Volume 52, number 2
OPTICS COMMUNICATIONS
15 November 1984
OPTICAL PUMPING OF A Ne(3P2) ATOM BEAM WITH A MULTIMODE LASER
K.W. GIBERSON, L.K. JOHNSON, M.W. HART, MS. HAMMOND, T.H. JEYS and F.B. DUNNING Departments of Space Physics and Astronomy Rice University, Houston, TX 77251, USA
and Physics, and the Rice Quantum Institute,
Received 8 August 1984
Stable optical pumping of a neon metastable atom beam is achieved using a multimode dye laser. The factors that give rise to this stability are discussed. The resultant polarized beam is suitable for the study of electron spin and orbital orientation dependences in a variety of metastable atom collision processes.
In recent years there has been increasing interest in the study of polarization dependences in atomic collision processes. Beams of polarized atoms for use in such studies can be obtained via optical pumping [l-8]. Repeated optical pumping cycles, which consist of the absorption of circularly polarized light followed by spontaneous emission, are used to transfer atoms to states for which the projection of the total angular momentum relative to the direction of propagation of the radiation is maximal. A number of studies have been reported that make use of single-mode, frequency-stabilized dye laser systems to optically pump beams of alkali atoms [4-71. Cusma and Anderson [8], however, recently demonstrated that a sodium atom beam can be successfully pumped using an inexpensive multimode dye laser. In the present paper we describe the use of a similar laser to optically pump a beam of Ne(3P2) metastable atoms [ 1,2] for use in an ongoing series of experiments which probe surface electronic structure through the study of polarization dependences in metastable atom-surface interactions [9]. Efficient , stable optical pumping is obtained and the factors that give rise to this stability are discussed. The present apparatus, parts of which have been described in detail elsewhere [3], is shown schematically in fig. 1. A beam of ground state neon atoms is formed by effusion through a multichannel array. A fraction of these atoms are then excited to 3Po,2 * metastable levels by electron impact using a coaxial 0 030-4018/84/$03.00 0 Elsevier Science Publishers B.V. (North-Holland Physics Publishing Division)
electron gun. Any charged particles, or long-lived, highlyexcited atoms formed by electron impact are removed from the beam by a transverse electric field. The atoms then enter a weak magnetic field that is perpendicular to the beam and that defines the quantization axis. The 3P2 atoms are optically pumped using circularly-polarized 640.2 nm 2p5 3s[3/212 (3P2) -+ 2p5 3p[5/213 (3D3) radiation, incident parallel to the magnetic field, from a dye laser. The laser beam is reflected so as to make two passes through the metastable atom beam, illuminating it for a distance of -1 cm. Since the mean reciprocal metastable atom velocity is -2 X 10-5 s cm-l [lo], atoms will, on average, be illuminated for -20 ps. The Doppler width resulting from laser and atom beam divergences is -30 MHz. Following optical pumping the beam polarization is determined by spatial separation of the MJ components using a Stem-Gerlach (SG) analyzer [3]. The optical pumping radiation was provided by a Spectra-Physics Model 375 dye laser, which has a cavity mode spacing of -400 MHz. The laser was tuned using a three element birefringent filter and a single 2 mm intracavity etalon; the dye employed was DCM. The laser was pumped by the 5W all-lines output of an argon ion laser. Initially the dye laser was adjusted to obtain maximum output power. This required use of a high dye circulator pressure and resulted in output powers of -500 mW at 640 nm, sufficient to provide an intensity of -2 W cm-* in the optical pumping 103
Volume
52, number
OPTICS COMMUNICATIONS
2
I5 Novcmbcr 1984
DYE
I
CIRCULAR POLARIZER
SOLENOID
LASER
I
II
DEFLECTION PLATES
MOVABLE CHANNELTRON
&
STERiGERLACH MULTICHANNEL ARRAY
VACbbM \ COAXIAL ELECTRON
GUN
Fig. 1. Schematic
diagrams
region. The majority of the data presented here were taken with the laser operating under these conditions. The stability of the laser, however, was poor, possibly a result of the relatively low viscosity of the dye solvent (40% propylene carbonate, 60% ethylene glycol) that is used with DCM. Large, rapid fluctuations in output power were observed and no stable output mode structure was evident. The laser switched rapidly back and forth between different cavity modes within an envelope of -6 GHz total width. The mode and temporal stability of the laser could be greatly improved, at the expense of a decrease in output power, by reducing the pressure in the dye circulator. Under these conditions only minor fluctuations in output power were observed and the laser appeared to operate stably on three cavity modes, consisting of a strong central mode and two weaker, possibly hole-burning [ 111, modes separated by -*1.6 GHz. The spatial profile at the output of the SG analyzer in the absence of optical pumping is shown in fig. 2a. Electron impact results in the excitation of atoms to both 3P0 and 3P, metastable states and these each contribute to the centralMJ = 0 peak. The 3P, contribution was determined by illuminating the metastable atom beam with 626.7 nm 2p5 3s’[1/2] ,, (3Po) -+ 2~~ 3~‘]3/23 1 radiation from the DCM laser. As discussed elsewhere [12], the 3p’ level can decay to other than the original 3P, state and thus excitation to this level results in selective removal of 3P0 atoms from the 104
MAGNET
WALLS
of the apparatus.
