This journal is© the Owner Societies 2014 Phys. Chem. Chem. Phys., 2014, 16, 22373--22381 | 22373

Cite this:Phys.Chem.Chem.Phys.,

2014, 16, 22373

Characterization of RbSr molecules: spectralanalysis on helium droplets

Gunter Krois, Florian Lackner,* Johann V. Pototschnig, Thomas Buchsteiner andWolfgang E. Ernst*

We report an experimental investigation of RbSr molecules attached to helium nanodroplets. The molecules

are prepared on the surface of helium droplets by utilizing a sequential pickup scheme. We provide a detailed

analysis of the excitation spectrum in the wavelength range 11 600–23 000 cm�1. The spectrum has been

recorded by resonance enhanced multi-photon ionization time-of-flight spectroscopy. The inherent mass

sensitivity of the method allows for an unraveling of the RbSr spectrum, which is influenced by Rb and

Sr dimer contributions, because of the proximity of their respective isotopologues. In addition, the

vibrationally resolved 42S+ band was investigated using laser induced fluorescence spectroscopy. The

vibronic transitions exhibit a lambda-shaped peak form, which is characteristic of excitations on helium

droplets and indicative of strong coupling of the molecule to the superfluid helium environment.

Furthermore, the vibrationally resolved 42S+ state enables the determination of molecular parameters,

which are in excellent agreement with previously measured dispersed fluorescence spectra, originating

from bare RbSr molecules. The assignment of recorded transitions is based on calculated transition dipole

moments and potential energy curves. The theoretical results allow for the identification of transitions

from the vibronic X2S+ ground state to the 22P, 32S+, 42S+, 32P, 42P and 62S+ states. The detailed

investigation of RbSr on helium droplets provides a solid basis for further high resolution gas phase

studies of this diatomic molecule that holds promise in the area of cold molecular physics.

1 Introduction

Research on ultracold molecules1,2 can help to gain new insightsinto ultracold chemistry,3 high-resolution spectroscopy4 or mole-cular Bose–Einstein condensation5,6 and technologies such asquantum computation7 may be advanced. The preparation ofmolecules in their absolute ground state is experimentally achallenging procedure, which is typically achieved by photo-association or magnetoassociation (via Feshbach resonances)of ultracold atoms and further cooling with coherent popula-tion transfer methods (e.g. STIRAP).8–11

As both polar and paramagnetic molecules, alkali–alkalineearth (Ak–Ake) diatomics have recently attracted increasingattention. A promising candidate among these molecular speciesis RbSr,12–15 because both Rb and Sr16 can be prepared in anultracold state, even in a quantum degenerate gas mixture.14

While there have been theoretical calculations of RbSr potentialenergy curves,15,17,18 no experimental spectroscopy has beenreported for these molecules. Due to the rather different tem-peratures required for the evaporation of the two atomic species,the gas phase preparation of RbSr has turned out to be difficult.

Therefore, an approach using matrix isolation spectroscopymay be appropriate for a first observation of electronic statesas well as the corresponding transition probabilities.

This is where the method of helium nanodroplet isolation(HENDI) spectroscopy comes into play, a method which hasadvanced to a powerful spectroscopic tool since its develop-ment in 1992.19,20 Helium nanodroplets have been described asthe ‘ultimate matrix for spectroscopy’,21,22 because they providea unique, soft and cold (T = 0.37 K)22 matrix for the isolation andspectroscopic investigation of atoms and tailored molecules.Insights into the onset of superfluidity,23 and investigations ofhigh-spin alkali molecules24–29 and Ak–HeN Rydberg complexeswith electron orbits much larger than the helium droplet30–33 areonly a few examples of the tremendous successes achieved withhelium nanodroplet isolation spectroscopy (HENDI).

Both Ak as well as Ake doped droplets have been previouslyinvestigated34–39 and several experiments and theoretical inves-tigations on homo- and heteronuclear alkali metal dimers onHe nanodroplets have been performed.40–42 Lately, Ak–Akecomplexes have been prepared and investigated on He nano-droplets. In the example of the LiCa molecule, it was shownthat molecular parameters of the free LiCa molecule canbe retrieved from a spectroscopic investigation on a heliumnanodroplet.43

Institute of Experimental Physics, Graz University of Technology, Petersgasse 16,

A-8010 Graz, Austria. E-mail: florian.lackner@tugraz.at, wolfgang.ernst@tugraz.at

Received 16th July 2014,Accepted 8th September 2014

DOI: 10.1039/c4cp03135k

www.rsc.org/pccp

PCCP

PAPER

Publ

ishe

d on

08

Sept

embe

r 20

14. D

ownl

oade

d by

TU

Gra

z on

30/

09/2

014

17:5

8:44

.

View Article OnlineView Journal | View Issue

22374 | Phys. Chem. Chem. Phys., 2014, 16, 22373--22381 This journal is© the Owner Societies 2014

Recently the first formation of RbSr on He droplets has beenshown,44 and emission spectra of bare RbSr, desorbed uponexcitation, elucidated the electronic structure of bare RbSr.Here we present a detailed investigation of the excitationspectrum of RbSr on He droplets over a wide energy range,thereby revealing insight into the interaction between RbSr andthe He droplet. The proximity of Rb2, Sr2 and RbSr isotopologuesdemands attention in order to unravel the RbSr spectrum and toidentify RbSr transitions in resonance enhanced multi-photonionization time-of-flight (REMPI-TOF) and laser induced fluores-cence (LIF) studies. Our assignment of molecular states is basedon multireference configuration interaction (MRCI) calcula-tions44 of potential energy curves (PECs) and transition dipolemoments (TDMs). A detailed theoretical perspective will bepublished elsewhere. The theoretical results can be comparedto the study presented by Zuchowski et al.15 Note that, while thelatter work15 and ref. 44 represent the only research dealing withexcited states of RbSr, a number of theoretical papers about theRbSr ground state have appeared recently.12,17,18

2 Experimental

The experimental setup and technique have been described indetail in ref. 28, 27 and 30. Utilizing the method of supersonicjet expansion, a beam of nano-sized helium droplets is formed.The cluster size follows a lognormal distribution and the maximumof this distribution (Np0,T0

) is determined by the helium pressure( p0 = 60 bar), the nozzle temperature (T0 = 15 K), and the nozzlediameter (d0 = 5 mm). With the experimental parameters used inour experiments, the maximum of the droplet size distribution isat N60,15 = 6000, corresponding to a droplet radius of R60,15 = 40 Å(assuming spherical droplets.22) For comparison with otherstudies, these parameters correspond to a mean droplet size of%N60,15 = 14 000. The cluster beam passes through a 400 mmskimmer into the pickup chamber, where it propagates throughtwo heated pickup cells containing the dopants Rb and Sr.

