ethylene glycol
PEO-b-PS	Polyethylene oxide-block-polystyrene
PEO–PPO–PEO	Poly(ethylene oxide)–poly(propylene oxide)–poly(ethylene oxide)
PS	Polystyrene
PVP	Polyvinyl pyrrolidone
SDBS	Sodium dodecylbenzenesulfonate
TBOT	Titanium butoxide
TEA	Triethanolamine
TEOS	Tetraethyl orthosilicate
TFA	Trifluoroacetic acid
THF	Tetrahydrofuran

Acknowledgements

The authors would like to thank Prof. Shaobin Wang for fruitful discussions.

Notes and references

1. A. P. Alivisatos, Science, 1996, 271, 933–937.
2. M. Khairy, S. A. El-Safty, M. Ismael and H. Kawarada, Appl. Catal., B, 2012, 127, 1–10.
3. H. Li, Z. Bian, J. Zhu, D. Zhang, G. Li, Y. Huo, H. Li and Y. Lu, J. Am. Chem. Soc., 2007, 129, 8406–8407.
4. P. Zhang, A. Li and J. Gong, Particuology, 2015, 22, 13–23.
5. D. Chen and R. A. Caruso, Adv. Funct. Mater., 2013, 23, 1356–1374.
6. W. Wu, Z. Wu, T. Yu, C. Jiang and W.-S. Kim, Sci. Technol. Adv. Mater., 2015, 16 DOI:10.1088/1468-6996/16/2/023501.
7. J.-H. Lee, Sens. Actuators, B, 2009, 140, 319–336.
8. X. Lai, J. E. Halpert and D. Wang, Energy Environ. Sci., 2012, 5, 5604–5618.
9. Z. Wang and L. Zhou, Adv. Mater., 2012, 24, 1903–1911.
10. J. Hu, M. Chen, X. Fang and L. Wu, Chem. Soc. Rev., 2011, 40, 5472–5491.
11. J. Qi, X. Lai, J. Wang, H. Tang, H. Ren, Y. Yang, Q. Jin, L. Zhang, R. Yu, G. Ma, Z. Su, H. Zhao and D. Wang, Chem. Soc. Rev., 2015, 44, 6749–6773.
12. X. W. Lou, L. A. Archer and Z. Yang, Adv. Mater., 2008, 20, 3987–4019.
13. K. Byrappa and T. Adschiri, Prog. Cryst. Growth Charact. Mater., 2007, 53, 117–166.
14. H.-W. Shim, A.-H. Lim, J.-C. Kim, E. Jang, S.-D. Seo, G.-H. Lee, T. D. Kim and D.-W. Kim, Sci. Rep., 2013, 3, 2325,  DOI:10.1038/srep02325.
15. X. Zou, H. Fan, Y. Tian, M. Zhang and X. Yan, Dalton Trans., 2015, 44, 7811–7821.
16. H. Cao, A. Yang, H. Li, L. Wang, S. Li, J. Kong, X. Bao and R. Yang, Sens. Actuators, B, 2015, 214, 169–173.
17. T. Zhang and D. D. Sun, Chem. Eng. J., 2013, 225, 271–279.
18. Z. Yuan, Y. Wang and Y. Qian, RSC Adv., 2012, 2, 8602–8605.
19. Y. Dai, S. Tang, S. Vongehr and X. Meng, ACS Sustainable Chem. Eng., 2014, 2, 692–698.
20. T. Lin, L. Yu, M. Sun, G. Cheng, B. Lan and Z. Fu, Chem. Eng. J., 2016, 286, 114–121.
21. J. Zhu, S. Tang, H. Xie, Y. Dai and X. Meng, ACS Appl. Mater. Interfaces, 2014, 6, 17637–17646.
22. M.-G. Ma, J.-F. Zhu, R.-C. Sun and Y.-J. Zhu, Mater. Lett., 2010, 64, 1524–1527.
23. F. Qin, H. Zhao, G. Li, H. Yang, J. Li, R. Wang, Y. Liu, J. Hu, H. Sun and R. Chen, Nanoscale, 2014, 6, 5402–5409.
24. Y. Li, J. Liu, X. Huang and G. Li, Cryst. Growth Des., 2007, 7, 1350–1355.
25. D. Ma, S. Huang, W. Chen, S. Hu, F. Shi and K. Fan, J. Phys. Chem. C, 2009, 113, 4369–4374.
26. X. Liang, J. Xiao, B. Chen and Y. Li, Inorg. Chem., 2010, 49, 8188–8190.
27. C. Sun, S. Rajasekhara, Y. Chen and J. B. Goodenough, Chem. Commun., 2011, 47, 12852–12854.
28. W. Hao, S. Chen, Y. Cai, L. Zhang, Z. Li and S. Zhang, J. Mater. Chem. A, 2014, 2, 13801–13804.
29. S. Yoon, J. Appl. Electrochem., 2014, 44, 1069–1074.
30. H. Ma, Y. Xu, Z. Rong, X. Cheng, S. Gao, X. Zhang, H. Zhao and L. Huo, Sens. Actuators, B, 2012, 174, 325–331.
31. L.-Y. Jiang, S. Xin, X.-L. Wu, H. Li, Y.-G. Guo and L.-J. Wan, J. Mater. Chem., 2010, 20, 7565–7569.
32. S. Yang, S. Zhang, H. Wang, H. Yu, Y. Fang and F. Peng, RSC Adv., 2014, 4, 43024–43028.
33. R. Liu, J. Yin, W. Du, F. Gao, Y. Fan and Q. Lu, Eur. J. Inorg. Chem., 2013, 1358–1362.
34. G. Liu, Q. Deng, H. Wang, D. H. Ng, M. Kong, W. Cai and G. Wang, J. Mater. Chem., 2012, 22, 9704–9713.
35. C. Wang, Y. Cui and K. Tang, Nanoscale Res. Lett., 2013, 8, 1–4.
36. M. Muruganandham, R. Amutha, M. S. A. Wahed, B. Ahmmad, Y. Kuroda, R. P. Suri, J. J. Wu and M. E. Sillanpaa, J. Phys. Chem. C, 2011, 116, 44–53.
