Interactions of water vapor with oxides at elevated temperatures

Article abstract of DOI:10.1016/j.jpcs.2004.06.044

Many volatile metal hydroxides form by reaction of the corresponding metal oxide with water vapor. These reactions are important in a number of high temperature corrosion processes. Experimental methods for studying the thermodynamics of metal hydroxides include: gas leak Knudsen cell mass spectrometry, free jet sampling mass spectrometry, transpiration and hydrogen-oxygen flame studies. The available experimental information is reviewed and the most stable metal hydroxide species are correlated with position in the periodic table. Current studies in our laboratory on the Si-O-H system are discussed.

Full text of DOI:10.1016/j.jpcs.2004.06.044

Journal of Physics and Chemistry of Solids 66 (2005) 471–478
www.elsevier.com/locate/jpcs
Interactions of water vapor with oxides at elevated temperatures
a,
Nathan Jacobson , Dwight Myers , Elizabeth Opila , Evan Copland
b
a
c
*
a
NASA Glenn Research Center, Cleveland, OH 44135, USA
b
East Central University, Ada, OK 74820, USA
c
Case Western Reserve University/NASA Glenn Research Center, Cleveland, OH 44135, USA
Accepted 12 June 2004
Abstract
Many volatile metal hydroxides form by reaction of the corresponding metal oxide with water vapor. These reactions are important in a
number of high temperature corrosion processes. Experimental methods for studying the thermodynamics of metal hydroxides include: gas
leak Knudsen cell mass spectrometry, free jet sampling mass spectrometry, transpiration and hydrogen–oxygen flame studies. The available
experimental information is reviewed and the most stable metal hydroxide species are correlated with position in the periodic table. Current
studies in our laboratory on the Si–O–H system are discussed.
q 2004 Elsevier Ltd. All rights reserved.
1. Introduction—importance of volatile hydroxides
in corrosion
hydrocarbon fuel combustion atmospheres contain w10%
water vapor [3]. For 1 bar total pressure, 0.1 bar water vapor
is formed; for higher pressure combustion processes an even
higher pressure of water vapor is formed. Further, combus-
tion processes in a heat engine typically involve rapidly
flowing gases. Together with reaction (2) this can lead to
substantial removal of material.
A number of elements form volatile hydroxides of the
general formula M(OH) or oxy-hydroxides of the general
n
formula MO (OH) . These form by either of two reaction
p
q
routes [1,2]:
There are numerous examples of metal hydroxide
formation in corrosion processes. Zaplatynsky [4] has
exposed a number of commercial Ni-base alloys to air at
Hydroxidesðs; lÞ Z HydroxideðgÞ
e:g: KOHðsÞ Z KOHðgÞ
(1)
(1a)
1200 8C. Depending on the alloy, he observed volatilization
of tungsten, molybdenum, niobium, manganese, and
chromium from surface oxides. Some of this is due to
volatile oxides; but Krikorian [5] points out that the
presence of moisture in laboratory air could also create
high volatility hydroxides and oxy-hydroxides during high-
temperature exposure.
Oxideðs; lÞ CH OðgÞ
2
Z HydroxideðgÞ or Oxy-hydroxideðgÞ
(2)
(2a)
(2b)
e:g: BeOðsÞ CH OðgÞ Z BeðOHÞ ðgÞ
2
2
A number of authors have studied chromia vaporization
experimentally and provided supporting thermodynamic
calculations. It is well-known that chromia vaporizes in an
oxidative environment [6–8]:
e:g: MoO ðsÞ CH OðgÞ Z MoO ðOHÞ ðgÞ
3
2
2
2
Here we shall focus on reactions of the second type,
which are important in a number of high-temperature
corrosion processes. Many high-temperature corrosion
processes occur in combustion environments. Generally
Cr O ðsÞ C3=2 O ðgÞ Z 2CrO ðgÞ
(3)
2
3
2
3
A plot of the vapor pressure of CrO (g) from Cr O with
3 2 3
*
Corresponding author.
21% O /79% Ar is shown in Fig. 1. The addition of 10%
2
0022-3697/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jpcs.2004.06.044

472
N. Jacobson et al. / Journal of Physics and Chemistry of Solids 66 (2005) 471–478
substantially more grain boundary etching on Al O in an
2
3
Ar/H O environment as compared to a pure Ar environment
2
at 1700 8C. This is attributed to Al(OH) (g) formation. Etori
3
et al. [18] have observed weight losses of Al O and ZrO in
3
2
2
a petroleum gas burner at 1500 8C, 1.8 bar total pressure,
and a gas velocity of 150 m/s. The gas atmosphere
contained 9.5% H O. They suggested the possibility of
2
metal hydroxide formation, although no further evidence
such as a downstream deposit was observed. Recently, Yuri
and Hisamatu [19] have done further studies on Al O
2
3
weight loss in a burner and observed a water vapor pressure
dependence of near 1.5, suggesting the reaction:
1
=2 Al O C3=2 H OðgÞ Z AlðOHÞ ðgÞ
(7)
2
3
2
3
In summary there are a number of high-temperature
Fig. 1. Calculated vapor pressures of dominant species over Cr
from Refs. [34,44,45].
2 3
O . Data is
corrosion situations where volatile hydroxides play a key
role. It is essential to understand the thermodynamics of
these volatile species so we can predict these corrosion
rates.
water vapor enhances volatility even further from the
reaction [8,9]:
Cr O ðsÞ C2H OðgÞ C3=2 O ðgÞ Z 2CrO ðOHÞ ðgÞ
(4)
2
3
2
2
2
2
The vapor pressure of CrO (OH) (g) from Cr O (s)C
2
2
2 3
2. Techniques for studying thermodynamics
of metal hydroxides
2
higher vapor pressure of CrO (OH) (g) than CrO (g).
