reliability assessment of a mems microphone under mixed flowing gas environment and shock impact...

Microelectronics Reliability xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Microelectronics Reliability

journal homepage: www.elsevier .com/locate /microrel

Reliability assessment of a MEMS microphone under mixed flowing gasenvironment and shock impact loading

0026-2714/$ - see front matter � 2014 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.microrel.2014.01.003

⇑ Corresponding author.E-mail address: [email protected] (J. Li).

Please cite this article in press as: Li J et al. Reliability assessment of a MEMS microphone under mixed flowing gas environment and shock impact lMicroelectron Reliab (2014), http://dx.doi.org/10.1016/j.microrel.2014.01.003

Jue Li ⇑, Mikael Broas, Jani Raami, Toni T. Mattila, Mervi Paulasto-KröckelElectronics Integration and Reliability, Department of Electrical Engineering and Automation, Aalto University School of Electrical Engineering, P.O.B. 13340, FIN-00076 Aalto, Finland

a r t i c l e i n f o

Article history:Received 24 June 2013Received in revised form 20 November 2013Accepted 2 January 2014Available online xxxx

a b s t r a c t

In this work the reliability of a Micro-Electro-Mechanical Systems (MEMS) microphone is studiedthrough two accelerated life tests, mixed flowing gas (MFG) testing and shock impact testing. The objec-tive is to identify the associated failure mechanisms and improve the reliability of MEMS devices. Failureanalyses are carried out by using various tools, such as optical microscopy, scanning electron microscopy(SEM), energy dispersive X-ray spectrometry (EDS). Finite element analysis is also conducted to study thecomplex contact behaviors among the MEMS elements during shock impact testing. The predicted failuresites are in agreement with the experimental findings.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

The applications of Micro-Electro-Mechanical Systems areincreasing rapidly, especially in the consumer electronics. MEMSaccelerometers, gyroscopes and microphones have already beenused in various portable electronic products, such as mobilephones, tablets, cameras and game controllers [1–4]. MEMS micro-phones are widely used in portable electronic devices due to theirsmall size, better acoustic performance, and compatibility withautomated assembly process.

Some MEMS devices like gyroscopes need to be hermeticallypackaged and their reliability is highly dependent on the qualityof hermetic bonding [5–8]. Different from hermetically packageddevices the sensing element of MEMS microphones are commonlyexposed to temperature changes, humidity and corrosive gasesduring field use [9–11]. The humidity and corrosive gases usuallycause wire bond corrosion and failures [12] and oxidation-inducedstress corrosion in polycrystalline silicon thin film membranes[13,14]. To assess the reliability of electronics exposed to variousenvironmental conditions mixed flowing gas tests are often used[15]. For instance, Zhang et al. studied the influence of H2S expo-sure on immersion-silver-finished PCBs via MFG testing [16]; Yanget al. used MFG tests to assess the reliability of land grid arraysockets [17]; Zhao et al. investigated the reliability of Ni/Pd/Au–Pd and Ni/Pd/Au–Ag pre-plated leadframe packages by usingMFG tests [18]. So far, however, there have been no reports onthe MFG tests of MEMS microphone or similar thin film structures.

In addition, MEMS microphones experience vibrations, shockimpacts or occasional drops during normal operations. The shockimpact reliability of MEMS microphones is crucial to the portable

electronic devices. The acceleration experienced by a mobile phoneduring normal operating environment is around several g (‘g’ is theEarth’s gravitational acceleration). However, a mobile phone acci-dentally falling from a height of about 1.2 m can experience aroughly 500–5000g shock load as measured on the printed wiringboard (PWB) [19,20]. The peak acceleration values vary with differ-ent impact surfaces, and orientations of collisions. Some recentinvestigations report that the maximum acceleration at the impactsite of the secondary impacts can be much higher than those of theprimary impact [21,22]. As a result, MEMS microphones mountedon the surfaces of PWBs need to be shock-resistant in order to en-sure the expected lifetimes of the electronic products. Besides, thereliability studies should be extended to high shock levels for thesake of a sufficient margin of safety.

