Vacuum annealing of nanocrystalline WC powders

Article abstract of DOI:10.1134/S0020168512060088

The effect of vacuum annealing temperature on the chemical and phase compositions, particle size, and lattice strain of nanocrystalline tungsten carbide (WC) powders with a particle size from 20 to 60 nm has been studied by X-ray diffraction and electron microscopy. The results demonstrate that vacuum annealing of WC nanopowders at t_ann ≤ 1400°C is accompanied by a marked decrease in carbon content and changes in phase composition due to carbon desorption from the surface of the powder as a result of the interaction of carbon with oxygen impurities. In addition, annealing leads to an increase in particle size due to coalescence of aggregated nanoparticles and reduces the lattice strain of the powder. Pleiades Publishing, Ltd., 2012.

Full text of DOI:10.1134/S0020168512060088

ISSN 0020ꢀ1685, Inorganic Materials, 2012, Vol. 48, No. 7, pp. 680–690. © Pleiades Publishing, Ltd., 2012.
Original Russian Text © A.S. Kurlov, A.I. Gusev, 2012, published in Neorganicheskie Materialy, 2012, Vol. 48, No. 7, pp. 781–791.
Vacuum Annealing of Nanocrystalline WC Powders
A. S. Kurlov and A. I. Gusev
Institute of Solid State Chemistry, Ural Branch, Russian Academy of Sciences,
Pervomaiskaya ul. 91, Yekaterinburg, 620990 Russia
eꢀmail: kurlov@ihim.uran.ru
Received November 24, 2011
Abstract—The effect of vacuum annealing temperature on the chemical and phase compositions, particle
size, and lattice strain of nanocrystalline tungsten carbide (WC) powders with a particle size from 20 to 60 nm
has been studied by Xꢀray diffraction and electron microscopy. The results demonstrate that vacuum annealꢀ
ing of WC nanopowders at tann
≤
1400 C is accompanied by a marked decrease in carbon content and changes
°
in phase composition due to carbon desorption from the surface of the powder as a result of the interaction
of carbon with oxygen impurities. In addition, annealing leads to an increase in particle size due to coalesꢀ
cence of aggregated nanoparticles and reduces the lattice strain of the powder.
DOI: 10.1134/S0020168512060088
INTRODUCTION
were stored in air: microcrystalline powder manufacꢀ
tured at the Kirovgrad Hard Alloys Plant (OAO KZTS,
Kirovgrad, Russia), with an average particle size of
Currently, the use of nanocrystalline tungsten carꢀ
bide (WC) powders is regarded as the most promising
approach to the fabrication of nanostructured hard
alloys with a reduced sintering temperature and
enhanced hardness, strength, and fracture toughness
[1, 2].
Many processes for the preparation of nanocrystalꢀ
line WC powders have been proposed to date [3, 4].
There is particular interest in highꢀenergy milling of
microcrystalline tungsten carbide (WC) powder [5],
which is a “topꢀdown” nanotechnology, and in plasma
synthesis from a tungsten oxide and hydrocarbon [6],
which is a “bottomꢀup” nanotechnology.
9 m, which will be referred to as WC(micro); nanocꢀ
µ
rystalline powder with an average particle size of
20 nm, produced by grinding the WC(micro) powder
in a Retsch PM 200 planetary ball mill (the milling
conditions were similar to those described previously
[5, 9, 10]), WC(mill); and nanocrystalline powder with
an average particle size of 60 nm synthesized at OAO
VNIIETO (Moscow, Russia) through a plasma process
using tungsten(III) oxide (WO3) and propane (C3H8),
which were reacted in a lowꢀtemperature hydrogen
plasma flow, followed by thermochemical processing
(the synthesis procedure was similar to that described in
Tsvetkov and Panfilov [6]), WC(plasma). In addition,
we annealed a mixture of the WC(mill) nanopowder
and 2 wt % MTꢀ900 carbon black (the mixture will be
referred to as WC(mill) + C).
WC powders are most widely used in the fabricaꢀ
tion of hard alloys, the most widespread of which are
WC–Co alloys. The highest sintering temperature of
microcrystalline WC–Co alloys in conventional proꢀ
cesses in vacuum or under a protective or inert atmoꢀ
Loose WC powders (2–3 g) were annealed in a vacꢀ
sphere is 1380 to 1420 С. As shown in previous studꢀ
°
5
uum of 0.0013 Pa (10⎯ mm Hg) in an SShVꢀ
ies concerned with the oxidation of WC powder [7,
8], a decrease in WC powder particle size is accompaꢀ
nied by a reduction in oxidation onset temperature
and increase in oxidation rate [7, 8]. It is, therefore,
important to assess the thermal stability of the size
and composition of WC nanoparticles at temperaꢀ
1.2.5/25I1 laboratory furnace with tungsten lining and
heaters. The annealing schedule was as follows: slow
(1–2 h, depending on the highest annealing temperaꢀ
ture) heating to
cooling of the powder. The annealing temperatures
were 400, 600, 800, 1000, 1200, and 1400 . In each
tann, holding at tann for 1 h, and furnace
°C
tures of up to 1400 С.