The optimum pickup temperature for a maximum RbSr signalhas been found to be around TRb D 80 1C and TSr D 410 1C.

Two different experimental techniques were used to obtainthe results below. For the method of REMPI-TOF spectroscopy,a dye laser (Lambda Physik FL 3002) was used to excite thedopant molecules, and a fraction of the pump laser (RadiantDyes RD-EXC 200 XeCl laser, 26 ns pulse duration, 100 Hz) wasapplied to ionize it. From an energy of 14 700 cm�1 upwards, atwo-photon ionization scheme, utilizing only the dye laser, wassufficient. The ion yield was recorded using a time-of-flightmass spectrometer (Jordan D-850 AREF) with angular reflectronas a function of the dye laser wavenumber.

The RbSr transition around 14 000 cm�1 was also investigatedby LIF spectroscopy, using a cw ring dye laser. The photon signalwas monitored as a function of the dye laser wavenumber usinga photomultiplier tube (Hamamatsu R943-01).

3 Results and discussion3.1 Mass spectrum

Both Rb and Sr possess several isotopes with a high abundancein a narrow mass range; consequently some mass windows arecomprised of a superposition of dimer and RbSr ion yields. Thedimer contribution can affect the REMPI spectra, and thereforewe discuss the mass spectrum at the beginning of the resultssection. As an example, we show the sum of the mass spectrarecorded while the excitation laser was scanned across the42S+ ’ X2S+ transition (a full mass spectrum is acquired forevery laser scan-step) in Fig. 1. This transition represents aprime example, because it is superimposed by Rb and Sr dimersignals, and consequently all possible masses (including monomers)can be seen and their influence on the excitation spectra can bediscussed using this mass spectrum.

The left portion in Fig. 1 shows the monomer masses ofRb and Sr, where the contributions of Rb (green bars) and Sr(magenta bars) are denoted in the figure according to their

Fig. 1 An example of a mass spectrum, showing Rb and Sr monomers (left) and dimers (right), as well as the RbSr molecule, recorded during a REMPI-TOFexperiment. The black line denotes the ion yield recorded with a time of flight mass spectrometer.

Paper PCCP

Publ

ishe

d on

08

Sept

embe

r 20

14. D

ownl

oade

d by

TU

Gra

z on

30/

09/2

014

17:5

8:44

. View Article Online

This journal is© the Owner Societies 2014 Phys. Chem. Chem. Phys., 2014, 16, 22373--22381 | 22375

natural abundance: 85Rb-72.18%, 87Rb-27.82%, 86Sr-9.86%,87Sr-7.0% and 88Sr-82.58%.45 Note that atoms and moleculesdesorb from the helium droplet upon excitation and can thusbe detected in their corresponding mass window. The rightportion of Fig. 1 shows the distribution of the dimers and theRbSr isotopologues. Most mass peaks contain contributionsfrom two or more different molecules (isotopologues of Rb2, Sr2

and RbSr); the color coding in Fig. 1 helps to identify thesecontributions. The percentage of Rb2 is denoted by a green bar, thatof Sr2 by a magenta bar, and the percentage of RbSr is denotedby a blue bar.

As can be seen, the formation of RbSr molecules on the Hedroplet is quite efficient. The binding energy of the RbSrground state is on the order of B1000 cm�1.12,15,17,18,44 Thisenergy is carried away by evaporating He atoms. Assuming5 cm�1 binding energy of a He atom to the droplet,20 this resultsin an approximate droplet shrinking of B200 atoms upon theformation of a RbSr molecule on the droplet surface, which isnegligible compared to the initial droplet size of N60,15 = 6000.

As shown in the right portion of Fig. 1, the masses of85Rb88Sr and 87Rb88Sr coincide with the masses of Sr dimers(86Sr87Sr and 87Sr88Sr). Even though the abundance of these Srdimer isotopologues is small, it can lead to an unwantedbackground in the REMPI-TOF spectrum, despite its advantageof providing mass resolved excitation spectra. This can beprevented to some extent by lowering the temperature of thepickup cells, thus decreasing the probability of a two-atompickup. 85Rb88Sr has the highest abundance among the RbSrisotopologues and is only weakly influenced by a 86Sr87Srcontribution. From the example shown in Fig. 1, the ratio of87Rb88Sr to the 87Sr88Sr dimer can be estimated as 9 : 1, while theratio of 85Rb88Sr to 86Sr87Sr is 99 : 1. Consequently, in order toguarantee an unambiguous identification of RbSr transitions,we only show the 85Rb88Sr ion yield in the presented spectra,

if not denoted otherwise. The exclusion of the other isotopologuesfrom the spectra and the corresponding loss in total signal arecompensated for by the better signal-to-dimer backgroundratio, because the appearance of dimers can be neglected,and the pickup can be optimized for a maximum RbSr signal.Note that, in contrast to LiCa on He droplets,43 isotope shiftscould not be observed in the RbSr spectra, due to the heavierconstituents of the molecule.

