Impact of network topology on cationic diffusion and hardness of borate glass surfaces

Morten M. Smedskjaer, John C. Mauro, Sabyasachi Sen, Joachim Deubener, and Yuanzheng Yue
Citation: The Journal of Chemical Physics 133, 154509 (2010); doi: 10.1063/1.3497036
View online: http://dx.doi.org/10.1063/1.3497036
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/133/15?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Ionic diffusion and the topological origin of fragility in silicate glasses
J. Chem. Phys. 131, 244514 (2009); 10.1063/1.3276285
Magnetotransport properties of a percolating network of magnetite crystals embedded in a glass-ceramic matrix
J. Appl. Phys. 105, 083911 (2009); 10.1063/1.3110202
Mssbauer Effect Study of Bi2O3. Na2O. B2O3. Fe2O3 Glass System
AIP Conf. Proc. 765, 294 (2005); 10.1063/1.1923672
Transient nucleation in oxide glasses: The effect of interface dynamics and subcritical cluster population
J. Chem. Phys. 111, 737 (1999); 10.1063/1.479353
Photostimulated luminescence of Ce 3+ -doped alkali borate glasses
Appl. Phys. Lett. 71, 43 (1997); 10.1063/1.119463

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
141.214.17.222 On: Sun, 08 Mar 2015 11:24:36

THE JOURNAL OF CHEMICAL PHYSICS 133, 154509 2010

Impact of network topology on cationic diffusion and hardness

of borate glass surfaces
Morten M. Smedskjaer,1 John C. Mauro,2 Sabyasachi Sen,3 Joachim Deubener,4 and
Yuanzheng Yue1,5,a
1

Section of Chemistry, Aalborg University, DK-9000 Aalborg, Denmark

Science and Technology Division, Corning Incorporated, Corning, New York 14831, USA
3
Department of Chemical Engineering and Materials Science, University of California-Davis, Davis,
California 95616, USA
4
Institute of Non-Metallic Materials, Clausthal University of Technology, D-38678 Clausthal-Zellerfeld,
Germany
5
Key Laboratory for Glass & Ceramics, Shandong Institute of Light Industry, 250353 Jinan, China
2

Received 5 August 2010; accepted 14 September 2010; published online 19 October 2010
The connection between bulk glass properties and network topology is now well established.
However, there has been little attention paid to the impact of network topology on the surface
properties of glass. In this work, we report the impact of the network topology on both the transport
properties such as cationic inward diffusion and the mechanical properties such as hardness of
borate glasses with modified surfaces. We choose soda lime borate systems as the object of this
study because of their interesting topological features, e.g., boron anomaly. An inward diffusion
mechanism is employed to modify the glass surface compositions and hence the surface topology.
We show that accurate quantitative predictions of the hardness of the modified surfaces can be made
using topological constraint theory with temperature-dependent constraints. Experimental results
reveal that Ca2+ diffusion is most intense in glasses with lowest BO4 fraction, whereas Na+ diffusion
is only significant when nonbridging oxygens start to form. These phenomena are interpreted in
terms of the atomic packing and the local electrostatic environments of the cations. 2010
American Institute of Physics. doi:10.1063/1.3497036
I. INTRODUCTION

Alkali and alkaline earth ions are mobile in oxide glasses

near the glass transition temperature Tg. A fundamental understanding of the phenomenon of alkali and alkaline earth
ionic diffusion in glasses is of great scientific and technological importance, e.g., for understanding of glass transition,
electrical conductivity, and fragility and for developing battery materials, sensors, and the ion exchange strengthening
process.1,2 Ionic sites in crystalline materials are mostly well
defined and diffusion occurs via a hopping process among
these sites. In contrast, ionic diffusion in glassy materials is
much more complicated, determined by an interplay among
network structure, topology, energetics, and cooperativity.
The local environments and spatial distributions of cations in
oxide glasses silicate, borate, and phosphate have been extensively studied in the past.3,4 Among these systems, borate
glasses are particularly special due to the presence of the
boron anomaly, i.e., the transformation of trigonal BO3 units
into tetrahedral BO4 units with the initial addition of network
modifying alkali and/or alkaline earth oxides.57 The transformation between BO3 and BO4 units can also be achieved
by varying temperature, pressure, or thermal history.5,812
Therefore, borate glasses are interesting model systems for
investigating the impact of glass structure and topology on
ionic transport.
a

Electronic mail: yy@bio.aau.dk.