°
annealing experiment, only the asꢀprepared WC powꢀ
ders were used.
The phase composition of the WC powders before
and after annealing was determined by Xꢀray diffracꢀ
The purpose of this work was to examine the temꢀ
perature effect on the phase composition and size of
tungsten carbide (WC) nanoparticles vacuumꢀ
annealed at temperatures from 400 to 1400 С.
°
tion (XRD) in the angular range
scan step of 0.03 and a counting time per data point
of 2 s (DRONꢀUM1 diffractometer, Cu radiaꢀ
2θ
= 30
°
–125
°
with a
°
EXPERIMENTAL
K
α1, 2
In our vacuum annealing experiments, we used tion) XRD patterns were analyzed numerically using
three WC powders differing in particle size, which Philips Analytical X’Pert Plus (Version 1.0) software.
680

VACUUM ANNEALING OF NANOCRYSTALLINE WC POWDERS
681
The particle size distributions in the WC powders were relatively broad, so neighboring reflections in the
were assessed using a Horiba Partica LAꢀ950V2 laser angular range θ > 60 overlapped and merged. Analꢀ
2
°
diffraction particle size analyzer. Prior to measureꢀ ysis of the broadening of the diffraction peaks from the
ments, the powder to be analyzed was suspended in nanopowders showed that it was due to both the small
water. The average size of the powder particles was also particle size,
evaluated from diffraction data and by scanning elecꢀ uated from the width of XRD line profiles by a proceꢀ
tron microscopy (SEM) on a JEOL JSMꢀ6390LA. In dure described elsewhere [3, 11, 12]. We obtained
addition, the crystallite size and lattice strain of the 20 10 nm and = 0.83 .05% for the WC(mill)
WC(mill) and WC(plasma) nanocrystalline powders powder and = 60 10 nm and = 0.43 .05% for
were determined from the width of XRD line profiles. the WC(plasma) powder. The average size of aggreꢀ
To a first approximation, the average crystallite size gated particles, aggr, evaluated by laser diffraction on
thus determined can be regarded as the average partiꢀ the Horiba Partica LAꢀ950V2 (Fig. 2) was almost one
cle size of the nanopowders. order of magnitude greater than the average crystallite
size of the WC nanopowders. This was due to the
D, and lattice strain,
ε.
D
and
ε
were evalꢀ
D
=
D
ε
ε
0
D
ε
0
D
D
D
The WC powders were analyzed for total carbon
Ctotal and free (uncombined) carbon Cfree on a Metaꢀ
vakꢀCSꢀ30 analyzer. Oxygen impurities were deterꢀ
mined by a carrier gas method with a CoulomaticꢀO
analyzer and by energyꢀdispersive Xꢀray (EDX) analꢀ
ysis on a JEOL JSMꢀ6390LA electron microscope
equipped with a JEDꢀ2300 spectrometer system.
Adsorbed water was determined from the weight loss
of a powder sample calcined for 1 h in a vacuum of
formation of large aggregates of nanoparticles. The
SEM micrographs in Fig. 3 indicate that the particle
size of the WC(mill) and WC(plasma) nanopowders
differs little from the average crystallite size, but the
nanoparticles indeed form large aggregates, up to sevꢀ
eral hundred nanometers in size. In the WC(micro)
powder, the average particle size is 9
both laser diffraction (Fig. 2, curve
microscopy (Fig. 3a) data, which are in good agreeꢀ
ment. The narrow diffraction peaks of this powder
µ
m according to
) and electron
1
1
×
10–4 Pa at 400
C. Spectral analyses were carried
out on a PerkinElmer SCIEXꢀELAN 9000 mass
spectrometer.
°
(Fig. 1, scan 1) indicate that it consists of large partiꢀ
cles and is free of lattice strain.
The specific surface area of powder can be found as
RESULTS AND DISCUSSION
S
D
=
6/
is the average particle size. The density of WC is
15.67 g/cm3. From this value and the above average
particle sizes, we find that the of the WC(micro),
WC(mill), and WC(plasma) powders before vacuum
annealing was 0.04, 19.2, and 6.4 m2/g. The
6.4 m2/g obtained for the WC(plasma) powder is in
reasonable agreement with the
= 6.0–9.5 m2/g
inferred from BET measurements.
