A new quantum RAM system is much more hardware efficient than previous models because it reads and writes data using audible electromagnetic pulses and superconducting resonators.
A computer's random access memory, also known as RAM, acts as a short-term memory bank from which data can be quickly recalled. RAM is used by apps on your phone or computer so you can switch between tasks instantly. In theory, comparable quantum RAM components could speed up the execution of a quantum algorithm or increase the density of information that can be stored in a quantum processor, according to researchers working on quantum computers. James O'Sullivan of the London Center for Nanotechnology has demonstrated a hardware-efficient method for storing and retrieving quantum information in atomic spins using chirped microwave pulses, an important step towards realizing quantum RAM.
Quantum memory device experimental demonstrations, like quantum computing, are still in their infancy. Superconducting metal-based circuits are used in one of the most popular chip-based quantum computing platforms. The superconducting qubits used in the central processing unit of this system transmit and receive data via microwave photons. However, there is currently no quantum memory technology that can stably store these photons for long periods of time. Fortunately, scientists have some suggestions.
The use of spins of impurity atoms in the chip of the superconducting circuit is one of these concepts. One of the fundamental quantum properties of an atom is spin. It is aligned with or against an applied magnetic field, such as an internal compass needle.
Storing Quantum Information
These two alignments, which can be used to store quantum information, can be compared to the 0s and 1s of a classical bit. The spins of the impurity atoms can act as a "multimodal" memory device that simultaneously stores information from many photons if there are many of them on the chip.
The information storage times of atomic spins can be much longer compared to superconducting qubits. For example, studies have shown that silicon devices containing bismuth atoms can retain quantum information for longer than one second.
One may wonder why spin qubits are not used instead of superconducting qubits. There are indeed research teams developing atom-based quantum computers, but manipulating and measuring atomic spins presents special challenges.
Superconducting qubits and atomic spins are hybridly combined, but in this case using microwave photons to transport data between the two systems has proven difficult. Researchers have already demonstrated that an atomic spin ensemble can absorb and receive information from microwave photons, but these experiments required the use of special superconducting circuitry or strong magnetic field gradients, both of which complicate the hardware for a quantum memory.
O'Sullivan and colleagues present a sophisticated, hardware-efficient approach to microwave photon information storage and retrieval. A superconducting circuit resonator mounted on a silicon chip coated with bismuth atoms is the team's invention.
The team inserted weak microwave excitations containing about 1000 photons into the resonator; these photons were absorbed by the spins of the bismuth atoms. They then used frequency-ramped electromagnetic microwave pulses to strike the resonator, creating a chirping effect. In this way, a special "phase" identifier, which captures the relative beacon positions of nearby spins, was imprinted on the spins by the quantum information contained in the photons. The team then regained this information by applying an identical pulse to the spin collection, which they discovered reversed this imprinted phase and transmitted the photons back to the superconducting circuit.
Data Storage with Microwave Pulses
O'Sullivan and colleagues show that memory systems can simultaneously store many packets of photonic data as four weak microwave pulses. More importantly, they show that data can be read back in any order, proving that their tool truly functions as RAM.
The team claims to have achieved an efficiency of 3% in this first test, suggesting that the memory lost most of the information. As a result, their technology is still far from providing the reliable storage and retrieval required for a future quantum computer. According to the review of possible causes, the reason for this low efficiency is not the transfer process but rather the improvement potential of the device.
The group believes they can greatly increase the efficiency of the device by increasing the spins.
Besides storing information, quantum RAM components can also help increase the qubit density in a quantum processor. IBM introduced Project Goldeneye, a massive dilution refrigerator, in September. This super-cold beast, which will house IBM's next-generation superconducting quantum computer, has a larger volume than three standard freezers. Given that currently available superconducting quantum computers have a density of less than 100 qubits per square millimeter, it's clear why IBM needs such a large refrigerator. This size problem may one day be solved by the spin-based quantum memory device developed by O'Sullivan and colleagues, which can theoretically store several qubit states in the space currently occupied by only one.
📩 23/01/2023 16:26