Typical Packaged AB Performance - ferrodevices.comTypical Type AB Performance 8 Typical Performance • Typical performance data for the 400µ2 “AB” packaged part are plotted on

Radiant Technologies, Inc. 1

Typical Performance of Packaged“AB” Capacitors

Joe T. Evans, Jr.Radiant Technologies, Inc.

August 29, 2008

Radiant Technologies, Inc.

Typical Type AB Performance

2

Introduction• The Type AB capacitors were designed to demonstrate all of the

performance and reliability properties found in most ferroelectriccapacitors, including fatigue and imprint.

• The data plotted in the following pages came from one dataset:“typical type ab performance 07_07_08.dst” which can bedownloaded from www.ferrodevices.com/components2.html. Visioncan be downloaded from www.ferrodevices.com/tdownload.html.

• The data shown in this document represents a cross-section of thetypes of tests that may be executed on ferroelectric capacitors.



3

Description of Samples• Capacitor Structure (from bottom to top):

– 5000Å silicon dioxide on <100> silicon– 400Å titanium dioxide– 1500Å polycrystalline <111> &<200> platinum bottom electrode– 2550Å 20/80 MOD PZT deposited in seven layers– 1000Å polycrystalline platinum top electrode– 2200Å 20/80 PZT passivation topped by 400Å titanium dioxide– Metal interconnect using 200Å chromium &5000Å gold

Silicon Wafer 550 micrometers thick

Half of amicrometer of

SiliconDioxide as afoundation

One quarter ofa micrometer

of PZT asprotection

Top ElectrodeBottom

Electrode Ferroelectricmaterial



4

Description of Samples

10,000 square micron capacitorContact Pad in Gold

The 400µ2 and 100µ2

capacitors are “cross”type devices with thetop electrode contactto the side of thecapacitor. The largercapacitors are “plate”capacitors where thetop electrode is fullyenclosed by thebottom electrode andthe top electrodecontact is over theactive capacitor area.



5

Packaging

• The capacitors are bonded into TO-18 transistor type cans with fourleads.

• See the document “Ferroelectric Component Technical Description - RC2-AAA” for more information

Top view

COMMON CAP A

CAP B CASECap A Cap BCommon

TO-18 Can



6

Socket Board

• Most test data were collected from a single “AB” capacitor mountedon the TO-18 Socket Board shown below.

• Data at temperature was collected from unpackaged die on asemiconductor probe station with a hot chuck.

Cap A

Cap B

Common

GND



7

Recovery• As has been reported in the past, a DC bias applied to a ferroelectric

capacitor can reverse some imprint effects.

• The “AB” capacitors are capable of withstanding long periods at 9Vacross their 2550Å of thickness.

• Imprint effects may be almost 100% reversed using the recoveryprocedure recommended below.

• A significant amount of fatigue effects may also be reversed using thesame recovery procedure.

• Recommended Recovery Procedure:– 9V 1Hz square wave for 100 seconds at room temperature. Longer

periods may be executed if necessary.



8

Typical Performance• Typical performance data for the 400µ2 “AB” packaged

part are plotted on the following pages.

• All test data were taken on a Precision LC or PrecisionPremier II.

– All plots were formatted using plotting tools in Vision and thencopied directly into this presentation.

• The dataset containing the plotted data may be found atwww.ferrodevices.com/components2.html.



9

Before Recovery(10ms Period)

-50

-40

-30

-20

-10

0

10

20

30

40

50

-7.5 -5.0 -2.5 0.0 2.5 5.0 7.5

Hysteresis Before Recovery[ Type AB WHITE ]

Pol

ariz

atio

n (µ

C/c

m2)

Voltage



10

After 100 Second Recovery(10ms Period)

-50

-40

-30

-20

-10

0

10

20

30

40

50

-7.5 -5.0 -2.5 0.0 2.5 5.0 7.5

Hysteresis Before and After Recovery[ Type AB WHITE]

Pol

ariz

atio

n

Voltage

Hysteresis After Recovery: Polarization (µC/cm2)

Initial Hysteresis: Polarization (µC/cm2)



