Microstructure of Al-Si Slurry Coatings on Austenitic High-Temperature Creep Resisting Cast Steel

,is paper presents the results of microstructural examinations on slurry aluminide coatings using scanning electron microscopy, X-ray microanalysis, and X-ray diffraction. Aluminide coatings were produced in air atmosphere on austenitic high-temperature creep resisting cast steel. ,e function of aluminide coatings is the protection of the equipment components against the high-temperature corrosion in a carburising atmosphere under thermal shock conditions. ,e obtained coatings had a multilayered structure composed of intermetallic compounds. ,e composition of newly developed slurry was powders of aluminium and silicon; NaCl, KCl, and NaF halide salts; and a water solution of a soluble glass as an inorganic binder. ,e application of the inorganic binder in the slurry allowed to produce the coatings in one single step without additional annealing at an intermediate temperature as it is when applied organic binder. ,e coatings were formed on both: the ground surface and on the raw cast surface. ,e main technological parameters were temperature (732–1068°C) and time of annealing (3.3–11.7 h) and the Al/Si ratio (4–14) in the slurry. ,e rotatable design was used to evaluate the effect of the production parameters on the coatings thickness. ,e correlation between the technological parameters and the coating structure was determined.


Introduction
High-temperature corrosion is a common problem in various industrial processes, such as heat treatment, chemical processing, refining, etc. in, for example, petrochemical, aerospace, and gas turbine industries.Furnaces for thermochemical treatment, for example, carburising or nitriding furnaces, are often made of high-temperature creep resisting cast steels, especially Fe-Cr-Ni based alloys.A high mechanical strength at elevated temperatures and a resistance to thermal shocks are required from these materials.It is difficult to keep a satisfactory compromise between the mechanical strength and the surface degradation resistance in such severe working conditions.On the surface of Cr-rich materials, a protective chromium oxide layer (Cr 2 O 3 ) is formed and provides the carburisation resistance at temperatures below 1050 °C [1].e chromia is also liable to spalling and cracking during creep and thermal shocks.erefore, the other protective layers such SiO 2 and Al 2 O 3 are considered.For oxidation resistance above 1200 °C, alloys that rely on the protective Al 2 O 3 scale formation are preferred over those forming chromia.Unfortunately, the addition of aluminium and silicon to the high-temperature creep resisting alloys to develop full protection involves metallurgical compromise in strength and ductility.erefore, silicon is usually limited to 2% and aluminium to less than 4% in alloys [2].
One of the most frequent solutions to introduce aluminium and silicon is an application of appropriate coatings.Inward diffusion of aluminium and outward diffusion of iron, nickel, and chromium results in the formation of intermetallic phases with very interesting properties.ey are widely used for the improvement of the corrosion resistance of iron-and nickel-based alloys.Among these intermetallics, Fe and Ni aluminides improve the resistance to many aggressive environments including oxidising [3,4], sulphidising [5], and carburising [6,7] agents.ese coatings extend the lifetime of the equipment as a result of the carburisation reduction of the surface layer.Coatings, that protect the base material in industrial processes, need not only to inhibit the diffusion of elements, but also must offer resistance to thermal shocks that occur during quenching following the carburising process [8].
ese shocks cause accelerated destruction of the protective coatings.In addition to aluminium in the coating, very important is the presence of silicon, which decreases the carbon solubility in austenite and is one of the main elements improving the resistance to carbon absorption at high temperatures.Moreover, it has been observed that SiO 2 is present as a subscale beneath chromium oxide [1].
e manufacturing method of the coating should be easy to apply on complex-shaped parts, and the cost should be as low as possible.One of these methods is the slurry cementation.Slurry aluminisation can provide advantages in the process, that is, simplicity and economical efficiency.Typical slurries manufactured on the base of aluminium powder (as an active component) and of the solution of an organic binder are widely used [9][10][11][12].When the polyvinyl alcohol is used as the organic binder in the slurry, slurrycoated specimens have to be held at a temperature of 300 °C for 4 h [12] or 400 °C for 3 h [10] to expel the binder and only after that heat treated.e use of the other type of binder instead of the organic one could allow to produce the coatings in an one single step without additional annealing at an intermediate temperature.
In this work, a single-step slurry cementation was applied to create the protective Al-Si coatings.e method used by the author is a modification of the conventional one.In this study, the inorganic binder, that is, a water solution of sodium silicate, was applied instead of the organic one.e simultaneous use of the sodium silicate and the silicon powder in the slurry causes dehydration and self-hardening of the slurry.is means that silicon can be beneficial in the slurry both from a technological point of view and later in the diffusion-forming process.e innovative composition of the slurry has forced the determination of optimum production parameters of the protective coatings on austenitic high-temperature creep resisting cast steel.
e slurry usually comprises metal powders with a grain size below 40 μm [13], but very often the grain size is in the order of a few micrometers [14,15].In this work, powders of larger granularity were used. is has reduced the cost of the slurry.In addition, a larger grain of aluminium powder reduces the risk of an explosion during production.
Additionally, in this work, a mixture of three selected salts was used.ese salts applied in the slurry are a flux and have good properties from an ecological and economic point of view.e existence of oxides on both aluminium particles in the slurry and the surface of samples is a barrier for diffusion processes.e oxide shell is dissolved by this flux during annealing.e coating is formed in a liquid state (the temperature of annealing is above the temperature of aluminium melting).e liquid state guarantees high rates of the diffusion reaction, and the interruption of the oxide shell accelerates processes in the liquid state.

