Coherent Control of Ultrafast Bond Making and Subsequent Molecular Dynamics: Demonstration of Final-State Branching Ratio Control Liat Levin,1 Daniel M. Reich,2 Moran Geva,1 Ronnie Kosloff,3 Christiane P. Koch,2 and Zohar Amitay1, a) 1)The Shirlee Jacobs Femtosecond Laser Research Laboratory, Schulich Faculty of Chemistry, Technion-Israel Institute of Technology, Haifa 32000, Israel 2)Dahlem Center of Complex Quantum Systems and Department of Physics, Freie Universit¨atBerlin, Berlin, Germany 3)Fritz Haber Research Centre and The Department of Physical Chemistry, Hebrew University, Jerusalem 91904, Israel (Dated: 27 April 2021) Quantum coherent control of ultrafast bond making and the subsequent molecular dynamics is crucial for the realization of a new photochemistry, where a shaped laser field is actively driving the chemical system in a coherent way from the thermal initial state of the reactants to the final state of the desired prod- ucts. We demonstrate here coherent control over the relative yields of Mg2 molecules that are generated via photoassociation and subsequently photodriven into different groups of final states. The strong-field pro- cess involves non-resonant multiphoton femtosecond photoassociation of a pair of thermally hot magnesium atoms into a bound Mg2 molecule and subsequent molecular dynamics on electronically excited states. The branching-ratio control is achieved with linearly chirped laser pulses, utilizing the different chirp dependence that various groups of final molecular states display for their post-pulse population. Our study establishes the feasibility of high degree coherent control over quantum molecular dynamics that is initiated by femtosecond photoassociation of thermal atoms.

PACS numbers: 42.65.Re, 82.50.Nd, 82.53.Eb, 82.53.Kp

A driving force for the field of quantum coherent con- of photoassociation was also demonstrated at ultralow trol has been the idea of generating a new type of photo- temperatures with a timescale five orders of magnitude chemistry1–4 in which a shaped laser field actively drives longer.10–13 Quantum control of bond formation was also the irradiated quantum system in a coherent way from observed in laser-induced catalytic surface reactions14,15 the initial state of the reactants to the final state of the without, however, full insight into the reaction mecha- desired products. This is conceptually different from the nism. The key element is distillation of an entangled common photochemistry approach, where the reactants pair from a thermal ensemble. In strong contrast, many are individually photo-excited and then react without studies have successfully demonstrated photodissociation further interaction with external fields. The key tool for control of bound molecules, controlling both the total the coherent control of binary reactions are shaped fem- yield of fragments as well as their branching ratio to dif- tosecond laser pulses.5 Typically, coherent control of a ferent channels.16–24 binary reaction proceeds in several steps. First, ultrafast The strong-field process in our previous photoassocia- coherent bond making (i.e., free-to-bound femtosecond tion investigations6–9 involves non-resonant multiphoton photoassociation) is photo-induced between the free col- femtosecond photoassociation of a pair of thermally hot liding reactants, and an excited bound molecule is gener- magnesium atoms into a bound excited Mg2 molecule ated. Then, subsequent photo-control over the molecu- and subsequent molecular dynamics on electronically ex- lar dynamics directs the system into target intermediate cited states. At the first stage, we have demonstrated the states. Last, these intermediate states serve as a plat- generation of purity and rovibrational coherence in the form for further coherent photo-control in order to selec- ensemble of photoassociated Mg2 molecules via the mech- tively break bonds (i.e., femtosecond photodissociation) anism of Franck-Condon filtering.6,7 Then, we have uti- for generating the desired products. lized the generated purity and coherence to demonstrate Despite the significance, there is still no realization coherent control over the resulting yield of molecules arXiv:2104.11786v1 [physics.atom-ph] 23 Apr 2021 of this full scheme of binary reaction control from reac- populating a single group of final states.8,9 The molec- tants to products. A major reason is the lack of success- ular yield was