• For Contributors +
• Journal Search +
Journal Search Engine
ISSN : 1225-0562(Print)
ISSN : 2287-7258(Online)
Korean Journal of Materials Research Vol.27 No.7 pp.362-366
DOI : https://doi.org/10.3740/MRSK.2017.27.7.362

# Improving Interface Characteristics of Al2O3-Based Metal-Insulator-Semiconductor(MIS) Diodes Using H2O Prepulse Treatment by Atomic Layer Deposition

Hogyoung Kim1, Min Soo Kim2, Sung Yeon Ryu2, Byung Joon Choi2
1Department of Visual Optics, Seoul National University of Science and Technology, Seoul 01811, Republic of Korea
2Department of Materials Science and Engineering, Seoul National University of Science and
Technology, Seoul 01811, Republic of Korea
Corresponding author : hogyoungkim@gmail.com (H. Kim, Seoul Nat'l Univ. Sci. Thch.)
May 18, 2017 June 7, 2017 June 9, 2017

## Abstract

We performed temperature dependent current-voltage (I-V) measurements to characterize the electrical properties of Au/Al2O3/n-Ge metal-insulator-semiconductor (MIS) diodes prepared with and without H2O prepulse treatment by atomic layer deposition (ALD). By considering the thickness of the Al2O3 interlayer, the barrier height for the treated sample was found to be 0.61 eV, similar to those of Au/n-Ge Schottky diodes. The thermionic emission (TE) model with barrier inhomogeneity explained the final state of the treated sample well. Compared to the untreated sample, the treated sample was found to have improved diode characteristics for both forward and reverse bias conditions. These results were associated with the reduction of charge trapping and interface states near the Ge/Al2O3 interface.

## 초록

Seoul National University of Science and Technology

This is an Open-Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

## 1.Introduction

The modulation of Schottky barrier height(SBH) has been demonstrated by inserting a thin dielectric layer such as Al2O3,1,2) Ge3N4,3) GeOx,4) MgO,5) Si3N4,6) and TiO27) between the metal and Ge contacts. When the thin interfacial layer is present, the Fermi level of the metal is released toward the conduction band of Ge, alleviating the Fermi level pinning(FLP) effect and yielding a lower SBH. As a possible mechanism, it was proposed that the inserted insulating layer can block the electron wave function from metal to semiconductor and therefore decrease the number of metal induced gap states(MIGS).8) The dipole formed at the metal/semiconductor(MS) interface has been proposed to cause the potential drop to modulate the SBH.9) It was also proposed that the fixed charges in the non-ideal dielectric layers would cause an extra potential drop across the dielectric layer.10) As a method to deposit such interfacial layers, atomic layer deposition(ALD) can be used to obtain high-quality dielectric layer due to the accurate thickness control and reproducibility.

Lee et al. investigated the electronic passivation of a Ge (100) surface, via the chemisorption of H2O at room temperature(RT) using scanning tunneling microscopy (STM) and scanning tunneling spectroscopy(STS) and found that a majority of the surface is covered by H2O chemisorbed dimer (−OH and −H terminations of the dangling bonds on the dimer), single and double dangling bond sites.11) The H2O chemisorption can passivate the Ge surface by terminating dangling bonds. However, the coverage of H2O is limited at elevated temperature. As suggested by Papagno et al., the H2O adsorption on Ge (100) can be enhanced when the sample is cooled down at liquid nitrogen temperature.12) The density functional theory(DFT) calculations showed that the reaction of trimethylaluminum(TMA) is more favorable on the Ge- OH sites than on the Ge-H sites, although both reactions are favorable from a thermodynamic point of view.13) In either case, H2O prepulsing in ALD process can enhance the possibility to form thermally stable Ge-O-Al bonds.