3.2 Excitation spectrum

A survey spectrum of 85Rb88Sr, recorded using REMPI-TOF spectro-scopy, is shown in Fig. 2. We focus here on a detailed analysis of therecorded transitions of RbSr on He nanodroplets. In addition to theexperimental data, a stick spectrum of theoretical Franck–Condonfactors (FCFs)46 multiplied by the squared transition dipolemoments (TDMs) for the denoted transitions has been calculated,which is shown in the following along with the spectra. The productof FCF*TDM2 is hereafter referred to as ‘transition probability’. Eachband of FCFs is calculated for J = 0 rotational states. The transitionsare starting at the vibrational ground state X2S+(n00 = 0) of RbSr,because the molecules are efficiently cooled to their vibronic groundstate due to the cold He environment (0.37 K). The transitionprobabilities have been scaled to fit the experimental data, but theirrelative heights with respect to each other reflect the FCF*TDM2

distribution (all bands have been scaled by the same factor). Thecorresponding potential energy curves have been presented inref. 44, and the details of the theoretical methods and results willbe presented elsewhere. Experimentally, the 42S+ ’ X2S+ transitionis vibrationally resolved and each band shows a characteristiclambda-shaped peak form.47,48 All other states denoted in thespectrum could not be vibrationally resolved and appear as broa-dened, featureless and partly overlapped structures. The assignmentof the transitions is based on a comparison of experimental andtheoretical spectra, as will be shown in detail in the following.

Fig. 2 The excitation spectrum of 85Rb88Sr in the range of 11 600–23 000 cm�1. The signal was recorded by exciting the RbSr molecule on the Hedroplet using a tunable dye laser and ionizing it with a fraction of the XeCl pump laser (up to 14 700 cm�1) or a second photon of the dye laser (above14 700 cm�1). The ion yield was recorded using a time of flight mass spectrometer. The data have been offset corrected; the gray lines show the originaldata points and the red curve represents the data smoothed by convolution with a Gaussian. The colored bars denote the calculated transitionprobabilities (FCF*TDM2) for each transition; blue indicates transitions to a P state and magenta to a S state.

PCCP Paper

Publ

ishe

d on

08

Sept

embe

r 20

14. D

ownl

oade

d by

TU

Gra

z on

30/

09/2

014

17:5

8:44

. View Article Online

22376 | Phys. Chem. Chem. Phys., 2014, 16, 22373--22381 This journal is© the Owner Societies 2014

3.3 22P ’ X2R+ and 32R+ ’ X2R+ transitions

Fig. 3 shows a detailed view of the lowest energy transitions thatcould be observed using our laser systems. Both correlate with theRb 5s, 2S + Sr 5s5p, 3P atomic asymptote. The two different data setsin the figure denote the ion yield upon excitation with a dye laserand ionization with a fraction of XeCl pump laser radiation (twophoton ionization – R2PI), and the ion yield upon excitation andionization with the pulsed dye laser alone (three photon ionization– R3PI). The structure in Fig. 3 shows a very broad double peakfeature in the R2PI (blue) and in the R3PI (red) signals. On the lowenergy side, the structure is cut off, because the end of the tuningrange of the dye laser optics is reached at 11 600 cm�1, accom-panied by a rapid decrease of the Rh800 pulse energy.

On the high energy side the R2PI signal shows a constantbackground, which is caused by the XeCl laser. Also shown inFig. 3, in the form of colored bars, are the calculated transitionprobabilities (FCF*TDM2); these suggest two overlappingtransitions in this wavelength range, which coincide well withthe experimental data. Based on these calculations we assignthe transition at lower energy to 22P ’ X2S+ and the higherenergy transition to 32S+ ’ X2S+.

Note that the MRCI potential energy curves, which havebeen used for the calculation of the transition probabilities,do not consider relativistic effects. It has been shown that fora proper description of dispersed fluorescence (DF) spectra oftransitions originating from molecular states that correspondto the Rb 5s, 2S + Sr 5s5p, 3P asymptote, spin–orbit coupling hasto be taken into account.44 However, for the excitation spectrum,the broadened and overlapping transitions can be well explainedwith a simple non-relativistic picture. The correspondingrelativistic Hund’s case (c) potentials with a considerable transi-tion dipole moment are of 2S+ and 2P character at short range.The main effect of spin–orbit coupling for these states willresult in a slight shift (two O = 1/2 states and one O = 3/2 state

are relevant).44 However, for the sake of consistency, we assign theobserved double peak structure on the basis of Hund’s case (a)potentials without SO-coupling which is, in light of the precedingdiscussion, valid within our experimental resolution.

A comparison of the R3PI and the R2PI signals shows thatthe latter is in better accordance with the calculations, becausethe transition dipole moment for 32S+ ’ X2S+ is higher thanfor the 22P’ X2S+ transition. Note that both signals have beenrecorded with approximately the same laser pulse energy. Thisbehavior is reasonable, because the R3PI signal involves anintermediate state above the 22P and 32S+ states. The transitionprobability to this intermediate state as well as its lifetime willaffect the signal. In the R2PI signal these additional effects donot occur, because the RbSr molecule is directly ionized fromthe 22P or 32S+ excited state.

3.4 42R+ ’ X2R+ transition

Fig. 4 shows several aspects of the RbSr 42S+ ’ X2S+ transitionin detail. The vibrational resolution of this transition allows amore refined study. The calculated transition probabilities,which facilitate the assignment of the transition, are shownin panels (a) and (b) of Fig. 4.

According to our calculations, the 42S+ PEC is associated withthe Rb 5s, 2S + Sr 5s4d, 3D asymptote. The 42S+ PEC experiencesan avoided crossing with the 52S+ PEC.

Fig. 3 Detailed view of the 32S+ ’ X2S+ and 22P ’ X2S+ transitions.Blue denotes the R2PI signal recorded by two-photon ionization with thedye and the XeCl laser, while the red signal was obtained by R3PI with thedye laser alone. The gray lines show the original data points, the red andblue curves represent the data smoothed by convolution with a Gaussian.The colored bars denote the calculated transition probabilities(FCF*TDM2); blue indicates transitions to the 22P state and magenta tothe 32S+ state. The gap in the spectrum is due to a small wavelength rangenot covered by the two dyes Rh800 and Sty9.

