RESEARCH ARTICLE www.advancedscience.com Kinetic Control of Ultrafast Transient Liquid Assisted Growth of Solution-Derived YBa2Cu3O7-x Superconducting Films Silvia Rasi,* Albert Queraltó, Juri Banchewski, Lavinia Saltarelli, Diana Garcia, Adrià Pacheco, Kapil Gupta, Aiswarya Kethamkuzhi, Laia Soler, Julia Jareño, Susagna Ricart, Jordi Farjas, Pere Roura-Grabulosa, Cristian Mocuta, Xavier Obradors, and Teresa Puig* 1. Introduction Transient liquid assisted growth (TLAG) is an ultrafast non-equilibrium growth process mainly governed by kinetic parameters, which are only accessible The discovery of high temperature super- conductivity (HTS) in complex oxides such through fast in situ characterizations. In situ synchrotron X-ray diffraction as the cuprates[1] has enormously boosted (XRD) analysis and in situ electrical resistivity measurements are used to the interest in growing an incredible range derive kinetic diagrams of YBa2 Cu3 O7−x (YBCO) superconducting films of transition metal oxides displaying a wide prepared via TLAG and to reveal the unique peculiarities of the process. In span of remarkable functional properties particular, diagrams for the phase evolution and the YBCO growth rates have (magnetic, semiconducting, ferroelectric, been built for the two TLAG routes. It is shown that TLAG transient liquids multiferroicity, catalytic, ionic conductivity, etc.). A formidable progress has been made can be obtained upon the melting of two barium cuprate phases (and not just in creating novel techniques for epitaxial one), differentiated by their copper oxidation state. This knowledge serves as thin film growth characterized by differ- a guide to determine the processing conditions to reach high performance ent degrees of crystalline quality, composi- films at high growth rates. With proper control of these kinetic parameters, tional control, growth rate, etc.[2–4] Actually, films with critical current densities of 2–2.6 MA cm−2 at 77 K and growth rates the selection of the more adequate growth technique depends on the specific demands between 100–2000 nm s−1 are reached. These growth rates are 1.5–3 orders of each application. The discovery of HTS of magnitude higher than those of conventional methods. opened the possibility of developing many unique electrical, magnetic, and electronic applications.[5] For that reason, enormous efforts are being made to develop practical superconductors hav- S. Rasi, A. Queraltó, J. Banchewski, L. Saltarelli, D. Garcia, A. Pacheco, K. Gupta, A. Kethamkuzhi, L. Soler, J. Jareño, S. Ricart, X. Obradors, ing at the same time high performance and low production cost T. Puig to be able to meet the industrial demands.[6,7] Institut de Ciència de Materials de Barcelona In this context, HTS coated conductors (CCs), that is, HTS ICMAB-CSIC REBa2 Cu3 O7 (REBCO, “RE”: rare earth) epitaxial films grown Campus UAB, Bellaterra, Catalonia 08193, Spain E-mail: silviara@buffalo.edu; teresa@icmab.es on buffered metallic substrates, appear to be a very promising D. Garcia approach to fabricate kilometers long superconductors. Several Departament de Química growth techniques of REBCO films are pursuing these goals, for Universitat Autònoma de Barcelona instance: Pulsed Laser Deposition (PLD), Metalorganic Chemical Bellaterra, Catalonia 08193, Spain Vapor Deposition (MOCVD), e-beam evaporation (EV), Chemi- J. Farjas, P. Roura-Grabulosa cal Solution Deposition (CSD).[8] One of the parameters strongly GRMT promoting cost reduction is achieving high throughput manufac- Department of Physics Universitat de Girona turing, and this can be achieved through high growth rate meth- Campus Montilivi, Edif. PII, Girona, Catalonia E17003, Spain ods. In this work, we investigate the recently reported ultra-fast C. Mocuta Transient Liquid Assisted Growth (TLAG-CSD),[9–11] based on a Synchrotron SOLEIL liquid mediated growth process that, as we demonstrate, is able to L’Orme des Merisiers Saint-Aubin BP 48, Gif-sur-Yvette 91192, France obtain epitaxial films even at 2000 nm s−1 growth rates (which is The ORCID identification number(s) for the author(s) of this article between 1.5 and 3 orders of magnitude larger than conventional can be found under https://doi.org/10.1002/advs.202203834 methods). Several attempts have been made to use liquid-based © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH. mediated growth of YBCO[12–17] taking advantage of the high Y This is an open access article under the terms of the Creative Commons diffusivity in the liquid (typically two orders of magnitude larger Attribution License, which permits use, distribution and reproduction in than in solids). However, with the non-equilibrium TLAG pro- any medium, provided the original work is properly cited. cess, we obtain between 1.5 and 3 orders of magnitude larger DOI: 10.1002/advs.202203834 growth rates than in the methods reported so far. The novelty Adv. Sci. 2022, 9, 2203834 2203834 (1 of 12) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com with respect to other liquid-assisted techniques lies in the tran- other words, we need a thorough understanding of the kinetic sient nature of the barium cuprate liquids induced in TLAG, and factors governing TLAG-CSD, and of their influence in the nu- in the tunability of the liquid itself as well as of the Y supersatu- cleation and growth processes. ration through kinetic parameters. To this end, we used in situ synchrotron XRD analysis to de- In addition, TLAG-CSD benefits from being a cost-effective termine several kinetic phase diagrams, which show the coexis- CSD method.[9,18–20] The combination of ultra-fast growth tence of the different phases at specific temperature—PO2 con- rates and cost-effective CSD is extraordinary to achieve high ditions, depending on the heating ramp, liquid stoichiometry throughput fabrication of CCs at reduced costs for industrial and Cu oxidation state. Compared to our first report of TLAG,[9] production.[21–23] Moreover, among the CSD methodologies to here we study both the ternary (Ba–Cu–O) and the quaternary grow REBCO films, TLAG is based on fluorine-free precursors (Y–Ba–Cu–O) kinetic phase diagrams. The existence of the liq- which, up to now, have only achieved low growth rates.[12,24–26] uid is directly demonstrated by investigating the Ba–Cu–O sys- Compared to processes that use BaF2 as an intermediate com- tem via XRD under the same conditions as YBCO film growth. pound, TLAG has the further advantage of simplifying furnace Special attention is given to the main kinetic factors determin- designs since no toxic hydrofluoric acid is generated, so a large ing the BaCO3 reaction and subsequent phase evolutions. A rele- growth area can be easily implemented.[20,27–29] vant result of this analysis is that we experimentally generate, for To determine the conditions of ultra-high growth rates and to the first time, kinetic phase diagrams for both systems as com- transfer this technology to thick films and long-length commer- pared to the equilibrium phase diagram used in our first report cial metallic substrates, one needs to advance in the understand- of TLAG, and which has served as reference so far. Noticeably, we ing of the working principles of TLAG-CSD and of the transient demonstrate that the TLAG transient liquid can be reached upon liquid, which we tackle in this article with in situ X-ray diffraction the melting of two, and not just one, barium cuprate phases, (XRD) and resistivity probes. whose formation is governed by the copper oxidation state. For In TLAG-CSD, the transient liquid can be reached through two the first time, we show that by modifying the value of each kinetic main routes: the temperature route and the oxygen partial pres- parameter we can change the phase evolution (formation of crys- sure route, or T-route and PO2 -route, respectively.