Optimized Bose-Einstein-condensate production in a dipole trap based on a 1070-nm multifrequency laser: Influence of enhanced two-body loss on the evaporation process

We present an optimized strategy for the production of tightly confined Bose-Einstein condensates (BEC) of ${}^{87}$Rb in a crossed dipole trap with direct loading from a magneto-optical trap. The dipole trap is created with light of a multifrequency fiber laser with a center wavelength of 1070 nm. Evaporative cooling is performed by ramping down the laser power only. A comparison of the resulting atom number in an almost pure BEC to the initial atom number and the value for the gain in phase space density per atom lost confirm that this straightforward strategy is very efficient. We observe that the temporal characteristics of evaporation sequence are strongly influenced by power-dependent two-body losses resulting from enhanced optical pumping to the higher-energy hyperfine state. We characterize these losses and compare them to results obtained with a single-frequency laser at 1030 nm.

Figure 1

Schematic top view of the dipole trap laser setup. The beam from the fiber laser (FL) is split in two parts at beam splitter PBS1. In each beam line an acousto-optic modulator (AOM1 and AOM2) controls the intensity. The initial foci are prepared with the lenses L1 and L2 and relayed into the vacuum chamber with achromatic lens pairs (L3, L5 and L4, L6). The polarization of the two beams is fixed with two perpendicularly oriented polarizing beam splitter cubes (PBS2 and PBS3). Monitoring of the beam intensities behind the chamber is achieved by photo diodes INT1 and INT2.

Figure 2

Time sequence of the total laser power during evaporative cooling. The inset gives the start and end values of the laser power together with the duration for each linear evaporation ramp.

Figure 3

Cross sections of the atom distribution at different final trap depths after the transition to BEC. (a) Small BEC and mainly thermal cloud right below the phase transition, (b) partially condensed cloud, and (c) almost pure BEC with a condensate fraction above 80$\%$. All data are taken after the same time-of-flight of 20 ms. The dotted lines represent bimodal fits to the density distribution while the dashed lines indicate the thermal fraction.

Figure 4

Atom number decay in the crossed dipole trap for different total final laser powers. The decay can be well fitted by a sum of two exponentials (see text) as plotted for the case of $1.6$ W of laser power where the data (crosses and short dashed curve) are compared to the fit (black line). The decay curves for a total power below $0.2$ W show the same behavior as the one for $0.2$ W and are not plotted.

Figure 5

Time constants ${\tau }_1$ and ${\tau }_2$ for atom loss and the time constant ${\tau }_{\mathrm{ramp}}$ representing the duration found experimentally for each ramp reducing the laser power by a factor of 2 as a function of laser power. For the ramp constant, the corresponding laser power is given as the average power during the respective linear ramp.

Figure 6

Temporal evolution of the total atom number and the number in state $F=2$ together with the resulting fit of the solution of the rate equations (2) for a total laser power of (a) $9.2$ W, (b) $3.0$ W, and (c) $0.76$ W.

Figure 7

Temporal evolution of the fraction of atoms in state $F=2$ for different total final laser powers.

Figure 8

Extracted pump rate $p$ and calculated rate of spontaneous Raman scattering events causing a change in the hyperfine state of an atom as function of trap laser power.

Figure 9

Power dependent two-body loss coefficients $\beta$ extracted from atom loss measurements as shown in Fig. 6 by fitting the solution of the rate equations (2). The error bars only represent statistical uncertainties. Expected systematic uncertainties are not included (see text).