TERASWITCH - Towards low dissipation THz-induced switching of magnetic materials

磁化在两个稳定的比特状态（1和0）之间切换是现代存储技术的关键原理。随着“始终连接”设备数量的爆炸式增长以及消费者对多媒体和社交媒体内容的需求，全球存储和处理的数据量以前所未有的速度增长，新建数据中心的数量（例如Facebook在新加坡的新数据中心）就是明证。全球产生的大量数据正在导致新市场的出现，公司以不同的方式利用和交易数据-欧盟估计，到2020年，欧洲的数字经济价值将达到7390亿欧元[1]。对数据存储的需求不断增长，这提出了几个大问题：这些数据将存储在哪里？如何使对数据存储和处理的日益增长的需求与能源责任的政治和社会需求以及理想的碳中和相兼容？据估计，到2025年，全球20%的电力需求将用于为数据中心供电[2]，这一数字无疑将继续增长，因此任何降低数据处理和存储能源需求的技术都具有重要的国家和国际意义。该提案涉及减少数据存储和处理所涉及的能源使用的研究。磁性硬盘驱动器仍然构成服务器（因此云）级别的大部分数据存储，因为它们的每比特成本低。然而，在磁盘驱动器中写入信息的过程由于在两种状态之间切换位所需的磁场而使用相对大量的能量。使用飞秒（1飞秒是十亿分之一秒的百万分之一）激光脉冲的超快磁化动力学研究表明，低能量开关是可能的，使用数量级更少的能量。在这些研究中，转换发生在两皮秒（一皮秒是十亿分之一秒的千分之一）内，这使得每秒写入高达10^12（百万分之一）比特的可能性成为可能，比传统的记录方法快一千倍，这是一个非常有吸引力的途径，可以实现更快、更灵敏的设备，需要研究投资。然而，强激光脉冲的使用通常会导致大量的加热，并且可以激发许多非线性动力学。对此的一个可能的解决方案是使用具有高强度的THz范围内的频率的光。从历史上看，产生这样的光脉冲是非常困难的，但最近的实验发展使这成为可能，在过去的十年中，THz科学领域已经引起了人们的极大关注，最近又被用来控制磁性。初步研究表明，与传统记录相比，切换磁化状态所需的能量要低得多，这可能会彻底改变我们存储和处理信息的方式。该提案旨在发展这些想法，目的是了解潜在的物理过程以及我们如何设计有效，低能量的磁性控制。这项工作将与世界领先的实验小组一起进行，以提供重要的验证和与理论工作的比较。1-https://ec.europa.eu/digital-single-market/en/news/final-results-european-data-market-study-measuring-size-and-trends-eu-data-economy2-https://data-economy.com/data-centres-world-will-consume-1-5-earths-power-2025/

Magnetisation switching between two stable bit states (1 and 0) is the key principle of modern-day storage technology. With the explosion in the number of "always connected" devices, and the consumer desire for multimedia and social media content, the volume of data being stored and processed globally has risen at an unprecedented rate, as evidenced by the number of new data centres being built (e.g. Facebook's new data centre in Singapore). The vast quantities of data being generated globally is leading to the emergence of new markets with companies exploiting and trading data in diverse ways - an EU estimate values the digital economy in Europe will be worth 739bn euros by 2020[1]. This growing demand for data storage poses several big questions: where is this volume of data going to be stored? How can the growing demand for data storage and processing be made compatible with the political and social imperative for energy responsibility and, ideally, carbon neutrality? It is estimated that 20% of the world's electricity demand will be used to power data centres by 2025[2], a figure that will undoubtedly grow, and therefore any technology that reduces the energy requirements of data processing and storage is of great national and international importance. This proposal concerns research into reducing the energy use involved in data storage and processing.Magnetic hard disk drives still form most of the data storage at the server (and hence cloud) level due to their low cost per bit. However, the process of writing information in disk drives uses a relatively large amount of energy due to the magnetic field needed to toggle bits between the two states. Studies in ultrafast magnetization dynamics using femtosecond (1 femtosecond is one millionth of a billionth of a second) laser pulses have demonstrated that low-energy switching is possible, using orders of magnitude less energy. Switching in these studies occurs within two picoseconds (one picosecond is a thousandth of a billionth of a second) opening up the possibility of writing up to 10^12 (a million million) bits per second, one thousand times faster than conventional recording methods, an extremely attractive avenue to realise much faster and more responsive devices that requires research investment. However, the use of strong laser pulses often results in a large amount of heating and can excite a lot of non-linear dynamics. One possible solution to this is to use light at frequencies that are in the THz range with high intensities. Historically, it has been very difficult to generate such light pulses, but recent experimental developments have made this possible and the area of THz science in general has attracted significant attention over the past decade and more recently to control magnetism. Initial studies have shown that significantly lower amounts of energy are required to switch the magnetisation state than in conventional recording, which could revolutionise the way we store and process information. This proposal is aimed at developing these ideas with the goal of understanding the underlying physical processes and how we can engineer efficient, low energy control of magnetism. The work will be carried out alongside world-leading experimental groups to provide important validation and comparisons with theoretical work.1 - https://ec.europa.eu/digital-single-market/en/news/final-results-european-data-market-study-measuring-size-and-trends-eu-data-economy2 - https://data-economy.com/data-centres-world-will-consume-1-5-earths-power-2025/