Open Access
Add to BibSonomy
Abstract
Atomic layer deposition (ALD), an emerging method of thin film fabrication, has recently witnessed a surge of applications in the optoelectronics field. However, reliable processes capable of controlling film composition have yet to be developed. In this work, the effect of precursor partial pressure and steric hindrance on the surface activity was presented and analyzed in detail, which led to the development of a component tailoring process for ALD composition control in intralayer for the first time. Further, a homogeneous organic/inorganic hybrid film was successfully grown. The component unit of the hybrid film under the joint action of EG and O plasma could achieve arbitrary ratios by controlling the EG/O plasma surface reaction ratio via varied partial pressures. Film growth parameters (growth rate per cycle and mass gain per cycle) and physical properties (density, refractive index, residual stress, transmission, and surface morphology) could be modulated as desired. Moreover, the hybrid film with low residual stress was effectively used in the encapsulation of flexible organic light-emitting diodes (OLEDs). Such a component tailoring process is an important step forward in ALD technology, and allowing for in-situ control of thin film components at the atomic level in intralayer.
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
The atomic layer deposition (ALD) technique is based on the cyclic use of self-limited chemical reactions for precise control of layer thickness and composition at the atomic level, as well as conformal growth mode, [1] and has tremendous applications in ultrathin and hyperfine film preparation. Typical thin film materials such as Al2O3, SiO2, and ZnO have been widely used in a variety of electronic fields [2–5]. In recent years, thin film deposition and component control have been extensively used in micro/nanofabrication technologies, such as mechanical structure, electrical isolation, and connection, etc [6–9]. However, the reliable component control process for ALD is still in its infancy.
Super-cycle is a commonly used component control method for ALD, performed by two individual surface reactions as the dopant layer and host layer, respectively [10–14]. The process cycle ratio of the two individual surface reactions is approximated as the component ratio. As shown in Supplement 1, Fig. S1, the steps of two normal ALD processes are combined, where m cycles of the first process are followed by n cycles of the second process in a super-cycle process. The variables m and n can be chosen to obtain the desired component and structure of the film.
There have been several reports about the super-cycle process. For instance, Sun et al. prepared Al-doped ZnO (AZO) film by super-cycle of ZnO/Al2O3 laminates using diethylzinc (DEZ), trimethylaluminum (TMA) and water at a deposition temperature of 150 °C. To broaden the component control range, the total ALD cycles for all the AZO samples were set to 1,090, with a thickness of about 200 nm as a conversion, which is unusual for ALD technology known for its ultra-thin film growth capability [15]. A similar super-cycle process was performed by Kim et al. to prepare AZO films with various doping concentrations. To achieve the control of the component ratio, the cycle number of the host material had to be raised from 9 to 49, and this number would rise further as the demand for a more refined control increased [16]. Super-cycles of indium oxide ALD and indicone molecular layer deposition (MLD) are used to grow organic-inorganic hybrid thin films by Park et al. The film performance gradually increases as the ALD/MLD cycle ratio increases from 39 to 99 but at the expense of an increase in the film's required thickness due to the structural characteristics of the super-cycle process. The further increase of the cycle ratio to >99 is not demonstrated in this work, most probably due to the thickness limitation of the application scenario [17].
So far, the hybridization growth of ALD-film can only be achieved by super-cycle due to the lack of an effective component control process. However, as the film grown by super-cycle is essentially a stacked structure with interlayer hybridization, [18] the film properties are theoretically characterized by periodic changes in the direction perpendicular to the substrate plane, and the hybridization may not affect the surface area of the hybrid film because the surface was terminated by host layer rather than dopant layer [19]. In addition, fine component control is always achieved at the cost of an exponential increase in required thickness, so the super-cycle process cannot achieve ultra-thin film hybridization. For instance, realizing a 100:1 component ratio using an ALD process with 50 cycles is impossible with a super-cycle. The aforementioned technical barriers limit the application of ALD in the field of ultrathin and hyperfine film preparation, and super-cycle can only be used as an emergency tool rather than a long-term solution for component control. It would be more appropriate to refer to the films grown by super-cycle as laminated films rather than hybrid films. Therefore, novel component control processes are required for ALD to achieve intralayer component control.
Herein, a component tailoring process for ALD composition control in intralayer was proposed for the first time. A homogeneous organic-inorganic hybrid film was grown as a typical case by this novel growth model. The precise regulation of process temperature and precursor partial pressure effectively changed the substrate surface activity, allowing for various component tailoring. Furthermore, the surface active state induced by the steric hindrance effect during film growth was also utilized for introducing desired components with small steric hindrance. An unsaturated surface caused by the regulation of EG precursor partial pressure or EG induced-steric hindrance effect was identified and served as an intermediate state for film growth. Therefore, the O Plasma followed by EG injection oxidized the unreacted active sites -CH3 exposed to the substrate surface caused by this unsaturated surface state. Next, the generated -OH reacted with the subsequently arriving TMA, resulting in the introduction of Al-O component, confirming the achievement of inorganic component tailoring. The component unit of the hybrid film under the joint action of EG and O plasma could achieve arbitrary ratios in intralayer by controlling the EG/O plasma surface reaction ratio via varied partial pressures. Film growth parameters (growth rate per cycle and mass gain per cycle) and physical properties (density, refractive index, residual stress, transmission, and surface morphology) could be modulated as desired. Moreover, as one of the important applications, the hybrid film with low residual stress preserved intact barrier performance after 10,000 bends with a 3 mm bending radius, and was effectively used to extend the lifetime of flexible organic light-emitting diodes (OLEDs) as an ideal encapsulation material. The realization of autonomously controllable growth of films by the component tailoring process made the application of ALD more multifaceted, which demonstrated the potential technological value.
2. Result and discussion
All sample preparation processes were carried out in a clean room. ALD-Al2O3 and MLD-alucone were grown by alternating pulses of TMA and O plasma, TMA and EG, respectively. Various hybrid films were grown using sequential pulses of TMA, EG, and O plasma with different process recipes. The Experimental Section contains a detailed description of the typical process parameters.
