Basic Characteristics and Applications of Aerosil

30. The Chemistry and Physics of the Aerosil Surface

J O H A N N M A T H I A S AND G E R H A R D W A N N E M A C H E R
Anwendungstechnik Kieselsauren und Pigmente und Forschung Chemie-Physikalisch der Degussa AG,
Rodenbacher Chaussee 4, D-6450 Hanau 11, West Germany
Received April 10, 1987;accepted September 18, 1987
With the aid of special measuring techniques, it was possible to determine the molar absorptivity e of
the OH-stretching vibration band of the free silanol groups at 3750 cm -~ and the total silanol group
density of hydrophilicAerosil. In most casesthe silanol group density is independent of the BET surface
area (exception: AerosilOX 50); the density of the free silanol groupsdecreases as the BET surfacearea
increases (a deviation from the linear courseoccurswith AerosilOX 50 and Aerosil 380). A temperature
increase up to 600°C results only in the elimination of bridged silanol groups; higher temperatures are
required before the isolated silanol groups are eliminated. Freshly manufactured Aerosil has a high
content of isolated silanol groups; during storage, this content of isolated silanol groups decreases, and
the silanol group density increases. The bonding of dimethylsilyl groups on the surface of Aerosil R 972
is discussed and the temperature stability of the surface groups of Aerosil R 972 and Aerosil R 812 is
demonstrated. © 1988 Academic Press, Inc.

1. INTRODUCTION (1). In order to achieve effective thickening, it

Aerosil is the registered trade mark for a
fumed silica from Degussa A G (Frankfurt,
F R G ) , produced by high-temperature hydro-
lysis. Many application properties of this
product are influenced by the chemical nature
of the surface of the primary particles. For this
reason, it is desirable to obtain further infor-
mation on chemical structures of the Aerosil
surface. Silanol groups exist in both an isolated
state and bridged by hydrogen bonds. They
may be detected by means of IR spectroscopy
based on the stretching vibration bands of the
O H groups at 3750 cm-1 (free silanol groups)
and between 3800 and 2800 cm -1 (bridged
silanol groups). The silanol groups impart the
Aerosil with hydrophilic character. When they
react with silanes such as dimethyl dichloro-
silane or hexamethyl disilazane or are elimi-
nated by thermal treatment, the particles are
rendered hydrophobic.
An important property of Aerosil for ap-
plications is the ability to raise the viscosity of
liquid systems and to render them thixotropic
61
Journal of Colloid and Interface Science, Vol. 125, No. 1, September 1988
is necessary to exceed a m i n i m u m dispersion
energy. This is closely linked with surface pa-
rameters such as the specific surface area, the
silanol group density, degree of hydrophobic-
ity, or reactivity of the silanol groups. Addi-
tional important factors in adjusting the theo-
logical properties of liquids, in addition to the
silica concentration used, are the nature of the
organochemical surface groups and the ra-
tio of the different surface groups to one an-
other (2).
2. EXPERIMENTAL
Materials
Aerosil fumed silicas (Degussa AG, Frank-
furt, F R G ) with the following specific surface
areas (according to D I N 66 131) were used:
Aerosil OX 50 (50 m 2 / g ) , 130 (129 m 2 / g ) ,
150 (155 m 2 / g ) , 200 (196 m 2 / g ) , 300 (303
m 2 / g ) , 380 (372 m 2 / g ) , R 972 (102 m 2 / g ) ,
R 812 (245 m2/g). The physical and chemical
data of these products are published in (1).
0021-9797/88 $3.00
Copyright © 1988 by Academic Press, Inc.
All fights of reproduction in any form reserved.
62 MATHIAS AND WANNEMACHER

