2022 年 63 巻 3 号 p. 329-334
Iron-based powder mixtures for the powder metallurgy process commonly contain solid lubricants. Both zinc stearate (ZnSt) and N,N′-ethylenebis(stearamide) (EBS), which are conventional lubricants in this field, exhibit similarly adequate lubrication performance. However, their effects on powder mixture flowability are different; that is, powder mixtures containing ZnSt exhibit better flowability than ones containing EBS. In this study, the adhesive and frictional properties of five surface combinations (iron–iron, iron–EBS, iron–ZnSt, EBS–EBS, and ZnSt–ZnSt) were investigated using surface force and resonance shear measurements. The adhesive forces obtained for all combinations were almost the same. On the other hand, the frictional forces for the iron–EBS and EBS–EBS combinations obtained from resonance shear measurement were larger than the others under low applied loads (<ca. 1.0 mN). This result suggests that the frictional properties of lubricants under low applied loads determine the powder mixture flowability.
This Paper was Originally Published in Japanese in J. Jpn. Soc. Powder Metallurgy 66 (2019) 554–559.
In powder metallurgy, small amounts of solid lubricants such as zinc stearate (ZnSt) or N,N′-ethylenebis(stearamide) (EBS) are commonly added to iron-based powder mixtures to protect the dies and compacts from seizure during compression and ejection. Solid lubricants in a powder mixture are fully attached to the surfaces of the iron powder particles, exist in a free state apart from iron powders, or in a mixed state of them.
ZnSt is a representative lubricant with a long history of use in powder metallurgy, and is known to show good lubricity and have little effect on the flowability of the powder mixture, but as a drawback, it causes fouling of the furnace and sintered parts by forming zinc oxide during sintering. EBS has been used as an alternative to ZnSt and is known to be effective avoiding the fouling problem because it contains no metallic components, however, it causes poor flowability of the powder mixture.1)
Uenosono et al. attempted to clarify the reason why EBS had a negative effect on powder flowability.1) They considered three kinds of forces acting in a powder mixture containing EBS, capillary force caused by liquid bridging, electrostatic force and van der Waals force, and concluded from calculated results that the van der Waals force between solid lubricants was at least ten times greater than the other forces, and thus seemed to have a predominant effect on powder flowability.
However, it is not easy to compare the forces that act between pairs of EBS particles and those between pairs of ZnSt particles because the commercially available EBS and ZnSt powders have different particle sizes and shapes, and it would be difficult to eliminate the seemingly obvious influence of those factors. Using atomic force microscopy (AFM), Artieda-Guzmán et al. measured the surface adhesive and viscoelastic properties of EBS and ZnSt tablets which were prepared by compacting EBS and ZnSt lubricant powders.2) They discussed the relevance of those physical properties of the tablet surfaces to the compressibility and ejection properties of the iron-based powder mixtures, but did not refer to its relevance to powder flowability. In addition, the adhesive forces they measured were the forces acting between the silicon cantilever and the lubricant surfaces, and were not those between two lubricant surfaces or between iron and lubricant surfaces.
Therefore, in the present study, we prepared thin films of iron, ZnSt, and EBS with smooth surfaces to investigate the adhesion and friction properties of five surface combinations (iron–iron, iron–EBS, iron–ZnSt, EBS–EBS, and ZnSt–ZnSt).
As evaluation methods, we employed surface forces and resonance shear measurements (RSM) using a surface forces apparatus (SFA).3) SFA has the advantage of enabling direct measurement of the interaction forces between two surfaces in air as well as in liquids as a function of the surface separation distance. Surface force measurement can also be used to evaluate the force required to separate two surfaces in contact, i.e., adhesion force. The resonance shear measurement (RSM), developed based on SFA4–8) for applying shear between two surfaces, can evaluate the viscosities and frictional properties of liquids confined between two surfaces as a function of separation distance (D) and normal load (L). The RSM can also evaluate the viscoelastic and frictional properties between solid surfaces in contact.9–14)
In this study, the measurements of the adhesive and frictional properties of the five surface combinations were conducted while observing the contact surfaces with an optical microscope installed in the SFA,4) and the relevance of the obtained results to the flowabilities of the powder mixtures were discussed.
Powder (A) is an atomized pure ion powder (JIP301A, JFE Steel Corp., D50 = 75 µm). Powders (B) and (C) were prepared by mixing the atomized pure iron powder with 0.8 mass% ZnSt powder (D50 = 14 µm) and 0.8 mass% EBS powder (D50 = 25 µm), respectively, using a V-type mixer (Triple V-type mixer 1101-1.5, Yoshida Seisakusho Co., Ltd.) for 15 min.