[9] In T-route, talline intermediate Ba–Cu–O phases) and change the growth the system is heated up at constant PO2 in the region of YBCO rate by several orders of magnitude. Therefore, we offer quan- stability.[30–32] As the BaCO3 reacts with CuO,[9,10] an intermedi- titative tools to control phase evolution and growth rate. In ad- ate BaCuO2 -CuO transient liquid[12,33] is formed and YBCO crys- dition, we used the ultra-fast acquisition of in situ XRD at syn- tallizes from the Y2 O3 dissolution in the liquid and Y diffusion chrotron facilities and of in situ electrical resistivity measure- to the substrate interface. Conversely, in the PO2 -route, the sys- ments to determine the growth rates in different regions of the ki- tem is heated up at lower PO2 where YBCO is not thermodynam- netic phase diagrams, identifying conditions for ultrafast epitax- ically stable, and where the BaCO3 elimination leads to an in- ial TLAG growth even at 2000 nm s−1 for the first time worldwide, termediate BaCu2 O2 solid phase.[30,34,35] Subsequently, a fast PO2 while obtaining high superconducting performances. These re- increase into the region where YBCO is stable leads to the trans- sults, which represent both a 20-fold increase in growth rate and formation of the BaCu2 O2 solid phase into the transient liquid, a fourfold increase in film thickness with comparable critical cur- through which again Y2 O3 is dissolved and YBCO crystallizes. rent densities compared to our first report of TLAG, together with Notice how all these phase transformations until REBCO growth the progress reported here in understanding the fundamental as- occur in a really short time (i.e., for a 400 nm film only 200 ms pects of TLAG that made them possible, help establish TLAG as are needed at 2000 nm s−1 ). In both routes, the key factor is the a consistent and reproducible growth methodology. kinetically favored formation of the transient liquid,[9] which re- quires fast heating rates in the T-route, as opposed to the direct 2. Results and Discussion reaction of the Ba–Cu–O solid phases (or vanishing amounts of liquid) with Y2 O3 to form YBCO and which results in a much Pyrolyzed films of Y2 O3 -BaCO3 -CuO and BaCO3 -CuO were ana- slower growth.[25,36] lyzed by in situ synchrotron XRD to identify the kinetics of the Each route presents its own peculiar advantages: the PO2 -route phase evolution and YBCO growth in the two TLAG routes (PO2 offers the possibility to decouple the crystallization of YBCO from and T-route). In situ XRD offers the unique opportunity to follow the BaCO3 reaction with the copper oxides (Cux O), which enables the ultra-fast phase evolution characteristic of YBCO films grown a separate optimization of each step. Conversely, the strength via TLAG, while the synchrotron radiation enables the detection of the T-route lies in the fact that the reaction between BaCO3 of these phase transformations at fast time scales (few tens of and Cux O can be assisted/accelerated by the concomitant for- milliseconds of acquisition times) even with a low amount of ma- mation of the Ba–Cu–O transient liquid. Therefore, TLAG is a terial, such as in the case of films. non-equilibrium process able to grow REBCO films governed by The Ba–Cu–O ternary system is discussed first for its relative the RE solubility and diffusivity in the Ba–Cu–O transient liq- simplicity (Section 2.1) while, relying on this understanding, the uids that up to now has demonstrated ultrafast growth rates of Y–Ba–Cu–O quaternary system is developed later (Section 2.2). 100 nm s−1 .[9] In this article, we show that much higher growth The specific phase diagrams here reported refer to the Cu-rich rates than previously published, of even 2000 nm s−1 , are possi- initial solution composition of YBa2 Cu4.66 Ox (i.e., 3BaO-7CuO, ble using TLAG. However, we need to control, first, the ability to 3–7 stoichiometry), unless otherwise stated. The main interest form the transient liquid in a region where REBCO is the stable in the copper-rich composition comes from its lower supersatu- thermodynamic phase, and, second, the instantaneous Y super- ration, which is able to widen the epitaxial growth window toward saturation conditions of the liquid to induce c-axis nucleation. In lower temperatures. Results on the YBa2 Cu3 Ox (i.e., 2BaO-3CuO, Adv. Sci. 2022, 9, 2203834 2203834 (2 of 12) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com ing of this reaction product above a certain temperature (non- equilibrium eutectic melting). In the first process, the precursors (BaCO3 and CuO) react to form a barium cuprate phase. This phase corresponds to BaCuO2 at the highest PO2 (2 × 10−1 bar), and to BaCu2 O2 at the lowest PO2 (2.5 × 10−4 bar), distinguished by the copper oxidation state (Cu+2 and Cu). At intermediate PO2 (10−3 bar), both the formation of BaCuO2 and its reduction to BaCu2 O2 at higher temperature are detected. In the second process, the disappearance of the barium cuprate and the copper oxide phases (Cux O) reveals that melt- ing of BaCuO2 -CuO[34] (at 0.2 bar) and BaCu2 O2 [35] (at 10−3 and 2.5 × 10−4 bar) has occurred in very short times (a few seconds). The melting temperature is defined as the temperature where the BaCuO2 or BaCu2 O2 XRD integrated signal decreases by 10% of its original value (Figure 1). Although the melting of both of these phases has already been comprehensively studied in equilibrium conditions for the Ba–Cu–O and Y–Ba–Cu–O sys- tem in powders,[30,35] here we intend to characterize the non- equilibrium transient liquid employed in the TLAG process. As a consequence, the transient liquid is able to originate from both the melting of BaCuO2 -CuO and BaCu2 O2 phases. It is known that the barium cuprate liquid originating from Cu(II) is (2BaCuO2 -CuO)L ;[35] while that originated from Cu(I) comes from the melting of BaCu2 O2 which can occur without an addi- tional uptake of copper.[30,35] However, in Figure 1, we show that some Cu2 O can also be incorporated in the Ba–Cu(I)–O based liq- Figure 1. Phase evolution for Ba–Cu–O films of (3–7) composition at three uid (BaCu2 O2 +xCu2 O)L . In particular in this case, a liquid with different oxygen partial pressures (PO2 = 2 × 10−1 bar, 10−3 bar and 2.5 × 10−4 bar) and heating ramp of 4.5 °C s−1 . For enhanced visibility, curves the stoichiometry (3BaCu2 O2 +½Cu2 O)L is formed. However, the are normalized to their maximum peak area value. The black arrow indi- study of the maximum amount of Cu2 O that can be incorporated cates the non-equilibrium melting temperature of the Ba–Cu–O phases in the BaCu2 O2 liquid and its influence on the TLAG process according to the defined criterium. would require further studies. 2–3 stoichiometry) are rather similar as shown in the Support- ing Information. In Section 2.3 the growth rates associated to the 2.1.2. Ba-Cu-O Kinetic Phase Diagram quaternary Y–Ba–Cu–O kinetic phase diagram are described and correlated with those obtained from in situ resistivity results. The time-resolved phases of Figure 1 are represented in the T– PO2 kinetic phase diagram of Figure 2. We have identified regions with the coexistence of different phases. The symbols (closed cir- 2.1. Ba-Cu-O Kinetic Phase Diagrams: Understanding the cles) at the three different PO2 values, connected by dotted lines, Formation of Transient Liquids correspond to the crystalline reactions observed during the XRD experiments of this work and are shown in the legend of Figure 2; 2.1.1. Ba-Cu-O Phase Evolution the solid lines represent the main equilibrium processes from literature experiments performed on mechanically mixed pow- Although some understanding already exists on the BaCuO2 - ders of the same precursors (BaCO3 /BaO, CuO, Y2 O3 ): eutectic CuO thermodynamic phase diagram under equilibrium condi- melting of BaCuO2 -CuO (TE ), CuO-Cu2 O reduction (a), BaCuO2 - tions and certain values of PO2 ,[34,35] here we intend to under- BaCu2 O2 reduction (b), and BaCu2 O2 melting (TM ). stand the kinetic