Realization of highly-dense Al₂O₃ gas barrier for top-emitting organic light-emitting diodes by atomic layer deposition

Abstract

In this paper Al₂O₃ films are prepared with a method of atomic layer deposition (ALD) as the thin film encapsulation technology for top-emitting organic light-emitting diodes (TE-OLED). Time-of-flight secondary ion mass spectrometry (TOF-SIMS), X-Ray Reflectometry (XRR) and X-ray photoelectron spectroscopy (XPS) are used to analyze the effect of different chemical precursors and behavioral factors on the performance of the Al₂O₃ thin films. The analyses disclosed that Al₂O₃ films prepared with a trimethylaluminum (TMA) and H₂O (TMA + H₂O) process contained more unreacted –CH₃ groups, and the films with a TMA + O₃ process show a large number of carbon-based impurities. However, the Al₂O₃ films prepared using H₂O and O₃ in turn in a deposition cycle as oxygen sources exhibited higher density and purity, leading to a superior water vapor transmission rate (WVTR) as low as 5.43 × 10⁻⁵ g per m² per day estimated with the calcium (Ca) corrosion method at 40 °C/100%. The TE-OLED with Al₂O₃ (TMA + H₂O + O₃) thin film as an encapsulated layer has longer lifetimes, and produces no black spots under operational times up to 400 hours.

---

1. Introduction

A big challenge for organic light-emitting diode (OLED) development in widespread application is their vulnerability to ambient moisture and oxygen.^1–3 The requirement of water vapor transmission rate (WVTR) in OLEDs, that needs to be less than 10⁻⁶ g per m² per day, is more rigorous than that of any other packaging application.⁴ Glass to glass encapsulation technology is the state-of-the-art scheme for organic electronics, but the thin film encapsulation (TFE) technique is indispensable for the realization of flexible OLEDs. Nowadays, atomic layer deposition (ALD) is considered as the most promising method to produce a gas barrier for the encapsulation of OLEDs, because of its special features, such as being a low-temperature process, conformal film deposition, pinhole-free, a high quality layer, and its excellent uniformity.^5–8

For the deposition of gas barrier layers by ALD, aluminum oxide (Al₂O₃) is the most investigated and used material in the field of thin film encapsulation. The deposition is usually performed using TMA and H₂O as precursors and reactants, which is considered as an ideal self-limiting process for ALD growth. Since the first report by G. S. Higashi and C. G. Fleming in 1989,⁹ the low temperature process of Al₂O₃ has been extensively studied and characterized regarding the details of precursor chemistry and performance evaluation for TFE application.^10,11 Besides water, ethylene glycol and ozone or oxygen plasma have been widely used as oxygen reactants with TMA to obtain a dense film.^12–14 Most recently, Duan’s group have optimized the pumping gas time (PGT) to purge the excess H₂O or O₃ as a way to fabricate an Al₂O₃ gas barrier with a lower WVTR.¹⁵ In addition, Al₂O₃ films as building blocks were employed in the laminated structure for the configuration of a highly impermeable gas barrier, such as Al₂O₃/TiO₂ and Al₂O₃/ZrO₂ stacked structures.^16,17

In this study, we investigate the gas barrier performance of ALD thin film encapsulations with a single layer of Al₂O₃. For this investigation, the influence of the gas species and its way of being imported into the substrate on the barrier properties is examined. The highlight of this work is that with H₂O and O₃ as basic reactants in the ALD process, an Al₂O₃ film with a lower WVTR can be achieved by dividing the introduction of the oxygen source into two steps, that is to say, the H₂O reactant was firstly pulsed into the substrate with TMA chemisorbed, and then O₃ was pulsed into the chamber for further reaction with the residual methyl group in TMA. This ternary reaction system (TMA + H₂O + O₃) can realize an Al₂O₃ film with high density, which achieves a superior WVTR as low as 5.43 × 10⁻⁵ g per m² per day. Additionally, the deposition rate of this reaction can be greatly improved. The top-emitting OLED (TE-OLED) encapsulated with ternary reacted Al₂O₃ exhibits excellent storage properties without any visible black spot growth under the particular conditions (temperature = 40 °C, humidity = 100%).