In this work, the reliability of a MEMS microphone under harshenvironment and shock impact loading is studied by experimentsand simulations. The paper is organized as follows. Firstly, the de-tails of the experimental setup, failure analysis, and finite elementanalysis are reported. Secondly, the observed failure modes arepresented followed by simulation results and discussions of thefailure mechanisms. Thirdly and finally, conclusions drawn fromthe reliability assessment are given.

2. Materials and methods

2.1. MEMS microphone

In the package of the MEMS microphone, the MEMS and theapplication specific integrated circuit (ASIC) are placed side by sideon a laminate substrate, connected electrically with wire bondingand sealed under a metal lid. Fig. 1(a) shows an SEM image ofthe MEMS microphone with the lid removed. The MEMS micro-phone is based on a typical parallel capacitor membrane design

oading.

http://dx.doi.org/10.1016/j.microrel.2014.01.003

mailto:[email protected]


http://www.sciencedirect.com/science/journal/00262714

http://www.elsevier.com/locate/microrel


Fig. 1. (a) SEM image of the MEMS microphone without lid. (b) Schematic drawingof the MEMS structure cut along the diagonal line; backplate, diaphragm, and bulksilicon are indicated by the arrows, respectively. (c) Enlarged view of the diaphragmrunner.

2 J. Li et al. / Microelectronics Reliability xxx (2014) xxx–xxx

[23] and fabricated by surface micromachining techniques [24]. Itssensing element consists of two polycrystalline silicon thin filmmembranes, namely a backplate and a diaphragm (see Fig. 1(b)).As shown in Fig. 1(c), the diaphragm is only connected to the basestructure via a beam-like structure, called ‘diaphragm runner’. Thefree space below the membranes is the acoustic chamber, which isexposed to external air flow through a hole. The diaphragm vi-brates due to the air pressure changes, resulting in the capacitancechange between the two electrodes.

2.2. Experimental setup

There were in total thirty-two MEMS microphone samplestested by the mixed flowing gas test. The test board, corrosion test

Fig. 2. (a) Test board with two MEMS microphones. (b) Corrosion test chamber. (c)Jig and shock impact orientations. (d) Pneumatic shock impact tester.

Table 1Parameters used in the mixed flowing gas test and typical Nordic outdoor environments [

Environment �C % RH Cl2 (lg/m3)

Test chamber 30 70 19Outdoor (Nordic) 5 78 0.9

Please cite this article in press as: Li J et al. Reliability assessment of a MEMS miMicroelectron Reliab (2014), http://dx.doi.org/10.1016/j.microrel.2014.01.003

chamber, aluminium jig, and pneumatic shock impact tester arepresented in Fig. 2. The MFG test was carried out in the corrosiontest chamber according to the Telcordia Technologies GR-63-COREstandard [25,26], an industrial standard for telecommunicationsequipment and systems. The volume exchange rate and the airflow rate were set to be 8 times per hour and 28.7 l per minute,respectively. The relatively humidity (RH) and concentration ofgaseous pollutants were monitored at the inlet tubes. Table 1 givesthe temperature, relative humidity, and concentration of gaseouspollutants used in MFG test. The MFG test is a highly acceleratedtest as the concentrations of gaseous pollutants in the test chamberare about 4–50 times higher than those in typical Nordic outdoorenvironments (presented in Table 1).

The MFG test setup and monitoring system are the same aswhat have been reported in Ref. [27]. As shown in Fig. 2(b), a loud-speaker is placed inside of the test chamber to monitor the healthof the microphones. During the monitoring process, firstly a knownacoustic stimulus (a logarithmic sine sweep over the specified fre-quency range of the microphone) is produced by the loudspeaker,secondly the microphone records the signal and the recorded sig-nal is send back to the computer, and thirdly the control programcompares the magnitude frequency response of measured signalwith a reference sample. The similar monitoring system can alsobe used for other reliability tests.