IR Spectroscopy mm, and 6 m g / c m 2 in weight. All measure-

ments were performed using the spectrometer
An apparatus, illustrated in Fig. 1, was con-
at a resolution of 2.5 cm -1. The absorbance
structed for the purpose of investigating solid
maximum o f the band denoting the presence
surfaces and adsorption layers. The IR reactor
of free silanol groups increases somewhat if
cells (cells A and B) are the main part of the
the resolution of the equipment is improved
apparatus. The cells contain the sample in the
(by approximately 25% at a resolution of 0.5
form of a self-supporting disk and are posi-
cm-1). Nevertheless, the measurements were
tioned in the IR beam of a spectrophotometer
carried out as mentioned above, because, at a
(Perkin-Elmer 325 ) for measurement.
very high resolution, the signal to noise rela-
Cell A is constructed according to (3) and
tion leads to unfavorable values as a result of
allows in situ investigation of surface reactions
the reduced density of energy of the IR beam
up to a temperature of 550°C. In cell B, ac-
(radiation losses due to cell installation). (The
cording to (4), samples can be treated up to
SiOH group densities given in this paper are
1200°C in a platinum furnace outside the IR
not influenced as a result).
beam. In order to record the spectrum, the
It was observed that the absorbance of the
sample is pulled back into the spectrometer
band of free silanol groups decreases at high
sample compartment with the aid of a chain,
compacting pressures (5). This is due to an
without disturbing the gas atmosphere.
increased probability of hydrogen bridge in-
The gas flow to the reactor cells is controlled
teractions arising from the reduced spaces be-
by mass flow meters. Liquids are transported
tween the particles. Accordingly, a pressure of
after evaporation in a saturator with nitrogen
only 2000 bar was used to form the disks.
as a carrier gas. Quadrupole MS is used for
gas analysis. The vacuum equipment ( P
< 10 -7 mbar), consisting of a rotary pump Lithium Alanate Method
and a turbo pump, is also used for preparation
of the sample or for vacuum supply of the mass The lithium alanate method (6) was mod-
spectrometer. ified as illustrated in Fig. 2. A freshly prepared
The IR investigations were performed on solution of 2% by weight of LiA1H4 in Diglyme
self-supporting disks with a diameter of 13 is added at room temperature. Before mea-

L___xz-~# ',7~ CellA

2 8 = 8 ~ I ,'\11

,/
P1

N2 02 N2

FIG. 1. IR equipment for surface analysis. 1. Pressure gauge. 2. Pressure gauge (high vacuum). 3. Mass
spectrometer. 4. Oil trap. 5. Adsorption trap. 6. Pressure converter. 7. High vacuum valves. 8. Gate valves.
9. Saturator. 10. Capillary tubes. 11. IR windows. P1, P2. Rotary pumps. P3. Turbo pump.
Journal of Colloid and Interface Science, Vol. 125,No. 1, September1988
 AEROSIL SURFACE
LiALH4- solution
FIG. 2. Device for determining silanol group density
with LiA1H4.
surement, the solution is degassed with a sec-
ond vacuum pump. The increased pressure,
as a result of the hydrogen development, is
recorded and the amount of hydrogen can be
calculated by considering the volume of the
added solution, the volume of the vessel, and
the vapor pressure of the solvent.
H-D Exchange
A tube reactor containing Aerosil is supplied
with a nitrogen flow which contains D 2 0 (75%
humidity referred to 20°C). The reaction is
carried out at 100°C until the hydroxyl groups
are completely removed. Later, the sample is
vacuum-dried at the same temperature to re-
move adsorbed D20. In the next step, deute-
rium is quantitatively eliminated by adding
water vapor in the same way. The H 2 0 / H D O /
D 2 0 mixture leaving the reactor is condensed
in a trap with liquid nitrogen and the mass of
the mixture is gravimetrically determined. The
concentration of deuterium atoms in the con-
densate is determined from the infrared gas
spectrum of the evaporated product. The ab-
sorbency of the Q branch of the HDO mole-
cule is used for evaluation (calibration with
mixtures of H 2 0 and D 2 0 ) .
63
Preparation of the Sample
To perform IR and LiA1H4 investigations,
the samples are dried for 1 h at 100°C and at
a pressure of less than 10 -2 mbar to completely
remove all physically adsorbed water. The ex-
istence of remaining water in the sample is
indicated in the IR spectrum by the combi-
nation band at 5200 cm -1.
3. RESULTS AND DISCUSSION
Determining the Total Number
of Silanol Groups
The total number of isolated and bridged
silanol groups can be determined by the re-
action with lithium alanate. The ability of
bridged silanol groups to react with LiA1H4
and simultaneously to release hydrogen may
be demonstrated by means of a simple exper-
iment with hydrophobic silica, for instance
Aerosil R 812. Although free silanol groups
may no longer be detected by IR spectroscopy
(see Fig. 3), the residue of silanol groups can
be determined with LiA1H4, as will be evalu-
ated later on.
It is well known that the surface silanol
groups are accessible for H - D exchange (7).
It is thus possible to determine the SiOH den-
sity via two independent chemical methods.
Furthermore, with H - D exchange it is possible
to control the degree of conversion using IR
spectroscopy (see Fig. 4). After the isolated
silanol groups are completely exchanged, only
1004 tO0
gO
40t ~ 40-
O~ 0
4000 3000 4000 3000
wove number v[cm-1] wave number v [cm-1]
FIG. 3. IR spectra of Aerosil R 972 (left) and Aerosil
R 812 (right).
Journal of Colloid and Interface Science, Vol. 125, No. 1, September 1988
 64 MATHIAS AND WANNEMACHER

