¿ìè¶ÌÊÓƵ

Quantum theory is one of the most successful theories of nature ever
devised. It describes accurately the behaviour of the microscopic entities
constituting the subatomic world. And yet physicists and philosophers have
been arguing about what quantum theory means for more than 60 years. Now,
new experiments have been proposed to test the theory’s most widely accepted
interpretation. At stake is our understanding of the very nature of physical
reality.

Problems with quantum theory arise because it has to take account of
the peculiar properties of microscopic objects such as photons, electrons
and atoms. We can do one type of experiment to show that these objects behave
like tiny projectiles, with well-defined paths or trajectories through space.
Another type of experiment will show that they can also behave like waves,
with their peaks and troughs adding or cancelling to give interference effects.
However, it seems that we can find no experiment to show both types of behaviour
at the same time. What you get depends on what you look for and how you
look for it.

In the late 1920s, the Danish physicist Niels Bohr argued that the dual
wave-particle properties of objects obeying the laws of quantum mechanics
are not contradictory, but complementary (see ‘What is light?’, ¿ìè¶ÌÊÓƵ,
2 November, 1991). He believed that this kind of behaviour places a fundamental
limit on what we can know about nature. We determine what kind of reality
can be revealed (waves or particles) by the type of experiment we choose
to set up. Before we make a measurement, the reality of a quantum object
is not determined.

This interpretation is totally counter to our intuition, which leads
us to expect that an object will continue to exist in some determined form
even though we cannot perceive it. Try closing your eyes tight and taking
your hands away from this magazine. Does it continue to exist in the same
form even though you can’t see or feel it? If your common sense tells you
that it does, why shouldn’t the same be true for photons, electrons or atoms?

There is an alternative interpretation which many find a little more
comforting. This is based on the German physicist Werner Heisenberg’s original
interpretation of his famous uncertainty principle, which he produced early
in February 1927, 65 years ago. The principle states that we cannot determine
simultaneously two so-called complementary properties of a quantum object
with absolute precision. Examples of complementary properties are an object’s
position in space and its momentum. If we can measure an object’s position
with absolute certainty, then there is an infinite uncertainty in its momentum,
and vice versa.

Originally, Heisenberg believed that his principle placed a limit not
on what we can know, but on what we can measure. To make his point he devised
a ‘thought’ experiment involving a hypothetical gamma-ray microscope. Using
gamma rays instead of light, such a microscope allows us, in principle,
to ‘see’ an electron. The more gamma-ray photons we bounce off the electron,
the sharper the image. But each photon gives the electron a severe jolt,
so that the momentum of the electron is changed in an unpredictable way
by the very act of trying to make the measurement.

We might be able to get a fix on the electron’s position at some instant
in time, but if we do, we will not be able to measure its momentum. If we
want to measure the electron’s momentum, we could use photons of lower energy.
But then we must give up hope of measuring its position. The measurement
of one complementary property excludes simultaneous measurement of the other.

Heisenberg’s interpretation implies that the position and momentum of
the electron are defined all along (one might say that they are always real),
but they cannot be measured simultaneously with absolute precision because
of the disturbance caused by the act of measurement. Our measuring devices
are simply too clumsy to detect reality without changing it.

However, Bohr was convinced that the position and momentum of the electron
are not real until they are measured. Bohr and Heisenberg argued intensely
over this point. Their argument comes down to this: which is more fundamental,
complementarity or the uncertainty principle? Does the uncertainty principle
arise from the complementary wave and particle properties of quantum objects,
or is it the other way around? Heisenberg eventually backed down and accepted
Bohr’s argument.

You might think that the difference between being able to know something
and being able to measure it is trivial. After all, how can we acquire knowledge
of something if there is a limit to our ability to measure it in the first
place? The difference depends on how we view reality. Are things real only
when we look? This is a question that some would prefer to leave to the
philosophers. But Marlan Scully at the Center for Advanced Studies, University
of New Mexico and Herbert Walther at the Max Planck Institute for Quantum
Optics in Germany, want to answer this question by doing experiments.