2.1 Surface activity modulated by process temperature and precursor partial pressure
ALD or MLD processes are performed by alternating half-reactions of two precursors on the substrate surface [20,21]. During the half-reactions, the microstructure of the film was influenced by chemisorption and physisorption, which are two major actions of the precursors on the substrate. The precursors chemisorbed on the substrate surface spatially block the surrounding active sites, causing structural steric hindrance [22]. Similarly, transient blocking by physical adsorption of precursors causes transient steric hindrance [23]. Because of these two growth steric hindrances, the ideal surface saturation reaction during the ALD process is ambitious. However, the residual active sites can continue to participate in chemical reactions, creating the conditions for the intralayer component control process to be achieved. Therefore, a thorough investigation of surface adsorption processes is required. In this study, typical cases of alucone films grown by MLD were examined. In-situ quadrupole mass spectrometry (QMS) and in-situ quartz crystal microbalance (QCM) were employed to investigate the effect of process temperature and precursor partial pressure on film growth.
Equations (1) and (2) represent two surface half-reactions of alucone films grown via MLD, where the chemical bonds exposed on the surface of the substrate are marked with a “*” [24].
The growth process of alucone grown by MLD at different temperatures (40, 80, and 120 °C) was monitored in real-time by in-situ QMS and in-situ QCM, where M/Z = 31, M/Z = 15, and M/Z = 16 represented the evolution of EG, TMA and CH4 molecular, respectively. As shown in Fig. 1(a-c), various partial pressures of precursors were observed at different process temperatures using the same EG pulse time. The EG molecule condensed heavily at a low temperature of 40 °C, [25] and low EG partial pressure led to an inadequate surface chemical reaction with less CH4 by-product generation, as shown by QMS curve. In addition, as depicted in the QCM curve, a negative mass gain following the maximum mass gain occurred after EG injection at 40 °C, revealing that physically adsorbed EG molecules after half-reaction underwent desorption during the purge process [23]. The lower the temperature, the greater the condensation and the longer the cleaning time after EG injection (33.1 s @ 40 °C, 23.2 s @ 80 °C, 19.9 s @ 120 °C), in which the cleaning time is determined by the moment when the partial pressure of the EG molecule is reduced to 10%. Meanwhile, a small amount of CH4 by-product was produced during the EG purge process at 40 °C, which indicates that the EG molecules condensed on the substrate surface temporarily blocked the active sites -CH3 beneath them. These blocked -CH3 sites would be re-active after the EG molecules were removed from the substrate surface during purge process [21]. In contrast, the condensation behavior of EG molecules was non-existent at 120 °C, resulting in a high partial pressure of EG. However, the alucone growth process with high EG partial pressure was accompanied by a small amount of CH4 by-product generation, which was attributed to alucone decomposition induced by a high process temperature of 120 °C. This reverse reaction catalysis resulted in an unsaturated surface chemical reaction during the growth of alucone at a high process temperature.
Fig. 1. Mass gain and molecular evolution results converted by in-situ QCM and In-situ QMS during MLD at (a) 40 °C, (b) 80 °C, and (c) 120 °C. (d) Molecular evolution results converted by In-situ QMS and (e) mass gain results converted by in-situ QCM during MLD at 40 °C with the EG pulse time of 1 s, 3 s, and 5 s. (f) The MPGC of alucone and Al2O3 grown by MLD and ALD with various oxidant precursor pulse times.
The partial pressure of the precursor also influenced the film growth process in addition to the process temperature. Alucone films were grown by MLD at 40 °C with various EG pulse times of 1 s, 4 s, and 7 s, respectively. The different EG partial pressures determined by the various pulse times were depicted in the QMS curves of Fig. 1(d). The partial pressure of TMA was selected as a reference intensity, and the pulse time of EG was increased from 1 s to 4 s, accompanied by a similar increase in the production of by-product CH4 during the respective half-reactions of EG and TMA, which indicated that the surface chemisorption state of the film growth was gradually saturated and the surface reaction was in equilibrium. However, when the EG pulse time was increased from 4 s to 7 s, the increase of by-product CH4 generation from the TMA half-reaction process was much greater than that from the EG half-reaction process, indicating that the substrate surface maintained an excess EG state prior to the TMA injection due to the easy condensation property of EG. The excess EG cannot be completely detached from the substrate surface within a fixed purge time, and the resulting surface CVD increased by-product CH4 production during the subsequent TMA half-reaction [26,27]. In contrast, the high activity of TMA rendered it almost non-condensing on the substrate surface, and its surface reaction was nearly saturated, allowing for an almost constant amount of by-product CH4 generation regardless of the length of the EG pulse. Figure 1(e) summarized the mass gain curves corresponding to 20 cycles at three different EG pulse times. The surface chemistry gradually reached saturation or even CVD state with the increase of EG pulse time, resulting in a corresponding increase in mass gain per cycle (MGPC) [26].
Similar to EG, O plasma was an oxidant precursor capable of growing Al2O3 films by chemically reacting with the surface -CH3 induced by TMA injection. Equation (3) described the surface half-reaction between -CH3* and O plasma, where the chemical bonds exposed on the substrate surface, were marked with a “*” [28].
To investigate the surface half-reaction saturation of the two oxidant precursors, the MGPC corresponding to different pulse times were obtained, respectively, as shown in Fig. 1(f). The TMA pulse time was set to 0.04 s to ensure a saturated surface half-reaction. According to the relationship between EG pulse time and MGPC, the alucone growth mode could be divided into three regions: the unsaturated reaction region, the saturation region, and the CVD region. In the unsaturated reaction region, the insufficient EG caused by a shorter pulse time made the surface half-reaction unsaturated, resulting in a large number of unreacted active sites remaining on the surface, and MGPC increased accordingly with the increase of EG pulse time. When the EG pulse time was increased to 4 s, the surface reaction was saturated and reached the saturation region. At this point, only unreacted active sites caused by the steric hindrance effect remained on the substrate surface. A small increase in the EG pulse time did not affect the surface reaction because sufficient purge time kept the reaction in equilibrium. By further increasing the pulse time of EG, the reaction equilibrium state was broken and a large number of EG molecules were physically adsorbed on the substrate surface, causing a surface CVD reaction and a significant increase in MGPC. Similarly, the surface oxidation reaction of Al2O3 gradually saturated with increasing O plasma pulse time, and the corresponding MGPC attained a maximum value of 47.86 ng/cm2·cycle at a pulse time of 16 s. Unlike the alucone growth process, the MGPC did not increase significantly when the O plasma pulse time was further increased during Al2O3 growth, indicating that O plasma did not physically adsorb on the substrate surface due to its high reactivity, thus, avoiding the CVD process.