TABLE I
~oot lOO

Comparison of Silanol Group Densities Determined

Using the Modified LiA1H4 Method and H - D Exchange
60 fi0 (Fresh Samples)

40 40 LiAIH4 D20
_g
20 2a Sample SiOH (nm2) SD SiOH (nm2) SD

#OOO 3000 4OOO 3000 2000 Aerosil 200 2.16 0.04 2.04 0.1
wovenumberv[cm -1] wovenumberv[cm'1] Aerosil 200 a 1.8 0.03 1.7 0.1
Aerosil 300 2.2 0.04 2.08 0.12
FIG. 4. H-D exchange for Aerosil 200 basic material
(left) and exchangedmaterial (right). a A second, independent sample.

a small number of hardly accessible bridged
silanol groups remain (inner silanol groups).
The accuracy of the two methods is demon-
strated by the conformity of the results (Ta-
ble I).
Figure 5 presents a general view of the sil-
anol group content in different types of Aerosil
as a function of the specific surface area
(BET). The silanol group density is virtually
independent of the specific surface area and
amounts of about 2.5 S i O H / n m 2 are obtained
L (ln Io/I = ~cd). The product cd is in accor-
I
1.5t
with sufficiently aged samples (see "Aging of
Aerosil"). An exception here is OX 50, which
has a somewhat diminished SiOH density.
This is caused by the modified method of pro-
duction, where higher flame temperatures are
applied in comparison with other Aerosil
grades.
Determination of the Molar Absorptivity
of the OH-Stretching Vibration Band at
3750 cm -j (Free Silanol Groups)
The transmittance I/Io of the sample in the
beam can be described by Lambert-Beer's law
dance with the density of the silanol groups in
the IR beam (mole SiOH / cm 2). In order to
3 determine ~, this value must be measured re-
liably. For this purpose, a reaction of hexa-
methyl disilazane ( H M D S ) with silanol groups
was performed according to Eq. [1].
2
~Si--O--H + ½(CHs)sSi--NH
- - S i ( C H 3 ) 3 "~ ~ S i - - O - - Si(CH3)3

0,5 -1
+ ½NH3 [11
The conversion was carried out in cell A at a
temperature of 100°C. HMDS was added with
N2 as a carrier. The conversion was controlled
~o zoo 3oo 400
by means of IR spectroscopy. As pointed out
by Hertl and Hair (8), a second-order kinetic
specific surfoce[m2/g] was obtained for this reaction. Even at a tem-
perature of 200°C and under conditions cho-
FIG. 5. Total number of silanol groups (LiAIH4 sen by the authors (8), no conversion of
method). bridged silanol groups was observed.
Journal of Colloid and Interface Science, Vol. 125, No, 1, September 1988
 AEROSIL SURFACE 65

