Thermal Annealing

8/3/2019 Thermal Annealing

1/15

Integrated Ferroelectrics, 95: 316, 2007

Copyright Taylor & Francis Group, LLC

ISSN 1058-4587 print / 1607-8489 online

DOI: 10.1080/10584580701755971

Effects of Rapid Thermal Annealing on

Microstructures and Properties of PZT-Pt/Ti Stacks

for MEMS Application

Jian Lu,1* Yi Zhang,1 Tsuyoshi Ikehara,1 Ryutaro Maeda,1

and Takashi Mihara2

1Networked MEMS Technology Group, National Institute of Advanced Industrial

Science and Technology (AIST), 1-2-1 Namiki, Tsukuba, Ibaraki, 305-8564, Japan2Future Creation Lab., Olympus Corporation, Shinjuku Monolith, 2-3-1

Nishi-Shinjuku, Shinjuku-ku, Tokyo, 163-0914, Japan

ABSTRACT

The effects of rapid thermal annealing (RTA) on microstructures, properties, and resid-ual stress of Pt/Ti electrodes and sol-gel derived lead zirconate titanate (PZT) films

were investigated. It was found that when heating rate was 3.5C/s, the Pt/Ti electrode

conglomerated to form a typical hillock surface morphology, companied by remarkable

degradations in both surface roughness and electric conductivity. When using 10.5C/s

as the heating rate, the Pt/Ti conglomeration and degradation can be effectively retarded.

Accordingly, the PZT film exhibited different properties at various RTA conditions. Fur-

thermore, Ti diffusion was found mainly happened in a short period after heat-treatment

started.

Keywords: Pt/Ti bottom electrode; Ti diffusion; PZT film; RTA; sol-gel; heating rate;AFM

1. INTRODUCTION

Micro electromechanical systems (MEMS) provide us a significant opportunity

to miniaturize the conventional sensors and actuators to micrometer or even

to nanometer range by thin film materials and silicon technology. Miniaturiza-

tion makes it possible to develop integrated devices or systems which com-bine both complementary metal-oxide-semiconductor (CMOS) circuits and


2/15

4/[310] Jian Lu et al.

micromechanical components, and then leads to attractive improvements in

device efficiency, functionality, performance, and stability [1]. Among var-

ious principles for device sensing and actuation, piezoelectric transductionusing lead zirconate titanate (Pb(Zrx ,Ti1x )O3, PZT) film is competitive be-

cause of the low power consumption, self-sensing self-actuation capability and

impedance-matching with electronic circuits [2, 3]. However, up to date, the

application of piezoelectric MEMS is still delayed by PZT integration diffi-

culties. One of the critical issues is the high-temperature annealing and the

resulting inter-diffusion, residual stress and property-degradation in bottom

electrode layer and PZT film, especially in sol-gel derived thick PZT films

[46]. Besides, different PZT electric, piezoelectric and mechanical properties

are required by various MEMS applications, e.g. micro actuator needs thickerPZT film with excellent piezoelectric properties for the pursuit of large driving

force and great displacement [7]. For micro resonators, Youngs modulus and

dielectric loss of the PZT film are crucial to obtain high device quality factor [8,

9]. It is another important issue need to be addressed for piezoelectric MEMS

integration.

Recent studies have suggested the possibilities to control the nucleation

and grain-growth behavior of the PZT film towards preferred film properties by

heat-treatment process [1012]. It is practically difficult because the underneath

