growing continuously, and has become essential in several contexts. For this reason, it is necessary and urgent to find better ways to exploit the electromagnetic (EM) spectrum. This is, in fact, the target of new, sophisticated modulation techniques, like OFDM, TDM, etc., which aim at increasing the spectral efficiency. It is also the goal of MIMO systems, which aim, through more and more complex antenna arrays, at better exploiting spatial diversity [1]. An alternative way of exploiting the spatial features of EM waves has been proposed and tested: a channel multiplexing scheme based on superimposing waves that carry orbital angular momentum (OAM) [2] and [3].
From basic physics, it is well known that an EM wave can carry both linear and angular momentum [4], [5], [6] and [7]. The linear momentum, proportional to the wave power density (Poynting vector), has always been the basic ingredient of radio engineering. The angular momentum consists of two contributions, spin angular momentum (SAM) and orbital angular momentum (OAM). While SAM, related to wave polarization, is widely used in present systems to double the number of channels over a given frequency range, OAM has been, until now, far less exploited. Being associated with helicity of the wave phase front, it is a degree of freedom of an EM wave completely independent of frequency and polarization. Hence, it can be exploited to increase the overall system capacity, because EM fields at the same frequency but with different OAM values propagate, in free space, without mutual interference, and can be discriminated by suitable receivers. In fact, the dramatic potential of this approach at optical wavelengths has been successfully demonstrated [8] and [9]. Whether it is practical also for radio communications, is still an open question. The present paper reports on a new experiment, within this framework.
The basic reason why OAM-based channel multiplexing has drawn interest, is because waves with different OAM integer values can form an orthogonal basis. This, in principle, allows channel discrimination directly at the physical layer, without post processing of the received signals. However, it must be underlined that, to fully exploit the orthogonality between different OAM waves, one has to receive the whole wave front, in order to correctly recognize the transmitted OAM value [10] and [11]. This requirement is reasonable at optical wavelengths, as proved in [8], but very difficult to cope with at microwave frequencies. It can be by-passed, as proved recently in [3], but it is not yet clear to what extent. For this reason, we focused, recently, on the possibility of discriminating different OAM values by measuring phase gradients over a limited part of the wave front. Our preliminary results [12] indicate that this approach is comparable in performance to conventional MIMO, in agreement with what had been predicted in a broad-scope previous paper [13].
Here we focus on another, completely different approach to the OAM-channel discrimination problem. It relies on the dramatic simplification of the field phase distribution in space that is obtained if one radiates, instead of one wave characterized by a single integer OAM value, two superimposed waves with equal amplitudes and opposite values of their OAMs.
The two counter-rotating helices, both propagating along View the MathML source in a cylindrical reference system (r, ϕ, z), generate a “standing wave” in the azimuthal direction. Different OAM values can then be discriminated simply by counting the number of peaks (and/or nulls) around a closed path concatenated with the propagation axis z – e.g., a circle lying on a constant-z plane [11].
The theoretical equivalence, in terms of capacity, between this approach and the use of individual OAM modes, is quite evident, and was already stated in [13]. Still, we believe that this approach is worth an experiment – to the best of our knowledge, unprecedented – since there are in it some potential advantages, to be assessed in practice. First of all, it liberates the receiver from the burden of accurate phase measurements. Another benefit can be grasped if we think of a circular antenna array as the radiating element. In the single-OAM-mode approach, signals feeding the individual antennas must obey strict phase requirements; hence, the bandwidth of the array is limited by that of the phase shifters encompassed in its feeding network. In the “plus and minus ℓ” configuration, one phase shift of π radians is enough; as well known, this can be a geometry-based, built-in property of suitable microwave junctions, what can increase significantly the array bandwidth. A third potential benefit may stem from the fact that the field of a standing wave vanishes identically over |ℓ| longitudinal planes (i.e., planes passing through the propagation axis). If the antennas are designed so that one of these planes coincides with a discontinuity in the medium