LU93244B1 - Coating comprising a layer of TIO2 allowing SNO2 atomic layer deposition - Google Patents

Abstract

The first object of the invention is directed to a coating comprising at least one layer of TiO2. Said coating is remarkable in that said at least one layer of TiO2 has a crystalline orientation with Miller indices equivalent to (001), (110) or (111). The second object is directed to process for depositing TiO2 on a substrate, comprising the steps of (a) providing a substrate; (b) performing said deposition with TiCI4 and water by atomic layer deposition on said substrate; and (c) performing an annealing step. Said step (b) is carried out at a temperature comprised between 50°C and 200°C, said step (c) is carried out at a temperature comprised between 450°C and 600°C in ambient air. Further layer(s) can be afterwards deposited on to the TiO2 coating.

Description

COATING COMPRISING A LAYER OF TlO2 ALLOWING SnO2 ATOMIC LAYER DEPOSITION

Description

Technical field [0001] The invention is directed to a coating made of T1O2 and obtained by atomic layer deposition. Said coating is designed to be the basis for a second deposition, said second deposition being performed in a smooth manner thanks to the crystalline orientation of the T1O2 layer.

Background art [0002] It is known that the major problem of T1O2 ALD synthesis from TiCU and water is the important roughness of the crystalline films (see Cheng H.-E., et a!., Morphological and Photoelectrochemical properties of ALD T1O2 films, J. Electrochem. Soc., 2008, 155 (9), D604). Therefore it is mandatory to optimize the process growth in order to get the best trade-off between the film quality (that would trigger the photocatalytic properties) and its roughness that needs to be compatible with a 40 nm pore diameter.

[0003] Earlier research on T1O2 ALD deposition from TiCU and H2O determined typical deposition temperatures in a range of 200°C - 400°C (see Aarik J. et ai, Morphology and structure of T1O2 thin film grown by ALD, J. Cryst. Growth., 1995, 148(3), 268-275). However, the high reactivity of the precursor also enables film formation at lower temperature (see Tan L. K., et ai., Free-Standing Porous Anodic Alumina Templates for ALD of Highly Ordered Τ1Ό2 Nanotube Arrays an Various Substrate, 2008, 112(1), 69-73).

[0004] However, when porous substrates are used, deposition at low temperature is known to cause an inhomogeneous deposition into the porous space, leading to rough surface. This is a problem when further coating must be performed, for example for providing particular surface properties to substrate presenting a complexed surface.

Summary of invention

Technical Problem [0005] The invention has for technical problem to alleviate at least one of the drawbacks present in the prior art. More particularly, the invention has for technical problem to provide a coating which allows further coating to be performed in a smooth manner.

Technical solution [0006] The first object of the invention is directed to a coating comprising at least one layer, said at least one layer being a layer of TiCte. Said coating is remarkable in that said at least one layer of TiCte has a crystalline orientation with Miller indices, i.e. (hkl), equivalent to (001), (110) or (111).

[0007] According to a preferred embodiment, said at least one layer of TiCte comprises a surface energy superior to 0.60 J/m2.

[0008] According to a preferred embodiment, said at least one layer of TiCte has a thickness comprised between 20 nm and 200 nm, preferentially between 20 nm and 90 nm, more preferentially equal to 35 nm.

[0009] According to a preferred embodiment, coating further comprises a second layer, said second layer being preferentially a layer of SnCte.

[0010] The second object is directed to process for depositing TiCte on a substrate, comprising the steps of (a) providing a substrate; (b) performing said deposition with TiCU and water by atomic layer deposition on said substrate; and (c) performing an annealing step. Said step (b) is carried out at a temperature comprised between 50°C and 200°C, said step (c) is carried out at a temperature comprised between 450°C and 600°C in ambient air.

[0011] According to a preferred embodiment, said step (b) is carried out at a temperature comprised between 50°C and 100°C.

[0012] According to a preferred embodiment, said substrate is curved glass.

[0013] According to a preferred embodiment, said substrate is porous.

[0014] According to a preferred embodiment, said step (b) comprises pulsing

TiCU on said substrate so as to form a functionalized substrate, purging with an inert gas, pulsing H2O on said functionalized substrate, and purging with an inert gas, said inert gas being preferentially nitrogen gas.

[0015] According to a preferred embodiment, said purging with an inert gas lasts between 5 seconds and 30 seconds.

[0016] According to a preferred embodiment, said step (c) is carried out for at least 2 hours.

