Laser cooling and trapping of 224 Ra +

,


I. INTRODUCTION
Radium, the heaviest alkaline earth element, has favorable electronic properties for laser cooling and trapping in both neutral and singly ionized forms [1,2].Ra + has a narrow-linewidth electric quadrupole (E2) transition, which is advantageous for trapped-ion optical clocks [3][4][5].The Ra + ion clock operates with wavelengths that are compatible with integrated photonic technologies, which makes Ra + appealing for a transportable optical clock.Certain isotopes of radium, such as 225 Ra (I = 1/2), have an octupole deformed nucleus which when paired with their nuclear spin enhances sensitivity to time-reversal symmetry violations [6].A challenge with radium is that there are no stable isotopes.The longest lived, 226 Ra (I = 0), has a 1600-yr half-life, while 225 Ra has only a 15-day half-life.For all radium isotopes, radioactivity limits their use to small quantities.
Previous atomic and molecular experiments used a variety of mechanisms to work with various radium isotopes: Spectroscopy of trapped 209−214 Ra + ions was performed at the TRIµP nuclear facility, where radium atoms were produced when a lead beam impinged on a carbon target [7].An optical atomic clock was demonstrated with a 226 Ra + ion, where the Ra + was produced via laser ablation of a radium chloride target in a vacuum system [5].The atomic electric dipole moment of neutral 225 Ra was measured in an optical dipole trap, where 225 Ra was directly loaded into an effusive oven and heated out for laser cooling and trapping [6].Radium isotopes were produced at the European Organization for nuclear Research (CERN) by using 1.4-GeV protons * mingyufan212@gmail.com to irradiate a uranium carbide target which was then heated to release radium atoms to form RaF for spectroscopy [8].When working with all but 226 Ra, these techniques require specialized facilities and or breaking vacuum on timescales incommensurate with typical atomic and molecular experiments.Fortunately, thorium may be used as a generator to continuously produce in vacuo 224 Ra, 226 Ra, and the desirable 225 Ra isotope, relieving the need for nuclear facilities or opening vacuum systems.This method was used for spectroscopy of neutral 225 Ra from an effusive oven [9,10].
Thorium has a vapor pressure that is more than 10 12 times lower than that of radium [11]; therefore when an oven is heated, it will produce a radium beam while a negligible quantity of thorium will leave the oven.An oven based on this mechanism should provide radium for several thorium half-lives.We use an effusive oven based on this mechanism to realize the first photoionization loading of radium ions into an ion trap and the first laser cooling of 224 Ra + ions.Radium-224, half-life 3.6 days [12], is continuously produced by the α-decay of 228 Th, half-life 1.9 yr, in the oven's crucible.The effectiveness of the oven for ion trap experiments is demonstrated by measuring the trapping rates for several oven temperatures.When the oven is depleted of radium that has built up over several days, the continual decay of thorium generates a sufficient radium flux for ion trapping.

II. OVEN DESIGN
Effusive atomic ovens are a common means to generate atomic beams for laser cooling and trapping of both neutral and ionized atoms [13,14].The oven reported here is based on an effusion cell design commonly used for molecular beam epitaxy (MBE) [15].The effusion FIG. 1.A schematic of the effusive oven.A titanium crucible was loaded with 228 Th (purple circles).The crucible is resistively heated to emit a thermal beam of radium atoms (gold circles).The crucible's temperature is measured with a thermocouple in contact with its outer surface.
cell has heater wires that can heat a titanium crucible up to 1470 K.The crucible's interior is a 59-mm-long cylinder with a 7-mm diameter.The crucible cap has a 12.7-mm-long, 2-mm-diameter aperture; see Fig. 1.We transferred 40(20) µCi of 228 Th(NO 3 ) 4 in 0.1 M HNO 3 into the crucible and dried the solution in the oven by heating the crucible with a hot plate to ∼350 K. Once dried, we added ∼1 mg of strontium, attached the cap, put the crucible in the effusion cell, and installed it in a vacuum chamber.Initially, a strontium beam from the heated oven was used for laser alignment and ion trap testing.This strontium may play a role in reducing radium compounds which might form due to reactions with contaminants.