enhanced with linearly-chirped femtosec- ful coherent control of ultrafast bond making and the ond pulses having a positive chirp, reaching a maximum subsequent molecular dynamics. For binary gas phase with a specific optimal chirp value, as well as with posi- photo-reactions where the reaction mechanism is most tively chirped pulses split into sub-pulses with a temporal easily unraveled, such control has been demonstrated so 6–9 structure that fits the photo-induced coherent vibrational far only in a few studies we have conducted. Control dynamics. The different excitation and control mecha- nisms have been qualitatively explained by our ab initio calculations. a)E-mail: [email protected] In view of controlling the full scheme of a binary 2

nesium dimer7–9 together with the ultrafast excitation scheme. The starting point is an ensemble of magnesium atoms at a temperature of 1000 K that thermally pop- ulates scattering states above the van-der-Waals ground 1 + electronic state X Σg of Mg2, indicated by gray shad- ing in Fig.1. The ensemble is irradiated with shaped femtosecond laser pulses having a central wavelength of 840 nm, a transform-limited (TL) duration of 70 fs, linear polarization, and an energy that corresponds to a TL peak intensity of 7.2×1012 W/cm2. The pulse photo-associates pairs of magnesium atoms and gener- ates Mg2 molecules via a free-to-bound non-resonant 1 + two-photon transition from X Σg scattering states to 1 rovibrational levels of the Πg state. The correspond- ing Franck-Condon window is located at short internu- clear distances. Then, the pulse further induces subse- 1 quent molecular dynamics on the Πg and higher elec- tronically excited states, resulting in a post-pulse pop- ulation that spans a manifold of final molecular states. Our present objective is to control the population branch- ing ratio among different final states of the photoassoci- ated Mg2 molecules that cover an extended energy band 1 located below and above the Πg asymptote, indicated also by gray shading in Fig.1. These levels belong to 1 + 1 1 1 + the A Σu , (1) Πu, (2) Πu or (2) Σu electronically ex- 1 cited states, which are all dipole-coupled to the Πg and FIG. 1. Potential energy curves of Mg2 and ultrafast excita- 1 + 1 1 + 1 X Σ states. The Πg, A Σ and (1) Πu states share tion scheme corresponding to an intense phase-shaped near- g u the same asymptote of Mg(1P )+Mg(1S), having with infrared linearly polarized femtosecond pulse. The strong- 1 1 field ultrafast excitation involves bond making (free-to-bound the ground-state asymptote Mg( S)+Mg( S) an atomic −1 photoassociation) and subsequent molecular dynamics on the transition energy of 35051 cm corresponding to a mea- electronically excited states. The experimental observable is sured wavelength of λa=285.2 nm. For the control, we the spectrum of the post-pulse UV radiation spontaneously employ linearly-chirped femtosecond laser pulses, all hav- emitted from the final molecular states. The potential en- ing the same spectrum but different spectral phase. Their ergy curves are plotted for the rotational quantum number 1 2 spectral phase is of the form Φ(ω) = 2 k(ω − ω0) , where J=70, which corresponds roughly to the maximum of the ini- ω is a given frequency, ω0 the central frequency of the tial thermal rotational population. The gray shades indicate pulse spectrum (corresponding here to a wavelength of the initial thermal population of scattering states and the 840 nm) and k the linear chirp parameter. The latter electronically excited post-pulse population which decays by (k) is the control variable. The unchirped, or TL, pulse UV emission. corresponds to k = 0. Experimentally, magnesium vapor with a pressure of reaction from reactants to products, the main limita- about 5 Torr is held in a heated static cell at 1000 tion of our previously demonstrated bond-making con- K with Ar buffer gas, and the sample is irradiated at trol has been the control target, i.e., the overall molecu- 1 kHz repetition rate with the intense shaped femtosec- lar yield.6–9 Here, we extend our approach to coherently ond pulses described above. They are phase shaped us- controlling the relative yields of molecules that are gener- ing a liquid-crystal spatial light modulator.5 The post- ated via photoassociation and subsequently photo-driven pulse population of the final molecular states is probed into multiple groups of final states. The investigated