In summary, the ideal saturated surface chemisorption in the ALD process was hard to achieve, and the degree of surface chemistry could be controlled by the partial pressure of the precursor and the process temperature. As shown in Fig. 2, after the unsaturated chemisorption between the “precursor 1” and the active sites on the substrate surface, the residual active sites were uniformly distributed on a macroscopic scale on the substrate surface and had the ability to react with the “precursor 2”, which enabled the component tailoring process for ALD. Therefore, the active state of the substrate surface played a decisive role in the component tailoring process, and both unsaturated surface chemisorption and steric hindrance effects could leave active sites on the substrate surface. Here, the component tailoring process was divided into two categories. The low partial pressure time of “precursor 1” resulted in more unreacted active sites residue on the substrate surface, and these active sites could then react with “precursor 2” via intralayer chemical reaction, resulting the first category called “surface saturated reaction component tailoring” process. In contrast, the self-limiting steric hindrance effects induced by the long-chain “precursor 1” could also make the unreacted active sites residue on the substrate surface, and the short-chain “precursor 2” could effectively use these active sites to achieve the intralayer growth due to their smaller space volume or higher reactivity, resulting in the second category called “steric hindrance effect component tailoring” process.
Fig. 2. Schematic diagram of component tailoring process: (a) Unsaturated surface adsorption of precursor 1. (b) Surface reaction of precursor 2. (c) Formation of hybridized states.
2.2 Surface saturated reaction component tailoring process
EG and O plasma, as two different oxidant precursors, could both chemically react with TMA to form the corresponding chemical structures. The surface chemical reaction model of “surface saturated reaction component tailoring” was described in Fig. 3(a) and represented as a three-step typical cycle.
(1) TMA pulse of 0.04 s reacted with the active sites -OH exposed to the substrate surface to form Al-O bonds. The excess TMA and by-product CH4 were removed in the subsequent purge process, while exposing -CH3 to the substrate surface as new active sites; (2) The pulse time of the subsequently injected EG was set to less than or equal to 4 s to achieve the transition of the surface chemical state from saturated to unsaturated, corresponding to the above-mentioned “saturation region” and “unsaturated reaction region”, which resulted in some unreacted active sites -CH3 remaining on the substrate surface, providing reaction conditions with O plasma; (3) After the residual EG and by-product CH4 were completely removed from the substrate surface, O plasma was injected with different pulse times and oxidized the residual -CH3 to re-expose the -OH to the substrate surface. The component tailoring in intralayer could be accomplished by regulating the partial pressure of EG and O plasma. Here, Table 1 displayed the partial pressure ratios developed based on the saturation conditions of EG and O plasma as typical recipes to analyze the component tailoring process.
Fig. 3. Surface saturated reaction component tailoring: (a) Schematic diagram of three-step typical cycle. (b) Molecular evolution results converted by In-situ QMS and (c) mass gain results converted by in-situ QCM during component tailoring processes with different partial pressure ratios of EG and O plasma. (d) FTIR spectra and (e) XPS high-resolution spectra (C 1s) of hybrid film grown by component tailoring processes.
Table 1. Typical recipes consist of various partial pressure ratios
The evolutions of TMA (M/Z = 15), EG (M/Z = 31), O2 (M/Z = 16), CH4 (M/Z = 16), and CO2 (M/Z = 44) molecules during component tailoring MLD process with different partial pressure proportions were monitored by in-situ QMS. As shown in Fig. 3(b), the pulse time of the TMA was set to 0.04 s to ensure its corresponding surface half-reaction saturation, and the precursor pulse time was positively associated with the partial pressure in the reaction chamber. The evolution of the by-product CH4 corresponding to different partial pressure ratios of EG and O plasma was marked with dashed lines of different colors, reflecting the surface reaction degree of the different precursors, respectively. As the partial pressure ratio of EG and O plasma gradually decreased (curves from left to right), the by-product CH4 produced by the surface half-reaction of EG gradually decreased (green dashed line), while the by-product CO2 produced by the surface half-reaction of O plasma gradually increased (blue dashed line), indicating that the surface reaction degree of the two oxidant precursors was effectively modulated by partial pressure. Furthermore, the by-product CH4 produced by the TMA surface half-reaction also showed a slow increase (black dashed line), due to the steric hindrance effect of EG resulting in a lower -OH saturated density after its surface half-reaction than that produced by O plasma. During partial pressure modulation, the shift of the oxidation half-reaction process from EG-dominated to O plasma-dominated caused an increase in the surface -OH density, which promoted the positive proceeding of the TMA half-reaction. This phenomenon was also confirmed by the trend of mass gain after TMA injection, as depicted in Fig. 3(c). O plasma had almost no steric hindrance effect during film growth due to its smaller spatial volume, and in addition, the higher reactivity prevented condensation on the substrate surface. Therefore, O plasma was able to impart a greater density of active sites (-OH) on the substrate surface compared to EG. The mass gain after TMA injection increased significantly as the partial pressure ratio of EG and O plasma decreased, indicating that more -OH was involved in the surface half-reaction of TMA. Also, the negative mass gain after EG injection, which represented the desorption of physically adsorbed EG molecules, decreased with decreasing EG partial pressure due to condensation mitigation. Meanwhile, a slight negative mass gain occurred during the purge process after TMA injection in processes where both EG and O plasma are involved, which indicated that the desorption process of EG molecules was still occurring slowly under the combined action of O plasma and Ar plasma carrier gas. More importantly, during the 3s-EG and 4s-O plasma process, a significant mass gain was generated after the O plasma injection, and decreased as the O plasma partial pressure increased. This phenomenon was attributed to the low density of the residual active sites -CH3 on the surface under the condition of a sufficient EG half-reaction, when the excess O plasma would oxidize with C in the EG, -C-H became -C-OH, resulting in a significant mass gain. As the EG partial pressure decreased, more active sites -CH3 remained on the substrate surface, consuming a large amount of O plasma, resulting in a less O plasma residual, which inhibited the oxidation reaction of C in EG and resulted a decrease in mass gain. In the absence of EG injection (0s-EG), sufficient active sites -CH3 on the surface were able to produce considerable mass gain after O plasma oxidation.