Samples and Slurry Composition.
e substrate used for the investigation was austenitic high-temperature creep resistant cast steel (GX20NiCrSi30-18-1) containing approximately 0.2 wt.% C, 18% Cr, 30% Ni, 0.7% Mn, 1.5% Si, 0.03% P, 0.03% S, and Fe balance.Dimensions of the samples were 10 × 20 × 6 mm.e surface of samples used in the experiment was finished by grinding.However, several samples of the raw cast surface were subjected to produce the coatings as well (to verify the possibility of the coating formation).
e wall thickness of the casting was about 9 mm; therefore, the dimensions of this sample were 10 × 20 × 9 mm.e active slurry was prepared from silicon and aluminium powders.e grain size was approximately 100 µm for the aluminium powder and within the range 200-250 µm for the silicon powder.e binder used in the slurry consisted of an aqueous solution of sodium silicate (soluble glass).e flux was used of composition of sodium and potassium salts (NaF, NaCl, and KCl).e ratio of the sodium silicate to (Al + Si) powders was 15/100, and the ratio of the salts to (Al + Si) powders was 9.5/100.e quantitative composition of the slurry was determined experimentally and patented [16].
e samples after degreasing with acetone were immersed several times in the water slurry and dried until the value of the covering mass amounted to 0.3 g•cm −2 and then annealed in air atmosphere (without any protective atmosphere).e slurry residues were removed after annealing, and the samples were washed in an ultrasonic bath in acetone.

Design of Experiment.
e purpose of the application of the plan is the determination of the optimum parameters of coatings producing.e annealing temperature, time, and the ratio Al/Si in the slurry were adopted as three independent variables in the design of the experiment.e determination of the equations describing the thickness as a function of these three parameters was the objective in the chosen experiment design.A rotatable central composite design was chosen.us, rotatable designs do not favour one direction over another during exploring.is is reasonable when the knowledge about the response surface before experimentation is insufficient.is design requires twenty interactions (total points in the design) if there are three variable parameters [17].e various combinations of levels of these interactions lie on the edge (factorial) points (±1, ±1, ±1) and the centre point (0, 0, 0) of a cubic experimental region, as illustrated schematically in Figure 1.
e plan assumes the existence of additional (axial) points (±α, 0, 0), (0, ±α, 0), and (0, 0, ±α).In the rotatable design, the variance of the predicted response is the same at all points that have the same distance from the centre of the design.
e plan of the rotatable design for the coded variables (x i ) together with all the real values (X i ) of the interaction between the three variables is shown in Table l.It was assumed that the unknown characteristics can be approximated by the second-order model which describes 2 Advances in Materials Science and Engineering quadratic surfaces, and quadratic surfaces can take several shapes [17].Second-order model for three independent variables takes the form (1) A second-order design has a uniform precision which is a reasonable criterion, because we want to know precision how close to the origin a maximum or other surface feature may be [17].
e central temperature, the central annealing time, and the central ratio Al/Si were correspondingly 900 °C, 7.5 hours and 9 : 1, respectively.e remaining points (±1, ±α) were Table 1: e rotatable design for three-factor interactions with coded and real variables.Advances in Materials Science and Engineering calculated assuming an appropriate step.e steps (levels coded as ±1) for the temperature, time, and ratio Al/Si were correspondingly ±100, ±2.5, and ±3, respectively.e ratio Al/Si in the slurry was changed from 4 : 1 to 14 : 1. e samples covered with the slurry were annealed at corresponding temperatures between 732 °C and 1068 °C for corresponding times between 3.3 and 11.7 hours, respectively, in the air atmosphere.e design predicted the coating properties throughout the area, even at parameter combinations actually not working.e properties of coatings are predicted for the whole range of the three parameters using only the parameter combinations considered.

