Al₂O₃–Gd₂O₃ double-films grown on graphene directly by H₂O-assisted atomic layer deposition

Abstract

We demonstrate the direct Al₂O₃–Gd₂O₃ double-films growth on graphene by H₂O-assisted atomic layer deposition (ALD) using a hexamethyl disilazane precursor {Gd[N(SiMe₃)₂]₃}. No defects are brought into graphene as shown by Raman spectra; the surface root-mean-square (RMS) roughness of the Al₂O₃–Gd₂O₃ double-films is down to 0.8 nm, comparable with the morphology of pristine graphene; the films are compact and continuous, and the relative permittivity is around 11, which indicate that H₂O-assisted ALD can prepare high quality dielectric films on graphene.

---

Introduction

The outstanding electronic transport properties of graphene, including extremely high carrier mobilities (200 000 cm² V⁻¹ s⁻¹),^1–4 have generated significant interest in graphene-based nanoelectronics. In order to achieve graphene-based field-effect transistors (GFETs), the growth of high dielectric constant (high-κ) materials to act as the gate insulator is required.^5–7 Over the last decades, rare earth (RE) oxide based materials have been extensively studied because of their wide range of applications in optical devices, microelectronics, and magnetic devices.^8–10 The application of RE oxides as high-κ dielectrics in complementary metal oxide semiconductor (CMOS) devices has been investigated in detail, and Gd₂O₃ has shown great potential in this respect, exhibiting high dielectric constants (12–17), high thermal stabilities, and large band gaps (6.0 eV).^11–13 For applications in microelectronics, Gd₂O₃ films should be ultrathin, conformal, and pinhole-free with minimal disorder or traps. Atomic layer deposition (ALD) is the method of choice, as it allows excellent layer thickness control, low-temperature growth and uniform step coverage on complex device geometries. Nevertheless, due to the chemical inertness of graphene, the direct ALD growth of dielectric layers on bare graphene is rather challenging.

Several surface functionalization have been pursued by researchers to improve the uniformity of gate dielectric grown on graphene by ALD, including the deposition and oxidation of metal films, functionalization of graphene via ozone or nitrogen dioxide, and the spin-coating of polymer films as seeding layers.^14–18 However, these methods either introduce undesired impurities or break the chemical bonds in the graphene lattice, which results in significant degradation in carrier mobility.

In our previous work, we tried to deposit Al₂O₃ onto graphene directly by ALD.¹⁹ The Al₂O₃ films were compact and continuously covered graphene surface with a relative permittivity of 7.2 and a breakdown critical electrical field of 9 MV cm⁻¹. However, the permittivity of Al₂O₃ (7–9) is low, compared with RE oxide such as Gd₂O₃. Up till now, no related study on the investigation of direct Gd₂O₃ deposition onto graphene has been reported. This is an urgent issue to be addressed since it facilitates the achievement of graphene-based nano-electronic devices.

In this paper, we report on the direct H₂O-assisted ALD growth of both Gd₂O₃ films and Al₂O₃–Gd₂O₃ double-films on graphene surface without any functionalization. Raman spectra were performed to indicate both the thickness and quality of graphene before and after Gd₂O₃ films and Al₂O₃–Gd₂O₃ double-films deposition. Atomic force microscope (AFM) and high resolution transmission electron microscopy (HRTEM) were implemented to investigate the surface morphology and microstructure of Al₂O₃–Gd₂O₃ double-films on graphene, respectively. X-ray photoelectron spectroscopy (XPS) was applied to confirm all of elements present in the films. Spectroscopic ellipsometer and Capacitance–voltage (C–V) measurements were also carried out to show the optical and electrical quality of Al₂O₃–Gd₂O₃ double-films on graphene, respectively.