ρ
D
[3], where
ρ
is the density of the material and
ρ
=
The effect of vacuum annealing on the WC(mill)
and WC(plasma) nanocrystalline powders was studied
in comparison with the WC(micro) microcrystalline
powder. According to spectral analysis data, the WC
powders before vacuum annealing were similar in
impurity content: within 0.1 wt % Na and within
0.01 wt % Fe, Ni, Mo, and Sn. The content of other
metallic impurities did not exceed 0.005 wt %. The
adsorbed water content of the WC(micro), WC(mill),
and WC(plasma) powders before vacuum annealing
was 0.1, 1.0, and 0.9 wt %, respectively. The three WC
powders were analyzed chemically and characterized
by particle size measurements.
S
S
=
S
Note that the WC(micro) powder, which had the
smallest specific surface area, had the lowest adsorbed
water content, 0.1 wt %, whereas the asꢀprepared
WC(mill) nanopowder, which had the largest specific
surface area, contained the largest percentage of
adsorbed water, 1.0 wt %.
Figure 1 shows XRD patterns of the WC(micro),
WC(mill), and WC(plasma) powders before vacuum
annealing and the WC(micro) powder vacuumꢀ
The chemical and phase compositions and other
characteristics of the asꢀprepared and annealed WC
powders are presented in the table. According to
chemical analysis data for total and free carbon, the
asꢀprepared powders contained 6.14 0.15 wt % Ctotal
and 0.15 0.10 wt % Cfree (table), which corresponds
to the tungsten carbide WC. It is worth noting that the
unannealed WC(mill) nanopowder, produced by millꢀ
ing the WC(micro) powder, contained more Ctotal by
annealed at 1200 C. The asꢀprepared WC(micro)
°
microcrystalline powder contained, in addition to the
higher tungsten carbide WC (hexagonal structure,
6
sp. gr. P m
bide (hexagonal structure, sp. gr.
scan ). The WC(mill) nanopowder was prepared by
milling the microcrystalline powder, so it also conꢀ
tained a small amount of W2C (Fig. 1, scan ). The asꢀ
prepared WC(plasma) plasmaꢀsynthesized nanopowꢀ
der (Fig. 1, scan ) consisted only of the higher tungꢀ
sten carbide WC.
Note that impurity phases in the WC nanopowders
were rather difficult to detect by XRD (Fig. 1, scans
2),
2
wt % W2C, the lower tungsten carꢀ
P63/mmc) (Fig. 1,
1
3
0.3 wt % compared to the WC(micro) powder.
ꢀ
Almost all of this difference is accounted for by free
carbon, which was essentially missing in the
WC(micro) powder before vacuum annealing. This
seems to be related to the powder preparation process.
4
3,
The oxygen content of the WC(micro), WC(mill),
4) because even reflections from the major phase WC and WC(plasma) powders before vacuum annealing
INORGANIC MATERIALS Vol. 48
No. 7 2012

682
KURLOV, GUSEV
D
= 9 μm
WC (98.0%)
W2C (2.0%)
1
D
= 9 μm
WC (96.5%)
W2C (2.5%)
W (1.0%)
2
3
4
D
D
= 20 nm,
ε
= 0.83%
WC (98.0%)
W2C (2.0%)
= 60 nm, ε = 0.43%
WC (100.0%)
40
60
80
, deg
100
120
2θ
Fig. 1. XRD patterns of WC powders: (
1
) asꢀprepared WC(micro) microcrystalline powder, average particle size of
; ( ) asꢀprepared WC(mill) nanocrystalline powder, = 20 10 nm; (
= 60 10 nm.
9
1
4
µ
m;
(2
) WC(micro) powder vacuumꢀannealed at 1200
°
C
3
D
) asꢀ
prepared WC(plasma) nanocrystalline powder,
D
was 0.11, 0.91, and 1.23 wt %, respectively, according 1.5 wt % and increased the W2C content by 0.5 wt %.
to chemical analysis data and 0.12 0.05, 1.52 0.05
,
In addition, it produced 1.0 wt % W (table). Accordꢀ
ing to chemical analysis data, the annealing reduced
the carbon content of the WC(micro) powder by no
more than 0.01 wt %.
and 1.80 0.1 wt % according to EDX analysis data.
After annealing, no EDX analysis was performed
because, when exposed to air, the annealed WC powꢀ
ders reabsorbed water and oxygen, and their concenꢀ
trations returned to their original level in a few hours.
Vacuum annealing of the WC(mill) powder and
WC(mill) + C mixture. After vacuum annealing of the
WC(mill) nanopowder at 400 and 600 C, no changes
°
Vacuum annealing of the WC(micro) powder. The
in its XRD pattern were detected. At the same time,
chemical analysis showed that raising tann to 600
WC(micro) powder was vacuumꢀannealed at 1200 C.
