The h-BN films were grown using a home-built hybrid CVD setup. Figure 3.1 (a) shows the schematic of the growth setup. The setup comprises of a 52 mm inner diameter (I.D.) horizontal split tube furnace (MTI Corporation). The solid precursor ammonia-borane (NH3-BH3) powder, (97%
purity, Sigma-Aldrich) is contained in a home-made quartz container, attached to the main growth chamber (22 mm I.D. quartz tube) via a leak valve and heated separately from the quartz tube via use of a resistive heating belt. Cu foils (25 µm, 99.999% pure, Alfa Aesar, item no. 10950) are used as the catalytic growth substrates. Copper foil was soaked and sonicated in acetone and isopropyl alcohol (IPA) for 30 min consecutively to remove organic impurities. Then, it was washed with deionized water and dried with nitrogen gas. The pressure in the growth chamber can be independently controlled via an angle valve at the vacuum pump while the pressure in the precursor bubbler can be controlled via the leak valve and carrier gas flow rates. The monolayer and multilayer h-BN was synthesized using a pressure controllable CVD system. The copper foil was inserted into the center of a 22 mm I.D. quartz tube, heated by a horizontal split-tube furnace. The Ammonia borane (NH3-BH3) (97% purity, from Sigma-Aldrich), stable in an atmospheric environment, was used as the precursor. It was loaded in a homemade quartz container which is isolated from the main
CVD system with a leak valve to control the flow rate. The quartz tube inlet and outlet were blocked by filters to prevent the BN nanoparticles from diffusing into the gas line. First, the quartz tube was pumped down to 5×10-3 torr, and then ultrahigh purity grade hydrogen gas was introduced during the temperature ramp-up of the furnace (pressure ~ 100 mtorr, flow rate ~ 50 sccm). The copper foil was annealed at 950°C in hydrogen for 60 min to obtain a smooth surface. After annealing, the ultrahigh purity argon gas (300 sccm) was introduced into the system and waited for 30 min to stable the tube environment. The precursor was heated to 130°C and decomposed to hydrogen gas, monomeric aminoborane, and borazine gas. After the precursor temperature reached 130°C, the manual valve between the quartz tube outlet and the pump was slowly closed and stopped until the pressure reached 20 torr. When the desired pressure was achieved, the leak valve to the precursor was opened. The typical growth time is 3 min for monolayer h-BN layer and 20 min for the 20 nm thickness h-BN layer. To atomically control the thickness of the h-BN layer, it is very important to have a leak valve to control the flow rate of the precursor. Also, we can change the growth rate of the h-BN by changing the pressure of the growth environment. After growth, the tube furnace was cooled down with the cooling rate ~16°C/min.
To grow ultrathin (< 1.5 nm) h-BN films, we use a growth pressure of ~ 2 Torr. Figure 3.1(b) and (c) show an optical photograph and atomic force microscopy (AFM) analysis of a wafer-scale h-BN monolayer which has transferred onto a 285 nm SiO2/Si substrate. Raman spectra acquired over six different spots on the sample (denoted by x), as seen in the adjacent plot suggests that the sample thickness and quality is uniform over the entire cm2 scale. Likewise, the growth mode can be switched from a slow growth rate of ~ 0.3 nm/min in catalytically controlled CVD at low pressures to a high growth rate of 1 nm/min at near atmospheric pressures (~500 Torr). Figure 1c illustrates the large area uniformity of ~ 15 monolayer h-BN as inferred from the optical micrograph and corresponding Raman spectra. To validate the precise control of layer thickness and growth uniformity, a series of growth experiments were performed for varying times at both low and ambient pressures to produce h-BN films of varying layer number/thicknesses. These h-BN films were then characterized with several spectroscopic techniques to characterize their structure and chemical composition. Atomic scale structure and areal homogeneity of h-BN films were revealed by scanning tunneling microscopy (STM) for CVD-grown monolayer h-BN. Despite numerous reports on CVD synthesis of h-BN, little is known about the atomic scale electronic structure for layers grown on
microscopy of sub-monolayer growth directly on the copper substrate. A sub-monolayer growth would provide strong contrast in the SEM between the insulating h-BN and conductive copper foil.