Microstructural Characterization.
Metallographic preparations of the cross sections of all coatings were conducted according to the standard methods.All samples have been examined using the field emission scanning electron microscopy (FESEM) (Hitachi SU-70) equipped with the energy dispersive X-ray microanalysis (EDS; ermo Scientific NORAN System 7).Accelerating voltage of 20 kV was applied for SEM and EDS analyses.e COMPASS software was used by collecting a complete, deadtime-corrected spectrum at every point of the sample.COMPASS is a statistical-based software applied in the ermo NORAN 7 software which enables quantitative analysis of each elemental phase [18].
e COMPASS software finds image locations that have the same spectral "fingerprint," with no assumptions about the chemical composition of the sample.Such created elemental phase maps show the spatial distribution of regions that have statistically similar spectral features (peak locations and relative peak heights) and cumulative spectra of these regions.
e XRD analysis was conducted with Panalytical X'Pert-PRO X-ray diffractometer using Co-Kα radiation and the HighScore software to analyse the diffraction patterns.Consecutive layers of the coating material were removed by grinding.Such a protocol enabled to identify phase components present in particular layers.e microhardness tests were performed using a Buehler hardness tester using a Knoop indenter at the load of 0,098 N during 15 s.

Results and Discussion
e coatings were successfully produced for each set of assumed technological parameters.
e microscopic observations proved that the microstructures of the coatings depended on the production parameters.In general, the typical coating consisted of three layers: the internal layer (1), the middle layer (2), and the external layer (3) (Figure 2(a)).
e total thickness of the resulting coatings for particular cases was rather uniform, and the standard deviations did not exceed 5%, except the coatings produced at a low temperature, especially in a long time, which were less uniform (the standard deviations did not exceed 10%).
e microstructure of the coating obtained with the central technological parameters (900 °C, 7.5 h, 9 Al/Si) is shown in Figure 2. Figure 2(a) shows the Al-Si coating obtained on the surface after grinding and Figure 2(b) on the raw cast surface.No significant differences in microstructure between the coatings on the machined and the raw cast surface were observed.us, the coatings can be produced on the cast surface without further mechanical processing like grinding [19].e possibility to obtain coatings on the surface of the raw cast is very important from the technological point of view.e presence of the flux in the slurry promotes the formation of the coating despite the existence of a thick oxide layer on the surface of the casting.
Different structures and properties of the coatings were obtained depending on the technological parameters in the established plan of the experiment.For example, the structures of the two coatings prepared with the parameters belonging to set numbers 7 and 2 are shown in Figures 3(a) and 3(b), respectively.e main differences in the structure of the coatings were the thickness and the chemical composition.
e relations (as regression models) between the thickness and various parameters were determined using Statistica 12 software, and the equations thus obtained are presented below.e following notation was used: Y x , the thickness of the x sublayer (µm); Y t , the total thickness (µm); X 1 , the annealing temperature ( °C); X 2 , the annealing time (h); X 3 , the Al/Si ratio; R, multiple correlation coefficient; R 2 , unbiased estimator; F, value of Snedecor test.Advances in Materials Science and Engineering R � 0.78; R 2 � 0.60; F � 12.80. ( e determined regression models for the thickness were used to show the dependence on the annealing time and the temperature in the following graphs (Figure 4).e total thickness (5) of the coatings and the thickness of each layer depended mainly on the temperature and the time of the annealing.e ratio of aluminium to silicon in the slurry only affected the thickness of the second layer (3).Hence, the Al/Si ratio had no significant influence on the thickness of the first and third layers (in the applied range).e graph of the second layer thickness has been drawn for the Al/Si ratio X 3 � 6.It was found that the total thickness of the coating increased with increasing temperature and time of annealing (Figure 4(d)).e same holds for the thickness of the third layer (Figure 4(c)).
