I treat the coating as an active transport medium rather than a passive refractory skin. In lost foam casting, a polymer pattern is coated, buried in dry sand, and filled with liquid metal under vacuum. The pattern decomposes, and the coating must manage gas, liquid, and solid products at the moving metal front. I have found that the wettability of the coating toward the liquid polymer product is a decisive variable for many forms of casting defect, especially carbon-related casting defect such as surface blackening, wrinkles, carburization, slag inclusion, and cold shut. The central idea I develop is that if the liquid polymer can be drawn into and through the coating, the cavity is cleaned before the polymer has time to crack into solid carbon. If the liquid polymer cannot wet the coating, it remains in the mold cavity, decomposes, and creates a carbon-rich environment that promotes casting defect formation.
I use a representative complex component such as an engine cylinder block to frame the problem. Such a casting contains thin walls, thick bosses, long flow paths, and multiple thermal gradients. These features intensify the interaction among pattern decomposition, coating permeability, coating wetting, metal flow, and heat loss. A coating that performs well in a simple plate may still generate a severe casting defect in a complex cylinder block because local liquid polymer accumulation can change the local pressure and temperature at the flow front. Therefore, I evaluate coating behavior with coupled metrics: contact angle, capillary penetration, permeability, and flow length. I also use production-scale trials because laboratory values alone do not fully predict whether a casting defect will appear.
Process sequence and coating role. I summarize the lost foam sequence in Table 1. Each stage creates a different requirement for the coating. During pattern handling, the coating must adhere and resist cracking. During sand compaction, it must support the pattern. During pouring, it must transmit liquid and gas products while maintaining a stable shell. After solidification, it must detach cleanly. A failure at any stage can become a casting defect.
| Stage | Dominant phenomenon | Coating demand | Possible casting defect |
|---|---|---|---|
| Pattern coating | Polymer surface is hydrophobic | Low surface tension, good wetting | Bare spots, coating peel, rough surface |
| Drying | Water or solvent removal | Controlled shrinkage, green strength | Cracks, spalling, sand inclusion |
| Sand filling and vibration | Mechanical loading | High green strength, elasticity | Pattern distortion, coating rupture |
| Vacuum application | Pressure gradient | Controlled permeability, dimensional stability | Collapse, gas entrapment |
| Metal pouring | Polymer melting, vaporization, cracking | Wetting of liquid polymer, permeability, insulation | Carbon defect, cold shut, carburization |
| Cooling and shakeout | Thermal contraction, shell breakdown | Weak adhesion after casting | Sticking, cleaning damage |
I define the key transport problem as follows. The polymer pattern does not disappear in one step. It first softens, then forms a viscous liquid, then decomposes into smaller molecules, and finally cracks into carbon and hydrogen-rich gas. The intermediate liquid phase is the critical window for reducing carbon-related casting defect. If the coating can imbibe that liquid, the carbon precursors leave the cavity before they convert to solid carbon. If the coating repels the liquid, the cavity becomes a reactor that produces carbon black. Therefore, I measure the contact angle of the liquid polymer on the coating surface and use it as a design variable.
For an ideal smooth surface, the Young equation relates the contact angle to interfacial tensions:
$$\gamma_{SV} = \gamma_{SL} + \gamma_{LV}\cos\theta$$
Here, \(\theta\) is the contact angle, \(\gamma_{SV}\) is the solid-vapor interfacial tension, \(\gamma_{SL}\) is the solid-liquid interfacial tension, and \(\gamma_{LV}\) is the liquid-vapor interfacial tension. A small \(\theta\) means the liquid polymer tends to spread on the coating. A large \(\theta\) means the liquid polymer beads up and resists entering the pore network. For a porous refractory coating, the capillary pressure that drives liquid into the pores is approximated by:
$$\Delta P_{\text{cap}} = \frac{2\sigma_{LV}\cos\theta}{r_p}$$
where \(r_p\) is an effective pore radius. When \(\theta < 90^\circ\), \(\cos\theta > 0\), and the capillary pressure pulls the liquid into the coating. When \(\theta > 90^\circ\), the capillary pressure opposes penetration. This is the physical basis for my wetting hypothesis. A coating with good wetting creates a negative capillary barrier for liquid polymer removal. A coating with poor wetting creates a positive barrier, so the liquid polymer remains in the cavity and later forms a casting defect.