[0017] According to a preferred embodiment, said step (b) is performed between 100 times and 1000 times.

[0018] The third object of the invention is directed to a process for depositing at least one component on a substrate covered with at least one layer of T1O2, said process comprising the steps of (a’) performing the process in accordance with the process of the second object of the invention, (b’) performing a deposition of said at least one component on said substrate covered with at least one layer of T1O2; and (c’) performing a postdeposition treatment.

[0019] According to a preferred embodiment, said component is SnC>2 and wherein said step (b’) is the step of performing the deposition with SnCL and water by atomic layer deposition.

[0020] According to a preferred embodiment, said atomic layer deposition is carried out at a temperature comprised between 200°C and 350°C.

[0021] According to a preferred embodiment, said step (b’) is performed 3000 times.

[0022] In general, the particular embodiments of each object of the invention are also applicable to other objects of the invention. To the extent possible, each object of the invention is combinable with other objects.

Advantages of the invention [0023] The invention is particularly interesting in that it provides a coating that allows to be itself coated with another component, in particular SnC>2, in order to form a smooth deposition of this component which is difficult to deposit on substrate presenting a complexed surface. Such surface might be curved, porous and/or both.

Brief description of the drawings [0024] Figure 1: Optical (SEM) images of T1O2 films deposited in temperature range of RT-400°C.

[0025] Figure 2: X-Ray Diffraction (XRD) T1O2 anatase (101) peak evolution as a function growth of temperature (200°C-400°C).

[0026] Figure 3: Diffractogram of the T1O2 film synthesized at 400°C.

[0027] Figure 4: Growth rate of ALD T1O2 films deposited at RT-400°C.

[0028] Figure 5: Possible paths of TiCU chemisorption on oxide surfaces.

[0029] Figure 6: QCM as a function of deposition time recorded profile at 200°C using purge time of 2 s, 5 s, 30 s and 600 s for both precursors.

[0030] Figure 7: QCM as a function of deposition time as a recorded profile at 100°C using purge time of 600 s for both precursors.

[0031] Figure 8: QCM as a function of deposition time as a recorded profile at 20°C using purge time of 1800 s for both precursors.

[0032] Figure 9: QCM as a function of deposition time as a recorded profile at 20°C using purge time of 5s and 30s for both precursors.

[0033] Figure 10: XRD of T1O2 deposited at 200°C after annealing at 600°C.

[0034] Figure 11: Optical (SEM) images of T1O2 after annealing at 600°C. T1O2 films were deposited at 200°C after 500, 1000 and 2000 cycles with purge time of 30 s and after 500, 1000 cycles with purge time of 5 s annealed.

[0035] Figure 12: XRD of T1O2 deposited at 100°C after annealing at 600°C.

[0036] Figure 13: Optical (SEM) images of T1O2 films deposited at 100°C after 350, 700 and 1650 cycles with purge time of 5s and 30s annealed at 600°C.

[0037] Figure 14: SEM (a) and band contrast (b) and EBSD (c) images for T1O2 samples of thickness of 35 nm (I) and 70 nm (II).

[0038] Figure 15: XRD of ΤΊΟ2 deposited at room temperature after annealing at 600°C.

[0039] Figure 16: Optical (SEM) images of annealed T1O2 films deposited at RT after 160, 320 and 965 cycles with purge time 5s and 30s.

[0040] Figure 17: AFM topography images of annealed T1O2 films deposited at RT after 160, 320 and 965 cycles with purge time 5s and 30s.

[0041] Figure 18: XPS on Ti 2p, O 1s and Cl 2p spectra on amorphous and annealed samples, grown in short (5 s) and long (30 s) purge time regimes at RT.

[0042] Figure 19: TiO2 film surface topography and contact potential difference measurements on the same area.

[0043] Figure 20: EBSD map and associated pole figures.

[0044] Figure 21: Optical (SEM) images of SnO2 film deposited on Si/TiO2 substrate.

[0045] Figure 22: Atomic structure of anatase TiO2 (001) and (101) surfaces and their water-adsorption behaviour.

[0046] Figure 23: XRD on SnO2 grown with and without TiO2 buffer layer on Si (100) substrate after 3000 ALD cycles.

[0047] Figure 24: SnO2 deposition after 3000 ALD cycles on AAO with TiO2 buffer layer.

[0048] Figure 25: Deposition in porous template steps: (a) AAO without deposition, (b) deposition 20 nm of TiO2, (c) SnO2 deposition.

