Improved Resistive Switching Characteristics by Al2O3 Layers Inclusion in HfO2-Based RRAM Devices

(1)

ECS Solid State Letters, 2 (8) P63-P65 (2013) P63 2162-8742/2013/2(8)/P63/3/$31.00©The Electrochemical Society

Improved Resistive Switching Characteristics by Al

2

O

3

Layers

Inclusion in HfO

2

-Based RRAM Devices

Chun-Yang Huang,a_{Jheng-Hong Jieng,}a_{Wen-Yueh Jang,}b_{Chen-Hsi Lin,}b and Tseung-Yuen Tsenga,z

a_{Department of Electronics Engineering and Institute of Electronics, National Chiao Tung University,}

Hsinchu 30010, Taiwan

b_{Winbond Electronics Corporation, Hsinchu 30010, Taiwan}

A series of complex HfO2/Al2O3layer by layer resistive random access memory(RRAM) structure grown by atomic layer deposition

are investigated. The modulation of forming voltage can be achieved by controlling the number of Al2O3layers in HfO2devices.

In addition, the crystallization temperature of HfO2based RRAM devices can also be improved by insetting Al2O3layers in HfO2

film. Compared with pure HfO2device, a significant improvement in resistive switching properties such as forming voltage variation

and the distribution of HRS/LRS during resistance switching is demonstrated in the HfO2/Al2O3layer by layer devices. Moreover,

Manuscript submitted April 15, 2013; revised manuscript received May 20, 2013. Published May 31, 2013.

Resistive random access memory (RRAM) is a promising candi-date to replace the currently flash memory device due to its simple structure, low voltage operation, high scalability, and multibit data storage.1,2_{It is found that transition metal oxides (TMOs) can be}

uti-lized in RRAM devices, such as ZrO2,3,4NiO,5and HfO2.6–11Among

those various TMOs, HfO2 is one of the appealing materials that

had considerable attention owing to its high dielectric constant (k), superior resistive switching (RS) performance, and compatible stan-dard complementary metal oxide semiconductor (CMOS) technology process.6_{However, the thermal stability of HfO}

2thin film is a serious

issue for memory characteristics due to the low crystallization temper-ature (<400◦C).12_{The RRAM devices with crystalline phase HfO}

2

film suffer the RS behaviors variation due to location dependent con-ductive filament (CF) formation. For example, if CF grows along the grain boundaries of polycrystalline HfO2film, low forming voltage

and good RS behaviors are observed, whereas high forming voltage is detected in amorphous phase and crystalline phase HfO2 RRAM

devices.7 _{On the other hand, crystalline phase HfO}

2 film, which is

stoichiometric structure, prevents creating enough oxygen vacancies in RRAM devices for RS behaviors.8 _{Moreover, the high forming}

voltage crystalline phase HfO2 film may cause RRAM devices hard

breakdown during forming process. Besides, for the scaled RRAM devices, the larger non-uniformity grain boundaries for cell by cell with crystalline phase HfO2 film will cause the variation of device

performance.9_{Therefore, the amorphous phase HfO}

2exhibiting

uni-formity film quality is suitable for RRAM development. According to previous literature, the RS behaviors are dependent on the degree of crystalline of HfO2film, which critically influences the device yield.

Hence, this phenomenon of low crystallization temperature in HfO2

is not allowed existence in further RRAM applications.

In this letter, we fabricate HfxAlyO films, which are the

architec-ture with a series of complex HfO2/Al2O3layer by layer structure by

using atomic layer deposition system (ALD), for HfO2-based RRAM

devices. Here, we show the insetting Al2O3layers would significantly

increase the crystallization temperature of HfO2film. Moreover, the

forming voltage variation of HfO2film can be reduced by the insetting

Al2O3layers in the film during ALD deposition. The memory

perfor-mances such as endurance, retention, and multibit storage properties are also discussed.

