Using JPF1 format

JOURNAL OF APPLIED PHYSICS

VOLUME 89, NUMBER 8

15 APRIL 2001

Laser-cooled fluorescence mass spectrometry using laser-cooled barium ions in a tandem linear ion trap Takashi Babaa) and Izumi Waki Central Research Laboratory, Hitachi, Ltd., Hatoyama, Saitama, 350-0395, Japan

共Received 27 October 2000; accepted for publication 18 January 2001兲 We report on a mass spectrometry that uses laser-cooled barium ions to optically detect sample ions in a linear radio-frequency-quadrupole ion trap. To obtain a mass spectrum, we observe the intensity modulation of laser-induced fluorescence emitted by the barium ions, as we electrically excite the secular motions of sample ions that are sympathetically cooled by the barium ions. In a fluorescence mass spectrum of sample xenon ions, we identified the major isotopes of xenon. To estimate the absolute amount of the sample ions that contribute to the mass spectrum, we detect the ions using an electron multiplier that is absolutely calibrated by crystallized laser-cooled barium ions. We enhance transfer efficiency to the electron multiplier using a linear radio-frequency-quadrupole ion guide that is placed in tandem with the ion trap. Typically 30 ions of xenon isotope 131Xe⫹ were detected with a signal-to-noise ratio of ⬃30 in the fluorescence mass spectrum. © 2001 American Institute of Physics. 关DOI: 10.1063/1.1354638兴

the asymmetry of the profiles.6 In a previous report,6 we did not have the means to measure the absolute amount of the molecular ions. Though we placed an electron multiplier by the ion trap 共miniature racetrack RFQ storage ring兲, the transfer efficiency to the electron multiplier was too poor for estimation of the trapped ion number because we extracted the trapped ions from a narrow gap space between the linear quadrupole electrodes. In this article, we perform LCFMS using a tandem linear radio-frequency-quadrupole ion trap. Here we measure the absolute amount of sample ions that contribute to the fluorescence mass spectrum using an absolutely calibrated electron multiplier that is calibrated by crystallized laser-cooled barium ions, as a standard of the absolute amount of trapped ions. We achieve high transfer efficiency from the ion trap to the electron multiplier using an ion guide placed between the ion trap and the electron multiplier.

I. INTRODUCTION

Sympathetic cooling1 was proposed as a tool to cool ions which are not directly laser cooled. In atomic physics, several experimental and theoretical works were reported for frequency standards2 and for quantum computers.3 In molecules,4,5 we reported the detection of sympathetically cooled molecular ions using a mass spectrometric technique, which we call sympathetically laser-cooled fluorescence mass spectrometry 共or LCFMS兲.6 To sympathetically cool molecular ions that originate from the vacuum residual gas, we stored them in a radio-frequency-quadrupole 共RFQ兲 linear ion trap2,6–8 simultaneously with laser-cooled magnesium ions 24Mg⫹ . When we excited secular motions9 of the molecular ions, the intensity of laser-induced fluorescence emitted by the magnesium ions was modulated. We identified the ⫹ ⫹ molecular ions H3 O⫹ , NH⫹ and 4 , COH , and CO2 H atomic ions of magnesium isotopes that were sympathetically cooled. We crudely estimated that about 80 25Mg⫹ ions contributed to the fluorescence mass spectrum. Because most of the ions remained in the ion trap after several consecutive analyses, we expect that LCFMS can be used as an in situ tracing multistage mass spectroscopy 共MS/MS兲 for a small amount of sample. In this article we present LCFMS using laser-cooled barium ions,10–12 which are about six times heavier than the laser-cooled magnesium ions, where the amounts of sample and barium ions are absolutely calibrated. To be usable for mass spectrometry, it is necessary to understand the mechanism of LCFMS, which is determined by energy transfer from the molecular ions to the laser-cooled ions through Coulomb collisions when we excite the secular motion of the sample ions. Since mass and the number of ions are key parameters in such a multiparticle Coulomb system, it is necessary to understand how they affect the line profiles, such as

