In my study, I focused on improving the microstructure and mechanical properties of QT400-18 ductile iron by introducing mechanical vibration during the solidification stage of lost foam casting. I selected this route because lost foam casting is a near-net-shape process with strong advantages in dimensional accuracy, surface quality, and process flexibility, but its dry sand mold often causes slower heat extraction. In my experiments, this slower cooling produced coarse dendrites and relatively low mechanical performance. To address this, I applied vertical mechanical vibration to the sand box and the solidifying melt. My goal was to fragment dendrites, increase heterogeneous nucleation, refine ferrite grains, modify pearlite morphology, and improve graphite spheroidization. I designed my work around two independent vibration variables: frequency and amplitude. I kept amplitude constant at 1.5 mm while changing frequency from 0 Hz to 30 Hz, 40 Hz, and 50 Hz. I then kept frequency constant at 40 Hz while changing amplitude from 0 mm to 1 mm, 1.5 mm, and 2 mm. I compared the resulting graphite morphology, ferrite grain size, pearlite spacing, tensile properties, low-temperature impact toughness, and friction and wear behavior.
I based my first-person experimental reasoning on a simple principle: during lost foam casting, the metal front decomposes the foam pattern, the liquid metal fills the cavity, and the dry sand remains around the casting. The sand does not extract heat as quickly as a metallic mold or a dense green sand mold. As a result, the solidification time becomes longer, the thermal gradient becomes lower, and dendrites have more time to coarsen. When I superimposed vertical vibration, I changed the local pressure, flow, and nucleation conditions in the melt. I observed that this change was not merely a stirring effect. It also affected the graphite growth environment, the austenite shell formation, the eutectic transformation, and the final pearlite morphology. My central hypothesis was that a carefully selected vibration frequency and amplitude in lost foam casting would refine the graphite and matrix structures and thereby improve strength, elongation, low-temperature toughness, and wear resistance.
Experimental Materials and Vibration Procedure
I prepared QT400-18 ductile iron from pig iron, scrap steel, and ferrosilicon. The chemical composition of the charge materials is summarized in Table 1. I used FeSiCaMgRE as the spheroidizing agent and CaBa-FeSi as the inoculant. Their compositions are given in Table 2. I deliberately kept the base carbon and silicon levels consistent among the trials so that the measured differences could be attributed mainly to mechanical vibration rather than to large composition shifts.
| Charge material | C (wt.%) | Si (wt.%) | Mn (wt.%) | S (wt.%) | P (wt.%) | Fe (wt.%) |
|---|---|---|---|---|---|---|
| Pig iron | 4.07 | 0.72 | 0.06 | 0.012 | 0.026 | Balance |
| Scrap steel | 0.05 | 0.01 | 0.26 | 0.005 | 0.014 | Balance |
| Ferrosilicon | 0.035 | 72.15 | — | 0.003 | 0.008 | Balance |
| Additive | Si (wt.%) | Al (wt.%) | Mg (wt.%) | Ca (wt.%) | RE (wt.%) | Ba (wt.%) | Fe (wt.%) |
|---|---|---|---|---|---|---|---|
| Spheroidizing agent | 45.96 | 0.72 | 5.69 | 1.09 | 0.89 | — | Balance |
| Inoculant | 72.99 | 1.15 | — | 1.73 | — | 2.25 | Balance |
I used a two-motor one-dimensional vibration table. The two motors drove eccentric blocks and generated vertical periodic vibration. I selected vertical vibration because it compacts the dry sand and couples strongly to the solidifying melt in lost foam casting. I controlled frequency with a frequency converter and amplitude with the angle between the eccentric blocks. The working frequency range was 0 Hz to 50 Hz, and the working amplitude range was 0 mm to 2 mm. I fixed the sand box to the vibration platform. Before pouring, I started the vibration table and allowed it to reach a stable stage. I poured the melt while the sand box was under a negative pressure of 0.06 MPa. I maintained this negative pressure for 10 min after pouring and then shut down the vacuum system.
For pattern preparation, I cut foam with a resistance wire, bonded the foam into the desired shape, coated it with refractory paint, and dried it three times to ensure uniform coverage. I embedded the coated foam pattern in dry quartz sand with a grain size of 20 to 40 mesh. I compacted the sand through vibration. I covered the sand surface with a plastic film, applied negative pressure, and then covered the film with a thin dry sand layer to prevent liquid metal splashing from damaging the film. I melted the charge in a medium-frequency induction furnace. I added ferrosilicon after the melt reached about 1450 °C to reduce silicon loss. I tapped the melt at about 1500 °C. Before tapping, I checked carbon and silicon with a carbon-silicon analyzer and adjusted the composition if necessary. I placed the spheroidizing agent, inoculant, and iron chips in the ladle in that order from bottom to top, and then I poured the melt into the ladle for spheroidization and inoculation. I then poured the treated melt into the foam pattern under vibration and negative pressure.