Fourier-transform infrared spectroscopy (FTIR) was used for characterizing the chemical composition of films grown by various processes (Fig. 3(d)). The FTIR intensity of the bending and stretching vibration bands of Al-O bonds (490–950 cm-1) was normalized for comparison of the relative contents of the various chemical structures [29,30]. The bending and stretching vibration bands of C-O bonds were recorded at wavelength of 1000–1200 cm-1, [31] indicating the characteristic structure introduced by the EG precursor. The bending and stretching vibration bands of O-C-O bonds recorded at wavelengths of 1300–1780cm-1 [32,33] represented the O-C-OH structure due to the oxidation of O-C-H by the O plasma injection following the EG pulse [31]. Briefly, the C-O and O-C-O structures represented almost the entire C environment. As the EG partial pressure time decreased, the chemisorption of EG molecules on the substrate surface was weakened, leading to less O-C-H structures and more active sites -CH3 remaining on the substrate surface. Therefore, the oxidation of O-C-H structures by O plasma became weaker and weaker, resulting in a consequent reduction of O-C-O structures. Meanwhile, the partial pressure regulation of the EG and O plasma precursors caused the film structure to shift between the films corresponding to each of the two precursors individually, the component tailoring process was realized. Furthermore, as shown in Fig. 3(e), X-ray photoelectron spectroscopy (XPS) was employed to compare the compositions and structures of various processes. The C 1s peak of films grown by the two recipes consisting of 4s-EG & 0s-O plasma and 3s-EG & 4s-O plasma, respectively, could be deconvoluted into three different bond configurations with C-C, C-O, and O-C-O at 284.8 eV, 286.4 eV, and 289.8 eV, respectively [34,35]. The characteristic peak of O-C-O bonds was unique to component tailoring process and represented the oxidation of O-C-H by O plasma, which was consistent with the FTIR analysis data.
In summary, the saturation degree of the chemical half-reaction on the substrate surface was controlled by the partial pressure regulation of different oxidant precursors, and the component tailoring of films was effectively achieved in intralayer. Such a process was known as “surface saturated reaction component tailoring”.
2.3 Steric hindrance effect component tailoring
The self-limiting steric hindrance effect of long-chain precursors prevented saturation of their surface half-reactions, resulting in some unreacted active sites beneath the molecular chains, which created opportunities for chemical reactions with short-chain precursors. The surface chemical reaction model of “ steric hindrance effect component tailoring” was described in Fig. 4(a) and represented as a three-step typical cycle.
(1) TMA pulse of 0.04 s reacted with the active sites -OH exposed to the substrate surface to form Al-O bonds. The excess TMA and by-product CH4 were removed in the subsequent purge process, while exposing -CH3 to the substrate surface as new active sites; (2) The pulse time of the subsequently injected EG was set to 4 s to achieve a saturated surface chemical state, corresponding to the “saturated region” mentioned above. However, some unreacted active sites -CH3 still remained on the substrate surface due to steric hindrance effect, providing reaction conditions with O plasma; (3) After the residual EG and by-product CH4 were completely removed from the substrate surface, O plasma was then injected and oxidized the residual -CH3 to re-expose the -OH to the substrate surface. More importantly, the surface chemisorption saturation degree of EG could be adjusted by process temperature and recipe, as verified in the section 2.1 and our previous work [23]. The high active surface state produced by a suitable process temperature may reduce the steric hindrance effect and promote the surface chemisorption saturation degree. Similarly, the multiple short pulse process (MSP) was performed by splitting the single EG pulse into two consecutive EG short pulses, which prolonged the EG partial pressure time while ensuring a certain amount of EG, minimized the film defects caused by the steric hindrance effect, and increase the degree of surface chemisorption saturation [23]. The higher the degree of saturation, the lower the residual amount of active sites on the substrate surface providing reaction conditions, resulting in component modulation of the film growth process. The residual amount of active sites on the substrate surface that could provide reaction conditions was modulated by controlling the surface reaction saturation degree, enabling the component tailoring of the film growth process.
Fig. 4. Steric hindrance effect component tailoring: (a) Schematic diagram of three-step typical cycle. (b) Mass gain results converted by in-situ QCM and (c) molecular evolution results converted by In-situ QMS during component processes with different EG saturation states. Effect of precursor spatial volume on the composition tailoring process: (d) Schematic diagram of three-step typical cycle. (e) Mass gain results converted by in-situ QCM and (f) molecular evolution results converted by In-situ QMS during component processes with and without O plasma saturation pre-reaction.
Process temperature and recipe were regulated as a single variable to achieve various substrate surface activity states. (1) The process temperature was increased to 80 °C to provide a better growth condition, while the process recipe remained unchanged. (2) MSP process was used to grow the film under temperature-constant conditions. As shown in Fig. 4(b-c), the mass change and by-product evolution of films grown by various processes were monitored in real-time by in-situ QMS and in-situ QCM. Less active site -CH3 remained on the substrate surface after EG half-reaction under the process performed by 80 °C and MSP. The less mass gain after O plasma injection obtained by these two steric hindrance optimized processes represented less O plasma was involved in the surface reaction when compared to the conventional process. Subsequently, the focus was shifted to the TMA injection. The inhibition of the steric hindrance effect by these two processes enhanced the surface half-reaction efficiency of EG. The higher the EG/O adsorption ratio, the lower the density of active sites -OH on the saturated substrate surface, resulting in less TMA participating in the surface reaction. As a result, when compared to the conventional process, both optimized processes exhibited less mass gain after TMA injection. Further, M/Z = 16 and M/Z = 44 represented the evolution of CH4 and CO2 molecules as the by-product for TMA/EG and O plasma were detected by in-situ QMS. The decrease in by-product CO2 production after O plasma injection indicated that the consumption of O plasma decreased with the decrease in active sites residuals. Notably, the by-product CH4 production also decreased after TMA injection, indicating that the oxidation of the carbon chains by O plasma reached saturation, and this decrease in mass gain was only caused by the decrease in the chemisorption ratio of O plasma to the active sites.
