January 2015
Spotlight Summary by Andrea Reale e Alessio Gagliardi
Low-power communication with a photonic heat pump
A LED can work as a thermodynamic heat pump: under certain conditions, it can use externally supplied electrical work to pump heat from its own semiconductor lattice to the light it emits. This process manifests itself in the following fact: the energy of the photon of the emitted light, ħω , is larger than the electrical work used to generate it, eV, where e is the charge of an electron and V is the voltage applied to the LED junction. A heat pump (by Carnot’s theorem) can be more efficient if it works between two reservoirs whose temperature difference is small; in the case of the LED, it is necessary to keep the bias of the junction low: in this case the temperature of the emitted light (that can be derived from the statistics of occupation of the electromagnetic modes) is lower. It is then possible for an LED to generate light with a wall-plug efficiency greater than one.
Huang and co-authors demonstrate in this Optics Express article that the light generated in this way can be used for communicating information. They set up a PSK/OFDM (phase-shift keying/orthogonal frequency division multiplexing) scheme and obtain a low energy per bit figure of merit and a bit-error rate of 3x10-3. The transmission-reception apparatus is tested up to 90 kb/s only due to the limited performance (50kHz bandwidth) of the electronics used in the receiver. An upper limit of the heat-pumped LED communication system is related to the switching time required for LED modulation. A (reasonable) 20 nanosecond switching time of the IR LED sets 50Mb/s as an upper limit for the maximum bit rate. This limitation is linked to the spontaneous radiative recombination time in the LED, suggesting that this ultra-low power scheme will not intrinsically reach the Gb/s range in transmission speed.
A point raised by the authors, that certainly stimulates one’s intellectual curiosity, is that the light field generated from a heat pump contains the entropy of the reservoir from which the heat comes (and is therefore disordered), while information corresponds to order; the possibility of transmitting information through the use of heat-generated light should therefore be limited. The formal expression of this idea is the Landauer limit, first proposed by Von Neumann, that states that a computation should cost at least kT ln(2) Joule in energy per information bit, where k is the Boltzmann constant and T the temperature of the system.
However, the Landauer limit is related, as showed by Landauer himself, to the bit erasure from the memory. The erasure process is the only one which requires in the Landauer argument (in the limit of an ideal computation) an unavoidable dissipation. All the other processes: computation, writing on the memory and communication, can be made arbitrarily small in terms of energy consumption. It has even been shown that computation can be performed in a classical way without making any irreversible process, hence without dissipating energy into heat. Obviously, this concept is more theoretical than practical, as a reversible computer would be extremely sensitive to any external perturbation; of course some feedback can be added to correct possible errors, but at the cost of introducing new dissipative elements and thus invalidating the assumption of reversibility. A recent generalization of the Landauer principle says that the kT ln(2) value should be split between "measurement" of the information and erasure of the memory of the feedback system. In the work of Huang and co-authors, the communication can be considered as part of the "measurement" of the state of some signal source.
The current status of our knowledge makes it in general clear that the information process, i.e. the bit communication, can cost less than kT ln(2). The interest of the paper consists in providing a physical system with which, in a future perspective, tests of the Landauer limit could be carried out; the recent theoretical developments in the field just outlined make this line of enquiry all the more attractive - one should still though keep in mind that the present realization of the device does not yet come close to the violation of the Landauer limit.
The implementation of a communication channel with “heat-pump” LEDs can represent the base for future developments. On one side, there is the potential interest as an experimental “platform” to test fundamental questions of thermodynamics and information science; from the technological point of view, this use of LED allows the exploration of extremely low power, thresholdless communications. Possible future developments on low power transmission schemes of this kind might come from thresholdless stimulated emission, as in the case of polariton lasers.
You must log in to add comments.
Huang and co-authors demonstrate in this Optics Express article that the light generated in this way can be used for communicating information. They set up a PSK/OFDM (phase-shift keying/orthogonal frequency division multiplexing) scheme and obtain a low energy per bit figure of merit and a bit-error rate of 3x10-3. The transmission-reception apparatus is tested up to 90 kb/s only due to the limited performance (50kHz bandwidth) of the electronics used in the receiver. An upper limit of the heat-pumped LED communication system is related to the switching time required for LED modulation. A (reasonable) 20 nanosecond switching time of the IR LED sets 50Mb/s as an upper limit for the maximum bit rate. This limitation is linked to the spontaneous radiative recombination time in the LED, suggesting that this ultra-low power scheme will not intrinsically reach the Gb/s range in transmission speed.
A point raised by the authors, that certainly stimulates one’s intellectual curiosity, is that the light field generated from a heat pump contains the entropy of the reservoir from which the heat comes (and is therefore disordered), while information corresponds to order; the possibility of transmitting information through the use of heat-generated light should therefore be limited. The formal expression of this idea is the Landauer limit, first proposed by Von Neumann, that states that a computation should cost at least kT ln(2) Joule in energy per information bit, where k is the Boltzmann constant and T the temperature of the system.
However, the Landauer limit is related, as showed by Landauer himself, to the bit erasure from the memory. The erasure process is the only one which requires in the Landauer argument (in the limit of an ideal computation) an unavoidable dissipation. All the other processes: computation, writing on the memory and communication, can be made arbitrarily small in terms of energy consumption. It has even been shown that computation can be performed in a classical way without making any irreversible process, hence without dissipating energy into heat. Obviously, this concept is more theoretical than practical, as a reversible computer would be extremely sensitive to any external perturbation; of course some feedback can be added to correct possible errors, but at the cost of introducing new dissipative elements and thus invalidating the assumption of reversibility. A recent generalization of the Landauer principle says that the kT ln(2) value should be split between "measurement" of the information and erasure of the memory of the feedback system. In the work of Huang and co-authors, the communication can be considered as part of the "measurement" of the state of some signal source.
The current status of our knowledge makes it in general clear that the information process, i.e. the bit communication, can cost less than kT ln(2). The interest of the paper consists in providing a physical system with which, in a future perspective, tests of the Landauer limit could be carried out; the recent theoretical developments in the field just outlined make this line of enquiry all the more attractive - one should still though keep in mind that the present realization of the device does not yet come close to the violation of the Landauer limit.
The implementation of a communication channel with “heat-pump” LEDs can represent the base for future developments. On one side, there is the potential interest as an experimental “platform” to test fundamental questions of thermodynamics and information science; from the technological point of view, this use of LED allows the exploration of extremely low power, thresholdless communications. Possible future developments on low power transmission schemes of this kind might come from thresholdless stimulated emission, as in the case of polariton lasers.
Add Comment
You must log in to add comments.
Article Information
Low-power communication with a photonic heat pump
Duanni Huang, Parthiban Santhanam, and Rajeev J. Ram
Opt. Express 22(S7) A1650-A1658 (2014) View: Abstract | HTML | PDF