Testing of Critical Features of Polysilicon MEMS
- PDF / 2,218,743 Bytes
- 6 Pages / 412.92 x 637.2 pts Page_size
- 106 Downloads / 199 Views
(a) (b) Figure 1. (a) First generation sample (b) Modified sample design, with a more robust pivot and larger radii in the gage length fillets. Samples designed to directly measure, under uniform tensile loading, the fracture strength and Young's modulus of polysilicon have been fabricated using Sandia National Laboratories' Ultra-planar Multi-level MEMS Technology (SUMMiT) process. The pull-ring end of a sample is engaged by a 35 gtm diameter flat-tipped diamond using a nanoindenter. Normal force, lateral force and displacement are recorded. A normal force must be maintained 19 Mat. Res. Soc. Symp. Proc. Vol. 605 0 2000 Materials Research Society
throughout the test to prevent the conical shaped diamond from being pushed upwards by the pull-ring engagement reaction. Once on the surface, the tip moves laterally, which loads the thin polysilicon ligament in tension. Each sample has a freely moving pivot at the fixed end and a pull-ring. The gage section is nominally 2 gim wide and 2.5 pim thick. After fabrication, the width is typically 1.8 gim. Figure 1 shows two generations of samples after release; the modified sample, on the right, has a more robust pivot and larger gage-to-ring fillets. The first generation samples tended to fail at regions outside of the gage length; statistical and post-mortem analysis of those samples indicated multiple failure modes with about 20% breaking in the gage section. The samples are free to rotate about the pivot; the bumpers in the later design limit the range of travel; the samples still have to be aligned in the proper position with a probe tip before testing. The samples and recorded force-displacement data are analyzed after testing to calculate the stress-strain response and to identify the sample failure mode.
Data Analysis Raw load-displacement data from samples of multiple lengths, shown in Figure 2, is processed in two steps to evaluate the strength and modulus of the material. First, a force balance is solved to correct for the losses due to the frictional sliding of the tip along the substrate. The tip is a flat-bottomed conical diamond with an included angle, 20, of 59'. The normal force, N, and lateral force, L, are recorded and must be resolved into the tensile force on the sample, T, the reaction due to the tip angle normal to the surface, Nr, and the frictional losses,
Lf: Lf =/J.N, L=T+Lf
(1) (2)
N N, +N1
(3)
anT50 tanO
(4)
N,
1
exp =mahine
1
(5)
0
-9
20000 -
The measured displacement must be corrected for machine compliance. The total measured compliance is the sum of the machine compliance (including grips, etc) and the compliance from the tensile gage length. Testing samples with identical cross sectional dimensions and several lengths provides a simple and accurate way to evaluate the machine compliance. 1
30000 25000
,Z 15000 100o00:
300
--
1 U: i}
ii -
-
5000 0: -5000
0
5
10
15
Displacement 4m) Figure 2. Load-displacement curves from samples with gage lengths of 15 to 1000 jim
long.
Kmchiie is evaluated from the intercept of
Data Loading...
