Mass spectrometric identification of Si-O-H(g) species from the reaction of silica with water vapor at atmospheric pressure

Article abstract of DOI:10.1111/j.1151-2916.1997.tb02935.x

A high-pressure sampling mass spectrometer was used to detect the volatile species formed from SiO₂ at temperatures between 1200° and 1400°C in a flowing water vapor/oxygen gas mixture at 1 bar total pressure. The primary vapor species identified was Si(OH)₄. The fragment ion Si(OH)₃⁺ was observed in quantities 3 to 5 times larger than the parent ion Si(OH)₄⁺. The Si(OH)₃⁺ intensity was found to have a small temperature dependence and to increase with the water vapor partial pressure as expected. In addition, SiO(OH)⁺, believed to be a fragment of SiO(OH)₂, was observed. These mass spectral results were compared to the behavior of silicon halides.

Full text of DOI:10.1111/j.1151-2916.1997.tb02935.x

J. Am. Ceram. Soc., 80 [4] 1009–12 (1997)
Mass Spectrometric Identification of Si–O–H(g) Species from the
Reaction of Silica with Water Vapor at Atmospheric Pressure
,
†
Elizabeth J. Opila,* Dennis S. Fox,* and Nathan S. Jacobson*
National Aeronautics and Space Administration, Lewis Research Center, Cleveland, Ohio 44135
The volatile species enter this orifice, undergo a free jet expan-
sion, and form a molecular beam. The beam passes through a
series of differentially pumped vacuum chambers and finally
enters a chamber which has a pressure sufficiently low to oper-
ate a quadrupole mass filter. An ionization energy of 40 eV
was used. This sampling system preserves the chemical and
dynamic integrity of the volatile species, even if they are con-
densable or reactive. Although absolute vapor pressures cannot
be measured with this system, important qualitative trends can
be observed.9–12
The SiO2–H2O reaction took place in a 22 mm ID fused
quartz furnace tube, typically at 1300ЊC. An oxygen stream
with a flow rate of 233 sccm was saturated with water vapor
and subsequently flowed through a disk-shaped fused quartz
frit, chosen to maximize the reactive surface area, placed in the
hot zone of the furnace. Saturation of the oxygen gas with water
vapor was accomplished by bubbling the gas through a bath of
water heated to temperatures between 85Њ and 100ЊC. Incom-
plete saturation of the oxygen was achieved in the bubbler. The
vapor pressure of water at 85Њ to 100ЊC should vary between
0.57 to 1.00 bar13 but was actually found to vary from 0.18 to
A high-pressure sampling mass spectrometer was used to
detect the volatile species formed from SiO2 at temperatures
between 1200؇ and 1400؇C in a flowing water vapor/oxygen
gas mixture at 1 bar total pressure. The primary vapor
species identified was Si(OH)4. The fragment ion Si(OH)3؉
was observed in quantities 3 to 5 times larger than the
؉
parent ion Si(OH)4 . The Si(OH)3؉ intensity was found to
have a small temperature dependence and to increase with
the water vapor partial pressure as expected. In addition,
SiO(OH)؉, believed to be a fragment of SiO(OH)2, was
observed. These mass spectral results were compared to the
behavior of silicon halides.
I. Introduction
T HAS been known for some time that silica will form volatile
Ihydroxides or oxyhydroxides in water-vapor-containing
environments;1–8 however, the identity of the volatile species at
atmospheric pressure has not been determined. The following
three volatilization reactions are generally considered the most
likely to occur:
SiO2(s) ϩ H2O(g) ϭ SiO(OH)2(g)
SiO2(s) ϩ 2H2O(g) ϭ Si(OH)4(g)
2SiO2(s) ϩ 3H2O(g) ϭ Si2O(OH)6(g)
(1)
(2)
(3)
A recent mass spectrometric study8 of SiO2 with water vapor
pressures of 5 ϫ 10Ϫ5 bar and below found evidence of the
gaseous species SiO(OH)2, indicating that reaction (1) is
favored under these conditions. A transpiration study7 at 1 bar
total pressure showed that the amount of Si condensate was
proportional to the square of the water vapor partial pressure,
indicating that reaction (2) is favored. The purpose of this
communication is to report the direct high-pressure mass spec-
trometric observation of Si–O–H species, thereby confirming
the occurrence of reactions (1) and (2) at 1 bar total pressure.
