Characterizing Piezoelectric MEMs Displacement using an AFM Joe T. Evans, Jr., Scott Chapman Radiant Technologies, Inc. Presented July 22 at the 2013 International Symposium on the Application of Ferroelectrics in Prague, Czech Republic

Radiant Technologies, Inc.

Summary • Equipment Architecture • MEMs Measurement • Conductive Cantilevers • Conclusion

All of the test samples shown below were fabricated by Radiant.

Radiant Technologies, Inc.

System Architecture Digital microscope camera

Light lever Sample holder Z-piezo actuator

Manual sample positioning Radiant Technologies, Inc.

Sample Mount

Sample

Radiant Technologies, Inc.

Light Lever Control System Quad Cell

Laser Light Lever

 +

Chuck GPID

HVA PZT Actuator

Reference

• Negative feedback to the piezoelectric actuator under the chuck maintains the position of the reflected laser spot at the reference position on the Quad Cell. Radiant Technologies, Inc.

Large Signal Piezoelectric Test Quad Cell

Laser

Feedback Signal (pMEMs)

Precision Tester



Chuck

GPID

HVA

+ Reference

PZT Actuator

Error Signal (Piston)

• For Ångstrom-level displacements, the PNDS looks directly at the Error Signal. The test is run faster than the GPID can respond. • For MEMs-level displacements, the PNDS looks at the Feedback Signal. The test is run slowly so the GPID can respond.

Radiant Technologies, Inc.

Piston Displacement The photosensor error signal measured at high speed (~1kHz) is used to capture Ångstrom-level displacements.

Radiant Technologies, Inc.

piezoMEMs Displacement Large displacements such as the 2µ-thick PNZT/Pt/Glass membrane in the photo below are measured at 1Hz or slower while monitoring the feedback signal controlling the chuck vertical position.

The feedback signal is the position of the chuck and must be multiplied by “-1” to properly orient the loop. Radiant Technologies, Inc.

Mechanically Coupled Capacitors

1.2mm

Radiant Technologies, Inc.

Mechanically Coupled Capacitors DRIVE Actuator Sensor

RETURN

1.2mm

Sensor Actuator

GND Radiant Technologies, Inc.

Mechanically Coupled Capacitors S ensor C apacitor H ysteresis [ pMEMs-1201 RS1 on TO-18 ]

P o l a r i z a t i o n (µ C / c m 2 )

H y s ter es i s D ata

D r i v e Vo l ts

50

25

0

M echanically G enerated C harge [ pMEMs-1201 RS1 on TO-18 ]

P o l a r i z a t i o n (µ C / c m 2 )

10 0 -10 -20 0

1

2

3

4

5 6 Ti m e (m s )

7

8

9

10

Mechanically generated charge arises by executing a hysteresis loop on the actuator capacitors while measuring the charge produced by the sensor capacitors with zero volts across them.

H y s ter es i s D ata

0.15

D r i v e Vo l ts

0.10 0.05 0.00 -0.05 -0.10 20

D r iv e V o lts

D r iv e V o lts

20

10 0 -10 -20 0

1

2

3

4

5 6 Ti m e (m s )

7

8

9

~1:400 coupling coefficient Radiant Technologies, Inc.

10

Cantilever Motion A c tu a to r E d g e D is p la c e m e n t [ p M E M s-1 2 0 1 R S 1 o n T O -1 8 ]

100

A n g s tro m s

75 50 25 0 -2 5 -5 0 -1 5

-1 0

-5

0 Vo lt s

5

10

Average of ten 1-second loops.

PNDS cantilever capturing butterfly motion of resonator edge.

Radiant Technologies, Inc.

15

Complex Cantilever Motion

By measuring at different parts of the resonator as it is flexed in pseudostatic piezoelectric motion, more complex behavior is revealed. Radiant Technologies, Inc.

Complex Cantilever Motion 100

75

75

50

50

50

50

50

25

0

-25

25

0

-25

25

0

-25

P rocessed P iezo Displacement

100

75

P rocessed P iezo Displacement

100

75

P rocessed P iezo Displacement

100

75

P rocessed P iezo Displacement

100

25

0

-25

25

0

-25

-50

-50

-50

-50

-75

-75

-75

-75

-75

-100

-100

-100

-100

-100

0.0

2.5

5.0

7.5

10.0

12.5

15.0

0.0

2.5

5.0

7.5

10.0

12.5

15.0

0.0

2.5

5.0

7.5

10.0

12.5

15.0

-50

0.0

2.5

5.0

7.5

10.0

12.5

15.0

0.0

2.5

5.0

7.5

10.0

12.5

15.0

Construct of cantilever motion. Singlesided voltage application should make the cantilever bend upwards in a smooth curve. This cantilever does not. Radiant Technologies, Inc.

