- Email: [email protected]
Densification behavior, mechanical properties and thermal shock resistance of tungsten alloys fabricated at low temperature
Materials and Design 58 (2014) 290–297
Contents lists available at ScienceDirect
Materials and Design journal homepage: www.elsevier.com/locate/matdes
Densification behavior, mechanical properties and thermal shock resistance of tungsten alloys fabricated at low temperature He Zhang a,⇑, Liang Ge a, Mingjiang Shi a, Pan Ren b a b
School of Electronic and Information Engineering, Southwest Petroleum University, Chengdu 610500, China PetroChina Sichuan Petrochemical Co., Ltd., Nanchong 637000, China
a r t i c l e
i n f o
Article history: Received 1 November 2013 Accepted 8 January 2014 Available online 21 January 2014 Keywords: Tungsten alloy Low temperature sintering Mechanical properties
a b s t r a c t The low-temperature shrinkage of tungsten was greatly accelerated by the addition of trace Nb and Ni, and the addition of trace Nb and Ni also significantly promoted the final sintering density. The 99.1% of theory density for W–0.1 wt.%Nb–0.1 wt.%Ni material sintered at 1600 °C was obviously greater than 93.7% of theory density for W material sintered at 2000 °C. Ball milling treatment played an important role in promoting the sintering densification of W–0.1 wt.%Nb–0.1 wt.%Ni powder, and the powder milled for 10 h (W10) could be sintered to near full density (99.4% of theory density) at 1600 °C. The ball milling for 15 h has no effect in improving the sintering density, but it induced rapid growth of tungsten grains. The microhardness and tensile strength of the sintered tungsten alloys were highly dependent on its sintering density and grain size. Improving the sintering density while controlling the grain growth could effectively promote the microhardness and tensile strength. Furthermore, the improvement of thermal shock resistance of the W10 alloy was due to good microstructure and the increase in the tensile strength. Crown Copyright Ó 2014 Published by Elsevier Ltd. All rights reserved.
1. Introduction Tungsten (W) alloys are widely used as an important candidate material for various structural applications at high temperature [1] due to its excellent properties such as high melting point, high density, high modulus and thermal shock resistance, low coefficient of thermal expansion and good high temperature strength [2]. Various grades of W alloys are classified [3]: pure W, dispersion-strengthened W, and W alloys produced by different fabrication technologies, e.g. sintered, forged, rolled and hot-worked (deformed). Besides the fabrication, the raw powder materials, the alloying elements and dopants/impurities, treatments, and the final shape/geometry have a strong effect on the achieved properties of W and W alloys [4]. Since W alloy is extensively applied in the highly sophisticated field, it is required that the material should have near full densification (>99% of theory density), so as to make full use of its excellent properties [5]. However, an extreme temperature as high as 2700 °C is required for traditional microsized tungsten powder to be sintered to near full density since it has a very high melting point of 3422 °C [6]. In order to fabricate high density W alloys at a low temperature, in recent years, the researchers have used nanotechnology to
⇑ Corresponding author. Tel./fax: +86 028 6600 9495. E-mail address: [email protected] (H. Zhang). http://dx.doi.org/10.1016/j.matdes.2014.01.023 0261-3069/Crown Copyright Ó 2014 Published by Elsevier Ltd. All rights reserved.
synthesize nanometer tungsten powder which can provide a large driving force for sintering due to its abundant surface energy and grain-boundary energy, which is called nanometer activation [7]. Although preparing nanometer tungsten powder can theoretically improve its sinterability, the fact that nanometer tungsten powders can be directly sintered to near full density (>97% of theory density) has never been reported. Furthermore, transition element niobium (Nb) is a ductile and soft refractory metal with low melting point (2468 °C) compared with W [8], low vapor pressure, good chemical stability, and good strength retention at elevated temperatures [9]. It has been extensively applied in aerospace, electronic devices, steel industry, nuclear industry, and chemical engineering industry. In all applications, approximately 75% of all Nb metal is used as an addition to low-alloyed steels. Another 20–25% is used as an additive in nickel base superalloys and heat resisting steels. Only 1–2% is used in the form of pure niobium and Nb-based high temperature alloys [10]. The Nb alloy is made by melting and mixing two or more metals and the mixture has properties different from those of the individual metals [11]. In addition, transition element nickel (Ni, melting point is 1455 °C) is used primarily for the alloys it forms. It is used for making stainless steel and many other corrosion resistant alloys. Nevertheless, up to date, there are only a few papers devoted to the study on addition of Nb and/or Ni to W alloys. Furthermore, niobium and nickel have been theoretically proved
H. Zhang et al. / Materials and Design 58 (2014) 290–297
to be an excellent accelerator for the sintering densification of tungsten, which is called chemical activation, but there has also rarely been reported that tungsten can be sintered to near full density by adding trace niobium and nickel [12]. In the present work, the W alloys with high relative density were produced at low temperature by the combination of the nanometer activation method and chemical activation method. The nanometer tungsten powder (i.e. W–0.1 wt.%Nb–0.1 wt.%Ni composite powder) was first prepared by the physical and chemical method and then the composite powder was ball milled for different times (0 h, 10 h and 20 h), and the activated composite powders were finally achieved. The effects of ball milling time on the densification behavior of the activated composite powder were investigated in detail. Most of W alloys, because of inherently low fracture toughness, are susceptible to catastrophic fracture caused by the thermal stresses in applications at high temperature. Therefore, the thermal shock resistance of the W alloys was evaluated by water quenching method. The purpose of this work is to report a potential method for fabrication of W alloys at low temperature, which can be applied to aid materials engineering design for the development of W alloys, quality assurance, and characterization assessment of durability. 2. Experimental details The ammonium metatungstate, niobium nitrite (Nb(NO2)5) and nickel nitrite (Ni(NO2)2) were used as raw materials for the synthesis of nanometer tungsten powder. A sol was prepared by dissolving ammonium metatungstate, niobium nitrite and nickel nitrite with a small amount of polyethylene glycol (PEG-2000) into the distilled water. The nanometer W–0.1 wt.%Nb–0.1 wt.%Ni composite powder was then gained by a multi-step process, consisting of sol–spraying–drying of the solution at 250–350 °C, calcining at 300 °C for 2 h and a subsequent two-step reduction process (600 °C and 750 °C) in hydrogen atmosphere for 1.5 h and 2.5 h, respectively. The calcined composite powder was ball milled in a W alloys bottle for different times, and the milling conditions are listed in Table 1. The ball mill activated nanometer tungsten powders were obtained and a pure tungsten powder without addition of niobium and nickel was also synthesized using the same production process for comparison, the pure tungsten powder, W–0.1 wt.%Nb–0.1 wt.%Ni composite powder, W–0.1 wt.%Nb– 0.1 wt.%Ni composite powder ball-milled for 10 h and 15 h were denoted as W, W0, and W10 and W15, respectively. X-ray diffraction meter (D/ruax2550PC, Japan) was employed to identify the phases of these powders, and ultra high resolution field emission scanning electron microscopy (SEM, NOVA TM NanoSEM 230, Czech) was used to characterize the morphology of these powders. The oxygen content in these powders were determined by a nitrogen/oxygen/hydrogen determination (TCH-600, USA), the specific surface area of each powder was measured by a BET surface area analysis instrument (Monosorb Autosorb-1, USA). The crystalline phase was determined using the X-ray diffraction (XRD) (Rigaku, Japan). The broad–scan analysis was typically conducted within the 2-h range of 10–80° using the Cu Ka (k = 1.542 Å) radiation. The narrow scan analysis was conducted within the 2-h range of 20–30° and was subsequently Table 1 Ball milling parameters used in the present work. Milling medium Grinding medium Ball to powder ratio Liquid to solid ratio Plate and bowl speed Milling time
Ethanol Tungsten balls (3–10 mm in diameter) 3:1 (mass ratio) 3:1 (volume ratio) 240 rpm 10 h, 15 h
291