[0049] Figure 26: Optical (SEM) images of fabricated TiO2/SnO2 periodic hétérostructures. Top-view.

[0050] Figure 27: Optical (SEM) images of fabricated TiO2/SnO2 periodic hétérostructures. Tilted 20°.

Description of an embodiment [0051] Thin films of -25-40 nm are obtained using standard conditions provided by the ALD equipment supplier: precursors pulse 0.2 s and purge time 2 s for both precursors (for temperature range 200-400°C).

[0052] In the low temperature range (20°C (RT) -200°C) depositions of thicker films (-100 nm) are obtained with longer purge time as 5 s and 30 s. The SEM and AFM measurement (see figure 1 and table 1) show that the roughness of films increases with the deposition temperature.

[0053] Table 1: Summary on the thickness and roughness of T1O2 films (as deposited)

[0054] Samples grown in the low temperature range (RT-200°C) show a smooth and highly conformal films, while films obtained at 350°C show the highest RMS value (16.98 nm).

[0055] It is worth noting that the roughness decreases suddenly for the films deposited at 400°C. These results until 350°C are in agreement with the literature (see Huang. Y, et ai, Characterization of low temperature deposited ALD Ti02 for MEMS application, J. Vac. Sci. Technoi. A Vacuum, Surfaces, Films, 2012, 31, 01A148). Therefore, from the morphological point of view, for further deposition into porous substrate, the low temperature range (RT-200°C) is privileged.

[0056] However, for better understanding of the morphology variation, the growth mechanisms have to be discussed. The formation of random nanoparticles for the deposition at 200°C and purge time 2 s has been observed, similarly reported by Luka et ai. (Kinetics of anatase phase formation in Τ1Ό2 films during ALD and post-deposition annealing, CrystEngComm., 2013, 15(46), 9949). Such nanoparticles formation may unfortunately cause an inhomogeneous deposition into the porous space. The increase of the purge time to 5 and 30 s has solved this problem.

[0057] The XRD analysis demonstrates that films synthesised at a temperature below 300°C do not show any crystalline structure, therefore

diffractograms for films deposited at RT and 100°C are not presented (see figure 2). The anatase crystalline phase is detected for the depositions beyond 300°C and the increase of deposition temperature up to 400°C logically improves the film crystallinity (see figure 3).

[0058] The growth rate (growth per cycle, GPC) in whole temperature range (RT-400°C) represented on figure 4 for TiO2 films.

[0059] Three growth regimes can be distinguished: (i) low temperature regime (RT-200°C) where the growth rate decreases with the temperature increase that corresponds to the ALD condensation regime; (ii) intermediate regime (200°C-350°C) demonstrate a growth rate increase followed by the standard ALD plateau (iii) at high temperature regime (>350°C).

[0060] It is generally established that the surface reaction with the metal-oxide precursor mainly involves the OH functional groups (see Aarik J. et a!., ALD of Τ1Ό2 from TiCU and H2O: investigation of growth mechanism, Appi. Surf. Sci., 2001, 172, 148-158). Therefore, the typical TiO2 growth reaction follows equations 1 and 2 and is schematically represented on Figure 5.

[0061] [Pulse TiCU]: x(-OH)x + TiCU— (-O-)xTiCI4-x + xHCI (x=1, 2) (equation 1)

[0062] [Pulse H2O]: (-O-)xTiCI4-x + (4 - x)H2O (-O-)xTi(OH)4.x + (4-x)HCI (equation 2) [0063] Nevertheless, the surface chemistry is highly dependent on the temperature. At low temperatures (<300°C), the isolated hydroxyl groups OH and H-bonded OH groups are dominate on the surface of substrates such as SiO2 or AI2O3. The increase of temperature (>300°C) may lead to the dehydroxilation and the formation of oxygen bridges (equation 3).

[0064] Ti(-OH)-O-Ti(OH) -+Ti(-O-)2Ti +H2O (equation 3) [0065] At high temperature, the oxygen bridge terminated surface promotes also the direct TiCU chemisorption according to lla-llb reaction paths showed on figure 5 and used to improve the film quality because of the lack of HCI by-product.

[0066] The GPC increase in the intermediate regime can be related to the temperature dependent surface functionalities and also the important role of the HCI by-product on the growth mechanisms.

