Proceedings of the ASME 2011 International Mechanical Engineering Congress & Exposition IMECE2011 November 11-17, 2011, Denver, Colorado, USA
IMECE2011-65001
Performance Degradation of the MEMS Vibratory Gyroscope in Harsh Environments 1
Chandradip Patel and 2Patrick McCluskey 1,2
Department of Mechanical Engineering A. James Clark School of Engineering University of Maryland, College Park, MD, USA, 20740 Phone: (301) 405-0279 1 Email:
[email protected] 2 Email:
[email protected]
ABSTRACT The use of MEMS gyroscopes in a wide range of applications requiring then to function from medium to harsh environments make it necessary to evaluate the performance of MEMS gyroscopes under those conditions. This paper focuses on the effects of elevated temperature and humidity on the performance of MEMS vibratory gyroscopes. Performance of the MEMS gyroscope was evaluated by conducting Highly Accelerated Stress Testing (HAST) on a COTS (commercialoff-the-shelf) single axis MEMS vibratory gyroscope having an operating temperature range from -40˚C to +105˚C. The gyroscope sensors were exposed to 130°C and 85% relative humidity with a pressure of 33.3 psia or 230 kPa for 96 hours. Pre-baking and post-baking tests were conducted before and after HAST at 125˚C for 24 hours respectively. Also, stationary baseline testing (SBT) and rotary baseline testing (RBT) were performed before and after the pre-baking, HAST and post-baking tests to measure any permanent shift during the respective test. A preliminary result shows that the MEMS gyroscope output degraded in the pre-baking test and HAST; while it showed a recovery in post-baking test. After completing the entire test procedure, it was observed that MEMS gyroscope output didn’t come back to the original position, and resulted in a permanent output shift of 1.85deg/s.
INTRODUCTION MEMS vibratory gyroscopes are increasingly used in applications ranging from consumer electronics to aerospace and are now one of the most common MEMS products after accelerometers [1][2]. Despite their widespread use, the performance of MEMS gyroscopes in harsh environments is still under question. While some studies have been conducted to understand the temperature dependent performance of MEMS gyroscopes, the effects of sustained exposure to temperature combined with other harsh environment stresses have not been well researched [3][4][5][6]. Thus, it is necessary to quantify MEMS vibratory gyroscope performance under such conditions. A MEMS gyroscope is a sensor that measures the rate of change in the angular position of an object. The majority of the reported MEMS gyroscopes use vibrating mechanical elements to sense angular velocity [7]. A MEMS vibratory gyroscope can be simply visualized as a two degree-offreedom spring-mass-damper system as shown in Figure 1. The proof mass is suspended above the substrate by use of flexible beams which also work as mechanical springs. In most of the work on reported MEMS vibratory gyroscopes, the proof mass is subjected to vibration at resonance frequency by use of electrostatic force causing movement in the drive direction. When the gyroscope sensor experiences an angular rotation, a Coriolis force is induced in the direction orthogonal to both drive direction (x) and angular rotation axis (z). This rotation induced Coriolis force causes energy transfer between drive mode and sense mode. The proof mass movement caused by Coriolis force in the sense direction (y) is proportional to the angular rotation applied and can be measured with differential capacitance techniques by use of interdigitated comb electrodes.
KEY WORDS: Angular velocity bias, Single axis MEMS vibratory gyroscope, Highly Accelerated Stress Test (HAST), Zero rate output (ZRO). NOMENCLATURE ˚C Degree Celsius ˚/s Degree per Second HAST Highly Accelerated Stress Test SBT Stationary Baseline Test RBT Rotary Baseline Test RH Relative Humidity ZRO Zero Rate Output 1
Copyright © 2011 by ASME
Figure 3: Two proof mass-spring-damper system. One of the specific applications of these gyroscopes is for inertial navigation and tracking where single axis MEMS vibratory gyroscopes are combined and mounted on three Cartesian directions (i.e. X, Y and Z) with other electronics. This tracking system requires robust and accurate gyroscopes, the performance of which should be maintained within specific desired range. The unit tested in this study is a single axis MEMS gyroscope which is used to develop an inertial navigation and tracking unit, such as the unit is shown Figure 4.
Figure 1: Schematic illustration of MEMS Vibratory Gyroscope. Typical drive and sense mode resonant frequencies range from 10-20 kHz, and sense amplitude is an order of magnitude less than drive amplitude causing a small change in gyroscope parameters leads to large variation in device output. Thus, it is necessary to evaluate the effects of harsh environment stresses on the performance of MEMS vibratory gyroscopes. In this study, we have examined the combined effects of severe temperature and humidity conditions on the performance of MEMS vibratory gyroscope. The device used in this study had a manufacturer’s suggested operating temperature range from -40˚C to +105˚C with bias temperature coefficient of 0.005˚/sec/˚C within that range. It was fabricated from polysilicon material by surface micromachining processes. The MEMS wafer was packaged in a cavity package and sealed by use of a metal cap. The simple construction of the gyroscope package is shown schematically in Figure 2.