Ti can easily diffuse into the Pt layer during high-temperature annealing. Thishas been demonstrated to degrade the electric properties of the Pt/Ti bottom

electrode as well as the PZT film by an interfacial layer having low dielectric

constant [13]. K. Sreenivas et al. studied the stabilities of the Pt/Ti bilayer met-

allizations in an oxidizing atmosphere and concluded that the thermodynamic

driving force for the Ti diffusion was the formation of oxide (TiOx ) in Pt-grain

boundaries [14]. The inter-diffusion can be effectively suppressed by an inter-

mediate oxidation treatment or by using other adhesive materials instead of Ti,

such as Zr, Ta, and TiO2 [15, 16]. However, the Ti is still attractive because

the out-diffused Ti can facilitate the formation of the perovskite nucleationsites on metallizations [13], which greatly enhances the hetero-nucleation at

the PZT and Pt/Ti interface and induces preferential PZT orientation by lattice

match. To promote the application of piezoelectric MEMS, it is essential to

understand the Ti diffusion behavior under different heat-treatment processes

and the effects of Ti diffusion on the properties of PZT-Pt/Ti stacks.

Compared with conventional furnace annealing (CFA), rapid thermal an-

nealing (RTA) can shorten the high-temperature annealing to the shortest pos-

sibility for PZT crystallization. RTA also offers many degrees of freedom,

i.e. heating rate, platform time and platform temperature, for PZT annealing.Changing the annealing temperature is a straight way to control the Ti diffusion,

but it is limited by the crystallization activation energy for phase transforma


3/15

Effects of RTA on PZT-Pt/Ti Properties [311]/5

Table 1

Sputter conditions for Ti and Pt deposition

DC Sub.

Power Ar Temp. Base Pres. Depo. Pres. Depo. Rate Thickness

(W) (sccm) (C) (Torr) (Torr) (nm/min) (nm)

Ti 100 25 300 6.7 107 1.5 103 4.3 50

Pt 100 25 300 6.7 107 1.5 103 17.8 200

layer and the sol-gel derived PZT film. The development of the Pt/Ti electric

conductivity, surface morphology and residual stress against numbers of heat-

treatment cycles were also studied. The essential aspects of the RTA which

affect Ti diffusion were discussed.

2. EXPERIMENTAL

(l00)-oriented silicon wafers with 2 m-thick thermal SiO2 layer on both sides

were used as the starting substrates. Then Ti (target purity: 99.99%) was de-posited by DC magnetron sputter on the SiO2 /Si substrate at temperature of

300C and Ar working pressure of 1.5 mTorr, followed by in situ sputter de-

position of Pt (target purity: 99.99%) at the same temperature and working

pressure. Table 1 lists the detailed sputter conditions. Since thin Ti (10 nm)

was found resulted in depletion of the interfacial bonding-layer causing serious

adhesion problems, whereas thicker Ti (100 nm) caused the encapsulation of

the Pt surface with an insulating TiO2 [14], the thickness of the Ti and Pt was

set at 50 nm and 200 nm respectively in this study.

PZT film with molecular ratio of Zr:Ti = 52:48 was prepared on thePt/Ti/SiO2/Si substrate using the sol-gel technique as we reported before [6].

After spin-coating, the obtained PZT wet film was subsequently dried at 120C

for 2 min and baked at 250C for 5 min to remove organic compounds, and

finally annealed under atmospheric conditions by RTA at 650C. Thick PZT

film can be obtained by repeating the above spin-coating and heat-treatment

processes. To investigate the effects of the RTA on PZT-Pt/Ti stacks, the heat-

ing rate during RTA was set at 10.5C/s and 3.5C/s (heating time from 20C

to 650C was 1 min and 3 min), and the platform time at 650C was set at

2 min and 4 min for different samples. The cooling rate after RTA was identicalto each sample. After deposition, some of the PZT films were remove by wet

chemical etching [17] to identify the effects of the heat treatment on properties


4/15


The film thickness was measured using a surface profiler (Dektak3, Sloan

Technology Inc.). The film microstructures were observed using an atomic

force microscope (AFM) (SPA501, Seiko Instruments Inc.). The electric con-ductivity of the Pt/Ti bottom electrode layer was measured using a 4-points

resistivity processor (-5, NPS Inc.). The residual stress of the Pt/Ti layer was

analyzed by a thin film stress measurement system (FLX-2320-S, Toho Tech-

nology Corp.). X-ray diffraction (XRD) (RINT 2000, Rigaku) measurement

was performed to investigate the effects of the RTA on Pt/Ti crystallization.