Fig. 4 Closeup of the 42S+ ’ X2S+ transition. (a) The gray line shows theoriginal data points recorded using REMPI-TOF spectroscopy. The red linecorresponds to the smoothed data, green denotes the fitted signal (accordingto eqn (1)), and the black vertical lines mark the onset of the rising edges,corresponding to free molecule transitions. (b) The 42S+ ’ X2S+ transitionrecorded using LIF spectroscopy. (c) Shows the smoothed ion yield for atomsand molecules during the REMPI-TOF measurement. The cyan line corre-sponds to Sr2 recorded with only Sr doped droplets. All signal curves have beensmoothed by convolution with a Gaussian. The magenta bars in panel (a) and(b) denote the calculated transition probabilities (FCF*TDM2).

Paper PCCP

Publ

ishe

d on

08

Sept

embe

r 20

14. D

ownl

oade

d by

TU

Gra

z on

30/

09/2

014

17:5

8:44

. View Article Online

This journal is© the Owner Societies 2014 Phys. Chem. Chem. Phys., 2014, 16, 22373--22381 | 22377

As shown in panel (a) of Fig. 4, vibronic transitions fromn0 = 0–4 ’ n00 = 0 can be identified in the REMPI-TOF signal.Considering the droplet temperature of 0.37 K, the RbSrmolecule is cooled to its vibrational ground state, n00 = 0, uponformation on the droplet and all excitations start from thevibronic ground state. The asymmetric broadening of thevibrational bands is caused by an interaction of the moleculewith the droplet. As explained in ref. 48 and 47 for alkali tripletmolecules on He droplets, a strong coupling of the molecularvibrations to the surface of the He droplet causes the mergingof the zero phonon line with the phonon wing, resulting in alambda-shaped peak form. Consequently, the onset of therising edges of the broadened lines coincides with the bandorigins of the free molecule within B10 cm�1, as has beenshown e.g., for Li2 (ref. 26) and LiCa.43 Similar to RbSr, a strongcoupling of the LiCa molecule to the He droplet surfacehas been reported. This suggests that a strong coupling of theAk–Ake molecules with the He droplet is a common feature. Incomparison with Ak dimers on He droplets, where zero phononlines have been observed for singlet transitions,41 the strongcoupling of Ak–Ake molecules seems reasonable, because Akeatoms interact more strongly with the helium droplet.37 Thedifferent interactions of the atomic constituents with the heliumdroplet raise the question of the orientation of the molecule withrespect to the surface. In the recent work on LiCa on heliumnanodroplets it was argued that the observed, free-molecule like,spin–orbit splitting on the helium droplet can be explained by theconservation of the cylindrical symmetry in the case of a perpen-dicularly oriented molecule, in contrast to triplet state alkali-dimers lying flat on the droplet surface, where the spin–orbitsplitting was strongly changed.49 Unfortunately, we could notobserve a vibrationally resolved spin–orbit split state for RbSr, butthe strong spin–orbit interaction in the lower states correlatingwith the Rb (5p) 2S asymptote (below our laser tuning range) mayoffer a possibility for the investigation of the molecule’s orienta-tion to the droplet surface in future experiments.49 As the Akepartner is more strongly attracted by the He droplet, the diatomicmolecule may point towards the droplet with the Ake end.

In addition to REMPI-TOF spectroscopy, the 42S+ ’ X2S+

transition was investigated by LIF spectroscopy, and the result-ing spectrum is shown in panel (b) of Fig. 4. Again, thevibrational levels n0 = 0–4 can be identified. Compared to theREMPI-TOF signal the n0 = 0 ’ n00 = 0 line is much higher withrespect to the other vibronic transitions and corresponds better tothe relative heights of the calculated transition probabilities shownas vertical bars in the figure. The reason for this is a saturation ofthe transition in the REMPI-TOF signal due to the relatively highpulse energies necessary to obtain a reasonable count-rate. The LIFsignal also shows an underlying background around 14 100 cm�1,which originates from the Sr2 (11Su

+ ’ X1Sg+)50 transition

and probably also a very weak contribution of a Rb3 transition(24E0 ’ 14A2

0 and 14A100 ’ 14A2

0)28,51 in this wavelength range.The background was minimized by lowering the pickup tempera-tures, thus decreasing the probability for the dimer and especiallythe trimer pickup, but was not suppressed completely inthis experiment.

Panel (c) of Fig. 4 shows the REMPI-TOF signal for 173 amu(red), 85Rb (blue) and 88Sr (black) of one measurement and the86Sr87Sr dimer signal (cyan) of a different experiment, where theHe droplets were solely doped with Sr and the pickup wasoptimized for Sr2. As can be seen, a Sr dimer transition lies in therange of the RbSr transition, hence the 86Sr87Sr isotopologuecontributes to the 85Rb88Sr (173 amu) signal and causes a smallbackground in the range of 14 100 cm�1. The Sr dimer signal wassubtracted from the combined signal (red), resulting in thesignal shown in panel (a). This signal was then fitted in orderto extract molecular parameters of the 42S+ state. As a fittingfunction a sum of asymmetric 2s-functions was used:26,43,52

L(�n) = I(1 + e((�n��n0+w1/2)/w2))�1[1 � (1 + e((�n��n0�w1/2)/w3))�1](1)

The onset of each rising edge is determined by finding themaximum of the second derivative in the corresponding range.The band origins are denoted in Fig. 4 as vertical black lines andthe values are given in Table 1. The molecular parameters Te, oe

and oexe for the 42S+ state are obtained by a least-squares fitto the function given in eqn (2).53 Results are given in Table 1,including the results for the REMPI data reported in ref. 44.

