Synthesis and Characterization of BiFeO3 Thin Films for Multiferroic

Oct 22, 2015 - ... Engineering, ‡Department of Electrical Engineering, and §Department of Materials Science and Engineering, University of Californ...
3 downloads 17 Views 10MB Size
Article pubs.acs.org/cm

Synthesis and Characterization of BiFeO3 Thin Films for Multiferroic Applications by Radical Enhanced Atomic Layer Deposition Calvin D. Pham,† Jeffrey Chang,† Mark A. Zurbuchen,‡,§ and Jane P. Chang*,† †

Department of Chemical and Biomolecular Engineering, ‡Department of Electrical Engineering, and §Department of Materials Science and Engineering, University of California, Los Angeles, California 90095, United States ABSTRACT: A radical-enhanced atomic layer deposition (RE-ALD) process is described for the synthesis of BiFeO3. Metalorganic precursors β-diketonate, tris(2,2,6,6-tetramethyl3,5-heptanedionato) iron(III) (Fe(TMHD)3), and Bi(TMHD)3 are coreacted with oxygen radicals (produced by a coaxial microwave cavity) as the oxidation source. Stoichiometric BiFeO3 films were deposited on SrTiO3 (001) substrates at 210 °C and crystallized from 450 to 750 °C. The crystallinity and phases in the films were characterized using X-ray diffraction (XRD) and transmission electron microscopy (TEM). The BiFeO3 was determined to be phase-pure and epitaxial when annealed at 650 °C. To study the functional properties, ferroelectric switching was demonstrated using piezoresponse force microscopy (PFM), while weak ferromagnetic behavior was measured using superconducting quantum interference device (SQUID) magnetometry. The films have properties comparable to those of films grown by other methodsgrowth rates are superior in comparison to other ALD methods, demonstrating the potential of RE-ALD for the synthesis of multiferroic complex−oxide thin films.



Atomic layer deposition (ALD) is a thin film growth technique which has quickly become useful in both science and technology for the fabrication of high-quality ultrathin films. The technique consists of alternating half-reactions between a cation precursor and an anion precursor which are both self-limiting.11,12 Ideally, this self-limiting nature allows the precursors to uniformly saturate a surface with a single monolayer even if the surface is three-dimensional and not in line-of-sight from the vapor source and can therefore conformally coat high aspect-ratio features. These aspects make the growth of BiFeO3 via ALD attractive because one can potentially coat three-dimensional nanoscale geometries with a conformal film, enabling unique composite multiferroic structures that could yield increased magnetoelectric coupling, such as embedded nanoparticles, embedded nanopillars, and nanolaminatesdenoted by their configurations as 0-3, 1-3, and 2-2, respectively.13−17 Despite this, there have been only a few demonstrated examples of BiFeO3 growth via ALD18−21 and no reports of plasma or radical-assisted ALD thus far. Therefore, in this work we report the growth of high-quality BiFeO3 films using RE-ALD and their multiferroic properties.

INTRODUCTION To meet the increasingly stringent demand for memory with improved energy efficiency, speed, and robustness, devices based on multiferroic complex metal oxides have been proposed as a potential route for spintronic memory.1,2 Multiferroics are defined within this context for materials in which there is a coexistence and coupling of ferroelectric and magnetic ordering parameters. Such magnetoelectrically coupled materials exhibit the ability to switch the magnetic state upon the application of an electric field or vice versa. This class of materials could create new opportunities for nonvolatile memories with lower power consumption, fast switching speed, and more robust cycling characteristics, as well as other novel spintronic-based applications such as antennae, motors, or sensors. For the majority of research on multiferroic materials, efforts have been focused on those which demonstrate their functional properties at or above room temperature, with an emphasis on the development of devices and technological applications. BiFeO3 has been so widely studied because it is the only known room-temperature single-phase multiferroic material with an antiferromagnetic Néel temperature of ∼643 K and a ferroelectric Curie point of ∼1103 K.3,4 However, BiFeO3 can only be characterized as a G-type antiferromagnet in bulk and is weakly ferromagnetic in thin-film form as a result of the Dzyaloshinskii− Moriya interaction.5−8 In most literature on BiFeO3 thin film growth, methods such as pulsed laser deposition (PLD), sputtering, or chemical vapor deposition (CVD) are reported, but these all require deposition temperatures in excess of 600 °C.4,9,10 On the other hand, atomic layer deposition has been shown to be a viable method for complex oxide growth at greatly reduced temperatures and is compatible for integration in largearea wafers of 300 mm and larger. © XXXX American Chemical Society



