In my work on sand mold casting with external chills, I treated the coating layer as a multifunctional interface rather than as a simple refractory wash. The external chill accelerates local solidification, controls thermal gradients, and improves feeding in heavy-section castings. At the same time, the chill creates a difficult interfacial environment: the metal surface is smooth, non-porous, and thermally conductive, while the sand mold surface is porous, chemically active, and mechanically fragile. A coating for sand mold casting with external chills must therefore bond to both surfaces, resist erosion by molten iron, release gases quickly, and detach cleanly after shakeout. My central objective was to formulate an alcohol-based coating that satisfies these contradictory requirements without sacrificing suspension stability, brushing performance, or drying speed.
Conventional water-based coatings for sand mold casting dry slowly, absorb moisture, and increase gas-related defects. Conventional alcohol-based coatings dry quickly, but many of them adhere weakly to chill surfaces or form brittle layers that crack under thermal shock. In my approach, I used a composite refractory aggregate, a modified lithium-based bentonite suspension system, and a phenolic resin–polyvinyl butyral binder system. I then evaluated the formulation through orthogonal experiments and production trials. The result was a coating that improved surface quality, reduced sand adhesion, minimized metal penetration, and increased chill recovery in sand mold casting with external chills.
Performance criteria. I defined the coating performance for sand mold casting by five measurable indicators: condition viscosity, 24 h suspension rate, gas evolution, coating wear resistance, and brushability. Condition viscosity reflects flow and leveling behavior. The 24 h suspension rate reflects storage stability and the ability of the suspension network to hold refractory particles. Gas evolution reflects the volume of gas released per unit mass of coating at high temperature. Wear resistance reflects the strength of the dried coating after ignition. Brushability reflects the ability of the coating to form a uniform layer on both sand and metal surfaces. The suspension rate was calculated as
$$S_{24}=\frac{V_s}{V_0}\times100\%$$
where \(S_{24}\) is the 24 h suspension rate, \(V_s\) is the volume of the still-suspended coating after 24 h, and \(V_0\) is the initial coating volume. Gas evolution was calculated as
$$G=\frac{V_g}{m_c}$$
where \(G\) is gas evolution in mL/g, \(V_g\) is the measured gas volume, and \(m_c\) is the coating mass. For adhesion, I considered the thermodynamic work of adhesion:
$$W_{ad}=\gamma_L(1+\cos\theta)$$
where \(W_{ad}\) is the work of adhesion, \(\gamma_L\) is the surface tension of the liquid coating, and \(\theta\) is the contact angle. A lower contact angle on both sand and metal corresponds to better wetting and more uniform coverage. Because the external chill and sand mold have different surface energies, the coating must maintain a balanced wetting state. If the coating wets the metal too well, it may adhere too strongly after solidification; if it wets the sand too poorly, it will penetrate insufficiently and form a weak layer. I therefore designed the coating for sand mold casting with external chills to achieve intermediate wetting, strong green adhesion, and controlled high-temperature release.
Formulation strategy. I used bauxite and white corundum as the composite refractory aggregate. Bauxite provides high refractoriness, moderate thermal expansion, and good cost efficiency. White corundum provides high hardness, chemical resistance, and resistance to basic slag and iron oxide. The mass ratio of bauxite to white corundum was fixed at 4:1. The total refractory mass was set to 100 parts. The suspension agent was a composite-modified bentonite prepared by a two-step intercalation–grafting method. The binder was a phenolic resin–polyvinyl butyral system. Industrial ethanol was used as the solvent. The target coating density was \(1.60\pm0.02\) g/cm³. Table 1 summarizes the main formulation variables and their functions.
