**4.3 Vacuum system**

For optimal conditions to form an atomic cloud it is necessary to reach an ultra high vacuum level with pressures lower than 10-7 Pa (10-9 Torr). Our vacuum system (Fig. 10) was built with pipes with nominal 2.75 inch diameter conflate type flanges made of 308 steel. The connections between the pipes and other devices were sealed with cooper gaskets. Our system consisted in a rotary vane pump, followed by a turbomolecular pump 4 and an ion pump 5. To measure the low vacuum level up to 1.33x10-2 Pa (10-4 Torr) we used a Convectron 6 gauge. To measure vacuum pressures lower than 1.33x10-3 Pa (10-5 Torr) we used a Bayard Alpert gauge 7. Both gauges were connected to a multi-gauge controller 8. The ultra high vacuum was measured alternatively with the indicator of ionic pump controller. The vacuum process started with the onset of the rotary vane pump to obtain a vacuum close to 1.33x10-2 Pa (10-4 Torr). After obtaining this vacuum pressure we started the turbomolecular pump, to obtain a vacuum close to 10-5 Pa (10-7 Torr). To obtain lower vacuum pressures the system was heated in a process called baking to evaporate the water molecules embedded inside the pipes and chamber. For this we rolled around the pipes and flanges along the vacuum line a heater that was made of a nearly 10 m long AWG26 nichrome wire. To electrically isolate the nichrome wire from the pipes we inserted it into a series of 1 m fiber glass spaghettis that were coupled one by one. To do this we slide the outer part at end of one spaghetti into the inner part of the following. The ionic pump was heated with its own heater, when the pump was switched off. The temperature used in the vacuum process was 120 ºC. To reach this temperature we increased the temperature 10 ºC every 30 minutes with a Variac transformer by increasing the current along the nichrome wire. The complete baking process took at least 5 days. The first day was used to reach the 120 ºC baking temperature. This temperature was kept constant during the next 3 days. In the fifth day we initiated the decrease of the temperature at the same rate as at the heating stage, that is a decrease of 10 ºC every 30 minutes. This was a precaution to protect the glass and the glue, because all have different temperature expansion coefficients. To obtain a homogeneous temperature along the vacuum line we made a temperature measurement at different places. For this we installed several thermocouples in some points between the heating wires and the pipes. We also covered the heater with aluminium foil. With the baking of the vacuum line we could reduce the pressure by more than one order of magnitude. The ultimate vacuum was less than 100 nPa (1nTorr).

## **4.4 Observation optical cell: discussion of different methods**

Three versions of observation cells were used in our trap. In the first case we bored a 30 mm hole in the centre of a 2.75 inch conflate type blank flange 9. On the flat side we constructed a

<sup>4</sup> Varian, Model TurboVac V50

<sup>5</sup> Varian, Model VacIon Plus 20 StarCell

<sup>6</sup> Granville Phillips, Model 275238

<sup>7</sup> Varian, 580 Nude ion gauge thoria iridium

<sup>8</sup> Varian, Model L8350301

<sup>9</sup> MDC-Vacuum, Model 110008

Fig. 10. Vacuum system a: optical table, b: turbomolecular pump, c: reduction nipple CF 4.5 to 2.75 inch 10, d, j: tee CF 2.75 flange 11 , e: Convectron vacuum sensor, f, p: nipple 12, g, o: manual valve for ultra high vacuum 13 , h: Bayard-Alpert UHV sensor, i: short nipple CF 2.75 flange 14, k: window 15, l: six way cross 16, blank flange for back side 17, m: 8 pins electrical feedthrough, n: bottle with seven horizontal windows and one vertical window, q: ion pump, r: aluminium plate support for ionic pump with dimensions 30x19.5x1 cm mounted in 4 rods of 2 inch diameter, s: L form mount for tubing.

cell that uses four optical glass plates with 4 mm wall thickness and dimensions 35 x 50 mm. On the top of the cell we glued a 35 x 35 mm optical glass plate. The cell was glued to the flat side of the flange. The plates were glued with high vacuum Torr seal 18. The second version consisted in an optical glass cell with outer wxlxh wall dimensions 55x55x52.5 mm and 2.5 mm wall thickness 19. The cell was glued on the 4.5 inch side of a zero length reducer from nominal conflate flange 4.5 inch to 2.75 inch. We did not remove the edge of the 4.5 inch side so the cell was installed very tight. This caused that the glass broke after some heat up vacuum procedures. The cell could be repaired several times with the vacuum Torr seal. The first two versions of cells are shown in Fig.11.