reaction of the BaCO3 -CuO phases in non- The solid green area represents the region of the phase dia- equilibrium conditions, and for a wide T–PO2 range, where YBCO gram where only solid phases coexist. It covers the initial state of TLAG films can be grown. the system after pyrolysis, consisting of BaCO3 and the yttrium The time and temperature resolved phase evolution of BaCO3 - and copper oxides, their recrystallization and coarsening, until CuO films are reported in Figure 1, for the conditions of the T- their reaction starts forming a liquid. route at a fixed heating rate (4.5 °C s−1 ) and three different PO2 Region Where Liquid Ba-Cu-O Coexists with Solid Phases: The values: 2 × 10−1 bar, 1 × 10−3 bar and 2.5 × 10−4 bar. The ap- striped areas correspond to regions where the Ba–Cu–O liquid plied heating ramp (being the maximum available with the XRD can be formed and coexist with solid phases. It starts where the synchrotron furnace employed) is much faster than the equilib- BaCO3 starts to be eliminated (10% has decomposed) and ends rium conditions ever investigated. However, it is well below the at the full melting (90% has melted) of the barium cuprates, as 20–30 °C s−1 employed in the laboratory studies.[9] detected by in situ XRD. They represent regions where the oc- Two processes are observed for the Ba–Cu–O system in T- currence of kinetically-driven processes in the quaternary system route: 1) the reaction of BaCO3 with Cux O; and 2) the melt- leads to the coexistence of solid and liquid phases facilitating the Adv. Sci. 2022, 9, 2203834 2203834 (3 of 12) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com (Figure 2). As it is shown later, this full-liquid region is of primary importance for the ultra-fast growth of TLAG films. In several binary equilibrium phase diagrams (Pb/Al, Al/In, Cu/Pb, etc.), at certain conditions of composition and tempera- ture (and PO2 for non-metallic systems such as Cr/Cr2 O3 ) two im- miscible liquids have been reported.[38] However, there is so far no indication that our two ternary melts are immiscible.[35] Con- sequently, in the absence of any evidence, we depict one single liquid region with a continuous Cu-valence variation in TLAG ki- netic phase diagrams, rather than a miscibility gap. Correspond- ingly, we might also expect the properties of the liquid to change with the progressive change of its oxidation state, which should probably be reflected in the TLAG YBCO properties as well. 2.2. Y-Ba-Cu-O Kinetic Phase Diagrams: the Routes to YBCO TLAG With this understanding on the ternary Ba–Cu–O system, and with the aim of defining the most interesting growth regions for both PO2 -route and T-route, T–PO2 kinetic phase diagrams are de- fined from the in situ XRD experiments performed on BaCO3 - Figure 2. Corresponding kinetic phase diagrams based on the experi- Y2 O3 -CuO pyrolyzed films. ments shown in Figure 1, where the phases coexisting in each region are labelled based on the reactions indicated in the legend. Solid lines refer to equilibrium reactions from literature a) CuO(s) -Cu2 O(s) ;[37] b) 2.2.1. T-Route BaCuO2(s) -BaCu2 O2(s) ;[35] TE : BaCuO2(s) +CuO(s) – [BaCuO2 +CuO](l) ;[35] Tm : BaCu2 O2(s) – [BaCu2 O2 ](l) .[30] The different regions and symbols are Since the main kinetic parameter for the T-route is the heating explained in the text. rate, two kinetic phase diagrams are constructed, each one for a different heating rate (0.4 and 4.5 °C s−1 in Figure 3). The results fast TLAG YBCO growth. The idea behind these striped (solid illustrate the relevant effect of this parameter. + liquid) regions hinges upon the fact that the transient liquid The kinetic diagrams of Figure 3 identify the phases evolution can appear due to the out-of-equilibrium conditions imposed by of quaternary Y–Ba–Cu–O obtained from experiments at two rep- the high heating ramps of the TLAG system. In fact, as we will resentative PO2 values for two different heating ramps. The exper- show in the Y–Ba–Cu–O phase diagrams of Section 2.3, along imental points are plotted on the background colored diagram of this region YBCO grows at rates of 5–30 nm s−1 under heating the ternary kinetic phase diagram of Figure 2. The phase evolu- ramps of 4.5 °C s−1 , which is the highest heating rate achievable tion of these four experiments is represented by solid symbols during in situ XRD experiments, but about five times below our connected with dotted lines. The white region of the figure de- optimal laboratory heating rates of 20–30 °C s−1 . Additionally, the notes the area where YBCO is not stable, either because peritec- homogeneous distribution of small nanocrystalline precursors of tic melting occurs (TP ) or because decomposition to other phases the pyrolyzed film obtained from solution deposition contributes takes place upon crossing the instability line (Tins ). The region to speed up the reactions. This can be helped by the presence surrounded by the brown dotted-dashed line is where we have of some of the precursor phases in amorphous state (see, for seen the crystallization of YBCO, which may coexist with all the example, in Figure S1, Supporting Information, the increase in other labelled phases. The extension of the diagram to PO2 higher the peak intensity of BaCO3 sometimes observed before its de- than the 2 × 10−3 bar (region without solid symbols) is based on crease, which indicates that some BaCO3 , initially in the amor- the knowledge acquired from the Ba–Cu–O system (Figure 2), phous state, crystallizes). Notice that the non-equilibrium tran- although further confirmation would be required in the low tem- sitions occurring in TLAG are strongly shifted from the equilib- perature region. The solid lines are the equilibrium transitions rium solid lines from literature,[30,35] indicating the kinetic power also shown in Figure 2 and used as guidelines. The transition of the process. from the solid + liquid regions to the full liquid is color graded Full-Liquid Region: Beyond the solid + liquid regions of Fig- to indicate that the exact position is not identified. ure 2, Ba–Cu–O and Cux O solid phases have disappeared indi- If we compare the quaternary Y–Ba–Cu–O with the ternary cating that only liquid phases exist. Within this region, the liquid Ba–Cu–O system at the same heating rate, (4.5 °C s−1 , Figure 3) composition in terms of Ba–Cu–O stoichiometry will vary to ac- similar regions of phase coexistence are identified, although commodate for the variations in the Cu oxidation state observed some relevant differences appear. In the quaternary system, the in the delimiting regions. We believe this is a continuous varia- intermediate BaCuO2 and BaCu2 O2 crystalline phases are not al- tion from high PO2 (Ba–Cu–O liquid with a Cu(II) oxidation state) ways observed in those striped regions of coexistence of solid and to low PO2 (Ba–Cu–O melt with Cu(I) oxidation state), thus the liquid barium cuprate. Thus, BaCO3 elimination is immediately full liquid region is color graded in between those two extremes followed by YBCO crystallization. Conversely, in the Ba–Cu–O Adv. Sci. 2022, 9, 2203834 2203834 (4 of 12) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com Figure 4. XRD phase evolution (integrated signal of the respective char- acteristic XRD peak) as a function of time and temperature at a fixed PO2 (10−3 bar) for the Y–Ba–Cu–O system of 2Ba-3Cu liquid composition, comparing three values of heating rate. Each signal is normalized to its maximum value for better visibility. The dashed, red x-axis at the bottom corresponds to temperature for all three cases, while the continuous black x-axes correspond to the individual times. Inset: not-normalized phase in- Figure 3. T-route phase diagrams for the Y–Ba–Cu–O system with Ba/Cu tensity showing that the YBCO (005) contribution is majoritarian. = 3/7 ratio, measured at two different heating rates: 0.4 (upper panel) and 4.5 °C s−1 (lower panel). Solid lines from literature refer to a) CuO(s) -Cu2 O(s) ;[37] b) BaCuO2(s) -BaCu2 O2(s) ;[35] TE : BaCuO2(s) +CuO(s) – [BaCuO2 +CuO](l) ;[35] Tm : BaCu2 O2(s) – [BaCu2 O2 ](l) ;[30] TP : YBCO peri- s−1 , where YBCO crystallizes in few seconds without the forma- tectic melting, YBCO – Y2 BaCuO5 + L;[31] Tins : YBCO instability line.