1. Introduction
Optical instrumentation is an important element in modern communication systems. Fibre optic systems, in particular, have revolutionized data transmission in terms of speed and bandwidth, and such devices are ideally-suited for point-to-point and long-haul data transmission. For applications with shorter scales and variability in user endpoints, an interesting subset of optical communication has emerged in the form of free-space optical (FSO) communication. FSO systems have the same high-speed benefits as their fibre optic counterparts [
1
W. S. Rabinovich, R. Mahon, H. R. Burris, G. C. Gilbreath, P. G. Goetz, C. I. Moore, M. F. Stell, M. J. Vilcheck, J. L. Witkowsky, L. Swingen, M. R. Suite, E. Oh, and J. Koplow, “Free-space optical communications link at 1550 nm using multiple-quantum-well modulating retroreflectors in a marine environment,” Opt. Eng.
44(5), 056001 (2005). [CrossRef]
] and have additional benefits in terms of signal distribution, as FSO signals can be broadcasted in the communication environment for optical wireless networking [
2
S. Junique, D. Agren, Q. Wang, S. Almqvist, B. Noharet, and J. Y. Andersson, “A modulating retroreflector for free-space optical communication,” IEEE Photon. Technol. Lett.
18(1), 85–87 (2006). [CrossRef]
–
5P. B. Chu, N. R. Lo, E. C. Berg, and K. S. J. Pister, “Optical communication using micro corner cube reflectors,” in Proceedings of IEEE Micro Electro Mechanical Systems Workshop (IEEE, Nagoya, 1997), pp. 350–355.
]. Challenges associated with this optical signal distribution are addressed in this work.
The primary requirement for FSO communication is the capability of the system to realize and control multi-directional communication. This can be accomplished with active downlink communication [
6
W. Mao and J. M. Kahn, “Free-space heterochronous imaging reception of multiple optical signals,” IEEE Trans. Commun.
52(2), 269–279 (2004). [CrossRef]
] using modulated laser transceivers broadcasting signals to distributed remote receivers. This requirement can also be met with passive uplink communication [
6
W. Mao and J. M. Kahn, “Free-space heterochronous imaging reception of multiple optical signals,” IEEE Trans. Commun.
52(2), 269–279 (2004). [CrossRef]
] using distributed receivers that redirect the incoming light back to its source and provide optical encoding (via multi-quantum well (MQW) absorbers [
1
W. S. Rabinovich, R. Mahon, H. R. Burris, G. C. Gilbreath, P. G. Goetz, C. I. Moore, M. F. Stell, M. J. Vilcheck, J. L. Witkowsky, L. Swingen, M. R. Suite, E. Oh, and J. Koplow, “Free-space optical communications link at 1550 nm using multiple-quantum-well modulating retroreflectors in a marine environment,” Opt. Eng.
44(5), 056001 (2005). [CrossRef]
], mechanical beam deflection [
6
W. Mao and J. M. Kahn, “Free-space heterochronous imaging reception of multiple optical signals,” IEEE Trans. Commun.
52(2), 269–279 (2004). [CrossRef]
], or liquid crystal (LC) polarization modulation [
7
D. C. O’Brien, W. W. Yuan, J. J. Liu, G. E. Faulkner, S. J. Elston, S. Collins, and L. A. Parry-Jones, “Optical wireless communications for micromachines,” Proc. SPIE
6304, 63041A (2006). [CrossRef]
]). For applications requiring signal broadcasting and even motion of nodes, the passive configuration becomes advantageous as it is inherently distributive in nature, characterized by low power consumption, and more suited to bi-directional communications (compared to the active configuration which has independent nodes and a reliance on the point-to-point laser transceiver/receiver alignment).
The standard structure employed in passive FSO communications is the three-mirrored corner-cube retroreflector (CCR). Such a structure redirects incoming light back to its source as each mirror reverses its corresponding light ray component [
8G. Chartier, Introduction to Optics (Springer, New York, 2005).