ul- by detecting and spectrally resolving the ultraviolet (UV) trafast strong-field process is, as before, photoassociation radiation emitted in their spontaneous radiative decay 1 + of thermal magnesium atoms followed by excited-state to the ground electronic state X Σg . The spectral dynamics of the generated Mg2 molecules. The demon- range of the detected UV emission is λuv,S=281.5 nm −1 strated branching-ratio control is achieved with linearly- to λuv,L=287.5 nm (34783−35524 cm ). In the setup, chirped pulses, utilizing the different chirp dependence the UV radiation emitted toward the laser-beam entrance that different groups of final molecular states display for to the cell is collected at a small angle from the laser- their post-pulse population. Our results establish the beam axis using an appropriate optical setup, spectrally feasibility of high degree coherent control over quantum resolved by a spectrometer with 0.1-nm resolution, and molecular dynamics that is initiated by femtosecond pho- then intensity-measured using a time-gated camera sys- toassociation of thermal atoms. tem with a post-pulse gate of 20 ns. The spectral range Figure1 shows the potential energy curves for the mag- and resolution of the UV emission are the major differ- 3

FIG. 2. Experimentally detected UV emission intensity as a function of emission wavelength λuv and chirp parameter k of the linearly chirped femtosecond pulses. The intensity values are color coded, and the black dashed line indicates the FIG. 3. Examples of experimental UV emission spectra 1 wavelength λa corresponding to emission at the Πg atomic measured following the ultrafast bond-making excitation by asymptote. linearly chirped femtosecond pulses for different values of the chirp parameter k. Each spectrum corresponds to a cut of the emission map of Fig.2 at the corresponding chirp. For each ences to our earlier studies6–9 which measured only the emission spectrum, panel (a) presents the measured spectrum and panel (b) the spectral shape obtained by normalizing the total UV emission intensity integrated over the range of measured spectrum by its total intensity. See text for details. wavelengths longer than λa. Figure2 shows the experimentally measured post-pulse UV emission spectrum, obtained upon irradiation of the (f) tal intensity [Iem,tot(k)] for each state is proportional magnesium vapor with linearly chirped pulses. The re- to the number of Mg2 molecules populating that state sults are presented as a color-coded map of the mea- (f) [Nmol(k)]. The corresponding proportionality constant sured intensity, Iem(λuv, k), as a function of the UV emis- (f) [αem] depends on the different radiative channels that sion wavelength λuv and the chirp parameter k. The the state can spontaneously decay through, and stands wavelength value of λuv = λa, corresponding to emis- sion at the atomic transition, is indicated by a dashed for the intensity fraction of the state’s emission at the black line in Fig.2. Out of the full map, Fig.3(a) UV spectral range detected here out of the full emission from that state. Hence, a change in the shape of the total presents examples of the measured spectrum Iem(λuv; k) at several chirps. Figure3(b) shows the same data nor- measured spectrum [Sem(λuv; k)] can result only from a corresponding change in the relative individual contribu- malized to the total integrated intensity Iem,tot(k), i.e., tions of the different final states to the total spectrum. Sem(λuv; k) = Iem(λuv; k)/Iem,tot(k) with Iem,tot(k) = R λuv,L R λuv,L Such a change can occur due to a change in the rela- Iem(λuv; k)dλuv (with Sem(λuv; k)dλuv = 1 (f) (f) λuv,S λuv,S (f) P tive population [p (k)=Nmol(k)/ f Nmol(k)] of mul- by definition). tiple final states . In other words, the observed chirp- As seen in Figs.2 and3, the emission spectrum dependent changes of Sem(λuv; k) directly imply corre- Iem(λuv; k) changes with the chirp k, such that both sponding chirp-dependent changes in the relative yields its shape Sem(λuv; k) and total intensity Iem,tot(k) ex- of photoassociated molecules populating the different fi- hibit a strong chirp dependence. We focus here on the nal molecular states, i.e., coherent chirp control over the change in the spectral shape. The total emission results corresponding yield branching ratio. from a summation over the emission from the different The associated quantitative analysis is as follows. final states, where each state f contributes to the total After the excitation with a pulse of chirp k, the (f) measured spectrum an individual spectrum [Iem (λuv; k)] emission spectrum from a final state f is given by (f) (f) (f) (f) (f) with a given state-specific shape [Sem (λuv)]. The to- Iem (λuv; k)=αem·Nmol(k)·Sem (λuv), such that the state- 4

