Laser spectroscopy

Group Leaders:

What do we do?

In brief, we perform high resolution spectroscopy of atoms in order to measure energy shifts or phase effects due to symmetry violations or quantum electrodynamics.

Research Interests

Symmetry violations

Atomic QED and symmetry violations

Most atomic spectroscopy still concentrates on the quantum mechanical framework in place before the Second World war (WWII). In particular, we rely on three discrete symmetries of space-time which were formulated at this time. Initially, Wigner introduced space-inversion with respect to an origin, which became known as parity (P), which divided all quantum states into one of two alternatives, odd or even parity. Hot on the heals came charge-conjugation (C) symmetry, a result found by Dirac as he introduced anti-particles to explain his "negative energy" solutions. Finally, Kramer and Heisenberg formulated the idea of time reversal (T) symmetry. These symmetries helped to formulate numerous selection rules in spectroscopy, as well as provide a framework for understanding of nuclear and high energy particle physics.

Quantum mechanics did not stop here, however. After WWII, two major extensions of quantum theory appeared: the completion of quantum electrodynamics (QED) and the prospect of P-, C- and even T-violations in nature. These theories are often regarded as existing solely in the domain of high energy physics, but they can be felt at much lower energy scales. A simple example is the Lamb shift, which is a QED effect observable in radio-frequency measurements on a hydrogen atom. Our group attempts to study both QED effects and symmetry violations in atomic systems using lasers.

Parity and Time Reversal nonconservation

Symmetry has been a powerful tool in simplifying atomic spectra. However, by the early 1950s a number of theoreticians began to believe that C-, P- and T-violations may be possible. The only true invariant of nature appeared to be the product CPT, and by the 1956 the CPT theorem was a central pillar in physics. However, violations of the individual symmetries had not been recognised at this stage (though P-violation had actually been observed by accident in 1928, but no-one really noticed, much to the chagrin of the original discoverers). Within months, however, the decay of K+ mesons had been shown to violate P-conservation, whilst a year later (1957) beta decay was also shown to be a violation of parity. The cause of this violation lay beyond conventional quantum mechanics and required the introduction of a new force in nature: the weak interaction.

Feynman and Gell-Mann (and independently, Sudarshan and Maslak) demonstrated that the K+ meson decay was also a violation of C-conservation. T-symmetry, however, always felt like it must be conserved, and this possibly influenced Lev Landau to formulate the theory of CP-invariance (1957). However, in 1964, Fitch and Cronin observed the tiny production of pion pairs from K-meson decay, and thus CP-invariance was dead, whilst Fitch and Cronin said "thank you very much" and picked up a Noble Prize. But if CPT invariance is truth, and CP-violation reality, then T-violation does not belong to fantasy. Since the 1960s, T-violation has been sought in a number of experiments, including nuclear fission, beta decay and neutron scattering, but it has remained beyond view.
Parity non-conservation in atoms (PNC)

In order to study PNC, we require a transition that is electric dipole forbidden. The transition strength of such a resonance line will generally be composed of three components: the magnetic dipole (M1), the electric quadrupole (E2) and the electric dipole (E1) PNC. The M1 and E2 transitions have broadly the same selection rules but a couple of differences remain and thus it is possible to eliminate one of these contributions for some atoms: in lead, the E2 component can be eliminated. The PNC component manifests itself strongly in the dispersion of the resonance line.

Few calculations on expected PNC effects exist in the literature. These theories have concentrated on M1 transitions in lead, thallium, bismuth and cesium. This creates an unusual experimental difficulty: the possible lack of a suitable laser source operating at the transition frequency desired. Nick Edward's experiments on PNC in thallium were limited by the poor signal to noise ratio, a direct result of the low output of the semiconductor diode laser used. Thus, parallel to these experiments have been attempts to produce spectrally bright laser sources for these M1 transitions. To this end, Gareth Scourfield has been working on a new solid state ring laser which uses Forsterite as the lasing medium. This has a tuning range of between 1200 and 13500 nm, which enables it to probe the M1 transition in thallium. In addition, it overlaps the current vogue wavelengths for optical telecommunications.
Time-reversal non-conservation (TNC)

Direct experimental observation of T-violation has been proposed using a number of experimental schemes. The observation of an electric dipole moment (EDM) in an atom may be one method, but this EDM would possess a number of contributions, only half involving CP- or T- violation. Accounting for the all the additional contributions may be difficult, especially as there is no real consensus on what the magnitude of the electron EDM is.