°
°C
Chemical analysis and XRD data showed that annealꢀ
ing reduced the carbon content of the powder, but reduced the carbon content of the powder by almost
without significant changes in its phase composition. 1 wt % and that the annealed powder contained free
As seen in Fig. 1, the XRD patterns of the WC(micro) carbon (table). Diffraction peak broadening analysis
powder before and after annealing (scans
1,
2
) are indicated that annealing at 600
C had no effect on the
°
almost identical. The only distinction is that the XRD particle size of the powder but reduced the lattice
pattern of the annealed powder contains one, weak strain from 0.83 to 0.64%. Annealing of the WC(mill)
reflection from tungsten. Quantitative phase analysis nanopowder at 800
showed that annealing at 1200 reduced the content tent by 1.5 wt % (relative to the original level) and the
of the major phase (WC) in the WC(micro) powder by free carbon content to zero. After this anneal, the
°C reduced the total carbon conꢀ
°
C
INORGANIC MATERIALS Vol. 48
No. 7
2012

VACUUM ANNEALING OF NANOCRYSTALLINE WC POWDERS
683
18
16
14
12
10
8
Dav = 9 μm
1
D
aggr = 230 nm
2
3Daggr = 510 nm
6
4
2
10 μm
(а)
(b)
(c)
0
0.01
0.1
1
10
100
Particle size
, μm
Fig. 2. Particle size distributions (laser diffraction data) in
the WC powders before annealing: (
1
) WC(micro),
D =
av
9
µ
m; (
2
) WC(mill),
D
= 230 nm; ( ) WC(plasma),
3
aggr
D
= 510 nm.
aggr
XRD pattern of the powder (Fig. 4, scan 2) showed
3
sharp peaks of metallic tungsten (sp. gr. Im m). The
diffraction peaks of WC became narrower, which corꢀ
responded to an increase in average particle size to
30 nm and a decrease in lattice strain to 0.6%. W2C
was missing, even though the unannealed WC(micro)
powder, which was milled to produce the WC(mill)
powder, contained 2 wt % W2C. Annealing of the
1 μm
WC(mill) powder at 800 and 1000 C yielded almost
°
identical phase compositions, which only differed in
WC to W ratio. Raising the annealing temperature to
1000 C was accompanied by a decrease in the total
°
carbon content and an increase in the W content of the
WC(mill) nanopowder. In addition, the WC particle
size increased to 50 nm. After annealing at 1200
the WC(mill) powder contained, in addition to WC
and W, about 7 wt % W2C (Fig. 4, scan ). As ann was
raised to 1400 , the W2C content increased to
36 wt %, whereas W disappeared (Fig. 4, scan ).
Annealing at 1400 reduced the total carbon content
of the WC(mill) powder by about 1.5 wt % relative to
the original level, increased the average particle size by
four times (to 80 nm), and reduced the lattice strain to
0.26%.
°C,
3
t
°C
4
°C
1
μ
m
Fig. 3. SEM micrographs of the tungsten carbide powders
before vacuum annealing: (a) WC(micro), (b) WC(mill),
(c) WC(plasma).
To prevent annealingꢀinduced decarburization of
the WC(mill) powder and maintain its phase purity,
2 wt % carbon (MTꢀ900 carbon black) was added to
the asꢀprepared powder. The WC(mill) + C mixture
annealed WC(mill) + C mixture contained 6.28 wt %
total carbon and 1.28 wt % free carbon.
was vacuumꢀannealed at 1200 C, like the WC(mill)
°
SEM examination showed that vacuum annealing
powder with no carbon additions. The XRD pattern of
the annealed WC(mill) + C powder mixture showed
rather narrow reflections, which were due to only one
of the WC(mill) powder at tann
on the particle size and shape (Fig. 5a), which
remained stable. Annealing at 1200 and above proꢀ
≤
800°
C had little effect
°
C
phase: the higher tungsten carbide WC (Fig. 4, scan 5).
duced coarser particles (Figs. 5b, 5c), which resulted
from the coalescence of aggregated nanoparticles.
According to SEM data, the WC(mill) + C mixture
Diffraction peak broadening analysis indicated that
the anneal increased the average particle size to
110 nm and reduced the lattice strain to almost zero
(0.05%). According to chemical analysis data, the annealed at 1200 contained, in addition to fine partiꢀ
°
INORGANIC MATERIALS Vol. 48
No. 7 2012

684
KURLOV, GUSEV
D
= 110 nm,
ε
= 0.05%
WC (100%)
5
D
= 80 nm, ε = 0.26%
WC (64.0%)
W C (36%)
2
4
D
= 50 nm, ε = 0.40%
WC (78.0%)
W C (7%)
2
W (15%)
3
2
1
D
D
= 30 nm, ε = 0.59%
WC (80.5%)
W (19.5%)
= 20 nm, ε = 0.83%
WC (98%)
W C (2%)
2
30
40
50
2θ, deg
60
70
Fig. 4. XRD patterns of the (
1
) asꢀprepared and (
2
–
4
) vacuumꢀannealed WC(mill) nanopowder and (
5) vacuumꢀannealed
WC(mill) + 2 wt % carbon powder mixture; annealing temperature of (
2
) 800, ( ) 1200, and ( ) 1400 C.