(f) R λuv,L (f) specific shape Sem (λuv) satisfies Sem (λuv)dλuv=1. results in a different shape for the total emission spec- λuv,S The chirp-dependent shape of the total spectrum trum [Sem(λuv; k)]. This is the control mechanism of the emitted from a group of states is then given by present chirp-dependent branching ratio among the dif- P (f) (f) (f) ferent final states. S (λ ; k)= W (k) · Sem (λ ) where W (k)= em uv f uv 8 α(f)·p(f)(k) (f) (f) Furthermore, our previous results also indicate that em . Both α and S (λ ) don’t depend on (vib.region) P (f) (f) em em uv αem·p (k) the chirp kmax , which maximizes the population f 1 the chirp k, while p(f)(k) might have such a dependence. transfer into a given Πg vibrational region, corresponds to the positively-chirped pulse having the longest tempo- Thus, indeed, the changes observed for Sem(λuv; k) when tuning the chirp k necessarily reflect changes in the rela- ral segment with instantaneous intensities Jpulse(t) that (f) (vib.region) tive population p (k) of multiple final states. are all above a certain threshold intensity Jth . The latter is associated with the vibrational region. In- The chirp dependence of the measured emission spec- (vib.region) trum seen in Figs.2 and3 includes several prominent creasing the chirp k from zero up to this kmax characteristics. In terms of the total spectral intensity lengthens this segment up to a maximal duration, and I (k), the positively-chirped pulses enhance it with a further increase in k leads to its shortening (eventu- em,tot ally down to zero duration). For a given pulse energy, respect to the TL pulse, while the negatively-chirped (vib.region) (vib.region) pulses attenuate it. In terms of the spectral shape kmax is inversely proportional to Jth . In Sem(λuv; k), from chirp zero up to a positive chirp k of other words, in order to increase the strong-field de- 1 0.006-0.007 ps2, it exhibits only weak chirp dependence excitation efficiency into a specific Πg vibrational re- and its dominant part is the longer-wavelength part of gion, it is beneficial to temporally stretch the positively- λuv>λa. Then, as the chirp k increases to larger positive chirped pulse as much as possible up to a certain limit values, the spectral shape shifts to shorter wavelengths, (that depends on the vibrational-region’s location). This with an enhancement of its shorter-wavelength part and fits the strong-field non-perturbative nature of the de- an attenuation of its longer-wavelength part. In terms of excitation dynamics. Since an efficient population trans- 1 the complementary characteristic of the emission inten- fer to a higher Πg vibrational region requires a smaller sity Iem(k; λuv) at a given wavelength λuv, for all the de- number of efficient Raman transitions, the corresponding (vib.region) tected wavelengths, it is enhanced with positively-chirped threshold intensity Jth is smaller and the cor- (vib.region) pulses over an extended range of positive chirps, and has responding kmax is larger. As emitting states of a maximum at an emission wavelength-dependent posi- higher excitation energy are linked with a higher vibra- tive chirp kmax(λuv). The value of kmax(λuv) stays un- tional region, their population and emitted radiation will 2 changed (as 0.006 ps ) for λa<λuv<λuv,L, and then con- thus be maximally enhanced with a larger chirp value tinuously increases as λuv continuously decreases from λa as compared to emitting states of lower excitation en- down to λuv,S. ergy. Hence, the present experimental chirp dependence Our interpretation for the mechanism facilitating the of the total spectral shape Sem(λuv; k) and the emission branching-ratio control is built upon the dynamics we wavelength-dependent optimal chirp kmax(λuv) point to- have previously8,9 identified to be induced by the posi- ward the following characteristic for the states that dom- 1 + 1 inantly contribute to the detected emission: The higher tively chirped pulses. Subsequent to the X Σg -to- Πg free-to-bound two-photon transition, the strong-field dy- the energy of the emitting state, the shorter are