[30] tion of the BaCuO2 (s) intermediate. This is further illustrated in The corresponding XRD phase evolution can be found in Figure S2, Sup- Figure 4, where the time and temperature phase evolution for a porting Information. 2BaO-3CuO liquid stoichiometry is considered. Notice that, at the slower heating ramp, crystallization of the BaCu2 O2 phase is ob- served, but it progressively disappears at higher heating ramps, system, both BaCuO2 and BaCu2 O2 are detected in these regions until at 4.5 °C s−1 this phase is not detected anymore, again con- by XRD. firming that its crystallization is not mandatory for the YBCO This is further understood with the kinetic study of the quater- growth. Therefore, we conclude that the heating rate is a very nary phase evolution at different heating rates (Figure 3, upper relevant kinetic parameter in the TLAG T-route. These results versus lower panels). For instance, if we fix PO2 at 1 × 10−3 bar, also demonstrate that crystalline BaCuO2 or BaCu2 O2 are not BaCO3 elimination time (region between the two dark green tri- required intermediate phases in TLAG reaction paths, because angles) is reduced from ≈100 s to just 10 s; similarly, the region amorphous or nanocrystalline Ba–Cu–O phases are enough to of Ba–Cu–O(s+l) after BaCO3 elimination is reduced from ≈200 form the transient liquid. s to almost zero to enter directly into the full liquid region. In Remarkably, YBCO nucleation is controlled through the heat- other words, the system at slower heating rates (0.4 °C s−1 ) has ing rate (Figure 4). At lower ramps, homogeneous nucleation time to crystallize some intermediate BaCuO2 contrarily to 4.5 °C (random nuclei) is favored, whereas by increasing the heating Adv. Sci. 2022, 9, 2203834 2203834 (5 of 12) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com ramp, the region where c-axis nucleation is not favorable is crossed much faster, thus avoiding random nucleation and pro- moting heterogeneous nucleation at higher temperatures. For in- stance, at 4.5 °C s−1 (lowest panel of Figure 4), very few random nuclei are formed and most of the YBCO phase nucleates epitax- ially (YBCO (005)), in this case already at the full liquid region. This can be seen in the inset of Figure 4, where the raw inten- sity has not been normalized and the epitaxial YBCO phase is majoritarian at the end of the heating ramp. In other words, su- persaturation is indirectly modified by the heating ramp. Thus, higher heating rates are also responsible for more epitaxial nucle- ation. In the same inset of Figure 4, it is also shown that reorien- tation of the few YBCO (103) grains formed at low temperatures takes place at a higher temperature, where the region of full liq- uid is reached. This confirms that the liquid is able to reorient the YBCO random nuclei to epitaxial YBCO (005), as previously suggested.[9] 2.2.2. PO2 -Route The kinetic phase diagram for TLAG PO2 -route (Figure 5) is also developed from in situ phase evolution (integrated intensity of the corresponding XRD peaks). The lower panel represents the heating process at a constant PO2 = 10−5 bar, while the upper panel shows the region of YBCO growth upon the PO2 jump at a constant temperature. They are to be read following the arrows, which represent the direction of change in the kinetic parameter. The phase evolution of the in situ XRD experiments, at different PO2 values and temperatures, is represented by solid symbols, while the solid lines again represent the equilibrium transforma- tions, as in Figures 2 and 3. Since the PO2 -route consists of two steps, we will focus on a different kinetic parameter for each one. At the first stage, when heating the sample at a fixed PO2 = 10−5 bar, the main kinetic pa- rameter is the heating rate (Figure 5, bottom panel). Notice that the CuO to Cu2 O reduction is strongly shifted to higher tem- peratures as compared to its equilibrium position (blue line of Figure 5, solid and empty green hexagons), indicating the strong contribution of kinetics in all this process. In this region of low PO2 , we notice that the reaction of BaCO3 with the copper ox- ides is delayed to higher temperatures as compared to the T-route case. It is important to note that, depending on the furnace used, a different compromise between temperature, heating ramp, gas flow and dwell times is required to ensure the completion of this reaction before the PO2 jump. This BaCO3 -Cux O reaction is Figure 5. PO2 -route kinetic phase diagrams for the Y–Ba–Cu–O system shifted to high temperatures (850 °C, light brown pentagons) es- with a 3BaO-7CuO liquid composition. The different regions of the phase pecially in the case of small, conduction-type furnaces like the diagram are shown in different colors, the striped regions indicate co- existence of solid and liquid phases. The symbols indicate actual exper- one used in this in situ XRD study, but the reaction is much iments and the color meaning is specified in the legend. The first step, faster for the tubular radiation furnaces. Despite that, the system at low PO2 (<10−5 bar), is shown in the lower panel for two different is converted to BaCu2 O2 (s) and Cu2 O(s) for both heating ramps. heating ramps. The second step, upon the PO2 -jump, is displayed in In this step, the microstructure and grain size of the intermediate the upper panel. The black arrows indicate the direction of the kinetic phases will depend on this compromise: a good balance between parameter. Solid lines from literature refer to a) CuO(s) -Cu2 O(s) ;[37] b) small coarsening and BaCO3 elimination is needed to speed up BaCuO2(s) -BaCu2 O2(s) ;[35] TE : BaCuO2(s) +CuO(s) – [BaCuO2 +CuO](l) ;[35] Tm : BaCu2 O2(s) – [BaCu2 O2 ](l) ;[30] TP : YBCO peritectic melting, YBCO – the transient liquid formation and YBCO growth rate after the Y2 BaCuO5 + L;[31] Tins : YBCO instability line.[30] PO2 jump. Finally, also in this region the system deviates from the equilibrium phases expected from previously reported ther- modynamic YBCO phase diagrams: neither BaCu2 O2 melting is detected at this low PO2 , nor Y2 BaCuO5 (211 phase) or YBa3 Cu2 Oy Adv. Sci. 2022, 9, 2203834 2203834 (6 of 12) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com Figure 6. Phase evolution of two XRD experiments used to build the dia- gram of Figure 5 at 820 °C, showing two different situations upon the PO2 jump: immediate BaCu2 O2 conversion to YBCO in the full liquid region (upper panel), and (lower panel) the BaCu2 O2 oxidation to BaCuO2 prior to conversion to YBCO in the liquid + solid region. Figure 7. (Upper panel) growth rate phase diagram for T-route, with the instantaneous YBCO growth rate (Ginst ) superimposed on the T-route (132 phase) are observed.