The self-assembling computer chips of the future | Karl Skjonnemand

翻译人员: 校对人员: Lipeng Chen

Computers used to be as big as a room.

过去，计算机和房间一样庞大。

But now they fit in your pocket,

但是如今你可以把计算机揣进兜里，

on your wrist

戴在手腕上，

and can even be implanted inside of your body.

甚至是嵌入身体中。

多棒啊！

How cool is that?

And this has been enabled by the miniaturization of transistors,

这些都得益于晶体管的微型化，

which are the tiny switches in the circuits

晶体管是电路中的小开关，

at the heart of our computers.

位于计算机的核心区域。

And it's been achieved through decades of development

晶体管经过数十年的研发、

and breakthroughs in science and engineering

科学工程上的突破

and of billions of dollars of investment.

和数十亿美元的投入之后取得成功。

But it's given us vast amounts of computing,

它赋予了我们强大的计算能力、

huge amounts of memory

海量的记忆功能

and the digital revolution that we all experience and enjoy today.

以及我们共同经历的数字革命。

But the bad news is,

但是坏消息是，

we're about to hit a digital roadblock,

随着晶体管小型化的速率不断下降，

as the rate of miniaturization of transistors is slowing down.

我们即将迎来数字化的瓶颈。

And this is happening at exactly the same time

与此同时，

as our innovation in software is continuing relentlessly

我们在软件方面不断创新，

with artificial intelligence and big data.

人工智能和大数据蓬勃发展。

And our devices regularly perform facial recognition or augment our reality

我们的设备可以进行 面部识别以及现实增强，

or even drive cars down our treacherous, chaotic roads.

可以在危险、混乱的道路上 进行无人驾驶。

It's amazing.

简直不可思议！

But if we don't keep up with the appetite of our software,

但如果我们跟不上软件发展的速度，

we could reach a point in the development of our technology

就可能会达到科技发展的瓶颈，

where the things that we could do with software could, in fact, be limited

软件发展会受到限制，

by our hardware.

来自硬件发展的限制。

We've all experienced the frustration of an old smartphone or tablet

我们都经历过 在不断增多的软件更新

grinding slowly to a halt over time

和新功能的重压下，

under the ever-increasing weight of software updates and new features.

老版智能手机和平板带来的失望感， 加载缓慢甚至是停滞卡顿。

And it worked just fine when we bought it not so long ago.

我们刚买这些设备的时候， 它们运转得还不错。

But the hungry software engineers have eaten up all the hardware capacity

但是随着软件的更新，

硬件渐渐跟不上了。

over time.

The semiconductor industry is very well aware of this

半导体行业已经意识到了这一点，

and is working on all sorts of creative solutions,

并且致力于摆脱这一困境。

such as going beyond transistors to quantum computing

比如说超越晶体管到量子计算，

or even working with transistors in alternative architectures

或者在替代架构中使用晶体管，

such as neural networks

比如在神经网络中，

to make more robust and efficient circuits.

创造出更坚固有效的电路。

But these approaches will take quite some time,

但是这些方法都很耗时，

and we're really looking for a much more immediate solution to this problem.

我们正在寻找解决这个问题的捷径。

The reason why the rate of miniaturization of transistors is slowing down

晶体管小型化速率下降，

is due to the ever-increasing complexity of the manufacturing process.

是由制造过程日益复杂导致的。

The transistor used to be a big, bulky device,

过去，晶体管是 很大、很笨重的设备，

until the invent of the integrated circuit

直到基于纯晶硅片的

based on pure crystalline silicon wafers.

集成电路的问世， 晶体管才不断变小。

And after 50 years of continuous development,

在持续五十年的发展后，

we can now achieve transistor features dimensions

如今我们可以使晶体管的特性尺寸

down to 10 nanometers.

达到10纳米以下。

You can fit more than a billion transistors

你可以把超过十亿个的晶体管

in a single square millimeter of silicon.

放在一个一平方毫米的硅片中。

And to put this into perspective:

为了更形象地描述这一点， 我将提供一些数据：

a human hair is 100 microns across.

人的头发直径是100微米。

A red blood cell, which is essentially invisible,

一个肉眼几乎看不见的血红细胞，

is eight microns across,

直径是8微米。

and you can place 12 across the width of a human hair.

头发的宽度几乎是血红细胞的12倍。

But a transistor, in comparison, is much smaller,

但是相比之下，晶体管更小，

at a tiny fraction of a micron across.

直径远小于1微米。

You could place more than 260 transistors

晶体管的宽度，

是一个血红细胞的260分之一，

across a single red blood cell

or more than 3,000 across the width of a human hair.

是一个头发丝宽度的三千分之一。

It really is incredible nanotechnology in your pocket right now.

这个不可思议的纳米科技 现在就被你揣在兜里。

And besides the obvious benefit

除了显而易见的好处，

of being able to place more, smaller transistors on a chip,

即我们可以放置更多、 更小的晶体管在芯片中，

smaller transistors are faster switches,

更小的晶片还意味着更快的转换速度，

and smaller transistors are also more efficient switches.