I define the vertical vibration displacement as
$$x(t)=A\sin(2\pi f t)$$
where \(A\) is the amplitude and \(f\) is the frequency. The corresponding acceleration is
$$a(t)=-A(2\pi f)^2\sin(2\pi f t).$$
This expression shows that the acceleration magnitude grows with the square of frequency. Therefore, a small increase in frequency can produce a much larger inertial force in the melt. I used this relationship to interpret why 50 Hz produced a stronger effect than 40 Hz, and why the 50 Hz condition began to disturb graphite spheroidization. I also considered the local pressure variation in the melt. During vibration, the melt experiences alternating compression and tension. When the local tensile stress exceeds the cavitation threshold, small cavities form and collapse. The collapse generates local pressure pulses that can detach solid nuclei from the mold wall or from suspended particles. This mechanism is one reason why mechanical vibration promotes nucleation in lost foam casting.
Microstructural Response to Vibration Frequency
I first held amplitude at 1.5 mm and varied frequency. I examined graphite morphology, ferrite grain size, pearlite morphology, and tensile properties. I found that frequency had a strong but non-monotonic effect. In the range from 0 Hz to 40 Hz, increasing frequency improved graphite spheroidization, increased graphite content, reduced average graphite size, refined ferrite, and increased the amount of granular pearlite. At 50 Hz, the average graphite size continued to decrease, but spheroidization decreased. This indicated that excessive vibration energy disturbed the environment required for graphite to grow as spheres. Table 3 summarizes the graphite measurements I obtained with image analysis. I used the relevant graphite evaluation standards and the spheroidal graphite metallographic inspection method as my basis for classification.
| Frequency (Hz) | Spheroidization rate (%) | Spheroidization class | Average graphite diameter (μm) | Graphite content (%) |
|---|---|---|---|---|
| 0 | 78.8 | 4 | 33.86 | 13.58 |
| 30 | 83.1 | 3 | 29.61 | 15.21 |
| 40 | 85.9 | 3 | 27.09 | 16.26 |
| 50 | 82.3 | 3 | 26.44 | 16.41 |
At 0 Hz, I observed relatively few graphite nodules, a larger average size, and many irregular graphite particles. At 30 Hz, graphite content increased and the average size decreased. At 40 Hz, the graphite distribution became more uniform, the number of spherical graphite particles increased, and the average diameter decreased further. At 50 Hz, the graphite became even finer, but the spheroidization rate dropped from 85.9% to 82.3%. I explain this as a competition between nucleation enhancement and growth disturbance. Vibration increases the number of nuclei and promotes detachment of graphite nuclei from heterogeneous substrates. However, if the vibration is too strong, the local flow and pressure pulses disturb the austenite shell that surrounds each graphite particle. The austenite shell is essential for spherical growth because it controls carbon diffusion to the graphite. When the shell is unstable, graphite can grow irregularly.
I also evaluated ferrite grain size. I used the three-circle intercept method to measure the average intercept and then calculated the grain size. I used the following relation for the mean linear intercept of the ferrite phase:
$$\bar l_\alpha=\frac{V_V^\alpha L}{M N_\alpha}$$
where \(V_V^\alpha\) is the volume fraction of ferrite in the matrix, \(L\) is the total measuring line length, \(M\) is the magnification, and \(N_\alpha\) is the average number of intersections. I then calculated the grain size number from
$$G=-6.643856\log_{10}(\bar l_\alpha)-3.288.$$
The results are listed in Table 4. I calculated the ferrite volume fraction in the matrix by correcting for graphite content. For example, at 0 Hz, the matrix fraction was \(1-0.1358=0.8642\), and the ferrite fraction in the matrix was about 0.7901, giving \(V_V^\alpha \approx 0.6828\). I repeated this correction for each condition. The average intercept decreased as frequency increased from 0 Hz to 40 Hz, and it decreased slightly further at 50 Hz. The grain size number increased from 6.87 to 7.41. This confirmed that mechanical vibration refined the ferrite grains in lost foam casting.
| Frequency (Hz) | Intersections \(N_\alpha\) | \(V_V^\alpha\) | Mean intercept \(\bar l_\alpha\) (mm) | Grain size number \(G\) | Average ferrite size (μm) |
|---|---|---|---|---|---|
| 0 | 58 | 0.6828 | 0.029598 | 6.87 | 33.76 |
| 30 | 65 | 0.7057 | 0.026951 | 7.14 | 31.85 |
| 40 | 72 | 0.7142 | 0.024843 | 7.37 | 29.93 |
| 50 | 74 | 0.7201 | 0.024542 | 7.41 | 28.62 |
The ferrite grain size distribution shifted to smaller sizes as frequency increased. At 0 Hz, the distribution was broad, with many grains between 20 μm and 60 μm and a small number of coarse grains between 70 μm and 90 μm. At 30 Hz, the fraction of 30 to 60 μm grains decreased, and the fraction of 20 to 30 μm grains increased. At 40 Hz, most grains were between 20 μm and 30 μm, and the fraction of 40 to 60 μm grains decreased. At 50 Hz, the average size reached 28.62 μm. I attribute this refinement to two effects. First, vibration fragments dendrites and creates new nuclei. Second, vibration increases the effective undercooling and promotes nucleation. The critical radius for nucleation can be written as
$$r^*=-\frac{2\gamma T_m}{\Delta H_f \Delta T}$$
where \(\gamma\) is the interfacial energy, \(T_m\) is the melting point, \(\Delta H_f\) is the latent heat of fusion, and \(\Delta T\) is the undercooling. When vibration increases \(\Delta T\), \(r^*\) decreases. Therefore, more nuclei become stable, and the final grain size becomes smaller.
