Research ArticlePHYSICS

An atom interferometer inside a hollow-core photonic crystal fiber

See allHide authors and affiliations

Science Advances  19 Jan 2018:
Vol. 4, no. 1, e1701723
DOI: 10.1126/sciadv.1701723


  • Fig. 1 Experimental configuration.

    (A) Experimental setup. NPBS, nonpolarizing beam splitter; PBS, polarizing beam splitter; DM, dichroic mirror; HP, half-wave plate; PM fiber, polarization-maintaining fiber; EOM, electro-optical modulator; AOM, acousto-optical modulator; APD, avalanche photodiode; HCF, hollow-core fiber. (B) Timing sequence for atom interferometer (not drawn to scale). Each experimental cycle takes 800 ms, including 580 ms for cooling atoms. The external magnetic fields Bx, By, and Bz are for canceling the ambient magnetic field, shifting the atomic ensemble, and defining the quantization axis for atom-light interaction. (C) Relevant energy-level diagram showing frequencies of laser beams used.

  • Fig. 2 Atoms inside the fiber.

    (A) OD versus probe time with push beams on and off. The probe time at tp = 0 corresponds to the time of releasing atoms from the MOT. Atoms are reloaded for each data point, and the optical dipole trap is off when the probe field is on. The error bars represent the standard error of five experimental runs. The inset shows an image of the cross section of the hollow-core fiber used in the experiment. (B) Measurement of atom temperature Tp in the radial direction at different probe times (see Materials and Methods). The release time in the x axis is the duration where the dipole trap is off for atoms to expand. The error bars indicate the standard error of seven experimental runs. The inset is a plot of temperature at different probe times.

  • Fig. 3 Mach-Zehnder interferometer.

    (A) OD versus phase with different interferometer times T. The fringes are fitted to a sinusoidal function. Each data point is an average of 60 runs with the standard error. (B) Phase shift versus time T. Each data point is extracted from the fitting of a sinusoidal function as in (A) with the Raman beam direction along and opposite to the gravity. The zero point in the y axis corresponds to the phase at T = 4 μs. The curves are the fits with Eq. 1. The error bars are the uncertainty of the sinusoidal fits. The red circles are the phase shifts based on the estimation of Eq. 1. (C) Contrast versus time. A comparison of the decay time with optical dipole trap on until tp and off during the interferometer sequence. The error bars are from the uncertainty of the sinusoidal fits.

  • Fig. 4 Interferometer at different locations of the fiber.

    (A) OD versus phase at different probe times. Data are fitted to a sinusoidal function. The standard error is an average of 60 experimental runs. (B) Velocity width measurement of atoms selected in the interferometer measurement. We use the sequence as shown in the inset. Each data point is from a sinusoidal fitting of interference fringe similar to Fig. 3A with T = 4 μs. The contrast decay is fitted to a Gaussian function. The error bars are from the uncertainty of the sinusoidal fitting.

Stay Connected to Science Advances

Navigate This Article