2. Experimental procedures

For this study, the Al₂O₃ layer was grown at 80 °C with a BENEQ TFS500 system, using TMA, H₂O, and O₃ as the precursor gases. The ozone precursor was produced with an ozone generator BMT Messtechnik 803N using oxygen and catalytic nitrogen to gain approximately 8 g h⁻¹ ozone production. In this experiment, we fabricated three kinds of Al₂O₃ thin film as a comparison to identify the role of the oxidizing precursors reacting with TMA. Films A, B and C were denoted as the Al₂O₃ films fabricated with the TMA + O₃, TMA + H₂O, and TMA + H₂O + O₃ precursors. The pulse time for TMA and H₂O was fixed at 0.25 s, and the pulse durations for O₃ were varied from 0 to 50 s. The purge time for TMA, H₂O, and O₃ was kept the same as 10.0 s. In the ternary reaction systems, the import of the O₃ precursor was following the completion of H₂O purge. The schematic illustration of the technical flow process is shown in Fig. 1.

Fig. 1 Illustration of technical flow process for Al₂O₃ film with TMA + H₂O + O₃ ternary reaction system.

The film thickness and reflective index of the deposited films were measured by spectroscopy ellipsometry (SE). X-Ray Reflectometry (XRR, Bruker D8 advance diffractometer) as an effective approach to evaluate the film density was carried out. The dielectric performance was characterized by a semiconductor parameter analyzer (Agilent B1500A). For the film with different oxygen reactants, a secondary ion mass spectroscope (SIMS, PHYSICAL ELECTRONICS, TRIFT II Model 2100) was used to confirm the incorporation of elements and chemical groups in the Al₂O₃ films. The chemical bonding states of the bulk Al₂O₃ films were observed with an X-ray photoelectron spectroscopy (XPS, Thermo, ESCALAB 250). The film interfacial properties of Al₂O₃ used as a barrier film were investigated with high-resolution transmission electron microscopy (HR-TEM, JEOL, JEM-2100F).

In this study, the structure of the green TE-OLED was glass/Ag (100 nm)/IZO (7.6 nm)/MeO-TPD:F4-TCNQ (156 nm), 4 wt%/NPB (20 nm)/VOM1511:GD-5 (40 nm)/Bebq2: (25 nm)/LiF (1 nm)/Mg : Ag (9 : 2, mass ratio, 15 nm)/ZnSe (24 nm), where MeOTPD:F4-TCNQ is tetrafluoro-tetracyanoqino dimethane doped into N,N,N,N-tetrakis(4-methoxyphenyl)-benzidine as the hole injection layer (HIL), NPB is N,N-bis(naphthalen-1-yl)-N,N-bis(phenyl)benzidine as the hole transport layer (HTL), VOM1511:GD-5 is the green emitting layer, Bebq2 is a bis(10-hydroxybenzo[h]quinolinato)beryllium complex as the electron transport layer (ETL) and LiF/Mg : Ag is the transparent cathode. ZnSe is the capping layer. MeO-TPD, F4-TCNQ, NPB, Bebq2 were home-made, VOM1511 and GD-5 were commercial materials and purchased from Visionox Corporation and ZnSe was purchased from Alfa Aesar.