The dielectric properties and ferroelectric properties of the as-deposited PZT

film were characterized using an impedance analyzer (HP4294A, Hewlett-

Packard) and a standard ferroelectric test system (RT-60A, Radiant Tech.),

respectively. The piezoelectric properties of the as-deposited PZT film werecharacterized using a laser Doppler vibrometer (MLD-821, Neoark Inc.) and

a lock-in amplifier (LI5630, NF Inc.). The mechanical properties of the PZT

film were studied by a nanoindentation method (Tribolndenter, Hysitron).

3. RESULTS AND DISCUSSION

Figure 1 shows AFM images of the as-sputtered Ti adhesive layer (Fig. 1(a))

and Pt bottom electrode layer (Fig. 1(b)). The Ti layer exhibits dense, smoothand uniform morphology with nano-structured crystallite of about 20 nm in

diameter. The surface roughness Ra and RMS of the Ti layer over 25 m2 was

measured as 0.59 nm and 0.75 nm, respectively. After covering the Ti by a 200

nm-thick Pt, Ra and RMS increased to 1.30 nm and 1.60 nm due to larger Pt

crystallite of about 70 nm in diameter. XRD pattern reveals that the as-sputtered

Pt was strongly oriented in Pt (111) with 2 degree of 39.78. Sheet resistance

Ks of the Pt/Ti layer was 0.66 /sq..

Considering that 1 m is a commonly used PZT thickness for MEMS

device, 8 layers of PZT film (120 nm to 130 nm thick per layer) were preparedon above Pt/Ti bottom electrodes at different RTA conditions. Figure 2 shows

AFM images of the Pt/Ti layers after removing PZT film by wet chemical

etching (marked as PtTi-Al, PtTi-Bl and PtTi-Cl from Fig. 2(a) to Fig. 2(c)

respectively). It was found that when using 10.5C/s as the heating rate, the

Pt/Ti layers exhibited similar surface morphologies as the as-sputtered Pt/Ti

(Fig. 1(b)). The surface roughnesses Ra /RMS of PtTi-Al and PtTi-Bl over

25 m2 were 1.36 nm/1.70 nm and 1.51 nm/1.93 nm, and the sheet resistance

Ks of PtTi-Al and PtTi-Bl were 0.68 /sq. and 0.70 /sq., respectively. When

using 3.5

C/s as the heating rate, the PtTi-Cl tend to conglomerate to form ahillock surface morphology, companied by an increase in both Ra (1.6 nm),

RMS (2 01 nm) and K (0 75 /sq ) The Ti diffusion and oxidation along the


5/15


Figure 1. AFM image of the as-sputtered (a) Ti layer and (b) Pt layer. (See Color

Plate I)

underneath SiO2 layer due to dissociation. Accordingly, the reactions and their

thermodynamic driving force for such surface morphology change are expected

due to the Ti diffusion and oxidation. Figure 2 reveals that the Ti diffusion and

oxidation is more sensitive to the heating rate rather than platform time. When

keep the total annealing time at a fixed period (PtTi-Bl and PtTi-Cl), a higher

heating rate can suppress the Pt/Ti conglomeration and degradation to a great

extent.

Figure 3 shows AFM images of the Pt/Ti layers after 8 cycles of heat-

treatment without PZT film deposition (marked as PtTi-A2, PtTi-B2 and PtTi-C2 from Fig. 3(a) to Fig. 3(c), respectively). Table 2 lists all the measured

surface roughness electric conductivities and 2 degree of Pt (111) In this


6/15


Figure 2. AFM images of the Pt/Ti bottom electrode after remove 1 m-thick (8 layers)


7/15



8/15


Table 2

Electric properties, surface roughness, and 2 X-ray diffraction angle of the Pt/Ti

electrodes as shown in Figs. 2 and 3

Sheet resistance

Ks (ohm/sq.)