EXPERIMENTAL METHOD

A detailed overview of the experimental apparatus used in this work for the deposition of metal oxide thin films using β-diketonate precursors and oxygen radicals has been described previously so only a brief description is presented here.22,23 The multibeam system used in this study consists of a 10 in.-outer diameter stainless steel main chamber along with a load-lock assembly to facilitate sample insertion and Received: June 14, 2015 Revised: August 17, 2015

A

DOI: 10.1021/acs.chemmater.5b02162 Chem. Mater. XXXX, XXX, XXX−XXX

Article

Chemistry of Materials removal without exposing the entire system to atmosphere. Affixed to the main chamber ports are several components: a coaxial microwave cavity radical beam source used to introduce highly reactive oxygen radicals to the sample, a six precursor doser array used to introduce evaporated organometallic precursor fluxes to the sample, a temperature-controlled sample stage, and a hot filament ion gauge which was used to measure the chamber pressure. Pressure in the chamber was maintained between 1 × 10−6 Torr at base and 2 × 10−5 Torr during operation by a CTI 4000 L/s cryogenic pump. The beams of the atom source and the precursor dosers converged onto a heated sample stage where the substrate was mounted. Oxygen radicals were produced from the atom source using a 2.46 GHz Sairem microwave power supply at 25 W and ∼0.6 sccm O2 gas. Full details are reported elsewhere.24 The metal β-diketonate precursors used were tris(2,2,6,6-tetramethyl-3,5-heptanedionato) bismuth(III) [Bi(TMHD)3] (99.9%-Bi, Strem Chemicals, Inc.) and Fe(TMHD)3 (99.9%-Fe, Strem Chemicals, Inc.). These are solid at room temperature and were sublimed at 150 and 130 °C, respectively. Depositions were carried out from 190 to 230 °C. The substrates used were Si (001) during process optimization, followed by SrTiO3 (001) (MTI Corp.) and Nb:SrTiO3 (001) (0.7 wt %) (MTI Corp.) for the piezoresponse force microscopy (PFM) experiments. For BiFeO3 films deposited on SrTiO3 and Nb:SrTiO3, specimens were removed from the chamber after deposition and immediately rapid thermal annealed in order to facilitate crystallization inside an AccuThermo AW 610 RTP furnace using oxygen gas flowed at 5 sccm between 450 and 750 °C. The film thicknesses were estimated with a J.A. Woollam M-88 spectroscopic ellipsometer with data modeled using a Lorenz oscillator fit that was calibrated by cross-section SEM imaging. The films’ composition and chemical bonding information were analyzed using an Axis Ultra DLD (Kratos Analytical Ltd.) X-ray photoelectron spectroscopy (XPS) instrument using monochromatized Al Kα radiation (1486.6 eV). Crystal structure was characterized using a X’Pert Pro (PANalytical B.V.) Powder X-ray diffractometer (XRD) using Cu Kα radiation. An FEI Titan transmission electron microscope (TEM) was used for cross-section microstructure imaging and selected area electron diffraction (SAED) at 300 keV. Cross-sectioned lamella for TEM analysis were prepared by the standard trench and pluck approach, with successive finishing at 30, 10, and 5 kV using an FEI Nova 600. The piezoelectric properties were examined with a Dimension Icon (Bruker) AFM instrument in PFM mode using a conductive tip, a drive frequency of 100 kHz, and an amplitude of 9000 mV. Magnetic properties were investigated using a MPMS (Quantum Design Inc.) Superconducting Quantum Interference Device (SQUID) magnetometer, between ±3 T.

Figure 1. Precursor pulsing sequence for BiFeO3.



RESULTS AND DISCUSSION A single ALD cycle for deposition of a binary oxide consisted of the following sequence: a metalorganic precursor dose, a pumpdown to prevent gas-phase reactions, the oxygen radicals pulse, and a subsequent pump-down before repeating for x number of cycles until the desired film thickness was reached. For a global cycle of BiFeO3, the film composition was controlled by pulsing precursors in the sequence as shown in Figure 1 which consisted of a cycles of Bi(TMHD)3:O and b cycles of Fe(TMHD)3:O. To develop the protocol for the ternary oxides, it was first necessary to study the growth of the binary oxides Bi2O3 and Fe2O3. Synthesis of Binary Oxides. To confirm the initial synthesis of the desired constituent oxides Fe2O3 and Bi2O3 XPS was used. Figure 2 shows results for the films deposited at 210 °C on Si (001), consisting of 100 cycles of the ALD sequence with parameters of 90 s metalorganic precursor, 5 s pump-down, 20 s oxygen radical exposure, and 5 s pump-down. XP spectra were corrected to the accepted literature value of 284.8 eV for C 1s. As expected, the high-resolution XPS spectra for Fe 2p, seen in Figure 2a, displayed the spin orbital doublets 2p3/2 and 2p1/2, with peak positions at approximately 710.7 and 724.0 eV, respectively. In addition, the detailed spectra also featured