,
9
M. Scholl, “Complex reflectivity of a corner cube retroreflector,” Proc. SPIE
2268, 422–430 (1994). [CrossRef]
]. Individual CCRs can accomplish this redirection over a π/2 Steradians solid angle (with the azimuthal angle
ϕ and polar angle
θ both ranging from 0° to 90°). For orientations beyond these limits, retroreflection will fail. Attempts at improving the directionality of passive communication systems have largely made use of multi-CCR configurations, with Zhou
et al. having demonstrated a quadruplet assembly with 2π Steradians solid angle coverage [
10L. Zhou, K. S. J. Pister, and J. M. Kahn, “Assembled corner-cube retroflector quadruplet,” in Proceedings of IEEE Micro Electro Mechanical Systems (IEEE, Las Vegas, 2002), pp. 556–559.
] and Collier
et al. having demonstrated an octuplet assembly with 4π Steradians solid angle coverage [
11
C. M. Collier, X. Jin, J. F. Holzman, and J. Cheng, “Omni-directional characteristics of composite retroreflectors,” J. Opt. A, Pure Appl. Opt.
11(8), 085404 (2009). [CrossRef]
].
In this work, a new approach to bi-directional FSO link reception is introduced. The presented device, defined here as a retro-detection photocell, is an integrated package that meets the fundamental requirements for bi-directional FSO communication (retroreflection, detection, and modulation), while allowing for tracking of the optical signal alignment. An active corner-cube architecture is used to retroreflect and simultaneously sample the incident light intensity with a architecture that provides an enhanced level of directivity–being concentrated within a π/2 Steradians solid angle. Differential selection of the three constituent photodiodes (PDs) can then make use of this heightened directivity through real-time triangulation and optimization of the optical signal alignment. This established and optimized optical link will be particularly advantageous for environments with multiple communication nodes and spurious reflections. Operation of the FSO retro-detection photocell is demonstrated here for both active laser and passive Pi-cell LC bi-directional communication.
2. Device structure
A schematic of the investigated corner-cube-based retro-detection photocell is presented in
Fig. 1
. The integrated device is fabricated out of three mutually-orthogonal silicon PDs with 9.7 × 9.7 mm
2 electrically-isolated active regions. Attention is paid to the design and the alignment of these PDs to optimize its reflectance. A vertical (i.e. multi-layered) electrode-semiconductor-electrode structure is chosen for the PDs to maintain a uniform reflectivity with minimal light scattering from the surface. This is in contrast to metal-semiconductor-metal surface layouts, for which there would be diffraction, scattering, and losses on the reflected beam. Furthermore, the structure is cured during the alignment and calibration assembly process to ensure that the required level of retroreflection (over a distance of approximately 5 m) is achieved by the orthogonal surfaces. A high-speed Pi-cell LC optical modulator [
11
C. M. Collier, X. Jin, J. F. Holzman, and J. Cheng, “Omni-directional characteristics of composite retroreflectors,” J. Opt. A, Pure Appl. Opt.
11(8), 085404 (2009). [CrossRef]
] is then optically bonded to the retro-detection photocell entrance interface to modulate the incident and retroreflected optical beams.
Fig. 1 SolidWorks schematic of the retro-detection photocell. The structure consists of three mutually-orthogonal silicon PDs arranged in an interior corner. A Pi-cell LC modulator is mounted at the entrance interface for optical modulation. Differential combinations of PD 1, PD 2, and PD 3 are used to triangulate the incident laser’s ϕ and θ incident angle.
