Please mind the construction work. This site is under construction as of April 27, 2015.

What is this?

This is a web site dedicated to the generation of a "continuous atom laser". It contains results of a now defunct project at the University of Utrecht, ran by the owner of this site, Johnny Vogels. The purpose of this site is to advertise this possibilty. Without feigning modesty, it is still the best way around to improve the available flux of degenerate cold atomic matter. The machine imagined here would supersede the flux of all other machines for generating ultra cold degenerate atomic matter together.

Who else was involved?
Louise Kindt, PhD candidate
Prof. Peter van der Straten

What type of continuous atom laser?

A continuous atom laser based on the cooling of a subsonic atom beam in a magnetic guide.

What is this not?

This project is about cooling a beam of atoms to degeneracy in a more effective way. It is not about creating continuous beams of degenerate matter from already degenerate matter.

Why is this a good idea?

(To my knowledge) nobody has succeeded in the cooling of a supersonic beam to condensation. We tried at LKB/ENS. A supersonic beam has a too lower internal collision rate and to high of a velocity.

Why is this a good idea?

It is a new way, more efficient way of making Bose Einstein condensates. Bose Einstein condensates are the playground of physicists. Many new states and phenomena have been discovered. With this technique, op to 107- 108 atoms per second can be cooled down and condensed. With conventional techniques, up to 106 atoms per second can be cooled down.

Why is it so much more efficient?

An optimized Zeeman slower can slow down over 1012 atoms per second. However, a magneto optical trap can load only 1010 or so. Evaporative cooling of large condensates occurs in over a minute. Therefor, if evaporative cooling is done on the contents of a single magneto optical trap, most atoms from a Zeeman slower are necessarily lost. By loading the contents of a magneto optical trap into a guide multiple times per second, a much higher portion of the Zeeman slowed atoms can be used for evaporative cooling. Guided fluxes of 2 1010 atoms per second have bean demonstrated.

How do you get a subsonic beam?

We use a geometry which generates a shock wave in the beam, converting a supersonic beam into a subsonic beam.

Isn't this equivalent to a ramjet?

Yes! In a scramjet engine the incoming air stays supersonic. In a ramjet, the incoming supersonic beam goes through a diffuser where a shock wave occurs. In a scramjet engine the same problems as in cooling a supersonic beam: A small time for traversing the system, and low compression. The classical example of a supersonic jet enige is the Mach 3 Lockheed A-12 and SR-71 blackbird with Pratt & Withney J58 (partial) ramjet. Currently, tests on scramjet engines are conducted, reaching Mach 9+.

Isn't this equivalent to a ramjet?

No! We used gravity as much as possible. We used uphill gravity to slow down the beam when the beam was supersonic. We used downhill gravity to slow down the beam when the beam was subsonic. We could do this because the temperature was just below a mK. Without gravity, one is limited in compression to the stagnation values for the density and collision rate. To do this in a ramjet engine, one would need to make it as high at an atmosphere.

Huh? On a downhill slope, the atoms will accelerate, will they not?

The beam will not accelerate due to a pressure gradient. A subsonic beam is comparable to a column in the atmosphere: The density is higher below. If you make a vertical tube as high as the atmosphere, and induce a flow, the velocity will be lowest at sea-level due to the 'atmospheric compression'.

Is this a regular shock wave?

It was actually better. A shock wave in a ramjet can become unstable, stopping the flow of air into the sytem. It's called inlet unstart. It can be caused by a low incoming beam velocity, generating an unstable shock wave in the converging geometry. It can also be caused by too much pressure downstream, pushing the shock wave into the converging geometry where it becomes unstable. Sophisticated 'bleeing' systems exist to stabilize the shock wave by letting air escape. In our system, bleeding occurred due to atoms being able to escape back to the atom source (a magneto optical trap).

What is special about magnetically guided atoms?

Typically, the guide is almost cylindrically symmetric. The mean free path is much larger than the radius of the beam. Therefor, the atoms maintain their axial velocity, angular momentum and radial energy between collisions. In the collisional theory, these parameters replace the three components of the velocity.

What is special about a magnetically guided atom beam?

A magnetically guided atoms have a high viscosity and especially a high thermal conductivity. This is due to a small high energy and high angular momentum part of the atoms spiraling around most of the atoms, which gives them a very long mean free path. In fact, for a linear trapping potential, the cross sectional area for thermal conductivity of the beam is 99 times the surface of the beam where the density is more than 1/e of the peak density. Without special suppression techniques, the optimal mach number to gravitationally slow down to is not very low.

What is special about a magnetically guided atom beam?

Magnetically guided atoms are so cold that gravity is always a consideration.

Are your atom touching the guide?

Our guide was constructed from 4 of 6 mm tubing with 4 mm internal diameter, with currents of 2x 300 to 2x ~850 A, generated by 2 of 15 kW ESS 20-750 Lambda EMI power supplies. The spacing between the wires was 2 mm, and atoms could be observed through a glass tube with low helium permeability. The structure was as open as possible. The minimum radius from the center of the trap to the guide was maintained. Nevertheless, we estimate that some pre-evaporative cooling was taking place on the guide wires of the guide.

What is special about loading the atom guide this way?

Even without evaporative cooling, this 'shock wave loading' is a really nice way to load a conservative trap.

How does the shock wave start?

We turn on a magnetic barrier downstream while injecting the supersonic beam. The magnetic barrier reflects the supersonic beam. When the two beams collide, a subsonic region develops, with the shock wave between them. Over time, the shock wave moves upstream until it reaches the converging part of the guide where bleeding starts to occur. In the mean time, the density downstream increases and increases, especially on a downward slope. The density is increased by gravity and pre-evaporative cooling.