0021-9606/2010/13315/154509/6/$30.00

Ionic transport properties have been experimentally studied in various binary alkali borate,1325 ternary alkali
borate,2631 and quaternary alkali borate32 glasses. The correlation between transport and structure in alkali borate
glasses has also been studied by means of molecular dynamics simulations.3336 However, to the best of our knowledge,
studies about alkaline earth diffusion in borate glasses have
not been reported in literature. Recently, we performed a
detailed study about the relationships among structure, topology, dynamics, and properties for a series of soda lime borate
Na2O CaO B2O3 glasses.37 In that study, we found that
addition of either soda Na2O or lime CaO to boric oxide
B2O3 results in structural and topological changes that are
attributed to the conversion of BO3 to BO4. This conversion
takes place until the total modifier content is around 33%,
and further addition of modifier results in formation of nonbridging oxygens NBOs. In this paper, we study another
important aspect of the same series of glasses, namely, the
influence of the network structure and topology on the surface transport properties, specifically, the inward diffusion of
both Na+ and Ca2+ ions by means of the reduction-inward
diffusion route.3840 Reduction of a polyvalent element from
a higher to a lower valence state e.g., Fe3+ to Fe2+ takes
places when the glass is exposed to certain types of gases at
the glass transition temperature. The reduction proceeds via
two simultaneous processes: gaseous e.g., H2 permeation
and outward flux of electron holes h. The gaseous permeation results in the introduction of some structurally bonded

133, 154509-1

2010 American Institute of Physics

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
141.214.17.222 On: Sun, 08 Mar 2015 11:24:36

154509-2

J. Chem. Phys. 133, 154509 2010

Smedskjaer et al.

hydroxyl groups in the surface layer e.g., H2 + 2Fe3+

+ 2Si O 2Fe2+ + 2Si OH.38,39 The electron holes are
generated through the internal reduction e.g., Fe3+ Fe2+
+ h.40 To maintain charge neutrality, the outward flux of
electron holes requires an inward diffusion of mobile cations,
which in turn leads to the formation of a surface layer rich in
network formers. The mechanism of the inward diffusion has
been described in more detail elsewhere.3840
In this work, we also attempt to determine how the inward diffusion and the resulting surface modification influence the glass mechanical performances, viz., by measuring
the microhardness before and after the diffusion process.
Since hardness of materials is of great importance in the
materials science and engineering communities, prediction of
hardness has attracted much interest.4144 We attempt to predict the change in hardness due to inward diffusion by considering the change in the number of network constraints in
the surface layer. If the prediction becomes possible, constraint theory may be used as a tool for quantitatively designing the mechanical properties of glasses.

FIG. 1. Composition dependence of density and atomic packing factor

for the xNa2O 10CaO 89 xB2O3 1Fe2O3 glasses. APF is calculated using Eq. 1. The inset shows the composition dependence of boron speciation expressed as the fraction of tetrahedral to total boron N4. These fractions are taken from Ref. 37.

these measurements, the degree of oxygen incorporation due

to oxidation of Fe2+ was determined and used to estimate the
Fe3+ / Fetot ratio, where Fetot = Fe2+ + Fe3+.45

