IoT Applications Moshe Avraham 1 , Gady Golan 1 , Michele Vaiana 2 , - PowerPoint PPT Presentation

Wafer Level Packaged CMOS-SOI-MEMS Thermal Sensor at Wide Pressure Range for IoT Applications Moshe Avraham 1 , Gady Golan 1 , Michele Vaiana 2 , Giuseppe Bruno 2 , Maria Eloisa Castagna 2 , Sara Stolyarova 3 , Tanya Blank 3 , Yael Nemirovsky 3,4 1. Ariel University, Ariel, 40700, Israel 2. STMicroelectronics, Stradale Primosole, 50 – 95121 Catania, Italy 3. Electrical Engineering Dept., Technion- Israel Institute of Technology, Haifa 32000, Israel 4. TODOS TECHNOLOGIES Ltd., Israel. Presented at the 7th Electronic Conference on Sensors and Applications, 15 － 30 November 2020; Available online: https://ecsa-7.sciforum.net/. 1

RESEARCH MOTIVATION IR sensors have huge markets: IoT, Smart homes, Automotive, etc ▪ Thermal sensors detect temperature changes induced by remote ▪ sensing of IR radiation and provide uncooled IR sensors MEMS enable high performance miniature thermal sensors ▪ From: https://www.todos-technologies.com/ 2

RESEARCH INNOVATION: TMOS ▪ The TMOS (Thermal-MOS) is a thermal sensor developed at the Technion ▪ Achieved by CMOS-SOI-MEMS process ▪ CMOS transistor is the standard building block of CMOS CHIPS manufactured in FABs ▪ By applying backend machining – TMOS becomes the highest performance thermal sensor (compared to bolometers, PYRO ’ s and thermopiles) ▪ Operation at subthreshold requires very low power Suspended Transistor Holding arm Oxides Buried oxide Silicon Micro-machined cavity bulk Fabricated TMOS sensor Schematic cross-section of the TMOS layers 3D TMOS pixel Schematic 3

TMOS OPERATION PRINCIPLE The micro-machined Absorbed photons thermally insulated increase the TMOS Transistor voltage transistor has very low temperature and modify detects temperature thermal mass and very the current-voltage changes at subthreshold low thermal conductivity characteristics IR Δ𝑈 Suspended Holding arm radiation Transistor Oxides Buried oxide Silicon Micro-machined cavity t bulk 4

WAFER LEVEL PROCESSING AND VACUUM PACKAGING Currently on 8-inch wafers and 0.13 𝝂𝒏 CMOS-SOI PROCESS ▪ 3 silicon wafers are bonded ▪ Top silicon optical window wafer ▪ vacuum of 10¯ ⁵ atm ▪ stable over 5 years MEMS and TMOS wafer Bottom silicon wafer 5

THE RESEARCH QUESTION ▪ Residual pressure determines the thermal conductance - 𝑯 𝒖𝒊 𝑫 𝒖𝒊 ▪ thermal time constant 𝝊 𝒖𝒊 = 𝑯 𝒖𝒊 ▪ What is the effect of the residual vacuum upon performance? ▪ What pressure is critical for the proper performance of the device? Δ𝑈 ∆𝑈 𝑡𝑡 t 𝜐 𝑢ℎ 6