1% O /10% H O/69% Ar is also shown in Fig. 1. Note the
2 2
2
2
3
Similar oxy-hydroxides from Mo and W are thought to
play a role in the degradation of Mo emitter materials [10]
and W filaments in lamps [2].
Thermodynamic studies of metal hydroxides require
highly oxidizing environments and the most common
method for studies of high-temperature vapors, based on
the Knudsen cell, requires a more reducing environment.
Hence the Knudsen cell technique must be adapted for water
vapor studies. There are several studies in the literature of
hydroxides using gas-leak Knudsen cell mass spectrometry
Boron is commonly proposed as a constituent in high-
temperature materials. Transition metal borides have
extremely high melting points [11], boron is used as an
oxidation inhibiter in carbon [12], and boron nitride (BN) is
used as a fiber coating in continuous fiber reinforced
ceramic matrix composites [13]. However, the oxy-
hydroxide of boron forms readily from boric oxide:
[20–27]. Here a small amount of water vapor is admitted to
a Knudsen cell to react with an oxide. Alternatively H (g) or
2
D (g) may be admitted which reacts with the oxide to form
2
B O ðsÞ CH OðgÞ Z 2BOðOHÞðgÞ
(5)
H O(g) or D O(g) and the metal hydroxides. The volatile
2
2
3
2
2
products are then characterized with the mass spectrometer.
There are pressure limitations on this technique as the
ionizer and detector of the instrument cannot tolerate
The BN fiber coating in composites has been observed to
vaporize [13] in moisture-containing environments at
temperatures as low as 500 8C, very likely by BN oxidation
to B O and subsequent vaporization to BO(OH)(g).
K5
P(H O) greater than w10 bar.
2
2
3
A free jet-sampling mass spectrometer (FJSMS) can
directly sample a 1 bar chemical process in an oxidizing
environment. This type of instrument has been described in
detail elsewhere [28,29] and the principle will be briefly
summarized here. The system consists of a series of
differentially pumped vacuum chambers. In our system,
the reaction occurs in a tube furnace adjacent to a small
orifice in a Pt–Rh sampling cone. The gas species enter the
orifice and undergo a free jet expansion. An abrupt
transition to molecular flow occurs and the molecular
beam is directed to a mass spectrometer. The actual
expansion process is quite complex and dependent on the
mass of the vapor species. We use a quadrupole mass
spectrometer, which further introduces mass discrimination
effects. Hence conversion of ion intensities to absolute
pressures cannot be easily done. We use our instrument only
Silicon-based ceramics, such as silicon carbide (SiC) and
silicon nitride (Si N ), and composites of these are
promising high-temperature materials. These materials
3
4
rely on a thin film of thermally grown silica (SiO ) for
2
corrosion protection. In dry oxygen, this film is remarkably
durable, however, in high-temperature water-vapor contain-
ing environments, this film vaporizes according to:
SiO ðsÞ C2H OðgÞ Z SiðOHÞ ðgÞ
(6)
2
2
4
Oxidation of the SiC or Si N substrate occurs
3
4
concurrently with the volatilization reaction (6) and can
lead to substantial material loss over long periods of time
[
14–16].
There is also evidence of volatile hydroxide formation
with Al O and ZrO . Tai et al. [17] have observed
2
3
2

N. Jacobson et al. / Journal of Physics and Chemistry of Solids 66 (2005) 471–478
473
for qualitative determinations of the amounts of volatile
hydroxides and oxy-hydroxides [30].
with a residual gas analyzer to detect any leaks in the
system.
The most valuable quantitative technique for obtaining
thermodynamic data on hydroxides is the transpiration
method [31]. A carrier gas entrains an equilibrium vapor and
transports it to a low temperature portion of the system
where it condenses. This amount of condensate is accurately
determined by an appropriate analytical technique.
Metal hydroxides have also been studied in hydrogen–
oxygen flames using spectroscopic techniques. The metals
often form mono-hydroxides such as CuOH, GaOH, and
InOH [40,41]. These studies yield data over a wide range of
temperature, but proper identification of spectral lines may
be difficult.
Thermal functions for many metal hydroxides have been
estimated treating the hydroxyl group as a pseudo halide
In many cases the dependence of the amount of vapor
species on water vapor pressure can be used to infer the
chemical formula for the vapor species. From the amount of
condensate and the identity of the vapor species, the vapor
pressure of that species can be calculated. This vapor
pressure as a function of temperature is used to obtain
thermodynamic data. Flow rates are set to avoid surface
reaction kinetic limitations and gas phase diffusion
limitations on the reaction rate. Among the first studies of
hydroxides with this technique are those of Glemser and
colleagues [32–34]. Belton and colleagues [35–37] have
also used transpiration to study transition metal hydroxides.
More recently Hashimoto [38] has used the transpiration
method to study Ca, Si, and Al hydroxides. He used a Pt/Rh
transpiration cell.
[2,23,42–46]. Various correlations have been established
between fluoride or chloride and corresponding hydroxide
bond strengths. Vibrational frequencies and molecular
shapes for hydroxides have also been taken from fluorides
and chlorides.
Ab initio methods of quantum chemistry should yield
more accurate thermal functions for hydroxides [47,48].
Baushlicher et al. [47] have calculated dissociation energies
and shapes of the alkali and alkaline-earth mono-hydroxides.
