Titanium Oxide Nanorods Extracted From Ilmenite Sands

Titanium Oxide Nanorods Extracted From Ilmenite Sands

Jun Yu,Ying Chen,*,?and Alexey M.Glushenkov?

Department of Electronic Materials and Engineering,Research School of Physical Sciences and

Engineering,the Australian National Uni V ersity,Canberra,ACT0200,Australia

Recei V ed October9,2008;Re V ised Manuscript Recei V ed No V ember24,2008

ABSTRACT:One dimensional titanium oxides(TiO2)nanorods and nanowires have substantial applications in photocatalytic, nanoelectronic,and photoelectrochemical areas.These applications require large quantities of materials and a production technique suitable for future industry fabrication.We demonstrate here a new method for mass production of TiO2nanorods from mineral ilmenite sands(FeTiO3).In this process,powder mixtures of ilmenite and activated carbon were?rst ball milled;the milled samples were then heated twice at two different temperatures.First high-temperature annealing produced metastable titanium oxide phases, and subsequent second low-temperature annealing in N2-5%H2activates the growth of rutile nanorods.This solid-state growth process allows large-quantity production of rutile nanorods.

Introduction

Titanium oxides(TiO2)nanorods and nanowires have sub-

stantial applications in photocatalytic,nanoelectronic,and

photoelectrochemical areas.1These applications require a mass

quantity of materials,but most current production methods can

only produce a small amount of samples,suitable for laboratory

research purposes only.Titanium oxide nanorods and nanowires

have been produced using wet chemistry methods including

electrodeposition,2sol-gel electrophoresis,3,4hydrothermal,5

and solvothermal methods.6Physical thermal deposition7-9and

metal-organic chemical vapor deposition10methods produce

a thin layer of samples on a substrate surface.For example,

Wu et al.reported catalyst-assisted growth of TiO2nanowires

on Ti-coated silicon substrates.7,11Amin et al.described the

growth of TiO2nanowires and nanoribbons by using thermal

annealing of nickel coated TiO powders at760Torr and

850-920°C in argon(Ar).12There is no report of a mass

production technique,which could be scaled up in industries.

Here we demonstrate a new method for mass production of TiO2

nanorods from mineral ilmenite sands.Ilmenite is a naturally

occurring iron titanate(nominally FeTiO3)and thus signi?cantly

reduces the production cost.In addition,FeTiO3contains both

titanium oxide and catalytic iron(Fe)required for nanorod

growth.In the new process,high-energy ball milling was used

to assist the reduction of ilmenite into an intermediate structure containing metastable titanium oxide phases by a high-temper-ature annealing and the subsequent low-temperature annealing activates the growth of rutile nanorods.This solid-state growth process allows large-quantity production of rutile nanorods.

Experimental Section

The starting ilmenite used in this study is a naturally occurring iron titanate(FeTiO3)and abundant in nature.High-purity ilmenite(99%) was provided by Consolidated Rutile Limited located in Australia. Chemical composition(wt.%)of the ilmenite are TiO2(dry basis)49.6, iron(total)35.1,FeO32.8,Fe2O313.7,Al2O30.47,Cr2O30.25,SiO2 0.45.Several grams of the mixture of ilmenite(FeTiO3)and active carbon(weight ratio of4:1)were milled in a Fritsch planetary ball mill with10steel balls(diameter1cm)for50h in vacuum atmosphere at room temperature.Isothermal annealing was conducted in a horizontal tube furnace at different temperatures(700-1200°C)and in different atmospheres(Ar or N2-5%H2)at a?ow rate of100mL/min.

Milled and annealed samples were characterized by using an X-ray diffraction(XRD)spectrometer with a cobalt K R radiation(λ)0.1789 nm).Scanning electron microscopy(SEM)was conducted using Hitachi S4500and S4300instruments.X-ray energy dispersive spectroscopy (EDS)system was used for chemical composition analysis.Transmis-sion electron microscopy(TEM)analysis was carried out using a Philips CM300microscope operating at300kV.For TEM study,a powder sample was ultrasonicated in ethanol for a few minutes to form a diluted suspension solution,and a few drops of the suspension was dripped on a copper grid with carbon?lm.Backscattered electron imaging (BSE)using in-lens Energy Selective Backscatter(EsB)detector and X-ray intensity distribution maps were conducted with a Zeiss ULTRA plus SEM microscope.Powder sample was coated with carbon to increase conductivity.Shimadzu TA50instrument was employed to conduct thermogravimetric analysis(TGA).

