Atomic Unique Atomic Spectroscopy Aims at Answering a Universal Question

Richard Gaughan

There is far more matter cover the detailed properties of the than antimatter in our antiproton. In 1930 Paul Dirac theoretically pre- universe, but scientists dicted the existence of these antimatter don’t know enough particles, which are the exact opposite of about the properties of common particles. The positron was ex- perimentally verified in 1932, while the antimatter to understand _ antiproton (p) was not observed until why. By spectroscopically 1955. Why did it take so long to experi- analyzing atoms created mentally identify the antiproton? One reason is that it is essentially nonexistent when antiprotons collide in our earthly environment, and it can with helium, physicists only be produced in particle accelerators at CERN are measuring more powerful than those required to produce positrons. the properties of The absence of antimatter was not a antimatter with subject of much concern until physicists unprecedented accuracy. began to improve our understanding of the origin of the universe. The early uni- verse, filled with dense energy, almost instantaneously expanded to the point where matter condensed from the initial PHOTODISC INCORPORATED PHOTODISC sea of energy. So where is the problem? e live in a universe con- The problem is that our current un- structed from atoms com- derstanding predicts that as the universe posed of light, negatively cooled, both matter and antimatter charged electrons orbiting should have been produced in roughly Wpositively charged protons and un- equivalent quantities. Something has to charged neutrons (both heavy). We account for the observed matter pre- know now that our universe could have dominance in today’s universe, and sci- been predominantly composed of anti- entists around the world are searching matter — atoms with light, positively for possibilities. The researchers of the charged positrons orbiting a nucleus “Atomic Spectroscopy and Collisions containing negatively charged antipro- Using Slow Antiprotons” (ASACUSA) Richard Gaughan tons and antineutrons, which, although experiment at CERN are seeking the an- is a freelance science uncharged, differ from neutrons in hav- swer within the details of antiproton journalist and optical systems engineer at ing opposite magnetic properties. properties. Mountain Optical The question of why we live in a uni- If the antiproton is a perfect analogue Systems Technology, P.O. verse of matter as opposed to antimatter to the proton, then its charge will be ex- Box 1589, Boulder Creek, is being addressed in what may seem to actly opposite, and its mass and mag- CA 95006. He may be be an unlikely place, using an unlikely netic moments will be identical. There contacted by e-mail at rgaughan@ technique. Physicists at CERN, the Eu- is, of course, a significant obstacle to mountainoptical.com. ropean Laboratory for Nuclear Physics measuring these quantities: antiprotons in Geneva, Switzerland, are using inno- don’t hang around very long. When a vative spectroscopic techniques to un- proton and antiproton meet, they mutu-

32 Spectroscopy 18(1) January 2003 www.spectroscopyonline.com Atomic

ally annihilate, leaving only a burst of L pions. And, as previously noted, our universe has plenty of protons around, Ϫ55 eV so quick annihilation is the rule. Pro- tons and antiprotons are used as probes to investigate the structure of matter and antimatter, which is one reason that Ϫ p ␣ p ␣ e