11

Longer Recovery(5 minutes, room temperature)

-50

-40

-30

-20

-10

0

10

20

30

40

50

-7.5 -5.0 -2.5 0.0 2.5 5.0 7.5

After 5 Minutes Total Recovery[ Type AB WHITE ]

Pol

ariz

atio

n (µ

C/c

m2)

Voltage0

50

100

150

200

250

-7.5 -5.0 -2.5 0.0 2.5 5.0 7.5

Normalized CV of Recovered Capacitor[ Type AB WHITE ]

Nor

mal

ized

Cap

acita

nce

(µF/

cm2)

Voltage



12

Opposing Hysteresis Loops

The 20/80 PZT slides back and forth depending on the value of its remanent polarizationbefore the loop starts. Other compositions do not have this effect to such a magnitude.

-40

-30

-20

-10

0

10

20

30

40

-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6

Opposing 6V Loops[ Type AB WHITE ]

Pol

ariz

atio

n

Voltage

10ms -6V Hyst: Polarization (µC/cm2): 1

10ms 6V Hyst: Polarization (µC/cm2): 1



13

Nested Hysteresis Loops(10ms Period)

-50

-40

-30

-20

-10

0

10

20

30

40

50

-7.5 -5.0 -2.5 0.0 2.5 5.0 7.5

Hysteresis Loop vs Voltage[ Type AB WHITE, 10ms period ]

Pol

ariz

atio

n

Voltage

Nested Loops: P olarization (µC/cm2): 1 Nested Loops: P olarization (µC/cm2): 2 Nested Loops: P olarization (µC/cm2): 3 Nested Loops: P olarization (µC/cm2): 4 Nested Loops: P olarization (µC/cm2): 5 Nested Loops: P olarization (µC/cm2): 6 Nested Loops: P olarization (µC/cm2): 7

Nested Loops: P olarization (µC/cm2): 8 Nested Loops: P olarization (µC/cm2): 9 Nested Loops: P olarization (µC/cm2): 10 Nested Loops: P olarization (µC/cm2): 11 Nested Loops: P olarization (µC/cm2): 12 Nested Loops: P olarization (µC/cm2): 13 Nested Loops: P olarization (µC/cm2): 14



14

Nested nCV Loops

The normalized CV is the mathematical derivative of a hysteresis loop. It has the units ofcapacitance per unit area and represents the instantaneous capacitance of the capacitor ateach point in the loop. The plot above is the derivative of the nested hysteresis loops on theprevious page.

0

50

100

150

200

250

300

-7.5 -5.0 -2.5 0.0 2.5 5.0 7.5

nCV vs Voltage[ Type AB WHITE, 10ms Period ]

uF/c

m^2

Voltage

Nested Loops: Polarization (µC/cm2): 1 Nested Loops: Polarization (µC/cm2): 2 Nested Loops: Polarization (µC/cm2): 3 Nested Loops: Polarization (µC/cm2): 4 Nested Loops: Polarization (µC/cm2): 5 Nested Loops: Polarization (µC/cm2): 6 Nested Loops: Polarization (µC/cm2): 7

Nested Loops: Polarization (µC/cm2): 8 Nested Loops: Polarization (µC/cm2): 9 Nested Loops: Polarization (µC/cm2): 10 Nested Loops: Polarization (µC/cm2): 11 Nested Loops: Polarization (µC/cm2): 12 Nested Loops: Polarization (µC/cm2): 13 Nested Loops: Polarization (µC/cm2): 14



15

Capturing Data During Loops

• The Vision data management software which controls theRadiant testers is capable of executing measurement tasksin a loop while changing one or more parameters in eachloop.

• The Vision Library provides data filters which may beinserted in a loop to read and then plot the measurementresults as a function of the loop parameters.

• The parameters plotted on the next few pages werecaptured in this manner during the same test executionthat captured the nested loops plotted on the previouspages.