II. EXPERIMENTAL

III. RESULTS AND DISCUSSION

A series of glass compositions xNa2O 10CaO

1Fe2O3 89 xB2O3 with x = 5, 10, 15, 20, 25, 30, and 35
mol % were prepared by melting mixtures of high purity
H3BO3, Na2CO3, CaCO3, and Fe2O3 in a Pt90Rh10 crucible
under atmospheric air in an induction furnace. The melts
were held at 1050 1150 C for 15 min and then cast onto
a brass plate. Tg of the glasses was determined by differential
scanning calorimetry at 10 K/min for both upscan and prior
downscan as described in detail elsewhere.37 The glasses
were annealed at their respective Tg for 2 h to diminish internal stresses. After annealing, the glass samples were cut to
dimensions of approximately 10 10 2 mm3. One surface
of each sample was then ground by a six-step procedure with
SiC paper, followed by polishing with 3 m diamond suspension. To prevent the chemical attack of water during
grinding, ethanol 96 vol % was used as grinding media.
The inward diffusion process was then induced by heattreating the polished glasses for 8 h at their respective Tg and
1.1Tg in an electric furnace under a flow of H2 / N2 1/99 v/v
gas. The change in surface composition due to the inward
diffusion was determined by means of secondary neutral
mass spectroscopy SNMS on an INA3 Leybold AG,
Alzenau, Germany instrument equipped with a Balzers
QMH511 quadrupole mass spectrometer Balzers Instruments, Balzers, Liechtenstein and a Photonics SEM.
XP1600/14 amplifier JEOL USA Inc. Peabody, USA. The
Vickers microhardness HV was measured using a Duramin
5 indenter Struers A/S, Ballerup, Denmark in air and at
room temperature. Thirty indentations were performed on
each sample at a load of 98.12 mN applied for a duration of
5 s. The density of the samples was determined using a helium pycnometer Porotech GmbH, Hofheim, Germany. The
initial iron redox state in the as-prepared glasses was estimated by thermogravimetry TG measurements in air on a
simultaneous thermal analyzer STA 449C Jupiter, Netzsch,
Selb, Germany using the method established earlier.45 From

A. Atomic packing

We first consider the measured density values of our

soda lime borate glasses since this provides information
about the atomic packing of the glass network. Based on the
density of each glass, we calculate the atomic packing factor
APF, i.e., the ratio between the minimum theoretical volume occupied by the ions and the corresponding molar volume of the glass, as46
APF =

f iVi
,
f iM i

for the ith constituent with the formula AxBy. f i is the molar
fraction, M i is the molar mass, and Vi = 4 / 3NaxrA3
+ yrB3 is the theoretical volume, where rA and rB are the
ionic radii and Na the Avogadros number. APF may then be
calculated by using the effective ionic radii given by
Shannon47 for the appropriate coordination number.
Both density and APF increase as the boric oxide is replaced by sodium oxide. These results are shown in Fig. 1
and in Table I, along with Tg and the ratio of tetrahedral to
TABLE I. Glass transition temperature Tg, ratio of tetrahedral to total
boron N4, density , and the atomic packing factor of
xNa2O 10CaO 1Fe2O3 89 xB2O3 glasses.
x
mol %

Tg
Ka

N4 1%a

g / cm3

APF b

5
10
15
20
25
30
35

693
756
771
768
756
740
711

16
24
36
40
46
43
42

2.14
2.27
2.35
2.45
2.52
2.57
2.61

0.538
0.562
0.573
0.588
0.595
0.597
0.595

Reference 37.
APF is calculated using Eq. 1.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
141.214.17.222 On: Sun, 08 Mar 2015 11:24:36

154509-3

J. Chem. Phys. 133, 154509 2010

Impact of topology on glass surface properties

sion of these ions, and consequently a surface layer enriched

in boric oxide is created. This implies that the motion of the
borate network is decoupled from that of the alkali and alkaline earth ions, i.e., borate units such as BO4 are immobile
in the time window of the experiments. The inward diffusion
process leads to a lower concentration of the modifier ions
near the surface and an enrichment of these ions in the interior to maintain mass balance. For example, this is clearly
seen from the depth profile of sodium in Fig. 2d. However,
the enrichment in the interior of calcium is not clearly visualized. This is because the enrichment occurs in a broader
depth range and the depth profiles are not shown in a sufficiently large range. It should be mentioned that prior to the
heat-treatment, the glasses did not show any concentration
variations as a function of depth. Furthermore, the surface
depletion of modifier ions is not due to the evaporation of
those ions during heat-treatment. First, the heat-treatments
were carried out at rather low temperatures, i.e., near Tg. At
such low temperatures, the evaporation of the alkali and alkaline earth ions will hardly occur. Second, the surface
depletion is not observed during heat-treatment of the glasses
near Tg in an inert atmosphere such as argon. If evaporation
had occurred at this temperature, the surface depletion
should have been observed.
In order to analyze the composition profiles quantitatively, we have calculated the diffusion depths of the sodium
Na and calcium Ca ions. For the glasses treated at
1.1Tg, we have also calculated the average depletion extents
i.e., concentrations in the modified surface layers. These
results are summarized in Table II. In Figs. 3a and 3b, the
diffusion depths Na and Ca are plotted against the Na2O
content for heat-treatment temperatures of Tg and 1.1Tg, respectively. At both temperatures, the same qualitative trends
are observed. Ca decreases with increasing Na2O content
until x 25 and then increases. Hence, as more boron is converted from the three- into the four-coordinated state, the
extent of Ca2+ diffusion decreases. This happens even when
Tg increases in this region Table I. The diffusion extent was
compared among different glasses at their respective Tg in
order to ensure the same viscosity, i.e., the same relaxation
rate of the glass structure for these glasses. Na seems to be
constant until x 25 and then increases. The composition
with 25 mol % Na2O marks the onset of NBO formation.37
Based on a topological model and experimentally deter-