Kozlov and Porslev proposed an experiment using counter-proporgating beams through an atomic gas in the presence of a strong electric field. Because the beams are travelling in opposite directions along a well defined axis of symmetry, they are the time reverse of each other (as far a quantum mechanics goes). The T-violating matrix elements are imaginary, and thus there is no energy shift, but there can be a phase shift between the two beams. This phase shift may be detected simply by homodyning the two beams. In the experiments of David Hodgkinson, the atomic vapour is thallium and he searched for evidence of T-violation about an M1 resonance.

Quantum Electrodynamics (QED)

QED measurements in atoms
QED background

Old quantum theory stated that the 2S and 2P states of hydrogen were degenerate. True to form, this was subsequently proven false by experiment. Lamb and Retherford measured a small splitting of the levels that is known today as the Lamb shift. It is the result of the emission and reabsorption of virtual photons by the electron ("self energy") in the atom. Thus, for a true picture of atomic energy states one must include the quantum theory of light and matter known by the moniker quantum electrodynamics or QED. Legend has it that the theoretical expression for the Lamb shift in hydrogen was formulated by Bethe whilst taking a train journey.

A false belief peddled when a student is introduced to quantum theory is that the hydrogen atom is the simpliest, and in fact only, multi-body system that can be sloved exactly. However, this is not the case: in many ways, the simpliest quantum system is muonium, the pairing of a positive muon with an electron. This is because both muon and electron are leptons, and when we enter the world of QED the physics of leptons is much better understood than that of hadrons, the family of particles to which the proton belongs. A simple example of the difference between leptons and hadrons is that leptons are essentially point like particles, where as hadrons have a finite (if small) size. When increasingly accurate determinations of the Lamb shift appeared, it soon became apparent that the size of the proton (which has still to be accurately determined) was embarassingly unknown.
1S-2S Spectroscopy in muonium

In collaboration with the University of Heidelberg, we are conducting experiments to measure the 1S-2S transition in muonium, in an attempt to determine the Lamb shift of the transition (1S and 2S shifts together). We attempt to do this via 2 +1 Resonance Enhanced Multiphoton Ionisation (REMPI) of the atom via the 2S state. The muonium is stripped of the electron, and the muon decays to a positron, electron and neutrino. The decay products of the muon are detected in a Time Of Flight (TOF) analyser. The coincidence between the laser pulse and the detection of decay products indicates that the positrons and electrons are from the muonium 2S state. The muonium is produced using a pulsed muon source (50 Hz) at the ISIS facility at the Rutherford Appleton Laboratory (RAL). A fraction of these muons are stopped in a silicon dioxide powder target and forms muonium by electron capture. Around 80 atoms leave the surface per muon pulse. They interact with two conterpropagating laser beams, which allow Doppler free spectroscopy of the 1S-2S transition.

The 2S state of hydrogen has a forbidden one photon electric dipole transition to the ground state (essentally, two photon emission is necessary) and thus the 2S state is long lived and the two photon transition line is very narrow. However, the muonium atom has a lifetime of only a few microseconds (the decay time of the positive muon), so there is a fundamental limit on the accuracy of the shift. The lifetime of the muon is 2.2 microseconds, and since both the ground and excited states decay with this time constant, the fundamental linewidth is 145 kHz. Nevertheless, experiment has yet to approach the natural linewidth of muonium. In the 1991 experiment, 28 ns uv pulses were produced from seeded dye laser amplifier, that unfortunately possessed a large amount of chirp- dispersion of the output radiation wavelength. Because of instabilities in the dye flow, the chirp also varied from pulse to pulse. Finally, the short pulse length limited the frequency accuracy to at most 5MHz. The most recent determination of the 1S-2S transition frequency has an experimental uncertainty of 46 MHz. Our next experiment hopes to reduce this error by a factor of at least ten.

In order to improve the accuracy, the dye laser is replaced by a Ti:Sapphire solid state laser seeding an Alexandrite laser amplifier at 409 254 660 MHz. To lock the laser to an absolute frequency, we are developing at Oxford, in collaboration with the University of Novosibirsk, a new frequency standard based on hot iodine. The technique of Frequency Modulation spectroscopy is used to produce derivative-like hyperfine spectra suitable for locking. (Click on the laser spectroscopy section for more details).

