Interfacial induction sustained ferromagnetism from solution chemistry to ceramics 5

reported room temperature ferromagnetism of anatase Co-TiO2 film deposited by laser molecular beam epitaxy technique and the same property was revealed in the thermodynamically more stab

Titanium oxide exists in three crystalline structures, namely anatase, rutile and brookite as shown in Figure 7.1 Earlier work by Matsumoto et al reported room temperature ferromagnetism of anatase Co-TiO2 film deposited by laser molecular beam epitaxy technique and the same property was revealed in the thermodynamically more stable rutile state using the same fabrication technique (Matsumoto, et al., 2001a; Matsumoto, et al., 2001b)

168

169

room temperature ferromagnetism of the thin films fabricated by the solution chemistry approach

Many studies have been carried out to determine the origin of ferromagnetism in doped TiO2 films Our study in this work shows unanimously no metallic cobalt clusters in the films, and therefore the ferromagnetism should not be due to the zero valence cobalt Researchers like S.A Chambers explained the origin in terms of ferromagnetic exchange interaction, attributing this room temperature ferromagnetism

Co-to intrinsic stimulation (Chambers, 2002; Chambers, et al., 2003; Toyosaki, et al., 2004) Cui et el also reported that the ferromagnetism comes from the exchange mechanism between Co2+-O2--Co3+ (Cui, et al., 2004; Shutthanandan, et al., 2004) Coey and coworkers proposed that the ferromagnetic exchange here (in dilute ferromagnetic oxides and nitrides) is different from the conventional super-exchange

or double-exchange interaction as such interactions cannot produce long-range magnetic order at such low magnetic cationic concentrations They thus proposed that the ferromagnetic exchange is due to the indirect exchange via shallow donors (associated with oxygen vacancies) (Coey, et al., 2005)

From the earlier chapters on heterogeneous metallic oxide composite ½(1-x)La2O3

-xSrO/⅓Co3O4, it was found that such oxide composite manifested ferromagnetism at room temperature with the stipulation that the oxides must be mixed in nano-domains

On the contrary, Sr-doped lanthanum cobalt oxide (namely perovskite LSCO structure), studied by many materials scientists, has a Curie temperature much lower than room temperature (Raccah and Goodenough, 1968; Rao, et al., 1977; Petrov, et al., 1995) The uniqueness of the heterogeneous composite system lies in its vast interfacial phase

170

between SrO and spinel Co3O4 nano-domains In order to maintain the oxide phases in

a few nanometers range during the thermal curing process required for cultivating crystal structures in the respective oxide phases, applying a low calcination temperature is imperative The generation of this ferromagnetism is attributed to interfacial stimulation happening at the boundary between SrO and Co3O4 domains and presumed via the interfacial “Jahn-Teller” effect An attempt to implement this interfacial stimulation mechanism into the Co-TiO2 matrix for enhancing its room temperature ferromagnetism has been found viable Through our research on this tri-component (SrO-Co-TiO2) composite, it has been found that SrO exhibits a surfactant-like effect on maintaining an extremely high dispersion of the three oxide phases in the matrix of thin film in addition to the induction effect

7.2 Experimental

7.2.1 Chemicals

Titanium (IV) isopropoxide (C12H28O4Ti, 97%, Aldrich), cobalt (II) methoxyethoxide (Co(OCH2CH2OCH3)2, 99%, Alfa Aesar), strontium nitrate (Sr(NO3)2, >99%, Acros Organic), hydrochloric acid (37%, Fisher Chemicals), triethanolamine (98%, Aldrich), ethanol (≥ 99.9%, Merck), isopropanol (≥ 99.5%, Tedia), Brij 30 (Aldrich) and polyethylene glycol 400 (Merck) were used as received

2-7.2.2 Preparation of precursors for thin film fabrication

Two different sol-gel processes, which employ the acidic and the basic catalysts respectively, were designed to fabricate the dense TiO2-based films

