Chinese Journal of Chemical Physics  2017, Vol. 30 Issue (6): 614-618

The article information

Tian-gang Yang, Long Huang, Tao Wang, Chun-lei Xiao, Dong-xu Dai, Xue-ming Yang
杨天罡, 黄龙, 汪涛, 肖春雷, 戴东旭, 杨学明
Efficient Preparation of D2 Molecules in v=2 by Stimulated Raman Pumping
Chinese Journal of Chemical Physics, 2017, 30(6): 614-618
化学物理学报, 2017, 30(6): 614-618

Article history

Received on: November 15, 2017
Accepted on: December 8, 2017
Efficient Preparation of D2 Molecules in v=2 by Stimulated Raman Pumping
Tian-gang Yanga,c, Long Huanga,b, Tao Wanga, Chun-lei Xiaoa, Dong-xu Daia, Xue-ming Yanga     
Dated: Received on November 15, 2017; Accepted on December 8, 2017
a. State Key Laboratory of Molecular Reaction Dynamics, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116023, China;
b. University of Chinese Academy of Sciences, Beijing 100049, China;
c. School of Physics and Optoelectric Engineering, Dalian University of Technology, Dalian 116024, China
*Author to whom correspondence should be addressed. Chun-lei Xiao,; Xue-ming Yang,
Part of the special issue for "the Chinese Chemical Society's 15th National Chemical Dynamics Symposium"
Abstract: We report the preparation of D2 molecules in v=2 level in molecular beam condition. A single longitudinal mode laser system was used for excitation of D2 from (v=0, j=0) to (v=2, j=0) with the scheme of stimulated Raman pumping. An excitation efficiency of 25.2% has been achieved, which was determined by the scheme of resonance-enhanced multiphoton ionization. Dependence of relative excitation efficiency on laser energy has been measured. We found that the increasing rate of excitation efficiency became slower as pulse energy of Stokes laser increase, while the excitation efficiency still increases approximately linearly with pump pulse energies up to 60 mJ. The spectral line shapes of Raman transition was also measured at different laser energies and considerable dynamical Stark effect was observed. A single peak was found on the three dimension surface of relative excitation efficiency, indicating the process occurred in the present study is a process of stimulated Raman pumping instead of stimulated adiabatic Raman passage.
Key words: Stimulated Raman pumping    D2    Vibrational excitation    Molecular beam    

The dynamical process of a chemical reaction is largely determined by the initial state of reactants. So it is of great importance to develop techniques for preparing reactant molecules with well-defined motions, such as vibration, rotation and translation, as the precondition to investigate the relation between reactivity and motion of reactants, which is the key to understand and regulate chemical reactions [1-4].

In molecular beam experiments, a reactant molecule is isolated from others so it is possible to prepare it with well-defined velocity and internal quantum state(s). After several decades of development, it is now rather routine to control the velocity of molecules in the beam by changing the nozzle temperature, and/or by premixing carrier gases. In recent years, new schemes of controlling the translational motion of molecules in the gas phase that rely on the use of inhomogeneous electric or magnetic fields are rapidly developed. Two schemes in particular standing out are Stark deceleration and Zeeman deceleration [5, 6], which are capable of not only precisely tuning the velocity but also selecting the internal quantum state at the same time. In order to prepare molecules in specific internal states, especially rovibrationally excited state, a variety of optical pumping techniques have been developed, including infrared absorption (IR) [7], stimulated Raman pumping (SRP) [8], and stimulated emission pumping (SEP) [9]. These techniques provide us the opportunity to examine quantum state resolved reactivity in bimolecular and molecule-surface reactions. For example, Zare's and Crim's group prepared vibrationally excited H$_2$O/HOD with overtone IR excitation and pioneered the bond-selective chemistry [10, 11]. Kopin and colleagues prepared CH$_4$ and its isotopic variants in the first vibrationally excited state by nanosecond IR laser and investigated the influence of molecular vibration on the dynamics of polyatomic reactions [2, 12]. Recently, Beck and coworkers developed the methods of rapid adiabatic passage with a high-power single-longitudinal mode continuous-wave infrared optical parametric oscillator (OPO) to transfer H$_2$O/CH$_4$ to a specific rovibrationally excited state and then explore the quantum state resolved reactivity and steric effects in the chemisorption on single crystal metal surface [13, 14]. In the last few years, we took advantage of single-longitudinal mode nanosecond pulsed OPO to efficiently transfers hydrogen molecules and its isotopic variants to $v$=1 level [15, 16] and examined the effects of vibrational excitation on dynamical resonances in Cl+HD reactions [17]. We found the effect of chemical bond softening may introduce potential wells in the transition state region and therefore result in reaction resonances. The mechanism of chemical bond softening is expected to be general in the reaction with a barrier when the reactant molecules are vibrationally excited. It should be noted if the reactant starts from a higher vibrational level, the effect of chemical bond softening is anticipated to be more evident. So it is of great interest to develop the technique for transferring molecules to a selected high vibrational energy level.