They conclude that the more ionic hydroxides are linear and
the morecovalent hydroxides are bent. This is consistent with
experimental data [49,50]. Allendorf et al. [48] use ab initio
methods to calculate thermal functions for a variety of Si–O–
H species. Their results will be compared to our experimental
data in a later section.
Our transpiration apparatus [39] is shown schematically
in Fig. 2. We also use a Pt/Rh transpiration cell. We use a
peristaltic pump to inject water into the gas stream. An
argon blanket gas flows in the region between the furnace
tube and the transpiration cell. The blanket gas is monitored
3. Review of thermodynamics of metal hydroxides
There are only a few reviews in the literature on the
thermodynamics of volatile metal hydroxides [1,2,42,
1,52]. The authors of these reviews look for periodic
5
trends in thermodynamic stabilities; however, given the
limited experimental data it is difficult to find these trends.
Some element groups in the periodic table do lead to highly
stable volatile hydroxides and/or oxy-hydroxides.
Table 1 presents available experimental enthalpies
D H8(298)) and entropies (D S8(298)) for a series of
r r
(
hydroxide formation reactions from the condensed phase
oxide, i.e. reaction (2). The condensed phase oxide with the
highest metal oxidation state was selected, e.g. Fe O for Fe.
2
3
Only data for experimentally observed gaseous hydroxide or
oxy-hydroxide species are listed in the table. When the data
appeared to be estimated, they are not listed. Note that the
group IA hydroxides, with the exception of Li(OH)(g) and
Li(OH) (g) [21], form by direct vaporization only (reaction
2
(1)) and this reaction is given.
For all the reactions, a more favorable change in free
energy is attained by a lower D H8(298) and a higher
r
D S8(298). For this approximation, we assume a constant
r
D H8 and D S8 with temperature. Note also that at higher
r
r
temperatures the TD S8 term in free energy becomes more
r
important and the particular reaction will be more
important [2].
Table 1 also gives an estimate of metal–hydroxide group
bond energies at 298.15 K from these experimental data.
Fig. 2. Schematic of our transpiration system.

474
N. Jacobson et al. / Journal of Physics and Chemistry of Solids 66 (2005) 471–478
Table 1
Thermodynamic data on metal hydroxides
o
298
o
298
o
Group
IA
Reaction
D H
r
DrS
D
298
ðMKOHÞ
Geometry of
M–OH bond
Reference
kJ/mol
J/mol K
kJ/mol
Li
2
O(s)CH
2
O(g)Z2Li(OH)(g)
186
228
192
177
157
97
164
157
156
156
433
344
361
361
374
Linear
Linear
Linear
Linear
Linear
JANAF [45]
JANAF [45]
JANAF [45]
IVTAN [46]
JANAF [45]
NaOH(s)ZNaOH(g)
KOH(s)ZKOH(g)
RbOH(s)ZRbOH(g)
CsOH(s)ZCsOH(g)
IIA
BeO(s)C1/2H
BeO(s)CH O(g)ZBe(OH)
MgO(s)C1/2H O(g)ZMg(OH)(g)C1/4O
MgO(s)CH O(g)ZMg(OH) (g)
CaO(s)C1/2H O(g)ZCa(OH)(g)C1/4O
CaO(s)CH O(g)ZCa(OH) (g)
SrO(s)C1/2H O(g)ZSr(OH)(g)C1/4O
SrO(s)CH O(g)ZSr(OH) (g)
BaO(s)C1/2H O(g)ZBa(OH)(g)C1/4O
BaO(s)CH O(g)ZBa(OH) (g)
(s)C1/2H O(g)ZCrOH(g)C1/2O
2
O(g)ZBe(OH)(g)C1/4O
2
(g)
(g)
(g)
(g)
(g)
(g)
(OH) (g)
614
174
557
271
562
266
507
238
443
163
153
31
477
539
351
399
411
433
408
419
444
442
See text
See text
Linear
Linear
Linear
Linear
Linear
Linear
Linear
Linear
JANAF [45]
JANAF [45]
JANAF [45]
JANAF [45]
JANAF [45]
JANAF [45]
JANAF [45]
JANAF [45]
JANAF [45]
JANAF [45]
2
2
(g)
2
2
156
52
2
2
2
2
154
59
2
2
2
2
148
61
2
2
2
2
138
54
2
2
VIB
1/2Cr
2
O
3
2
2
759
214
227
K26
89
380
358
364
460
530
Linear
Bent
Gorokhov et al. [23]
7
81
1
/2Cr
MoO
WO (s)CH
2
O
3
(s)CH
O(g)ZMoO
O(g)ZWO (OH)
O(g)ZMn(OH)(g)C3/4O
2
O(g)C3/4O
(OH)
(g)
2
(g)ZCrO
2
2
61
Bent
Ebbinghaus [44]
JANAF [45]
3
(s)CH
2
2
2
(g)
136
179
3
2
2
2
87
JANAF [45]
VIIB
MnO
MnO
MnO
2
2
2
(s)C1/2H
2
2
(g)
642
254
322
Linear
Hildenbrand and Lau
[
24]
6
64
268
206
300
372
Bent
(s)C1/2H
2
O(g)ZMnO(OH)(g)C1/4O
2
(g)
470
Linear
Hildenbrand and Lau
24]
[
4
70
219
289
373
361
Bent
(s)CH
2
O(g)ZMn(OH)
2
(g)C1/2O
2
(g)
363
Linear
Hildenbrand and Lau
24]
[
4
42
653
69
311
322
Bent
VIII
IB
1/2Fe
2
2
O
3
3
(s)C1/2H
(s)CH O(g)ZFe(OH)
O(g)ZCu(OH)(g)C1/4O