What is special about this geometry?

You have to imagine that we were creating cold atom clouds of over a meters, with densities and collision rates sufficient to cool to Bose Einstein condensation. Usually one starts with trapped atoms of at most 10 mm long to cool to Bose Einstein condensation. We were loading the contents of hundreds of Magneto-Optical traps into a big magnetic trap, rather than the contents of just one.

What is special about this geometry?

When the length of the guide is pushed significantly further than we did, theoretically an instability can occur to high density, i.e. straight to Bose-Einstein condensation! (Especially if three-body decay is not taken into account.)

Is this a regular shock wave?

No. The mean free path of the atoms was much higher than the radius of the beam. The width of a shock wave is associated with the reciprocal of the viscosity of the beam, and is pretty long. When the shock wafer reaches the region where the wire tubes converge, it starts to bleed. In the hydrodynamic limit, the performance of system is not determined by the internal details of the shock wave, only by the mach number at which it occurs. The density, velocity, temperature and change before and after the shock wave are fixed by conservation of flux, momentum and energy and the mach number.

What do I need?

You need a a supersonic magnetically guided beam with optical density of over 0.5 to 1 (with the fields on, detuning optimized). Extend it to a a couple of meters in ultra high vacuum with an atom lifetime of >300 s, and you are in the ballpark. We generated a superb vacuum using ST707 non-evaporative getter strips, supported by 50 l/s ion pumps on either end of a 4 m guide for He and NH4, and long 250 C baking and leak testing. The ST707 getter strips could be heated in-situ using current.

What are the steps the atoms go through?
  1. An oven with Rubidium
  2. An increasing field Zeeman slower for the Rb-87 atoms.
  3. Capturing atoms in a magneto optical trap. The duty cycle of this step is about 50%.
  4. Pre-shaping the cloud to an elliptical shape with a 2D quadrupole field
  5. Optical molasses cooling and upward launching
  6. Optical pumping to the magnetically trapped |F=1, mF=-1> state with a transient transverse bias field
  7. Magnetic pre-guiding in a 2D quadrupole field
  8. Compression due to an increasing 2D quadrupole pre-guiding field
  9. Radial compression due to the 2D quadrupole field of the guide as the guide is entered
  10. Radial compression due to the 2D quadrupole field of the guide increases as the atoms move further into it. The current circuits of the guide is not only completed at the entrance of the guide, but also "shorted" further down.
  11. Deceleration and compression due to uphill gravity.
  12. The guide bends to horizontal.
  13. Radial compression due to the guide wires getting from 8 mm to 2 mm spacing. At this point, the velocity is optimized to where only a small fraction of the atoms return due to deceleration due to radial compression and gravity. The packets axially merge into a beam. As a beam, the mach number is close to unity.
  14. At this point, the beam should ideally have an optical density of about unity when probed in-situ at optimal detuning. The atoms are in a 1000 G/cm field (@ 400 A) with a small axial bias field, and at about 1 mK.
  15. The beam goes through the shock wave through collisions with other atoms.
  16. The beam slows down due to gravitational compression.
  17. The beam pre-cools against the wires of the guide.
  18. At this point, the beam is virtually at a stand still, and we demonstrated that the collision rates which are high enough for evaporative cooling to Bose Einstein condensation. This is how far we got, really. We got to shoot the BEC's further into the guide, technically making a new type of non-mode locked 'pulsed atom laser'.
  19. The beam should be put in a magnetic conveyor belt, which is moving at the required beam velocity. Such a magnetic conveyor belt consists of a series of moving magnetic barriers which are temporarily inserted adiabatically into the beam. This compresses the beam some more, suppresses the thermal conductivity from upstream to downstream, and may allow for a bit more efficient 3D cooling instead of 2D cooling. The conveyor belt moves as slow as is necessary for regular evaporative cooling. Note the inherent parallelism!
  20. Cooling of the beam through RF, with which one should be careful for RF leakage, or through shrinking apertures, which allow for only 2D cooling.
  21. The barriers can be lowered throughout the evaporation to perform decompression to decrease three-body decay.
  22. Depending on when the barriers are removed altogether, the beam emerges as a subsonic or supersonic beam from the conveyor belt.
  23. The final cooling to condensation can occur as a beam, because the mach number went up, suppressing the thermal conductivity.

Why didn't you complete the project?

The Rule of pi.

Where can I find technical details and pictures?

Louise Kindt's Thesis

What can you do with a continuous atom laser or microKelvin atomic beams.

  1. Create artificial black holes and white holes (through shock waves and supersonic expansions) of degenerate beams.
  2. Create black hole lasers
  3. Detect amplified hawking radiation
  4. Sympathetically cool other species
  5. Really, any experiment one would like to do with Bose Einstein condensates
  6. Atom interferometry
  7. Use your imagination!

Can this work in outer space?

Yes, with modifications. Ignoring power and space requirements, the hill in the guide can be, less effectively, replaced by a bias field or localized increased confinement.

Who else is/was working on this?

LKB/ENS under David-Guery Odelin and Jean Dalibard
University of Michigan under George Raithel.

Any viable alternatives?

The future will tell.


Is this website done?

No, I plan to add calculations about the effectivity of evaporative cooling in a guide and references.

2015 Copyrights by Dr. Ir. J.M. Vogels. All rights Reserved.

Please contact j-m-v-o-g-e-l-s-a-t-l-i-v-e-dot-c-o-m for questions, details, implementations, manufacturing, help, etc.