Geometry of Chemical Beam Vapor Deposition System for Efficient

Research Article pubs.acs.org/acscombsci

Geometry of Chemical Beam Vapor Deposition System for Efficient Combinatorial Investigations of Thin Oxide Films: Deposited Film Properties versus Precursor Flow Simulations Estelle Wagner,*,†,‡ Cosmin S. Sandu,†,‡ Scott Harada,‡ Cedric Pellodi,§ Marc Jobin,§ Paul Muralt,⊥ and Giacomo Benvenuti†,‡ †

ABCD Technology, 12 Route de Champ-Colin, CH-1260 Nyon, Switzerland 3D-Oxides, Technoparc, 130 rue Gustave Eiffel, F-01630, Saint Genis Pouilly, France § HEPIA, University of Applied Sciences (HES-SO), 4 rue de la Prairie, CH-1202 Genève, Switzerland ⊥ Laboratoire de Céramique, Ecole Polytechnique Fédérale de Lausanne, Station 12, CH-1015 Lausanne, Switzerland ‡

S Supporting Information *

ABSTRACT: An innovative deposition system has been developed to construct complex material thin films from single-element precursors by chemical beam vapor deposition (CBVD). It relies on well distributed punctual sources that emit individually controlled precursor beams toward the substrate under high vacuum conditions combined with well designed cryo-panel surfaces that avoid secondary precursor sources. In this configuration the impinging flows of all precursors can be calculated at any substrate point considering the controlled angular distribution of the emitted beams and the ballistic trajectory of the molecules. The flow simulation is described in details. The major advantage of the deposition system is its ability to switch between several possible controlled combinatorial configurations, in which the substrate is exposed to a wide range of flow compositions from the different precursors, and a uniform configuration, in which the substrate is exposed to a homogeneous flow, even on large substrates, with high precursor use efficiency. Agreement between calculations and depositions carried out in various system configurations and for single, binary, or ternary oxides in mass transfer limited regime confirms that the distribution of incoming precursors on the substrate follows the theoretical models. Additionally, for some selected precursors and in some selected conditions, almost 100% of the precursor impinging on the substrate is incorporated to the deposit. The results of this work confirm the potentialities of CBVD both as a research tool to investigate efficiently deposition processes and as a fabrication tool to deposit on large surfaces. KEYWORDS: thin film deposition, oxides, chemical beams, combinatorial

■

Despite these advantages and the fabrication of devices,7−10 the expected breakthroughs did not take place and industrial CBE equipment11 disappeared from the market. Today, III−V semiconductor synthesis with CBE is mainly performed for nanowires12,13 or quantum dot deposition.14,15 In parallel to the main work on III−V semiconductors materials, which require a particular chemistry based mainly on alkyl compounds and hydrides, attempts were made to extend the technology to organometallic precursors for other material deposition. These included materials for Si technologies (Si, SixGe1−x,5 FeSi216), single oxides (Al2O3,17 CdO,18 CeO2,19 CuO,20 HfO2,21 MgO,22 SiO2,23 TiO2,24 Y2O3,25 ZnO,26 ZrO227), binary oxides (ErSiO,28 LiNbO3,29 LiTaO3,30 PbTiO331), and complex superconductive oxides (YBCO,32 La2−xBaxCuO433). These

INTRODUCTION Chemical beam epitaxy (CBE) and the related techniques of metalorganic molecular beam epitaxy (MOMBE) and gas source molecular beam epitaxy (GSMBE)1,2 originated in the 1980s3 as thin film deposition techniques for III−V semiconductors at the interface between chemical vapor deposition (CVD) and Molecular Beam Epitaxy (MBE). These technologies combine the chemical reaction of evaporated precursor molecules that decompose on heated substrates (CVD) with line-of-sight molecule trajectories of precursor beams from source to substrate (MBE). Applied to III−V semiconductor deposition, CBE demonstrated a wide range of advantages4−6 including easy deposition and doping of complex materials from various precursors, high composition and thickness uniformity even on large substrates, high precursor conversion rate, high growth rates, high reproducibility, compatibility with UHV characterization techniques, compatibility with laser beam structuring, and surface selective growth. © XXXX American Chemical Society

Received: September 14, 2015 Revised: January 28, 2016

A

DOI: 10.1021/acscombsci.5b00146 ACS Comb. Sci. XXXX, XXX, XXX−XXX

Research Article

ACS Combinatorial Science

sputtering,58 and CVD (low pressure,59 atmospheric,60 aerosolassisted61). Compared to these methods, CBVD has the unique advantage of allowing the reactant flows at each substrate point to be determined precisely and easily. In conjunction with characterization mapping, it is thus possible to know which flow conditions were responsible for a given physicochemical or functional property. The optimal value can then be reproduced homogeneously across an entire large substrate in a subsequent deposition, which compared to other system, reduces the gap between laboratory research and large scale production. A number of papers have been published showing deposition results of combinatorial study with the CBVD system (LiNbO3,62 Hf-doped LiNbO3,63 HfO2−Nb2O5 mixtures,64 Nb2O565) or with the HV-CVD system (Er-doped Al2O2,42 BaTiO3,66 TiO267) but none of them describes precisely the system conception and its geometry, or details the formalism that allows to calculate quantitatively the incoming flows on the substrate which is the aim of the present paper for the CBVD system (only the basics of the calculations were previously given36,68). Emphasis is put here on the good agreement between flow calculations and some selected model combinatorial experiments, such a “calibration” being a required step to analyze quantitatively combinatorial results referring to calculated impinging precursor flow.