16

Hysteresis Coercive Voltage vs Volts

-3

-2

-1

0

1

2

3

1 2 3 4 5 6 7 8 9

Coercive Voltage vs Volts[ Type AB WHITE ]

Coe

rciv

e V

olta

ge

Volts

Nested Loops: Vc (Volts) Nested Loops: -Vc (Volts)



17

Hysteresis Pr vs Volts

-40

-30

-20

-10

0

10

20

30

40

1 2 3 4 5 6 7 8 9

Pr vs Volts[ Type AB WHITE ]

Pol

ariz

atio

n

Volts

Nested Loops: Pr (µC/cm2) Nested Loops: -Pr (µC/cm2)



18

Hysteresis Pmax vs Volts

0

5

10

15

20

25

30

35

40

45

50

55

60

1 2 3 4 5 6 7 8 9

Pmax vs Volts[ Type AB WHITE ]

Pol

ariz

atio

n

Volts



19

Hysteresis vs Period

As the period of the hysteresis test gets longer, the hysteresis loop gets more square with alower coercive voltage. The red loop had a 1ms period. The blue loop had a10 secondperiod.

-40

-30

-20

-10

0

10

20

30

40

-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6

6V Hysteresis vs Period 1ms->10s[ Type AB WHITE ]

Pol

ariz

atio

n

Voltage

Hysteresis vs Period: Polarization (µC/cm2): 1 Hysteresis vs Period: Polarization (µC/cm2): 2 Hysteresis vs Period: Polarization (µC/cm2): 3

Hysteresis vs Period: Polarization (µC/cm2): 4 Hysteresis vs Period: Polarization (µC/cm2): 5



-100

-75

-50

-25

0

25

50

75

100

0 1 2 3 4 5 6 7 8 9

6V 1ms PUND[ Type AB WHITE ]

Pol

ariz

atio

n (µ

C/c

m2)

Pulse Top/Pulse Bottom Sequence

20

Standard PUND Results

The PUND test consists of five pulses where one polarization measurement is made at thetop of each pulse and one at the bottom of each pulse. The first pulse presets the remanentpolarization. The next four measure it. The PUND test simulates ferroelectric RAMoperation and is a quick way to measure remanent polarization.

Pulse 2 Pulse 3

Pulse 4 Pulse 5

P*

P*r

P^P^r

-P*

-P*r

-P^

-P^r



21

PUND vs Voltage(±P* vs ±P^ at 1ms Pulsewidth)

-100

-75

-50

-25

0

25

50

75

100

0 1 2 3 4 5 6 7 8 9

PUND P*/P^ vs Volts[ Type AB WHITE ]

Pol

ariz

atio

n

Volts

1ms PUND 0.2V->9V: P* (µC/cm2) 1ms PUND 0.2V->9V: P^ (µC/cm2)

1ms PUND 0.2V->9V: -P* (µC/cm2) 1ms PUND 0.2V->9V: -P^ (µC/cm2)



22

PUND vs Voltage(±dP at 1ms Pulsewidth)

-100

-75

-50

-25

0

25

50

75

100

0 1 2 3 4 5 6 7 8 9

PUND deltaP vs Volts[ Type AB WHITE ]

Pol

ariz

atio

n

Volts

1ms PUND 0.2V->9V: dP 1ms PUND 0.2V->9V: -dP

The “deltaP” value is the difference in polarization generated by a switching pulse and thefollowing non-switching pulse of the PUND test. It is the remanent portion of the hysteresisloop and is two times the value of the switchable spontaneous polarization.

dP = p* - P^-dP = (-P*) – (-P^)



23

PUND vs Pulsewidth(±P* vs ±P^ at 6V)

-100

-75

-50

-25

0

25

50

75

100

10-3 10-2 10-1 100 101 102 103

PUND P*/P^ vs Pulsewidth[ Type AB WHITE ]

Pol

ariz

atio

n

Pulsewidth (ms)

6V PUND 100us->1s: P* (µC/cm2) 6V PUND 100us->1s: P^ (µC/cm2)

6V PUND 100us->1s: -P* (µC/cm2) 6V PUND 100us->1s: -P^ (µC/cm2)



24

PUND vs Pulsewidth(±dP at 6V)

-100

-75

-50

-25

0

25

50

75

100

10-3 10-2 10-1 100 101 102 103

PUND deltaP vs Pulsewidth[ Type AB WHITE ]

Pol

ariz

atio

n

Pulsewidth (ms)

6V PUND 100us->1s: dP 6V PUND 100us->1s: -dP



25

Measuring the Remanent Hysteresis Loop

• The PUND test can be executed with half-hysteresis loopsinstead of pulses.