On the other hand, the pulse order of the EG and O plasma was exchanged to investigate the effect of precursor spatial volume on the composition tailoring achieved by the steric hindrance effect. As shown in Fig. 4(d), an O plasma pulse of 16 s followed by TMA pulse saturated the substrate surface chemistry, and a further 2 s EG was injected and expected to react with the residual active sites -CH3. In addition, the process without O plasma pre-reaction was performed as a reference to compare and analyze the surface reaction behavior of EG under various processes. The mass change and by-product evolution of films grown by various processes were monitored in real-time by in-situ QMS and in-situ QCM. The EG pulse followed by O plasma produced almost no by-product CH4 (M/Z = 16) (Fig. 4(f)), but maintained a small mass gain after desorption (Fig. 4(e)), indicating that the O plasma had completely consumed the active sites - CH3 that were expected to participate in the subsequent chemical reaction with EG. The injected EG condensed on the substrate surface and led to a mass gain instead of being involved in the corresponding surface chemical half-reaction. The additional bending and stretching vibration bands of C-O bonds compared to ALD-Al2O3 was recorded in the FTIR curve as depicted in Supplement 1, Fig. S2, which is caused by the physisorption of EG molecules. The by-product CH4 after TMA injection exhibited a reduced partial pressure as depicted in QMS curves due to the condensed EG molecules reduced the density of active sites -OH on the substrate surface.
2.4 Component ratio, film properties, and encapsulation application
To facilitate the analysis of film component tailoring ratios, “unit 1” and “unit 2” were defined as the chemical composition units introduced by O plasma and EG, respectively. XPS was employed to compare the compositions and structures of various component tailoring processes. The elements O and Al were observed as the constituent elements to analyze the component ratio between unit 1 and unit 2. The film composition of O plasma or EG as the only oxidant precursor had a fixed Al:O ratio, which could be used accordingly to deduce the unit 1:unit 2 ratio in various component tailoring processes. The specific calculation procedure was shown in Eq. (4)–(7) as follows:
Table 2. Summary of Al:O ratios and the derived unit 1:unit 2 composition ratios of films grown by various component tailoring processes
Based on the growth mechanism and film composition of component tailoring process, the next step was to calculate the growth rate per cycle (GPC), MGPC, and density of the relevant film processes. To this end, QCM (Supplement 1, Figs. S4-S5) and ellipsometry tests were used, and the results are summarized in Table 3 together with the refractive index values at 632.8 nm wavelength (Supplement 1, Fig. S6). As the proportion of O plasma surface reaction in these typical process recipes increased from “surface saturated reaction component tailoring”, both GPC and MGPC underwent a reasonable turning. The film growth mode under the joint action of EG and O plasma gradually changed from EG-dominated to O plasma-dominated. The film density also increased as the proportion of unit 1 increased, the same trend was reflected in the refractive index [36,37]. Further, the less produced GPC during the “steric hindrance effect component tailoring” process at 40 °C demonstrated that O plasma could effectively promote the desorption of EG molecules with surface reaction regaining its self-limiting nature. The less EG molecule physisorption caused by the establishment of ideal temperature and MSP processes resulted in a decrease of GPC and MGPC through inhibiting the generation of CVD processes on the substrate surface. The film density and refractive index gradually decreased with the increase of the EG surface reaction ratio.
Table 3. Summary of GPC, MGPC, density, and refractive index of films grown by various component tailoring processes
Moreover, as shown in Fig. 5(a-g), the film was grown on a 4-inch silicon wafer by component tailoring process to measure the residual stress according to the change in wafer’s curvature before and after deposition of the film. The compressive stress of the film gradually increased as the proportion of unit 1 increased, which was summarized in Fig. 5 h. The transmission in visible wavelength and surface morphology of hybrid films grown by various component tailoring processes were given in Supplement 1, Figs. S7 and S8, respectively.
Hybrid film grown by process g (2 + 2 s EG pulse / 4 s O plasma pulse) with high density and lower residual stress was selected as an encapsulation material for flexible OLEDs (Supplement 1, Fig. S9). As shown in Fig. 6(a), the WVTR value of hybrid film showed only some enhancement from 1.18 × 10−5 g·m-2·day-1 to 1.23 × 10−5 g·m-2·day-1 after 10,000 bends with a 3 mm bending radius. The uniformly distributed carbon backbone inside the film contributed to this excellent flexibility. In addition, the mild encapsulation process did not affect the performance of flexible OLEDs, which was depicted in Fig. 6(b). The flexible OLEDs with and without encapsulation were bent 1000 times at a bending radius of 3 mm for damage testing, and stored at 60 °C and 90% relative humidity subsequently for 600 h to investigate the luminance degradation, as shown in Fig. 6(c). The unencapsulated flexible OLEDs had significant luminance degradation due to accelerated moisture penetration during the bending process. The unencapsulated flexible OLEDs completely failed after 50 h, while the encapsulated flexible OLEDs preserved 97.8% of its initial luminance after 600 hours.
Fig. 5. Residual stress analysis: (a-g) Surface profile of different wafers before (top picture) and after (below picture) deposition of the films grown by various component tailoring processes. (h) Summary of residual stress of films grown by various component tailoring processes.
Fig. 6. (a) WVTR of hybrid film before and after 10000 bends with a 3 mm bending radius from 1/R-Time curve via electrical calcium test. (b) Current density and luminance-volage curve of flexible OLEDs before and after encapsulation. (c) Time evolution of normalized luminance for the flexible OLEDs with and without encapsulation under 60 °C and 90% relative humidity during and after bending.
3. Conclusion
The effect of precursor partial pressure and steric hindrance on the surface activity of films was thoroughly studied and successfully applied for the first time in the modulation of film components using the process called “component tailoring process”. The film component modulation in intralayer through tuning the precursor partial pressure and matching the substrate surface reaction saturation was called “surface saturated reaction component tailoring”. In contrast, the introduction of short-chain precursor components using the self-limiting property of the long-chain precursor steric hindrance effect was called “steric hindrance effect component tailoring”. The film growth processes associated with various recipes were characterized by their own reactions and were thoroughly analyzed via in-situ testing. It was demonstrated that the EG/O plasma surface reaction ratio could be modulated by varying partial pressure, and the component unit of the hybrid film under the joint action of EG and O plasma could achieve arbitrary ratios. Film growth parameters (GPC and MGPC) and physical properties (density, refractive index, residual stress, transmission, and surface morphology) could be modulated as desired. Moreover, as one of the important applications, the hybrid film with low residual stress preserved intact barrier performance after 10,000 bends with a 3 mm bending radius, and was effectively used in the encapsulation of flexible OLEDs without any damage. The proposed component tailoring process broke the traditional film growth process via ALD, and realized the in-situ control of thin film components at atomic level in intralayer without thickness limitation, which filled the technical gap in ALD component control, and made the application of ALD more multifaceted.