2.1.2 Measurement of powder flow ratesThe powder flow rates of (A), (B) and (C) were determined according to JIS Z 2502. 50 g of the each powder prepared in 2.1.1 was put into a funnel with an orifice of 2.5 mm in diameter, and the time (s/50 g) until all the powder had flowed out of the orifice was measured.
2.2 Preparation of thin films 2.2.1 Preparation of iron filmsIron films were prepared on mica sheets, glued on cylindrical quartz disks with epoxy resin (Epikote 1004, Hexion Specialty Chemicals Inc.), by radio frequency (RF) magnetron sputtering (pressure 0.7 MPa, RF power 100 W, argon flow rate 5 cm3/min, deposition time 30 min) using a vacuum sputtering apparatus (SPV2-TMP-T1-RF1/R, Toei Scientific Industrial Co., Ltd.).
2.2.2 Preparation of lubricant filmsLubricant films were prepared on the iron films described in 2.2.1 by spin-coating the lubricant solutions using a spin-coater (1H-D7, Mikasa Co., Ltd.). The solutions were prepared by dissolving the lubricant powder in an organic solvent (mixture of chloroform:benzene:ethanol = 63:17:20 (v/v) for EBS, chloroform:benzene = 64:36 (v/v) for ZnSt). The concentrations of the EBS and ZnSt solutions were 0.25 mg/mL and 0.06 mg/mL, respectively. After spin-coating, the lubricant films were dried under a vacuum for 1 h at room temperature.
2.2.3 Measurement of surface smoothness and thicknesses of thin filmsThe surface smoothness of the prepared films were evaluated by imaging with an atomic force microscope (AFM) (SPI-3800-SPA400, Seiko Instruments Inc.) with a silicon nitride cantilever (OMCL-TR800PSA-W, Olympus Corp., spring constant k = 0.57 N/m). The film thicknesses were measured by an ellipsometer (DHA-XA/S3-T, Mizojiri Optical Co., Ltd.).
2.3 Adhesive force measurement using SFAFigure 1(a) shows a schematic illustration of the prepared surfaces on cylindrical disks. The adhesive force between the two surfaces was measured using an SFA equipped with an optical microscope (Fig. 1(b)).4) As shown in Fig. 1(b), one quartz disk was fixed to an upper lens holder, and the other disk was fixed to a lower lens holder in a crossed cylinder geometry. The lower lens holder was connected to a pulse motor stage via a horizontal spring (spring constant k = 266 N/m), and the surface separation distance and the normal load between the two surfaces were controlled by the pulse motor drive. The Newton’s ring pattern (Fig. 1(c)) generated by the interference of the light reflected between the upper and lower surfaces was monitored using an optical microscope placed above the chamber. The changes in the Newton’s rings were recorded as a video during one cycle of adhesive force measurement, that is, approach, contact, loading, retraction, and separation of the two surfaces. The moments when contact and separation of the two surfaces occurred were determined based on this video. The deflection of the cantilever required to separate the surfaces (Δd) was determined as the difference in the pulse motor positions at which the surfaces were contacted and separated, and the adhesive force required to separate the surfaces Fad was obtained as Fad = kΔd. The obtained Fad values were normalized by a curvature radius(R) of the cylindrical quartz disk (Fad/R).
Schematic illustration of (a) surfaces on cylindrical disks and (b) a surface forces apparatus equipped with optical microscope, and (c) a microscope image of newton’s ring.
The moments when the surfaces were contacted and separated were determined as follows. The bright and dark Newton’s rings pattern move continuously during the approaching process, and the move of the rings stops when the surfaces were contacted. When additional pressure was applied, the contact area increases, and as a result, the region with the same brightness as the center of the Newton’s rings expanded. When the load applied to the two surfaces reached a certain arbitrary value, the separation process was started. In the separation process, changes occurred in the opposite sequence; first, the region with the same brightness as the center of the Newton’s rings decreased, and when the surfaces separated, the Newton’s ring pattern started to change again. When an adhesive force acts between the surfaces, a discontinuous change of the pattern appears at the instant of separation due to the jump-out of the surface depending on the adhesive force.
In this study, we investigated the relationship between the adhesive force and maximum applied load with the five surface combinations (iron–iron, iron–EBS, iron–ZnSt, EBS–EBS, and ZnSt–ZnSt).
