Research Papers

Thermal and Structural Analysis of a Suspended Physics Package for a Chip-Scale Atomic Clock

[+] Author and Article Information
A. D. Laws

Department of Mechanical Engineering, University of Colorado, Boulder, CO 80309alexander.laws@colorado.edu

R. Borwick

 Teledyne Scientific Company, Thousand Oaks, CA 91360rborwick@teledyne.com

P. Stupar

 Teledyne Scientific Company, Thousand Oaks, CA 91360pstupar@teledyne.com

J. DeNatale

 Teledyne Scientific Company, Thousand Oaks, CA 91360jdenatale@teledyne.com

Y. C. Lee

Department of Mechanical Engineering, University of Colorado, Boulder, CO 80309leeyc@colorado.edu

J. Electron. Packag 131(4), 041005 (Oct 21, 2009) (9 pages) doi:10.1115/1.4000211 History: Received July 31, 2008; Revised June 03, 2009; Published October 21, 2009

The power dissipation for chip-scale atomic clocks (CSAC) is one of the major design considerations. 12 mW of the 30 mW power budget is for temperature control of the vertical-cavity-surface-emitting laser (VCSEL) and the alkali-metal vapor cell. Each of these must be maintained at 70+/0.1°C even over large ambient temperature variations of 050°C. Thus the physics package of a CSAC device, which contains the vapor cell, VCSEL, and optical components, must have a very high thermal resistance, greater than 5.83°C/mW, to operate in 0°C ambient temperatures while dissipating less than 12 mW of power for heating. To create such a high level of insulation, the physics package is enclosed in a gold coated vacuum package and is suspended on a specially designed structure made from Cirlex, a type of polyimide. The thermal performance of the suspended physics package has been evaluated by measuring the total thermal resistance from a mockup package with and without an enclosure. Without an enclosure, the thermal resistance was found to be 1.07°C/mW. With the enclosure, the resistance increases to 1.71°C/mW. These two cases were modeled using finite element analysis (FEA), the results of which were found to match well with experimental measurements. A FEA model of the real design of the enclosed and suspended physics package was then modeled and was found to have a thermal resistance of 6.28°C/mW, which meets the project requirements of greater than 5.83°C/mW. The structural performance of the physics package was measured by shock-testing, a physics package mockup and recording the response with a high-speed video camera. The shock tests were modeled using dynamic FEA and were found to match well with the displacement measurements. A FEA model of the final design, not the mockup, of the physics package was created and was used to predict that the physics package will survive a 1800 g shock of any duration in any direction without exceeding the Cirlex yield stress of 49 MPa. In addition, the package will survive a 10,000 g shock of any duration in any direction without exceeding the Cirlex tensile stress of 229 MPa.

Copyright © 2009 by American Society of Mechanical Engineers
Your Session has timed out. Please sign back in to continue.



Grahic Jump Location
Figure 1

Chip-scale atomic clock physics package designed by scientists at Teledyne Scientific Co. The cell and VCSEL in this assembly must be maintained at 70°C using heaters on either side of the cell with less than 12 mW of power. The physics assembly is suspended on a specially designed Cirlex suspension to limit conductive heat losses, coated in low emissivity coatings to limit radiation losses, and packaged in a vacuum to limit convective losses.

Grahic Jump Location
Figure 2

(a) Suspension design to minimize the physics package volume. (b) The suspensions are made from two identical 30 mil thick Cirlex 3000 CL parts. Notice that this design will tend to unwind when a vertical force is applied. To stiffen the structure, the top suspension is reversed so that the two suspensions unwind against each other and utilize the strength of the solid physics package.

Grahic Jump Location
Figure 3

Solid model of the mockup physics package provided to the University of Colorado by Teledyne Scientific Co. The thermistor was attached, using thermally conductive adhesive, to the cell in order to heat and sense the temperature of the package.

Grahic Jump Location
Figure 4

Mockup physics package suspended on two Cirlex pieces and mounted in a polished copper enclosure. A thermistor mounted to the side of the physics package was used to heat the physics package and sense the temperature of the package.

Grahic Jump Location
Figure 5

Power dissipated by the thermistor mounted to the mockup physics package versus the temperature difference created between the thermistor and room temperature. The thermal resistances are calculated for maximum power points by dividing the temperature difference by the power and are found to be 1.07°C/m W with no enclosure, 1.35°C/m W for one of the enclosed package tests, and 1.71°C/m W for the other enclosed package test.

Grahic Jump Location
Figure 6

3D thermal finite element models created to simulate the two experimental cases: (a) without an enclosure and (b) with a polished copper enclosure.

Grahic Jump Location
Figure 7

Typical temperature contours for the two modeled experimental cases: (a) no enclosure and (b) polished copper enclosure. The resistance of the thermally conductive adhesive allows a large temperature gradient between the thermistor and the package.

Grahic Jump Location
Figure 8

3D thermal finite element model of the design case physics package and suspension. Low emissivity coatings are used to minimize radiation. Vacuum packaging eliminates convection. Conduction occurs through the suspension and through wires on the arms of the bottom suspension. Heat is applied on either side of the cell using clear resistance heaters. The overall thermal resistance is 6.28°C/m W, which meets the project goal.

Grahic Jump Location
Figure 9

Mockup physics package built by Teledyne Scientific Co., mounted between two 30 mil thick Cirlex suspensions.

Grahic Jump Location
Figure 10

Four frames taken from high-speed video of the mockup physics package, shot at 6000 fps, during a 2000 g, 0.32 ms vertical impact. Frame 0 is highlighted to show the physics package, microruler, and clamping assembly. The divisions on the microruler are 640 μm from center to center.

Grahic Jump Location
Figure 11

Displacement measurements made from a high-speed video clip of a 2000 g, 0.32 ms vertical impact of the mockup physics package. The time between frames is 0.0303 ms and 1 mm on screen is equal to 11.8±0.2 μm. By measuring the displacement of the ruler, fixed to the impact table, with respect to the frame and the displacement of the package with respect to the frame, the displacement of the package with respect to the impact table can be found.

Grahic Jump Location
Figure 12

Dynamic finite element models created to study the (a) mockup physics package subjected to vertical and horizontal shocks and the (b) final design physics package and suspension subjected to vertical shocks, horizontal shocks, and modal analysis.

Grahic Jump Location
Figure 13

Displacement of the suspended mockup physics package, subjected to a 2000 g, 0.32 ms shock, measured from high-speed video footage compared with a FEA model. The uncertainty in the video measurements corresponds to +/−1.5 mm measured on screen, which corresponds to +/−0.0177 mm of displacement.

Grahic Jump Location
Figure 14

Typical stress distributions for (a) vertical or (b) horizontal accelerations of the suspended mockup physics package. In vertical loading all eight suspension beams are loaded as fixed-fixed bending beams. The maximum stresses occur at the beam-ends. In horizontal loading, two beams are loaded under tension, two beams are loaded in compression and the other four are loaded in bending. The maximum stress occurs at the neck of the beams that are in tension.

Grahic Jump Location
Figure 15

Contours of the Von Mises stresses developed in the structure for either (a) vertical accelerations or (b) horizontal accelerations. For similar amplitude accelerations the stress is higher in the vertical case because the beams are in bending as opposed to tension in the horizontal case.

Grahic Jump Location
Figure 16

Radius at stress concentration point was increased to reduce stress caused by a 1500 g, 0.5 ms shock. The maximum stress must be below the static yield stress of Cirlex σy=49 MPa. This change will have a negligible impact on the total thermal resistance and first natural frequency of the package.




Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In