3. Results and discussion

In order to characterize the growth mechanism of the Al₂O₃ films, the thickness variations with different oxidizing reactants were measured using a spectroscopic ellipsometer. The films were prepared using 500 cycles of ALD deposition grown on a Si wafer. The results show that the growth rate of the TMA + H₂O and TMA + O₃ Al₂O₃ films was approximately 0.796 Å per cycle and 1.189 Å per cycle, which means that O₃ is more effective to acquire thicker Al₂O₃ films than H₂O during one cycle, owing to the stronger oxidizing capacity of the O₃ gas precursor.¹⁸ Meanwhile, by increasing the pulse time of H₂O from 0.25 s to 2.0 s (the data is not shown here) for TMA + H₂O Al₂O₃, the film growth rate will be increased gradually but with an almost saturated value of 0.943 Å per cycle, indicating that the H₂O reactants could not completely exhaust the molecule group in the TMA precursor at the deposition temperature. But with the introduction of O₃ after H₂O in the cycle process, the growth rate started to increase rapidly. As shown in Fig. 2, the increment of time for the O₃ import will sharply improve the reaction rate, and the values are 1.157 and 1.356 Å per cycle for 2 s and 30 s O₃ pulse times. The improvement of deposition rate indicates that the residual TMA precursor in the TMA + H₂O process at the low processing temperature can be consumed by a subsequent O₃ cycle.

Fig. 2 The film growth rate per cycle (GRC) and refractive index of Al₂O₃ films fabricated with different oxygen reactants.

The refractive index measured using spectroscopic ellipsometers are 1.632 and 1.582 for TMA + H₂O and TMA + O₃ Al₂O₃ films, respectively, which are similar to previous reports.¹⁹ But after the H₂O reactant, when the following O₃ was introduced into the reaction system, the refractive index gradually dropped from 1.602 to 1.580 by increasing the O₃ pulsing time from 2.0 s to 50.0 s. Conventionally, the decrease of refractive index implies the degradation of the density of the film. And the density of barrier films influences the barrier performance crucially.

To further confirm the film density of the Al₂O₃ films, XRR as an effective approach to evaluate the density was carried out. Through the critical angle analysis by XRR measurement, the fitting value of density could be obtained, as shown in Fig. 3(a). The XRR density for the Al₂O₃ films deposited with TMA + H₂O and TMA + O₃ is 2.95 g cm⁻³ and 2.80 g cm⁻³. The maximum XRR density (3.00 g cm⁻³) is observed for the film with the introduction of ozone after H₂O reactants, implying that the decrease of refractive index with the addition of ozone pulses in the TMA + H₂O process cannot be explained by the decreased density of the Al₂O₃ films. The lowest XRR density is obtained with the TMA + O₃ Al₂O₃ film, which is due to the nature of the ozone species at the specific temperature. The number of effective radicals generated by ozone at low temperature is inadequate, accompanied with the traits of lower ozone concentration and short radical lifetimes. This determines that ozone can’t effectively penetrate into the internal groups in every mono-molecular deposition to form a dense Al₂O₃ framework. As another important clue, Fig. 3(b) shows the leakage behavior of Al₂O₃ dielectrics with different oxygen reactant. The leakage current decreases as the H₂O and O₃ precursor is introduced into the chamber stepwise due to a more complete reaction by the more drastic oxygen species. The test results prove that TMA + H₂O + O₃ Al₂O₃ shows a superior dielectric property in comparison to the single reactant of H₂O or O₃. Hence the film density or quality of Al₂O₃ increases with the additional introduction of ozone in the process. This improvement could be understood as a compaction of the Al₂O₃ film through removal of holes/voids as well as a reduction of hydrogen impurities.

Fig. 3 (a) The film density measured with XRR and (b) insulator characteristics of Al₂O₃ films fabricated with different oxygen reactants.