37. H. Yang, L. Liu, H. Liang, J. Wei and Y. Yang, CrystEngComm, 2011, 13, 5011–5016.
38. B. Tao, Y. Zhang, D. Han, Y. Li and Z. Yan, J. Mater. Chem. A, 2014, 2, 5455–5461.
39. X. Jin, C. Liu, J. Xu, Q. Wang and D. Chen, RSC Adv., 2014, 4, 35546–35553.
40. C. C. Li, J. Dou, L. Chen, J. Lin and H. C. Zeng, ChemCatChem, 2012, 4, 1675–1682.
41. A. K. Mondal, D. Su, Y. Wang, S. Chen, Q. Liu and G. Wang, J. Alloys Compd., 2014, 582, 522–527.
42. T. Zhu, J. S. Chen and X. W. Lou, J. Phys. Chem. C, 2012, 116, 6873–6878.
43. X. Li, S. Xiong, J. Li, J. Bai and Y. Qian, J. Mater. Chem., 2012, 22, 14276–14283.
44. D. Han, X. Jing, J. Wang, P. Yang, D. Song and J. Liu, J. Electroanal. Chem., 2012, 682, 37–44.
45. P.-P. Jin, X. Zou, L.-J. Zhou, J. Zhao, H. Chen, Y. Tian and G.-D. Li, Sens. Actuators, B, 2014, 204, 142–148.
46. Z. Li, Q. Zhao, W. Fan and J. Zhan, Nanoscale, 2011, 3, 1646–1652.
47. X. Wang, S. Qiu, C. He, G. Lu, W. Liu and J. Liu, RSC Adv., 2013, 3, 19002–19008.
48. X. Wang, S. Qiu, J. Liu, C. He, G. Lu and W. Liu, Eur. J. Inorg. Chem., 2014, 863–869.
49. X. W. Lou, J. S. Chen, P. Chen and L. A. Archer, Chem. Mater., 2009, 21, 2868–2874.
50. C. Niu, M. Huang, P. Wang, J. Meng, X. Liu, X. Wang, K. Zhao, Y. Yu, Y. Wu and C. Lin, Nano Res., 2015, 1–11.
51. C. Zhang, Z. Chen, Z. Guo and X. W. D. Lou, Energy Environ. Sci., 2013, 6, 974–978.
52. S. Jeon and K. Yong, J. Mater. Chem., 2010, 20, 10146–10151.
53. W. Zeng, C. Dong, B. Miao, H. Zhang, S. Xu, X. Ding and S. Hussain, Mater. Lett., 2014, 117, 41–44.
54. S. A. K. Leghari, S. Sajjad and J. Zhang, RSC Adv., 2013, 3, 15354–15361.
55. M. K. Devaraju, S. Yin and T. Sato, Inorg. Chem., 2011, 50, 4698–4704.
56. X. Liu, J. Zhang, L. Wang, T. Yang, X. Guo, S. Wu and S. Wang, J. Mater. Chem., 2011, 21, 349–356.
57. W. Wang, Y. Tian, X. Wang, H. He, Y. Xu, C. He and X. Li, J. Mater. Sci., 2013, 48, 3232–3238.
58. A. Lei, B. Qu, W. Zhou, Y. Wang, Q. Zhang and B. Zou, Mater. Lett., 2012, 66, 72–75.
59. W. Guo, T. Liu, J. Wang, W. Yu, R. Sun, Y. Chen, S. Hussain, X. Peng and Z. Wang, Ceram. Int., 2013, 39, 5919–5924.
60. Y. Lai, M. Meng and Y. Yu, Appl. Catal., B, 2010, 100, 491–501.
61. X. Fu, X. Yang, Z. Qiu, F. Zhao, J. Zhuang, A. He, L. Chen, C. Wu, X. Duan and C. Liang, CrystEngComm, 2013, 15, 3334–3340.
62. C. Wang, Y. Le and B. Cheng, Ceram. Int., 2014, 40, 10847–10856.
63. L. Nie, A. Meng, J. Yu and M. Jaroniec, Sci. Rep., 2013, 3, 3215,  DOI:10.1038/srep03215.
64. Z. Dong, T. Ye, Y. Zhao, J. Yu, F. Wang, L. Zhang, X. Wang and S. Guo, J. Mater. Chem., 2011, 21, 5978–5984.
65. S.-H. Hsieh, G.-J. Lee, C.-Y. Chen, J.-H. Chen, S.-H. Ma, T.-L. Horng, K.-H. Chen and J. J. Wu, Top. Catal., 2013, 56, 623–629.
66. Y. Huo, Y. Jin and Y. Zhang, J. Mol. Catal. A: Chem., 2010, 331, 15–20.
67. Z. Shen, J. Liu, F. Hu, S. Liu, N. Cao, Y. Sui, Q. Zeng and Y. Shen, CrystEngComm, 2014, 16, 3387–3394.
68. D. Zhang, Y. Qian, L. Shi, H. Mai, R. Gao, J. Zhang, W. Yu and W. Cao, Catal. Commun., 2012, 26, 164–168.
69. Y. Jiao, F. Wang, X. Ma, Q. Tang, K. Wang, Y. Guo and L. Yang, Microporous Mesoporous Mater., 2013, 176, 1–7.
70. Z. Yang, J. Wei, H. Yang, L. Liu, H. Liang and Y. Yang, Eur. J. Inorg. Chem., 2010, 3354–3359.
71. S. Yuan, Q. Zhang, B. Xu, Z. Jin, Y. Zhang, Y. Yang, M. Zhang and T. Ohno, RSC Adv., 2014, 4, 62255–62261.
72. G. Shen, H. Liu, Q. Wang, Z. Wang and Y. Chen, J. Nanopart. Res., 2012, 14, 1–8.
73. H. Jiu, Y. Sun, L. Zhang, C. Zhang, J. Zhang and J. Liu, Ceram. Int., 2014, 40, 3149–3154.
74. C. Chen, X. Fang, B. Wu, L. Huang and N. Zheng, ChemCatChem, 2012, 4, 1578–1586.
75. H. Du, L. Jiao, Q. Wang, J. Yang, L. Guo, Y. Si, Y. Wang and H. Yuan, Nano Res., 2013, 6, 87–98.
76. J. Zhao, Y. C. Zou, X. X. Zou, T. Y. Bai, Y. P. Liu, R. Q. Gao, D. J. Wang and G. D. Li, Nanoscale, 2014, 6, 7255–7262.