There are a few mature methods for preparing highly vibrationally excited molecules. For molecules with IR-active vibrational mode such as H$_2$O, HCl, CH$_4$ etc., direct overtone excitation or sequential IR excitation are applicable [7, 11]. For molecules with suitable electronically excited state such as OH, NO, CN, very high vibrational levels could be reached by SEP [9, 18, 19]. However, neither IR excitation nor SEP could be applied to molecular hydrogen. Because it is IR-inactive, and the lowest electronically excited state is about 11 electron volt above the electronically ground state, it is too high to reach in SEP scheme, there is no a suitable light source in the vacuum ultraviolet regime. In the past few years, we have developed the scheme of stimulated Raman pumping for $v$=1 excitation [15, 16]. In this work, we will describe the preparation of D$_2$ ($v$=2) by direct Raman overtone excitation.


The experimental apparatus and procedures for vibrational excitation of hydrogen molecules and its isotopic variant have been described in detail previously [15, 16]. The D$_2$ gas sample (Cambridge Isotope Laboratories, 99.8% purity) was used without further treatment. A D$_2$ molecular beam was formed by supersonic expansion through a pulsed valve and collimated by a circular skimmer (Beam Dynamics, orifice diameter=1 mm) and then went into a detection chamber in which it was intersected by three laser beams at the interaction region, as shown in FIG. 1 of Ref.[16]. Two lasers were used for optical pumping of D$_2$ molecules. A red laser at 655 nm which serves as the pump laser for Raman excitation was generated by a home-made single-longitudinal mode OPO [20], which was pumped by an injection-seeding Nd:YAG laser (Continuum, Powerlite 8000 DLS). An IR laser at 1064 nm was produced by another injection-seeding Nd:YAG laser (Continuum, Powerlite 8000 DLS) and served as Stokes laser for Raman excitation. The pulse width (FWHM) of the Raman pump and Stokes laser are about 6 and 8 ns, respectively. With a dichroic mirror the pump laser and Stokes laser were combined collinearly and then perpendicularly intersected with the D$_2$ molecular beam. A spherical lens of 300-mm focal length was used to focus the two laser beams before entering the vacuum chamber. Before combination with Stokes laser, the pump laser passed through a telescope system for separately adjusting the spot size of the visible laser at the interaction region. Both laser beams were vertically polarized. Because they originated from two independent Nd:YAG laser, the pulse energy and time delay could be changed separately. A dye laser (Sirah GmbH, PESC-G-24) was pumped by the second harmonic output of a third Nd:YAG laser (Spectra-Physics, Quanta-Ray Pro 290) and generated a tunable laser radiation in the range of 603$-$641 nm, which was converted to deep UV in the range of 201$-$213 nm by tripling. The UV laser beam was separated from the fundamental and the second harmonic beams by a triangular prism and then guided to the chamber using three quartz right-angle prisms. The UV laser was tightly focused on a spot with a diameter of 0.1 mm at the interaction region and served as the probe laser. Using the scheme of (2+1) resonance-enhanced multi-photon ionization (REMPI) via the $E$/$F^1\Sigma_{\rm{g}}^+$ electronic state [21], the UV beam ionized D$_2$ in a specific rovibrational level and the resulting D$_2$$^+$ ions were repelled by a DC electric field to a micro channel plate (MCP) detector. By integrating the area of the peak corresponding to D$_2$$^+$ in the time-of-flight spectrum, the relative population of D$_2$ in the selected state was obtained. In order to avoid saturation, the pulse energy of the UV probe laser was kept below 0.7 mJ. In the experiment, wavelengths of involving lasers were monitored by a high resolution wavemeter (HighFinesse, WS7 with the multichannel option). The timing of laser pulses was measured with a fast photodiode and an oscilloscope (Tektronix TDS5054).