the UV 1 wavelengths at which its state-specific emission spectral namics taking place on the Πg state involves multiple (f) resonant Raman transitions (via higher electronically ex- shape Sem (λuv) is intensified. Considering this charac- cited states) that lead to a vibrational de-excitation, i.e., teristic in combination with the results of the theoretical a population transfer into vibrational levels that are lower emission analysis described next allows us to identify the than the levels accessed by the photoassociative two- dominant emitting states. photon transition. Then, final perturbative one-photon Using the potential curves presented in Fig.1 and the 1 7–9 absorption projects the vibrationally de-excited Πg pop- corresponding electronic transition dipole moments , ulation onto the final molecular states that emit UV light we have assigned the origin of the experimentally de- 1 + when decaying to the X Σg state. The higher the ener- tected UV emission to different electronic states and vi- getic location of the emitting states, the higher is the en- brational levels. First, we employ energetic considera- 1 ergetic location of the Πg vibrational region from which tions to group the different states according to their emis- they are effectively accessed by the final projection. Fol- sion wavelength. Second, analysis of the corresponding lowing their different locations, the population transfer to vibrationally-averaged transition dipole moments (VT- 1 + distinct vibrational regions is induced by a different num- DMs) with the electronic ground state X Σg allows us ber of Raman transitions. So, with a given pulse energy, to calculate the full spectrum of spontaneous emission each positively chirped pulse, due to its unique tempo- from each state in these groups. In the experimentally ral intensity profile (i.e., a Gaussian of unique width), detected range of λuv,S<λuv<λuv,L, four such distinct generates a different distribution of relative population sets of states can be identified: (I) high-lying vibra- 1 1 + 1 1 among the various Πg vibrational regions and, thus, also tional levels of the A Σu state near the Mg( P )+Mg( S) 1 among the various final emitting states. This, in turn, asymptote, (II) vibrational levels of the two coupled Πu 5 states below the Mg(1P )+Mg(1S) asymptote, (III) vi- that the detected emission at the spectral range of 1 brational levels of the two coupled Πu states above the λuv,S<λuv<λa is dominated by the states of group (III). Mg(1P )+Mg(1S) asymptote but below the avoided cross- This is also consistent with the larger state-specific emis- 1 ing of the Πu’s potentials, and (IV) low-to-moderately sion intensities seen in Fig. 4 for group (III) as compared 1 + high-lying vibrational levels of the (2) Σu state. States to group (IV). Overall, we attribute the detected emis- from groups (I) and (II) contribute exclusively to the sion at λuv>λa primarily to high-lying vibrational lev- 1 + 1 1 emission at wavelengths longer than λa. States from els of the A Σu state located near the Mg( P )+Mg( S) groups (III) and (IV) contribute both to the emission asymptote, whereas the detected emission at λuv<λa pre- at wavelengths longer and shorter than λa. Even though dominantly appears to originate from vibrational lev- 1 it would be energetically possible for even higher-lying els of the two coupled Πu states located above the states to add to the detected emission window by de- Mg(1P )+Mg(1S) asymptote but below the avoided cross- caying into vibrationally high-lying levels in the elec- ing. tronic ground state, the VTDMs for such transitions In summary, coherent control of ultrafast bond making are effectively zero. Moreover, vibrational levels on the and subsequent molecular dynamics in experiment and 1 two coupled Πu states above the avoided crossing are theory has been extended here to demonstrating branch- subject predissociation which is rapid compared to the ing ratio control. The relative yields of molecules, which spontaneous-emission timescale. This implies that their are coherently generated in thermal femtosecond pho- contribution to UV emission can be neglected, and group toassociation and subsequently photo-driven into differ- (III) only contains levels below the avoided crossing. ent target states, are coherently controlled using linearly- Figure4 shows the group-specific UV emission spectra chirped pulses of positive chirp. The control knob