[0067] In the literature, the role of HCI on the T1O2 growth mechanism was fully established by Leem et at., Rote of HCi in ALD of Τ1Ό2 thin films from TiCU and water, 2014, 35(4), 1195-1201). The authors showed that in the temperature range 150-300°C, the volontary addition of the HCI after the TiCL pulse significantly lowers the growth rate. For the high temperature range (>300°C), the additional HCI increased the growth rate. The authors postulated that at lower temperatures (150-300°C), the surface reaction occurs via OH-groups and forms intermediate or non-stable products, according to the reactions depicted on equations 1 and 2. HCI may induce the reversible reaction and promote the desorption of TiCL-x or Ti(OH)2Cl2 volatile compounds. The overall reaction being partially reversible, the growth rate of T1O2 is lowered. Beyond 300°C, the oxygen bridge terminated surfaces are induced and promote the TiCL adsorption as previously described without any release of HCI. The oxygen bridges are mainly triggering the T1O2 growth reaction according to equation 4.

[0068] Ti(-O-)2Ti +TiCL - Ti(CI)-O-Ti(OTiCI3) (equation 4) [0069] This investigation highlights that the mono- (figure 5 la) or bifunctional (Figure 5 lb) absorption of TiCL on hydroxyl groups is reversible under HCI atmosphere, while dissociative absorption on oxygen bridges is irreversible. Interestingly, while HCI was added at the end of the ALD cycle (after water pulse), it significantly decreases the growth rate for the whole studied temperature range (150 - 400°C). When oxygen bridges are exposed to HCI (equation 5 and the schematic lllb-lllc represented on figure 5). T1O2 surfaces are functionalised with -OH, -Cl, =OHCI. That retards the growth of T1O2. This fundamental investigation highlights the predominant role of HCI on the T1O2 growth mechanism. The time of residence of HCI in our ALD processes is strongly dependent on the purge time.

[0070] Ti(-O-)2Ti + HCI -+ Ti(CI)-O-Ti(OH) (equation 5) [0071] Therefore, further investigations require the optimisation of purge time using the quartz crystal microbalance (QCM). The QCM being integrated into the reactor lid cannot be used at deposition temperature above 250°C. Henceforth, the low temperature TiO2 deposition regime was principally considered: at RT, 100 and 200°C, so more as these regimes conduct to the low surface roughness of TiO2 films.

[0072] The importance of the purge time is logically more significant in the low temperature regime, where the excess of precursors and by-products desorption is not assisted by heat.

[0073] The QCM measurements realised first at 200°C using pulse time 0.2 s and purge time 2, 5, 30 s for both precursors (see figure 6). The pulse of TiCU induces the mass increase following stabilisation during the purge.

[0074] The addition of a water pulse enables the surface reaction and the mass drop due to the release of HCI by-product (figure 6B). The completed mass stabilisation was not achieved in the case of a 2 s purge time.

[0075] The increase of the purge time up to 5 (figure 6C) and 30 s (figure 6D) improved the mass stabilisation. However, for a purge time of 30 s after the TiCU pulse, an anomalous mass drop after 12-15 s of purge time is observed (figure 6D). To further study the “relaxation” time after the pulse of each precursor, an exaggerated purge of 600 s is applied (figure 6E).

[0076] This study shows that after the TiCU pulse the mass signal decreases continuously during the first 5-8 s of purging, but then the mass reincreases significantly without any visible stabilisation. The water pulse promotes a slow mass decrease within the first 10 s of purging, then the sensor signal increases again.

[0077] Such QCM behaviour of mass increases within the long purge time is intriguing. It may be associated to the re-adsorption of volatile species or surface reaction by-products from the reactor walls and re-deposits on the QCM surface.

[0078] Moreover, after the completed 50 cycles deposition with purge time 2 and 5 s, the mass sensor signal continues to decrease; that could be characteristic of the desorption of volatile compounds (Figure 6 A zoom).

[0079] Aarik and co-workers (Aarik J. et ai, ALD of T1O2 from TiCU and H2O: investigation of growth mechanism, Appi. Surf. Sci., 2001, 172, 148-158) also pointed out on the continuous decrease of mass sensor signal during the purge time for depositions realised at 100°C. The authors have suggested the continuous dehydration of the surface after exposure to water pulse at low temperature, or desorption/decomposition of formed Ti(OH)xCl4-x volatile species.

[0080] The QCM control on the mass evolution during the ALD deposition demonstrated that, at 200°C, the purge time after the water injection should be longer than 600 s.

