ХИМИЧЕСКАЯ ФИЗИКА, 2004, том 23, № 2, с. 24-28
ЭЛЕМЕНТАРНЫЕ ^^^^^^^^
ФИЗИКО-ХИМИЧЕСКИЕ ПРОЦЕССЫ
УДК 539.184
OBSERVATION OF ELEMENTARY NONADIABATIC PROCESSES IN ATOM-ATOM AND ATOM-MOLECULE COLLISIONS
© 2004 r. C. Figl*, R. Goldstein*, J. Grosser*, O. Hoffmann*, F. Rebentrost**
* Institut für Atom- und Molekülphysik, Universität Hannover, 30167 Hannover, Germany **Max-Planck-Institut für Quantenoptik, 85748 Garching, Germany Received 16.11.2002
Optical excitation of collision pairs into a definite molecular state and analysis of the final atomic state gives a direct access to the nonadiabatic processes which occur at the convergence of the molecular potential curves before the collision partners separate. We describe the method, give experimental examples and discuss different nonadiabatic mechanisms.
1. INTRODUCTION
Collisions between atoms and molecules are frequently discussed in terms of a single electronic state and therefore a single potential surface of the collision-al molecule. The transitions between electronic states which occur in the entrance channel of a collision at large distance can however be of decisive importance for the subsequent pathway of the process; this may even cover the possible outcomes of chemical reactions [1]. Nonadiabatic transitions can be observed in standard scattering experiments. However the interpretation may be complicated because the experimental data usually reflect the combined action of different colli-sional mechanisms. The isolated study of elementary nonadiabatic events has become possible by the optical excitation of collision pairs [2-4]; beam experiments of this type have turned out to be particularly useful [5]. We report here differential scattering studies of the processes.
Na(3s) + M + hv — Na(3p) + M,
with atomic and molecular targets M. The method permits the isolated observation of nonadiabatic transitions in the exit channel of collisions. Transitions in the entrance channel of full collisions are simply the time inverse of these processes.
2. EXPERIMENTAL METHOD AND RESULTS
The experimental setup [6] consists two particle beams, the laser beam for the excitation of the collision pairs, and the differential detection system, see Fig. 1. After the collisional excitation, the excited atomic products are transferred to a Rydberg state by a second laser as the first step in the detection process,
Na(3pi/2, 3/2) + hv' —- Na(nd),
with n between 25 and 40. After travelling to a rotatable detector at 70 mm distance, the Rydberg atoms are counted using field ionization and single particle detec-
tion. The transfer to a Rybderg state serves to stabilize the excited Na atoms, which otherwise would decay to the ground state before reaching the detector.
Of particular significance is the fact that the experimental procedure permits to separately detect the two fine-structure levels by tuning the detection laser to a corresponding transition. Fig. 2 shows our experimental results for different collision partners M. The experimental data points show the fraction of the excited Na atoms which are found in the Na(3p1/2) state. They are plotted as a function of the relative kinetic energy of the atoms after the collision. The measurements were carried out for detunings of 120, 360, and 480 cm1 (we use the detuning A = (v - vres)/c, where vres is the Na(3s1/2 - 3p1/2) transition frequency to characterize the laser frequency). The laboratory scattering angle was fixed at 10.8° for the Na + rare gas collisions and at 18° for the rest.
We turn first to a more detailed discussion of the experimental accuracy. The finestructure state populations were obtained by scanning the wavelength of the detection laser over the corresponding lines and taking the area as the measure of the population. The same upper state Na(nd) was used for the detection of both fine-
Fig. 1. The principle of the experimental set-up with two particle beams, two laser beams, and a rotatable differential detector.
P1/2 fraction, %
Fig. 2. The data points show the experimental Na(3p1/2) fraction for different targets as a function of the relative energy after the collision. The numbers in the boxes indicate the detuning of the pump wave length form the atomic 3s-3pi/2 resonance and the laboratory scattering angle. The lines are theoretical results.
