Building a Quantum Computer, One Fragile Qubit at a Time
P ractically all modern computers, from the cheap microcontroller in your dishwasher to high-tech hardware crunching numbers for artificial intelligence systems, rely on versions of the same technology: slabs of silicon patterned with microscopic structures called transistors.
几乎所有现代计算机，从洗碗机中廉价的微控制器到人工智能系统中处理数字的高科技硬件，都依赖于同一技术的不同版本：带有称为晶体管的微观结构的硅片。
Electronic circuits containing transistors can rapidly and reliably toggle between two states, usually labeled “0” and “1.” That enables them to store and manipulate bits, the basic units of information.
包含晶体管的电子电路可以快速可靠地在两种状态之间切换，通常标记为“0”和“1”。这使它们能够存储和操作位（信息的基本单位）。
Quantum computers have the potential to process information in new and more powerful ways beyond mere 1s and 0s, and to solve certain problems that are too hard for their ordinary “classical” cousins.
量子计算机有潜力以新的、更强大的方式处理信息，而不仅仅是 1 和 0，并解决某些对其普通“经典”同类来说太难的问题。
But building a machine powerful enough to fulfill that promise remains a formidable challenge.
但建造一台足够强大的机器来实现这一承诺仍然是一项艰巨的挑战。
Quantum computing hasn’t yet had its transistor moment, and researchers are still exploring many different approaches to developing quantum hardware.
量子计算尚未出现晶体管时代，研究人员仍在探索开发量子硬件的许多不同方法。
Current approaches differ first and foremost in which physical systems they use as qubits, the elementary building blocks of quantum computers.
当前方法的不同之处首先在于它们用作量子位的物理系统，量子位是量子计算机的基本构建块。
Unlike circuits that store classical bits, qubits can exhibit strange phenomena like superposition and entanglement that give them extra computational power.
与存储经典位的电路不同，量子位可以表现出奇怪的现象，例如叠加和纠缠，从而赋予它们额外的计算能力。
But these quantum effects are also very fragile, easily disrupted by stray interactions between qubits and the surrounding environment.
但这些量子效应也非常脆弱，很容易被量子位与周围环境之间的杂散相互作用破坏。
Each proposed qubit technology tries to reconcile two properties that are hard to achieve simultaneously: Qubits must be isolated from outside disturbances, but easy for researchers to manipulate.
每一种提出的量子位技术都试图协调两个难以同时实现的特性：量子位必须与外界干扰隔离，但又易于研究人员操纵。
Some researchers have placed their bets on natural quantum systems such as atoms.
一些研究人员将赌注押在原子等自然量子系统上。
To use a single atom as a qubit, you must first isolate and trap it in a vacuum chamber, and researchers have pursued two distinct approaches to doing so.
要使用单个原子作为量子位，必须首先将其隔离并捕获在真空室中，研究人员已经采用了两种不同的方法来实现这一目标。
In trapped-ion quantum computing, researchers knock one electron off each atom to get positively charged ions that can be held in place by electric fields.
在俘获离子量子计算中，研究人员将每个原子上的一个电子击落，以获得带正电的离子，这些离子可以通过电场固定在适当的位置。
The other approach uses arrays of tightly focused laser beams, called optical tweezers, to trap neutral atoms.
另一种方法使用紧密聚焦的激光束阵列（称为光镊）来捕获中性原子。
Other researchers are pursuing an alternative approach, called superconducting quantum computing, which involves the design of artificial qubits.
其他研究人员正在寻求一种称为超导量子计算的替代方法，其中涉及人工量子位的设计。
Using modified versions of microfabrication processes developed for classical computing, researchers assemble tiny circuits made of metals like aluminum and niobium that become superconductors when cooled to very low temperatures.
研究人员利用为经典计算开发的微加工工艺的改进版本，组装了由铝和铌等金属制成的微型电路，这些金属在冷却到极低的温度时会变成超导体。
Housed in special cryogenic systems called dilution refrigerators, these superconducting circuits can act like qubits.
这些超导电路安装在称为稀释冰箱的特殊低温系统中，可以像量子位一样工作。
Many other qubit candidates have been explored, from electron spins to photons to more exotic quantum systems .
