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

▶
assessment of thick film resistors for manufacturing
SMART_READER_LITE
LIVE PREVIEW

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

Apr 10, 2023 156 likes •414 views

Assessment of TFRs for piezoresistive sensors Assessment of thick-film resistors for manufacturing piezoresistive sensors Thomas Maeder, Caroline Jacq and Peter Ryser cole Polytechnique Fdrale de Lausanne (EPFL), Switzerland IMAPS/ACerS 11

slide-1
SLIDE 1

1 IMAPS/ACerS 11th CICMT, Dresden, 20-23.4.2015

Thomas Maeder, Caroline Jacq and Peter Ryser

École Polytechnique Fédérale de Lausanne (EPFL), Switzerland

Assessment of TFRs for piezoresistive sensors

Assessment of thick-film resistors for manufacturing piezoresistive sensors

slide-2
SLIDE 2

2 IMAPS/ACerS 11th CICMT, Dresden, 20-23.4.2015

Outline

  • 1. Introduction – manufacturing & trimming issues
  • 2. Resistor study
  • 3. Overglazing, trimming, etc.
  • 4. Conclusions & outlook

Outline

slide-3
SLIDE 3

3 IMAPS/ACerS 11th CICMT, Dresden, 20-23.4.2015

Outline

  • 1. Introduction – manufacturing & trimming issues
  • 2. Resistor study
  • 3. Overglazing, trimming, etc.
  • 4. Conclusions & outlook

Outline

slide-4
SLIDE 4

4 IMAPS/ACerS 11th CICMT, Dresden, 20-23.4.2015

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

1 - Introduction

slide-5
SLIDE 5

5 IMAPS/ACerS 11th CICMT, Dresden, 20-23.4.2015

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

1 - Introduction

slide-6
SLIDE 6

6 IMAPS/ACerS 11th CICMT, Dresden, 20-23.4.2015

Examples – pressure cell

1 - Introduction

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

slide-7
SLIDE 7

7 IMAPS/ACerS 11th CICMT, Dresden, 20-23.4.2015

Examples – cantilever force cell

1 - Introduction

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) Top Bottom

Rj2

–

+ Rj1

–

Rj2

+ Rj1 +

Ro1

–

Ro1

+

slide-8
SLIDE 8

8 IMAPS/ACerS 11th CICMT, Dresden, 20-23.4.2015

Examples – glass-sealed pressure cell

1 - Introduction

n All-active bridge n Discrete offset trim cuts only

  • n cell

n All other trims on separate

module

slide-9
SLIDE 9

9 IMAPS/ACerS 11th CICMT, Dresden, 20-23.4.2015

Trimming of sensor electronics

1 - Introduction

n Normally passive & active part n High resistor values often problematic n Harsh post-processing (breaking, soldering, ultrasound, …) Top

slide-10
SLIDE 10

10 IMAPS/ACerS 11th CICMT, Dresden, 20-23.4.2015

Factors for offset, TCO & stability

1 - Introduction

n Resistor interactions

n Substrate (Al2O3, 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

slide-11
SLIDE 11

11 IMAPS/ACerS 11th CICMT, Dresden, 20-23.4.2015

Outline

  • 1. Introduction – manufacturing & trimming issues
  • 2. Resistor study
  • 3. Overglazing, trimming, etc.
  • 4. Conclusions & outlook

Outline

slide-12
SLIDE 12

12 IMAPS/ACerS 11th CICMT, Dresden, 20-23.4.2015

Resistor study

1 - Introduction

n (Substrate = alumina) n Termination material n Resistor material & length n Overglaze material

slide-13
SLIDE 13

13 IMAPS/ACerS 11th CICMT, Dresden, 20-23.4.2015

Processing parameters

1 - Introduction

n Resistor under…overfired

n See whether this changes its interactions with overglaze

n Overglaze under…overfired

n Extent of effect on resistor

slide-14
SLIDE 14

14 IMAPS/ACerS 11th CICMT, Dresden, 20-23.4.2015

As-fired 10 kΩ – spread of values

1 - Introduction

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

slide-15
SLIDE 15

15 IMAPS/ACerS 11th CICMT, Dresden, 20-23.4.2015

As-fired 100 Ω – spread of values

1 - Introduction

n Less difference seen in 100 Ω compositions n Not dominant – used for fine trimming

slide-16
SLIDE 16

16 IMAPS/ACerS 11th CICMT, Dresden, 20-23.4.2015

As-fired 10 kΩ – effect of process

1 - Introduction

n Process dependence of value & TCR different n Strong length effects on TCR -> TCO for short resistors

slide-17
SLIDE 17

17 IMAPS/ACerS 11th CICMT, Dresden, 20-23.4.2015

Outline

  • 1. Introduction – manufacturing & trimming issues
  • 2. Resistor study
  • 3. Overglazing, trimming, etc.
  • 4. Conclusions & outlook

Outline

slide-18
SLIDE 18

18 IMAPS/ACerS 11th CICMT, Dresden, 20-23.4.2015

Overglazing resistors

3 –Overglazing, trimming, …

n Overglazing above nominal temperature – : strong drift n Length dependence on ∆TCR: leads to TCO

slide-19
SLIDE 19

19 IMAPS/ACerS 11th CICMT, Dresden, 20-23.4.2015

Trimming problems

n Behaviour mostly normal: slight value increase n Decrease of value for 100 kΩ composition!

3 – Overglazing, trimming, …

slide-20
SLIDE 20

20 IMAPS/ACerS 11th CICMT, Dresden, 20-23.4.2015

Trimming & stability of DP 2041 bridges

3 – Overglazing, trimming, …

n Au initially ~2× better

than Ag:Pd

n After trimming n Trim + ultrasound

n Advantage lost upon

  • verglazing

n Trim-overglaze

interactions dominant

n Temperature not so

dominant (anneals)

n Better: refire

  • verglaze or glaze

again

  • 6
  • 4
  • 2

2 4 6 8 40% AgPd 45% AgPd 50% AgPd 40% Au 45% Au 50% Au ca.

  • 40 ± 20

mV/V

  • 10
  • 8
  • 6
  • 4
  • 2

2 4 6 8

slide-21
SLIDE 21

21 IMAPS/ACerS 11th CICMT, Dresden, 20-23.4.2015

Outline

  • 1. Introduction – manufacturing & trimming issues
  • 2. Resistor study
  • 3. Overglazing, trimming, etc.
  • 4. Conclusions & outlook

Outline

slide-22
SLIDE 22

22 IMAPS/ACerS 11th CICMT, Dresden, 20-23.4.2015

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 Radj = same paste as bridge,


long meander (value ~10× bridge)

n Little to no effect on TCO


(if DAC reasonably good)

4 - Conclusions & outlook

slide-23
SLIDE 23

23 IMAPS/ACerS 11th CICMT, Dresden, 20-23.4.2015

Questions?

4 - Conclusions & outlook

THANK YOU!

slide-24
SLIDE 24

24 IMAPS/ACerS 11th CICMT, Dresden, 20-23.4.2015

Gauge factor measurement

Measurement of gauge factor

n Alumina cantilever n Effective signal ~independent of loading errors