171

7.2.2.1 The acidic method

For formulating the sol-solution to synthesize Co-TiO2 film (CTO), the sol solution was first prepared by adding titanium tetra-isopropoxide (TTIP, 6.384-7.728 mmol) and cobalt methoxyethoxide (CMEO, 0.672-2.016 mmol) into an acidified ethanolic solution, which was made up of 0.3ml concentrated hydrochloric acid (37%) in 9.8ml ethanol Three different compositions were attempted, namely CoxTi1-xO2-y with x =

0.08, 0.16 and 0.24 A given amount (5.4 % by volume) of polyethylene glycol (PEG 400) was then added into it, and the resulting mixture was stirred for 1hr at ambient temperature (~25°C) The sol solution generated in this controlled hydrolysis step contains nano CTO particles (< 5nm) as shown in Figure 7.2

Figure 7.2 HR-TEM image of nano CTO particles in the sol solution

172

7.2.2.2 The alkaline method

For formulating the sol-solution for fabrication of SrO-Co-TiO2 film (SCTO), an aqueous solution containing the desired amount of strontium nitrate was added dropwise to an isopropanol solution containing titanium tetraisopropoxide, cobalt methoxyethoxide (with the stoichiometry Ti/Co = 1.0 / 0.16) and an organic base Triethanolamine (TEA, 9.8% by vol), which acts as the coordination ligand to stabilize the metal ions (in particular Sr2+ and Ti4+ ions), were used as the organic base here Two drops of the non-ionic surfactant, Brij-30, was then added into the solution Similar to the acidic based method, the same amount of PEG 400 was introduced to the solution, and the resulting solution was stirred for 1h at ambient temperature (~25°C) to obtain a transparent sol solution For comparison purposes, it was found that the acidic sol-gel approach is inappropriate in formulating the sol solution This is because Sr2+ and Cl- are likely to form an electron donor-acceptor bridged complex and such bridged complex can be extended in a 2 dimensional network, leading to the heavy gelation in the acidic medium

7.2.3 Development of film

The substrate used here is silicon wafer The silicon substrate was cleaned and degreased in acetone in an ultrasonic cleaner Spin coating was performed with a spin rate of 2000 rpm for 2 min to develop a liquid film of the sol solution The resultant liquid film was then dried for 30 min at 80°C before heated up to 500oC using a heating rate of 2°C/min, where the curing sample was dwelled for 1 h before allowed

to cool at 2°C/min The coating and curing procedure was repeated several times to get

a thicker film

173

obtain the morphology of the samples Valence states of Co 2p, Ti 2p and Sr 3d were

analysed with the X-ray photoelectron spectroscopy using C 1s (284.6eV) as reference (XPS, Kratos Axis HSi System) and Vibrating Sample Magnetometer (VSM, Lake Shore 735, noise floor 10-7 emu for 10 sec/pt) was used to measure the magnetic properties of the films at room temperature

7.3 Results and Discussion

7.3.1 Fabrication of continuous dense Co-TiO 2 films

From FESEM, it can be observed that both the CTO films (x = 0.08 and 0.16) have a

uniform membrane morphology and good adhesion to the substrate shown by the intimate interface (Figure 7.3 a and b) The PEG present in the sol solution yields a lubricating effect on sol particles such that they are able to maneuver among themselves when the solvent is being removed By adding small amount of PEG, conducting a gradual combustion of the organic entity and setting a slow calcination rate to consolidate the amorphous oxide composite generated from the combustion, a crack-free film could be ultimately fabricated

174

a

b

175

c

Figure 7.3 FESEM images of CTO films with (a) x = 0.08, (b) x = 0.16 and its

cross sectional view (c)

Besides the control of processing conditions, the stoichiometry of cobalt is also crucial

to the quality of film Severe cracks, as can be observed from Figure 7.4, were resulted

with the usage of high dopant content (x = 0.24) When the Co content is too high,

crystallites (CoTiO3) with high cobalt oxide content will be formed As these high content crystallites have different coefficients of thermal expansion from the TiO2

