{"id":3215,"date":"2013-10-09T13:08:02","date_gmt":"2013-10-09T13:08:02","guid":{"rendered":"https:\/\/www.ophiropt.com\/blog\/laser-measurement\/?p=3215"},"modified":"2024-12-17T11:51:21","modified_gmt":"2024-12-17T11:51:21","slug":"using-power-meters-to-help-keep-moores-law-alive","status":"publish","type":"post","link":"https:\/\/www.ophiropt.com\/blog\/using-power-meters-to-help-keep-moores-law-alive\/","title":{"rendered":"Using Power Meters to Help Keep Moore\u2019s Law Alive"},"content":{"rendered":"

In 1965, Intel\u2019s Gordon Moore made a prediction about the future of technology that remained accurate far longer than even Moore himself expected. \u00a0It has come be known as: \u201cMoore\u2019s Law.\u201d<\/p>\n

Moore observed that the number of components that fit on a chip (at minimum cost) had been doubling each year.\u00a0 He simply predicted that it would continue to do so.\u00a0\u00a0\"\"<\/a><\/p>\n

Processors supporting Moore’s Law; Author: Wgsimon<\/em><\/p>\n

Over the course of the last 50 years, there have been multiple times when this exponential growth seemed to be approaching an inevitable slow-down.<\/p>\n

Every day, electronics get smaller.\u00a0 This trend allows for Moore\u2019s law to continue, but more importantly, it means that we can accomplish more and more with less and less.\u00a0 For example, shown below is a Datapoint 2200 (one of the first PCs) and a Mac mini.\u00a0 The old computer, a Datapoint 2200 is dozens of times heavier and larger (the pictures below are approximately to scale), while the Mac\u2019s processor speed and memory are orders of magnitude greater (about 6).\u00a0 On top of all that, the Datapoint would cost over $25,000 today (after inflation adjustment), while the Mac goes for $800.<\/p>\n

\"\"<\/a><\/p>\n

Datapoint 2200 (1971); Author: Ecksemmess<\/em>\u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0 \u00a0Mac mini (2011); Author: Wiki637<\/em><\/p>\n

Gordon Moore observed this trend decades ago, but by-and-large it has continued well past anyone \u2013 even Moore \u2013 could have expected.\u00a0 Whenever the exponential seems to approach an insurmountable ceiling, a new technology is developed or some novel approach is designed that makes the impossible seem downright simple.<\/p>\n

One example of a new technology extending Moore\u2019s \u201cprophecy\u201d beyond the expected was Kanti Jain\u2019s use of ultraviolet laser light to improve lithography<\/a> in 1982.\u00a0 Shorter wavelengths mean smaller feature sizes, and currently a 193nm excimer laser is capable of producing chips with a 40nm dimension.\u00a0 (I know, this seems impossible, but clever methods are employed to actually create structures shorter than the wavelength of the laser itself.)<\/p>\n

Once again, however, we\u2019re starting to face an (seemingly) unavoidable obstacle. As chip designers attempt to create features smaller than 40nm, the natural next step is smaller wavelength UV lasers.\u00a0 If a 193nm can produce 40nm chips, why can\u2019t a 13.5nm laser produce, say, 10nm chips?\u00a0 Perhaps it can, but there are proving to be several issues with using extreme UV (or EUV) lasers for this:<\/p>\n