Surface

Ra (nm)

Surface

RMS (nm)

2 of Pt

(111)

As-Sputtered PtTi 0.66 1.30 1.69 39.78

Remove PZT

PtTi-Al 0.68 1.36 1.70 39.90

PtTi-Bl 0.70 1.51 1.93 39.88

PtTi-Cl 0.75 1.60 2.01 39.88

Without PZT

PtTi-A2 0.88 4.82 6.31 39.86

PtTi-B2 0.89 5.91 7.64 39.87

PtTi-C2 0.93 8.12 9.89 39.86

phenomenon is the significant Pt/Ti degradation in sheet resistance Ks from

original 0.66 /sq. to 0.880.93 /sq. It is believed due to move severe Ti

diffusion and oxidation because the Pt/Ti surface was surrounded by abundant

oxygen atoms in this case. However, in Fig. 3, similar dependence on heating

rate and platform time as shown in Fig. 2 can be observed. To reduce the effects

of the high-temperature annealing on Pt/Ti properties, higher heating rate is

recommended.The sheet resistance Ks of the Pt/Ti layer after different cycles of heat-

treatment without PZT film deposition was shown in Fig. 4. Ks was found

increased by 50% of the original value after the first cycle heat-treatment, and


9/15


Figure 5. Surface roughness Ra of the Pt/Ti layers against cycle times of the heat-

treatment without PZT film deposition.

then recovered gradually form 0.950.99 /sq. to 0.880.93 /sq. in the

following cycles. The formation of the insulating TiOx in Pt grain-boundaries

as well as on Pt/Ti surface [14, 15] is likely to deteriorate the Pt/Ti electric

conductivity. Figure 4 indicates that the Ti diffusion and oxidation mainly

happened in a short period after high-temperature annealing started. As the

heat-treatment proceeded, the Pt re-crystallized, the out-diffused Ti enhanced

the Pt grain growth [18], and then the Ks recovered by a few percent. In

addition, the Ks of PtTi-A2 and PtTi-B2 were found 0.3 /sq. lower than that

of PtTi-C2. This is well consistent with our above discussion on the heatingrate and platform time effects.

Surface roughness of the bottom electrode layer has been proved to affect

the ferroelectric properties of the PZT film [19]. Figure 5 shows Ra of the

Pt/Ti layer after different cycles of heat-treatment process without PZT film

deposition. Half of the Ra degradation happened after the first cycle heat-

treatment for all the Pt/Ti layers. This exhibits the similar Ti diffusion and

oxidation behavior as indicated in Fig. 4. Differently, Ra of the Pt/Ti layer

was almost saturated in following cycles when using 10.5C/s as the heating

rate. When using lower heating rate of 3.5

C/s, Ra increased gradually as theheat-treatment proceeded. Their effects on PZT films will be clarified later in

this paper


10/15


Figure 6. Residual stress of the Pt/Ti layers against cycle times of the heat-treatment

without PZT film deposition.

electrode layer rather than PZT itself [6]. Figure 6 shows the development of thePt/Ti residual stress with the numbers of heat-treatment cycles. All the samples

exhibit tensile stress of about 320 MPa and increased notably after the first cycle

heat-treatment, and then saturated at 700900 MPa when the heat-treatment

proceeded. Accordingly, 2 degree of Pt (111) increased to 39.8639.90

(listed in Table 2). The Pt/Ti residual stress strongly depends on the duration

of high-temperature annealing in each cycle. Furthermore, figure 6 supported

the stress balance method we suggested before [6].

To identify the effects of the RTA and the Pt/Ti degradation on PZT film,

AFM images of the 1 m-thick PZT films were shown in Fig. 7. The filmshowed dense microstructure with uniform crystallite diameters of 34 nm in

Fig. 7(a) (marked as PZT-a) and 40 nm in Fig. 7(b) (marked as PZT-b). While

in Fig. 7(c) (marked as PZT-c) when heating rate was 3.5C/s, large crystallites

with diameter of 200 to 300 nm and height of 20 to 30 nm were grown and

distributed arbitrarily among those small crystallites of 36 nm in diameter.