used to determine ‘‘Bragg’’ grain size. The BET grain size was determined using the surface area measurement technique. It was found that these tungsten powders were difficult to be shaped because of their nanometer particle size, so they were mixed with 0.5 wt.% paraffin prior to forming, then they were pressed into standard tensile samples by two-direction cold pressing with pressure of 250 MPa. The green compacts were pre-sintered at 1000 °C in atmospheric pressure of 5 Pa for 2 h in order to eliminate the paraffin, later pre-sintered compacts were sintered in tungsten rod furnace at different temperatures (1500 °C, 1600 °C, 1700 °C, 1800 °C, 1900 °C, 1950 °C, 2000 °C and 2030 °C) for 2 h, heating rate was 2 °C/min and flowing H2 was employed as the protective atmosphere. The densities of the sintered specimens were measured by Archimedes principle with deionized water as the immersing medium. The quasi–static mechanical properties of these specimens were measured by a standard Instron 3369 material test machine (USA) and the tensile speed was about 1 mm/min according to ASTM:A356 (6 mm 3 mm 60 mm). The hardness of specimen was tested by a nanoindentation method and load of 50 mN. The thermal shock resistance can be appraised by water quenching method based on the definition of the critical thermal shock temperature difference (DTcrit). The DTcrit can be measured experimentally by quenching specimens from various elevated temperatures and determining the quenching temperature that results in a reduction of strength for a given specimen geometry [13]. The DTcrit value is defined as 70% of the room temperature strength, which was determined using linear interpolation of the retained strength values as described in ASTM:C1525-04. Before the water quenching, all samples were ground and polished with diamond slurries down to a 1 lm finish and the edges of all samples were chamfered to minimize the effect of stress concentration due to machining flaws. At least ten samples were tested for each experimental condition and all samples were from same billet. The polished rectangular bars for thermal shock testing were heated in the vacuum up to the desired temperature difference and held for 10 min to eliminate any temperature gradient effect before quenching by dropping parallel to their tensile surface into the water bath. The temperature of the water bath was controlled to about 25 °C by adjusting the cooling water flow. The water quenching temperature differences were 200, 300, 400, 500, 600 and 700 °C. The 5 Pa was kept for the samples heated in vacuum and the time taken for the transfer from the furnace to the water bath was less than 1 s. A digital microhardness tester (HXD-1000T) was employed to determine the microhardness of the sintered tungsten bulk, and the tensile strength was tested by means of mechanical testing machine (Instron3369, USA). 3. Results and discussions 3.1. Phase and performances of these powders The XRD patterns of four kinds of powders (W, W0, and W10 and W15) are shown in Fig. 1, which indicated an increase in peak broadening with progress of milling. The peaks of Nb and Ni were not detected due to low amounts of Nb and Ni. Using a combination of W(1 1 0) and W(2 1 1) peaks, the grain size and lattice distortion were calculated according to the XRD patterns as shown in the following equation [14]:
b cos h ¼ 0:94ðk=dÞ þ 4e sin h
ð1Þ
where b is the full-width at half-maximum (FWHM), h is Bragg angle, k is the X-ray wavelength, d is grain size and e is lattice distortion. The broadening factors that are not induced by ball milling were taken out by the following equation [14,15]:
292
H. Zhang et al. / Materials and Design 58 (2014) 290–297
Fig. 1. XRD patterns of four kinds of powder.
b2 ¼ b2M b2S
ð2Þ
where bS and bM are FWHM of initial powder and milled powders, respectively. Dislocation densities of these powders were calculated by taking internal stress e into the following equation [16,17]:
q¼
6pE e 2 G lnðr=r 0 Þ b
ð3Þ
where E is Young’s modulus, G shear modulus, b the magnitude of the Burgers vector and r and r0 the outer and inner cut-off radius, respectively. The modulus was evaluated from the slopes of load deflection curves of strength tests. A static extensometer was used to measure the deflection with an error in the measurement of 0.1%. Parameters b, r and r0 are automatically given by X-ray diffraction meter. With E = 452 GPa, G = 176 GPa, b = 4.37 1010 m and ln(r/ r0) = 4. The particle sizes of four kinds of powders were calculated based on the BET specific surface area using the following equation [18]:
dBET ¼ 6=ðSBET qTheory Þ
ð4Þ
where dBET is the particle size, SBET is the powder specific surface area and qTheory is the theoretical density of tungsten, i.e. 19.3 g/ cm3. Table 2 lists the characteristic parameters of the four kinds of powders based on XRD and calculation results. The X-ray peak broadening was due to the combined effect of the structural refinement (finer grain size). Furthermore, the increase in lattice distortion, residual strain and dislocation density, which also was confirmed by data listed in Table 2. BET grain size was too larger than ‘‘Bragg’’ grain size, which was because the specific surface area obtained by the surface area measurement technique was the specific surface area of agglomerated particles, was not the specific surface area of single nanoparticles. The increase in lattice distortion, residual strain and dislocation density could significantly increase the sinterability of these tungsten powders [19]. The impurities
Table 2 Characteristic parameters of the four kinds of powders. Powder
W W0 W10 W15
Grain size (nm)
Lattice distortion (%)
Specific surface area (m2/ g)
dBET (nm)
56.9 51.3 44.1 39.8
0.1521 0.1431 0.1874 0.2002
1.51 2.26 1.44 1.59
210 140 210 200
Oxygen content (wt.%)
Dislocation density (1014 m2)
0.0667 0.2213 0.3613 0.4132
1.57 1.39 2.26 2.59
resulted from polyethylene glycol were carbon, hydrogen and oxygen. Furthermore, the impurities were inevitably included nitrogen element. Literature has confirmed that coarsening in nonoxides bulk materials, including tungsten alloys, is promoted generally by oxygen present as oxide impurities on the particle surfaces, and if the oxygen existed, the carbon element would be depleted. Therefore, the oxygen content was calculated, whereas other elements were ignored. It could be seen from Table 2 that the oxygen content had been rising all the time during milling, but its amount was still less than 0.4 wt.%. In the present work, maximum content of oxygen was 0.4132, which was a much higher level than that of the dopants. The formation of the nano-scale powders indicated that the effects of oxygen on grain size were less than that of the dopants. The grain size was mainly affected by the combination of the lattice distortion, residual strain and dislocation density. Furthermore, the melting point of Nb is 2468 °C and maximum sintering temperature is 2030 °C, so a partial loss of Nb is not considered. The melting point of Ni is 1455 °C and Ni has a very low vapor pressure under the protective atmosphere, so a partial loss of Ni is negligible. It is expected that the content of Ni and Nb is constant. The content of Ni and Nb in the specimen sintered at 2030 °C was analyzed and the results confirmed the obvious change in the content of Ni and Nb in the specimen was not detected. Fig. 2 shows SEM images of four kinds of tungsten powders. It could be seen from Fig. 2(a) and (b) that W and W0 powder particles are of irregular polyhedron in shape, and there is serious aggregation among particles. As shown in Fig. 2(b), (c) and (d), the W–Nb–Ni composite powder broke in the mode of brittle fracture was measured during ball milling, and the BET particle size increased significantly when the W–Nb–Ni powder was milled for 10 h and then decreased slightly after being milled for 15 h. Such a change process could be explained as the following: at the initial stage of ball milling (ball milled for 10 h), a large amount of small particles were produced and they were readily to aggregate because of the generation of new surfaces [20], which resulted in the reduction in the specific surface area and the improvement of the sinterability of these tungsten powders [19]. At the final stage of ball milling (ball milled for 15 h), the ball milling energy is mainly consumed in the process of broking the aggregation. At last, the particle refinement and aggregation process reached a balance point, and the particle dispersion became better, which led to the increase in specific surface area and the improvement of the sinterability of these tungsten powders [19]. 3.2. Densification process The geometrical density of the test samples in the state ‘‘as pressed’’ as well as after driving out the paraffin at 1000 °C of four kinds of powders are listed in Table 3, which revealed that the linear shrinkages of W–Nb–Ni composite was obviously greater than that of the pure W powder, indicating that the addition of Nb and Ni had played a significant activation role for the sintering of tungsten at this stage. At the same time, the linear shrinkages of W–Nb–Ni composite were also greater than that of W15, which was further greater than that of W10. The apparent density of W–Nb–Ni composite powder increased greatly after ball milling and the green compacts of W10 and W15 had greater density compared to W0 when they were pressed under the same pressure. As for the activated sintering, the shrinkage DL/L0 varies as [21]:
DL g XdC cSV DA t ¼ L0 D4 kT
ð5Þ
where X is the atomic volume, d is the width of the second-phase activator, g is a collection of geometric terms, C is the solubility of the base phase in the activator, cSV is the solid–vapor surface energy, DA is the diffusivity of the base phase in the activator, D is
H. Zhang et al. / Materials and Design 58 (2014) 290–297
293
Fig. 2. SEM morphologies of the four kinds of powder (A) W; (B) W0; (C) W10; (D) W15.