FIG. 2. SNMS depth profiles of xNa2O 10CaO 1Fe2O3 89 xB2O3

glasses heat-treated at various temperatures for 8 h in H2 / N2 1/99. a
Glass with x = 25 treated at its Tg. b Glass with x = 5 treated at 1.1Tg. c
Glass with x = 25 treated at 1.1Tg. d Glass with x = 35 treated at 1.1Tg. The
curves are plotted as concentration of a given element at given depth divided
by the concentration of that element in the bulk of the glass c / cbulk.

total boron N4 as determined in our previous study.37 The

composition dependence of N4 is shown in the inset of Fig.
1. The observed changes in and APF are identical to those
previously reported for binary alkali borates,13,4850 binary
alkaline earth borates,48,51,52 and ternary soda lime borates.48
Since the BO4 unit has the maximal packing density among
the borate units,51 the increase in the atomic packing density
up to x 25 is explained by the increase in the fraction of
four-coordinated boron inset of Fig. 1. In this region, the
fraction of BO4 units increases at the expense of BO3 units
that have a lower packing density. For x 25, addition of
sodium oxide results in formation of NBOs in the glass network at the expense of BO4 units. The borate units with
NBOs have smaller packing fraction than the BO4 units,51
and our data show that the atomic packing density becomes
nearly constant for x 25.
B. Inward diffusion

Figures 2a2d show the SNMS depth profiles of various glasses heat-treated at their Tg or 1.1Tg for 8 h. The
depth profiles reveal that a surface depletion of sodium, calcium, and iron has occurred as a result of the inward diffu-

TABLE II. Diffusion depths of sodium Na and calcium Ca and the average depletion extents of sodium
cNa / cNa,bulkav and calcium cCa / cCa,bulkav in the modified surface layers.
Na
nm

Ca
nm

x
mol %

Th = Tg

Th = 1.1Tg

Th = Tg

Th = 1.1Tg

5
10
15
20
25
30
35

21
28
15
30
29
49
76

66
49
41
73
62
110
170

320
311
273
250
231
254
285

736
716
655
632
603
629
639

cNa / cNa,bulkav

0.660
0.685
0.691
0.632
0.645
0.716
0.763

cCa / cCa,bulkav

0.257
0.243
0.239
0.241
0.245
0.251
0.267

cNa / cNa,bulkav and cCa / cCa,bulkav are given at heat-treatment temperatures Th of 1.1Tg.

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
141.214.17.222 On: Sun, 08 Mar 2015 11:24:36

154509-4

J. Chem. Phys. 133, 154509 2010

Smedskjaer et al.

FIG. 3. Composition dependence of the extent of calcium Ca and sodium Na diffusion in the xNa2O 10CaO 1Fe2O3 89 xB2O3 glasses.
To induce the diffusion, the glasses have been heat-treated for 8 h in H2 / N2
1/99 at a their respective Tg and b 1.1 times their respective Tg. The
lines are drawn as guides for the eyes.

