Introduction

The machining of particleboards, a widely used material in the furniture industry, is known to be very difficult because of extremely short tool life (Bridges 1971; Boehme and Münz 1987; Porankiewicz 1997a; Porankiewicz and Grönlund 1991; Stühmeier 1989). For a number of years, the content of mineral contamination has been discussed as the major factor involved in the wearing process of cutting edges during the machining of particleboards (Bridges 1971; Boehme and Münz 1987; Porankiewicz 1997a; Stühmeier 1989). In some studies, high-temperature corrosion (HTC) was pointed out as a second important factor impacting tool wear during cutting of particleboards (Mouseev 1981; Porankiewicz 1997b, 1998). Based on thermal analysis and spectral surface analytical methods, evidence supporting the presence of HTC in the wearing process has been given (Porankiewicz 1998; Porankiewicz and Wagner 1997; Stühmeier 1989). It is well known how to estimate physical and mechanical properties of particleboard influencing tool wear, but it is not known how to do this in the case of chemical properties. The literature shows important developments in the metallurgy of cemented carbides, with the aim of decreasing the HTC of tool materials during the processing of wood and secondary wood products (Porankiewicz 1997a; Wysiecki 1997). However, high levels of variation in tool wear using different modified cemented carbide grades when machining different particleboards shows that our knowledge of this area is still fragmentary (Porankiewicz 1997a).

The objective of the present work was to examine the method of assessment of the HTC of cemented carbide material in contact with particleboard under simulated conditions. This work also attempted to apply the results obtained to check the role of HTC in the wearing process of the cutting edge when milling particleboard, as well as to find a possibility to improve the metallurgy of cemented carbides for particleboard processing.

Experimental materials and procedures

The evaluation of the HTC of cemented carbides by particleboard materials in simulated conditions was performed using thermal gravimetric analysis (TGA, Shimadzu TGA-50). In order to accelerate corrosion processes in thermal analysis, a fine cobalt powder (2.5 μm diameter, according to the FSS method) and five different three-layer melamine-coated particleboards (designated a–e), commercially manufactured in different countries, were applied. The use of less-broken cemented carbide particles for the study of HTC in previous experiments ended without results. The experiments were conducted in a porcelain crucible, in ambient atmosphere, with a temperature gradient of 50°C/min with about 3–4 mg of particleboard or about 16 mg of Co powder in each test. The size of the particleboard specimen was limited by the volume of the crucible used. In interfered analyses, the cobalt and particleboard were mixed in a porcelain mortar. In order to decrease the variation of evaluation of HTC properties it is recommended that samples for thermal analysis be obtained from different places in the particleboard. Because thermal analysis is sensitive to specimen macroparameters, it is important to keep all test conditions unchanged during analysis. X-ray diffraction (XRD, Philips, equipped with an Fe Kα X-ray radiation source and APD software) was used to examine the samples after thermal analysis with respect to cobalt and tungsten compounds. The minimum detectable concentration by mass was 0.1%.

The cutting edge wear was measured after particleboard peripheral milling on a milling machine at a cutting speed of 64 m/s. The same three-layer particleboards designed for the furniture industry were used for thermal analysis and machining tests, with an average density of 630–770 kg/m3 and moisture content of 4– 6%. Cutter heads of diameter 125 mm with 11 cemented carbide turnover inserts were used. Because of the inserts limitation, for two grades only two tests were performed. Altogether, 49 milling tests were carried out. The rake and sharpness angles were γ F=20°, β F=55°, respectively. The bevel angle was equal to zero. The average tungsten carbide (WC) grain size was estimated according to the ASTME method on a SEM images, using a magnification of 3000:1. The chemical composition of the cemented carbide grades was examined using an inductively coupled plasma (ICP) spectrometry method to determine the content of tungsten, cobalt, nickel, chromium, vanadium and iron.

