Assessment of thick-film resistors for manufacturing piezoresistive - PowerPoint PPT Presentation

Assessment of TFRs for piezoresistive sensors Assessment of thick-film resistors for manufacturing piezoresistive sensors Thomas Maeder, Caroline Jacq and Peter Ryser É cole Polytechnique Fédérale de Lausanne (EPFL), Switzerland IMAPS/ACerS 11 th CICMT, Dresden, 20-23.4.2015 1

Outline 1. Introduction – manufacturing & trimming issues 2. Resistor study 3. Overglazing, trimming, etc. 4. Conclusions & outlook Outline IMAPS/ACerS 11 th CICMT, Dresden, 20-23.4.2015 2

Outline 1. Introduction – manufacturing & trimming issues 2. Resistor study 3. Overglazing, trimming, etc. 4. Conclusions & outlook Outline IMAPS/ACerS 11 th CICMT, Dresden, 20-23.4.2015 3

Typical thick-film piezoresistive sensor n Typical elements n Sensing bridge n Offset trim n TCO trim n Differential amplifier n Typical values (±) n Offset ~30 mV/V n Response ~3 mV/V n TCO ~1 µV/V/K 
 (50 K : ~0.05 mV/V) n For 0.1% F.S.: n Offset reduction ~10'000 × n Stability (bridge) ~10 ppm IMAPS/ACerS 11 th CICMT, Dresden, 20-23.4.2015 1 - Introduction 4

Why trim? n Modern digital chips n Input stage usually PGA (programmable-gain amplifier) n Gain limited by signal n In raw state, offset dominates signal, >> response n For optimal use, reduce offset to < response n With typical raw offset ~30 mV/V, max. gain ~30 × n With typical response ~3 mV/V, typ. gain required ~200 × n Reduce offset typically by ~10…30 × n Trimming of TCO usually not necessary with chips n Typically, temperature error <10% of piezoresistive response n Can be done digitally n Laser trim: large-scale production; better temperature sensing IMAPS/ACerS 11 th CICMT, Dresden, 20-23.4.2015 1 - Introduction 5

Examples – pressure cell Ceramic: classical layout n All-active bridge n Coarse offset trim on cell n Direct TCO trim n Need good amplifier – usually not accessible after mounting of electronics Steel: changes n Issue: trim on dielectric n Coarse offset trim off-cell n Indirect TCO trim n PTC resistor on cell n Normal resistor in parallel IMAPS/ACerS 11 th CICMT, Dresden, 20-23.4.2015 1 - Introduction 6

Examples – cantilever force cell n All-active bridge n Discrete offset trim (stable, active, ~no TCO change) n Coarse classical trim (more precise) n No TCO trim (on base, with fine trim) R j2 + R j1 + R o1 – Top Bottom R o1 + R j2 – R j1 – + IMAPS/ACerS 11 th CICMT, Dresden, 20-23.4.2015 1 - Introduction 7

Examples – glass-sealed pressure cell n All-active bridge n Discrete offset trim cuts only on cell n All other trims on separate module IMAPS/ACerS 11 th CICMT, Dresden, 20-23.4.2015 1 - Introduction 8

Trimming of sensor electronics n Normally passive & active part n High resistor values often problematic n Harsh post-processing (breaking, soldering, ultrasound, …) Top IMAPS/ACerS 11 th CICMT, Dresden, 20-23.4.2015 1 - Introduction 9

Factors for offset, TCO & stability n Resistor interactions n Substrate (Al 2 O 3 , dielectric, LTCC…) n Terminations n Overglaze n TCO ≠ TCR; TCO determined by TCR tracking n Trimming n Discrete (stable) or classical (precise) n Trimming resistor used (coarse: use same as bridge) n Terminations (material near terminations ≠ away) n Parameters & resistor material n Post-processing IMAPS/ACerS 11 th CICMT, Dresden, 20-23.4.2015 1 - Introduction 10

