Throughout physics, stable composite objects are usually formed via attractive forces, which allow the constituents to lower their energy by binding together. Repulsive forces separate particles in free space. However, in a structured environment such as a periodic potential and in the absence of dissipation, stable composite objects can exist even for repulsive interactions. Here we report on the first observation of such an exotic bound state, comprised of a pair of ultracold atoms in an optical lattice. Consistent with our theoretical analysis, these repulsively bound pairs exhibit long lifetimes, even under collisions with one another. Signatures of the pairs are also recognised in the characteristic momentum distribution and through spectroscopic measurements. There is no analogue in traditional condensed matter systems of such repulsively bound pairs, due to the presence of strong decay channels. These results exemplify on a new level the strong correspondence between the optical lattice physics of ultracold bosonic atoms and the Bose-Hubbard model [1,2], a correspondence which is vital for future applications of these systems to the study of strongly correlated condensed matter systems and to quantum information.Cold atoms loaded into a 3D optical lattice provide a realisation of a quantum lattice gas [1,2]. An optical lattice can be generated by pairs of counterpropagating laser beams, where the resulting standing wave intensity pattern forms a periodic array of microtraps for the cold atoms, with period a given by half the wavelength of the light, λ /2. The periodicity of the potential gives rise to a bandstructure for the atom dynamics with Bloch bands separated by band gaps, which can be controlled via the laser parameters and beam configuration. The dynamics of ultracold atoms loaded into the lowest band of a sufficiently deep optical lattice is well described by the BoseHubbard model with Hamiltonian[1, 3]are destruction (creation) operators for the bosonic atoms at site i. J/h denotes the nearest neighbour tunnelling rate, U the on-site collisional energy shift, and ε i the background potential. The high degree of control available over the parameters in this system, e.g., changing the relative values of U and J by varying the lattice depth, V 0 , has led to seminal experiments on strongly correlated gases in optical lattices, e.g., the study of the superfluidMott insulator transition[4], the realisation of 1D quantum liquids with atomic gases [5,6] (see also [7, 8]), and the investigation of disordered systems [9]. 3D optical lattices have also opened new avenues in cold collision physics and chemistry [10,11,12,13].A striking prediction of the Bose-Hubbard Hamiltonian (1) is the existence of stable repulsively bound atom pairs. These are most intuitively understood for strong repulsive interaction |U| ≫ J, U > 0, where an example of such a pair is a state of two atoms occupying a single site,This state has a potential energy offset U with respect to states where the atoms are separated (see Fig. ...

We produce Bose-Einstein condensates of two different species, 87 Rb and 41 K, in an optical dipole trap in proximity of interspecies Feshbach resonances. We discover and characterize two Feshbach resonances, located around 35 and 79 G, by observing the three-body losses and the elastic crosssection. The narrower resonance is exploited to create a double species condensate with tunable interactions. Our system opens the way to the exploration of double species Mott insulators and, more in general, of the quantum phase diagram of the two species Bose-Hubbard model. Ultracold atomic gases seem uniquely suited to experimentally realize and investigate physics long studied in the domain of condensed matter and solid state physics, with the distinct advantage that specific effects are better isolated from unnecessary complications often present in condensed samples. The paradigmatic superfluid to Mott insulator transition of a condensate in an optical lattice [1] confirmed the predictions of the Bose-Hubbard model [2,3], originally introduced to describe superfluid Helium. With two species, the zero-temperature diagram of quantum phases is much richer than the simple duplication of the single species ' [4]. Indeed it has been proposed that two species obeying an extended BoseHubbard model can mimick the physics of lattice spins described by the Heinsenberg model [5,6] and give rise to yet unobserved quantum phases, like the double Mott insulator and the supercounterflow regime [7], with peculiar transport properties. Therefore, a double species condensate in an optical lattice stands as a promising candidate system for quantum simulations. Recently, the investigation of the two-species BH was started from the regime where one species exhibits the loss of phase coherence usually associated with the Mott insulator transition, while the other is completely delocalized [8]. Already at this stage, the presence of two species leads to a surprising shift of the critical point, which is now object of intense theoretical work [9].In addition, a double Mott insulator is expectedly extremely useful to produce heteronuclear polar molecules [10], since the association efficiency strongly depends on the phase space overlap of the two species [11]. The rapid losses of associated molecules observed for bosonic systems could be largely suppressed by the presence of the three-dimensional optical lattice [12], if most of the sites are occupied with only one atom per species. Both these research avenues require the dynamic control of interspecies interactions, along with a well established collisional model.