These experiments are related to another ‘thought’ experiment, originally
devised by Albert Einstein in October 1927 at the beginning of his legendary
debate with Bohr over the meaning of quantum theory. Einstein imagined what
might happen to a quantum object, such as an electron, if it were allowed
to pass through a narrow slit in a plate which is free to recoil in the
vertical plane (see Figure 1a). As the electron bounces off an inside edge
of the slit, the plate recoils and its movement – up or down – combined
with the law of conservation of momentum tells us the direction in which
the electron must have been deflected.

The electron now passes through a second plate which has two slits,
and is eventually detected on a piece of photographic film. The spot that
forms on the film indicates that an electron ‘struck here’. Einstein argued
that this position on the film can be combined with the movement of the
first plate, allowing the experimenter to trace the path that the electron
must have followed. Most important, the experimenter can find out which
slit in the second plate the electron must have passed through.

Now, suggested Einstein, suppose we let a large number of electrons
pass through this apparatus, one by one. In terms of the wave picture, the
plate with the two slits must produce an interference pattern of alternating
bright and dark fringes (see Figure 1b). Such a pattern is expected to build
up even though each electron is detected as a tiny spot on the film. However,
measuring the recoil direction of the first plate allows us to trace the
path of each electron through the apparatus. The experiment appears to show
both the particle (defined path or trajectory) and the wave (interference
pattern) properties of electrons at the same time (see Figure 1c). This
is something that Bohr’s interpretation of quantum theory simply does not
allow.

Bohr’s reaction was to take the thought experiment a step further. He
imagined that the first plate would be mounted on two weak springs, as shown
opposite. A pointer and scale are used to measure the amount of movement
caused by the impact of each electron as it passes through the slit. Bohr
argued that in order for the experimenter to measure the recoil of the first
plate, the pointer and scale must be illuminated with light. This means
that photons must be bounced off the first plate, which will cause it to
recoil in unpredictable directions. This ‘clumsiness’ in reading the position
of the pointer against the scale means that there is an uncertainty in the
position of the plate, and hence the slit, consistent with Heisenberg’s
principle.

Bohr was able to show that the uncertainty in the position of the slit
corresponds to the spacing between fringes in the interference pattern.
In trying to follow the path of each electron, the movement of the first
plate causes the interference pattern to become ‘washed out’. The result
is instead a random pattern that can be interpreted as the scatter of the
electrons as they ‘squeeze’ through the two slits.

By allowing the first plate to move, we can reveal the electron’s particle
properties, but not its wave properties. If we want to show its wave properties,
we must keep the first plate fixed in position, and so give up hope of measuring
the path of the electron.

Now Bohr’s argument rests on the ‘clumsiness’ of the measuring device,
in exact analogy with Heisenberg’s gamma-ray microscope experiment, rather
than his own principle of complementarity. Perhaps, like many physicists
after him, he felt that the uncertainty principle couldn’t be ‘beaten’,
leaving open the more fundamental questions about knowledge, measurement
and reality.

The American physicist Richard Feynman expressed this view in his famous
Lectures on Physics. He wrote: ‘If an apparatus is capable of determining
which (slit) the electron goes through it cannot be so delicate that it
does not disturb the (interference) pattern in an essential way. No one
has ever found (or even thought of) a way around the uncertainty principle
. . . if a way to beat the uncertainty principle were ever discovered, quantum
mechanics would give inconsistent results and would have to be discarded
as a valid theory of nature.’

Now, Scully and Walther have met Feynman’s challenge; they have thought
of a way around the uncertainty principle. However, they think that Bohr
was right when he argued that complementarity is the more fundamental principle.
As a result, they believe that quantum theory will not be invalidated by
their experiments.