Operation of the retro-detection photocell is based on its constituent photocurrents. The three photocurrents i1
(t), i2
(t), and i3
(t), correspond to PD 1, PD 2, and PD 3, respectively, and differential combinations of these photocurrents can be used to give the required directionality for triangulation. The differential photocurrents are defined here as i1-2
(t) = i1
(t) - i2
(t), i1-3
(t) = i1
(t) - i3
(t), and i2-3
(t) = i2
(t) - i3
(t). These differential signals are related to the azimuthal ϕ and polar θ orientations and can be used to optimize the beam alignment as follows. Typically, ϕ is optimized first during a 0° to 90° rotation. Over this range, the PD 1 photocurrent diminishes from its full-illumination maximum down to its negligible-illumination minimum, while the PD 3 photocurrent rises from its negligible-illumination minimum up to its full- illumination maximum. A careful tracking of the differential photocurrent between PD 1 and PD 3 during this rotation will show that the ϕ = 45° midpoint gives a balanced differential photocurrent with i1-3
(t) = 0. This condition indicates an optimal alignment for ϕ. In the next control phase, the same optimization process is employed to balance θ while monitoring both i1-2
(t) and i2-3
(t). When both these differential photocurrents are zero, the system is completely balanced with ϕ = 45° and θ = cos−1(1/√3) ≈54.7°. The retro-detection photocell is now oriented with its (x,y,z) = (1,1,1) coordinate aligned directly toward the illumination source.
3. Experimental and theoretical results
To investigate this retro-detection system, a laser-based FSO setup is built. The setup covers a 10 m propagation range, and testing is carried out with a 1 mW, 650 nm laser diode as the illumination source. The photocell is rotated in a gyroscope with independent
ϕ and
θ orientations. The laser beam is expanded to a sufficiently large and collimated spot size to uniformly illuminate the retro-detection photocell, and a digital data acquisition system is employed to monitor the differential photocurrents and apply the required orientation optimization and control. The retro-detection photocell is tested first with 1 kHz modulation on the laser diode source and an on-state (open) condition for the Pi-cell LC modulator [
12
Y. Sun, H. Ma, Z. Zhang, and G. Fu, “Rapid response mechanism of pi cell,” Appl. Phys. Lett.
92(11), 111117 (2008). [CrossRef]
]. The three photocell photocurrents are tracked as a function of time with the laser source sampled at three different locations in the facility. Each of these locations corresponds to a propagation distance of 9 m, with orientations of
ϕ = 20°, 30°, and 45° and a polar angle of
θ ≈54.7°. The results are shown in
Figs. 2(a)
–
2(c), respectively, with the resulting differential photocurrents
i1-2
(
t),
i1-3
(
t), and
i2-3
(
t) shown in the lower half of each figure. It is apparent that the largest differential signal levels are exhibited in
Fig. 2(a). This is the result of the significant misalignment for this orientation, as PD 1 has a disproportionately large illumination power. (This is qualitatively evident from the photocell alignment, whose orientation, as viewed from the laser, is shown in the figure inset.) The alignment balancing is improved for the results of
Fig. 2(b), as the orientation exhibits only a partial misalignment and reduced differential photocurrents. The system becomes completely balanced for the results of
Fig. 2(c), as this alignment has negligible differential photocurrent levels. It is in this final balanced orientation that the retro-detection photocell is aligned for optimal transmission and reception along the (
x,
y,
z) = (1,1,1) direction.
Fig. 2 Experimental results for the retro-detection photocell with three orientations. Time-varying photocurrents i1
(t), i2
(t), and i3
(t) and differential photocurrents, i1-2
(t), i1-3
(t), and i2-3
(t), are shown for a polar angle θ ≈54.7° and (a) ϕ = 20°, (b) ϕ = 30°, and (c) ϕ = 45°. The insets show the retro-detection photocell at the respective orientations viewed from the laser source.