2
O(g)ZFe(OH)(g)C1/2O
(g)C1/4O (g)
(g)
2
(g)
213
229
102
334
318
411
Linear
Bent
Murad [22]
6
1/2Fe
O
2
2
2
324
Bent
Belton and Richard-
son [35] JANAF [45]
CuO(s)C1/2H
2
2
400
145
161
260
230
Linear
Bent
Belyaev et al. [40]
4
29
IIB
ZnO(s)CH
2
O(g)ZZn(OH)
(s)C1/2H O(g)ZBO(OH)(g)
(s)CH O(g)ZB(OH) (g)C1/4O
(s)C3/2 H O(g)ZB(OH) (g)
2
(g)
201
55
300
Bent
Glemser et al. [32]
IIIA
1/2B
2
2
2
O
O
O
3
3
3
2
196
401
118
84
600
557
556
549
566
458
487
JANAF [45]
JANAF [45]
JANAF [45]
IVTAN [45]
JANAF [45]
IVTAN [46]
Hashimoto/IVTAN
1
1
1
1
1
1
/2B
2
2
2
(g)
/2B
2
3
6.4
K15
199
/2Al
/2Al
/2Al
/2Al
2
2
2
2
O
O
O
O
3
3
3
3
(s)C1/2H
(s)C1/2H
(s)CH
(s)C3/2 H
2
O(g)ZAl(OH)(g)C1/2O
O(g)ZAlO(OH)(g)
2
(g)
779
498
572
188
2
134
2
O(g)ZAl(OH)
2
(g)C1/4O
(g)
2
(g)
121
2
O(g)ZAl(OH)
3
K7.3
[
38,46]
Kelly and Padley
41]
1/2Ga
2
O
3
(s)C1/2H
2
O(g)ZGa(OH)(g)C1/2O
2
(g)
550
211
428
Linear
[
5
70
224
215
408
369
Bent
2
1/2In O
3
(s)C1/2H
2
O(g)ZIn(OH)(g)C1/2O
2
(g)
494
Linear
Kelly and Padley
41]
[
5
13
228
190
349
297
Bent
IVA
SiO
SiO
SiO
2
2
2
(s)C1/2H
2
O(g)ZSiO(OH)(g)C1/4O
2
(g)
675
Linear
Hildenbrand and Lau
25,26]
[
7
18
188
62
254
436
Bent
(s)CH
2
O(g)ZSiO(OH)
2
(g)
260
Linear
Hildenbrand and Lau
25,26]
[
3
17
64
408
487
Bent
Bent
(s)C2 H
2
O(g)ZSi(OH)
4
(g)
45
K76
This study

N. Jacobson et al. / Journal of Physics and Chemistry of Solids 66 (2005) 471–478
475
These bond energies were estimated [5] from:
the enthalpy, entropy, and bond energy for Mn(OH) (g),
2
shown in Table 1.
This brief review has several findings:
MOHðgÞ Z MðgÞ COHðgÞ
(8a)
(8b)
(8c)
(8d)
MðOHÞ ðgÞ Z MðgÞ C2OHðgÞ
2
1. The only complete data sets are for groups IA, IIA, VIB,
and IIIA.
MOðOHÞ ðgÞ Z MOðgÞ C2OHðgÞ
2
2
3
4
. With the exception of the first row, the bond energies for
groups IA and IIA are fairly constant.
MO ðOHÞ ðgÞ Z MO ðgÞ C2OHðgÞ
2
2
2
. For group IIIA, the hydroxide/metal bond energy
increases with increasing atomic number.
The bond energies are useful to compare the strengths of
the metal/hydroxide group bond for the various species.
For the data taken from the JANAF [45] and IVTAN [46]
tables, the D H(298), D S(298) and bond energy could be
. Group VIB tends to form highly stable oxy-hydroxides.
Here with increasing atomic number the hydroxide/metal
bond energy decreases.
r
r
5
. There are also several exceptionally stable hydroxides
and oxy-hydroxides. These are Be(OH) , BO(OH),
readily calculated. For the data taken from other sources,
these quantities were calculated from the vapor pressures and
partition functions. Vibrational frequencies, interatomic
distances, and bond angles were taken from the sources cited.
In surveying the literature, we found some controversy in
the shape of the metal–oxygen–hydrogen bond. Consider
first the simple mono-hydroxides. As noted, ab initio
calculations [47] indicate that metal mono-hydroxides
with ionic metal/hydroxide bonds are linear; whereas the
metal/hydroxide bond is bent if it has a larger degree of
covalent character. This ab initio study [47] shows that all
alkali and alkaline-earth mono-hydroxides are linear, except
for BeOH, which has a larger degree of covalency.
2
B(OH) and the group VIB oxy-hydroxides.
2
6
. Theoretical evidence and spectroscopic evidence indi-
cate that the mono-hydroxides with ionic bonding are
linear; whereas the mono-hydroxides with more covalent
bonding are bent. Recalculation of the partition functions
for the transition metal hydroxides assuming a bent
metal/hydroxide bond will yield more accurate thermal
functions.
4
. Experimental study of the Si–O–H system
Recent spectroscopic studies on CuOH and AgOH
indicate a bent structure [50]. It is likely that all the transition
elements form hydroxides with bent M–O–H structures.
Ebbinghaus [44] assumes a bent structure for Cr–O–H in all
the various chromium hydroxides and oxy-hydroxides for
which he estimates thermal functions. In our literature
search, we found some calculations done for linear structures
of transition metal hydroxides and some for bent structures.