For the films with different gas precursors, the TOF-SIMS was measured to confirm the residual unreacted methyl and the incorporation of hydrogen related species in the Al₂O₃ films. As seen in Fig. 4, the TMA + O₃ Al₂O₃ film shows the lowest amount of –CH₃ groups in the internal film, resulting from the fact that ozone has better oxidation properties. But in the traditional TMA + H₂O Al₂O₃ film, a larger amount of –CH₃ was embedded, indicating that H₂O reactants could not fully exhaust the group in the every TMA pulsing cycle at the low process temperature. But introducing O₃ with a stronger oxidizing capacity could effectively consume part of the unreacted –CH₃ group, thus making the film more dense as manifested in the TMA + H₂O + O₃ Al₂O₃. Moreover, the amount of –OH and –H species which survived in the TMA + H₂O + O₃ Al₂O₃ film was reduced as compared to TMA + H₂O Al₂O₃. These species are speculated to be one of the reasons for the lower water and oxygen permeability, which seems to speed up the formation of a permeable channel for water and oxygen molecules. The lower concentration of –OH and –H species could be used for prolonging the diffusion paths of permeated molecules.

Fig. 4 TOF-SIMS results of –CH₃, –OH, and –H groups in Al₂O₃ films with different oxygen reactants, and the value of –CH₃ content was amplified by 50 times as a comparison.

The chemical composition and bonding states of the Al₂O₃ films fabricated with different precursors were investigated in detail with the C1s, O1s and Al2p core level spectra measured using X-ray photoelectron spectroscopy. To avoid surface contamination of the samples during the XPS measurements, the films were sputtered off the extreme surface of samples exposed to ambient air. Fig. 5(a)–(c) show the high resolution of the C1s core level spectra for the Al₂O₃ films. The C1s core level was fitted with two peaks assigned to –CH₂–CH₂– bonds (C1sA peak set at 285.3 ± 0.1 eV), C=O bonds (carboxyl groups and carbonates, C1sB peak at 290.4 ± 0.1 eV).²⁰ As shown in Fig. 5(a), it can be seen that a lot of C=O bonds survived in the Al₂O₃ film, owing to the stronger oxidizing capacity of ozone with the result of methyl oxidation. The carbonate related impurity C=O induced by the ozone process may lead to the generation of irregular Al–O–Al bridges, ultimately deteriorating the density of the Al₂O₃ film proven by XRR measurement. The inferior density would lead directly to a poor gas barrier performance. But, as shown in Fig. 5(b), the absence of a carboxyl-related signal and a smaller peak area indicates no carbonate impurities and less carbon residual in the TMA + H₂O film. On the other hand, interestingly, the amount of C=O bonds and the total carbon residual is remarkably increased by inducing an extra following ozone pulse in the TMA + H₂O + O₃ ternary reacted system. The existence of C=O bonds can be explained by the reaction between residual –CH₃ groups, that could not be triggered during the H₂O pulse, and the following O₃ precursors. Moreover, the increasing content of carbon compared with the TMA + H₂O process should be attributed to the residual oxidation functional groups (hydroxyl, epoxide etc.) induced by the O₃ precursor. These residual groups could further react with –CH₃ groups in the following cycle of the overdosed TMA precursor, thus leaving carbon-related impurities in the Al₂O₃ film. The latter required a bit of explanation that the reduction of the reflective index obtained in Fig. 2 can be well explained by this introduction of carbon related impurity in the ternary reacted Al₂O₃ film, but it seems not to degrade the film density.

Fig. 5 The XPS spectra of C1s, O1s and Al2p in (a, d, g) TMA + O₃, (b, e, h) TMA + H₂O, and (c, f, i) TMA + H₂O + O₃ Al₂O₃ films.

