† Corresponding author. E-mail:
We demonstrate the direct loading of cold atoms into a microchip 2-mm Z-trap, where the evaporative cooling can be performed efficiently, from a macroscopic quadrupole magnetic trap with a high loading efficiency. The macroscopic quadrupole magnetic trap potential is designed to be moveable by controlling the currents of the two pairs of anti-Helmholtz coils. The cold atoms are initially prepared in a standard six-beam magneto-optical trap and loaded into the macroscopic quadrupole magnetic trap, and then transported to the atom chip surface by moving the macroscopic trap potential. By means of a three-dimensional absorption imaging system, we are able to optimize the position alignment of the atom cloud in the macroscopic trap and the microchip Z-shaped wire. Consequently, with a proper magnetic transfer scheme, we load the cold atoms into the microchip Z-trap directly and efficiently. The loading efficiency is measured to be about 50%. This approach can be used to generate appropriate ultracold atoms sources, for example, for a magnetically guided atom interferometer based on atom chip.
Atom chips[1–4] are very promising tools for the precise control and manipulation of ultracold atoms. By applying modest electric currents on the chip wires, large magnetic field gradients and curvatures can be produced in close proximity to the chips.[5] A variety of trapping,[6] guiding,[7] transporting,[8] and the evaporative cooling[9–11] of cold atoms have been realized on the atom chips. The atoms can be cooled to hundreds of nK after evaporative cooling. And the ultracold atoms are ideally suited for high precision measurement experiments, such as atomic clocks[12] and atom interferometers,[13,14] as well as quantum statistics studies in Bose–Einstein condensate (BEC)[9,10,15] and related degenerate phenomena. In addition, the ultracold atoms held in close proximity to the chip surface are versatile probes for the atom-surface interactions such as the Casimir–Polder interaction,[16] local magnetic fields,[17] and current flow irregularities.[18]
In order to prepare the ultracold atoms, a large number of cold atoms are required to be loaded into a small-volume microchip trap, where the evaporative cooling can be performed efficiently. A typical method is that the atoms are firstly cooled and trapped in a standard six-beam MOT, then magnetically transferred close to the chip surface, and then loaded into the small-volume microchip trap.[15,19] However, various complex transfer wire configurations[10] or intermediate magnetic traps[15,20] are often required during loading the atoms from the macrocopic magnetic trap into the small-volume microchip trap. Therefore, we develop an approach of directly loading cold atoms from a macroscopic quadrupole magnetic trap into a 2 mm-scale microchip Z-trap. Since the evaporative cooling can be performed efficiently in the 2 mm-scale microchip Z-trap even to achieve BEC as shown in the literatures.[9,21,22]
In this paper, we report a direct loading of cold 87Rb atoms into a microchip 2 mm-scale Z-trap from a macroscopic quadrupole magnetic trap (QMT) with a high loading efficiency, and the evaporative cooling in the Z-trap is estimated to be highly efficient. The cold atoms are loaded into the microchip Z-trap by switching the trapping potential from the macroscopic quardrupole potential to the microscopic chip potential gradually, instead of using an additional intermediate magnetic trap. The cold 87Rb atoms are prepared in a standard six-beam MOT firstly, and then loaded into the initial QMT generated by the MOT coils. Then the atoms are moved vertically with the quadrupole magnetic field minimum towards the position near the chip surface, and the detailed process can be seen in our previous work.[23] By switching the trapping potential from the macroscopic quardrupole potential to the microscopic chip potential, the atoms are directly loaded into the microchip 2 mm-scale Z-trap from the final QMT. In addition, we use a three-dimensional absorption imaging system combining a grazing incidence imaging setup and an orthogonal-angle-of-incidence imaging setup to precisely optimize the position alignment between the atom cloud and the chip Z-trap. It is beneficial for improving the atom loading efficiency of the microchip Z-trap.
The experimental setup is illustrated in Fig.
We use two pairs of partly overlapping anti-Helmholtz coils, which are the MOT and transfer coils, respectively, to transport the atoms from the MOT center to the chip surface.[26–28] The center of the MOT coils is located at 28 mm below the chip surface so that the atom chip does not hinder the six-beam MOT configuration. The QMT can be moved vertically towards the chip surface by increasing the currents in the transfer coils while keeping the currents in the MOT coils constant. We use the three-dimensional printing technology to make the nylon skeleton for the coils, which can eliminate the effect of eddy currents that is always an intractable problem for the traditional metal skeleton. In addition, there are also three pairs of Helmholtz coils (not shown in Fig.
The whole experimental procedure mainly consists of three stages, the MOT loading stage, the QMT transport stage, and the microchip trap loading stage. The time sequence is shown in Fig.