For these situations where there is some controversy about
the shape of the molecule, we have done the ‘third law’
calculation for both a linear and bent molecule, using the
spectroscopic data discussed above. The two different
molecular shapes result in two different calculated moments
of inertia, which in turn leads to different rotational partition
functions. Compare the linear and bent molecules in Table 1.
Note that heat of reaction is generally increased about
In the last part of this paper, we discuss our recent
experimental work on the Si–O–H system. We use both
transpiration and free jet expansion mass spectrometry, as
discussed in the experimental section. Transpiration studies
were done over a range of temperatures and pressures.
Thermodynamic quantities were derived from both the
second and third law methods.
There are a number of possible reactions in this system
[
43]:
SiO ðsÞ C1=2 H OðgÞ Z SiOðOHÞðgÞ C1=4 O ðgÞ
(9a)
(9b)
(9c)
(9d)
2
2
2
SiO ðsÞ CH OðgÞ Z SiOðOHÞ ðgÞ
2
2
2
SiO ðsÞ C2H OðgÞ Z SiðOHÞ ðgÞ
2
2
4
2
2
0 kJ/mol, the entropy of reaction is increased about
0 J/mol K and the calculated bond energy is decreased
2
SiO ðsÞ C3H OðgÞ Z Si OðOHÞ ðgÞ
2
2
2
6
about 20 kJ/mol in changing from a linear to a bent molecule.
The possibility of a bent M–O–H bond on the di-
hydroxides is less well-studied. There does not seem to be
experimental data on the shapes of any of these species. In
general if the M–O–H bond is bent in the mono-hydroxide; it
is assumed to be bent in the di-hydroxide. The presence of an
additional hydroxide group suggests an internal rotation—
either free or hindered. Ebbinghaus [44] assumes that
There are two sets of calculated data for these species.
One is based on treatment of the hydroxyl group as a
pseudo-halide [43]; the other is based on ab initio methods
[48]. These results are plotted as function of temperature in
Fig. 3. The vapor pressures of Si(OH) show reasonable
4
agreement; however, there is clear disagreement on the
vapor pressures of SiO(OH) and SiO(OH)2.
As noted, Hashimoto [38] has done a precise transpira-
tion study of the Si–O–H system from 1173 to 1773 K. His
pressure dependent experiments indicate that Si(OH) is
Cr(OH) (g) has free internal rotations and the higher
2
hydroxides have hindered rotations. So now the altered
moments of inertia and internal rotations have a larger impact
on the partition function. Note the large effect on
4
the dominant vapor species. He derives second law
enthalpies and entropies from his measurements. We did

476
N. Jacobson et al. / Journal of Physics and Chemistry of Solids 66 (2005) 471–478
Fig. 4. Pressure dependence of Si–OH formation.
QSi is calculated from the amount of Si in the deposit
collected.
In these transpiration experiments the identity of the vapor
species are best determined by the dependence of pressure of
Si–O–H species on the partial pressure of water vapor. For
Fig. 3. Calculated vapor pressure of Si–OH species over SiO
2
with
x(H O)Z0.37 and P(total)Z1 bar. The lines labeled K were calculated
2
the formation of Si(OH) according to reaction (9c) a plot of
4
from thermodynamic functions taken from Krikorian’s estimates [43] based
on the pseudo halide behavior of the hydroxyl group. The lines labeled A
were calculated from the thermodynamic functions taken from Allendorf’s
et al. ab initio calculations [48]. Note that the vapor pressure of SiO(OH)(g)
from Allendorf’s calculations are too low to appear on this graph.
log P(Si–OH) vs log P(H O) should yield a slope of 2. For
2
the formation of SiO(OH) according to reaction (9b) such a
2
plot should yield a slope of 1. The results are shown in Fig. 4
for 1173, 1373, and 1673 K. The lower temperatures have a
slope close to 2; but the high temperature has a slope of 1.69.
a transpiration and mass spectrometry study to look further
at the identity of the vapor species and also obtain third law
enthalpies, based on the ab initio thermal functions of
Allendorf et al. [48].
Thus at lower temperatures, Si(OH) appears to be the
4
dominant specie; whereas at higher temperatures a second
specie, very likely SiO(OH) , is also important.
2
These results are consistent with studies in our free-jet
expansion mass spectrometer. Hydroxides also behave like
pseudo-halogens in mass spectrometer fragmentation pro-
cesses. Thus a typically observed ion is formed by the
removal of one hydroxyl group from the parent. The major
ions observed correspond to Si(OH) and possibly SiO(OH) .
Our transpiration apparatus has been described earlier in
this paper. The deposits of Si-containing species on the
Pt/Rh collection tube were dissolved in a solution of 4% HF
at 50 8C. The solution was then analyzed with plasma
emission spectroscopy. The lower limit of detection was
about 20 mg of Si. The amount of Si-containing condensate
was converted to vapor pressure by considering the flow rate
of the argon carrier gas, water vapor, and the Si vapor
species [38]. Consider first the molar flow rate of Ar, QAr,
entering the furnace (position 1) before water is introduced:
4
2
Having identified Si(OH) as the major specie to about
4
1373 K, we can then obtain a second law heat and entropy
from the van’t Hoff equation:
ꢀ
1ꢁ
o
o
KDH
DS
PðSiðOHÞ Þ
4
ln K Z
C
with K Z
(13)
P f
1
2
1
R
T
R
½PðH OÞꢀ
QAr
Z
(10)
2
RT1
Here P is the pressure, f is the volume flow rate, R is the gas
constant, and T is the absolute temperature. The volume flow
rate entering the reaction chamber (position 2) is given by
T2
f2 Z R½Q CQ ꢀ
(11)
w
Ar
P2
Here Q is the molar flow rate of water. The volume flow
w
rate leaving the reaction chamber (position 3) is simply
expression (11) with the addition of Q . This is small in
Si
comparison to Q CQ , so we can take f Zf . Finally
w
Ar
3
2
the pressure of Si species leaving the reaction chamber
(
position 3) is given by:
PSi
QSi
QAr CQ CQ
PSi
Q RT3
Si
Z
or
Z
(12)
PT
w
Si
PT
f P
3 3
p 2 2 4
Fig. 5. Plot of log K vs 1/T for SiO (s)C2H O(g)ZSi(OH) (g).