T(n0) = T + oe(n0 + 12) � oexe(n0 + 1

2)2 (2)

In eqn (2), T(n 0) labels the energy between the zero pointenergy of the ground state and the vibrational levels of theexcited state, 42S+(n 0) ’ X2S+(n 00 = 0), and T denotes the energyfrom X2S+(n 00 = 0) to the minimum of the 42S+ PEC. In thefollowing, we estimate Te (the distance between the PECminima) for the sake of a better comparison with theoreticalresults, and the value is listed in Table 1. As an approximationfor the zero point energy of the ground state, we take the valueof (n 00 = 1 � n 00 = 0)/2 and add it to T which gives Te. The resultsof dispersed fluorescence experiments are included in Table 1and show that the values obtained for RbSr on the droplet arewell within B10 cm�1 of the free RbSr values. While the resultsobtained with different experimental approaches are in excellentagreement, the absolute energy of the transition is overestimated bythe calculation by B280 cm�1. Considering the complex electronic

Table 1 Vibrational bands and molecular parameters of 85Rb88Sr for the42S+ ’ X2S+ transition. One standard deviation uncertainties of the fit aregiven in parentheses

Band Energy (cm�1)

n0�n 00REMPI-TOF,this work

LIF,this work

Theory,this work DF44

0–0 14 028(1) 14 032(1) 14 300.6 140 14(5)1–0 14 114(3) 14 119(4) 14 376.72–0 14 192(1) 14 201(2) 14 452.73–0 14 272(1) 14 285(3) 14 528.44–0 14 348(21) 14 358(4) 14 603.9

Te 14 006(4) 14 006(4) 14 282.2oe 86(3) 92(4) 76.4oexe 1.2(0.7) 1.9(0.7) 0.12

PCCP Paper

Publ

ishe

d on

08

Sept

embe

r 20

14. D

ownl

oade

d by

TU

Gra

z on

30/

09/2

014

17:5

8:44

. View Article Online

22378 | Phys. Chem. Chem. Phys., 2014, 16, 22373--22381 This journal is© the Owner Societies 2014

structure of the RbSr molecule and the proximity of atomicasymptotes, this still provides a good result. The calculatedharmonic constant is slightly underestimated.

Panel (c) of Fig. 4 shows that the mass signals of the 85Rb (blue)and 88Sr (black) atoms roughly follow the RbSr signal. A similarbehavior has been reported for LiCa, where predissociating statesenhance the fragmentation process. In general, dynamic processesof atoms or molecules on helium droplets induced by laser excita-tion are governed by non-radiative relaxation mechanisms,enhanced by the helium environment. For RbSr on He droplets asubstantial relaxation upon excitation of the 42S+ state into lowerelectronic states has been reported.44 This considerably attenuatesthe direct transition into the ground state 42S+ - X2S+ and themajority of the fluorescence has been found to originate from lowerstates. This has been attributed to the 12D state, which crosses the42S+ state at its potential minimum and provides a relaxationchannel to lower states. The observation of fragments which corre-late with the 42S+ ’ X2S+ transition in Fig. 4(c) suggests that notonly are bound states populated during the relaxation, but alsorepulsive states that lead to a fragmentation of the molecule. Thecomplex electronic structure of RbSr makes it difficult to give acomplete description of the observed fragmentation of molecules,but weakly bound states which have a quartet character in the non-relativistic picture have to be considered in this process. We alsonote that the signature of excited atomic Rb fragments has also beenobserved in the dispersed fluorescence spectra44 as D-line emission.

Additionally, this fragmentation process may be super-imposed by fragmentation caused by a dissociation of the RbSr+

molecule by a third photon of one of the two lasers in theREMPI-TOF experiments.

3.5 32P ’ X2R+ transition

Fig. 5 shows the 32P ’ X2S+ transition for the RbSr isotopolo-gues 85Rb88Sr and 85Rb86Sr. The transition shows a broad feature-less structure with a steep rising edge on the low energy side.Based on the calculations, as shown in the form of transitionprobabilities (FCF*TDM2) in the figure, we assign this structure to

the 32P’ X2S+ transition. The signal is reasonably well reproducedby the theoretical results, in terms of the spectral position as well asthe relative intensity distribution. Of course, the stick spectrumwould have to be modeled using appropriate line width functions toaccount for the helium broadening. The 32P state correlates withthe same atomic asymptote as the 42S+ state (Rb 5s, 2S + Sr 5s5p, 3D)and experiences an avoided crossing with the 42P state. Note thatwe only show isotopologues which are not affected by a contribu-tion of Sr2 (see discussion above). The two signals are reliable inthis spectral range and the difference in the signal intensityreflects approximately the natural abundance.

We note that for energies above 14 700 cm�1, only the dyelaser has been used to excite and ionize the RbSr molecule. Inthe energy range of this transition, ionization of RbSr with twophotons of the dye laser becomes possible. Considering a depthof the RbSr ground state of B1000 cm�1 and a depth of theRbSr+ ground state54 of 4285 cm�1, two photon ionization forbare RbSr is possible above an energy of B15 200 cm�1. Experi-ments have shown that the vertical ionization threshold (IT) ofspecies on and in He droplets is lowered as a consequence ofpolarization effects, ranging from several tens of wavenumbersfor surface-located alkali atoms55 up to B1000 cm�1 for mole-cules inside the droplet.56 According to this estimation weexpect the spectral signature of the IT to appear in the spectralregion of the 32P ’ X2S+ transition, but the weak structureforbids an unambiguous assignment of the small signal increaseat B15 160 cm�1 to the IT.

3.6 Higher excited states

Fig. 6 shows the REMPI-TOF spectrum of RbSr in the wave-length range of 16 000–23 000 cm�1. Theoretical transitionprobabilities are plotted as vertical bars, blue for transitionsto a 2P state and magenta for transitions to a 2S+ state. In thisenergy range the assignment of peaks to molecular transitionsbecomes more difficult due to the large number of molecularpotential energy curves.