III. LASER COOLING AND TRAPPING
Photoionization (PI) and subsequent laser cooling and trapping of 224 Ra + was realized with the trap depicted in Fig. 2(a).The trap is a linear Paul trap, described in Ref. [2]; the diagonal radio-frequency (rf) electrodes are separated by 6 mm, and the end cap electrodes are separated by 15 mm.The trap center is 44 mm from the oven aperture, and the rf drive frequency is 990 kHz.Permanent magnets generate a 5.0(5) G static magnetic field at approximately 90 • with respect to the laser cooling beams.
Radium atoms from the oven are photoionized in a two-stage process; see Fig. 2(b).This process is similar to the scheme used for other alkaline earth atoms [16,17].First, neutral radium is excited on the 1 S 0 → 1 P 1 transition with 1.1 mW of 483-nm light.A photoionizing beam then drives the population from the 1 P 1 level to the continuum with 1 mW of 450-nm light.The photoionizing beam waist is at the trap center and is approximately 150 µm.Laser cooling is realized with a 468-nm cooling laser red detuned from the 7s 2 S 1/2 → 7p 2 P 1/2 transition and a 1079-nm repump laser that drives the 6d 2 D 3/2 → 7p 2 P 1/2 transition; see Fig. 2(c).Scattered 468-nm light is collected by the imaging system and focused onto a photomultiplier tube (PMT) and a camera [2].
Three PI laser wavelengths (405-, 422-, and 450-nm) were tested for the second photoionization step.These three wavelengths are all above the 458-nm ionization energy threshold from the 1 P 1 state.The number of 224 Ra atoms photoionized and detected in the trap per minute (the ion capture rate) was measured for each wavelength.The three PI wavelengths produced comparable ion capture rates for similar powers.In practice, we used 450-nm light because it had the most available power (∼1 mW at the trap).
Rydberg autoionization was explored using 461-and 468-nm lasers as the second PI stage to excite Rydberg states.The PI rate with 461-nm light was lower than that with 450-nm light, and no ions were trapped when using 468-nm light.We verified that radium could still be ionized and trapped with the 450-nm PI light before, during, and after the Rydberg tests.
We characterized the reliability and longevity of the oven source by measuring the ion capture rate for extended periods of operation.The ion capture rate was determined by an automated process which monitored PMT counts from laser cooled radium ions.When the PMT counts exceeded a detection threshold, the trap's rf power was switched off to dump the ion and then turned back on to trap the next ion.The time between turning on the rf and loading a new ion was recorded to determine the ion capture rate.The photoionization beams were applied continuously.The ion capture rate at several oven temperatures as a function of time is shown in Fig. 3.The high initial capture rate, a consequence of the flux of radium atoms that have built up prior to turning on the oven, increases rapidly with oven temperature.On the order of 10 11 radium atoms accumulate in one 224 Ra half-life [18].After the initial surge, the continual decay of thorium is sufficient to maintain a flux of radium atoms for trapping.A steady-state capture rate of ∼0.13(1) ions/min is reached after approximately 3 h for each oven operating temperature that was tested.At higher temperatures, the steady state capture rate increases due to a combination of increased rates of radium desorption from surfaces and effusion out of the crucible's titanium walls [19,20].
IV. NEUTRAL RA 224 SPECTROSCOPY The 7s 2 1 S 0 → 7s7p 1 P 1 ( 1 S 0 → 1 P 1 ) radium transition at 483 nm is useful for photoionization loading of radium into ion traps.The 1 S 0 → 1 P 1 transition was first measured for 226 Ra by Rasmussen in 1933 [21].Subsequent spectroscopy of this transition has been performed for 226 Ra and 225 Ra [10,22].We measured the 1 S 0 → 1 P 1 transition of the 224 Ra frequency and compared our value with the previous results using the isotope shift measurements of Wendt et al. [23].Saturated absorption spectroscopy of molecular tellurium ( 130 Te 2 ) was used as a frequency reference for the 224 Ra spectroscopy [10].
For saturated absorption spectroscopy, two parallel 483-nm probe beams and a counterpropagating pump beam which overlaps with one of the probe beams are passed through a cell containing 130 Te 2 at 870( 20   224 Ra was performed in the vacuum apparatus depicted in Fig. 4. A thermal beam of 224 Ra atoms is generated by heating the oven to 880(10) K. Two counterpropagating 483-nm laser beams [each 1.5(1) mW] are perpendicular to the radium beam 25 mm from the oven aperture.The beam waists are 3.4(1) mm.The geometry is chosen to minimize the effect of Doppler broadening.Radium atoms are excited on the 1 S 0 → 1 P 1 transition, and scattered light is collected by an imaging objective onto a PMT.

Spectroscopy of
The 483 nm laser frequency is scanned continuously while the two counterpropagating beams are alternately shuttered and the two spectra are recorded.The reported frequency of the 224 Ra 1 S 0 → 1 P 1 transition is the average frequency of both spectra using a wavemeter (10 MHz resolution) [24] calibrated with 130 Te 2 reference line 0; see Fig.We determine the 226 Ra 1 S 0 → 1 P 1 transition frequency, 621 037 830±60 MHz, by the isotope shifts of FIG. 6.
A comparison of the reported 226 Ra 7s 2 1 S0 → 7s7p 1 P1 transition frequencies where ν226 is our value.Frequencies reported by Rasmussen [21] and Trimble et al. [22] are direct measurements.The frequency reported by Santra et al. [10] is an isotope-shifted value from their measurement of the 225 Ra transition frequency.
224 Ra and 226 Ra with respect to 214 Ra [23].There is a 660-MHz discrepancy between our value for the 1 S 0 → 1 P 1 transition frequency and the value reported in Ref. [25]; see Fig. 6.