It is easy to control the reaction under our

conditions in such a way that the less acid
bridged silanol groups (9) do not participate
in the conversion, so that the IR band of these £180
free sil0n0l group density
groups remains constant. After final degassing
(100°C) every free silanol group in the starting
product is replaced by a Me3 Si group. Later
on, the CH3 groups are converted with oxygen
(cell B, 800°C) to water and CO2. The amount

liI
of CO2 is detected by "on-line" measurements
with the mass spectrometer. Within this pro-
cedure the conversion to CO2 occurs quanti- 80 =:
tatively, because no CO remains and, further-
60 ~0
more, formation of carbide can be excluded,
because oxidation with ozone instead of ox-
ygen at 300°C leads to the same results.
Figure 6 illustrates the spectrum of the orig-
inal material with the sharp peak of isolated
silanol groups at 3750 cm -1 and the broad 0 I00 200 30o 40o
absorption band of bridged SiOH groups. The specific surfoce[m21g ]
spectrum on the right shows the hydrophobic
product. The free silanol groups are completely
FIG. 7. N u m b e r of silanol groups as a function of the
converted, and the remaining CH3 groups specific surface area.
cause the CH-stretching vibration band at
2970 cm -1. The number of bridged silanol
groups did not change. The absorbance was 9%. Korn et al. (10) found a molar absorptivity
determined from the IR spectrum by drawing of 1.79 × 105 c m 2 / m o l e with Aerosil 200 sus-
a baseline using the transmittance at 3760 and pended in carbon tetrachloride (pretreatment:
3700 cm -1 as reference points (Fig. 6). I and 300°C for 3 h, p = 0.1 mbar). It is assumed
I0 are evaluated at the wavenumber of the ab- that a reason for this deviation is a peak
sorbance maximum. On the basis of Lambert- broadening due to the interaction of the hy-
Beer's law, a value of~ = 4.4 × 105 c m 2 / m o l e droxyl groups with the solvent, which causes
was calculated with a standard deviation of a lower value for e. The absorptivity of the CH
band at 2970 cm -1 is calculated in the same
manner: ~ = 8.6 × 10 4 cm2/mole.

~oot 100. Density of Free Silanol Groups

The density of isolated SiOH groups can be
determined by IR spectroscopy with a given
60-
molar absorptivity value. The correlation be-
Z~ 4o tween the density and the specific surface area
is illustrated in Fig. 7 (samples stored for 1
~ 20
year). The density of isolated silanol groups
o~ D
(SiOH / n m 2) diminishes with increasing spe-
4000 3obo 4000 3000 cific surface area, whereby the deviation from
wove oumer v [cm-1] - wove numberv[cm -t]
the linear course toward higher density is
FIG. 6. IR spectrum of Aerosi1300 (left) and after treat- caused mainly by Aerosil OX 50 and toward
m e n t with H M D S (fight). lower density by Aerosil 380.
Journal of Colloid and Interface Science, Vol. 125, No. 1, September 1988
66 MATHIAS A N D W A N N E M A C H E R