• The half-hysteresis loops can be subtracted from eachother to produce the remanent hysteresis loop, which isthe switching loop for the remanent polarization in thecapacitor.

• Measuring the remanent hysteresis loop eliminates thecontribution by the dielectric and paraelectric effects aswell as resistive leakage.

• The Remanent Hysteresis Task in Vision performs thismeasurement under automatic control.



26

Switching Half-Loops

The two plots above are the two switching half-loops measured after preset half-loops in theopposite direction. The measurement above appears to be a single loop but it is actually thepositive and negative measurements plotted against each other by aligning each with theorigin.

-50

-40

-30

-20

-10

0

10

20

30

40

50

-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6

Switching Half-Loops[ Type AB WHITE ]

Pol

ariz

atio

n (µ

C/c

m2)

Voltage

Switched - Logic 1



27

Non-switching Half-Loops

The non-switching half loops are plotted above so they line up with their respectiveswitching half-loops.

-50

-40

-30

-20

-10

0

10

20

30

40

50

-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6


Pol

ariz

atio

n (µ

C/c

m2)

Voltage

Unswitched - Logic 0



28

Switching vs Non-switching

Note how the shape of the return trace of the non-switching loops matches that of theswitching loops.

-50

-40

-30

-20

-10

0

10

20

30

40

50

-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6


Pol

ariz

atio

n (µ

C/c

m2)

Voltage

Unswitched - Logic 0 Switched - Logic 1



29

Remanent Hysteresis Loop

Subtracting the respective non-switching half-loops from their associated switching loopsgives the remanent hysteresis loop. The remanent polarization value of the remanenthysteresis loop can also be plotted against voltage and frequency.

-50

-40

-30

-20

-10

0

10

20

30

40

50

-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6


Pol

ariz

atio

n (µ

C/c

m2)

Voltage

Unswitched - Logic 0 Switched - Logic 1 Remanent



30

Nested Remanent Hysteresis Loops(10ms)

-50

-40

-30

-20

-10

0

10

20

30

40

50

-7.5 -5.0 -2.5 0.0 2.5 5.0 7.5

Rhyst vs Volts[ Type AB WHITE ]

Pol

ariz

atio

n

Voltage

Rhyst 10ms 1V->: Polarization (µC/cm2): 1 Rhyst 10ms 1V->: Polarization (µC/cm2): 2 Rhyst 10ms 1V->: Polarization (µC/cm2): 3 Rhyst 10ms 1V->: Polarization (µC/cm2): 4 Rhyst 10ms 1V->: Polarization (µC/cm2): 5

Rhyst 10ms 1V->: Polarization (µC/cm2): 6 Rhyst 10ms 1V->: Polarization (µC/cm2): 7 Rhyst 10ms 1V->: Polarization (µC/cm2): 8 Rhyst 10ms 1V->: Polarization (µC/cm2): 9



31

Remanent Hysteresis vs Period(6V from 1ms to 10s)

-50

-40

-30

-20

-10

0

10

20

30

40

50

-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6

Rhyst vs Period[ Type AB WHITE ]

Pol

ariz

atio

n

Voltage

Rhyst 6V 1ms->10s: Polarization (µC/cm2): 1 Rhyst 6V 1ms->10s: Polarization (µC/cm2): 2 Rhyst 6V 1ms->10s: Polarization (µC/cm2): 3

Rhyst 6V 1ms->10s: Polarization (µC/cm2): 4 Rhyst 6V 1ms->10s: Polarization (µC/cm2): 5



32

Leakage Through Capacitor(100ms Soak - 10s Integration)

The leakage through the capacitor is measured by applying a fixed voltage across thecapacitor and integrating over time the charge that passes through the capacitor. Normally,for PZT capacitors, the leakage decreases as the length of DC stress increases.