| Component | Function in sand mold casting coating | Key requirement |
|---|---|---|
| Bauxite, 250 mesh | Primary refractory skeleton; increases permeability and thermal shock resistance | High refractoriness, controlled particle size |
| Bauxite, 400 mesh | Fills voids among coarse particles; reduces iron penetration | Fine particle size, uniform dispersion |
| White corundum, 250 mesh | Hard phase; improves wear resistance and chemical stability | High hardness, low impurity |
| Composite-modified bentonite | Suspension agent; forms card-house network in ethanol | High swelling index, alcohol compatibility |
| Phenolic resin | High-temperature binder; provides carbon residue and hot strength | Controlled residue, low free phenol |
| Polyvinyl butyral | Low-temperature binder; improves flexibility and adhesion | Compatible with ethanol and resin |
| Industrial ethanol | Solvent; enables fast drying and ignition curing | High purity, controlled evaporation |
Modification of bentonite. Natural calcium bentonite has poor suspension stability in alcohol. Simple lithium exchange improves swelling in water, but the resulting gel can dehydrate in ethanol and lose its network structure. I therefore used a lithium–intercalation–grafting route. First, calcium bentonite was lithium-exchanged with lithium carbonate. Then, organic intercalation was performed with cetyltrimethylammonium bromide, octadecyltrimethylammonium chloride, 5-mercapto-1H-tetrazole acetic acid, and sodium carboxymethyl cellulose. Finally, grafting was performed with γ-aminopropyltriethoxysilane and 1-BOC-2-trifluoroacetylhydrazine. The interlayer spacing was calculated from the Bragg equation:
$$d=\frac{n\lambda}{2\sin\theta}$$
where \(d\) is the interlayer spacing, \(n\) is the diffraction order, \(\lambda\) is the X-ray wavelength, and \(\theta\) is the incident angle. The measured interlayer spacing and swelling index are summarized in Table 2. Lithium exchange slightly reduced the basal spacing from 1.26 nm to 1.22 nm. After intercalation, the spacing increased to 3.12 nm. After grafting, the spacing remained at 3.09 nm. The swelling index increased from 5.8 mL/g to 20.1 mL/g. These results confirmed that the two-step modification expanded the interlayer space and improved alcohol compatibility.
| Bentonite state | Basal spacing \(d_{001}\) / nm | Swelling index / (mL·g⁻¹) | Expected effect in sand mold casting |
|---|---|---|---|
| Natural calcium bentonite | 1.26 | 5.8 | Poor suspension in alcohol |
| Lithium-exchanged bentonite | 1.22 | — | Improved ion exchange but limited alcohol stability |
| Intercalated bentonite | 3.12 | — | Expanded layers, better dispersion |
| Grafted bentonite | 3.09 | 20.1 | Stable network, high suspension rate |
Binder design. The binder must provide green strength before ignition, hot strength during pouring, and controlled brittleness after cooling. Phenolic resin contributes high carbon residue and hot strength, but it can be brittle. Polyvinyl butyral contributes flexibility and adhesion, especially on metal chill surfaces. I fixed the mass ratio of phenolic resin to polyvinyl butyral at 2.5:1. The total binder addition was varied from 2% to 6% based on the refractory mass. The composite binder forms a microgel in ethanol, which increases viscosity and improves particle suspension. However, excessive binder increases gas evolution and may generate porosity in sand mold casting. The optimal binder content therefore required a compromise between coating strength and gas volume.
Refractory aggregate grading. Particle packing controls permeability, surface finish, and resistance to metal penetration. A single particle size leaves large voids. A properly graded mixture reduces porosity and increases packing density. I estimated the packing behavior using
$$\phi_{pack}=\frac{V_s}{V_s+V_v}$$
where \(\phi_{pack}\) is the packing fraction, \(V_s\) is the solid volume, and \(V_v\) is the void volume. For a binary mixture, the void fraction depends on the size ratio and volume fraction. In my formulation, 250 mesh bauxite particles form the coarse skeleton. 400 mesh bauxite particles fill the interstitial spaces. White corundum particles strengthen the skeleton. The coarse fraction improves permeability and reduces cracking. The fine fraction reduces iron penetration and improves surface smoothness. The orthogonal experiment varied the mass ratio of 250 mesh bauxite to 400 mesh bauxite as 60:20, 53:27, and 40:40, while the white corundum content remained part of the 4:1 bauxite-to-corundum ratio.
Orthogonal experimental design. I selected three factors: refractory aggregate grading (A), composite-modified suspension agent content (B), and composite binder content (C). Each factor had three levels. The response variables were condition viscosity, 24 h suspension rate, gas evolution, wear mass loss, and brushability grade. I used an \(L_9(3^3)\) orthogonal array. Table 3 shows the factor levels.