3
,
5
4
°
INORGANIC MATERIALS Vol. 48
No. 7
2012

VACUUM ANNEALING OF NANOCRYSTALLINE WC POWDERS
685
INORGANIC MATERIALS Vol. 48
No. 7 2012

686
KURLOV, GUSEV
1 μm
1 μm
(а)
(b)
1
μ
m
1 μm
(c)
(d)
Fig. 5. SEM micrographs of the WC(mill) nanopowder vacuumꢀannealed at 800 (a), 1200 (b), and 1400
WC(mill) + 2 wt % carbon powder mixture vacuumꢀannealed at 1200 C (d).
°
C (c) and the
°
cles 200 nm in size, large, wellꢀfaceted particles up to Significant changes were only detected after annealing
ꢀ
1.0–1.5
µ
m in size (Fig. 5d).
of the WC(plasma) powder at 1400
the powder contained 33 wt % W2C, whereas the
tungsten content sharply dropped, to 2 wt % (the
presence of tungsten was only evidenced by a weak
reflection at = 40.27 ). Diffraction peak broadenꢀ
°
C (Fig. 6, scan 4):
Ӎ
Thus, the addition of carbon allows WC nanopowꢀ
der to retain its phase purity during annealing but leads
to marked particle growth even at moderate annealing
temperatures.
Ӎ
2θ
°
ing analysis indicated that the anneal increased the
particle size by 20 nm and reduced the lattice strain by
almost a factor of 2 relative to the asꢀprepared powder.
According to chemical analysis data, the anneal
reduced the total carbon content by about 1.5% relaꢀ
tive to the asꢀprepared WC(plasma) powder.
Vacuum annealing of the WC(plasma) powder.
According to XRD data, vacuum annealing at 400 and
600 C produced no changes in the phase composition,
°
average particle size, or lattice strain of the
WC(plasma) nanopowder. Compositional stability of
this powder was also evidenced by chemical analysis
data, which indicated that the total carbon content
SEM examination showed that vacuum annealing
was sufficient for WC formation and that the powder had little effect on the shape of the WC(plasma) powꢀ
contained free carbon. At the same time, annealing at der particles but slightly increased their size (Fig. 7), in
800 C slightly increased the particle size, to 70 nm, agreement with the above diffraction data. The
°
and had a significant effect on the phase composition annealingꢀinduced increase in the particle size of the
of the powder: in addition to the major phase WC, it WC(plasma) powder was less significant than that of
contained about 21 wt % metallic tungsten (Fig. 6, the WC(mill) powder, whose particle size after annealꢀ
scan
carbon content decreased by about 1 wt % and that the powder. The WC(plasma) powder particles were
free carbon content dropped to zero. As ann was raised rounded in shape both before and after annealing: at
from 800 to 1200 , the particle size of the carbide least, SEM examination failed to detect any aggreꢀ
2). Chemical analysis data showed that the total ing at 1400
°C was four times that of the asꢀprepared
t
°C
phase slightly increased (to 80 nm) and the lattice gated particles similar to those resulting from sintering
strain and total carbon content slightly decreased, in the annealed WC(mill) powder. The WC(plasma)
which, however, had little effect on the phase compoꢀ nanoparticles were already highly aggregated before
sition of the powder (79.5 wt % WC and 20.5 wt % W). annealing. Even after annealing at 1400 C—the temꢀ
°
INORGANIC MATERIALS Vol. 48
No. 7
2012

VACUUM ANNEALING OF NANOCRYSTALLINE WC POWDERS
687
D
= 80 nm, ε = 0.18%
WC (65%)
W2C (33%)
W (2%)
4
D
= 80 nm, ε = 0.35%
WС (79.5%)
W (20.5%)
3
D
= 70 nm, ε = 0.38%
WС (79%)
W (21%)
2
D
= 60 nm, ε = 0.43%
WС (100%)
1
30
40
50
, deg
60
70
2
θ
Fig. 6. XRD patterns of the WC(plasma) nanopowder (prepared by plasma synthesis) (
1) before and (2–4) after vacuum annealing
at ( ) 800, ( ) 1200, and ( ) 1400 C.
2
3
4
°
INORGANIC MATERIALS Vol. 48
No. 7 2012

688
KURLOV, GUSEV
the powder with tann (table), which can be understood
in terms of the specific features of the W–C phase diaꢀ
gram [13]. As shown earlier [13, 14], the lower tungꢀ
sten carbide W2C forms at
pattern of the WC(mill) powder annealed at 1200
shows reflections from W2C Annealing the
WC(micro) powder at 1200 also increased the perꢀ
t
≥
1250
°
C, but the XRD
°
C
.
