ХИМИЧЕСКАЯ ФИЗИКА, 2004, том 23, № 2, с. 41-45
ЭЛЕМЕНТАРНЫЕ ^^^^^^^^^^
ФИЗИКО-ХИМИЧЕСКИЕ ПРОЦЕССЫ
УДК 539.1
FEMTOSECOND CROSSED BEAM STUDY OF ATOM-ATOM COLLISIONS
© 2004 r. R. Goldstein, J. Grosser, O. Hoffmann, H.-J. Schmidtke, D. Wößner
Institut für Atom- und Molekülphysik, Universität Hannover, 30167 Hannover, Germany
Received 16.11.2002
We discuss femtosecond pump-probe observations of thermal Na + Kr collisions using a crossed atomic beam arrangement. We report the results of a first experiment and describe possible future applications.
1. INTRODUCTION
The usual experimental approach to collision processes in indirect. The collision partners before and after the collision are the subject of observation, but not the collisional event itself. In the last decades, two powerful tools were developed which both aim at the direct observation of atomic or molecular processes. First, femtosecond pump-probe techniques are widely applied to observe the time evolution of molecular processes [1]. Second, optical excitation of collision pairs in a crossed beam arrangement [2] allows a detailed observation of collision processes and can in particular yield geometric images of a collision event [3]. We describe here an experiment which combines the two methods by applying the femtosecond pump-probe technique to an atom-atom collision, using atomic beams and a differential detection of the collision products [4]. This new approach brings time resolution and the capability for geometric observations together and should ultimately provide experimental results with the character of a motion picture of a collision. We report on the first observation of a detectable signal, we describe the major experimental problems, and we discuss the possibilities for future experiments.
2. PRINCIPLE
Figure 1a shows the potential curve diagram [5] of the Na + Kr collision pair and the pump-probe scheme. The system starts with the atoms at large distance on the lowest X2E potential curve. At shorter distance, the collision pair is excited to the A2n state by the first femtosecond pulse and subsequently to the CX state by the second one, leading to the production of a Na atom in the 4s state after the collision, which we detect with a scattering angle sensitive detector. Fig. 1b shows a classical trajectory representation of the collision based on numerical computation for the actual experimental conditions. The dots on the trajectories correspond to time intervals of 50 fs. The velocity directions before and after the collision are fixed in our experiment. The four trajectories in the figure are the only ones which connect the given initial and final velocity directions, and only these trajectories contribute to the experimental
signal. The optical transitions are subject to resonance conditions, hu> = V(r) - V(r), with Vf and V the potential curves. The corresponding transition points are shown in the figure. The travelling times between the transition points are indicated as numbers. Actually, transitions can occur in ranges with typical widths of 140 fs around the transition points. A detectable signal is expected only when the delay between the pump and
30000
i? 20000
<D С
<D
s 10000
iß о
. 1 1 C2z a Na(4s) + Kr
А2П hv2 Na(3p) + Kr
\---- X2Z hv1 Na(3s) + Kr
с
А
1.0
0.5 0
-0.5 -1.0
0.2 0.4 0.6 0.8
Atom-atom, nm
b 42 fs
..............*P'".o-"" 330 fs
.........■•.....-*
..............• 0......... 1 1 358 fs 74 fs
-0.5
0
х, nm
0.5
1.0
Fig. 1. The scheme of the pump-probe experiment. a - The potential curves of the NaKr collision pair with the pump and probe transitions. The horizontal lines show the total (potential + relative kinetic) energy of the collision pair. b -Calculated trajectories with transition points. The numbers indicate the travelling time between the pump and the probe transition.
0
42
GOLDSTEIN h «p.
Fig. 2. The experimental setup with two atomic beams, pump and probe fsec laser beams, the nsec detection laser beam, and a rotatable differential detector for the excited Na atoms. Na atom velocities before and after the collision are measured by time of flight techniques.
probe pulses coincides with one of the travelling times within the corresponding range.
3. EXPERIMENT
Figure 2 shows the experimental setup with the crossed atomic beams, the femtosecond laser beams, and a rotatable detector. As a part of the detection system, we use a third (nanosecond) laser, which stabilizes the Na(4s) atoms immediately after the collision by transferring them to a Rydberg state Na(np), with typically n ~ 30. In contrast to Na(4s) atoms, Na Rydberg atoms live long enough to survive the 7 cm trip to the detector. The Rydberg detection scheme was devel-opped earlier, see Ref. [6]. The characteristic features of the five beams are collected in Table 1.