Experimental

The 5-nm thin HfO2, HfxAlyO, and Al2O3 RS layers were

de-posited on Pt/Ti/SiO2/Si substrates by using ALD at 250◦C and 0.2

Torr Ar ambient with Hf[N(C2H5](CH3)]4, (CH3)3Al, and H2O

precur-sors. The HfxAlyO films were the t-series of complex m-cycle HfO2

z_E-mail:_{[email protected]}

layers and n-cycle Al2O3 layers structure, or{(HfO2)m/(Al2O3)n}t multilayer architecture. The HfO2 layers were firstly deposited and

then Al2O3 layers during every series deposition. For example, the

5-nm Hf0.7Al0.3O film was composed of t= 8 series of mixing m

= 6 cycles HfO2layers and n= 1 cycle Al2O3layer. The material

compositions of HfxAlyO films were modified by different m and n

values. Besides, all the films’ thicknesses controlled by t values were kept in 5-nm for comparison, as listed in TableI. Subsequently, the post deposition annealing (PDA) processes were carried out for X-ray diffraction (XRD) analysis at different temperatures in N2 ambient

for 30 s. Finally, a 50-nm thick Ti top electrode and a 20-nm thick Pt capping layer with a diameter of 150μm were deposited by electron beam evaporation. All the electrical characteristics were performed using an Agilent 4156C semiconductor parameter analyzer.

Results and Discussion

Fig.1areveals the crystallization temperature of HfxAlyO films as

a function of different Al percentage (Al%= y/[x + y]). By utilizing Al2O3 layer inclusion in HfO2 film, the crystallization temperature

increases with an increase of the Al content from 400◦C to 1200◦C. It can be explained that Al distributes uniformly in HfxAlyO films and Al

acts as a network modifier to suppress the crystallization of HfO2film.

The inset of Fig.1ashows the XRD patterns of HfO2and Hf0.7Al0.3O

films with as-deposited and 400◦C PDA processes, respectively. The Hf0.7Al0.3O film is still in amorphous state after 400◦C PDA process.

However, the HfO2 film shows crystalline phase. Besides, due to

the crystallization temperature improved by inserting Al2O3layers in

HfO2film, the as-deposited and 400◦C PDA Hf0.7Al0.3O samples both

exhibit amorphous state. A typical cross-section TEM image of the amorphous Ti/Hf0.7Al0.3O/Pt RRAM device is shown in Fig.1b.

The forming voltages (VF) in the HfO2and HfxAlyO RRAM

de-vices are also dependent on Al percentage, as shown in Fig.2a. Al incorporated in the HfO2devices induce the increasing of the VFfrom

2.4 V for HfO2device to 4.1 V for Al2O3device. On the other words,

the VFcan be modulated by different Al percentage inclusion in HfO2

RRAM devices. It is due to that the breakdown strength in dielectric material is dependent on (k)−1/2.13_{Besides, the dielectric constants}

in HfxAlyO RRAM devices are a function of Al percentage, which

Table I. Components of 5-nm {(HfO2)m/(Al2O3)n}t multilayer

architecture.

Composition HfO2 Hf0.7Al0.3O Hf0.55Al0.45O Hf0.11Al0.89O Al2O3

m-cycle 1 6 3 1 0

n-cycle 0 1 1 3 1

t-series 56 8 14 14 53

) unless CC License in place (see abstract). ecsdl.org/site/terms_use

address. Redistribution subject to ECS terms of use (see 140.113.38.11

(2)

P64 ECS Solid State Letters, 2 (8) P63-P65 (2013)

Figure 1. (a) Crystallization temperature of different Al percentage in HfxAlyO films. The inset shows XRD patterns of HfO2and Hf0.7Al0.3O films

with as deposited and 400◦C PDA, respectively. (b) Typical cross-section TEM image of the Ti/Hf0.7Al0.3O/Pt RRAM device.

k value decreases from∼2514 _(HfO

2) to ∼915 (Al2O3). Therefore,

the HfO2device shows the smallest VFthan other HfxAlyO devices.

The inset of Fig.2ashows the comparison of VFdistribution in HfO2

device with that in Hf0.7Al0.3O device. The Hf0.7Al0.3O device reveals

a much narrower VFdistribution than HfO2device. This phenomenon

is due to that the oxygen vacancies are easier to generate and assem-ble along Al atoms from bottom to top electrodes and the CF grows stably along Al atoms in RS layer.16_{In addition, the Ti top electrode}

contacts with Al2O3 layer in Hf0.7Al0.3O film. Comparing to HfO2,

the top thin Al2O3layer can enhance the electric field due to lower

dielectric constant.17_{The formation and rupture of CF can be confined}

stably at the thin Al2O3layer. Therefore, the statistical distributions

of high resistance state (HRS) and low resistance state (LRS) dur-ing resistance switchdur-ing cycles are greatly improved, as shown in Fig. 2b. Table II lists the comparison of RS properties of various HfxAlyO components. Obviously, the standard deviations of operation

voltages in multilayer structure are smaller than pure HfO2and Al2O3

devices. Besides, the I-V RS properties of Hf0.7Al0.3O, Hf0.55Al0.45O,

Figure 2. (a) Forming voltage of HfxAlyO devices as a function of Al

per-centage. The inset shows the statistical distributions of forming voltages of HfO2and Hf0.7Al0.3O devices. (b) Resistance distributions for 100 dc sweep

cycles of HfO2and Hf0.7Al0.3O devices. The resistances are measured at a

read voltage of 0.3 V.

