Characterization of Fe-doped In-Sb-Te (Fe

Characterization of Fe-doped In-Sb-Te (Fe: 10 at.%) material with individual electrical- phase-change and magnetic properties

Young Mi Lee, Dang Duc Dung, Sunglae Cho, Min Sang Jung, Duck Kyun Choi, Docheon Ahn, Min Kyu Kim, Jae-Young Kim, and Min-Cherl Jung

Citation: AIP Advances 1, 022150 (2011);

View online: https://doi.org/10.1063/1.3609265

View Table of Contents: http://aip.scitation.org/toc/adv/1/2 Published by the American Institute of Physics

Characterization of Fe-doped In-Sb-Te (Fe: 10 at.%) material with individual electrical-phase-change and magnetic

properties

Young Mi Lee,¹Dang Duc Dung,¹Sunglae Cho,¹Min Sang Jung,²Duck Kyun Choi,²Docheon Ahn,³Min Kyu Kim,³Jae-Young Kim,³and Min-Cherl Jung^4,a

1Energy Harvest-Storage Research Center (EHSRC) and Dept. of Physics, University of Ulsan, 680-749, Republic of Korea

2Department of Material Science and Engineering, Hanyang University, Seoul 133-791, Republic of Korea

3Pohang Accelerator Laboratory, Pohang University of Science and Technology (POSTECH), Pohang 790-784, Republic of Korea

4Center for Atomic Wires and Layers, Dept. of Physics, Pohang University of Science and Technology (POSTECH), Pohang 790-784, Republic of Korea

(Received 12 May 2011; accepted 9 June 2011; published online 28 June 2011)

We propose a new electrical-phase-change magnetic material, namely Fe-doped In- Sb-Te (FIST), for possible non-volatile multi-bit memory applications. FIST was formed by typical co-sputter method with Fe 10 at.% doping in In₃Sb₁Te₂. FIST offers the electrical-phase-change and magnetic properties by way of the change of In 4d chemical bonding density and embedded Fe nanoclusters with the size of 4∼5 nm, respectively. It maintained the amorphous phase on the electrical-phase- change. Chemical state of In was only changed to increase the density of In-In chemical bonding during the electrical-phase-change without Fe nanoclusters con- tribution. Also, the magnetic property by Fe nanoclusters was not changed by the electrical-phase-change. On this basis, we propose the FIST material with the indi- vidual electrical-phase-change and magnetic properties for the multi-bit nonvolatile memory materials. Copyright 2011 Author(s). This article is distributed under a Creative Commons Attribution 3.0 Unported License. [doi:10.1063/1.3609265]

I. INTRODUCTION

Chalcogenides are basic materials suitable for rewritable optical media and non-volatile random access memory device (PRAM) applications.^1–5The main and key properties of these phase-change materials are the extreme changes of both optical reflectivity and electrical resistance during the amorphous-to-crystalline phase-change occurring within the 100∼400^oC temperature range.^4,5 Ge2Sb2Te5ternary alloy (GST) is used especially widely as the main material in both applications, owing to its proper phase-change temperature of 180^oC and its high phase-change speed.^4,5 Many studies have been undertaken to improve the application properties of the phase-change temperature and speed by means of dopants and variable stoichiometry.^6–8 The GST amorphous-to-crystalline phase-change mechanism, with Ge atom movement from the tetrahedral to the octahedral site, or with an Sb atom bonding change from the anisotropic (3-short bonds) to the isotropic (3-short and 3-elongated bonds) state, has been reported on in the literature.^4,9–11Samsung Electronics Co. Ltd.

officially announced several years ago a GST-based 512 Mb PRAM.¹²Recently as well, In-Sb-Te- based chalcogenide materials were found to have a multi-level structural phase-change property.¹³ Eun Tae Kim et al. reported that In₃Sb₁Te₂ (IST)-based PRAM showed the phase-change with

aElectronic mail:mcjung@postech.ac.kr

2158-3226/2011/1(2)/022150/9 1, 022150-1 ^C Author(s) 2011

022150-2 Leeet al. AIP Advances1, 022150 (2011)

three times as increasing the temperature. Studies such as these have demonstrated the highly promising applications of phase-change materials to multi-level optical media and multi-bit memory devices.