Experimental section

Graphene films were grown on Cu foil (0.025 mm, 99.8%) in a low pressure CVD system. During the graphene growth process, the quartz tube was maintained at 1050 °C for 60 min under the flow of 50 sccm H₂ and 10 sccm CH₄. After the growth, graphene samples were transferred onto Si substrates covered by 300 nm thickness of SiO₂. Acetone was used to remove PMMA and graphene was annealed at 200 °C for 3 hours under the flow of 10 sccm Ar and 50 sccm H₂ to remove the photoresist residue before ALD processes. The H₂ shielding could admirably prevent graphene from being oxidized by O₂ existing at the interface between graphene and substrates. The graphene flakes were monolayers characterized by Raman spectra and all the graphene samples were grown and transferred at the same condition.

The Gd₂O₃ films were deposited from Gd[N(SiMe₃)₂]₃ (purchased from J&K) and H₂O in a commercial ALD reactor (BENEQ TFS 200-124) maintained at a low level of base pressure by a vacuum pump (Adixen). Gd[N(SiMe₃)₂]₃ was preheated to 177 °C while H₂O was kept at room temperature. Four cycles of H₂O were introduced into ALD chamber before Gd₂O₃ films growth and acted as deposition sites on graphene. Details of the optimal H₂O dosage choice were discussed in our previous work.¹⁹ Nitrogen gas (99.999% in purity) was used as a carrier gas at a flow rate of 200 sccm. The deposition process involved a pulse of Gd[N(SiMe₃)₂]₃ for 1s followed by a 10s purge. Subsequently, a pulse of H₂O for 1s followed by a 10s purge was applied. This process was repeated 100 times at 200 °C. In this reaction, Gd₂O₃ formed and NH(SiMe)₂ was purged away (see eqn (1)).

2Gd[N(SiMe)₂]₃ + 3H₂O → Gd₂O₃ + 6NH(SiMe)₂ (1)

The thickness of Gd₂O₃ films was 20 nm confirmed by spectroscopic ellipsometer. For comparison, another sample of Gd₂O₃ films on graphene with an Al₂O₃ seed-like layer (Al₂O₃–Gd₂O₃ double-films) was also fabricated. Likewise, 4 cycles of pre-H₂O treatment were performed before Al₂O₃ growth and the Al₂O₃ seed-like layer (5 nm) was grown from Al(CH₃)₃ (TMA) and H₂O on graphene at 100 °C directly by ALD (see eqn (2)); subsequently, the chamber temperature was elevated to 200 °C and 11 nm of Gd₂O₃ films were deposited.

2Al(CH₃)₃ + 3H₂O → Al₂O₃ + 6CH₄ (2)

After growth of Gd₂O₃ films and Al₂O₃–Gd₂O₃ double-films, the samples were characterized by Raman spectra. In addition, AFM, TEM, XPS, spectroscopic ellipsometry and C–V measurements were also applied to estimate the dielectric films.

Results and discussion

Raman spectra were firstly utilized to determine both the quality and thickness of the graphene films with and without high-κ dielectrics growth. As shown in Fig. 1a, the pristine graphene had a weak D band peak at 1350 cm⁻¹, a G band peak at 1562 cm⁻¹, a sharp 2D band peak at 2675 cm⁻¹ with a full width at half maximum (FWHM) of 50 cm⁻¹ and an I_2D/I_G ratio greater than 1.3, indicating the graphene sample was monolayer with few defects. The weak D band peak was due to a little wrinkle generated during the transferring process. After high-κ dielectrics deposition, no raise of defect-related D-band was detected for both Gd₂O₃ (Fig. 1b) and Al₂O₃–Gd₂O₃ (Fig. 1c) samples, implying no defects or disorder were introduced into graphene by ALD processes. Raman spectra also revealed the phenomenon of G band blueshift and G peak upshift. The blueshift of G band was due to the compressive strain in graphene developed during the ALD process^20,21 and the upshift of G peak was due to the non-adiabatic removal of the Kohn anomaly from the Γ point.^22–24

Fig. 1 Raman spectra of graphene with and without dielectric films. (a) Pristine graphene. (b) Graphene with 20 nm Gd₂O₃ deposited at 200 °C. (c) Graphene with 5 nm Al₂O₃ deposited at 100 °C and 11 nm Gd₂O₃ deposited at 200 °C.