particle accelerators such as Fermilab B,n (Argonne, IL) and CERN create high- ␣ energy particle beams and observe the products of their collision with various targets. 39,38 ps fs ps In 1991 researchers at KEK, the accel- 38,37 erator located in Tsukuba, Japan, noted 37,36 an anomaly in the time spectrum of Ϫ 80 eV 36,35 annihilations following the collision of antiparticles with cryogenic- 35,34 temperature helium atoms. The annihi- 34,33 lation spectrum showed the expected nearly instantaneous burst — over in 33,32 nanoseconds — but in addition, a 20- s-long tail was observed in the annihi- 32,31 lation spectrum. About 3% of the anti- protons incident on the helium target n,L ϭ 31,30 were not immediately annihilating. In- stead, the annihilation was somehow Figure 1. Antiprotonic helium is created when an antiproton replaces an electron in a helium atom. delayed. The experimental accuracy was The initial state is far from the ground state, with an antiproton orbit that keeps it far from the _ limited by the presence of other parti- nuclear protons. Laser excitation at energies resonant with energy level differences within the p cles in the incident beam, so the Japan- helium induce a transition to less stable states, where the antiproton rapidly collides with a nuclear ese team moved their experiment to proton, leading to a peak in the annihilation rate. CERN, where a higher-purity antipro- ton source was available. With the higher-quality CERN data, the explana- teractions of the particles within the system: the alpha particle of the helium tion was soon confirmed: the negatively atom. “The challenge is always to reach nucleus, the antiproton, and the re- charged antiprotons were replacing one ever-higher precision, ever-tighter com- maining electron. The three-body of the electrons in the helium atoms, parisons between the proton and an- quantum electrodynamics (QED) cal- creating a metastable antiprotonic he- tiproton, the world and the antiworld,” culation, although complex, is a tenable lium atom molecule, or atomcule (see said John Eades, a physicist on the problem, and extremely accurate theo- Figure 1). ASACUSA team. “The universe is a very retical predictions_ can now be made. The creation of this relatively long- big place and has been around for a The initial p helium atom is not a lived exotic atom was a fascinating re- very long time. There is time and space ground state system. Although the re- sult in its own right, but even more in- enough for extremely tiny violations of maining electron is typically in the triguing were the experimental the symmetry laws that we believe to ground state, the antiproton starts out 1/2 possibilities inherent in the new atom. govern its behavior to have produced in a state with n (M*/me) 38, In the same way that the simplest atom, very large effects, like the observed where n is the principal quantum num- hydrogen, with its single proton and matter–antimatter imbalance.”To de- ber (for the ground state_ n 1), M* is single electron, served as the principal tect differences between protons and the reduced mass of the p helium sys- testbed for investigating the properties antiprotons, a set of theoretical predic- tem, and me is the mass of the electron. of the electron and proton (and indeed, tions must be developed. The initial orbit is nearly circular, with for discovering the fundamental laws of l, the antiproton angular momentum quantum_ mechanics), antiprotonic he- The Theoretical Model quantum number, n 1. The lium (p helium) could serve as a test- If antiprotons are slowed enough, and metastability is a consequence of the bed for investigating the characteristics helium atoms are dense enough, then large angular momentum difference be- of the antiproton. Again, similar to the all the antiprotons will be captured in tween the antiproton and electron or- investigations into the hydrogen atom the target by helium atoms, each_ replac- bits, which makes it difficult for the en- of the last century, spectroscopy would ing one of the orbital electrons. p he- ergy to be transferred directly from the serve as a unique window into the in- lium can be thought of as a three-body antiproton to the electron. In addition,

January 2003 18(1) Spectroscopy 33 Atomic

ton mass is identical to the proton 0.84 mass, within 60 ppb. Eades noted that this accuracy is equivalent to “measur- 0.82 ing the height of two mountains the 0.80 height of Everest, and finding them the same to within about half a millimeter.” 0.78 Extending the Investigation 0.76 With the first-order transition energies (arbitrary units) 0.74 confirmed, the ASACUSA group turned ϩϩ R itself toward measuring higher-order 0.72 effects. Each (n,l) state is modified by spin-orbit interactions. The antiproton 0.70 total magnetic moment is a function of both angular momentum quantum 0.68 number and spin quantum number, where the angular momentum depends 12.87 12.88 12.89 12.90 12.91 12.92 12.93 12.94 12.95 on orbital characteristics, while spin refers to the intrinsic magnetic moment MW Frequency (GHz) of the respective particles. In the initial orbit, the antiproton’s angular momen- _ tum is on the order of 38, while its spin Figure 2. In addition to being split by electron spin magnetic interactions, each energy level of the p 1 He is further split by the smaller interactions of the antiproton spin. By tuning a laser to the is ⁄2, as indeed is that of the electron. transition between one super hyperfine structure level and the energy level associated with the next However, the electron’s spin moment principal quantum number the population of the SHFS levels is indicated by peaks in the results in a much stronger magnetic annihilation resonance. Frequency scanning a microwave beam repopulates the SHFS levels at the moment, so that the predominant in- resonance energies, which are again detected by annihilation peaks. teraction is between the antiproton’s or- bital angular momentum and the elec- tron’s spin. This hyperfine structure _ (HFS) is further complicated by the en- the initial p helium state is also well- third harmonic of a Coherent Laser Di- ergy splitting of antiproton spin-orbit shielded from collisions, minimizing vision (Santa Clara, CA) Q-switched interactions, which creates a super hy- the interaction with electric fields from Nd:YAG laser pumps a Lambda Physik perfine structure (SHFS). The end re- other_ nearby atoms. The consequence is (Göttingen, Germany) dye laser to pro- sult is that each (n,l) state is first split that p helium deexcites by radiating duce 4–6 ns pulses at wavelengths from into a doublet with level spacing of optical-frequency photons as the an- 529–745 nm. To produce pulses from 10–15 GHz, then each of those levels is tiproton cascades through a series of about 264–372 nm, the dye laser output split into a doublet with 100 to 300 constant n 1 1 metastable states. At is doubled in a -barium borate crystal. MHz level spacing. Again, if the energy n 32 and lower, collisions with other As the laser wavelength is varied, ab- level spacing created by the HFS and atoms in the target chamber rapidly sorbed_ photons induce transitions in SHFS can be compared with prediction, bring the antiproton orbit close enough the p helium. If the final state is not values of both the antiproton’s orbital to the nucleus that close-range, strong- metastable, resonances in the annihila- and spin magnetic moments can be de- force interactions dominate, and the tion rate are detected (see Figure 2). termined. The antiproton’s orbital mag- antiproton subsequently annihilates Varying the laser pulse wavelength netic moment is known as the antinu- with a nuclear proton. provides information_ for a detailed clear magneton._ The annihilation products, the burst spectrum of p helium. Already more When p helium is created, each of of pions, are detected by Cerenkov radi- than 20 resonances have been identified the four quadruplet substates is equally ation emitted as the particles travel and compared with predictions. The represented within the population. Al- through transparent sheets of Lucite. transition energy is a function of the though transitions between the quadru- Approximately 97% of the incident an- Rydberg constant, which is propor- plet states can be induced by irradiating _ 2_ _ tiprotons immediately annihilate, while, tional to QpMp,where Qp is the charge with microwave radiation of the proper _ without intervention, the remainder of the antiproton and Mp is the an- frequency, all this does is redistribute annihilate along one of several possible tiproton mass. The agreement between specific atoms in different energy states, 1/e distributions of about 3 s. The theoretical and experimental transition ending with the quadruplets still ASACUSA scientists, however, do pro- energies combined with previous meas- equally populated. To make the energy vide additional intervention, in the urements of the antiproton charge to splitting detectable, the ASACUSA re- form of laser irradiation. The second or mass ratio confirmed that the antipro-

