跳到主要內容
訪客人次:570415
:::
訪客人次:570415
:::
:::

科技部新聞稿:不失憶的記憶體—全球首例自旋流解密MRAM關鍵瓶頸

是什麼樣的記憶體技術,讓全球各半導體大廠如三星、東芝、英特爾等摩拳擦掌紛紛投入,準備在後摩爾定律世代一較高下?答案是磁阻式隨機存取記憶體(MRAM)。它是一種非揮發性記憶體技術,也就是斷電時,利用奈米磁鐵所儲存的資料並不會流失,是「不失憶」的記憶體。由國立清華大學賴志煌教授與林秀豪教授所帶領的研究團隊,在科技部長期的支持下,研究MRAM的特性、製程與操控,獨步全球,成功以自旋流操控鐵磁-反鐵磁奈米膜層的磁性翻轉,研究成果於今(108)年2月19日刊登於材料領域頂尖期刊《自然材料》(Nature Materials)。

 

電腦、平板、手機在生活中處處可見,其背後運作的核心功能,不外乎資訊的處理與儲存。開發適當的電子元件,既可以同時快速處理資訊,又能夠穩定儲存資訊,吸引學界業界各個領域的專家投入,而其中MRAM是極被看好的後摩爾定律世代的記憶體。其結構有如三明治,上層是自由翻轉的鐵磁層,可快速處理資料,底層則是釘鎖住的鐵磁層,可用作儲存資料,兩層中則有氧化層隔開。當此二鐵磁層的磁化方向相同,是低電阻態,代表「1」 ;此二鐵磁層的磁化方向相反,是高電阻態,代表「0」。有別於目前的主流記憶體(SRAM 與 DRAM),MRAM兼具處理與儲存資訊的功能,且斷電時資訊不會流失,電源開啟可即時運作,耗能低、讀寫速度快,是多方看好的明日之星。

 

其中一個技術關鍵,就是如何操控釘鎖住的鐵磁層。若想要將鐵磁層的磁矩方向釘鎖住,簡單卻神奇,只需「黏」上一層反鐵磁層即可,製成的鐵磁-反鐵磁膜層即可應用在磁記憶體上。此現象稱為「交換偏壓」,雖發現至今已超過60年,其應用性極廣,但背後的物理機制未明。而且交換偏壓的操控性極為有限。必須將元件升溫,然後於外加磁場下降溫,才能改變鐵磁層磁矩的釘鎖方向。

 

無論是外加磁場或是升降溫度,都與現有電子元件的操作格格不入。世界各研究團隊莫不希望突破此困境,尋求嶄新的操控技術。其中一個突破點,就是善用自旋流。電子具有電荷,也具有自旋:當電荷流動時,即會產生熟悉的電流,若有辦法驅動自旋流動,即可產生自旋流。賴教授與林教授的團隊利用自旋流通過鐵磁-反鐵磁膜層,率先展示操控元件「交換偏壓」方向與大小的里程碑。此技術可與現有電子元件的操控與製程無縫接軌,是MRAM的大突破,為自旋電子學的發展帶來嶄新視野。

 

利用自旋流操控交換偏置乃全球首見,賴教授表示當初投稿時引發諸多質疑,審稿委員懷疑是元件溫度升高所致,與自旋流無關。由於團隊橫跨材料與物理領域,兼具實驗與理論的專業能力,在面對高難度的質疑與挑戰時,能夠跳脫框架思考,以極高的效率與執行力,清楚精確地回應相關的質疑與挑戰。舉例來說,在經過理論分析與實驗操作評估後,研究團隊開發新的測量技術,首創該領域微秒等級的即時溫標。此技術可偵測元件任一時刻的溫度,藉此明確排除熱效應,成功消弭審稿委員的質疑。

 

目前研究團隊將此突破性的發現,應用到其它結構的奈米膜層,陸續發現更多具影響力的結果,除了學術的貢獻外,經由科技部半導體射月計畫的連結,將對於國內記憶體產業發展有決定性的影響力。這項技術在學理上的存取速度接近 SRAM,具快閃記憶體的非揮發性特性,平均能耗遠低於 DRAM,應用於嵌入式記憶體(Embedded Memory)極具潛力,隨著人工智慧、物聯網裝置與更多的資料收集與感測需求,MRAM的市場也將迅速成長。

 

A Memory that Remembers

 

The Ministry of Science and Technology and the National Tsing Hua University have long supported Professor Chih-Huang Lai and Professor Hsiu-Hau Lin’s research team to study characteristics, process and control of magnetic random access memory (MRAM). Setting the world’s first record, they succeeded in manipulating the magnetic switch of the ferromagnetic-antiferromagnetic nanolayers by spin current. This influential breakthrough was published in the prestigious journal “Nature Materials” recently.