Pearlite morphology also changed with frequency. At 0 Hz, pearlite was mainly lamellar. At 30 Hz and 40 Hz, the lamellar spacing decreased, and short rod and granular pearlite became more common. At 50 Hz, granular pearlite increased further, but the average pearlite spacing increased again. I measured the pearlite layer spacing and obtained 0.34 μm at 0 Hz, 0.27 μm at 30 Hz, 0.22 μm at 40 Hz, and 0.29 μm at 50 Hz. This trend is summarized in Table 5. The minimum spacing occurred at 40 Hz. I explain the granular pearlite formation through the colloidal equilibrium concept. The carbon concentration at the tip of a cementite lamella is higher than at the flat side. Vibration accelerates carbon diffusion in ferrite, destroys the local concentration balance, and causes the lamella to dissolve at the tip and grow on the flat side. As a result, lamellar pearlite gradually changes into short rod and granular pearlite.
| Frequency (Hz) | Pearlite layer spacing (μm) | Dominant morphology |
|---|---|---|
| 0 | 0.34 | Lamellar |
| 30 | 0.27 | Lamellar with short rods |
| 40 | 0.22 | Fine lamellar and granular |
| 50 | 0.29 | Granular with coarser spacing |
Tensile Behavior under Frequency Variation
I performed tensile tests at a crosshead speed of 1 mm/min. I tested three specimens for each condition and used the average. The elongation was calculated as
$$A=\frac{L_u-L_0}{L_0}\times 100\%$$
where \(L_u\) is the final gauge length and \(L_0\) is the original gauge length. The tensile results are listed in Table 6. At 0 Hz, the tensile strength was 419.5 MPa and the elongation was 18.3%. At 30 Hz, the tensile strength increased to 446.5 MPa and the elongation increased to 21.1%. At 40 Hz, the tensile strength reached 457.1 MPa and the elongation reached 22.2%. At 50 Hz, the tensile strength decreased slightly to 452.3 MPa and the elongation decreased to 20.6%. The maximum tensile strength and elongation occurred at 40 Hz.
| Frequency (Hz) | Tensile strength (MPa) | Elongation (%) | Change in tensile strength vs. 0 Hz (MPa) | Change in elongation vs. 0 Hz (%) |
|---|---|---|---|---|
| 0 | 419.5 | 18.3 | — | — |
| 30 | 446.5 | 21.1 | +27.0 | +2.8 |
| 40 | 457.1 | 22.2 | +37.6 | +3.9 |
| 50 | 452.3 | 20.6 | +32.8 | +2.3 |
The improvement in tensile strength came from several microstructural changes. First, spheroidization increased, so the stress concentration caused by irregular graphite decreased. Second, graphite content increased while average graphite size decreased. Fine spherical graphite is less damaging to the matrix than large irregular graphite. Third, ferrite grains were refined. According to the Hall-Petch relationship,
$$\sigma_y=\sigma_0+k_y d^{-1/2}$$
where \(\sigma_y\) is the yield strength, \(\sigma_0\) is the friction stress, \(k_y\) is the strengthening coefficient, and \(d\) is the grain size. A smaller ferrite grain size increases strength. Fourth, the pearlite spacing decreased and granular pearlite increased. The interfaces in granular pearlite are smaller than those in lamellar pearlite, so dislocations can move more easily during plastic deformation. This contributed to higher elongation. At 50 Hz, the spheroidization rate decreased, and irregular graphite reappeared. These irregular particles cut the matrix and reduced both strength and elongation.
I also examined the tensile fracture surfaces. At 0 Hz, the fracture surface contained large exposed graphite nodules, a small number of shallow dimples, and some cleavage facets with river patterns. This indicated a mixed fracture mode with a significant brittle component. At 30 Hz, the graphite nodules on the fracture surface became smaller and more numerous, dimples increased, and cleavage features decreased. At 40 Hz, the fracture surface was dominated by dimples centered on graphite nodules or on the pits left by detached graphite. Tear ridges became more obvious. These tear ridges form when dislocation motion and slip bands are blocked during plastic deformation. At 50 Hz, the number of dimples decreased slightly, and some river patterns reappeared, but the fracture still had a large plastic deformation component.
I used laser confocal microscopy to reconstruct the tensile fracture surfaces in three dimensions. The height difference reflects the degree of plastic deformation. At 0 Hz, the height difference was 711.9 μm. At 30 Hz, it increased to 838.7 μm. At 40 Hz, it reached 934.2 μm. At 50 Hz, it decreased to 825.3 μm. I also measured the surface roughness using
$$S_a=\frac{1}{A}\iint_A |z(x,y)|\,dx\,dy.$$
The roughness index \(S_a\) followed the same trend as elongation: it increased from 0 Hz to 40 Hz and then decreased at 50 Hz. This consistency supported my conclusion that the 40 Hz condition produced the best combination of strength and ductility under a constant amplitude of 1.5 mm in lost foam casting.
