Quantitative Analysis of Nanoscale Step Dynamics in High

Subscriber access provided by ORTA DOGU TEKNIK UNIVERSITESI KUTUPHANESI

Article

Quantitative Analysis of Nanoscale Step Dynamics in High-temperature Solution-grown Single Crystal 4HSiC via in situ Confocal Laser Scanning Microscope Aomi Onuma, Shingo Maruyama, Naoyoshi Komatsu, Takeshi Mitani, Tomohisa Kato, Hajime Okumura, and Yuji Matsumoto Cryst. Growth Des., Just Accepted Manuscript • DOI: 10.1021/acs.cgd.7b00325 • Publication Date (Web): 14 Apr 2017 Downloaded from http://pubs.acs.org on April 18, 2017

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 free 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 accessible to all readers and 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.

Crystal Growth & Design 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 27

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

Crystal Growth & Design

Quantitative Analysis of Nanoscale Step Dynamics in High-temperature Solution-grown Single Crystal 4H-SiC via in situ Confocal Laser Scanning Microscope Aomi Onuma†, Shingo Maruyama†, Naoyoshi Komatsu‡, Takeshi Mitani‡, Tomohisa Kato‡, Hajime Okumura‡ and Yuji Matsumoto†,* †

Department of Applied Chemistry, School of Engineering, Tohoku University, Sendai 980-

8579, Japan ‡

National Institute of Advanced Industrial Science and Technology, Ibaraki 305-8569, Japan

*Correspondence to: [email protected]

Abstract: Nanoscale understanding of high-temperature crystal growth dynamics in solution has been a challenge to be tackled by many researchers engaged in investigating solution processes for bulk single crystal growth. Here we propose a new approach to in situ observation at a buried solid/liquid interface in high-temperature solution using a conventional confocal laser scanning microscope. In the solution growth of 4H-SiC with Si-Ni based alloy flux as a model system, we show the ability to quantitatively analyze step motions at the growing SiC crystal on the

ACS Paragon Plus Environment

1


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

Page 2 of 27

nanoscale at high temperatures up to 1700 °C in a vacuum. The temperature-dependent stepadvance rates for various steps with different step heights demonstrated the advantageous effect of adding Al to the flux on the step-flow growth of SiC: addition of just 4 at% Al effectively suppressed step-bunching. These experiments point to the importance of in situ nanoscale observation in understanding solution growth mechanisms, and hence the potential to accelerate the development of solution growth processes for high-quality bulk single crystals.

In situ observation of the solution growth of crystals on the nanoscale is a powerful approach to look into the details of the growth processes and thereby understand the growth mechanisms involved. In fact, so far there have been many attempts at developing experimental microscopic techniques for such in situ nanocale observations; however, they are limited to low-temperature solution growth systems (typically less than 300 °C). Examples include phase contrast microscopy,1,2 laser microscopy combined with differential interference contrast microscopy,3–7 phase shift interferometry,8–10 and atomic force microscopy (AFM),6,11–15 for various techniques for growing bulk crystals of inorganic salts and organic compounds from solutions. On the other hand, the nanoscale observation of crystal growth dynamics in a high-temperature solution (e.g. >1000 °C for oxides and carbides) has been hampered by the lack of any suitable experimental technique. While there have been some attempts to observe high-temperature liquid–solid interfaces by in situ optical microscopy techniques,16–24 most of them have focused on large scale crystal shape dynamics typically up to several hundred micrometers, which are limited by the optical setting. Thus, nanoscale quantitative analysis of the step dynamics at high-temperature has scarcely been reported.


2

Page 3 of 27

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


Silicon carbide (SiC) has attracted much attention for applications in high-temperature and highpower electric devices owing to its specific properties.25 Recently a solution growth method for SiC bulk single crystals has been developed, and this method is a promising, alternative growth method to the existing sublimation process, resulting in improved crystal quality. With this method, a lower dislocation density is expected because the growth process should proceed under conditions much closer to thermal equilibrium.26,27 However, one of the most serious drawbacks in the solution growth of SiC crystals is that the carbon solubility in Si melt is quite low. Thus, with the primary aim of enhancing the carbon solubility, multi-component flux systems including various rare earth and transition metals etc. in the Si melt have been the subject of intense investigation. As a result, additional, positive roles of these additives in the flux sometimes can be found in the control of, for example, polytypes and/or surface morphology of grown SiC crystals.28 In fact, for example, our recent study on the solution growth of SiC crystals revealed that a small addition of Al to the Si-Cr flux markedly improved the surface flatness of the grown SiC crystal as compared with the case where only the Si-Cr flux was used.28 However, the mechanism of such a morphological change due to the Al addition has not been sufficiently clarified. Here, we report a novel approach to quantitative analysis of nanoscale step dynamics in the liquid phase homoepitaxial growth of SiC crystals with “non-transparent” Si-Ni flux and 4 at% Al-added Si-Ni flux through in situ confocal laser scanning microscope (CLSM)29 observation up to 1700 °C and the quantitative analysis of nanoscale step dynamics.


