Detonation synthesis of ultrafine alumina

Article abstract of DOI:10.1007/s10789-005-0153-6

The main steps in the detonation synthesis of ultrafine alumina are analyzed, and the factors determining the key features of each step are identified. The experimentally determined yield of ultrafine particles is interpreted in relation to the process pa

Full text of DOI:10.1007/s10789-005-0153-6

Inorganic Materials, Vol. 41, No. 5, 2005, pp. 468–475. Translated from Neorganicheskie Materialy, Vol. 41, No. 5, 2005, pp. 548–556.
Original Russian Text Copyright © 2005 by Chiganova.
Detonation Synthesis of Ultrafine Alumina
G. A. Chiganova
Department of the Physics of Nanophase Materials, Krasnoyarsk Scientific Center, Siberian Division,
Russian Academy of Sciences, Akademgorodok, Krasnoyarsk, 660036 Russia
e-mail: chiganov@akadem.ru
Received April 15, 2004; in final form, December 1, 2004
Abstract—The main steps in the detonation synthesis of ultrafine alumina are analyzed, and the factors deter-
mining the key features of each step are identified. The experimentally determined yield of ultrafine particles is
interpreted in relation to the process parameters. The effect of synthesis conditions on the size of ultrafineAl O
2
3
particles is examined.
INTRODUCTION
particles showed that they consisted of δ- and θ-Al O
2
3
in the ratio ~7 : 1 [1].
The rapidly growing development of new processes
for the fabrication of high-performance functional
ceramics has stimulated the search for new and optimi-
zation of known techniques for the synthesis of
ultrafine metal oxides. The phase composition, particle
size, and properties of ultrafine oxides depend strongly
on the preparation conditions. Ultrafine alumina pre-
pared using explosion energy is successfully employed
in the fabrication of enhanced-performance composite
materials [1] and high-strength oxide ceramics [2].
Bukaemskii et al. [5] reported the particle size dis-
tribution for one of the synthesized powder samples,
which had three maxima corresponding to ultrafine
spherical particles with a number-average size of
7
0 nm; dense, near-spherical particles ranging in size
from 1 to 50 µm; and a coarser fraction consisting of
cellular aggregates, hollow spherical particles, and their
fragments. The polydispersivity of the powder was
interpreted under the assumption that aluminum com-
bustion may occur in different regimes. The ultrafine
Detonation synthesis and properties of ultrafine alu- oxide resulted from combustion in the vapor phase,
mina have been the subject of extensive studies [1, 3– characteristic of small metal particles. The regular
1
0]. Detonation synthesis is carried out in a hermeti- spherical shape of the particles attests to temperatures
cally sealed chamber filled with oxygen-containing above the melting point of alumina. The spherical
gases (initial pressure p ). The explosive is placed in the shape of the particles and lognormal particle size distri-
0
bution indicate that the particles grow from oxide vapor
by the liquid drop coalescence mechanism. The dense
particles 5–50 µm in size result from partial sintering of
aluminum particles at the shock-wave front, followed
by combustion [5]. The particles larger than 50 µm in
size originate from aluminum melting near the explo-
sive. As a result of explosion, the liquid aluminum layer
breaks into drops whose size correlates with the size of
the large alumina particles [5]. The morphology of
these particles suggests that, during oxidation, the
oxide layer is in the liquid state and expands under the
action of metal vapor to form hollow spheres and cellu-
lar aggregates [6].
center of a cardboard assembly consisting of two coax-
ial cylinders and is surrounded by a layer of PAP-1 alu-
minum powder (flaky particles with average dimen-
sions of 10.6 × 7.2 × 1 µm and volume-average size of
4
.3 µm [5]). Explosion is initiated from the upper face of
the cylinder [3, 6]. According to Beloshapko et al. [3, 4],
the particles of the powder thus prepared are mul-
tiphase: the central part consists of α-Al é , and the
2
3
surface layer consists of δ-Al é and Al
O
N .
2
3
8/3 + x/3 4 – x x
As the particle size decreases to below 100 nm, the
α-Al é content rises sharply. Selective dissolution of
2
3
metastable alumina phases shows, however, that
ultrafine particles consist of metastable phases, while
the larger particles contain the α phase [7]. Taking into
account the large particle size of corundum and its den-
sity, we developed a procedure for classifying the syn-
thesized powders according to their phase composition
and particle size using sedimentation [8]. It was also
found that the nitrogen content of the ultrafine powders,
Some of the synthesized material forms a hard crust
on the chamber wall. Some of the metal drops were
assumed to reach the wall, adhere tenaciously to it, and
then oxidize. From the weight of ultrafine alumina,
Bukaemskii and Beloshapko [6] determined the yield
of ultrafine Al O in terms of the amount of alumina
2
3
determined by the Kjeldahl method (tenths of a per- calculated from the stoichiometry of the oxidation reac-
cent) was far lower than the minimum nitrogen content tion. The experimentally determined dependences of
of the nitrogen-containing phase (2.24 wt %). Detailed the yield on the amount of the explosive (M ) at a con-
e
x-ray diffraction (XRD) examination of the ultrafine stant amount of Al (M ) and on M at constant Me
Al
Al
0
020-1685/05/4105-0468 © 2005 Pleiades Publishing, Inc.