[0081] The decrease in the deposition temperature to 100°C does not significantly modify the mass sensor profiles compared to the previously recorded for 200°C. The study of the relaxation time at 100°C demonstrates that a purge time of 600s is still not sufficient to stabilise the mass sensor signal after the pulse of TiCU, however after injection of water, the mass sensor signal stabilisation is interestingly achieved after a purge time of 30 s of purge (see figure 7).

[0082] The mass change, as recorded by QCM, for depositions realised at room temperature is significantly different from previously discussed QCM profiles recorded for thin-films grown at 100 and 200°C. An anomalous behaviour is noted while the water pulse is supplied. Instead of the mass decrease due to the HCI release, the mass sensor shows a mass increase and then a continuous decrease of the mass increment without stabilisation, even after 30 min of purge (figure 8).

[0083] The mass increment per cycle is strongly dependent on the purge time. The mass increments per cycle of ~ 0.14, 0.06, and 0.02 pg/cm2 is determined for depositions realised with 5, 30 and 1800 s purge time. The final mass increment, after 50 cycles, is twice larger for 5 s purge time than for 30 s purge time (see figure 9) at room temperature.

[0084] For high temperature (100 and 200°C), the final mass increment is not significantly different for the different purge time. This demonstrates that the unreacted precursors or by-product thus accumulate within the film at room temperature. At a low temperature, the surface is mainly functionalised by hydroxyl groups and H-bonded OH. Furthermore at temperatures below 100°C, the presence of molecularly absorbed water on the surface is inevitable. Thus, TiCL molecules could react not only with functional groups but also with residual water. The reaction of TiCL hydrolysis at room temperature generally leads to the formation of TiOCL, TiOCI, T12O3 and/or Ti(OH)xCL-x with generation of HCI. All of these chemistries may be expected in the films.

[0085] Several theoretical studies have reported that the first half-reaction of TiCL and H2O is endothermic and thus the desorption of HCI is hampered. That is in line with our hypothesis about an accumulative growth mechanism at room temperature.

[0086] Nevertheless, it is important to consider the slow film growth rate (~6 Â/ cycle for purge 30 s) for which a further increase in the purge time (to reach the ideal ALD regime) will lead to an unreasonably long deposition time (dozens of hours), for the considered TiCL thickness.

[0087] Consequently, TiCL thin-films deposited with purge times of 5 s and 30 s were chosen and since ALD thin films grown at low temperature are amorphous, the post-deposition annealing for engineering the thin film crystallisation was used. The use of anodized aluminium oxide (AAO) membranes for further T1O2 nanowires fabrication also imposes the annealing temperature limit at 600°C.

[0088] The used ALD pulse/purge parameters are 0.2 s pulse /5s and 30 s nitrogen purge for both precursors and result in different thicknesses of 25, 50, 90 nm corresponding to 500, 1000, 2000 cycles. The TiCL films grown with the purge time 30s are much less crystallised than the films grown with 5 s purge time (figure 10).

[0089] The 25 nm and 50 nm samples after 500 and 1000 cycles respectively for both purge time show an enhanced crystallinity with the increase of the film thickness. Indeed, the typical crystallite size increases from 18 nm to 25 nm when the thickness is increased to 50 nm. The thickest sample grown with 30 s purge time exhibits a broadening of the (101) anatase peak, while any other diffraction plan are detected. The morphology of these particular samples is significantly smoother than the thinnest film.

[0090] The thick sample grown with long purge time (30 s) does not demonstrate any morphology change after annealing (figure 10). The estimated crystallite sizes reduce significantly for 90 nm thick samples. These overall results tend to demonstrate that the thermal budget of annealing is not enough to properly recrystallise samples thicker than 50 nm. For 50 nm samples, the calculated lattice parameter from the position of the diffraction peak indicates the compressive stress: for 30 s purge time a=b=3.7521 Â, 0=9.3937 À and for 5 s purge time a=b=3.7612 Â, c=9.3937 Â vs. reference values being: a=b=3.7845 À, c=9.514 Â.

[0091] The morphologies of 25 and 50 nm samples are very similar: they exhibit "hand-fan"-like domains (figure 11). These results are compared with the ones obtained for TiO2 films grown at 100°C.

[0092] The ALD growth at deposition temperature of 100°C with two purge times 5, 30 s results in films with thicknesses of 25, 50 and 120 nm (350, 700, 1650 cycles). The obtained amorphous samples were also annealed at 600°C in ambient air. No significant thickness variation is observed according to the purge time.

