756
chrotron [5] or laser [13, 14] excitation in combination with
photon counting detection techniques. Polarization-sensitive
fluorescence detection is often used to gather information on
rotational motion by investigation of the anisotropy decay.
While temporal resolution is thus ensured, most experiments
provide fluorescence information at a single or a few in-
dividual emission wavelengths, where bandpass filters may
often be used in a wavelength-integrative manner to en-
sure high photon count rates [15–18]. Important information,
i.e. on the detailed photophysical processes responsible for
the emission, may be lost by this approach. This has been
demonstrated by several researchers [5, 13, 14], who have ob-
served a pronounced change of fluorescence lifetime with
wavelength which may devalue results measured at a single
emission wavelength or with broadband filters [5].
In the present investigation, a novel approach providing
simultaneous time and wavelength resolution is demonstrated
in picosecond laser fluorescence experiments using trypto-
phan as intrinsic probe. This two-dimensional approach pro-
vides direct correlations between the temporal and spectral
fluorescence behavior. In addition, a second two-dimensional
technique is used, excitation-emission spectroscopy, which
provides complementary information and allows us to de-
tect energy transfer processes. With a combination of these
two powerful approaches, four independent discriminative
variables are addressed which govern the photophysical pro-
cesses – excitation and emission wavelength as well as in-
tensity and temporal decay, and unparalleled information is
obtained from the observed correlations. To demonstrate the
potential of these techniques, several examples are given in
biochemical systems which have been well investigated be-
fore. Conformational rearrangements are observed in an ex-
periment involving the Ras protein and several nucleotides,
energy transfer is studied in tryptophan/tyrosine mixtures,
and static and dynamic quenching is investigated in calse-
questrin at different Ca2+ concentrations.
From this, between 50 µJ and 250 µJ are split off by a quartz
plate and are used for the actual measurements; the remain-
der is directed to an energy meter and serves as reference
for the pulse energy. The temporal shape of the pulse is ob-
tained with an autocorrelator, and the wavelength position
is controlled with a dual-bandwidth wavemeter (Burleigh).
A more detailed description of the optical setup can be found
elsewhere [19, 20].
The fluorescence emission is collected with a spherical
mirror ( f = 250 mm, d = 250 mm) to avoid chromatic aber-
rations when using a large spectral detection range. A cut-off
filter (Schott WG 320, 3 mm) at 320 nm is used to block
Rayleigh scattering and scattered light from optical surfaces.
The emission is spectrally dispersed with a 275-mm imag-
ing spectrometer (Acton Research Corp., SpectraPro 275i)
with a 150-l/mm grating (blaze wavelength 500 nm). Tempo-
ral resolution is achieved with a streak camera (Hamamatsu
C2830), the output of which is amplified by a MCP, im-
aged on a phosphor screen and coupled to a CCD camera
(C4880) where the data is stored and transmitted to a com-
puter. The image detected by the streak camera has a size of
1000×1018 pixels, it is decoded by the HPD-TA software
supplied by Hamamatsu. Available time intervals are 0.5, 1, 2,
5, and 10 ns.
1.2 Calibration and evaluation of time-resolved spectra
For a typical fluorescence measurement using the picosec-
ond laser/streak camera setup, the emission of 1800 con-
secutive laser pulses is integrated corresponding to a time
period of 3 min for each experiment. From each measure-
ment, a two-dimensional image is obtained with a time and
a wavelength axis; intensities are stored with 16 bit resolution
by the CCD chip. The time and wavelength axes correspond
to the selected spectrometer dispersion and streak interval;
for most of the experiments, a wavelength range of 200 nm
(λem = 300–500 nm) and a time interval of 10 ns is chosen.
The wavelength span and the absolute spectral position are
determined using a Hg calibration lamp and Rayleigh and Ra-
man scattering with different excitation wavelengths between
250 and 350 nm; the Raman signature of water (Raman shift
of 3450 cm−1) is used as a reference in all images. Since the
efficiency of the spectrograph grating (blaze 500 nm) and of
the camera are not constant over this large wavelength re-
gion, a calibration is performed using a deuterium lamp, and
a halogen lamp with a blackbody temperature of 3200 K. To
investigate potential damaging effects on the biomolecular
samples by the irradiation of laser light, a series of experi-
ments is conducted with attenuating the laser beam by one
order of magnitude. Also, measurements are performed re-
peatedly with the same sample. No indication of deterioration
of the samples is found in any experiment.
1 Experimental
1.1 Time- and wavelength-resolved measurements
Time-resolved measurements are performed using a short-
pulse laser system in conjunction with an intensified streak
camera. The laser system (Spectra Physics) consists of
a Ti:sapphire laser (Tsunami) pumped by a 7-W Ar+ laser. It
is capable of producing pulses with either 3 ps or 80 ps du-
ration in the near infrared (between 750 and 900 nm) with
a repetition frequency of 80 MHz, individual pulses have an
energy of approximately 10 nJ. For the present application,
80-ps pulses with considerably higher energy in the UV are
desired. A regenerative amplifier TSA-50 and two linear am-
plifiers are used; they are pumped with the second-harmonic
radiation of two Nd:YAG lasers. With amplification factors
of 105, 2.5 and 4, respectively, the resulting pulses have typ-
ical energies of 15 mJ at about 850 nm and a repetition rate
of 10 Hz. Consecutive frequency-doubling and mixing of the
second harmonic with the fundamental wavelength provides
UV radiation tunable between 250 nm and 300 nm. For the
experiments described in this paper, wavelengths near 280 nm
are used; the pulse length was 80 ps, and a bandwidth of about
0.7 cm−1 and energies between 1 mJ and 2 mJ are reached.
Two-dimensional raw data images are processed by as-
signing the corresponding time and wavelength scales. Back-
ground luminescence is measured with the respective pure
buffer solution and subtracted from the fluorescence images.
Furthermore, a smoothing procedure is used averaging in-
formation from 3×3 pixels. To account for the wavelength
dependence of the detection efficiency, the images are divided
by the appropriate calibration file. For the evaluation of spec-
tra and of temporal decays from the processed fluorescence