3


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

Page 4 of 27

EXPERIMENTAL Sample Preparation. We choosed Si-Ni based flux to grow single crystal SiC because it was found to be as a good candidate for fabricating high quality SiC thin films in the previous our studies.30–32 To ensure smooth wetting of bulk powder flux on the flux/SiC interface when it melted, a 150 nm-thick Si-Ni (Si0.67Ni0.33) or Al-added Si-Ni (Si0.65Ni0.31Al0.04) alloy thin film was first deposited on a semi-insulating 4H-SiC (0001) Si-face seed substrate by pulsed laser deposition (PLD; Pascal Co., LTD.) with a base pressure of ~10-9 Torr. Next, a powder of similar composition Si-Ni (Si0.67Ni0.33) or Al-added Si-Ni (Si0.65Ni0.31Al0.04) serving as bulk flux was prepared on another 4H-SiC (0001) substrate, which was used as a SiC source. The seed substrate was then placed on the source substrate, with the thin film and powder fluxes between the substrates. The distance between the source and seed SiC substrates was at least a few of hundred micrometers. In situ CLSM Experiments. In situ CLSM observation of a flux/SiC interface was carried out in a high-vacuum chamber (Kitano Seiki Co., LTD.) with a base pressure of ~10-8 Torr. The in situ CLSM observation was done through a quartz viewport. The CLSM was a Keyence VKX120/130 equipped with a long-working-distance lens (×20, working distance = 20.5 mm, numerical aperture = 0.35). Since 4H-SiC is transparent at the wavelength of the CLSM incident probe laser light (658 nm) even at the high temperatures used in this study,33 we could observe the flux/SiC interface through the SiC substrate. Our full experimental setup of the chamber is shown in Figure S1. The tempearture of the carbon suceptor was measured by a pyrometer. The sample heating rate was 30 °C/min for a sample temperature range below 1000 °C and 10 °C/min above 1000 °C, and the cooling rate was 100 °C/min until the temperature was lowered to 300 °C. All the in situ CLSM observations were made while heating up.


4

Page 5 of 27

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


Calculation of the Carbon Solubility. The equilibrium solubility of carbon in solution was evaluated from the liquidus curve in the phase diagram of the Si–Ni-C and Si–Ni–Al–C systems calculated using the CALPHAD method.34 The phase diagrams were calculated by using Thermo-Calc (Thermo-Calc Software, Sweden). In this modeling, the liquid phase was described as a substitutional solution, and the Gibbs energy functions for Si, Ni, Al, and C were taken from the SGTE database (SSOL5) implemented in Thermo-Calc.

RESULTS AND DISCUSSION In situ Laser Scanning Microscope Observation. Typically for the solution growth of SiC single crystals, a top-seeded solution growth method is employed. In this method, a SiC seed crystal is dipped into a Si-X flux melt in a graphite crucible serving as a carbon source. In our experiments, instead we carried out SiC growth using a travelling solvent method with a sandwich configuration, as shown in Figure 1a, as a model system of the solution growth of SiC, which is often used for liquid phase epitaxial growth.35,36 The sample was heated by a Nd:YAG laser from the back side of the source substrate, to which a carbon susceptor was attached so as to efficiently absorb the laser light. This heating configuration made the source substrate temperature higher than the seed substrate one, and the resultant small temperature gradient between these two substrates was a driving force for the SiC mass transfer from the source to seed via the flux. The probe laser light (658 nm) of the CLSM was introduced through the transparent seed SiC substrate and was focused on the growth interface.


5


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

Page 6 of 27

Figure 1b and 1c show a set of CLSM images captured in a temperature range of 1400 to 1700 °C for Si-Ni (Si0.67Ni0.33) flux and Al-added Si-Ni (Si0.65Ni0.31Al0.04) flux, respectively. During the first heating step from room temperature, a Si-Ni flux thin film pre-deposited on the seed SiC started to melt at ~930 °C, followed by the complete melting of bulk flux powders prepared on the source SiC. The melting temperature range was around the bulk melting point of Si0.67Ni0.33 alloy (966 °C).37 These fluxes then became fused together between the SiC substrates (see Video S1 and Figure S2), accompanied by the formation of some Si precipitates below 1100 °C. Subsequently, many 2D islands with random shapes (indicated by the black arrow in Fig. 1b) were seen to grow at temperatures up to 1480 °C, as shown in Fig. 1b. After that, shrinkage and disappearance of the 2D islands due to instability was observed during the transition process from the 2D-island growth to step flow growth around 1500 °C (see Video S1 and Figure S2). Note that the step movement during the heating step, whose direction is denoted by the white arrows in Figure 1b and 1c, results from the growth of SiC layers on the seed SiC substrate, but not from the etching back process. In fact, the thickness of the grown SiC layer, carefully estimated after the CLSM observation by the CLSM with z-scan height mode, was ~360 nm for Si-Ni and ~120 nm for Al-added Si-Ni. Above ~1500 °C, clear step and terrace structures appeared all over the seed crystal surface, and the steps started to advance in a stepflow manner. With increasing temperature, two neighboring steps sometimes merged to form a new bunched step as a result of one step catching up with the other. It should be pointed out that the contrast of the merged step edge line became higher than those of the individual steps before they were merged. On the other hand, in the Al-added Si-Ni flux system, the frequency of the step merging was obviously less than in the Si-Ni flux system (details will be discussed later), resulting in step


6

Page 7 of 27

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


edge lines with a relatively small contrast in the CLSM image, even at 1700 °C (Figure 1c and Video S2).