II. TANDEM LINEAR ION TRAP A. Configuration

The ion trap that we constructed is a four-sectioned tandem linear ion trap 共Fig. 1兲, which has an end section 共a兲, an ion source section 共b兲, a linear ion trap section 共c兲, and an ion guide section 共d兲 with an absolutely calibrated electron multiplier tube. All sections have the same quadrupole cross section. The distance from the quadrupole center to the electrodes r 0 is 3 mm. We made the electrodes of copper and plated the surface with gold to prevent oxidization. The ion trap section 共c兲 is 50 mm in length. Laser-cooling beams pass along the center axis of the ion trap section. An ac dipole electric field for mass analysis is applied to the section. Other sections are curved so that the electron multiplier would not obstruct the laser beam path. Each section is separated with a 1 mm gap in order to electrostatistically insulate each other. The mechanical precision is 10 ␮m.

a兲

Electronic mail: [email protected]

0021-8979/2001/89(8)/4592/7/$18.00

4592

© 2001 American Institute of Physics

Downloaded 18 Jun 2002 to 152.2.239.51. Redistribution subject to AIP license or copyright, see http://ojps.aip.org/japo/japcr.jsp

J. Appl. Phys., Vol. 89, No. 8, 15 April 2001

T. Baba and I. Waki

4593

The equation of motion of an ion with mass m and electric charge e in the potential ␺ (x,y,z,t) is expressed by the well-known Mathieu equation14 using dc parameters a x and a y and rf parameters q x and q y , a x ⫽⫺a y ⫽

q x ⫽q y ⫽

FIG. 1. The tandem ion trap is composed of an end section 共a兲, an ion source section 共b兲, an ion trap section 共c兲, and an ion guide section 共d兲. The barium and xenon ions are created in the ion source section 共b兲 using an electron gun and a barium oven. We perform mass analysis in the ion trap section 共c兲. When we measure the amount of trapped ions, the trapped ions are transferred to the electron multiplier through the ion guide section 共d兲.

Figure 2 shows the electric circuit to drive the foursectioned tandem linear ion trap. The circuit contains an rf circuit for ion trapping, dc bias circuits for control of the ion position, and an ac circuit for mass analysis. We apply nearly the same rf voltage to every section, whose frequencies are identical and whose phases and amplitudes are nearly the same.

␺ 共 x,y,z,t 兲 ⫽ 共 U⫺V cos ⍀t 兲

x 2 ⫺y 2 2r 20

.

共1兲

m⍀ 2 r 20

2eV m⍀ 2 r 20

⬅a,

⬅q.

共2兲

共3兲

When q⭐0.5 we can approximate ␺ (x,y,z,t) by a timeaveraged harmonic potential, i.e., pseudopotential,9 q x 2 ⫹y 2 x 2 ⫺y 2 ⌿ 共 x,y,z 兲 ⫽ V ⫹U . 8 r 20 2r 20

共4兲

¯ is given by The depth of the pseudopotential D ¯⫽ D

qV , 8

共5兲

when U is set to 0. The harmonic motions in the pseudopotential are called secular motions. The frequency of the secular motion, called the secular frequency, is given by

␻ x⫽

⍀ 2

␻ y⫽

⍀ 2

B. Operating parameters

Since the theory of the linear ion trap is reported elsewhere,2,7,8,13 we note some parameters used in this article. In the formulation, we define the z axis along the quadrupole centerline of the linear electrode. When a quadrupole rf voltage with amplitude V and frequency ⍀/2␲ and a quadrupole dc voltage U are applied to the linear quadrupole electrodes, the ideal linear quadrupole potential ␺ is given by

4eU

冑冑

a⫹

q2 , 2

⫺a⫹

q2 . 2

共6兲共7兲

We set a⫽0 in the LCFMS, so that two secular frequencies ␻ x /2␲ and ␻ y /2␲ degenerate to the same value,

␻⫽

eV

冑2m⍀r 20

.

共8兲

In LCFMS we measure the secular frequency of sample ions, which is proportional to the specific charge of the ion 关Eq. 共8兲兴, by detecting laser-induced fluorescence emitted by the laser-cooled ions as we oscillate the sample ions by a supplemental ac electric field. We fix the rf frequency at 1.76 MHz with an amplitude of 162 V so that q⫽0.2, pseudopotential depth is 3.6 V, and secular frequency is 120 kHz for barium ions. C. Operation

FIG. 2. The driving electric circuit of the tandem ion trap is shown.