e situation was different using the regression models representing the thickness of the first and the second layer.e increase in the thickness of these layers is apparent with increasing temperature and the time, but only at temperatures above 900 °C.e thickness of the first and the second layer decreased significantly with the increasing times at temperatures below 800 °C.e Knoop hardness measurement results performed in layers 1 and 2 and in the base material were correspondingly 784 ± 137 HK0.01 in the first layer, 1251 ± 224 HK0.01 in the second layer, and 436 ± 68 HK0.01 in the base material.e results of the coating hardness measurements showed that the first layer has an intermediate hardness value in between those of the substrate and the second layer reducing the stress gradient between the substrate and the coating.e third layer was brittle at room temperature, and hardness tests have not been successfully performed.e spalling of the third layer occurring probably during metallographic preparation was observed, especially in coatings where the thickness of first and the second layer was small and third layer was big (Figure 4(c)-in a long time of annealing).
Figure 5 shows the elemental mapping of the coating obtained at central parameters of the experiment design.e results of a quantitative point analysis (EDS) for the first layer and the second layer of the coatings are presented in Figures 6(a) and 6(b) and Table 2.
Quasiphase maps by the COMPASS procedure showed the spatial distribution of regions that have statistically similar spectral features (Figure 7).
e results of the chemical quantitative analysis of each phase suggested by the COMPASS software are given in Table 3.
XRD phase analysis was performed on sections parallel to the surface of the coating, and subsequent XRD patterns are shown in Figure 8.
AlN, Al 5 FeNi, and β-NiAl phases were found near the surface of the coating formed at 1000 °C.e β-NiAl phase was identified in the second layer of this coating, and the α-(Fe, Cr, Si) solution and CrSi 2 were identified in the first layer.Alloyed austenite was identified in the substrate.XRD studies were also carried out for the coatings formed at a temperature of 800 °C, because the morphology of these coatings (formed at 800 °C) was slightly different from those formed at 1000 °C (Figures 3(a) and 3(b)).It was also observed that the kinetics of growth of the first and the second layer at a temperature of 800 °C differs from the kinetics of growth of these layers annealed at higher temperature (Figures 4(a AlN phase was observed in the surface layer of the coating produced from the slurry with the ratio of Al/Si � 6 (the amount of the silicon in the slurry is higher).In the    [20], with a certain content of nickel and aluminium.e solubility of aluminium in c-Fe is rather small contrary to that of α-(Fe, Cr) [21].e aluminised coatings obtained at 1000 °C in this work consisted of similar phase compositions to coatings with interfacial layers of the α-(Cr, Fe) and β-FeAl layers and the β-FeAl layer and a thick outer layer of δ-Fe 2 Al 5 reported by Hu et al. [22].ey also observed [22] spherical β-FeAl precipitates present in the α-(Cr, Fe) grains.In the current work, a ne-grained intermetallic β-Al(Ni, Fe) phase was formed in the matrix of the rst layer containing  2, point 2).e amount of nickel in the β-Al(Ni, Fe) precipitates from first layer was significantly higher than that in the β-Al(Fe, Ni) phase from the second layer (Table 3, phase 3 and phase 4).
It was found that, with increasing production temperature, the thickness of the first layer increased, and simultaneously, a thin sublayer rich in nickel was formed at the boundary with the second layer (Figure 7, Table 3, phase 4).
is sublayer had nearly the same chemical and phase