| Level | A: \(m_{250}:m_{400}\) bauxite | B: suspension agent / wt% | C: binder / wt% |
|---|---|---|---|
| 1 | 60:20 | 3 | 2 |
| 2 | 53:27 | 4 | 4 |
| 3 | 40:40 | 5 | 6 |
Test methods. I measured condition viscosity with a flow cup. I measured suspension rate after 24 h by volume. I measured gas evolution by heating a fixed coating mass and collecting the released gas. I measured wear resistance by abrasion mass loss. I measured brushability by applying two consecutive layers on sand and metal specimens, then evaluating flow, brush marks, and thickness uniformity. The brushability grades were defined as follows: Grade I, uniform coating on both surfaces, no brush marks, thickness 0.5–1.2 mm; Grade II, basically uniform coating, slight brush marks, thickness 0.5–1.2 mm; Grade III, non-uniform thickness, obvious brush marks, or thickness outside 0.5–1.3 mm. For numerical analysis, I assigned values of 1, 2, and 3 to Grades I, II, and III, respectively. Lower values therefore indicated better brushability.
Orthogonal results. Table 4 presents the measured results for the nine formulations. The condition viscosity ranged from 5.0 s to 8.3 s. The 24 h suspension rate ranged from 88% to 95%. Gas evolution ranged from 11.0 mL/g to 15.9 mL/g. Wear mass loss ranged from 0.049 g to 0.070 g. Brushability ranged from Grade I to Grade III. These results showed that the formulation variables strongly affected the balance among flow, stability, gas release, and adhesion.
| Run | A | B | C | Viscosity / s | 24 h suspension / % | Gas evolution / (mL·g⁻¹) | Wear loss / g | Brushability grade |
|---|---|---|---|---|---|---|---|---|
| 1 | 1 | 1 | 1 | 5.0 | 88 | 11.0 | 0.067 | III |
| 2 | 1 | 2 | 2 | 6.5 | 94 | 13.4 | 0.056 | I |
| 3 | 1 | 3 | 3 | 7.3 | 91 | 15.9 | 0.052 | II |
| 4 | 2 | 1 | 2 | 6.2 | 90 | 13.8 | 0.063 | I |
| 5 | 2 | 2 | 3 | 8.3 | 93 | 14.2 | 0.049 | III |
| 6 | 2 | 3 | 1 | 5.9 | 95 | 12.6 | 0.069 | I |
| 7 | 3 | 1 | 3 | 8.0 | 89 | 13.5 | 0.050 | III |
| 8 | 3 | 2 | 1 | 6.0 | 94 | 12.0 | 0.070 | II |
| 9 | 3 | 3 | 2 | 7.1 | 93 | 13.2 | 0.058 | II |
Range analysis. I analyzed the orthogonal results using the mean response at each level and the range. For each factor \(j\) and level \(i\), the mean response was
$$\bar{k}_{i,j}=\frac{1}{n_{i,j}}\sum_{r=1}^{n_{i,j}}y_{i,j,r}$$
where \(y_{i,j,r}\) is the response value for the \(r\)-th run at level \(i\) of factor \(j\), and \(n_{i,j}\) is the number of runs at that level. The range was
$$R_j=\max_i(\bar{k}_{i,j})-\min_i(\bar{k}_{i,j})$$
A larger range indicates a stronger influence of the factor on the response. Table 5 summarizes the range analysis for viscosity and 24 h suspension rate. For viscosity, the influence order was binder > aggregate grading > suspension agent. For 24 h suspension rate, the influence order was suspension agent > aggregate grading > binder. This showed that the binder strongly controls flow, while the modified bentonite strongly controls storage stability.
| Response | Factor | \(k_1\) | \(k_2\) | \(k_3\) | Range \(R\) | Influence order |
|---|---|---|---|---|---|---|
| Viscosity / s | A | 6.27 | 6.80 | 7.03 | 0.76 | C > A > B |
| Viscosity / s | B | 6.40 | 6.93 | 6.77 | 0.53 | C > A > B |
| Viscosity / s | C | 5.63 | 6.60 | 7.87 | 2.24 | C > A > B |
| 24 h suspension / % | A | 91.0 | 92.7 | 92.0 | 1.7 | B > A > C |
| 24 h suspension / % | B | 89.0 | 93.7 | 93.0 | 4.7 | B > A > C |
| 24 h suspension / % | C | 92.3 | 92.3 | 91.0 | 1.3 | B > A > C |
Table 6 summarizes gas evolution, wear mass loss, and brushability. For gas evolution, the influence order was binder > suspension agent > aggregate grading. The binder decomposes at high temperature and releases carbon monoxide, carbon dioxide, and light hydrocarbons. Phenolic resin and polyvinyl butyral both contribute to gas evolution. For wear resistance, the influence order was binder > aggregate grading > suspension agent. A higher binder content strengthens the ceramic skeleton, while proper aggregate grading produces a dense “skeleton–filler” structure. For brushability, the influence order was binder > aggregate grading ≈ suspension agent. The binder controls viscosity and adhesion, which directly affect leveling and thickness uniformity.