To determine whether an Al₂O₃ seed-like layer grown directly by ALD was beneficial for the quality of Gd₂O₃ films on graphene, samples were measured by AFM. As shown in Fig. 2a, the surface RMS roughness of Gd₂O₃ films directly deposited on graphene was 2.7 nm. Although Gd₂O₃ films could be grown directly on graphene with assistant of physically absorbed H₂O molecules, pin-holes were detective and the surface morphology was rough. In order to grow Gd₂O₃ films from Gd[N(SiMe₃)₂]₃ by ALD, the chamber temperature should be over 177 °C due to the high sublimation temperature of Gd[N(SiMe₃)₂]_3. However, the surface of graphene had no dangling bonds and high temperature prevented uniform distribution of physically absorbed H₂O molecules on graphene, which resulted in insufficient deposition sites for Gd₂O₃ and led to pinholes in Gd₂O₃ films. When an Al₂O₃ seed-like layer was deposited directly on graphene by ALD at 100 °C, the subsequent Gd₂O₃ films grown at 200 °C were continuous and pinhole-free. As shown in Fig. 2b, the surface RMS roughness of Al₂O₃–Gd₂O₃ double-films on graphene was down to 0.8 nm, which was comparable with the morphology of pristine graphene with a RMS roughness of 0.6 nm (Fig. 2c). As our previous work reported,¹⁹ 100 °C was suitable for direct ALD Al₂O₃ growth on graphene with 4 cycles of pre-H₂O treatment. In addition, Al(CH₃)₃ was more active than Gd[N(SiMe₃)₂]₃ and Al₂O₃ molecules were smaller than Gd₂O₃ molecules; it was more easy to directly deposit continuous Al₂O₃ than Gd₂O₃ onto graphene. Therefore, for Gd₂O₃ films directly deposited on graphene by ALD, an Al₂O₃ seed-like layer was feasible and it could enhance the property of Gd₂O₃ films.

Fig. 2 AFM images (500 × 500 nm²) of graphene with and without high-κ films. (a) 20 nm Gd₂O₃ deposited at 200 °C on graphene. (b) 5 nm Al₂O₃ deposited at 100 °C and 11 nm deposited Gd₂O₃ at 200 °C on graphene. (c) Pristine graphene.

Moreover, in order to further indicate the beneficial effect of an Al₂O₃ seed-like layer, HRTEM was performed to character the cross-sectional structure of Al₂O₃–Gd₂O₃ double-films on graphene deposited directly by H₂O-assisted ALD. As shown in Fig. 3, the thicknesses of Al₂O₃ and Gd₂O₃ films were 5 nm and 11 nm, respectively. The Al₂O₃ films were amorphous while parts of the Gd₂O₃ films were polycrystalline (red dotted circles in Fig. 3). The easy crystalline nature of Gd₂O₃ was due to its low crystallization temperature. It was particularly worth mentioning that in spite of the loose Al₂O₃ films, the Gd₂O₃ films were compact. During the ALD process, Al₂O₃ could act as a seed-like layer, which was beneficial to the compactness of the subsequent Gd₂O₃ films. In our previous study, for purpose of investigating whether the thickness of an Al₂O₃ seed-like layer could be reduced, we prepared five controlled samples with an Al₂O₃ seed-like layer of 1 nm, 2 nm, 3 nm, 4 nm and 5 nm, respectively. Through contrast experiments, we found that the Al₂O₃ seed-like layer could be reduced to 4 nm. Details of the discussion about how thin the Al₂O₃ seed-like layer can go could be found in our previous work.¹⁹

Fig. 3 A HRTEM image of Al₂O₃–Gd₂O₃ double-films on graphene. Al₂O₃ deposited at 100 °C was 5 nm and Gd₂O₃ deposited at 200 °C was 11 nm.

Fig. 4 demonstrates the results of XPS analysis to confirm all of elements present in the films. The binding energy was calibrated by centering the C 1s peak at 284.5 eV. The Gd 3d_5/2 peak located at 1187.6 eV (Fig. 4a), separated by 1045.8 eV from the Gd 4d peak located at 141.8 eV (Fig. 4b), indicating the existence of Gd³⁺. In addition to Gd³⁺, the Si 2p peak located at 103.4 eV was also detected. The existence of Si in Gd₂O₃ films was probably generated from the oxidizing of by-products in eqn (1) during the ALD process.