4. Experimental section
4.1 Fabrication of ALD/MLD films
The pressure in the reaction chamber was maintained at 0.25 Torr during the process, and high-purity Ar (99.999%) with a flow rate of 100 sccm was adopted as the carrier gas and cleaning gas of the precursor. First, ALD-Al2O3: Trimethyl aluminum (TMA, 99.9999%) and O plasma were used as the aluminum precursor and oxidizer of ALD-Al2O3, respectively. Throughout the ALD process, TMA was kept at room temperature, and O plasma was generated by 15 sccm O2 (99.999%) at a radio frequency power of 100 W. The substrate temperature was set to 40 °C, and typical process parameters for the growth of ALD-Al2O3 consisted of O plasma pulse of 4 s - 20 s as required, O plasma purge of 100 s, TMA pulse of 0.04 s, and TMA purge of 100 s. Second, MLD-alucone: TMA and ethylene glycol (EG, 99.8%) were used as the aluminum precursor and oxidizer of MLD-alucone, respectively. The TMA was maintained at room temperature, while the EG was heated to 80 °C for better reactivity. The substrate temperature was set to 40 °C, 80 °C, or 120 °C as required, and typical process parameters for the growth of MLD-alucone consisted of TMA pulse of 0.04 s, TMA purge of 100 s, EG pulse of 1 s–9 s as required, EG purge of 120 s. Third, component tailoring process: TMA, EG, and O plasma were used as the aluminum precursor, oxidizer A, and oxidizer B of hybrid film, respectively. The O plasma was generated by 15 sccm O2 (99.999%) under a radio frequency power of 100 W. The TMA was maintained at room temperature, while EG was heated to 80 °C for better reactivity. The substrate temperature was set to 40 °C in “surface saturated reaction component tailoring” and 40 °C or 80 °C as required in “steric hindrance effect component tailoring”, and typical process parameters for the growth of hybrid film consisted of TMA pulse of 0.04 s, TMA purge of 100 s, EG pulse of 0 s - 4 s as required, EG purge of 120 s, O plasma pulse of 0 s - 16 s as required, and O plasma purge of 100 s. For MSP process, and typical process parameters consisted of TMA pulse of 0.04 s, TMA purge of 100 s, first EG pulse of 2 s, first EG purge of 30 s, second EG pulse of 2 s, second EG purge of 90 s, O plasma pulse of 4 s, and O plasma purge of 100 s.
4.2 Film characterization
In-situ quartz crystal microbalance (QCM) (SQM160, INFICION) and quadrupole mass spectrometry (QMS) measurements were simultaneously performed during film growth by ALD, MLD, and component tailoring process to study evolutions of the film mass and by-products in the chamber. Hybrid films, ALD-Al2O3 and MLD-alucone were deposited on KBr tablets for 300 cycles before being analyzed using Fourier transform infrared spectroscopy (FTIR, PerkinElmer Spectrum One) to determine the chemical components based on the bond bending and stretching vibration. In the meantime, X-ray photoelectron spectroscopy (XPS, Thermo ESCALab 250 analyzer) was used in a constant energy mode to determine the elemental composition, elemental ratio (C, O, Al), and chemical bond structures (C-O bonds, O-C-O bonds, and C-C bonds) of both films grown on 10 × 10 mm2 silicon wafers for 300 cycles. To obtain film thickness and refractive indexes, various films grown on 25 × 25 mm2 silicon wafers were measured at three ellipsometry angles (55°, 65°, and 75°) using variable-angle spectroscopic ellipsometry (J. A. Woolam), and the experimental data were fitted using the Cauchy model, for thickness testing, films were grown for 300 cycles, and the growth rate per cycle (GPC) was calculated by the total thickness and total cycles (GPC = total thickness / total cycles), for refractive indexes testing, films were grown at 50 nm according to the respective GPC. The optical transmittance of films in the visible wavelength was measured by UV–vis–IR spectrophotometer (UV3600, Shimadzu), atomic force microscopy (AFM, using ICON-PT, Bruker) was used to analyze surface topography of films under a tapping mode, all samples were grown on glass substrates at 50 nm.
The film was grown on silicon wafer substrate (4-inch, 650 µm) at 50 nm to measure the residual stress. The wafer surface was scanned for the difference in the curvature change before and after deposition of the film. The residual stress of the film can be calculated using the Stony equation:
Here, Es is the Young's modulus and vs is the Poisson ratio of the silicon wafer, ts and tf are the thicknesses of silicon wafer and deposited film, respectively. r1 and r2 are the radius of curvature of silicon wafer before and after deposition of the film, respectively.
Electrical calcium corrosion testing was used for the detection of WVTR of barrier film with 50 nm thickness. The calcium film with a thickness of 200 nm was evaporated to a 10 × 2.25 mm region, and the Al film with a thickness of 100 nm was deposited on both sides of the calcium film as contact electrodes. Agilent B2902A Precision Source (Agilent Technologies, Inc., Santa Clara, CA, USA) was used to measure the electrical conductance (1/R) of the calcium film. The WVTR was calculated according to Eq. (9):
4.3 Device preparation and characterization
50 nm-hybrid film was placed below PEN substrate and above the device for the bottom and top encapsulation, respectively, to avoid moisture penetration.
The organic layers were deposited on cleaned PEN substrates with the ITO anode. The typical stack structure contained 5 nm thick 1,4,5,8,9,11-hexaazatriphenylene hexacarbonitrile (HAT-CN) as the hole injection layer, 45 nm thick 1,1-bis[(di-4-tolylamino) phenyl] cyclohexane (TAPC) as the hole transport layers, 20 nm thick 4,4’-bis(N-carbazolyl)-1,1’-biphenyl (CBP) doped with 8% bis[2-(2-pyridinyl-N)phenyl-C](acetylacetonato)iridium(III) (Ir(ppy)2(acac)) as the emitting layer, 30 nm thick 2,2’,2''-(1,3,5-Benzinetriyl)-tris(1-phenyl-1-H-benzimidazole) (TPBi) as the electron transport layer, 1 nm 8-hydroxyquinolinolatolithium (Liq) as the electron injection layer, and 15 nm Ag as the cathode.