II. Experimental Procedure
The identification of the Si–O–H species was made using
a high-pressure sampling mass spectrometer. This instrument
has been used extensively to identify the volatile species in
atmospheric pressure gas–solid reactions.9–12 A diagram of this
instrument is shown in Fig. 1. It has been described in detail
elsewhere9,12 and will only be briefly summarized here. The
reaction, described in more detail below, occurs in a vertical
tube furnace adjacent to a small orifice in the sampling cone.
A. W. Searcy—contributing editor
Manuscript No. 191424. Received October 29, 1996; approved February 6, 1997.
Supported in part under NASA Cooperative Agreement NCC-3-205.
*
Member, American Ceramic Society.
†Resident Research Associate from Cleveland State University, Cleveland, OH
44115.
Fig. 1. Mass spectrometer sampling system.
1009

1010
Communications of the American Ceramic Society
Vol. 80, No. 4
Fig. 2. Ionization efficiency curve for Si(OH)3ϩ produced from Si(OH)4. An appearance potential of 16.5 Ϯ 0.6 eV was calculated from the linear
regression of the data shown by the solid line.
0.94 bar, as determined by weight gains of a Mg(ClO4)2 desic-
cant in separate calibration runs. However, the degree of water
saturation was adequate for this study.
m/e 78; however, a small peak corresponding to SiO(OH)ϩ at
m/e 61 was observed. We believe this ion is a fragment of
SiO(OH)2ϩ for several reasons. First, the fragmentation pattern
of Si(OH)4 to predominantly Si(OH)3ϩ ions suggests that the
loss of one OHϪ group occurs readily under the ionizing condi-
tions used in this study. Also, Hildenbrand and Lau8 observed
III. Results and Discussion
SiO(OH)2 in their gas inlet Knudsen cell study at 10Ϫ5 bar
ϩ
Mass spectra were observed with the flowing H2O/O2 mix-
ture and compared to spectra obtained in flowing O2. Peaks
observed only in the H2O/O2 mixture were attributed to the
formation of volatile hydroxides of silicon. The most intense
peak was observed at m/e 79 corresponding to the Si(OH)3ϩ
species. A smaller peak was observed at m/e 96 corresponding
to the Si(OH)4ϩ species. The ratio of intensities at m/e 79 to 96
was between 3 and 5. This fragmentation pattern is similar to
water vapor partial pressure with a much lower ionization
energy, which limits fragmentation processes. Finally, the esti-
mated thermodynamic data of Krikorian5 predict that SiO(OH)2
should exist at pressures near the detectability limit of our mass
spectrometer under the conditions of this experiment, whereas
SiO(OH) is predicted to have a much lower vapor pressure.
No evidence of the parent ion Si2O(OH)6ϩ at m/e 174 or the
fragment ion Si2O(OH)5ϩ at m/e 157 was observed within the
limits of the quadrupole mass filter sensitivity at high mass
numbers. Krikorian’s data5 predict this species becomes
important only at water vapor pressures of hundreds of bars.
The ionization efficiency curve for the Si(OH)3ϩ ion is shown
in Fig. 2. The shape of this curve is consistent with an ionization
process involving the production of a single positive ion.16 In
this case the appearance potential will be equal to the sum of
the dissociation energy, the ionization energy, and some finite
amount of excess energy given by the reaction Si(OH)4 ϩ e ϭ
Si(OH)3ϩ ϩ OH ϩ 2e. The value of 16.5 Ϯ 0.6 eV (95%
ϩ
the pattern observed for SiCl4(v), where the SiCl3 fragment
ion is observed in quantities about 3 times the parent ion
ϩ 14
SiCl4 . Similarities between SiCl4 and Si(OH)4 are to be
expected since the hydroxyl group, OH, is often considered a
pseudohalogen.10 We have observed SiClx (x ϭ 1–4), Si2OCl6,
and Si3OCl8 vapor species in studies of the Si–O–Cl system in
our laboratory with this instrument.15 SiCl4 and Si2OCl6 are
the chloride species analogous to the volatile hydroxides in
Eqs. (2) and (3).