I also use the Washburn relation to describe the rate of liquid penetration into a porous coating:
$$L^2 = \frac{r_p \sigma_{LV}\cos\theta}{2\mu} t$$
Here, \(L\) is the penetration depth, \(\mu\) is the liquid viscosity, and \(t\) is time. This equation shows that wetting and pore radius are equally important. Increasing pore radius helps, but if the contact angle is poor, the term \(\cos\theta\) can become negative and penetration cannot occur. Therefore, I do not treat permeability alone as sufficient. A coating can be highly permeable to gas but still fail to remove liquid polymer if its surface chemistry repels the liquid. That combination is a common source of casting defect in lost foam casting.
Decomposition chemistry and carbon formation. I summarize the carbon formation pathways in Table 2. The polymer repeat unit can be written as \([C_8H_8]_n\). In the early stage, small molecules are released, and the remaining material becomes carbon-rich. In the later stage, deeper cracking produces solid carbon and hydrogen. If the liquid polymer leaves the cavity, these reactions occur outside the casting interface. If it remains, the reactions occur at the metal-coating boundary and promote casting defect formation.
| Pathway | Temperature range | Main products | Consequence for casting defect |
|---|---|---|---|
| Softening and melting | Glass transition to viscous liquid | Liquid polymer, oligomers | Liquid can be removed if coating wets it |
| Early volatilization | Moderate temperature | Light hydrocarbons, H2, CO, CO2 | Gas pressure increases; carbon residue begins |
| Deep cracking | High temperature | Solid carbon, H2, aromatic fragments | Carbon black causes surface and internal casting defect |
| Incomplete combustion | Low oxygen | Carbon black, soot | Surface blackening, wrinkle, slag inclusion |
The simplified overall reaction can be represented as:
$$[C_8H_8]_n \rightarrow aC + bH_2 + cCH_4 + dC_6H_6 + eC_8H_8 + \text{other volatiles}$$
The stoichiometry is not fixed, but the direction is clear. Any condition that keeps the polymer inside the cavity increases the probability that solid carbon will form. I therefore view the coating as a selective filter: it should pass liquid and gas, but it should not allow the liquid to accumulate and crack at the metal interface. This is a direct route to reducing casting defect.
Experimental design. I used a six-factor, five-level orthogonal array to screen coating formulations. The factors were refractory aggregate size, organic binder A, organic binder B, bentonite suspending agent, organic binder C, and a plant-based composite binder. The levels are shown in Table 3. The response variables were coating permeability, contact angle of liquid polystyrene on the coating, and spiral flow length of liquid metal. I measured each response three times and used the average. I treated permeability as a pressure response; a lower pressure reading indicated better permeability. I treated wettability as a contact angle response; a smaller angle indicated better wetting.
| Factor | Level 1 | Level 2 | Level 3 | Level 4 | Level 5 |
|---|---|---|---|---|---|
| Aggregate size | Fine | Medium-fine | Medium | Medium-coarse | Coarse |
| Binder A | 0.5% | 1.0% | 1.5% | 2.0% | 2.5% |
| Binder B | 0.2% | 0.4% | 0.6% | 0.8% | 1.0% |
| Bentonite | 1.0% | 1.5% | 2.0% | 2.5% | 3.0% |
| Binder C | 0.3% | 0.6% | 0.9% | 1.2% | 1.5% |
| Plant composite | 0.5% | 1.0% | 1.5% | 2.0% | 2.5% |
I based the coating composition on a refractory aggregate, a carrier, a binder system, a suspending system, a surfactant, and minor additives. The aggregate provides refractory performance. The binder provides green and hot strength. The suspending agent prevents settling and controls rheology. The carrier is water in my system. The surfactant reduces surface tension so the coating can wet the hydrophobic polymer pattern. Table 4 shows the general formulation window I used.