Figure 2. High resolution XPS spectra for (a) the Fe 2p peak from the Fe2O3 film and (b) the Bi 4f peak from the Bi2O3 film grown by ALD, both on Si (001).

shakeup satellite peaks at 718.9 and 732.8 eV, as typically expected for the Fe3+ oxidation state.25 For Bi2O3 seen in Figure 2b, the peak positions for Bi 4f7/2 and 4f5/2 were 158.6 and 163.9 eV, respectively, which indicated the Bi3+ oxidation state.26 After confirming the initial deposition of the films, it was then possible to optimize the growth parameters. The film thickness was characterized using a combination of SEM cross-section micrographs and ellipsometry measurements fitted using a Lorenz oscillator calibrated by the thickness measured by SEM. Growth rates for both Fe2O3 and Bi2O3 were determined as a function of the deposition temperatures (190−230 °C), shown in Figure 3a. For Fe2O3, the deposition rate was relatively consistent, increasing slightly from 0.40 to 0.49 Å/cycle within the range of temperatures studied. For the composition of our Fe2O3 as a function of the deposition temperature, shown in Figure 3b, the proportions between Fe, O, and C did not vary significantly, with the Fe B

DOI: 10.1021/acs.chemmater.5b02162 Chem. Mater. XXXX, XXX, XXX−XXX

Article

Chemistry of Materials

For the Bi2O3 films, the deposition rate was approximately 0.70−0.74 Å/cycle, as seen in Figure 3a. At a lower temperature of 190 °C, the deposition rate was drastically higher at ∼1.4 Å/ cycle, indicating a condensation effect in which the Bi(TMHD)3 precursor was believed to be undergoing physisorption and not in the ideal self-saturated ALD growth. This effect was also reported by Shen et al. during thermal ALD optimizations using Bi(TMHD)3 but with H2O as the oxidation source.28 On the other hand, at temperatures of up to 230 °C beyond our growth window of 200−210 °C, the growth rate decreased, indicating the precursor desorption from the surface, or a lowered sticking coefficient, a result mirrored in the report of Shen et al., who reported a growth window between 270 and 300 °C; however, for this work, the ALD window is much lower and narrower compared to the thermal ALD process. The Bi2O3 films’ compositions were determined within the temperature range of 190−230 °C by XPS as shown in Figure 3c. The relative percentage of bismuth decreased and that of both oxygen and carbon increased with temperature. At 190 °C, the bismuth content was the highest at ∼30% whereas the carbon content was the lowest at ∼23%. This indicated that although the Bi(TMHD)3 was increasingly condensed on the surface, the oxygen radicals were still effective at removing the organic ligands from the precursor and not causing uncontrolled incorporation of carbon into the film. At higher temperatures, the carbon percentage was greater while the deposition rate diminished due to desorption of precursor from the surface. Because the deposition rate of Bi2O3 was lower, the relative fraction of Bi/C decreased. Nonetheless, the amount of carbon contamination was not substantially changed over the temperature range studied, increasing from ∼23% at 190 °C to ∼27% at 230 °C. As ALD is a self-limiting technique, it was necessary to determine the effect of precursor pulse time upon its growth rate, as it coincides with ALD’s ability for conformal deposition over nonplanar (3-D or rough) surfaces. The growth rates of Fe2O3 and Bi2O3 as a function of the pulse time are shown in Figure 4a; the fitted curves are shown as a guide for the eye. The data indicated that 90 s precursor pulse time for both Bi(TMHD)3 and Fe(TMHD)3 was sufficient to achieve saturation and was kept constant as a process parameter for the remainder of the experiments. To demonstrate a linear growth rate, the thicknesses of the Fe2O3 and Bi2O3 films were measured as a function of number of cycles as shown in Figure 4b. According to the fitted linear curves, the Fe2O3 growth rate was ∼0.49 Å/cycle and the Bi2O3 growth rate was ∼0.65 Å/cycle, with no apparent nucleation delay for the precursor absorption. When comparing our growth rates to literature using the same TMHD-based precursors and similar temperature ranges, it is clear that the RE-ALD growth rates are 4−6 times higher; the reported growth rates using thermal ALD were 0.124 Å/cycle for Fe2O3 using ozone27 and 0.1 Å/cycle for Bi2O3 using H2O.28 In addition, a report using another bismuth-containing precursor Bi(OCMe2iPr)3 and water as the coreactant demonstrated a growth rate of ∼0.36 Å/cycle. Finally, Zhang et al. reported thermal ALD growth rates, using H2O as the anion source, of 0.07 and 0.04 Å/cycle for Bi2O3 and Fe2O3 using Bi(tmhd)3 and Fe(tmhd)3, respectively.19 In summary, the radical enhanced process yields a substantially higher growth rate. Synthesis and Structure of BiFeO3 Films. Once the baseline growth rates of Fe2O3 and Bi2O3 were established, it was possible to integrate the two to produce the BiFeO3 ternary oxide. As mentioned previously, the precursors were pulsed