THERMAL MODELING OF THE PACKAGED TMOS SENSOR 𝑄 𝑝𝑞𝑢 ▪ The power balance equation: Τ 𝜃𝑄 𝑝𝑞𝑢 = 𝐷 𝑢ℎ 𝑒 ∆𝑈 𝑢 𝑒 𝑢 + 𝐻 𝑢ℎ ∆𝑈(𝑢) 𝜃𝑄 𝑝𝑞𝑢 In steady-state: ∆𝑈 𝑡𝑡 = ▪ 𝐻 𝑢ℎ h 𝐻 𝑏𝑠𝑛 The thermal conductance is determined by three mechanisms: ▪ 𝐻 𝑢ℎ = 𝐻 𝑡𝑝𝑚𝑗𝑒𝑡 + 𝐻 𝑕𝑏𝑡 + 𝐻 𝑠𝑏𝑒𝑗𝑏𝑢𝑗𝑝𝑜 The thermal capacitance: 𝐷 𝑢ℎ = 𝜍𝑑′𝐵 𝑡𝑢𝑏𝑕𝑓 ℎ ▪ Δ𝑈 The thermal time constant: 𝜐 𝑢ℎ = 𝐷 𝑢ℎ ▪ ∆𝑈 𝑡𝑡 𝐻 𝑢ℎ t 𝜐 𝑢ℎ 𝒒𝒃𝒔𝒃𝒏𝒇𝒖𝒇𝒔𝒕 𝒃𝒐𝒆 𝒅𝒑𝒐𝒕𝒖𝒃𝒐𝒖𝒕: 𝜍 kg J m 3 − mass density, c ′ kg ∙ 𝐿 − specific heat capacitance, 𝜃 − 𝑝𝑞𝑢𝑗𝑑𝑏𝑚 𝑓𝑔𝑔𝑗𝑑𝑗𝑓𝑜𝑑𝑧, h − stage height [m] 7

THERMAL MODELING – SOLID CONDUCTION AND RADIATION Body emits radiation according to its temperature: ▪ 4 ≈ 𝜁𝜏 2 ∙ 𝐵 𝑡𝑢𝑏𝑕𝑓 𝑈 4 𝑄 𝑠𝑏𝑒 = 𝜁𝜏𝐵 𝑢𝑝𝑢𝑏𝑚 𝑈 𝑡 𝑡 𝑩 𝒕𝒖𝒃𝒉𝒇 ▪ Hence, the thermal conductance due to thermal radiation: 3 𝐻 𝑠𝑏𝑒 = 𝑒 𝑄 𝑠𝑏𝑒 𝑒 𝑈 = 8𝜁𝜏𝐵 𝑡𝑢𝑏𝑕𝑓 𝑈 Τ 𝑡 𝑴 𝑩𝑺𝑵 ▪ The thermal conduction through a material is derived from Fourier law and equals to: TMOS top view schematic ∆𝑈 ∙ ∆𝑢 = 𝑙 ∙ 𝐵 Q 𝑩 𝑩𝑺𝑵 𝐻 = 𝑀 For example, in our device, the thermal solid conduction is governed by the ▪ holding arm: 𝐻 𝑏𝑠𝑛 = 𝑙 𝑏𝑠𝑛 ∙ 𝐵 𝑏𝑠𝑛 𝑀 𝑏𝑠𝑛 𝒒𝒃𝒔𝒃𝒏𝒇𝒖𝒇𝒔𝒕 𝒃𝒐𝒆 𝒅𝒑𝒐𝒕𝒖𝒃𝒐𝒖𝒕: Arm cross-section 𝑋 W m ∙ 𝐿 − thermal conductivity, A stage − stage area m 2 , L arm − arm length m 𝜁 − body emmisivity, 𝜏 = 5.67 ∙ 10 −8 𝑛 2 ∙ 𝐿 4 − Stefan– Boltzmann constant, k 𝑅 − 𝑢ℎ𝑓𝑠𝑛𝑏𝑚 𝑓𝑜𝑓𝑠𝑕𝑧 𝐾 , 𝑈 − 𝑢𝑓𝑛𝑞𝑓𝑠𝑏𝑢𝑣𝑠𝑓 𝐿 , 𝑢 − 𝑢𝑗𝑛𝑓 𝑡 , 𝐵 − 𝑏𝑠𝑓𝑏, 𝑀 − 𝑚𝑓𝑜𝑕𝑢ℎ[𝑛] 8