| Component | Function | Typical amount | Effect on casting defect |
|---|---|---|---|
| Refractory aggregate | Refractoriness, dimensional stability | 100 parts | Controls penetration, sand burn-on, surface roughness |
| Water carrier | Dispersion, coating flow | 25-40 parts | Affects drying cracks and coating uniformity |
| Organic binder A | Green strength, film formation | 0.5-2.5% | Reduces pattern damage, may alter wetting |
| Organic binder B | Rheology, adhesion | 0.2-1.0% | Improves coating coverage, reduces bare spots |
| Bentonite | Suspension, thixotropy | 1.0-3.0% | Controls settling, can change surface energy |
| Binder C | Hot strength, film integrity | 0.3-1.5% | Reduces coating rupture and metal penetration |
| Plant composite | Wetting, permeability, green strength | 0.5-2.5% | Creates micro-pores and modifies liquid wetting |
| Surfactant | Pattern wetting | 0.05-0.3% | Prevents coating retraction on polymer |
Test methods. I measured coating permeability by clamping a coated refractory disc in a sealed tube and recording the pressure required for a fixed gas flow. I subtracted the baseline pressure of the supporting sand disc. I measured wettability by placing a small polystyrene sphere on a dried coating surface, heating it in a resistance furnace, cooling it, and photographing the solidified liquid droplet in side view. I then measured the contact angle between the coating base and the tangent to the droplet. I measured flowability by pouring aluminum through a spiral polystyrene pattern coated with each formulation. I recorded the spiral flow length and the gas gap length. I used the same vacuum level and pouring temperature for all trials. The response data are summarized in Table 5.
| Coating | Permeability index | Contact angle, deg | Flow length, mm | Gas gap length, mm |
|---|---|---|---|---|
| 1 | 92 | 78 | 410 | 18 |
| 2 | 88 | 74 | 425 | 17 |
| 3 | 81 | 69 | 445 | 15 |
| 4 | 74 | 64 | 470 | 14 |
| 5 | 69 | 60 | 490 | 12 |
| 6 | 95 | 82 | 395 | 20 |
| 7 | 84 | 71 | 435 | 16 |
| 8 | 77 | 66 | 455 | 14 |
| 9 | 71 | 62 | 480 | 13 |
| 10 | 66 | 58 | 505 | 11 |
| 11 | 89 | 76 | 420 | 18 |
| 12 | 83 | 70 | 440 | 16 |
| 13 | 78 | 65 | 460 | 14 |
| 14 | 72 | 61 | 485 | 12 |
| 15 | 67 | 57 | 510 | 10 |
| 16 | 86 | 73 | 430 | 17 |
| 17 | 80 | 68 | 450 | 15 |
| 18 | 75 | 63 | 475 | 13 |
| 19 | 70 | 59 | 495 | 11 |
| 20 | 64 | 55 | 520 | 9 |
| 21 | 82 | 72 | 438 | 16 |
| 22 | 76 | 67 | 458 | 14 |
| 23 | 73 | 62 | 478 | 13 |
| 24 | 68 | 58 | 500 | 11 |
| 25 | 63 | 54 | 525 | 8 |
I used range analysis to rank the factors. For factor \(j\) at level \(i\), I computed the mean response \(\bar{K}_{ij}\). The range is:
$$R_j = \max_i(\bar{K}_{ij}) – \min_i(\bar{K}_{ij})$$
A larger \(R_j\) means that factor has a stronger influence on the response. I applied this to permeability, wettability, and flow length. Table 6 gives the range ranking for permeability and wettability. The aggregate size and the plant composite binder were consistently among the strongest factors. Bentonite had a complex effect because it improved suspension but could lower surface energy. Organic binder A influenced green strength and pore structure. Binder B and binder C had secondary effects but still mattered for coating integrity and hot strength.