Co-phase, the material stress that was generated during the cooling course (after calcination at 500°C) led to the cracking of film This conclusion can be supported

from the IR spectra of the dried precursor gels (Figure 7.5) From the IR spectrum b,

which still contains the characteristic vibration absorption peaks of organic moiety that could be assigned to methoxyethoxide ligand, it can be inferred that the sol-gel reaction (hydrolysis and condensation) rate of the cobalt-containing precursor (CMEO)

is slower than that of the titanium precursor (TTIP) Therefore, with increasing CMEO content, the CoO-enriched oxide phase will be easily generated

176

The thermal degradation graphs of the dried precursor gel (Figure 7.6) were collected

to analyze the dehydration and condensation course of how the dried gel was

converted to the crystalline CTO From the thermal degradation graph for CTO with x

= 0.08, it was observed that the first major weight loss took place in the range from 200°C to 325°C which was followed by a small weight loss from 325°C to 450°C These two weight-loss stages can be accounted for the elimination of water (due to condensation of surface OH groups among the nano-sol particles and then micro-domains) The third weight loss stage was due to the fast transition from the

amorphous phase to the anatase phase This step sets off at 459.4°C for CTO with x = 0.08 and 496.9°C and 556.3°C respectively for x = 0.16 and x = 0.24 Such increasing trend with x value shows that the formation of CoTiO3 phase retards the growth of anatase phase (which will be shown in Figure 7.7), and thus a higher transition temperature was required

Figure 7.4 SEM image of CTO film of x = 0.24

177

400 600

800 1000

1200 1400

1600 1800

2000 2200

459.4°C

496.9°C 556.3°C

Figure 7.6 Thermal degradation chart of the precursor (a) x = 0.08, (b) x = 0.16 and (c) x = 0.24

Figure 7.7 XRD chart of TiO2 and CTO films after calcination at 500°C for 1 hr, (a) TiO2, (b) x = 0.08, (c) x = 0.16 and (d) x =

With a curing (calcination) temperature of 500°C, it can be observed from Figure 7.7 that the major phase of the three films is anatase Though TGA analysis of the dried gel showed that a temperature above 500°C is needed for the generation of anatase phase (for system with higher dopant content), the sample stayed much longer in the curing process than in TGA scan at 500°C, and therefore, a longer calcination period enables the completion of this crystallization For the two films with higher Co

contents, x = 0.16 and 0.24, CoTiO3 phase, as aforementioned, is observed from the XRD chart (JCPDS 2004b) Another notable point is that no trace amount of Co metal can be found from the three XRD patterns Magnetic measurements of these three CTO films were taken using VSM at room temperature The coercivity and remanence

of the films are shown in Table 7.1 and the data are apparently dependent on x The optimal composition is when x = 0.16, where two factors, which are the content of

CoTiO3 phase that weakens ferromagnetism and the extent of Co-doping anatase TiO2

that stimulates ferromagnetism, seem to be balanced

Table 7.1 Magnetic properties of the CTO and SCTO films

Remanence (% wrt saturation magnetization)

182

2 / 3

p

2 / 3

p

2 / 3

2 p

binding energy of the Co main peak (Co 2 = 780.0 eV) appears for all three films

Since this peak is also present for CTO with x = 0.08 which contains only negligible

CoTiO3 phase in contrast to the XRD of x = 0.16 (Figure 7.7), this XPS peak could

thus be attributed to the cobalt species that doped TiO2 phase The peak could be fitted satisfactorily by adding a shoulder peak at 781.5eV, which is indicative of the presence

of Co3+ ion According to the areas of the two fitting peaks, the ratio of Co2+ to Co3+ is about 0.8 Since anatase TiO2 has the tetragonal bipyramids structure with an elongated distortion, when Ti4+ (r = 0.75Å) is substituted by both Co2+ and Co3+, to minimize structural distortion, Co2+ ion would take the low spin state (r = 0.79 Å) rather than the high spin state (r = 0.89 Å) and Co3+ ion would take the high spin state (r = 0.75 Å) (Cotton and Wilkinson, 1999) Hence, the magnetism might principally come from the doping site by Co3+ With the increase in x, Co 2 becomes more intense and the Co2+/Co3+ increases, and this corresponds to the growth of CoTiO3

