Tiny Chiplets: A New Level of Micro Manufacturing

Tiny Chiplets: A New Level of Micro Manufacturing

JOHN MARKOFF April 13, 2013

 

 

微型芯片：開啟數碼製造的技術革命

JOHN MARKOFF 報道 2013年04月13日

Amy Sullivan/PARC

微小的硅片經放大後的模樣，這種被稱為「微型芯片」的硅片大小不及一粒沙子。An enlarged view of small slivers of silicon, each no larger than a grain of sand, called chiplets. Using laser printers, Xerox’s Palo Alto Research Center may one day be able to create desktop manufacturing plants that use chiplets to “print” the circuitry for a wide array of electronic devices.

加利福尼亞州帕洛阿爾托——在顯微鏡下，四個硅片——被稱為微型芯片(chiplets)的電子電路——表演了一支複雜動感的舞蹈，彷彿是受控於一個隱藏的木偶大師。接着，它們按照指令精準地排列出一個電路圖案，每一個硅片都位於正確的接觸點。
在施樂公司(Xerox)的帕洛阿爾托研究中心（Palo Alto Research Center，簡稱PARC）展示的這項技術，是一種新型電子產品製造系統的一部分。該系統借鑒了施樂公司在上世紀70年代的發明：激光打印機。

這項技術一旦完善，就可以成就一個桌面製造廠，為各種的電子設備“打印”電路，這些設備包括有彈性的智能手機，即使你坐在上面也不會破；還有新型機器手的柔軟、敏感皮膚；以及具備智能感應功能的醫用繃帶，能收集你的健康數據，用後即可丟棄。

PALO ALTO, Calif. — Under a microscope, four slivers of silicon — electronic circuits called chiplets — perform an elaborate, jerky dance as if controlled by a hidden puppet master. Then on command, they all settle with pinpoint accuracy, precisely touching a pattern of circuit wires, each at just the right point of contact.
The technology, on display at Xerox’s Palo Alto Research Center, or PARC, is part of a new system for making electronics, one that takes advantage of a Xerox invention from the 1970s: the laser printer.
今天的芯片是用大片的晶圓(wafer)製造的，每片晶圓上都有數百個指甲蓋大小的晶粒(die)，每個晶粒里都有相同的電路。這些晶圓被切割成獨立的晶粒後會分別封裝，然後再在印刷電路板上重新組合，每個電路板都可能容納幾十甚至幾百個芯片。

PARC的研究人員則在設想一種截然不同的模型。在美國國家科學基金(National Science
Foundation)和國防高級研究計劃局（Defense Advanced Research Projects Agency，簡稱Darpa）的資助下，他們設計出一個類似激光打印機的機器，能精確地將幾萬乃至幾十萬個微型芯片準確地放在一個平面的正確位置上，每 個微型芯片的大小不超過一粒沙子。

這些芯片可以是微處理器或計算機內存，還可以是製造完整計算機系統所需的其他電路。它們也可以是微機電系統（microelectromechanical systems，簡稱MEMS）一類的模擬器件，用於感應溫度、壓力或運動。

在PARC研究人員的設想里，這種新型製造系統可被用來製造定製的計算機，每次僅製造一台，也可以被用作3D打印系統的一部分，以製造直接內置計算機系統的智能設備。

這仍是一項未來技術。研究人員還需要多年才能在比一秒還要短很多的時間裡同時把數萬或數十萬個電路放在準確的位置。他們也承認，要想製造出可堪商用的系統，這只是第一步。

然而，如果PARC的研究人員獲得成功，他們將顛覆硅谷50年來的傳統智慧。
生產射頻標識(RFID)設備的硅谷企業Alien Technology在另一項類似但相對簡單的技術上佔據着前沿地位。這種被稱為流體自組裝(Fluidic Self Assembly)的基本概念是讓一種稱為“納米塊”的小型集成電路懸浮在液體中，然後在一個平面上流淌，納米塊隨即掉入跟自己的形狀相符的小孔中。

五十年來，計算機生產的傳統思維是以每兩年翻一番的速度把更多的晶體管塞到指甲蓋大小的芯片中，從而讓計算機更快、更強大。而上述的兩種思路都徹底顛覆了這種思維。
這種新生的印刷技術提出了一個離經叛道的理念：為什麼一定要把儘可能多的晶體管塞到一個小空間里，為什麼不把晶體管塗抹在一個更大的平面上呢？

此外，這項研究也許會對經濟產生極其巨大的影響——在製造業開啟一個全新的數碼時代，堪比30年前激光打印技術在出版業掀起的革命。

這種技術可以取代現在需要工廠才能組裝的電路板，對現在這個遍布全球、需要僱傭成千上萬工人的供應鏈進行大幅壓縮。

同時它也是眾多和3D打印相關的技術之一，後者是公眾極為關注的領域，已經有人開始擔心今後一切都可以在家裡製造——從工具到槍支。

去年12月，麻省理工學院(MIT)比特和原子研究中心(Center for Bits and Atoms)負責人、物理學家尼爾·葛申菲爾德(Neil Gershenfeld)在《外交事務》(Foreign Affairs)雜誌上撰文稱：“數碼製造技術可以根據個人需求，在任何地點、任何時間去設計和製造實體物品。”
在原型設計領域，固體、機械物件的3D打印已經開始被大規模應用，這種技術在小規模生產領域也越來越普及，而PARC的科學家們認為，最終將會有一系列的製造技術，將微電子設備和機械元件無縫融合在一起。