Microstructural Response to Vibration Amplitude
After identifying 40 Hz as the best frequency at 1.5 mm amplitude, I fixed the frequency at 40 Hz and varied the amplitude from 0 mm to 1 mm, 1.5 mm, and 2 mm. I again examined graphite morphology, ferrite grain size, pearlite morphology, and tensile properties. The graphite measurements are shown in Table 7. At 0 mm, the spheroidization rate was 78.8%, the average graphite diameter was 33.86 μm, and the graphite content was 13.58%. At 1 mm, the spheroidization rate increased to 81.3%, the average diameter decreased to 30.79 μm, and the graphite content increased to 14.66%. At 1.5 mm, the spheroidization rate reached 85.9%, the average diameter decreased to 27.09 μm, and the graphite content increased to 16.26%. At 2 mm, the spheroidization rate increased slightly to 86.2%, the average diameter decreased to 26.87 μm, and the graphite content increased to 16.32%. The improvement from 1.5 mm to 2 mm was small, indicating that the beneficial effect of amplitude began to level off.
| Amplitude (mm) | Spheroidization rate (%) | Spheroidization class | Average graphite diameter (μm) | Graphite content (%) |
|---|---|---|---|---|
| 0 | 78.8 | 4 | 33.86 | 13.58 |
| 1 | 81.3 | 3 | 30.79 | 14.66 |
| 1.5 | 85.9 | 3 | 27.09 | 16.26 |
| 2 | 86.2 | 3 | 26.87 | 16.32 |
I observed that increasing amplitude increased the force acting on the graphite nuclei. At constant frequency, the vibration acceleration is proportional to amplitude, as shown by
$$a_{\max}=A(2\pi f)^2.$$
Therefore, a larger amplitude means a larger inertial force and a stronger disturbance in the melt. This stronger disturbance promotes compositional homogenization, increases the concentration difference between the graphite-liquid interface and the bulk liquid, and enhances graphite nucleation. As a result, graphite content increases and average size decreases. However, once the amplitude becomes large enough, the growth environment can be disturbed, and further improvement becomes limited. In my experiments, 1.5 mm and 2 mm amplitudes both produced good graphite morphology, but 2 mm gave only a slight additional benefit.
I also measured ferrite grain size under amplitude variation. The results are listed in Table 8. At 0 mm, the grain size number was 6.87, and the average ferrite size was 33.76 μm. At 1 mm, the grain size number increased to 7.09, and the average size decreased to 32.24 μm. At 1.5 mm, the grain size number reached 7.37, and the average size decreased to 29.93 μm. At 2 mm, the grain size number reached 7.47, and the average size decreased to 29.14 μm. The grain refinement was consistent with the graphite refinement. When graphite content increases, carbon is consumed by more graphite particles, so the amount of carbon available to form cementite decreases. Since pearlite is a mixture of ferrite and cementite, less cementite favors more ferrite. At the same time, the increased number of graphite nuclei promotes heterogeneous nucleation of ferrite. Both effects refine the ferrite matrix.
| Amplitude (mm) | Intersections \(N_\alpha\) | \(V_V^\alpha\) | Mean intercept \(\bar l_\alpha\) (mm) | Grain size number \(G\) | Average ferrite size (μm) |
|---|---|---|---|---|---|
| 0 | 58 | 0.6828 | 0.029598 | 6.87 | 33.76 |
| 1 | 65 | 0.7078 | 0.027416 | 7.09 | 32.24 |
| 1.5 | 72 | 0.7142 | 0.024843 | 7.37 | 29.93 |
| 2 | 75 | 0.7195 | 0.024023 | 7.47 | 29.14 |
Pearlite morphology also responded to amplitude. At 0 mm, lamellar pearlite was dominant. At 1 mm, the lamellar spacing decreased slightly. At 1.5 mm, the spacing decreased further, and short rod and granular pearlite became more common. At 2 mm, the pearlite became finer, and granular pearlite increased further. Because vibration increases heat transfer and promotes faster solidification, iron and carbon atoms have less time to diffuse over long distances. This reduces the pearlite spacing. In addition, the carbon concentration balance at the cementite tips is disturbed, which promotes dissolution at the tips and growth on the flat sides. This transforms lamellar pearlite into granular pearlite. The trend is summarized in Table 9.
| Amplitude (mm) | Pearlite layer spacing trend | Dominant morphology |
|---|---|---|
| 0 | Largest | Lamellar |
| 1 | Slightly reduced | Lamellar with short rods |
| 1.5 | Further reduced | Fine lamellar and granular |
| 2 | Smallest | Fine granular with some lamellar |
Tensile Behavior under Amplitude Variation
The tensile results under amplitude variation are shown in Table 10. At 0 mm, the tensile strength was 419.5 MPa and the elongation was 18.3%. At 1 mm, the tensile strength increased to 440.3 MPa and the elongation increased to 20.8%. At 1.5 mm, the tensile strength reached 457.1 MPa and the elongation reached 22.2%. At 2 mm, the tensile strength reached 462.8 MPa and the elongation reached 22.8%. The improvement from 1.5 mm to 2 mm was small, which matched the graphite and ferrite measurements. I concluded that 40 Hz and 2 mm gave the best tensile properties among the conditions I tested, but the 1.5 mm condition already captured most of the benefit.
