Effects of shutter transients in molecular beam epitaxy
© Gozu et al.; licensee Springer. 2012
Received: 19 July 2012
Accepted: 19 October 2012
Published: 12 November 2012
We have studied the effects of shutter transients (STs) in molecular beam epitaxy (MBE). Two series of samples were grown by MBE and evaluated by X-ray diffraction (XRD) and X-ray reflectivity (XRR) measurements. The effects of STs were evaluated by growth rate (GR) analysis using a combination of growth time (GT) and thickness evaluated by XRD and XRR measurements. We revealed two opposite effects of STs: (1) overshoot of GR and (2) increase in GR with GT and subsequent saturation. Each effect was consistent with the previous studies; however, the previous studies showed no relationships between these two effects. By considering closing time of the shutter, the two opposite effects were well understood.
KeywordsMolecular beam epitaxy Shutter transients X-ray diffraction X-ray reflectivity
Molecular beam epitaxy (MBE) is an ideal method to grow nano-structures. MBE allows for the controlled growth of films with sharp doping profiles and different chemical compositions changing over a spatial depth of several angstroms. Multi-layer structures with alternating doping (n, p, or intrinsic) or alternating band gaps can be grown. In addition, self-organized effects can be applied for growing the nano-structures with controlled dimensions that are not only in the growth direction (thickness) but also laterally in the plane of the growth surface. These nano-structures are well known as quantum wells (QWs), quantum wires, and quantum dots (QDs). When their sizes are decreased, the quantum confinement effect becomes the dominant contribution to their electric and optical properties. Because quantum confinement energy is inversely proportional to size (1/L), accuracy of size is very important to fabricate the nano-structures. However, MBE cannot avoid the effects of shutter transients (STs) by which a prominent inaccuracies are caused by strong flux transients after opening the shutter due to temperature gradients in the cells. In an MBE system, beam sources are provided by thermal evaporation of solid source materials in the cells, and the beam sources are individually controlled by mechanical shutters to obtain the intended structures. The mechanical shutters also reflect heat back into the cells. Therefore, a temperature gradient occurs when the mechanical shutters are moved. The studies conducted for STs were contrasting in two ways: (1) growth rate (GR) decreased and was saturated after a characteristic time constant (CTC) [1–4], and (2) the GR was below the desired GR and then approached the target GR within the CTC [5, 6]. Therefore, the effects of STs seemed to be contradictory or to have a machine dependence. Therefore, we have studied the effects of STs for our MBE system.
Evaluation method of shutter transients
In order to evaluate STs for MBE, growth time (GT) dependence of GR must be evaluated. Reflection high-energy electron diffraction (RHEED) in MBE is a powerful tool to evaluate GR. During growth, the intensity of oscillation at a specular spot can be observed, and 1 period of the oscillation corresponds to the growth of 1 mono-layer (ML). Evolution of the oscillation can be used to evaluate the GT dependence of the GR. Previous studies have used this method to evaluate STs [2, 4, 5]. However, GR evaluation using RHEED has some drawbacks. The oscillation is damped. Although substrate rotation is required to obtain good thickness uniformity on the samples, substrate rotation is not possible during RHEED oscillation measurements. This means that only one position in the sample is measured, and thus, some error is involved in this method. In addition, observation of the oscillation is strongly dependent on the growth condition. To overcome these drawbacks, in the present study, some samples were grown, and then the thickness of the samples was evaluated using X-ray diffraction (XRD) and reflectivity (XRR) measurements. XRD measurements can be used to evaluate the layer thickness when the sample has a periodic structure. Because of satellite peak patterns, separation that corresponds to the period can be distinguished. On the other hand, XRR measurements do not require a periodic structure to evaluate thickness. To evaluate the thickness of individual layers by XRR, model-based fitting is required. Because the GT of the samples is known, the GR can be calculated. In this study, InGaAs layers were targeted to evaluate the GT dependence of the GR.
Results and discussion
Coupled double quantum well samples
XRR simulation results for the two CDQW samples
InP buffer 10 nm
InGaAs 30 s
AlAs 3.7 s
InGaAs 30 s
InP buffer + sub.