NH(SiMe)₂ + 2H₂O → SiO₂ + NH₅Me₂ (3)

Fig. 4 XPS analysis of Al₂O₃–Gd₂O₃ double-films on graphene: Gd 3d_5/2 (a), Gd 4d (b), Si 2p (c) and O 1s (d).

As shown in eqn (3), the by-product, NH(SiMe)₂, in eqn (1) could react with H₂O, leading to the formation of SiO₂ and this result could also be concluded from the analysis of O 1s peak illustrated in Fig. 4d. The O 1s peak was asymmetric and further deconvolution revealed three distinct components. The two stronger peak locating at 531.2 eV and 532.7 eV originated from Gd–O and Si–O bonds, respectively, and the weak peak at 529.6 eV associated with OH⁻ hydroxyl groups due to the incomplete reaction of Gd[N(SiMe₃)₂]₃/H₂O. The atomic ratio of Si/Gd is about 0.96, indicating high concentration doping of Si in Gd₂O₃ film.

Spectroscopic ellipsometry was applied for the investigation of the optical constants (refractive index n, absorption coefficient α, and complex permittivity ε) of Al₂O₃–Gd₂O₃ double-films on graphene. The refractive index n and extinction coefficient k of Al₂O₃–Gd₂O₃ double-films were directly obtained from ellipsometry measurements. As shown in Fig. 5a, the refractive index n of Al₂O₃–Gd₂O₃ double-films on graphene presented increasing tendency with increase of wavelengths λ and it was approximately 3.2 in the visible region. The absorption coefficients α was calculated from the extinction coefficients k by the following formula: α = 4πk/λ. As illustrated in Fig. 5b, the absorption coefficients α of Al₂O₃–Gd₂O₃ double-films demonstrated increasing tendency with increase of the photon energy from 1.0 eV to 3.0 eV (1100 nm to 400 nm) and had an absorption peak in the range of 3.0 eV to 5.0 eV (400 nm to 250 nm). The complex permittivity ε (ε = ε_r − iε_i) of Al₂O₃–Gd₂O₃ double-films on graphene could also be calculated from refractive index n and extinction coefficients k according to the following formulas: ε_r = n² − k² and ε_i = 2nk, where ε_r and ε_i are the real part and imaginary part of the complex permittivity, respectively (Fig. 5c). It was interesting to note that the real part of the complex permittivity ε_r was negative in the near ultraviolet region (from 220 nm to 400 nm), indicating that graphene/Al₂O₃–Gd₂O₃ might act as a meta material and applied to novel optical devices. The negative value of ε_r was due to the sudden enhancement of the extinction coefficients k in the near ultraviolet region as shown in the inset of Fig. 5c, and this enhancement was determined by the intrinsic properties of graphene/Al₂O₃–Gd₂O₃.

Fig. 5 Optical properties analysis of Al₂O₃–Gd₂O₃ double-films on graphene by spectroscopic ellipsometry: refractive index (a), absorption coefficient (b) and complex permittivity (c).

C–V measurements were also employed to estimate the electrical properties of Al₂O₃–Gd₂O₃ double-films on graphene. A metal–oxide–graphene (MOG) structure was adopted in the electrical analysis shown in Fig. 6a. The areas of each electrode were 0.01 mm² and the distance between two electrodes was 0.8 mm. The actual capacitance value was double of the measured one due to the series connection of two same capacitors. Fig. 6b showed the accurate model of MOG structure (left) and the series circuit model²⁵ (middle) employed in the C–V measurements. The impedance of a MOG capacitor was given by

where R_c included the ac conductance induced resistances arising from interface traps, and R_s included all of the series resistances. The impedance could also be obtained from the following formula when a series circuit model was performed:where R_m and C_m referred to measured values. Equating the imaginary parts of eqn (4) and (5), one could obtain

Fig. 6 (a) The MOG structure. (b) The accurate model (left) and the series model (right) of MOG structure. (c), (d) C–V measurements of Gd₂O₃ films and Al₂O₃–Gd₂O₃ double-films performed at different frequencies: 200 kHz (curve a), 500 kHz (curve b), and modified by dual-frequency method (curve c), respectively.