Element
Chemical composition (%at.), according to Figure 6(a) Chemical composition (% at.), according to Figure 6(b  compositions as those of β-Al(Ni, Fe) precipitates from the rst layer (Figure 6(a), Table 2, point 2 and point 3).In the vicinity of this Ni-rich sublayer, a self-forming Cr-rich layer (Figure 6(a), Table 2, point 1) was observed.is sublayer can be used as a barrier blocking aluminium di usion, because of a low solid solubility of aluminium in Cr-rich layer, as reported by Gao et al. [23].
Alumina (α-Al 2 O 3 ) and AlN were also formed near the surface.Additionally, aluminide coatings have an ability to form a protective oxide Al 2 O 3 scale at higher temperatures during oxidation [24].Aluminium oxide has good protective properties at high temperature, which is advantageous.AlN phase is oxidized to α-Al 2 O 3 , but not directly.e transition oxides are transformed into amorphous Al 2 O 3 gradually and nally to α-Al 2 O 3 [25].erefore, the presence of the AlN phase on the surface could be also very advantageous.

Summary
e aluminide protective coatings by the slurry method were successfully formed on austenitic creep-resisting cast steel.
e coatings consisted of three-layers having di erent thicknesses and properties.
e thickness of the coatings from the high-temperature processes were very uniform, contrary to the coatings produced using the low temperature, especially in a long time, which were less uniform.e use of the long time of annealing caused growing of the third layer, which was very brittle.e structure and, as a consequence, the properties of the coatings depended on the production parameters.e thickness was the function of time and temperature of the annealing.
e growth of the second layer is mainly induced by the aluminium di usion from an active mixture and nickel and Advances in Materials Science and Engineering iron di usion from the substrate.e activity of the di using elements during the heat treatment is the parameter which determines the type of coating that is obtained.ere are two variants: outwards and inwards di usion coatings.e rst one is produced in a low activity of aluminium and leads to the β-Al(Fe, Ni) phase forming. is case was observed in coatings produced at 900, 1000, and 1068 °C (Figure 8).In the second layer also, the Cr(Al, Si) 2 phase formed in the matrix of the β-Al(Fe, Ni) phase was observed (Figures 3(b) and 5).
e inwards di usion coating is produced in a high activity of aluminium and leads to the Ni 2 Al 3 or Al 5 FeNi phase creation.is case was observed in coatings produced at 732 and 800 °C.Phases which contain a higher amount of aluminium were present, for example Al 5 FeNi (sample no. 3, Figure 9) and Ni 3 Al 2 (sample no. 3 and no. 7, Figure 10).e Cr(Al, Si) 2 phase was also observed in both the investigated cases.
For various Al to Si ratios in the slurry, di erences in the structure of the coatings were signi cant in coatings manufactured at the lower temperature, that is, 800 °C.e Al 8 Cr 5 phase was created, when the amount of silicon was smaller in the slurry (Al/Si 12), and Al 5 FeNi phase was observed, when the amount of silicon in the slurry was higher (Al/Si 6).Analysis of the results allowed to develop a simpli ed model of the coatings showing the in uence of the silicon in the slurry (Figure 11).
e Al-Si coatings were designed to protect the hightemperature creep resisting cast steel against carburisation.Assessment of the coatings' behaviour in the carburising atmosphere under thermal shock conditions will be the nal veri cation of their protective properties.
e promising results of the preliminary investigations of carburisation and oxidation of coated cast steel have been carried out and presented in the article [26].e selection of the production parameters leads to the formation of the coating having a suitable structure and properties.e phase composition of the coatings, which is Al/Si ratio dependent, can have a signi cant in uence on the resistance against carburisation.
Coatings can be produced on the raw cast surface without any further surface nishing processes.is is a very   10 Advances in Materials Science and Engineering important advantage of the applied method, because it simpli es the process of equipment manufacturing and reduces costs.e advantageous slurry method is more convenient and less harmful than the classic powder methods. is method is also easy to apply and could be used for a local application.Other advantages of this method are a low expenditure of the slurry materials and possibility of producing uniform and high-quality coatings on equipment with a large and complex shape.A very important aspect from an industrial conditions perspective is the increase of the mechanical strength combined with the decreased brittleness of the dried slurry (before annealing), caused by the soluble glass additive.Moreover, the use of the powder of coarser granularity reduces the risk of an explosion.