The XPS survey spectra presented in Fig. 5 also show the high resolution O1s and Al2p core level spectra. Three peaks were used for fitting the O1s spectra as shown in Fig. 5(d)–(f). The main component (O1sA) at the BE of 530.8 ± 0.1 eV is attributed to the oxide ions of the alumina matrix. The O1sB peak observed at 531.7 ± 0.1 eV can be identified as oxygen vacancies or hydroxyl groups bonded with Al atoms. The third peak, O1sC at 532.7 ± 0.1 eV, is associated with water molecules due to surface adsorption.²¹ The TMA + O₃ Al₂O₃ film has a lower percentage of oxygen vacancies because of the strong oxidability generated by ozone, but a higher proportion of oxygen/water adsorption indicating the loose structure in the film. In the TMA + H₂O Al₂O₃ film, H₂O as the oxygen reactant cannot effectively form an ideal Al–O ratio leading to the existence of a certain amount of oxygen vacancies. However, the oxygen in the lattice greatly rises in the TMA + H₂O + O₃ ternary reacted system, while the adsorption of water on the surface reduces. This improvement occurred after the introduction of ozone, which made the alumina structure more dense. Two peaks were used for fitting the Al2p core level spectra as shown in Fig. 5(g)–(i). The Al2pA peak at 74.1 ± 0.2 eV is assigned to the Al(III) ions of an oxide matrix.²² The Al2pB peak at higher BE (75.0 ± 0.3 eV) is assigned to the Al(III) ions of a hydroxide matrix, as shown in Table 1. The O : Al atomic ratio calculated with the O1sA and Al2pA components is 1.42, 1.41, and 1.44 for the TMA + O₃, TMA + H₂O, and TMA + H₂O + O₃ Al₂O₃ samples. This means that the ternary reacted film is in more excellent agreement with the stoichiometry of Al₂O₃ (O : Al ratio of 1.5) and confirms the assignment of the O1sA and Al2p peaks. This peak information also shows that the binding energy of the Al2pA peak in the TMA + H₂O + O₃ film shifts to a lower position (73.9 eV) as a comparison with the other two samples, indicating that the ternary reacted system is more essential for obtaining an Al₂O₃ film with stoichiometry. This near-perfect structure will greatly strengthen the barrier performance of the TMA + H₂O + O₃ Al₂O₃ film. Moreover the amount of –OH and –H species will be significantly reduced owning to the ozone oxidation, which will finally enhance the film quality to repel the molecule diffusion. As a result, the TMA + H₂O + O₃ ternary reacted system will be more effective for the realization of an Al₂O₃ film with better barrier performance.

Table 1 The summary of the C1s, O1s, and Al2p peak properties for the Al₂O₃ films

Element precursor	C1s		O1s			Al2p		Atomic ratio (C/O/Al)
	C1sA	C1sB	O1sA	O1sB	O1sC	Al2pA	Al2pB
O3	23.34%	76.66%	31.13%	36.29%	32.56%	40.64%	50.36%	2.01/57.5/40.49
H2O	100.00%	0.00%	26.61%	46.85%	26.55%	43.86%	56.14%	0.41/58.27/41.32
H2O/O3	40.01%	59.99%	35.97%	44.71%	19.31%	46.44%	53.56%	0.90/58.49/40.62

---

The film structure of the Al₂O₃ layer used as a barrier film in the TFE structure was investigated by high-resolution transmission electron microscopy (HR-TEM). The four layers of Al₂O₃ were continuously deposited into the glass substrate with the sequence as TMA + H₂O + O₃, TMA + O₃, TMA + H₂O, and TMA + H₂O + O₃ based Al₂O₃ films. For this measurement, our starting point was to consider the quality difference of the bulk Al₂O₃ film that was fabricated with variable oxygen reactants, while slightly having shed light on the influence of the surface properties that induce the initial grown film inconsistencies on another surface. Fig. 6 shows the interface between the adjacent layers of any two of these materials. Especially, the HR-TEM images clearly show that the TMA + O₃ Al₂O₃ film is quite different from the two other films. Experimental facts demonstrate that the contrast ratio of the electron image depends on the directions of the incident electron beam relative to that of the crystals, which exhibits a lower film density. Therefore, the influence of the TMA + O₃ Al₂O₃ film structure on the performance of the gas-diffusion barrier can be included as one of the main causes. In another respect, the bulk characteristics between the TMA + H₂O + O₃ and TMA + H₂O Al₂O₃ are not very obvious, which can be also manifested in XRR measurements that have a close value of the film density.