[0093] However, the purge time of the deposition process does noticeably impact the thin-films crystallinity after the post-deposition annealing (figure 12). All samples demonstrate the presence of anatase phase. For samples grown with a purge time of 5 s, the peak intensity increases monotonously with the sample thickness while samples grown with a purge time of 30 s show an anomalous crystallisation behaviour. Indeed, the thinnest film grown with 30 s purge time seems to be poorly crystallised after annealing at 600°C. At intermediate thickness of 50 nm in both depositions (5 and 30 s), the anatase peak intensity corresponding to the (101) and (200) orientations appear with similar intensity. While the peak corresponding to the (004) orientation was not detected in case of longer purge time deposition (30 s). The increase of the film thickness to 120nm makes this purge time influence even more remarkable. The film deposited with longer purge time forms smallest crystallite size (12 nm), while the deposition with (5 s) of the purge time leads to the formation of larger crystallites (38 nm). The calculated lattice parameters from the obtained XRD peak positions for short purge regime demonstrate a compressive stress compare to the TiO2 anatase reference (a=b=3.7660 Â, c=9.4460 Â vs. reference: a=b=3.7845 Â, c=9.514 Â). These results highlight that the recrystallisation behaviour is significantly dependent on the purge time used for the growth of the amorphous thin film. From secondary electronic microscopy (SEM) observations on figure 13, thin films deposited with 5 s purge time demonstrate a large grain morphology being significantly changed according to the film thickness. The 25 nm thin film shows an unusual “hand fan”-like crystallisation with contrasted stripes going from the centre of grain to the border of it.

[0094] Such morphology corresponds to the one that is formed by an explosive crystallisation (see Pore V., et at., Explosive crystallization in ALD mixed titanium oxides, Cryst. Growth. Des., 2009, 9(7), 2974-2978).

[0095] Such morphology is obtained when the ALD process is carried out at a temperature comprised between 50°C and 200°C, preferentially at a temperature comprised between 50°C and 100°C, followed by an annealing step carried out at temperature comprised between 450°C and 600°C, preferentially during 2 hours.

[0096] The explosive crystallisation of amorphous or liquid materials is an autocatalysed process that occurs when the applied heat induces a release of the latent heat that makes this crystallisation very fast as “explosive”. Such morphology has never been reported for TiO2 ALD films.

[0097] The electron back scatter diffraction (EBSD) characterisations are realised on similar samples (35 nm and 70 nm); both samples are from different deposition batches with 30 s purge time; the 35 nm sample is annealed at 500°C for 1 h while the 70 nm sample is annealed at 450°C for 2 h (figure 14).

[0098] The EBSD map and associated pole figures (pole figures are not presented) further confirm anatase grains with a main (001) orientation. It is known that anatase (001) facets have high surface energy. The EBSD analysis also demonstrates an important disorientation (<8°) within the same grain. This morphology type should present an important strain. Indeed, lattice parameters calculated from the XRD peak positions for this sample (data are not presented) are lower than reference values, which also indicates compressive stress (Aa=b=0.0234 Â, Ac=0.08 Â). Thus, crystallisation is likely to start from a nucleus and propagate laterally. EBSD band contrast images also demonstrate that the brighter stripes correspond to the better electron diffraction patterns, which indicate better crystallisation. It was also found that within the large “hand fan”-like domains, the crystallographic orientation changes and may include smaller grains. The increase of the thickness leads to a more homogeneous grain distribution, where the stripes are less marked. The crystal disorientation <5° is still present within the domain.

[0099] The non-ideal regime of ALD depositions at RT using 5 s and 30 s purge time results in conformal films. A thickness variation is found versus the number of cycles (160, 320 and 965) grown with two purge time. It is worth noticing that the samples thickness was reduced by 10-25% after annealing. Table 2 summarises the thickness values of as deposited and annealed films.

[00100] Table 2: RT TiCL film thicknesses as deposited and after annealing

[00101] The film crystallinity is analysed by XRD in the 0.5°grazing incidence configuration, which provides the information about all crystallographic

orientations of the film (figure 15). The obtained diffractograms evidence anatase phase in all samples annealed at 600°C for 2 h; the intensity of the detected peaks usually increases with the film thickness. According to these XRD data, we note a slight increase in the peak intensity for samples grown the shortest purge time (5 s). Taking into account the QCM and GPC data, which show a more important mass increment for the depositions with shorter purge time, such XRD intensity difference can be attributed to the small variation of the film thickness.