| Amplitude (mm) | Tensile strength (MPa) | Elongation (%) | Change in tensile strength vs. 0 mm (MPa) | Change in elongation vs. 0 mm (%) |
|---|---|---|---|---|
| 0 | 419.5 | 18.3 | — | — |
| 1 | 440.3 | 20.8 | +20.8 | +2.5 |
| 1.5 | 457.1 | 22.2 | +37.6 | +3.9 |
| 2 | 462.8 | 22.8 | +43.3 | +4.5 |
I explained the tensile improvement using the same mechanisms as in the frequency study. The graphite became more spherical, smaller, and more numerous. The ferrite grains became finer. The pearlite spacing decreased, and granular pearlite increased. All these changes delayed crack initiation, reduced stress concentration, and allowed more plastic deformation before fracture. In the fracture surfaces, I observed more dimples and deeper dimples at higher amplitude. At 2 mm, the fracture surface was dominated by dimples centered on graphite nodules or on pits left by detached graphite. Cleavage facets were almost absent. This indicated a ductile fracture mode. The three-dimensional fracture reconstruction also showed that the height difference increased with amplitude. At 0 mm, the height difference was 711.9 μm. At 1 mm, it increased to 801.3 μm. At 1.5 mm, it reached 934.2 μm. At 2 mm, it reached 951.3 μm. The roughness index \(S_a\) followed the same trend as elongation. These results confirmed that increasing amplitude at 40 Hz improved the tensile ductility of QT400-18 in lost foam casting.
Graphite Precipitation under Mechanical Vibration
I used thermal analysis to understand why vibration changed graphite precipitation. I compared the 40 Hz, 2 mm condition with the no-vibration condition. I found that the eutectic temperature in the vibration condition was lower, and the eutectic transformation time was shorter. The lower eutectic temperature means that solidification began at a higher effective undercooling. The undercooling can be expressed as
$$\Delta T=T_m-T.$$
A larger \(\Delta T\) reduces the critical radius \(r^*\), as shown earlier. Therefore, more graphite nuclei can survive and grow. A shorter eutectic transformation time means that atoms have less time to diffuse over long distances. This favors fine graphite and fine matrix structures. In lost foam casting, the slow cooling caused by dry sand usually gives atoms more time to diffuse and dendrites more time to coarsen. Mechanical vibration counteracts this tendency by increasing undercooling and shortening the transformation time.
I also considered the solute balance during graphite growth. In ductile iron, the partition coefficient of carbon is greater than unity, so carbon is rejected from the solid into the liquid during solidification. The graphite nucleus grows by absorbing carbon atoms that diffuse through the liquid and across the solid-liquid interface. The carbon flux can be described qualitatively by Fick’s first law:
$$J_C=-D_C\nabla C.$$
When vibration increases the concentration gradient \(\nabla C\) near the graphite-liquid interface and reduces the diffusion boundary layer thickness, carbon transport to the graphite nucleus becomes more uniform. This favors spherical growth. In addition, vibration can detach weakly attached graphite nuclei from heterogeneous substrates. These detached nuclei become new nucleation sites. The result is more graphite nodules, a smaller average size, and a higher spheroidization rate. However, if the vibration is too strong, the austenite shell around the graphite may not close quickly or may be disturbed. The austenite shell controls carbon diffusion to the graphite. A stable shell promotes uniform spherical growth. An unstable shell leads to irregular graphite. This explains why 50 Hz at 1.5 mm amplitude produced finer graphite but lower spheroidization.
Low-Temperature Impact Behavior
I selected the 40 Hz, 2 mm condition as the best vibration condition and compared it with the no-vibration condition for low-temperature impact testing. I used Charpy V-notch specimens with dimensions of 55 mm × 10 mm × 10 mm, a notch depth of 2 mm, a notch angle of 45°, and a notch root radius of 0.25 mm. I tested at 20 °C, 0 °C, -20 °C, -40 °C, -50 °C, and -60 °C. I added the -50 °C condition because the impact energy dropped sharply between -40 °C and -60 °C. The impact energy results are summarized in Table 11.
| Temperature (°C) | No vibration, average impact energy (J) | Vibration at 40 Hz and 2 mm, average impact energy (J) | Improvement (J) |
|---|---|---|---|
| 20 | 18.23 | 19.86 | 1.63 |
| 0 | 16.17 | 17.18 | 1.01 |
| -20 | 14.16 | 15.46 | 1.30 |
| -40 | 12.19 | 13.17 | 0.98 |
| -50 | 8.14 | 11.74 | 3.60 |
| -60 | 4.48 | 7.67 | 3.19 |
I observed that the impact energy decreased as temperature decreased for both conditions. The vibration condition had higher impact energy at every temperature. The ductile-to-brittle transition temperature was about -40 °C for the no-vibration condition and about -50 °C for the vibration condition. This shift is important because it means that the vibration-treated QT400-18 produced by lost foam casting can serve at lower temperatures without brittle failure. At -60 °C, both conditions became brittle, and the difference in impact energy decreased, but the vibration condition still showed higher energy absorption.