This also causes the slope of the concentra-

tion curve representing the free silanol groups
(in mmole/g) to decrease. The low content of
isolated silanol groups in the case of Aerosil
380 is the result of a special rough surface 60
(primary particle size equivalent to that of
Aerosil 300) where many opportunities for
obsorbonce ol 4500cm-I
hydrogen bonds exist. Generally, the decrease
in the density of isolated silanol groups with 40
decreasing particle diameter is justified by the
higher probability of finding a suitable neigh-
bor group to form a hydrogen bond. The ab-
sorbance of the band of bridged SiOH groups
is also presented in Fig. 7 (sample weight 6
mg/cm 2). For calculation purposes, a second
baseline was drawn between the transmittance
at 3800 and 2800 c m -1 (see Fig. 6) and in-
ZOO 400 600 800 I000
tegration was carried out over the shaded re-
temperalure [o C}
gion after transformation from transmittance
to absorbance. The integral absorbance ob-
tained is a measure for the concentration of FIG. 8. Elimination of silanol groups by thermal treat-
ment.
bridged silanol groups. Nevertheless, direct
conclusions to the concentration are impos-
sible, because the wavelength dependency of Aging of Aerosil
must be considered.
When the IR spectra of freshly manufac-
Influence of Temperature tured Aerosil and Aerosil that has been stored
under dry conditions are compared, the stored
Aerosil 200 was heated at well-defined tem-
sample is found to show a decrease in the
peratures in vacuo (cell B) for 1 h and the
number of isolated silanol groups and an in-
resulting IR spectrum was recorded. The re-
crease in the concentration of bridged groups
suits of the investigations are illustrated in
(see Fig. 9). This alteration is attributed to a
Fig. 8.
cleavage of strained siloxane groups by ab-
The number of bridged silanol groups de-
sorbed water.
creases as the temperature increases, due to
the reaction to water and siloxane groups. This ~---Si--O--Si~ + H - - O - - H --~
process is completed at about 600°C. At the
2 ~-SiOH [2]
same time, the concentration of isolated sil-
anol groups increases with the temperature Table II shows a comparison between the sil-
until a maximum of about 1.8 SiOH/nm 2 is anol group densities of freshly manufactured
attained at 600°C. Further thermal treatment Aerosil 200 and the same product which has
causes elimination of the free silanol groups. been stored for a year. Table II points out that
Additionally, the integrated absorbance of it is not valid to make statements on the silanol
the combination band of SiOH at 4500 cm -1 group density without taking into considera-
is illustrated in Fig. 8 (sample weight 25 mg/ tion the aging of the sample. This is because
cm 2 due to the low extinction coefficient). It the amount of physically adsorbed water is suf-
may be concluded from the temperature de- ficient to change this parameter during storage,
pendence that both free and bridged silanol immediately after the production process has
groups are indicated by this absorption band. been concluded.
Journal of Colloid and Interface Science, Vol. 125,No. 1, September 1988
 AEROSIL SURFACE 67

The higher density of the total number of TABLE II

silanol groups is attained by the reaction of Silanol Group Density of Freshly Manufactured and
2.5 mg water with 1 g Aerosil 200 (specific Stored Aerosi1200 (Size of Sample: 1 Sack (Original Pack-
surface area: 200 m 2/g); this is equivalent to aging); Storage: Standard Warehouse)
a drying loss of only 0.25%. It is evident that Stored for
this low amount of humidity in the form of 3 days old I year Test method
physically adsorbed water is still contained in
the Aerosil just after production. The decrease Isolated SiOH/nm 2 1.5 1.15 IR
Total SiOH/nm 2 1.8 2.65 LiA1H4
in the isolated silanol group density is obtained
by higher probability for hydrogen bridge
linkage with increasing silanol group density.
Silanol group densities of about 2.5 S i O H /
area of 130 m=/g and has dimethylsilyl groups
nm 2 were determined with dry samples of
on the surface, whereas trimethylsilyl groups
Aerosil which had been stored for approxi-
are fixed on the surface of Aerosil R 812 (basic
mately 1 year. This value is comparable with
silica 300 m 2 / g ) . The IR spectra of the two
the values published in (6), though one should
silicas are compared in Fig. 3. (The differing
bear in mind that there is some "uncertainty"
band intensity of the samples, which are of
due to the age of the samples.
equal weight, arises from the products having
The relevance of this process to Aerosil ap-
different specific surface areas). Aerosil R 972
plications is difficult to predict because of ad-
contains a small residue of isolated silanol
ditional parameters such as the chemical
groups (about 0.1 SiOH / nm 2 ) , whereas that
structure of the liquid components, the type
sort of SiOH group is not detected in the case
and intensity of the dispersion, the order in
ofAerosil R 812. Total silanol group densities
which the components of a formulation are
(Aerosil R 972:0.53 S i O H / n m 2, Aerosil R
added, and the specific surface area, tapped
812:0.54 S i O H / n m 2 related to the specific
density, etc., of the silica.
surface area of the basic silicas) were deter-
mined using the LiA1H4 method.
Hydrophobic A erosil
With the help of the absorbance of the CH
Aerosil R 972 and Aerosil R 812 are hy- bond at 2970 cm -t of Aerosil R 972 and
drophobic silicas which differ both in the spe- Aerosil R 812, the kind of bonding of the di-
cific surface area of the basic silica and in the methylsilyl group with the Aerosil surface can
nature of the surface groups. Aerosil R 972 is be determined assuming that the silanol group
manufactured from silica having BET surface densities of the basis silica are identical. There