-1410

-1310

-1210

-1110

-1010

-910

-810

-710

-610

-510

-410

-310

-210

-110

010

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

DC Leakage in Amps @ 9V[ Type Ab WHITE ]

Cur

rent

(A

mps

)

Time (ms)



33

Remanent Polarization vs Leakage(100ms Soak - 10s Integration)

The state of the remanent polarization affects the leakage through the PZT capacitor. Thetwo leakage measurements above were taken with opposite DC Bias but with the remanentpolarization in the same direction for both measurements.

-1210

-1110

-1010

-910

-810

-710

-610

-510

-410

-310

-210

-110

010

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

Leakage @1V With and Against +9V Remanent State[ Type AB WHITE ]

Cur

rent

(am

ps)

Time(ms)

-1V 10s Leakage: Current (Amps) +1V 10s Leakage: Current (Amps)



34

Current vs Voltage(100ms Soak - 1s Integration)

The current leakage of the capacitor can be plotted as a function of DC bias as well as theremanent polarization direction over a range of DC bias values. There are two IVmeasurements plotted above: 1) switching and 2) non-switching. Note the hump in theleakage for the “switching” measurement as the test voltage approaches the coercive voltageof the hysteresis loop. Leakage is highest at the coercive voltage.

-1110

-1010

-910

-810

-710

-610

-510

-410

-310

-210

-110

010

-5 -4 -3 -2 -1 0 1 2 3 4 5

Switched vs Unswitched 1s IV[ Radiant Type AB BLUE ]

Cur

rent

(am

ps)

Volts

4V 1s nSW IV: Current (Amps) 4V 1s SW IV: Current (Amps)

The Type AB BLUEcapacitor is the samecomposition and hasthe same thickness asthe WHITE capacitor.The BLUE has an areaof 100,000 cm2, tentimes larger than theWHITE. The largearea increases the netcurrent through thedevice during leakagetests.



35

Small Signal Capacitance vs Voltage(100ms Soak – 1kHz Test)

The small signal capacitance test evaluates the capacitance of the sample using voltageamplitudes so small that the remanent polarization is not changed by the test. Like the IVtest, the small signal capacitance test can be executed over a range of voltages and it too isaffected by the remanent polarization.

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

-4 -3 -2 -1 0 1 2 3 4

1KHz SW vs nSW CV[ Radiant Type AB BLUE, 9V preset ]

uF/c

m^2

Volts

1ms 4V CV nSW: Capacitance (nF) 1ms 4V CV SW: Capacitance (nF)



36

IV vs Capacitance vs Hysteresis

By re-scaling each of the measurements and plotting them together on the same graph, therelationship between the leakage vs voltage, the small signal capacitance vs voltage, and thenCV of the remanent polarization vs voltage may be examined. Interesting!

0

25

50

75

100

125

150

175

200

225

-4 -3 -2 -1 -0 1 2 3 4

nCV vs ssCV vs IV[ Radiant Type AB BLUE ]

uF/c

m^2

and

uA

/cm

^2

Volts

Scale nCV of Rhyst: Single-Trace Math Filter: 1 Scale IVs x1E10: Single-Trace Math Filter: 1 Scale IVs x1E10: Single-Trace Math Filter: 1

Scale CVs x40: Single-Trace Math Filter: 1 Scale CVs x40: Single-Trace Math Filter: 1



37

Fatigue

Fatigue is defined as the loss of polarization from repeated cycling of the capacitor aroundits loop. Experience indicates that the domains must switch direction for fatigue to occur.Fatigue does not occur with mono-polar pulsing where remanent polarization does notswitch (non-switching half-loops). The 20/80 PZT of the Type AB capacitors fatiguesstrongly.

-40

-30

-20

-10

0

10

20

30

40

-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6

Hysteresis BEFORE and AFTER Fatigue[ Radiant Type AB WHITE ]

Pol

ariz

atio

n

Voltage

Hyst AFTER Fatigue: Polarization (µC/cm2)

Hyst BEFORE Fatigue: Polarization (µC/cm2)



Fatigue

From the nCV of the hysteresis loop before and after fatigue, it appears that the switchingpeak evaporated as the fatigue progressed. The expanded view on the right indicates thatthe linear capacitance and leakage, already small before the test began, changed little.