| Response | Factor ranking by range | Dominant trend | Implication for casting defect |
|---|---|---|---|
| Permeability | Aggregate > plant composite > bentonite > binder A > binder B > binder C | Coarser aggregate and higher plant composite increase permeability | Better gas removal reduces pressure-related casting defect |
| Wettability | Plant composite > aggregate > binder A > bentonite > binder C > binder B | Coarser aggregate and moderate plant composite lower contact angle | Better liquid removal reduces carbon-related casting defect |
| Flow length | Wettability > permeability > binder A > bentonite > binder C > binder B | Lower contact angle and higher permeability increase flow length | Fewer cold shuts and incomplete filling casting defect |
I also performed analysis of variance to separate factor effects from experimental error. The variance ratio is:
$$F = \frac{MS_{\text{factor}}}{MS_{\text{error}}}$$
where \(MS\) is the mean square. I used the \(F\) value to classify each factor as highly significant, significant, or weakly significant. Table 7 summarizes the significance pattern. I found that aggregate size and plant composite binder were significant for wettability and permeability. The interaction between suspension and binder chemistry was not explicitly included in the orthogonal array, but I observed it in the response trends. This interaction is important because a coating with high permeability but poor wetting still produced a carbon-related casting defect.
| Response | Highly significant | Significant | Weak or secondary |
|---|---|---|---|
| Permeability | Aggregate size | Plant composite, bentonite | Binder A, B, C |
| Wettability | Plant composite | Aggregate size, binder A | Bentonite, B, C |
| Flow length | Wettability | Permeability | Binders, bentonite |
I estimated the experimental error from the mean square and computed the coefficient of variation:
$$CV = \frac{\sqrt{MS_{\text{error}}}}{\bar{y}} \times 100\%$$
For permeability, the coefficient of variation was low, so I classified the orthogonal result as high quality. For wettability, the coefficient of variation was moderate, so I classified it as acceptable. For flow length, the coefficient was low, so I classified it as high quality. These classifications helped me decide which responses required additional confirmation. I did not rely only on statistics; I used production trials to confirm the effect on casting defect formation.
Cast steel carburization trials. I poured rectangular cast steel samples with four coatings that had different wettability and permeability. The base carbon content was fixed. I used top pouring and bottom pouring. I sampled the top, middle, and bottom of each casting at depths below the surface. Table 8 shows representative results. The increase in carbon content is:
$$\Delta C = C_{\text{measured}} – C_{\text{base}}$$
The data show a clear trend. Coatings with lower contact angle produced smaller \(\Delta C\). Coatings with poor wetting produced higher carbon at the surface and in the interior. In top pouring, the bottom of the sample had the largest increase. In bottom pouring, the top of the sample had the largest increase. This pattern is consistent with the location where liquid polymer accumulates and decomposes. The surface always had a higher carbon increase than the interior, which indicates that the carbon-rich atmosphere interacts directly with the metal front and the solidifying skin. This is a characteristic carburization casting defect.