人们已经探索了许多其他候选量子位，从电子自旋到光子再到更奇特的量子系统。
Scaling up from small prototypes to much larger systems is one of the biggest challenges facing all these approaches.
从小型原型扩展到更大的系统是所有这些方法面临的最大挑战之一。
It’s not enough to make a few good qubits: Researchers will ultimately need at least tens of thousands, by even the most optimistic estimates, and perhaps even millions.
仅仅制造几个好的量子位是不够的：即使是最乐观的估计，研究人员最终也将需要至少数万个量子位，甚至可能需要数百万个。
More qubits also mean larger and more complex control and measurement systems.
更多的量子位也意味着更大、更复杂的控制和测量系统。
While it’s too early to say which technology, if any, will win out, the following images offer a glimpse inside the ambitious efforts required to build reliable quantum computers.
虽然现在说哪种技术（如果有的话）会胜出还为时过早，但下面的图片让我们了解了构建可靠的量子计算机所需的雄心勃勃的努力。
This striking 2017 photograph, taken in a trapped-ion quantum computing lab at the University of Oxford, shows a single strontium ion suspended in a vacuum chamber.
这张引人注目的照片是 2017 年在牛津大学的俘获离子量子计算实验室拍摄的，显示了悬浮在真空室中的单个锶离子。
Electric fields generated by needle- and blade-shaped steel electrodes confine the ion to the center of a 2-millimeter region between needle tips.
针状和刀片状钢电极产生的电场将离子限制在针尖之间 2 毫米区域的中心。
The ion is visible because it’s continuously absorbing photons from a laser beam and emitting photons to shed the excess energy.
离子是可见的，因为它不断吸收激光束中的光子并发射光子以释放多余的能量。
Relatively bulky “blade traps” like this one are often used in academic labs to test new quantum computing techniques on a few qubits.
像这样相对庞大的“刀片陷阱”经常在学术实验室中使用，以在几个量子位上测试新的量子计算技术。
David Nadlinger/University of Oxford
大卫·纳德林格/牛津大学
To scale up trapped-ion quantum computers, researchers will need to use more compact “surface traps,” in which the ions are levitated above a chip by electric fields generated by thin gold electrodes on the chip surface.
为了扩大捕获离子量子计算机的规模，研究人员需要使用更紧凑的“表面陷阱”，其中离子通过芯片表面的薄金电极产生的电场悬浮在芯片上方。
This surface trap, the subject of a 2021 paper by researchers at the National Institute of Standards and Technology in Boulder, Colorado, features an integrated photon detector (tiny solid black dot in center of chip) used to measure the state of beryllium ions that serve as qubits.
这种表面陷阱是科罗拉多州博尔德国家标准与技术研究所的研究人员在 2021 年发表的一篇论文的主题，其特点是集成光子探测器（芯片中心的微小实心黑点），用于测量用作量子位的铍离子的状态。
Operating a neutral-atom quantum computer requires many different laser systems for moving, trapping, measuring, and manipulating the atoms that serve as qubits.
操作中性原子量子计算机需要许多不同的激光系统来移动、捕获、测量和操纵充当量子位的原子。
This photograph of an optical table in a Harvard University neutral-atom lab shows a few of the mirrors, lenses, optical fibers, and detectors used along with other components to route, split, and stabilize the laser beams.
这张哈佛大学中性原子实验室光学台的照片显示了一些镜子、透镜、光纤和探测器，它们与其他组件一起用于路由、分裂和稳定激光束。
The heart of a neutral-atom quantum computer is the glass vacuum cell in which atoms are trapped in optical tweezer arrays.
中性原子量子计算机的核心是玻璃真空室，原子被捕获在光镊阵列中。
This vacuum cell, from the quantum computing start-up QuEra, is surrounded by coiled wires that enable researchers to precisely control the magnetic field in the trap region.
这个真空室来自量子计算初创公司 QuEra，周围环绕着线圈，使研究人员能够精确控制陷阱区域的磁场。
Also visible are components used to focus laser beams into the cell.
还可以看到用于将激光束聚焦到电池中的组件。
With the optical tweezer arrays used in neutral-atom quantum computing, it’s easy to rearrange atoms to bring any pair of qubits into contact.