;
i
1
‘-4
I.--,
CHANNELTRQN POSITION
Fig. 2. (a) SG profile with no optical pumping; -mixed 3Po 2 beam, --- following removal of the 3Po atoms. (b) SG profile ’ following optical pumping with RHCP 640.2 nm 3Pz --t 3D3 radiation; -mixed 3Pn 2 beam; --- following subtraction of the 3P,, contribution. ’ beam. The MJ = 0 SG profile obtained following removal of 3Po atoms is included in fig. 2a, the shaded region indicating the 3Po contribution to the original
OPTICS COMMUNICATIONS
Volume 52, number 2
profile. The 3P2 ikfJ = +1(-l) andMJ = +2(-2) features are not resolved because the velocity distribution of the metastable atoms in the beam leads to a range of deflection angles for eachMJ. Initial optical pumping experiments were undertaken using high dye circulator pressures in the laser and an intensity of -2 W cms2 in the optical pumping region. The SG profile obtained for a mixed 3P0 2 beam following optical pumping with RHCP radiation is shown in fig. 2b, together with the profile expected for a pure 3P2 beam, obtained by subtracting the 3P0 contribution. (Since only one dye laser system was available, it was not possible to optically pump the 3P2 atoms and to simultaneously remove the 3P0 atoms from the beam.) Optical pumping leads to efficient transfer of atoms to states of positive MJ. Use of LHCP radiation leads to similarly efficient transfer to states of negative MJ. Despite the observed instabilities in the laser, no problems were encountered in tuning the laser to the 3P2 + 3D3 transition and efficient optical pumping was routinely obtained for periods of several hours without making any adjustments to the laser. The intensitydependence of the optical pumping efficiency is illustrated in fig. 3 which shows, for RHCP pumping radiation, the total number ofMJ = -1 and -2 atoms remaining in the beam as a function of laser intensity in the optical pumping region. To ensure that
2 P 0-I
’ “1””0.1 ’ ’ “““’ 1.0 ’ ’ INTENSITY
I’
(W~rn-~)
Fig. 3. The total number of MJ = -1 and -2 atoms remaining in the beam as a function of the intensity in the optical pumping region for RHCP radiation. The vertical extent of each data point indicates the changes in this number observed at each intensity over a period of -30 min. The pointer indicates the number of MJ = -1 and -2 atoms in the beam with no optical pumping.
15 November 1984
the spectral characteristics of the optical pumping radiation did not change during these measurements, the laser operating conditions were held constant and the intensity varied by use of neutral density filters. The vertical extent of each data point indicates the range of variation in the total number of remaining -1 and -2 atoms observed at each intensity over a period of 30 min, and provides a measure of the temporal stability of the optical pumping. Although reductions in the intensity of the optical pumping radiation lead to reduced optical pumping efficiencies, relatively stable pumping is observed at all intensities. It is apparent from figs. 2 and 3 that, even at the highest intensities, optical pumping is incomplete and not all atoms are transferred to +2 (or -2) states. This may result, at least in part, from the amplitude fluctuations present in the laser output. Any large decrease in laser output power of a duration longer than, or comparable to, that for which an atom is illuminated will enable groups of atoms to pass through the pumping region without being efficiently pumped, leading to a decrease in the average beam polarization. A number of optical pumping experiments were therefore conducted using a low dye circulator pressure in the laser to obtain a stable output. Efficient optical pumping was again achieved, resulting in beam polarizations comparable to those obtained previously. Laser tuning, however, proved to be extremely critical and considerable difficulty was experienced in tuning to the 3P2 + 3D3 transition. Furthermore the optical pumping stability was poor and continual adjustment of the laser was required to maintain a high beam polarization. Thus, paradoxically, optical pumping is least stable with the dye laser operating under its most stable conditions. The present results can be rationalized by considering how the 3P2 -+ 3D3 photoexcitation rate depends on laser frequency. The photoexcitation rate R resulting from illumination of unpolarized 3P2 atoms by monochromatic radiation of intensity I and frequency v is
MJ
=
hfJ
R = aI/hv,
=
(1)
where u is the photoabsorption cross section. This may be expressed in terms of the Einstein A coefficient (5 X 107 s-1) [ 131 and the photoabsorption cross section uc at line center vc as [14] u = UC(1 + [4n(v - v,)/A] 2)-l )
(2) 105
Volume
52, number
OPTICS COMMUNICATIONS
2
where ‘c =
(c2/2nv&,/~a
(3)
and g, and gb are the statistical weights associated with the 3P, and 3D3 levels, respectively. Numerical substitution yields R=
2.9 X lo91
1 + [0.35(v- VJI”
(s-l ).