°C
centage of W2C, along with the formation of metallic
tungsten. According to the phase diagram of the W–C
system [13], such changes in phase composition can
only be caused by partial decarburization of the WC
powders, as observed in our experiments.
Even though similar in particle size, the WC(mill)
and WC(plasma) nanopowders differed in some
respects, which showed up in our annealing experiꢀ
1
1
1
μ
m
(а)
(b)
(c)
ments. With increasing tann, the particle size or the
WC(mill) nanopowder increases and its lattice strain
decreases considerably more rapidly than those of the
WC(plasma) powder. The same refers to phase transꢀ
formations. In the WC(mill) powder, W2C was present
even after annealing at 1200
disappeared upon annealing at 1400
WC(plasma) nanopowder contained no W2C after
annealing at 1200 , whereas a small amount of tungꢀ
sten persisted even after annealing at 1400
°C
and metallic tungsten
°
C.
The
°C
°C.
It is known that, during vacuum evaporation, WC
loses predominantly carbon and passes into the W +
WC or W2C + WC twoꢀphase region, depending on the
evaporation temperature [15–18]. The carbon partial
pressure over tungsten carbide is low even at high
μ
m
temperatures (1730–2130
°
C
): 3.7
×
10–6 to 4.2
×
10–3 Pa [15, 17]. The rate of carbon vaporization,
v
,
C
from tungsten carbide at 1983
°
C
is as low as 9.24
×
10–9 g/(cm2 s) [15]. According to Bolgar et al. [16], the
vaporization rate in the temperature range 1975–
2650
°
C
(2248–2923 K) can be represented in the form
. For 1400 , the rate of
carbon vaporization from WC is hundreds of thouꢀ
sands of times lower than at 2000 and cannot
account for even a hundredth of the experimentally
observed carbon loss, CC, during vacuum annealing
log C = 6.317 – 34310/
T
t
≤
°C
v
°C
v
C
μ
m
Δ
of WC. Thus, the observed decrease in carbon content
as a result of the vacuum annealing of WC at
Fig. 7. SEM micrographs of the WC(plasma) nanopowder
vacuumꢀannealed at (a) 800, (b) 1200, and (c) 1400 C.
tann
≤
1400 C cannot be accounted for by direct carꢀ
°
°
bon vaporization.
A more likely cause of the decrease in carbon conꢀ
tent during vacuum annealing of the WC nanopowders
is oxygen impurities. Even the first nanocrystalline
carbide studied, vanadium carbide nanopowder, was
found to contain 5 wt % oxygen [19]. Most of the oxyꢀ
gen was present in the surface oxide film and adsorbed
water. Krasovskii et al. [20] studied plasmaꢀsyntheꢀ
perature at which hard alloys are commonly sinꢀ
tered—no wellꢀdefined coalescence (agglomeration,
association, or fusion) of the nanoparticles was
detected. This points to their poor sinterability, which
may have an adverse effect on the quality of hard
alloys.
Causes of the decarburization of the WC powders. sized WC nanopowder and showed it to contain
The present experimental data demonstrate that vacꢀ 2 wt % oxygen. More than 80% of the oxygen was
Ӎ
uum annealing leads to decarburization of the WC present in Xꢀray amorphous tungsten oxide phases,
powders studied, especially in the case of the WC nanꢀ and the rest, in water and in adsorbed form. Accordꢀ
opowders and tann
≥
800
°C. The decarburization, in ing to Warren et al. [21] and Brillo et al. [22], the surꢀ
turn, leads to variations in the phase composition of face layer of WC in an oxygenꢀcontaining atmoꢀ
INORGANIC MATERIALS Vol. 48
No. 7
2012

VACUUM ANNEALING OF NANOCRYSTALLINE WC POWDERS
689
sphere is similar in electronic structure to WO3. As powders, 1.4 and 1.5 wt %, found after annealing at
shown in a study of the interaction of oxygen with the 1400
(0001) surface of a WC single crystal by Auger elecꢀ
tron spectroscopy [22], oxygen chemisorption under
high vacuum leads to the formation of surface tungꢀ
°
C.
CONCLUSIONS
Vacuum annealing of nanocrystalline WC powders
at tann 1400 is accompanied by a marked decrease
sten oxycarbides, WO, and WO3
.
≤
°
C
According to previous results [21–23], heat treatꢀ
ment of WC nanopowder at temperatures from 900 to
1100 C removes at least 80% of the oxygen in the form
of CO. Clearly, the removal of O in the form of CO
reduces the carbon content of the WC powder as well.
In metal and metal carbide nanopowders, oxygen is
present predominantly on the surface of the particles
[3, 19–24], so the oxygen content is proportional to
in carbon content, changes in phase composition, and
an increase in particle size due to coalescence of
aggregated nanoparticles. In addition, annealing
reduces the lattice strain of such powders. Phase purity
of WC nanopowder can be maintained during annealꢀ
ing by adding an excess of free carbon, but annealing is
then accompanied by significant particle growth.
