We present an unambiguous characterization of the rotation group SO(3) biconnectedness topology using two-qubit maximally entangled states. We show how to generate cyclic evolutions of these states, which are in one-to-one correspondence to closed paths in SO(3). The difference between the well known two classes of such paths translates into the gain of a global phase of π for one class and no phase change for the other. We propose a simple quantum optics interference experiment to demonstrate this topological phase shift. [2]. In addition, it is known that any quantum computing protocol involving many qubits can be implemented by concatenation of one or two-qubit gates [2]. This is why much attention has been paid to two-qubit systems, their properties and characterizations.In the present work, we intend to show that two-qubit MES can also be used to display, even experimentally, the well known double-connectedness of the SO(3) rotation group. Since the latter is often evoked to explain the minus sign multiplying a spin 1/2 state after a 2π rotation, we shall first argue that this is to some respect an ambiguous statement. Indeed, take one qubit state |Ψ(t = 0) = α |0 + β |1 subject to the Hamiltonian H =h ω 2σ z (an effective magnetic field along the z axis). At time t, the state reads |Ψ(t) = e −iωt 2 α |0 + e iωt 2 β |1 . Let us turn to the Bloch sphere representation of |Ψ(t) . The three coordinates are given by the expectation values of the Pauli matrices. As is well known, the expectation value of the spin precesses around the effective magnetic field. At t = 2π/ω it has experienced a full 2π rotation, and a π phase for the wave function. The π phase has been clearly demonstrated on several systems, starting with beautiful experiment on spin 1/2 neutrons [3]. On general grounds, the change of global phase γ under a cycle decomposes into its dynamical part, γ d (as derived from the time dependent Schrödinger equation) and its geometrical part, the Berry phase γ g , whose origin relies on the Hilbert space peculiar geometry. The latter phase, γ g , equals half of the area bounded by the representing part of the Bloch sphere [4]. The dynamical phase γ d is simply related to the time average of the Hamiltonian.In the above case of one qubit precessing in a magnetic field, we have γ d = −π cos θ and γ g = −π(1 − cos θ), so the expected global phase γ = γ d + γ g = −π for an initial state with |α| = cos θ/2 and |β| = sin θ/2. Let us remark here that this π phase is present even for θ = 0, where it is of pure dynamical nature (with no precession and therefore no rotation at all). We may therefore question the geometrical interpretation of the π phase. It would be more correct to state that this phase has a geometrical and a dynamical component, being purely geometrical only when θ = π/2, when the initial state is orthogonal to the magnetic field. However, even in that case, we find problems in relating the phase to the double connectedness of SO(3). Indeed, the latter property relates paths on the SO(3) manifol...