Mass scans for the oxyhydroxides in reactions (1) and (3)
were also made. No evidence was found for SiO(OH)2ϩ at
Fig. 3. Variation of Si(OH)3ϩ intensity (filled circles) with water vapor pressure. Theoretical values for the water vapor pressure variation with bath
temperature are also shown (solid line). Actual values of water vapor pressure in the experiments were always less than these values because
complete saturation of the oxygen stream was not achieved.

April 1997
Communications of the American Ceramic Society
1011
Fig. 4. Variation of Si(OH)4 and SiO(OH)2 pressure with temperature and water vapor partial pressure as calculated from the results of Hashimoto7
and Krikorian.5
confidence interval) obtained here falls in the same range as the
value of 12.9 eV obtained for SiCl3ϩ from SiCl4 (Ref. 14) and
16.2 eV obtained for SiF3ϩ from SiF4 (Ref. 17). Krikorian18 has
shown that average metal hydroxide M–OH bond strengths are
typically between the values found for M–Cl and M–F bond
strengths. It is therefore expected that the appearance potential
found for Si(OH)3ϩ from Si(OH)4 would lie between the values
found for SiCl3ϩ and SiF3ϩ. The value obtained here is statisti-
cally the same as that for SiF3ϩ.
Our mass spectrometer measurements are consistent with the
recent transpiration study of Hashimoto7 in which he obtained
precise thermodynamic data for Si(OH)4. His data show a very
small temperature dependence of P(Si(OH)4), as shown in
Fig. 4. In our laboratory, the temperature dependence of the
ϩ
Si(OH)3 intensity was monitored between 1200Њ and 1400ЊC
and found to vary by less than a factor of 2, as expected from
Hashimoto’s data. However, variations in both gas density and
gas velocity with furnace temperature lead to small changes in
ion intensity. This precludes a quantitative measurement of reac-
tion enthalpy. The temperature dependence of the SiO(OH)2
species as predicted from the estimates of Krikorian is also
shown in Fig. 4. The temperature dependence of this species is
expected to be greater than that of Si(OH)4 but was not mea-
sured in the present study.
The partial pressure of Si(OH)4 under these conditions is
predicted to be about 10Ϫ5 to 10Ϫ6 bar from Hashimoto’s data.
The partial pressure of SiO(OH)2 is predicted to be in this
same range using the estimations of Krikorian. Krikorian’s
estimations also predict the vapor pressure of SiO(OH) to be
lower, less than 10Ϫ8 bar. The limit of sensitivity of the instru-
ment used here is about 10Ϫ6 bar.9,12 Although absolute vapor
pressure measurements are not possible with this instrument,
From reaction (2) it can be seen that the pressure of Si(OH)4
should increase with the square of the water vapor partial pres-
sure. The intensity of Si(OH)3ϩ was therefore monitored as the
water bath temperature was cooled from 100Њ to 85ЊC. It can be
ϩ
seen in Fig. 3 that the Si(OH)3 intensity does vary with the
bath temperature. The fully saturated water vapor partial pres-
sure drops off rapidly with bath temperature, as is also shown
on this plot. As mentioned previously, complete saturation of
the oxygen was not achieved. Therefore, water vapor pressures
were always below the values shown on this plot and an exact
dependence of intensity on the square of the water vapor pres-
sure cannot be shown. Nevertheless, it is convincingly shown
ϩ
that the Si(OH)3 intensity depends strongly on the partial
pressure of water vapor.