We create sample ions and barium ions in the ion source section 共b兲 by impact of a 150 eV electron beam. The creation speed of xenon ions is controlled by the partial pressure of the xenon gas in the vacuum chamber, which we introduce through a leakage valve. We control creation speed of barium ions by the temperature of the atomic oven that generates the barium atomic vapor. The total amount of the trapped ions is controlled by the electron beam duration. We prevent ion loss from the end of the ion source by applying a dc voltage to the end electrode 共a兲. The dc bias V b of the ion source section is kept at 0.1 V higher than the ion trap section V c (⯝0) so that the ions are transferred easily to the ion trap section. After the ion creation, V b is kept at 4 V, nearly


4594


FIG. 3. The storage lifetime was measured. After the ions are stored in the ion trap section for a certain storage period, they are transferred to the electron multiplier. The storage lifetime 共1/2 decay time兲 is 1620 s.

the pseudopotential depth, higher than Vc , so that ions are localized in the ion trap section. By separating the ion source from the ion trap section, we prevent the coating of barium metal onto the surface of the ion trap section 共c兲, because contact potential 共2.7 V兲 between gold and barium distorts the pseudopotential of the ion trap section. The ion guide section 共d兲 functions as an end electrode of the ion trap section 共c兲 as well. During ion storage, we keep the dc bias of the section the same as the pseudopotential depth 4 V. When we eject the trapped ions, we decrease the dc bias to ⫺1 V. The ions are transferred to the electron multiplier through the ion guide section. The effect of curvature of the section can be neglected because the kinetic energy of the transferred ions is much less than the pseudopotential depth of the ion guide section. We place the electron multiplier at the end of the ion guide section with a 5 mm gap space. The window of the electron multiplier is 6⫻8 mm, which is enough to cover the exit area 共6 mm⫻6 mm兲 of the ion guide section. We apply ⫺2 kV to the first dynode of the electron multiplier, which accelerates and focuses the ejected ions, giving nearly 100% detection efficiency. We remove impurity ions lighter than xenon ions, such as Ba2⫹ and CO⫹ 2 , by increasing the rf voltage so that q reaches ⬃0.7 for xenon ions. We did not remove ions heavier than barium ions, because such impurity ions are negligible in our vacuum. D. Basic performances of the tandem linear ion trap

We measured storage lifetime, which is a basic performance of the tandem linear ion trap. We define the storage lifetime as the period during which the total amount of trapped ions decreases to half, which needs to be much longer than the time scale of LCFMS measurements, which is currently about 100 s. After storage of the ions for a certain time, we count the stored ions by the electron multiplier. We did not apply laser cooling or sympathetic cooling to the trapped ions in this storage lifetime measurement. Figure 3 shows the decay of the 共relative兲 amount of stored ions after the storage time. In the figure, we show an exponential fit, whose decay constant, i.e., storage lifetime, is 1620 s, which is long enough to

T. Baba and I. Waki

FIG. 4. The transfer efficiency between sections was measured. We repeat round trip transfers between the sections 共c兲 and 共b兲. The remaining ions are counted by the electron multiplier. Transfer efficiency per single transfer is 94%.

perform LCFMS. The storage lifetime of the ion source section 共b兲 was the same as the ion trap section 共c兲. Figure 4 shows a measurement of ejection efficiency from the ion trap section, which is another basic performance of the tandem linear ion trap. The efficiency affects the detection efficiency of trapped ions. We repeat round trip transfers between the ion trap section 共c兲 and the ion source section 共b兲 for N times. The remaining ions detected by the electron multiplier decreases by a factor ␩ 2N , where ␩ is the single transfer efficiency. Since each stay in a section takes 5 s, ion loss within a section is negligible because the stay is much shorter than the storage lifetime. The figure shows an exponential fit with 94% single transfer efficiency. Therefore, we assume that we can transfer the ions trapped in the ion trap section 共c兲 with 94% efficiency through the ion guide section 共d兲 toward the electron multiplier. III. LASER COOLING OF BARIUM IONS