Using in-situ H2O prepulsing on the bottom electrode prior to ALD, Lin et al. showed that the electrical performance of high-k metal-insulator-metal(MIM) capacitors can be improved.14) Swaminathan et al. reported that H2O prepulsing can be used to prepare less defective Ge metal-oxide-semiconductor(MOS) devices with ALD-Al2O3 as a gate insulator.15) In Au/Al2O3/InP metal-insulatorsemiconductor( MIS) Schottky structures, the increased SBH was observed when the Al2O3 thickness is larger than 5 nm.16) SBH larger than the Ge band gap has been observed, which was explained through the formation of an inversion layer at the NiGe/n-Ge interface.17) This observation was regarded to be technologically important because NiGe is an ideal Schottky source/drain material in Ge-based p-MOSFETs due to the absence of energy barrier between NiGe and p-Ge. Liu et al. demonstrated that when the thickness of amorphous-Ge reaches above 10 nm, the Al/amorphous-Ge/n-Ge shows ohmic characteristics, associated with electron hopping through localized states of amorphous-Ge layer as well as the termination of dangling bonds at the amorphous-Ge/n-Ge interface.18) Hence, it will be meaningful to investigate the physical properties governing the metal/n-Ge diodes with an Al2O3 interlayer (thickness greater than 10 nm). Here, we comparatively investigated the electrical properties of Au/ Al2O3/n-Ge MIS diodes with and without H2O prepulse treatment prepared by ALD.

## 2.Experimental

Sb-doped Ge (100) wafer (thickness: 500 μm, carrier concentration: 5 × 1015 cm−3) grown by Czochralski method, was used in this investigation. The wafer was cut into small pieces (about 5 × 10 mm2) and were loaded into an ALD chamber within a minute of the cleaning process to minimize exposure to air. Then, the deposition temperature ramped up to 250 °C for 5 min in an N2 ambient (200 SCCM and 100 mTorr). Before deposition, some of the pieces were subjected to H2O prepulse treatment for 3 min. Then Al2O3 thin film was deposited at 250 °C. TMA and deionized water were used as the precursors with a purging gas of nitrogen (N2). Pulse sequence was composed of TMA (1s), N2 purge (5s), H2O (0.5s) and N2 purge (25s). Using spectroscopic ellipsometer, the Al2O3 thickness was measured to be about 15 nm. After solvent cleaning, gold(Au) Schottky contacts with thicknesses of 50 nm were deposited by using radio-frequency (RF) magnetron sputtering through a shadow mask onto the Al2O3 layer of all the samples. For ohmic contacts, Al metal with a thickness of 150 nm was deposited and then In metal was rubbed over the entire back surface of the samples. Current-voltage (I-V) and capacitance-voltage (C-V) measurements were carried out with a Keithley 238 current source and an HP 4284A LCR meter. I-V measurements under various temperatures were performed using a hot chuck connected with a temperature controller.

## 3.Results and Discussion

The forward bias I-V characteristics measured at room temperature were analyzed based on the TE model.19) The effective barrier heights (φB) were determined to be 0.71 (± 0.06) and 0.70 (± 0.03) eV, for the samples with and without H2O prepulse treatment, respectively. The ideality factors were calculated to be 1.50 (± 0.19) and 1.93 (± 0.20), for the samples with and without H2O prepulse treatment, respectively. Note that the barrier height and the ideality factor for the Au/n-Ge Schottky diodes (i.e., without Al2O3 interlayer) measured at room temperature were found to be 0.59 (± 0.03) eV and 1.54 (± 0.23), respectively. This indicates that the insertion of Al2O3 interlayer mainly enhanced the barrier height. Fig. 1 shows the typical semi-logarithmic I-V plots for both samples measured at different temperatures, which reveal rectifying characteristics over the entire temperature range. For the sample with H2O prepulse treatment, the current values increase monotonically with increasing the temperature. In contrast, the current values do not increase monotonically for both the forward and reverse bias conditions for untreated sample. Similar results were observed in Au/ZnO Schottky diodes,20) which was associated with the fact that the surface compensation around the contact periphery due to acceptor-like adsorbates occurred to varying extents at each temperature.21) In this work, the configuration of Au/Al2O3 contact periphery (i.e., circles of a diameter of 300 μm) is the same for both samples. Hence, the compensation due to acceptor-like defects, if present, may occur near the Al2O3/Ge interface. The intersection of I-V curves under forward bias was attributed to the interfacial layer near the interface and the interface states and the effect of series resistance.22,23) These I-V behaviors in Fig. 1 indicate that the diode characteristics were improved after H2O prepulse treatment due to the suppression of interface defects at Al2O3/Ge interface.