2.4 Resonance shear measurement using SFAResonance shear measurement of the prepared surfaces was performed as schematically illustrated in Fig. 2(a), in the same manner as in the previous reports.5–13) The upper surface was connected to a four-sectored piezo tube and hung by a pair of vertical leaf springs. The lower surface was connected to a pulse motor stage via a supporting horizontal spring (spring constant k = 266 N/m) and could be driven vertically by the motor. Using the piezo tubes, the upper unit was laterally oscillated by sinusoidal voltages (amplitude Uin = ±1 V) at an angular frequency ω (= 2πf). The deflection of the vertical leaf springs (Δx) was detected as the output voltage (Uout) of a capacitance probe with an amplifier. Uout was measured while changing ω, and the normalized amplitude (Uout/Uin) was plotted as a function of ω, which represents the resonance curve.
Schematic illustrations of (a) a resonance shear measurement instrument, (b) typical resonance curves, (c) a mechanical model for analyzing resonance curves.
Typical resonance curves are shown in Fig. 2(b). When the two surfaces are separated in air (air separation, AS), a resonance peak appeared at around 30 Hz, depending only on the physical parameters (mass (m1), damping parameter (b1), and spring constant (k1)) of the upper unit. On the other hand, when the two surfaces with large adhesive and frictional forces (e.g. mica) are in contact in air, i.e., in a nearly non-slip condition (solid contact, SC), the two surfaces oscillate together, and the resonance peak appeared at the higher resonance frequency depending on the physical parameters of both the upper and lower units, especially the high spring constant of the lower unit. When two surfaces are in contact but slip occurs, the peak position shifts from the SC position to a lower frequency, and its intensity decreases and its width broadens. As the applied load increased, the peak position shifts gradually to higher frequency, its intensity increases, and finally the peak position and intensity approach asymptotically to those of the SC peak. The intensity and resonance frequency of the resonance peak reflect the frictional property between the two surfaces. The increase in the intensity and frequency of the peak towards the SC peak corresponds to the increase in the friction between the two surfaces. Furthermore, we analyzed the resonance curves based on the mechanical model7) shown in Fig. 2(c) to obtain the viscous and elastic parameters (b2, k2) between the surfaces, as in the previous reports,7,9–13) and then calculated the friction force Ffriction according to the following formula.
| \begin{align*} F_{\text{friction}} &= \max|F_{\text{elastic}} + F_{\text{viscous}}| \\ &= \max | - k_{2} (x_{1} - x_{2}) - b_{2} (V_{1} - V_{2})| \end{align*} |
As shown in Fig. 3, the flow rates of the pure iron powder (A) and the iron-based powder mixture with 0.8 mass% ZnSt (B) were almost the same, while flow rate of the powder mixture with 0.8 mass% EBS (C) was higher than the others. This result indicated that EBS has a larger negative effect on the flowability of the powder mixture compared with ZnSt.
Powder flowability obtained for iron powder only, iron powder + 0.8% ZnSt, and iron powder + 0.8% EBS.
Figure 4 shows the AFM images, the surface roughness obtained from the AFM images, and the film thicknesses determined by ellipsometry. The thickness of the iron film was around 96 nm, and those of the EBS and ZnSt films were 31.0 nm and 29.5 nm, respectively. The maximum molecular lengths of both EBS and ZnSt, which have two C18 alkyl chains, were estimated as 4–5 nm.15) Thus, a film thickness of around 30 nm corresponds at least 6 molecular layers in both cases. In all surfaces, an analysis of the AFM images showed that the maximum height difference (P-V) was less than 10 nm and the root mean square roughness Rms was 0.5 nm or less. Thus, the prepared films were free of large agglomerates that could influence the adhesive and frictional measurements.
AFM images of a sputter prepared Fe film, a spin-cast EBS film, and a spin-cast ZnSt film. The surface roughness (P-V (peak to valley) and Rms (root mean square)) obtained from AFM images, and the thicknesses obtained by ellipsometry were shown below the images.
The obtained adhesive forces of all surface combinations were plotted with respect to the applied maximum load as shown in Fig. 5. We had expected that the adhesive forces of the combination containing EBS (EBS–EBS and/or Fe–EBS) would be stronger than those of the other combinations because the EBS-containing powder mixture showed poor flowability. However, contrary to our expectation, there were no significant differences among the five surface combinations at 4 to 7 mN/m, and the adhesion did not depend on the maximum load. This result indicates that adhesive force had no direct relationship with the flowabilities of the powder mixtures, and especially the deterioration of flowability in the EBS mixtures.