W.-D. Songet al.have commended the utility of Fe-doped Ge₂Sb₂Te₅ (FGST) phase-change magnetic material with multi-functional properties.¹⁴ They highlighted the variable properties in one material, along with a new degree of freedom associated with the spin carrier and resistance change.¹⁴ It can be associated with a growing list of possible and promising applications of both PRAM and spintronics. If this potential is to be realized to any significant degree, the research on phase-change magnetic materials will have to focus on key materials with multi-functional properties.

We here propose a new electrical-phase-change magnetic material, namely Fe-doped IST (FIST), for possible non-volatile multi-bit memory material applications. This material offers resistance change and magnetic information properties by way of electrical-phase-change and embedded Fe nanoclusters. In the present study, we observed this new electrical-phase-change system without the structural phase-change.

II. EXPERIMENTAL DETAILS

Fe-doped In₃Sb₁Te₂(FIST) was formed by co-sputtering with Fe and In₃Sb₁Te₂(IST) targets onto a Si substrate. The deposition rates of Fe and IST were 0.8 and 9.5 Å/sec, respectively. (Fe:

10 at.% in IST) The deposition power, time, and pressure were 500 W, 120 sec, and 10 mTorr, respectively. The thickness of the formed thin film was 100 nm. We measured the resistance as increasing temperature by using a halogen lamp in vacuum chamber with Ar gas under the pressure of 1 mTorr. The ratio of temperature in time was 3^oC/min. Prior to the synchrotron experiment, the FIST thin film was fabricated by Ne⁺ion mild sputtering in an ultra-high-vacuum chamber for 1 hour at the beam energy of 1 kV. This process removes native oxide from the film surface without the broken of stoichiometry.^6–8,10 Utilizing XPS measurement, we confirmed a clean FIST thin film lacking the O 1score-level peak.¹⁹In order to confirm the phases of the samples, their high-resolution synchrotron x-ray powder diffraction data were measured at 8C2 beamline of Pohang Accelerator Laboratory (PLS). The incident x-rays were vertically collimated by a mirror, and monochromatized to the wavelength of 1.5490 Å using a double-crystal Si(111) monochromator. The Fe L-edge absorption and core-level spectra were obtained by NEXAFS and HRXPS, respectively, at the 8A1 beamline of PLS. The magnetic properties were obtained using a superconducting quantum interface device (SQUID) magnetometer. The x-ray magnetic circular dichroism (XMCD) measurements were carried out at the 2A beamline of PLS. The energy resolution of incident light was set to 0.3 eV and the circular polarization ratio was higher than 95%. The spectra were obtained in total electron yield mode with an∼100 angstrom probe depth. All of the measurements were performed in a UHV environment so as to exclude surface contamination. An electromagnet was used to flip the 0.9 T magnetic field at every photon energy point. All of the spectra were normalized by the intensities of the incident photon beam, which were measured by means of an Au grid positioned in front of the XMCD chamber.

III. RESULTS AND DISCUSSION

A. Resistance change with XRD and TEM observation

The resistance change was clearly shown by the annealing process at the temperature of 49^oC (Fig.1(a)). The resistance was changed from 1.0×10³ to 2.0×10² /sq. Also, the resistance was almost unchanged with a temperature increase to 100^oC, and there was stable resistance of 2.0×

10² /sq. Before and after annealing, both of the A and B structural phases in Fig. 1(a) were amorphous measured by x-ray diffraction (XRD) with synchrotron radiation. Both the as-received (a-FIST_1) and resistance change samples (a-FIST_2) showed that the resistance change was not due to the amorphous-to-crystalline phase-change.

FIG. 1. (a) Sheet resistance with increasing temperature. We identified the resistance change temperaure as 49^oC. However, both structures are the amorphous phase. (b) and (c) are TEM images of thea-FIST_2 and high temperature process, respectively. After the annealing process, the size of the Fe nanocluster was unchanged. However, an Fe cluster size increase from 4∼5 to 22∼26 nm was evident after the high temperature process.