Fig. 5 The 32P ’ X2S+ transition in detail. The colored lines show twodifferent isotopologues of RbSr (black-171 amu, red-173 amu); for easiercomparison the signal for the 85Rb86Sr (black) isotopologue has been multi-plied by a factor of 5. The signal was obtained by ionization with the dye laseralone. The signal curves have been smoothed by convolution with a Gaussian.The blue bars denote the calculated transition probabilities (FCF*TDM2).

Fig. 6 REMPI-TOF signal of the highest RbSr transitions recorded. The graylines show the original data points and the red curve the smoothed signal (byconvolution with a Gaussian). Calculated transition probabilities (FCF*TDM2)are shown as vertical bars, blue bars denote a transition into a 2P state andmagenta bars into a 2S+ state. The arrows assign the correlated experimentaland theoretical transitions. Around 22 000 cm�1 the 62P’ X2S+ is expectedbut not unambiguously assigned (see text).

Paper PCCP

Publ

ishe

d on

08

Sept

embe

r 20

14. D

ownl

oade

d by

TU

Gra

z on

30/

09/2

014

17:5

8:44

. View Article Online

This journal is© the Owner Societies 2014 Phys. Chem. Chem. Phys., 2014, 16, 22373--22381 | 22379

The calculations enable an assignment of the two broadenedand featureless structures at 16 600 cm�1 and 19 700 cm�1 to thetransitions 42P’ X2S+ and 62S+ ’ X2S+, respectively. Both statesshow roughly the same structure, where the steep rising edge ismore pronounced in the 62S+ ’ X2S+ transition. Both features arenarrow and vibrational states are not resolved. A possible explana-tion for this is provided by the theoretical calculations. For bothelectronic transitions only the n 0 = 0 ’ n 00 = 0 vibronic transitionhas a significant transition probability, due to the similarity of theexcited and ground state PEC.

Beyond 16 000 cm�1, theoretical calculations become increas-ingly challenging because of the proximity of atomic Rb and Srenergy levels and numerous avoided crossings related to theincreasing density of molecular states. In this region, resultsobtained from different theoretical methods can differ on theorder of 1000 cm�1.12,44 Consequently, the differences betweenexperiment and calculation increase.

The 42P state dissociates to the Rb 4d, 2D + Sr 5s2, 1Sasymptote and is hence affected by the avoided crossing withthe 32P state described above. The calculated values for the 42Ppotential energy curve deviate for large internuclear separationby B1000 cm�1 from atomic values.57 A comparison with theexperimental data suggests that this deviation is transferred tothe potential energy curve for short ranges, because the experi-mental value deviates from the theoretical value by B1250 cm�1.

The 52S+ ’ X2S+ transition, which is expected to lie also inthe range of 16 000–17 000 cm�1, has a transition dipole momentthat is four orders of magnitude lower than the TDM for the42S+ ’ X2S+ transition. Hence we excluded an association of theobserved transitions with the 52S+ state and explain the absenceof this state in the excitation spectrum by this prediction.

The transition at B19 700 cm�1 is assigned to the 62S+ ’ X2S+

transition, and the theoretical result lies B400 cm�1 below thisvalue. The corresponding asymptote of the 62S+ state (Rb 6s, 2S + Sr5p2, 3D) is underestimated by B200 cm�1 in the theoreticalcalculations. Assuming that this deviation is transferred to thePEC at short ranges, the displacement between the theoreticaland experimental values corresponding to the 62S+ ’ X2S+

transition is significantly reduced.Above an energy of 20 500 cm�1, the density of excited states

and the number of avoided crossings increase, resulting in anincreased deviation for the asymptotic values. The experimentaldata show a broad structure extending from 20 500 to 23 000 cm�1.An unambiguous assignment is not possible, but our calculationssuggest that the observed structure consists of several overlappingtransitions. Contributions to the signal most probably arise fromthree transitions into the 62P, 72P and 82S+ states, where thecalculated transition probability for the 62P’ X2S+ is the highest,as shown in Fig. 6.

4 Conclusion

In summary, we have presented a thorough experimental studyof the RbSr molecule on helium nanodroplets. The utilizationof REMPI-TOF spectroscopy enabled the measurement of an

excitation spectrum in a large spectral range from 11 600 to23 000 cm�1. Based on high-level ab initio calculations,44 sixtransitions originating from the X2S+ (n00 = 0) vibronic groundstate to excited states could be identified. The correspondingstates were assigned as 22P, 32S+, 42S+, 32P, 42P and 62S+. Theinteraction of the RbSr molecule with the helium dropletmanifests itself in the form of broadened and overlappingvibrational lines. On the basis of the observed peak shapeswe can conclude that the coupling of the vibrational motion ofthe RbSr molecule with the liquid helium environment isstrong. This appears to be a general trend for Ak–Ake moleculeson helium nanodroplets.43 The vibrationally resolved 42S+ statewas investigated in more detail with LIF and REMPI-TOFspectroscopy. The fact that vibrational levels could be resolvedfor the 42S+ state enabled the determination of molecular para-meters. A comparison with recent experiments,44 where 42S+ -

X2S+ florescence of bare RbSr molecules (formed on heliumdroplets) could be observed, revealed that these extracted para-meters agree within a few wavenumbers with the free moleculeparameters. This indicates that, despite the relatively large dropletinduced broadening, the influence of the droplet on the absoluteenergies of the transitions is very small. In addition, we presentmass spectrometric studies of the RbSr molecules formed onhelium droplets, which is very important for the correct interpreta-tion of the recorded spectra, because of the proximity of the variousisotopologues of Rb and Sr dimers to those of RbSr. Furthermore,we found that the complex relaxation dynamics upon laser excita-tion leads to a substantial fragmentation of molecules.