For the shock impact tests, there were in total twenty-four sam-ples tested in four different impact orientations (X, Y, Z+ and Z�),i.e. six samples for each orientation. The pneumatic shock impacttester shown in Fig. 2(d) is capable of producing shock impact lev-els up to 100,000g. The PWB (34 � 30 � 1.7 mm3) is made fromaluminium alloy 7075 in order to minimize the influence of thePWB motion on dynamic response of the MEMS microphone undershock impacts. The test boards can be attached to the jig in differ-ent orientations, X, Y, Z+ and Z�. The finite element simulation wasemployed to help design the jig, which can survive shock impactsup to 65,000g with negligible plastic deformation. The MEMSmicrophones were in a power-off state during the shock impactsand their state of health was tested by a health monitoring systemas previously discussed after every five impacts. During the shockimpacts the acceleration histories were measured by a lightweightaccelerometer mounted on the top of the jig. These measuredacceleration pulses were used as the loading conditions in the fi-nite element simulations.

2.3. Finite element model setup

The finite element analyses were conducted by employing thefinite element software, ABAQUS v.6.9-EF. The structural detailsin the MEMS chip are much smaller in scale than the componentboard, and therefore, a submodeling approach was used in orderto achieve both accuracy and efficiency. In the global model thecomponent board with two MEMS microphones was considered(see Fig. 3). A cross-section view of the MEMS microphone in theglobal model is presented in Fig. 3(c) where most of the structuraldetails in the MEMS chip are simplified and the wire bonds are notconsidered either. The focus of this simulation is on the reliabilityof the MEMS chip since the reliability of the wire bonds has beenpreviously investigated [21,22].

As shown in Fig. 4, the submodel with fine mesh was establishedon the level of the MEMS chip to build a realistic representative of

25,26].

H2S (lg/m3) NO2 (lg/m3) SO2 (lg/m3)

262 188 1364.6 28 30

crophone under mixed flowing gas environment and shock impact loading.


Fig. 3. Details of the global model. (a) Test board with two MEMS microphones. (b)Model of one MEMS microphone. (c) Cross-section view of the MEMS microphone.

Fig. 4. Challenges in the MEMS microphone simulation. (a) Top view of the MEMSstructure. (b) SEM image showing the thickness changes in the backplatemembrane. (c) Structural details in the backplate membrane or in the vicinity ofit. (d) Part of the model meshed with shell elements from the finite elementsubmodel.

J. Li et al. / Microelectronics Reliability xxx (2014) xxx–xxx 3

the real structure. However, there are still some structural detailsthat need to be simplified with special treatments before a soundfinite element model can be built. As shown in Fig. 4(b) and (c),the challenges include the thickness changes of the porous back-plate membrane, the distance stoppers, the large number of fingers,and the bowed edge of the backplate. The details of the modelingprocedure are addressed as follows. Shell elements were used formeshing the backplate membrane and thickness changes were con-sidered in the shell section definition. The distance stoppers de-signed on the fringe of the backplate membrane cannot beneglected as they are crucial for the interaction between the two


membranes under shock impact. The geometric feature of the dis-tance stoppers was simplified by the rectangular shell elementsand the height of the distance stopper was included in the shell sec-tion definition (see Fig. 4(d)). The bowed edge is responsible fortransferring boundary conditions from the base to the backplatemembrane, and therefore this feather was remained in the modeland meshed with solid elements. Furthermore, the porous back-plate membrane was modeled as a homogeneous material with cal-culated equivalent material properties. More details of theequivalent material properties calculation and the finite elementmodel setup can be found in Ref. [28]. The calculated equivalentmaterial properties as well as other material properties used inthe simulation are listed in Table 2.

3. Results and discussion

The experimental and simulation results are presented in thefollowing sections.

3.1. Mixed flowing gas test failures

The MEMS microphone samples were removed from the corro-sion chamber after 90 days testing. No obvious failures were de-tected by the health monitoring system except for a slightdecrease (about 0.5 decibel difference) in the magnitude frequencyresponse. After careful sample preparation and failure analysis,two distinctive failure modes were observed, namely wire bondcorrosion and membrane brittle fracture. Wire bond corrosionwas found in every sample experienced 90 days MFG test. Six outof thirty-two samples showed membrane fracture after the metal-lic lid was removed for inspection.

3.1.1. Wire bond corrosionFirst of all, the composition of the wire bond was investigated

and found to be gold wire ball on top of gold, nickel and chromium.The composition of the wire bonds was measured with energy dis-persive X-ray spectrometry as shown in Fig. 5. The Au/Au bondinginterface and Ni/Au deposition interface can also be seen in theSEM micrograph.