structure components. The lifetime with respect to spontaneous decay is the same for both fine-structure levels; similarly, the optical transition probabilities for the detection transitions are identical, except for polarization effects. Except for the last point, the procedure gives therefore a correct measure for the population, for low light intensities, where the signal increases linearly with the intensity of the detection laser, as well as for higher intensity, where the signal becomes saturated. The results in Fig. 2 were obtained at high intensity under conditions in which a 10% variation of the detection laser intensity leads to a typical variation of the signal by 3-4%. A possible orientation or alignment of the products can lead to a systematic error in the measured population, which should however not exceed the order of a few percent. The probability for a Rydberg atom to survive the 7 cm trip to the detector depends on numerous effects like optical decay, collisions, and transitions bye thermal background radiation, which are not completely under control. The Rydberg atoms produced from the 3p1/2 or 3p3/2 states have identical quantum numbers n and 1, but the total angular momentum j may be different. However, the spin-orbit interaction in the Rydberg states is extremely small. We expect therefore the stability to be identical in both cases. Fine structure
changing secondary collisions, which occur in the time between the production and the detection of the collision products, form a problem in gas cell optical collision experiments [4]; as our target density is considerably lower, this is without significance at present. The laser intensities are monitored outside the vacuum chamber, and the signal is corrected for intensity variations. We checked the agreement of the laser intensity measurement with the intensity in the scattering volume separately. It was indeed necessary to use anti-reflex coated windows to achieve a satisfactory performance. A background signal of typically 5 to 20% of the signal had to be subtracted from the measured line intensities. There are various background contributions
[7]. The background can be measured with the target beam switched off. However, it must be expected that the background intensity varies with the target density. An elaborate procedure for the background subtraction is therefore required; details are discussed elsewhere
[8]. The error bars in Fig. 2 contain the statistical error and a contribution corresponding to 50% of the subtracted background signal to cover possible systematic errors in the background correction. Further systematic errors should be below 3%, altogether.
XHMHHECKA£ OH3HKA tom 23 № 2 2004
Potential energy, cm 1 400
200
0
-200 -
10 20 30
Atom-atom distance, a.u.
Potential energy, cm 1 400
200
-200
10 20 30
Na-N2 distance, a.u.
Fig. 3. Convergence region of the NaAr potential curve system. The arrows indicate the excitation and detection transitions.
Fig. 4. Convergence region of the NaN2 potential curve system; the data shown refer to the T-shaped geometry. The arrows indicate the excitation and detection transitions.
3. DISCUSSION: ATOM-ATOM COLLISIONS
Figure 3 shows the relevant region of the potentials for Na + Ar as a representative example; the data were calculated from spin-free potentials [12], assuming a constant spin orbitinteraction. As the photon energy is above the atomic resonance, we excite the B2X state during the collision. Completely adiabatic behaviour (on transitions between different potentials) would lead after the collision to an exclusive population of the Na(3p3/2) state, unlike experimentally observed. However, when the particles leave, they pass through a region where the potential curves converge and where nonadiabatic transitions between the electronic states are possible, leading to population of both the Na(3p1/2) and Na(3p3/2) states after the collision, in accordance with the experimental findings. The results for atom-atom collisions (the upper row in Fig. 2) demonstrate an astonishing variability, with nonadiabatic transition probabilities ranging from 20 to 70%.
The nonadiabatic coupling region lies between the regime of Hund's cases a and b which apply at small atom-atom distance, and that of cases c and e, applying at large distance [9, 10]. There is a unique high energy limit for this case, a Na(3p1/2) population of 1/3 of the total Na(3p) population [5], corresponding to a sudden (completely diabatic) passage through the coupling region. The low energy limit is zero, corresponding to completely adiabatic behaviour. The curves in Fig. 2 will therefore go from 0 at low energy to 1/3 at high energy. The NaAr potentials in Fig. 3 show two different coupling mechanisms. First, there occurs an avoided crossing between the two curves with angular momentum projection Q = 1/2, and second, transitions are possible once again near the ultimate convergence. At in-
termediate energy, the avoided crossing is very effective in transferring the system from the uppermost to the lower potential, leading to a large N(3p1/2) population at the end. As a whole, the Na(3p1/2) population curve goes therefore from 0 at low energy over an intermediate maximum to the high energy limit of 1/3; experimentally, we see only the decreasing part of the curve. NaKr shows a quite similar behaviour and has indeed a similar structure of the potential curve system. NaNe has no avoided crossing; there is no intermediate maximum, therefore, only a monotonous increase from 0 to 1/3. KAr shows a similar behaviour; the fact that the limit 1/3 is
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.