| Response | Factor | \(k_1\) | \(k_2\) | \(k_3\) | Range \(R\) | Influence order |
|---|---|---|---|---|---|---|
| Gas evolution / (mL·g⁻¹) | A | 13.43 | 13.53 | 12.90 | 0.63 | C > B > A |
| Gas evolution / (mL·g⁻¹) | B | 12.77 | 13.20 | 13.90 | 1.13 | C > B > A |
| Gas evolution / (mL·g⁻¹) | C | 11.87 | 13.47 | 14.53 | 2.66 | C > B > A |
| Wear loss / g | A | 0.0583 | 0.0603 | 0.0593 | 0.0020 | C > A > B |
| Wear loss / g | B | 0.0600 | 0.0583 | 0.0597 | 0.0017 | C > A > B |
| Wear loss / g | C | 0.0687 | 0.0590 | 0.0503 | 0.0184 | C > A > B |
| Brushability grade | A | 2.00 | 1.67 | 2.33 | 0.66 | C > A ≈ B |
| Brushability grade | B | 2.33 | 2.00 | 1.67 | 0.66 | C > A ≈ B |
| Brushability grade | C | 2.00 | 1.33 | 2.67 | 1.34 | C > A ≈ B |
Discussion of individual responses. For condition viscosity, the best level combination was A1B1C1, which gave low viscosity and good flow. However, this combination had a low 24 h suspension rate. For 24 h suspension rate, the best combinations were A2B2C1 or A2B2C2. The modified bentonite formed a card-house network that supported refractory particles in ethanol. The 5% suspension agent level gave the highest suspension rate in most runs. For gas evolution, the best combination was A3B1C1 because lower binder and lower suspension agent contents reduced volatile release. For wear resistance, the best combination was A1B2C3 because higher binder content produced a stronger coating. For brushability, the best combination was A2B3C2 because moderate aggregate grading and higher suspension agent content improved uniformity on both sand and metal.
These individual optima conflicted. A coating for sand mold casting with external chills cannot be optimized for only one property. If I selected A1B1C1, the coating would flow well but settle quickly. If I selected A2B3C2, the coating would brush well but release more gas. If I selected A1B2C3, the coating would be strong but too viscous and gas-prone. I therefore used a weighted normalization method to identify a compromise formulation. I assigned weights based on the practical importance of each property in sand mold casting with external chills: condition viscosity 10%, 24 h suspension rate 25%, gas evolution 20%, wear resistance 10%, and brushability 35%. The suspension rate was treated as a positive indicator. Viscosity, gas evolution, wear mass loss, and brushability grade were treated as negative indicators. For a positive indicator, the normalized value was
$$x’=\frac{x-x_{\min}}{x_{\max}-x_{\min}}$$
For a negative indicator, the normalized value was
$$x’=\frac{x_{\max}-x}{x_{\max}-x_{\min}}$$
The weighted score was calculated as
$$Q=100\sum_{k=1}^{m}w_kx’_k$$
where \(Q\) is the weighted score, \(w_k\) is the weight of the \(k\)-th indicator, \(x’_k\) is its normalized value, and \(m\) is the number of indicators. Table 7 shows the normalized weighted scores for the nine formulations. The highest score was 20.28 for A2B3C1. This formulation used a 53:27 ratio of 250 mesh to 400 mesh bauxite, 5% composite-modified suspension agent, and 2% composite binder. It provided low viscosity, high suspension, good brushability, acceptable gas evolution, and adequate wear resistance.
| Run | Formulation | Weighted normalized score |
|---|---|---|
| 1 | A1B1C1 | 18.24 |
| 2 | A1B2C2 | 19.81 |
| 3 | A1B3C3 | 18.13 |
| 4 | A2B1C2 | 18.76 |
| 5 | A2B2C3 | 18.52 |
| 6 | A2B3C1 | 20.28 |
| 7 | A3B1C3 | 17.70 |
| 8 | A3B2C1 | 19.79 |
| 9 | A3B3C2 | 19.44 |
Optimal formulation. The optimal formulation for sand mold casting with external chills was A2B3C1. Its key properties were a condition viscosity of 5.9 s, a 24 h suspension rate of 95%, a gas evolution of 12.6 mL/g, a wear mass loss of 0.069 g, and a brushability Grade I. The suspension rate and brushability were excellent. The gas evolution was below the typical limit for alcohol-based coatings, and the wear mass loss was far below the accepted maximum of 0.5 g. The formulation met the general requirements for sand mold casting coatings and was specifically adapted to the dual surface environment of sand and external chill. Table 8 summarizes the optimal formulation.