The cutting edge wear VB W, represented by the edge recession in the plane of the bisector of the sharpness angle, was measured using the contact method. Initial sharpness, described as the radius of the cutting edge roundness, was r N= 5.4 µm (SD=1.7 µm). The average cutting edge wear VB W was calculated for the area of contact with the particleboard skin with a width of about 1 mm. Very high random damage to the cutting edge was excluded. The average feeding rate per tooth f Z and the cutting depth g S were f Z=0.2 mm and g S=1.5 mm, respectively. The cutting path L S was calculated from Eq. (1) as the sum of the length contact of single arcs of the cutting edge and working material on a feeding length L F

$$ L_{\rm{S}} = {{L_{\rm{F}} } \mathord{\left/ {\vphantom {{L_{\rm{F}} } {f_{\rm{Z}} \cdot w_{{\rm{AC}}} \cdot R_{\rm{S}} \cdot \left[ {arc\sin \left( {\phi _{\rm{D}} } \right) + arc\cos \left( {\phi _{\rm{G}} } \right)} \right]}}} \right. \kern-\nulldelimiterspace} {f_{\rm{Z}} \cdot w_{{\rm{AC}}} \cdot R_{\rm{S}} \cdot \left[ {arc\sin \left( {\phi _{\rm{D}} } \right) + arc\cos \left( {\phi _{\rm{G}} } \right)} \right]}},\;\;\;\left( {\rm{m}} \right) $$
(1)

where for a parametric cycloid equation w AC=0.993189,

$$ R_{\rm{S}} = {\rm{cutting radius }}\;\left( {{\rm{mm}}} \right){\rm{,}} $$
$$ \phi _{\rm{L}} = arc\cos \left[ {{{\left( {R_{\rm{S}} - g_{\rm{S}} } \right)} \mathord{\left/ {\vphantom {{\left( {R_{\rm{S}} - g_{\rm{S}} } \right)} {R_{\rm{S}} }}} \right. \kern-\nulldelimiterspace} {R_{\rm{S}} }}} \right]\;\;\left( {{\rm{rad}}} \right), $$
(2)
$$ \phi _{\rm{U}} = arc\,sin\left( {{{f_{\rm{Z}} } \mathord{\left/ {\vphantom {{f_{\rm{Z}} } {2 \cdot R_{\rm{S}} }}} \right. \kern-\nulldelimiterspace} {2 \cdot R_{\rm{S}} }}} \right)\;\;\left( {{\rm{rad}}} \right). $$
(3)

Angles φ L, φ U are the upper and lower angle of contact of the cutting edge and working material, respectively, and g S is the cutting depth (mm).

The following properties of the particleboard were analyzed: the content of hard mineral contamination of the particleboard skin C MC, the weighted average size of mineral contamination particles W AS and the relative two-dimensional macroporosity P S. To assess the content of particleboard hard mineral contamination, a burning method was used. The remaining ash was treated with hot muriatic acid in order to dissolve all soft soluble fractions. Hard contamination particles were separated into fractions using the following steel mesh sizes: 1500, 600, 400, 200, 100 and 50 µm. The two biggest fractions of contamination particles were excluded for evaluation of C MC and W AS. The weighted average size W AS was estimated for six fractions. The porosity P S was assessed for the particleboard skin by means of a tool microscope.

An experimental model of the relation between cutting edge wear VB W after particleboard cutting and the cutting conditions was developed using an optimization program created by the author (Porankiewicz 1988) with further modifications. Calculations were performed at the Poznań Supercomputer Center (PCSS) using an IBM PS-2 computer. Applying the developed model, the synergistic effect (SE) between mechanical and chemical wear of cemented carbide cutting edges during particleboard cutting was analyzed. The SE can be defined as (Kotlyar et al. 1988)

$$ SE = {{\left( {VB_{\rm{W}} - VB_{\rm{W}}^{\rm{C}} - VB_{\rm{W}}^{\rm{M}} } \right)} \mathord{\left/ {\vphantom {{\left( {VB_{\rm{W}} - VB_{\rm{W}}^{\rm{C}} - VB_{\rm{W}}^{\rm{M}} } \right)} {VB_{\rm{W}} \cdot 100}}} \right. \kern-\nulldelimiterspace} {VB_{\rm{W}} \cdot 100}}\;\;\left( {\rm{\% }} \right), $$
(4)

where VB W C is the wear obtained by maximum HTC (chemical wear), and VB W M is the wear obtained by minimum HTC (mechanical wear). The assumptions made allowed for the analysis of the influence of SE in particleboard milling. The simultaneous action of different kinds of mechanical wear and HTC with increasing temperature in a contact region during cutting were not included in the SE simulation.