Vacuum annealing of microcrystalline WC powder
°
the specific surface area
bon loss CC during vacuum annealing is proportional at tann
to the oxygen content of the powder, CO, it should also
be proportional to the initial of the powder.
From the unitꢀcell parameters of the hexagonal
S of the powder. Since the carꢀ
Δ
≤
1400 C leads to a very small decrease in carꢀ
°
bon content, because free carbon disappears, and has
little effect on the phase composition of the powder.
S
6
carbide WC (sp. gr. P m
2
), we find that the surface
ACKNOWLEDGMENTS
We are grateful to V.A. Moldaver† for supplying the
plasmaꢀsynthesized WC powder.
density of W atoms, s, is 1.4
when a WO3 monolayer is formed on WC, the weight
of surface oxygen per square meter of the surface is
n
×
1019 m–2. Therefore,
This work was supported by the Russian Foundaꢀ
tion for Basic Research (grant nos. 10ꢀ03ꢀ00023a and
12ꢀ08ꢀ00016a), the Presidium of the Russian Acadꢀ
emy of Sciences (project no. 12ꢀPꢀ234ꢀ2003: Syntheꢀ
sis and Stabilization of Hybrid Nanoparticles for Variꢀ
ous Applications; program no. 21: Fundamental
Issues in the Development of Nanotechnologies and
Nanomaterials), and the Presidium of the Ural
Branch of the Russian Academy of Sciences (project
no. 11ꢀ3ꢀNPꢀ290).
3
nsmаAO, where ma
unit and AO is the atomic weight of O. If the surface of
particles is covered with monolayers of the oxide
phase, the relative oxygen content of the powder (with
=
1.66
×
10–24 g is the atomic mass
p
a specific surface area
given by
S) is to a first approximation
C
O = 3pnsmаAOS = 0.00111pS
.
(1)
The number of monolayers of the oxide phase is
then
p
=
CO/3nsmаAOS
=
CO/(0.00111S).
(2)
REFERENCES
The specific surface area
S
of the asꢀprepared
1. Klyachko, L.I., Fal’kovskii, V.A., and Khokhlov, A.M.,
Tverdye splavy na osnove karbida vol’frama s tonkodisꢀ
persnoi strukturoi (FineꢀGrained Tungsten Carbide
Based Hard Alloys), Moscow: Ruda i Metally, 1999.
2. Kurlov, A.S. and Rempel, A.A., Effect of WC Nanoparꢀ
ticle Size on the Sintering Temperature, Density, and
WC(mill) and WC(plasma) powders is 19.2 and
6.4 m2/g, respectively (table), and CO is 0.0152 and
0.0180 (1.52 and 1.80 wt%) as determined by EDX
analysis. Therefore, the number of monolayers of the
oxide phase,
p, on the surface of the asꢀprepared
Microhardness of WC–8 wt % Co Alloys, Inorg. Mater.
2009, vol. 45, no. 4, pp. 380–385.
,
WC(mill) and WC(plasma) powders is 0.7 and 2.5
.
This agrees well with the report by Brillo et al. [22] that
the oxide film on WC ranges in thickness from 0.6 to
2.8 monolayers.
Heating causes the oxygen to desorb from WC in
the form of CO [20–22]. Therefore, the relative carꢀ
3. Gusev, A.I., Nanomaterialy, nanostruktury, nanotekhꢀ
nologii (Nanomaterials, Nanostructures, and Nanoꢀ
technologies), Moscow: Fizmatlit, 2007, 2nd ed.
4. Rempel’, A.A., Nanostructured Materials: Nanotechꢀ
nologies, Properties, and Applications, Usp. Khim.
2007, vol. 76, no. 5, pp. 474–500.
,
bon loss is proportional to CO
C = (AC AO O = 3pnsmаACS = 0.00083pS
From (3) and the above values = 0.7 and = 2.5, the
expected carbon loss, CC, of the asꢀprepared
:
ΔC
/
)C
.
(3)
5. Kurlov, A.S. and Gusev, A.I., Model for Milling of
Powders, Tech. Phys., 2011, vol. 56, no. 7, pp. 975–980.
p
p
6. Tsvetkov, Yu.V. and Panfilov, S.A., Nizkotemperaturnaya
plazma v protsessakh vosstanovleniya (LowꢀTemperaꢀ
ture Plasma in Chemical Reduction Processes), Mosꢀ
cow: Nauka, 1980.
7. Kurlov, A.S. and Gusev, A.I., Particle Size Effects on
the Oxidation of Tungsten Carbide Nanopowders,
Δ
WC(mill) and WC(plasma) nanopowders is 0.011 and
0.013 (1.1 and 1.3 wt %), respectively. Given the
uncertainty in the O and C contents and the approxiꢀ
mate character of the above estimates, the CC values
obtained are in reasonable agreement with the maxiꢀ
mum carbon loss of the WC(mill) and WC(plasma)
Δ
†
Deceased.