| Parameter | Optimal value |
|---|---|
| Bauxite-to-white corundum mass ratio | 4:1 |
| 250 mesh to 400 mesh bauxite mass ratio | 53:27 |
| Composite-modified suspension agent | 5 wt% of refractory |
| Composite binder | 2 wt% of refractory |
| Phenolic resin to polyvinyl butyral mass ratio | 2.5:1 |
| Solvent | Industrial ethanol |
| Coating density | \(1.60\pm0.02\) g/cm³ |
| Condition viscosity | 5.9 s |
| 24 h suspension rate | 95% |
| Gas evolution | 12.6 mL/g |
| Brushability grade | I |
Mechanism of performance improvement. The improved performance can be explained by three coupled mechanisms. First, the modified bentonite forms a robust suspension network in ethanol. The interlayer spacing increases from 1.26 nm to 3.09 nm, and the swelling index increases to 20.1 mL/g. The expanded layers exfoliate into thin platelets. These platelets interact through edge-to-face and face-to-face contacts, forming a card-house structure. This structure increases the yield stress of the coating and prevents refractory particles from settling. The yield stress can be approximated by a Bingham model:
$$\tau=\tau_0+\mu\dot{\gamma}$$
where \(\tau\) is shear stress, \(\tau_0\) is yield stress, \(\mu\) is plastic viscosity, and \(\dot{\gamma}\) is shear rate. A moderate yield stress is desirable because it supports particles at rest but allows flow during brushing. The 5% suspension agent level provided sufficient yield stress without making the coating too viscous.
Second, the graded refractory aggregate increases packing density and reduces iron penetration. The coarse 250 mesh bauxite particles create a permeable skeleton. The fine 400 mesh particles fill the gaps. White corundum particles strengthen the skeleton and resist chemical attack. The packing fraction can be expressed as
$$\phi_{pack}=\frac{\sum_i V_{s,i}}{\sum_i V_{s,i}+V_v}$$
where \(V_{s,i}\) is the solid volume of particle size class \(i\) and \(V_v\) is the void volume. A higher packing fraction reduces the pore size and the probability of molten iron penetration. It also improves surface finish because the coating surface becomes more uniform. In sand mold casting with external chills, the chill surface is smooth, so any coarse particle protruding from the coating can create a local high point. The 53:27 ratio balanced permeability and smoothness.
Third, the phenolic resin–polyvinyl butyral binder provides complementary properties. Phenolic resin gives high carbon residue and hot strength. Polyvinyl butyral gives flexibility and adhesion. The binder ratio of 2.5:1 produced a continuous film that bonded to sand grains and to the metal chill. During ignition, ethanol burns off and the binder cures rapidly. The coating reaches a dry, strong state within seconds. During pouring, the binder begins to decompose. The decomposition gases must escape through the porous coating and the sand mold. Because the coating is thin and permeable, gas evolution does not cause significant porosity. After cooling, the coating becomes brittle and releases from the metal surface. This release behavior is critical for chill recovery. The thermal stress that drives cracking can be estimated as
$$\sigma=E\alpha\Delta T$$
where \(\sigma\) is thermal stress, \(E\) is elastic modulus, \(\alpha\) is thermal expansion coefficient, and \(\Delta T\) is the temperature change. A moderate elastic modulus and a controlled thermal expansion mismatch help the coating crack and detach without damaging the chill surface.