利用中性原子量子计算中使用的光镊阵列，可以轻松地重新排列原子以使任何一对量子位接触。
In a 2018 paper , researchers at Paris-Saclay University showed off this capability by arranging atoms in many striking three-dimensional shapes, including one modeled after the Eiffel Tower.
在 2018 年的一篇论文中，巴黎萨克雷大学的研究人员通过将原子排列成许多引人注目的三维形状（包括模仿埃菲尔铁塔的形状）展示了这种能力。
In a paper published in 2025, a team of researchers at the California Institute of Technology trapped a record 6,100 individually controlled cesium atoms in an optical tweezer array 1 millimeter in diameter.
在 2025 年发表的一篇论文中，加州理工学院的一组研究人员在直径 1 毫米的光镊阵列中捕获了创纪录的 6,100 个单独控制的铯原子。
They showed that the qubits encoded in these atoms had long coherence times, a prerequisite for quantum computing.
他们表明，这些原子中编码的量子位具有较长的相干时间，这是量子计算的先决条件。
But they didn’t do any computations — the record for the largest number of qubits used to run a quantum algorithm remains far lower.
但他们没有进行任何计算——用于运行量子算法的最大量子比特数的记录仍然低得多。
The recipe for crafting superconducting qubits borrows much from the process for making classical computer chips.
制造超导量子位的方法很大程度上借鉴了制造经典计算机芯片的过程。
Thin strips of metal and other materials are layered on pristine silicon disks to make the qubits, and the disks are then diced into separate chips.
将金属和其他材料的细条层叠在原始硅盘上以形成量子位，然后将硅盘切成单独的芯片。
This chip fabrication process takes place in a clean room, like this one in Albany, New York (top), to prevent contamination from dust and other particulates.
该芯片制造过程在洁净室中进行，例如纽约奥尔巴尼的洁净室（上），以防止灰尘和其他颗粒物的污染。
The chips (bottom) were produced from an IBM design code-named Nighthawk.
这些芯片（下）是根据代号为 Nighthawk 的 IBM 设计生产的。
Each chip hosts 120 qubits along with other superconducting circuits for controlling interactions between them.
每个芯片拥有 120 个量子位以及其他超导电路，用于控制它们之间的相互作用。
The recipe for crafting superconducting qubits borrows much from the process for making classical computer chips.
制造超导量子位的方法很大程度上借鉴了制造经典计算机芯片的过程。
Thin strips of metal and other materials are layered on pristine silicon disks to make the qubits, and the disks are then diced into separate chips.
将金属和其他材料的细条层叠在原始硅盘上以形成量子位，然后将硅盘切成单独的芯片。
This chip fabrication process takes place in a clean room, like this one in Albany, New York (left), to prevent contamination from dust and other particulates.
该芯片制造过程在洁净室中进行，例如纽约奥尔巴尼的洁净室（左），以防止灰尘和其他颗粒物的污染。
The chips shown here (right) were produced from an IBM design code-named Nighthawk.
此处（右）所示的芯片是根据代号为 Nighthawk 的 IBM 设计生产的。
Each chip hosts 120 qubits along with other superconducting circuits for controlling interactions between them.
每个芯片拥有 120 个量子位以及其他超导电路，用于控制它们之间的相互作用。
This chip from the early 2010s, designed by researchers at the University of California, Santa Barbara, provides an illustration of superconducting qubit design.
这款 2010 年代初的芯片由加州大学圣塔芭芭拉分校的研究人员设计，为超导量子位设计提供了例证。
It measures 6 millimeters on each side and hosts four superconducting qubits (gold-colored squares in middle of green and yellow regions).
它的每边尺寸为 6 毫米，拥有四个超导量子位（绿色和黄色区域中间的金色方块）。
Each qubit is paired with a meandering superconducting strip, called a resonator, that serves as memory for storing quantum information.
每个量子位都与一个称为谐振器的蜿蜒超导带配对，用作存储量子信息的存储器。
Qubits interact with each other through the longer zigzagging resonator in the center of the chip.
量子位通过芯片中心较长的锯齿形谐振器相互作用。
External connections deliver microwave pulses for control and measurement.