DETONATION SYNTHESIS OF ULTRAFINE ALUMINA
469
were presented in [6, 9] as plots of the yield versus taining explosives, independent of their density and the
M /M . According to Bukaemskii and Beloshapko [6], charge weight; in contrast, in a chamber filled with CO2
e
Al
ultrafine alumina results from the vapor-phase combus- the amount of carbon was larger than that in nitrogen,
tion of the solid layer of aluminum heated by the shock argon, and helium atmospheres. This indicates that CO2
wave to a temperature below its melting point but suffi- has at least a low oxidizing power in the explosion step.
cient for its ignition. This layer consists of isolated and
The far smaller average size of the ultrafine particles
partially sintered particles of the original aluminum
obtained in CO in comparison with air (35 and 75 nm,
2
powder. They determined the range of process parame-
respectively) was attributed by Bukaemskii et al. [12]
ters suitable for the synthesis of ultrafine alumina:
to the fact that the process in CO is multistep. The
2
M /M < 0.2 [6]. Nevertheless, at M = const the yield
e
Al
e
metal first reacts with CO to form oxide particles close
2
passes through a maximum at M /M = 0.2–0.5 [9].
e
Al
in size to those forming in oxygen-containing atmo-
spheres. Next, carbon or CO reduces the alumina to Al
in the vapor phase. The Al vapor is then oxidized, and
alumina particles grow [12]. Given that detonation syn-
thesis is a very fast, far-from-equilibrium process, this
sequence of consecutive steps appears unlikely.
The high yield of ultrafine alumina in the experiments
with M /M ≥ 0.2 was attributed to the improved oxi-
e
Al
dation conditions and vapor-phase combustion of liquid
aluminum near the explosive layer, and the marked
reduction in the yield of ultrafine particles at M /M ≤
e
Al
0
.2 was interpreted under the assumption that there was
The proposed physical models for the formation of
ultrafineAl O [6, 10] do not take into account the pres-
a third, outer aluminum layer heated to below the igni-
tion temperature, from which oxide particles about
2
3
ence of the surface oxide film and the exothermic pro-
cess of aluminum oxidation. Moreover, the existing
models fail to explain some experimental data. In par-
ticular, the amount of ultrafine alumina forming in pure
oxygen is always insignificant. To oxidize aluminum in
air, the initial air pressure in the chamber must ensure a
threefold excess of oxygen [6]. It is not clear why the
yield of ultrafine alumina remains constant at 63–64% as
M /M decreases from 0.2 to 0.05 at constant M [6, 9]
4
µm in size and hollow spheres were formed [6].
Later, Bukaemskii [10] assumed that ultrafine parti-
cles were not formed from the liquid layer. The synthe-
sis of ultrafine alumina is only possible from solid lay-
ers, including the outer layer of aluminum particles
heated to a temperature insufficient for subsequent igni-
tion. Note that, in the studies in question [6, 10], the
ignition temperature of aluminum powder was assumed
to be lower than the melting point of aluminum (933 K)
even though the ignition temperature of aluminum par-
ticles is known to depend on both the particle size and
the state of the surface oxide film [11].
e
Al
Al
even though this decrease in the amount of the explo-
sive (by a factor of 4) leads to a substantial decrease in
the fraction of aluminum in contact with the explosive.
In a number of works [5, 13], ultrafine particles were
Bukaemskii [9] also investigated the effect of the assumed to grow from oxide vapor. Aluminum subox-
M /M ratio on the average size of ultrafine particles. ides are known to be gases under normal conditions.
e
Al
In experiments with constant M , the dependence was Davydov et al. [15], however, refute the assumption
Al
that the oxidation of aluminum powder in positive-oxy-
gen-balance explosive mixtures may lead to the forma-
tion of gaseous suboxides.
rather weak; at constant M , the particle size was found
e
to decrease with increasing M /M . Bukaemskii et al.
e
Al
[
12] reported that the particle size is a linear function of
Al O vapor concentration, which was governed by the
Clearly, the specifics of shock-compression experi-
2
3
ratio of the weight of ultrafine alumina to the volume of ments makes it difficult to gain detailed insight into the
the chamber. The significant deviation of several data mechanism of the synthesis process, which remains a
points was attributed to unfavorable synthesis condi- challenging problem.
tions. Note that the conclusion made in that work that
This issue is also of practical importance, in partic-
ular, in the context of the preparation of Cr-doped
the particle size varies linearly with ultrafineAl O con-
2
3
centration is inconsistent with the liquid drop coales- Al O , a new promising optical material [16]. Advances
2
3
cence mechanism because, in the case of drop coales-
cence, only the weight (volume) of the resulting parti-
cle is a linear function of concentration [13].
in this area of research are highly dependent on knowl-
edge of the mechanism of detonation synthesis.