experiments were published under the names CBE, CBD (Chemical Beam Deposition), MOMBE or HV-CVD (HighVacuum CVD), and mostly demonstrated the availability of more or less well behaving precursors for a wide range of elements. A recent publication reviews the oxide deposition work carried out with CVD under high vacuum conditions and discusses the chemical precursor requirements.34 As the precursor chemical decomposition reaction is purely a surface mechanism, with no gas phase reaction, these deposition techniques require precursors with additional properties with respect to standard CVD precursors (as discussed in20,35), including a high sticking coefficient and a high decomposition probability. It was already shown previously that titanium tetraisoproxide36 and niobium tetraethoxide dimethyl diamino ethoxide37 fulfill these requirements respectively for titanium and niobium oxides deposition with almost complete incorporation in the growing film of the incoming precursor molecules on the substrate. The present work confirms these results, expanding them to the case of binary and ternary oxide deposition, and additionally shows that zirconium tetraisobutoxide behaves similarly in some selected conditions. In the early 2000s, a new concept of deposition reactor developed, based on the geometric distribution of punctual Knudsen precursor sources around the substrate, surrounded by cryopanels, with the aim of shaping the precursor flow distribution on the substrate (chemical beam vapor deposition, CBVD). A first geometry, with a single ring of sources placed on top of a large volume prechamber, was optimized and proposed to deposit theoretically 1% homogeneous films (homogeneity defined as (maximum flow-minimum flow)/ minimum flow) on 6 in. wafers with 10% of effused precursor reaching the substrate38,39 (and was successfully applied to TiO2 deposition36). This geometry was later on modified to conceive a system with additional combinatorial facility for 3 precursors,40 which is presented in this Research Article, and allows switching between several combinatorial configurations and the homogeneous one. In parallel to this CBVD system with optimized flow design, another more compact system labeled HV-CVD41,42,34 was developed for smaller 4 in. wafers, replacing the prechamber ring by 3 tubes (1 tube per precursor), each with a ring of tilted sources on top, with a higher rate of precursor reaching the substrate (6−23%) but much lower homogeneity (9−28%), no easy conversion between combinatorial and homogeneous modes and in all cases, different distributions for the 3 different precursor lines.41 The CBVD system described in the present paper was particularly designed to target the expanding field of oxide material deposition,43−46 in which wide screening of material composition, crystallinity, morphology, and thickness is required to optimize multifunctional film properties. Inspired by pharmaceutical techniques, combinatorial methods47−50 have become a must since the pioneering work of Hanak51 for efficient material development studies. They encompass both deposition techniques (usually deposits with continuous composition gradients, obtained using the general principle of different material sources spatially distributed around the substrate) and high-throughput characterization techniques (with methods developed to rapidly screen chemical composition,52 crystallinity,53 thickness,54 or functional properties55 across large surface areas). A recent paper reviews combinatorial oxide deposition technologies and their applications.56 The continuous composition spread deposition techniques mainly include pulsed laser deposition (PLD),57

■

RESULTS AND DISCUSSION Chemical Beam Vapor Deposition Gas Delivery Principle. A schematic layout of the Sybilla 150 CBVD reactor is presented in Figure 1. The substrate is radiatively heated very homogeneously (which is a fundamental point of a CBE system69). Liquid nitrogen cryo-panels, surrounding the growth chamber, condense all impinging molecules (molecules effusing from the sources out of the substrate, molecules which

Figure 1. Schematic layout of the chemical beam vapor deposition system (Sybilla 150 from ABCD Technology). Three different precursors (labeled A, B, and C) are evaporated from thermostatically controlled reservoirs into a compartmentalized prechamber, with 6 areas per precursor. Precursor molecules effuse from the prechamber into the deposition chamber through Knudsen holes with a cosine angular distribution and have ballistic trajectories to the heated substrate where they decompose. Liquid nitrogen cooled walls trap all unreacted molecules and decomposition byproducts to maintain the deposition chamber under High-Vacuum conditions. B