+U
lO~|

60
40
I°o
~
100
80---'k~

40-
0
I o l

~ -s,-o-s,-~
0
• I

C HS
0

C H3
0
l_-

z01 g 2o ell% f% "s;-o, "s;-o.

s; cHEI c%" I
oi 0 01 ~0 0 0
4000 3[]00 4000 300o
wove number v[cm -1] wovenumDervtcm-1]
FIG. 9. IR spectra of Aerosil 200; fresh sample (left) FIG. 10. Bonding possibilities of dimethyldichlorosilane
and stored for 1 year (right). (a-c) and HMDS (d) with the Aerosil surface.
Journal of Colloid and Interface Science, Vol. 125, No. 1, September 1988
68 MATHIAS AND WANNEMACHER
R B12
o R 97Z
W
~oo zoo 3~o 4bo 5DO 600 7bO
Temperolure[oC]
FIG. 1I. Thermal stabilityof silyl groups on the surface
of Aerosil R 972/Aerosil R 812.
are three different bonding possibilities for a
dimethylsilyl group (see Fig. 10a-10c).
Based on the absorbance value normalized
with respect to a surface area unit, a ratio of
0.7 is obtained (absorbance R 972:absorbance
R 812), which corresponds to a methyl group
ratio of 4:6, as results by comparison of Figs.
10a and 10d ( R 812). Bonding according to
Fig. 10c can be excluded based on the methyl
group concentration determined and the res-
idue of silanol groups measured. This result is
in accordance with (1 1).
Due to the determined residue of silanol
groups on Aerosil R 972 and Aerosil R 812
(mainly bridged SiOH groups), one can sup-
pose that, during the continuous production
of these products and also during the conver-
sion in the IR cell, only the isolated silanol
Journal of Colloid and Interface Science, Vol. 125, No. 1, September 1988
groups participate in the reaction. The silyl
groups are strongly chemically fixed at the
surface. This is confirmed by the stability of
the methyl groups at high temperature in air,
which is illustrated in Fig. 11. The concentra-
tion of methyl groups remains constant under
these conditions up to a temperature of 400°C,
where decomposition occurs.
Despite the density of the remaining silanol
groups of the two silicas being the same, Aero-
sil R 812 displays a more pronounced hydro-
phobic effect in applied engineering practice.
This m a y be caused by the m u c h higher
shielding effect of the trimethylsilyl groups and
(with respect to the differences in the specific
surface area) the higher methyl group content
of 1 g Aerosil R 812 compared to Aerosil
R 972.
ACKNOWLEDGMENT
We thank Mr. G. Kunz for carrying out the IR mea-
surements.
REFERENCES
1. "TechnicalBulletin PigmentsNo. 23," 3rd ed. Degussa
AG, Frankfurt/M., 1987.
2. "TechnicalBulletinPigmentsNo. 27," 4th ed. Degussa
AG, Frankfurt/M., 1986.
3. Knoezinger, H., Stolz, H., Clement, G., and Meye,
W., Chem. Ing. Tech. Z. 42, 548 (1979).
4. Gallei, E., and Schadow, E., Rev. Sci. Instrum. 45,
1504 (1974).
5. Ryason, P. R., and Russel, B. G., J. Phys. Chem. 79,
1276 (1975).
6. "Technical Bulletin Pigments No. 11." DegussaAG,
Frankfurt/M, 1980.
7. Davidov,V. Ya., Kiselev,A. V., and Zhuavlev,L. T.,
Trans. Faraday Soc. 60, 2254 (1964).
8. Hertl, W., and Hair; M. L., J. Phys. Chem. 75, 2181
(1971).
9. Fowkes,F. M., Tribol. Set. 119 (1982).
10. Korn, M., Killmann, E., and Eisenlauer, J., J. Colloid
Interface Sci. 76, 7 (1980).
11. "Technical Bulletin Pigments No. 6." Degussa AG,
Frankfurt/M, 1980.