The performance of OLEDs devices was measured with a home-built I-V-L test system using a Photo Research Inc. PR-680 L SpectraDuo Spectroradiometer/Photometer and a Keysight B2902A Precision Source/Measure Unit. The performance of the device was only measured according to the emission on ITO side. The luminance of OLEDs with and without encapsulation was measured at 9 V during damage test and stability test.
Funding
National Natural Science Foundation of China (61675088, 61974054); International Science & Technology Cooperation Program of Jilin (20190701023GH); Scientific and Technological Developing Scheme of Jilin Province (20200401045GX); Project of Science and Technology Development Plan of Jilin Province (20190302011G); Jilin Provincial Science and Technology Development Plan Project (20230201040GX, 20230508058RC).
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
Supplemental document
See Supplement 1 for supporting content.
References
1. S. M. George, “Atomic Layer Deposition: An Overview,” Chem. Rev. 110(1), 111–131 (2010). [CrossRef]
2. O. Graniel, M. Weber, S. Balme, P. Miele, and M. Bechelany, “Atomic layer deposition for biosensing applications,” Biosens. Bioelectron. 122, 147–159 (2018). [CrossRef]
3. J. Sheng, J.-H. Lee, W.-H. Choi, T. Hong, M. Kim, and J.-S. Park, “Review Article: Atomic layer deposition for oxide semiconductor thin film transistors: Advances in research and development,” J. Vac. Sci. Technol., A 36(6), 060801 (2018). [CrossRef]
4. Y. Yu, Z. Zhang, X. Yin, A. Kvit, Q. Liao, Z. Kang, X. Yan, Y. Zhang, and X. Wang, “Enhanced photoelectrochemical effciency and stability using a conformal TiO2 film on a black silicon photoanode,” Nature Energy 2(2017).
5. E. Song, Y. K. Lee, R. Li, J. Li, X. Jin, K. J. Yu, Z. Xie, H. Fang, Y. Zhong, H. Du, J. Zhang, G. Fang, Y. Kim, Y. Yoon, M. A. Alam, Y. Mei, Y. Huang, and J. A. Rogers, “Transferred, ultrathin oxide bilayers as biofluid barriers for flexible electronic implants,” Adv. Funct. Mater. 28(12), 1702284 (2018). [CrossRef]
6. M. R. Islam, M. Rahman, S. F. U. Farhad, and J. Podder, “Structural, optical and photocatalysis properties of sol-gel deposited Al-doped ZnO thin films,” Surf. Interfaces 16, 120–126 (2019). [CrossRef]
7. S. D. Ponja, B. A. D. Williamson, S. Sathasivam, D. O. Scanlon, I. P. Parkin, and C. J. Carmalt, “Enhanced electrical properties of antimony doped tin oxide thin films deposited via aerosol assisted chemical vapour deposition,” J. Mater. Chem. C 6(27), 7257–7266 (2018). [CrossRef]
8. J. Multia, J. Heiska, A. Khayyami, and M. Karppinen, “Electrochemically active in situ crystalline lithium-organic thin films by ALD/MLD,” ACS Appl. Mater. Interfaces 12(37), 41557–41566 (2020). [CrossRef]
9. J. Sheng, T. Hong, D. Kang, Y. Yi, J. H. Lim, and J.-S. Park, “Design of InZnSnO semiconductor alloys synthesized by supercycle atomic layer deposition and their rollable applications,” ACS Appl. Mater. Interfaces 11(13), 12683–12692 (2019). [CrossRef]
10. P. Banerjee, W.-J. Lee, K.-R. Bae, S. B. Lee, and G. W. Rubloff, “Structural, electrical, and optical properties of atomic layer deposition Al-doped ZnO films,” J. Appl. Phys. (Melville, NY, U. S.) 108(4), 043504 (2010). [CrossRef]
11. R. E. Agbenyeke, S. Song, H. Choi, B. K. Park, J. H. Yun, T.-M. Chung, Y. K. Lee, C. G. Kim, and J. H. Han, “Atomic layer deposition of Cu2SnS3 thin films: effects of composition and heat treatment on phase transformation,” Chem. Mater. 33(20), 8112–8123 (2021). [CrossRef]
12. J.-G. Song, G. H. Ryu, S. J. Lee, S. Sim, C. W. Lee, T. Choi, H. Jung, Y. Kim, Z. Lee, J.-M. Myoung, C. Dussarrat, C. Lansalot-Matras, J. Park, H. Choi, and H. Kim, “Controllable synthesis of molybdenum tungsten disulfide alloy for vertically composition-controlled multilayer,” Nature Communications 6, (2015). [CrossRef]
13. M. E. Donders, W. M. Arnoldbik, H. C. M. Knoops, W. M. M. Kessels, and P. H. L. Notten, “Atomic Layer Deposition of LiCoO2 Thin-Film Electrodes for All-Solid-State Li-Ion Micro-Batteries,” J. Electrochem. Soc. 160(5), A3066–A3071 (2013). [CrossRef]
14. H. Liu, Y.-F. Liu, P.-P. Xiong, P. Chen, H.-Y. Li, J.-W. Hou, B.-N. Kang, and Y. Duan, “Aluminum-doped zinc oxide transparent electrode prepared by atomic layer deposition for organic light emitting devices,” IEEE Trans. Nanotechnol. 16(4), 634–638 (2017). [CrossRef]
15. Q. Q. Hou, F. J. Meng, and J. M. Sun, “Electrical and optical properties of Al-doped ZnO and ZnAl2O4 films prepared by atomic layer deposition,” Nanoscale Res. Lett. 8 (2013).