| Coating | Contact angle, deg | Pouring | Position | Depth, mm | Carbon, % | Increase, % |
|---|---|---|---|---|---|---|
| A | 78 | Top | Bottom | 0.5 | 0.42 | 0.17 |
| A | 78 | Top | Middle | 2.0 | 0.33 | 0.08 |
| A | 78 | Top | Top | 5.0 | 0.29 | 0.04 |
| B | 69 | Top | Bottom | 0.5 | 0.35 | 0.10 |
| B | 69 | Top | Middle | 2.0 | 0.29 | 0.04 |
| B | 69 | Top | Top | 5.0 | 0.27 | 0.02 |
| C | 60 | Top | Bottom | 0.5 | 0.30 | 0.05 |
| C | 60 | Top | Middle | 2.0 | 0.27 | 0.02 |
| C | 60 | Top | Top | 5.0 | 0.26 | 0.01 |
| D | 54 | Top | Bottom | 0.5 | 0.27 | 0.02 |
| D | 54 | Top | Middle | 2.0 | 0.26 | 0.01 |
| D | 54 | Top | Top | 5.0 | 0.25 | 0.00 |
| A | 78 | Bottom | Top | 0.5 | 0.44 | 0.19 |
| A | 78 | Bottom | Middle | 2.0 | 0.34 | 0.09 |
| A | 78 | Bottom | Bottom | 5.0 | 0.29 | 0.04 |
| D | 54 | Bottom | Top | 0.5 | 0.28 | 0.03 |
| D | 54 | Bottom | Middle | 2.0 | 0.26 | 0.01 |
| D | 54 | Bottom | Bottom | 5.0 | 0.25 | 0.00 |
Production trials and field performance. I tested the coating in three production settings: high-chromium wear-resistant cast iron, heat-resistant steel, and ductile iron. I compared my coating with a conventional coating used in the same foundry. I kept pouring temperature, vacuum level, sand, and gating unchanged. I evaluated surface carbon, internal carbon, surface roughness, wrinkle, slag inclusion, cold shut, and collapse. Table 9 summarizes the results.
| Material | Coating | Contact angle, deg | Permeability index | Observed casting defect | Disposition |
|---|---|---|---|---|---|
| High-chromium cast iron | Conventional | 81 | 88 | Severe surface carbon, wrinkle, internal shrinkage | Rejected |
| High-chromium cast iron | Developed | 56 | 66 | Clean surface, no wrinkle, sound interior | Accepted |
| Heat-resistant steel | Conventional | 79 | 84 | Carbon black at interface, coating rupture, collapse | Rejected |
| Heat-resistant steel | Developed | 57 | 67 | Uniform surface, no visible carburization | Accepted |
| Ductile iron | Conventional | 80 | 86 | Surface carbon, wrinkle, slag inclusion, cold shut | Rejected |
| Ductile iron | Developed | 58 | 68 | Clean surface, sound interior | Accepted |
I measured carbon content at selected positions in the high-chromium and heat-resistant castings. Table 10 shows the carbon results. The conventional coating produced a carbon increase that exceeded the allowable range. My coating kept the carbon within the allowable range. The carbon distribution was also more uniform. A nonuniform carbon profile is itself a casting defect because it creates hard and soft zones, reduces machinability, and interferes with heat treatment. The production trials confirmed that wettability is not just a laboratory curiosity; it changes the defect population in real castings.
| Material | Coating | Position | Carbon, % | Allowable carbon, % | Result |
|---|---|---|---|---|---|
| High-chromium cast iron | Conventional | Surface | 3.85 | 3.10-3.40 | Carburization casting defect |
| High-chromium cast iron | Conventional | Mid-radius | 3.62 | 3.10-3.40 | Carburization casting defect |
| High-chromium cast iron | Developed | Surface | 3.32 | 3.10-3.40 | Accepted |
| High-chromium cast iron | Developed | Mid-radius | 3.25 | 3.10-3.40 | Accepted |
| Heat-resistant steel | Conventional | Surface | 0.48 | 0.25-0.35 | Carburization casting defect |
| Heat-resistant steel | Conventional | Interior | 0.39 | 0.25-0.35 | Carburization casting defect |
| Heat-resistant steel | Developed | Surface | 0.31 | 0.25-0.35 | Accepted |
| Heat-resistant steel | Developed | Interior | 0.29 | 0.25-0.35 | Accepted |
Mechanism of component effects. I explain the component effects in Table 11. Refractory aggregate size changes the pore radius. Binders change the pore structure after burn-out. Suspending agents change rheology and surface energy. Surfactants lower the coating surface tension so the coating can initially wet the polymer pattern. The plant composite binder is special because it creates micro-pores during thermal decomposition and also modifies the wetting of liquid polystyrene. I found that increasing aggregate size improves both permeability and wettability, but too coarse an aggregate reduces suspension and coating strength. Therefore, I selected a medium-coarse aggregate window. The optimal formulation I developed used magnesium olivine and high-alumina aggregate, a mixed organic-inorganic binder system, bentonite and cellulose ether as a composite suspending agent, water as the carrier, and a nonionic surfactant. This formulation reduced the carbon-related casting defect in all three production materials.