Figure 4: Navigation and Tracking Unit EXPERIMENTAL SET-UP AND PROCEDURE It is shown that the effects of severe temperature and humidity on semiconductor device can be rapidly determined by performing Highly-Accelerated Temperature and Humidity Stress Test (HAST). This accelerated test indicates potential gross failure sites in a relatively short time span compared to Temperature Humidity Bias (THB) test [8]. Figure 2: MEMS gyroscope package construction In order to conduct highly accelerated stress test (HAST), currently, there is no define test guideline or standard specifically for MEMS application. Every MEMS manufacturer follows their own defined test procedure [9] [10]. Thus, we used the available information and established our own test procedure which is followed in this study as shown below.
The device has two proof masses placed on either side that are driven in opposite directions causing the Coriolis force induced on the two masses to be in opposite directions as shown in Figure 3. This special arrangement helps to nullify the external inertial inputs caused by undesirable ambient vibration and shock.
2
Copyright © 2011 by ASME
Stationary Baseline Test (SBT) & Rotary Baseline Test (RBT) [Room Condition for 5 min]
(measurement) time within 10 to 15 minutes. This is to make sure that package internal condition doesn’t change during baseline testing.
Pre-Baking test [125˚C for 24 hours]
Stationary Baseline Test (SBT) & Rotary Baseline Test (RBT) [Room Condition for 5 min]
Figure 5: Stationary Baseline Test
HAST [130˚C, 85% RH, 230 kPa, bias for 96 hours]
Stationary Baseline Test (SBT) & Rotary Baseline Test (RBT) [Room Condition for 5 min]
Figure 6: Rotary Baseline Test
Post-Baking test [125˚C for 24 hours]
Stationary Baseline Test (SBT) & Rotary Baseline Test (RBT) [Room Condition for 5 min] The primary purpose of conducting the stationary baseline (SBT) and the rotary baseline (RBT) tests was to observe the device characteristic before and after baking and HAST tests to allow for later comparison. The SBT and RBT were performed by placing the MEMS gyroscope on stationary table and in-house fabricated rotary table respectively as shown in Figure 5 and Figure 6. In both baseline tests, MEMS gyroscope output was recorded with help of palmtop. During rotary baseline test, rotary table speed was kept at 60deg/s (10 RPM). Both the baseline tests were conducted for 5 minutes at room condition with data acquisition rate of 40 Hz. It is essential to notice that the SBT and RBT performed after baking test and HAST, care was exercise to keep the test
Figure 7: Highly Accelerated Stress Test Setup The HAST test was performed for 96 hours excluding 30 minutes each for ramp up and ramp down. The unit was placed inside the HAST chamber and connected to internal terminals for bias and data acquisition purpose. The external terminals were connected to an evaluation board which was 3 Copyright © 2011 by ASME
The summary of all sub-test, degradation or recovery pattern and their values are summarized in Table 1. After accumulating the effects of pre-baking, HAST and postbaking, it can be concluded that MEMS gyroscope had a permanent degradation of about 1.85deg/s.
further connected to desktop PC that collected gyroscope insitu data at 1 Hz. The HAST setup is shown in Figure 7. Both, pre-baking and post-baking test were performed by placing the unit in a thermal convection chamber at 125˚C for 24 hours. The unit was kept unbiased during these tests. RESULTS AND ANALYSIS The test was completed by following the entire test procedure as shown in a flow chart previously. As mentioned earlier, the SBT and RBT were performed to measure any permanent change caused by particular subtests like pre-baking, postbaking test or HAST. Results of SBT, RBT and their trend are shown in Figure 8, Figure 9 and Figure 10 respectively. Both SBT and RBT results were plotted after minimizing noise by plotting a 50 sample moving average. The offset between the lines in Figure 8 and Figure 9 represents a permanent or nonrecoverable shift in MEMS gyroscope output that resulted from respective test condition. The SBT and RBT trend plot, Figure 10, was established by taking average of entire raw data collected for 5 minutes for each sub SBT and RBT test.
Sub-Test
Degradation or Recovery
Value (Deg/s)
Pre-Baking
Degradation
-1.04
HAST
Degradation
-1.46
Post-Baking
Recovery
+0.65
Overall
Degradation
-1.85
Table 1: Summary of the sub-tests.