Table 3 Geometrical density of the test samples in the state ‘‘as pressed’’ as well as after driving out the paraffin at 1000 °C of four kinds of powders. Specimen
W (%)
W0 (%)
W10 (%)
W15 (%)
As pressed As pre-sintered at 1000 °C
40.1 40.4
40.3 52.9
42.9 46.7
43.2 52.1
the particle size, k is Boltzmann’s constant, t is sintering temperature and T is the absolute temperature. The greater diffusivity DA in the tungsten material was reduced due to the addition of trace Nb and Ni compared with pure W when there is no activator. Furthermore, the particle size D was reduced by ball milling. Therefore, Eq. (5) could well explain why the activated tungsten powders have a much higher shrinkage extent than the pure W powder. Densification behavior of four kinds of powders at different sintering temperatures is presented in Fig. 3. Compared with the results of other groups [3,4,8], the nano-size of the pure W powder was sintered at 1600 °C for 2 h, and the relative density was about 90%. In the present work, the relative density of the nano-size of the pure W powder was about 87%, which was attributed to the different sintering conditions and treatment approach of powders. It is clear that the W–Nb–Ni composite powder sintered at 1900 °C reached maximum relative density as high as 99.1% of theory density and the pure W material sintered at 2000 °C reached maximum relative density as high as 93.7 of theory density. This indicated that the final sintering densification of tungsten was greatly improved due to the addition of trace Nb and Ni. At the same time, ball milling treatment could also promote the further densification of W–Nb–Ni composite powder to near full density. In consequence, W10 could be sintered to 99.3% of theory density at 1600 °C while W15 could be sintered to 99.4% of theory density at this temperature. This revealed that the activated tungsten pow-
Fig. 3. Change curves of relative density of four kinds of powder versus sintering temperature.
ders were superior to the pure tungsten powder in improving the sintering densification rate as well as sintering densification extent, which could also be observed from Fig. 4. Fig. 4 shows micrographs of the sintered bulk of the four kinds of powder sintered at 1600 °C (a) W; (b) W0; (c) W10; (d) W15. No obvious flaws such as pores, gaps and cracks were detected, which confirmed the results of density. The densification rate of the compact was calculated by [22]:
@q DB ¼ @t G3 kT
ð6Þ
where D is the diffusion coefficient, B is a collection of material and geometric constants, k is Boltzmann’s constant, G is the grain size,
294
H. Zhang et al. / Materials and Design 58 (2014) 290–297
Fig. 4. Micrographs of the sintered bulk of the four kinds of powder sintered at 1600 °C (A) W; (B) W0; (C) W10; (D) W15.
and T is the temperature. This equation indicated that the densification rate would increase as the diffusion coefficient increased due to addition of trace Nb and Ni, and grain size decreased due to ball milling. It could be also seen from Fig. 3 that the densities of W10 and W15 sintered at all temperatures were similar and the densities did not increase further when the temperature was more than 1600 °C. It revealed that increasing ball milling time from 10 h to 15 h has slight beneficial effect on the sintering densification of W–Nb–Ni composite powder. At the same time, 1600 °C was an optimum sintering temperature for W10 and W15 powders because a further increasing of temperature could provide little effects on their sintering densities. 3.3. Grain size change behavior of the sintered tungsten materials Fig. 5 displays the change behavior of grain sizes versus sintering temperatures. It could be observed that the grain growth of the sintered compacts of W0, W10 and W15 increased with increasing sintering temperature. The grain size of raw W was very small whereas that of sintered W was much larger, which was still less than the grain size of W0, W10 and W15. The grain growth of tungsten was attributed to the introduction of oxygen impurities, not attributed to the addition of trace Nb and Ni. The grain growth of the sintered compact further increased as the ball milling time increased from 10 h to 15 h, which also confirmed due to the grain growth of tungsten due to the introduction of oxygen impurities [23]. Grain growth in a sintered compact has been found to follow a cubic relationship with time [22]:
G3 ¼ G30 þ Mt
ð7Þ
where G0 is the initial grain size and M is the grain boundary mobility. Compared with small difference in original grain size, the great
Fig. 5. Change curves of average grain size of the sintering bulk of four kinds of powder versus sintering temperature.
difference in grain size after sintering was attributed to the grain boundary mobility. Therefore, the addition of trace Nb and Ni was favorable to improve the mobility of tungsten grain boundaries, and ball milling further accelerated this mobility. Compared with the original grain size, the obvious grain growth was measured and grain size was initially limited by pore pinning, the grain size could be approximated by:
G3 ¼ Mt
ð8Þ
The effect of grain growth on the densification rate was calculated by substituting Eq. (8) into Eq. (6):
@q DB ¼ @t MkTt
ð9Þ
H. Zhang et al. / Materials and Design 58 (2014) 290–297
295
where H is microhardness, qs is fractional sintered density, H0 and
H are constants and G is grain size. The change curves of micro-
Fig. 6. Change in microhardness with sintering temperature of the four kinds of sintered compacts.
It could be seen from Eq. (9) that the increase in diffusivity and the decrease in grain boundary mobility would increase the densification rate. 3.4. Microhardness and tensile strength From previous studies on sintered products, mechanical properties had been found to be highly dependent on density and grain size [24,25]. In the present work, the change behavior of microhardness and tensile strength of four kinds of sintered tungsten materials with sintering temperature were investigated in detail. The hardness for a brittle material such as pure tungsten at room temperature followed the Hall–Petch relation dependence on inverse square-root grain size [26]:
H H ¼ qs H0 þ pffiffiffiffi G
ð10Þ
hardnesses versus sintering temperatures are shown in Fig. 6. It was clear that the microhardness of sintered W mainly increased with the increasing temperature, which was consistent with its increasing sintering densification. The microhardness of W0 increased as the sintered temperature increased to 1800 °C and then the microhardness decreased due to the increase in grain size. The maximum microhardness of W10 and W15 sintered at 1600 °C was because of their near full density and then the microhardness decreased with the increasing temperature due to the increase in the grain size. The fractography of four kinds of tungsten powders sintered at 1600 °C are shown in Fig. 7. The brittle rupture of four kinds of sintered tungsten was readily observed. In order to eliminate the effect of hydrogen brittleness on the tensile strength, the tensile strength values of four kinds of sintered compacts after the samples were subjected to vacuum treatment at 1350 °C are listed in Table 4. The tensile strength values of four kinds of tungsten powders sintered at 1600 °C were 833 ± 80, 1450 ± 130, 1570 ± 139 and 1430 ± 143 MPa for W, W0, W10 and W15, respectively. The lower tensile strength for pure W specimen was due to low relative density. The tensile strength values of W0, W10 and W15 specimens were slightly greater than the data reported in the literature [2], which was attributed presumably to ultrofined grain due to the addition of trace Nb and Ni. The relative density of sintered W was 93.7% of theory density, which was much lower than 99.1% for W0, 99.3% for W10 and 99.4% for W15, whereas the tensile strength of sintered W was statistically equivalent to that of W10 and greater than the tensile strength of sintered W0 and W15 compacts. This indicated that the tensile strength of sintered materials was significantly affected by the addition of trace Nb and Ni as well as the ball milling. Many researchers have confirmed that the strength of sintered materials have a close relationship with the relative density [27]. Furthermore, the effect of grain size on brittle materials at higher densities
Fig. 7. Fractography of the sintered bulk of four kinds of powder sintered at 1600 °C (A) W; (B) W0; (C) W10; (D) W15.
296
H. Zhang et al. / Materials and Design 58 (2014) 290–297
Table 4 Tensile strength of the sintered compacts of four kinds of powders sintered at 1600 °C.
Tensile strength (MPa)
W
W0
W10
W15
833 ± 80
1450 ± 130
1570 ± 139
1430 ± 143
and production technology it would be more meaningful to compare the pure W sample sintered to its highest density (here at 2030 °C) to the alloyed sample sintered already at 1600 °C to a superior density. The pure W sample sintered at 2030 °C has mechanical properties equivalent to one of W10 sintered at 1600 °C. However, W10 has better thermal shock resistance than pure W sample, which was due to the microstructure because the gain size of pure W sample sintered at 2030 °C was a much greater than that of W10.