The TE model was applied again to the forward bias I-V data in Fig. 1 to determine the barrier heights and the results of which are presented in Fig. 2(a). The increase of barrier height with increasing temperature was attributed to an inhomogeneous Schottky barrier.24) The relation between barrier height and temperature can be described as $ϕ B = ϕ ¯ B − q σ 0 2 / 2 k T$, where $ϕ ¯ B$ is a zero-bias mean barrier height and σ0 is a standard deviation. The σ0 value for the sample with prepulse treatment (0.104 V) was smaller than that for the sample without prepulse treatment (0.189 V), indicating modification of the local barrier height.

With the lateral barrier inhomogeneity, the modified Richardson plot can be given by(1)

$ln ( I 0 / T 2 ) − q 2 σ 0 2 / 2 k 2 T 2 = ln ( A A * * ) − q ϕ ¯ B / k T$
(1)

where I0 is the reverse bias saturation current, A is the contact area and A** is the Richardson constant. Fig. 2(b) shows plots of $ln ( I 0 / T 2 ) − q 2 σ 0 2 / 2 k 2 T 2$ vs. 1/kT. The intercepts at the ordinate produced modified Richardson constants of 125.1 and 344.5 Acm−2K−2, respectively, for the samples with and without H2O prepulse treatment. The value for the treated sample is comparable to the theoretical value of 140.0 Acm−2K−2 for n-type Ge, which implies that the TE model along with the barrier inhomogeneity can explain the current transport properties of the Au/Al2O3/n-Ge MIS diodes. As shown in Fig. 3, the C-V curves measured at 500 kHz reveals that the capacitance values for treated sample are lower than that for treated sample. This can be due to the reduced contribution of interface states to total capacitance for the treated sample. In addition, the flat band voltage(VFB) of prepulse treated sample seems to shift positively compared to that of untreated sample, associated with the compensation of positive charges in the oxide.25) Meanwhile, the dielectric constants of Al2O3 and GeO2 are known as 8-926) and 6.0,27) respectively. During the H2O prepulse treatment, Ge dangling bonds can be passivated, forming GeO2 layer. Hence, Al2O3 layer is grown on this GeO2 layer. When the very thin GeO2 layer is present between Al2O3 and Ge layers, the average dielectric constant is lowered compared to the case of pure Al2O3 layer. This will reduce the measured capacitance values. However, further investigation is required to clarify the exact mechanism.

Still now, the thickness of Al2O3 interlayer was not considered for the analysis. Considering the interfacial layer, the forward bias current density-voltage (J-V) characteristics for the MIS Schottky diodes can be described as follows19)

$J = A * * T 2 exp ( − ζ δ ) exp ( − q ϕ B / k T ) [ exp ( q V / n k T ) − 1 ]$
(2)

where n is the ideality factor, ζ (in eV) and δ (in Å) are the effective barrier and effective thickness of the interfacial layer, respectively. For values of V greater than 3kT/q, Eq. (2) can be expressed as(3)

$J ≈ A * * T 2 exp ( − ζ δ ) exp [ − q / k T ( ϕ B − V / n ) ]$
(3)