[00102] The SEM and AFM pictures show that the morphology of the samples undergoes a significant modification after annealing (figures 16 and 17). The 15 nm thin-films fabricated with purge time 30s with a low number of ALD cycles (160 cycles) do not show any particular change in the morphology compared to the amorphous one. When the film thickness is increased to 90 nm, the post-deposition annealing induces a porous structure. When the purge time is 5 s, the porous structure is formed even at low film thickness (~i.e. 15 nm). Such kind of structure is likely to be formed by degassing some of volatile compounds from the amorphous film since films are thermally shrunk during annealing. A priori, the longer purge time ensures better evacuation of reaction by-products during the growth process. However, the thick samples deposited with purge time 30 s are likely to also undergo a certain accumulation of by-products into the film, which is releasing during the annealing and leads to porous structure.

[00103] The by-products desorption upon annealing is further corroborated by the important mass increment per cycle at RT, indicating the lack of desorption of by-products and therefore their preferential accumulation into the film. The chemical characterisation of the films is achieved by performing an XPS analysis on the amorphous and annealed films. The XPS spectra are calibrated by placing the main Ti 2p3/2 peak at 458.8eV and applying a constant shift of the remaining peaks.

[00104] The O 1s peak of the amorphous samples is found to be asymmetric and a shift of 0.3 eV of the peak position is observed after the annealing (figure 18).

[00105] Such a shift of O 1s was also noticed by Park and coworkers on TiCte powders treated by HCI (see Park S. K., et at., Effect of HCI and H2SO4 treatment of TiO2 powder on the photosensitized degradation of aqueous rhodamine B under visible tight, J. Nanosci. Nanotechnot, 2014, 14(10), 8122-8128).

[00106] However, this shift, equivalent to 0.75 eV was found for both Ti 2p3/2 and O 1s peaks and was attributed to a charge effect due to the TiO2 protonation. Moreover the binding energy difference (ΔΕ) between Ti 2p3/2 and O 1s for bare TiCte and HCI-treated TiCte was the same at 71.25 eV. In the analyzed samples, this binding energy gap ΔΕ (Ti 2p3/2 ,O1s) is found at 71.5 eV and 71.2 eV on amorphous and annealed films, respectively. The Ti 2p3/2 peak on the surface demonstrates that in all samples, titanium seems to be in Ti4+ coordination.

[00107] In this study, SnO2 deposition is realized on anatase-TiO2 film grown by ALD at 100°C and annealed. 35 nm TiO2 film consists of large grains having different crystalline orientations. Therefore, the corresponding local surface energy varies according to this crystalline orientation (see table 3).

[00108] Table 3: TiO2 surface energy according to crystalline orientation

[00109] The SnO2 thin-film is very smooth when grown on certain grain orientations, its growth mimics the grain shapes of the TiO2 under-layer. This growth behaviour of SnO2 is mainly obtained over the whole TiO2 surface. However, on other few grains, having probably different crystalline

orientation, rough deposition of highly-dense SnO2 nanoparticles is obtained.

[00110] The Kelvin probe force microscope (KPFM) measurements realized on the TiO2 film confirm the surface potential variation between the formed grains (figure 19). Unfortunately the KPFM data have only indicative character and does not allow determining the exact crystalline orientation of each grain.

[00111] The EBSD characterisation of TiO2 buffer layer (figure 20) reveals the presence of important amount of high surface energy orientation as (001). According to the EBSD map, the percentage of this orientation was estimated as -47% of the whole map area. In the same way we estimate the percentage of smooth area of SnO2 film from the SEM image (figure 21) which is -50% of the image area. That supposes that (001) facets are responsible for the smooth SnO2 growth.

[00112] The (001) facets are usually difficult to obtain because of their high surface energy. This is mainly due to the atomic coordination of Ti and O. The (001) orientation is characterised by a high density of undercoordinated Ti atoms and the Ti-O-Ti bond angle is very large (figure 22). Such arrangement distabilises the 2p states of the surface oxygen atoms and makes them very reactive. Numerous theoretical and experimental studies demonstrate that the (001) promote the dissociative adsorption of various precursors (water, methanol etc.) while on the low energy orientation (101) these molecules are adsorbed without dissociation (see Selloni A., Crystal growth: anatase shows its reactive side, Nat. Mater., 2008, 7(8), 613-615).