Results and Discussion

The SEM image in Figure1a shows large grains of starting ilmenite sands in the range of100μm.Mixture of the ilmenite

*Corresponding author.E-mail:ying.chen@https://www.sodocs.net/doc/dc18766902.html,.au.

?Present address:Institute for Technology Research and Innovation,Deakin University,Victoria3217,

Australia.

Figure1.(a)SEM image of ilmenite sands;(b)SEM image of ball

milled ilmenite and activated carbon powder mixture;(c)XRD pattern

of the milled sample.

CRYSTAL

GROWTH

&DESIGN

2009

VOL.9,NO.2

1240–1244

10.1021/cg801125w CCC:$40.75 2009American Chemical Society

Published on Web

01/05/2009

powder and activated carbon at weight ratio of 4:1was ?rst ball milled in a high-energy planetary ball mill for 50h;both FeTiO 3and C have been reduced down to small particles of about 100nm as revealed by the SEM image in Figure 1b.The X-ray diffraction (XRD)pattern of the milled sample in Figure 1c displays a full set of the diffraction peaks of FeTiO 3phase (PDF 29-733).The relatively broadened peak shapes are caused by the crystal size reduction induced by intensive ball milling treatment.13The broad,weak peak around 21degrees is associated with the amorphous structure of activated carbon.The SEM image in Figure 1b also shows that ?ne particles of FeTiO 3and C were homogenously mixed and form large aggregates indicating close contacts between these two materials,which are bene?cial to reduction reactions in the subsequent annealing.14

The ball milled powder sample was then annealed at 1200°C for 8h in Ar -5%H 2gases ?owing at 100mL/min to induce carborthermic reductions.The XRD diffraction pattern in Figure 2a was recorded from the heated sample,and it reveals the presence of monoclinic Ti 3O 5,Ti 2O 3and R -Fe phases,indicating the partial reduction of FeTiO 3by C and H 2.

The BSE image of the heated sample in Figure 2b shows a mixture of large gray Ti containing particles and small bright Fe dominant particles.The different contrasts are due to different atomic masses of Ti and Fe.EDS mapping was used to examine chemical compositions of these particles.The Ti and Fe elemental mapping images in Figure 2,panels c and d,respectively,con?rm the chemical nature of the Ti and Fe particles in the BSE image.Oxygen mapping shows most O in Ti particles.These results indicate that FeTiO 3has been reduced into large titanium oxide particles and small iron particles.A small amount of Fe could exist in Ti particles.Because the weight ratio of ilmenite to carbon was 4:1,the available C is only suf?cient for reduction of FeTiO 3to TiO 2and Fe.14The further reduction of TiO 2to Ti 3O 5and Ti 2O 3is believed to happen due to the presence of hydrogen gas.

The heated sample was annealed again at a lower temperature of 700°C for 4h in N 2-5%H 2to activate the one-dimensional growth of rutile structure.The SEM image in Figure 3a shows the early stage of the nanorod growth,and we can see many nanorods appearing from the surfaces of large titanium oxide particles.After heating for 8h,all titanium oxide surfaces were covered by a layer of nanorods of around 100nm in diameter and a few micrometers in length (Figure 3b).Typical rectangular cross sections and sizes of the nanorods can be seen clearly in Figure 3c,which will be discussed later.The XRD pattern of the ?nal sample in Figure 3d shows only TiO 2rutile and Fe phases.Ti 3O 5and Ti 2O 3phases are no longer detected.A selective chemical leaching treatment using 3M HCl solution was found to effectively remove most Fe.

