Fundamental Limits for Coherent Manipulation on Atom Chips

The limitations for the coherent manipulation of neutral atoms with fabricated solid state devices, so-called ‘atom chips’, are addressed. Specifically, we examine the dominant decoherence mechanism, which is due to the magnetic noise originating from the surface of the atom chip. It is shown that the contribution of fluctuations in the chip wires at the shot noise level is not negligible. We estimate the coherence times and discuss ways to increase them. Our main conclusion is that future advances should allow for coherence times as long as 1 second, a few μm away from the surface. PACS: 03.75.-b Matter waves – 32.80.Lg Mechanical effects of light on atoms and ions – 03.67.Lx Quantum computation – 05.40.-a Fluctuation phenomena and noise In the quest for physical implementations of quantum information processing, “atom chips” are currently of great interest. This is because they promise well-controlled quantum optical manipulations of neutral atoms in integrated and scalable microtrap arrays. In these traps, atoms are strongly confined by electromagnetic fields close to nanostructured solidstate substrates. Microtraps used in current experiments are magnetic traps produced by current-carrying wires [1,2,3,4, 5,6,7,8,9,10] and periodically magnetized substrates [11], or hybrid traps involving optical or electric fields [12,13]. In this paper, we discuss the limitations that wire-based magnetic traps on atom chips may have to face when they are downscaled into the micron range. Recently, both theoretical and experimental indications have been found that the ‘hot’ chip substrate — typically held at room temperature — is not a quiet environment: at distances below a few 100μm from the chip, the trap lifetime is shorter than in free space and the atom temperature increases [5,6,14,15,16]. While it is not excluded that strong compression in these microtraps plays a role due to enhanced collisional interactions (see [17] for a review), noisy magnetic fields may also be involved. They provide a coupling to the environment that may cause loss, heating and decoherence and are elaborated upon in this paper. We review the sources of magnetic fields and quote estimates for trapping and coherence times. In particular, we discuss the contribution of electric current noise at the shot noise level and evaluate its spectral and spatial properties. This is compared to the noise due to the thermal chip substrate. 1 Atom chip ‘building block’: the side guide In the 1930’s, Frisch and Segré realized that when a homogeneous magnetic field (‘bias field’) is superimposed with the field of a straight wire current, the magnetic field vanishes on a line parallel to the current (see figure 1) [18]. In the vicinity of this line, the field increases in a quadrupolar fashion. The height of the field zero is given by