4. Conclusions
Fig. 8. Strength values of pure W and W10 before and after water quenching with increasing the thermal shock temperature difference.
has been noted. In such case, the strength r could be calculated in terms of grain size d and porosity e as following [28]:
r ¼ r0 da exp½be
ð11Þ
where r0 is a strength constant, a and b are empirical constants. It could be seen from the Eq. (11) that the strength of sintered material would decrease if the porosity e and/or grain size d increased, and this would be used to explain the strength change of four kinds of sintered activated tungsten powders. Although sintered composite powder had the smallest grain size, it had the lowest tensile strength compared with that of W10 and W15, due to its lowest sintering density. The sintered W10 and W15 materials had the similar porosity (sintering density), but the grain size of sintered W10 material was much smaller than that of W15 material. Therefore, the strength of sintered W10 material was higher than that of the sintered W15 material. Due to W10 sintered at 1600 °C had best mechanical properties, the W10 was used to evaluate the thermal shock resistance of W alloys and the thermal shock resistance of pure W material was also tested for comparison. The strength values of pure W and W10 before and after water quenching are shown in Fig. 8 with increasing the thermal shock temperature difference. The measured DTcrit for W10 as high as 541 °C was greater than 440 °C for pure W, which indicated that W10 had good thermal shock resistance than pure W. Furthermore, the critical temperature difference is also calculated by a triaxial modification of Hooke’s law in terms of the average strength of the material (r), Young’s modulus (E), Poisson’s ratio (v) and coefficient of thermal expansion (a) as shown in the following equation [18].
R ¼ DT crit ¼
rð1 tÞ Ea
ð12Þ
Therefore, the improvement of thermal shock resistance of the W10 alloy was due to the increase in the flexural strength. Furthermore, the pure W sample displayed here exhibits a density of only 87%, and thus appears ill suited for such a technological comparative test. To underline the potential of the developed special alloy
Activated nanometer tungsten powder (i.e. W–0.1 wt.%Nb– 0.1 wt.%Ni composite powder) was prepared by physical and chemical method and subsequent ball-milling treatment, and the characteristics of these powders milled by different times (0 h, 10 h and 15 h) were investigated. The effect of ball-milling time on the densification behavior and mechanical properties was investigated, which was further compared with that of the nanometer pure tungsten powder. The results indicated that the sintering shrinkage rate and the sintering densification rate were greatly accelerated by the activation treatment, and the activated tungsten powder milled by 10 h could be sintered to near full density at 1600 °C, whereas the activation treatment also could promote quick growth of tungsten grains. Although sintered composite powder had the smallest grain size, it had the lowest tensile strength compared with that of W10 and W15, due to its lowest sintering density. The sintered W10 and W15 materials had the similar porosity (sintering density), but the grain size of sintered W10 material was much smaller than that of W15 material. Therefore, the strength of sintered W10 material was higher than that of the sintered W15 material. Furthermore, the improvement of thermal shock resistance of the W10 alloy was due to the increase in the flexural strength. Acknowledgements This work is supported by Foundation of Sichuan Educational Committee (Research on pattern recognition, No. 12ZB342) and Oil & Gas Pipeline Detection Technology Research Based on Multi-sensor Data Fusion (No. 2012XJZ021), the project is founded by Technology Research Dept. CNOOC Research and 2012 Youth Foundation of Southwest Petroleum University. References [1] Schade P. 100 years of doped tungsten wire. Int J Refract Met Hard Mater 2010;28:648–60. [2] Ravi Kiran U, Panchal A, Sankaranarayana M, Nandy TK. Tensile and impact behavior of swaged tungsten heavy alloys processed by liquid phase sintering. Int J Refract Met Hard Mater 2013;37:1–11. [3] Çalisßkan NK, Durlu N, Bor Sß. Swaging of liquid phase sintered 90W–7Ni–3Fe tungsten heavy alloy. Int J Refract Met Hard Mater 2013;36:260–4. [4] Liu WS, Ma YZ, Zhang JJ. Properties and microstructural evolution of W–Ni–Fe alloy via microwave sintering. Int J Refract Met Hard Mater 2012;35:138–42. [5] Ravi Kiran U, Panchal A, Sankaranarayana M, Nandy TK. Tensile and impact behavior of swaged tungsten heavy alloys processed by liquid phase sintering. Int J Refract Met Hard Mater 2013;37:1–11. [6] Telu S, Mitra R, Pabi SK. High temperature oxidation behavior of W–Cr–Nb alloys in the temperature range of 800–1200 °C. Int J Refract Met Hard Mater 2013;38:47–59. [7] Wu ZJ, Wang Z, Shi GD, Sheng J. Effect of surface oxidation on thermal shock resistance of the ZrB2–SiC–ZrC ceramic. Compos Sci Technol 2011;71:1501–6. [8] Ryu T, Hwang KS, Choi YJ, Sohn HY. The sintering behavior of nanosized tungsten powder prepared by a plasma process. Int J Refract Met Hard Mater 2009;27:701–4. [9] Wang HT, Fang ZZ, Hwang KS, Zhang HB, Siddle D. Sinter-ability of nanocrystalline tungsten powder. Int J Refract Met Hard Mater 2010;28:312–6. [10] Pusavec F. Porous tungsten machining under cryogenic conditions. Int J Refract Met Hard Mater 2012;35:84–9.
H. Zhang et al. / Materials and Design 58 (2014) 290–297 [11] Ma YZ, Zhang JJ, Liu WS, Yue P, Huang BY. Microstructure and dynamic mechanical properties of tungsten-based alloys in the form of extruded rods via microwave heating. Int J Refract Met Hard Mater 2014;42:71–6. [12] Hu P, Wang Z, Sun X. Effect of surface oxidation on thermal shock resistance of ZrB2–SiC–G composite. Int J Refract Met Hard Mater 2010;28:280–5. [13] Wang Z, Qu Q, Wu ZJ, Shi GD. The thermal shock resistance of the ZrB2–SiC–ZrC ceramic. Mater Des 2011;32:3499–503. [14] Zhang ZW, Zhou JE, Xi SQ, Ran G, Li PL. Phase transformation and thermal stability of mechanically alloyed W–Ni–Fe composite materials. Mater Sci Eng A 2004;379:148–53. [15] Wang Z, Wu ZJ, Shi GD. The oxidation behaviors of a ZrB2–SiC–ZrC ceramic. Solid State Sci 2011;13:534–8. [16] Grahle P, Arzt E. Microstructural development in dispersion strengthened NiAl produced by mechanical alloying and secondary recrystallization. Acta Mater 1997;45:201–11. [17] Zhang L, Ukai S, Hoshino T, Hayashi S, Qu XH. Y2O3 evolution and dispersion refinement in Co-base ODS alloys. Acta Mater 2009;57:3671–82. [18] Wang Z, Xie B, Zhou WY, Shi GD, Wu ZJ. Thermophysical properties of TiB2–SiC ceramics from 300 °C to 1700 °C. Int J Refract Met Hard Mater 2013;41:609–13. [19] Zhang XH, Wang Z, Hu P, Han WB, Hong CQ. Mechanical properties and thermal shock resistance of ZrB2–SiC ceramic toughened with graphite flake and SiC whiskers. Scripta Mater 2009;61:809–12. [20] Wu ZJ, Wang Z, Qu Q, Shi GD. Oxidation mechanism of a ZrB2–SiC–ZrC ceramic heated through high frequency induction at 1600 °C. Corros Sci 2011;53:2344–9.
297
[21] German RM. Sintering theory and practice. New York: John Wiley; 1996. [22] Zovas PE, German RM. Retarded grain boundary mobility in activated sintered molybdenum. Metall Trans A 1984;15A:1103–10. [23] Zhou SB, Wang Z, Zhang W. Effect of graphite flake orientation on microstructure and mechanical properties of ZrB2–SiC–graphite composite. J Alloys Compd 2009;485:181–5. [24] Wang Z, Wang S, Zhang XH, Hu P, Han WB, Hong CQ. Effect of graphite flake on microstructure as well as mechanical properties and thermal shock resistance of ZrB2–SiC matrix ultrahigh temperature ceramics. J Alloys Compd 2009;484:390–4. [25] Zhang XH, Wang Z, Sun X, Han WB, Hong CQ. Effect of graphite flake on the mechanical properties of hot pressed ZrB2–SiC ceramics. Mater Lett 2008;62:4360–2. [26] Wang Z, Wu ZJ, Shi GD. Fabrication, mechanical properties and thermal shock resistance of a ZrB2–graphite ceramic. Int J Refract Met Hard Mater 2011;29:351–5. [27] Wang Z, Wu ZJ, Shi GD. Effect of annealing treatment on mechanical properties of a ZrB2–SiC–graphite ceramic. Mater Sci Eng A 2011;528:2870–4. [28] Wang Z, Li SC, Wang M, Wu GQ, Sun XM. Effect of SiC whiskers on microstructure and mechanical properties of the MoSi2–SiCw composites. Int J Refract Met Hard Mater 2013;41:489–94.