Consideration of shutter transients
We observed two opposite types of STs in the previous two sections. First was the increase in GR with GT, and second was the overshoot of GR. These behaviors were inconsistent with each other. The difference between the two experiments was the sample structure and the evaluation method. The first was the stacked SLs with XRD measurements, and the second was the single CDQW with XRR measurements. Because XRD measurement of the stacked SLs revealed an average GR for the whole structure, while the XRR measurements of the single CDQW revealed an initial state of growth, the two opposite behaviors could be attributed to long or short shutter closing times. Therefore, the effects of STs in MBE involved two opposite behaviors depending on the shutter closing time. However, Celii et al.  reported that STs for their MBE system only showed the overshoot of GR whose amplitude and decay time were related with the shutter closing time. On the other hand, other studies [1–3, 5, 6] reported one type of STs without the consideration of the shutter closing time. Due to the fact that the effects of STs were strongly dependent on the MBE system, our observations of the two types of STs were characteristics of our MBE system because an amount of the heat back into the cell must depend on the whole structure of the cells. However, the origin of STs arose from the temperature gradient in the cell when the mechanical shutter was moved. This origin is not avoided; therefore, the effects of STs can be said that there is universal phenomena in the MBE system. It should be noted here that our evaluation methods for the effects of STs were different with respect to the previous studies [1–6]. The main difference was that XRR measurement was used to evaluate STs for the initial state of growth. The previous studies used to evaluate STs for their MBE systems as following: beam flux measurements , RHEED measurements [2, 5, 6], optical absorption measurements of QW samples, reflection mass spectroscopy , and XRD measurements of SL samples . These measurements as well as XRR measurement are powerful methods to reveal the effects of STs, but each measurement has some drawbacks. The drawbacks for RHEED measurements have already been pointed out in the previous section. However, XRD and XRR measurements which were used in this study have a merit of high accuracy of thickness evaluation. In addition, XRR measurement does not require the periodic structure like SLs which are used for XRD measurement. This feature is very important to reveal the initial state of growth as mentioned before. Therefore, a combination of the evaluation methods used in this study is well balanced and suitable to reveal an entire feature of STs. On the other hand, our experimental results revealed that 5% to 10% of the error in the thickness easily occurred for the designed structure at around 3 nm. Therefore, the effects of STs should be carefully taken into account if the size of the intended structure is as small as a few nanometers.
We have studied the effects of STs in MBE, and two opposite effects were found. Each effect was consistent with previous studies; however, the previous studies showed no relationships between them. Our experimental results could be categorized into two situations: long and short closing times of the shutters. By categorizing these situations, the two opposite effects were understood. Finally, we pointed out that the effects of STs should be carefully taken into account if the size of the intended structure is as small as a few nanometers.
- Miller JN: Flux noise in effusion cells: a key to understanding oval defects. J Vac Sci Technol B 1992, 10: 803. 10.1116/1.586120View ArticleGoogle Scholar
- Heyn Ch, Cunis S: Shutter transients during solid-source epitaxy. J Vac Sci Technol B 2014, 25: 2005.Google Scholar
- Cristea P, Fedoryshyn Y, Holzman JF, Robin F, Jäckel H, Müller E, Faist J: Tuning the intersubband absorption in strained AlAsSb/InGaAs quantum wells towards the telecommunications wavelength range. J Appl Phys 2006, 100: 116104. 10.1063/1.2400794View ArticleGoogle Scholar
- Celii FG, Kao YC, Beam EA, Duncan WM, Moise TS: Molecular-beam epitaxy flux transient monitoring and correction using in situ reflection mass spectrometry. J Vac Sci Technol B 1993, 11: 1018. 10.1116/1.586860View ArticleGoogle Scholar
- Roch T, Andrews AM, Fasching G, Benz A, Schrenk W, Unterrainer K, Strasser G: High-quality MBE growth of AlxGa1-xAs-based THz quantum cascade lasers. Cent Eur J Phys 2007, 5: 244. 10.2478/s11534-007-0004-yGoogle Scholar
- Andrews AM, Roch T, Benz A, Fasching G, Schrenk W, Unterrainer K, Strasser G: Optimization of MBE growth parameters for GaAs-based THz quantum cascade lasers. AIP Conf Proc 2007, 893: 51.View ArticleGoogle Scholar
- Colombi P, Agnihotri DK, Asadchikov VE, Bontempi E, Bowen DK, Chang CH, Depero LE, Farnworth M, Fujimoto T, Gibaud A, Jergel M, Krumrey M, Lafford TA, Lamperti A, Ma T, Matyi RJ, Meduna M, Milita S, Sakurai K, Shabel’nikov L, Ulyanenkov A, Van der Lee A, Wiemer C: Reproducibility in X-ray reflectometry: results from the first world-wide round-robin experiment. Appl Cryatallography 2008, 41: 143. 10.1107/S0021889807051904View ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.