Research Article


polar coordinate (R, αi, 0) that reaches a point M of polar coordinate (r, β, h) is given by eq 1. This flow is proportional to a constant Io76 (itself proportional to the prechamber pressure as the deposition chamber pressure is negligible) and to a constant C, which is related to the hole shape and dimension.77−79 In this work, the hole is a circular orifice with a diameter much larger than its thickness so C was approximated as 1. As all gas phase interactions are negligible, the flow emitted by a group of sources (eq 3) is the sum of the flows of emitted by the each source (eq 1). Integrating eq 1 for a full ring of hole sources lead to eq 4, which describes the flow at a point M emitted by a ring of N sources regularly distributed (for a large N number) . This equation is actually identical as the one that is obtained for a single source emission onto a rotating substrate.80,81 The dimensions of the system are chosen to maximize the fraction of emitted precursor that reaches the substrate, Eff (calculated with eq 2) and, simultaneously obtain a homogeneous flow over the whole substrate. In the Sybilla 150 system, the distances are optimized to provide a 1% flow homogeneity on a 6 in. wafer, resulting in R = 115 mm, h = 147.5 mm, Eff = 10.5%. In these conditions, N = 6 sources is sufficient to obtain a good approximation of eq 4. Consequently, to deposit homogeneously complex oxides, for instance for a ternary compound from 3 precursors, the 3 prechambers for the 3 precursors may be divided into 6 segments and alternated as shown in Figure 1. Homogeneous single element deposition with a simple compartment chamber for a single element deposition was reported elsewhere.36 In the present Research Article, we will focus on graded deposits of single and multielement oxides and show that the deposited thickness and the chemical composition are in good agreement with calculations, assuming that all precursor molecules reaching the substrate decompose. This imposes to work in a mass transfer limited deposition regime, which is usually obtained at intermediate surface temperature for single element deposition.82,34 Graded Thickness Single Element Oxide Films. For a single element deposition, the Sybilla 150 reactor allows the independent activation of each of the six effusive segments. TiO2 deposits (from Ti(OiPr)4, CAS = 546-68-9), Nb2O5 (from Nb(OEt)4 dmae, SAFC research product), HfO2 (from Hf(O-tBu)4, CAS = 2172.02.3), ZrO2 (from Zr(OtBu)4, CAS = 2081-12-1), and Ta2O5 (from Ta(OEt)5, CAS = 150747-55-0) were deposited varying the number of active segments. For these precursors, at sufficient substrate temperature to overcome the activation energy of the precursor decomposition reaction,82,83 the deposition takes place in the mass transfer limited regime, and the deposited thickness is directly proportional to the precursor flow, with a conversion efficiency factor of precursor into deposit close to 1. (Also possibly named “reactive sticking coefficient”, the conversion efficiency factor is defined as the ratio between the number of metal atoms incorporated in the deposited film (calculated from thin film thickness measurements) and the number of metal atoms impinging on the substrate (calculated from precursor flow simulation). The quite large uncertainty (up to 30%) on its value depends simultaneously on film density, on the measuring precision of precursor pressure in the prechamber and on the control of Knudsen effusing hole dimensions. Examples of such deposits are presented in Figure 3 in comparison with theoretical calculations obtained with eq 3. Excellent agreement is observed between calculated flows and

have touched the substrate without undergoing decomposition and deposition reaction byproducts). As pointed out already long ago,70 such cryo-panels are actually the detail that differentiates CBVD from a simpler high vacuum-CVD reactor, as they ensure that chemical beams travel with oriented (nonisotropic) ballistic trajectories from the sources to the substrate without gas phase collisions. In our system, they also ensure that the incoming flow rate on each substrate point can be calculated only considering precursor emitted by the sources. It is worth noting that these cryo-panels forbid the use of any carrier or oxidative gases that are not condensed at liquid nitrogen temperature. Precursors are evaporated continuously from thermostatically controlled reservoirs without any carrier gas to a prechamber with a constant pressure (prechamber pressure is typically related to reservoir temperature with an Arrhenius law) and the precursor molecules are then emitted to the growth chamber through orifices drilled in a plate. Such an evaporation system was shown to be particularly suitable for metalorganic oxide precursors.30 The question of which orifices to choose and how to distribute them to maximize flow homogeneity on the substrate and the fraction of emitted precursor reaching the substrate was debated.71−74 The innovative solution proposed in the Sybilla 150 model presented here is to place Knudsen holes on a circle (see Figure 2). Providing the holes are drilled

Figure 2. Geometric scheme of the system and flow calculation. Geometrical data of the system (distance source plane-substrate, h, source ring diameter R, substrate diameter, rsub, source area, A) are indicated to calculate the surface flow F (molecules/cm2/s) from the source Si (polar coordinates R, αi) that reaches a substrate point M (polar coordinates r, β) considering a prechamber pressure Pprec below the source Si, a prechamber temperature T, and a deposition chamber pressure PGR such that PGR ≪ Pprec. Rg is the gas constant (similar equations for the case of tilted sources used in other systems are reported in Supporting Information).

in a thin enough sheet, that their dimensions are small compared to the molecule mean free path and that the source plane-substrate distance is large with respect to the hole diameter,75 these holes can be considered as point sources and provide a cosine effusion of molecules in all directions per solid angles.76 Figure 2 details the flow calculations. On the basis of the indicated notations, the precursor flow emitted by a source Si of C


Research Article


Figure 3. Comparison of experimentally deposited thickness (normalized) and theoretical flow calculation as a function of the number of active prechamber segments along a 6 in. wafer diameter (in the main gradient direction), for a single element deposition. The main graph represents normalized flow data (continuous lines) and normalized experimental thickness (dots) for TiO2 or Nb2O5 deposits in the mass transfer limited regime, for a different number of active prechamber segments. Photographs of deposits on 6 in. wafers in which the interference colors correlate with deposited thickness have been inserted. Each curve corresponds to a given number of active segments and each situation is represented schematically in a rectangle below the main graph. The position of the active segment(s) is highlighted in black on the prechamber scheme and the main gradient direction diameter is represented in red on the 3D deposit shape simulation (in yellow). The flow ratio indicated corresponds to the ratio of higher flow to lower flow along this diameter.

The deposition was carried out on a 4 in. glass wafer at 440 ± 15 °C for 4 h. A photograph of the deposit, the measured thickness and the chemical composition, are presented in Figure 4. Optical thickness (converted into geometrical thickness assuming an homogeneous film refractive index) measured by reflectivity with an automated procedure on 94 × 108 points,55 is in excellent agreement with simulated thickness assuming that all the incoming precursor molecules decompose. An example of reflectivity measurement on a film point showing high transparency of the film is presented in Figure S1. The chemical composition was measured by EDX along 5 different diameters, and the Nb content was estimated as (Nb signal/(Nb signal+Ti signal)). A very good agreement is also obtained with the simulated Nb content calculated assuming again that all impinging precursor molecules decompose.