V. NULL RESULTS
Short-lived radioisotopes are challenging to work with, particularly due to the difficulty of producing a sufficient atom flux for neutral spectroscopy and ion trapping.Radium poses further difficulties as it is reactive.Different techniques were tested in the neutral spectroscopy setup of Fig. 4 with indeterminate results or uncertain effectiveness, some of which are described here as paths for future exploration.
Argon gas with a pressure of ∼100 Torr was flowed through the neutral spectroscopy apparatus in an effort to slow the radium atoms after nuclear decay and prevent them from becoming deeply buried in the titanium walls of the crucible [26].No increase in PMT counts on the 1 S 0 → 1 P 1 transition was observed.
We also investigated reducing agents.When 228 Th was loaded into the crucible without any reducing agents, no 1 S 0 → 1 P 1 transition peak was observed.After adding in ∼1 mg of strontium, 200 mg Zr powder, and 50 mg BaCO 3 to the crucible to reduce radium compounds [10], there was an increase in the neutral radium spectroscopy signal compared with when only strontium was used.However, it is unclear if the lack of signal without reducing agents was due to a low pressure of reactive background gas molecules rather than the lack of reducing agents.The pressure of the neutral spectroscopy chamber [∼10 −5 Torr with an oven temperature of 880(10) K] was three orders of magnitude higher than that of the ion-trapping apparatus.At the lower pressures achieved in the ion trap, reducing agents may not be necessary.

VI. CONCLUSION
This work lowers the barrier to using the short-lived 224 Ra isotope in cold-atom experiments.An effusive oven based on the decay of thorium is a reliable source of radium atoms for ion-trapping experiments and could be used for cold-neutral-atom experiments depending on atom number requirements and acceptable activity.The source efficiency may be increased with more advanced oven nozzle geometries [27].An oven based on the same principle may be used to laser cool and trap 225 Ra ions, produced via the decay of 229 Th (7825-yr half-life [28]), or 226 Ra ions, produced via the decay of 230 Th (75 400-yr half-life).Such ovens are robust to radium depletion, e.g., due to overheating the oven, as radium is continuously repopulated by thorium.

FIG. 2 .
FIG. 2. (a) A schematic of the apparatus for photoionization and laser cooling and trapping of 224 Ra + .The trap is a linear Paul trap depicted by two rf rods.The effusive oven geometry is given in Fig. 1.(b) The transitions used for photoionization of 224 Ra; 405-, 422-, and 450-nm lasers were tested on the 1 P1 transition to the continuum.(c) The 224 Ra + level structure with the laser cooling transitions highlighted.

FIG. 3 .
FIG.3.224Ra + ion capture rates at different oven temperatures as a function of time.The 450-nm photoionization laser has 1 mW optical power and a 150-µm beam waist at the ion trap.
) K. All three beams are derived from a single laser.The probes, each 60(10) µW, are collected on a balanced photodiode, and their signals are subtracted.The pump beam, shifted 80 MHz by an acousto-optic modulator (AOM) relative to the probe light, has 1.3(1) mW of power and a 0.8(1)-mm beam waist at the center of the 130 Te 2 cell.The AOM serves as a shutter for the pump beam, turning it on and off at a modulation frequency of 10 kHz.The same 10-kHz modulation frequency is mixed with the split photodiode output with a lock-in amplifier to measure the Te 2 spectrum.

FIG. 4 .
FIG.4.A schematic of the vacuum apparatus used for neutral spectroscopy of 224 Ra.The setup uses an effusive oven with the same geometry as in Fig.1.A gas inlet valve allows for the introduction of gas, such as argon, into the chamber.A pair of counterpropagating 483-nm beams are perpendicular to the atomic beam.A fraction of the fluorescence from atoms excited on the 1 S0 → 1 P1 transition is collected by an imaging objective and focused onto a PMT.

5 .
The fit uncertainties for the 224 Ra and 130 Te 2 transitions are their fitted half-width-at-half-maximum values, which account for model imperfections.Wavemeter drift is determined from the difference between the 130 Te 2 saturated absorption spectrum measured before and after the 224 Ra spectroscopy.The observed wavemeter drift within the measurement time of ∼2 h is 2 MHz.The latter 130 Te 2 spectrum and the 224 Ra spectra are shown in Fig. 5. Imperfect alignment of the 483-nm beams results in a 1(9) MHz first-order Doppler shift.The reported 1 S 0 → 1 P 1 transition frequency for 224 Ra is 621 043 830±60 MHz.

FIG. 5 .
FIG. 5. Spectroscopy of the 224 Ra 7s 2 1 S0 → 7s7p 1 P1 transition and saturated absorption spectroscopy of 130 Te2 peaks covering line -1 to line 6 as labeled in Ref. [10], where ν224 is our measured 224 Ra transition frequency.The amplitude of the 130 Te2 saturated absorption signal is normalized to Te2 line 0. The measured 224 Ra 1 S0 → 1 P1 transition frequency is calibrated from the Te2 spectrum.The difference in peak height between beam 1 (blue) and beam 2 (orange) is due to the decay in atomic flux during the measurement.The left inset shows the Lorentzian fit of Doppler-free Te2 line 0. The right inset shows the Voigt fit of the fluorescence from the thermal 224 Ra beam.