Conclusions
(1) Experimental rotatable designs are useful tools for identifying signi cant process parameters for the aluminide coatings by the slurry method.(2) ree-layer Si-rich aluminide coatings were obtained on high-temperature creep resisting cast steel using a modi ed slurry method with soluble glass as the binder in the air atmosphere.Application of the aqueous solution of sodium silicate in the slurry allowed for annealing in an one single step.(3) e outer layer of the coatings contained Al-rich phases.e middle layer contained intermetallic phases, mainly β-Al(Ni, Fe). e crystal lattice of the rst layer was di erent from the substrate and from the other layers of the coating.Ferrite was formed from Ni-depleted austenite.is occurred because nickel di uses from the substrate to the second layer, and Ni content was lower in the rst layer than in the substrate.(4) ickness of the coatings depended on the production parameters (mainly on time and temperature).e ratio Al/Si in the slurry (in the applied range) did not have a signi cant impact on the thickness of the coating.e content of aluminium and silicon in the slurry a ected the phase components formed in the coatings.(5) Hardness value of internal layer was intermediate between the hardness of the middle layer and the base material, which is favourable from the thermal shock resistance viewpoint.(6) e slurry method is also applicable for materials of a raw cast surface thanks to the presence of a ux in the slurry.

Figure 1 :
Figure 1: Graphical representation of the rotatable design for the three variables.

Figure 2 :
Figure 2: Microstructure (BEI) of Al-Si coating on sample (a) number 16 (surface after grinding), (b) with a rough cast surface (technological parameters in both cases: 900 °C, 7.5 h, Al/Si � 9; central point of the experiment).

4
) and 4(b)).A decrease in thickness with time was observed at annealing temperature of 800 °C.Figures9 and 10show XRD diffraction patterns of two coatings produced at 800 °C for 10 hours but using different slurry compositions.
second and the third layers, Al 5 FeNi, Cr(Al, Si) 2 , Al 3 Ni 2 , and β-AlNi phases were identi ed. e Al 5 FeNi phase dominated in the third layer and the Al 3 Ni 2 and β-AlNi phases in the second layer.e α-(Fe, Cr, Si) solution in the rst layer was identi ed.Cr(Al, Si) 2 , Al 8 Cr 5 , Al 3 Ni 2 , and β-AlNi phases were observed near the surface of the coating produced from the slurry with the ratio of Al/Si 12. Cr(Al, Si) 2 , Al 3 Ni 2 , and β-AlNi were in the third and second layers.e α-(Fe, Cr, Si) solution was observed in the rst layer similar to the previously described cases.e presence of M 7 C 3 carbides from the substrate was also visible in the coating.e rst layer was a di usion layer formed by outward iron di usion and inward di usion of silicon.Nickel di usion from this area to the second layer led to Ni content decrease in the rst layer, consequently promoting a transition of f.c.c. to b.c.c.crystal lattice.e c-(Fe, Ni, Cr) solution transformed into the α-(Fe, Cr, Si) solution

Figure 4 :
Figure 4: Graph of the correlation between the technological parameters and the thickness of (a) rst layer, (b) second layer, (c) third layer, and (d) coatings.

Figure 5 :
Figure 5: Elemental mapping in a cross section in coating number 17 (technological parameters 900 °C, 7.5 h, Al/Si � 9; central point of the experiment).

Figure 6 :
Figure 6: Microstructure (BEI) of (a) the first layer and (b) the second layer of the coating no. 4 obtained at temperature 1000 °C in 10 h (Al/Si � 6) with points of EDS X-ray microanalysis.

Figure 7 :
Figure 7: Mapping of phase components (according to Figure 5) with cumulative EDS spectrums.

Figure 11 : 3 αFigure 10 :
Figure 11: A simpli ed model of the coatings.E ect of the silicon content in the slurry on the phase composition of the coatings.

Table 2 :
Chemical composition according to Figures6(a

Table 3 :
Average content of chemical elements in phases according to Figure7.