experimentally measured thickness. The data are normalized to the homogeneous flow that is obtained when all 6 segments are activated. With decreasing the number of active segments, the film growth rate is reduced and the flow ratio of the maximum flow to the minimum flow obtained on the wafer increases. With a single experiment, depending on the chosen number of active segment used, flow ratios from 1 to 6 can be obtained on a 6 in. wafer in this setup. Graded Composition Binary Oxide Films. Such an example of a multielement deposit in the mass transfer limited regime will be presented here for binary oxide depositions of Ti−Nb. A deposit was realized using a source segment for TTIP (partial pressure in the prechamber = 115 × 10−3 mbar) and a source segment for Nb(O-Et)4 dmae (partial pressure in the prechamber: 5.5 × 10−3 mbar), situated geometrically at 100°. D


Research Article


Figure 5. Ternary deposit Zr−Ti−Nb from Zr(O-tBu)4, Ti(O-iPr)4, and Nb(OEt)4 dmae on a 4 in. SiO2 substrate. (a) Theoretical mapping of the expected deposit. The positions of the 3 precursor sources used are highlighted (one segment per precursor). The colored diameters correspond to directions along which experimental EDX characterizations were carried out. (b) Measured chemical compositions by EDX are reported on the ternary diagram, comparing experimental measurements (dots) to calculated values (lines). The colors on the ternary diagram correspond to the color of the diameter in panel a on which they were measured. The yellow colored area corresponds to the calculated zones of chemical composition that should be present in the deposit on the full wafer.

either to a substrate misalignment or to a limited precursor interaction at the substrate surface. Additionally to metallic element concentration, EDX measurements show an almost constant oxygen content of 67 ± 1% (Figure S3), in good agreement with expected oxide composition. XRD diffractogram at the film center shows the films is quasi amorphous, although the presence of some small crystallites is evidenced by TEM cross section images in the Zr richer region (Figure S4).

Figure 4. Example of a binary oxide deposition, namely, Ti(1−x)NbxO2 under deposition conditions for which element incorporation is proportional to precursor flow composition. A photograph of the deposit is shown in panel a, highlighting the positions of the Ti and Nb precursor sources that are used. In panel d, thickness mapping realized by UV−vis spectroscopy, is compared to the theoretical deposit thickness calculated assuming that all incoming precursor molecules decompose, represented in panel c. The plot of both values as a function of the position on the wafer along the main gradient diameter (highlighted in pink in panel a) is shown in panel b (red line = calculation, blue dots = experimental points). In panel e, a calculated map of the Nb content in the deposit is shown (Nb content increasing from black to white). The colored diameters correspond to the directions along which chemical composition was measured experimentally by EDX. The graph in panel f compares calculated Nb content along these diameters (continuous lines) to experimental measurements (dots). Each line/dot color corresponding to that of the diameter drawn with the same color in panel e.

■

CONCLUSION Provided depositions are undertaken in conditions of flow corresponding to the mass transfer limited regime, we have shown that the presented CBVD reactor geometry enables the deposition of graded and homogeneous thickness single element thin films, with excellent agreement between the experimentally measured and the calculated thickness. Similarly, using alkoxide precursors (Ti(O-IPr)4, Nb(O-Et)4 dmae, Zr(O-tBu)4), it is possible, under certain conditions of flow and substrate temperature, to obtain multielement deposits in which the precursor incorporation follows quantitatively the flow composition. CBVD as a combinatorial deposition technique has an advantage compared to its competitors in that the impinging precursor flow on the substrate can be easily modeled. Therefore, it is possible to correlate deposit properties to deposition conditions. It enables fast and efficient material screening, as a functional property can be evaluated on a single wafer that corresponds to a wide range of chemical compositions or thicknesses. The deposition conditions of the region exhibiting the most desirable properties can then be reproduced homogeneously on the whole substrate by simply activating a full ring of precursor sources. The multielement deposition presented in this work corresponds to “ideal” conditions in which all impinging precursor molecules for every element are incorporated in the deposited films. However, under other conditions, precursor decomposition reactions are affected by the presence of other compounds, resulting in preferential incorporation of one element over another, in the activation of a precursor decomposition reaction by another precursor, which would

The deposited film has an almost constant O content (on all measured points of the wafers) as measured by EDX (O/(Ti + Nb + O) = 0.71 ± 0.1) and exhibits a nanocrystalline anatase structure, with substitution of Nb in the TiO2 lattice as observed on the XRD pattern (Figure S2). Graded Composition Ternary Oxide Films. A ternary oxide, Ti−Nb−Zr, was deposited using a source segment by precursor, from Ti(O-iPr)4, (partial pressure in the prechamber 11 × 10−3 mbar), Nb(O-Et)4dmae (partial pressure in the prechamber 11 × 10−3 mbar), and Zr(O-tBu)4 (partial pressure in the prechamber 31 × 10−3 mbar) on a 4 in. glass substrate at 425 ± 5 °C for 1 h. Chemical composition on 4 different diameters was measured by EDX. Experimental results compared to calculated chemical composition are presented in Figure 5. Reasonable agreement is obtained between calculation and experiment, although a slight discrepancy is observed, that may be due E