A dual-frequency method²⁶ was introduced to correct for series resistance effects and to determine accurately the capacitance of Al₂O₃–Gd₂O₃ films. Measuring the capacitance at two different frequencies, substituting into (6) for each frequency, subtracting, and solving for C, one could obtain.

where C₁ (C₂) referred to the values measured at the frequency f₁ (f₂).

The measured capacitance (C_m) was consisted of three components: the oxide capacitance (C_ox), the quantum capacitance of graphene (C_q) and the capacitance induced by interface states (C_it), as shown in Fig. 6b (right). C–V measurements of Gd₂O₃ films and Al₂O₃–Gd₂O₃ double-films were both implemented at two frequencies (200 kHz and 500 kHz) and modified by eqn (7) as shown in Fig. 6c and d, respectively. All the C–V measurements showed the expected broad V-shape induced by the quantum capacitance (C_q) of graphene,^27,28 indicating C_it did not play a major role in the measured capacitance (C_m), or the V-shape curve would not be detected. In addition, C_q was in parallel to C_it, while both of them were in series to the oxide capacitance (C_ox), and C_ox was approximately an order of magnitude lower than C_q at the bias away from the Dirac point of graphene. Thus, the measured capacitance at ±2 V was close to C_ox. The capacitances of Gd₂O₃ films and Al₂O₃–Gd₂O₃ double-films were 0.64 μF cm⁻² and 0.58 μF cm⁻², while the relative permittivities were 15 and 11, respectively. The permittivity reduction of Al₂O₃–Gd₂O₃ double-films was due to the relatively small permittivity of Al₂O₃ (7–9). However, the introduction of an Al₂O₃ seed-like layer was beneficial to both the surface morphology and compactness of subsequently deposited Gd₂O₃ films.

Conclusions

In summary, we have directly deposited Gd₂O₃ films and Al₂O₃–Gd₂O₃ double-films onto pristine graphene by ALD. No additional defects are introduced into graphene after high-κ films deposition. The introduction of an Al₂O₃ seed-like layer benefits both the surface morphology and compactness of subsequently deposited Gd₂O₃ films. The surface RMS of Al₂O₃–Gd₂O₃ double-films is down to 0.8 nm and the relative permittivity is around 11, which indicate high quality of high-κ films. This technique provides scientific guidance in fabricating novel graphene-based electronic devices.

Acknowledgements

This work is funded by the National Natural Science Foundation of China (Grant no. 11175229).We would like to thank Prof. Zengfeng Di, Dr Gang Wang, Dr Xiaohu Zheng and Dr Haoran Zhang for their generous help.