Figures 1(b) and 1(c) show transmission electron microscopy (TEM) images of a-FIST_2 and high temperature process, respectively. In Fig. 1(b), we found nanoclusters with the size of 4∼5 nm. Nanoclusters were randomly distributed in thea-FIST_2 sample with no change the size of a-FIST_1. It is very stable during the resistance change. We performed the high temperature annealing process to confirm the structural stability of nanoclusters. The high temperature process

022150-4 Leeet al. AIP Advances1, 022150 (2011)

entailed heating at over 300^oC for 10 seconds. However, after the high temperature process, the nanocluster size was observed to increase to 22∼26 nm, as shown in Fig.1(c). We assumed that nanoclusters had been made to combine with each other at the high temperature. Also, we confirmed that nanoclusters were not structurally stable at high temperature.

B. Atomic structure of nanoclusters

The nanoclusters’ atomic and chemical information was analyzed by near-edge x-ray absorption of fine structure (NEXAFS) and extended x-ray absorption of fine structure (EXAFS) experiments using synchrotron radiation. The x-ray absorption spectra of the FeL-edge for both thea-FIST_1 anda-FIST_2 are plotted in Fig.2(a). These show that the two peaks (L2andL3) shared the same shape and energy position, indicating thus that the chemical state of Fe is only-metallic Fe (i.e. Fe-Fe bonding). The EXAFS data in Figs.2(b)–2(d)constitutes additional evidence of only-metallic Fe bonding. We obtained the FeK-edge absorption spectra to confirm the points of difference among Fe foil,a-FIST_1, anda-FIST_2. However, the absorption spectra ofa-FIST_1 anda-FIST_2, as shown in Figs.2(b)and2(c), were identical. As shown in Fig.2(d)and as obtained by the Fourier transformation from the absorption spectra, we did not observe the other bondings except Fe-Fe bonding with 2.2 Å. In these TEM image results, we assumed that the nanoclusters were composed of only-Fe atoms. We confirmed that the formed thin film was the Fe-nanoclusters-embedded IST system.

C. Magnetic property of Fe nanoclusters

We investigated the magnetic properties of the samples using a superconducting quantum interference device (SQUID; Quantum Design) magnetometer. Figures 3(a) and 3(b) show the temperature-dependent magnetization (M-T) data under an applied magnetic field of 1000 Oe along with the magnetization hysteresis loops measured at the selected temperatures of 10, 110, 200, 250 and 300 K, respectively. The magnetization monotonically decreased with temperature, from 150 emu/cm³at 10 K to 45 emu/cm³at 300 K. At all of the measured temperatures, the film showed hysteric M vs. H curves with small coercive fields, which is characteristics of ferromagnetic ordering of Fe atoms. The spectra in Figs.3(c)and3(d)clearly show that the ferromagnetism of these samples originated from the Fe clusters. Even though there were several fine structures on the higher-energy side of the x-ray absorption spectra (XAS), which suggested the existence of non-metallically bonded Fe ions, they did not contribute to the x-ray magnetic circular dichroism (XMCD) spectra. Indeed, Figs.3(c)and3(d)show nearly identical line shapes for both, confirming that the magnetic origin was the Fe metal. We confirmed that the magnetic information appeared by Fe nanoclusters was not changed during the resistance change. Therefore, we assumed that the resistance change and magnetic properties were not correlated in this system.

D. Analysis of chemical states for the electrical-phase-change

The core-level spectra of Fe 3p, Te 4d, Sb 4d, and In 4dduring the resistance change are plotted in Figs. 4(a)–4(d). In the Fe 3p core-level spectra, we observed that the binding energies were 52.7 eV for botha-FIST_1 anda-FIST_2. This shows the typical metallic Fe-Fe bonding, corrobo- rating the TEM, NEXAFS and EXAFS results.¹⁵ As shown in Fig.4(b), the Te 4dcore-levels did not change the peak shape or the 4d5/2binding energy of 40.1 eV on thea-FIST_1 anda-FIST_2 phases. These peaks were similar to those for the Ge₂Sb₂Te₅, GeTe, and N-doped Ge₂Sb₂Te₅ systems.¹⁰