phase as depicted by the XRD chart (Figure 7.7) In addition, on the three XPS spectra, the representative peak for Co metal (Co = 778.3eV) is absent This excludes the possibility that metallic cobalt clusters are responsible for the room temperature ferromagnetism and is consistent with the XRD result mentioned earlier The other pair of peaks with binding energy at 785 ~ 786.5 and 802.5 eV is most likely to be due

to Ar sputtering done for surface cleaning before analysis or shake-up satellite peaks (Moulder, 1992; Yang, et al., 2003; Chai, et al., 2005) From the above XRD and XPS characterizations, it could be concluded that both TiO2 and CoTiO3 phases compete for

cobalt ions and the majority of them goes to join the latter phase with increasing x

value The distribution of cobalt ions in anatase TiO2 phase is responsible for the room

temperature ferromagnetism The optimal cobalt content happens at x = 0.16, and the

excess of cobalt ions would go to the CoTiO3 phase

Binding energy (eV)

7.3.2 Implanting the SrO nano domains into the matrix of CTO film

Our previous finding of the interfacial induction, which takes place between SrO and

Co3O4 nano-domains and triggers the unique room temperature ferromagnetism, has led us pursue the same type of induction mechanism in the CTO film via implanting of

nano-size SrO domains into its matrix The above CTO film (x = 0.16) was selected as

the host matrix for lodging the SrO as it exhibits the highest ferromagnetism in the CTO series Besides this, an equivalent stoichiometry of Sr(II) and Co(II) was used Thus, the film’s composition is SrxCoxTi1-2xO2 (SCTO) where x = 0.16 The film with a

relatively lower Sr content, SrxCo2xTi1-3xO2 (x = 0.08) was also attempted Based on

the alkaline-based sol-solution formulated in this work, a dense and crack-free thin film was realized (Figure 7.9) via spin coating and curing process, in which each coating contributes to about 150nm in thickness

a

184

b

~800nm

Figure 7.9 SEM images of SCTO film after calcination at 500oC for 1h, (a) one layer; (b) five layers

The TGA profile of the dried CTO gel (x = 0.16, Figure 7.6) shows that the reduction

in mass eases after 627°C However, the TGA profile of the dried SCTO gel levels out

at 800°C (Figure 7.10) Ostensibly, the presence of Sr2+ in the gel has a significant impact to its thermal stability Furthermore, the XRD investigation of the SCTO prepared by calcining its dry gel at 800°C for 1 h exhibits a heterogeneous metallic oxide composite comprising of rutile, a trace amount of anatase and an oxide phase that can be credited to SrCoO2 (labeled by s-c on Figure 7.11) since it shows the characteristic peaks of s-c (with 2θ = 29°, 33°, 44.5°, 56° and 59°) (JCPDS 2004c)

The SrCoO2 generated in SCTO film should be in crystallite form as the peaks are rather broad

In contrast, if the SCTO gel is subjected to calcination at 500°C, no distinct XRD pattern is obtained (inset of Figure 7.10) Indeed, the sample prepared at this

185

186

temperature corresponds to an intermediate state on the TGA profile There is still a mass difference of around 14% between this intermediate state and the oxide composite left behind at 800°C This intermediate state, stable at 500°C, is predominated by interfacial phases since the mass eliminated while it is heated to 800°C is primarily caused by elimination of water through condensation reaction of surface –OH groups on tiny domains of the individual oxides Nevertheless, it is noteworthy that when the content of SrO in SCTO is halved, a sharper XRD pattern

was obtained (Figure 7.11, x = 0.08 for Sr xCo2xTi1-3xO2); it shows both anatase and rutile TiO2 as the two main phases From this, it can then be inferred that the SrO component obstructs crystallization in SCTO In addition, CoTiO3 phase, rather than SrCoO2, was formed when the system (x = 0.08 for Sr xCo2xTi1-3xO2) was subjected to

calcination at 800°C Compared with the system (x = 0.16 for Sr xCoxTi1-2xO2), it seems that the SrO phase also retarded formation of CoTiO3 via reacting with cobalt oxide