and Hf0.11Al0.89O devices are almost the same(not shown here), but

Hf0.55Al0.45O and Hf0.11Al0.89O devices have higher forming voltages

with larger standard deviation. Therefore, we choose the Hf0.7Al0.3O

device for more detailed studies.

To further confirm the RS performance, the electrical properties of the Hf0.7Al0.3O device are also studied. Fig.3adepicts the endurance

characteristic of the Hf0.7Al0.3O device after post metal annealing

at 400◦C for 30 min in vacuum ambient. The resistance ratios of HRS/LRS can be well retained after more than 11000 switching cy-cles under set voltage (Vset) of 0.6 V and reset voltage (Vreset) of

−0.5 V applied on Ti top electrode. Fig.3bshows the read distur-bance property of the Hf0.7Al0.3O device under a positive voltage stress

(0.3 V) at room temperature. It is clearly shown that both HRS and LRS do not exhibit any degradation for more than 104_s.

According to the CF model, CF will form when a set voltage applied on the device.18_{The multistates in RRAM device are}

depen-dent on the CF size.5,19_{Therefore, the level of dissolution of CF can}

be achieved by different stopping voltage (Vstop) in the reset sweep.

Hence, the multibit storage operation of the Hf0.7Al0.3O device can

be achieved by changing Vstop, as shown in Fig.4a. In addition, to

Table II. Resistive switching characteristics of different components of HfxAlyO devices.μ is the mean value and σ is the standard deviation.

Composition HfO2 Hf0.7Al0.3O Hf0.55Al0.45O Hf0.11Al0.89O Al2O3

VF(μ/σ) 2.42 / 0.34 2.69 / 0.09 2.78 / 0.16 3.51 / 0.16 4.09 / 0.24

Vset(μ/σ) 0.92 / 0.11 0.69 / 0.03 0.99 / 0.05 0.75 / 0.05 1.25 / 0.21

|Vreset| (μ/σ) 0.65 / 0.08 0.52 / 0.04 0.77 / 0.09 0.58 / 0.04 0.62 / 0.07

(3)

ECS Solid State Letters, 2 (8) P63-P65 (2013) P65

Figure 3 (a) Endurance characteristic of Hf0.7Al0.3O device for 11000

switch-ing cycles. (b) Read disturbance behavior for the device at room temperature.

Figure 4 (a) Multibit characteristic obtained by different Vstopvalue. (b)

Re-tention characteristic of multibit storage states of Hf0.7Al0.3O device at room

temperature.

clearly identify the different states of multibit storages in nonvolatile memory application, the interval of each storage state should be large enough (∼10 times) to detect by operation system. Therefore, the resistance state of bit 1 or on state is achieved by Vsetat 1 mA

compli-ance current, and the other off states from bit 2 to bit 4 are achieved by changing Vstop from−1 V to −1.5 V and −2.5 V, respectively.

Fig.4bdemonstrates the retention measurement of multibit storages of the Hf0.7Al0.3O device with a read voltage of 0.3 V for memory

performance. All bits of storage are almost kept at the same resistance values without observable degradation after 104 _{s. Based on above}

results, the Hf0.7Al0.3O device exhibits more stable and uniform RS

characteristics than pure HfO2device.

Conclusions

The thermal stability can be improved by inserting Al2O3 layers

in HfO2 film. Base on experimental results, the crystallization

tem-perature and forming voltage of HfO2based RRAM devices can be

modulated by changing the number of Al2O3 layers in HfO2 film

during ALD deposition. In addition, the Hf0.7Al0.3O device shows

superior thermal stability and less variation of resistive switching op-erations than pure HfO2device. Especially, the device exhibits good

memory performances, including low operation voltage, reproducible endurance, reliable read disturbance, and multibit storage character-istics. Above results suggest that the Hf0.7Al0.3O device is promising

for next generation nonvolatile memory application.