Further investigation reveals that the wire bonds were detachedfrom the pads and the corrosion products had grown at the Au/Auinterface. Fig. 6(a) presents an SEM image of a detached wire bondas well as the corrosion products. In Fig. 6(b), the EDS analysis of atested wire bond shows that the nickel had diffused verticallythrough the pad metallization and reacted with the gaseous pollu-tants in the corrosion chamber. The nickel diffusion and the growthof the corrosion products in volume are believed to be the reasoncausing the wire bond detaching. The composition of the corrosionproducts was found to be nickel, sulfur, and oxygen. Unfortunately,the accurate composition ratio cannot be measured as EDS is notaccurate enough for detecting lighter elements such as oxygen.In order to improve the reliability, Ni/Au deposit should be re-placed by Ni/Pd/Au deposit so that palladium can be the diffusionbarrier for nickel.

3.1.2. Membrane brittle fractureAs shown in Fig. 7, cracks were observed in both diaphragm and

backplate membranes after removing the metallic lid of the micro-phone samples, which had experienced 90 days MFG test. Since nosignificant electrical performance changes as monitored by thehealth monitoring system, the cracking of the membranes proba-bly took place during the sample handling, e.g. taking the samplesout of the test chamber or opening the lid for failure analysis.Membrane embrittlement is believed to be the cause of the crack-ing. The explanation of the observed fractures would be the stresscorrosion cracking of surface thin film [5].



Table 2Material properties used in the simulation [29,30].

Material Young’s modulus (GPa) Poisson’s ratio Density (g/cm3)

Polysilicon 158 0.22 2.33Porous membrane (calculated) 48.93 0.33 1.39Single-crystal silicon (orthotropic) E1 = E2 = 169 G13 = G23 = 79.6 c12 = 0.064 2.33

E3 = 130 G12 = 50.9 c13 = 0.28c23 = 0.36

SnAgCu 52.4 0.37 7.2

Fig. 5. (a) SEM image of a virgin wire bond. (b) Composition of the wire bonds measured by EDS.

Fig. 6. (a) SEM image of a detached wire bond which had experienced 90 days mixed flowing gas test. (b) A cross-section image of a tested wire bond and EDS analysis.


The embrittlement of the silicon membranes is probably associ-ated with the growth of surface SiO2 thin film. The surface of siliconis usually covered by a thin oxide layer, namely silicon dioxide(SiO2) layer. The growth of the SiO2 layer is highly dependent on


the temperature and the exposure time to air as the process is dif-fusion controlled. The oxide growth can also be accelerated by thehumid environments as reported in [24]. The membrane embrittle-ment might be caused by the environmentally-assisted oxide layer



Fig. 7. (a) Top view of a fractured diaphragm membrane. (b) Bottom view of thesame sample in (a). (c) Top view of a microphone sample with fractured diaphragmand backplate membranes. (d) Top view of a microphone same with fracture only inthe backplate membrane.

Fig. 8. Micrographs of the failed MEMS microphones. (a) The whole diaphragmmembrane was missing. (b) most of the diaphragm was cracked. (c) partialdiaphragm was cracked and (d) the cracking backplate and runner.

Fig. 9. Six snapshots of the strain distributions of the deformed diaphragm anddiaphragm runner under Z-orientation shock impact with 65 kG peak acceleration.


growth and the segregation of impurities to the grain boundaries.Muhlstein et al. reported a failure mechanism of the surface oxidethin film (termed reaction-layer fatigue) in Ref. [13]. In their study,the endurance strength of their polycrystalline silicon samples wasfound to be around 2 GPa, i.e. roughly half of the fracture strength.However, the reaction-layer fatigue cannot be employed to explainthe observed fractures in the current study since the stress level inthe MEMS microphone during MFG testing or sample handlingwould never exceed 2 GPa. Compared to the test environment in[13], the situation in the mixed flowing gas test is more complicatedas the combined effect of the humidity and the gaseous pollutants isstill unclear.

