Spectroscopic Analysis of
3+
Tb
Doped Sol-Gels
rare earth
sol gel research
Departments of Chemistry & Physics
Hamilton College
Brendan W. Sullivan ‘07, Ann J. Silversmith, Karen S. Brewer
Departments of Chemistry and Physics, Hamilton College
350000
300000
This individual project was intended to examine the quenching
processes in Terbium (Tb) doped sol-gel glasses. Neighboring
Tb3+ ions in the sol-gel matrix can undergo a process called
cross-relaxation, where one ion’s electron drops down in energy
and donates this energy to the other ion’s electron (Figure B).
This is a problem because that drop in energy of the first electron
yields a photon with a wavelength outside of the visible
spectrum, whereas other energy level changes result in visible
fluorescence that can be examined easily (Figure C).
5D
3
5D
4
7F
10
2
3
4
5
6
Fig. B: cross-relaxation
0% Al, 900C/6hrs
1400000
Fig. A: Dieke diagram
Intensity (arbitrary units)
1200000
0.01% Tb
0.02% Tb
0.05% Tb
0.1% Tb
0.2% Tb
0.5% Tb
1% Tb
2% Tb
1000000
800000
600000
400000
200000
Fig. F: concentration
dependence
0
400
450
500
550
600
650
Wavelength (nm)
30
200000
28
5D
3
26
24
150000
5D
4
20
50000
18
400
590nm
620nm
545nm
14
490nm
414nm
16
436nm
460nm
0
Intensity (arbitrary units)
fresh
1hr
2hrs
5hrs
150000
6hrs
450
500
550
600
650
Fig. D
12
10
25000
8
7F
0
1
2
3
4
5
7F
6
6
4
2
0
0.02% Tb
2% Al
900C/6hrs
20000
15000
10000
Fig. C: Tb3+ energy levels
These are the problems we sought to
examine with this project, in hopes
of finding Tb to be a good probe for
learning about rare-earth clustering
and sol-gels in general. Along the
way, we discovered other interesting
phenomena as well.
Another factor that affects 5D3 emission is the concentration
of Tb3+ ions. We hypothesize that higher concentrations of RE
ions result in more clustering and thus more quenching of
fluorescence, and that lower concentrations will tend to show
stronger peaks in the 5D3 range. This can be seen in Figure F.
However, we have noticed that the differences between
subsequent concentration changes become smaller for very
low concentrations. The 5D3 to 5D4 peak ratios (see Figure I
below) are very similar for 0.01% Tb and 0.02% Tb, and both
are typically much better than any other concentration, so it’s
possible that we have found the lower limit of RE
concentration, meaning that the ions are dispersed through the
matrix well enough such that any further reductions in
concentration will not reduce clustering effects.
300000
200000
D3 Emission
100000
22
Energy (1000cm-1)
5
0.02% Tb, 2% Al, 900C/6hrs
250000
D4 Emission
250000
The most interesting phenomenon we noticed was the decay of
5D emission over time after reheating. Heating to 900°C yielded
3
great results with strong 5D3 peaks, but unfortunately, when tested
the next day, there was significantly less fluorescence in that
region. Figure H shows how a sample containing 0.02% Tb and
2% Al decayed over time after being heated to 900°C for 6 hours.
5000
0
400
450
500
550
600
650
Fig. E
One phenomenon that we examined
and found to be consistent with
current literature is the effect of
Aluminum co-doping on the
fluorescence of rare-earth ions in solgels. It is believed that the addition
of Al3+ ions to the matrix will
increase the inter-ion separation
between RE ions and thus reduce the
effects of clustering.
Our experimentation confirmed this
idea, as show in the two graphs on
the left. Figure D is an emission
spectra of a sol-gel pellet containing
0.02% Tb and 0% Al (all percentages
are by molar ratio relative to TMOS).
Figure E is an emission spectra
where the pellet also contains 2% Al.
The 5D3 emission of the sol-gel
containing 2% Al is clearly much
stronger, thus showing that codoping with Al3+ ions helps reduce
quenching effects on 5D3 emission.
(all spectroscopic data obtained on Fluoromax)
0.02% Tb, 2% Al
Intensity (arbitrary units)
30
28
26
24
22
20
18
16
14
12
10
8
6
4
2
Another problem has
to do with residual
hydroxyl (OH-) groups
that remain even after
the annealing process
in sol-gel synthesis.
These ion groups can
affect the fluorescence
of the Tb3+ ions by
releasing and
absorbing energy into
phonons, or tiny
vibrational energies, as
opposed to light.
Energy
(1000cm-1)
In recent years, there has been much research
in the field of rare-earth (RE) dope sol-gel
glasses. These glasses are formed by mixing
water, hydrochloric acid, and either TMOS
(tetramethylorthosilicate)
or
TEOS
(tetraethylorthosilicate),
and
gradually
heating over the span of a few days. This
allows the mixture to gel, and then by letting
it cool, a solid piece of glass is formed.
During the formation of the gels, we can add
ions of rare-earth metals from the lanthanide
series of the periodic table. These elements
frequently have very strong fluorescence in
the visible spectrum.