N. Jacobson et al. / Journal of Physics and Chemistry of Solids 66 (2005) 471–478
477
Table 2
the latter is that the ionic mono-hydroxides tend to have
linear M–O–H bonding and the more covalent hydroxides
tend to have bent M–O–H bonding.
2
Enthalpies and entropies for the formation reaction Si(cr)C2O (g)C
2 4
2H (g)ZSi(OH) (g)
Study
Hashimoto—second law
38]
T (K)
D
r
H (kJ/mol)
D
r
S (J/mol K)
Available experimental data on metal hydroxides have
been discussed. From these data, enthalpies and entropies of
formation from water vapor and the most stable oxide are
calculated as well as metal/hydroxide bond energies.
Although experimental data on many hydroxides are
unavailable, some trends can be observed. With the
exception of the first row, groups IA and IIA have fairly
constant metal hydroxide bond energies. For group IIIA, the
metal hydroxide bond energy decreases with increasing
atomic number. There are also several exceptionally stable
hydroxides and oxy-hydroxides. These are Be(OH)2,
1600
K1342.7G2.7
592.5G1.0
[
This study—second law
Allendorf et al. [48]
Allendorf et al. [48]
Krikorian [5])
1200
1200
298
0
K1354G2.7
K1342
K1342
544.1G2.1
539.5
K1348
K1344.2G1.6
This study—third law
298
The results are shown in Fig. 5 and Table 2. Values from
the calculations of Allendorf et al. [48] and experimental
measurements of Hashimoto [38] are shown for comparison
and the agreement is very good.
BO(OH), B(OH) and the group VIB oxy-hydroxides.
2
Studies from our laboratories on the Si–O–H system are
discussed. Transpiration and free-jet sampling mass spec-
trometry are used. It appears that Si(OH) is the dominant
The calculations of Allendorf et al. [48] allow the
derivation of the free energy function for a third law
calculation of D H8(298). His numbers for the enthalpy of
formation at 298.15 K and free energy of formation were
fitted to a polynomial. This was combined with the JANAF
4
r
vapor species to about 1373 K; above that SiO(OH) may be
2
important. A second law enthalpy and entropy and a third law
enthalpy for the reaction of water vapor and SiO to form
2
[45] and IVTAN [46] data for Si, O , H to yield the free
2 2
energy function of Si(OH)4:
Si(OH) are reported. These compare favorably with theo-
4
retical calculations [48] and previous experimental data [38].
FEFðSiðOHÞ4Þ
Z 291:214 K0:056846T Cð6:6727!10ꢁ ÞT
6
2
Acknowledgements
K32124:72=T K89:9145 lnðTÞ
This was used with the standard third law equation to
calculate an enthalpy of reaction for reaction (9c):
(14)
Helpful discussions with Drs L. Gorokhov (Russian
Academy of Sciences), M. Allendorf (Sandia National
Laboratories), M. Zehe (NASA Glenn) are very much
appreciated. Thanks are also due to D. Simon and G. Blank
(both of NASA Glenn) for the design and fabrication of the
transpiration cell.
ꢀ
ꢁ
o
o
G ðTÞ KH ð298Þ
o
o
D H ð298Þ Z DG ðTÞ KTD
r
T
o
ZKRT ln K KTDðFEF ð298ÞÞ
(15)
p
Twenty-nine data points from our transpiration study
were used to calculate an enthalpy and the results are shown
in Table 2. The agreement with Allendorf’s calculations is
excellent.
References
[
1] O. Glemser, H.G. Wendlandt, in: H.J. Emel e´ us, A.G. Sharpe (Eds.),
Advances in Inorganic Chemistry and Radiochemistry vol. 5 (1963),
pp. 215–254.
[
[
[
2] J.W. Hastie, High Temperature Vapors Science and Technology,
Academic Press, New York, 1975. Ch 2 and 5.
5
. Summary and conclusions
3] N.S. Jacobson, Corrosion of silicon-based ceramics in combustion
environments, J. Am. Ceram. Soc. 76 (1993) 3–28.
Volatile metal hydroxides are important in a number
4] I. Zaplatynsky, Volatilization of oxides during oxidation of some
superalloys at 1200 8C, Oxid. Met. 11 (1977) 289–305.
of high-temperature corrosion processes. Examples of
these have been discussed. Thermodynamic data on these
species are limited, in part due to the complexities of
thermodynamic measurements in oxidizing environments.
Gas leak Knudsen cell mass spectrometry, free-jet sampling
[5] O.H. Krikorian, Predictive calculations of volatilities of metals and
oxides in steam-containing environments, High Temp-High Pressure
1
4 (1982) 387–397.