The major experimental problem is the small signal level, which is primarily due to the extremely small density of collision pairs. On the average there are typically only 100 pairs of Na and Kr atoms in the scattering volume, which have a distance of 1 nm or less and are therefore candidates to be excited by the fsec laser pulses. Out of these, a fraction below 10-4 is close enough to the first transition point when the first fsec pulse arrives, and has the right impact parameter to be scattered to the differential detector. Furthermore, the efficiency of the detection scheme is of the order of a few percent. In order to obtain a reliable estimate of the expected signal, independent of the limited precision to
Table 1. The intensities and pulse durations of the five beams. The laser intensities refer to the temporal and spatial beam maxima
Na beam 3 ■ 1010 atoms/cm3 20^s
Krbeam 1014 atoms/cm3 continuous
pump laser 1.2 GW/cm2 100 fs
probe laser 0.8 GW/cm2 100 fs
nsec laser 0.6 MW/cm2 15 ns
which we know quantities like detection probabilities and beam intensities, we scaled from a similar experiment which differs from the present one only in one detail, namely by using a single nsec laser pulse for the excitation of the collision pairs in place of the pumpprobe pair of femtosecond laser pulses [6]. All other details are so similar that scaling is straightforward. For the experimental conditions applied here, we expect on this basis a signal intensity of 10-4 counted atoms for every laser shot. The number should apply within a factor of 3; this reflects uncertainties in the scaling procedure as well as the typical variations in the performance of the apparatus.
4. COMPETING SIGNALS
The low signal level makes the experiment extremely sensitive to competing signals. Competing signals originate mainly from optical excitation processes which are different from the desired collisional mechanism. In particular, there are 3 ■ 108 Na beam atoms in the interaction volume at any time. A small probability to excite these atoms to a Rydberg state may result in a number of Rydberg atoms which is large compared to the expected signal. According to our present experience, there are mainly two undesired pathways to excite Na atoms from the ground state to a Rydberg state. The first one is the subsequent absorption of three photons, one from each laser beam, leading from the 3s state to 3p, 4s and finally Rydberg atoms. The fsec laser pulses are detuned from the corresponding atomic resonance frequencies, but they contain the resonance frequencies in their spectra. It is well known [7] that under the conditions of high laser intensity, it is not only the resonance frequency itself which can excite a free atom, but also neighbouring frequencies in a band with the width of the Rabi frequency Q = Ed/ft, where E is the amplitude vector of the electric field and d = = (TJXerJ^) the transition dipole moment between the states in question. It is therefore necessary to suppress a correspondingly broad spectral range. The intensities of the two lasers and the corresponding Rabi frequencies are collected in Table 2. Filtering is achieved for both pump and probe pulses by passing them through a pair of dispersion prisms. Filtering makes the pulses longer, typically we have 50 fs pulses before and 100 fs pulses after filtering. We can easily monitor this type of undesired signal by running the experiment with a large delay (above 1 ps) between pump and probe pulses. It turns out that this competing signal can be reduced practically to zero by adequate filtering of both pulses. A second competing signal is observed only under the conditions of a temporal overlap of the fsec laser pulses and has therefore clearly another origin. We ascribe it to the direct excitation of Rydberg states by three photons from the fsec laser pulses. This mechanism forms a severe limitation at present, which prohibits in particular measurements with small pumpprobe delay. The scattering experiment was carried out
with the detector at an angle of 11°. Atoms from the atomic beam, which are excited by a competing mechanism of the sort discussed above, will normally not arrive at the detector under these conditions. However, about 1% of the beam intensity is scattered towards the detector by the target gas. Excitation by a competing mechanism plus scattering can give rise to undesired signals even under the conditions of the scattering experiment.
5. EXPERIMENTAL RESULTS
The collision pair looses an amount of kinetic energy by the pump-probe process which is equal to the sum of the fsec laser detunings, 700 cm1. Excited atoms produced by optical excitation of Na beam atoms keep their kinetic energy. When they are scattered to the detector, a small energy change of purely kinematic origin occurs. We measure the velocities of the Na atoms before and after the collision by time of flight techniques. Starting from a measured velocity distribution in the beam and applying a straightforward kinematic calculation, we can predict the velocity distributions after the collision, both for atoms pro
Для дальнейшего прочтения статьи необходимо приобрести полный текст. Статьи высылаются в формате PDF на указанную при оплате почту. Время доставки составляет менее 10 минут. Стоимость одной статьи — 150 рублей.