77. S. Li, A. Li, R. Zhang, Y. He, Y. Zhai and L. Xu, Nano Res., 2014, 7, 1116–1127.
78. C. Wang, Q. Li, F. Wang, G. Xia, R. Liu, D. Li, N. Li, J. S. Spendelow and G. Wu, ACS Appl. Mater. Interfaces, 2014, 6, 1243–1250.
79. X. Guan, L. Li, G. Li, Z. Fu, J. Zheng and T. Yan, J. Alloys Compd., 2011, 509, 3367–3374.
80. S. Wang, P. Li, H. Zhu and W. Tang, Powder Technol., 2012, 230, 48–53.
81. D. Li, Q. Qin, X. Duan, J. Yang, W. Guo and W. Zheng, ACS Appl. Mater. Interfaces, 2013, 5, 9095–9100.
82. Z. Wu, K. Yu, S. Zhang and Y. Xie, J. Phys. Chem. C, 2008, 112, 11307–11313.
83. B. Wang, J. S. Chen, H. B. Wu, Z. Wang and X. W. Lou, J. Am. Chem. Soc., 2011, 133, 17146–17148.
84. L. Wang, Z. Lou, J. Deng, R. Zhang and T. Zhang, ACS Appl. Mater. Interfaces, 2015, 7, 13098–13104.
85. S.-W. Cao, Y.-J. Zhu, M.-Y. Ma, L. Li and L. Zhang, J. Phys. Chem. C, 2008, 112, 1851–1856.
86. W. Li, Y. Deng, Z. Wu, X. Qian, J. Yang, Y. Wang, D. Gu, F. Zhang, B. Tu and D. Zhao, J. Am. Chem. Soc., 2011, 133, 15830–15833.
87. Z.-G. Wu, Y.-J. Zhong, J.-T. Li, X.-D. Guo, L. Huang, B. Zhong and S.-G. Sun, J. Mater. Chem. A, 2014, 2, 12361–12367.
88. L.-P. Zhu, L.-L. Wang, N.-C. Bing, C. Huang, L.-J. Wang and G.-H. Liao, ACS Appl. Mater. Interfaces, 2013, 5, 12478–12487.
89. F. X. Ma, H. Hu, H. B. Wu, C. Y. Xu, Z. Xu and L. Zhen, Adv. Mater., 2015, 27, 4097–4101.
90. B. Wang, H. Wu, L. Yu, R. Xu and T. T. Lim, Adv. Mater., 2012, 24, 1111–1116.
91. X. Li, C. Tang, M. Ai, L. Dong and Z. Xu, Chem. Mater., 2010, 22, 4879–4889.
92. Y. Teng, L. X. Song, A. Ponchel, Z. K. Yang and J. Xia, Adv. Mater., 2014, 26, 6238–6243.
93. T. Zhang, M. Chen, F. Gu, D. Han, Z. Wang and G. Guo, Integr. Ferroelectr., 2012, 138, 117–122.
94. S.-J. Kim, I.-S. Hwang, C. W. Na, I.-D. Kim, Y. C. Kang and J.-H. Lee, J. Mater. Chem., 2011, 21, 18560–18567.
95. L.-X. Li, D. Xu, X.-Q. Li, W.-C. Liu and Y. Jia, New J. Chem., 2014, 38, 5445–5452.
96. S. Yang, P. Huang, L. Peng, C. Cao, Y. Zhu, F. Wei, Y. Sun and W. Song, J. Mater. Chem. A, 2016, 4, 400–406.
97. J. Ma, Q. Cheng, V. Pavlinek, P. Saha and C. Li, New J. Chem., 2013, 37, 722–728.
98. S.-W. Bian, Y.-P. Zhao and C.-Y. Xian, Mater. Lett., 2013, 111, 75–77.
99. D. Li, X. Wu and Y. Chen, J. Phys. Chem. C, 2013, 117, 11040–11046.
100. G. Wang, H. Xu, L. Lu and H. Zhao, J. Mater. Chem. A, 2015, 3, 1127–1132.
101. H. Zhang, Y. Li, Z. Hong and M. Wei, Mater. Lett., 2012, 79, 148–151.
102. X. Zhang, X. Song, S. Gao, Y. Xu, X. Cheng, H. Zhao and L. Huo, J. Mater. Chem. A, 2013, 1, 6858–6864.
103. D. Xie, Q. Su, Z. Dong, J. Zhang and G. Du, CrystEngComm, 2013, 15, 8314–8319.
104. L. Chu, M. Li, Z. Wan, L. Ding, D. Song, S. Dou, J. Chen and Y. Wang, CrystEngComm, 2014, 16, 11096–11101.
105. L. Shen, L. Yu, X. Y. Yu, X. Zhang and X. W. D. Lou, Angew. Chem., Int. Ed., 2015, 54, 1868–1872.
106. X. M. Yin, C. C. Li, M. Zhang, Q. Y. Hao, S. Liu, L. B. Chen and T. H. Wang, J. Phys. Chem. C, 2010, 114, 8084–8088.