The Ra for PZT-a, PZT-b and PZT-c over 25 m2 was measured as 0.65 nm,

0.54 nm and 10.1 nm. The RMS for PZT-a, PZT-b and PZT-c was measured as

0.86 nm, 0.69 nm and 13.9 nm. The effects of the heating rate and Pt/Ti electrode

on PZT morphologies and microstructures were clear. Under high annealingtemperature, PZT film tends to grow towards minimum surface energy. Figure 7

indicates that by properly lengthening the platform time during RTA surface


11/15


Figure 7. AFM images of the as-deposited 1 m-thick PZT films annealed at different

RTA conditions: (a) PZT-a: heating rate 10.5C/s, platform time 2 min; (b) PZT-b:


12/15


Figure 8. PE hysteresis loops of the PZT films annealed at different RTA condi-

tions.

microstructures in PZT-c always makes PZT cracking as we found in our

experiments. During PZT deposition, hetero-nucleation and homo-nucleation

coexisted and competed with each other. The investigations on PZT nucleationand growth behavior at different RTA conditions are essential to understand the

results in Fig. 7, while its still undergoing and will be discussed in our future

publications.

Figure 8 shows P E hysteresis loops of the PZT films annealed at the

different conditions. The electric, piezoelectric and mechanical properties of

the PZT films were summarized in Table 3 for comparison. Clearly, the Ti diffu-

sion and oxidation at lower heating rate (PZT-c) resulted in poor electric quality

and poor surface roughness of the bottom electrode, and then degraded the re-

manent polarization (Pr ) and the dielectric constant () of the PZT film. A.I.Mardare et al. observed similar results when PZT film was deposited by laser

ablation technique [20]. Besides, large crystallite in PZT-c is preferred to ob-

tain highpiezoelectric coefficient d33. However, such an asymmetric structure

Table 3

Electric, piezoelectric, mechanical properties and surface roughness of the

as-deposited PZT films as shown in Fig. 7

Pr Ec d33 E Ra RMS (C/cm2) (kV/cm) (pm/V) (GPa) (nm) (nm)


13/15


apparently resulted in a low Youngs modulus E. This is likely to be an-

other reason for ferroelectric degradation since electric charge can easily trans-

fer through the grain boundary and induces large leakage-current in thosePZT films.

4. CONCLUSIONS

This paper studies the effects of RTA on PZT-Pt/Ti stacks. It was found that

the microstructures, electric and mechanical properties of the Pt/Ti bottom

electrode layer as well as the sol-gel derived PZT film strongly depend on Ti

diffusion and oxidation behavior. The Ti diffusion and oxidation mainly occursin a short time after heat-treatment starts and was more sensitive to heating rate

rather than platform time during RTA. To suppress the Pt/Ti degradation and to

yield better PZT ferroelectric, mechanical properties with a smoother surface

morphology, higher heating rate is recommended. In addition, it was found that

PZT film easier to crack at lower heating rate. The arbitrarily distributed large

crystallites might be the reason because of stress-accumulation.

REFERENCES

1. S. D. Lyshevski,Nano- and Microelectromechanical Systems. Boca Raton:

CRC Press; 2001.

2. D. L. Devoe and A. P. Pisano, Surface micromachined piezoelectric ac-

celerometers (PiXLs), Journal of Microelectromechanical Systems 10,

180186 (2001).

3. T. Kobayashi, J. Tsaur, and R. Maeda, Fabrication of optical micro scanner

driven by PZT actuators, Jpn J. App. L. Phys. 44, 70787082 (2005).

4. S. T. Kim, C. Y. Kim, K. H. Park, K. Y. Kim, J. S. Lee, Y. W. Jeong,

and H. J. Kwon, Study on microstructures and interdiffusion behaviour

in Pt/Ti/SiO2/Si and Pb(Zr,Ti)O3/Pt/Ti/SiO2/Si multilayer systems, Jpn.