Research Article


(13) García Núñez, C.; Brana, A. F.; Pau, J. L.; Ghita, D.; Garcia, B. J.; Shen, G.; Wilbert, D. S.; Kim, S. M.; Kung, P. Pure zincblende GaAs nanowires grown by Ga-assisted Chemical Beam Epitaxy. J. Cryst. Growth 2013, 372, 205−212. (14) Zribi, J.; Morris, D.; Ares, R. Formation and morphological evolution of InAs quantum dots grown by Chemical Beam Epitaxy. J. Vac. Sci. Technol. B 2012, 30 (5), 051207. (15) Gong, Q.; Notzel, R.; van Veldhoven, P. J.; Eijkemans, T. J.; Wolter, J. H. Wavelength tuning of InAs quantum dots grown on InP (100) by Chemical-Beam Epitaxy. Appl. Phys. Lett. 2004, 84 (2), 275− 277. (16) Crumbaker, T. E.; Natoli, J. Y.; Berbezier, I.; Derrien, J. Growth of β-FeSi2 on silicon substrates by Chemical Beam Epitaxy. J. Cryst. Growth 1993, 127 (1−4), 158−164. (17) Iizuka, H.; Yokoo, K.; Ono, S. Growth of single crystalline γAl2O3 layers on silicon by Metalorganic Molecular-Beam Epitaxy. Appl. Phys. Lett. 1992, 61 (25), 2978−2980. (18) Ashrafi, A. B. M. A.; Kumano, H.; Suemune, I.; Ok, Y. W.; Seong, T. Y. CdO epitaxial layers grown on (001) GaAs surfaces by Metalorganic Molecular-Beam Epitaxy. J. Cryst. Growth 2002, 237, 518−522. (19) Ikegawa, S.; Motoi, Y. Growth of CeO2 thin films by MetalOrganic Molecular Beam Epitaxy. Thin Solid Films 1996, 282 (1−2), 60−63. (20) Fritsch, E.; Machler, E.; Arrouy, F.; Berke, H.; Povey, I.; Willmott, P. R.; Locquet, J. P. Schiff base precursor compounds for the Chemical Beam Epitaxy of oxide thin films 0.1. Deposition of CuO on MgO[001] using copper(II) bis(benzoylacetone)-ethylendiimine. J. Vac. Sci. Technol., A 1996, 14 (6), 3208−3213. (21) Hong, J. H.; Moon, T. H.; Myoung, J. M. Microstructure and characteristics of the HfO2 dielectric layers grown by Metalorganic Molecular Beam Epitaxy. Microelectron. Eng. 2004, 75 (3), 263−268. (22) Niu, F.; Hoerman, B. H.; Wessels, B. W. Epitaxial thin films of MgO and Si using Metalorganic Molecular Beam Epitaxy. J. Vac. Sci. Technol., B: Microelectron. Process. Phenom. 2000, 18 (4), 2146−2152. (23) Bonzel, H. P.; Pirug, G.; Verhasselt, J. Low temperature growth of SiO2 films on Si(100) using a hot molecular beam of tetraethoxysilane. Chem. Phys. Lett. 1997, 271 (1−3), 113−117. (24) Taylor, C. J.; Gilmer, D. C.; Colombo, D. G.; Wilk, G. D.; Campbell, S. A.; Roberts, J.; Gladfelter, W. L. Does chemistry really matter in the chemical vapor deposition of titanium dioxide? Precursor and kinetic effects on the microstructure of polycrystalline films. J. Am. Chem. Soc. 1999, 121 (22), 5220−5229. (25) Fritsch, E.; Machler, E.; Arrouy, F.; Orama, O.; Berke, H.; Povey, I.; Willmott, P. R.; Locquet, J. P. Benzoylpivaloylmethanide precursors for the Chemical Beam Epitaxy of oxide thin films 0.1. Synthesis, characterization, and use of yttrium benzoylpivaloylmethanide. Chem. Mater. 1997, 9 (1), 127−134. (26) Terasako, T.; Yura, S.; Azuma, S.; Shimomura, S.; Shirakata, S.; Yagi, M. Comparative study on structural and optical properties of ZnO films grown by metalorganic molecular beam deposition and metalorganic chemical vapor deposition. J. Vac. Sci. Technol. B 2009, 27 (3), 1609−1614. (27) Kim, M. S.; Ko, Y. D.; Hong, J. H.; Jeong, M. C.; Myoung, J. M.; Yun, I. Characteristics and processing effects of ZrO2 thin films grown by metal-organic molecular beam epitaxy. Appl. Surf. Sci. 2004, 227 (1−4), 387−398. (28) Isshiki, H.; Masaki, K.; Ueda, K.; Tateishi, K.; Kimura, T. Towards epitaxial growth of ErSiO nanostructured crystalline films on Si substrates. Opt. Mater. 2006, 28 (6−7), 855−858. (29) Joshkin, V. A.; Moran, P.; Saulys, D.; Kuech, T. F.; McCaughan, L.; Oktyabrsky, S. R. Growth of oriented lithium niobate on silicon by alternating gas flow chemical beam epitaxy with metalorganic precursors. Appl. Phys. Lett. 2000, 76 (15), 2125−2127. (30) Bellman, R.; Raj, R. Design and performance of a new type of Knudsen cell for chemical beam epitaxy using metal-organic precursors. Vacuum 1997, 48 (2), 165−173.

otherwise not occur, or in the prevention of a precursor being incorporated into the film. In these cases, the combinatorial approach is a valuable tool to discover and understand precursor interaction mechanisms at the substrate surface, and consequently study.precursor decomposition kinetics.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acscombsci.5b00146. Modified flow equations in case of tilted sources; UV−vis and XRD measurements of Nb:TiO2 deposited films; EDX, XRD, and TEM characterization of (Zr,Ti,Nb) ternary oxide deposits (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors wish to acknowledge the FEDER (Fonds Européen de Développement Economique et Régional) (5887/2013-2011) for financing the Nanobium project through the Interreg IVA program.

