Subscriber access provided by UNIV OF SCIENCES PHILADELPHIA
Article
Understanding the growth mechanisms of multilayered systems in atomic layer deposition process. Christoph W. Wiegand, René Faust, Alexander Meinhardt, Robert H. Blick, Robert Zierold, and Kornelius Nielsch Chem. Mater., Just Accepted Manuscript • DOI: 10.1021/acs.chemmater.7b05128 • Publication Date (Web): 23 Feb 2018 Downloaded from http://pubs.acs.org on February 23, 2018
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Chemistry of Materials is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 10 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Chemistry of Materials
Understanding the growth mechanisms of multi-layered systems in atomic layer deposition process. Christoph W. Wiegand1,2, René Faust2, Alexander Meinhardt2, Robert H. Blick2, Robert Zierold2,*, Kornelius Nielsch1,3. 1 Institute for Metallic Materials, Leibniz Institute for Solid State and Materials Research Dresden (IFW Dresden), Helmholtzstraße 20, 01069 Dresden, Germany. 2 Center for Hybrid Nanostructures, Universität Hamburg, Luruper Chaussee 149, 22607 Hamburg, Germany. 3
Insitute of Material Science, Technical University Dresden, Helmholtzstraße 10, 01069 Dresden, Germany.
ABSTRACT: In atomic layer deposition the initial growth is of particular interest since it defines the nucleation behavior and determines the minimum number of cycles to achieve a closed layer. The growth rate is quantified by the growth per cycle (GPC). Due to nucleation inhibition or enhancement, the initial GPC for ALD processes of a given material system onto a specific substrate may differ from its (steady-state) literature value, since the GPC is mostly noted as an average value of at least a few hundred cycles. However, the knowledge of the growth behavior within the first few cycles is of particular importance in context of super cycles and nanolaminates. Individual ALD cycles of the host material (e.g. TiO2) are replaced by those for the deposition of another (e.g. Al2O3) compound to infiltrate the host material with a dopant. For a precise dosage/doping and tailor-made synthesis, knowledge of the individual GPC is crucial. Herein, precise quartz crystal microbalance (QCM) studies were used to investigate the initial growth of TiO2 onto Al2O3 (deposited by ALD) (1) and vice versa the initial growth of Al2O3 onto TiO2 (2). In case (1), an enhanced initial GPC of the TiO2 deposition was observed that was close to the equilibrium value of Al2O3 deposition. In the 2nd case, the initial GPC of Al2O3 was found to be close to the equilibrium value of TiO2 deposition. The growth process itself can be simply modeled by a superposition of parallel growth onto itself and onto the foreign species. We attribute our observations to an ALD process intrinsic inhibition of the TiO2-growth.
INTRODUCTION: The physical properties of thin films, such as surface roughness1, hardness2, lattice constant3, dielectric constant4, refractive index5, and phase-transition temperatures6 to name a few of them, mainly depend on their composition as well as crystallinity. The need for tailor-made materials whose properties are solely predetermined by external (working) conditions has risen with respect to an economically and ecologically usage of raw materials. Example giving, component parts of various devices and application areas are subjected to the most adverse environmental conditions during their use. A thermal barrier is essential for turbines, which work at temperatures up to 1650 °C, to prevent the raw material from overheating and corrosion, elongating the turbine lifetime.7 Novel promising approaches like inverse titania opals, which act as photonic crystals in the near-infrared, lack structural stability at temperatures higher than 1000 °C, mostly due to crystallization and loss of microstructure8. The use of a dual material systems, as nanolaminates or by doping can shift the phase-transition temperature to higher temperatures, making the inversed opals much more stable at elevated temperatures.6,9 Hence, the preparation method of choice has to be versatile, capable of processing various materials and material combinations and able to produce a conformal coating. Specifically, Atomic Layer Deposition (ALD) is one method which enables the experimentalist to synthesize multilayer systems by utilizing supercycle approaches10. In detail, adjusting the number of ALD
cycles of the combined – two or more – individual processes according to the known growth rates, the synthesis of tailormade ternary materials is possible but also doping of a specific species into a host material. A wide variety of novel, tailormade, composite materials have been employed utilizing ALD methods, including alloys3,11–23, nanolaminates5,24–32, doped materials33–37, and mixed oxides38,39. For multilayer systems as well as nanolaminates with individual layer thicknesses in the lower nm-range – corresponding to at least several tenth of ALD cycles – the individual layers and thus the total film growth can be described by the equilibrium growth rates. However, the depicted deposition behavior of one material onto the surface of the host material becomes important when ALD is used for doping of only a small amount of material into a host or into extremely thin layers. As a proof-ofconcept, we herein investigate the interplay of the first pulses of Al2O3 and TiO2 depositions onto each other and show that surface-dependent inhibition can be observed which may drastically alter the physical properties of compounds deposited by a super-cycle approach. EXPERIMENTAL SECTION: General ALD deposition parameters. A QCM from Colnatec ® was modified to fit into an in-house build ALD reaction chamber for measuring the change in frequency, which corresponds to a change in mass. The depositions were conducted using continuous flow mode by utilizing trimethyl-
ACS Paragon Plus Environment
Chemistry of Materials 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
aluminium (TMA) (Sigma Aldrich) and DI-water at room temperature and titanium tetrakis isopropoxide (Ti(OiPr)4) (Sigma Aldrich) heated to 85 °C as precursors. The pulse times corresponds to 0.4 s (TMA), 0.4 s (H2O), and 2 s(Ti(OiPr)4), respectively. The purge time was kept between 60 s (after the metal-precursor pulses) and 90 s (after the H2O pulses) for the general depositions. The pulse lengths and purge times were chosen based on results in growth studies in the exposure mode40–42 and according to literature.43,44 The flow of Ar(5.0)-carrier gas was set to 2.5 l/h. Deposition of alternating Al2O3 and TiO2 thin-films. A sequential deposition of 50 cycles TiO2 onto prior deposited TiO2 and 50 cycles of Al2O3 onto prior deposited Al2O3 served as reference. In the following, the growth of 50 cycles TiO2 (Al2O3) was analyzed as a function of the number n (m) ALD cycles of the prior deposited material namely Al2O3 (TiO2). The experimental procedure is schematically sketched in Figure 1 a).