107. S. Liu, G. Huang, J. Yu, T. W. Ng, H. Y. Yip and P. K. Wong, ACS Appl. Mater. Interfaces, 2014, 6, 2407–2414.
108. X. W. Lou, D. Deng, J. Y. Lee and L. A. Archer, Chem. Mater., 2008, 20, 6562–6566.
109. S. Ding and X. W. D. Lou, Nanoscale, 2011, 3, 3586–3588.
110. H. X. Yang, J. F. Qian, Z. X. Chen, X. P. Ai and Y. L. Cao, J. Phys. Chem. C, 2007, 111, 14067–14071.
111. P. Wu, N. Du, H. Zhang, C. Zhai and D. Yang, ACS Appl. Mater. Interfaces, 2011, 3, 1946–1952.
112. Z. Zheng, B. Huang, X. Qin, X. Zhang and Y. Dai, J. Colloid Interface Sci., 2011, 358, 68–72.
113. M. Sharma, D. Das, A. Baruah, A. Jain and A. K. Ganguli, Langmuir, 2014, 30, 3199–3208.
114. E. Uchaker, N. Zhou, Y. Li and G. Cao, J. Phys. Chem. C, 2013, 117, 1621–1626.
115. D. Su, S. X. Dou and G. Wang, J. Mater. Chem. A, 2014, 2, 11185–11194.
116. H. B. Wu, A. Pan, H. H. Hng and X. W. D. Lou, Adv. Funct. Mater., 2013, 23, 5669–5674.
117. A. Pan, H. B. Wu, L. Yu and X. W. D. Lou, Angew. Chem., 2013, 125, 2282–2286.
118. J. Yu and X. Yu, Environ. Sci. Technol., 2008, 42, 4902–4907.
119. X. Ma, X. Zhang, L. Yang, K. Wang, K. Jiang, Z. Wei and Y. Guo, CrystEngComm, 2014, 16, 7933–7941.
120. Y. Sun, L. Zhang, J. Zhang, P. Chen, S. Xin, Z. Li and J. Liu, Ceram. Int., 2014, 40, 1599–1603.
121. H.-S. Roh, G. K. Choi, J.-S. An, C. M. Cho, D. H. Kim, I. J. Park, T. H. Noh, D.-W. Kim and K. S. Hong, Dalton Trans., 2011, 40, 6901–6905.
122. Z. Zhang, H. Chen, H. Che, Y. Wang and F. Su, Mater. Chem. Phys., 2013, 138, 593–600.
123. W. Luo, Y. Li, J. Dong, J. Wei, J. Xu, Y. Deng and D. Zhao, Angew. Chem., Int. Ed., 2013, 52, 10505–10510.
124. C. Yuan, X. Zhang, L. Su, B. Gao and L. Shen, J. Mater. Chem., 2009, 19, 5772–5777.
125. P. Manjula, R. Boppella and S. V. Manorama, ACS Appl. Mater. Interfaces, 2012, 4, 6252–6260.
126. G. Shang, J. Wu, M. Huang, J. Lin, Z. Lan, Y. Huang and L. Fan, J. Phys. Chem. C, 2012, 116, 20140–20145.
127. S. Wang, Z. Lu, D. Wang, C. Li, C. Chen and Y. Yin, J. Mater. Chem., 2011, 21, 6365–6369.
128. J. Shao, X. Li, Z. Wan, L. Zhang, Y. Ding, L. Zhang, Q. Qu and H. Zheng, ACS Appl. Mater. Interfaces, 2013, 5, 7671–7675.
129. J. Huang, X. Xu, C. Gu, G. Fu, W. Wang and J. Liu, Mater. Res. Bull., 2012, 47, 3224–3232.
130. S. Bai, K. Zhang, L. Wang, J. Sun, R. Luo, D. Li and A. Chen, J. Mater. Chem. A, 2014, 2, 7927–7934.
131. J. Huang, X. Xu, C. Gu, M. Yang, M. Yang and J. Liu, J. Mater. Chem., 2011, 21, 13283–13289.
132. P. Zhao, Y. Zhu, X. Yang, K. Fan, J. Shen and C. Li, RSC Adv., 2012, 2, 10592–10597.
133. F. Wang, X. Wang, D. Liu, J. Zhen, J. Li, Y. Wang and H. Zhang, ACS Appl. Mater. Interfaces, 2014, 6, 22216–22223.
134. H. Liu, G. Shao, W. Jia, Z. Zhang, Y. Zhang, J. Liang, X. Liu, H. Jia and B. Xu, CrystEngComm, 2013, 15, 3615–3622.
135. P. Ma, Y. Wu, Z. Fu and W. Wang, J. Alloys Compd., 2011, 509, 3576–3581.
136. J. He, J. Chen, L. Ren, Y. Wang, C. Teng, M. Hong, J. Zhao and B. Jiang, ACS Appl. Mater. Interfaces, 2014, 6, 2718–2725.
137. J. H. Pan, Y. Bai and Q. Wang, Langmuir, 2015, 31, 4566–4572.
138. W. Zhang, Z. X. Chi, W. X. Mao, R. W. Lv, A. M. Cao and L. J. Wan, Angew. Chem., 2014, 126, 12990–12994.
139. B. Bai, W. Guan, Z. Li and G. Li Puma, Mater. Res. Bull., 2011, 46, 26–31.
140. Y. Chen and J. Lu, J. Porous Mater., 2012, 19, 289–294.
141. C.-A. Wang, S. Li and L. An, Chem. Commun., 2013, 49, 7427–7429.
142. N. Kang, J. H. Park, M. Jin, N. Park, S. M. Lee, H. J. Kim, J. M. Kim and S. U. Son, J. Am. Chem. Soc., 2013, 135, 19115–19118.
143. Y. Qin, F. Zhang, Y. Chen, Y. Zhou, J. Li, A. Zhu, Y. Luo, Y. Tian and J. Yang, J. Phys. Chem. C, 2012, 116, 11994–12000.
144. C. Deng, H. Hu, X. Ge, C. Han, D. Zhao and G. Shao, Ultrason. Sonochem., 2011, 18, 932–937.
145. G. Tian, Z. Gu, X. Liu, L. Zhou, W. Yin, L. Yan, S. Jin, W. Ren, G. Xing and S. Li, J. Phys. Chem. C, 2011, 115, 23790–23796.