Fig. 6 The interface characteristics of Al₂O₃ films with variable oxygen reactants.

The WVTR measurements were carried out to evaluate the gas permeability of the Al₂O₃ film using a traditional calcium (Ca) corrosion method, which involves a Ca sensor at 40 °C and 100% RH. The 200 nm-thick Ca film with length/width (L/W) as 10/20 mm was deposited on the patterned titanium electrodes (100 nm) with the shape of two narrow bars. The electrical measurements were performed using two electrodes connected by a SMU probe to the Keithley 2400 source meter. It is assumed that the Ca film follows a laterally homogeneous corrosion during the erosion induced by the water or oxygen molecule, as a result we can see that the amount of calcium left shows a directly proportional relationship between the calcium left and current measured. The WVTR value was determined using the following equation:²³

where d(G)/d(t) is the change of Ca conductance as a function of time, and the n is the molar equivalent of the degradation reaction. In this equation, δ_Ca and ρ_Ca are the Ca resistivity (3.4 × 10⁻⁸ Ω m) and density (1.55 g cm⁻³), where M (H₂O) and M (Ca) are the molar masses of water and Ca. According to the geometry of the experimental setup, the ratio of the area of the Ca sensor to the area of the window for water permeation is consistent.

From the Ca test, as shown in Fig. 7, the WVTR value for the TMA + O₃ Al₂O₃ film is 4.15 × 10⁻⁴ g per m² per day, while it is 9.64 × 10⁻⁵ g per m² per day for the TMA + H₂O Al₂O₃ film. It seems that the H₂O reactant is more effective than the O₃ to fabricate high quality Al₂O₃ films, although the oxidizing capacity of O₃ is much stronger than H₂O. But in the ternary reaction system of TMA + H₂O + O₃, the predicted permeation rate of the Al₂O₃ film calculated from this equation was 5.43 × 10⁻⁵ g per m² per day, which has a better comparability to other single Al₂O₃ films. These results indicate that the O₃ oxidizing gas precursor is not alone in affecting the film properties of the Al₂O₃ film and its barrier performance, indicating that it is quite different from those of a single source of oxygen species; therefore, we speculate that the introduction of O₃ after H₂O oxidizing gas precursors will be used as an effective way to further react with the TMA precursor which, combined with the formation of a dense film, seems to strengthen the resistance of water and oxygen molecules. Actually the permeation mechanism and dynamic penetration process of water and molecular oxygen in the Al₂O₃ film is more complex. Therefore, much more in-depth investigation of this ternary reacted film is necessary when used in thin film encapsulation technology.

Fig. 7 The normalized change of electrical conductance of Ca film with different Al₂O₃ as a function of time.

The WVTR results have implied to us that the TFE structure based on the TMA + H₂O + O₃ Al₂O₃ film may have excellent barrier properties when integrated on TE-OLEDs. The electroluminescent (EL) spectra of the encapsulated TE-OLEDs are firstly investigated. Fig. 8(a) shows that the EL spectra almost do not depend on the Al₂O₃ film architecture, which means that there are only slight differences in the EL spectra behavior between the devices with Al₂O₃ fabricated using different precursors can be observed, indicating that the performance of the TE-OLEDs did not change appreciably with the process of thin-film encapsulation. To further verify the quality of the encapsulation layers, lifetime tests were performed. Fig. 8(b) shows the typical plots of normalized luminance versus operating time for the TE-OLEDs in a nonstop constant-voltage mode with a starting luminance of 10 000 cd m⁻², which were measured at constant temperature and humidity using a luminance meter (KONICA MINOLTA CS-2000). For this study, the lifetime is defined as the decay time that the luminance decreases to 5000 cd m⁻². Actually it is hard to make a distinction between the internal mechanism of OLED degradation and external permeated water/oxygen degradation. Herein, we put forward the difference in the Al₂O₃ films used in encapsulation with the assumption that the decay of the OLED itself is the same, so the luminance of the OLED device measured is merely a response of the Al₂O₃ barrier performance for deterioration. As shown in Fig. 8(b), the luminance of the device with TMA + H₂O + O₃ Al₂O₃ encapsulation measured in an oven deteriorated slower (with lifetime exceeding over 400 hours) than the elapsed luminance of the devices with TMA + O₃ or TMA + H₂O Al₂O₃ film capping. This suggests that the degradation induced by the H₂O or O₂ gas permeation into the TE-OLEDs was blocked off effectively in the ternary reacted film, which could be attributed to the better WVTR performance.

