Chapter 5: Intrinsic Defect Induced Room Temperature
5.8. Magnetization studies
The magnetic properties of as-synthesiz ed samples were investigated using V S M . The field dependent magnetic (M – H) measurements are shown in F ig . 5.8. F ig . 5.8 (a ) shows a clear ferromagnetic hysteresis loops for samples D 50 0 , D 9 0 0 , and G 50 0 . The samples D 50 0 and D 9 0 0 show large magnetic moments compared to sample G 50 0 . M oreover, D 9 0 0 shows a higher magnetic moment than that of D 50 0 . F ig . 5.8 (b ) shows the M – H loops of the samples A 50 0 , A 7 0 0 and A 9 0 0 . These samples show hysteresis loops at low field, but lack s the saturation magnetiz ation, perhaps due to paramagnetic contribution from the conduction electron.50 N ote that the magnetic moment is relatively high for A 7 0 0 and this may be due to the relatively large concentration of or may be related to nanobrick -lik e surface morphology of A 7 0 0 and/or due to mixed phase of TiO 2(B )– anatase. It has been reported that the FM is sensitive to the shape and siz e of the nanostructures.2 0 , 51 , 52 N ote that the samples (D 50 0 , D 9 0 0 , G 50 0 ) grown at 18 0 °C after 24 h and 48 h reactions are nanoporous N R bs with nanobrick -lik e structures on the surface. S uch complex structures may contain a large concentration of defects, supported by our P L analysis. The concentration of oxygen vacancies are found to depend on the different surface morphologies and siz e of the nanostructures, which is the reason for the observed different magnetic moments. N anosiz e TiO 2 usually shows weak FM at R T and the FM depends very much on the synthesis
procedure, due to the different defects, especially the oxygen vacancies in TiO 2
nanomaterials. O ne of the most important reasons of relatively strong ferromagnetism in our samples may come from the formation of porous lik e structures, since we are using ethylene glycol as a cosolvent which has chelating properties. D ue to the chelating properties of ethylene glycol, it forms several bonds with metal ions and form cage lik e structures which helps the formation of porous structure creating more oxygen vacancies. Further, it has been reported that the oxygen vacancies induce lattice distortion in rutile TiO 2 and induces strong FM in undoped TiO 2 films due to charge redistribution.2 3 The magnitude of the magnetic moment in the rutile phase was predicted to be about four times higher than that in anatase TiO 2. This is consistent with the observation of highest magnetic moment seen in sample D 9 0 0 with a mixed rutile-anatase phase as compared to D 50 0 and G 50 0 with TiO 2(B ) phase.
It is well k nown that the structure of TiO 2 is very sensitive to oxygen and can be easily reduced under an oxygen deficient environment. Therefore, the sample A 50 0 was annealed at 30 0 °C under moderate vacuum (1.2 × 10-2 mbar) in order to induce a higher concentration of oxygen vacancies in the sample. Interestingly, the vacuum annealed sample A 50 0 V shows enhanced ferromagnetism with a well-defined hysteresis loop having saturation magnetiz ation (M s) of 0 .19 1 emu g-1, remanent (M r) of 0 .0 2 emu g-1 and coercive field (Hc) of 9 9 O e, as shown in F ig . 5.8 ( c ). S uch enhancement of FM in oxygen deficient TiO 2 N R bs strongly suggests that R TFM and concentration of oxygen vacancies are directly correlated. Indeed the samples exhibiting a high intensity of visible P L (see inset of F ig . 5.5(a )) only show clear ferromagnetic hysteresis loops with saturation magnetiz ation plateau.