The purpose of this work was to gain further insight
The powder obtained at M /M ≤ 0.2 in a chamber into the mechanism underlying the detonation synthesis
e
Al
filled with CO consisted predominantly of original of ultrafine alumina.
2
aluminum particles, separate or partially sintered; the
amount of alumina was insignificant as determined by
EXPERIMENTAL
Detonation synthesis products were classified into
XRD [6]. At M /M = 0.41 and 0.75, the powder con-
e
Al
sisted largely of alumina. According to Bukaemskii and
Beloshapko [6], at M /M = 0.41 and 0.75, CO is close size ranges in aqueous suspensions prepared by soni-
e
Al
2
in oxidizing power to air. At the same time, according cating powders in water. To separate particles greater
to Petrov et al. [14], no carbon was detected in air upon than 50 µm in size, the suspension was passed through
the detonation of negative-oxygen-balance carbon-con- a sieve. Next, the suspension was left standing. After
INORGANIC MATERIALS Vol. 41 No. 5 2005

4
70
CHIGANOVA
p, T
Shock compression of aluminum powder. The
variations in pressure and temperature during shock com-
pression are represented schematically in Fig. 1 [17].
In region I, the pressure and temperature at the
–
12
–9
shock-wave front increase in a time τ ≈ 10 to 10 s.
–
6
The compression step lasts ~10 s (region II). In
region III, the pressure is removed by isoentropic
unloading waves (τ ≈ 10 to 10 s), and the tempera-
ture drops (at ~10 K/s) to a residual temperature T '. In
T
–
5
–6
8
region IV, cooling is governed by usual heat conduc-
tion. In the expansion step, the particles fly apart,
together with the detonation products. The density ρ and
pressure p in the shock-compressed layer can be evalu-
ated using a shock adiabat, the mass velocity of the par-
ticles behind the shock-wave front (u), and the wave-
T'
p
I
II
III
IV
p0, T0
front velocity (D): ρ = ρ D/(D – u), p = ρ Du [17].
τ
0
0
The shock adiabat of aluminum powder with a bulk
Fig. 1. Variations in pressure and temperature during shock
compression and unloading; (I–IV) see text.
3
density of 0.45 g/cm used in detonation synthesis was
reported in [18]. The average mass velocity is u ≈
1
.83 km/s, and the wave-front velocity is D ≈ 2.34 km/s.
3
micron-sized particles settled on the bottom, ultrafine Under these conditions, ρ ≈ 2.07 g/cm (below the true
3
particles were separated by decantation.
density of aluminum, 2.7 g/cm ) and p ≈ 1.93 GPa.
The content of unreacted aluminum was determined
It is well known that particles of commercial Al
gravimetrically, from the weight gain during calcina- powders do not fuse together into an ingot even well
tion of the powder at 1270 K to a constant weight (with above the melting point of Al: each melt drop is encap-
allowance made for the weight change during heat sulated in alumina [19].
treatment at 870 K owing to the removal of sorbed
gases, moisture, and the constitutional water). The
The increase in the melting points (T ) of Al and
m
Al O caused by shock compression at 1.93 GPa can be
2
3
fixed-nitrogen content of the ultrafine particles was
determined by a modified Kjeldahl method. To this end,
alumina was dissolved as described elsewhere [7]. To a
dilute (1 : 10) aliquot of the resultant solution was
added, in the cold, an excess of an alkali. The solution
was boiled, and the released ammonia was collected in
a receiver containing dilute sulfuric acid. The nitrogen
content was evaluated from titration data.
evaluated by the Clapeyron–Clausius equation,
dT /dp = T ∆V/∆H ,
m
m
m
where ∆V is the change in molar volume upon fusion
and ∆ H is the molar heat of fusion. The melting point
m
of aluminum is ~1180 K, and that of Al O is ~2950 K.
2
3
The temperature of the detonation products of hexogen,
the explosive most frequently used in the detonation
synthesis of Al O , is ~3000 K [20]. It seems likely that
Control experiments were carried out with ammo-
nium chloride, reagent-grade alumina, and aluminum
nitride. Thermal analysis (DTA + TG) was carried out
in a MOM thermoanalytical system in air. Bulk-density
powder samples (1 g) were heated to 1500 K at a rate of
2
3
the melting of the oxide film on the Al particle surface,
which prevents molten Al from forming a continuous
layer, is only possible in the material in contact with the
explosive. Indeed, in CO atmosphere at M /M ≤ 0.2,
1
5 K/min.
2
e
Al
Our experimental data were analyzed in comparison
the fraction of particles greater than 50 µm in size is
insignificant [6]. This also indicates that there is no sig-
nificant damage to the oxide film on aluminum particles
in the course of shock compression, flying apart of the
products, and subsequent cooling. During shock com-
with the results reported in [6, 9, 12].
RESULTS AND DISCUSSION
The main steps in the preparation of ultrafine alu- pression, aluminum does not oxidize even if oxygen-
mina are
rich explosives are used: according to published data,
the energy release associated with Al oxidation in
explosive–aluminum mixtures begins 2–10 µs after the
passage of the wave front [21]. With allowance made
for the characteristic time scale of the detonation syn-
thesis of alumina, Al oxidation is likely to occur at
residual temperatures. At a relatively low power and a
(
(
(
(
1) shock compression of aluminum powder;
2) oxidation of aluminum particles;
3) formation of primary Al O particles;
2
3
4) formation of ultrafine oxide particles.