Fig. 5. Thermogravimetric results19 showing the volatility of silica at 1300ЊC in 0.5 bar H2O/0.5 bar O2 flowing at 4.4 cm/s.

1012
Communications of the American Ceramic Society
Vol. 80, No. 4
Table I. Transport Data Used in Eq. (4) for Calculating Fluxes of Volatile Species
Si(OH)4, 1300ЊC
SiO(OH)2, 1300ЊC
␳Ј (25ЊC), ␳Ј (1300ЊC) (g/cm3)
v (25ЊC) (cm/s)
L (cm)
1.3 ϫ 10Ϫ3, 2.5 ϫ 10Ϫ4
4.4
1.3 ϫ 10Ϫ3, 2.5 ϫ 10Ϫ4
4.4
2.5
2.5
␩ (g/(cmиs))
6.4 ϫ 10Ϫ4
1.9
6.4 ϫ 10Ϫ4
1.8
D (cm2/s)
␳ (g/cm3)
6.7 ϫ 10Ϫ10
22.4
1.2 ϫ 10Ϫ9
22.4
†
Re ϭ ␳ЈvL/␩
Sc ϭ ␩/␳ЈD‡
1.40
1.43
†Re is the dimensionless Reynolds number. ‡Sc is the dimensionless Schmidt number.
Si(OH)3ϩ, Si(OH)4ϩ, and SiO(OH)ϩ were observed only at the
highest sensitivity settings. These results support the presence
of Si(OH)4 and SiO(OH)2.
SiO(OH)ϩ ion was observed, suggesting the SiO(OH)2 species
may also form at atmospheric conditions.
Further support for the existence of Si–O–H volatile species
at atmospheric pressures is obtained from some recent thermo-
gravimetric studies in our laboratories.19 A coupon of fused
quartz was exposed to 0.5 bar of H2O/0.5 bar of O2 flowing at
4.4 cm/s at 1300ЊC. The volatilization of the coupon resulted in
the characteristic linear weight loss shown in Fig. 5. This is
correlated with the expression for the boundary layer limited
diffusion of vapor species:
References
1G. W. Morey and J. M. Hesselgesser, “The Solubility of Quartz and Some
Other Substances in Superheated Steam at High Pressures,” Trans. ASME, 865–
75 (1951).
2G. R. B. Elliot, “Gaseous Hydrated Oxides, Hydroxides, and Other Hydrated
Molecules,” UCRL-1831, University of California, Berkeley, CA, June, 1952.
3E. L. Brady, “Chemical Nature of Silica Carried by Steam,” J. Phys. Chem.,
57, 706–10 (1953).
4O. Glemser and H. G. Wendlandt, “Gaseous Hydroxides”; pp. 215–58 in
Advances in Inorganic Chemistry and Radiochemistry, Vol. 5. Academic Press,
New York, 1963.
5O. H. Krikorian, “Thermodynamics of the Silica–Steam System”; p. 481 in
Symposium on Engineering with Nuclear Explosives (Las Vegas, NV, January
14–16, 1970), Vol. 1, unpublished.
␳ЈvL 1/2
␩
1/3 D␳
J ϭ 0.664
(4)
΂ ΃ ΂ ΃
␩
␳ЈD
L
6M. C. Cheng and I. B. Cutler, “Vaporization of Silica in Steam Atmosphere,”
J. Am. Ceram. Soc., 62 [11–12] 593–96 (1979).
Here, ␳Ј is the concentration of the major gas species, v is the
linear gas velocity, L is the sample length parallel to the flow
direction, ␩ is the gas viscosity, D is the interdiffusion coeffi-
cient for the diffusing gas species in the major gas species, and
␳ is the concentration of the diffusing gas species at the solid–
gas interface. Typical values used to calculate these fluxes are
shown in Table I. The major difference in the calculated fluxes
for Si(OH)4 and SiO(OH)2 arises from the difference in concen-
tration of each species in the boundary layer, which is directly
calculated from the vapor pressure of each species. For
P(Si(OH)4) ϭ 9 ϫ 10Ϫ7 bar (from Hashimoto’s data at 1300ЊC)
the calculated flux is 6 ϫ 10Ϫ3 mg/(cm2иh). For P(SiO(OH)2) ϭ
2 ϫ 10Ϫ6 bar (from Krikorian’s estimate) the calculated flux is
1.1 ϫ 10Ϫ2 mg/(cm2иh). The sum of these fluxes is equal to
1.7 ϫ 10Ϫ2 mg/(cm2иh) which is within about an order of
magnitude of the measured flux, 1.8 ϫ 10Ϫ3 mg/(cm2иh). Given
the numerous estimates of both thermodynamic and transport
data, this is reasonable agreement.