Production validation. I prepared the optimal coating A2B3C1 at a density of 1.60 g/cm³ and applied it in a large foundry trial. The sand mold was a furan resin-bonded sand mold. The external chills had an exposed area of 25 cm × 15 cm per block and a total surface area of about 1.2 m². The casting was an HT300 machine tool bed. I applied two layers with a total thickness of 0.7 mm. The coating spread uniformly on both the sand mold and the external chill surfaces. After ignition, the coating surface cured within 5 s. The pouring temperature was 1420 °C. After pouring and holding, the mold was shaken out at 150–200 °C. The coating detached easily. The casting surface was clean, and the cleaning time was greatly reduced. The external chills separated easily from the mold, which improved recovery and reuse. Table 9 compares the performance before and after the coating improvement.
| Item | Before improvement | After improvement |
|---|---|---|
| Surface roughness \(Ra\) / μm | 25.0 | 12.5 |
| Sand adhesion area / % | 8.5 | <2 |
| Cleaning time / min | 45 | 25 |
| Chill recovery rate / % | 65 | >90 |
| Gas pore defects with diameter >2 mm | Present | None |
The production results confirmed that the coating solved several problems in sand mold casting with external chills. Sand adhesion decreased from 8.5% to less than 2%. Surface roughness decreased from 25.0 μm to 12.5 μm. Cleaning time decreased from 45 min to 25 min. Chill recovery increased from 65% to more than 90%. No obvious gas pores larger than 2 mm were observed. Metal penetration was also eliminated. These improvements came from the balanced combination of suspension stability, refractory grading, binder chemistry, and fast alcohol drying. The coating did not simply form a barrier; it created a controlled interface that tolerated thermal shock and released cleanly after solidification.
Role of the external chill interface. In sand mold casting with external chills, the chill surface is a heat sink. Heat transfer across the coating–chill interface can be described by
$$q=h_cA(T_m-T_c)$$
where \(q\) is heat flow, \(h_c\) is the interfacial heat transfer coefficient, \(A\) is the contact area, \(T_m\) is the metal temperature, and \(T_c\) is the chill temperature. The coating adds a thermal resistance but also prevents direct contact between the molten metal and the chill. If the coating is too thick, it reduces the chilling effect. If it is too thin, it fails to prevent sand adhesion and metal penetration. The optimal thickness was found to be 0.5–1.2 mm. In production, 0.7 mm provided sufficient protection without significantly reducing the cooling rate. The coating must also maintain a stable interface during filling. The erosion resistance of the coating can be related to its mechanical strength:
$$\tau_{ero}=c+\sigma_n\tan\phi$$
where \(\tau_{ero}\) is the erosion resistance, \(c\) is cohesion, \(\sigma_n\) is normal stress, and \(\phi\) is the internal friction angle. A dense packed aggregate and a strong binder increase both \(c\) and \(\sigma_n\), which improves resistance to molten iron flow. The 250 mesh coarse particles and 400 mesh fine particles together created a compact layer that resisted erosion. The white corundum particles further increased hardness and chemical stability.
Gas evolution and defect prevention. Gas defects are a major risk in alcohol-based coatings for sand mold casting. Gas evolution depends on binder content, solvent residue, and refractory surface chemistry. I measured gas evolution as 12.6 mL/g for the optimal coating, which was below the threshold of 20 mL/g. The low gas evolution resulted from the low binder content of 2% and the use of ethanol rather than water. Ethanol evaporates quickly and burns off during ignition. The phenolic resin and polyvinyl butyral decompose at high temperature, but their total amount is small. The graded refractory aggregate also provides escape paths for gas. If the coating were too dense, gas would be trapped at the interface and form pores. If the coating were too permeable, molten iron could penetrate. The optimal formulation balanced these two effects. The gas pressure at the interface can be approximated by
$$P_g=\frac{nRT}{V}$$
where \(P_g\) is gas pressure, \(n\) is gas amount, \(R\) is the gas constant, \(T\) is temperature, and \(V\) is the available volume. A high \(P_g\) increases the risk of blowing defects. By reducing binder content and maintaining permeability, I kept the local gas pressure below the critical value for pore formation.
Suspension stability and storage. The 24 h suspension rate of 95% indicated that the coating remained homogeneous after storage. This is important for sand mold casting because coatings are often prepared in batches and used over several hours. Sedimentation can cause concentration gradients, leading to uneven coating thickness and inconsistent chilling. The modified bentonite network increased the viscosity at low shear rate and prevented settling. The rheology can be described by the Casson model:
$$\sqrt{\tau}=\sqrt{\tau_0}+\sqrt{\mu_p\dot{\gamma}}$$
where \(\tau_0\) is yield stress and \(\mu_p\) is plastic viscosity. The modified bentonite increased \(\tau_0\) without excessively increasing \(\mu_p\). As a result, the coating had a low condition viscosity of 5.9 s but still maintained a high suspension rate. This combination is difficult to achieve with conventional bentonite. The intercalation–grafting modification was therefore a key factor in the success of the coating for sand mold casting with external chills.