Figure 5 shows the system for laser cooling of barium ions.10–12 Our transition for laser cooling is 6 2 S 1/2 ⫺6 2 P 1/2 :607 425 GHz 共blue-green兲. The pump back transition from 5 2 D 3/2 to 6 2 P 1/2 is 461 312 GHz 共red兲. The bluegreen laser beam is generated by a standing wave dye laser 共modified Coherent 899-21兲 with coumarin 101 dye pumped by a long UV beam 共270 nm multiline兲 of an argon-ion laser 共Coherent Innova 400兲. The red laser beam is generated by a ring dye laser 共Coherent 899-21兲 with kiton red dye pumped by a 514.5 nm beam of an argon ion laser 共Coherent Innova 100兲. We stabilized the frequencies of the lasers by locking to molecular hyperfine spectra measured by Doppler free spectroscopy.15 The blue-green reference is tellurium vapor and the red reference is iodine vapor. We tune the laser frequencies by acousto-optical frequency shifters so that the frequencies match the barium ion transitions. We apply a magnetic field parallel to the laser beam in order to eliminate the optical pumping to the D-state sublevels.11 We observe the blue-green fluorescence either by a photomultiplier tube or by an image-intensified charge coupled device 共CCD兲 camera.



FIG. 5. We use two dye lasers for laser cooling. Frequencies of the lasers are stabilized by referring to the hyperfine spectra of tellurium and iodine vapor. The laser frequencies are fine-tuned by acousto-optical frequency shifters 共AOM兲. Both laser beams are coupled coaxially and focused at the center of the linear ion trap section 共c兲. We observed laser-induced fluorescence by a photomultiplier tube or an image intensified CCD camera.

IV. CALIBRATION OF THE ION DETECTION SYSTEM

Figure 6 shows schematically how we measure the absolute amount of trapped ions destructively by the electron multiplier, which is absolutely calibrated by crystallized laser-cooled barium ions,16–19 whose absolute amount can be counted optically 共Fig. 7 inset兲. When a stable ion crystal is formed, we assume that the absolute amount of the trapped ions is equal to the number of the crystal lattice points in the crystal image. This assumption is justified because ions in a rf trap are known to exhibit a clear phase transition between a crystal phase and a gaseous phase.20 Though the lattice contains 135,137Ba⫹ ions that do not emit laser-induced fluorescence, all lattice points emit nearly equal intensity of fluorescence when we integrate the frames of CCD video pictures for a few seconds, because the trapped ions swap their positions within a fraction of a second due to collisions with neutral particles.17,21 It is difficult to extract laser-cooled crystallized ions from the ion trap section to calibrate the electron multiplier because cooled ions can be bound in a weak stray field in the trap. It would be, in principle, possible to take the lasercooled ion string, heat them to approximate one electron volt energy, and then dump these ions into the electron multiplier. The ions could be heated either by a supplemental ac field or the blue-detuned laser beam. We found it, however, difficult to control the amount of heating, so the heated ions were often lost. Therefore the cooled ionic crystal was first used to calibrate the absolute rate of barium ion creation. The absolute rate of creation was then used to calibrate the absolute

T. Baba and I. Waki

4595

FIG. 6. Method for the measurement of the absolute amount of trapped ions. 共a兲 Linear ion trap with an ion guide and an electron multiplier 共SEMT兲. In 共b兲–共e兲, the solid lines represent dc potential along the ion trap electrodes. 共b兲 Since laser-cooled ions are crystallized when we laser cool them strongly, we are able to estimate the absolute amount of trapped ions when we count the number of lattice points. Here, we calibrate the ionization rate of barium ions. 共c兲 Another set of freshly created barium ions are transferred to the electron multiplier. We calibrate the electron multiplier by using the ions whose absolute amount is estimated using the above measurement in 共b兲共d兲 and 共e兲. To examine the absolute amount of trapped ions, we transfer the ions to the electron multiplier that is absolutely calibrated.

sensitivity of the electron multiplier using noncooled ions. Once the electron multiplier was calibrated, it was used to estimate the amount of trapped sample ions that cannot be laser cooled, as described below. Figure 7 shows the calibration of the absolute rate of barium ion creation. We plot the absolute amount of the trapped ions as a function of the ionization time, which is the duration of the electron beam. We observed crystals when the ionization time was between 2 and 4 s. When the ioniza-

FIG. 7. Image of the crystallized barium ion 共inset兲, and ionization rate of barium ions. The number of the lattice points of the crystallized ions is approximately equal to the amount of trapped ions. The ionization rate derived from a linear fit is derived 9.1⫾0.4 关ions/s兴 by linear fitting of the figure. The error bars in the figure show the standard error of measurements of 10–20 sets of crystals.