3.2. Shock impact acceleration tolerance

The shock impact tests were carried out in all the impact orien-tations with peak accelerations up to 80,000g. No failures werefound in tests with X and Y impact orientations, and only failuresin the Z+ and Z� orientations were detected under the accelerationlevel of about 65,000g. All six samples were failed after the Z orien-tation shock impact tests and two failure modes were distin-guished, the cracking of the diaphragm and the cracking ofbackplate and runner. Micrographs of three failed microphonesare shown in Fig. 8. In Fig. 8(a) the diaphragm had completely dis-appeared after the shock impact. In the case of the second micro-phone, most of the diaphragm had disappeared and only a smallfraction remained (see Fig. 8(b)). In the third microphone shownin Fig. 8(c), a corner of the cracked diaphragm remained after theshock impact, and in the same microphone the cracking of thebackplate and runner was detected (see Fig. 8(d)).

3.3. Stress–strain analysis

The simulation results are used to explain the observed failuremodes and predict potential failure locations. As shown in Fig. 9,six snapshots of the deformation and strain distribution of the dia-phragm and diaphragm runner are presented. The calculation wasconducted in the Z impact orientation with a peak acceleration le-vel of 65,000g. All the deformed shapes are related to the excitedvibrational modes of the structure under the shock impact and


the contact interaction between the diaphragm and the distancestoppers of the backplate (not shown in Fig. 9). The strain concen-trations are located at the edge of the diaphragm as well as thevicinity of the connection between the diaphragm and the runner.Besides, the end of the runner is also found to be a critical part as itexperienced cyclic bending during the shock impact.

In order to predict the potential failure locations in the MEMSstructure, the stress concentrations are studied. As shown inFig. 10, three potential failure locations are predicted. Fig. 10(a)presents the stress concentration at the edge of the diaphragmbeneath the backplate runner. This is probably the location wherethe crack initiated and finally led to the cracked diaphragm asshown in Fig. 8(b). Another evidence is presented in Fig. 11(a)where the crack diaphragm is visible beneath the backplate andthe predicted location of crack nucleation is indicated by an ar-row. Fig. 10(b) presents the stress concentration near the connec-tion between the diaphragm and runner. This location is inagreement with the observed diaphragm cracking shown inFig. 11(b), in which the predicted location of crack nucleation isindicated by an arrow. The last potential failure location is atthe end of the diaphragm runner (see Fig. 10(c)). However, noevidences of this failure mode have been found during the phys-ical failure analysis in this study.

During shock impacts the central region of the backplate anddiaphragm are prone to contact each other, which is a necessary



Fig. 10. 1st principal stress distributions and potential failure locations. (a) Stressconcentration at the edge of the diaphragm beneath the backplate runner. (b) Stressconcentration at the vicinity of the connection between the diaphragm and runner.(c) Stress concentration at the end of the diaphragm runner.

Fig. 11. Possible locations of crack nucleation. (a) At the edge of the diaphragmbeneath the backplate runner. (b) At the connection between the diaphragm anddiaphragm runner.


condition for any stiction failures. The surface-to-surface contactbetween the membranes was considered in the finite elementmodel. Two comparison cases were calculated, i.e. the Z+ and Z�orientation shock impact with a peak acceleration level of

Fig. 12. Calculated time history of the distanc


65,000g. The calculated time history of the distance between thebackplate and diaphragm is shown in Fig. 12. The distance is de-rived from the center of the two membranes. As shown in the fig-ure, the initial distance is 6 lm, and zero distance indicates theoccurrence of contact. In the initial stage of the Z+ orientation im-pact, both the backplate and diaphragm bent towards the PWBside. The displacement of the backplate was less than that of thediaphragm as the boundary of the backplate was constrained,resulting in an increasing distance between the two membranes.The first contact took place when the diaphragm bounced backfrom the lowest point and moved towards the backplate. In thecase of the Z� orientation impact, the backplate and diaphragmbent towards the lid side right after the shock impact. Due to theboundary constraint, the backplate bent less than the diaphragm,which allowed the diaphragm to catch up with the backplate andled to the contact. This contact analysis can be employed for fur-ther stiction failure analysis in the future.