By exciting the ion’s electrons with
ultraviolet light, the electrons will jump to a
higher energy level, and subsequently drop
down to a lower level, in the process emitting
a photon. These photons are what cause the
gels to glow.
Figure A shows a Dieke diagram depicting
energy levels for all the RE elements:
5
0.02% Tb
0% Al
900C/6hrs
Fig. G:
annealing
temperature
dependence
120000
100000
80000
750C/24hrs
60000
900C/6hrs
40000
20000
0
400
450
500
550
Wavelength (nm)
600
650
Typically our sol-gels are annealed by gradually heating them to 750°C for 24
hours. As an experiment, we tried reheating sol-gel samples to higher
temperatures for 6 hours. For instance, we tried reheating to 900°C, 950°C,
and 1000°C for 6 hours. In terms of qualitative data, heating to 950°C and
1000°C yielded sol-gels that were slightly more cloudy, and less transparent.
Heating to 900°C, though, seemed like a perfect temperature, as it yielded
optically clear samples, and also showed strong 5D3 emission when tested.
Figure G shows a comparison between annealing at 750°C and reheating to
900°C.
23hrs
30hrs
49hrs
100000
5D3 to 5D4 peak ratios chart
50000
3.0
0% Al, 900C/6hrs
0
400
420
440
460
480
500
520
2% Al, 750C/24hrs
2.5
Wavelength (nm)
2% Al, 900C/6hrs
Fig. H: 5D3 emission time decay
2% Al, 900C/6hrs,
aged 3 days
2% Al, 900C/6hrs,
vac uum 2 days
To compare the 5D3 emission of different samples, we created a 5D3 to
5D ratio. We calculated the heights of each 5D peak and one 5D peak,
4
3
4
and then found their ratio. Figure I shows the ratios for 5 sets of
samples that we tested.
It is clear that 2% Al shows more activity than 0% Al, and 900°C/6hrs
shows more activity than 750°C/24hrs. Also, the samples that had aged
3 days since reheating to 900°C show very little activity.
5D3 to 5D4
2.0
1.5
0.5
0.0
0.01% Tb
Our project has examined many characteristics of
sol-gels. We
have confirmed the idea that Al3+ co-doping reduces clustering effects
and improves 5D3 emission. Concentration effects, as well, have been
confirmed, and in fact we believe we have found the lower limit of Tb3+
concentration. Annealing temperature dependence, and the effects of
decay over time after reheating, however, can be explored much further.
Our experiments have shown that after reheating to 900°C for 6 hours,
5D emission gradually decreases (see Figure K), but we are unsure as
3
to why exactly this happens. Perhaps future experiments will be able to
examine the mechanism for this process and discover ways to prevent
it. Finally, the effects of Al/Na co-doping can also be explored further
by trying other combinations of the two ions.
0.02% Tb
0.05% Tb
0.1% Tb
% Tb
Fig. I: 5D3 to 5D4 ratios
5D3 to 5D4 Peak Ratios Over Time (2% Al, 900C/6hrs)
8
7
ratio of 5D3 peak to 5D4 peak
Tb3+
When annealed at 750°C for 24 hours, the samples showed very little
activity. However, when reheated to 900°C for 6 hours, they showed
significantly stronger 5D3 emission (see Figure J). The 2% Na and 1.5%
Na samples, though, still showed little activity. The best co-dopant amount
that we tried appears to be 0.5% Na and 1.5% Al. Perhaps in the future
other combinations of Al3+/Na+ can be tested.
1.0
Finally, we tried reheating a set of samples to 900°C for 6 hours and
then immediately placing them in a secure environment. To do this we
used Schlenkware and a vacuum to create an airtight environment and
to keep the samples unaffected by the atmosphere for 2 days. Then we
removed them and tested their fluorescence. The results showed that
they behaved almost as well as completely fresh samples, and clearly
much better than other aged samples.
Conclusions:
We also experimented with other co-dopants in the sol-gel recipe. One
paper mentioned their results when co-doping with both Al3+ and Na+ ions,
saying that “when both Al3+ and Na+ ions are co-doped at the same time,
the composition dependent PL spectra…revealed that one of them
compensates the spectral change due to the other.”1 We attempted to obtain
similar results by co-doping such that the total concentration of Al3+ and
Na+ ions summed to 2%.
6
5
0.01% Tb
0.02% Tb
4
0.05% Tb
0.1% Tb
3
2
Fig. J: Al3+/Na+ co-doping
1
0
0
10
20
30
40
50
60
# of hours since heating
Fig. K: 5D3 to 5D4
ratios over time
References:
1. K. Itoh, N. Kamata, T. Shimazu, C. Satoh, K. Tonooka, K. Yamada. Journal of Luminescence 87-89 (2000) 676-678.
2. T. Ishizaka, R. Nozaki, Y. Kurokawa. Journal of Physics and Chemistry of Solids 63 (2002) 613-617.
3. R. Reisfeld, T. Saraidarov, E. Ziganski, M. Gaft, S. Lis, M. Pietraskiewicz. Journal of Luminescence 102-103 (2003) 243-247.
Thanks to:
Profs. Silversmith and Brewer, Greg Armstrong, Helena Grabo, Kate Schirmer, Peter Burke