As shown in Fig. 4(a) for the treated sample, with plotting ln(J/T2) vs. 1/kT (Richardson plot) at different forward biases, a set of different Ea values (Ea = φBV/ n) can be obtained from the slopes of the Richardson plots. Then, the barrier height can be determined from the linear fitting to the Ea vs. VF plot shown in Fig. 4(b), which resulted in 0.61 eV. This value is similar to the barrier heights of 0.59 eV for the Au/n-Ge Schottky diodes from the I-V data. The results indicate that we can enhance the barrier height from ~0.6 to ~0.7 eV by adding an Al2O3 interlayer. Note that this method was not applied to the untreated sample because the current values do not vary monotonically as shown in Fig. 1(a). Because the barrier height was about 0.71 eV with considering the Al2O3 interlayer(effective barrier height), the effective energy barrier across the Al2O3 interlayer can be obtained using Eq. (2), which is given as $ζ = [ ( ϕ B e f f − ϕ B ) / ( δ × k T / q ) ] 2$. This effective barrier at room temperature was calculated to be about 0.07 eV.

Fig. 5(a) shows the cross-sectional scanning transmission electron microscope(STEM) image around the Al2O3 interlayer region for the treated sample, indicating that the Al2O3 interlayer was grown uniformly on Ge surface. The thickness of Al2O3 measured by TEM is approximately 15 nm. In order to investigate the distribution of each component near the Al2O3/Ge interface, energydispersive X-ray spectroscopy(EDS) mapping was conducted for the elements of Au, Al, O and Ge. From the EDS mapping images shown in Fig. 5(b), Al2O3 interlayer prepared with H2O prepulse treatment shows that diffusion of Ge and Au atoms into the Al2O3 interlayer is insignificant.

It was shown that the leakage characteristic of the near-stoichiometric Al2O3 film is better than that of the oxygen-deficient one.25) The dominant bonding state on the H2O prepulsed Ge (100) prior to TMA exposure was found to be Ge-OH and this Ge-OH interfacial bonding is regarded to be beneficial in the reduction of charge trapping and fast interface state density, which eventually enhanced the concentration of adsorbed Al and thermally stable Ge-O-Al bonds serving as an ALD nucleation layer on Ge (100) surface.15) Consequently, the oxygendeficient Al2O3 before H2O prepulse treatment was transformed into near-stoichiometric Al2O3 and the quality of the Al2O3 film was improved after H2O prepulse treatment.

## 4.Conclusion

Using current-voltage (I-V) measurements, the temperature dependent electrical properties of Au/Al2O3/n-Ge MIS diodes prepared with and without H2O prepulse treatment by ALD were investigated. Compared to the untreated sample, the treated sample was found to have improved diode characteristics for both forward and reverse bias conditions. This was associated with the reduction of charge trapping and interface states near the Ge/Al2O3 interface. Even though the thickness of Al2O3 interlayer was 15 nm, the TE model with barrier inhomogeneity explained well for the treated sample. Considering the thickness effect of Al2O3 interlayer, the barrier height was found to be 0.61 eV, similar to those from the Au/n- Ge Schottky diodes. The result suggests that MIS diodes with improved performances were obtained using H2O prepulse treatment.

## Acknowledgments

This study was supported by the Research Program funded by the Seoul National University of Science and Technology (Seoultech).

## Figure

Semilogarithmic current-voltage (I-V) characteristics for the Au/n-Ge Schottky diodes (a) without and (b) with H2O prepulse treatment.

(a) Barrier height vs. temperature and (b) modified Richardson plot of In (I0/T2)−q2σ02/2k2T2 vs. 1/kT

Capacitance-voltage (C-V) characteristics for both samples measured at 500 kHz.

(a) Richardson plot of the H2O prepusle treated sample under different forward biases and (b) plot of the activation energy (Ea) vs. the corresponding forward biases.

(a) Cross-sectional scanning transmission electron microscope (STEM) image across the Al2O3 interlayer and (b) energydispersive X-ray spectroscopy (EDS) color mappings for each element scanned over the Au/Al2O3/Ge full structure.