Contents lists available at ScienceDirect
Materials and Design journal homepage: www.elsevier.com/locate/matdes
Densification behavior, mechanical properties and thermal shock resistance of tungsten alloys fabricated at low temperature He Zhang a,⇑, Liang Ge a, Mingjiang Shi a, Pan Ren b a b
School of Electronic and Information Engineering, Southwest Petroleum University, Chengdu 610500, China PetroChina Sichuan Petrochemical Co., Ltd., Nanchong 637000, China
a r t i c l e
i n f o
Article history: Received 1 November 2013 Accepted 8 January 2014 Available online 21 January 2014 Keywords: Tungsten alloy Low temperature sintering Mechanical properties
a b s t r a c t The low-temperature shrinkage of tungsten was greatly accelerated by the addition of trace Nb and Ni, and the addition of trace Nb and Ni also significantly promoted the final sintering density. The 99.1% of theory density for W–0.1 wt.%Nb–0.1 wt.%Ni material sintered at 1600 °C was obviously greater than 93.7% of theory density for W material sintered at 2000 °C. Ball milling treatment played an important role in promoting the sintering densification of W–0.1 wt.%Nb–0.1 wt.%Ni powder, and the powder milled for 10 h (W10) could be sintered to near full density (99.4% of theory density) at 1600 °C. The ball milling for 15 h has no effect in improving the sintering density, but it induced rapid growth of tungsten grains. The microhardness and tensile strength of the sintered tungsten alloys were highly dependent on its sintering density and grain size. Improving the sintering density while controlling the grain growth could effectively promote the microhardness and tensile strength. Furthermore, the improvement of thermal shock resistance of the W10 alloy was due to good microstructure and the increase in the tensile strength. Crown Copyright Ó 2014 Published by Elsevier Ltd. All rights reserved.
1. Introduction Tungsten (W) alloys are widely used as an important candidate material for various structural applications at high temperature [1] due to its excellent properties such as high melting point, high density, high modulus and thermal shock resistance, low coefficient of thermal expansion and good high temperature strength [2]. Various grades of W alloys are classified [3]: pure W, dispersion-strengthened W, and W alloys produced by different fabrication technologies, e.g. sintered, forged, rolled and hot-worked (deformed). Besides the fabrication, the raw powder materials, the alloying elements and dopants/impurities, treatments, and the final shape/geometry have a strong effect on the achieved properties of W and W alloys [4]. Since W alloy is extensively applied in the highly sophisticated field, it is required that the material should have near full densification (>99% of theory density), so as to make full use of its excellent properties [5]. However, an extreme temperature as high as 2700 °C is required for traditional microsized tungsten powder to be sintered to near full density since it has a very high melting point of 3422 °C [6]. In order to fabricate high density W alloys at a low temperature, in recent years, the researchers have used nanotechnology to
⇑ Corresponding author. Tel./fax: +86 028 6600 9495. E-mail address: [email protected] (H. Zhang). http://dx.doi.org/10.1016/j.matdes.2014.01.023 0261-3069/Crown Copyright Ó 2014 Published by Elsevier Ltd. All rights reserved.
synthesize nanometer tungsten powder which can provide a large driving force for sintering due to its abundant surface energy and grain-boundary energy, which is called nanometer activation [7]. Although preparing nanometer tungsten powder can theoretically improve its sinterability, the fact that nanometer tungsten powders can be directly sintered to near full density (>97% of theory density) has never been reported. Furthermore, transition element niobium (Nb) is a ductile and soft refractory metal with low melting point (2468 °C) compared with W [8], low vapor pressure, good chemical stability, and good strength retention at elevated temperatures [9]. It has been extensively applied in aerospace, electronic devices, steel industry, nuclear industry, and chemical engineering industry. In all applications, approximately 75% of all Nb metal is used as an addition to low-alloyed steels. Another 20–25% is used as an additive in nickel base superalloys and heat resisting steels. Only 1–2% is used in the form of pure niobium and Nb-based high temperature alloys [10]. The Nb alloy is made by melting and mixing two or more metals and the mixture has properties different from those of the individual metals [11]. In addition, transition element nickel (Ni, melting point is 1455 °C) is used primarily for the alloys it forms. It is used for making stainless steel and many other corrosion resistant alloys. Nevertheless, up to date, there are only a few papers devoted to the study on addition of Nb and/or Ni to W alloys. Furthermore, niobium and nickel have been theoretically proved
H. Zhang et al. / Materials and Design 58 (2014) 290–297
to be an excellent accelerator for the sintering densification of tungsten, which is called chemical activation, but there has also rarely been reported that tungsten can be sintered to near full density by adding trace niobium and nickel [12]. In the present work, the W alloys with high relative density were produced at low temperature by the combination of the nanometer activation method and chemical activation method. The nanometer tungsten powder (i.e. W–0.1 wt.%Nb–0.1 wt.%Ni composite powder) was first prepared by the physical and chemical method and then the composite powder was ball milled for different times (0 h, 10 h and 20 h), and the activated composite powders were finally achieved. The effects of ball milling time on the densification behavior of the activated composite powder were investigated in detail. Most of W alloys, because of inherently low fracture toughness, are susceptible to catastrophic fracture caused by the thermal stresses in applications at high temperature. Therefore, the thermal shock resistance of the W alloys was evaluated by water quenching method. The purpose of this work is to report a potential method for fabrication of W alloys at low temperature, which can be applied to aid materials engineering design for the development of W alloys, quality assurance, and characterization assessment of durability. 2. Experimental details The ammonium metatungstate, niobium nitrite (Nb(NO2)5) and nickel nitrite (Ni(NO2)2) were used as raw materials for the synthesis of nanometer tungsten powder. A sol was prepared by dissolving ammonium metatungstate, niobium nitrite and nickel nitrite with a small amount of polyethylene glycol (PEG-2000) into the distilled water. The nanometer W–0.1 wt.%Nb–0.1 wt.%Ni composite powder was then gained by a multi-step process, consisting of sol–spraying–drying of the solution at 250–350 °C, calcining at 300 °C for 2 h and a subsequent two-step reduction process (600 °C and 750 °C) in hydrogen atmosphere for 1.5 h and 2.5 h, respectively. The calcined composite powder was ball milled in a W alloys bottle for different times, and the milling conditions are listed in Table 1. The ball mill activated nanometer tungsten powders were obtained and a pure tungsten powder without addition of niobium and nickel was also synthesized using the same production process for comparison, the pure tungsten powder, W–0.1 wt.%Nb–0.1 wt.%Ni composite powder, W–0.1 wt.%Nb– 0.1 wt.%Ni composite powder ball-milled for 10 h and 15 h were denoted as W, W0, and W10 and W15, respectively. X-ray diffraction meter (D/ruax2550PC, Japan) was employed to identify the phases of these powders, and ultra high resolution field emission scanning electron microscopy (SEM, NOVA TM NanoSEM 230, Czech) was used to characterize the morphology of these powders. The oxygen content in these powders were determined by a nitrogen/oxygen/hydrogen determination (TCH-600, USA), the specific surface area of each powder was measured by a BET surface area analysis instrument (Monosorb Autosorb-1, USA). The crystalline phase was determined using the X-ray diffraction (XRD) (Rigaku, Japan). The broad–scan analysis was typically conducted within the 2-h range of 10–80° using the Cu Ka (k = 1.542 Å) radiation. The narrow scan analysis was conducted within the 2-h range of 20–30° and was subsequently Table 1 Ball milling parameters used in the present work. Milling medium Grinding medium Ball to powder ratio Liquid to solid ratio Plate and bowl speed Milling time
Ethanol Tungsten balls (3–10 mm in diameter) 3:1 (mass ratio) 3:1 (volume ratio) 240 rpm 10 h, 15 h
291