Figure 3.2 below shows some SEM micrographs of sub monolayer growths and analysis of domain sizes. Figure 3.2 (a) shows a large scale SEM image with different h-BN domain sizes. The sharp geometric and triangular edge features suggest the highly crystalline nature of the growth as verified by STM, TEM, and Raman in the manuscript. Figures 3.2 (b) and (c) show the higher magnification images of triangular h-BN domains with edge length of ~ 500 nm and ~ 1 μm, respectively. The statistical distribution of domain size is plotted in Figure 3.2 (d) which shows that the average domain size of h-BN is ~ 750 nm.
Figure 3.1. (a) Schematic diagram of hybrid atmospheric pressure and low pressure CVD system used for h-BN grow. (b) Photograph of a large and uniform monolayer h-BN film on a 285 nm thick SiO2/Si substrate, and the corresponding AFM image and Raman spectra. (c) Photograph of a large and uniform multilayer h-BN film on a 285 nm thick SiO2/Si substrate, and the corresponding AFM image and Raman spectra.
Figure 3.2. (a)-(c) SEM images showing triangular ad-layer with different domain size. (d) Histograms for the h-BN domain size.
As shown in Figure 3.3 below, we can systematically vary the thickness of grown h-BN films from monolayer to 30 layers. We show the optical photograph and atomic force microscopy (AFM) analysis of these h-BN samples that have been transferred onto 285 nm SiO2/Si substrates.
Figure 3.3. (a) - (d) Photograph of 2-layer h-BN to 5-layer thinner h-BN on SiO2/Si substrate within the white dashed line and corresponding AFM images of bi-layer h-BN to 5-layer h-BN on SiO2/Si substrate and the height profile. (e) - (h) Photograph of thicker multilayer h-BN (10 layers ~ 30 layers) on SiO2/Si and the corresponding substrate AFM images of multilayer h-BN (10 layers ~ 30 layers) on SiO2/Si substrate and the height profile. (i) A representative atomic force microscopy (AFM) image below shows a large scale 40 40 topography image of a 5-layer h-BN sample to show the thickness uniformity of the film. This h-BN film was grown on Cu and transferred onto 285 nm SiO2/Si substrate. The root mean square roughness extracted from this AFM micrograph is 0.69 nm.
borazine, hydrogen, and argon gases on the surface of a Cu catalyst at a synthesis temperature (typically 950 °C). The main thermos-kinetics of the CVD process and reaction of borazine molecules include several steps:
The borazine species first (1) get transported to the reaction region, followed by (2) adsorption on Cu surfaces, then (3) dehydrogenation of B3H6N3 to form active boron and nitrogen species.
Therefore, (4) surface diffusion of B and N species occurs on the Cu surface leading to (5) growth of h-BN domains, and ultimately, (6) inactive species (such as hydrogen) diffuse away from the surface through the boundary layer. The above process can be classified into two regimes: the mass transport region (hg ≪ Ks), primarily involving diffusion through the boundary layer, and the surface reaction region (catalysis; hg ≫ Ks). Here hg denotes the mass transfer coefficient, and Ks represents the surface growth rate. In these growth regimes, the total pressure PTOT determines whether boron nitride growth proceeded by catalysis or otherwise. For h-BN growths on Cu in LPCVD mode, we note that hg increases as a result of lowering the total pressure, which promotes surface catalysis (hg ≫ Ks), tending to grow from monolayer to thin multi-layer h-BN. In APCVD growth conditions, we note an increase in the surface growth rate Ks and decrease in the mass transfer coefficient hg, which makes the mass transport flux dominant over surface reactions (hg ≪ Ks). This means the h- BN growth will continue to happen until elimination of all the high temperature pyrolyzed species of ammonia-borane, which ultimately leads to growth of thick multilayer h-BN without the need of surface catalysis of Cu. Therefore, we can systematically vary the thickness of grown h-BN films by using a leak valve to control the flow rate of the borazine (or partial pressure of the borazine). In Figure 3.4 below, we show the quantitative dependence of h-BN film thickness on the partial pressure of the precursor (borazine). The growth temperature was fixed at 950 ºC, while the precursor temperature was fixed at 130 ºC.
Figure 3.4. (a) Processes involved during h-BN synthesis on Cu in a CVD process. (b) A plot of film thickness with respect to the Borazine partial pressure.