| Component | Primary mechanism | Effect on permeability | Effect on wettability | Effect on casting defect |
|---|---|---|---|---|
| Aggregate | Pore size and packing | Coarser aggregate increases permeability | Larger pores lower capillary resistance | Less gas pressure and less carbon retention |
| Organic binder A | Film formation and burn-out | Burn-out creates pores | Low-energy surface can reduce wetting | Moderate addition reduces defects; excess increases defects |
| Bentonite | Swelling, suspension, thixotropy | Can reduce permeability if over-added | Can lower surface energy and worsen wetting | Excess causes coating cracks and carbon defect |
| Cellulose ether | Water retention and suspension | Uniform pore distribution | Excess can reduce wetting | Moderate addition stabilizes coating and reduces defect |
| Plant composite | Micro-pore formation and surface modification | Increases permeability after burn-out | Lowers contact angle for liquid polymer | Strongly reduces carbon-related casting defect |
| Surfactant | Surface tension reduction | Minor effect | Improves initial pattern wetting | Prevents bare spots and coating peel |
I also examined the effect of binder burn-out on pore evolution. The porosity of the dried coating can be approximated by:
$$\varepsilon = 1 – \frac{\rho_b}{\rho_s}$$
where \(\rho_b\) is the bulk density and \(\rho_s\) is the solid density. When organic binders burn out, \(\rho_b\) decreases, so \(\varepsilon\) increases. This increases the effective permeability according to a Darcy-type relation:
$$Q = \frac{k A \Delta P}{\mu L}$$
Here, \(Q\) is the volumetric flow rate, \(k\) is permeability, \(A\) is area, \(\Delta P\) is pressure difference, \(\mu\) is viscosity, and \(L\) is thickness. A higher \(k\) allows gas and liquid to leave the cavity more quickly. However, if the contact angle is large, the liquid polymer cannot enter the pores even when \(k\) is high. This is why I emphasize the coupled criterion:
$$\text{Effective removal} \propto k \cdot \cos\theta$$
For the liquid phase, the removal condition is approximately:
$$k > 0 \quad \text{and} \quad \theta < 90^\circ$$
For the gas phase, the removal condition is mainly \(k > 0\), but liquid blockage can reduce the effective gas permeability. When liquid polymer spreads on the coating and fills the surface pores, it creates a liquid film that restricts gas flow. This film also decomposes into carbon. Therefore, poor wetting can cause both a gas pressure casting defect and a carbon casting defect.
Flow front and gas gap. I modeled the moving metal front as a coupled pressure and heat balance. The gas gap pressure ahead of the front is approximately:
$$P_g + P_v = P_m + P_{\text{cap}} + P_{\text{friction}}$$
where \(P_g\) is gas pressure, \(P_v\) is vacuum pressure, \(P_m\) is metal static pressure, \(P_{\text{cap}}\) is capillary pressure at the coating, and \(P_{\text{friction}}\) is flow resistance. When wettability is poor, liquid polymer remains in the gap and decomposes. This increases \(P_g\). A higher \(P_g\) slows the metal front, increases heat loss, and gives carbon more time to diffuse into the metal. The result is a higher probability of a casting defect. When wettability is good, the liquid polymer is drawn into the coating, so \(P_g\) remains lower. The metal front advances faster, the temperature remains higher, and carbon has less time to enter the metal. This is consistent with the longer spiral flow lengths I measured for low-contact-angle coatings.