Acknowledgment

This work is supported by National Science Council, Taiwan, under Project No. NSC 99-2221-E-009-166 - MY3.

References

1. R. Waser, R. Dittmann, G. Staiko, and K. Szot,Adv. Mater., 21, 2632 (2009). 2. J. J. Yang, D. B. Strukov, and D. R. Stewart,Nat. Nanotechnol., 8, 13 (2013). 3. M. C. Wu, Y. W. Lin, W. Y. Jang, C. H. Lin, and T. Y. Tseng,IEEE Electron Device

Lett., 32, 1026 (2011).

4. C. Y. Lin, C. Y. Wu, C. Y. Wu, T. C. Lee, F. L. Yang, C. Hu, and T. Y. Tseng,IEEE Electron Device Lett., 28, 366 (2007).

5. D. Ielmini, F. Nardi, C. Cagli, and A. L. Lacaita, in Proc. International Reliability Physics Symposium., 5D.1.1 (2010).

6. H. Y. Lee, P. S. Chen, T. Y. Wu, Y. S. Chen, C. C. Wang, P. J. Tzeng, C. H. Lin, F. Chen, C. H. Lien, and M. J. Tsai, Tech Dig Int Electron Device Meet., 297 (2008).

7. M. Lanza, G. Bersuker, M. Porti, E. Miranda, M. Nafria, and X. Aymerich,Appl. Phys. Lett., 101, 193502 (2012).

8. M. G. Sung, W. G. Kim, J. H. Yoo, S. J. Kim, J. N. Kim, B. G. Gyun, J. Y. Byun, T. W. Kim, W. Kim, M. S. Joo, J. S. Roh, and S. K. Park, in Proc. International Reliability Physics Symposium., 6B.5.1 (2011).

9. B. Govoreanu, G. S. Kar, Y-Y. Chen, V. Paraschiv, S. Kubicek, A. Fantini, I. P. Radu, L. Goux, S. Clima, R. Degraeve, N. Jossart, O. Richard, T. Vandeweyer, K. Seo, P. Hendrickx, G. Pourtois, H. Bender, L. Altimime, D. J. Wouters, J. A. Kittl, and M. Jurczak, Tech Dig Int Electron Device Meet. 729 (2011).

10. C. Walczyk, D. Walczyk, T. Schroeder, T. Bertaud, M. Sowinska, M. Lukosius, M. Fraschke, D. Wolansky, B. Tillack, E. Miranda, and C. Wenger,IEEE Trans. Electron Devices, 58, 3124 (2011).

11. Y. Y. Chen, L. Goux, S. Clima, B. Govoreanu, R. Degraeve, G. S. Kar, A. Fantini, G. Groeseneken, D. J. Wouters, and M. Jurczak,IEEE Trans. Electron Devices, 60, 1114 (2013).

12. A. Salaun, H. Grampeix, J. Buckley, C. Mannequin, C. Vallee, P. Gonon, S. Jeannot, C. Gaumer, M. G. Jean, and V. Jousseaume,Thin Solid Films, 525, 20 (2012). 13. J. W. McPherson, J. Kim, A. Shanware, H. Mogul, and J. Rodriguez,IEEE Trans.

Electron Devices, 50, 1771 (2003).

14. T. Bertaud, C. Bermond, T. Lacrevaz, C. Vallee, Y. Morand, B. Flechet, A. Farcy, M. Gros-Jean, and S. Blonkowksi,Microelectronic Engineering, 87, 301 (2010). 15. K. C. Chiang, C. H. Lai, A. Chin, T. J. Wang, H. F. Chiu, J. R. Chen, S. P. McAlister,

and C. C. Chi,IEEE Electron Device Lett., 26, 728 (2005).

16. S. Yu, B. Gao, H. Dai, B. Sun, L. Liu, X. Liu, R. Han, J. Kang, and B. Yu,Electrochem. Solid-State Lett., 13, H36 (2010).

17. Q. Q. Sun, J. J. Gu, L. Chen, P. Zhou, P. F. Wang, S. J. Ding, and D. W. Zhang,IEEE Electron Device Lett., 32, 1167 (2011).

18. S. Y. Wang, D. Y. Lee, T. Y. Tseng, and C. Y. Lin,Appl. Phys. Lett., 95, 112904 (2009).

19. M. C. Wu, W. Y. Jang, C. H. Lin, and T. Y. Tseng,Semicond. Sci. Technol., 27, 065010 (2012).