■

REFERENCES

(1) Foord, J. S.; Davies, G. J.; Tsang, W. T. Chemical Beam Epitaxy and Related Techniques; Wiley, 1997. (2) Tsang, W. T. Advances in MOVPE, MBE, and CBE. J. Cryst. Growth 1992, 120, 1−4. (3) Veuhoff, E.; Pletschen, W.; Balk, P.; Luth, H. Metalorganic CVD of GaAs in a molecular-beam system. J. Cryst. Growth 1981, 55 (1), 30−34. (4) Benchimol, J. L.; Alexandre, F.; Lamare, B.; Legay, P. Benefits of Chemical Beam Epitaxy for micro and optoelectronic applications. Prog. Cryst. Growth Charact. Mater. 1996, 33 (4), 473−495. (5) Luth, H. Chemical Beam Epitaxy - a child of surface science. Surf. Sci. 1994, 299 (1−3), 867−877. (6) Martin, T.; Whitehouse, C. R.; Lane, P. A. Growth reactions and mechanisms in Chemical Beam Epitaxy (CBE). J. Cryst. Growth 1992, 120 (1−4), 25−32. (7) Choa, F. S.; Tsang, W. T.; Logan, R. A.; Gnall, R. P.; Koch, T. L.; Burrus, C. A.; Wu, M. C.; Chen, Y. K.; Kapre, R. InGaAs/InGaAsP integrated tunable detector grown by chemical beam epitaxy. Appl. Phys. Lett. 1993, 63 (13), 1836−1838. (8) Nutsch, A.; Dahlheimer, B.; Dohr, N.; Kratzer, H.; Lukas, R.; Torabi, B.; Trankle, G.; Abstreiter, G.; Weimann, G. Chemical Beam Epitaxy of integrated 1.55 μm lasers on exact and misoriented (100)InP substrates. J. Cryst. Growth 1998, 188 (1−4), 275−280. (9) Veuhoff, E. Potential of MOMBE/CBE for the production of photonic devices in comparison with MOVPE. J. Cryst. Growth 1998, 188 (1−4), 231−246. (10) Lelarge, F.; Dagens, B.; Cuisin, C.; Le Gouezigou, O.; Patriarche, G.; Van Parys, W.; Vanwolleghem, M.; Baets, R.; Gentner, J. L. GSMBE growth of GaInAsP/InP 1.3 μm-TM-lasers for monolithic integration with optical waveguide isolator. J. Cryst. Growth 2005, 278 (1−4), 709−713. (11) Tsang, W. T. Chemical Beam Epitaxy. IEEE Circuit Devic. 1988, 4 (5), 18−24. (12) Persson, A. I.; Froberg, L. E.; Jeppesen, S.; Bjork, M. T.; Samuelson, L. Surface diffusion effects on growth of nanowires by Chemical Beam Epitaxy. J. Appl. Phys. 2007, 101 (3), 034313. F


Research Article

ACS Combinatorial Science (31) Hsieh, K. Y.; Rou, S. H.; King, L. L. H.; Kingon, A. I. Synthesis and characterization of PbTiO3 thin-films grown by chemical beam deposition. MRS Online Proc. Libr. 1990, 204, 551−556. (32) Endo, K.; Moriyasu, Y.; Teherani, F. H.; Yoshizawa, T.; Badic, P.; Abe, K.; Itoh, J.; Kajimura, K. Preparation of YBCO superconducting films by MOMBE. Phys. C 2002, 372, 604−607. (33) Machler, E.; Locquet, J. P.; Fritsch, E.; Williams, E. J.; Willmott, P.; Lingenauer, M.; Felder, P.; Huber, J. R.; Berke, H. Superconducting La2‑xBaxCuO4 thin-films prepared by chemical beam epitaxy. Phys. C 1994, 235, 705−706. (34) Kuzminykh, Y.; Dabirian, A.; Reinke, M.; Hoffmann, P. High vacuum chemical vapor deposition of oxides: A review of technique development and precursor selection. Surf. Coat. Technol. 2013, 230, 13−21. (35) Machler, E.; Arrouy, F.; Fritsch, E.; Bednorz, J. G.; Berke, H.; Huber, J. R.; Locquet, J. P. Using phthalocyanine precursors to prepare oxide thin films: Decoupling the growth rate from the evaporation rate. Appl. Phys. Lett. 1997, 71 (5), 710−712. (36) Benvenuti, G.; Halary-Wagner, E.; Brioude, A.; Hoffmann, P. High uniformity deposition with chemical beams in high vacuum. Thin Solid Films 2003, 427 (1−2), 411−416. (37) Dabirian, A.; Kuzminykh, Y.; Wagner, E.; Benvenuti, G.; Rushworth, S.; Hoffmann, P. Evaluation of niobium dimethylaminoethoxide for chemical vapour deposition of niobium oxide thin films. Thin Solid Films 2014, 571, 94−101. (38) Benvenuti, G. Chemical beam deposition of titanium dioxide thin films. PhD Thesis, Ecole Polytechnique Fédérale de Lausanne (EPFL), Switzerland, 2003. DOI:10.5075/epfl-thesis-2744. (39) Benvenuti, G.; Halary-Wagner, E.; Amorosi, S.; Hoffmann, P. (Ecole Polytechnique Fédérale de Lausanne (Switzerland)) Large area deposition in high vacuum with high thickness uniformity. Eur. Patent EP1504136, 2003. (40) Benvenuti, G.; Halary-Wagner, E.; Petit, C. (ABCD Technology S.A. (Switzerland)) Vapor phase deposition system. Eur. Patent EP2347030 (A1), 2008. (41) Multone, X. High vacuum chemical vapor deposition (HVCVD) of alumina thin films. PhD Thesis, Ecole Polytechnique Fédérale de Lausanne (EPFL), Switzerland, (2009. DOI:10.5075/epflthesis-4485. (42) Multone, X.; Luo, Y.; Hoffmann, P. Er-doped Al2O3 thin films deposited by high-vacuum chemical vapor deposition (HV-CVD). Mater. Sci. Eng., B 2008, 146 (1−3), 35−40. (43) Blamire, M. G.; MacManus-Driscoll, J. L.; Mathur, N. D.; Barber, Z. H. The materials science of functional oxide thin films. Adv. Mater. 2009, 21 (38−39), 3827−3839. (44) Heber, J. Materials science: Enter the oxides. Nature 2009, 459 (7243), 28−30. (45) Koinuma, H. Chemistry and electronics of oxides from carbon dioxide to perovskite. Thin Solid Films 2005, 486 (1−2), 2−10. (46) Posadas, A. B.; Lippmaa, M.; Walker, F. J.; Dawber, M.; Ahn, C. H.; Triscone, J.-M. Growth and novel applications of epitaxial oxide thin films. Top. Appl. Phys. 2007, 105, 219−304. (47) Green, M. L.; Takeuchi, I.; Hattrick-Simpers, J. R. Applications of high throughput (combinatorial) methodologies to electronic, magnetic, optical, and energy-related materials. J. Appl. Phys. 2013, 113 (23), 231101. (48) Potyrailo, R.; Rajan, K.; Stoewe, K.; Takeuchi, I.; Chisholm, B.; Lam, H. Combinatorial and high-throughput screening of materials libraries: Review of state of the art. ACS Comb. Sci. 2011, 13 (6), 579− 633. (49) Xiang, X. D. Combinatorial materials synthesis and screening: An integrated materials chip approach to discovery and optimization of functional materials. Annu. Rev. Mater. Sci. 1999, 29, 149−171. (50) Zhao, J. C. Combinatorial approaches as effective tools in the study of phase diagrams and composition-structure-property relationships. Prog. Mater. Sci. 2006, 51 (5), 557−631. (51) Hanak, J. J. Multiple-sample-concept in materials research synthesis, compositional analysis and testing of entire multicomponent systems. J. Mater. Sci. 1970, 5 (11), 964−971.