Wear resistance and handling. Coatings for sand mold casting must survive handling after drying. The coated sand mold and chills may be moved, assembled, and cleaned before pouring. Abrasion can remove the coating and expose the sand or metal. The optimal coating had a wear mass loss of 0.069 g, which was much lower than the allowed limit. The low wear loss came from the dense aggregate packing and the phenolic resin–polyvinyl butyral binder. The resin formed a continuous carbonaceous network, while the polyvinyl butyral improved toughness. The white corundum particles increased hardness. The wear resistance can be related to the Archard equation:
$$V_w=K\frac{F_nL}{H}$$
where \(V_w\) is wear volume, \(K\) is a wear coefficient, \(F_n\) is normal load, \(L\) is sliding distance, and \(H\) is hardness. A higher hardness and a lower wear coefficient reduce wear volume. The white corundum and densely packed bauxite increased the effective hardness of the coating. The flexible binder reduced brittle fracture. The result was a coating that remained intact during mold assembly and pouring.
Comparative advantages. Compared with conventional water-based coatings, the developed alcohol-based coating for sand mold casting with external chills dried faster and produced fewer gas defects. Compared with conventional alcohol-based coatings, it adhered better to both sand and metal. The dual adhesion was achieved by the composite binder and the modified bentonite. The phenolic resin provided strong bonding to sand and metal at high temperature. The polyvinyl butyral provided flexibility and green adhesion. The modified bentonite provided suspension stability and controlled rheology. The graded refractory aggregate provided thermal stability and erosion resistance. The combined effect was a coating that could be applied to a complex mold containing both sand and metal chills. Table 10 summarizes the functions of the main components in relation to defect prevention.
| Defect in sand mold casting with external chills | Main cause | Coating function | Key component |
|---|---|---|---|
| Sand adhesion | Molten iron penetrates sand pores and reacts with silica | seals sand surface and reduces wetting | Graded bauxite and white corundum |
| Metal penetration | Large pores in coating or sand | Fills voids and increases packing density | 400 mesh bauxite and modified bentonite |
| Gas pores | High binder or solvent residue | Reduces gas evolution and allows gas escape | Low binder content and ethanol solvent |
| Coating spalling | Thermal expansion mismatch | Provides flexible binder and controlled thermal stress | Polyvinyl butyral and graded aggregate |
| Poor chill adhesion | Low surface energy of metal | Improves wetting and green strength | Phenolic resin–PVB binder |
| Difficult cleaning | Excessive sintering or strong bonding | Forms brittle, releasable layer | Optimized binder ratio and aggregate grading |
Scale-up and production reliability. I validated the formulation in a large casting with a total chill surface area of about 1.2 m². The coating was mixed in a wheel mixer for 2.5–3 h and diluted to the target density. The brushing operation covered vertical and horizontal surfaces. No sagging was observed. The coating remained uniform on both sand and metal. After ignition, the coating cured within 5 s, which is suitable for high-throughput sand mold casting. The pouring temperature of 1420 °C was high enough to test the refractory performance. The coating resisted erosion and did not react with the molten iron. After shakeout, the coating detached from the casting and the chills. The chills were recovered and reused. The surface roughness of the casting was 12.5 μm, which met the requirement for the machine tool bed. The sand adhesion area was less than 2%. No gas pores larger than 2 mm were found. These results showed that the coating is not only a laboratory formulation but also a practical solution for sand mold casting with external chills.
Economic and operational benefits. The improved coating reduced cleaning time by 20 min per casting. It increased chill recovery from 65% to more than 90%. It reduced sand adhesion and metal penetration, which lowered rework and scrap. It also reduced the consumption of new chills and the labor required for surface cleaning. The raw materials are commercially available and compatible with existing alcohol-based coating processes. The mixing procedure is similar to conventional coating preparation, so no major equipment change is required. The formulation can be adjusted by changing the aggregate grading, suspension agent content, and binder content. This flexibility makes it suitable for different sand mold casting conditions, including different casting sizes, pouring temperatures, and chill geometries.