The operation of the retro-detection photocell is fundamentally based on the measurement of asymmetric illumination levels for the constituent PDs. This point was demonstrated in the previous experimental results. To quantify the relationship between the illumination asymmetry and differential photocurrents, however, it is necessary to include the effects of internal reflections, as the light incident on any particular photodiode is a combination of directly-incident light and light reflecting off neighbouring PDs. This response is modeled with a ray-tracing approach, whereby the structure is illuminated from a uniform intensity grid at various
ϕ and
θ directions. The intensity of each ray is tracked through the multiple absorption and reflection processes of each silicon PD. The resulting photocurrents are extracted as a function of the model’s incident
ϕ and
θ directions, and the summed differential photocurrents magnitudes |
i1-2
(
t)| + |
i1-3
(
t)| + |
i2-3
(
t)| is shown for this simulation in
Fig. 3(a)
with angular ranges defined by 0 ≤
ϕ ≤ 45° and 0 ≤
θ ≤ 90°. The fundamental goal of optimizing the beam alignment can be viewed on this curve by traveling from any initial
ϕ and
θ location to the absolute minimum at
ϕ = 45° and
θ = cos
−1(1/√3) ≈54.7°. This procedure is carried out through the independent
ϕ and
θ differential photocurrent minimization processes described above. A representative optimization procedure is shown by the line ABC in
Fig. 3(a) as
ϕ is optimized from 10° to 45° and
θ is then optimized from 5° to 54.7°. Note that the internal reflections lessen the steepness of figure surface, but they do not prevent the desired optimization process. This is due to the fact that the optimal (
x,y,z) = (1,1,1) alignment is characterized by balanced incident light levels
and balanced internal reflection light levels. The azimuthal and polar angle optimization procedures can be carried out to establish the optimal link for optical signal detection, displayed as the summed photocurrent in
Fig. 3(b), and optical signal retroreflection, displayed as the retroreflected power in
Fig. 3(c). The optimization procedure carried out through the differential photocurrents in
Fig. 3(a) brings the system to a point of optimized alignment for both optical detection and retroreflection.
Fig. 3 Ray-tracing model results for the retro-detection photocell are shown. The (a) summed differential photocurrent, (b) summed photocurrent, and (c) retroreflected power are shown as a function of the azimuthal angle ϕ and polar angle θ. A representative optimization process is shown by the black trace along the ABC curve in (a).
4. Applications
Given the above optimization procedure, an FSO communication link is demonstrated here for this control structure. The laser diode is directed at the retro-detection photocell, and the
ϕ and
θ angles are optimized. The photocell (
x,
y,
z) = (1,1,1) orientation becomes aligned toward the laser source, and the system becomes optimized for bi-directional communication. Two communication modes are then tested. The active downlink mode is investigated first with a 2 kHz modulated laser. The Pi-cell LC modulator remains in its on-state (open), and the summed photocurrent,
i1
(
t) +
i2
(
t) +
i3
(
t), is shown in
Fig. 4(a)
as a function of time. The modulated characteristics of the laser transmitter are seen to be effectively detected by the photocell, as the incident light is sampled by each of the three balanced PDs and subsequently summed. As a second FSO investigation, a passive uplink mode is studied next using continuous laser illumination and the Pi-cell LC modulator. The LC device is modulated at 100 Hz, and a photodetector at the laser transmitter is used to record the returning (retroreflected) light. The signal from this photodetector is shown in
Fig. 4(b) as a function of time. Note that the Pi-cell LC modulator effectively maps the remotely encoded information onto the retroreflected beam. The signal level and modulation depth are seen to appreciable, as the system demonstrates a signal-to-noise ratio (SNR) beyond 40 dB. This ability is due in large part to the previous beam alignment optimization procedure.
Fig. 4 Photocurrents for the (a) active downlink mode and (b) passive uplink mode are shown as a function of time for the retro-detection photocell with modulated laser illumination and continuous laser illumination with a Pi-cell LC modulator (not shown), respectively. The photocell is in the optimal orientation with ϕ = 45° and θ ≈54.7° (1 m propagation). The figure inset shows the retro-detection structure at this optimal orientation as viewed by the source.
5. Conclusion
In summary, theoretical and experimental results were presented for a new corner-cube-based retro-detection photocell. It was shown that differential photocurrent levels can be used to probe the alignment states of incident beams. These differential photocurrents can then be used in an actively-controlled optimization/triangulation procedure. Experimental results for this optimization were presented along with a model, and the aligned system was demonstrated with both active downlink and passive uplink communication modes. The presented retro-detection elements will be well-suited to integrated directional sensing applications when used in isolation or in future arrayed/tiled implementations requiring enhanced optical signal levels.