4. Conclusions

In this work the reliability of a MEMS microphone was studiedby two accelerated tests, i.e. mixed flowing gas testing and shockimpact testing. The concentrations of gaseous pollutants in thecorosion test chamber are about 4–50 times higher than those intypical Nordic outdoor environments. Wire bond corrosion andmembrane embrittlement induced membrane cracking were de-tected after the 90 days mixed flowing gas test. The nickel diffusionthrough the pad metallization and the growth of the corrosionproducts are believed to be the reason causing the wire bonddetaching. It is believed that using Ni/Pd/Au deposit instead ofNi/Au can improve the reliability as palladium prevents nickelfrom diffusing into gold.

The membrane embrittlement is found to be associated withthe growth of surface silicon dioxide thin film as well as the com-plex interaction with the gaseous pollutants. This failure mecha-nism is different from the normal fatigue failure ofpolycrystalline silicon thin film. Further studies are required toinvestigate the mechanism of the membrane embrittlement.

The shock impact tests covered all three orthogonal axes as wellas various shock levels ranging from 1500g to 80,000g. No failureswere found in the tests with X and Y impact orientations, and onlyfailures in the Z+ and Z� orientations were detected under theacceleration level of 65,000g or higher (which was around tentimes greater than the acceleration levels that were measured infield use for consumer electronics devices). The failure modes werefound to be the cracking of the diaphragm and the cracking ofbackplate and runner. The finite element analysis predicted the

e between the backplate and diaphragm.




deformation and stress-strain state of the MEMS structure. Thepredicted locations of stress concentrations are in agreement withthe observed failure modes. The simulation results help explain thecomplex interactions between the diaphragm and the backplate.Furthermore, the surface-to-surface contact at the central regionof the membrane was predicted, which shows the existence of po-tential stiction failure.

The presented experimental and simulation results deepen ourunderstanding of corrosion reliability and shock impact reliabilityof the MEMS microphone. The conclusions drawn from the shockimpact testing are also applicable to other MEMS sensors with sim-ilar thin membrane structure.

Acknowledgements

The authors would like to thank the Finnish Funding Agency forTechnology and Innovation (TEKES), Nokia, Okmetic, and MurataElectronics for their financial support. Joonas Makkonen and JussiHokka are acknowledged for their help in the experiments. TheFinnish IT Center for Science (CSC) is also acknowledged for provid-ing computer facilities and software.

References

[1] Davies BR, Bateman VI, Brown FA, Montague S, Murray JR, Rey D, et al.Micromachined accelerometer design, modeling and validation. Proc ConfModeling Simulation Microsyst 1998; 552–5.

[2] Li GX, Shemansky FA. Drop test and analysis on micromachined structure. SensActuators 2000;A85:280–7.

[3] Maluf N. An introduction to microelectromechanical systemsengineering. Boston (MA): Artech House; 2000.

[4] ITRS International technology roadmap for semiconductors 2011 Edition.Micro-Electro-Mechanical Systems (MEMS); 2011.

[5] Huang Y, Vasan A, Doraiswami R, Osterman M, Pecht M. MEMs reliabilityreview. IEEE Trans Dev Mater Reliab 2012;12(2):482-512.

[6] Patel C, Patrick M. Simulation of MEMS vibratory gyroscope through Simulink.In: 8th International conference and exhibition on device packaging; 2012.

[7] Jang C, Goswami A, Han B. Hermeticity evaluation of polymer-sealed MEMSpackages by gas diffusion analysis. J Microelectromech Syst2009;18(3):577–611.

[8] Jang C, Goswami A, Han B, Ham S. In situ measurement of gas diffusionproperties of polymeric seals used in MEMS packages by optical gas leak test. JMicro-Nanolith MEM 2009;8(4):043025.

[9] Hartzell AL, Da Silva MG, Shea HR. MEMS reliability. Berlin: Springer; 2010.[10] Hartzel A, Woodilla D. Reliability methodology for prediction of

micromachined accelerometer stiction. In: Proc IEEE int reliability physicssymp; 1999. p. 202–4.


[11] Tanner DM, Walraven JA, Helgesen K, Irwin LW, Brown F, Smith NF, Masters N.MEMS reliability in shock environments. In: Proc 38th IEEE int reliab physsymp; 2000. p. 129–10.