Figure 1. In situ CLSM observation of the flux/SiC interface. (a) A schematic view of travelling solvent growth of SiC with a sandwich configuration, i.e., the source 4H-SiC substrate/flux/seed 4H-SiC substrate. The sample was attached on a carbon susceptor for efficiently absorbing the Nd:YAG laser to heat the sample. CLSM observation of the flux/seed substrate interface was carried out by introducing the probe laser light (658 nm) of the CLSM through the transparent


7


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

Page 8 of 27

seed SiC substrate, focusing on the growth interface. Above the melting point of the flux, the meniscus of the flux will be formed between the source and seed substrates. (b and c) A set of CLSM images recorded in a temperature range of 1400 to 1700 °C for (B) Si-Ni flux and (c) Aladded Si-Ni flux.

Step Height Analysis. Figure 2a shows an AFM topographic image of the seed SiC substrate surface taken at room temperature after removing the flux by etching with a HF:HNO3 (1:1 vol) solution. Steps with various heights in the range of about 6 to 100 nm were clearly observed. Figure 2b shows an in situ CLSM image (mirror-reversed) taken at 995 °C during the cooling step at exactly the same sample position as that in the AFM image. Excellent agreement between the CLSM and AFM images was found in terms of the shape of step edges, suggesting that there was almost no further step movement during the cooling step from 1614 °C to 995 °C under the present experimental conditions. In fact, no additional growth or etching of SiC occurred during the cooling step over a wide temperature range (see Figure S3). The relative intensity line-profile (black = 0 and white = 1) of the CLSM image was plotted along the solid line and compared with the corresponding height line-profile in the AFM image. The larger step height seemed to give a lower intensity in the CLSM image. More quantitatively, for various steps observed with the CLSM at 560, 995 and 1614 °C during the cooling step, Figure 2c shows the relationship between the actual step height, h, and the CLSM intensity at the corresponding step edge lines, Inorm, which is normalized by that at their neighboring terrace surfaces (Inorm=Ie/It, as shown in Figure 2b). A good linear relationship was found between them in the step height range below 50 nm, as expressed by the equation Inorm (h) = 1.05(1) –


8

Page 9 of 27

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


0.084(4) × h. The minimum step height observable in the CLSM image here was as low as ~6 nm (see Figure S4). The contrast in the CLSM image is basically determined by how much of the incident probe laser light is reflected at the confocal point, and the low intensity at a step edge line indicates that the incident probe laser was efficiently reflected at the step edge in a direction far from the position of the detector. The linearity can be roughly understood by assuming that each of the bunched steps observed in the present experiment consists of small steps, and the step density is proportional to the bunched step height (see Supporting Information for details).38

Figure 2. Empirical relationship between actual step heights in AFM and step edge intensities in CLSM. (a) An AFM topographic image of the seed SiC substrate surface taken at room temperature after removing the flux by etching with a HF:HNO3 (1:1 vol) solution, along with a height-profile on the solid line. (b) An in situ CLSM image (mirror-reversed) taken at 995 °C during the cooling step at exactly the same sample position as that in the AFM image (a), along with the corresponding intensity line profile. (c) Correlation between actual step heights and


9


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

Page 10 of 27

normalized step edge intensities in CLSM images, sampled from the CLSM images at three different temperatures during the cooling stage. Step Dynamics During Growth. The calibration curve empirically deduced between the AFM step height and the corresponding CLSM intensity allowed us to quantitatively analyze the nanoscale step height at the flux/SiC interface during the CLSM observation. Figure 3a, for example, shows a set of CLSM images taken at high temperatures above 1600 °C during the heating step for the Si-Ni flux system. From the bunched step as indicated by the white arrow, the birth of a new step with a height of 12±2 nm was observed at ~1630 °C, whereas a 9±2 nmheight step indicated by the yellow arrow was seen to advance faster and to merge with the anterior step with a height of 18±2 nm indicated by the red arrow at ~1640 °C as shown in Figure 3b. It should be emphasized that the sum rule is satisfied exactly in this step height analysis. For example, in the bunched step formation from the two steps, as already explained above, the height of the bunched step that was newly formed by merging the 9±2 nm and 18±2 nm steps together was 28±3 nm indicated by the orange arrow, confirming the validity of the calibration curve in Figure 2c for real-time step height analysis.