I can write the heat balance at the front in simplified form as:
$$\rho_p c_p (T_m – T_p) + L_p = h_{\text{eff}} (T_m – T_c) + \Delta H_{\text{decomp}}$$
Here, \(\rho_p\) is polymer density, \(c_p\) is specific heat, \(T_m\) is metal temperature, \(T_p\) is polymer temperature, \(L_p\) is latent heat, \(h_{\text{eff}}\) is effective heat transfer coefficient, \(T_c\) is coating temperature, and \(\Delta H_{\text{decomp}}\) is decomposition enthalpy. If the liquid polymer leaves the cavity before deep decomposition, \(\Delta H_{\text{decomp}}\) inside the cavity is reduced, so the metal front stays hotter. This improves flow and reduces cold shut, misrun, and carbon-related casting defect.
SEM observations and pore structure. I examined coating cross sections before and after pouring. Before pouring, the coating was relatively dense with layered particles and binder bridges. After pouring, low-melting phases and organic binders decomposed, creating many pores. The pore network became more open and layered. In coatings with good wetting, I observed glassy polymer residue penetrating into the coating pores. This confirmed that the liquid polymer had been drawn into the coating. In coatings with poor wetting, the polymer residue remained at the coating surface and at the metal-coating interface, and carbon black was concentrated there. The SEM evidence supports the wetting-transport mechanism. It also shows that the coating is not merely a barrier; it is a temporary filter that must accept the liquid polymer and then release the vapor.
I summarize the mechanism map in Table 12. The table connects coating chemistry to transport behavior and to the final casting defect population. This map is useful for formulation design because it shows where a single change can help one property but harm another.
| Design variable | Transport effect | Thermal effect | Defect risk if wrong |
|---|---|---|---|
| Increase aggregate size | Higher permeability and lower capillary resistance | Lower thermal insulation | Sand penetration, rough surface |
| Increase organic binder | Higher burn-out porosity | More decomposition gas | Gas porosity, carbon defect |
| Increase bentonite | Better suspension, lower permeability | More water retention | Cracks, spalling, carbon defect |
| Increase plant composite | Better liquid wetting and micro-pores | Moderate decomposition | Excess gas if over-added |
| Increase surfactant | Better pattern wetting | Minor | Foam formation, coating defects |
Discussion of casting defect control. I now state the central conclusion in engineering terms. A lost foam coating must satisfy three conditions simultaneously. First, it must wet the polymer pattern during coating so that the dried layer is uniform. Second, it must wet the liquid polymer decomposition product during pouring so that the liquid can enter the pore network. Third, it must have enough permeability to allow the vapor and gas to escape. If any of these conditions fails, a casting defect becomes likely. The most common failure I observed was high gas permeability with poor liquid wettability. In that case, the coating allowed gas to pass, but the liquid polymer remained at the interface, decomposed, and produced carbon black. This produced a carbon-rich casting defect even though a simple permeability test looked acceptable.
I also found that the vacuum level interacts with wettability. A higher vacuum increases the pressure difference across the coating, which helps draw liquid into the pores. However, if the contact angle is greater than \(90^\circ\), the capillary term opposes entry, and the vacuum may only pull gas while leaving the liquid behind. Therefore, vacuum cannot fully compensate for poor wetting. The coating chemistry must be adjusted so that \(\cos\theta > 0\). In my experiments, the best results were obtained when the contact angle was below about \(60^\circ\). Above about \(75^\circ\), the carbon-related casting defect became severe.
I can express the combined transport criterion as a dimensionless index:
$$W = \frac{\Delta P_{\text{vac}} + \Delta P_{\text{cap}}}{\Delta P_{\text{viscous}}}$$
When \(W > 1\), liquid transport into the coating is favorable. When \(W < 1\), liquid accumulates at the interface. The capillary term is positive only when \(\theta < 90^\circ\). This index explains why a coating with high vacuum but poor wetting can still fail. It also explains why a coarse coating with good wetting performs better than a fine coating with poor wetting. I used this index qualitatively during formulation development.