Consider the characteristic times and the principal small amount of the explosive, oxidation occurs prima-
factors determining the process in each step.
rily on the surface of isolated Al particles.
INORGANIC MATERIALS Vol. 41 No. 5 2005

DETONATION SYNTHESIS OF ULTRAFINE ALUMINA
471
Oxidation of aluminum particles. The oxidation
of Al powder under nonisothermal conditions was stud-
ied by thermal analysis using bulk-density powder (a
1
2
1
: 10 mixture ofAl powder and ultrafineAl O powder,
2 3
added in order to heat-insulate the Al particles and pre-
vent them from fusing together) [22]. The results are
displayed in Fig. 2. The DTA curve shows an endother-
mic peak due to Al melting (T = 935 K) and exother-
0
mic peak corresponding to Al oxidation (T = 1320 K,
0
Tmax = 1410 K). Under the conditions of this study, Al
oxidation does not reach completion. On the whole, the
DTA curve is similar to that of ASD-4 Al powder with
an average particle size of 10 µm, which is charac-
terized by surface oxidation by a diffusion mechanism
[
22]. At the same time, in contrast to that in ASD-4, the
3
00
600
900
1200
1500
weight gain due to Al oxidation is insignificant below
the melting point. Further heating results in a slight
weight gain, which, however, does not exceed 3.5% up
to 1240 K. Thus, we are led to conclude that, under the
conditions in question, Al melting is accompanied by
only a slight damage to the oxide film (severe damage
T, ä
Fig. 2. (1) TG and (2) DTA curves of a 1 : 10 mixture of Al
powder and ultrafine Al O .
2
3
to the oxide film and exposure of the metal to air are try of the oxidation reaction. Therefore, under condi-
accompanied by a considerable increase in oxidation tions corresponding to the oxidation of separate alumi-
rate [22]). The oxidation rate of the Al powder is deter- num particles, the amount of the explosive must have
mined by the diffusional permeability of the surface no effect on the yield of ultrafine particles, as confirmed
oxide layer.
Under the conditions of detonation synthesis (high
residual temperatures and fast reaction rate), the heat of
Al oxidation increases the temperature of the particles,
resulting in the melting of the growing oxide layer
by experimental data: at constant M and M /M ≤ 0.2,
Al
e
Al
the yield of ultrafine particles is constant at 64% [6, 10],
which is very close to the calculated value (the maxi-
mum yield of the detonation synthesis of alumina).
For spherical particles of molten Al and Al O , one
2
3
(
which facilitates diffusion of the reactants), melt
can estimate the thickness L of the oxide layer formed.
At a conversion α = 0.33–0.34 (and provided that the
released vapor does not expand the liquid shell), we
obtain
vaporization within the oxide shell, and (after the rup-
ture of the oxide layer) release of Al vapor. Taking into
account the temperature of the detonation products
8
(
3000 K), the cooling rate during unloading (10 K/s),
–
5
1/3
1/3
and the unloading time (10 s), we estimate the resid-
ual temperature at about 2000 K. One can evaluate the
L = r (V /2V ) [1 – (1 – α) ],
0
1
2
fraction of aluminum (α) whose oxidation provides where r is the initial radius of the Al particles (evalu-
0
enough heat (∆ H ) to compensate for the heating of the
r
T
ated from the particle volume), and V and V are the
1 2
remaining aluminum (1 – α) to the boiling point (T ), its
b
molar volumes of Al O and Al, respectively (calcula-
2 3
vaporization (∆ H), and the heating and melting of the
v
tions were performed for solid phases in order to com-
oxide layer formed:
pare with the measured particle size). At r = 2.63 µm,
0
3
L is 0.37 µm and the volume of the layer is 35.7 µm .
Tb
After the release of the Al vapor, the liquid shell may
contract to form a drop. After cooling, its radius will be
(
α/2)∆ H = (1 – α) C dT + ∆ H
r
T
p
v
∫
~
2 µm. In the zone of high residual temperatures, such
T
Al
drops may coalesce. The expansion of melt drops under
the action ofAl vapor (or dissolved gas) may lead to the
formation of hollow spheres (and their fragments). As
shown by Bukaemskii et al. [5], the amount of such
spheres is not very large under the optimal synthesis
conditions. Micron-sized particles cool far more slowly
than ultrafine particles and crystallize in the most stable
Tm
+
(α/2) C dT + ∆ H
.