Figure 3. (a) Temperature dependence of ALD deposition rate Å/cycle for Fe2O3 (circles) and Bi2O3 (triangles) films grown on Si (001) substrates. (b) and (c) show the compositions of Fe2O3 and Bi2O3 films, respectively, as a function of deposition temperature, as calculated using normalized integration of XPS peaks.

content fairly consistent between ∼30−32% and the C content between ∼17−20% with no obvious trend corresponding to the deposition rate. According to prior research using our experimental setup, the amounts of adventitious oxygen and carbon were estimated to be 15% and 13%, respectively.23 But, it was observed that as the temperature increased beyond 250 °C, the precursor decomposed on the surface, engaging in CVD growth. This observation is relatively consistent with the work of Lie et al., who also used Fe(TMHD)3 precursor but with ozone as the coreactant and reported a growth window of 160−310 °C above which the precursor similarly decomposed on the surface.27 The difference in the maximum allowable temperature is believed to be due to the difference in oxidation sources, as the oxygen radicals used in this work are more reactive than ozone. When comparing the growth rate within the temperature window, the previously reported thermal ALD process had a growth rate of 0.124 Å/cycle which is substantially slower than the RE-ALD process in this work.27 C

DOI: 10.1021/acs.chemmater.5b02162 Chem. Mater. XXXX, XXX, XXX−XXX

Article

Chemistry of Materials

sequentially in a (a(Bi(TMHD)3:O) + b(Fe(TMHD)3:O) manner while controlling the values of a and b to achieve the Bi/Fe = 1 stoichiometry. Thus, the composition of the mixed ternary films could be measured by XPS shown in the survey scan of Figure 5a with degree of stoichiometry control in Figure 5b. Initially, a 2:2 pulsing ratio was attempted using a Si (001) substrate, indicated by the data point at 50% cycle ratio, and was marked by a Bi/Fe content ratio close to unity. Additional data points were collected to confirm the control of the film stoichiometry by adding or removing cycles from a or b, respectively. However, when transferring the process directly to the SrTiO3 (001) substrate, a cycle ratio of a:b = 2:2 showed only 5% iron content, suggesting a nucleation period. The value of b to control the iron content was thus iterated upon to achieve the 1:1 stoichiometry, resulting in the optimal cycle ratio of a:b = 2:7. The growth rate of films using this cycle ratio was 5.5 Å/ supercycle. We speculate that the discrepancy is caused by differences in the surface atomic arrangement when depositing films on SrTiO3, but this requires additional investigation. In addition, detailed spectra for Fe 2p and Bi 4f are shown in Figure 5c,d, respectively. The peak positions and satellite peaks (sat.) for Fe 2p indicated in the detailed spectra suggested that the iron and bismuth ions were in the 3+ oxidation state. Carbon content was approximately 25 atom %; however, it should be reiterated that the XPS measurement was performed ex situ and, when ignoring the adventitious carbon content of 13%, is estimated to be ∼12 atom % carbon contained in the film. With the protocol optimized, stoichiometric BiFeO3 was then formed by depositing the Bi2O3−Fe2O3 (Fe/Bi = 1) films on closely lattice matched SrTiO3 (001) substrates immediately followed by rapid thermal annealing (RTA) to crystallize the films.