[30] This deviation is the consequence of phase diagram, and (bottom panel) maximum values, Gmax inst of YBCO as a function of the heating rate for two PO2 values. Ginst values are determined heating the initial precursor phases very fast. from in situ XRD. The kinetic phase diagram upon a fast (<500 ms) PO2 jump is shown in the upper panel of Figure 5, and the XRD phase evolution of two specific experiments used to build this diagram 2.3. Growth Rate and Kinetic Control of TLAG-YBCO are shown in Figure 6. Similar regions as in T-route are iden- tified based on the phases observed after the PO2 jump, which The interest of studying the kinetic phase diagrams lies in the op- are labelled in the legend of Figure 5: a region where no YBCO portunity to control the nucleation and growth steps of the TLAG is formed upon a small PO2 jump, corresponding to the light process, and especially to identify the best conditions for the ul- brown pentagons of BaCu2 O2 ; a region of BaCu2 O2 oxidation to trafast growth rate of c-axis oriented REBCO films. In this section, BaCuO2 with little to negligible YBCO formation, at low temper- we will analyze the correlations of the YBCO growth rates in the atures (gray region); and a region with the possibility of coexis- different regions of the kinetic phase diagrams. tence of solid and liquid barium cuprate at high temperatures The kinetic study being our main aim, we will compile the in- (empty triangles and red triangles). This area, in PO2 -route, is de- stantaneous growth rates, Ginst , calculated from the time deriva- fined by the phase evolution observed upon the pressure jump: tive of the in situ YBCO (005) reflection integrated signal (ex- if some crystalline BaCuO2 intermediate is observed (Figure 6, amples in Figure S6, Supporting Information). Results of Ginst , bottom panel), or if BaCu2 O2 solid phase elimination is slowed expressed in nanometers per second, are plotted on the kinetic down because there is little or no liquid produced, the system phase diagrams of Figure 7, upper panel, for the T-route. Notice is in a solid and liquid region. Conversely, in the color-graded that Ginst increases until it reaches a maximum value and then it regions with blue triangles (Figure 5), crystalline intermediate decreases when approaching full YBCO conversion. It is also no- BaCuO2 is never observed upon the PO2 jump, and BaCu2 O2 dis- ticeable that the largest values are achieved at higher PO2 , where appearance is virtually instantaneous. We attribute it to full liquid YBCO is still stable at higher temperatures. In the case of the formation (or with some residual Cux O depending on the ini- PO2 -route (Figure 8), Ginst is determined from the time deriva- tial solution composition). This difference is clearly revealed by tive of the inverse of the in situ electrical resistance curve. No- the XRD phase evolution plotted in Figure 6: in the upper panel, tice, that the inverse of the in situ resistance increases over time where the jump happened in the full liquid region (blue sym- during the pressure jump, until it reaches saturation when the bols), YBCO crystallization can occur from hundreds of millisec- film is fully grown (see Figure S7, Supporting Information). In onds to seconds, while in the lower panel, where the jump hap- situ resistivity is a very useful technique to follow the growth of pened in the gray region where BaCuO2 crystallized, the growth the superconducting YBCO phase given the large difference in of YBCO takes tens of seconds. More details on the criteria used resistivity between the precursor phases (insulating) and YBCO to define these regions, as well as the raw XRD scans for both (rather metallic at growth conditions).[39,40] This technique was T-route and PO2 -route, are reported in Figures S4 and S5, Sup- previously used by several groups to study nucleation and growth porting Information). of TFA-YBCO films.[40–42] In this work, it allowed us to monitor Adv. Sci. 2022, 9, 2203834 2203834 (7 of 12) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com orders of magnitude, reaching 100 nm s−1 at 4.5 °C s−1 (Figure 7). Therefore, we conclude that previous YBCO growth studies using fluorine-free precursors but applying small heating rates, were very likely only achieving small growth rates,[12,24–26,36] in agree- ment with our trend. Additionally, it suggests that, at our labora- tory heating rates of 20–30 °C s−1 , the growth rate could be much higher, possibly reaching 1000 nm s−1 . In PO2 -route, the intrinsic kinetic parameter is the fast oxygen pressure increase which brings the system to cross the YBCO in- stability line from the BaCu2 O2 solid phase. The PO2 jumps asso- ciated to the points in Figure 8 (at a fixed temperature) range from 100 to just 20 ms. Consequently, in PO2 -route, ultrafast growth rates are associated to fast PO2 jumps to the region of full liq- uid. Here, very small changes in the final PO2 (by just doubling or tripling the PO2 value at a fixed temperature, Figure 8, bottom panel) can result in one order of magnitude differences in growth rates. The importance of acknowledging these kinetic parameters is such that much faster growth rates than previously claimed are actually demonstrated in TLAG when the kinetic state of the system is properly tuned. In fact, YBCO instantaneous growth rates as high as 2000 nm s−1 have been demonstrated with in situ resistivity measurements in the PO2 -route. We should remind here that similar YBCO growth rates have been already previ- ously demonstrated when growing melt textured YBCO ceramics based on similar Ba–Cu–O liquids.[43,44] Figure 8. (Upper panel) growth rate phase diagram for PO2 -route, with the instantaneous YBCO growth rate (Gmax ) superimposed on the PO2 -route To confirm these ultra-fast growth processes observed with an inst phase diagram, and (bottom panel) Gmax of YBCO as a function of the indirect method, a sample grown at ≈1000 nm s−1 in PO2 -route, inst PO2 , at a fixed temperature of 840 °C. Ginst values are determined from in was quenched at the very same moment in which the pressure situ electrical resistivity. jump was performed. This was achieved by cooling down the sys- tem at 100 °C s−1 (for the first 10 s). The TEM and XRD analysis of this sample confirmed that the YBCO layer was fully epitaxi- the YBCO growth kinetics in our in-house optimized furnaces, ally grown (Figure S8, Supporting Information), indicating that without the need to use in situ XRD radiation at a synchrotron the in situ resistance measurement is indeed representative of facility. From this point of view, it is a fast tool to complement the the YBCO growth process. information retrieved from in situ XRD analysis. Finally, results from fully grown and oxygenated films at For the T-route, Ginst values, obtained from in situ synchrotron growth rate conditions in the range 100–2000 nm s−1 (shown XRD, are plotted in Figure 7 as a function of temperature, in Figure 9), confirmed that epitaxial and high performance and its corresponding time scale. The maximum value, Gmax inst , YBCO films are obtained at these high growth rates. This con- corresponds to the maximum growth rate of the experiment firms the outstanding opportunities of TLAG and the interest for which is also plotted in Figure 7, lower panel, as a function Coated Conductors manufacturing. In particular, the XRD anal- of the heating rate. For the PO2 -route case, Ginst values are ob- ysis showed that epitaxial films are achieved in this PO2 range tained from the inverse of the electrical resistance curves ob- (see Figure 9b,c for two cases), which are also highly textured in- tained during growth, and the maximum values Gmax inst are plot- plane (full width half maximum Δϕ = 0.61, not shown here) and ted in Figure 8, lower panel, as a function of the final PO2 of the out-of-plane (Δ𝜔 = 0.18°, shown in Figure S9, Supporting Infor- experiment. mation). Additionally, a sharp and high superconducting tran- In both routes, the highest heating rates (yellow symbols in sition temperature and high critical current densities were ob- the upper panels of Figures 7 and 8) are achieved in the full liq- tained (see Figure S9, Supporting Information, for a particular uid region of the Ba–Cu–O phase diagram, while more moder- case with Tc = 91.7 K and ΔT = 5 K, and critical current densi- ate, though still very high growth rates, are achieved in the re- ties of Jc (77 K) = 2.6 and Jc (5 K) = 23 MA cm−2 ). In this case, a gions were both liquid and solid barium cuprate phases coexist, charge carrier density, nH (100 K) = 3.61 × 1021 cm−3 , was mea- confirming again the idea that enough liquid is formed in these sured which is representative of an optimally doped film.