used to determine ‘‘Bragg’’ grain size. The BET grain size was determined using the surface area measurement technique. It was found that these tungsten powders were difficult to be shaped because of their nanometer particle size, so they were mixed with 0.5 wt.% paraffin prior to forming, then they were pressed into standard tensile samples by two-direction cold pressing with pressure of 250 MPa. The green compacts were pre-sintered at 1000 °C in atmospheric pressure of 5 Pa for 2 h in order to eliminate the paraffin, later pre-sintered compacts were sintered in tungsten rod furnace at different temperatures (1500 °C, 1600 °C, 1700 °C, 1800 °C, 1900 °C, 1950 °C, 2000 °C and 2030 °C) for 2 h, heating rate was 2 °C/min and flowing H2 was employed as the protective atmosphere. The densities of the sintered specimens were measured by Archimedes principle with deionized water as the immersing medium. The quasi–static mechanical properties of these specimens were measured by a standard Instron 3369 material test machine (USA) and the tensile speed was about 1 mm/min according to ASTM:A356 (6 mm 3 mm 60 mm). The hardness of specimen was tested by a nanoindentation method and load of 50 mN. The thermal shock resistance can be appraised by water quenching method based on the definition of the critical thermal shock temperature difference (DTcrit). The DTcrit can be measured experimentally by quenching specimens from various elevated temperatures and determining the quenching temperature that results in a reduction of strength for a given specimen geometry [13]. The DTcrit value is defined as 70% of the room temperature strength, which was determined using linear interpolation of the retained strength values as described in ASTM:C1525-04. Before the water quenching, all samples were ground and polished with diamond slurries down to a 1 lm finish and the edges of all samples were chamfered to minimize the effect of stress concentration due to machining flaws. At least ten samples were tested for each experimental condition and all samples were from same billet. The polished rectangular bars for thermal shock testing were heated in the vacuum up to the desired temperature difference and held for 10 min to eliminate any temperature gradient effect before quenching by dropping parallel to their tensile surface into the water bath. The temperature of the water bath was controlled to about 25 °C by adjusting the cooling water flow. The water quenching temperature differences were 200, 300, 400, 500, 600 and 700 °C. The 5 Pa was kept for the samples heated in vacuum and the time taken for the transfer from the furnace to the water bath was less than 1 s. A digital microhardness tester (HXD-1000T) was employed to determine the microhardness of the sintered tungsten bulk, and the tensile strength was tested by means of mechanical testing machine (Instron3369, USA). 3. Results and discussions 3.1. Phase and performances of these powders The XRD patterns of four kinds of powders (W, W0, and W10 and W15) are shown in Fig. 1, which indicated an increase in peak broadening with progress of milling. The peaks of Nb and Ni were not detected due to low amounts of Nb and Ni. Using a combination of W(1 1 0) and W(2 1 1) peaks, the grain size and lattice distortion were calculated according to the XRD patterns as shown in the following equation [14]:
b cos h ¼ 0:94ðk=dÞ þ 4e sin h
ð1Þ
where b is the full-width at half-maximum (FWHM), h is Bragg angle, k is the X-ray wavelength, d is grain size and e is lattice distortion. The broadening factors that are not induced by ball milling were taken out by the following equation [14,15]:
292
H. Zhang et al. / Materials and Design 58 (2014) 290–297
Fig. 1. XRD patterns of four kinds of powder.
b2 ¼ b2M b2S
ð2Þ
where bS and bM are FWHM of initial powder and milled powders, respectively. Dislocation densities of these powders were calculated by taking internal stress e into the following equation [16,17]:
q¼
6pE e 2 G lnðr=r 0 Þ b
ð3Þ
where E is Young’s modulus, G shear modulus, b the magnitude of the Burgers vector and r and r0 the outer and inner cut-off radius, respectively. The modulus was evaluated from the slopes of load deflection curves of strength tests. A static extensometer was used to measure the deflection with an error in the measurement of 0.1%. Parameters b, r and r0 are automatically given by X-ray diffraction meter. With E = 452 GPa, G = 176 GPa, b = 4.37 1010 m and ln(r/ r0) = 4. The particle sizes of four kinds of powders were calculated based on the BET specific surface area using the following equation [18]:
dBET ¼ 6=ðSBET qTheory Þ
ð4Þ
where dBET is the particle size, SBET is the powder specific surface area and qTheory is the theoretical density of tungsten, i.e. 19.3 g/ cm3. Table 2 lists the characteristic parameters of the four kinds of powders based on XRD and calculation results. The X-ray peak broadening was due to the combined effect of the structural refinement (finer grain size). Furthermore, the increase in lattice distortion, residual strain and dislocation density, which also was confirmed by data listed in Table 2. BET grain size was too larger than ‘‘Bragg’’ grain size, which was because the specific surface area obtained by the surface area measurement technique was the specific surface area of agglomerated particles, was not the specific surface area of single nanoparticles. The increase in lattice distortion, residual strain and dislocation density could significantly increase the sinterability of these tungsten powders [19]. The impurities
Table 2 Characteristic parameters of the four kinds of powders. Powder
W W0 W10 W15
Grain size (nm)
Lattice distortion (%)
Specific surface area (m2/ g)
dBET (nm)
56.9 51.3 44.1 39.8
0.1521 0.1431 0.1874 0.2002
1.51 2.26 1.44 1.59
210 140 210 200
Oxygen content (wt.%)
Dislocation density (1014 m2)
0.0667 0.2213 0.3613 0.4132
1.57 1.39 2.26 2.59
resulted from polyethylene glycol were carbon, hydrogen and oxygen. Furthermore, the impurities were inevitably included nitrogen element. Literature has confirmed that coarsening in nonoxides bulk materials, including tungsten alloys, is promoted generally by oxygen present as oxide impurities on the particle surfaces, and if the oxygen existed, the carbon element would be depleted. Therefore, the oxygen content was calculated, whereas other elements were ignored. It could be seen from Table 2 that the oxygen content had been rising all the time during milling, but its amount was still less than 0.4 wt.%. In the present work, maximum content of oxygen was 0.4132, which was a much higher level than that of the dopants. The formation of the nano-scale powders indicated that the effects of oxygen on grain size were less than that of the dopants. The grain size was mainly affected by the combination of the lattice distortion, residual strain and dislocation density. Furthermore, the melting point of Nb is 2468 °C and maximum sintering temperature is 2030 °C, so a partial loss of Nb is not considered. The melting point of Ni is 1455 °C and Ni has a very low vapor pressure under the protective atmosphere, so a partial loss of Ni is negligible. It is expected that the content of Ni and Nb is constant. The content of Ni and Nb in the specimen sintered at 2030 °C was analyzed and the results confirmed the obvious change in the content of Ni and Nb in the specimen was not detected. Fig. 2 shows SEM images of four kinds of tungsten powders. It could be seen from Fig. 2(a) and (b) that W and W0 powder particles are of irregular polyhedron in shape, and there is serious aggregation among particles. As shown in Fig. 2(b), (c) and (d), the W–Nb–Ni composite powder broke in the mode of brittle fracture was measured during ball milling, and the BET particle size increased significantly when the W–Nb–Ni powder was milled for 10 h and then decreased slightly after being milled for 15 h. Such a change process could be explained as the following: at the initial stage of ball milling (ball milled for 10 h), a large amount of small particles were produced and they were readily to aggregate because of the generation of new surfaces [20], which resulted in the reduction in the specific surface area and the improvement of the sinterability of these tungsten powders [19]. At the final stage of ball milling (ball milled for 15 h), the ball milling energy is mainly consumed in the process of broking the aggregation. At last, the particle refinement and aggregation process reached a balance point, and the particle dispersion became better, which led to the increase in specific surface area and the improvement of the sinterability of these tungsten powders [19]. 3.2. Densification process The geometrical density of the test samples in the state ‘‘as pressed’’ as well as after driving out the paraffin at 1000 °C of four kinds of powders are listed in Table 3, which revealed that the linear shrinkages of W–Nb–Ni composite was obviously greater than that of the pure W powder, indicating that the addition of Nb and Ni had played a significant activation role for the sintering of tungsten at this stage. At the same time, the linear shrinkages of W–Nb–Ni composite were also greater than that of W15, which was further greater than that of W10. The apparent density of W–Nb–Ni composite powder increased greatly after ball milling and the green compacts of W10 and W15 had greater density compared to W0 when they were pressed under the same pressure. As for the activated sintering, the shrinkage DL/L0 varies as [21]:
DL g XdC cSV DA t ¼ L0 D4 kT
ð5Þ
where X is the atomic volume, d is the width of the second-phase activator, g is a collection of geometric terms, C is the solubility of the base phase in the activator, cSV is the solid–vapor surface energy, DA is the diffusivity of the base phase in the activator, D is
H. Zhang et al. / Materials and Design 58 (2014) 290–297
293
Fig. 2. SEM morphologies of the four kinds of powder (A) W; (B) W0; (C) W10; (D) W15.