Figure 1. (a) Schematic illustration of the ALD cycles. TiO2 and Al2O3 have been deposited onto itself, separated with 900 s purging-time, as a reference, whereas the further growth of TiO2 and Al2O3 was performed on top of previously deposited Al2O3 and TiO2, respectively. (b) Schematic illustration of deposition-cycles of TiO2 (blue) separated by blocks (highlighted in white) with varying purging-time. Influence of purging time on TiO2 growth behavior. The time between individual blocks of 30 TiO2 ALD cycles was varied between 300 s and 7200 s and the initial growth of the subsequent cycle was observed. A schematic sketch is shown in Figure 1 b).
Page 2 of 10
Validation experiment with multiple H2O-pulses onto a TiO2 surface. A TiO2 thin film, deposited by 50 cycles of Ti(OiPr)4 and H2O was treated within one hour with 2 to 40 H2O-pulses (2; 4; 8; 16; 20; 40). The variation in frequency of the following TiO2-deposition, especially during the first cycle was measured. RESULTS AND DISCUSSION: We utilized a quartz crystal microbalance (QCM) setup, which possesses a resolution high enough to differentiate between the single pulses of each precursor used in the ALD deposition process1,45. The resonance frequency of the QCM changes with additionally deposited mass onto the crystal. A typical frequency pattern during an ALD deposition of Ti(OiPr)4 + H2O onto previously deposited TiO2 is shown in Figure 2 a).
Figure 2. (a) The frequency variation (black) and temperature (red) during a Ti(OiPr)4+H2O process recorded by a QCM. (b) Negative frequency variation for the first few pulses of Ti(OiPr)4+H2O onto an Al2O3 surface as a function of process time. (c) Influence of previous Al2O3 cycle(s) on the frequency variation during the following deposition of TiO2.
ACS Paragon Plus Environment
Page 3 of 10 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Chemistry of Materials
The stepwise variation of the frequency corresponds to Ti(OiPr)4 as well as H2O pulses, which is read as mass (ng/cm2) deposition during each of the individual pulses. By calculating the slope of such a curve, it is possible to identify the growth rate of the deposited material. The high resolution in frequency and time of the QCM setup permits the study of only a few ALD cycles onto different surfaces. For depositing TiO2 onto itself (which is shown in Figure 2 a), the first pulse has the same variation of frequency – corresponding to a mass uptake – of about ∆f=~2 Hz/cycle, as the following pulses, indicating a negligible influence of the surface onto the first pulses and leading to a constant growth rate over the entire process. The same applies to the deposition of Al2O3 from TMA and H2O (see supporting information Figure S1) onto an Al2O3 surface with a frequency change of ∆f=~9 Hz/cycle. Conversely, when the surface is changed from previously deposited TiO2 to previously deposited Al2O3, the first pulses of Ti(OiPr)4 + H2O differ dramatically in amplitude from the following ones, which is displayed in Figure 2 b). To ease the comparison of the frequency variation during TiO2 cycles (Ti(OiPr)4 + H2O), the frequency before the first pulse is named f0 and the frequencies after the nth cycle as fn (e.g. f1 after the first cycle, f2 after the second cycle etc.). The change in frequency – corresponding to the mass uptake during the cycle – is then obtained from: ∆ . The negative change in frequency is plotted in Figure 2 b) – due to the inversely proportional correlation of the frequency and mass of the crystal – ascending curve reveals a mass uptake and therefore film growth. The gained frequency variation during the first pulses of Ti(OiPr)4 + H2O is plotted in Figure 2 c): ∆fn decreases from the 1st cycle with approx. 9 Hz down to already approx. 4.5 Hz during the 2nd cycle and continues to decrease below 3 Hz. This trend points to the conclusion that the first cycles of TiO2 deposition onto an Al2O3 surface leads to a higher mass deposition than the deposition onto a TiO2 surface. This difference of deposition behavior, reflected in the mass gain, can only be explained by the different surfaces presented to the gaseous precursors. As demonstrated previously by Kim et al.46 no reaction between TMA and Ti(OiPr)4 is expected, hinting to a general difference in growth behavior of the surfaces of TiO2 and Al2O3. To further investigate the influence of different surfaces or modifications of the surfaces, a variable amount of Al2O3 pulses were deposited in between two TiO2 depositions and the frequency variation during the first 50 cycles was observed for each block. The frequency variation per pulse is depicted in Figure 3. The inset in Figure 3 a) displays the frequency variation when a variable amount of Al2O3 layers is deposited in between two blocks of TiO2 depositions. The change in frequency of each TiO2 cycle for 1, 2, 4, and 8 cycles of prior Al2O3 deposition is presented in Figure 3 a). The effect of a higher mass deposition rate caused by the prior Al2O3 deposition vanishes after approximately 10 cycles of TiO2 deposition. However for the first few pulses (