The XRD data presented above has led to a conclusion that the chemical environment

of Ti4+ in SCTO is different from that in CTO when both were subjected to calcination

at 500°C It has been clear that the latter consists of anatase and CoTiO3 phases But for the SCTO system, its XRD pattern has not yet been fully developed as shown on the inset of Figure 7.10 Despite this, it is believed that the oxide domains of the SCTO contained crystallite in nanoscale according to its room temperature magnetic properties With such calcination temperature, the cobalt species in this system was likely to form Co3O4 nano domains instead of doping TiO2 since TiO2 phase is still in miniature size This reasoning can be further verified by comparing the XPS of spectra

36%

0 0.005 0.01 0.015 0.02 0.025

Figure 7.11 XRD of SCTO film calcined at 800°C with r, a and s-c represent rutile, anatase and SrCoO2 phase respectively

This work has fabricated two types of doped-TiO2 thin films, which are of Co-TiO2

(CTO) and SrO-Co-TiO2 (SCTO), by employing the sol-gel coating technique because

it provides the versatility of incorporating multiple dopants in the lattice of TiO2 film with precisely controlled dose Colloidal stability of TiO2 sol-solution and the dopant concentration in it are the two crucial factors affecting the quality of the Co-doped TiO2 films The CTO film (x = 0.16) gives an optimal magnetic coercivity of 68 Oe

and remanence 13.9% of its saturation magnetization and is the highest among the three doping levels It is also found that this magnetic response does not originate from the presence of metallic Co This magnetic response is deemed to originate from the occupation of high spin Co3+ ions in the cationic sites of anatase TiO2 Alternatively,

189

of Ti 2p of the CTO and SCTO composites As shown in Figure 7.12, the two CTO compositions (x = 0.08 and 0.16) display the same binding energy doublet peaks of

Ti4+, namely 2p 2 and2 p 2 In contrast, the Ti 2p doublets of SCTO (after calcination

at 500°C) decreased by around 1.5 eV The lowering of B.E value could be attributed

to the presence of abundant surface species of Ti-OH on anatase nano-domains

(crystallites) Alternatively, looking at the XPS spectra of Sr 3d as shown in Figure 7.13, it can be observed that the 3d5/2 occurs at 133.02eV and Δ = 1.70 eV Both 3d5/2

and Δ are lower than the standard values given in handbook that is based on the measurement of pristine SrO (Moulder, 1992) As noted above, the SCTO is a mixture containing of highly dispersed SrO nano-domains The downward shift of XPS binding energy of Ti4+ and Sr2+ in SCTO displays strong interfacial characteristics, which in turn verify that SCTO is composed of nano-domains of the three oxides

7.4 Conclusion

460465

470475

Binding energy (eV)

130132

134136

138140

Binding energy (eV)

the introduction of Sr2+ into the Co-TiO2 system significantly deterred the crystallization extent of TiO2 as well as the formation of unwanted CoTiO3 phase The calcination of SCTO at 500°C leads to an interfacial (or tiny individual oxide domains) predominant system because of the dispersing effect of SrO SCTO manifests greater magnetic properties than CTO This enhancement is originated from the interfacial induction between SrO and Co3O4 oxide domains

192

temperature ferromagnetism exhibited by the nano-composite, ½(1-x)Ln2O3

-xSrO/⅓Co3O4, originated from the interfacial induction effect between the dispersed SrO and Co3O4 domains The theoretical interpretation of this unique physical phenomenon could satisfactorily fit the experimental data and is also proven to be applicable in other systems, such as the doped TiO2 system From this study, several conclusions can be drawn

1 It was discovered that by playing around with the fabrication procedure, in

which the two key factors are the dopant concentration and calcination temperature, a heterogeneous oxide composite, composing of La2O3, SrO and Co3O4, can be obtained instead of the homogeneous La1-xSrxCoO3-δ

crystals By using a calcination temperature of 600°C and with a high dopant concentration (x ≥ 0.8), a heterogeneous complex oxide of ½(1-

x)La2O3-xSrO/⅓Co3O4 is obtained and what is interesting is that while the homogenous La1-xSrxCoO3-δ possessed Curie temperature much lower than ambient condition, this interesting heterogeneous material manifested room temperature ferromagnetism Further analysis proved that such room

193