146. W. J. Tseng, T. T. Tseng, H. M. Wu, Y. C. Her and T. J. Yang, J. Am. Ceram. Soc., 2013, 96, 719–725.
147. Y. Liu, Z. Chen, C.-H. Shek, C.-M. L. Wu and J. K.-L. Lai, ACS Appl. Mater. Interfaces, 2014, 6, 9776–9784.
148. S. Ding, T. Zhu, J. S. Chen, Z. Wang, C. Yuan and X. W. D. Lou, J. Mater. Chem., 2011, 21, 6602–6606.
149. D. Xie, W. Yuan, Z. Dong, Q. Su, J. Zhang and G. Du, Electrochim. Acta, 2013, 92, 87–92.
150. L. Shi and H. Lin, Langmuir, 2010, 26, 18718–18722.
151. F. Gyger, M. Hubner, C. Feldmann, N. Barsan and U. Weimar, Chem. Mater., 2010, 22, 4821–4827.
152. S. Ding, J. S. Chen, G. Qi, X. Duan, Z. Wang, E. P. Giannelis, L. A. Archer and X. W. Lou, J. Am. Chem. Soc., 2010, 133, 21–23.
153. G. Xi, Y. Yan, Q. Ma, J. Li, H. Yang, X. Lu and C. Wang, Chem. – Eur. J., 2012, 18, 13949–13953.
154. Y. Liu, Q. Li, S. Gao and J. Shang, CrystEngComm, 2014, 16, 7493–7501.
155. G. Jia, H. You, Y. Song, Y. Huang, M. Yang and H. Zhang, Inorg. Chem., 2010, 49, 7721–7725.
156. Z. Xu, Y. Gao, T. Liu, L. Wang, S. Bian and J. Lin, J. Mater. Chem., 2012, 22, 21695–21703.
157. H. Jiu, Y. Fu, L. Zhang, Y. Sun, Y. Wang and T. Han, Micro Nano Lett., 2012, 7, 947–950.
158. F. Meng, J. Yin, Y.-Q. Duan, Z.-H. Yuan and L.-J. Bie, Sens. Actuators, B, 2011, 156, 703–708.
159. Y. Wei, Y. Huang, J. Wu, M. Wang, C. Guo, D. Qiang, S. Yin and T. Sato, J. Hazard. Mater., 2013, 248, 202–210.
160. H. Yao, D. Jia and H. Zhang, Ceram. Int., 2015, 41, 1531–1534.
161. Y. Xia and R. Mokaya, J. Mater. Chem., 2005, 15, 3126–3131.
162. X. Sun, J. Liu and Y. Li, Chem. – Eur. J., 2006, 12, 2039–2047.
163. M.-M. Titirici, M. Antonietti and A. Thomas, Chem. Mater., 2006, 18, 3808–3812.
164. P. M. Arnal, C. Weidenthaler and F. Schüth, Chem. Mater., 2006, 18, 2733–2739.
165. D. G. Shchukin and R. A. Caruso, Chem. Mater., 2004, 16, 2287–2292.
166. H. Ren, J. J. Sun, R. B. Yu, M. Yang, L. Gu, P. R. Liu, H. J. Zhao, D. Kisailus and D. Wang, Chem. Sci., 2016, 7, 793–798.
167. X. Zeng, J. Yang, L. Shi, L. Li and M. Gao, Nanoscale Res. Lett., 2014, 9, 1–10.
168. G. Zhang and X. W. D. Lou, Angew. Chem., Int. Ed., 2014, 53, 9041–9044.
169. X. Lai, J. Li, B. A. Korgel, Z. Dong, Z. Li, F. Su, J. Du and D. Wang, Angew. Chem., Int. Ed., 2011, 50, 2738–2741.
170. Z. H. Dong, X. Y. Lai, J. E. Halpert, N. L. Yang, L. X. Yi, J. Zhai, D. Wang, Z. Y. Tang and L. Jiang, Adv. Mater., 2012, 24, 1046–1049.
171. J. Y. Wang, N. L. Yang, H. J. Tang, Z. H. Dong, Q. Jin, M. Yang, D. Kisailus, H. J. Zhao, Z. Y. Tang and D. Wang, Angew. Chem., Int. Ed., 2013, 52, 6417–6420.
172. S. M. Xu, C. M. Hessel, H. Ren, R. B. Yu, Q. Jin, M. Yang, H. J. Zhao and D. Wang, Energy Environ. Sci., 2014, 7, 632–637.
173. Z. Dong, H. Ren, C. M. Hessel, J. Wang, R. Yu, Q. Jin, M. Yang, Z. Hu, Y. Chen, Z. Tang, H. Zhao and D. Wang, Adv. Mater., 2014, 26, 905–909.
174. H. Ren, R. Yu, J. Wang, Q. Jin, M. Yang, D. Mao, D. Kisailus, H. Zhao and D. Wang, Nano Lett., 2014, 14, 6679–6684.
175. J. Y. Wang, H. J. Tang, H. Ren, R. B. Yu, J. Qi, D. Mao, H. J. Zhao and D. Wang, Adv. Sci., 2014, 1 DOI:10.1002/advs.201400011.
176. J. Wang, H. Tang, L. Zhang, H. Ren, R. Yu, Q. Jin, J. Qi, D. Mao, M. Yang, Y. Wang, P. Liu, Y. Zhang, Y. Wen, L. Gu, G. Ma, Z. Su, Z. Tang, H. Zhao and D. Wang, Nat. Energy, 2016, 1, 16050.