4596


FIG. 8. Linearity of amount of the trapped ions to the ionization time of barium ions. The coefficient of the electron multiplier signal per unit ionization time is 112⫾9 关counts/s兴. The error bars in the figure show the standard deviation of ten measurements.

tion time was shorter than 1 s, we did not observe any fluorescence after laser cooling for a few minutes. Our imaging system did not resolve individual ions when the ionization time was longer than 4 s. A linear fit of Fig. 7 gives barium ion creation of 9.1⫾0.4 ions/s. Then, we count the trapped barium ions, which are not laser cooled, by the electron multiplier for absolute calibration of its sensitivity 关Fig. 6 共c兲兴. After storage of barium ions using the same ion-creation condition as the above calibration of absolute ionization rate, they were ejected from the ion trap section 共c兲, transferred through the ion guide section 共d兲 and detected by the electron multiplier. Figure 8 shows a linear fit of the electron multiplier signal as a function of the ionization time. The coefficient of the electron multiplier signal per unit ionization time was 112⫾9 counts/s. We calibrated the absolute sensitivity of the ion-guided ion detection system using Figs. 7 and 8, where the condition of ionization and ion trapping were identical in both measurements. The calibrated absolute efficiency of the total detection system was 12.3⫾0.9 关counts/ion兴, which is reasonable because the kinetic energy of injected ions is 2 keV.22

T. Baba and I. Waki

FIG. 9. Fine tuning of the secular frequencies. Using quadrupole dc voltage Va, we compensated the quadrupole component of the dc stray potential in order to make single dip of the secular frequency of barium ions (V a ⫽0.13). Without compensation, the dip was split in two (V a ⫽0).

condition, the trap contained 91⫾10 barium ions and 136⫾12 xenon ions. The errors are mainly determined by Poisson statistics. After we laser cooled the ions and observed strong fluorescence, we decreased the laser-cooling efficiency by detuning the blue-green laser to ⫺10 MHz in order to make a gas phase of ions. When we applied the analyzing ac voltage with 1 mV amplitude and scanned the frequency from 122.5 to 132.5 kHz, we observed intensity modulations of laser-induced fluorescence 共Fig. 10兲. Since the secular frequency of 138Ba⫹ was 122.3 kHz 共Fig. 9兲, we calculated the masses of the observed dips using the inverse relation between mass and the secular frequency 关Eq. 共8兲兴.

V. LASER-COOLED FLUORESCENCE MASS SPECTROMETRY

When we apply the mass analyzing ac voltage to lasercooled 138Ba⫹ ions, we observed two dips that correspond to the secular frequencies in the spectrum (V a ⫽0 in Fig. 9兲. When we apply a quadrupole dc voltage V a (⬅V1⫺V2, V3 ⫽V2, V4⫽V1兲 to the ion trap section 共c兲, the dip frequencies shift and overlap at V a ⫽0.13 关V兴. Thus the origin of the difference is a weak stray dc potential in the ion trap, which makes a⫽0 in Eq. 共7兲. After the compensation, the secular frequency of 138Ba⫹ is measured as 122.3 kHz. We mass analyzed xenon isotope ions using LCFMS, in which the absolute amount of the sample ions was estimated using the calibrated electron multiplier. We created xenon and barium ions with the electron gun with a duration of 10 s in 2⫻10⫺7 Pa partial pressure of xenon gas. In this

FIG. 10. A laser-cooled fluorescence mass spectrum of xenon isotopes. The dips are identified as xenon isotopes compared with natural abundance of xenon. The estimated amount of 131Xe⫹ ions is 29⫾5.4.