[45] Elec- regions to enable TLAG growth. tron microscopy results (HR-STEM and scanning electron mi- In T-route (Figure 7, lower panel), the growth rate determined croscopy [SEM] images) of this representative sample, shown in from the intensity of the YBCO (005) diffraction peak strongly Figure 9d and Figure S9a, Supporting Information, respectively, increases with increasing heating rate. This is understood as a confirmed a smooth surface below the CuO surface segregated consequence of the fact that the system is driven further into the precipitates (which do not disrupt current percolation), a high full liquid region. For example, by increasing the heating ramp by quality epitaxial growth and very low porosity, all being charac- one order of magnitude, the growth rate can be increased by two teristic features of liquid assisted growth. Adv. Sci. 2022, 9, 2203834 2203834 (8 of 12) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com Figure 9. a) Jc values for a set of 400-nm YBCO samples grown via TLAG in the range 100–2000 nm s−1 at different conditions of final PO2 , showing the high performance of the process; b,c) Ex situ 2D-XRD of two of the samples shown in (a), specifically for the conditions of PO2 = 10−3 bar and PO2 = 2.8 × 10−3 bar; d) Cross-sectional HR-STEM image of c-YBCO/STO interface. More characterization is shown in Figure S9, Supporting Information. 3. Discussion cess. Second, high supersaturation conditions, usually ascribed to high nucleation densities, are expected to induce high growth TLAG is an innovative growth process that takes place in condi- rates.[46,47] In liquid assisted processes to grow YBCO (like liq- tions very far from equilibrium, contrary to many other materials uid phase epitaxy and derived techniques[14–16] ), supersaturation processing techniques. To be able to control such a kinetically- is determined by the amount of Y dissolved in the liquid, given driven process, it is crucial to identify the nature of the liquid by Equation (1): phase and understand the effects of the main kinetic parame- ters that bring the system out of equilibrium. Additionally, since C𝛿 TLAG is defined by the formation of a transient liquid, it is Δ𝜇 = kTln = kTln (𝜎 + 1) (1) Ce mandatory to characterize this melt across the wide T–PO2 region where it forms. where Ce is the equilibrium concentration of Y in the liquid, The transient liquid of TLAG is a Ba–Cu–O melt whose ini- C𝛿 is the actual concentration of Y in the liquid, 𝜎 is the rela- tial composition is determined by the initial solution composition tive supersaturation and Δ𝜇 is the chemical potential change. (3–7 and 2–3 for this article), but whose oxidation state varies de- However, being TLAG a non-equilibrium process, Y dissolution pending on the T—PO2 region. Thus, the liquid phase is able to in the melt is strongly influenced by the kinetic parameters change in a continuous way from [BaCuO2 + CuO] (l) to BaCu2 O2 like heating rate (in T-route) or PO2 -jump speed (in PO2 -route). (l), with the Cu oxidation state continuously going from Cu(II) to For instance, high heating ramps may shift Y dissolution to Cu(I). We realized that all these melts accept large quantities of higher temperatures, thus shifting the high growth rates to copper oxides (Cux O), enabling to obtain a full liquid in the wide higher temperatures. Third, the growth rate should also be in- range of precursor composition employed. Therefore, we envis- fluenced by the liquid properties like viscosity and Y diffusivity. age a large number of factors influencing the ultrafast growth of Fourth, and perhaps with a major role, the continuous change epitaxial TLAG films. First, the amount of liquid generated at the of the liquid composition through the T–PO2 region (i.e., its Cu growth conditions (T and PO2 ) is seen to have a crucial role, since oxidation state) will accommodate changes in the liquid prop- all the highest growth rate values are reached in the full liquid erties (viscosity, Y-dissolution, and diffusivity) but also different region; thus, solid phases probably slow down the growth pro- intrinsic capabilities for nucleation (being different liquids). Adv. Sci. 2022, 9, 2203834 2203834 (9 of 12) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com Thus, one cannot exclude that a Cu(II) liquid may also induce netically driven growth, we have identified some robust kinetic a higher growth rate than the Cu(I) liquid, as Figures 7 and 8 parameters (heating rate and PO2 jump). These kinetic param- would suggest. Additionally, these kinetic phase diagrams are eters, together with the equilibrium thermodynamic parameters built starting from a particular solution composition and film (T and PO2 ) and the initial solution composition, are combined to thickness. These initial conditions will affect film composition, achieve the TLAG unique performance. We have proposed spe- porosity and crystallinity after pyrolysis and during the TLAG cific rules to understand the underlying mechanism guiding the process, so that the exact value of T and PO2 limiting these achievement of ultrafast growth rates, and we have shown that kinetic regions might shift when these parameters are changed. the growth rate can be modified by more than 3 orders of magni- We also think that the particular combination of substrate and tude depending on the value of the selected processing parame- YBCO precursors can play a role on the exact phase evolution ters (PO2 , heating rate). In conclusion, we have demonstrated that of intermediate phases, due to the different interface and strain by proper tuning of all these parameters, ultra-fast growth rates of energies for islands nucleation, which modify the nucleation even 2000 nm s−1 (20 times higher than previously demonstrated barriers. Therefore, in the interest of defining the fastest regions with TLAG and 1.5–3 orders of magnitude higher than other for growth for a given substrate, one has to take into considera- growth methods) are reached while growing epitaxial and high- tion the amount of liquid formed, the kinetics employed to reach performance YBCO films of 0.4–0.5 μm. Future work should the specific region of the T–PO2 diagram, the kinetic properties show that the novel nature of this non-equilibrium, liquid as- of the specific liquid and the initial solution composition, which sisted, and thin film growth process can be extended to other are, all in all, connected in a rather complex manner since some functional materials. processing parameters may influence several of these factors. Finally, there are two additional aspects to be considered. On 5. Experimental Section the one hand, the parameters chosen should be compatible with epitaxial growth. For example, in T-route, faster ramps lead to less Sample Preparation: The precursors’ chemical solutions were pre- crystalline intermediate phases and higher epitaxial nucleation. pared starting from acetate salts of yttrium, copper and barium, according to Refs. [9, 10]. in a 1:1 mixture of propionic acid and anhydrous methanol In the case of the PO2 -route, given the fact that two separate steps with a total metal concentration of 1.5 m. Two types of solutions were pre- take place, it is crucial to control both: the first stage to ensure pared, with and without the addition of Y salt to the barium and copper BaCO3 elimination while avoiding phase coarsening, and the sec- mixture. The Ba–Cu ratio was either YBa2 Cu4.66 O7−x (for the 3–7 com- ond stage to ensure ultra-fast growth by tuning the PO2 jumps. position) or YBa2 Cu3 O7−x (for the 2–3 composition). In some cases, a On the other hand, the desirable conditions should be compati- pure propionate-based solution with an amine additive was used, which ble with the substrates chosen. It is known that these Ba–Cu–O enabled the authors to reach a higher thickness with nanoscale homogene- liquids are very reactive and corrosive[48] especially at high tem- ity