Results and discussion

Results of the thermal analysis show an interaction of the particleboard skin (including melamine coating) with Co (Figs. 1 and 2), which was the result of intensive thermal degradation of the particleboard analyzed. These interactions, which are well distinguished in the first derivative of the TGA plot (samples a, c, d, and e for the skin in Fig. 1, and samples b, c, and e for the coating in Fig. 2), show rapid mass increases (peaks). These mass increases can be associated with the corrosion of cobalt, because in the TGA plots for particleboard skin and melamine film without cobalt, no rapid mass increases can be seen (Fig. 3). The interaction closely follows the particleboard mass loss in the temperature range 350–380°C. The corrosion peaks (Figs. 1 and 2) are located in the starting phase of cobalt mass increase, simply due to the oxidation process caused by the oxygen present in ambient air. The comparison of actual and previous observations (Porankiewicz 1993) shows that the use of highly broken cobalt binder speeds up the HTC to a level where it is possible to distinguish corrosion peaks (Porankiewicz 1997b, Porankiewicz 1998). The peaks appear at different temperatures for different particleboards, in the range 420–460°C for particleboard skin, and 470–530°C for melamine coating. Results of the XRD examination show that during the thermal analysis, the cobalt present in the particleboard corroded and the final reaction products are CoO and Co3O4. No oxide phase of the cobalt was found in the case when the heating process of cobalt with particleboard was stopped below the corrosion peak, and large amounts of cobalt oxides CoO and Co3O4 were observed when the heating process was stopped just above the corrosion peak (Fig. 4). This shows that cobalt oxides are final products of HTC, which occurs near the corrosion peaks on TGA plots.

Fig. 1.
figure 1

TGA plots acquired for particleboard skins a, b, c, d and e mixed with cobalt. Lines indicate corrosion peaks

Full size image
Fig. 2.
figure 2

TGA plots acquired for melamine coatings of particleboards a, b, c, d and e mixed with cobalt. Lines indicate corrosion peaks

Full size image
Fig. 3.
figure 3

TGA plots: a Melamine coating. b Particleboard skin

Full size image
Fig. 4.
figure 4

XRD plots obtained for cobalt heated with melamine coating below and above the corrosion peak temperature for Bragg deflection angle 2θ

Full size image

Results of TGA show that for particleboards that are more difficult to machine (cases b, c, d and e in Figs. 1 and 2), the corrosion peaks are higher. Most of the corrosion peaks for the coating are higher in comparison to those for the particleboard skin. These facts allowed a quantitative characterization of HTC peaks on the TGA plots. Among many others, the descriptor R MSMI, based on the relative maximum speed of cobalt mass rapid increase, was calculated from first derivative of mass versus time on the TGA plots. This descriptor, R MSMI, contains the most important information about the HTC peaks. Because the width of the cutting edge wear zone caused by peripheral milling of coated particleboards is larger than the thickness of the melamine film, R MSMI was calculated as the sum of values estimated for particleboard skin and melamine film separately. The values of the R MSMI descriptor for the particleboards tested are given in Table 1.

Table 1. The value of estimated R MSMI (min−1) descriptor for particleboards tested
Full size table

Other descriptors evaluated using an area or average value characterizing HTC peaks in the TGA plot were less correlated with the cutting edge wear than was the R MSMI descriptor. This shows that the HTC of the tool material in contact with particleboard is a dynamic process that cannot be analyzed in any static experiments.

Results of observations of the cutting edge wear show that the same grade of cemented carbide is not appropriate to cut all particleboards (Fig. 5) because cemented carbides have very different corrosivity towards tool materials and hard mineral contamination (Tables 1 and 2). The relative wear of the cutting edge for different cemented carbide grades and the same particleboard, as well as for different particleboards and the same grade of cemented carbide, is as great as 5.8 times and 14.8 times, respectively. The maximum difference observed is as great as 22.5 times. These extreme differences show the importance of the problem of the proper choice of a grade of cemented carbide for machining a particular type of particleboard. The generalized experimental relation VB=f(CW, C Co, C Ni, C Cr, C V, C Fe, C MC, W AS, R MSMI, R WC, L S, P S) between the edge wear of the cutting edge while peripheral milling melamine-coated particleboard and the cutting conditions is modeled as