mined Tg values, we have shown in our previous study37 that

there is no preference between sodium or calcium to cause
the boron coordination change. The diffusion results presented here provide additional evidence for this conclusion
since both sodium and calcium diffusion becomes faster as
soon as NBOs start to form. If there were a preference, then
the sodium and calcium ions would display different onset
thresholds for where the diffusivity starts to increase.
To understand the results presented in Fig. 3, it is important to recall that the inward diffusion process is driven by
the reduction of Fe3+ to Fe2+, and hence, the initial redox
state of iron is an important factor affecting the extent of the
inward diffusion.38 The TG measurements indicate that the
Fe3+ / Fetot ratio varies slightly with the sodium content in
the glasses. Specifically, Fe3+ / Fetot varies between 82
and 95 at. % for sodium contents of 5 and 35 mol %, respectively. However, such relatively small differences among the
glasses cannot explain the observed composition dependence
of diffusion.38 Therefore, there must be other factors causing
this dependence as discussed below.
Similar to the case of inward diffusion we have reported
for silicate glasses,53 in borates the diffusivity of alkaline
earths here Ca2+ is faster than that of alkalis here Na+. We
have attributed this to the fact that the alkaline earth ions are
able to carry more positive charges than the alkali ions to
charge-balance the outward flux of electron holes.53 In the
soda lime borate glasses, an additional striking phenomenon
has been found, i.e., the sodium ions set up a electrostatically
rigid environment of NBOs around themselves, whereas this

is not the case for calcium ions.37 This gives a locally floppy
region around calcium with a lower activation barrier for ion
hopping compared to that for sodium.54,55 Furthermore, according to far-IR spectroscopic studies of alkali borates56 and
alkaline earth borates,57 the modifier ions can occupy two
types of sites, one of higher optical basicity negative charge
donating ability of oxygen and one of lower. The sites of
higher basicity provide a lower residual electronic charge on
the cation, and hence the cations in these sites are more
mobile.58,59 At low modifier concentration there are no NBOs
and the cations occupy sites of low optical basicity and
hence of low mobility. Therefore, a higher field strength cation such as Ca2+ with its stronger affinity for oxygen should
be able to be get charge-neutralized more easily than Na+
and will move faster. Once NBOs are formed, both Na+ and
Ca2+ will be charge neutralized more easily and move even
faster compared to the situation in the absence of NBOs. The
present study provides experimental evidence for these ideas.
The decrease in Ca2+ diffusivity with increasing x in the
range of x 25 may be attributed to the increase in atomic
packing density as BO3 units are converted into BO4 groups,
i.e., the mechanical strain imposed by ionic transport upon
the borate network increases with increasing x for x 25. In
contrast to our findings, some researchers reported another
scenario in which an increase in alkali content and hence,
BO4 concentration enhances the ionic mobility in binary
alkali borate glasses.13,1719 This may be explained by a
shortening of the average ionic jump distance and an increase in the number of nearby well-matched potential target
sites with increasing alkali content. In our compositions,
however, we keep the concentration of Ca2+ constant and
vary that of Na+. Thus, Ca2+ cannot take advantage of the
abovementioned factors despite the fact that the ionic radius
of Ca2+ 1.00 is very similar to that of Na+ 1.02 .47
This could indicate that the Na+ ions are virtually immobile
in comparison to the Ca2+ ions in this composition range
Fig. 3. Therefore, the additional Na+ sites formed due to the
increasing Na2O content have very low probability to become vacancies into which the Ca2+ ions can jump, and
hence, diffusion of the Ca2+ ions does not get facilitated.
C. Hardness

Now another question arises, namely, how the surface

modification caused by the inward diffusion process affects
the mechanical properties. This can be partly answered by
Fig. 4, where the Vickers hardness HV data are plotted as a
function of the Na2O content. As discussed in detail
elsewhere,60 the hardness of the untreated glasses increases
as BO3 units are converted into BO4 units and then decreases
as NBOs are created. At the utilized load of 98.12 mN, the
penetration depth of the indenter is around 0.6 0.7 m, i.e.,
in the same range as the diffusion depths for samples heattreatments at 1.1Tg Fig. 3b. Thus, the hardness measurements can provide information on the impact of the inward
diffusion on surface structure and topology. Figure 4 shows
that HV is lower for the heat-treated compared to the untreated samples for x 25. This might be ascribed to the
removal of network modifiers from the surface layer, which

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
141.214.17.222 On: Sun, 08 Mar 2015 11:24:36