(4)
where I is expressed in watts per square centimeter and v - vc in megahertz. Thus, near line center, intensities of only a few milliwatts per square centimeter provide very large photoexcitation rates. The intensity required to obtain large photoexcitation rates, however, increase rapidly with increasing frequency separation v - v~. Consider initially the 3P2 + 3 Dj photoexcitation rate expected with the laser operating stably on three cavity modes separated by -1.6 GHz. The intensity in the optical pumping region due to the central mode is -0.8 W cmm2. Near line center this intensity will result in photoexcitation rates 2 1O6 s-l , and will thus make possible several optical pumping cycles during the 20 MSthat an atom is typically illuminated, resulting in high polarizations. The photoexcitation rate falls rapidly, however, as the frequency of the central mode moves from line center, and is only -1.5 X lo5 s-l at a separation of 500 MHz. (The other cavity modes remain too far separated in frequency to provide any significant photoexcitation.) Thus efficient optical pumping can be obtained with the laser operating stably, but only if the frequency of the central mode is close to line center. This makes laser tuning critical and, if efficient long-term optical pumping is required, imposes severe constraints on the frequency stability of the laser. The situation is very different with the laser operating unstably and switching rapidly back and forth between different cavity modes. In this case frequency instabilities are present in the output of the laser that result not only from mode switching but also from instabilities in the frequencies of the invidivual modes themselves. If changes in frequency occur rapidly, atoms will, on average, be irradiated by a quasi-continuous distribution of frequencies within the 6 GHz overall laser linewidth. Thus laser tuning is no longer critical. Furthermore, if the laser is initially tuned so that its output is centered on the transition frequency 106
15 November
19114
vc, long-term frequency drifts of even one gigahertz will not produce large changes in the optical pumping intensity in the vicinity of vc. Thus stable optical pumping will be obtained at all optical pumping intenslties. However, only at the highest optical pumping intensities used in this work is the average 3P3i + “D, photoexcitation rate sufficient to provide a high polanzation. Both neon isotopes will, however, be efficiently pumped because the isotope shift, -1.2 GHz, is well within the laser linewidth. In conclusion. efficient. stable optical pumping of a Ne(3P2) atom beam can be obtained using a multimode dye laser. Because collisional relaxation does not occur in the beam, the present Ne(3P2) polarizations are much greater than those obtained by optical pumping in weak discharges [ 1.21. Dye lasers can also provide radiation at the wavelengths required to optically pump beams of other heavy rare gas 3P2 metastable atoms [ 151. Again the resultant polarizations should be large and more than sufficient to permit study of polarization dependences not only in metastable atom---surface interactions, but also in interactions involving gas phase targets. It is a pleasure to acknowledge valuable discussions with G.K. Walters, J.P. Hannon and L.W. Anderson during the course of this work. This research was supported by the Materials Sciences Section, Office of‘ Basic Energy Sciences, U.S. Department of Energy. The Robert A. Welch Foundation and the Donors of the Petroleum Research Fund administrated by the American Chemical Society.
References [l]
L..D. Schearer,
Phys. Rev. 1X0 (1969)
83; 188 (1969)
505
[2] J.-P. Lemoigne, I:. Sage and D. Lecler, J. Phys. 39 (197X) 125. [3] ‘T.W. Riddle, M. Onellion, I:.B. Dunning and G.K. Walter\. Rev. Sci. Instrum. 52 (1981) 797. 141 G. Baum, C.D. Caldwell and W. Schrodcr, Appl. Phys. 31 (1980) 121. [5] D. Hils, W. Jitschin and H. Kleinpoppen, Appl. Phys. 25 39 (1981). [6] A. Fischer and I.V. Hertel, Z. Phys. 304 (1982) 103. [7] W. Dreves, H. Jansch, E. Koch and D. Fick, Phys. Rev. Lett. 50 (1983) 1759. [8] J.T. Cusma and L.W. Anderson, Phys. Rev. A28 (1983) 1195.
Volume 52, number 2
OPTICS COMMUNICATIONS
[9] M. OneIlion, M.W. Hart, F.B. Dunning and G.K. Walters, Phys. Rev. Lett. 52 (1984) 380. [lo] R.D. Rundel, F.B. Dunning and R.F. Stebbings, Rev. Sci. Inst. 45 (1974) 116. [ 1 l] L.A. WestIing, M.G. Raymer, M.G. Sceats and D.F. Coker, Optics Comm. 47 (1983) 212. [ 121 F.B. Dunning, T.B. Cook, W.P. West and R.F. Stebbings, Rev. Sci. Instrum. 46 (1975) 1072.
15 November 1984
[13] J. Bakos and J. Szigeti, Opt. Spectrosc. 23 (1967) 255 [Opt. Spectrosk. 23 (1967) 478. [ 14 ] J.J. Sakurai, Advances quantum mechanics (AddisonWesley, London, 1976). [15] K.W. Giberson, M.W. Hart, M.S. Hammond, F.B. Dunning and G.K. Walters, Rev. Sci. Inst. 55 (1984) 1357.
107