These physicists have proposed a new experiment which can be carried
out using the new techniques of quantum optics (see Figure 2). Instead of
electrons, they propose to use atoms of rubidium. The wave properties of
atoms have been amply demonstrated in many recent experiments (see ¿ìè¶ÌÊÓƵ,
Science, 22 June 1991). The atoms are formed into two narrow beams and are
excited by a beam of light from a laser before passing through two slits.
Each slit makes the atoms act as spherical ‘atom waves’, which are described
mathematically in quantum theory in terms of ‘wavefunctions’. These wavefunctions
combine in the mathematical equations of quantum theory to produce a characteristic
two-slit interference pattern.

In front of each slit is placed a device called a micromaser cavity
which can detect which slit an excited atom passes through without significantly
disturbing the atom. This allows the researchers to follow the path of each
atom without it falling foul of the uncertainty principle. The cavities
can be used to provide ‘which way’ information characteristic of particles.

All this becomes technically possible through new developments in quantum
optics and one-atom masers . As each excited atom enters one or other of
the micromaser cavities, it falls back to a lower energy state, emitting
a photon of light with a wavelength in the microwave region. If the conditions
are right, a standing electromagnetic wave is produced and maintained inside
the cavity, not unlike a standing sound wave inside an organ pipe. We could
say that one of the cavities is ‘switched on’ by interacting with an excited
atom. By looking to see which cavity is switched on, the researchers can
tell which slit the atom must have gone through. Careful calculations show
that this can all be done without disturbing the atom’s passage through
the cavity. The physicists have thus found a delicate way of measuring which
slit an atom goes through in a two-slit interference experiment.

What should happen to the interference pattern? The pattern cannot be
washed out, as in Einstein’s thought experiment, because there is nothing
to recoil in the apparatus and so no uncertainty is introduced. You cannot
base an argument on the ‘clumsiness’ of the measurement, and so these experiments
provide a straight test of Bohr’s interpretation.

In fact, quantum theory makes a clear prediction. Interference is produced
when the atom wavefunctions from the two slits combine. Now, however, we
must take into account the fact that a standing electromagnetic wave has
been created in one of the micromaser cavities. Both cavities can be described
in terms of wavefunctions, one representing a cavity which is switched on,
and one representing a cavity which is switched off. When these cavity wavefunctions
are included in the mathematical equations, they erase the interference
pattern in the wavefunction of the combined atom.

The mere fact that ‘which way’ information has been obtained is enough
to wipe out the interference pattern (waves), leaving only the scatter pattern
(particles). Quantum theory predicts that, just as Bohr claimed, reality
is determined by the way the apparatus is set up, and the measurements it
makes.

Are these experiments technically feasible? Scully and Walther believe
that they are, but have set out to do some simpler but closely related experiments.
Instead of passing the excited rubidium atoms through two slits, they are
passing them through two micromaser cavities, one after the other. Both
cavities are set up so that an atom is much more likely to emit a microwave
photon inside the cavities than in free space , with the probability of
emission much greater in the first cavity than in the second cavity.

The basic principles of this kind of ‘two-field’ experiment were first
described in the 1950s by the physicist Norman Ramsey, and the outcome is
readily predicted by quantum theory. If it is not possible to tell in which
cavity the photon is emitted, the result is an interference effect, exactly
analogous to interference in the two-slit experiment, demonstrating the
wave-like characteristics of the atoms. But how can this experiment provide
‘which way’ information characteristic of particles? Interference fringes
are seen when the cavities already contain a large number of microwave photons
before they interact with a rubidium atom. The larger the number of photons,
the greater the amplitude of the standing electromagnetic wave in each cavity.
Both cavities are, in effect, already switched on and so it is impossible
to tell in which one the emission happens.

