N. Guskos et al. / Journal of Alloys and Compounds 606 (2014) 32–36
35
Lorentzian functions and are presented in Fig. 2b. From the relative
integrated intensity, calculated as I = Aꢀ
H2, where A is line ampli-
tude and H the linewidth, the amount of trivalent titanium and
a
D
D
FeTiO3 can be estimated. For Ti3+ the following intensity ratios
were calculated: 1:1.6:0.3, for samples n = 1 wt.%, 5 wt.% and
10 wt.%, respectively. For FeTiO3 the following intensity ratios were
obtained: 1:0.8:2.9 for sample n = 1 wt.%, 5 wt.% and 10 wt.%,
respectively. The amount of free radicals was almost the same in
samples n = 1 wt.% and 5 wt.%, while it was two times smaller in
sample n = 10 wt.%.
The FMR spectra for nFe,N-TiO2 samples were satisfactory fitted
by using two (for n = 5 wt.%) and three (n = 1 wt.% and 10 wt.%)
components of the Callen lineshape. These components represent,
in a very simplified way, lines that are produced by magnetic
anisotropy of the spin system. The following equation for the Cal-
len lineshape is obtained in case of the linear polarization of the
microwave field [31,33]:
b
H20½ðH20 þDB2ÞðH2DB þ2H0jHjdBÞþH2ðH2 þd2ÞDB
ꢁ
0
0
B
IðHÞ /
2
2
2
2
½ðH ꢂH0Þ H2 þðjHjDB þH0dBÞ ꢁ½ðH þH0Þ H2 þðjHjDB þH0dBÞ ꢁ
0
0
ð1Þ
where H0 is the intrinsic resonance field,
DH is the linewidth con-
nected with relaxation of the Landau–Lifshitz type, and dH the line-
width connected with the Bloch–Bloembergen relaxation. Under
certain circumstances the Landau–Lifshitz relaxation can be identi-
fied with the longitudinal (spin–lattice) and the Bloch–Bloember-
gen with the transverse (spin–spin) relaxations. The result of
fitting is presented in Fig. 1 by solid lines. All parameters strongly
depend on the initial concentration of iron (Fig. 3).
Fig. 4. Comparison of magnetic and photocatalytic properties of three investigated
samples: (upper panel) Temperature dependence of the FMR integrated intensity.
Inset shows the magnetic intensities at RT. (bottom panel) CO2 evolution rate for
three investigated samples.
Fig. 3 presents the temperature dependence of the FMR param-
eters, namely the resonance field, two types of linewidth and the
FMR amplitude. As the resonance field of all components is con-
cerned it shows generally weak temperature dependence, increas-
ing in value as the temperature is decreasing from RT (Fig. 3).
Assuming that the overall magnetic anisotropy field Banis scales
Photocatalytic activity of our samples was investigated in Ref.
[27]. It was found that the co-modified photocatalysts exhibit
much higher activity than the single-modified materials and the
most desirable iron concentration is 5 wt.%. In Fig. 4, as an illustra-
tion, the CO2 photocatalytic evolution during acetic acid decompo-
sition under mercury lamp visible light (>400 nm) irradiation in
the presence of our samples, is presented. According to previous
findings by Ohtani et al. [35] in the applied conditions, CO2
evolution is an effect of complete acetic acid oxidation: CH3COOH +
2O2 ? 2CO2 + 2H2O. The amount of CO2 was, therefore, used for
photocatalytic activity measurements.
Comparison of both graphs in Fig. 4 allows to draw the conclu-
sion that the higher the FMR integrated intensity the smaller the
photocatalytic activity of the (Fe,N)-TiO2 sample. It is generally ac-
cepted that the presence of FeTiO3 phase in TiO2 structure im-
proves its photocatalytic performance under visible light
irradiation. This explains why sample n = 5 wt.% is better in this re-
spect than sample n = 1 wt.%. On the other hand, it is known that
high amount of Fe3O4 could be responsible for low photoactivity
of TiO2, accounting for the strongest FMR signal and the lowest
photocatalytic activity of n = 10 wt.% sample.
with the difference in magnetic fields of the components, Banis
ꢃ
BiꢂBj, the largest Banis field is measured for sample n = 10 wt.%
(Banis ꢃ 16 kG), the smallest for n = 5 wt.% (ꢃ2 kG). This is consis-
tent with the need to use only two components to fit the line-
shape of sample n = 5 wt.%, while three components were need
to properly describe the FMR spectra of n = 1% and n = 10% sam-
ples. With decreasing temperature the values of Banis for all sam-
ples slightly increase.
As the temperature dependences of the components linewidths
are concerned, for samples n = 1 wt.% and n = 5 wt.% there are only
rather weak changes observed in the whole temperature range, but
for sample n = 10 wt.% at temperatures below 10 K a strong varia-
tion of both types of linewidths is registered (Fig. 3). For the stron-
gest component in that sample (designated as component 3) with
the higher amplitude, the Landau–Lifshitz linewidth decreases
while the Bloch–Bloembergen linewidth increases sharply on cool-
ing below 10 K. This behavior may be indicative of the reorienta-
tion processes of the spin system that involve essential changes
of the dipole–dipole interactions.
4. Conclusions
Another important parameter, the FMR integrated intensity, is
calculated as the product of the amplitude and squared linewidth
and is assumed to reflect the nanoparticles concentration. Temper-
ature dependence of the FMR integrated intensity of our samples is
presented in Fig. 4. The most intense FMR signal is registered for
sample n = 10 wt.% in the whole temperature range. The smallest
integrated intensity at 290 K was obtained for sample n = 5 wt.%,
in contrast to the nominal iron concentration. Below 200 K sample
n = 1 wt.% displays the smallest integrated intensity. This indicates
that the skin effect due to variations of the electrical conductivity
could play an important role at higher temperatures [34].
FMR/EPR analysis of three co-modified nFe,N-TiO2 samples
leads to the conclusion that characterization of the materials mag-
netic properties provide valuable insight to their photocatalytic
activity. The presence of Ti3+ ions, FeTiO3 and Fe3O4 phases, free
radicals is crucial in this respect. In all three investigated samples
the FeTiO3 phase and free radicals were identified. In samples
n = 1 wt.% and n = 5 wt.% Ti3+ ions were observed, with higher con-
centration for the latter sample, whereas no signal from trivalent
titanium ions could be traced for the n = 10 wt.% composite. All
samples are conductive; at RT the best conductivity was inferred
Guskos
