[Quantum Column] Understanding Quantum Computer Hardware & Qubits short introduce

簡志勳
4 min readJul 9, 2023

--

At the end of 2022, a groundbreaking (for non-researchers) technological advancement shook various industries in the technology sector. The emergence of chatGPT language training model AI led to an AI frenzy, triggering an AI war across browsers, GPUs, and various technology industries.

As early as 1943, scientists proposed the concept of artificial neurons (neural networks), followed by various types of neural networks such as RNN and CNN. The first-generation GPT language training model was introduced in 2018 and underwent subsequent improvements. In November 2022, OpenAI presented the first commercialized GPT3.5 model.

Within 40 years after the introduction of the concept of neurons, physicist Richard Feynman proposed the idea of quantum computers, utilizing computers to simulate physical phenomena. In today’s Taiwan network, apart from laboratory websites briefly mentioning what quantum computers are, the instruments used for measurement, and how to operate them, the remaining 80% of information leans towards quantum information and quantum algorithm-related content. This series of quantum columns will delve into what quantum computers are, how they are constructed, the necessary equipment, and provide introductions on measurement and implementation. It will also share the latest research from major companies (Google, IBM, Intel, etc.) and laboratories.

As someone engaged in research, my focus is on superconducting quantum bits, as superconductors serve as the material. The main knowledge shared will cover the equipment required, the development and various types of superconducting quantum bits, chip design, and simulation. So let’s start by discussing the types of quantum bits and the equipment required to manipulate superconducting quantum bits.

The basic unit of a quantum chip is a quantum bit (qubit), and currently, it is mainly realized through the following approaches:

Quantum dot (or spin qubit):

Single-electron control is achieved using silicon (Si) and gallium arsenide (GaAs). Notably, Intel is researching in this direction.

Trapped ion:

Atoms are controlled in specific locations in space using magnetic and electric fields, and further controlled through lasers. IonQ is a startup company focusing on this aspect.

Superconducting qubit:

Implemented using Josephson junctions, which can be seen as nonlinear inductors. When combined with capacitors, they form a harmonic oscillator system and can be viewed as artificial atoms. They are controlled using microwave systems. Google, IBM, Amazon, and other major companies are researching in this area.

Diamond center (NV center):

Nitrogen atoms replace carbon atoms in the lattice structure of diamond. As nitrogen (from Group 5A) has one extra electron compared to carbon (from Group 4A), this electron with spin is tightly trapped within the vacancy and can be used as a qubit.

Topological qubit:

Using semiconductor technology, topological states are realized to form quantum bits. Computations are performed by exchanging quasi-particles in either a clockwise or counterclockwise manner. Microsoft is a well-known company in this field.

To provide a more fundamental introduction to superconducting quantum computers, we first need a device called a dilution fridge. The dilution fridge is used to create an environment as low as 10 millikelvin (equivalent to -273.1 degrees Celsius) by cooling down the system. Superconducting qubits require extremely low temperatures to minimize thermal population and maintain them in their ground state. Therefore, an ultra-low temperature environment is necessary.

The current mainstream superconducting qubits, such as Transmon/Xmon, operate at frequencies ranging from 4 to 8 gigahertz (GHz). According to the energy level formula,

we can deduce that for a 4 GHz energy level, the environment temperature needs to be below 200 millikelvin. In recent years, new architectures like Fluxonium have even lower energy levels, reaching around 500 MHz or lower. For these cases, the environment temperature needs to be below 20 millikelvin. Thus, a dilution fridge is an essential device for superconducting qubits.

Bluefors and Oxford are leading companies in the field of dilution fridges, with prices starting at over 300 thousand USD

Bludfors XLD dilution refrigerator

ref:https://www.hpcwire.com/2020/08/19/intel-connects-the-quantum-dots-in-accelerating-quantum-computing-effort/

--

--