N ote that the observed M s at room temperature (R T) in our undoped TiO 2 N R bs is more than three times higher than that reported for Fe and N co-doped TiO 2 nanorods53 and other reported literatures.2 7, 53, 54 A chievement of such a high magnetic moment in defect engineered undoped TiO 2 is remark able and our studies provide clinching evidence that oxygen vacancies mediate the ferromagnetic interaction. The magnetiz ation parameters for different samples are listed in T a b le 5.1. F ig . 5.8 (d ) shows the M -H measurement of B 50 0 , B 7 0 0 and B 9 0 0 samples. The M -H curve indicates almost paramagnetic nature (or very weak ferromagnetic behavior at low field). N ote that these samples show the Ti interstitial-rich defects with negligible defects as observed from the P L studies (see F ig . 4.5(d ), Chapter 4). This result again supports the oxygen vacancy mediated FM in undoped TiO 2
nanostructures system and further indicates that Ti interstitial defects are not responsible for the observed FM in our TiO 2 N R bs.
T a b le 5.1 . M agnetiz ation parameters of the as-synthesiz ed and vacuum annealed samples: saturation magnetiz ation (), remanent magnetiz ation (), and Coercive field () were determined from M – H loops; , , , were evaluated from fitting of the M – H curve with B M P model.
S a m p le n a m e
!
(e m u /g )
"
(e m u /g )
#$ (O e )
%
(e m u /g )
&'(( × 1 0-1 7(e m u )
& × 1 0-6 (c g s )
) × 1 01 5 (c m -3)
A 50 0 V 0 .19 1 0 .0 2 9 9 0 .19 0 8 .8 6 1.32 8 .8 3
D 50 0 0 .0 61 0 .0 0 4 9 9 0 .0 7 0 6.21 0 .22 4.63
D 9 0 0 0 .10 9 0 .0 0 7 8 7 0 .0 9 7 8 .28 2.17 4.7 7
G 50 0 0 .0 54 0 .0 0 3 9 0 0 .0 48 6.54 1.0 2 3.0 2
F ig . 5.8 .M agnetic field versus magnetiz ation (M – H) loop at room temperature showing hysteresis in as-synthesiz ed samples: (a) D 50 0 , D 9 0 0 and G 50 0 ; (b) A 50 0 , A 7 0 0 and A 9 0 0 ; (c) A 50 0 and A 50 0 V ; (d) B 50 0 , B 7 0 0 and B 9 0 0 . The insets show the magnified M – H loop showing clear ferromagnetic hysteresis behaviour.
It is understood that for use in a wide range of applications without temperature control, the ferromagnet should have a transition Curie temperature (TC) enough above R T (30 0 K ). F ig . 5.9 shows the temperature dependent magnetiz ation (M -T) of the sample G 50 0 in the temperature range 29 0 to 8 60 K . S ince the effect of high temperature V S M measurement in atmospheric conditions is equivalent to the post annealing of the as-grown samples in air that can destroy the ferromagnetic coupling, we performed the high temperature V S M measurement in a nitrogen atmosphere. From the differential plot of the M -T curve, we obtained a TC at ~ 7 9 9 K . Interestingly, the gradual increase in magnetiz ation from 29 0 K to around 334 K is observed in the M -T curve,2 8 as shown in the inset of F ig . 5.9. This indicates that the oxygen vacancies are increased during the early stages of the heating process in a nitrogen atmosphere and these defects contribute to the observed FM . N ote that defect mediated bound magnetic polaron (B M P ) model has been invok ed to explain the R TFM . D ue to increased vacancy concentration, more B M P s are lik ely to form, which include electrons locally trapped by the oxygen vacancy, giving rise to gradual increase in magnetiz ation. The thermal fluctuations of the localiz ed spins may have comparatively less effect in this temperature range. V ery recently, Tian et al.55 reported nearly temperature- independent saturation magnetiz ation up to 60 0 K , which strongly favored the B M P model.
The strength of exchange interaction is stronger in nanostructured magnetic semiconductors where the mean distance between the localiz ed spins is small, which may enhance the thermal stability.
F ig . 5.9 . Temperature dependent magnetiz ation (M – T) curve of sample G 50 0 , showing a ferromagnetic to paramagnetic transition at 7 9 9 K .