However, we observed changes of the Sb and In 4d core-level spectra during the resistance change. The chemical shift of the Sb 4d_5/2 core-level betweena-FIST_1 (31.6 eV) anda-FIST_2 (31.7 eV) was 0.1 eV. However, the peak shape and full-width at the half maximum (FWHM) of the Sb 4dcore-level was not altered during the resistance change. The energy shift at the higher binding energy signalled that the Sb atoms were located in a more chemically tight environment.¹⁶In the case of the In 4dcore-levels, we observed the new shoulder at around 16.8 eV after the resistance change.

700 705 710 715 720 725 730 735

In te n s ity (a rb . u n it)

Photon Energy (eV)

a -FIST_2 a -FIST_1

Fe L -edge

7120 7160 0 2 4 6 8 10

0 1 2

0 2 4 6 8 10 12 -1

0 1 2 3 4

a-FIST_2

a-FIST_1

N o rm al iz ed A b so rp ti o n ( a . u .)

Photon energy (eV) Fe-foil

o

Four ie r- tr a n s for m e d m a gni tude ( a . u. )

Interatomic distance (A)

o

Fe-Fe metallic interaction a-FIST_2

a-FIST_1

Fe-foil

a-FIST_2

a-FIST_1

Fe-foil

k

2

-w e ight e d EXAFS s igna l ( a . u. )

k (A^-1)

(a)

(b) (c) (d)

FIG. 2. (a) FeL-edge absorption spectra ofa-FIST_1 anda-FIST_2. These are the typical Fe metal absorption peaks. (b), (c), and (d) are the EXAFS spectra. Also, in botha-FIST_1 anda-FIST_2, the atomic structure and chemical state of Fe showed evidence of Fe-Fe metallic bonding.

In these analyses of the core-level data, we assumed that the chemical states of the embedded Fe nanoclusters had not been changed by the chemical bonding to the other atoms during the resistance change. In fact, this bonding had not contributed to the resistance change. The contribution of the resistance change was assumed to be only the change of the chemical bonding states of the Sb and In 4dcore-levels.

022150-6 Leeet al. AIP Advances1, 022150 (2011)

700 705 710 715 720 725 730 735 -0.5

0.0 0.5 1.0

1.5

XMCD : a -FIST1 at RT

σ⁺ (XAS with left circular) σ^- (XAS with right circular) σ⁺-σ^- (Dichorism) Σ(σ⁺-σ^-)

Intensity (arb. unit)

Photon Energy (eV)

700 705 710 715 720 725 730 735 -0.5

0.0 0.5 1.0 1.5

σ⁺ (XAS with left circular) σ^- (XAS with right circular) σ⁺-σ^- (Dichorism)

Σ(σ⁺-σ^-)

Intensity (arb. unit)

Photon Energy (eV)

XMCD : a -FIST_2 at RT

0 100 200 300 400

0 20 40 60 80 100

Fe-doped IST

1 kOe M (emu/cm3)

T (K)

ZFC FC

-10 -5 0 5 10

-200 -150 -100 -50 0 50 100 150 200

Fe-doped IST

M (emu/cm3)

H (kOe)

10K 110K 200K 250K 300K

(a) (b)

(c) (d)

FIG. 3. (a) Magnetic properties of FIST thin film in zero-field-cooled (ZFC) and field-cooled (FC) magnetizations at 1 kOe from 5 to 350 K. (b) Magnetization curves for the 10∼300 K temperature range. (c) and (d) are the XMCD spectra of a-FIST_1 anda-FIST_2, respectively. In these results, the origin of the magnetic properties was assumed to be the metallic Fe.

In order to analyse the spectra in detail, the In 4d core-level spectra of both a-FIST_1 and a-FIST_2 were curve-fitted using Doniach-S˘unji´c curves convoluted as shown in Fig.5(a).¹⁷ The background due to inelastic scattering was subtracted by the Shirley (or integral) method.¹⁸We found that the three peaks were convoluted in the In 4dcore-level spectra of botha-FIST_1 anda-FIST_2.