Conductive Cantilever • Butterfly loops require contact mode with a top electrode. • In the examples above, the top electrodes were connected to the tester using bond pads. • Where simple capacitors are used, a conductive cantilever must make the electrical contact with the top electrode. • Conductive cantilevers have had a poor success record in recording full hysteresis loops.

Radiant Technologies, Inc.

Conductive Cantilever • The traditional 40nm-diameter probe tip coated with metal cannot last long when switching ferroelectric capacitors.  The assumed mechanism is that current density is too high during ferroelectric switching, causing the metal coating to evaporate.

• A better choice is to use a tip with a wider diameter and coat with thicker metal.

• Radiant uses the 3m “Plateau” tip from App Nano coated with 1000Å of platinum. •

This tip will execute hundreds of hysteresis loops on large area capacitors without burning out. Radiant Technologies, Inc.

Conductive Cantilever P o l a r i z a ti o n ( µ C /c m 2 )

H y steresis an d B u tterfly L o o p s Proc. Hyst

Raw Hyst 1

Raw Hyst 2

Raw Hyst 3

Raw Hyst 4

Raw Hyst 5

Proc. Disp

Raw Disp 1

Raw Disp 2

Raw Disp 3

Raw Disp 4

Raw Disp 5

20 10 0 -10 -20

A n g s tr s o m s

150 100 50 0

-75

-50

-25

0 Volts

25

50

75

• After 40 hysteresis loops on a capacitor with an area of 0.0007 cm2.  The last five are shown for polarization and displacement of this antiferroelectric sample. Radiant Technologies, Inc.

Current of a Loop • The current required during the execution of a polarization hysteresis loop can be derived from the hysteresis loop. 1500Å 3/20/80 PNZT – 1 millisecond loop 1m s 4.2V H ysteresis - Type A D 103 1m s 4.2V H ysteresis - Type A D 103 P o la r iz a t io n ( µ C /c m 2 ) P o la r iz a t io n ( µ C /c m 2 )

30 30

20

Peak current at +Vc is 1A/cm2.

20

10 10

0 0

1000

-1 0

750

-2 0

500

-3 0

250

-1 0 -2 0

-4

-3 -4

-2 -3

-1 -2

0 1 -1V o lta g e0 1 V o lta g e

2

3 2

4 3

4

The use of a triangle wave stimulus makes V/t constant. This makes it simple to calculate current density from the polarization hysteresis loop.

m A /c m 2

-3 0

C urrent D ensity Loop

0 -2 5 0 -5 0 0 -7 5 0 -4

-3

-2

-1

0 V o lta g e

1

2

3

Radiant Technologies, Inc.

4

Current Density – 40nm Tip d2 Tip diameter

• Traditional 40nm diameter tip  Assume 20nm metal coating at tip even though 500Å deposited.  Tip Cross-section area 3.8x10-11 cm2 ( ~3800nm2 )  Current density for 4µm2 capacitor

Metal coated diameter d1

Cross-section =  * (r12 – r22)

Current / Cross-section [1A/cm2 x 4x10-8cm2] / 3.8x10-11cm2 1060 Amps/cm2

Radiant Technologies, Inc.

Current Density – 3µm Tip d2 Tip diameter

• 3µm plateau tip  Assume 100nm metal coating around the edge of the plateau  Tip Cross-section area 9.7x10-9 cm2  Current density for 4µm2 capacitor

Metal coated diameter d1

Cross-section =  * (r12 – r22)

Current / Cross-section [1A/cm2 x 4x10-8cm2] / 9.7x10-9cm2 4.1 Amps/cm2

Radiant Technologies, Inc.

Analysis • The rule of thumb in integrated circuit design is that currents must remain below 100kA/cm2 to prevent damage. • The current densities for a 300µ-diameter TE dot (~7x10-4cm2) are

 40nm tip

18.6MA/cm2

 Plateau tip

72.9kA/cm2

• The traditional 40nm-diameter metal-coated tip will not work for large area capacitors while the 3µ-diameter tip will. • Experience shows that 40nm-diameter metal-coated tips do not work long even for micron-scale capacitors, indicating that the conduction geometry of the tip is more restricted than the model used above. Radiant Technologies, Inc.

Conclusions • The AFM configured to measure large displacements can be used to measure the complex movement of piezoMEMs. • Completed piezoMEMs should have built-in conduction paths for the two electrodes, allowing unlimited actuation by the tester. • Simple capacitors without built-in conduction paths can be reliably tested using cantilevers having large, dull points coated with up to 1000Å thick platinum.

Radiant Technologies, Inc.