也意味着更高的转换效率。

So this combination has given us

这个结合赋予我们

lower cost, higher performance and higher efficiency electronics

更低成本、更高性能 和更高效率的电子设备，

that we all enjoy today.

在今天为我们带来了极大的方便。

To manufacture these integrated circuits,

生产这些集成电路，

the transistors are built up layer by layer,

需要我们将晶体管 在一个纯晶硅片上

on a pure crystalline silicon wafer.

一层层地叠加起来。

And in an oversimplified sense,

简言之，

every tiny feature of the circuit is projected

电路的每一个微小特征 都被投射在

onto the surface of the silicon wafer

硅片表面，

and recorded in a light-sensitive material

被记录在光敏材料上，

and then etched through the light-sensitive material

然后被蚀刻在光敏材料上，

to leave the pattern in the underlying layers.

将图样留在底层。

And this process has been dramatically improved over the years

多年来，这一过程 得到了极大的改进，

to give the electronics performance we have today.

从而赋予了电子设备今日的表现。

But as the transistor features get smaller and smaller,

但是随着晶体管越变越小，

we're really approaching the physical limitations

我们迎来了制造技术的

of this manufacturing technique.

物理极限。

The latest systems for doing this patterning

最新制造底样的系统

have become so complex

变得十分复杂，

that they reportedly cost more than 100 million dollars each.

导致每件设备的成本 高达1亿多美金。

And semiconductor factories contain dozens of these machines.

而每家半导体工厂 都需要采购大量的这些设备。

So people are seriously questioning: Is this approach long-term viable?

于是人们开始正视这个问题： 这个方法是长期可行的吗？

But we believe we can do this chip manufacturing

但是我们相信我们可以 对芯片制造方法做出改变，

in a totally different and much more cost-effective way

用一种全新的、更划算的方式，

using molecular engineering and mimicking nature

使用分子工程和模拟自然的方法，

down at the nanoscale dimensions of our transistors.

在我们晶体管的纳米维度上。

As I said, the conventional manufacturing takes every tiny feature of the circuit

如我所说，传统制造方法将 电路的每一个微小特征

and projects it onto the silicon.

都投射到了晶片上。

But if you look at the structure of an integrated circuit,

但是如果你关注 一个集成电路的结构、

the transistor arrays,

晶体管的排列，

many of the features are repeated millions of times.

你会发现这些微小特征 被重复了数百万次。

It's a highly periodic structure.

这是一种高度周期性的结构。

So we want to take advantage of this periodicity

所以我们想在我们的替代生产技术中

in our alternative manufacturing technique.

利用这种周期性。

We want to use self-assembling materials

我们想使用自组装材料，

to naturally form the periodic structures

自然地组建周期性结构

that we need for our transistors.

来构建晶体管。

We do this with the materials,

我们用材料进行试验，

then the materials do the hard work of the fine patterning,

让这些材料完成 精细图案的制作工作，

rather than pushing the projection technology to its limits and beyond.

而不是试图在投射技术上寻找突破。

Self-assembly is seen in nature in many different places,

自组装原理在大自然中随处可见，

from lipid membranes to cell structures,

从脂质膜到细胞结构，

so we do know it can be a robust solution.

所以我们认为 这将会是有效的解决方法。

If it's good enough for nature, it should be good enough for us.

如果该方法可以应用于大自然， 同理可用于芯片产业。

So we want to take this naturally occurring, robust self-assembly

所以这一切就顺其自然了，

and use it for the manufacturing of our semiconductor technology.

将稳固的自组装方法 应用到半导体的生产中去。

One type of self-assemble material --

一种自组装材料——

it's called a block co-polymer --

名为嵌段共聚物——

consists of two polymer chains just a few tens of nanometers in length.

由两条长度只有 几十纳米的聚合物链组成，

But these chains hate each other.

但是这些聚合物链彼此排斥。

They repel each other,

它们彼此排斥，

very much like oil and water or my teenage son and daughter.

就像水油不相溶， 就像我青春期的儿女。

(Laughter)

（笑声）

But we cruelly bond them together,

但是我们强制使它们结合在一起，

creating an inbuilt frustration in the system,

在系统中创造一种嵌入式窘组，

as they try to separate from each other.

即便它们想要相互分离。

And in the bulk material, there are billions of these,

一块巨型材料， 包含着数十亿个这样的聚合物链，

相似的化合物会粘结在一起，

and the similar components try to stick together,

and the opposing components try to separate from each other

同时互斥的化合物则会

at the same time.

相互分离。

And this has a built-in frustration, a tension in the system.

这是嵌入式的窘组， 一种系统的张力。

So it moves around, it squirms until a shape is formed.

所以这些化合物四处移动， 蠕动直到形成一个形状。

And the natural self-assembled shape that is formed is nanoscale,

天然的自组装形状是纳米级的，

it's regular, it's periodic, and it's long range,

它有规律和周期性，还很长。

这就是我们在晶体管排列中所需要的。

which is exactly what we need for our transistor arrays.

So we can use molecular engineering

所以我们可以应用分子工程

to design different shapes of different sizes

来设计不同尺寸的不同形状，

and of different periodicities.