Figure 4. (a) Effect of ALD precursor pulse time on growth rate for Fe2O3 (circles) and Bi2O3 (triangles) films grown on Si (001) at 210 °C. (b) Postcalibration RE-ALD film thickness achieved, plotted as a function of number of cycles for Fe2O3 (circles) and Bi2O3 (triangles) films grown on Si (001) at 210 °C. Fitted lines are guides for the eye.

Figure 5. (a) XPS survey scan for BiFeO3 thin film grown by RE-ALD on Si (001). (b) Bismuth cation percentage (Bi/(Bi + Fe) %) of Bi2O3−Fe2O3 films on Si(001) and SrTiO3 (001) substrates as a function of pulse sequence percent ratio between Bi2O3 and Fe2O3 (defined as 100% × a/(a + b), where a and b are the number of Bi(TMHD)3:O and Fe(TMHD)3:O cycles, respectively). Detailed XPS spectra for (c) Fe 2p (Sat. indicates satellite peaks) and (d) Bi 4f photoelectron peaks. D

DOI: 10.1021/acs.chemmater.5b02162 Chem. Mater. XXXX, XXX, XXX−XXX

Article

Chemistry of Materials

Figure 6. (a) Diffraction patterns for 35 nm thick BiFeO3 films grown on SrTiO3 (001) substrates as prepared and as a function of annealing temperature. Vertical lines indicate references from powder samples (PDF: BiFeO3 071-2494, α-Bi2O3 071−-2274, Bi2Fe4O9 025-0090, SrTiO3 0860178). (b) Detailed look at BiFeO3 (001)pc reflection for samples annealed at 550 and 650 °C.

We then examined the interface in detail by HRTEM. The BiFeO3/SrTiO3 interface was smooth and showed no indication of any competitive grain growth having occurred during crystallization at 650 °C. Figure 7b shows a representative region that was consistent throughout the sample. Finally, to corroborate the XRD results, diffraction analysis was performed. Zone-axis diffraction patterns (ZADP) were collected of the substrate and film in immediate sequence, with the BiFeO3 [100]pc DP shown in Figure 7c. BiFeO3 is R3c, and peak splitting to the growth-twin domains was clearly visible, shown in the inset of Figure 7c for a 006pc reflection. The growth-twin domains are due to the pseudocubic cells of which the corners are displaced relative to a perfect cube ∼0.38° along ⟨110⟩pc (i.e., 0.27° along ⟨100⟩pc and ∼0.27° along ⟨010⟩pc). Using the SrTiO3 pattern as a reference, the out-of-plane BiFeO3 (001)pc lattice parameter was estimated to be 3.987 Å, i.e., compressive strain caused by the SrTiO3 substrate. Thus, epitaxy was confirmed, with a pseudocube-on-cube orientation relationship of BiFeO3(001)pc∥SrTiO3(001) and BiFeO3[100]pc∥SrTiO3[100]. When referring to the rhombohedral structure of BiFeO3, the three orthogonal zone axes corresponding to the pseudocubic ⟨100⟩ zone axes are [241], [4̅2̅1], and [2̅21̅]. Multiferroic Properties of BiFeO3 Films. Once high quality BiFeO3 films could be attained by RE-ALD, it was possible to characterize their functional ferroelectric and magnetic behavior. Because of the quality of the 650 °C film deposited on SrTiO3 (001), confirmed by XRD and TEM, the ferroelectric response was studied by PFM using a sister sample deposited using the same protocol on a Nb:SrTiO3 (001) (0.7 wt %) substrate to act as a bottom electrode. As shown in Figure 8, ferroelectric domain switching was demonstrated by locally applying ±12 V to reversibly write a square pattern into the domains as shown in Figure 8b. The phase contrast confirmed the film was ferroelectric. Nonetheless, it has been predicted that weak ferromagnetic ordering can occur in bulk antiferromagnetic systems, due to the interaction of canted spins known as the Dzyaloshinskii−Moriya interaction.5,29 Indeed, as shown in Figure 9, the same epitaxial BiFeO3 sample characterized by TEM and XRD was observed to be weakly ferromagnetic, as indicated by the well-formed magnetic hysteresis loop. The saturation magnetization, Ms, at room temperature (298 K) was ∼27 emu/cm3, whereas the magnetic coercivity HC was ∼0.063 kOe.