(52) Gamez, G.; Mohanty, G.; Michler, J. Ultrafast elemental mapping of materials combinatorial libraries and high-throughput screening samples via pulsed glow discharge optical emission spectroscopy. J. Anal. At. Spectrom. 2013, 28 (7), 1016−1023. (53) Long, C. J.; Bunker, D.; Li, X.; Karen, V. L.; Takeuchi, I. Rapid identification of structural phases in combinatorial thin-film libraries using x-ray diffraction and non-negative matrix factorization. Rev. Sci. Instrum. 2009, 80 (10), 103902. (54) Leon, M. D.; Adomaitis, R. A. Full-wafer mapping and response surface modeling techniques for thin film deposition processes. J. Cryst. Growth 2009, 311 (13), 3399−3408. (55) Jobin, M.; Jotterand, S.; Pellodi, C.; dos Santos, S.; Sandu, C. S.; Wagner, E.; Benvenuti, G. Full wafer metrology for chemically graded thin films. P. Proc. SPIE 2012, 8430, 843009. (56) Kafizas, A.; Parkin, I. P. Inorganic thin-film combinatorial studies for rapidly optimizing functional properties. Chem. Soc. Rev. 2012, 41 (2), 738−781. (57) Green, M. L.; Schenck, P. K.; Chang, K. S.; Ruglovsky, J.; Vaudin, M. ″higher-kappa″ dielectrics for advanced silicon microelectronic devices: A combinatorial research study. Microelectron. Eng. 2009, 86 (7−9), 1662−1664. (58) Heo, G. S.; Gim, I. G.; Park, J. W.; Kim, K. Y.; Kim, T. W. Effects of substrate temperature on properties of ITO-ZnO composition spread films fabricated by combinatorial rf magnetron sputtering. J. Solid State Chem. 2009, 182 (10), 2937−2940. (59) Xia, B.; Smith, R. C.; Moersch, T. L.; Gladfelter, W. L. Balancing reactor fluid dynamics and deposition kinetics to achieve compositional variation in combinatorial chemical vapor depositions. Appl. Surf. Sci. 2004, 223 (1−3), 14−19. (60) Kafizas, A.; Hyett, G.; Parkin, I. P. Combinatorial atmospheric pressure chemical vapour deposition (cAPCVD) of a mixed vanadium oxide and vanadium oxynitride thin film. J. Mater. Chem. 2009, 19 (10), 1399−1408. (61) Knapp, C. E.; Kafizas, A.; Parkin, I. P.; Carmalt, C. J. The use of combinatorial aerosol-assisted chemical vapour deposition for the formation of gallium-indium-oxide thin films. J. Mater. Chem. 2011, 21 (34), 12644−12649. (62) Dabirian, A.; Harada, S.; Kuzminykh, Y.; Sandu, S. C.; Wagner, E.; Benvenuti, G.; Brodard, P.; Rushworth, S.; Muralt, P.; Hoffmann, P. Combinatorial chemical beam epitaxy of lithium niobate thin films on sapphire. J. Electrochem. Soc. 2011, 158 (2), D72−D76. (63) Dabirian, A.; Kuzminykh, Y.; Sandu, S. C.; Harada, S.; Wagner, E.; Brodard, P.; Benvenuti, G.; Rushworth, S.; Muralt, P.; Hoffmann, P. Combinatorial high-vacuum chemical vapor deposition of textured hafnium-doped lithium niobate thin films on sapphire. Cryst. Growth Des. 2011, 11 (1), 203−209. (64) Dabirian, A.; Kuzminykh, Y.; Afra, B.; Harada, S.; Wagner, E.; Sandu, C. S.; Benvenuti, G.; Rushworth, S.; Muralt, P.; Hoffmann, P. Combinatorial discovery and optimization of amorphous HfO2-Nb2O5 mixture with improved transparency. Electrochem. Solid-State Lett. 2010, 13 (7), G60−G63. (65) Dabirian, A.; Kuzminykh, Y.; Harada, S.; Parsons, C.; Sandu, S. C.; Wagner, E.; Benvenuti, G.; Rushworth, S.; Muralt, P.; Hoffmann, P. Efficient optimization of high vacuum chemical vapor deposition of niobium oxide on full wafer scale. IOP Conf. Ser.: Mater. Sci. Eng. 2010, 8, 012026. (66) Reinke, M.; Kuzminykh, Y.; Malandrino, G.; Hoffmann, P. Precursor adsorption efficiency of titanium tetra isopropoxide in the presence of a barium beta-diketonate precursor. Surf. Coat. Technol. 2013, 230, 297−304. (67) Reinke, M.; Ponomarev, E.; Kuzminykh, Y.; Hoffmann, P. Combinatorial characterization of TiO2 chemical vapor deposition utilizing titanium isopropoxide. ACS Comb. Sci. 2015, 17 (7), 413− 420. (68) Reinke, M.; Kuzminykh, Y.; Hoffmann, P. Limitations of patterning thin films by shadow mask high vacuum chemical vapor deposition. Thin Solid Films 2014, 563, 56−61. G