T. Baba and I. Waki

VI. DISCUSSION

In Fig. 10 the dip positions and the depths in the spectrum agreed with the natural abundance of xenon isotopes. Since the estimated total amount of xenon ions is 136⫾12 ions in Fig. 10, the amount of each isotope ion according to the natural abundance of xenon should be: 3⫾1.7128Xe⫹ ions, 36⫾6.0 129Xe⫹ , 6⫾2.5 130Xe⫹ , 29⫾5.4 131Xe⫹ , 37⫾6.1 132Xe⫹ , 14⫾3.7 134Xe⫹ , and 12⫾3.5 136Xe⫹ ions. Significant dips at m/e⫽134 and 136 in Fig. 10 may contain a contribution from barium isotope ions 134Ba⫹ and 136Ba⫹ , because we did not purify the barium isotope. Moreover, since 134Ba⫹ and 136Ba⫹ may emit laser-induced fluorescence by the laser-cooling beam tuned to 138Ba⫹ , the dips at m/e⫽134 and 136 might be enhanced when excited by the ac analyzing field. Other dips should be pure xenon isotope ions. To estimate the absolute detection limit, we focus on the 131Xe⫹ isotope. The noise 共one standard deviation of fluorescence fluctuation兲 around m/e⫽131 is about onethirtieth of the equivalent width of the dip. When we define the absolute detection limit of LCFMS as three standard deviations of the noise, the absolute detection limit is ⬃3 ions. Since the amount of 128Xe⫹ and 130Xe⫹ ions were about the same as this absolute detection limit, it is reasonable that we did not observe evident dips in Fig. 10 at m/e 128 and 130. Thus the position and the depth of the dips are consistent to the masses and the abundance of the xenon isotope ions. Asymmetry of the line profiles is observed in the figure, i.e., the lighter mass side is steeper than the heavy mass side. This asymmetry is independent of the sweep direction of the mass analyzing ac frequency. This asymmetry can be explained by space charge.23 Space charge makes the pseudopotential shallower and the secular frequency of the sample ions shift to a lower frequency. Figure 10 shows a gradual increase of the fluorescence intensity in the direction of frequency scan. A probable reason is the loss of the sample ions when the ions are on resonance with the supplemental ac voltage due to insufficient damping by the sympathetic cooling. When the total amount of trapped ions decrease due to the ion loss, the fluorescence intensity can be increased because rf heating20,24,25 is decreased. Shimizu et al. suggested that ions lighter than the lasercooled ions are not cooled by sympathetic cooling in a threedimensional RFQ ion trap 共Paul trap兲 when the temperature of the trapped ions is higher than ⬃1 K.26,27 In their model of ion–ion collision in the rf field, the one-dimensional sympathetic heating rate is proportional to ⫺

2m 共 m s ⫺ ⑀ m 兲共 m s ⫹m 兲 2

4597

cooled by magnesium ions in the previous report.6 The inconsistency may be solved by the difference between a Paul trap 共Shimizu’s case兲 and our linear ion trap. In the z direction of the linear ion trap, parameter ⑀ is always 0 because of the absence of a radio-frequency field. When ⑀ ⫽0, the sample ions can be cooled even when m⭓m s , because the z-direction heating rate is always negative 共i.e., cooling兲 though the x- and y-direction heating rate is positive. In order to understand the mechanism of LCFMS further, such as the difference between linear traps and Paul traps, the asymmetry in line profiles, and the loss of ions during mass scan, we should perform experiments in which we change masses and the absolute amounts of the sample and the laser-cooled ions. These LCFMS characteristics should be compared with molecular dynamical simulation and modeling of LCFMS. We expect several advantages of laser-cooled barium ions for future applications of LCFMS. When we take q⫽0.4 for barium ions, for example, we can trap sample ions whose masses extend from m/e⬃50 to 1000. Mass range is important for organic molecules. In contrast, it is difficult to simultaneously trap the sample ions whose mass is 1000 together with lighter laser-cooled ions, such as Mg⫹ , because space charge effect23 can limit the mass range. In addition, barium ions can be laser cooled using diode lasers,12 which provide better stability of power and frequency as well as smaller size and easier operation than conventional dye lasers. A LCFMS mass spectrometer may be useful for low temperature molecular chemistry.28,29 VII. CONCLUSIONS

We successfully mass analyzed isotopes of xenon ions by laser-cooled fluorescence mass spectrometry using lasercooled barium ions in a four-sectioned tandem linear ion trap. We measured the absolute amount of the sample ions that contributed to a mass spectrum using an ion detection system composed of an electron multiplier and an ion guide section, which was calibrated using crystallized laser-cooled barium ions as a standard amount of trapped ions. Thirty ions of a xenon isotope were observed with a signal-to-noise ratio of ⬃30. ACKNOWLEDGMENT

We acknowledge valuable discussions with Dr. Yasuharu Hirai. 1

,

共9兲

where, m s is the mass of the sample ion, m is the mass of laser-cooled ions, and ⑀ ⬃1 for q⭐0.5. When m⭓m s , i.e., the sample ions are lighter than the laser-cooled ions, the one-dimensional heating rate is positive and the sample ions are sympathetically heated although the laser-cooled ions are strongly cooled. In contrast, we observed sympathetic cooling of ions lighter than laser-cooled ions in a linear RFQ ion trap. Xenon ions were sympathetically cooled by barium ions in this article, and H3 O⫹ and NH⫹ 4 were sympathetically