Because the multiferroic phase of BiFeO3 has a rhombohedrally distorted perovskite structure (a = 5.63 Å, α = 59.4°, PDF 071-2494), single-crystal cubic SrTiO3 (001) (a = 3.9 Å, PDF 86178) was used for subsequent experiments for its ∼1% lattice mismatch, as well as electrically conducting Nb:SrTiO3 (0.7 wt %) to enable through-thickness ferroelectric and piezoelectric measurements of BiFeO3. BiFeO3 is known to grow epitaxially on SrTiO3 substrates with a (012) orientation, using methods such as PLD, which corresponds to a (001) orientation for the pseudocubic (pc) equivalent.4 The BiFeO3(001)pc/SrTiO3 (001) crystallization process by RTA was optimized by varying temperatures (450−750 °C) under oxygen for 60 s and a 50 °C/s ramp. Crystallinity, phasepurity, orientation, and texture were then analyzed as a function of the crystallization temperature using XRD as shown in Figure 6a. BiFeO3 films deposited at 210 °C before RTA consisted of a small amount of α-Bi2O3, as indicated by its (012) reflection (28.01°, indicated on plot) (PDF 71-2274), which persisted up to crystallization temperatures of 450 °C. Low temperature crystallization of α-Bi2O3 has been reported previously in literature.28 Above 550 °C, the α-Bi2O3 reflection diminished and the reflections from BiFeO3 became prominent, as indicated in Figure 6 using the pseudocubic indexing (2θ = 22.4° for (001)pc, 45.8° for (002)pc). Their highest intensity was observed for the 650 °C film. RTA above 750 °C resulted in a phase transformation to polycrystalline Bi2Fe4O9 (PDF 025-0090) likely due to bismuth volatilization, as indicated by multiple orientation peaks (indexed in Figure 6a). A detailed view of the BiFeO3 (001)pc peak is shown in Figure 6b for the 550 and 650 °C films. A change in the peak position as a function of film thickness was not detected up to 93.5 nm. The surface roughness was analyzed using AFM, and it was found that the RMS roughness increased from 3.8 Å before RTA to 44.8 Å after RTA at 650 °C. XRD data for the 550 and 650 °C films had no unknown reflections and indicated that the films were oriented with BiFeO3(001)pc∥SrTiO3(001), implying they were, at minimum, oriented plane-on-plane and possibly epitaxial. We note that although ALD films are typically polycrystalline, we achieved epitaxial BiFeO3 by RE-ALD. To further characterize the orientation and crystallinity of the 650 °C sample, TEM was employed. The film is fully dense and reasonably smooth, as shown in Figure 7a, with a nominal thickness of 93.5 nm. No secondary phases were detected by either imaging or diffraction. After aligning to the SrTiO3 [100] zone axis, it was clear that the sample consisted of a single grain. E

DOI: 10.1021/acs.chemmater.5b02162 Chem. Mater. XXXX, XXX, XXX−XXX

Article

Chemistry of Materials

Figure 8. Piezoresponse force microscopy phase images of BiFeO3 film on Nb:SrTiO3 (001) substrate (a) before and (b) after ±12 V applied using PFM tip in square pattern.

Figure 7. (a) Low magnification TEM image of BiFeO3 film crystalized 650 °C on SrTiO3 (001) substrate. (b) TEM micrograph of BiFeO3 film, indicating epitaxial nature, arrows signifying interface. (c) Zone-axis diffraction pattern for BiFeO3 film and (inset) enlarged 006pc reflection indicating growth-twin domains.



CONCLUSIONS Epitaxial BiFeO3 thin films were synthesized using a radical enhanced atomic layer deposition (RE-ALD), with stoichiometry controlled toward unity by adjusting the precursor cycle sequence based on XPS. Crystallization of BiFeO3/SrTiO3 (001) was obtained by RTA from 550 to 650 °Cepitaxy of the BiFeO3 was confirmed by combined XRD and TEM with an orientation relationship of BiFeO3 {001}pc∥SrTiO3 (001) and BiFeO3 ⟨100⟩pc∥SrTiO3 (100). The surface roughness increased from 3.8 Å before RTA to 44.8 Å after RTA at 650 °C.

Figure 9. Magnetic hysteresis of BiFeO3 film annealed at 650 °C on SrTiO3 substrate with applied magnetic field oriented in-plane (black squares) and out-of-plane (red circles) relative to the film plane. Inset indicates zoomed view of magnetic coercivity.