以及不同周期性的不同形状。

So for example, if we take a symmetrical molecule,

比如说，如果我们 选用一种对称分子，

where the two polymer chains are similar length,

它的两条聚合物链长度相似，

the natural self-assembled structure that is formed

则自然的自组装结构就会是

is a long, meandering line,

长的曲线形，

very much like a fingerprint.

像指纹一样。

And the width of the fingerprint lines

指纹线的宽度

and the distance between them

和其间的距离，

is determined by the lengths of our polymer chains

不仅取决于聚合物链的长度，

but also the level of built-in frustration in the system.

还取决于系统内嵌窘组的级别。

And we can even create more elaborate structures

我们还可以创造更复杂的结构。

if we use unsymmetrical molecules,

如果我们使用非对称分子，

where one polymer chain is significantly shorter than the other.

其中一条聚合物链显著短于另一条。

And the self-assembled structure that forms in this case

这种情况下的自组装结构是这样的：

is with the shorter chains forming a tight ball in the middle,

短链在中间形成一个牢固的圆球，

and it's surrounded by the longer, opposing polymer chains,

被包围在更长的、 相互排斥的聚合物链中，

forming a natural cylinder.

形成一个自然的圆柱体。

And the size of this cylinder

这个圆柱体的尺寸

and the distance between the cylinders, the periodicity,

以及圆柱体之间的距离、周期性，

is again determined by how long we make the polymer chains

取决于我们选用的聚合物链的长度，

and the level of built-in frustration.

以及内嵌窘组的水平。

So in other words, we're using molecular engineering

换言之，我们在利用分子工程

to self-assemble nanoscale structures

获得自组装的纳米结构。

that can be lines or cylinders the size and periodicity of our design.

这些结构可以是线形的、圆柱形的， 同时也符合我们设计的周期性。

We're using chemistry, chemical engineering,

我们在使用化学、化学工程

to manufacture the nanoscale features that we need for our transistors.

来制造我们晶体管 所需的纳米级特征。

But the ability to self-assemble these structures

但是自组装这些结构的能力

only takes us half of the way,

只解决了一半的问题，

because we still need to position these structures

因为我们还需要排列这些结构，

where we want the transistors in the integrated circuit.

使得晶体管们可以形成集成电路。

But we can do this relatively easily

但是这些东西相对更简单，

使用宽导向结构来固定自组装结构，

using wide guide structures that pin down the self-assembled structures,

将它们锚定到位，

anchoring them in place

and forcing the rest of the self-assembled structures

使剩余的自组装结构

to lie parallel,

可以平行排列，

aligned with our guide structure.

从而与我们的导向结构保持一致。

For example, if we want to make a fine, 40-nanometer line,

比如，如果我们想制作一个 精细的、40纳米长的线形，

which is very difficult to manufacture with conventional projection technology,

这对传统的投射技术 而言是非常困难的，

we can manufacture a 120-nanometer guide structure

我们可以先制作 一个120纳米的导向结构，

with normal projection technology,

使用普通的投射技术，

and this structure will align three of the 40-nanometer lines in between.

这个结构将把 3个40纳米长的线形排列在一起。

So the materials are doing the most difficult fine patterning.

所以这些材料在进行 最困难的精细复写。

And we call this whole approach "directed self-assembly."

我们称这种方法为： 直接自组装法。

The challenge with directed self-assembly

这种方法的挑战在于，

is that the whole system needs to align almost perfectly,

整个系统都需要完美地排列，

because any tiny defect in the structure could cause a transistor failure.

因为结构中任何微小的缺陷 都会导致晶体管的失效。

And because there are billions of transistors in our circuit,

因为我们电路中存在数十亿个晶体管，

we need an almost molecularly perfect system.

我们需要一个无比精细完美的系统。

But we're going to extraordinary measures

但我们需要付出非凡的努力，

来达到这一目标。

to achieve this,

from the cleanliness of our chemistry

从我们的化学清洁

to the careful processing of these materials

到在半导体工厂中的

in the semiconductor factory

这些材料的精细处理

to remove even the smallest nanoscopic defects.

从而消除纳米级别的最小失误。

So directed self-assembly is an exciting new disruptive technology,

所以直接自组装法是一种 全新的，令人激动的颠覆性技术。

but it is still in the development stage.

但是它还在发展阶段。

But we're growing in confidence that we could, in fact, introduce it

但是我们有信心在未来的几年里，

to the semiconductor industry

在半导体行业中

as a revolutionary new manufacturing process

引入这种全新的

in just the next few years.

变革型制造方法，

And if we can do this, if we're successful,

如果我们成功了，

we'll be able to continue

我们将能够继续进行

with the cost-effective miniaturization of transistors,

低成本的晶体管小型化、

continue with the spectacular expansion of computing

计算能力的快速发展

and the digital revolution.

以及数字的变革。

And what's more, this could even be the dawn of a new era

除此之外，这是将会是

of molecular manufacturing.

分子制造新纪元的曙光。

How cool is that?

听上去相当不错吧！

Thank you.

谢谢。

(Applause)

（掌声）