Understanding the effect of helium droplets on the spectra ofdopant molecules is the key for the establishment of the heliumnanodroplet isolation (HENDI) approach as a method for thecharacterization of novel and tailored molecules, which are difficultto form by conventional methods. The great advantage of thismethod is the exceptional doping possibility,20 which allows one todope the droplets with virtually any gaseous, liquid, solid as well ashighly reactive atomic and molecular building blocks. This hasalready been exploited in the past, for example, for the formationof various, high-spin alkali molecules.27–29,41,42,51 With our workon Ak–Ake molecules on helium droplets (see also ref. 43) weextend the family of surface bound molecules and expand theHENDI approach towards a new class of strongly surface-coupledmolecules, thereby offering novel perspectives for the deliberatemolecule formation with helium nanodroplets.

The results for RbSr demonstrate the potential of the HENDIapproach for the preparation and characterization of moleculeswith relevance for experiments in the ultracold temperatureregime. The recorded excitation spectrum represents the firstexperimental investigation of RbSr and may aid the search foroptimum pathways for the formation of ultracold ground stateRbSr molecules.

The next promising candidate for the characterization onhelium droplets among the Ak–Ake molecules is RbCa, becauseboth Rb and Ca are well under control in ultracold atomicphysics, and, for both atoms, Bose–Einstein Condensation hasbeen achieved.58,59 As a further example with relevance to ultra-cold molecular physics, the formation of molecules containing

PCCP Paper

Publ

ishe

d on

08

Sept

embe

r 20

14. D

ownl

oade

d by

TU

Gra

z on

30/

09/2

014

17:5

8:44

. View Article Online

22380 | Phys. Chem. Chem. Phys., 2014, 16, 22373--22381 This journal is© the Owner Societies 2014

Cr as a constituent may also be explored by the helium dropletisolation approach.60,61

We thank P. Zuchowski and O. Dulieu for stimulatingdiscussion and for exchanging their unpublished calculations.We also thank F. Schreck for fruitful discussion. This researchhas been supported by the Austrian Science Fund (FWF) underGrant FWF-E-P19759 & FWF-E-P22962 and the ERDF Programof the European Union and the Region of Styria.

References

1 L. D. Carr, D. DeMille, R. V. Krems and J. Ye, New J. Phys.,2009, 11, 055049.

2 O. Dulieu and C. Gabbanini, Rep. Prog. Phys., 2009,72, 086401.

3 R. V. Krems, Int. Rev. Phys. Chem., 2005, 24, 99–118.4 T. Zelevinsky, S. Kotochigova and J. Ye, Phys. Rev. Lett., 2008,

100, 043201.5 L. Pitaevski and S. Stringari, Bose-Einstein Condensation,

Oxford Science Publications, 2003.6 G. Quemener and P. S. Julienne, Chem. Rev., 2012, 112,

4949–5011.7 D. DeMille, Phys. Rev. Lett., 2002, 88, 067901.8 N. Vitanov, M. Fleischhauer, B. Shore and K. Bergmann,

in Advances In Atomic, Molecular, and Optical Physics,ed. B. Bederson and H. Walther, Elsevier, 2001, vol. 46,pp. 55–190.

9 S. Ospelkaus, A. Pe’er, K.-K. Ni, J. J. Zirbel, B. Neyenhuis,S. Kotochigova, P. S. Julienne, J. Ye and D. S. Jin, Nat. Phys.,2008, 4, 622–626.

10 J. G. Danzl, E. Haller, M. Gustavsson, M. J. Mark, R. Hart,N. Bouloufa, O. Dulieu, H. Ritsch and H.-C. Nagerl, Science,2008, 321, 1062–1066.

11 K.-K. Ni, S. Ospelkaus, M. H. G. de Miranda, A. Pe’er,B. Neyenhuis, J. J. Zirbel, S. Kotochigova, P. S. Julienne,D. S. Jin and J. Ye, Science, 2008, 322, 231–235.

12 P. S. Zuchowski, J. Aldegunde and J. M. Hutson, Phys. Rev.Lett., 2010, 105, 153201.

13 T. Aoki, Y. Yamanaka, M. Takeuchi, Y. Torii and Y. Sakemi,Phys. Rev. A: At., Mol., Opt. Phys., 2013, 87, 063426.

14 B. Pasquiou, A. Bayerle, S. M. Tzanova, S. Stellmer,J. Szczepkowski, M. Parigger, R. Grimm and F. Schreck,Phys. Rev. A: At., Mol., Opt. Phys., 2013, 88, 023601.

15 P. Zuchowski, R. Guerout and O. Dulieu, Phys. Rev. A: At.,Mol., Opt. Phys., 2014, 90, 012507.

16 S. Stellmer, M. K. Tey, B. Huang, R. Grimm and F. Schreck,Phys. Rev. Lett., 2009, 103, 200401.

17 R. Guerout, M. Aymar and O. Dulieu, Phys. Rev. A: At., Mol.,Opt. Phys., 2010, 82, 042508.

18 G. Gopakumar, M. Abe, M. Hada and M. Kajita, J. Chem.Phys., 2014, 140, 224303.

19 S. Goyal, D. L. Schutt and G. Scoles, Phys. Rev. Lett., 1992, 69,933–936.

20 C. Callegari and W. E. Ernst, in Handbook of High-ResolutionSpectroscopy, ed. M. Quack and F. Merkt, Helium Droplets

as Nanocryostats for Molecular Spectroscopy – from theVacuum Ultraviolet to the Microwave Regime, John Wiley& Sons, Chichester, 2011, pp. 1551–1594.

21 K. K. Lehmann and G. Scoles, Science, 1998, 279, 2065–2066.22 J. P. Toennies and A. F. Vilesov, Angew. Chem., Int. Ed., 2004,

43, 2622–2648.23 S. Grebenev, J. P. Toennies and A. F. Vilesov, Science, 1998,

279, 2083–2086.24 J. Higgins, W. E. Ernst, C. Callegari, J. Reho, K. K. Lehmann,

G. Scoles and M. Gutowski, Phys. Rev. Lett., 1996, 77,4532–4535.

25 J. Higgins, C. Callegari, J. Reho, F. Stienkemeier,W. E. Ernst, K. K. Lehmann, M. Gutowski and G. Scoles,Science, 1996, 273, 629–631.