Notes and references

1. K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang, S. V. Dobonos, I. V. Grigorieval and A. A. Firsov, Science, 2004, 306, 666–669.
2. A. A. Balandin, S. Ghosh, W. Bao, I. Calizo, D. Teweldebrhan, F. Miao and C. N. Lau, Nano Lett., 2008, 8, 902–907.
3. M. D. Stoller, S. Park, Y. Zhu, J. An and R. S. Ruoff, Nano Lett., 2008, 8, 3498–3502.
4. C. Lee, X. Wei, J. W. Kysar and J. Hone, Science, 2008, 321, 385–388.
5. B. H. Lee, J. Oh, H. H. Tseng, R. Jammy and H. Huff, Mater. Today, 2006, 9, 32–37.
6. L. Wang, J. J. Travis, A. S. Cavanagh, X. Liu, S. P. Koenig, P. Y. Huang, S. M. George and J. S. Bunch, Nano Lett., 2012, 12, 3706–3710.
7. W. C. Shin, T. Y. Kim, O. Sul and B. J. Cho, Appl. Phys. Lett., 2012, 101, 033507.
8. J. C. G. Buenzil and C. Piguet, Chem. Soc. Rev., 2005, 34, 1048–1077.
9. J. Kido and Y. Okamoto, Chem. Rev., 2002, 102, 2357–2368.
10. J. Robertson, Eur. Phys. J.: Appl. Phys., 2004, 28, 2357.
11. W. H. Chang, C. H. Lee, Y. C. Chang, P. Chang, M. L. Huang, Y. J. Lee, C. Hsu, J. M. Hong, C. C. Tsai, J. R. Kwo and M. Hong, Adv. Mater., 2009, 21, 4970–4974.
12. A. P. Milanov, K. Xu, A. Laha, E. Bugiel, R. Ranjith, D. Schwendt, H. J. Osten, H. Parala, R. A. Fischer and A. Devi, J. Am. Chem. Soc., 2010, 132, 36–37.
13. K. Xu, R. Ranjith, A. Laha, H. Parala, A. P. Milanov, R. A. Fischer, E. Bugiel, J. Feydt, S. Irsen, T. Toader, C. Bock, D. Rogalla, H. Osten, U. Kunze and A. Devi, Chem. Mater., 2012, 24, 651–658.
14. M. J. Hollander, M. LaBella, Z. R. Hughes, M. Zhu, K. A. Trumbull, R. Cavalero, D. W. Snyder, X. J. Wang, E. Hwang, S. Datta and J. A. Robinson, Nano Lett., 2011, 11, 3601–3607.
15. B. Lee, G. Mordi, M. J. Kim, Y. J. Chabal, E. M. Vogel, R. M. Wallace, K. J. Cho, L. Colombo and J. Kim, Appl. Phys. Lett., 2010, 97, 043107.
16. Y. Xuan, Y. Q. Wu, T. Shen, M. Qi, M. A. Capano, J. A. Cooper and P. D. Ye, Appl. Phys. Lett., 2008, 92, 013101.
17. D. B. Farmer, H. Y. Chiu, Y. M. Lin, K. A. Jenkins, F. N. Xia and P. Avouris, Nano Lett., 2009, 9, 4474–4478.
18. Y. M. Lin, K. A. Jenkins, A. V. Garcia, J. P. Small, D. B. Farmer and P. Avouris, Nano Lett., 2009, 9, 422–426.
19. L. Zheng, X. Cheng, D. Cao, G. Wang, Z. Wang, D. Xu, C. Xia, L. Shen, Y. Yu and D. Shen, ACS Appl. Mater. Interfaces, 2014, 6, 7014–7019.
20. T. M. G. Mohiuddin, A. Lombardo, R. R. Nair, A. Bonetti, G. Savini, R. Jalil, N. Bonini, D. M. Basko, C. Galiotis, N. Marzari, K. S. Novoselov, A. K. Geim and A. C. Ferrari, Phys. Rev. B: Condens. Matter Mater. Phys., 2009, 79, 205433.
21. J. A. Robinson, M. LaBella, K. A. Trumbull, X. Weng, R. Cavelero, T. Daniels, Z. Hughes, M. Hollander, M. Fanton and D. Snyder, ACS Nano, 2010, 4, 2667–2672.
22. S. Pisana, M. Lazzeri, C. Casiraghi1, K. S. Novoselov, A. K. Geim, A. C. Ferrari and M. Maur, Nat. Mater., 2007, 6, 198–201.
23. A. Das, S. Pisana, B. Chakraborty, S. Piscanec, S. K. Saha, U. V. Waghmare, K. S. Novoselov, H. R. Krishnamurthy, A. K. Geim, A. C. Ferrari and A. K. Sood, Nat. Nanotechnol., 2008, 3, 210–215.
24. H. Xu, Y. B. Chen, J. Zhang and H. L. Zhang, Small, 2012, 8, 2833–2840.
25. K. J. Yang and C. Hu, IEEE Trans. Electron Devices, 1999, 46, 1500–1501.
26. Z. Luo and T. P. Ma, IEEE Electron Device Lett., 2004, 25, 655–657.
27. T. Fang, A. Konar, H. Xing and D. Jena, Appl. Phys. Lett., 2007, 91, 092109.
28. J. Xia, F. Chen, J. Li and N. Tao, Nat. Nanotechnol., 2009, 4, 505–509.

---