Fig. 8 (a) The EL spectra and (b) the TE-OLEDs luminance versus continuous operation times for Al₂O₃ encapsulation with variable oxygen reactants.

The photography images of the encapsulated TE-OLEDs stored in an oven (40 °C, 100%) for 160 hours were recorded to verify the effect of TFE. The black spots are usually caused by cathode delamination due to reactions with ambient H₂O or O₂ gases. As shown in Fig. 9(c) and (f), the TMA + H₂O + O₃ Al₂O₃ film encapsulated OLEDs display a good image without obvious black spots. For further observation, the TE-OLEDs encapsulated with H₂O or O₃ derived Al₂O₃ films have been almost weakened with the generation of many large-sized black spots, and the proportion of the these black spots is thriving or having wider permeation paths to cause further damage. After being encapsulated with a dense Al₂O₃ barrier, the growth of initial black spots alleviates with a very slow rate, manifesting an excellent barrier performance for this kind of Al₂O₃ film.

Fig. 9 Photography of three comparative TE-OLEDs. Before aging: (a) TMA + O₃, (b) TMA + H₂O, (c) TMA + H₂O + O₃; after aging in oven for 160 hours: (d)–(f).

4. Conclusions

In summary, the barrier performance of Al₂O₃ films based on different oxygen reactants with TMA by the use of atomic layer deposition was investigated. The TMA + H₂O + O₃ ternary reacted systems would lead to a more dense film structure in comparison with the single H₂O or O₃ based Al₂O₃. The TOF-SIMS results demonstrated that the H₂O precursor could not completely exhaust the –CH₃ group in the TMA chemisorbed substrate without the further import of O₃ gas, but the film with only O₃ as an oxygen reactant (TMA + O₃) shows a certain amount of carbonate impurities, which will degrade the Al–O–Al bonding state. The produced Al₂O₃ thin film with the TMA + H₂O + O₃ process shows an O/Al atomic ratio of 1.44 and deposition rate of 1.33 A per cycle, determining that it will have great advantages in the industrial production for future wearable/flexible displays when integrated in OLED thin film encapsulation.

Acknowledgements

The authors are grateful to the National Basic Research Program of China (Grant No. 2015CB655000), the National Natural Science Foundation of China (Grant No. 51502093, 61574061, 61574062), China Postdoctoral Science Foundation (Grant No. 2015M582385), the Fundamental Research Funds for the Central Universities (2014ZM0034, 2015ZM018, 2015ZM072, 2015ZM075) and Guangdong Innovative Research Team Program (No. 201101C0105067115).