177. J. B. Joo, I. Lee, M. Dahl, G. D. Moon, F. Zaera and Y. Yin, Adv. Funct. Mater., 2013, 23, 4246–4254.
178. J. Chen, D. W. Wang, J. Qi, G. D. Li, F. Y. Zheng, S. X. Li, H. J. Zhao and Z. Y. Tang, Small, 2015, 11, 420–425.
179. H. M. Abdelaal and B. Harbrecht, C. R. Chim., 2015, 18, 379–384.
180. L. Zong, P. Xu, Y. Ding, K. Zhao, Z. Wang, X. Yan, R. Yu, J. Chen and X. Xing, Small, 2015, 11, 2768–2773.
181. X. Li, F. Chen, X. Lu, C. Ni, X. Zhao and Z. Chen, J. Porous Mater., 2010, 17, 297–303.
182. H. Liu, J. B. Joo, M. Dahl, L. Fu, Z. Zeng and Y. Yin, Energy Environ. Sci., 2015, 8, 286–296.
183. G. Yang, P. Hu, Y. Cao, F. Yuan and R. Xu, Nanoscale Res. Lett., 2010, 5, 1437–1441.
184. G. Veerappan, D.-W. Jung, J. Kwon, J. M. Choi, N. Heo, G.-R. Yi and J. H. Park, Langmuir, 2014, 30, 3010–3018.
185. G. Zhang, X. Shen and Y. Yang, J. Phys. Chem. C, 2011, 115, 7145–7152.
186. Y. Liu, R. Che, G. Chen, J. Fan, Z. Sun, Z. Wu, M. Wang, B. Li, J. Wei and Y. Wei, Sci. Adv., 2015, 1, e1500166.
187. D. Mao, J. X. Yao, X. Y. Lai, M. Yang, J. A. Du and D. Wang, Small, 2011, 7, 578–582.
188. J. S. Chen, Y. L. Tan, C. M. Li, Y. L. Cheah, D. Luan, S. Madhavi, F. Y. C. Boey, L. A. Archer and X. W. Lou, J. Am. Chem. Soc., 2010, 132, 6124–6130.
189. J. H. Pan, Z. Cai, Y. Yu and X. Zhao, J. Mater. Chem., 2011, 21, 11430–11438.
190. W.-G. Yang, F.-R. Wan, Q.-W. Chen, J.-J. Li and D.-S. Xu, J. Mater. Chem., 2010, 20, 2870–2876.
191. Z. Lin, M. Zheng, B. Zhao, G. Wang, L. Pu and Y. Shi, J. Solid State Electrochem., 2014, 18, 1673–1681.
192. D. Chen, F. Huang, Y. B. Cheng and R. A. Caruso, Adv. Mater., 2009, 21, 2206–2210.
193. H. Ming, Z. Ma, H. Huang, S. Lian, H. Li, X. He, H. Yu, K. Pan, Y. Liu and Z. Kang, Chem. Commun., 2011, 47, 8025–8027.
194. K. Del Ángel-Sánchez, O. Vázquez-Cuchillo, M. Salazar-Villanueva, J. Sánchez-Ramirez, A. Cruz-López and A. Aguilar-Elguezabal, J. Sol–Gel Sci. Technol., 2011, 58, 360–365.
195. H.-E. Wang, L.-X. Zheng, C.-P. Liu, Y.-K. Liu, C.-Y. Luan, H. Cheng, Y. Y. Li, L. Martinu, J. A. Zapien and I. Bello, J. Phys. Chem. C, 2011, 115, 10419–10425.
196. D. Chen, L. Cao, F. Huang, P. Imperia, Y.-B. Cheng and R. A. Caruso, J. Am. Chem. Soc., 2010, 132, 4438–4444.
197. Y. Cui, L. Liu, B. Li, X. Zhou and N. Xu, J. Phys. Chem. C, 2010, 114, 2434–2439.
198. C. Wang, J. Chen, X. Zhou, W. Li, Y. Liu, Q. Yue, Z. Xue, Y. Li, A. A. Elzatahry and Y. Deng, Nano Res., 2014, 1–8.
199. I. Lee, J. B. Joo, Y. Yin and F. Zaera, Angew. Chem., 2011, 123, 10390–10393.
200. J. B. Joo, Q. Zhang, M. Dahl, I. Lee, J. Goebl, F. Zaera and Y. Yin, Energy Environ. Sci., 2012, 5, 6321–6327.
201. X. Cheng, M. Chen, L. Wu and G. Gu, Langmuir, 2006, 22, 3858–3863.
202. H. Xu, X. Chen, S. Ouyang, T. Kako and J. Ye, J. Phys. Chem. C, 2012, 116, 3833–3839.
203. L. Cao, D. Chen and R. A. Caruso, Angew. Chem., Int. Ed., 2013, 52, 10986–10991.
204. W. Li, J. Yang, Z. Wu, J. Wang, B. Li, S. Feng, Y. Deng, F. Zhang and D. Zhao, J. Am. Chem. Soc., 2012, 134, 11864–11867.
205. B. Li, Y. Xie, M. Jing, G. Rong, Y. Tang and G. Zhang, Langmuir, 2006, 22, 9380–9385.
206. P. Gurunathan, P. M. Ette and K. Ramesha, ACS Appl. Mater. Interfaces, 2014, 6, 16556–16564.
207. H. Chun Zeng, Curr. Nanosci., 2007, 3, 177–181.
208. W. Q. Fang, X. H. Yang, H. Zhu, Z. Li, H. Zhao, X. Yao and H. G. Yang, J. Mater. Chem., 2012, 22, 22082–22089.
209. D. Zhang, J. Zhu, N. Zhang, T. Liu, L. Chen, X. Liu, R. Ma, H. Zhang and G. Qiu, Sci. Rep., 2015, 5, 8737,  DOI:10.1038/srep08737.
210. Y. Yin, R. M. Rioux, C. K. Erdonmez, S. Hughes, G. A. Somorjai and A. P. Alivisatos, Science, 2004, 304, 711–714.
211. J. Liu, S. Z. Qiao, J. S. Chen, X. W. D. Lou, X. Xing and G. Q. M. Lu, Chem. Commun., 2011, 47, 12578–12591.
212. Y.-H. Choi, D.-H. Kim, H. S. Han, S. Shin, S.-H. Hong and K. S. Hong, Langmuir, 2014, 30, 700–709.
213. M. Kong, W. Zhang, Z. Yang, S. Weng and Z. Chen, Appl. Surf. Sci., 2011, 258, 1317–1321.
214. S. Li and C.-A. Wang, J. Colloid Interface Sci., 2015, 438, 61–67.
215. J. H. Pan, X. Z. Wang, Q. Huang, C. Shen, Z. Y. Koh, Q. Wang, A. Engel and D. W. Bahnemann, Adv. Funct. Mater., 2014, 24, 95–104.