THERMAL MODELING – GAS CONDUCTION AT HIGH PRESSURE The thermal conductivity of gas at high pressure is independent on the pressure and equals to a constant for a given ▪ temperature (like solids): 𝑿 𝒍 𝒊𝒋𝒉𝒊−𝒒𝒔𝒇𝒕𝒕𝒗𝒔𝒇 = 𝒅𝒑𝒐𝒕𝒖𝒃𝒐𝒖 𝑳∙𝒏 ▪ The thermal conductivity of the gas is given by: 𝑙 𝑕𝑏𝑡 = 𝟐 ′′ ∙ 𝑞 𝑋 𝟒 𝝇𝒅′𝒘 𝒏𝒑𝒎 ∙ 𝒎 𝒏𝒈𝒒 = 𝐻 0 𝑞 0 ∙ 𝑚 𝑛𝑔𝑞 𝐿∙𝑛 At high pressure where the collision distance between two gas molecules is much smaller than the device typical ▪ dimensions. Therefore, the mean-free-path is governed by the molecule collision distance In this case, at high pressure, the mean-free-path is proportional to the inverse number of molecules - 𝒎 𝒏𝒈𝒒 ∝ 𝒐 −𝟐 , and ▪ the pressure is proportional to the number of molecules - 𝒒 ∝ n , where n is the number of molecule 𝑿 ▪ For air at high pressure, the value of 𝒍 𝐛𝐣𝐬 is well established and equals to 𝟏. 𝟏𝟑𝟕 𝒏∙𝑳 at 300 ° [K] 𝑙𝑕 𝐾 𝑛 3 , 𝑑 ′ − 𝑡𝑞𝑓𝑑𝑗𝑔𝑗𝑑 𝑢ℎ𝑓𝑠𝑛𝑏𝑚 𝑑𝑏𝑞𝑏𝑑𝑗𝑢𝑏𝑜𝑑𝑓 𝒅𝒑𝒐𝒕𝒖𝒃𝒐𝒖𝒕 𝒃𝒐𝒆 𝒒𝒃𝒔𝒃𝒏𝒇𝒖𝒇𝒔𝒕: 𝑞 − 𝑞𝑠𝑓𝑡𝑡𝑣𝑠𝑓 𝑄𝑏 , 𝜍 − 𝑛𝑏𝑡𝑡 𝑒𝑓𝑜𝑡𝑗𝑢𝑧 𝑙𝑕 ∙ 𝐿 , 𝑤 𝑛𝑝𝑚 − 𝑛𝑝𝑚𝑓𝑑𝑣𝑚𝑓 𝑤𝑓𝑚𝑝𝑑𝑗𝑢𝑧 𝑛 ′′ 𝑋 𝑡 , 𝑚 𝑛𝑔𝑞 𝑛 − 𝑛𝑓𝑏𝑜 𝑔𝑠𝑓𝑓 𝑞𝑏𝑢ℎ, 𝐻 0 𝐿 − 𝑜𝑝𝑠𝑛𝑏𝑚𝑗𝑨𝑓𝑒 𝑢ℎ𝑓𝑠𝑛𝑏𝑚 𝑑𝑝𝑜𝑒𝑣𝑑𝑢𝑏𝑜𝑑𝑓 𝑔𝑝𝑠 𝑞 0 = 1[𝑄𝑏] 9