154509-5

J. Chem. Phys. 133, 154509 2010

Impact of topology on glass surface properties

Now we consider the surface compositions of the heattreated samples as determined by SNMS and then calculate
the number of room temperature constraints in the depth
range that the indenter penetrates in order to calculate the
hardness of that layer. The penetration depth is calculated as
1/7 of the impression diagonal length as a consequence of
the Vickers indenter geometry. Based on the number of constraints per atom associated with each distinct network forming species, nx , y can be calculated by37
nx,y = 5NB4 + 3NB3 + 3NO +
FIG. 4. Composition dependence of the measured Vickers hardness HV for
the untreated xNa2O 10CaO 1Fe2O3 89 xB2O3 glasses and the measured
and predicted HV for the same glasses heat-treated for 8 h in H2 / N2 1/99
at 1.1Tg. The predicted HV values were determined by using Eqs. 2 and
3. The HV data points of the untreated glasses are taken from Ref. 60.

causes a decrease in the BO4 fraction in that layer, and hence

the layer becomes less load resistant. In the NBO regime
x 25, however, the heat-treatment in hydrogen gas increases the hardness of the glass. This trend may be explained as follows. The modifying ions associated with
NBOs are more mobile than those associated with fourcoordinated boron as charge compensators. As a result, the
modifying ions can easily be depleted from the surface layer
for the glass compositions with x 25%, and this causes a
decrease in the number of NBOs near the surface. Consequently the structure of the surface layer becomes rigid, and
hence, the hardness increases.
Finally, we attempt to quantitatively account for the
change in hardness due to the inward diffusion process. In
our previous study,37 we were able to predict the scaling of
glass transition temperature with composition based on a topological model incorporating temperature-dependent constraints e.g., BO linear constraints and BOB angular
constraints. The constraints become rigid as the liquid is
cooled from high to low temperatures.61,62 Since hardness is
measured at room temperature, additional constraints are important at room temperature in comparison to the glass transition range. Specifically those are the O-B-O angular
constraints.37,60 We have therefore also been able to establish
a correlation between the number of room temperature constraints nx , y and the hardness of the untreated glasses measured at 98.1 mN Ref. 60
HV = 12.6nx,y 31.4 GPa.

This correlation was established by assuming that a certain

critical number of constraints ncrit must be present in order
to produce a connected network that is required for the material to display mechanical resistance, i.e., for hardness to
become nonzero. As discussed in detail elsewhere,60 we set
ncrit = 2.5 since n = 2.5 corresponds to a rigid two-dimensional
network in the three-dimensional space, i.e., additional constraints are acting in the third dimension to provide mechanical resistance. We then assume that hardness is directly proportional to these number of additional constraints, i.e., the
constraints in excess of ncrit. Equation 2 was then obtained
by determining the proportionality constant empirically.60

4x
NM NB,
2x + y
3

where NB4, NB3, NO, and NM NB are the fractions of

four-coordinated boron, three-coordinated, oxygen including both bridging and nonbridging varieties, and network
modifiers Na and Ca that create nonbridging oxygens, respectively. The derivation of the coefficients in Eq. 3 is
given elsewhere.37 The fractions in Eq. 3 may be calculated
from the molar fractions of Na2O and CaO,37 that are in turn
determined from the SNMS data.
Using the above-mentioned procedure that combines
Eqs. 2 and 3 and the measured SNMS data, we have
calculated the hardness values of the samples heat-treated at
1.1 times their respective Tg Fig. 4. Excellent agreement
between the predicted and measured values is observed. This
suggests that temperature-dependent constraint theory61,62
may be used as a tool for predicting mechanical properties
such as hardness as a function of composition. The result
also supports our qualitative explanations above about the
changes in hardness. An effect not included in the current
model of Eq. 3 is the incorporation of hydroxyl groups in
the glass surface layer due to the partial permeation of H2
into the glass. These groups are expected to break linear and
angular constraints, and hence, lower the measured hardness
value. However, our previous studies38,39 and the good agreement between measurement and prediction in Fig. 4 suggest
that the degree of hydroxyl incorporation is rather limited at
the utilized hydrogen pressure of 0.01 bar.
IV. CONCLUSION