Table 3 Geometrical density of the test samples in the state ‘‘as pressed’’ as well as after driving out the paraffin at 1000 °C of four kinds of powders. Specimen
W (%)
W0 (%)
W10 (%)
W15 (%)
As pressed As pre-sintered at 1000 °C
40.1 40.4
40.3 52.9
42.9 46.7
43.2 52.1
the particle size, k is Boltzmann’s constant, t is sintering temperature and T is the absolute temperature. The greater diffusivity DA in the tungsten material was reduced due to the addition of trace Nb and Ni compared with pure W when there is no activator. Furthermore, the particle size D was reduced by ball milling. Therefore, Eq. (5) could well explain why the activated tungsten powders have a much higher shrinkage extent than the pure W powder. Densification behavior of four kinds of powders at different sintering temperatures is presented in Fig. 3. Compared with the results of other groups [3,4,8], the nano-size of the pure W powder was sintered at 1600 °C for 2 h, and the relative density was about 90%. In the present work, the relative density of the nano-size of the pure W powder was about 87%, which was attributed to the different sintering conditions and treatment approach of powders. It is clear that the W–Nb–Ni composite powder sintered at 1900 °C reached maximum relative density as high as 99.1% of theory density and the pure W material sintered at 2000 °C reached maximum relative density as high as 93.7 of theory density. This indicated that the final sintering densification of tungsten was greatly improved due to the addition of trace Nb and Ni. At the same time, ball milling treatment could also promote the further densification of W–Nb–Ni composite powder to near full density. In consequence, W10 could be sintered to 99.3% of theory density at 1600 °C while W15 could be sintered to 99.4% of theory density at this temperature. This revealed that the activated tungsten pow-
Fig. 3. Change curves of relative density of four kinds of powder versus sintering temperature.
ders were superior to the pure tungsten powder in improving the sintering densification rate as well as sintering densification extent, which could also be observed from Fig. 4. Fig. 4 shows micrographs of the sintered bulk of the four kinds of powder sintered at 1600 °C (a) W; (b) W0; (c) W10; (d) W15. No obvious flaws such as pores, gaps and cracks were detected, which confirmed the results of density. The densification rate of the compact was calculated by [22]:
@q DB ¼ @t G3 kT
ð6Þ
where D is the diffusion coefficient, B is a collection of material and geometric constants, k is Boltzmann’s constant, G is the grain size,
294
H. Zhang et al. / Materials and Design 58 (2014) 290–297
Fig. 4. Micrographs of the sintered bulk of the four kinds of powder sintered at 1600 °C (A) W; (B) W0; (C) W10; (D) W15.
and T is the temperature. This equation indicated that the densification rate would increase as the diffusion coefficient increased due to addition of trace Nb and Ni, and grain size decreased due to ball milling. It could be also seen from Fig. 3 that the densities of W10 and W15 sintered at all temperatures were similar and the densities did not increase further when the temperature was more than 1600 °C. It revealed that increasing ball milling time from 10 h to 15 h has slight beneficial effect on the sintering densification of W–Nb–Ni composite powder. At the same time, 1600 °C was an optimum sintering temperature for W10 and W15 powders because a further increasing of temperature could provide little effects on their sintering densities. 3.3. Grain size change behavior of the sintered tungsten materials Fig. 5 displays the change behavior of grain sizes versus sintering temperatures. It could be observed that the grain growth of the sintered compacts of W0, W10 and W15 increased with increasing sintering temperature. The grain size of raw W was very small whereas that of sintered W was much larger, which was still less than the grain size of W0, W10 and W15. The grain growth of tungsten was attributed to the introduction of oxygen impurities, not attributed to the addition of trace Nb and Ni. The grain growth of the sintered compact further increased as the ball milling time increased from 10 h to 15 h, which also confirmed due to the grain growth of tungsten due to the introduction of oxygen impurities [23]. Grain growth in a sintered compact has been found to follow a cubic relationship with time [22]:
G3 ¼ G30 þ Mt
ð7Þ
where G0 is the initial grain size and M is the grain boundary mobility. Compared with small difference in original grain size, the great
Fig. 5. Change curves of average grain size of the sintering bulk of four kinds of powder versus sintering temperature.
difference in grain size after sintering was attributed to the grain boundary mobility. Therefore, the addition of trace Nb and Ni was favorable to improve the mobility of tungsten grain boundaries, and ball milling further accelerated this mobility. Compared with the original grain size, the obvious grain growth was measured and grain size was initially limited by pore pinning, the grain size could be approximated by:
G3 ¼ Mt
ð8Þ
The effect of grain growth on the densification rate was calculated by substituting Eq. (8) into Eq. (6):
@q DB ¼ @t MkTt
ð9Þ
H. Zhang et al. / Materials and Design 58 (2014) 290–297
295
where H is microhardness, qs is fractional sintered density, H0 and
H are constants and G is grain size. The change curves of micro-
Fig. 6. Change in microhardness with sintering temperature of the four kinds of sintered compacts.
It could be seen from Eq. (9) that the increase in diffusivity and the decrease in grain boundary mobility would increase the densification rate. 3.4. Microhardness and tensile strength From previous studies on sintered products, mechanical properties had been found to be highly dependent on density and grain size [24,25]. In the present work, the change behavior of microhardness and tensile strength of four kinds of sintered tungsten materials with sintering temperature were investigated in detail. The hardness for a brittle material such as pure tungsten at room temperature followed the Hall–Petch relation dependence on inverse square-root grain size [26]:
H H ¼ qs H0 þ pffiffiffiffi G
ð10Þ
hardnesses versus sintering temperatures are shown in Fig. 6. It was clear that the microhardness of sintered W mainly increased with the increasing temperature, which was consistent with its increasing sintering densification. The microhardness of W0 increased as the sintered temperature increased to 1800 °C and then the microhardness decreased due to the increase in grain size. The maximum microhardness of W10 and W15 sintered at 1600 °C was because of their near full density and then the microhardness decreased with the increasing temperature due to the increase in the grain size. The fractography of four kinds of tungsten powders sintered at 1600 °C are shown in Fig. 7. The brittle rupture of four kinds of sintered tungsten was readily observed. In order to eliminate the effect of hydrogen brittleness on the tensile strength, the tensile strength values of four kinds of sintered compacts after the samples were subjected to vacuum treatment at 1350 °C are listed in Table 4. The tensile strength values of four kinds of tungsten powders sintered at 1600 °C were 833 ± 80, 1450 ± 130, 1570 ± 139 and 1430 ± 143 MPa for W, W0, W10 and W15, respectively. The lower tensile strength for pure W specimen was due to low relative density. The tensile strength values of W0, W10 and W15 specimens were slightly greater than the data reported in the literature [2], which was attributed presumably to ultrofined grain due to the addition of trace Nb and Ni. The relative density of sintered W was 93.7% of theory density, which was much lower than 99.1% for W0, 99.3% for W10 and 99.4% for W15, whereas the tensile strength of sintered W was statistically equivalent to that of W10 and greater than the tensile strength of sintered W0 and W15 compacts. This indicated that the tensile strength of sintered materials was significantly affected by the addition of trace Nb and Ni as well as the ball milling. Many researchers have confirmed that the strength of sintered materials have a close relationship with the relative density [27]. Furthermore, the effect of grain size on brittle materials at higher densities
Fig. 7. Fractography of the sintered bulk of four kinds of powder sintered at 1600 °C (A) W; (B) W0; (C) W10; (D) W15.
296
H. Zhang et al. / Materials and Design 58 (2014) 290–297
Table 4 Tensile strength of the sintered compacts of four kinds of powders sintered at 1600 °C.
Tensile strength (MPa)
W
W0
W10
W15
833 ± 80
1450 ± 130
1570 ± 139
1430 ± 143
and production technology it would be more meaningful to compare the pure W sample sintered to its highest density (here at 2030 °C) to the alloyed sample sintered already at 1600 °C to a superior density. The pure W sample sintered at 2030 °C has mechanical properties equivalent to one of W10 sintered at 1600 °C. However, W10 has better thermal shock resistance than pure W sample, which was due to the microstructure because the gain size of pure W sample sintered at 2030 °C was a much greater than that of W10.