0

50

100

150

200

250

-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6

nCV BEFORE and AFTER Fatigue[ Radiant Type AB WHITE ]

uF/c

m^2

Voltage



0.0

2.5

5.0

7.5

10.0

12.5

15.0

-5.0 -2.5 0.0 2.5 5.0

nCV BEFORE and AFTER Fatigue[ Radiant Type AB WHITE ]

uF/c

m^2

Voltage





Fatigue

The remanent hysteresis before and after fatigue indicates that remanent polarizationdecreased substantially but some still exists after fatigue.

-30

-20

-10

0

10

20

30

-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6

Rhyst BEFORE and AFTER Fatigue[ Radiant Type AB WHITE ]

Pol

ariz

atio

n

Voltage

Rhyst AFTER Fatigue: Polarization (µC/cm2)

Rhyst BEFORE Fatigue: Polarization (µC/cm2)



Fatigue

The classic fatigue test monitors the PUND values as a function of cycles. This capacitorwas cycled with a 3kHz triangle wave at 6V to produce the fatigue effect. Notice that thenon-switching parameters ±P^ did not change much while the switching parameters ±P*changed quite a bit. It is primarily the remanent polarization that fatigues.

-100

-75

-50

-25

0

25

50

75

100

100 101 102 103 104 105 106 107 108 109 1010 1011

3kHz Triangle Fatigue @ 6V[ Radiant Type AB WHITE ]

Pol

ariz

atio

n

Cumulative Cycles

P* (µC/cm2) P^ (µC/cm2) -P* (µC/cm2) -P^ (µC/cm2)



Fatigue

It is primarily the remanent polarization that fatigues.

-100

-75

-50

-25

0

25

50

75

100

100 101 102 103 104 105 106 107 108 109 1010 1011

3kHz Triangle Fatigue @ 6V[ Radiant Type AB WHITE ]

Pol

ariz

atio

n

Cumulative Cycles

dP (µC/cm2) dPr (µC/cm2) -dP (µC/cm2) -dPr (µC/cm2)



Imprint

Imprint is a reliability issue with memory capacitors. Its primary mechanism is the gradualgrowth of an internal DC bias over time that shifts the hysteresis loop horizontally on thevoltage axis. It is accelerated by temperature. The capacitor above saw 2300 seconds at155°C between the blue loop and the red loop.

-40

-30

-20

-10

0

10

20

30

40

-7.5 -5.0 -2.5 0.0 2.5 5.0 7.5

Hysteresis Before and After 155C Imprint[ Type AB WHITE Unpackaged Die ]

Pol

ariz

atio

n

Voltage

Hystersis AFTER: Polarization (µC/cm2)

Hysteresis BEFORE: Polarization (µC/cm2)



Imprint

1x109 seconds is equal to 30 years. 1x108.5 seconds equals 10 years. The imprint driftoccurs constantly as long as the capacitor remains in the same remanent polarization state.In the data above, the capacitor was in the “down” state during the 155°C imprint periods.

-5.0

-2.5

0.0

2.5

5.0

100 101 102 103 104 105 106 107 108

Coercive Voltage Shift due to 155C Imprint[ Type AB WHITE Unpackaged Die ]

Vc

Cumulative Time (s)

Vc -Vc



Imprint

Despite the fact that the capacitor has shifted almost a full volt to the right, the remanentpolarization of the capacitor, measured periodically during the imprint period, shows nodegradation. Nevertheless, once the hysteresis shifts far enough, the remanent polarizationwill begin to degrade.

-100

-75

-50

-25

0

25

50

75

100

100 101 102 103 104 105 106 107 108

Change in Remanent Polarization due to 155C Imprint[ Type AB WHITE Unpackaged Die ]

Pol

ariz

atio

n

Cumulative Time (s)

dP (µC/cm2) dPr (µC/cm2) -dP (µC/cm2) -dPr (µC/cm2)



Imprint

The Imprint Figure of Merit is defined by Radiant as the voltage shift divided by the initial,unimprinted coercive voltage. When the hysteresis loop translates the distance of onecoercive voltage (i.e. FOM = 1), FeRAM type memories using such a capacitor will begin tofail. At that point, the written data becomes permanently “imprinted” in the memory.Accordingly, the capacitor above will just imprint in 30 years at 155C.