I examined the impact fracture surfaces. At 20 °C, both conditions showed many dimples centered on graphite nodules or on pits left by detached graphite. As temperature decreased, the number of dimples decreased, and cleavage facets increased. At -40 °C, the no-vibration condition began to show large cleavage areas, while the vibration condition still retained a mixed ductile-brittle morphology. At -50 °C, the vibration condition showed a transition morphology with both dimples and cleavage facets, whereas the no-vibration condition had already entered a more brittle regime. At -60 °C, both conditions showed brittle fracture with river patterns and cleavage steps.
I explain the improved low-temperature toughness using two main factors. First, the vibration condition had a higher graphite spheroidization rate, smaller graphite size, and more uniform graphite distribution. Irregular graphite acts as a stress concentrator and a crack initiation site. Spherical graphite reduces stress concentration and delays crack initiation. Second, the vibration condition had finer ferrite grains. A coarse ferrite grain contains a larger effective cleavage fracture unit. In low-temperature impact, a larger cleavage unit has a higher probability of containing a critical crack nucleus. A finer ferrite grain reduces this probability and increases toughness. The three-dimensional fracture reconstruction supported this interpretation. At the same temperature, the vibration condition had a larger height difference, which indicated more plastic deformation before fracture. The surface roughness also decreased with decreasing temperature, and the vibration condition had higher roughness at each temperature, consistent with higher impact energy.
Friction and Wear Behavior
I also studied the friction and wear behavior of the no-vibration condition and the 40 Hz, 2 mm vibration condition. I used a load of 50 N, a frequency of 2 Hz, a sliding distance of 10 mm, and a wear time of 30 min. The counterbody was a 6 mm diameter tungsten carbide ball. I measured hardness before and after wear, weight loss, friction coefficient, and wear surface morphology. The hardness results are shown in Table 12. Before wear, the no-vibration specimen had a Brinell hardness of 149.1 HB, and the vibration specimen had a Brinell hardness of 144.4 HB. After wear, the no-vibration specimen increased to 155.4 HB, and the vibration specimen increased to 158.5 HB.
| Condition | Hardness before wear (HB) | Average hardness before wear (HB) | Hardness after wear (HB) | Average hardness after wear (HB) |
|---|---|---|---|---|
| No vibration | 148.3, 146.7, 152.2 | 149.1 | 153.4, 154.3, 158.5 | 155.4 |
| Vibration at 40 Hz and 2 mm | 141.6, 147.2, 144.3 | 144.4 | 154.4, 159.7, 161.3 | 158.5 |
The hardness difference before wear was small. The no-vibration specimen had more pearlite, which increased its hardness. The vibration specimen had less pearlite but finer pearlite spacing and more granular pearlite. These two effects nearly balanced each other. After wear, both specimens hardened because of plastic deformation and strain hardening at the worn surface. The vibration specimen showed a larger increase in hardness, from 144.4 HB to 158.5 HB. I attribute this to the more stable hardened layer formed on the vibration specimen. Because the graphite in the vibration specimen was more spherical, stress was distributed more uniformly, and the hardened layer was less likely to crack and spall.
The weight loss and friction coefficient results are shown in Table 13. The no-vibration specimen had an average weight loss of 0.04369 g and a friction coefficient of 0.912. The vibration specimen had an average weight loss of 0.04191 g and a friction coefficient of 0.803. The vibration specimen lost 0.00178 g less mass and had a lower friction coefficient. I calculated the wear rate as
$$W=\frac{\Delta m}{F_n s}$$
where \(\Delta m\) is the mass loss, \(F_n\) is the normal load, and \(s\) is the sliding distance. Because the load and sliding distance were the same, the lower mass loss directly indicated a lower wear rate. The friction coefficient was calculated as
$$\mu=\frac{F_t}{F_n}.$$
A lower \(\mu\) means lower tangential force for the same normal load, which indicates better lubricating conditions at the contact interface.
| Condition | Weight loss (g) | Average weight loss (g) | Friction coefficient |
|---|---|---|---|
| No vibration | 0.04331, 0.04357, 0.04418 | 0.04369 | 0.912 |
| Vibration at 40 Hz and 2 mm | 0.04145, 0.04261, 0.04167 | 0.04191 | 0.803 |
I explain the improved wear resistance by two mechanisms. First, graphite acts as a solid lubricant in ductile iron. During sliding, graphite is dragged out of the matrix and smears over the worn surface, forming a lubricating layer. The vibration specimen had a higher graphite content and a higher spheroidization rate. Therefore, more graphite was available to form a stable lubricating layer. Second, irregular graphite promotes stress concentration and crack initiation. The no-vibration specimen contained more irregular graphite, so microcracks formed more easily in the matrix during sliding. These microcracks caused the hardened layer to spall, exposing fresh metal and increasing wear. The vibration specimen had more spherical graphite, which distributed stress more uniformly and stabilized the hardened layer. As a result, the vibration specimen had a lower friction coefficient and less material loss.