Research Article

ACS Combinatorial Science (69) Marheineke, B.; Popp, M.; Heinecke, H. Growth of GaInAs(P) using a multiwafer MOMBE. J. Cryst. Growth 1996, 164 (1−4), 16− 21. (70) Fraas, L. M.; McLeod, P. S.; Cape, J. A.; Partain, L. D. Epitaxialfilms grown by vacuum MOCVD. J. Cryst. Growth 1984, 68 (1), 490− 496. (71) Herman, M. A. Physical problems concerning effusion processes of semiconductors in Molecular-Beam Epitaxy. Vacuum 1982, 32 (9), 555−565. (72) Campbell, C. T.; Valone, S. M. Design considerations for simple gas dosers in surface science applications. J. Vac. Sci. Technol., A 1985, 3 (2), 408−411. (73) Guevremont, J. M.; Sheldon, S.; Zaera, F. Design and characterization of collimated effusive gas beam sources: Effect of source dimensions and backing pressure on total flow and beam profile. Rev. Sci. Instrum. 2000, 71 (10), 3869−3881. (74) Kuhl, D. E.; Tobin, R. G. On the design of capillary and effusive gas dosers for surface science. Rev. Sci. Instrum. 1995, 66 (4), 3016− 3020. (75) Adamson, S.; Ocarroll, C.; McGilp, J. F. Monte-Carlo calculations of the beam flux distribution from Molecular-Beam Epitaxy sources. J. Vac. Sci. Technol., B: Microelectron. Process. Phenom. 1989, 7 (3), 487−490. (76) Steckelmacher, W. Knudsen-flow 75 years on - the current stateof-the-art for flow of rarefied-gases in tubes and systems. Rep. Prog. Phys. 1986, 49 (10), 1083−1107. (77) Grimley, R. T.; Wagner, L. C.; Castle, P. M. Angulardistributions of molecular species effusing from near-ideal orifices. J. Phys. Chem. 1975, 79 (4), 302−308. (78) Fenter, F. F.; Caloz, F.; Rossi, M. J. Paper II: Simulation of flow conditions in low-pressure flow reactors (Knudsen cells) using a Monte Carlo technique. Rev. Sci. Instrum. 1997, 68 (8), 3180−3186. (79) Gomez-Goni, J.; Lobo, P. J. Comparison between Monte Carlo and analytical calculation of the conductance of cylindrical and conical tubes. J. Vac. Sci. Technol., A 2003, 21 (4), 1452−1457. (80) Dobrowolski, J. A.; Ranger, M.; Wilkinson, R. L. Measurements of angular evaporation characteristics of sources. J. Vac. Sci. Technol., A 1983, 1 (3), 1403−1408. (81) Teeter, G. Conceptual design of a deposition system for uniform and combinatorial synthesis of multinary thin-film materials via openboat physical-vapor deposition. J. Vac. Sci. Technol., A 2006, 24 (4), 1119−1127. (82) Wagner, E.; Sandu, C. S.; Harada, S.; Benvenuti, G.; Savu, V.; Muralt, P. Fabrication of complex oxide microstructures by combinatorial chemical beam vapour deposition through stencil masks. Thin Solid Films 2015, 586, 64−69. (83) Dabirian, A.; Kuzminykh, Y.; Wagner, E.; Benvenuti, G.; Rushworth, S. A.; Hoffmann, P. Chemical vapor deposition kinetics and localized growth regimes in combinatorial experiments. ChemPhysChem 2011, 12 (18), 3524−3528.

H


Geometry of Chemical Beam Vapor Deposition System for Efficient

Recommend Documents