 

Computers, tablets, and mobile phones are modern necessities in everyday life. The key functions behind these devices are nothing but information processing and storage. How to process information quickly, while maintaining the stability of stored information simultaneously, is the billion-dollar question attracting experts from all fields. In the post Moore’s law era, MRAM is the most promising candidate among the competing materials.

 

The structure of MRAM is like a sandwich: the upper layer is a free ferromagnetic layer (quick for information processing), the bottom layer is a pinned ferromagnetic layer (stable for information storage) and an oxide layer in-between to separate them. When the magnetization of the two ferromagnetic layers are parallel, it is in the low-resistance state, representing “1”. When magnetizations in the two layers are opposite, it is in the high-resistance state, representing “0”. MRAM are capable of information processing and storage at the same time. The stored information will not be lost even when the power is suddenly off. When the power is on, it can operate immediately with no booting process required. It consumes less energy with fast read and write speeds. Despite some technical bottlenecks to conquer, MRAM is viewed as the rising star in the future generation.

 

One of the key technical bottleneck is to manipulate the pinned ferromagnetic layer. How to pin a ferromagnetic layer in a specific direction? Simple yet magical. You only need to “glue” an antiferromagnetic layer on top. This phenomenon was discovered back in 1956 by scientists at General Electric in the United States and is referred as “exchange bias”. Although it has been discovered for over 60 years, the underlying mechanism remains unknown despite of its extremely wide applications. And, it is hard to manipulate the exchange bias. The device must be heated up first, followed by cooling process in the presence of magnetic field to set the direction of the exchange bias. This procedure is incompatible with existing manufacture processes in electronic industry.

 

Research teams around the world are seeking resolutions to conquer the technical bottleneck. Prof. Lai and Prof. Lin’s group achieve the goal by utilizing spin current. An electron has both charge and spin: when the change flows, it leads to the familiar (charge) current in everyday life. And, by driving the spin to move, the spin current is generated. They inject the spin current through the ferromagnetic-antiferromagnetic film layer and demonstrate how the magnitude and direction of the exchange bias can be manipulated. This breakthrough provides a brand new technical approach, seamlessly integrated with the existing electronic components. It is a milestone in magnetic memory and opens up a new horizon for the development of spintronics.

 

Using spin current to control the exchange bias has never been achieved before. When their manuscript was submitted to Nature Materials, many questions were raised during review processes. The common suspicion was that the observed results are due to thermal heating, not spin current injection. In the face of difficult questions and challenges, the strength of the interdisciplinary collaboration pays off. Combining the scopes of materials science and physics, the research team found ways to resolve the challenges with creative thinking and effective execution. For example, combining theoretical analysis and experimental investigation, the research team developed a new measurement technique to detect temperature variations within microsecond. This technique measures the temperature of the device at any instant, thereby explicitly ruling out the thermal effect and successfully eliminating all doubts from the reviewer.

 

And, this is not the end of the story. At present, the research team applies the breakthrough to devices with similar layer structures and discovers exciting phenomena. In addition to academic impacts to fundamental science, this breakthrough may have decisive influence in industrial developments. The Ministry of Science and Technology encourages interdisciplinary collaborations with long-term supports. With supports from the Engineering Division and the Natural Science Division, Professor Lai and Professor Lin form the interdisciplinary team and conquer the long-standing challenges with dazzling research results. The power of collaborations across different fields is self-evident.

 

 

Media Contact:

Professor Chih-Huang Lai

Department of Materials Science and Engineering, National Tsing Hua University

+886-910- 506524      

chlai@mx.nthu.edu.tw

 

Professor Hsiu-Hau Lin

Department of Physics, National Tsing Hua University

+886-963-179613       

hsiuhau.lin@gmail.com

 

Dr. Chin-Wei Chen

Program Manager/Assistant Research Fellow,

Department of Natural Sciences and Sustainable Development,

Ministry of Science and Technology

+886-2-2737-8070

cwchen@most.gov.tw

 

更新日期 : 2019/03/16