4. Conclusions
Fig. 8. Strength values of pure W and W10 before and after water quenching with increasing the thermal shock temperature difference.
has been noted. In such case, the strength r could be calculated in terms of grain size d and porosity e as following [28]:
r ¼ r0 da exp½be
ð11Þ
where r0 is a strength constant, a and b are empirical constants. It could be seen from the Eq. (11) that the strength of sintered material would decrease if the porosity e and/or grain size d increased, and this would be used to explain the strength change of four kinds of sintered activated tungsten powders. Although sintered composite powder had the smallest grain size, it had the lowest tensile strength compared with that of W10 and W15, due to its lowest sintering density. The sintered W10 and W15 materials had the similar porosity (sintering density), but the grain size of sintered W10 material was much smaller than that of W15 material. Therefore, the strength of sintered W10 material was higher than that of the sintered W15 material. Due to W10 sintered at 1600 °C had best mechanical properties, the W10 was used to evaluate the thermal shock resistance of W alloys and the thermal shock resistance of pure W material was also tested for comparison. The strength values of pure W and W10 before and after water quenching are shown in Fig. 8 with increasing the thermal shock temperature difference. The measured DTcrit for W10 as high as 541 °C was greater than 440 °C for pure W, which indicated that W10 had good thermal shock resistance than pure W. Furthermore, the critical temperature difference is also calculated by a triaxial modification of Hooke’s law in terms of the average strength of the material (r), Young’s modulus (E), Poisson’s ratio (v) and coefficient of thermal expansion (a) as shown in the following equation [18].
R ¼ DT crit ¼
rð1 tÞ Ea
ð12Þ
Therefore, the improvement of thermal shock resistance of the W10 alloy was due to the increase in the flexural strength. Furthermore, the pure W sample displayed here exhibits a density of only 87%, and thus appears ill suited for such a technological comparative test. To underline the potential of the developed special alloy
Activated nanometer tungsten powder (i.e. W–0.1 wt.%Nb– 0.1 wt.%Ni composite powder) was prepared by physical and chemical method and subsequent ball-milling treatment, and the characteristics of these powders milled by different times (0 h, 10 h and 15 h) were investigated. The effect of ball-milling time on the densification behavior and mechanical properties was investigated, which was further compared with that of the nanometer pure tungsten powder. The results indicated that the sintering shrinkage rate and the sintering densification rate were greatly accelerated by the activation treatment, and the activated tungsten powder milled by 10 h could be sintered to near full density at 1600 °C, whereas the activation treatment also could promote quick growth of tungsten grains. Although sintered composite powder had the smallest grain size, it had the lowest tensile strength compared with that of W10 and W15, due to its lowest sintering density. The sintered W10 and W15 materials had the similar porosity (sintering density), but the grain size of sintered W10 material was much smaller than that of W15 material. Therefore, the strength of sintered W10 material was higher than that of the sintered W15 material. Furthermore, the improvement of thermal shock resistance of the W10 alloy was due to the increase in the flexural strength. Acknowledgements This work is supported by Foundation of Sichuan Educational Committee (Research on pattern recognition, No. 12ZB342) and Oil & Gas Pipeline Detection Technology Research Based on Multi-sensor Data Fusion (No. 2012XJZ021), the project is founded by Technology Research Dept. CNOOC Research and 2012 Youth Foundation of Southwest Petroleum University. References [1] Schade P. 100 years of doped tungsten wire. Int J Refract Met Hard Mater 2010;28:648–60. [2] Ravi Kiran U, Panchal A, Sankaranarayana M, Nandy TK. Tensile and impact behavior of swaged tungsten heavy alloys processed by liquid phase sintering. Int J Refract Met Hard Mater 2013;37:1–11. [3] Çalisßkan NK, Durlu N, Bor Sß. Swaging of liquid phase sintered 90W–7Ni–3Fe tungsten heavy alloy. Int J Refract Met Hard Mater 2013;36:260–4. [4] Liu WS, Ma YZ, Zhang JJ. Properties and microstructural evolution of W–Ni–Fe alloy via microwave sintering. Int J Refract Met Hard Mater 2012;35:138–42. [5] Ravi Kiran U, Panchal A, Sankaranarayana M, Nandy TK. Tensile and impact behavior of swaged tungsten heavy alloys processed by liquid phase sintering. Int J Refract Met Hard Mater 2013;37:1–11. [6] Telu S, Mitra R, Pabi SK. High temperature oxidation behavior of W–Cr–Nb alloys in the temperature range of 800–1200 °C. Int J Refract Met Hard Mater 2013;38:47–59. [7] Wu ZJ, Wang Z, Shi GD, Sheng J. Effect of surface oxidation on thermal shock resistance of the ZrB2–SiC–ZrC ceramic. Compos Sci Technol 2011;71:1501–6. [8] Ryu T, Hwang KS, Choi YJ, Sohn HY. The sintering behavior of nanosized tungsten powder prepared by a plasma process. Int J Refract Met Hard Mater 2009;27:701–4. [9] Wang HT, Fang ZZ, Hwang KS, Zhang HB, Siddle D. Sinter-ability of nanocrystalline tungsten powder. Int J Refract Met Hard Mater 2010;28:312–6. [10] Pusavec F. Porous tungsten machining under cryogenic conditions. Int J Refract Met Hard Mater 2012;35:84–9.
H. Zhang et al. / Materials and Design 58 (2014) 290–297 [11] Ma YZ, Zhang JJ, Liu WS, Yue P, Huang BY. Microstructure and dynamic mechanical properties of tungsten-based alloys in the form of extruded rods via microwave heating. Int J Refract Met Hard Mater 2014;42:71–6. [12] Hu P, Wang Z, Sun X. Effect of surface oxidation on thermal shock resistance of ZrB2–SiC–G composite. Int J Refract Met Hard Mater 2010;28:280–5. [13] Wang Z, Qu Q, Wu ZJ, Shi GD. The thermal shock resistance of the ZrB2–SiC–ZrC ceramic. Mater Des 2011;32:3499–503. [14] Zhang ZW, Zhou JE, Xi SQ, Ran G, Li PL. Phase transformation and thermal stability of mechanically alloyed W–Ni–Fe composite materials. Mater Sci Eng A 2004;379:148–53. [15] Wang Z, Wu ZJ, Shi GD. The oxidation behaviors of a ZrB2–SiC–ZrC ceramic. Solid State Sci 2011;13:534–8. [16] Grahle P, Arzt E. Microstructural development in dispersion strengthened NiAl produced by mechanical alloying and secondary recrystallization. Acta Mater 1997;45:201–11. [17] Zhang L, Ukai S, Hoshino T, Hayashi S, Qu XH. Y2O3 evolution and dispersion refinement in Co-base ODS alloys. Acta Mater 2009;57:3671–82. [18] Wang Z, Xie B, Zhou WY, Shi GD, Wu ZJ. Thermophysical properties of TiB2–SiC ceramics from 300 °C to 1700 °C. Int J Refract Met Hard Mater 2013;41:609–13. [19] Zhang XH, Wang Z, Hu P, Han WB, Hong CQ. Mechanical properties and thermal shock resistance of ZrB2–SiC ceramic toughened with graphite flake and SiC whiskers. Scripta Mater 2009;61:809–12. [20] Wu ZJ, Wang Z, Qu Q, Shi GD. Oxidation mechanism of a ZrB2–SiC–ZrC ceramic heated through high frequency induction at 1600 °C. Corros Sci 2011;53:2344–9.
297
[21] German RM. Sintering theory and practice. New York: John Wiley; 1996. [22] Zovas PE, German RM. Retarded grain boundary mobility in activated sintered molybdenum. Metall Trans A 1984;15A:1103–10. [23] Zhou SB, Wang Z, Zhang W. Effect of graphite flake orientation on microstructure and mechanical properties of ZrB2–SiC–graphite composite. J Alloys Compd 2009;485:181–5. [24] Wang Z, Wang S, Zhang XH, Hu P, Han WB, Hong CQ. Effect of graphite flake on microstructure as well as mechanical properties and thermal shock resistance of ZrB2–SiC matrix ultrahigh temperature ceramics. J Alloys Compd 2009;484:390–4. [25] Zhang XH, Wang Z, Sun X, Han WB, Hong CQ. Effect of graphite flake on the mechanical properties of hot pressed ZrB2–SiC ceramics. Mater Lett 2008;62:4360–2. [26] Wang Z, Wu ZJ, Shi GD. Fabrication, mechanical properties and thermal shock resistance of a ZrB2–graphite ceramic. Int J Refract Met Hard Mater 2011;29:351–5. [27] Wang Z, Wu ZJ, Shi GD. Effect of annealing treatment on mechanical properties of a ZrB2–SiC–graphite ceramic. Mater Sci Eng A 2011;528:2870–4. [28] Wang Z, Li SC, Wang M, Wu GQ, Sun XM. Effect of SiC whiskers on microstructure and mechanical properties of the MoSi2–SiCw composites. Int J Refract Met Hard Mater 2013;41:489–94.