If, however, the cavities contain no photons initially (both cavities
are switched off), it becomes possible to tell in which cavity the emission
occurs because only one will be switched on by the atom. This is equivalent
to finding out which slit the atom goes through in the two-slit experiment.
Instead of ‘which way’ information, the experimenter obtains ‘which cavity’
information. Quantum theory predicts that under these circumstances, the
interference fringes should disappear.

The physicists are doing these experiments right now. If they find that
the fringes remain, even though both cavities are switched off before the
atoms pass through them, then they will have demonstrated the atoms’ wave-like
and particle-like properties simultaneously, Scully and Walther eagerly
await the results, but they are both convinced that the fringes will disappear
and complementarity will survive the test.

These researchers have gone on to ask another question likely to send
a chill down the spine of anyone trying to make sense out of all this. Returning
now to the two-slit experiment with micromasers, imagine that information
about which slit an atom goes through is ‘stored’ in one of the cavities,
and the atom is detected on the film. What happens if this ‘which way’ information
is then ‘erased’ before the experimenter looks to find out which slit it
actually was? On the one hand, if the atom emits a microwave photon in one
of the micromaser cavities the interference pattern must be destroyed. On
the other hand, if we don’t actually look to see which slit it was before
erasing the information, we cannot measure the atom’s path through the apparatus
and so we cannot show its particle properties. The atoms must appear as
waves or particles, so does the interference pattern come back?

Scully and Walther examined this situation using another idealised experiment,
shown in Figure 4. Now each micro-maser cavity contains a shutter which,
when closed, screens the photons in the cavity from a detector. If an atom
passes through the apparatus, it must leave a microwave photon in one of
the cavities before being detected on the film. Instead of looking to see
which cavity the photon is in, the experimenter opens both shutters. The
photon may now be detected, but in such a way that there is no possibility
of learning which cavity the photon was actually in initially. The ‘which
way’ information is erased before the experimenter looks at it.

Quantum theory predicts that the interference pattern should indeed
come back. But how can it? How can each atom ‘know’ in advance that the
shutters will be opened? How does the atom ‘know’ that it must be detected
at a position on the film such that, after many atoms have been detected,
an interference pattern results rather than a scatter pattern? It seems
paradoxical that by choosing to open the shutters or to keep them closed,
the experimenter can somehow influence events that have already happened.

The physicists have resolved this apparent paradox by carefully analysing
the prediction. When the shutters are opened, the microwave photon interacts
with the detector. The physics of this interaction is governed by the properties
of the atoms in the detector material. Quantum theory predicts that there
will be a 50 per cent chance that the microwave photon will be absorbed
(and hence detected) and a 50 per cent chance that it will not.

If we could somehow monitor the rubidium atoms producing the microwave
photons that are subsequently detected, according to quantum theory the
result would be an interference pattern, shown by the red curve in Figure
3. If, on the other hand, we monitor those atoms producing photons that
are not detected, the result is a similar interference pattern which is
shifted when compared with the first (the blue curve in Figure 3). If we
make no attempt to differentiate between the atoms, these interference patterns
simply combine to produce what appears to be a scatter pattern.

Thus, the interference pattern only ‘comes back’ if we correlate the
atoms with the ultimate fate of the microwave photons they left behind in
one or other of the cavities. This avoids the paradox of being able to make
choices which influence events that have already happened.

The conceptual and philosophical problems of quantum theory will not
be completely solved by the kinds of experiments, real or imaginary, discussed
here. But for perhaps the first time, Bohr’s idea of complementarity will
be put to its most stringent test. A test, moreover, that beats Heisenberg’s
uncertainty principle in a way that was unimaginable only a few years ago.
The imaginary and the unimaginable are becoming real through some fantastic
developments in quantum optics. In the next few years we should see the
results of experiments that strike right at the heart of one of the most
successful of scientific theories.

Jim Baggott is a freelance science writer and author of The Meaning
of Quantum Theory, to be published in April by Oxford University Press.