J. Appl. Phys. 34, 49454949 (1995).

5. C. Millon, C. Malhaire, C. Dubois, and D. Barbier, Control of the Ti

diffusion in Pt/Ti bottom electrodes for the fabrication of PZT thin film

transducers, Materials Science in Semiconductor Processing 5, 243247

(2003).

6. J. Lu, T. Kobayashi, Y. Zhang, R. Maeda, and T. Mihara, Wafer scale lead

zirconate titanate film preparation by sol-gel method using stress balance

layer Thin Solid Films 515 15061510 (2006)


14/15


8. J. Lu, T. Ikehara, Y. Zhang, R. Maeda, and T. Mihara, Energy dissi-

pation mechanisms in lead zirconate titanate thin film transduced micro

cantilevers, Jpn. J. Appl. Phys. 45, 87958800 (2006).9. J. Lu, T. Ikehara, T. Kobayashi, R. Maeda, and T. Mihara, Quality Factor

of Micro Cantilevers Transduced by Piezoelectric Lead Zirconate Titanate

Film, Microsystem Technologies 13, 15171522 (2007).

10. C. K. Kwok and S. B. Desu, Formation kinetics of PZT thin films,

Journal of Material Research 29, 17281733 (1994).

11. D. Shim, J. Park, K. Nam, and G. Park, Effects of heat-treatment condi-

tions on electrical properties of sol-gel-derived ferroelectric Pb(Zr,Ti)O3thin films, Journal of the Korean Physical Society 47, S364S367 (2005).

12. T. Kobayashi, M. Ichiki, J. Tsaur, and R. Maeda, Effect of multi-coatingprocess on the orientation and microstructure of lead zirconate titanate

(PZT) thin films derived by chemical solution deposition, Thin Solid

Films 489, 7478 (2005).

13. S. T. Kim, H. H. Kim, M. Y. Lee, and W. J. Lee, Investigation of Pt/Ti

bottom electrodes for Pb(Zr,Ti)O3 films, Jpn J. Appl. Phys. 36, 294300

(1997).

14. K. Sreenivas, I. Reaney, T. Maeder, and N. Setter, Investigation of Pt/Ti

bilayer metallization on silicon for ferroelectric thin film integration,

J. Appl. Phys. 75, 232239 (1994).15. T. Maeder, L. Sagalowicz, and R. Muralt, Stablized platinum electrodes

for ferroelectric film deposition using Ti,Ta and Zr adhension layers, Jpn

J. Appl. Phys. 37, 20072012 (1998).

16. R. D. Klissurska, T. Maeder, K. G. Brooks, and N. Setter, Microstruc-

ture of PZT sol-gel films on Pt substrates with different adhesion layers,

Microelectronic Engineering 29, 297300 (1995).

17. K. Zheng, J. Lu, and J. Chu, A novel wet-etching process of Pb(Zr,Ti)O3thin films for applications in microelectromechanical system,Jpn J. Appl.

Phys. 43, 39343937 (2004).18. Z. Song, N. Chong, L. H. W. Chan, C. Choy, and C. Lin, Thermal stability

of electrode stacks for application in oxide film devices, Thin Solid Films

406, 268274 (2002).

19. H. S. Lee, W. S. Um, K. T. Hwang, H. G. Shin, Y. B. Kim, and K. H. Auh,

Ferroelectric properties of Pb(Zr, Ti)O3 thin films deposited on annealed

IrO2 and Ir bottom electrodes, J. Vac. Sci. Technol A, 17, 29392943

(1999).

20. A. I. Mardare, C. C. Mardare, E. Pinheiro, and E. Joanni, Bottom electrode

crystallization of PZT thin films deposited by laser ablation, Jpn J. Appl.Phys. 43, 15271531 (2004).


15/15

Thermal Annealing

Documents