THERMAL MODELING – GAS CONDUCTION AT LOW PRESSURE ▪ The thermal conductivity by the gas is given by: 𝒍 𝒉𝒃𝒕 = 𝟐 ′′ ∙ 𝒒 𝑿 𝟒 𝝇𝒅′𝒘 𝒏𝒑𝒎 ∙ 𝒎 𝒏𝒈𝒒 = 𝑯 𝟏 𝒒 𝟏 ∙ 𝒎 𝒏𝒈𝒒 𝑳∙𝒏 ▪ At low pressure where the collision distance between two gas molecules is much larger than the device typical dimensions. Therefore, the mean-free-path is governed by the device smallest typical dimension Collision Distance between two molecules ′′ ▪ In this study, 𝒎 𝒏𝒈𝒒 = 𝒉𝒃𝒒 = 𝟒𝝂𝒏 and 𝑯 𝟏 𝒒 𝟏 ≈ 𝟑 Collision Distance between molecule = 𝑚 𝑛𝑔𝑞 = 𝑕𝑏𝑞 to device surface 𝑿 Therefore: 𝒍 𝒎𝒑𝒙−𝒒𝒔𝒇𝒕𝒕𝒗𝒔𝒇 = 𝟕 ∙ 𝒒 ▪ 𝒏∙𝑳 𝑙𝑕 𝐾 𝑛 3 , 𝑑 ′ − 𝑡𝑞𝑓𝑑𝑗𝑔𝑗𝑑 𝑢ℎ𝑓𝑠𝑛𝑏𝑚 𝑑𝑏𝑞𝑏𝑑𝑗𝑢𝑏𝑜𝑑𝑓 𝒅𝒑𝒐𝒕𝒖𝒃𝒐𝒖𝒕 𝒃𝒐𝒆 𝒒𝒃𝒔𝒃𝒏𝒇𝒖𝒇𝒔𝒕: 𝑞 − 𝑞𝑠𝑓𝑡𝑡𝑣𝑠𝑓 𝑄𝑏 , 𝜍 − 𝑛𝑏𝑡𝑡 𝑒𝑓𝑜𝑡𝑗𝑢𝑧 𝑙𝑕 ∙ 𝐿 , 𝑤 𝑛𝑝𝑚 − 𝑛𝑝𝑚𝑓𝑑𝑣𝑚𝑓 𝑤𝑓𝑚𝑝𝑑𝑗𝑢𝑧 𝑛 ′′ 𝑋 𝑡 , 𝑚 𝑛𝑔𝑞 𝑛 − 𝑛𝑓𝑏𝑜 𝑔𝑠𝑓𝑓 𝑞𝑏𝑢ℎ, 𝐻 0 𝐿 − 𝑜𝑝𝑠𝑛𝑏𝑚𝑗𝑨𝑓𝑒 𝑢ℎ𝑓𝑠𝑛𝑏𝑚 𝑑𝑝𝑜𝑒𝑣𝑑𝑢𝑏𝑜𝑑𝑓 𝑔𝑝𝑠 𝑞 0 = 1[𝑄𝑏] 10

THERMAL MODELING – GAS CONDUCTION AT INTERMIDATE PRESSURE ▪ At intermediate pressure, the thermal conductivity is given by the parallel combined of the both mechanisms: 1 1 1 = + 𝑙 𝑕𝑏𝑡 𝑙 ℎ𝑗𝑕ℎ−𝑞𝑠𝑓𝑡𝑡𝑣𝑠𝑓 𝑙 𝑚𝑝𝑥−𝑞𝑠𝑓𝑡𝑡𝑣𝑠𝑓 11

THERMAL MODELING - SUMMARY ▪ The holding arm conduction does not depend on pressure ▪ Typical CMOS-SOI thermal properties required for thermal simulation: Air conduction is governed by two mechanisms: at low pressure and high pressure ▪ This study evaluates this pressure impact on the thermal conductance of the packaged device ▪ The air thermal properties calculated by the ideal gas law and by the method showed in the previous slides ▪ 12

THERMAL MODELING – BOUNDARY CONDITIONS ▪ 3D model of the device was generated in FEA software ▪ Materials thermal properties were assigned Applying boundary conditions to our packaged model: ▪ Constant temperature of 𝟑𝟏 𝟏 [𝑫] on the outer package 1 𝝂𝑿 𝒑𝒐 𝒇𝒃𝒅𝒊 𝒕𝒗𝒕𝒒𝒇𝒐𝒆𝒇𝒆 𝒕𝒖𝒃𝒉𝒇 ▪ Simulations for wide pressure range values were performed 13