FIG. 1 (a) REMPI spectral peaks of D$_2$($v$=2, $j$=0) level, with SRP laser on and off, respectively. (b) REMPI spectral peaks of D$_2$($v$=0, $j$=0) level, with SRP laser on and off, respectively. Solid circles and squares are experimental data points, and solid lines are B-spline interpolations. A 25.2% depletion of the peak area indicates that 25.2% population of the D$_2$($v$=0, $j$=0) is pumped to ($v$=2, $j$=0) level by SRP pulses.

By tuning the OPO, we changed the wavelength of the pump to meet the Raman transition of D$_2$ ($v$=0, $j$=0) to D$_2$ ($v$=2, $j$=0). The pulse energy of pump was kept at 60 mJ which is the maximal energy we could obtain from the OPO. The pulse energy of Stokes was 300 mJ. The temporal delay, and the difference of photon energy between pump and Stokes pules were carefully optimized. The maximal signal appeared when the two pulses were fully overlapped in time and the difference of photon energy between pump and Stokes pules equals to 5868.050 cm$^{-1}$. The probe laser scans though REMPI line of D$_2$ ($v$=2, $j$=0) level and REMPI spectrum was obtained, as shown in FIG. 1(a) (the red curve). If we blocked Stokes laser beam for Raman excitation, the REMPI signal disappeared, as showed in the FIG. 1(b) (the blue curve). So it was confirmed that some D$_2$ molecules had been transferred to the vibrationally excited state ($v$=2, $j$=0). In order to measure the absolute efficiency of vibrational excitation, the probe laser was tuned to REMPI line of D$_2$ ($v$=0, $j$=0). If both the Raman pump and Stokes lasers were applied, we observed a considerable depletion of the REMPI peak, as showed in FIG. 1(b) (the red curve) indicating that a significant portion of D$_2$ ($v$=0, $j$=0) molecules in the REMPI probe volume had been excited to the upper state. By integrating the area of the depletion and comparing it with the area of REMPI peak with Stokes off, an absolute excitation efficiency of 25.2% was determined.

In order to investigated the dependence of the excitation efficiency on the pulse energy of Raman excitation lasers, we have measured the integral area of D$_2$ ($v$=2, $j$=0) REMPI peak at different pulse energies. As shown in FIG. 2(a), the pulse energy of the pump laser was fixed at 60 mJ, while the pulse energy of Stokes was changed from 72 mJ to 307 mJ. An increase of relative excitation efficiency was observed, but the increasing rate of REMPI signal became slower as pulse energy of Stokes laser increase. In FIG. 2(b), the pulse energy of Stokes laser was fixed at 307 mJ, while the pulse energy of the pump laser was changed from 10 mJ to 61 mJ. A linear increase of the relative excitation efficiency was observed, indicating the excitation was not yet saturated at the pump energy of 60 mJ and further enhancement of excitation was accessible by increasing the pulse energy of the pump laser.

FIG. 2 The dependence of the integral area of REMPI line of D$_2$($v$=2, $j$=0) on the pulse energy of Raman excitation lasers. (a) The energy of the pump is kept at 60 mJ while increasing the energy of Stokes laser. Solid squares are experimental data points, and solid lines are B-spline interpolations. (b) The energy of Stokes is kept at 307 mJ while increasing the energy of the pump laser. Solid squares are experimental data points, and solid lines are the linear fitting.