[00113] The crystallographic structure of films deposited with and without TiO2 buffer layer is controlled by XRD. The diffractograms (figure 23) show SnO2 in cassiterite phases for both samples.

[00114] The optimisation of the SnO2 smooth layer deposition using TiO2 anatase buffer layer is applied in first in commercially available AnodiskTM porous membranes with 200 nm pores diameter. The 20 nm of TiO2 buffer layer is deposited by ALD at 200°C and annealed at 450°C in air and is followed by SnO2 ALD deposition (3000 cycles at 350°C). The SEM observations show the significant improvement of the SnCL thin-films morphology into the membrane when a TiCL layer is used (figure 24).

[00115] A homogeneous, continuous and conformal closely-packed SnCL thin-films is grown all along the membrane pores. This is a significant breakthrough in the field of growing continuous and conformal SnCL thin-film by ALD.

[00116] A similar deposition process is applied on membranes having 2 pm pores length and well distributed 180 nm pores diameters. Figure 25 shows SEM pictures highlighting the top-view pores architecture after each process step (as-received, after TiCL deposition, after SnCL deposition).

[00117] This finding allows the direct fabrication of TiCL/SnCL heterostructures using the template-assisted approach and opens new architectural engineering of complex heterostructured nanowires or nanotubes with a broad range of dimensions (size, length, spacing etc.). Similarly to the TiCL periodic nanowires fabrication, the heterostructured periodic nanowires are presented on figures 26 and 27.

[00118] The developed protocol is applicable not only with SnCL but with also any kind of component. The crystalline orientation of the layer of TiCL, with the Miller indices equivalent to (001), (110) or (111), confers an energy to the surface which is relatively high (superior to 0.60 J/m2, as indicated on table 3), and that allows the coating of layer of component (such as SnCL) in smooth manner. This favors the coating of such component on curved glass, porous substrate, porous curved glass and any other type of substrates presenting a complex surface.

[00119] Examples of substrates that are used are silicon (Si) substrates, silica (SiCL) substrates and/or silicon substrates with an oxidized layer of silica (Si/SiCL).

Claims (15)

A coating comprising at least one layer, said at least one layer being a TiCte layer, characterized in that said at least one TiCte layer has a crystalline orientation with Miller indices equivalent to (001), (110) or (111). 2. The coating of claim 1, characterized in that said at least one TiCte layer comprises a surface energy greater than 0.60 J / m2. 3. Coating according to any one of claims 1-2, characterized in that said at least one TiCte layer has a thickness between 20 nm and 200 nm, preferably equal to 35 nm. 4. A coating according to any one of claims 1-3, characterized in that the coating further comprises a second layer, said second layer being preferably a SnCte layer. A method for depositing TiCte on a substrate, comprising the steps of: a) providing a substrate; b) performing said deposition with TiCU and water by atomic layer deposition on said substrate; and c) performing an annealing step; wherein step (b) is carried out at a temperature between 50 ° C and 200 ° C, preferably at a temperature of between 50 ° C and 100 ° C, and wherein step (c) is performed at a temperature between 450 ° C and 600 ° C in the ambient air. The method of claim 5, wherein said substrate is curved glass. The method of claim 5 and / or claim 6, wherein said substrate is porous. The method according to any one of claims 5-7, wherein said step (b) is to draw TiCL on said substrate so as to form a functionalized substrate, to be purged with an inert gas, to draw H2O on said functionalized substrate. and purging with an inert gas, said inert gas preferably being nitrogen gas. The method of claim 8, wherein said purging with an inert gas lasts between 5 seconds and 30 seconds. The method of any one of claims 5-9, wherein step (c) is performed for at least 2 hours. The method of any one of claims 5-10, wherein step (b) is performed between 100 times and 1000 times. 12. A method of depositing at least one component on a substrate coated with at least one layer of T1O2, said process comprising the steps of: a ') carrying out the method according to any one of claims 5-11, b' ) performing a deposition of said at least one component on said substrate covered with at least one layer of TiC> 2; and c ') performing a post-depot treatment. The method of claim 12, wherein said component is SnC> 2 and wherein said step (b ') is the step of performing SnCL deposition and water by atomic layer deposition. The method of claim 13, wherein said atomic layer deposition is performed at a temperature between 200 ° C and 350 ° C. The method of any one of claims 13-14, wherein said step (b ') is performed 3000 times.