Fluorescence Lifetime Spectrometer

FluoTime 200

Spectral evolution on the sub-nanosecond time scale: Application of TRES measurement mode and global fitting

Pilot measurements of a naphtalimide derivative dissolved in n-buthanol (sample courtesy of Dr. Anatoly Metelitsa, Institute of Physical and Organic Chemistry, Rostov University, Russia) indicated that the decay kinetics strongly depends on the emission wavelength. As an example, three decay curves, measured at the blue edge (480 nm), peak (520 nm) and red edge (560 nm) of the steady-state emission spectrum are shown below:

The measurements were performed on the FluoTime 200 lifetime spectrometer equipped with an MCP-PMT detector, single monochromator, Glan laser polarizers and a PicoHarp 300 TCSPC module. The sample was excited with an LDH-P-C-375 pulsed laser diode head driven at 20 MHz with a PDL 800-B driver. The instrument response function (IRF, plotted with dark blue color in the picture above) has a 52 ps FWHM.

A closer look at the first few nanoseconds after the excitation reveals that there is a fast decay component present when the detection wavelength is set to the blue side of the spectrum. A similar fast rising component is visible at the red edge, at around 560 nm. The rest of the temporal behavior is dominated by an ordinary fluorescence decay on the nanosecond time scale.

In order to get a more complete picture of the wavelength dependence, 41 decay curves were collected in the so-called TRES measurement mode. The monochromator was controlled by the measurement software; the emission spectrum was scanned from 450 to 650 nm with 5 nm steps. At each step, a 20 second decay measurement was performed. The whole measurement thus took less than 15 min.

The result can be regarded as a matrix of wavelength and time dependent intensity values, as plotted below:

The same data set can be also visualized as a 2D false color intensity map. Using the standard software tool delivered with the spectrometer it is possible to slice the matrix data along the wavelength dimension. This gives us a quick overview on the time evolution of the emission spectrum:

The above picture shows two normalized spectra, separated in time by less than a nanosecond. It may look like that the early-time spectrum peaking at 490 nm is being gradually red-shifted as the time proceeds. However, a closer inspection shows that this is not the case. By plotting many such time-slices as they follow each other in time it becomes apparent that there is an interplay of two emission bands. The one at 490 nm quickly vanishes giving rise to a new band peaking at 530 nm as can be seen below:

A more quantitative description of this process can be obtained by global analysis of the decay curves, using the FluoFit Professional software package. In this case, a satisfactory description (global χ² = 1.103) of the whole dataset of 41 curves can be obtained using only three global lifetime parameters. The wavelength dependence of the exponential amplitudes are plotted below:

The black, blue and red curves correspond to the amplitudes A₁, A₂ and A₃ of the recovered global lifetime components, 86 ± 5 ps, 930 ± 100 ps and 3.19 ± 0.02 ns, respectively. The spectacular dependence of A₁ on the emission wavelength is compatible with the spectral evolution revealed by TRES data slicing.

Conclusions:

The observed consistent behavior of exponential amplitudes completely rules out the possibility that the complex decay behavior is caused by impurities. It supports the conclusion that an excited state process with a rate constant on the order of 10¹⁰ s^-1 changes the initially prepared blue emitting states to those with green emission. Although solvent relaxation certainly also contributes by an overall red shift, this process alone cannot cause the observed spectral changes.

Instrument related remarks:

A pre-requisite of fast automated TRES measurement is a motorized monochromator which can be controlled by the timing electronics. The PicoHarp 300 TCSPC module, as well as the NanoHarp 250 and TimeHarp 200 PCI boards are able to control various stepper motor driven monochromators, either through CAN-bus or via the conventional serial (COM) port.

In practice, the information content of a TRES data set is limited only by the available measurement time, which can be strongly limited e.g. by sample stability. To obtain the sufficient spectral resolution, small wavelength steps and consequently a dozens of histograms are needed. The time-resolution of the decay curves should be finer than the time scale (i.e. speed) of the expected spectral evolution. The finer the time resolution, the longer it takes to collect the necessary amount of counts to get a reasonable data statistics in the histogram channels. It turns out that the signal count rate must be maximized up to the limit of the TCSPC method, which is determined as 1-2% of the excitation repetition rate. In fact, the 20 MHz diode laser excitation is essential for obtaining the above reported TRES in real time. To get the same data quality with 1 MHz excitation would take almost 5 hours. And, theoretically, 5 days with a thyratron-gated nanosecond flash lamp.

Peter Kapusta, December 2006