Figure 4a shows a bright-?eld TEM image of a TiO 2nanorod,and its corresponding SAED pattern is displayed in Figure 4b.Electron diffraction pattern consists of a regular periodic array of dots and indicates that the nanorode has a single crystalline structure.The [002]direction of elongation was concluded from the SAED pattern (its orientation was carefully corrected for the rotation induced by the lenses of the microscope).The same direction of elongation is detected from other nanorods in the sample.The same growth direction has been observed by Yoo et al.in TiO 2nanowires.15The nanorod has a

nonuniform

Figure 2.(a)XRD pattern of the sample after ?rst annealing at 1200°C;(b)BSE image;(c)Ti mapping;(d)Fe mapping.

TiO 2Nanorods Extracted From Ilmenite Sands Crystal Growth &Design,Vol.9,No.2,2009

1241

contrast in a bright-?eld image,which is related to the complex shape of the cross section of the nanorods.The darker area in the center of the nanorod corresponds,most likely,to a larger thickness of its crystal in this area.Figure 4c shows a high resolution image taken from the edge of the same nanorod.Two types of lattice planes are resolved with the distances of 0.23and 0.25nm corresponding to (200)and (101)crystal planes of rutile TiO 2,respectively.TEM analysis reveals that the nanorod growth direction is parallel to the c axis of tetragonal cell of rutile structure,and the nanorod side-walls of (002)and (101)planes actually construct unit cells of tetragonal rutile.This can explain the rectangular cross sections observed in Figure 3c.

To examine possible reduction reactions during the ?rst annealing,TGA analysis of the milled sample was conducted in Ar-5%H 2gases (50mL/min)up to 1200°C with a heating rate of 10°C/min.Slow weight loss is observed in the TGA curve (Figure 5a)under 800°C with a total weight loss of only 6.30%.Previous investigation has found the slow weight loss is due to the slow reduction reactions by solid C:

14,16

Figure 3.SEM images and XRD patterns showing sample morphology changes during the second annealing at 700°C in N 2-5%H 2.(a)4h,(b)8h;(c)cross-sections of nanorods;(d)XRD pattern of the sample after second

annealing.

Figure 4.TEM image of a nanorod:(a)low magni?cation image;(b)SAED pattern;(c)high magni?cation image taken from the edge of the

rod.

Figure 5.(a)TGA curve up to 1200°C in Ar -5%H 2at a ?ow rate of 50mL/min;(b)TGA curve of sample (a)up to 700°C in N 2-5%H 2at a ?ow rate of 50mL/min.

1242Crystal Growth &Design,Vol.9,No.2,2009Yu et

al.

FeTiO 3+C f TiO 2+Fe +CO

Fast weight loss is found when the temperature is above 800°C,which is due to the following fast reduction reactions induced by CO gas:14

FeTiO 3+CO f TiO 2+Fe +CO 2

CO 2+C f 2CO 3TiO 2+CO f Ti 3O 5+CO 2

Because a small amount of hydrogen was introduced into the annealing atmosphere,the following reactions also possibly occur:17

2TiO 2+H 2f Ti 2O 3+H 2O 2FeTiO 3+3H 2f 3H 2O +2Fe +Ti 2O 3

3TiO 2+H 2f Ti 3O 5+H 2O

Several reactions could occur at the same time,and thus the weight loss has different speeds as the TGA curve indicates.Figure 2a XRD pattern proved these possible reactions because Ti 3O 5,Ti 2O 3,and Fe phases are present in the powder sample after ?rst step annealing.

The above reactions normally take place at a temperature at least above 1200°C.18,19Preball-milling treatments have an activation role and reduces reaction starting temperatures.High energy ball milling greatly reduces the particle size of ilmenite and active carbon (Figure 1a,b)and increases of the total external surface area,which presumably increases the contact area between them.The enhanced reduction process in premilled samples is attributed mainly to intimate mixing of carbon with disordered ilmenite nanocrystallites.This substantially reduces the diffusion length of oxygen to carbon in the solid-state reaction,thus increasing the rate of CO production for initiating the main gaseous reduction.The ef?ciency of the carbothermic reaction is initially greatly increased by ball milling treatment.14