外部连接提供用于控制和测量的微波脉冲。
Chandelier-like dilution refrigerators are perhaps the most striking examples of quantum computing infrastructure.
类似枝形吊灯的稀释冰箱可能是量子计算基础设施最引人注目的例子。
On top is “Badger,” used in an influential 2009 demonstration of simple quantum algorithms on two superconducting qubits by Yale University researchers.
最上面的是“Badger”，耶鲁大学研究人员在 2009 年对两个超导量子位上的简单量子算法进行了一次颇具影响力的演示中使用了它。
A plumbing system along its right side circulates a helium isotope mixture for cooling.
其右侧的管道系统循环氦同位素混合物以进行冷却。
During operation, a chip hosting the qubits would be anchored to the copper box below the lowest plate, and the system would be enclosed in nested cylindrical shields to block thermal radiation and provide vacuum insulation.
在操作过程中，承载量子位的芯片将固定在最低板下方的铜盒上，系统将被封闭在嵌套的圆柱形屏蔽中，以阻挡热辐射并提供真空绝缘。
On the bottom, a more modern dilution refrigerator from Google Quantum AI shows one of the engineering challenges involved in scaling up superconducting qubit systems.
底部，来自 Google Quantum AI 的更现代的稀释冰箱展示了扩大超导量子位系统所面临的工程挑战之一。
Many cables, with S-shaped bends to relieve strain from thermal contraction during cooling, are needed to control individual qubits and tune their interactions.
许多电缆具有 S 形弯曲，以减轻冷却过程中热收缩产生的应变，需要控制单个量子位并调整它们的相互作用。
Quantum computing technology is advancing rapidly, but there’s still a long way to go.
量子计算技术正在迅速发展，但仍有很长的路要走。
Chandelier-like dilution refrigerators are perhaps the most striking examples of quantum computing infrastructure.
类似枝形吊灯的稀释冰箱可能是量子计算基础设施最引人注目的例子。
On the left is “Badger,” used in an influential 2009 demonstration of simple quantum algorithms on two superconducting qubits by Yale University researchers.
左边是“Badger”，耶鲁大学研究人员在 2009 年对两个超导量子位的简单量子算法进行了一次颇具影响力的演示中使用了它。
A plumbing system along its right side circulates a helium isotope mixture for cooling.
其右侧的管道系统循环氦同位素混合物以进行冷却。
During operation, a chip hosting the qubits would be anchored to the copper box below the lowest plate, and the system would be enclosed in nested cylindrical shields to block thermal radiation and provide vacuum insulation.
在操作过程中，承载量子位的芯片将固定在最低板下方的铜盒上，系统将被封闭在嵌套的圆柱形屏蔽中，以阻挡热辐射并提供真空绝缘。
On the right, a more modern dilution refrigerator from Google Quantum AI shows one of the engineering challenges involved in scaling up superconducting qubit systems.
右侧，来自 Google Quantum AI 的更现代的稀释冰箱展示了扩大超导量子位系统所面临的工程挑战之一。
Many cables, with S-shaped bends to relieve strain from thermal contraction during cooling, are needed to control individual qubits and tune their interactions.
许多电缆具有 S 形弯曲，以减轻冷却过程中热收缩产生的应变，需要控制单个量子位并调整它们的相互作用。
Quantum computing technology is advancing rapidly, but there’s still a long way to go.
量子计算技术正在迅速发展，但仍有很长的路要走。
Jessica Smolinski (left); Google Quantum AI
杰西卡·斯莫林斯基（左）；谷歌量子人工智能
Get highlights of the most important news delivered to your email inbox
获取发送到您的电子邮件收件箱的最重要新闻的摘要
Quanta Magazine moderates comments to facilitate an informed, substantive, civil conversation.
《Quanta》杂志对评论进行审核，以促进知情、实质性、文明的对话。
Abusive, profane, self-promotional, misleading, incoherent or off-topic comments will be rejected.
辱骂、亵渎、自我推销、误导、语无伦次或偏离主题的评论将被拒绝。
Moderators are staffed during regular business hours (New York time) and can only accept comments written in English.
版主在正常工作时间（纽约时间）有工作人员，并且只能接受英文评论。