The 4d_5/2core-level binding energies of In 4d-1, In 4d-2, and In 4d-3 were 16.9, 17.5 and 17.9 eV, respectively. Also, the chemical states of In 4d-1, In 4d-2 and In 4d-3 were In-In, In-Sb and In-Te bondings, respectively. We performed a relative intensity area calculation of the curve fittings to find the chemical state concentrations during the resistance change. As Fig.5(b)illustrates, the number of In-Te and In-Sb bondings during the resistance change was decreased by 2 and 8%, respectively.

However, the number of In-In bondings was increased by 10%. In these results, we assumed that the origin of the chemical phase-change (resistance change) without structural phase-change was the increase of the number of In-In metallic bondings.

IV. CONCLUSIONS

In summary, we proposed a new electrical-phase-change magnetic material formed by co- sputtering, Fe-doped IST, and measured its structural, magnetic and chemical properties by using synchrotron radiation. We observed the Fe-nanoclusters-embedded IST and the lack of change of Fe-Fe chemical bonding during the electrical-phase-change at the temperature of 49^oC. However, the size of the Fe nanoclusters was increased from 4∼5 nm to 22∼26 nm in the high temperature process at the temperature of 300 ^oC over the duration of 10 sec. We assumed that the origin of

48 50 52 54 56 58 60 62

a

-FIST_2

a

-FIST_1

Intensity (arb. unit)

Binding Energy (eV)

Fe 3 p

52.7 eV

36 38 40 42 44 46

a

-FIST_2

a

-FIST_1

Intensity (arb. unit)

Binding Energy (eV) 4 d

_5/2

Te 4 d

4 d

_3/2

40.1 eV

29 30 31 32 33 34 35 36

31.6 eV

a

-FIST_2

a

-FIST_1

Intensity (arb. unit)

Binding Energy (eV) 4 d

_5/2

Sb 4 d

4 d

_3/2 31.7 eV

15 16 17 18 19 20 21

Intensity (arb. unit)

Binding Energy (eV)

a

-FIST_2

a

-FIST_1

In 4 d

(a) (b)

(c) (d)

FIG. 4. Core-level spectra of (a) Fe 3p, (b) Te 4d, (c) Sb 4d, and (d) In 4das measured by HRXPS. During the phase-change, the core-level peaks of Fe and Te did not change with the binding energy and shape. However, the Sb 4dcore-level peaks shifted at the higher binding energy of 0.1 eV, and In 4dcore-level spectrum appeared a new shoulder at the binding energy of 16.8 eV.

the phase-change with the change of resistance was the increase of In-In chemical state without the effect of both the Fe nanoclusters and the structural amorphous-to-crystalline phase-change. From these results, we assumed that the magnetic and electrical-phase-change properties were individual origins with the Fe nanoclusters and the increase of the In-In chemical state, respectively. On this basis, we assumed that this FIST material with multi-functional properties showed a possibility of non-volatile multi-bit memory material applications.

ACKNOWLEDGEMENTS

M.-C.J. would like to acknowledge the financial support of the National Research Foundation (NRF) of Korea through the CRi program, as well as the experimental support of Dr. Jaeyoon Baik and Dr. Hyun-Joon Shin in the Pohang Accelerator Laboratory. Y.M.L. and S.C. would like to recognize the financial assistance of the Priority Research Center’s Program administered through the NRF and funded by the Ministry of Education, Science and Technology (2009-0093818). J.-Y.K and D.K.C. respectively would like to acknowledge the financial support of NRF grants (2010-0028044 and 2010-0014618, respectively) funded by the Korean Government.

022150-8 Leeet al. AIP Advances1, 022150 (2011)

In 4d-3(17.9 eV) In 4d-2(17.5 eV) In 4d-1(16.9 eV) 26

28 30 32 34 36 38 40 42 44 46

In-Te

In-Sb

10 % 8 %

R e la ti ve A rea R a ti o ( % )

Chemical states

a-FIST_1 a-FIST_2

2 %

In-In

(a)

(b)

22 21 20 19 18 17 16 15 14

Binding Energy (eV)

Intensit y (arb. unit)

In 4d a-FIST_1

a-FIST_2 In 4d-1

In 4d-3

In 4d-2

FIG. 5. (a) Curve fittings of In 4dcore-level spectra for botha-FIST_1 anda-FIST_2. (b) plots the relative area intensity as calculated by the curve fittings.