10

Page 11 of 27

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


Figure 3. In situ CLSM observation at high temperatures above 1600 °C. (a) A new step with a height of 12±2 nm, as indicated by the white arrow, was formed by advancing so as to separate from the posterior bunched step at ~1630 °C, whereas a 9±2 nm-height step, indicated by the yellow arrow, merged with the anterior step with a height of 18±2 nm (red arrow) to form a new bunched step with a height of 28±3 nm at ~1640 °C (orange arrow), where the sum rule in the step height analysis was satisfied. (b) Cross-sectional schematic illustration of these step motions on the yellow line in (a).

Next, we estimated the advance rates of various steps with different step heights from the CLSM images in the temperature range of 1600–1700 °C. In order to eliminate the effect of thermal drift during the observation at high temperature, we preprocessed the CLSM images to set a highest bunched step or a micro size point defect as a fixed point, relative to which the


11


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

Page 12 of 27

advance rates were calibrated (see Supporting Information). For example, the darkest step edge line, i.e., the highest step, observed in the CLSM image denoted by the white circle in Figure 3, was selected as a fixed point, whose step height changed from 56±4 nm at 1600 °C to >150 nm at 1680 °C. Figures 4a and 4b show the temperature dependence of the step-advance rate, v, for various steps with different step heights, for the Si-Ni and Al-added Si-Ni fluxes, respectively. The stepadvance rates in the growth with the Si-Ni flux were found to monotonically increase with the growth temperature, irrespective of their step heights. On the other hand, in the growth with the Al-added Si-Ni flux, such a monotonical temperature dependence seems to be not so clear, or rather more complicated, taking a minimum step-advance rate at a certain temperature for some particular steps. The different behavior of the temperature-dependent step-advance rate indicates that the factor that controls the step advance rate is different in the case of the SiC growth with the Al-added Si-Ni flux.


12

Page 13 of 27

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


Figure 4. Analysis of step-advance rates for different step heights in the temperature range between 1600 and 1700 °C. (a and b) Temperature dependence of the relative step-advance rates for different step heights during the growth with (a) Si-Ni flux and (b) Al-added Si-Ni flux. (c and d) Step-height dependence of the relative step-advance rates divided by the step distance v∞/y0 for different temperatures during the growth with (c) Si-Ni flux and (d) Al-added Si-Ni flux. For the Si-Ni flux, the plots for each temperature can be fitted well by the inverse proportional function of K/a, as shown by the solid lines, indicative of the diffusion regime of the


13


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

Page 14 of 27

SiC growth process. For the Al-added Si-Ni flux, the step-advance rates show little step-height dependence (dotted lines are the eye guide of the data), suggestive of the kinetic regime of the SiC growth process.

Next, we further discuss the step-height dependence of the step-advance rate. According to a theoretical model of crystal growth from solutions for the first order reaction, when the ratedetermining step of the growth is a mass transfer process in the solution on the crystal surface, but not a chemical reaction on the crystal surface, i.e., when the crystal growth proceeds in the diffusion regime but not the kinetic regime, the step-advance rate, v∞ , can be expressed as39 ‫ݒ‬ஶ =

గ஽஼బ ௩೎ ఙ

೤ ഏ೏ ௔ ୪୬ቂ బ ୱ୧୬୦ቀ ቁቃ ೌ

(1)

೤బ

where C0 is the equilibrium concentration of the solute at a given temperature, ‫ݒ‬௖ is the volume of a growth unit in the crystal, σ, is the supersaturation (=C∞/C0-1, where C∞ is the bulk concentration of the solute), D is the bulk diffusion coefficient, a is the step height, y 0 is the average step distance, and d is the thickness of the stagnant layer (~5×(xη/ϑρ)1/2), where x is the distance from the leading edge of the crystal surface, η is viscosity, ϑ is velocity, and ρ is density of the fluid flux. In the case of our experimental geometry, d would be as same as the distance between the source and seed SiC substrates, much larger than y0 (< 20 µm in the CLSM images up to 1700 °C) because the absence of stirring and only natural convection result in too small a value of ϑ, dominantly determining this thickness. It follows that in the present extreme πd >> y0, the step-advance rate, v∞, can be expressed as39 ‫ݒ‬ஶ ≈

஽஼బ ௩೎ ఙ௬బ ௔ௗ

(2)

Figures 4c and 4d are plots of v∞/y0 against the step height, a, at three different growth temperatures, for the Si-Ni and Al-added Si-Ni fluxes, respectively. For the Si-Ni flux, the values