26 F. Lackner, J. Poms, G. Krois, J. V. Pototschnig andW. E. Ernst, J. Phys. Chem. A, 2013, 117, 11866–11873.

27 J. Nagl, A. W. Hauser, G. Aubock, C. Callegari andW. E. Ernst, J. Phys. Chem. A, 2007, 111, 12386–12397.

28 J. Nagl, G. Aubock, A. W. Hauser, O. Allard, C. Callegari andW. E. Ernst, J. Chem. Phys., 2008, 128, 154320.

29 A. W. Hauser, G. Aubock and W. E. Ernst, in VibronicInteractions and the Jahn-Teller Effect: Theory and Applica-tions, ed. M. Atanasov, C. Daul and P. Tregenna-Piggot,Jahn-Teller Effect and Spin-Orbit Coupling in Heavy AlkaliTrimers, Springer Science + Business Media B.V., 2012,vol. 23, pp. 301–316.

30 F. Lackner, G. Krois, M. Theisen, M. Koch and W. E. Ernst,Phys. Chem. Chem. Phys., 2011, 13, 18781–18788.

31 F. Lackner, G. Krois, M. Koch and W. E. Ernst, J. Phys. Chem.Lett., 2012, 3, 1404–1408.

32 F. Lackner, G. Krois and W. E. Ernst, Mol. Phys., 2013, 111,2118–2125.

33 E. Loginov and M. Drabbels, Phys. Rev. Lett., 2011,106, 083401.

34 F. Stienkemeier, J. Higgins, C. Callegari, S. I. Kanorsky,W. E. Ernst and G. Scoles, Z. Phys. D: At., Mol. Clusters,1996, 38, 253–263.

35 M. Theisen, F. Lackner and W. E. Ernst, J. Chem. Phys., 2011,135, 074306.

36 F. R. Bruhl, R. A. Trasca and W. E. Ernst, J. Chem. Phys.,2001, 115, 10220–10224.

37 F. Stienkemeier, F. Meier and H. Lutz, J. Chem. Phys., 1997,107, 10816.

38 E. Loginov and M. Drabbels, J. Chem. Phys., 2012,136, 154302.

39 Y. Ren and V. V. Kresin, Phys. Rev. A: At., Mol., Opt. Phys.,2007, 76, 043204.

40 O. Allard, J. Nagl, G. Aubock, C. Callegari and W. E. Ernst,J. Phys. B: At., Mol. Opt. Phys., 2006, 39, 1169.

41 J. Higgins, C. Callegari, J. Reho, F. Stienkemeier,W. E. Ernst, M. Gutowski and G. Scoles, J. Phys. Chem. A,1998, 102, 4952–4965.

42 M. Mudrich, O. Bunermann, F. Stienkemeier, O. Dulieu andM. Weidemuller, Eur. Phys. J. D, 2004, 31, 291–299.

43 G. Krois, J. V. Pototschnig, F. Lackner and W. E. Ernst,J. Phys. Chem. A, 2013, 117, 13719–13731.

Paper PCCP

Publ

ishe

d on

08

Sept

embe

r 20

14. D

ownl

oade

d by

TU

Gra

z on

30/

09/2

014

17:5

8:44

. View Article Online

This journal is© the Owner Societies 2014 Phys. Chem. Chem. Phys., 2014, 16, 22373--22381 | 22381

44 F. Lackner, G. Krois, T. Buchsteiner, J. V. Pototschnig andW. E. Ernst, Phys. Rev. Lett., 2014, submitted article.

45 D. R. Lide, CRC Handbook of Chemistry and Physics, CRCPress, Boca Raton, 84th edn, 2003.

46 R. J. Le Roy, Level 8.0, University of Waterloo University ofWaterloo Chemical Physics Research Report CP-663, 2007.

47 V. Hizhnyakov, I. Tehver and G. Benedek, Eur. Phys. J. B,2009, 70, 507–512.

48 V. Hizhnyakov and G. Benedek, Chem. Phys. Lett., 2008, 460,447–450.

49 G. Aubock, J. Nagl, J. Callegari and W. E. Ernst, J. Phys.Chem. A, 2007, 111, 7404–7410.

50 A. Stein, H. Knockel and E. Tiemann, Eur. Phys. J. D, 2011,64, 227–238.

51 J. Nagl, G. Aubock, A. W. Hauser, O. Allard, C. Callegari andW. E. Ernst, Phys. Rev. Lett., 2008, 100, 063001.

52 G. Aubock, J. Nagl, C. Callegari and W. E. Ernst,J. Chem. Phys., 2008, 129, 114501.

53 G. Herzberg, Molecular Spectra and Molecular Structure I. Spectraof Diatomic Moleules, Van Nostrand Reinhold, New York, 1950.

54 M. Aymar, R. Guerout and O. Dulieu, J. Chem. Phys., 2011,135, 064305.

55 M. Theisen, F. Lackner, G. Krois and W. E. Ernst, J. Phys.Chem. Lett., 2011, 2, 2778–2782.

56 E. Loginov, D. Rossi and M. Drabbels, Phys. Rev. Lett., 2005,95, 163401.

57 Y. Ralchenko, A. Kramida, J. Reader and N. A. Team,NIST, 2011.

58 M. H. Anderson, J. R. Ensher, M. R. Matthews, C. E. Wiemanand E. A. Cornell, Science, 1995, 269, 198–201.

59 S. Kraft, F. Vogt, O. Appel, F. Riehle and U. Sterr, Phys. Rev.Lett., 2009, 103, 130401.

60 M. Ratschek, M. Koch and W. E. Ernst, J. Chem. Phys.,2012, 136, 104201.

61 A. Kautsch, M. Koch and W. E. Ernst, J. Phys. Chem. A, 2013,117, 9621–9625.

PCCP Paper

Publ

ishe

d on

08

Sept

embe

r 20

14. D

ownl

oade

d by

TU

Gra

z on

30/

09/2

014

17:5

8:44

. View Article Online