Measurement of Arrays of Dots Produced by Electron-Beam Lithography
- PDF / 3,308,494 Bytes
- 12 Pages / 612 x 792 pts (letter) Page_size
- 105 Downloads / 251 Views
0961-O19-02
Measurement of Arrays of Dots Produced by Electron-Beam Lithography Philip C. Hoyle and Ian Laidler Vistec Lithography Ltd., PO Box 87, 515 Coldhams Lane, Cambridge, CB1 3XE, United Kingdom
ABSTRACT Electron-beam mastering of templates for patterned media presents a challenge to the toolmaker to simultaneously meet throughput, resolution and placement requirements. Fundamental to tool development is the ability to measure the placement to true grid of shapes as small as 7 nm over the whole substrate. In this article we describe a technique, consisting of acquiring and analyzing scanning electron (SE) micrographs, for measuring the placement errors in lithography similar to that required for patterned media, albeit over a few square microns and without scale and orthogonality components. The method enabled the measurement of placement errors of dots in an array with accuracy down to about 2 nm. The technique was used to benchmark current X-Y tool performance and the smallest 3 × standard deviation of placement error was found to be 4.5 nm. A clearer understanding of the necessary tool improvements was obtained. The use of the technique as basis for measuring errors to true grid over the entire substrate is discussed. INTRODUCTION E-beam mastering of templates for patterned media presents a challenge to the toolmaker to simultaneously meet throughput, resolution and placement requirements. Depending on the data 1 density, magnetic islands are expected to have diameters in the range 50 nm to 7 nm , a th maximum position error of about 1/10 of the diameter and will be exposed as isolated shapes. A mastering tool will have a continuously moving rotary stage, as this is expected to outperform a tool with an X-Y stage. A rotary stage eliminates deflection field stitching, stage and deflection movement time overheads, and ensures any positional drift is spread in a more favorable pattern. Placement results from X-Y tools have typically been obtained using optical metrology tools, such as the Leica IPRO2, measuring marks several microns in size and therefore each measurement is an average of many hundreds or thousands of e-beam exposure elements (exels). The smallest marks that can be used take at least about 1 ms to expose. A mastering tool for 1 2 Tb/in must improve significantly on state-of-the-art placement accuracy for shapes that are made of perhaps only 1 to 10 exels. Such shapes will need to be exposed at rates of 1 to 10 MHz 12 so that the total time to expose the approximately 4x10 bits on a 2.5 inch diameter disk remains practical and economically viable. Therefore any noise, vibration and interference up to 10 MHz will reduce the shape placement accuracy. There are several potential sources of interference in current X-Y tools that fall into this extended band, such as the transmission of power at 125 kHz via the isolation transformer to the field emission tip supplies and various switched mode power supplies operating at tens of kHz. Any reduction in placement accuracy needs to be quantified to ass
Data Loading...