## Reference

1. Zhou Y , Ogawa M , Bao M , Han W , Kawakami R , Wang K (2009) Appl. Phys. Lett, Vol.94 ; pp.242104
2. Coss B , Smith C , Loh W , Majhi P , Wallace R , Kim J , Jammy R (2011) IEEE Electron Device Lett, Vol.32 ; pp.862
3. Lieten R , Degroote S , Kuijk M , Borghs G (2008) Appl. Phys. Lett, Vol.92 ; pp.022106
4. Nishimura T , Kita K , Toriumi A (2008) Appl. Phys. Exp, Vol.1 ; pp.051406
5. Lee D , Raghunathan S , Wilson R , Nikonov D , Saraswat K , Wang S (2010) Appl. Phys. Lett, Vol.96 ; pp.052514
6. Kobayashi M , Kinoshita A , Saraswat K , Wong H , Nishi Y (2009) J. Appl. Phys, Vol.105 ; pp.023702
7. Tsui B , Kao M (2013) Appl. Phys. Lett, Vol.103 ; pp.032104
8. Nishimura T , Kita K , Toriumi A (2007) Appl. Phys. Lett, Vol.91 ; pp.123123
9. Wager J , Robertson J (2011) J. Appl. Phys, Vol.109 ; pp.094501
10. Hu J , Nainani A , Sun Y , Saraswat K , Wong H (2011) Appl. Phys. Lett, Vol.99 ; pp.252104
11. Lee J , Kaufman-Osborn T , Melitz W , Lee S , Kumme A (2011) Surf. Sci, Vol.605 ; pp.1583
12. Papagno L , Frankel D , Chen Y , Caputi L , Anderson J , Lapeyre G (1991) Surf. Sci, Vol.248 ; pp.343
13. Lee J , Kaufman-Osborn T , Melitz W , Lee S , Delabie A , Sioncke S , Caymax M , Pourtois G , Kummel A (2011) J. Chem. Phys, Vol.135 ; pp.054705
14. Lin C , Chen Y , Lee C , Chang H , Chang W , Chang H , Liu C (2011) J. Electrochem. Soc, Vol.158 ; pp.H128
15. Swaminathan S , Oshima Y , Kelly M , McIntyre P (2009) Appl. Phys. Lett, Vol.95 ; pp.032907
16. Zheng S , Yang W , Sun Q , Chen L , Zhou P , Wang P , Zhang D , Xiao F (2013) Appl. Phys. Lett, Vol.103 ; pp.261602
17. Chi D , Lee R , Chua S , Lee S , Ashok S , Kwong D (2005) J. Appl. Phys, Vol.97 ; pp.113706
18. Liu H , Wang P , Qi D , Li X , Han X , Wang C , Chen S , Li C , Huang W (2014) Appl. Phys. Lett, Vol.105 ; pp.192103
19. Sze S (1981) Physics of Semiconductor Devices, Wiley,
20. Von Wenckstern H , Muller S , Giehne G , Hochmuth H , Lorenz M , Grundmann M (2010) J. Electron. Mater, Vol.39 ; pp.559
21. Kim H , Sohn A , Kim D (2012) Semicond. Sci. Technol, Vol.27 ; pp.035010
22. Pakma O , Serin N , Serin T , Altlndal Semicond (2008) Sci. Technol, Vol.23 ; pp.105014
23. Osvald J , Horvath Zs J (2004) Appl. Surf. Sci, Vol.234 ; pp.349
24. Tung R (2001) Mater. Sci. Eng. R, Vol.35 ; pp.1
25. Garg R , Misra D , Swain P (2006) J. Electrochem. Soc, Vol.153 ; pp.F29
26. Robertson J , Wallace R (2015) Mater. Sci. Eng, Vol.R 88 ; pp.1
27. Gu J , Liu Y , Xu M , Celler G , Gordon R , Ye P (2010) Appl. Phys. Lett, Vol.97 ; pp.012106