p
m
∫
Al O
3
T
2
In this way, we find α = 0.33. Taking into account
the heat of fusion of aluminum (under the assumption
that the shock compression time is too short for Al to
form, α-Al O . Thus, under the optimal synthesis con-
2
3
melt), we obtain α = 0.34. Accordingly, the maximum ditions, dense Al O particles 5–50 µm in size [5] orig-
2
3
fraction of vaporized Al is 66–67%, which places the inate from oxide shells forming on the surface of the
upper limit on the yield of ultrafine alumina in terms of aluminum particles before the transition to the combus-
the amount of alumina calculated from the stoichiome- tion regime. The largest Al O drops may reach the
2
3
INORGANIC MATERIALS Vol. 41 No. 5 2005

4
72
CHIGANOVA
MUF, g
20
sion. Therefore, to fully oxidize aluminum in a short
time, it is sufficient to ensure a 1.5- to 1.7-fold excess
of oxygen in the chamber. The fact that only insignifi-
cant amounts of ultrafine Al O can be obtained in pure
1
2
3
oxygen [6] is attributable to the rapid oxidation of Al in
the earliest stages of the expansion of the detonation
products, which leads to joining together of the liquid
oxide shells. As a result, the fraction of ultrafine parti-
cles in the oxidation products decreases. Increasing the
gas pressure in the chamber enhances the cooling rate
of liquid alumina, thereby reducing the alumina loss
because the melt drops solidify before reaching the
chamber wall.
9
6
3
0
0
0
The oxygen of the explosive participates in the oxi-
dation of aluminum even in the case of a large negative
oxygen balance [23]. According to our results, the frac-
tion of unoxidized aluminum in shock compression
synthesis in an oxygen-deficient atmosphere decreases
in the order trinitrotoluene > hexogen > PETN >
ammonite 6 GV, which corresponds to an increase in
0
40
80
120
M , g
Al
Fig. 3. Amount of ultrafine alumina powder as a function of
the initial amount of aluminum.
chamber wall without solidifying, forming a high- the oxygen content of the explosive, confirming that the
strength coating.
oxygen of the explosive also oxidizes the aluminum.
Accordingly, in inert atmospheres, increasing the rela-
tive amount of the explosive increases the fraction of
oxidized aluminum, which accounts for the increase in
the alumina content of the detonation products with
increasing M /M in a CO atmosphere [6].
A significant increase in the amount or power of the
explosive leads to an increase in the fraction of Al par-
ticles fused together. The surface oxidation of large par-
ticles leads to the formation of thick oxide layers. As a
result, the reaction rate is more strongly limited by the
diffusion of the reactants. This reduces the amount of
particles whose oxidation passes into the combustion
e
Al
2
Formation of primary Al O particles. Primary
2
3
Al O particles resulting from Al vaporization may be
2
3
regime and, accordingly, the yield of fine oxide parti- due to vapor-phase oxidation, followed either by the
formation and subsequent condensation of aluminum
suboxides or by the heterogeneous oxidation of con-
densed Al clusters. According to Morokhov et al. [24],
ultrafine Al O particles obtained by evaporating Al in
cles and increases the fraction of particles greater than
5
1
0 µm in size. The influence of M at M = const (M =
Al
e
e
2 g, p = 405.3 kPa) was analyzed in [6, 9] in the form
0
2
3
of the dependences of the yield of ultrafine powder on
M /M , which have an extremum. The effect is better
a mixture of an inert gas and oxygen are formed
through the oxidation of liquid aluminum. The results
reported by Il’in and Proskurovskaya [25] indicate that
the oxidation of aluminum in air typically leads to the
formation of a large amount of aluminum nitride: the
fixed-nitrogen content of the resulting powder may
attain 18 wt %. They attribute this high nitrogen content
to the reaction between the aluminum suboxide and
molecular nitrogen, leading to the formation of AlN or
AlON in the gas phase. If the reaction is run in a close
space or the amount of oxygen is limited, the fixed-
nitrogen content increases notably [22]. In ultrafine
Al O powders prepared under various conditions, the
e
Al
illustrated by the plot of the total weight of ultrafine
particles, M , versus M (Fig. 3). Clearly, above a cer-
UF
Al
tain amount of Al (60 g under the conditions of our
experiments) the yield of ultrafine oxide particles
depends on the net heat effect of oxidation. It seems
likely that, at a reduced cooling rate, some of the
ultrafine particles, the largest ones, reach the chamber
wall. On the whole, however, the effect of the exother-
mal events on the yield is rather weak, indicating that
most of the Al pass into the combustion regime, as
would be expected for particles flying apart after explo-
2
3
fixed-nitrogen content does not exceed 1.1 wt %
(
table), which correlates with the level typical of heter-
Fixed-nitrogen content of ultrafine Al O particles prepared
2
3
ogeneous oxidation of aluminum. Heterogeneous oxi-
dation is also characterized by the presence of unre-
acted aluminum [11]. Analysis of ultrafine Al O pow-
by detonation synthesis
Explosive Gas (p = 101.33 kPa) M /M
wt % N
0
e
Al
2
3
ders prepared under various conditions showed that all
of the samples contained unoxidized aluminum. The
minimumAl content was 2.5 wt %. Thermal analysis of
that sample, with an average particle size of 75 nm,
showed an endothermic peak due toAl melting (Fig. 4).