The Alexandrite ring amplifier produces 150-200 ns pulses, leaving a transform-limited pulse width of about 1 MHz. In addition, the chirp should be reproducable from pulse to pulse, and efforts are to be made to reduce this further. The output is frequency tripled to 1 227 760 GHz (around 244 nm) in LBO and BBO crystals. This radiation probes the cloud of transient muonium atoms formed just above the silicon dioxide target.

We have recently set up this experiment at RAL and performed a preliminary run. Several muonium "events" were recorded, but a final result will not be possible until the run is repeated next May.

Laser spectroscopy

Our group has always had a strong interest in atomic spectroscopy thanks to the work here on parity-violation in atoms. Recently, we have begun measurements on the molecular spectroscopy of iodine as part of the work towards a new frequency standard in the near infra-red (732 nm).
laser setup

The FM spectroscopy technique.

A Ti:sapphire laser (Microlase MB 101) is pumped by a 10 W Ar ion laser (Spectra Physics 6030), producing around 500 mW of 732 nm light with a 7 W pump. The output beam is horizontally polarised. A thick glass plate is used to separate 4% of the light for use as a probe beam in the FM spectroscopy. The remaining light (pump beam) is frequency shifted by 80 MHz in an AOM (Isle Optics LM 80) and amplitude modulated at 93 kHz. The polarisation of the pump beam is then rotated with a half-waveplate and reflected into the iodine cell by means of a polarising beam splitter. The counterpropagating probe beam passes through a home-made resonant Electro-Optic Modulator (EOM), where it acquires an FM modulation of 10.7 MHz and a modulation index of >0.5. The probe overlaps the pump beam and passes through the iodine cell and the polarising beam splitter, whereupon it is detected by a fast photodiode.

The Iodine Spectra

For the muonium experiment we required a sharp transition line within 1 GHz of 409 254 GHz (732 nm). No atomic line was suitable, so we turned to molecular candidates. Iodine is a popular molecular for frequency standards in the visible region. For example, iodine is a the basis of frequency standards at 532 nm (2nd harmonic of a Nd:YAG laser), 605nm, 612 nm and 632 nm, the latter used in iodine-stabilised He:Ne lasers. The iodine spectrum extends far into the infra-red, and lasers have been stabilised on iodine at 790 nm. However, the iodine atlas of Gerstenkorn and colleagues does not list a line close enough to our required frequency. The transitions listed in the atlas will be dominated by those from low lying vibrational levels in the electronic ground state, and it is possible that transitions from higher levels may be suitable. Thus, our colleagues in Novosibirsk investigated hot iodine in this frequency region, and discovered a weak transition (now identified as the 5 - 13, R(26) line) within the frequency window. This that time, both groups have concentrated on producing a suitable Doppler-free transition capable of stabilising the Ti:sapphire laser for the muonium spectroscopy.

To populate such a high vibrational state, a high temperature is required. Modelling using the Boltzmann distribution indicates that with a cell temperature of 650 degrees, less than 0.1% of the molecules will be in the v = 13 vibrational level. However, at such temperatures the iodine pressure in the cell would be enormous, and any attempt at Doppler free spectroscopy would be nullified by the broadening due to collisions. Thus, in order to maintain a low pressure in the cell one must restrict the supply of iodine molecules available. this is done by using a separate cold finger attached to the cell. The temperature of the cold finger determines the vapour pressure of the cell, whilst the wall temperature of the cell determines what fraction of the iodine vapour is in the required vibrational level. The drawback of this method is that you need to heating elements, one for the cell body and one for the cold finger. An oven was contructed here in Oxford capable of reaching 900 degrees celsius, whilst a Peltier cooler is used for the cold finger

Using the FM spectroscopic technique outlined above, one can produce Doppler-free spectra of the hyperfine structure of the 5 - 13, R(26) transition. The linewith of each individual transition is a result of collisions between the iodine molecules. This is a common problem when studying transitions from high lying vibrational levels because often a high vapour density is necessary to ensure a high molecular population in the vibrational state required. Thus, our primary concern is improving our signal/noise ratio so we can operate at the lowest possible cold finger temperature. For example, with the cold finger at 45 degrees celsius, the linewidth is 21 MHz, but reduce the temperature to 30 degrees and the linewidth falls to 14 MHz. Such a spectrum is shown in the picture below.