Ferroelectric switching with remanence was demonstrated by PFM, and ferromagnetic reponse was confirmed by SQUID F

DOI: 10.1021/acs.chemmater.5b02162 Chem. Mater. XXXX, XXX, XXX−XXX

Article

Chemistry of Materials magnetometry with Ms, at room temperature (298 K) at ∼27 emu/cm3, and HC of ∼63 Oe. Growth rates were 3−4 times higher than previously reported ALD approaches, with an estimated carbon content of ∼12 atom %, measured by ex situ XPS. The application of oxygen radicals as the anion source is believed to greatly decrease the carbon contamination, thus improving the film purity. RE-ALD is demonstrated to yield high quality and epitaxial BiFeO3 signifying RE-ALD’s great potential for large-scale deposition of BiFeO3 in microelectronics manufacturing.



(11) Leskelä, M.; Ritala, M. Atomic layer deposition (ALD): from precursors to thin film structures. Thin Solid Films 2002, 409, 138−146. (12) Leskelä, M.; Ritala, M. Atomic Layer Deposition Chemistry: Recent Developments and Future Challenges. Angew. Chem., Int. Ed. 2003, 42, 5548−5554. (13) Nan, C. W.; Bichurin, M. I.; Dong, S. X.; Viehland, D.; Srinivasan, G. Multiferroic magnetoelectric composites: Historical perspective, status, and future directions. J. Appl. Phys. 2008, 103, 031101. (14) Vaz, C. A. F.; Hoffman, J.; Ahn, C. H.; Ramesh, R. Magnetoelectric Coupling Effects in Multiferroic Complex Oxide Composite Structures. Adv. Mater. 2010, 22, 2900−2918. (15) Ma, J.; Hu, J.; Li, Z.; Nan, C.-W. Recent Progress in Multiferroic Magnetoelectric Composites: from Bulk to Thin Films. Adv. Mater. 2011, 23, 1062−1087. (16) Fiebig, M. Revival of the magnetoelectric effect. J. Phys. D: Appl. Phys. 2005, 38, R123. (17) Srinivasan, G. Magnetoelectric Composites. Annu. Rev. Mater. Res. 2010, 40, 153−178. (18) Liu, Y.-T.; Ku, C.-S.; Chiu, S.-J.; Lee, H.-Y.; Chen, S.-Y. Ultrathin Oriented BiFeO3 Films from Deposition of Atomic Layers with Greatly Improved Leakage and Ferroelectric Properties. ACS Appl. Mater. Interfaces 2014, 6, 443−449. (19) Zhang, F.; Sun, G.; Zhao, W.; Wang, L.; Zheng, L.; Liu, S.; Liu, B.; Dong, L.; Liu, X.; Yan, G.; Tian, L.; Zeng, Y. Atomic Layer Deposition of BiFeO3 Thin Films Using β-Diketonates and H2O. J. Phys. Chem. C 2013, 117, 24579−24585. (20) Akbashev, A. R.; Chen, G.; Spanier, J. E. A Facile Route for Producing Single-Crystalline Epitaxial Perovskite Oxide Thin Films. Nano Lett. 2014, 14, 44−49. (21) Jalkanen, P.; Tuboltsev, V.; Marchand, B.; Savin, A.; Puttaswamy, M.; Vehkamäki, M.; Mizohata, K.; Kemell, M.; Hatanpäa,̈ T.; Rogozin, V.; Räisänen, J.; Ritala, M.; Leskelä, M. Magnetic Properties of Polycrystalline Bismuth Ferrite Thin Films Grown by Atomic Layer Deposition. J. Phys. Chem. Lett. 2014, 5, 4319−4323. (22) Van, T. T.; Chang, J. P. Surface reaction kinetics of metal βdiketonate precursors with O radicals in radical-enhanced atomic layer deposition of metal oxides. Appl. Surf. Sci. 2005, 246, 250−261. (23) Van, T. T.; Chang, J. P. Radical-enhanced atomic layer deposition of Y2O3 via a β−diketonate precursor and O radicals. Surf. Sci. 2005, 596, 1−11. (24) Chang, J. P.; Arnold, J. C.; Zau, G. C.; Shin, H.-S.; Sawin, H. H. Kinetic study of low energy argon ion-enhanced plasma etching of polysilicon with atomic/molecular chlorine. J. Vac. Sci. Technol., A 1997, 15, 1853−1863. (25) Mills, P.; Sullivan, J. L. A study of the core level electrons in iron and its three oxides by means of X-ray photoelectron spectroscopy. J. Phys. D: Appl. Phys. 1983, 16, 723. (26) Uchida, K.; Ayame, A. Dynamic XPS measurements on bismuth molybdate surfaces. Surf. Sci. 1996, 357−358, 170−175. (27) Lie, M.; Fjellvåg, H.; Kjekshus, A. Growth of Fe2O3 thin films by atomic layer deposition. Thin Solid Films 2005, 488, 74−81. (28) Shen, Y. D.; Li, Y. W.; Li, W. M.; Zhang, J. Z.; Hu, Z. G.; Chu, J. H. Growth of Bi2O3 Ultrathin Films by Atomic Layer Deposition. J. Phys. Chem. C 2012, 116, 3449−3456. (29) Dzyaloshinsky, I. A Thermodynamic Theory of Weak Ferromagnetism of Antiferromagnetics. J. Phys. Chem. Solids 1958, 4, 241−255.