It seems, therefore, likely that, after detonation synthe-
Hexogen
Air
0.54
0.54
0.41
0.47
1.1
0.5
0.3
0.9
Nitrogen
Carbon dioxide
Air
Ammonite
INORGANIC MATERIALS Vol. 41 No. 5 2005

DETONATION SYNTHESIS OF ULTRAFINE ALUMINA
473
sis passes into the combustion regime, the Al vapor
condenses, and then heterogeneous oxidation occurs,
1
yielding liquid Al O , as evidenced by the regular
2
3
spherical shape and smooth surface of the alumina par-
ticles.
The fixed-nitrogen content of alumina powders pre-
pared via heterogeneous oxidation of ultrafine alumi-
num was reported to depend on the reaction rate [26]:
with increasing oxidation rate, the nitrogen content
rises. This finding fits well with the results presented in
the table. In particular, increasing the nitrogen pressure
and reducing the oxygen pressure in the chamber (dis-
placing air by nitrogen) reduces both the oxidation rate
and the nitrogen content of the resulting ultrafine Al O
2
2
3
300
600
900
1200
1500
powder (1.1 and 0.5 wt %, respectively). Reducing the
power of the explosive lowers the temperature in the
oxidation zone and, hence, the reaction rate. Accord-
ingly, the nitrogen content decreases slightly in going
from hexogen to ammonite.
T, ä
Fig. 4. (1) TG and (2) DTA curves of ultrafine Al O
2
3
particles.
Primary alumina particles subsequently coalesce to
form ultrafine particles.
d, nm
Formation of ultrafine oxide particles. Coales-
cence of liquid particles in the vapor phase, resulting in
spherical particles participating in the subsequent pro-
cess, was analyzed in detail by Smirnov [13] based on
Smoluchowski’s rapid coagulation theory. At time τ,
the average weight of particles resulting from coales-
cence is
1
1
60
20
80
m = m N Kτ,
0
0
4
0
0
where m is the average weight of the primary particles,
N is their concentration at time zero, and K is the rate
constant of particle association. Using the Einstein
equation and Stokes law, we obtain
0
0
40
80
120
M , g
Al
Fig. 5. Average particle size of Al O as a function of the
initial amount of Al.
m/m = 4kTN τ/(3η),
2 3
0
0
where k is the Boltzmann constant, T is the absolute
temperature, and η is the viscosity of the medium. The
Experimental data indicate that the amount of hexo-
gen has little effect on the size of the resulting ultrafine
particles [9]. Therefore, the effect of the residual temper-
ature in the coagulation zone is also insignificant. The
average alumina particle size versus M /M data [9] are
concentration of primary particles is M/Vm , where M
0
is the total weight of the substance in the volume under
consideration. Expressing m and m through the corre-
0
sponding radii R and R , we obtain a formula for the
0
e
Al
coalescence time of primary particles forming particles
of radius R and density ρ:
presented in Fig. 5 as the dependence on M . At low
Al
values of M , the particle size increases owing to the
Al
3
increase in the concentration of primary particles in the
τ = πR ηVρ/(kTM).
coagulation zone. The faster particle growth at M >
Al
6
0 g is due to the exothermic processes which raise the
The maximum temperature in the coagulation zone
temperature in the coagulation zone. The data in Fig. 5
provide additional evidence that the amount of alumi-
num influences the yield of ultrafine particles: the solid-
ification of relatively large particles occurs slower, and
the exothermic processes affect alumina deposition on
the chamber wall.
can be estimated at about 1270 K given that δ-Al O is
2
3
the major reaction product (at 1320 K, δ-Al O trans-
2
3
forms into the θ phase [27]). Knowing the weight of the
ultrafine particles, the volume of the chamber, and the
average particle size [12], one can evaluate the coales-
cence time. In air (the average particle diameter is
7
3
5 nm, and the true density of ultrafine Al O is
2 3
The liquid-drop coalescence time of primary parti-
cles is limited by their solidification time. As shown by
3
–3
.7 g/cm ), the coalescence time is about 4.7 × 10 s.
INORGANIC MATERIALS Vol. 41 No. 5 2005

4
74
CHIGANOVA
Petrov et al. [14], the cooling rate of detonation prod-
ucts flying apart in the chamber is proportional to the
heat capacity of the vapor phase, which can be found as
REFERENCES
1
. Bukaemskii, A.A., Tarasova, L.S., and Fedorova, E.N.,
Phase Composition and Stability of Ultrafine Al O Pro-
duced by Detonation Synthesis, Izv. Vyssh. Uchebn.
Zaved., Tsvetn. Metall., 2000, no. 5, pp. 60–63.
2
3
the product of the density, specific heat C , and relative
V
pressure p/p of the gas and the volume of the chamber.
0
Therefore, at a given volume and gas pressure, the ratio
2
. Lefevre, S., Monsot, D., and Fedorova, E., Microwave
Synthesis of Alumina Ceramics, Materialy vserossiiskoi
nauchno-tekhnicheskoi konferentsii po ul’tradispersnym
poroshkam, nanostrukturam i materialam (Proc. All-
Russia Sci. and Technol. Conf. on Ultrafine Powders,
Nanostructures, and Materials, Krasnoyarsk, 2003),
Krasnoyarsk: UPTs KGTU, 2003, pp. 172–174.
of the cooling rates in air and CO is about 0.66.