I examined the worn surfaces in the running-in stage and the stable wear stage. In the running-in stage, both specimens showed plastic deformation and some oxidation. The no-vibration specimen showed more iron oxide debris and more severe surface damage. In the stable wear stage, the no-vibration specimen showed deeper ploughing grooves, while the vibration specimen showed shallower grooves and a more continuous hardened layer. The three-dimensional reconstruction of the worn surfaces confirmed that the no-vibration specimen had a larger height difference and deeper wear tracks. The vibration specimen had a smoother worn surface and a shallower wear depth. These observations agreed with the weight loss and friction coefficient measurements.
Discussion of Mechanisms in Lost Foam Casting
I now summarize the mechanisms by which mechanical vibration improved QT400-18 in lost foam casting. The first mechanism is dendrite fragmentation. During solidification, the melt contains growing dendrites. Vibration causes relative motion between the solid and liquid phases. The shear stress at the dendrite root can exceed the critical value and break the dendrite. The broken dendrite arms become new nuclei. This increases the nucleation rate and refines the final structure. The second mechanism is cavitation. Alternating pressure in the melt can cause cavitation bubbles to form and collapse. The collapse generates local pressure pulses that detach nuclei from the mold wall or from suspended particles. This also increases the number of active nuclei. The third mechanism is increased undercooling. Vibration enhances heat transfer and can increase the effective undercooling. A larger undercooling reduces the critical radius for nucleation, so more nuclei survive. The fourth mechanism is homogenization. Vibration promotes solute mixing and reduces compositional gradients. This makes the graphite growth environment more uniform. The fifth mechanism is austenite shell stabilization. A stable austenite shell around graphite promotes spherical growth by controlling carbon diffusion. Moderate vibration helps the shell form quickly and uniformly. Excessive vibration disturbs the shell and leads to irregular graphite.
In lost foam casting, these mechanisms are especially important because the dry sand mold naturally slows heat extraction. Without vibration, the slow cooling produces coarse dendrites, large graphite, and coarse ferrite. With vibration, the process becomes more competitive with faster-cooling casting methods. My results show that the best condition among those I tested was 40 Hz and 2 mm. At this condition, the graphite spheroidization rate was high, the average graphite diameter was small, the ferrite grain size was refined, the pearlite spacing was reduced, and the tensile properties were improved. The impact toughness and wear resistance were also improved. However, I also found that too much frequency can be harmful. At 50 Hz, the spheroidization rate decreased, and the tensile properties decreased. This means that the vibration parameters must be optimized rather than simply maximized.
Summary of Quantitative Results
For clarity, I present a consolidated comparison in Table 14. The table compares the no-vibration condition with the best vibration condition, which was 40 Hz and 2 mm. The improvements include higher graphite spheroidization, smaller graphite size, higher graphite content, smaller ferrite grain size, lower pearlite spacing, higher tensile strength, higher elongation, higher impact energy, lower weight loss, and lower friction coefficient.
| Property | No vibration | Vibration at 40 Hz and 2 mm | Improvement or change |
|---|---|---|---|
| Graphite spheroidization rate (%) | 78.8 | 86.2 | +7.4 percentage points |
| Average graphite diameter (μm) | 33.86 | 26.87 | -6.99 μm |
| Graphite content (%) | 13.58 | 16.32 | +2.74 percentage points |
| Average ferrite size (μm) | 33.76 | 29.14 | -4.62 μm |
| Pearlite layer spacing trend | Largest | Smallest | Refined |
| Tensile strength (MPa) | 419.5 | 462.8 | +43.3 MPa |
| Elongation (%) | 18.3 | 22.8 | +4.5 percentage points |
| Impact energy at 20 °C (J) | 18.23 | 19.86 | +1.63 J |
| Ductile-to-brittle transition temperature (°C) | About -40 | About -50 | Lower by about 10 °C |
| Weight loss (g) | 0.04369 | 0.04191 | -0.00178 g |
| Friction coefficient | 0.912 | 0.803 | -0.109 |
I also summarize the frequency effect and amplitude effect in a compact form. Frequency has a non-monotonic effect: 40 Hz is better than 30 Hz and 50 Hz under 1.5 mm amplitude. Amplitude has a monotonic or nearly monotonic effect in the range I tested: increasing amplitude from 0 mm to 2 mm at 40 Hz improves the microstructure and properties, but the improvement from 1.5 mm to 2 mm is small. This suggests that 1.5 mm to 2 mm is near the practical saturation range for the process window I studied. If the amplitude were increased further, I would expect stronger disturbance, possible graphite degeneration, and possibly more casting defects. Therefore, I recommend 40 Hz and 2 mm as the best condition among those tested, with 40 Hz and 1.5 mm as a nearly equivalent condition with slightly lower energy input.