During the Raman excitation, because both the pump and Stokes lasers were very intense, Stark effect might not be ignored. So we investigated the relationship between Raman line shape and laser energy. The probe laser was fixed at the peak maximum of D$_2$ ($v$=2, $j$=0) REMPI line and we scanned wavelength of the OPO through Raman transition. In FIG. 3, we increased the energy of Stokes laser while kept the pulse energy of pump laser constant at 60 mJ. We found that Raman line extends to lower energy and becomes much broader. The linewidth (FWHM) increases from 0.05 cm$^{-1}$ at 72 mJ to 0.13 cm$^{-1}$ at 307 mJ. The broadening induced by the pump laser was also measured, but no significant change of the linewidth was observed because the pump laser was much weaker than Stokes laser. This broadening indicates that Stokes pulse have induced a significant Stark shift, which is comparable to Stark shift in the excitation of D$_2$($v$=1) with the scheme of SARP [16}. In an attempt to apply SARP scheme to D$_2$ $v$=0$-$2, we measured the dependence of relative excitation efficiency on both time delay and Raman shift. The relative excitation efficiency was measured as the integral area of D$_2$ ($v$=2, $j$=0) REMPI spectral peak. We scan the time delay between the pump and Stokes pulses with an interval of 1 ns from $-$10 ns to 10 ns, and the wavelength of OPO to go through the Raman line with an interval of 0.01 cm$^{-1}$ from 5867.95 cm$^{-1}$ to 5868.11 cm$^{-1}$. A three dimension surface was obtained by interpolating all these data points, as shown in FIG. 4. Only a single peak is observed at delay of 0 ns and Raman shift of 5868.05 cm$^{-1}$. It is quite different from the three dimension surface reported previously in the excitation of HD and D$_2$ with the scheme of SARP, which feature two peaks and a saddle point. As discussed in previous studies, the two peaks locating at the rising and falling edges of Stark-shifting laser, and the saddle point corresponding to the optimal temporal overlap of the pump and Stokes pulses, are signatures of SARP [15, 16]. The single peak in FIG. 4 indicates that SRP instead of SARP occurred in present study. Compared with fundamental transition, Raman polarizability of overtone transition is much smaller [22]. We believe this is the major reason why SARP cannot be applied. In future experiment it is necessary to increase pulse energy of Raman lasers, especially the pulse energy of the pump laser for achieving higher excitation efficiency.

FIG. 3 The line shape of Raman transition at different Stokes laser energy. Solid lines are B-spline interpolations.
FIG. 4 The relative excitation efficiency was measured as the integral area of D$_2$($v$=2, $j$=0) REMPI spectral peak at different Raman shift and time delay between pump and Stokes pulses. Data points are interpolate to generate a three dimension surface on which only a single peak is found.

In the present study, we report the preparation of D$_2$ ($v$=2, $j$=0) with stimulated Raman pumping. An excitation efficiency of 25.2% has been achieved. The dependence of excitation efficiency on pulse energies of Raman excitation laser has been measured. We found that Stokes laser is already saturated at 200 mJ, while the excitation efficiency still increases approximately linearly with pump pulse energies up to 60 mJ. Significant broadening has been observed from the measurement of Raman line shapes as a function of laser energy, which is mainly attributed to dynamical Stark shift induced by Stokes pulse. The dependence of excitation efficiency on time delay and detuning has been measured, exhibiting a three dimension surface with a single peak, which indicates that a SRP instead of SARP process happened in present study. Increasing the pulse energies of the Raman lasers, especially the pump laser, is necessary for transferring more D$_2$ molecules to $v$=2. This technique provides new opportunities for investigating the effect of high vibrational level in gas phase reactions and beam-surface interaction.


This work was supported by the National Natural Science Foundation of China (No.21573226 and No.21503206), Chinese Academy of Sciences (No.XDB17010100).

