The contents of this metal cylinder could either revolutionize technology or be completely useless— it all depends on whether we can harness the strange physics of matter at very, very small scales. To have even a chance of doing so, we have to control the environment precisely: the thick tabletop and legs guard against vibrations from footsteps, nearby elevators, and opening or closing doors. The cylinder is a vacuum chamber, devoid of all the gases in air. Inside the vacuum chamber is a smaller, extremely cold compartment, reachable by tiny laser beams. Inside are ultra-sensitive particles that make up a quantum computer.
這個金屬圓柱的內容, 可能可以帶來技術革命, 也可能完全沒用處—— 這全都要看我們能否 在非常非常小的尺度上 利用物質的奇特物理性質。 若要有機會能夠這麼做, 首先我們必須要能精準控制環境: 用很厚的桌面和桌腳對抗震動, 來自腳步、附近電梯 和開關門等等的震動。 這個圓柱是個真空腔體, 沒有空氣中的任何氣體。 在真空腔體中 有個較小、極冷的隔間, 用非常小的雷射光束可以到達。 裡面有特別敏感的粒子, 用來製作量子電腦。
So what makes these particles worth the effort? In theory, quantum computers could outstrip the computational limits of classical computers. Classical computers process data in the form of bits. Each bit can switch between two states labeled zero and one. A quantum computer uses something called a qubit, which can switch between zero, one, and what’s called a superposition. While the qubit is in its superposition, it has a lot more information than one or zero. You can think of these positions as points on a sphere: the north and south poles of the sphere represent one and zero. A bit can only switch between these two poles, but when a qubit is in its superposition, it can be at any point on the sphere. We can’t locate it exactly— the moment we read it, the qubit resolves into a zero or a one. But even though we can’t observe the qubit in its superposition, we can manipulate it to perform particular operations while in this state.
是什麼讓這些粒子 值得這麼大費功夫? 理論上,量子電腦可以超越 傳統電腦的計算限制。 傳統電腦處理資料時, 用的是位元的形式。 每一個位元會在兩種狀態中切換, 這兩種狀態被標為 0 和 1。 量子電腦用的是量子位元, 它可以在 0、1, 及所謂的疊加之間切換。 當量子位元在疊加狀態時, 它所具有的資訊遠超過 0 或 1。 可以把這些位置視為球體上的點: 球體的北、南極代表 1 和 0。 位元只能在兩極之間切換, 但當量子位元疊加時, 它能在球體上的任何位置。 我們無法精確地定位它—— 當我們讀量子位元時, 它就會拆解為 0 或 1。 儘管我們無法觀察疊加的量子位元, 我們能夠操控它,讓它在 這個狀態中執行特定的工作。
So as a problem grows more complicated, a classical computer needs correspondingly more bits to solve it, while a quantum computer will theoretically be able to handle more and more complicated problems without requiring as many more qubits as a classical computer would need bits.
所以,遇到更複雜的問題時, 傳統電腦需要相對 更多位元來解決它, 而理論上量子電腦 可以處理更複雜的問題, 傳統電腦需要許多額外的位元, 量子電腦卻不用那麼多量子位元。 量子電腦獨特的特性
The unique properties of quantum computers result from the behavior of atomic and subatomic particles. These particles have quantum states, which correspond to the state of the qubit. Quantum states are incredibly fragile, easily destroyed by temperature and pressure fluctuations, stray electromagnetic fields, and collisions with nearby particles. That’s why quantum computers need such an elaborate set up. It’s also why, for now, the power of quantum computers remains largely theoretical. So far, we can only control a few qubits in the same place at the same time.
源於原子和亞原子的粒子行為。 這些粒子的量子狀態 對應到量子位元的狀態。 量子狀態非常脆弱, 容易被溫度和壓力的波動、 雜散電場, 以及附近相撞的粒子所摧毀, 因而量子電腦需被精心設定。 正因如此,截至目前為止 量子電腦仍主要是理論上具有力量, 在同一時間、同一位置 只能控制少數幾個量子位元。
There are two key components involved in managing these fickle quantum states effectively: the types of particles a quantum computer uses, and how it manipulates those particles. For now, there are two leading approaches: trapped ions and superconducting qubits.
想要有效地管理 這些無常的量子狀態, 涉及到兩項關鍵要素: 量子電腦所使用的粒子類型, 以及它如何操控這些粒子。 現在有兩種主要的方法: 囚禁離子和超導量子位元。
A trapped ion quantum computer uses ions as its particles and manipulates them with lasers. The ions are housed in a trap made of electrical fields. Inputs from the lasers tell the ions what operation to make by causing the qubit state to rotate on the sphere. To use a simplified example, the lasers could input the question: what are the prime factors of 15? In response, the ions may release photons— the state of the qubit determines whether the ion emits photons and how many photons it emits. An imaging system collects these photons and processes them to reveal the answer: 3 and 5.
囚禁離子量子電腦 會把離子當作其粒子來使用, 並用雷射來操控它們。 離子會被放置在 電場製成的陷阱中。 來自雷射的輸入會造成 量子位元狀態在球體上轉動, 來告訴離子要做什麼工作。 用個簡化的例子來說明, 雷射可以輸入這個問題: 15 的質因數有哪些? 回應時,離子可能會釋出光子—— 量子位元的狀態 決定離子是否要釋出光子 和釋出多少光子。 成像系統收集這些光子, 處理它們,揭示答案: 3 和 5。
Superconducting qubit quantum computers do the same thing in a different way: using a chip with electrical circuits instead of an ion trap. The states of each electrical circuit translate to the state of the qubit. They can be manipulated with electrical inputs in the form of microwaves. So: the qubits come from either ions or electrical circuits, acted on by either lasers or microwaves. Each approach has advantages and disadvantages. Ions can be manipulated very precisely, and they last a long time, but as more ions are added to a trap, it becomes increasingly difficult to control each with precision. We can’t currently contain enough ions in a trap to make advanced computations, but one possible solution might be to connect many smaller traps that communicate with each other via photons rather than trying to create one big trap. Superconducting circuits, meanwhile, make operations much faster than trapped ions, and it’s easier to scale up the number of circuits in a computer than the number of ions. But the circuits are also more fragile, and have a shorter overall lifespan.
超導量子位元量子電腦 用不同的方式做同樣的事: 用電路晶片取代離子陷阱。 每個電路的狀態會被翻譯成 量子位元的狀態。 可輸入微波形式的電子來操控它們。 所以,量子位元來自離子或電路, 被雷射或微波操控。 兩種方法各有利弊。 離子的操控精確而且持久, 但當陷阱中的離子越來越多, 就越來越難以精確地控制各個離子。 我們目前還無法在一個陷阱中 放入足夠的離子來做進階的計算, 但有一個可能的解決方案: 改成連結許多較小的陷阱, 這些較小的陷阱會透過 光子來和彼此溝通, 這做法可以取代單個大陷阱。 超導電路的運作速度 比囚禁離子快很多, 而且,在電腦中增加電路的數目 會比增加離子的數目容易。 但其電路比較脆弱, 整體的壽命也比較短。
And as quantum computers advance, they will still be subject to the environmental constraints needed to preserve quantum states. But in spite of all these obstacles, we’ve already succeeded at making computations in a realm we can’t enter or even observe.
隨著量子電腦進步, 它們仍然會受制於 保存量子狀態所必要的環境限制。 儘管有上述這些障礙, 我們已經成功在一個我們無法進入 甚至無法觀察的領域中進行計算。