Practical Implications for Lost Foam Casting
My results have several practical implications for lost foam casting of QT400-18. First, mechanical vibration can be integrated into the existing compaction stage of lost foam casting. The same vibration table used for sand compaction can be used for solidification vibration if the frequency and amplitude are controlled. This reduces equipment cost and process complexity. Second, the negative pressure level of 0.06 MPa can be maintained during vibration. In my trials, this negative pressure improved foam decomposition gas removal and helped avoid gas defects. Third, the vibration should start before pouring and continue until the casting has solidified enough to avoid disturbance. In my process, I maintained vibration through pouring and then kept the negative pressure for 10 min after pouring. Fourth, the vibration parameters should be matched to the casting geometry and section thickness. Thick sections remain liquid longer and may benefit from longer vibration, while thin sections may solidify quickly and require less vibration energy. Fifth, excessive frequency should be avoided. In my study, 50 Hz produced finer graphite but lower spheroidization and lower tensile properties. Therefore, a frequency around 40 Hz appears to be a better choice for QT400-18 in lost foam casting.
I also note that the vibration amplitude and frequency jointly determine the vibration intensity. The maximum acceleration is
$$a_{\max}=A(2\pi f)^2.$$
For example, at 40 Hz and 2 mm, \(a_{\max}\) is about \(0.002 \times (2\pi \times 40)^2 \approx 126.3\ \text{m/s}^2\). At 50 Hz and 1.5 mm, \(a_{\max}\) is about \(0.0015 \times (2\pi \times 50)^2 \approx 148.0\ \text{m/s}^2\). The 50 Hz condition therefore has higher acceleration intensity than the 40 Hz, 2 mm condition. This helps explain why 50 Hz caused more disturbance. The acceleration scales with the square of frequency, so frequency has a stronger effect on vibration intensity than amplitude. This is an important practical point for process control in lost foam casting.
Conclusions
I conducted a systematic study of mechanical vibration during lost foam casting of QT400-18 ductile iron. Based on my microstructural observations, tensile tests, low-temperature impact tests, and friction and wear tests, I draw the following conclusions.
First, under a constant amplitude of 1.5 mm, increasing frequency from 0 Hz to 40 Hz improved graphite spheroidization, increased graphite content, decreased average graphite size, refined ferrite grains, and increased granular pearlite. At 50 Hz, graphite size continued to decrease, but spheroidization decreased because the strong vibration disturbed the spherical growth environment. The tensile strength and elongation increased from 0 Hz to 40 Hz and then decreased at 50 Hz. The best frequency under 1.5 mm amplitude was 40 Hz.
Second, under a constant frequency of 40 Hz, increasing amplitude from 0 mm to 2 mm improved graphite spheroidization, increased graphite content, decreased graphite size, refined ferrite grains, and increased granular pearlite. The tensile strength and elongation increased with amplitude. The improvement from 1.5 mm to 2 mm was small, indicating that the beneficial effect was approaching saturation. The best condition among those I tested was 40 Hz and 2 mm, which gave a tensile strength of 462.8 MPa and an elongation of 22.8%. Compared with the no-vibration condition, the tensile strength increased by 43.3 MPa and the elongation increased by 4.5 percentage points.
Third, the 40 Hz and 2 mm vibration condition improved low-temperature impact toughness. At 20 °C, the impact energy was 19.86 J for the vibration condition and 18.23 J for the no-vibration condition, an increase of 1.63 J. The ductile-to-brittle transition temperature was about -50 °C for the vibration condition and about -40 °C for the no-vibration condition. The improved toughness came from finer ferrite grains, more spherical graphite, and reduced stress concentration.
Fourth, the 40 Hz and 2 mm vibration condition improved wear resistance. In my friction and wear tests with a 50 N load, 2 Hz frequency, 10 mm sliding distance, and 30 min duration, the weight loss was 0.04191 g for the vibration condition and 0.04369 g for the no-vibration condition. The friction coefficient was 0.803 for the vibration condition and 0.912 for the no-vibration condition. The vibration specimen formed a more stable graphite lubricating layer and a more stable hardened layer, which reduced material loss and lowered the friction coefficient.
Fifth, the mechanisms responsible for the improvement include dendrite fragmentation, cavitation-induced nucleation, increased undercooling, compositional homogenization, and austenite shell stabilization. These mechanisms act together during solidification in lost foam casting. Moderate vibration promotes fine, spherical graphite and a refined matrix. Excessive vibration, especially at high frequency, can disturb the austenite shell and cause graphite degeneration. Therefore, process optimization is necessary. In my study, 40 Hz and 2 mm provided the best overall combination of microstructure and mechanical properties for QT400-18 produced by lost foam casting.
Overall, my first-person experimental work demonstrates that mechanical vibration is a practical, low-cost, and effective method for improving QT400-18 ductile iron in lost foam casting. It refines graphite, ferrite, and pearlite, increases tensile strength and elongation, lowers the ductile-to-brittle transition temperature, and improves wear resistance. The method can be integrated into the vibration compaction stage of lost foam casting without adding complex equipment, and it is consistent with green casting principles. I therefore recommend controlled vertical vibration at approximately 40 Hz and 1.5 mm to 2 mm amplitude for lost foam casting of QT400-18, with final parameters adjusted according to casting size, section thickness, and production requirements.