References

1. M. S. Weaver, L. A. Michalski, K. Rajan, M. A. Rothman, J. A. Silvernail, J. J. Brown, P. E. Burrows, G. L. Graff, M. E. Gross and M. Zumhoff, Appl. Phys. Lett., 2002, 81, 2929–2931.
2. R. Grover, R. Srivastava, M. N. Kamalasanan and D. S. Mehta, RSC Adv., 2014, 4, 10808.
3. E. Gautier, A. Lorin, J. M. Nunzi, A. Schalchli, J. J. Benattar and D. Vital, Appl. Phys. Lett., 1996, 69, 1071–1073.
4. P. E. Burrows, G. L. Graff, M. E. Gross, P. M. Martin, M. Hall, E. Mast, C. C. Bonham, W. D. Bennett, L. A. Michalski, M. S. Weaver, J. J. Brown, D. Fogarty and L. S. Sapochak, Proc. SPIE, 2001, 75–83.
5. A. A. Dameron, S. D. Davidson, B. B. Burton, P. F. Carcia, R. S. McLean and S. M. George, J. Phys. Chem. C, 2008, 112, 4573.
6. H. Kim, H. B. R. Lee and W. J. Maeng, Thin Solid Films, 2009, 517, 2563–2580.
7. J. Meyer, P. Goerrn, F. Bertram, S. Hamwi, T. Winkler, H. Johannes, H. T. Weimann, P. Hinze, T. Riedl and W. Kowalsky, Adv. Mater., 2009, 21, 1845–1849.
8. O. Sneh, R. B. Clark-Phelps, A. R. Londergan, J. Winkler and T. E. Seidel, Thin Solid Films, 2002, 402, 248–261.
9. G. S. Higashi and C. G. Fleming, Appl. Phys. Lett., 1989, 55, 1963.
10. M. D. Groner, F. H. Fabreguette, J. W. Elam and S. M. George, Chem. Mater., 2004, 16, 639–645.
11. Y. G. Lee, Y.-H. Choi, I. S. Kee, H. S. Shim, Y. W. Jin, S. Lee, K. H. Koh and S. Lee, Org. Electron., 2009, 10, 1352–1355.
12. X. Wang, Y. Duan, F. B. Sun, Y. Q. Yang, D. Yang, P. Chen, Y. H. Duan and Y. Zhao, RSC Adv., 2014, 4, 43850.
13. Y.-Q. Yang, Y. Duan, P. Chen, F.-B. Sun, Y.-H. Duan, X. Wang and D. Yang, J. Phys. Chem. C, 2013, 117, 20308–20312.
14. D. Cao, X. Cheng, L. Zheng, D. Xu, Z. Wang, C. Xia, L. Shen, Y. Yu and D. Shen, J. Vac. Sci. Technol., B: Nanotechnol. Microelectron.: Mater., Process., Meas., Phenom., 2015, 33, 2166–2746.
15. Y. Q. Yang and Y. Duan, J. Phys. Chem. C, 2014, 118, 18783–18787.
16. J. Meyer, H. Schmidt, W. Kowalsky, T. Riedl and A. Kahn, Appl. Phys. Lett., 2010, 96, 243308–243310.
17. A. Singh, H. Klumbies, U. Schröder, L. Müller-Meskamp, M. Geidel, M. Knaut, C. Hoßbach, M. Albert, K. Leo and T. Mikolajick, Appl. Phys. Lett., 2013, 103, 233302.
18. A. Singh, F. Nehm, L. Müller-Meskamp, C. Hoßbach, M. Albert, U. Schroeder, K. Leo and T. Mikolajick, Org. Electron., 2014, 15, 2587–2592.
19. Y.-Q. Yang, Y. Duan, Y.-H. Duan, X. Wang, P. Chen, D. Yang, F.-B. Sun and K.-W. Xue, Org. Electron., 2014, 15, 1120–1125.
20. B. Díaz, E. Häköen, J. Swiatowska, V. Maurice, A. Seyeux, P. Marcus and M. Ritala, Corros. Sci., 2011, 53, 2168–2175.
21. H. Jung, H. Jeon, H. Choi, G. Ham, S. Shin and H. Jeon, J. Appl. Phys., 2014, 115, 073502.
22. L. H. Kim, K. Kim, S. Park, Y. J. Jeong, H. Kim, D. S. Chung, S. H. Kim and C. E. Park, ACS Appl. Mater. Interfaces, 2014, 6, 6731–6738.
23. Y. C. Han, C. Jang, K. J. Kim, K. C. Choi, K. Jung and B.-S. Bae, Org. Electron., 2011, 12, 609–613.

---