and with high superconducting properties. Additional details about this solution preparation and amine composition cannot be disclosed due to peratures, but also the substrate texture quality should not be af- confidentiality. The solution was spin coated on 5 × 5mm2 single-crystal fected by the TLAG growth conditions (especially for the growth SrTiO3 (STO) substrates in a N2 glove box (humidity below 8%) and dried of Coated Conductors on metallic substrate architectures). De- at 70 °C for a few minutes. For the specific cases of Ba–Cu solutions, MgO spite the complicated kinetic nature of TLAG, we believe that the substrates were used to avoid liquid reactivity with Ti from STO. The sam- above considerations and the kinetic phase diagrams drawn con- ples were pyrolyzed in humid O2 [9,10] with ramps of 3–5 °C min−1 to 500 °C stitute a guideline toward fast TLAG growth rates, and that in and dwells of 5 min. The samples consist of 2 layers achieved through multi-deposition, to reach several targeted YBCO thicknesses (from 350 situ XRD and in situ resistivity measurements are fundamental to 500 nm), although thicknesses in the range of 1 μm have already been tools to advance in TLAG understanding and transfer to Coated demonstrated by increasing the number of multi-depositions.[49] Grown Conductors manufacturing. Moreover, we believe that with more films were oxygenated at 450 °C in a dry O2 atmosphere during 210 min. targeted efforts, ultrafast growth rates above 2000 nm s−1 could In Situ Analysis X-Ray Diffraction: In situ XRD was performed at the Dif- be demonstrated, since no intrinsic limits have been found yet. fAbs beamline of Soleil Synchrotron (Paris, France) with a 6 circle diffrac- tometer at a beam energy of 18 keV and equipped with an XPAD 2D hybrid pixel area detector,[50,51] which here allowed the authors to obtain fast ac- 4. Conclusion quisitions (100 ms/frame) and texture evaluation.[9,52] The detector was placed at a distance that enabled the visualization of all the phases of inter- For both T-route and PO2 -route of YBCO TLAG-CSD, kinetic est in a single 2D image (i.e., BaCO3 (111; 021), CuO (−111; 111), Cu2 O phase diagrams were constructed identifying the phase evolution (111), YBCO (005), YBCO (103), BaCu2 O2 (103; 211), BaCuO2 (600)). The that leads to ultrafast growth rates. In situ synchrotron XRD and 2D images were converted into “Intensity versus scattering angle” diffrac- in situ electrical resistivity measurements were used for this pur- tograms using a home-made Python code.[52] The sample was fixed on the heating plate of an Anton Paar furnace, model DHS1100, equipped with a pose. We have shown that the transient liquids used in TLAG graphite dome, capable of supplying heating ramps up to 4.5 °C s−1 . The have a varying composition: two barium cuprate phases are ac- furnace was connected to a set of valves and vacuum systems allowing tually melting in the YBCO TLAG growth region. The direct evi- for the operations in a controlled atmosphere of partial oxygen pressure dence of the existence of the transient liquid has come from the (PO2 ) and total pressure in each stage of the thermal treatment. The PO2 generation of a kinetic phase diagram for the Ba–Cu–O system was controlled with mixtures of high-purity dry N2 and synthetic air, pre- under the same processing conditions as the YBCO film growth. calibrated with an O2 -sensor. For the T-route, a gas flow of 0.6 L min−1 The properties of this liquid, as well as its amount, vary across PO2 was used, while in the PO2 -route a two-way vacuum system composed by a rotary pump and a turbomolecular pump was used to set the two neces- and T regions together with the copper valence (Cu-I or Cu-II). sary pressure states. Electrovalves were employed to execute the pressure Understanding the Ba–Cu–O kinetic phase diagram has been es- jump in times <500 ms. The image acquisition time varied from 100 ms sential to disentangle the growth mechanisms of the Y–Ba–Cu–O in Bragg–Brentano geometry (with respect to the YBCO (005) reflection), system. For the practical purpose of being able to control this ki- to 500 ms in grazing incidence (GIXRD) conditions. The latter was used Adv. Sci. 2022, 9, 2203834 2203834 (10 of 12) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com to enhance the signal of polycrystalline phases during the heating ramp at and A.K. acknowledge financial support from Spanish Ministry of Sci- PO2 = 10−5 bar in the P-route process. ence, Innovation and Universities through the FPI grant PRE2019-090621, In Situ Electrical Resistance Measurements: To measure the sample re- PRE2018-084537, and PRE2020-091817, respectively. L.So. and J.J. ac- sistance during the thermal treatment that leads to the growth of YBCO, knowledge financial support from Spanish Ministry for the FPU grants. STO substrates were prepared for two-probe contact resistance measure- S.R. thanks the Universitat de Girona for IFUdG grant and A.Q. thanks ments by gluing silver wires of 0.05 mm (GoodFellow) with silver paint the Spanish Ministry of Science, Innovation and Universities (“Juan de la (CDS electronique) up to 900 °C. Then, the precursor chemical solution Cierva” postdoctoral fellowship [Grant no. IJC2018-035034-I]). The authors was deposited by spin coating and a pyrolysis thermal treatment was per- thank the Scientific Services at ICMAB and ICN2 Electron Microscopy Di- formed as in not-contacted samples. Crystallization of these samples was vision. The authors acknowledge the use of instrumentation as well as the performed at ICMAB facilities, in a quartz tube inserted in a tubular fur- technical advice provided by the National Facility ELECMI ICTS, node “Lab- nace, connected to the same vacuum and gas system used for synchrotron oratorio de Microscopías Avanzadas” at University of Zaragoza. The au- experiments. Resistance measurements were performed through a Lab- thors also acknowledge SOLEIL Synchrotron for provision of synchrotron VIEW interface using a Keithley 2450 multimeter by applying a current of radiation facilities and are grateful for assistance while using the DiffAbs 100 μA during the whole growth process. Upon a fast PO2 increase in PO2 - beamline. route, the authors expect to observe a drop in the sample resistance (raw data in Supporting Information). A blank curve was also recorded for two- probe contact STO samples without YBCO precursors (Supporting Infor- Conflict of Interest mation), to confirm that the resistance drop observed after the P jump is only due to the growth of YBCO. The authors declare no conflict of interest. Characterization Techniques: Scanning electron microscopy (SEM) was carried out with a field emission QUANTA 200 FEG from FEI Com- panyTM, equipped with energy dispersive X-ray Spectroscopy (EDX). No Data Availability Statement further sample preparation is required because the final YBCO film is al- ready electrically conductive after oxygenation. The data that support the findings of this study are available from the cor- For transmission electron microscopy (TEM) analysis, cross-sectional responding author upon reasonable request. specimens of YBCO thin films were prepared by conventional methods: slicing, gluing, tripod polishing and Ar+ ion milling using Gatan PIPS. Cross-sectional high-resolution TEM (HR-TEM), high-angle annular dark Keywords field scanning transmission electron microscopy (STEM-HAADF), and en- ergy dispersive X-ray spectroscopy (EDX) were obtained using FEI Tecnai chemical solution deposition, growth from transient liquid, kinetic phase G2 F20 S-TWIN HR(S)TEM microscope, operated at 200 kV. Gatan digi- diagrams, superconducting YBa2 Cu3 O7-x tal micrograph software (Gatan, Inc., Pleasanton, CA, USA) was used for image analysis. HR-STEM images were obtained using a FEI Titan 60–300 Received: July 4, 2022 microscope equipped with an X-FEG gun, a CETCOR probe corrector, and Revised: August 13, 2022 a Gatan TRIDIEM 866 ERS energy filter operated in STEM mode at 300 kV. Published online: September 18, 2022 X-ray diffraction analysis of films ex situ after growth was carried out with a General Area Detector Diffraction System (GADDS) from Bruker- AXS (model D8 Advance) operating with a CuK𝛼 beam source. Tex- ture analyses were performed with a high resolution Discover D8 Bruker diffractometer. Inductive critical current densities (Jc at self-field) were ob- [1] M. K. Wu, J. R. Ashburn, C. J. Torng, P. H. Hor, R. L. Meng, L. Gao, Z. tained with a commercial superconducting quantum interference device J. Huang, Y. Q. Wang, C. W. Chu, Phys. Rev. Lett. 1987, 58, 908. (SQUID) magnetometer from Quantum Design, equipped with a super- [2] D. P. Kumah, J. H. Ngai, L. Kornblum, Adv. Funct. Mater. 