is the in the experimentally detected range, calculated for rota- chirp of the pulse. The results are explained by calcula- tional quantum number J=70 which corresponds roughly tions of the UV emission together with a model account- to the maximum of the initial thermal rotational popu- ing for vibrational de-excitation that is photo-induced lation. The results are shown as grayscale-coded maps, and controlled via strong-field chirped Raman transi- with each map corresponding to a different group. For tions. Our demonstrated control is a crucial element for each vibrational level in a group, the map shows the rel- the realization of coherent control over the complete pro- ative emission intensity at each wavelength assuming the cess of a binary photo-reaction. In the case of a reaction initial population of the level is 100%. with several product channels, the branching ratio con- A careful analysis of the calculated VTDMs and emis- trol to various target molecular states will be the first sion spectra for all rotational levels reveals that, for group part of an extended scheme, in which the different tar- (II), only very few vibrational contribute to the experi- get states will serve as intermediate platforms from which mental emission window. For the example of J = 70 subsequent selective photo-control will be applied toward shown in Fig.4, only a single contributing level re- different product channels. mains. Furthermore, the group-specific emission intensi- ties within the experimental emission window are larger for group (I) by two orders of magnitude larger com- ACKNOWLEDGMENTS pared to all other groups. Due to the fact that group (I) only contributes to emission at wavelengths longer Financial support from the Deutsche Forschungsge- than λa, we conclude that the detected emission at this meinschaft (DFG – German Research Foundation) un- range is dominated by group (I), with several dozen vi- der the DFG Priority Programme 1840, ”Quantum Dy- brational levels showing large VTDM values and high namics in Tailored Intense Fields (QUTIF)” is gratefully emission intensities. In particular, the density of vi- acknowledged. brational levels in group (I) is highest right below the 1 1 asymptote of Mg( P )+Mg( S), which leads to an ac- 1D. J. Tannor and S. A. Rice, “Control of selectivity of chemical cumulation of the emission intensity directly at wave- reaction via control of wave packet evolution,” The Journal of lengths slightly longer than λ . As seen in Fig.4, the Chemical Physics 83, 5013–5018 (1985). a 2 identification of group (I) as dominant in the emission R. Kosloff, S. A. Rice, P. Gaspard, S. Tersigni, and D. Tannor, “Wavepacket dancing: Achieving chemical selectivity by shaping spectral range of λa<λuv<λuv,L indeed fits the charac- light pulses,” Chemical Physics 139, 201–220 (1989). teristic identified above for the dominant emitting states 3S. A. Rice, M. Zhao, et al., Optical control of molecular dynamics with a correlation between higher state energy and in- (John Wiley, 2000). tensified emission at shorter wavelengths. Moreover, this 4M. Shapiro and P. Brumer, Principles of the quantum control of identification of group (I) agrees with our results from molecular processes (2003). 5A. M. Weiner, “Femtosecond pulse shaping using spatial light previous chirp-dependent experiments and quantum dy- modulators,” Review of scientific instruments 71, 1929–1960 namical calculations for the long-wavelength emission.8 (2000). 6 With regard to the emission at wavelengths shorter L. Rybak, S. Amaran, L. Levin, M. Tomza, R. Moszynski, R. Kosloff, C. P. Koch, and Z. Amitay, “Generating molecular than λa, Fig.4 shows that in this spectral range only rovibrational coherence by two-photon femtosecond photoasso- group (III) follows the aforementioned characteristic of ciation of thermally hot atoms,” Physical Review Letters 107, the dominant emitting states. Hence, we conclude 273001 (2011). 6

FIG. 4. State-specific UV emission spectra in the experimentally detected range calculated for the vibrational levels with rotational quantum number J=70 that belong to the four different groups of electronically excited states which may contribute to the detected emission. Each (grayscale-coded) map corresponds to a different group, and for each vibrational level in a group, the map shows the relative emission intensity at each wavelength assuming the initial population of the level is 100%.