2
Assuming that this ratio also applies to the coalescence
time of primary particles, we can find the radius of
–3
ultrafine alumina particles formed in 3.1 × 10 s in a
CO atmosphere (taking into account the amount of
2
ultrafine alumina obtained in CO ): R ≈ 20 nm. The
2
3
4
. Beloshapko, A.G., Bukaemskii, A.A., and Staver, A.M.,
Preparation of Ultrafine-Particle Compounds via Shock
Compression of PorousAluminum and the Characteriza-
tion of the Particles, Fiz. Goreniya Vzryva, 1990, vol. 26,
no. 4, pp. 93–98.
experimentally determined value is slightly lower: the
average particle diameter is ~35 nm [12]. The amount
of oxidized aluminum in CO is notably smaller and,
2
therefore, the temperature in the coagulation zone is
lower. Thus, the influence of inert gases on the particle
size of ultrafine alumina is only connected with their
thermophysical properties.
. Beloshapko, A.G., Bukaemskii, A.A., and Staver, A.M.,
Preparation of Ultrafine-Particle Compounds via Shock
Compression of PorousAluminum and the Characteriza-
tion of the Particles, II Vsesoznaya konferentsiya po fiz-
ikokhimii ul’tradispersnykh sistem (II All-Union Conf.
on the Physical Chemistry of Ultradisperse Systems,
The surface characteristics of ultrafine Al O parti-
2
3
cles prepared by detonation synthesis were studied
using macroelectrophoresis and potentiometric titra-
tion [28]. The results differ from those reported earlier
for alumina particles with different phase compositions
and sizes prepared by conventional techniques: the ini-
tial hydroxyl density on the surface of our samples is
lower. The obvious reason is that the particles were
formed from melt drops, which minimized the density
of surface defects acting as adsorption sites.
Jurmala, 1989), Riga: Inst. Neorg, Khim. Akad. Nauk
Latv. SSR, 1989, pp. 202–203.
5
. Bukaemskii, A.A., Beloshapko, A.G., and Puzyr’, A.P.,
Physicochemical Properties of Al O Powder Prepared
2
3
by Detonation Synthesis, Fiz. Goreniya Vzryva, 2000,
vol. 36, no. 5, pp. 119–125.
6
7
. Bukaemskii, A.A. and Beloshapko, A.G., Detonation
Synthesis of Ultrafine Alumina in Oxygen Atmosphere,
Fiz. Goreniya Vzryva, 2001, vol. 37, no. 5, pp. 114–120.
. Chiganova, G.A., Structure and Composition of
Ultrafine Alumina Particles from Chemical Phase Anal-
ysis Data, Zh. Anal. Khim., 1991, vol. 46, no. 7,
pp. 1439–1440.
CONCLUSIONS
The main role of shock compression in the synthesis
of alumina is to heat theAl particles, prevent them from
fusing together, and ensure temperatures high enough
to bring the oxidation reaction to the combustion
regime. To reduce the fusion of the particles in the com-
pression zone (as a result of melting or rupture of the
surface oxide film), the amount and power of the explo-
sive must be restricted.
8
. Beloshapko, A.G., Bukaemskii, A.A., Kuz’min, I.G.,
et al., Ultrafine Alumina Powders Synthesized by Shock
Compression, in Ul’tradispersnye materialy. Poluchenie
i svoistva (Ultradisperse Materials: Preparation and
Properties), Krasnoyarsk, 1990, pp. 92–95.
9. Bukaemskii, A.A., Preparation and Properties of New
Ultradisperse Materials, Extended Abstract of Cand. Sci.
(Phys.–Math.) Dissertation, Krasnoyarsk, 1995.
Under the optimal synthesis conditions, the maxi-
mum yield of ultrafine alumina is governed by the ther-
mal balance of aluminum oxidation, melting of the
residual oxide layer, and heating and vaporization of
the aluminum in the oxide shell.
1
1
1
0. Bukaemskii, A.A., A Physical Model for Detonation
Synthesis of Ultrafine Alumina Particles, Fiz. Goreniya
Vzryva, 2002, vol. 38, no. 3, pp. 121–126.
1. Gurevich, M.A., Ozerov, E.S., and Yurinov, A.A., Effect
of Oxide Film on the Ignition Behavior of Aluminum,
Fiz. Goreniya Vzryva, 1978, vol. 14, no. 4, pp. 50–54.
Ultrafine Al O particles result from heterogeneous
2
3
2. Bukaemskii, A.A., Avramenko, S.S., Vavilov, A.A.,
et al., Shock Synthesis of New Materials from Powders,
Materialy II mezhregional’noi konferentsii po ul’tradis-
persnym poroshkam, nanostrukturam i materialam
(Proc. II Regional Conf. on Ultrafine Powders, Nano-
structures, and Materials, Krasnoyarsk, 1999), Krasno-
yarsk: KGTU, 1999, pp. 35–41.
oxidation of condensing aluminum clusters, followed
by coalescence of the primary oxide particles. The size
of the resulting particles depends on the concentration
of primary Al O particles and the temperature in the
2
3
coagulation zone, i.e., on the amount of aluminum, the
volume of the chamber, the explosive power, and the
thermophysical properties of the gas filling the
chamber.