No further reduction appears to be detected during the second annealing at 700°C as the TGA curve in Figure 5b suggests,which was conducted in N 2and 5%H 2mixture gas up to 700°C with heating rate 20°C/min and ?ow rate 50mL/min.The sample weight does not reduce but increases slightly 2.36%from room temperature to 700°C,which might be due to the following possible reactions:

2Ti 2O 3+O 2f 4TiO 22Ti 3O 5+O 2f 6TiO 2

XRD pattern in Figure 3d has agreement with above two reactions.The oxygen needed for rutile formation comes from the furnace environment.On the other hand,both Ti 2O 3and Ti 3O 5phases are thermally metastable phases and readily transformed into more stable rutile phase,which may or may not require a very small amount of oxygen.20,21Hydrogen reduction may still exist,but the oxidation speed is probably faster than the reduction speed,leading to a slight weight increase in the TGA curve.Amin et al.also found that Ni-coated TiO particles become TiO 2nanowires in a low vacuum system due to the absorption of a small amount of oxygen.12The limited concentration of oxygen played a very important role,and to con?rm,the sample after ?rst step annealing was heated in oxygen at 700°C for 6h.All Ti 2O 3and Ti 3O 5have become TiO 2rutile,but the nanorods could not be found because Fe becomes Fe 2O 3which cannot catalyze nanorod formation.The limited concentration of the oxygen level needs further study.

The formation of special nanorod structure is because of one-directional growth of the rutile crystals in the second annealing treament.We observed that Fe particles produced by ?rst annealing act as catalysts to help nanorod growth.The SEM image in Figure 6a shows a small particle at the tip of one nanorod.The EDS analysis revealed that the particle is enriched with Fe and the nanorod is titanium oxide.The

corresponding

Figure 6.(a)SEM image indicating an iron particle at the tip of a nanorod;(b)and (c)EDS spectra recorded at point A and B of the nanorod in panel a,respectively;(d)SEM image showing possible hydrogen etched areas,indicated by the arrows.

TiO 2Nanorods Extracted From Ilmenite Sands Crystal Growth &Design,Vol.9,No.2,2009

1243

EDS spectra are displayed in Figure6,panels b and c, respectively.Although most iron forms large particles and they are not found to help nanorod growth,some?ne iron particles (diameter less than100nm)might be left among the TiO particles and act as catalysts.A similar catalytic role of Fe to the growth of TiO2nanorod and other one-dimensional structures has been observed.15,22But SEM images show that many nanorods do not have Fe particles at the tips,and therefore they are produced in different mechanisms.Figure6d shows irregular cross sections and etched traces on the rod external surfaces. SEM analysis of a number of nanorods formed at different stages found that nanorods with large cross sections seem to be etched during extended annealing in hydrogen containing gas.One large nanorod became several nanorods with smaller diameters.This phenomena might be caused by hydrogen gas reduction along [002]directions.It has been reported that open atomic channels exist between TiO6octahedra along the[002]direction of rutile crystals and hydrogen gas can have an etching effect along the channels.15,23Tsujiko et al.found inverse-pyramid holes on (001)surface after heating in hydrogen gas at700°C.24 Therefore,hydrogen gas has played important roles in the two annealing processes.Without hydrogen gas,rutile particles can be produced from ilmenite but not in nanorod form.Nitrogen gas cannot be used in the?rst annealing as nitrides(TiN and FeN)can be formed at1200°C.N2-5%H2gases were preferred the atmosphere compared to Ar-5%H2in the second annealing because nitrogen gas helps the transformation from Ti2O3to TiO2as reported by Berger25and Lyubimov et al.26

Conclusions

We have demonstrated a practical,cost-effective method for producing rutile nanorods from ilmenite sands,and this method can be scaled up for large quantity production.In this method, ilmenite and active carbon powders were?rst ball milled in a planetary milling device,followed by two-step annealing treatment.The?rst high-temperature annealing at1200°C in Ar-5%H2results in reduction of ilmenite into metastable titanium oxides and catalytic iron;the second low-temperature annealing at700°C in N2-5%H2atmosphere leads to the growth of nanosized rutile rods.Fe can be removed using selective acid leaching.These results show a promising future for large-quantity production of rutile nanorods required by various environmental,biomedical,transportation,and chemical manufacturing applications.