We demonstrate optical tuning of the scattering length in a Bose-Einstein condensate as predicted by Fedichev et al. [Phys. Rev. Lett. 77, 2913 (1996)]. In our experiment, atoms in a 87Rb condensate are exposed to laser light which is tuned close to the transition frequency to an excited molecular state. By controlling the power and detuning of the laser beam we can change the atomic scattering length over a wide range. In view of laser-driven atomic losses, we use Bragg spectroscopy as a fast method to measure the scattering length of the atoms.

We investigate experimentally the entropy transfer between two distinguishable atomic quantum gases at ultralow temperatures. Exploiting a species-selective trapping potential, we are able to control the entropy of one target gas in presence of a second auxiliary gas. With this method, we drive the target gas into the degenerate regime in conditions of controlled temperature by transferring entropy to the auxiliary gas. We envision that our method could be useful both to achieve the low entropies required to realize new quantum phases and to measure the temperature of atoms in deep optical lattices. We verified the thermalization of the two species in a 1D lattice.PACS numbers: 03.75. Hh, 67.85.Pq, 05.30.Jp In recent years an intense research of quantum phases common to condensed matter systems and atomic quantum gases has made remarkable progresses [1]. Some of these phases can only be reached provided that the temperature is suitably low. However, in strongly correlated quantum systems, even the temperature measurement can be a challenging task. If so, to ascertain whether a given quantum phase is accessible, it is convenient to focus on the critical value of entropy, rather than temperature. The advantage is especially clear when the strongly correlated regime is reached by sufficiently slow, entropy-preserving, transformations of the trapping potential, as it is often the case for atoms in deep optical lattices [2]. For these reasons, it is important to determine and grasp control of the entropy of degenerate quantum gases [3,4,5]. In this work, we demonstrate the reversible and controlled transfer of entropy between the two ultracold, harmonically trapped Bose gases, which is based on the use of a species-selective dipole potential (SSDP), i.e., an optical potential experienced exclusively by one species (Fig. 1) [6,7]. In particular, we drive the target gas across the threshold for Bose-Einstein condensation, by a reversible transfer of entropy to the auxiliary gas.The main idea can be understood from textbook thermodynamics. Let us consider two distinguishable gases filling an isolated box, exchanging neither particles nor energy with the outside, and imagine that only one gas (target) is compressed, e.g. through a piston permeable to the second gas (auxiliary). The temperature will increase and, in thermal equilibrium, heat, hence entropy, will transfer from the target to the auxiliary uncompressed gas. In the limit of the auxiliary gas containing a large number of particles, it stands as a thermal bath. In a more formal way, for an ideal gas of N particles, the entropy S is proportional to N log(Σ/N ), where the number of accessible single-particle states Σ increases with the energy density of states and with the average energy, i.e., the temperature. In an adiabatic compression of one single gas, the reduction of the energy density of states is compensated by a temperature raising such that Σ, hence S, remains constant. If we add the uncompressed auxiliary gas in thermal contact, the temperature rai...

Force spectroscopy has become an indispensable tool to unravel the structural and mechanochemical properties of biomolecules. Here we extend the force spectroscopy toolbox with an acoustic manipulation device that can exert forces from subpiconewtons to hundreds of piconewtons on thousands of biomolecules in parallel, with submillisecond response time and inherent stability. This method can be readily integrated in lab-on-a-chip devices, allowing for cost-effective and massively parallel applications.