INORGANIC MATERIALS Vol. 48
No. 7 2012

690
KURLOV, GUSEV
Russ. J. Phys. Chem. A, 2010, vol. 84, no. 12, pp. 2095– 16. Bolgar, A.S., Turchanin, A.G., and Fesenko, V.V., Terꢀ
2101.
modinamicheskie svoistva karbidov (Thermodynamic
Properties of Carbides), Kiev: Naukova Dumka, 1973,
pp. 97–99.
8. Kurlov, A.S. and Gusev, A.I., Effect of Particle Size on
the Oxidation of WC Powders during Heating, Inorg.
Mater., 2011, vol. 47, no. 2, pp. 133–138.
17. Kulikov, I.S., Termodinamika karbidov i nitridov (Therꢀ
modynamics of Carbides and Nitrides), Chelyabinsk:
Metallurgiya, 1988.
18. Svoistva, poluchenie i primenenie tugoplavkikh soedineꢀ
nii. Spravochnik (Properties, Preparation, and Applicaꢀ
tion of Refractory Compounds: A Handbook), Kosoꢀ
lapova, T.Ya., Ed., Moscow: Metallurgiya, 1986.
19. Rempel’, A.A. and Gusev, A.I., Nanostructure and
Atomic Order in Vanadium Carbide, Pis’ma Zh. Eksp.
Teor. Fiz., 1999, vol. 69, no. 6, pp. 436–442.
20. Krasovskii, P.V., Blagoveshchenskii, Yu.V., and Grigorꢀ
ovich, K.V., Determination of Oxygen in W–C–Co
Nanopowders, Inorg. Mater., 2008, vol. 44, no. 9,
pp. 954–959.
21. Warren, A., Nylund, A., and Olefjord, I., Oxidation of
Tungsten and Tungsten Carbide in Dry and Humid
Atmosphere, Int. J. Refract. Met. Hard Mater., 1996,
vol. 14, nos. 5–6, pp. 345–353.
22. Brillo, J., Kuhlenbeck, H., and Freund, H.ꢀJ., Interacꢀ
tion of O2 with WC (0001), Surf. Sci., 1998, vol. 409,
pp. 199–206.
23. Ribeiro, F.H., Dalla Betta, R.A., Guskey, G.J., and
Boudart, M., Preparation and Surface Composition of
Tungsten Carbide Powders with High Specific Surface
Area, Chem. Mater., 1991, vol. 3, no. 5, pp. 805–812.
24. Levinskii, Yu.V., Blagoveshchenskii, Yu.V., Vol’dꢀ
man, G.M., and Mel’nik, Yu.I., A Mechanism of Nioꢀ
bium, Molybdenum, and Tungsten Nanoparticle
Growth during Annealing in Vacuum and Hydrogen,
Materialovedenie, 2006, no. 10, pp. 38–47.
9. Gusev, A.I. and Kurlov, A.S., Production of Nanocrysꢀ
talline Powders by HighꢀEnergy Ball Milling: Model
and Experiment, Nanotechnology, 2008, vol. 19, no. 26,
paper 265 302.
10. Kurlov, A.S. and Gusev, A.I., Production of Nanocrysꢀ
talline Powder of WC via BallꢀMilling, 17th Plansee
Seminar/Int. Conf. on High Performance P/M Materials
Sigl, L.S. et al., Eds., Reutte: Plansee Group, 2009,
vol. 3, pp. GT24/1–GT24/11.
,
11. Kurlov, A.S. and Gusev, A.I., Determination of the Parꢀ
ticle Sizes, Microstresses, and Degree of Inhomogeneꢀ
ity in Nanostructured Materials from XꢀRay Diffracꢀ
tion Data, Glass Phys. Chem., 2007, vol. 33, no. 3,
pp. 276–282.
12. Gusev, A.I. and Kurlov, A.S., Particle (Grain) Size
Characterization of Nanocrystalline Materials, Metꢀ
allofiz. Noveishie Tekhnol., 2008, vol. 30, no. 5,
pp. 679–694.
13. Kurlov, A.S. and Gusev, A.I., Phase Equilibria in the
W–C System and Tungsten Carbides, Usp. Khim.
2006, vol. 75, no. 7, pp. 687–708.
,
14. Kurlov, A.S. and Gusev, A.I., Tungsten Carbides and
W–C Phase Diagram, Inorg. Mater., 2006, vol. 42,
no. 2, pp. 121–127.
15. Hoch, M., Blackburn, P.E., Dingledy, D.P., and
Johnston, H.L., The Heat of Sublimation of Carbon, J.
Phys. Chem., 1955, vol. 59, no. 2, pp. 97–99.
INORGANIC MATERIALS Vol. 48
No. 7
2012