[12] Hamidi A, Beck N, Thomas K, Herr E. Reliability and lifetime evaluation ofdifferent wire bonding technologies for high power IGBT modules.Microelectron Reliab 1999;39:1153–6.

[13] Muhlstein CL, Brown S, Ritchie R. High-cycle fatigue and durability ofpolycrystalline silicon thin films in ambient air. Sens Actuators A-Phys2001;94(3):177–212.

[14] Langfelder G, Longoni A, Zaraga F. Monitoring fatigue damage growth inpolysilicon microstructures under different loading conditions. Sens ActuatorsA-Phys 2010;159(2):233–8.

[15] Abbott WH. The development and performance characteristics of mixedflowing gas test environment. IEEE Trans Compon Hybrids Manuf Technol1988;11(1):22–35.

[16] Zhang S, Osterman M, Shrivastava A, Kang R, Pecht M. The Influence of H2Sexposure on immersion-silver-finished PCBs under mixed-flow gas testing.IEEE Trans Dev Mater Reliab 2010;10(1):71–111.

[17] Yang S, Wu J, Pecht M. Reliability assessment of land grid array socketssubjected to mixed flowing gas environment. IEEE Trans Reliab2009;58(4):634–47.

[18] Zhao P, Pecht M, Kang S, Park S. Assessment of Ni/Pd/Au–Pd and Ni/Pd/Au–Agpre-plated leadframe packages subject to electrochemical migration andmixed flowing gas tests. IEEE Trans Compon Package Technol2006;29(4):818–29.

[19] Tan LB, Ang CW, Lim CT, Tan, VBC, Zhang X, Modal and impact analysis ofmodern portable electronic products. In: Proc. 55th ECTC; 2005. p. 645–9.

[20] Karppinen JS, Li J, Pakarinen J, Mattila TT, Paulasto-Kröckel M. Shock impactreliability characterization of a handheld product in accelerated tests and useenvironment. Microelectron Reliab 2012;52:190–9.

[21] Meng J, Mattila TT, Dasgupt A, Sillanpää M, Jaakkol R, Anderson K, Hussa E.Testing and multi-scale modeling of drop and impact loading of complexMEMS microphone assemblies. In: Proc. 13th EuroSimE; 2012. p. 1–8.

[22] Meng J, Mattila TT, Dasgupta A, Sillanpää M, Jaakkola R, Luo, G, Anderson K.Drop Qualification of MEMS Components in Handheld Electronics at ExtremelyHigh Accelerations. In: Proc. 13th ITherm; 2012. p. 1020–8.

[23] Kaajakari V. Practical MEMS. Las Vegas: Small Gear Publishing; 2009.[24] Franssila S. Introduction to microfabrication. UK: John Wiley & Sons; 2010.[25] Hienonen R, Lahtinen R. Corrosion and climatic effects in electronics. Espoo

(Finland): VTT Automation; 2000.[26] GR-63-CORE, Chapter 5.5 Airborne contaminants test methods, Telcordia

Technologies; 2002.[27] Hokka J, Raami J, Hyvönen H, Broas M, Makkonen J, Li J, Mattila TT, Paulasto-

Kröckel M. Methods for reliability assessment of MEMS devices – case studiesof a MEMS microphone and a 3-Axis MEMS gyroscope. In: Proc. 62nd ECTC;2012. p. 62–8.

[28] Li J, Makkonen J, Broas M, Hokka J, Mattila TT, Paulasto-Kröckel M, Meng J,Dasgupta A. Reliability assessment of a MEMS microphone under shock impactloading. In: 14th EuroSimE; 2013. p. 1–6.

[29] Hopcroft MA, Nix WD, Kenny TW. What is the Young’s modulus of silicon. JMicroelectromech Syst 2010;19(2). 229–9.

[30] Jadaan OM, Nemeth NN, Bagdahn J, Sharpe JRWN. Probabilistic Weibullbehavior and mechanical properties of MEMS brittle materials. J Mater Sci2003;38:4087–127.


http://refhub.elsevier.com/S0026-2714(14)00007-9/h0010

















































reliability assessment of a mems microphone under mixed flowing gas environment and shock impact...

Documents