[
6] H.C. Graham, H. Davis, Oxidation/vaporization kinetics of Cr
J. Am. Ceram. Soc. 54 (1971) 89–93.
2 3
O ,
mass spectrometry, transpiration, and H /O flame studies
2
2
[7] C.E. Lowell, W.A. Sanders, Mach 1 Oxidation of Thoriated Nickel
are the commonly used experimental techniques. Theoreti-
cal predictions of thermodynamic quantities for these metal
hydroxides and oxy-hydroxides have been made using the
pseudo-halogen behavior of the hydroxyl group. More
recently, ab initio methods have been applied to obtain
thermodynamic quantities. An important result from
Chromium at 1204 8C (2200 8F), NASA Technical Note, TN D-6562,
1
971.
[
[
8] F.J. Kohl, C.A. Stearns, Vaporization of Chromium Oxides from
Surface of TD-NiCr Under Oxidizing Conditions, NASA Technical
Memorandum, TM X-52879, 1970.
9] G.C. Fryburg, R.A. Miller, F.J. Kohl, C.A. Stearns, Volatile products
in the corrosion of Cr, Mo, Ti, and four superalloys exposed to O
2

478
N. Jacobson et al. / Journal of Physics and Chemistry of Solids 66 (2005) 471–478
¨
containing H
738–1743.
10] G. Wahl, P. Batzies, Evaporation of Molybdenum in Vacuum, in O
and in H O-Atmospheres, in IEEE Conference Record of 1970
2
O and gaseous NaCl, J. Electrochem. Soc. 124 (1977)
[32] O. Glemser, H.G. V o¨ lz, B. Meyer, Gasf o¨ rmge Hydroxyde. II. Uber
gasf o¨ rmiges Zinkhydroxyd, Z. Anorg. Allgem. Chem. 192 (1957)
1
[
2
-,
311–324.
[33] O. Glemser, R.V. Haeseler, U¨ ber gasf o¨ rmiges Hydroxide des
2
Thermionic Conversion Specialist Conference, Miami Beach, FL,
IEEE, New York, pp. 119–21.
Molybd a¨ ns und Wolframs, Z. Anorg. Allgem. Chem. 316 (1962)
1
69–181.
[
[
[
11] R.L. Fleischer, High-temperature, high-strength materials—an over-
view, J. Met. Dec (1985) 16–20.
¨
[
[
[
[
34] O. Glemser, A. M u¨ ller, Uber ein gasf o¨ rmiges Hydroxid des Chroms,
Z. Anorg. Allgem. Chem. 334 (1964) 150–154.
12] J. Buchanan, J.A. Little, Glass sealants for carbon–carbon composites,
J. Mater. Sci. 28 (1993) 2324–2330.
35] G.R. Belton, F.D. Richardson, A volatile iron hydroxide, Trans.
Faraday Soc. 58 (1962) 1562–1572.
13] N.S. Jacobson, G.N. Morscher, D.R. Bryant, R.E. Tressler, High-
temperature oxidation of boron nitride: II, boron nitride layers in
composites, J. Am. Ceram. Soc. 82 (1999) 1473–1482.
14] E.J. Opila, R.E. Hann, Paralinear oxidation of CVD SiC in water
vapor, J. Am. Ceram. Soc. 80 (1997) 197–205.
36] G.R. Belton, R.L. McCarron, The volatilization of tungsten in the
presence of water vapor, J. Phys. Chem. 68 (1964) 1852–1856.
37] Y-W. Kim, G.R. Belton, The thermodynamics of volatilization of
[
[
chromic oxide: part I. The species CrO
1974) 1811–1816.
2
38] A. Hashimoto, The effect of H O gas on volatilities of planet-forming
3 2
and CrO OH, Met. Trans. 5
(
2
15] R.C. Robinson, J.L. Smialek, SiC recession caused by SiO scale
[
volatility under combustion conditions: part I, experimental results
and empirical model, J. Am. Ceram. Soc. 82 (1999) 1817–1825.
16] E.J. Opila, J.L. Smialek, R.C. Robinson, D.S. Fox, N.S. Jacobson, SiC
major elements: I. Experimental determination of thermodynamic
properties of Ca-, Al-, and Si-hydroxide gas molecules and its
application to the solar nebula, Geochim. Cosmochim. Acta 56 (1992)
[
2
recession caused by SiO scale volatility under combustion con-
5
11–532.
ditions: part II, thermodynamics and gaseous-diffusion model, J. Am.
Ceram. Soc. 82 (1999) 1826–1834.
[
[
39] E. Copland, D. Myers, E.J. Opila, N.S. Jacobson, in: M. McNallan,
E. Opila (Eds.), High Temperature Corrosion and Materials Chemistry
III Proceedings of the International Symposium, The Electrochemical
Society, Pennington, NJ, 2001, pp. 253–261.
[
[
17] W.-P. Tai, T. Watanabe, N.S. Jacobson, High-temperature stability of
alumina in argon and argon/water-vapor environments, J. Am. Ceram.
Soc. 82 (1999) 245–248.
40] V.N. Belyaev, N.L. Lebedeva, K.S. Krasnov, L.V. Gurvich, Compu-
tation of copper oxide and hydroxide dissociation energy by the flame
spectrophotometry method, Izvestiya vuzov. Khimiya i khim. tekhnol.
18] Y. Etori, T. Hisamatsu, I. Yuri, Y. Yasutomi, T. Machida, K. Wada,
Oxidation Behavior of Ceramics for Gas Turbines in Combustion Gas
Flow at 1500 8C, International Gas Turbine and Aeroengine Congress
and Exhibition, Orlando, FL, June 2–5, 1997 Paper 97-GT-355.
19] I. Yuri, T. Hisamatsu, Recession Rate Prediction for Ceramic
Materials in Combustion Gas Flow, Proceedings of ASME TURBO
EXPO 2003, June 16–19, Atlanta, GA, 2003 Paper GT2003-38886.