AUTHOR INFORMATION

Corresponding Author

*(J.P.C.) E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The authors would like to acknowledge the UCLA Molecular Instrumentation Center and the SPM facility at the Nano and Pico Characterization Lab at the California NanoSystems Institute (CNSI) for the use of characterization facilities. This work was supported in part by the Function Accelerated nanoMaterial Engineering (FAME) Center, one of six centers of Semiconductor Technology Advanced Research Network (STARnet), a Semiconductor Research Corporation (SRC) program sponsored by Microelectronics Advanced Research Corporation (MARCO) and Defense Advanced Research Projects Agency (DARPA).



REFERENCES

(1) Eerenstein, W.; Mathur, N. D.; Scott, J. F. Multiferroic and magnetoelectric materials. Nature 2006, 442, 759−65. (2) Bibes, M.; Barthelemy, A. Multiferroics: Towards a magnetoelectric memory. Nat. Mater. 2008, 7, 425−426. (3) Smolenskiĭ, G. A.; Chupis, I. E. Ferroelectromagnets. Soviet Phys. Usp. 1982, 25, 475. (4) Wang, J.; Neaton, J. B.; Zheng, H.; Nagarajan, V.; Ogale, S. B.; Liu, B.; Viehland, D.; Vaithyanathan, V.; Schlom, D. G.; Waghmare, U. V.; Spaldin, N. A.; Rabe, K. M.; Wuttig, M.; Ramesh, R. Epitaxial BiFeO3Multiferroic Thin Film Heterostructures. Science 2003, 299, 1719−1722. (5) Moriya, T. Anisotropic superexchange interaction and weak ferromagnetism. Phys. Rev. 1960, 120, 91. (6) Albrecht, D.; Lisenkov, S.; Ren, W.; Rahmedov, D.; Kornev, I. A.; Bellaiche, L. Ferromagnetism in multiferroic BiFeO3 films: A firstprinciples-based study. Phys. Rev. B: Condens. Matter Mater. Phys. 2010, 81, 140401. (7) Ramazanoglu, M.; Laver, M.; Ratcliff, W., II; Watson, S.; Chen, W.; Jackson, A.; Kothapalli, K.; Lee, S.; Cheong, S.-W.; Kiryukhin, V. Local Weak Ferromagnetism in Single-Crystalline Ferroelectric BiFeO3. Phys. Rev. Lett. 2011, 107, 207206. (8) Dzyaloshinskii, I. E., Thermodynamic theory of weak ferromagnetism in antiferromagnetic substances. Soviet Physics J. Exptl. Theoret. Phys. 1957, 5, 1259−1562. (9) Das, R. R.; Kim, D. M.; Baek, S. H.; Eom, C. B.; Zavaliche, F.; Yang, S. Y.; Ramesh, R.; Chen, Y. B.; Pan, X. Q.; Ke, X.; Rzchowski, M. S.; Streiffer, S. K. Synthesis and ferroelectric properties of epitaxial BiFeO3 thin films grown by sputtering. Appl. Phys. Lett. 2006, 88, 242904. (10) Yang, S. Y.; Zavaliche, F.; Mohaddes-Ardabili, L.; Vaithyanathan, V.; Schlom, D. G.; Lee, Y. J.; Chu, Y. H.; Cruz, M. P.; Zhan, Q.; Zhao, T.; Ramesh, R. Metalorganic chemical vapor deposition of lead-free ferroelectric BiFeO3 films for memory applications. Appl. Phys. Lett. 2005, 87, 102903. G

DOI: 10.1021/acs.chemmater.5b02162 Chem. Mater. XXXX, XXX, XXX−XXX