Stationary Baseline Test (SBT) Angular Velocity (Deg/s)
1
After performing initial SBT and RBT, pre-baking was conducted at 125˚C for 24 hours. The pre-baking was done at unbiased condition, so the effect of pre-baking on sample’s output (in-situ data) was not monitored. But, SBT and RBT conducted after pre-baking test revealed that pre-baking had caused non-recoverable or permanent degradation in gyroscope output of about 1.04deg/s from its initial output as shown in Figure 10. After completing SBT and RBT after pre-baking test, the sample was placed in the HAST chamber, and connected to internal and external terminals of the chamber for in-situ monitoring as shown in Figure 7. The in-situ HAST result, Figure 11, show that gyroscope output degraded from 0.68deg/s to -3.2deg/s during 96 hours. As test finished after 96 hours, the internal temperature, humidity and pressure of HAST chamber dropped from 130 ˚C, 85%RH and 230 kPa to ambient condition. Due to this change, a small recovery of 0.9deg/s was observed as shown in Figure 11 after 96 hours. So, the SBT and RBT conducted after HAST revealed that HAST had caused a shift in gyroscope output of about 1.46deg/s in-total as shown in Figure 10 and Figure 11. This degradation clearly shows the effects of severe temperature and humidity ingress on MEMS gyroscope which resulted into degradation. The potential cause of this degradation could be due to humidity penetration in the package that affects either MEMS die or electronic conditioning die, or it could be due to degradation of packaging materials due to high temperature exposure.
0.5 Initial SBT After Pre-Baking After HAST After Post-Baking
0
-0.5 -1
-1.5 -2
-2.5 0
1
2
3
4
5
Time (Mins) Figure 8: Stationary Baseline Test result at room condition and zero angular rotation.
Rotary Baseline Test (RBT) Angular Velocity (Deg/s)
61
The post-baking test was conducted similarly as pre-baking test. The SBT and RBT conducted after post-baking test revealed that post-baking had also caused an additional recovery in gyroscope output about 0.65deg/s as shown in Figure 10. This could be results of moisture egress from gyroscope package cavity caused by post-baking effect.
60.5 Initial RBT After Pre-Baking After HAST After Post-Baking
60 59.5 59 58.5 58 57.5 0.1
1
2
3
4
Time (Mins) Figure 9: Rotary Baseline Test at room condition and 60 deg/s rotary table angular rotation. 4
Copyright © 2011 by ASME
5
REFERENCES [1] N. Yazdi, F. Ayazi, K. Najafi "Micromachined inertial sensors," Proceedings of the IEEE, vol.86, no.8, Pages.1640-1659, Aug-1998. [2] S. Nasiri "A Critical Review of MEMS Gyroscopes Technology and Commercialization Status" InvenSense, Inc. [3] S. H. Choa "Reliability of vacuum packaged MEMS gyroscopes" Microelectronics and Reliability, Volume 45, Issue 2, Pages 361-369, February 2005. [4] C. Patel, F.P McCluskey, D. Lemus, A. Jones, J. Davis " The Temperature Effects on Performance and Reliability of MEMS Gyroscope Sensors" InterPACK'09, The Pacific Rim/ASME International Electronic Packaging Technical Conference, San Francisco, CA,USA July 1923, 2009.
Figure 10: SBT and RBT Trend for comparison.
In-Situ HAST Result 0.5 Die Temperature
Gyro Output
120
100 -0.5 80
-1
-1.5
60
-2 40 -2.5
Temperature (Deg C)
Angular Velocity (Deg/s)
0
20
0
10
20
30
40
50
60
70
80
90
[6] F.P McCluskey, C. Patel, D. Lemus "Performance and Reliability of MEMS Gyroscopes and Packaging at High Temperatures", High Temperature Electronic (HiTEC), May 2010. [7] C. Acar and A.M. Shkel “MEMS Vibratory Gyroscopes Structural Approaches to Improve Robustness”. MEMS Reference Shelf Series, Springer. 2009.
-3
-3.5
[5] C. Patel, F.P McCluskey, D. Lemus "Performance and reliability of MEMS gyroscopes at high temperatures," Thermal and Thermomechanical Phenomena in Electronic Systems (ITherm), 2010 12th IEEE Intersociety Conference, vol.1, no.5, Pages2-5, June 2010.
0 100
Time (Hrs) Figure 11: In-Situ Highly Accelerated Stress Test (HAST) gyro output and die temperature at 130°C, 85% RH and 230 kPa pressure.
[8] Highly-Accelerated Temperature and Humidity Stress Test (HAST), EIA/JEDEC STANDARD, JESD22-A110B. [9] Analog devices online reliability handbook.
SUMMARY AND CONCLUSIONS This paper shows the effects of both temperature and humidity by conducting tests like pre-baking, HAST and post-baking. During each sub-test, MEMS gyroscope output resulted into permanent or non-recoverable degradation. After completing entire test procedure, it was found that MEMS gyroscope had a permanent degradation of 1.85deg/s. It was hypothesized that the dominant cause of this permanent degradation was due to ingress and egress of moisture through the cavity package or degradation of packaging materials due to high temperature exposure. Thus, the effects of temperature and humidity can be minimized by selecting better packaging material.
[10] S. Nasiri, M Lim, and M. Housholde,” A Critical Review of the Market Status and Industry Challenges of Producing Consumer Grade MEMS Gyroscopes”, Invensense Inc.
ACKNOWLEDGEMENT The authors would like to acknowledge the financial and technical support of the Maryland Industrial Partnership Program and TRX Systems, Inc. The authors would also like to thank the CALCE Electronic Products and Systems Center for the use of their laboratories in the conduct in this work. 5
Copyright © 2011 by ASME