“你可以打印機械物件，但是如今的世界裡很多東西不全是機械性的，”PARC主管史蒂芬·胡沃爾(Stephen Hoover)說，“如何以低廉的成本將智能大規模嵌入到世界中，將是‘物聯網’中大有可為的一個領域。”

PARC構想的新製造系統會更易於定製。比如說，就像我們現在可以用軟件來定製一台電腦的用途，未來的電腦可以根據它即將併入的系統來確定每台電腦的形狀。或許在未來電腦只是一個部件而已，成為3D打印系統的部件編目中的一項。

PARC電機工程師尤金·周(Eugene Chow)帶領的團隊設計了一種新技術，他們稱之為“靜電干印微組裝”。周說這種技術把晶圓分成幾萬個微型芯片，像“墨水”一樣裝在瓶子里，然後“打印”，跟施樂激光打印機在紙面上留下色粉差不多。

在計算機輔助下，PARC的這項技術會利用一系列可以產生微電場的電極來控制微小電路的精確定位——不但要在正確的位置上，還要以正確的方向來擺放。
“這是一個瘋狂的革命性新工具，”周博士說。

翻譯：經雷、陶夢縈


If perfected, it could lead to desktop manufacturing plants that “print” the circuitry for a wide array of electronic devices — flexible smartphones that won’t break when you sit on them; a supple, pressure-sensitive skin for a new breed of robot hands; smart-sensing medical bandages that could capture health data and then be thrown away.
Today’s chips are made on large wafers that hold hundreds of fingernail-sized dies, each with the same electronic circuit. The wafers are cut into individual dies and packaged separately, only to be reassembled on printed circuit boards, which may each hold dozens or hundreds of chips.
The PARC researchers have a very different model in mind. With financing from the National Science Foundation and from Darpa, the Pentagon’s Defense Advanced Research Projects Agency, they have designed a laser-printer-like machine that will precisely place tens or even hundreds of thousands of chiplets, each no larger than a grain of sand, on a surface in exactly the right location and in the right orientation.
The chiplets can be both microprocessors and computer memory as well as the other circuits needed to create complete computers. They can also be analog devices known as microelectromechanical systems, or MEMS, that perform tasks like sensing heat, pressure or motion.
The new manufacturing system the PARC researchers envision could be used to build custom computers one at a time, or as part of a 3-D printing system that makes smart objects with computing woven right into them.
The technology is still in the future. The researchers are years from simultaneously placing tens or hundreds of thousands of circuits accurately in a fraction of a second. And they acknowledge that this would be only the first step in designing a commercially viable system.
Still, if the PARC researchers are successful, they will have thrown out 50 years of Silicon Valley conventional wisdom.
A related but simpler technology was pioneered by Alien Technology, a maker of RFID tags in Silicon Valley. Called Fluidic Self Assembly, it is based on suspending small integrated circuits called “nanoblocks” in a fluid and then flowing them over a surface where they drop into tiny holes of corresponding shapes.
Both approaches reverse a five-decade long tradition of making computers faster and more powerful by doubling every two years the number of transistors squeezed onto fingernail-sized computer chips.
The emerging printing technology poses a heretical idea: Rather than squeezing more transistors into the same small space, why not smear the transistors across a much larger surface?
Moreover, the research could have tremendous economic consequences — feeding the emergence of a new digital era in manufacturing, much as laser printing transformed publishing three decades ago.
By replacing the circuit boards now assembled in factories, the technology would vastly compress a supply chain that spans the globe and employs hundreds of thousands of workers.
It is one of a variety of technologies related to 3-D printers, which have captured the public’s imagination, raising the specter of homemade manufacturing of everything from tools to guns.
“Digital fabrication will allow individuals to design and produce tangible objects on demand, wherever and whenever they need them,” Neil Gershenfeld, a physicist who directs the Center for Bits and Atoms at M.I.T., wrote in December in the journal Foreign Affairs.
While there has already been an explosion in 3-D printing of solid and mechanical objects both for prototyping and increasingly for small production runs, PARC’s scientists believe that there will also ultimately be an ensemble of manufacturing technologies that seamlessly blend microelectronics with mechanical components.
“You can print mechanical objects, but a lot of things in the world today are more than mechanical,” said Stephen Hoover, PARC’s chief executive. “A lot of the opportunities we’re going to find in the ‘Internet of things’ are going to be about how to embed intelligence at very low cost in a distributed way into the world.”
The new manufacturing system the PARC researchers envision would allow easier customization. For example, just as today software customizes computers for each purpose, computers in the future could be individually shaped for each system they were added to. Or in the future the computer could just be another part, to be added to the component inventory of a 3-D printing system.
Eugene Chow is an electrical engineer who leads the PARC team that has designed the new technology that they have dubbed “Xerographic micro-assembly.” The technology breaks silicon wafers into tens of thousands of chiplets, bottles them as “ink” and then “prints” them, much as a Xerox laser printer puts toner on paper, he said.
The PARC technology is based on the ability to use computing and an array of electrodes that generate microscopic electrical fields to control the precise placement of tiny electronic circuits — not just in the correct position, but with the proper orientation as well.
“It’s a crazy new revolutionary tool,” Dr. Chow said.