![]() | Fig. 3. Time sequence of the atom transport process from the QMT into the chip Z-trap. The ZMT denotes the chip Z-trap. |
Initially, the 87Rb atoms are cooled and trapped in the standard six-beam MOT located at 28 mm below the chip surface. The cooling beam is red detuned 12 MHz from the
After the MOT loading stage, the atom cloud is compressed within 30 ms. The cooling laser red detuning is increased to 32 MHz and the repumping power is reduced to 1 mW, meanwhile the quadrupole field gradient is increased from 1.4 mT/cm to 4.0 mT/cm. Then we perform a polarization gradient cooling (PGC) process to further cool the atoms within 6 ms. We turn off the currents in the MOT coils, increase the cooling laser red detuning to 72 MHz, and decrease the power of the cooling laser beams to 10 mW simultaneously. The temperature of the atom cloud is about 20 μK after the PGC process. And then the atoms are optically pumped into the
Then, in the QMT transport stage, the trapped atoms are moved vertically with the magnetic field minimum from the MOT center towards the atom chip surface within 160 ms by increasing the currents in the transfer coils to 26 A linearly and keeping the currents in the MOT coils constant. Simultaneously, the x-bias magnetic field is increased to Bxbias1 linearly in the last 50 ms to align the quadrupole trap center with the chip Z-trap center. During the transport process, the atom cloud is compressed gradually, and no obvious loss of atoms is measured. The atom cloud is transferred to about 0.5 mm below the chip surface eventually. The number and temperature of the atoms in the final QMT are
At last, we perform the microchip Z-trap loading after the cold atoms are transferred close to the chip surface. Within a certain loading time t0, the currents in the MOT and transfer coils are linearly reduced to 0 A, the x-bias magnetic field is ramped from Bxbias1 to Bxbias2, and the current of the chip Z-shaped wire is increased to 3.5 A linearly. Thus, the trapping potential is switched from the macroscopic quardrupole potential to the microscopic chip potential gradually. As a result, the atoms are loaded into the microchip Z-trap directly.
In order to illustrate the feasibility of the atom loading scheme of the microchip 2 mm-scale Z-trap, we calculate and analyze the magnetic field distribution at different step in the loading process. We set
The atom cloud in either the macroscopic QMT or the microchip Z-trap is monitored by a three-dimensional absorption imaging system, which combines a grazing incidence imaging setup and an orthogonal-angle-of-incidence imaging setup, as shown in Fig.
The atom loading efficiency of the microchip 2 mm-scale Z-trap from the final QMT is influenced by the position deviation between the centers of the atom cloud trapped in the final QMT and the chip Z-trap. The number of atoms loaded into the microchip Z-trap can be approximately estimated to be the atom density times the overlapped volume of the two traps. The scale of the atom cloud is larger than the scale of the microchip trap, especially in the x-direction and z-direction. The scale of the chip Z-trap is 2 mm in the y-direction (also shown in Fig.
The atom cloud trapped in the final QMT is detected by the orthogonal-angle-of-incidence absorption imaging system. The absorption images of the atom cloud are shown in the Figs.
The dependence of the atom number in the chip Z-trap on Bxbias2 is shown in Fig.
The dependence of the number of atoms in the microchip Z-trap as a function of the chip Z-trap loading time is shown in Fig.
![]() | (1) |
![](cpb_26_3_033701/cpb_26_3_033701_ieqn033.gif)
The best fit of Eq. (
![]() | Fig. 8. Dependence of number of atoms in the chip Z-trap on the loading time. The solid curve is fit to the data using Eq. ( |
Finally, the atom transport efficiency achieves 50% from the macroscopic QMT to the microchip 2 mm-scale Z-trap after optimizing the experimental parameters. The number and temperature of the loaded atoms are
We have demonstrated a simple and efficient loading of cold atoms from a standard six-beam MOT into a 2 mm-scale microchip Z-trap. The atoms are firstly loaded into the standard six-beam MOT, then transferred vertically 27.5 mm up to the position near the chip surface after a simple QMT transport stage. Finally, the cold atoms are loaded into the microchip Z-trap by switching the trapping potential from the macroscopic quardrupole potential to the microscopic chip potential gradually, instead of using an additional intermediate magnetic trap. We have calculated and analyzed the magnetic field distribution at different steps in the microchip Z-trap loading process, and proved the feasibility of the direct atom loading scheme. We studied the mode-match between the macro QMT and the microchip 2 mm-scale Z-trap, and maximized the chip Z-trap loading efficiency by optimizing the transfer scheme and precisely adjusting the position alignment of the traps with the help of the three-dimensional absorption imaging system. Finally, we loaded
[1] | |
[2] | |
[3] | |
[4] | |
[5] | |
[6] | |
[7] | |
[8] | |
[9] | |
[10] | |
[11] | |
[12] | |
[13] | |
[14] | |
[15] | |
[16] | |
[17] | |
[18] | |
[19] | |
[20] | |
[21] | |
[22] | |
[23] | |
[24] | |
[25] | |
[26] | |
[27] | |
[28] | |
[29] | |
[30] |