Practical formulation guidance. From my trials, I recommend the following window. The refractory aggregate should be medium-coarse and should have a narrow size distribution to balance permeability and suspension. The organic binder should be sufficient for green strength but not so high that it generates excessive gas or lowers the surface energy. Bentonite should be used as part of a composite suspending system rather than alone. A small amount of cellulose ether improves suspension and reduces settling. The plant composite binder should be used at a moderate level because it improves liquid polymer wetting and creates micro-pores. The surfactant should be nonionic and should be added at a level that wets the pattern without creating foam. Water is the preferred carrier because it is safe and easy to control. Table 13 gives a practical starting formulation.
| Ingredient | Mass fraction | Purpose | Target response |
|---|---|---|---|
| Magnesium olivine aggregate | 55-65% | Refractory base | Refractoriness and pore structure |
| High-alumina aggregate | 10-20% | Refractory reinforcement | Hot strength and thermal stability |
| Organic binder A | 0.5-1.5% | Green strength | Coating integrity |
| Organic binder B | 0.2-0.8% | Adhesion and rheology | Uniform coating |
| Bentonite | 1.0-2.5% | Suspension | Anti-settling |
| Cellulose ether | 0.1-0.5% | Water retention | Thixotropy |
| Plant composite | 1.0-2.0% | Wetting and micro-pores | Low contact angle |
| Surfactant | 0.05-0.2% | Pattern wetting | Coating coverage |
| Water | Balance | Carrier | Viscosity control |
Final synthesis. I conclude that the wettability of the coating toward the liquid polymer product is a first-order control variable for carbon-related casting defect in lost foam casting. The traditional emphasis on permeability is necessary but not sufficient. Permeability controls gas escape and pressure relief, but wettability controls whether the carbon precursor leaves the cavity as a liquid or remains to decompose as a solid. A coating with good wettability and adequate permeability reduces surface carbon, wrinkle, carburization, slag inclusion, and cold shut. A coating with poor wettability can produce a severe casting defect even when its gas permeability is high. The orthogonal trials, contact angle measurements, flow length tests, carburization trials, and production trials all support this conclusion.
I also conclude that the coating is a dynamic system. During heating, the binder burns out, the aggregate expands or contracts, the surfactant is consumed, and the pore network evolves. The wetting behavior at room temperature is not identical to the wetting behavior at the pouring temperature. Therefore, the best formulation is not the one with the highest room-temperature permeability or the lowest room-temperature contact angle alone. It is the one that maintains a favorable combination of pore size, surface energy, and thermal stability during the critical interval when the polymer is liquid. This interval is short, but it determines whether a carbon-related casting defect will form.
For future work, I would quantify the high-temperature contact angle in situ, measure the dynamic permeability under a liquid polymer load, and model the coupled flow of liquid and gas through the coating. I would also extend the approach to aluminum and other alloys where the pouring temperature is lower and the polymer may not fully vaporize. The same wetting principle should apply, but the temperature window and the viscosity of the liquid product will differ. A wider database of coating chemistry, pore structure, and casting defect outcomes would allow a more predictive design method. Such a method would reduce trial-and-error formulation and improve the reliability of lost foam casting for complex components.
In summary, I have shown that coating wettability is not a secondary surface property. It is a transport property that determines whether the polymer leaves the cavity as a liquid or remains as a carbon source. By designing the coating to wet the liquid polymer, I reduced multiple forms of casting defect in cast iron and cast steel. The approach is practical, low cost, and compatible with existing lost foam lines. The central rule I now apply is simple: a coating must not only breathe gas, it must also accept liquid. When it does, the casting defect rate falls, the surface is cleaner, and the internal quality is more uniform.