0.25

0.50

0.75

1.00

1.25

1.50

100 101 102 103 104 105 106 107 108

Figure of Merit vs Time[ Type AB WHITE Unpackaged Die ]

Vqf

Cumulative Time (s)



Retention

Retention is a simple test: apply a write pulse, wait for the retention time, and apply a readpulse. Repeat for longer periods. In the test plotted above, the capacitor retained the“down” state so that it switched up each time it was read.

0

10

20

30

40

50

60

70

80

90

100

100 101 102 103 104 105 106 107 108

Capacitor A Retention of Switched State[ Type AB WHITE ]

Pol

ariz

atio

n

Absolute Time (s)



Retention

Retention is a complex test because imprint occurs during the measurement. The storedstate determines the direction of imprint which shifts in opposite directions for oppositestates. Retention cannot be measured by running sequential retention tests in opposite stateson the same capacitor. The second test would start from an imprinted condition! Retentionof opposing states must be measured on two new, identical capacitors.

0

10

20

30

40

50

60

70

80

90

100

100 101 102 103 104 105 106 107 108

Capacitor B Retention of Non-switched State[ Type AB WHITE ]

Pol

ariz

atio

n

Absolute Time (s)



10

20

30

40

50

60

70

100 101 102 103 104 105 106 107 108

Switched/Unswitched Retention @ Room Temperature[ Type AB WHITE ]

Pol

ariz

atio

n

Absolute Time (s)

Retain Cap A HIGH: Pulse 1 Top Retain Cap B LOW: Pulse 1 Top

Task A - Task B: Pulse 1 Top

10

20

30

40

50

60

70

100 101 102 103 104 105 106 107 108

Switched/Unswitched Retention @ Room Temperature[ Type AB WHITE ]

Pol

ariz

atio

n

Absolute Time (s)

Retain Cap A HIGH: Pulse 1 Top Retain Cap B LOW: Pulse 1 Top

Task A - Task B: Pulse 1 Top

Retention

The high and low states can be subtracted to see the trend of the retained state over theuseful life of the memory.

Cap A High – Cap B Lo = Retention



49

Capacitor Parameters vs Temperature

• The previous measurements in this document demonstratethe multitude of parameters that define deviceperformance.

• Most of these parameters vary with temperature.

• In the next section of this document, the same properties,hysteresis, PUND, leakage, etc., will be plotted as afunction of temperature.

• All of the temperature-based measurements wereconducted with individual dice on the hot chuck of asemiconductor probe station.



50

Hysteresis vs Temperature(10ms Period, 9V)

-40

-30

-20

-10

0

10

20

30

40

-7.5 -5.0 -2.5 0.0 2.5 5.0 7.5

Hysteresis vs Temperature[ Type AB WHITE die ]

Pol

ariz

atio

n

Voltage

Check Hysteresis: Polarization (µC/cm2): 1 Check Hysteresis: Polarization (µC/cm2): 8

Check Hysteresis: Polarization (µC/cm2): 13

Blue = 30°C Green = 100°C Red = 150°CNote the offset hysteresis loop due to imprint. The capacitor being tested wasintentionally imprinted for an hour prior to this test so it would not move during the test.



51

Pmax vs Temperature(10ms Period, 9V)

0

10

20

30

40

50

60

70

80

0 50 100 150 200 250 300

Pmax vs Temperature[ Type AB WHITE die ]

Pol

ariz

atio

n or

Vol

ts

Temperature (C)

Check Hysteresis: PMax (µC/cm2)



52

Vc vs Temperature(10ms Period, 9V)

-5

-4

-3

-2

-1

0

1

2

3

4

5

0 50 100 150 200 250 300


Pol

ariz

atio

n or

Vol

ts

Temperature (C)

Check Hysteresis: Vc (Volts) Check Hysteresis: -Vc (Volts)

The ±Vc values are asymmetrical from the beginning because of the initial imprint conditionof the sample.