14

Page 15 of 27

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


of v∞/y0 decreased with increasing step height, and the plots for each temperature can be fitted well by the inverse proportional function of K/a, as shown by the solid lines in Figure 4c. This result strongly implies that the surface reaction kinetics is the first order and the growth of SiC with the Si-Ni flux is in the diffusion regime. The fitting coefficients of K (=DC0v∞σ/d) for different growth temperatures were 46.6±7.0 (1620 °C), 43.3±3.9 (1655 °C), and 48.5±6.9 (1683 °C), respectively; they are almost constant within the experimental errors in the observed temperature range. In contrast, for the Al-added Si-Ni flux, there seems to be no clear step-height dependence of the step-advance rate (Figure 4d), though the step height of the data points was limited to a narrow range. Nevertheless, taking into account the different temperature dependence of the step-advance rate in the Al-added Si-Ni flux, as already pointed out, it is reasonable to expect that the crystal growth proceeds in the kinetic regime rather than the diffusion regime by the addition of the Al. If this is the case, as is theoretically predicted39, the step-advance rate will be constant and no longer dependent on the step height. Generally, the height information with a conventional CLSM is obtained by detecting or calculating the peak intensity around the focus position from a set of image data taken along the vertical direction. While lateral scanning of the probe laser beam can be done typically within a second, the vertical movement of the sample or optics takes longer (at least several seconds, sometimes up to several tens of seconds), and this will be a crucial factor in high-temperature observations due to the thermal drift. Even if such thermal drift were technically overcome, another problem would appear: namely, the vertical resolution of normal scanning height imaging in CLSM is governed by the axial point spread function, and thus the depth of field increases with the numerical aperture (NA). Since the NA tends to be low for a long-workingdistance lens, which is adopted to minimize radiation damage to the objective lens at high


15


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

Page 16 of 27

temperature, nanometer-scale height measurement is difficult. Therefore, normal height imaging using vertical scanning is not suitable for the real-time observation of nanoscale-step dynamics at high temperature. One of the advantages of the proposed technique is that the height information can be estimated without vertical scanning, enabling high-speed data acquisition (4 fps in this study). The minimum distance between steps for the height measurement is limited just by the lateral resolution of the microscope and not by any other factors, such as the direction or the distance of the interference fringes.23,24 The vertical resolution is basically limited by the dynamic range of the detector. The central significance of in situ CLSM observation of crystal growth interfaces in solution, the primary result of this paper, is illustrated by revealing the effect of the Al addition on the solution growth of SiC with quantitative analyses of step dynamics on the nanoscale over the course of the experiment. For the Si-Ni flux, the step-advance rates are found to decrease with increasing step height, resulting in frequent step-bunching (Figure 5a). In contrast, for the Aladded Si-Ni flux, the step-advance rates are less dependent of the step-heights, causing a marked suppression of such step-bunching (Figure 5b). In fact, as shown in Figure S11, as well as in Figure 1b and c, the measured step-heights for the Al-added Si-Ni flux were relatively smaller than those for the Si-Ni flux in the step-height distribution after the CLSM experiments; this is one of the possible mechanisms that explain the surface flatting effect of a small addition of Al in the bulk solution growth of SiC crystals with Si-Cr flux,28 though the flux composition and the growth temperature are different from those in the present case.


16

Page 17 of 27

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


Figure 5. Schematic illustration of the growth model of single crystal SiC in (a) Si-Ni flux and (b) Al-added Si-Ni flux.

It should be pointed out that this effect is brought about by adding only a few at%, e.g. 4 at%, of Al to the flux system in our case, which should not cause a significant change in the thermodynamics of the system. In fact, according to thermodynamic calculations, the carbon solubility of the Si-Ni flux is not greatly changed by the addition of 4 at% Al, and is at most ~15 % smaller at ~1650 °C (Figure S10). Furthermore, if any thermodynamic effect by decreasing the carbon solubility were dominant, the transition from the diffusion to the kinetic regime brought about by the Al addition would not be observed. This suggests that the Al addition results rather in kinetic effects, but not in thermodynamic ones, on the growth behavior of SiC. The kinetic effect brought about by the Al addition could also explain the temperature dependence of the step-advance rates for the Al-added Si-Ni flux; it did not monotonically


17


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

Page 18 of 27

increase with growth temperature, unlike the case for Si-Ni flux (Figure 4a), but sometimes took a minimum at around ~1650 °C (Figure 4b). In the diffusion regime, there are two temperature effects on the carbon solubility and supersaturation, both of which will be increased with the growth temperature, resulting in a monotonic increase in the step-advance rate. In contrast, in the kinetic regime, the growth reaction on the SiC surface, i.e., the step-advance rate, should be sensitive to the surface coverage of adsorbed carbon atoms or its related intermediates, which is in equilibrium with the carbon concentration in the solution. The carbon solubility monotonically increases, while the adsorbed carbon atoms or its related intermediates (e.g. Al-carbon) will be likely to desorb, at higher growth temperature. Thus, the temperature dependence of the stepadvance rates for the Al-added Si-Ni flux is rather complicated, resulting from competition between at least two opposing temperature effects. The change of growth regime found in our observation could be attributed to that the rate-limiting process of all the elementary steps in SiC growth was shifted to the possible adsorption process of Al-carbon intermediates on the growth interface. While an atomistic understanding of Al addition in the Si-Ni based flux still needs further investigation, this study suggests that not only the thermodynamics but also the kinetics in the solution growth are strongly affected by the small amount of added Al. Polytype control is one of the key issues in the crystal growth processes for single crystal SiC. The present in situ CLSM observation technique would also be useful for detecting and quantitatively analyzing the polytype change during the growth because of a six-fold symmetric morphology, as seen in the CLSM images, typically observed for hexagonal SiC polytypes, in contrast to a triangular domain for cubic ones.40 In our case, the SiC layers grown with the Si-Ni and Al-added Si-Ni fluxes on 4H-SiC seed crystal substrates have 4H polytypes, as confirmed by Raman investigation (Figure S12). Note that the addition of Al to the Si-Ni flux did not give any