$$ \displaylines{ VB_{\rm{W}} = - 41.320 - 4.998 \cdot Z_{\rm{W}} ^{1.152} - 2.204 \cdot 10^{ - 6} \cdot Z_{{\rm{Co}}} ^{10.140} + 1.372 \cdot \left( {15.985 + Z_{{\rm{Ni}}} } \right)^2 \cr + 217.93 \cdot \left( {Z_{{\rm{Cr}}} - 0.381} \right)^2 - 225.30 \cdot \left( {Z_{\rm{V}} \cdot C_{{\rm{MC}}} ^{ - 0.0421} - 0.333} \right)^2 \cr + 2.112 \cdot 10^{ - 5} \cdot \left( {C_{{\rm{MC}}} - 1981.4} \right)^2 - 3.123 \cdot 10^{ - 3} \cdot \left( {W_{{\rm{AS}}} - 50.771} \right)^2 \cr + 33.899 \cdot \left( {R_{{\rm{WC}}} - 0.921} \right)^2 + 14.249 \cdot L_{\rm{S}} \cdot R_{{\rm{MSMI}}} ^{2.201} + 8.374 \cdot 10^{ - 5} \cdot \left( {R_{{\rm{WC}}} ^{0.0321} \cdot W_{{\rm{AS}}} + 2420.5} \right)^2 \cr + 0.0155 \cdot Z_{{\rm{Cr}}} ^{0.0443} \cdot C_{{\rm{MC}}} + 0.0262 \cdot Z_{{\rm{Ni}}} ^{0.0613} C_{MC} - 0.153 \cdot C_{{\rm{MC}}} ^{1.130} \cdot P_{\rm{S}} \cr - 0.738 \cdot Z_{{\rm{Fe}}} \cdot C_{{\rm{MC}}} ^{0.675} - 12.114 \cdot Z_{{\rm{Co}}} \cdot Z_{{\rm{Ni}}} \cdot Z_{{\rm{Cr}}} {\rm{ }}\left( {{\rm{\mu m}}} \right), \cr VB_{\rm{W}} > 0. \cr} $$
(5)
Fig. 5.
figure 5

Relative cutting edge wear VB W W of cemented carbide grades G1–G9 after milling particleboards P1–P5

Full size image
Table 2. Range of variation of independent variables in Eq. (<equationcite>5</equationcite>)
Full size table

The correlation coefficient R between the observed and predicted edge wear using Eq. (5), the mean deviation in relation to regression curve SR, the sum of residuals square SK and the variance V are as great as R=0.98, SR=4.5 µm, SK=876 and V=28, respectively. The model in Eq. (5) was improved in comparison to previous work (Porankiewicz 1997b). The constant parameters were: sharpness angle β f=55°, rake angle γ f=20°, cutting speed v C=65 m/s. Equation (5) allows the prediction of the cutting edge wear of different cemented carbides after milling different particleboards.

The cutting edge wear increases with the increase of mineral contamination in the range of C MC>1500 mg/kg. For lower values of C MC, the cutting edge wear increase is nearly not observed. The cutting edge wear increases with the increase of weighted average size of hard mineral contamination particles W AS, until the W AS boundary value is as high as W AS=118 µm. The increase in the edge wear with an increase in W AS in the range of W AS<118 µm is probably due to the impact of HTC, represented in Eq. (5) by the R MSMI descriptor. This is caused by the more efficient removal of the products of HTC by small particles of hard mineral contamination, exposing metallic binder on corrosion environment. The border size of contamination particles is probably reached when the contamination particles are larger; then, because of the width of binder paths, their further contact with the binder no longer occurs. The decrease of wear with an increase of W AS for W AS>118 µm is caused by a decrease in the number of contamination particles, with an increase in their size, and because the effect of single contamination particle wearing increases more slowly with an increase in particle size.

The influence of the corrosivity of the machined material towards the tool material, represented by descriptor R MSMI on the edge wear, increases with increases in the cutting path length L S (Fig. 6). Along with the increase in L S, the impact of R MSMI on the cutting edge wear also increases the temperature increase in a cutting zone.

Fig. 6.
figure 6

Cutting edge wear VB W predicted from Eq. (5) versus R MSMI and L S

Full size image

The role of the R MSMI descriptor decreases with increases in W AS. The reason for this phenomenon is probably the creation of notches (that is, the loss of many WC grains at the same time) as a result of hard mineral contamination particles, which are bigger then WC grain size, causing indentation into the machined surface during cutting. In bottom of the notches, until they are leveled, the pressure on the cutting edge drops down, resulting in a reduction in the HTC. The R MSMI descriptor was estimated for pure cobalt. Because the binder in modified cemented carbides contain more elements, in the analysis performed some uncontrolled variation of the R MSMI descriptor may occur. This problem is now being examined. The increase in the porous share (P S) decreases the cutting edge wear, with an interaction term C MC∙P S. The pores cause less strengthening of particles of hard mineral contamination in a structure of particleboard.