53

±Pr vs Temperature(10ms Period, 9V)

-50

-40

-30

-20

-10

0

10

20

30

40

50

0 50 100 150 200 250 300


Pol

ariz

atio

n or

Vol

ts

Temperature (C)

Check Hysteresis: Pr (µC/cm2) Check Hysteresis: -Pr (µC/cm2)



54

Remanent Hysteresis vs Temperature(10ms Period, 9V)

0

5

10

15

20

25

30

35

40

45

50

0 50 100 150 200 250 300

Remanent Hysteresis Parameters vs Temperature[ Type AB WHITE die ]

Pol

ariz

atio

n or

Vol

ts

Temperature (C)

Remanent Hysteresis: PMax (µC/cm2)



55

PUND vs Temperature(1ms Period, 9V)

-100

-75

-50

-25

0

25

50

75

100

0 50 100 150 200 250 300

PUND vs Temperature[ Type AB WHITE die ]

Pol

ariz

atio

n

Temperature (C)

PUND: P* (µC/cm2) PUND: P^ (µC/cm2)

PUND: -P* (µC/cm2) PUND: -P^ (µC/cm2)



56

PUND vs Temperature(1ms Period, 9V)

-100

-75

-50

-25

0

25

50

75

100

0 50 100 150 200 250 300

PUND vs Temperature[ Type AB WHITE die ]

Pol

ariz

atio

n

Temperature (C)

PUND: dP PUND: dPr

PUND: -dP PUND: -dPr



57

Leakage vs Temperature(1ms Period, 9V)

-1210

-1110

-1010

-910

-810

-710

-610

-510

-410

-310

-210

-110

010

0 50 100 150 200 250 300

Leakage vs Temperature[ Type AB WHITE die ]

Cur

rent

(A)

Temperature (C)



58

Pyroelectric Coefficients• Vision has one task, Chamber, that will measure the

pseudo-static true and false pyroelectric properties of acapacitor.

• The true pyroelectric coefficient is the change in remanentpolarization (Remanent Hysteresis or PUND) withtemperature.

• The false pyroelectric coefficient is the change in thedielectric constant (Small Signal Capacitance) withtemperature.

• The Chamber task stabilizes the sample at eachtemperature in the profile and measures the remanentpolarization and the small signal capacitance.



59

Dielectric Constant vs Temperature(1ms Period, 9V)

0

100

200

300

400

500

600

0 25 50 75 100 125 150 175 200

Remanent Polarization & Dielectric Constant vs Temperature[ Type AB White Die ]

Pol

ariz

atio

n or

uF/

cm

Temperature °C

Dielectric

The dielectric constant usually increases with increasing temperature.



60

Remanent Polarization vs Temperature(1ms Period, 9V)

The remanent polarization usually decreases with increasing temperature. The slope of thedecrease is the true pyroelectric coefficient for that material.

0

5

10

15

20

25

30

35

40

45

50

0 25 50 75 100 125 150 175 200

Remanent Polarization & Dielectric Constant vs Temperature[ Type AB White Die ]

Pol

ariz

atio

n or

uF

/cm

Temperature °C

PMax (µC/cm2)

-43.7 nC/cm2 / °K



61

Conclusion• There are a multitude of performance parameters for ferroelectric

capacitors.

• A capacitor’s performance will vary according to its composition,electrodes, and fabrication process flow.

• The data plotted in this document are contained in a single dataset:“typical type ab performance 07_07_08.dst” which can bedownloaded from www.ferrodevices.com/components2.html.

• For those without a Radiant tester, Vision can be downloaded fromwww.ferrodevices.com/tdownload.html. After installing Vision, youcan open datasets and plot the data they contain.

• Please contact Radiant by e-mail with any questions or to ask ouropinion of your own measurements. Remember to send us the datasetin question with the e-mail.

Documents

Typical Packaged AB Performance - ferrodevices.comTypical Type AB Performance 8 Typical Performance • Typical performance data for the 400µ2 “AB” packaged part are plotted on