1T. Kyratsi, K. Chrissafis, J. Wachter, K. M. Paraskevopoulos, M. G. Kanatzidis,Adv. Mater.15, 1428 (2003).

2F. Jedema,Nat. Mater.6, 90 (2007).

3M. H. R. Lankhorst, B. W. S. M. M. Ketelaars, R. A. M. Wolters,Nat. Mater.4, 347 (2005)

4A. V. Kolobov, P. Fons, A. I. Frenkel, A. L. Ankudinov, J. Tominaga, and T. Uruga,Nat. Mater.3, 703 (2004)

5W. Welnic, A. Pamungkas, R. Detemple, C. Steimer, S. Blugel, and M. Wuttig,Nat. Mater.5, 56 (2006)

6M.-C. Jung, Y. M. Lee, H.-D. Kim, M. G. Kim, H. J. Shin, K. H. Kim, S. A. Song, H. S. Jeong, C. H. Ko, and M. Han, Appl. Phys. Lett.91, 083514 (2007).

7K. Kim, J. Park, J.-G. Chung, S. A. Song, M.-C. Jung, Y. M. Lee, H. J. Shin, B. Kuh, Y. Ha, and J.-S. Noh,Appl. Phys.

Lett.89, 243520 (2006).

8Y. M. Lee, M.-C. Jung, H. J. Shin, K. Kim, and S. A. Song,Appl. Phys. Lett.92, 211913 (2008).

9Z. Sun, J. Zhou, and R. Ahuja,Phys. Rev. Lett.96, 055507 (2006).

10M.-C. Jung, K. Kim, Y. M. Lee, J.-H. Eom, J. Im, Y.-G. Yoon, J. Ihm, S. A. Sonh, H. S. Jeong, and H.-J. Shin,J. Appl.

Phys.104, 074911 (2008).

11J. H. Eom, Y. G. Yoon, C. Park, H. Lee, J. Im, D. S. Suh, J. S. Noh, Y. Khang, and J. Ihm,Phys. Rev. B73, 214202 (2006).

12K.-J. Lee, B.-H. Cho, W.-Y. Cho, S. Kang, B.-G. Choi, H.-R. Oh, C.-S. Lee, H.-J. Kim, J.-M. Park, Q. Wang, M.-H. Park, Y.-H. Ro, J.-Y. Choi, K.-S. Kim, Y.-R. Kim, I.-C. Shin, K.-W. Lim, H.-K. Cho, C.-H. Choi, W.-R. Chung, D.-E. Kim, K.-S. Yu, G.-T. Jeong, H.-S. Jeong, C.-K. Kwak, C.-H. Kim, K. Kim,IEEE J. Solid-State Cir.(2007 IEEE International Solid-State Circuits Conference Cat. No.07CH37858),43, 150 (2008).

13E. T. Kim, J. Y. Lee, and Y. T. Kim,Phys. Status Solidi RRL,4, 103 (2009).

14W.-D. Song, L.-P. Shi, X.-S. Miao, and C.-T. Chong,Adv. Mater.20, 2394 (2008).

15J. F. Moulder, W. F. Stickle, P. E. Sobol, and K. D. Bomben,Handbook of X-ray Photoelectron Spectrscopy(Physical Electronics, Eden Prairie, MN) (1995).

16S. H¨ufner,In Photoelectron Spectroscopy : Principles and Applications, 2nd edition, Springer, Berlin, Germany,Ch.1.

(1995).

17S. Doniach and M. Sinji´c,J. Phys. C3, 285 (1970).

18D. A. Shirley,Phys. Rev. B5, 4709 (1972).

19M. -C. Jung, H. S. Shin, K. Kim, J. S. Noh, and J. Chung, Appl. Phys. Lett.89, 0435031 (2006).