Outline 1. Introduction – manufacturing & trimming issues 2. Resistor study 3. Overglazing, trimming, etc. 4. Conclusions & outlook Outline IMAPS/ACerS 11 th CICMT, Dresden, 20-23.4.2015 11

Resistor study n (Substrate = alumina) n Termination material n Resistor material & length n Overglaze material IMAPS/ACerS 11 th CICMT, Dresden, 20-23.4.2015 1 - Introduction 12

Processing parameters n Resistor under…overfired n See whether this changes its interactions with overglaze n Overglaze under…overfired n Extent of effect on resistor IMAPS/ACerS 11 th CICMT, Dresden, 20-23.4.2015 1 - Introduction 13

As-fired 10 k Ω – spread of values n Newer resistor compositions (DP 2041 / R314P) better n Thin Au (D) terminations = lowest spread n Low geometric disturbance of screen printing n Low diffusion with terminations IMAPS/ACerS 11 th CICMT, Dresden, 20-23.4.2015 1 - Introduction 14

As-fired 100 Ω – spread of values n Less difference seen in 100 Ω compositions n Not dominant – used for fine trimming IMAPS/ACerS 11 th CICMT, Dresden, 20-23.4.2015 1 - Introduction 15

As-fired 10 k Ω – effect of process n Process dependence of value & TCR different n Strong length effects on TCR -> TCO for short resistors IMAPS/ACerS 11 th CICMT, Dresden, 20-23.4.2015 1 - Introduction 16

Outline 1. Introduction – manufacturing & trimming issues 2. Resistor study 3. Overglazing, trimming, etc. 4. Conclusions & outlook Outline IMAPS/ACerS 11 th CICMT, Dresden, 20-23.4.2015 17

Overglazing resistors n Overglazing above nominal temperature – : strong drift n Length dependence on ∆ TCR: leads to TCO IMAPS/ACerS 11 th CICMT, Dresden, 20-23.4.2015 3 –Overglazing, trimming, … 18

Trimming problems n Behaviour mostly normal: slight value increase n Decrease of value for 100 k Ω composition! 3 – Overglazing, trimming, … IMAPS/ACerS 11 th CICMT, Dresden, 20-23.4.2015 19

Trimming & stability of DP 2041 bridges 8 40% AgPd 45% AgPd 50% AgPd n Au initially ~2 × better 6 40% Au 45% Au 50% Au than Ag:Pd 4 2 n After trimming 0 n Trim + ultrasound ca. -2 -40 ± 20 mV/V -4 n Advantage lost upon -6 8 overglazing 6 4 n Trim-overglaze 2 interactions dominant 0 -2 n Temperature not so -4 dominant (anneals) -6 -8 n Better: refire -10 overglaze or glaze again IMAPS/ACerS 11 th CICMT, Dresden, 20-23.4.2015 3 – Overglazing, trimming, … 20

Outline 1. Introduction – manufacturing & trimming issues 2. Resistor study 3. Overglazing, trimming, etc. 4. Conclusions & outlook Outline IMAPS/ACerS 11 th CICMT, Dresden, 20-23.4.2015 21

Conclusions & outlook n Thick-film piezoresistive sensors & laser trimming n Relatively low signal + harsh environments: difficult n High process temperatures -> materials interactions critical n Few alternatives to laser trimming (voltage?) for large series (cost) n Best stability: start with discrete coarse trims n Parameter development can be tedious n Must ensure access of beam to resistor (not always practical!) n Software offset trimming n R adj = same paste as bridge, 
 long meander (value ~10 × bridge) n Little to no effect on TCO 
 (if DAC reasonably good) 4 - Conclusions & outlook IMAPS/ACerS 11 th CICMT, Dresden, 20-23.4.2015 22

Questions? THANK YOU! 4 - Conclusions & outlook IMAPS/ACerS 11 th CICMT, Dresden, 20-23.4.2015 23

Gauge factor measurement n Alumina cantilever n Effective signal ~independent of loading errors IMAPS/ACerS 11 th CICMT, Dresden, 20-23.4.2015 Measurement of gauge factor 24