216. W. Yu, Q. Ma, C. Wang, X. Dong, J. Wang and G. Liu, Mater. Express, 2014, 4, 435–440.
217. D. Hwang, H. Lee, S.-Y. Jang, S. M. Jo, D. Kim, Y. Seo and D. Y. Kim, ACS Appl. Mater. Interfaces, 2011, 3, 2719–2725.
218. B. Liu, K. Nakata, M. Sakai, H. Saito, T. Ochiai, T. Murakami, K. Takagi and A. Fujishima, Langmuir, 2011, 27, 8500–8508.
219. Y. Yu, C. H. Chen and Y. Shi, Adv. Mater., 2007, 19, 993–997.
220. X. Jia, Z. Chen, X. Cui, Y. Peng, X. Wang, G. Wang, F. Wei and Y. Lu, ACS Nano, 2012, 6, 9911–9919.
221. D. Lehr, D. Großmann, W. Grünert and S. Polarz, Nanoscale, 2014, 6, 1698–1706.
222. G. Jian, L. Liu and M. R. Zachariah, Adv. Funct. Mater., 2013, 23, 1341–1346.
223. Z. Jin, F. Wang, J. Wang, J. C. Yu and J. Wang, Adv. Funct. Mater., 2013, 23, 2137–2144.
224. G. Jian, Y. Xu, L.-C. Lai, C. Wang and M. R. Zachariah, J. Mater. Chem. A, 2014, 2, 4627–4632.
225. S. Dilger, M. Wessig, M. R. Wagner, J. S. Reparaz, C. M. Sotomayor Torres, L. Qijun, T. Dekorsy and S. Polarz, Cryst. Growth Des., 2014, 14, 4593–4601.
226. Y. H. Cho, Y. C. Kang and J.-H. Lee, Sens. Actuators, B, 2013, 176, 971–977.
227. J. Huo, Y. Hu, H. Jiang, W. Huang, Y. Li, W. Shao and C. Li, Ind. Eng. Chem. Res., 2013, 52, 11029–11035.
228. A. B. D. Nandiyanto, O. Arutanti, T. Ogi, F. Iskandar, T. O. Kim and K. Okuyama, Chem. Eng. Sci., 2013, 101, 523–532.
229. Y. Huang, Z. Ai, W. Ho, M. Chen and S. Lee, J. Phys. Chem. C, 2010, 114, 6342–6349.
230. J. W. Overcash and K. S. Suslick, Chem. Mater., 2015, 27, 3564–3567.
231. Z. Padashbarmchi, A. H. Hamidian, H. Zhang, L. Zhou, N. Khorasani, M. Kazemzad and C. Yu, RSC Adv., 2015, 5, 10304–10309.
232. Y. J. Hong, M. Y. Son, B. K. Park and Y. C. Kang, Small, 2013, 9, 2224–2227.
233. Y. J. Hong, M. Y. Son and Y. C. Kang, Adv. Mater., 2013, 25, 2279–2283.
234. Y. N. Ko, S. B. Park, K. Y. Jung and Y. C. Kang, Nano Lett., 2013, 13, 5462–5466.
235. S. H. Choi, Y. J. Hong and Y. C. Kang, Nanoscale, 2013, 5, 7867–7871.
236. S. H. Choi, J.-K. Lee and Y. C. Kang, Nanoscale, 2014, 6, 12421–12425.
237. S. H. Choi and Y. C. Kang, Small, 2014, 10, 474–478.
238. C. Wang, M. Zhu, H. Liu, Y. Cui and Y. Chen, RSC Adv., 2014, 4, 24176–24182.
239. B. Fang, C. Zhang, G. Wang, M. Wang and Y. Ji, Sens. Actuators, B, 2011, 155, 304–310.
240. P. Si, P. Chen and D.-H. Kim, J. Mater. Chem. B, 2013, 1, 2696–2700.
241. M. Aghazadeh, A.-A. M. Barmi and H. M. Shiri, Russ. J. Electrochem., 2013, 49, 344–353.
242. X.-H. Xia, J.-P. Tu, X.-L. Wang, C.-D. Gu and X.-B. Zhao, Chem. Commun., 2011, 47, 5786–5788.
243. M. A. Hossain, G. Yang, M. Parameswaran, J. R. Jennings and Q. Wang, J. Phys. Chem. C, 2010, 114, 21878–21884.
244. C.-Y. Cao, W. Guo, Z.-M. Cui, W.-G. Song and W. Cai, J. Mater. Chem., 2011, 21, 3204–3209.
245. P. Chen, J.-D. Peng, C.-H. Liao, P.-S. Shen and P.-L. Kuo, J. Nanopart. Res., 2013, 15, 1–11.
246. H. Wang, A. Pyatenko, K. Kawaguchi, X. Li, Z. Swiatkowska-Warkocka and N. Koshizaki, Angew. Chem., 2010, 122, 6505–6508.
247. X. Li, Y. Shimizu, A. Pyatenko, H. Wang and N. Koshizaki, Nanotechnology, 2012, 23, 115602.
248. H. Wang, M. Miyauchi, Y. Ishikawa, A. Pyatenko, N. Koshizaki, Y. Li, L. Li, X. Li, Y. Bando and D. Golberg, J. Am. Chem. Soc., 2011, 133, 19102–19109.
249. X.-X. Zou, G.-D. Li, Y.-N. Wang, J. Zhao, C. Yan, M.-Y. Guo, L. Li and J.-S. Chen, Chem. Commun., 2011, 47, 1066–1068.
250. Z. Ju, C. Guo, Y. Qian, B. Tang and S. Xiong, Nanoscale, 2014, 6, 3268–3273.
251. Q. Jiao, M. Fu, C. You, Y. Zhao and H. Li, Inorg. Chem., 2012, 51, 11513–11520.
252. R. D. Hancock and A. E. Martell, Chem. Rev., 1989, 89, 1875–1914.
253. J. Yang, C. Lin, Z. Wang and J. Lin, Inorg. Chem., 2006, 45, 8973–8979.
254. B. Liu, L.-M. Liu, X.-F. Lang, H.-Y. Wang, X. W. D. Lou and E. S. Aydil, Energy Environ. Sci., 2014, 7, 2592–2597.
255. C. Guo, X. Wu, M. Yan, Q. Dong, S. Yin, T. Sato and S. Liu, Nanoscale, 2013, 5, 8184–8191.
256. X. Lü, F. Huang, J. Wu, S. Ding and F. Xu, ACS Appl. Mater. Interfaces, 2011, 3, 566–572.
257. S. Shang, X. Jiao and D. Chen, ACS Appl. Mater. Interfaces, 2012, 4, 860–865.
258. M. Pelaez, N. T. Nolan, S. C. Pillai, M. K. Seery, P. Falaras, A. G. Kontos, P. S. Dunlop, J. W. Hamilton, J. A. Byrne and K. O'Shea, Appl. Catal., B, 2012, 125, 331–349.