166 Quantum Optics and Laser Experiments

constant voltage to the PZT. This option is very useful for finding the resonances needed for

For optimal conditions to form an atomic cloud it is necessary to reach an ultra high vacuum level with pressures lower than 10-7 Pa (10-9 Torr). Our vacuum system (Fig. 10) was built with pipes with nominal 2.75 inch diameter conflate type flanges made of 308 steel. The connections between the pipes and other devices were sealed with cooper gaskets. Our system consisted in a rotary vane pump, followed by a turbomolecular pump 4 and an ion pump 5. To measure the low vacuum level up to 1.33x10-2 Pa (10-4 Torr) we used a Convectron 6 gauge. To measure vacuum pressures lower than 1.33x10-3 Pa (10-5 Torr) we used a Bayard Alpert gauge 7. Both gauges were connected to a multi-gauge controller 8. The ultra high vacuum was measured alternatively with the indicator of ionic pump controller. The vacuum process started with the onset of the rotary vane pump to obtain a vacuum close to 1.33x10-2 Pa (10-4 Torr). After obtaining this vacuum pressure we started the turbomolecular pump, to obtain a vacuum close to 10-5 Pa (10-7 Torr). To obtain lower vacuum pressures the system was heated in a process called baking to evaporate the water molecules embedded inside the pipes and chamber. For this we rolled around the pipes and flanges along the vacuum line a heater that was made of a nearly 10 m long AWG26 nichrome wire. To electrically isolate the nichrome wire from the pipes we inserted it into a series of 1 m fiber glass spaghettis that were coupled one by one. To do this we slide the outer part at end of one spaghetti into the inner part of the following. The ionic pump was heated with its own heater, when the pump was switched off. The temperature used in the vacuum process was 120 ºC. To reach this temperature we increased the temperature 10 ºC every 30 minutes with a Variac transformer by increasing the current along the nichrome wire. The complete baking process took at least 5 days. The first day was used to reach the 120 ºC baking temperature. This temperature was kept constant during the next 3 days. In the fifth day we initiated the decrease of the temperature at the same rate as at the heating stage, that is a decrease of 10 ºC every 30 minutes. This was a precaution to protect the glass and the glue, because all have different temperature expansion coefficients. To obtain a homogeneous temperature along the vacuum line we made a temperature measurement at different places. For this we installed several thermocouples in some points between the heating wires and the pipes. We also covered the heater with aluminium foil. With the baking of the vacuum line we could reduce the pressure by more than one order of

magnitude. The ultimate vacuum was less than 100 nPa (1nTorr).

**4.4 Observation optical cell: discussion of different methods** 

Three versions of observation cells were used in our trap. In the first case we bored a 30 mm hole in the centre of a 2.75 inch conflate type blank flange 9. On the flat side we constructed a

cooling.

4 Varian, Model TurboVac V50 5 Varian, Model VacIon Plus 20 StarCell 6 Granville Phillips, Model 275238

8 Varian, Model L8350301 9 MDC-Vacuum, Model 110008

7 Varian, 580 Nude ion gauge thoria iridium

**4.3 Vacuum system** 

<sup>10</sup> MDC-Vacuum, Model 402013

<sup>11</sup> MDC-Vacuum, Model 404002

<sup>12</sup> MDC-Vacuum, Model 402002

<sup>13</sup> MDC-Vacuum, Model 302001

<sup>14</sup> MDC-Vacuum, Model 468008

<sup>15</sup> MDC-Vacuum, Model 450020

<sup>16</sup> MDC-Vacuum, Model 407002

<sup>17</sup> MDC-Vacuum, Model 110008

<sup>18</sup> MDC-Vacuum, Model 9530001

<sup>19</sup> Hellma Cells, Model 704.003-OG

Cold Atoms Experiments: Influence of Laser Intensity Imbalance on Cloud Formation 169

repumping laser. The combined laser beams were simultaneously expanded by a laser beam expander consisting of a *f* = 50 mm lens followed by two *f* = 300 mm. The diameter of the three lenses was 25 mm. The diameter of the lasers was nearly 3 mm and at the exit it was 12

Fig. 13. Combination of repumping and cooling laser beams followed by simultaneous beam

After passing the iris diaphragm, both lasers were divided in a 0.3/0.7 divider. Most of the laser power (70%) was directed to the horizontal plane (Fig. 14). A polarizing beam splitter cube divided both lasers equally. Each pair of beams that leaved the polarizing beam splitter cube were divided again by means of two non polarizing beam splitter cubes. By this method it was possible to obtain two sets of counter propagating pairs of beams. In each leg of this arrangements quarter wave plates to with the correct circular polarizations. We installed a surveillance IR camera to observe the cloud and an IR CCD 22 with a 50 mm lens.

expansion. OGD = optical glass divider, L = lenses, HWP = half wave plate, PBSC =

polarizing beam splitter cube, ID = iris diaphragm.

22 Altec Vision, Model PL-B771U

mm giving an expansion of 4x. The laser disk was rounded by an iris diaphragm.

Fig. 11. Left: 55x55x55 mm glass cell, right: glass cell constructed with 35x50x4 mm plates.

The third version (Fig. 12) consisted in a cell prepared by a glass blower. The cell has a 2.75 inch conflate type adapter and 7 optical windows with 1 inch useful area 20.

Fig. 12. Side and top view of observation cell.