18

Page 19 of 27

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


polytype change from the 4H seed SiC, which is consistent with the results reported in the literature.28,41

CONCLUSIONS In summary, we achieved real-time quantitative analyses of nanoscale step dynamics during the solution growth of SiC at high temperature with a conventional CLSM. By using this technique, a transition of the SiC step dynamics by the addition of a small amount of Al to Si-Ni was clearly found. The main feature of this in situ observation technique is that it requires no special optical instruments, but merely conventional CLSM instruments that are commercially available, and thus it shows promise as a useful tool that will open up a new path to important information for understanding high-temperature solution growth mechanisms of not only SiC but also other crystal materials.

ASSOCIATED CONTENT Supporting Information. The following files are available free of charge. Full experimental setup of the chamber, Additional CLSM images, AFM images, CLSM reflection model, cross-sectional TEM images, additional data of the relationship between the normalized CLSM intensity and step height, details of the step motion analysis, thermodynamic simulation of carbon solubility, step height distribution, and Raman spectra. (PDF) Video S1: In situ CLSM image of SiC/Si-Ni flux interface during the heating step from 600 to 1700 °C (×120 speed). (MPG)


19


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

Page 20 of 27

Video S2: In situ CLSM image of SiC/Al-added Si-Ni flux interface during the heating step from 1400 to 1700 °C (×120 speed). (MPG)

AUTHOR INFORMATION Corresponding Author [email protected] Author Contributions A.O. and S.M. contributed equally to this work. A.O., S.M., T.K., H.O. and Y.M. designed the research. A.O. and S.M. conducted the experiments and analysis. N.K. and T.M. performed the thermodynamic calculations. A.O., S.M. and Y.M. wrote the manuscript with the assistance of N.K., T.M., T.K. and H.O. All authors contributed to the discussions of the results. Notes The authors declare no competing financial interest. ACKNOWLEDGMENT This work was supported by the Novel Semiconductor Power Electronics Project Realizing Low Carbon Emission Society under the New Energy and Industrial Technology Development Organization (NEDO), and by the Advanced Low Carbon Technology Research and Development Program (ALCA) of the Japan Science and Technology Agency (JST). We are grateful to Mr. T. Nakatsukasa, Mr. J. Yamashita and Mr. W. Masuoka (Keyence) for helpful discussions.


20

Page 21 of 27

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


REFERENCES (1)

Tsukamoto, K. In Situ Observation of Mono-Molecular Growth Steps on Crystals Growing in Aqueous Solution. I. J. Cryst. Growth 1983, 61, 199–209.

(2)

Dold, P.; Ono, E.; Tsukamoto, K.; Sazaki, G. Step Velocity in Tetragonal Lysozyme Growth as a Function of Impurity Concentration and Mass Transport Conditions. J. Cryst. Growth 2006, 293, 102–109.

(3)

Sazaki, G.; Matsui, T.; Tsukamoto, K.; Usami, N.; Ujihara, T.; Fujiwara, K.; Nakajima, K. In Situ Observation of Elementary Growth Steps on the Surface of Protein Crystals by Laser Confocal Microscopy. J. Cryst. Growth 2004, 262, 536–542.

(4)

Sazaki, G.; Zepeda, S.; Nakatsubo, S.; Yokoyama, E.; Furukawa, Y. Elementary Steps at the Surface of Ice Crystals Visualized by Advanced Optical Microscopy. Proc. Natl. Acad. Sci. 2010, 107, 19702–19707.

(5)

Van Driessche, A. E. S.; García-Ruiz, J. M.; Delgado-López, J. M.; Sazaki, G. In Situ Observation of Step Dynamics on Gypsum Crystals. Cryst. Growth Des. 2010, 10, 3909– 3916.

(6)

Van Driessche, A. E. S.; Sleutel, M. In Situ Measurement of Crystal Surface Dynamics in Pure and Contaminated Solutions by Confocal Microscopy and Atomic Force Microscopy. Cryst. Res. Technol. 2013, 48, 919–941.

(7)

Azhagurajan, M.; Kajita, T.; Itoh, T.; Kim, Y.-G.; Itaya, K. In Situ Visualization of Lithium Ion Intercalation into MoS 2 Single Crystals Using Differential Optical Microscopy with Atomic Layer Resolution. J. Am. Chem. Soc. 2016, 138, 3355–3361.