Limitations and control variables. The performance of the coating depends on several variables. The sand mold quality, including binder content and permeability, affects gas escape. The chill surface condition, including roughness and cleanliness, affects adhesion. The coating thickness must be controlled within 0.5–1.2 mm. If the coating is too thin, metal penetration may occur. If it is too thick, the chilling effect may be reduced. The ethanol quality and moisture content also affect drying and ignition. In my trials, the coating density was maintained at \(1.60\pm0.02\) g/cm³. The mixing time was 2.5–3 h. The ignition time was about 5 s. These parameters should be controlled in production to ensure consistent performance in sand mold casting with external chills.
Further optimization. The orthogonal experiment identified the optimal formulation within the tested range. Further improvement could focus on nanoparticle additives, such as nano-alumina or nano-silica, to increase coating strength and reduce gas evolution. Another direction is the use of hybrid solvents to control evaporation rate and improve adhesion on cold chill surfaces. The interaction between the coating and different chill materials, such as gray iron, ductile iron, and steel, could also be studied. The rheological model could be extended to include thixotropy and temperature dependence. The heat transfer coefficient of the coating–chill interface could be measured directly to optimize the chilling effect. These studies would further improve the performance of alcohol-based coatings for sand mold casting with external chills.
Summary of equations. The main equations used in my analysis are summarized below. The suspension rate was
$$S_{24}=\frac{V_s}{V_0}\times100\%$$
The gas evolution was
$$G=\frac{V_g}{m_c}$$
The work of adhesion was
$$W_{ad}=\gamma_L(1+\cos\theta)$$
The Bragg equation was
$$d=\frac{n\lambda}{2\sin\theta}$$
The mean response and range were
$$\bar{k}_{i,j}=\frac{1}{n_{i,j}}\sum_{r=1}^{n_{i,j}}y_{i,j,r}$$
$$R_j=\max_i(\bar{k}_{i,j})-\min_i(\bar{k}_{i,j})$$
The normalized values for positive and negative indicators were
$$x’=\frac{x-x_{\min}}{x_{\max}-x_{\min}}$$
$$x’=\frac{x_{\max}-x}{x_{\max}-x_{\min}}$$
The weighted score was
$$Q=100\sum_{k=1}^{m}w_kx’_k$$
The Bingham and Casson models were
$$\tau=\tau_0+\mu\dot{\gamma}$$
$$\sqrt{\tau}=\sqrt{\tau_0}+\sqrt{\mu_p\dot{\gamma}}$$
The thermal stress and heat flow equations were
$$\sigma=E\alpha\Delta T$$
$$q=h_cA(T_m-T_c)$$
The erosion and wear equations were
$$\tau_{ero}=c+\sigma_n\tan\phi$$
$$V_w=K\frac{F_nL}{H}$$
These equations provided a quantitative framework for understanding the behavior of the coating in sand mold casting with external chills. They also helped me interpret the orthogonal results and select the final formulation.
Final assessment. I developed an alcohol-based coating for sand mold casting with external chills by combining a modified lithium-based bentonite suspension agent, a graded bauxite–white corundum refractory aggregate, a phenolic resin–polyvinyl butyral binder, and industrial ethanol. The two-step intercalation–grafting modification increased the bentonite interlayer spacing and swelling index, which improved suspension stability in ethanol. The orthogonal experiment showed that binder content had the strongest effect on viscosity and gas evolution, while suspension agent content had the strongest effect on 24 h suspension rate. The weighted normalization method identified A2B3C1 as the optimal formulation. This formulation achieved a 24 h suspension rate of 95%, gas evolution of 12.6 mL/g, brushability Grade I, and excellent wear resistance. In production, the coating reduced surface roughness from 25.0 μm to 12.5 μm, reduced sand adhesion from 8.5% to less than 2%, reduced cleaning time from 45 min to 25 min, increased chill recovery from 65% to more than 90%, and eliminated visible gas pores and metal penetration. The coating is therefore a practical and effective solution for sand mold casting with external chills.
The broader significance of this work is that it treats the coating as an engineered interface for sand mold casting. The external chill creates a unique thermal and chemical environment. The coating must not only withstand high temperature but also bond to two dissimilar surfaces and release at the right time. By controlling the suspension network, particle packing, binder chemistry, and solvent evaporation, I achieved a coating that satisfies these requirements. The results demonstrate that a well-designed alcohol-based coating can improve casting quality, reduce cleaning labor, and increase chill reuse in sand mold casting with external chills. I expect that the formulation principles and evaluation method can be extended to other casting processes where external chills are used, including heavy machine tool beds, engine blocks, and large industrial castings.