7S. Amaran, R. Kosloff, M. Tomza, W. Skomorowski, tosecond laser pulses,” Phys. Chem. Chem. Phys. 14, 1185–1199 F. Paw lowski, R. Moszynski, L. Rybak, L. Levin, Z. Amitay, (2012). J. M. Berglund, et al., “Femtosecond two-photon photoassocia- 16R. J. Gordon and S. A. Rice, “Active control of the dynamics of tion of hot magnesium atoms: A quantum dynamical study using atoms and molecules,” Annual Review of physical chemistry 48, thermal random phase wavefunctions,” The Journal of Chemical 601–641 (1997). Physics 139, 164124 (2013). 17T. Brixner and G. Gerber, “Quantum control of gas-phase 8L. Levin, W. Skomorowski, L. Rybak, R. Kosloff, C. P. Koch, and liquid-phase femtochemistry,” ChemPhysChem 4, 418–438 and Z. Amitay, “Coherent control of bond making,” Physical (2003). Review Letters 114, 233003 (2015). 18M. Dantus and V. V. Lozovoy, “Experimental coherent laser con- 9L. Levin, W. Skomorowski, R. Kosloff, C. P. Koch, and Z. Ami- trol of physicochemical processes,” Chemical Reviews 104, 1813– tay, “Coherent control of bond making: the performance of ratio- 1860 (2004). nally phase-shaped femtosecond laser pulses,” Journal of Physics 19M. Wollenhaupt, V. Engel, and T. Baumert, “Femtosecond laser B: Atomic, Molecular and Optical Physics 48, 184004 (2015). photoelectron spectroscopy on atoms and small molecules: Pro- 10J. Carini, S. Kallush, R. Kosloff, and P. Gould, “Enhancement of totype studies in quantum control,” Annu. Rev. Phys. Chem. 56, ultracold molecule formation using shaped nanosecond frequency 25–56 (2005). chirps,” Physical Review Letters 115, 173003 (2015). 20O. K¨uhnand L. W¨oste, Analysis and control of ultrafast pho- 11A. Ciamei, A. Bayerle, C.-C. Chen, B. Pasquiou, and F. Schreck, toinduced reactions, Vol. 87 (Springer Science & Business Media, “Efficient production of long-lived ultracold Sr2 molecules,” 2007). Physical Review A 96, 013406 (2017). 21R. J. Levis, G. M. Menkir, and H. Rabitz, “Selective bond dis- 12 D. B. Blasing, J. P´erez-R´ıos,Y. Yan, S. Dutta, C.-H. Li, Q. Zhou, sociation and rearrangement with optimally tailored, strong-field and Y. P. Chen, “Observation of quantum interference and coher- laser pulses,” Science 292, 709–713 (2001). ent control in a photochemical reaction,” Physical Review Letters 22H. Rabitz, R. de Vivie-Riedle, M. Motzkus, and K. Kompa, 121, 073202 (2018). “Whither the future of controlling quantum phenomena?” Sci- 13 S. Kallush, J. L. Carini, P. L. Gould, and R. Kosloff, “Di- ence 288, 824–828 (2000). rectional quantum-controlled chemistry: Generating aligned ul- 23A. T´oth, A. Csehi, G. Hal´asz, and A.´ Vib´ok, “Control of tracold molecules via photoassociation,” Physical Review A 96, photodissociation with the dynamic stark effect induced by thz 053613 (2017). pulses,” Physical Review Research 2, 013338 (2020). 14 P. Nuernberger, D. Wolpert, H. Weiss, and G. Gerber, “Fem- 24N. Iwamoto, C. J. Schwartz, B. Jochim, K. Raju P, P. Feizollah, tosecond quantum control of molecular bond formation,” Pro- J. Napierala, T. Severt, S. Tegegn, A. Solomon, S. Zhao, et al., ceedings of the National Academy of Sciences 107, 10366–10370 + “Strong-field control of H3 production from methanol dications: (2010). Selecting between local and extended formation mechanisms,” 15 P. Nuernberger, D. Wolpert, H. Weiss, and G. Gerber, “Initia- The Journal of Chemical Physics 152, 054302 (2020). tion and control of catalytic surface reactions with shaped fem-