21


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

(8)

Page 22 of 27

Onuma, K.; Tsukamoto, K.; Nakadate, S. Application of Real Time Phase Shift Interferometer to the Measurement of Concentration Field. J. Cryst. Growth 1993, 129, 706–718.

(9)

Booth, N. A.; Chernov, A. A.; Vekilov, P. G. Characteristic Lengthscales of Step Bunching in KDP Crystal Growth: In Situ Differential Phase-Shifting Interferometry Study. J. Cryst. Growth 2002, 237–239, 1818–1824.

(10)

Tsukamoto, K.; Dold, P. Interferometric Techniques for Investigating Growth and Dissolution of Crystals in Solutions. AIP Conf. Proc. 2007, 916, 329–341.

(11)

Hillner, P. E.; Gratz, A. J.; Manne, S.; Hansma, P. K. Atomic-Scale Imaging of Calcite Growth and Dissolution in Real Time. Geology 1992, 20, 359.

(12)

Maruyama, M.; Tsukamoto, K.; Sazaki, G.; Nishimura, Y.; Vekilov, P. G. Chiral and Achiral Mechanisms of Regulation of Calcite Crystallization. Cryst. Growth Des. 2009, 9, 127–135.

(13)

Lupulescu, A. I.; Rimer, J. D. In Situ Imaging of Silicalite-1 Surface Growth Reveals the Mechanism of Crystallization. Science (80-. ). 2014, 344, 729–732.

(14)

Higgins, S. R.; Bosbach, D.; Eggleston, C. M.; Knauss, K. G. Kink Dynamics and Step Growth on Barium Sulfate (001): A Hydrothermal Scanning Probe Microscopy Study. J. Phys. Chem. B 2000, 104, 6978–6982.

(15)

Saldi, G. D.; Jordan, G.; Schott, J.; Oelkers, E. H. Magnesite Growth Rates as a Function of Temperature and Saturation State. Geochim. Cosmochim. Acta 2009, 73, 5646–5657.

(16)

Abe, T.; Sukamoto, K.; Sunagawa, I. Nucleation, Growth and Stability of CaAl2Si2O8 Polymorphs. Phys. Chem. Miner. 1991, 17, 473–484.


22

Page 23 of 27

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


(17)

Chikama, H.; Shibata, H.; Emi, T.; Suzuki, M. “In-situ” Real Time Observation of Planar to Cellular and Cellular to Dendritic Transition of Crystals Growing in Fe-C Alloy Melts. Mater. Trans. JIM 1996, 37, 620–626.

(18)

Fujiwara, K.; Nakajima, K.; Ujihara, T.; Usami, N.; Sazaki, G.; Hasegawa, H.; Mizoguchi, S.; Nakajima, K. In Situ Observations of Crystal Growth Behavior of Silicon Melt. J. Cryst. Growth 2002, 243, 275–282.

(19)

Aoyama, T.; Kuribayashi, K. Rapid Solidification Processes of Semiconductors from Highly Undercooled Melts. Mater. Sci. Eng. A 2001, 304–306, 231–234.

(20)

Kawanishi, S.; Yoshikawa, T.; Morita, K. Real-Time Observation of High Temperature Interface between SiC Substrate and Solution during Dissolution of SiC. Mater. Sci. Forum 2013, 740–742, 35–38.

(21)

Kangawa, Y.; Kusaba, A.; Sumiyoshi, H.; Miyake, H.; Boćkowski, M.; Kakimoto, K. Real-Time Observation System Development for High-Temperature Liquid/solid Interfaces and Its Application to Solid-Source Solution Growth of AlN. Appl. Phys. Express 2015, 8, 65601.

(22)

Fujiwara, K.; Obinata, Y.; Ujihara, T.; Usami, N.; Sazaki, G.; Nakajima, K. In-Situ Observations of Melt Growth Behavior of Polycrystalline Silicon. J. Cryst. Growth 2004, 262, 124–129.

(23)

Kawanishi, S.; Yoshikawa, T.; Morita, K.; Kusunoki, K.; Kamei, K.; Suzuki, H.; Sakamoto, H. Real-Time Observation of the Interface between SiC and a Liquid Alloy and Its Application to the Dissolution Behavior of SiC at 1573 K. J. Appl. Phys. 2013, 114, 214313.


23


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

(24)

Page 24 of 27

Kawanishi, S.; Kamiko, M.; Yoshikawa, T.; Mitsuda, Y.; Morita, K. Analysis of the Spiral Step Structure and the Initial Solution Growth Behavior of SiC by Real-Time Observation of the Growth Interface. Cryst. Growth Des. 2016, 16, 4822–4830.

(25)

Silicon Carbide, Volume 2: Power Devices and Sensors; Peter Friedrichs, Tsunenobu Kimoto, Lothar Ley, G. P., Ed.; Wiley-VCH, 2009.