D. J. Larson, J. C. Bergquist, J. J. Bollinger, W. M. Itano, and D. J. Wineland, Phys. Rev. Lett. 57, 70 共1986兲. 2 D. J. Wineland et al., IEEE Trans. Ultrason. Ferroelectr. Freq. Control 37, 515 共1990兲. 3 D. Kielpinski et al., Phys. Rev. A 61, 032310 共2000兲. 4 D. J. Wineland, R. E. Drullinger, and F. L. Walls, Phys. Rev. Lett. 40, 1639 共1978兲. 5 K. Mølhave and M. Drewsen, Phys. Rev. A 62, 011401 共2000兲. 6 T. Baba and I. Waki, Jpn. J. Appl. Phys., Part 2 35, L1134 共1996兲. 7 D. A. Church, J. Appl. Phys. 40, 3127 共1969兲. 8 M. Welling, H. A. Schuessler, R. I. Thompson, and H. Walther, Int. J. Mass Spectrom. Ion Processes 172, 95 共1998兲. 9 H. G. Dehmelt, Adv. At. Mol. Phys. 3, 53 共1967兲. 10 W. Neuhauser, M. Hohenstatt, P. E. Toschek, and H. Dehmelt, Appl. Phys. 17, 123 共1978兲.


4598 11


G. Janik, W. Nagourney, and H. Dehmelt, J. Opt. Soc. Am. B 2, 1251 共1985兲. 12 C. Rabb et al., Appl. Phys. B: Lasers Opt. B67, 683 共1998兲. 13 J. D. Prestage, G. R. Janik, G. J. Dick, and L. Maleki, IEEE Trans. Ultrason. Ferroelectr. Freq. Control 37, 535 共1990兲. 14 E. Fischer, Z. Phys. 156, 1 共1959兲. 15 T. W. Ha¨nsch, M. D. Levenson, and A. L. Schawlow, Phys. Rev. Lett. 26, 946 共1971兲. 16 R. Bru¨mel et al., Nature 共London兲 334, 309 共1988兲. 17 I. Waki, S. Kassner, G. Birkl, and H. Walther, Phys. Rev. Lett. 68, 2007 共1992兲. 18 G. Birkl, S. Kassner, and H. Walther, Nature 共London兲 357, 310 共1992兲. 19 M. G. Raizen, J. M. Gilligan, J. C. Bergquist, W. M. Itano, and D. J. Wineland, Phys. Rev. A 45, 6493 共1992兲. 20 R. Bru¨mel, C. Kappler, W. Quint, and H. Walther, Phys. Rev. A 40, 808 共1989兲.

T. Baba and I. Waki P. Bowe et al., Phys. Rev. Lett. 82, 2071 共1999兲. S. A. Fields, L. J. Burch, and W. A. Oran, Rev. Sci. Instrum. 48, 1076 共1977兲. 23 J. F. J. Todd, R. M. Waldren, and R. E. Mather, Int. J. Mass Spectrom. Ion Processes 34, 325 共1980兲. 24 J. D. Prestage, A. Williams, L. Maleki, M. J. Djomehri, and E. Harabetian, Phys. Rev. Lett. 66, 2964 共1991兲. 25 Y. Maeno, M. Tachikawa, Y. Moriwaki, and T. Shimizu, Jpn. J. Appl. Phys., Part 2 34, L174 共1995兲. 26 T. Shimizu, Y. Moriwaki, Y. Maeno, Y. Ohsima, and M. Tachikawa, Genshikaku Kenkyu 41, 49 共1996兲. 27 Y. Moriwaki, M. Tachikawa, Y. Maeno, and T. Shimizu, Jpn. J. Appl. Phys., Part 2 31, L1640 共1992兲. 28 A. N. Nikolov et al., Phys. Rev. Lett. 84, 246 共2000兲. 29 R. Wynar, R. S. Freeland, D. J. Han, C. Ryu, and D. J. Heinzen, Science 287, 1016 共2000兲. 21 22


Using JPF1 format

Using JPF1 format

Suggest Documents