2020, 30, conducting magnet of 7 T. In field transport critical current density, tran- 1901597. sition temperature Tc and Hall carrier density nH were measured with [3] J. H. Lee, G. Luo, I. C. Tung, S. H. Chang, Z. Luo, M. Malshe, M. a PPMS Quantum Design system equipped with a 9 T superconducting Gadre, A. Bhattacharya, S. M. Nakhmanson, J. A. Eastman, H. Hong, magnet. Tc and nH were obtained using the van der Pauw method. J. Jellinek, D. Morgan, D. D. Fong, J. W. Freeland, Nat. Mater. 2014, 13, 879. [4] J. L. MacManus-Driscoll, M. P. Wells, C. Yun, J.-W. Lee, C.-B. Eom, D. Supporting Information G. Schlom, APL Mater. 2020, 8, 040904. Supporting Information is available from the Wiley Online Library or from [5] H. Rogalla, P. H. Kes, 100 Years of Superconductivity, CRC Press, Boca the author. Raton, FL 2012. [6] D. Larbalestier, A. Gurevich, D. M. Feldmann, A. Polyanskii, Nature 2001, 414, 368. Acknowledgements [7] C. Yao, Y. Ma, iScience 2021, 24, 102541. [8] J. L. MacManus-Driscoll, S. C. Wimbush, Nat. Rev. Mater. 2021, 6, The authors acknowledge the European Research Council for the UL- 587. TRASUPERTAPE project (ERC-2014-ADG-669504), IMPACT project (ERC- [9] L. Soler, J. Jareño, J. Banchewski, S. Rasi, N. Chamorro, R. Guz- 2019-PoC-874964) and EU COST action for CA16218 (NANOCOHYBRI). man, R. Yáñez, C. Mocuta, S. Ricart, J. Farjas, P. Roura-Grabulosa, The authors also acknowledge financial support from Spanish Ministry X. Obradors, T. Puig, Nat. Commun. 2020, 11, 344. of Science and Innovation with PID2021-127297OB-C21 and PID2021- 127297OB-C22, and from Spanish Ministry of Science, Innovation and [10] S. Rasi, L. Soler, J. Jareño, J. Banchewski, R. Guzman, C. Mocuta, M. Universities through the “Severo Ochoa” Program for Centers of Excel- Kreuzer, S. Ricart, P. Roura-Grabulosa, J. Farjas, X. Obradors, T. Puig, lence in R&D (SEV-2015-0496 and CEX2019-000917-S), and from the J. Phys. Chem. C 2020, 124, 15574. Spanish Ministry of Economy and Competitiveness with the SUMATE [11] A. Queraltó, J. Banchewski, A. Pacheco, K. Gupta, L. Saltarelli, D. Gar- project (RTI2018-095853-B-C21, RTI2018-095853-B-C22, co-financed by cia, N. Alcalde, C. Mocuta, S. Ricart, F. Pino, X. Obradors, T. Puig, ACS the European Regional Development Fund, MCIU/AEI/FEDER, UE). They Appl. Mater. Interfaces 2021, 13, 9101. also thank the Catalan Government (2017-SGR-1519) and the Catalan en- [12] P. Vermeir, J. Feys, J. Schaubroeck, K. Verbeken, P. Lommens, I. Van ergy network XRE4S (2018 XARDI 00002) for their support. L.S., D.G., Driessche, Mater. Res. Bull. 2012, 47, 4376. Adv. Sci. 2022, 9, 2203834 2203834 (11 of 12) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH www.advancedsciencenews.com www.advancedscience.com [13] T. Ohnishi, J.-U. Huh, R. H. Hammond, W. Jo, J. Mater. Res. 2004, 19, [32] D. M. De Leeuw, C. A. H. A. Mutsaers, C. Langereis, H. C. A. 977. Smoorenburg, P. J. Rommers, Phys. C: Supercond. 1988, 152, 39. [14] J. L. MacManus-Driscoll, A. Kursumovic, B. Maiorov, L. Civale, Q. [33] P. Chu, R. C. Buchanan, J. Mater. Res. 1993, 8, 2134. X. Jia, S. R. Foltyn, H. Wang, IEEE Trans. Appl. Supercond. 2007, 17, [34] G. F. Voronin, S. A. Degterov, J. Solid State Chem. 1994, 110, 50. 2537. [35] T. B. Lindemer, E. D. Specht, Phys. C: Supercond. 1995, 255, 81. [15] X. Qi, J. L. MacManus-Driscoll, Curr. Opin. Solid State Mater. Sci. [36] J. Chu, Y. Zhao, M. Z. Khan, X. Tang, W. Wu, J. Shi, Y. Wu, H. Huhtinen, 2001, 5, 291. H. Suo, Z. Jin, Cryst. Growth Des. 2019, 19, 6752. [16] Y. Yamada, Supercond. Sci. Technol. 2000, 13, 82. [37] H. J. T. Ellingham, J. Soc. Chem. Ind. 1944, 63, 125. [17] Y. Wu, G. Jiang, J. Zhu, D. Wu, X. Quan, J. Shi, H. Suo, Y. Zhao, Su- [38] A. Block-Bolten, D. R. Sadoway, Metall. Mater. Trans. B 1983, 14B, percond. Sci. Technol. 2021, 34, 05LT01. 231. [18] X. Obradors, T. Puig, M. Gibert, A. Queraltó, J. Zabaleta, N. Mestres, [39] H. Chen, K. Zalamova, A. Pomar, X. Granados, T. Puig, X. Obradors, Chem. Soc. Rev. 2014, 43, 2200. Supercond. Sci. Technol. 2010, 23, 034005. [19] T. Schneller, R. Waser, (Eds: M. Kosec, D. Payne), Chemical Solution [40] C. F. Sánchez-Valdés, T. Puig, X. Obradors, Supercond. Sci. Technol. Deposition of Functional Oxide Thin Films, Springer, New York 2013. 2015, 28, 024006. [20] T. Araki, I. Hirabayashi, Supercond. Sci. Technol. 2003, 16, R71. [41] T. Qu, Y. Zhu, F. Feng, G. Lin, S. Deng, H. Lu, X. Zhang, Q. Fu, S. Xiao, [21] V. Chepikov, N. Mineev, P. Degtyarenko, S. Lee, V. Petrykin, A. P. Zeng, Z. Han, IEEE Trans. Appl. Supercond. 2016, 26, 6602205. Ovcharov, A. Vasiliev, A. Kaul, V. Amelichev, A. Kamenev, A. Molodyk, [42] Y. Li, Y. Gong, J. Tao, B. Wu, S. Zhou, J. Supercond. Novel Magn. 2020, S. Samoilenkov, Supercond. Sci. Technol. 2017, 30, 124001. 33, 1987. [22] J. H. Lee, H. Lee, J. W. Lee, S. M. Choi, S. I. Yoo, S. H. Moon, Super- [43] Y. Shiohara, A. Endo, Mater. Sci. Eng. R: Rep. 1997, 19, 1. cond. Sci. Technol. 2014, 27, 044018. [44] R. Yu, V. Gomis, N. Vilalta, F. Sandiumenge, B. Martínez, S. Piñol, X. [23] X. Obradors, T. Puig, Supercond. Sci. Technol. 2014, 27, 044003. Obradors, Phys. C: Supercond. 1997, 290, 161. [24] P. Vermeir, I. Cardinael, J. Schaubroeck, K. Verbeken, B. Michael, P. [45] A. Stangl, A. Palau, G. Deutscher, X. Obradors, T. Puig, Sci. Rep. 2021, Lommens, W. Knaepen, D. Jan, K. De Buysser, I. Van Driessche, Inorg. 11, 8176. Chem. 2010, 49, 4471. [46] J. Nývlt, J. Cryst. Growth 1968, 3–4, 377. [25] Y. Zhao, P. Torres, X. Tang, P. Norby, J. Grivel, Inorg. Chem. 2015, 54, [47] D. A. Porter, K. E. Easterling, M. Y. Sherif, Phase Transformations in 10232. Metals and Alloys, CRC Press, Boca Raton, FL 2009. [26] Y. Zhao, J. Chu, T. Qureishy, W. Wu, Z. Zhang, P. Mikheenko, T. H. [48] C. Krauns, M. Tagami, Y. Yamada, M. Nakamura, Y. Shiohara, J. Mater. Johansen, J.-C. Grivel, Acta Mater. 2018, 144, 844. Res. 1994, 9, 1513. [27] A. Gupta, R. Jagannathan, E. I. Cooper, E. A. Giess, J. I. Landman, B. [49] L. Saltarelli, et al. manuscript submitted for publication, n.d. W. Hussey, Appl. Phys. Lett. 1998, 52, 2077. [50] P. Pangaud, S. Basolo, N. Boudet, J.-F. Berar, B. Chantepie, P. [28] X. Obradors, T. Puig, S. Ricart, M. Coll, J. Gazquez, A. Palau, X. Grana- Delpierre, B. Dinkespiler, S. Hustache, M. Menouni, C. Morel, Nucl. dos, Supercond. Sci. Technol. 2012, 25, 123001. Instrum. Methods Phys. Res., Sect. A 2007, 571, 321. [29] Z. Li, M. Coll, B. Mundet, A. Palau, T. Puig, X. Obradors, J. Mater. [51] K. Medjoubi, T. Bucaille, S. Hustache, J.-F. Bérar, N. Boudet, J.-C. Chem. C 2019, 7, 4748. Clemens, P. Delpierre, B. Dinkespiler, J. Synchrotron Radiat. 2010, 17, [30] J. L. MacManus-Driscoll, J. C. Bravman, R. B. Beyers, Phys. C: Super- 486. cond. 1995, 241, 401. [52] C. Mocuta, M. I. Richard, J. Fouet, S. Stanescu, A. Barbier, C. Guichet, [31] F. Taïr, L. Carreras, J. Camps, J. Farjas, P. Roura, A. Calleja, T. Puig, X. O. Thomas, S. Hustache, A. V. Zozulya, D. Thiaudière, J. Appl. Crys- Obradors, J. Alloys Compd. 2017, 692, 787. tallogr. 2013, 46, 1842. Adv. Sci. 2022, 9, 2203834 2203834 (12 of 12) © 2022 The Authors. Advanced Science published by Wiley-VCH GmbH