(26)

Yakimova, R.; Tuominen, M.; Bakin, A. S.; Fornell, J. O.; Vehanen, A.; Janzen, E. Silicon Carbide Liquid Phase Epitaxy in the Si-Sc-C System. Silicon Carbide Relat. Mater. 1995 1996, 142, 101–104.

(27)

Hofmann, D.; Müller, M. Prospects of the Use of Liquid Phase Techniques for the Growth of Bulk Silicon Carbide Crystals. Mater. Sci. Eng. B 1999, 62, 29–39.

(28)

Mitani, T.; Komatsu, N.; Takahashi, T.; Kato, T.; Fujii, K.; Ujihara, T.; Matsumoto, Y.; Kurashige, K.; Okumura, H. Growth Rate and Surface Morphology of 4H–SiC Crystals Grown from Si–Cr–C and Si–Cr–Al–C Solutions under Various Temperature Gradient Conditions. J. Cryst. Growth 2014, 401, 681–685.

(29)

Maruyama, S.; Onuma, A.; Kurashige, K.; Kato, T.; Okumura, H.; Matsumoto, Y. HighThroughput Screening of Si-Ni Flux for SiC Solution Growth Using a High-Temperature Laser Microscope Observation and Secondary Ion Mass Spectroscopy Depth Profiling. ACS Comb. Sci. 2013, 15, 287–290.

(30)

Yonezawa, Y.; Ryo, M.; Takigawa, A.; Matsumoto, Y. Screening of Metal Flux for SiC Solution Growth by a Thin-Film Combinatorial Method. Sci. Technol. Adv. Mater. 2011, 12, 54209.

(31)

Onuma, A.; Maruyama, S.; Mitani, T.; Kato, T.; Okumura, H.; Matsumoto, Y. Uniform Growth of SiC Single Crystal Thin Films via a metal–Si Alloy Flux by Vapour–liquid–


24

Page 25 of 27

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


solid Pulsed Laser Deposition: The Possible Existence of a Precursor Liquid Flux Film. CrystEngComm 2016, 18, 143–148. (32)

ONUMA, A.; MARUYAMA, S.; MITANI, T.; KATO, T.; OKUMURA, H.; MATSUMOTO, Y. Effect of Al Addition to Si-Ni Flux on Pulsed Laser Deposition of SiC Thin Films. J. Ceram. Soc. Japan 2016, 124, 506–509.

(33)

Balmer, R. S.; Martin, T. Substrate Temperature Reference Using SiC Absorption Edge Measured by in Situ Spectral Reflectometry. J. Cryst. Growth 2003, 248, 216–221.

(34)

CALPHAD, Computer Coupling of Phase Diagrams and Thermochemistry; Kaufman, L., Ed.; Pergamon, Oxford.

(35)

Syväjärvi, M.; Yakimova, R.; Radamson, H. .; Son, N. .; Wahab, Q.; Ivanov, I. .; Janzén, E. Liquid Phase Epitaxial Growth of SiC. J. Cryst. Growth 1999, 197, 147–154.

(36)

Nishitani, S. R.; Kaneko, T. Metastable Solvent Epitaxy of SiC. J. Cryst. Growth 2008, 310, 1815–1818.

(37)

Binary Alloy Phase Diagrams; Massalski, T. B.; Okamoto, H.; Subramanian, P.; Kacprzak, L., Eds.; ASM International: Materials Park, OH, 1990.

(38)

Nakagawa, H.; Tanaka, S.; Suemune, I. Self-Ordering of Nanofacets on Vicinal SiC Surfaces. Phys. Rev. Lett. 2003, 91, 226107.

(39)

Markov, I. V. Crystal Growth. In Crystal Growth for Beginners; World Scientific, 2003; pp. 181–351.

(40)

Ujihara, T.; Maekawa, R.; Tanaka, R.; Sasaki, K.; Kuroda, K.; Takeda, Y. Stability Growth Condition for 3C-SiC Crystals by Solution Technique. Mater. Sci. Forum 2009, 600–603, 63–66.


25


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

(41)

Page 26 of 27

Mitomo, M.; Inomata, Y.; Kumanomido, M. The Effect of Doped Aluminium on Thermal Stability of 4H- and 6H-SiC. J. Ceram. Assoc. Japan 1970, 78, 224–228.


26

Page 27 of 27

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


For Table of Contents Use Only

Quantitative Analysis of Nanoscale Step Dynamics in High-temperature Solution-grown Single Crystal 4H-SiC via in situ Confocal Laser Scanning Microscope

Aomi Onuma, Shingo Maruyama, Naoyoshi Komatsu, Takeshi Mitani, Tomohisa Kato, Hajime Okumura and Yuji Matsumoto

Real-time observation of high-temperature solution growth of single crystal 4H-SiC was demonstrated by using a conventional confocal laser scanning microscope. From the quantitative analysis of nanoscale step dynamics, it was suggested that the addition of small amount of Al in Si-Ni flux changes the growth mode from diffusion regime to kinetic regime.


27

Quantitative Analysis of Nanoscale Step Dynamics in High

Recommend Documents