34 Spectroscopy 18(1) January 2003 www.spectroscopyonline.com searchers use a laser–microwave–laser initial analysis makes it appear that triple resonance method. great surprises are unlikely, the results The first laser pulse, triggered by the of this experiment will firm up our un- arrival of a packet of antiprotons into derstanding of the antiproton’s notori- the helium target chamber, is tuned to ously poorly known magnetic proper- excite a transition from a specific SHFS ties. Because protons do not orbit the pair of a metastable level to a short- nucleus of any known atom, this meas- lived level. This pulse does two things. urement of the ratio of antiproton First, it creates an annihilation rate magneton to proton magneton is the pulse with intensity proportional to the only direct evidence of their number of atoms excited into the un- equivalence. stable level. Second, it depletes one Consistent with Francis Bacon’s SHFS pair of the quadruplet states asso- statement that “if a man will begin with ciated with level (n,l). Then microwave certainties he shall end in doubts; but radiation of a specific frequency is di- if he will be content to begin with rected to the target chamber. If the mi- doubts, he shall end in certainties,”the crowave frequency is resonant with the ASAC-USA team is constantly striving energy required to flip the electron’s to achieve higher accuracy. Eades sum- spin direction relative to that of the an- marized this attitude: “Is the mass of tiproton’s orbital angular momentum, the antiproton the same as the mass of then the incident pulse will induce the proton to one part in a million? Yes! transitions, tending to equalize the pop- Then let’s see if at one part in ten mil- ulations of the lower and upper lion, they are different. Still the same? quadruplet pairs once again. But this Yes! Then how about one part in a hun- resonance is undetectable because it dred million?” does not induce a change in the annihi- lation rate. If, however, a second reso- More on the Way nant laser pulse is directed to the target The ASACUSA experimental plan_ in- chamber, then any atoms that have cludes measuring many more p helium transitioned to the lower doublet will spectra, including some measurements again be stimulated into an unstable of even more exotic atomcule configu- state, leading to a detectable pulse in the rations that have already begun. The re- annihilation rate. The measured quan- searchers also hope to improve the ex- tity is then I(t2)/I(t1), with I(t) repre- perimental accuracy, and extend the senting the Cerenkov radiation inten- precision of the comparison of protonic sity at time t. If the microwave energy is and antiprotonic quantities. The plan is off resonance, then the pulse at t2 will ambitious. The measurements being excite no atomcules, and there will be made on antiprotonic helium essen- no measurable annihilation rate in- tially repeat “many of the classic experi- crease. Peaks in the t2/t1 intensity ratio ments in the history of the hydrogen as the microwave frequency is scanned atom which took a century or more,” correspond to HFS energy level Eades said. “This we are doing in a few differences. years! The first result, which the ASACUSA “We're aiming at high-precision team recently published (Phys. Rev. measurements of as many of the an- Lett., 9 Dec. 2002), concerns the tiproton’s properties — mass, charge, strength of the antiproton’s orbital magnetic properties, et cetera — as we magnetic moment. This is the product can get at. The aim is ever-more-precise of the so-called antinuclear magneton comparisons of the antiproton with the _ _ (proportional to Qp/Mp), the angular proton, because even a minute differ- momentum quantum number, L, and ence would be tremendously important an additional orbital g-factor, gl, which to our picture of, and interpretation of, is defined equal to 1 for the proton. If the world.” ■ the antiproton is not the perfect oppo- site of the proton, then the antiprotonic g-factor will differ from 1. Although

Circle 29 36 Spectroscopy 18(1) January 2003 www.spectroscopyonline.com