Plots of adhesion forces vs. maximum applied load obtained for Fe–Fe, Fe–EBS, Fe–ZnSt, EBS–EBS, and ZnSt–ZnSt surfaces using the SFA.
Resonance curves (Uout/Uin vs. ω) were obtained for all surface combinations at various applied loads. A series of measurements was conducted at 3 to 6 different contact positions for each surface combination. In Fig. 6(a)–(e), the resonance peak intensities ((Uout/Uin)max; indicator of the friction force between the surfaces) at all of the measured contact points are plotted with respect to the applied load. The different symbols represent the data obtained at different contact positions. In the cases of (a) Fe–Fe and (e) ZnSt–ZnSt, the variations of the (Uout/Uin)max values depending on the contact point were small. By contrast, the (Uout/Uin)max valuesobtained for (b) Fe–EBS, (c) Fe–ZnSt, and (d) EBS–EBS varied significantly depending on the contact point. Especially with regard to (b) Fe–EBS and (d) EBS–EBS, the (Uout/Uin)max value increased drastically under low load range (<ca. 1.0 mN) at some contact positions, which are denoted by the arrows in the graphs. This indicated that the EBS films contained some areas where the friction force increased drastically in the low load range.
Plots of peak intensities (Uout/Uin)max vs. applied load obtained for (a) Fe–Fe, (b) Fe–EBS, (c) Fe–ZnSt, (d) EBS–EBS, and (e) ZnSt–ZnSt surfaces.
The load dependencies of the average peak intensities of all surface combinations are summarized in Fig. 7. Under the high load condition (above 1.5 mN (Fig. 7(a)), the (Uout/Uin)max values of Fe–Fe were higher than the other combinations which included EBS or ZnSt films. This result indicated that both of the EBS and ZnSt exhibited friction-reducing functions under the comparatively high loads (surface pressure) conditions. Conversely, in the low load range below 1.0 mN (Fig. 7(b)), which can be considered to correspond to powder flowability, the order of the (Uout/Uin)max values was as follows: Fe–EBS > EBS–EBS > Fe–Fe ≈ ZnSt–ZnSt ≈ Fe–ZnSt. In fact, the frictions obtained for the surface combinations containing EBS films were greater than the others under the low loads condition. This result was consistent with the poorer flowability of the powder mixture containing EBS compared with that of pure iron powder or the powder mixture containing ZnSt, as shown in Fig. 3.
Plots of average peak intensities (Uout/Uin)max vs. applied load, at the applied load of 0∼8 mN (a), and 0∼2 mN (b).
Figure 8 shows the friction forces (Ffriction) between the various surface combinations calculated using viscous (b2) and elastic (k2) parameters, obtained by analysing the resonance curves based on the physical model (Fig. 2(c)). Ffriction of Fe–Fe was the highest in the higher load range (>ca. 1.5 mN), while those of Fe–EBS and EBS–EBS were higher than the others at the lower load range (<ca. 1.0 mN). This result was the same as in the case of (Uout/Uin)max. In addition, at the low load range (<ca. 1.0 mN), the Ffriction of Fe–EBS was larger than that of EBS–EBS. This result indicated that main cause of the poorer flowability of the EBS-containing powder mixture was the friction between iron powder and EBS and not the friction between EBS–EBS.
Friction forces calculated using viscous (b2) and elastic (k2) parameters plotted against applied load, (a) load range of 0∼9 mN, and (b) load range of 0∼2 mN.
In this study, thin films of iron and lubricants (EBS and ZnSt) were prepared by the sputtering and spin-coating methods, respectively, and the adhesive and frictional properties of the five surface combinations (iron–iron, iron–EBS, iron–ZnSt, EBS–EBS, and ZnSt–ZnSt) were evaluated by SFA and compared to clarify the main reason for the different flowability of the ZnSt- and EBS-containing powder mixtures. The major conclusions of this study are summarized as below:
In this regard, however, the states of the iron and lubricant surfaces investigated in this study were possibly not the same as those in actual iron powders and iron-based powder mixtures. In order to verify the reliability of these conclusions, a more detailed investigation including the evaluation of the surface states of particles in actual powder mixtures and the comparison with those of the model systems used in this study is necessary. This is an issue for a future work.