[1] T. Wang, J. Chen, T. Yang, C. Xiao, Z. Sun, L. Huang, D. Dai, X. Yang, and D. Zhang, Science 342 , 1499 (2013). DOI:10.1126/science.1246546
[2] F. Wang, J. Lin, and K. Liu, Science 331 , 900 (2011). DOI:10.1126/science.1199771
[3] W. Dong, C. Xiao, T. Wang, D. Dai, X. Yang, and D. Zhang, Science 327 , 1501 (2010). DOI:10.1126/science.1185694
[4] T. Yang, L. Huang, T. Wang, C. Xiao, Y. Xie, Z. Sun, D. Dai, M. Chen, X. Yang, and D. Zhang, J. Phys. Chem. A 119 , 12284 (2015). DOI:10.1021/acs.jpca.5b06395
[5] S. Y. De Meerakker, and N. Vanhaecke, G. Meijer. Annu. Rev. Phys. Chem. 57 , 159 (2006). DOI:10.1146/annurev.physchem.55.091602.094337
[6] K. Dulitz, M. Motsch, N. Vanhaecke, and T. P. Softley, J. Chem. Phys. 140 , 104201 (2014). DOI:10.1063/1.4866906
[7] I. Rahinov, R. Cooper, C. Yuan, X. Yang, D. J. Auerbach, and A. M. Wodtke, J. Chem. Phys. 129 , 214708 (2008). DOI:10.1063/1.3028542
[8] P. Maroni, D. Papageorgopoulos, M. Sacchi, T. T. Dang, R. D. Beck, and T. R. Rizzo, Phys. Rev. Lett. 94 , 246104 (2005). DOI:10.1103/PhysRevLett.94.246104
[9] X. Yang, and A. M. Wodtke, J. Chem. Phys. 92 , 116 (1990). DOI:10.1063/1.458480
[10] M. J. Bronikowski, W. R. Simpson, B. Girard, and R. N. Zare, J. Chem. Phys. 95 , 8647 (1991). DOI:10.1063/1.461243
[11] R. B. Metz, J. D. Thoemke, J. M. Pfeiffer, and F. F. Crim, J. Chem. Phys. 99 , 1744 (1993). DOI:10.1063/1.465291
[12] W. Zhang, H. Kawamata, and K. Liu, Science 325 , 303 (2009). DOI:10.1126/science.1175018
[13] B. L. Yoder, R. Bisson, and R. D. Beck, Science 329 , 553 (2010). DOI:10.1126/science.1191751
[14] P. Morten Hundt, B. Jiang, M. E. van Reijzen, and H. Guo, R. D. Beck. Science 344 , 504 (2014).
[15] T. Wang, T. Yang, C. Xiao, and D. Dai, X. Yang. J. Phys. Chem. Lett. 4 , 368 (2013).
[16] T. Wang, T. Yang, C. Xiao, and D. Dai, X. Yang. Chin. J. Chem. Phys. 26 , 8 (2013). DOI:10.1063/1674-0068/26/01/8-12
[17] T. Yang, J. Chen, L. Huang, T. Wang, C. Xiao, Z. Sun, D. Dai, X. Yang, and D. Zhang, Science 347 , 6217 (2015).
[18] A. Geers, J. Kappert, F. Temps, and J. W. Wiebrecht, Chem. Phys. Lett. 155 , 614 (1989). DOI:10.1016/0009-2614(89)87483-4
[19] H. M. Lambert, T. Carrington, S. V. Filseth, and C. M. Sadowski, Chem. Phys. Lett. 169 , 185 (1990). DOI:10.1016/0009-2614(90)85185-F
[20] L. Velarde, D. P. Engelhart, D. Matsiev, J. LaRue, D. J. Auerbach, and A. M. Wodtke, Rev. Sci. Instrum. 81 , 063106 (2010).
[21] M. J. J. Vrakking, A. S. Bracker, T. Suzuki, and Y. T. Lee, Rev. Sci. Instrum. 64 , 645 (1993). DOI:10.1063/1.1144191
[22] W. Perreault, N. Mukherjee, and R. Zare, J. Chem. Phys. 145 , 154203 (2016). DOI:10.1063/1.4964938
杨天罡a,c, 黄龙a,b, 汪涛a, 肖春雷a, 戴东旭a, 杨学明a     
a. 中国科学院大连化学物理研究所, 分子反应动力学国家重点实验室, 大连 116023;
b. 中国科学院大学, 北京 100046;
c. 大连理工大学物理与光电工程学院, 大连 116024
摘要: 本文报道了在分子束条件下制备振动激发态v=2的氘分子.使用受激拉曼泵浦的方案,一套单纵模激光系统用于将氘分子从(v=0,j=0)激发至(v=2,j=0).用共振增强多光子电离技术对激发效率进行了测量,获得了25.2%的激发效率.测量了激发能量和相对激发效率之间的关系.随着斯托克斯激光能量增加,激发效率的增长趋于平缓;然而直至泵浦激光增加到60 mJ时,激发效率仍然表现出线性增加的趋势.测量了在不同激光能量下的拉曼跃迁的光谱线形,观测到显著的动态斯塔克效应.在相对激发效率的三维曲面上观测到了单个峰,说明目前实验中发生的是受激拉曼抽运,而不是受激绝热拉曼通道过程.
关键词: 受激拉曼泵浦    氘分子    振动激发    分子束