The Small Business Innovation Research (SBIR) Phase II project will design and construct prototype measurement systems based on near infrared (NIR) chromatic confocal sensor technology. Silicon is transparent in the NIR, and thus the sensor measures the distance to the front and back surfaces of the wafer simultaneously. The sensor will measure deep trenches and vias from the back side so that their aspect ratios are of no consequence. The proposed innovations lie in the sensor design and integration.

The proposed measurement systems will address the following semiconductor industry needs: 1) in situ wafer thickness measurement during wafer thinning operations; 2) wafer thickness and shape measurements of ultra-thin wafers; and 3) the measurement of deep, high aspect ratio, etched trenches and vias in silicon. Direct, in situ, measurements during wafer thinning are not currently possible. Neither is the nondestructive measurement of trench depth of many types of deep etched trenches and vias. The measurement of the thickness of ultra-thin wafers (<150 micron) requires greater accuracy for less cost than is currently available. Present technology does not have the resolution for measuring thickness in this thinner range, nor does it have sufficient spatial density on the wafer to accurately describe its shape.

Project Report

Tamar Technology built a white light sensor for measuring surface topography in a non-contact manner. An opportunity was seen in the semiconductor industry to measure the depth of high aspect ratio (larger depth to width ratio) features that were etched into silicon wafers. Present technology required the wafer to be cut through the feature of interest, and measured from the side with a scanning electron microscope (SEM). This was especially true of Through Silicon Vias (TSVs), which are essentially very small holes etched into wafers. Tamar understood that silicon is transparent to near infrared and also understood that the high aspect ratio features could be measured from the bottom of the wafer by "looking" through the wafer from the bottom and this turns the hole or trench into a "bump". So Tamar started to build a sensor that would move its Chromatic Confocal technology into the infrared region of the spectrum. Custom lenses were designed and the designs were sent out for fabrication, which takes 10 –12 weeks. During that time, a test set up was assembled to allow Tamar engineers to learn to use some of the electronic and optical components that were purchased. Tamar engineers know that this test set up would measure wafer thickness. What they discovered serendipitously, is that with a very small spot, the wafer thickness measurement could be made under a very small feature, including TSVs. Thus, measuring wafer thickness next to a TSV, and then measuring under the TSV, one measurement could be subtracted from the other to find etch depth. Further refinements to the optical train improved the measurement technique even more. Tamar ended up abandoning the first approach and followed the new path. It turned out that Tamar ended up with a completely unexpected sensor to solve the original problem, and that new method was faster, more accurate by a factor of about 10, and finished the project earlier. This measurement technology has been released and is now being sold in Tamar’s WaferScan 300 system. Sales have been realized at fortune 500 companies that make etching equipment and at customer sites that use the etching equipment. Tamar considers the SBIR project a great success. In the true nature of scientific research, a result was obtained that is not only completely different that the original plan, but also higher performance than the original plan.

Project Start
Project End
Budget Start
2008-04-01
Budget End
2011-03-31
Support Year
Fiscal Year
2007
Total Cost
$765,387
Indirect Cost
Name
Tamar Technology
Department
Type
DUNS #
City
Newbury Park
State
CA
Country
United States
Zip Code
91320