7A. Hashimoto, “The Effect of H2O Gas on Volatilities of Planet-Forming
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, 511–32 (1992).
8D. L. Hildenbrand and K. H. Lau, “Thermochemistry of Gaseous SiO(OH),
SiO(OH)2, and SiO2,” J. Chem. Phys., 101 [7] 6076–79 (1994).
9C. A. Stearns, F. J. Kohl, G. C. Fryburg, and R. A. Miller; pp. 303–55 in NBS
Special Publication, No. 561, Characterization of High Temperature Vapors and
Gases. National Bureau of Standards, Washington, DC, 1978.
10G. C. Fryburg, R. A. Miller, F. J. Kohl, and C. A. Stearns, “Volatile Products
in the Corrosion of Cr, Mo, Ti, and Four Superalloys Exposed to O2 Containing
H2O and Gaseous NaCl,” J. Electrochem. Soc., 124 [11] 1738–43 (1977).
11N. S. Jacobson, M. J. McNallan, and Y. Y. Lee, “Mass Spectrometric Obser-
vations of Metal Oxychlorides Produced by Oxidation–Chlorination Reactions,”
Metall. Trans. A, 17A, 1223–28 (1986).
12N. S. Jacobson, “Application of an Atmospheric Pressure Sampling Mass
Spectrometer to Chlorination Reactions,” NASA TM-87270, 1986.
13R. C. Weast (Ed.), CRC Handbook of Chemistry and Physics, 60th ed.;
p. D-196. CRC Press, Boca Raton, FL, 1979–80.
14J. E. Marra, E. R. Kreidler, N. S. Jacobson, and D. S. Fox, “Reactions of
Silicon-Based Ceramics in Mixed Oxidation Chlorination Environments,” J. Am.
Ceram. Soc., 71 [12] 1067–73 (1988).
15J. E. Marra, E. R. Kreidler, N. S. Jacobson, and D. S. Fox, “Direct Mass
Spectrometric Identification of Silicon Oxychloride Compounds,” J. Electro-
chem. Soc., 135 [6] 1571–74 (1988).
16R. T. Grimley, “Mass Spectrometry”; pp. 195–243 in The Characterization
of High-Temperature Vapors. Edited by J. L. Margrave. Wiley, New York, 1967.
17H. M. Rosenstock, K. Draxl, B. W. Steiner, and J. T. Herron, Journal of
Physical and Chemical Reference Data, Vol. 6, Supplement 1, Energetics of
Gaseous Ions. American Chemical Society and American Institute for Physics,
New York, 1977.
IV. Summary
Si(OH)4 has been found to be the primary volatile species
from the reaction of SiO2 with water vapor at atmospheric
pressure and a temperature of 1300ЊC. The partial pressure of
this species is highly dependent on the water vapor partial
pressure but has a small temperature dependence. The mass
spectra of this species are similar to those of SiCl4, demonstra-
ting the pseudo-halide behavior of Si(OH)4. In addition, the
18O. H. Krikorian, “Predictive Calculations of Volatilities of Metals and
Oxides in Steam-Containing Environments,” High Temp.—High Pressures, 14,
387–97 (1982).
19E. J. Opila and R. E. Hann, “Paralinear Oxidation of CVD SiC in Water
Ⅺ
Vapor,” J. Am. Ceram. Soc., 80 [1] 197–205 (1997).