20] D.J. Meschi, W.A. Chupka, J. Berkowitz, Heterogeneous reactions
2
1 (1978) 1698–1700.
[
[
[
[41] R. Kelly, P.J. Padley, Photometric studies in hydrogenCoxygenC
carbon dioxide flames, Trans. Faraday Soc. 67 (1971) 740–749.
[42] D.D. Jackson, Thermodynamics of the Gaseous Hydroxides, UCRL-
51137, Lawrence Livermore Laboratory, 1970.
studied by mass spectrometry. I. Reactions of B
J. Chem. Phys. 33 (1960) 530–533.
2
O
3
(s) with H
2
O(g),
[43] O.H. Krikorian, Thermodynamics of the Silica-Steam System,
Presented at Symposium on Engineering with Nuclear Explosives,
CONF-700101, May 1970.
21] J.A. Berkowitz, D.J. Meschi, W.A. Chupka, Heterogeneous reactions
studied by mass spectrometry. II. Reaction of Li
J. Chem. Phys. 33 (1960) 533–540.
2
O(s) with H
2
O(g),
[44] B.B. Ebbinghaus, Thermodynamics of gas phase chromium species:
the chromium oxides, the chromium oxyhydroxides, and volatility
calculations in waste incineration processes, Combust. Flame 93
[
[
22] E. Murad, Thermochemical properties of gaseous FeO and FeOH,
J. Chem. Phys. 73 (1980) 1381–1385.
(
1993) 119–137.
23] L.N. Gorokhov, J.I. Milushin, A.M. Emelyanov, Knudsen effusion
mass spectrometric determination of metal hydroxide stabilities, High
Temp. Sci. 26 (1990) 395–403.
[
[
45] M.W. Chase, NIST-JANAF Thermochemical Tables, Journal of
Physical and Chemical Reference Data, Monograph No. 9, Institute of
Physics, Woodbury, NY, 1999.
[
[
[
24] D.L. Hildenbrand, K.H. Lau, Thermochemistry of gaseous manganese
oxides and hydroxides, J. Chem. Phys. 1000 (1994) 8377–8380.
25] D.L. Hildenbrand, K.H. Lau, Thermochemistry of gaseous SiO(OH),
46] L.V. Gurvich, V.S. Iorish, D.V. Chekhovskoi, V.S. Yungman,
IVTANTHERMO—A Thermodynamic Database and Software
System for the Personal Computer, NIST Special Database 5, US
Department of Commerce, Gaithersburg, MD, 1993.
SiO(OH)
26] D.L. Hildenbrand, K.H. Lau, Comment on Thermochemistry of
gaseous SiO(OH), SiO(OH) , and SiO [J. Chem. Phys. 101, 6076
994], J. Chem. Phys. 108 (1998) 6535.
27] M. Farber, R.D. Srivastava, M.A. Frisch, S.P. Harris, Mass
Spectrometer Studies of AlCH O Reactions in Effusion Cells and
in Atmospheric H CO Flames, in High Temperature Studies in
2 2
, and SiO , J. Chem. Phys. 101 (1994) 6076–6079.
[
[
[
[
[
47] C.W. Bauschlicher Jr., S.R. Langhoff, H. Partridge, Ab initio study of
the alkali and alkaline-earth monohydroxides, J. Chem. Phys. 84
2
2
1
(
1986) 901–909.
[
[
48] M.D. Allendorf, C.F. Melius, P. Ho, M.R. Zachariah, Theoretical
study of the thermochemistry of molecules in the Si–O–H system,
J. Phys. Chem. 99 (1999) 15284–15293.
2
2
2
Chemistry, Faraday Symposia of the Chemical Society No. 8, 1973,
pp. 121–130.
49] R.L. Kuckowski, D.R. Lide Jr., L.C. Krisher, Microwave spectra of
alkali hydroxides: evidence for linearity of CsOH and KOH, J. Chem.
Phys. 44 (1966) 3131–3132.
28] C.A. Stearns, F.J. Kohl, G.C. Fryburg, R.A. Miller, in: J.W. Hastie
(Ed.), Characterization of High Temperature Vapors and Gases NBS
50] C.J. Whitham, J. Ozeki, S. Saito, Microwave spectroscopic detection
of transition metal hydroxides: CuOH and AgOH, J. Chem. Phys. 110
Special Publication No. 561, National Bureau of Standards,
Washington, DC, 1978, pp. 303–355.
(
1999) 11109–11112.
[
[
[
29] Applications of Free-Jet, Molecular Beam, Mass Spectrometric
Sampling Proceedings, Estes Park, CO, October 11-14, 1994,
NREL-CP_433-7748.
51] G. Elliot, Gaseous Hydrated Oxides, Hydroxides, and other Hydrated
Molecules, UCRL-1831, PhD Thesis, University of California at
Berkeley, 1952.
30] E.J. Opila, D.S. Fox, N.S. Jacobson, Mass spectrometric identification
of Si–O–H(g) species from the reaction of silica with water vapor at
atmospheric pressure, J. Am. Ceram. Soc. 80 (1997) 1009–1012.
31] U. Merten, W.E. Bell, in: J.L. Margrave (Ed.), The Characterization of
High Temperature Vapors, Wiley, New York, 1967, pp. 91–114.
[52] T. Greene, S.P. Randall, J.L. Margrave, Characterization of Unusual
Gaseous Oxide and Hydroxide Molecules in M-O-H Systems at High
Temperatures, in Thermodynamic and Transport Properties of Gases,
Am. Soc. Mech. Eng., 1959, pp. 222–225.