259. G. Zhu, J. Lian, M. Hojamberdiev and W. Que, J. Cluster Sci., 2013, 24, 829–841.
260. Y. Jia, X.-Y. Yu, T. Luo, J.-H. Liu and X.-J. Huang, RSC Adv., 2012, 2, 10251–10254.
261. Q. Zhu, J.-G. Li, X. Li, X. Sun and Y. Sakka, Sci. Technol. Adv. Mater., 2011, 12, 055001.
262. J. Zhang, X. Liu, S. Wu, M. Xu, X. Guo and S. Wang, J. Mater. Chem., 2010, 20, 6453–6459.
263. G. Han, Q. Lu, G. Liu, X. Ye, S. Lin, Y. Song, B. Liu, X. Yang and G. Li, J. Mater. Sci.: Mater. Electron., 2012, 23, 1616–1620.
264. H. Wang, Q. Liang, W. Wang, Y. An, J. Li and L. Guo, Cryst. Growth Des., 2011, 11, 2942–2947.
265. N. Sutradhar, A. Sinhamahapatra, S. K. Pahari, P. Pal, H. C. Bajaj, I. Mukhopadhyay and A. B. Panda, J. Phys. Chem. C, 2011, 115, 12308–12316.
266. P. Yang, S. Gai, Y. Liu, W. Wang, C. Li and J. Lin, Inorg. Chem., 2011, 50, 2182–2190.
267. R. Lv, P. Yang, F. He, S. Gai, G. Yang and J. Lin, Chem. Mater., 2015, 27, 483–496.
268. X. Wang, D. Chen, L. Cao, Y. Li, B. J. Boyd and R. A. Caruso, ACS Appl. Mater. Interfaces, 2013, 5, 10926–10932.
269. Y. A. Nor, L. Zhou, A. K. Meka, C. Xu, Y. Niu, H. Zhang, N. Mitter, D. Mahony and C. Yu, Adv. Funct. Mater., 2016, 26, 5408–5418.
270. X. Wu, K. Li and H. Wang, J. Hazard. Mater., 2010, 174, 573–580.
271. D. Xiang, F. Qu, X. Chen, Z. Yu, L. Cui, X. Zhang, J. Jiang and H. Lin, J. Sol-Gel Sci. Technol., 2014, 69, 370–377.
272. Q. Wang, S. Yu, Z. Tan, R. Zhang, Z. Li, X. Gao, B. Shen and H. Su, CrystEngComm, 2015, 17, 671–677.
273. Y. Hou, L. Wu, X. Wang, Z. Ding, Z. Li and X. Fu, J. Catal., 2007, 250, 12–18.
274. G. Li, H. Zhang, J. Lan, J. Li, Q. Chen, J. Liu and G. Jiang, Dalton Trans., 2013, 42, 8541–8544.
275. B. Liu, K. Nakata, M. Sakai, H. Saito, T. Ochiai, T. Murakami, K. Takagi and A. Fujishima, Catal. Sci. Technol., 2012, 2, 1933–1939.
276. H. Bai, Z. Liu, L. Liu and D. D. Sun, Chem. – Eur. J., 2013, 19, 3061–3070.
277. S. Ding, X. Yin, X. Lü, Y. Wang, F. Huang and D. Wan, ACS Appl. Mater. Interfaces, 2011, 4, 306–311.
278. H. Zhang, G. Du, W. Lu, L. Cheng, X. Zhu and Z. Jiao, CrystEngComm, 2012, 14, 3793–3801.
279. B. Chi, L. Zhao and T. Jin, J. Phys. Chem. C, 2007, 111, 6189–6193.
280. J.-M. Tarascon and M. Armand, Nature, 2001, 414, 359–367.
281. S. Yoon and A. Manthiram, J. Phys. Chem. C, 2011, 115, 9410–9416.
282. S. Yoon, C. A. Bridges, R. R. Unocic and M. P. Paranthaman, J. Mater. Sci., 2013, 48, 5125–5131.
283. W. Guo, W. Sun and Y. Wang, ACS Nano, 2015, 9, 11462–11471.
284. X. Yan, X. Tong, J. Wang, C. Gong, M. Zhang and L. Liang, Mater. Lett., 2013, 95, 1–4.
285. S. Ding, J. S. Chen, Z. Wang, Y. L. Cheah, S. Madhavi, X. Hu and X. W. Lou, J. Mater. Chem., 2011, 21, 1677–1680.
286. F. Wang, J. Y. Wang, H. Ren, H. J. Tang, R. B. Yu and D. Wang, Inorg. Chem. Front., 2016, 3, 365–369.
287. W. Yu, X. Jiang, S. Ding and B. Q. Li, J. Power Sources, 2014, 256, 440–448.
288. I. G. Yu, Y. J. Kim, H. J. Kim, C. Lee and W. I. Lee, J. Mater. Chem., 2011, 21, 532–538.
289. J.-Y. Liao, B.-X. Lei, D.-B. Kuang and C.-Y. Su, Energy Environ. Sci., 2011, 4, 4079–4085.
290. P. Cheng, S. Du, Y. Cai, F. Liu, P. Sun, J. Zheng and G. Lu, J. Phys. Chem. C, 2013, 117, 24150–24156.
291. S. Dadgostar, F. Tajabadi and N. Taghavinia, ACS Appl. Mater. Interfaces, 2012, 4, 2964–2968.

---