Results of experiments show that optimal the WC grain size (R WC) in a cemented carbide structure for coated particleboard machining is as high as R WC= 0.9 µm. The impact of the interaction term R WC∙W AS on the wear is low (Fig. 7). This disproves earlier expectations that at the minimum value of R WC used (R WC=0.48 µm), the lowest wear can be observed. The increase of tungsten content in the cemented carbide decreases cutting edge wear (Fig. 8), which is in agreement with the literature (Wysiecki 1997). In modern, isostatic condensed fine-grain cemented carbides, the increase in tungsten content causes an increase in hardness without any loss of impact strength, which makes it possible to increase the cutting edge wear resistance. The increase of cobalt content C Co in the range 2.5–4.79% causes a decrease in the wear of the cutting edge. The major influence, however, begins from about C Co=4% (Fig. 8). Very low C Co in the cemented carbide, in spite of observed HTC of the binder, did not result in any decrease in cutting edge wear.

Fig. 7.
figure 7

Cutting edge wear VB W predicted from Eq. (5) versus R WC and W AS

Full size image
Fig. 8.
figure 8

Cutting edge wear VB W predicted from Eq. (5) versus C W and C Co

Full size image

In order to limit the WC grain growth in the sintering of the cemented carbide, chromium in form of Cr3C2 is added (Bock et al. 1992). After sintering, the Cr3C2 combines with the binder as a solid-solution phase. The increase in chromium content in the cemented carbide up to C Cr=0.38% decreases the cutting edge wear (VB W) after milling coated particleboard with the highest mineral contamination (Fig. 9). The decrease in mineral contamination moves the minimum wear only insignificantly towards higher C Cr values; therefore it can be stated that the presence of chromium in cemented carbides in the amount of 0.38% increases its resistance against mineral contamination. Another way to limit the WC grain growth during the cemented carbide sintering process is the use of VC (Bock et al. 1992). Analysis results show that an increase in the vanadium content in cemented carbide increases the cutting edge wear in all ranges of variation (Fig. 10). VC does not seem to be the best solution to limiting WC grain growth in the cemented carbides for melamine-coated particleboard machining.

Fig. 9.
figure 9

Cutting edge wear VB W predicted from Eq. (5) versus C Cr and C MC

Full size image
Fig. 10.
figure 10

Cutting edge wear VB W predicted from Eq. (5) versus C Ni and C V

Full size image

The purpose of using nickel in the binder of cemented carbide was to decrease the HTC. However, an increase in the cutting edge wear can be observed with increasing nickel content in the cemented carbide in the whole range of variation for C Ni=0.01–1.12%. The addition of nickel to the binder significantly decreases its resistance against high mineral contamination of the particleboard. Because of the interaction term C Co∙C Cr∙C Ni in Eq. (5), a small positive role of the nickel in reducing the cutting edge wear can also be observed. With an increase in iron content in the binder of the cemented carbide in the range C Fe=0.01–0.25%, the cutting edge wear decreases. This is a quite unexpected observation. According to results of XRD analysis, the iron is present in the binder in small-grain form. According to the analysis performed, a change in the best cemented carbide grade composition for melamine-coated particleboard milling (namely CW=90%, C Co=3.1%, C Cr=0.24%, C V=0.20%) can be proposed: slightly increase the content of W, Co and Cr, decrease the content of Ni and V, and increase the average WC grain size to R WC=0.9 µm.

According to Eq. (5), the largest synergistic effect observed was 46% while milling particleboard with less mineral contamination with K05 cemented carbides for a cutting path of 800 m. The SE decreases with increasing cutting edge dullness and then disappears. For the highest mineral contaminated particleboard, abrasion wear is dominant and the SE cannot be observed. For modified cemented carbides, the SE was not determined. Modification of the composition of cemented carbides with Cr, Ni and V eliminated the SE. There were predictions of a larger role for the SE in the wearing process of cemented carbide cutting edges while particleboard machining.

Conclusion and final remarks

The R MSMI descriptor extracted from the TGA plots allows us to distinguish the different HTC properties of the examined particleboard and to develop an experimental model of the relation between cutting edge wear of modified cemented carbides and the cutting conditions. The wear of cemented carbide cutting edges by peripheral milling of melamine coated particleboard decreases with increases in the content of tungsten in the range C W=87.5–91.5% and with increases in the content of cobalt in the range C Co=2.5–4.8%.

The cutting edge wear increases with increases in the content of nickel in the range C Ni=0.1–1.12%. The wear also increases with increases in the content of vanadium (in the form of VC) in the range C VC=0.01–0.32%.

The lowest cutting edge wear can be obtained by using the optimal content of chromium (in the form of Cr3C2) of 0.38% and an average WC grain size of R WC=0.9 µm.