16. D.-J. Lee, H.-M. Kim, J.-Y. Kwon, H. Choi, S.-H. Kim, and K.-B. Kim, “Structural and electrical properties of atomic layer deposited Al-Doped ZnO films,” Adv. Funct. Mater. 21(3), 448–455 (2011). [CrossRef]
17. S. H. Lee, H. J. Jeong, K. L. Han, G. Baek, and J. S. Park, “An organic-inorganic hybrid semiconductor for flexible thin film transistors using molecular layer deposition,” J. Mater. Chem. C 9(12), 4322–4329 (2021). [CrossRef]
18. S. Lee, Y.-H. Kang, M.-S. Kim, H. Lee, Y.-H. Cho, M. Kim, T.-S. Yoon, H.-M. Kim, and K.-B. Kim, “Effect of the bilayer period of atomic layer deposition on the growth behavior and electrical properties of the amorphous In-Zn-O film,” ACS Appl. Mater. Interfaces 12(35), 39372–39380 (2020). [CrossRef]
19. N. P. Dasgupta, S. Neubert, W. Lee, O. Trejo, J.-R. Lee, and F. B. Prinz, “Atomic layer deposition of Al-doped ZnO films: effect of grain orientation on conductivity,” Chem. Mater. 22(16), 4769–4775 (2010). [CrossRef]
20. R. L. Puurunen, “Growth per cycle in atomic layer deposition: A theoretical model,” Chem. Vap. Deposition 9(5), 249–257 (2003). [CrossRef]
21. S. M. George, B. Yoon, and A. A. Dameron, “Surface Chemistry for Molecular Layer Deposition of Organic and Hybrid Organic-Inorganic Polymers,” Acc. Chem. Res. 42(4), 498–508 (2009). [CrossRef]
22. T. Muneshwar and K. Cadien, “Surface reaction kinetics in atomic layer deposition: An analytical model and experiments,” J. Appl. Phys. 124(9), 095302 (2018). [CrossRef]
23. H. Wang, Z. Wang, X. Xu, Y. Liu, C. Chen, P. Chen, W. Hu, and Y. Duan, “Multiple short pulse process for low-temperature atomic layer deposition and its transient steric hindrance,” Appl. Phys. Lett. 114(20), 201902 (2019). [CrossRef]
24. H. Wang, Z. Wang, X. Xu, W. Zhao, D. Wu, M. Muhammad, Y. Liu, C. Chen, B. Liu, and Y. Duan, “A Novel Nucleation Inducer for Ultrathin Au Anodes in High Efficiency and Flexible Organic Optoelectronic Devices,” Adv. Opt. Mater. 8(4), 1901320 (2020). [CrossRef]
25. A. A. Dameron, D. Seghete, B. B. Burton, S. D. Davidson, A. S. Cavanagh, J. A. Bertrand, and S. M. George, “Molecular layer deposition of alucone polymer films using trimethylaluminum and ethylene glycol,” Chem. Mater. 20(10), 3315–3326 (2008). [CrossRef]
26. X. Liang, M. Yu, J. Li, Y.-B. Jiang, and A. W. Weimer, “Ultra-thin microporous-mesoporous metal oxide films prepared by molecular layer deposition (MLD),” Chemical Communications, 7140-7142 (2009).
27. H. Pedersen, “Time as the Fourth Dimension: Opening up New Possibilities in Chemical Vapor Deposition,” Chem. Mater. 28(3), 691–699 (2016). [CrossRef]
28. H. Wang, Y. Liu, H. Liu, Z. Chen, P. Xiong, X. Xu, F. Chen, K. Li, and Y. Duan, “Effect of Various Oxidants on Reaction Mechanisms, Self-Limiting Natures and Structural Characteristics of Al2O3 Films Grown by Atomic Layer Deposition,” Adv. Mater. Interfaces 5(14), 1701248 (2018). [CrossRef]
29. B. H. Lee, B. Yoon, V. R. Anderson, and S. M. George, “Alucone Alloys with Tunable Properties Using Alucone Molecular Layer Deposition and Al2O3 Atomic Layer Deposition,” J. Phys. Chem. C 116(5), 3250–3257 (2012). [CrossRef]
30. N. Batra, J. Gope, J. Vandana, R. Panigrahi, P. K. Singh, and Singh, “Influence of deposition temperature of thermal ALD deposited Al2O3 films on silicon surface passivation,” AIP Adv. 5(6), 067113 (2015). [CrossRef]
31. D. N. Goldstein, J. A. McCormick, and S. M. George, “Al2O3 Atomic Layer Deposition with Trimethylaluminum and Ozone Studied by in Situ Transmission FTIR Spectroscopy and Quadrupole Mass Spectrometry,” J. Phys. Chem. C 112(49), 19530–19539 (2008). [CrossRef]
32. S. Sahoo, C. Chakraborti, P. Behera, and S. Mishra, “FTIR and Raman Spectroscopic Investigations of a Norfloxacin/Carbopol934 Polymeric Suspension,” Journal of young pharmacists : JYP 4(3), 138–145 (2012). [CrossRef]
33. J. Kwon, M. Dai, M. D. Halls, and Y. J. Chabal, “Detection of a formate surface intermediate in the atomic layer deposition of high-k dielectrics using ozone,” Chem. Mater. 20(10), 3248–3250 (2008). [CrossRef]
34. A. Jejurikar, X. T. Seow, G. Lawrie, D. Martin, A. Jayakrishnan, and L. Grondahl, “Degradable alginate hydrogels crosslinked by the macromolecular crosslinker alginate dialdehyde,” J. Mater. Chem. 22(19), 9751–9758 (2012). [CrossRef]
35. M. M. Hyland and A. A. Qaiser, “X-ray Photoelectron Spectroscopy Characterization of Polyaniline-cellulose Ester Composite Membranes,” Mater. Sci. Forum 657, 35–45 (2010). [CrossRef]
36. P. Jin, G. He, D. Q. Xiao, J. Gao, M. Liu, J. G. Lv, Y. M. Liu, M. Zhang, P. H. Wang, and Z. Q. Sun, “Microstructure, optical, electrical properties, and leakage current transport mechanism of sol-gel-processed high-k HfO2 gate dielectrics,” Ceram. Int. 42(6), 6761–6769 (2016). [CrossRef]
37. M. Jerman, Z. H. Qiao, and D. Mergel, “Refractive index of thin films of SiO2, ZrO2, and HfO2 as a function of the films’ mass density,” Appl. Opt. 44(15), 3006–3012 (2005). [CrossRef]