1
3. Smirnov, B.M., Fractal Clusters, Usp. Fiz. Nauk, 1986,
vol. 149, no. 2, pp. 177–219.
INORGANIC MATERIALS Vol. 41 No. 5 2005

DETONATION SYNTHESIS OF ULTRAFINE ALUMINA
475
1
4. Petrov, E.A., Sakovich, G.V., and Brylyakov, P.M.,
Retention of Diamond in the Course of Detonation Syn-
thesis, Dokl. Akad. Nauk SSSR, 1990, vol. 313, no. 4,
pp. 862–864.
Aluminum and Boron), Tomsk: Tomsk. Gos. Univ.,
2002.
2
2
3. Kostyukov, N.A., Kondensirovannye vzryvchatye
veshchestva (Condensed Explosives), Novosibirsk:
Novosibirsk. Gos. Univ., 1988.
1
5. Davydov, V.Yu., Grishkin, A.M., and Feodoritov, I.I.,
Experimental and Theoretical Studies ofAluminum Oxi-
dation in a Detonation Wave, Fiz. Goreniya Vzryva,
4. Morokhov, I.D., Trusov, L.I., and Lapovok, V.N., Fiz-
icheskie yavleniya v ul’tradispersnykh sredakh (Physical
Phenomena in Ultradisperse Media), Moscow: Ener-
goatomizdat, 1984.
1
992, vol. 28, no. 5, pp. 124–128.
1
1
1
6. Im Tkhek-de, Lyamkina, N.E., Lyamkin, A.I., et al.,
Shock Synthesis of Ultrafine Cr-Doped Al O Particles,
2
3
2
2
5. Il’in, A.P. and Proskurovskaya, L.T., Two-Step Combus-
tion of UltrafineAluminum Powder inAir, Fiz. Goreniya
Vzryva, 1990, vol. 26, no. 2, pp. 71–72.
Pis’ma Zh. Tekh. Fiz., 2001, vol. 27, no. 13, pp. 10–15.
7. Adadurov, G.A., Experimental Study of Dynamic-Com-
pression-Induced Chemical Processes, Usp. Khim.,
6. Shevchenko, V.G., Kononenko, V.I., Latosh, I.N., et al.,
Behavior of Ultrafine Aluminum-Based Powders during
Heating in Air, Rossiiskaya konferentsiya po polu-
cheniyu, svoistvam i primeneniyu energonasyshchen-
1
986, vol. 55, no. 4, pp. 555–578.
8. Beloshapko, A.G. and Bukaemskii, A.A., Shock Adiabat
of Porous Aluminum, in Ul’tradispersnye materialy.
Poluchenie i svoistva (Ultradisperse Materials: Prepara-
tion and Properties), Krasnoyarsk, 1990, pp. 28–32.
nykh ul’tradispersnykh poroshkov metallov
i ikh
soedinenii (Russian Conf. on the Preparation, Properties,
and Application of Energy-Rich Ultrafine Powders of
Metals and Compounds, Tomsk, 1993), Tomsk: NII VN,
1
2
9. Ugai,Ya.A., Neorganicheskaya khimiya (Inorganic Che-
mistry), Moscow: Vysshaya Shkola, 1989.
1
993, p. 106.
0. Voskoboinikov, I.M. and Apin, A.Ya., Temperature Mea-
surements in the Detonation Front of Explosives, Dokl.
Akad. Nauk SSSR, 1960, vol. 130, no. 4, pp. 804–806.
2
2
7. Khimicheskaya entsiklopediya (Chemical Encyclope-
dia), Moscow: Sov. Entsiklopediya, 1988, vol. 1.
2
1. Aniskin, A.I., Detonation of Explosive–Aluminum Mix-
tures, Materialy VIII vsesoyuznogo simpoziuma po
goreniyu i vzryvu “Detonatsiya i udarnye volny” (Proc.
VIII All-Union Symp. on Combustion and Explosion:
Detonation and Shock Waves), Chernogolovka, 1986,
pp. 26–32.
8. Chiganova, G.A. and Nafikova, O.N., Surface Electrical
Properties of Ultrafine Alumina Particles Prepared by
Detonation Synthesis, IV Mezhdunarodnaya konferen-
tsiya po khimii tverdogo tela i sovremennym mikro- i
nanotekhnologiyam (IV Int. Conf. on Solid-State Chem-
istry and Modern Micro- and Nanotechnologies, Kislo-
vodsk, 2004), Kislovodsk: SevKavGTU, 2004,
pp. 57−59.
2
2. Il’in, A.P. and Gromov, A.A., Gorenie alyuminiya i bora
v sverkhtonkom sostoyanii (Combustion of Ultradisperse
INORGANIC MATERIALS Vol. 41 No. 5 2005