Principles of Fluorescence Spectroscopy

PRINCIPLES OF FLUORESCENCE SPECTROSCOPY 129
The high time resolution of upconversion measurements
can be illustrated by quenching and exciplex formation
of 3-cyanoperylene (PeCN) in neat diethylaniline
(DEA). 167–168 The cyano group makes perylene a good electron
acceptor, and DEA is a good electron donor. PeCN in
the excited state undergoes photoinduced electron transfer
(PET) with DEA (Chapter 9). This reaction is very rapid
because in DEA the PeCN molecules are surrounded by
potential electron donors. Emission spectra of PeCN show
that it is highly quenched in DEA as compared to a nonelectron-donating
solvent (Figure 4.43). The lower panel
shows the wavelength-dependent decays of PeCN. The time
resolution is limited by the pulse width of the pump lasers,
which is near 100 fs. The decay time is about 1 ps for the
PeCN emission near 500 nm and about 20 ps for the exciplex
emission.
4.8.4. Microsecond Luminescence Decays
For decay times that become longer than about 20 ns, the
complexity of TCSPC is no longer necessary. In fact,
TCSPC is slow and inefficient for long decay times because
of the need to use a low pulse repetitive rate and to wait a
long time to detect each photon. In the past such decays
Figure 4.43. Emission spectra and time-resolved decays of 3-cyanoperylene
in chlorobenzene (CB) and diethylaniline (DEA). Revised
from [168].
Figure 4.44. Time-resolved intensities measured with a photon counting
multiscalar. Revised from [169].
would be measured using a sampling oscilloscope. At present
the preferred method is to use a multiscalar card. These
devices function as photon-counting detectors that sum the
number of photons occurring within a time interval. Figure
4.44 shows the operating principle. Following the excitation
pulse, photons arriving within a defined time interval are
counted. The arrival times within the interval are not
recorded. The time intervals can be uniform, but this
becomes inefficient at long times because many of the intervals
will contain no counts. A more efficient approach is to
use intervals that increase with time. The count rates for
multiscalars can be high: up to about 1 GHz. At present the
minimum width of a time interval is about 1 ns, with 5 ns
being more typical. Hence, the multiscalars are not yet
practical for measuring ns decays.
An example of a decay measured with a multiscalar is
shown in Figure 4.45. The fluorophore was [Ru(bpy) 3 ] 2+ ,
which has a lifetime of several hundred nanoseconds. A
large number of counts were obtained even with the long
lifetimes. However, the count rate is not as high as one may
expect: about 1 photon per 2 laser pulses. The data acquisition
time would be much longer using traditional TCSPC
with a 1% count rate (Problem 4.6).
4.9. DATA ANALYSIS: NONLINEAR LEAST
SQUARES
Time-resolved fluorescence data are moderately complex,
and in general cannot be analyzed using graphical methods.
Since the mid-1970s many methods have been proposed for
analysis of TCSPC data. These include nonlinear least
squares (NLLS), 3,16,170 the method-of-moments, 171–173
Laplace transformation, 174–177 the maximum entropy