Acknowledgment.This work is supported in part by the Australian Research Council under the Centre of Excellence and Discovery programs.The authors thank Mr.David Llewel-lyn,Dr.Frank Brink,and Dr.Cheng Huang for their assistance in microscopic analysis.

References

(1)Chen,X.;Mao,S.S.Chem.Re V.2007,107,2891.

(2)Lei,Y.;Zhang,L.D.;Fan,J.C.Chem.Phys.Lett.2001,338,231.

(3)Miao,L.;Tanemura,S.;Toh,S.;Kaneko,K.;Tanemura,M.J.Cryst.

Growth.2004,264,246.

(4)Miao,L.;Tanemura,S.;Toh,S.;Kaneko,K.;Tanemura,M.Appl.

Surf.Sci.2004,238,175.

(5)Zhang,Q.;Gao,https://www.sodocs.net/doc/dc18766902.html,ngmuir2003,19,967.

(6)Wen,B.;Liu,C.;Liu,Y.;New,J.Chem.2005,969.

(7)Wu,J.M.;Shih,H.C.;Wu,W.T.Chem.Phys.Lett.2005,413,490.

(8)Wu,J.M.;Shih,H.C.;Wu,W.T.;Tseng,Y.K.;Chen,I.C.J.Cryst.

Growth2005,281,384.

(9)Xiang,B.;Zhang,Y.;Wang,Z.;Wang,X.H.;Luo,X.H.;Zhang,

H.Z.;Yu,D.P.J.Phys.D2005,38,1152.

(10)Wu,J.J.;Yu,C.C.J.Phys.Chem.B2004,108,3377.

(11)Wu,J.M.;Wu,W.T.;Shih,H.C.J.Electrochem.Soc.2005,152,

G613.

(12)Amin,S.S.;Nicholls,A.W.;Xu,T.T.Nanotechnology2007,18,

445609.

(13)Chen,Y.;Bibole,M.;Le Hazif,R.And Martin,G.;Phys.Rev.B.

1993,48,14.

(14)Chen,Y.;Hwang,T.;Marsh,M.;Williams,J.S.Metall.Mater.Trans.

A1997,28A,1115.

(15)Yoo,S.;Akbar,S.A.;Sandhage,K.H.Ad V.Mater.2004,16,260.

(16)Chen,Y.;Hwang,T.;Williams,J.S.Mater.Lett.1996,28,55.

(17)Kellerman,D.G.;Zainulin,Y.G.;Perelyaev,V.A.Inorg.Mater.

1983,19,305.

(18)Becher,R.G.Aust.Pat.1963,241.

(19)Becher,R.G.;Canning,R.G.;Goodheat,B.A.;Uusna,S.Proc.Aust.

Inst.Min.Metall.1965,214,21.

(20)Zheng,L.Y.;Li,G.R.;Xu,T.X.;Yin,Q.R.J.Inorg.Mater.2002,

17,1253.

(21)Duyar,O.;Placido,F.;Durusoy,H.Z.J.Phys.D-Appl.Phys.2008,

41,9.

(22)Chen,Y;Chadderton,L.T.;Williams,J.S.;Fitz Gerald,J.Appl.Phys.

Lett.1999,74,2782.

(23)Sasaki,J.;Peterson,N.L.;Hoshino,K.J.Phys.Chem.Solids.1985,

46,1267.

(24)Tsujiko,A.;Kisumi,T.;Magari,Y.;Murakoshi,K.;Nakato,Y.J.

Phys.Chem.B2000,104,4873.

(25)Berger,L.M.J.Mater.Sci.Lett.2001,20,1845.

(26)Lyubimov,V.D.;Shveikin,G.P.;Shestakova,T.V.;Alyamovskii,

S.I.Inorg.Mater.1978,14,372.

CG801125W

1244Crystal Growth&Design,Vol.9,No.2,2009Yu et al.