* * *

THE AMAZING ONE-ATOM MASER

A conventional maser has two essential ingredients. The first is an
‘active medium’ of atoms or molecules which can be excited so that there
are more of them in a high energy state than in lower energy states. The
medium can then be stimulated to emit microwave radiation as the excited
atoms or molecules simultaneously relax back to lower energy states. The
first masers, which were built in the 1950s, used ammonia gas as the active
medium. The ammonia molecules in higher energy states were physically separated
from the rest by passing them through an electric field.

The second ingredient is a cavity which can trap the emitted radiation
as a standing electromagnetic wave. Interaction between the standing wave
and the active medium then stimulates further microwave emission, adding
to the amplitude of the wave through positive feedback. In this way, the
microwaves are amplified by the stimulated emission of radiation (hence
the acronym maser).

The idea of a maser based on an active medium consisting of just one
atom seems incredible, but so-called micromasers can be built in the laboratory
thanks to recent progress in quantum optics.

Scully and Walther and their colleagues first pass a beam of rubidium
atoms through a device that selects atoms with a narrow range of velocities.
This allows the physicists to control the time each atom spends in the cavity.
After being excited by the laser, the atoms enter a superconducting cavity
made of pure niobium which is cooled to a temperature of 0.5 K using liquid
helium.

Using a superconducting cavity allows the physicists to isolate the
interior of the cavity from any stray electromagnetic fields outside. They
also use coils to compensate for the Earth’s magnetic field. This is necessary
to control how the rubidium atoms interact with the standing wave inside
the cavity, and allows the photons to be ‘stored’ in the cavity long enough
to build up a standing wave with a measurable amplitude. Cooling is necessary
to reduce the number of ‘thermal’ photons emitted from the inside surface
of the cavity, and the cavity is carefully designed to prevent photons from
outside leaking in. Such thermal photons add to the background ‘noise’ which
would otherwise mask the changes in the numbers of photons present in the
cavity which result from interactions with the rubidium atoms.

The first excited atom to pass through the cavity gives up some of its
energy by releasing a microwave photon through spontaneous emission. It
establishes a standing wave inside the cavity. Each subsequent atom may
then be stimulated to emit a photon as it passes through, adding to the
amplitude of the standing wave. With their present arrangement, the physicists
can build up and maintain a standing wave corresponding roughly to 100 photons,
even though only a single atom at a time passes through the cavity.

Just as different organ pipes are tuned by their shape and size to different
sound frequencies, the cavity can be tuned to different microwave frequencies
by changing its dimensions. By squeezing the cylindrical cavity using piezoelectric
elements, its dimensions can be tuned over a range of microwave frequencies
which includes the frequency of the photon emitted by the rubidium atom.
When the frequencies of the cavity and photon are exactly matched, the cavity
is said to be in resonance with the atomic emission and a standing wave
can be produced.

The standing wave may exchange energy with the atom several times as
the atom passes through the cavity. The physicists can tell when this is
happening by looking at the ionisation spectrum of excited atoms that emerge
from the cavity. Ionising the excited atoms allows them to be detected very
easily, and the number of ions detected is directly proportional to the
number of excited atoms present in the beam.

As the frequency of the cavity is tuned into resonance, the emission
of photons inside the cavity leaves the atoms in the lower energy state.
This results in a depletion of excited atoms which is detected as a fall
in the ion signal.

With 1750 atoms passing through the cavity per second, each taking about
50 millionths of a second, there are 0.09 atoms in the cavity on average.
The fact that the physicists still see a fall in the ion signal (see the
Figure above) demonstrates that the standing wave is being maintained by
interacting with single atoms.

The researchers can use the micromaser to tell when, on average, a single
atom has emitted a microwave photon inside the cavity. However, no recoil
has been introduced into the system as a result of the emission: the atom
emerges with a slightly lower energy but its linear momentum is unchanged.
The micromaser is therefore ideal as a delicate device for finding out which
slit an atom goes through in a two-slit interference experiment.