The extent of ionic diffusion in soda lime borate glasses

depends on the boron speciation and the electrostatic environments of the mobile cations. The diffusion of Ca2+ becomes less pronounced as the BO4 concentration increases.
This is explained via the change in atomic packing of the
network since the timescale of the network relaxation is decoupled from that of modifier diffusion. Both Na+ and Ca2+
diffusivities increase when NBOs are present since the cations are more mobile in these sites with lower residual electronic charge on the cation. The inward diffusion processes
may be used to modify various surface properties. Hardness
can be tailored by controlling the extent of the inward diffusion and/or by designing the original glass composition. Specifically, the inward diffusion process causes hardness to increase and decrease in compositions with and without NBOs,
respectively. The change in hardness can be quantitatively
predicted by considering the number of room temperature

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
141.214.17.222 On: Sun, 08 Mar 2015 11:24:36

154509-6

constraints in the modified surface layer. Our result implies

that constraint theory may be used as a tool for predicting the
composition dependence of mechanical properties.
ACKNOWLEDGMENTS

The authors thank Thomas Peter Clausthal University of

Technology for performing SNMS measurements. This
work was financially supported by the International Doctoral
School of Technology and Science at Aalborg University under Ph.D. Stipend No. 562/06-FS-28045.
C. A. Angell, Solid State Ionics 1819, 72 1986.
G. N. Greaves and K. L. Ngai, Phys. Rev. B 52, 6358 1995.
A. K. Varshneya, Fundamentals of Inorganic Glasses Society of Glass
Technology, Sheffield, 2006.
4
A. Pedone, J. Phys. Chem. C 113, 20773 2009.
5
A. H. Silver and P. J. Bray, J. Chem. Phys. 29, 984 1958.
6
P. J. Bray and J. G. OKeefe, Phys. Chem. Glasses 4, 37 1963.
7
G. E. Jellison, S. A. Feller, and P. J. Bray, Phys. Chem. Glasses 19, 52
1978.
8
J. Zhong and P. J. Bray, J. Non-Cryst. Solids 111, 67 1989.
9
J. F. Stebbins and S. E. Ellsworth, J. Am. Ceram. Soc. 79, 2247 1996.
10
L. S. Du, J. R. Allwardt, B. C. Schmidt, and J. F. Stebbins, J. Non-Cryst.
Solids 337, 196 2004.
11
S. K. Lee, K. Mibe, Y. Fei, G. D. Cody, and B. O. Mysen, Phys. Rev.
Lett. 94, 165507 2005.
12
J. Wu, J. Deubener, J. F. Stebbins, L. Grygarova, H. Behrens, L. Wondraczek, and Y. Z. Yue, J. Chem. Phys. 131, 104504 2009.
13
F. Berkemeier, S. Voss, . W. Imre, and H. Mehrer, J. Non-Cryst. Solids
351, 3816 2005.
14
S. Murugavel and B. Roling, Phys. Rev. B 76, 180202 2007.
15
H. Mehrer, . W. Imre, and E. Tanguep-Nijokep, J. Phys. Conf. Series
106, 12001 2008.
16
. W. Imre, H. Staesche, S. Voss, M. D. Ingram, K. Funke, and H.
Mehrer, J. Phys. Chem. B 111, 5301 2007.
17
. W. Imre, F. Berkemeier, H. Mehrer, Y. Gao, C. Cramer, and M. D.
Ingram, J. Non-Cryst. Solids 354, 328 2008.
18
H. Doweidar, Y. M. Moustafa, G. M. El-Damrawi, and R. M. Ramadan,
J. Phys.: Condens. Matter 20, 035107 2008.
19
T. Matsuo, M. Shibasaki, and T. Katsumata, Solid State Ionics 154155,
759 2002.
20
O. Majrus, L. Cormier, G. Calas, and B. Beuneu, J. Phys. Chem. B 107,
13044 2003.
21
T. Nagasaki, R. Morishima, and T. Matsui, J. Nucl. Sci. Technol. 39, 386
2002.
22
U. Schoo and H. Mehrer, Solid State Ionics 130, 243 2000.
23
G. D. Chryssikos, L. P. Liu, C. P. Varsamis, and E. I. Kamitsos, J. NonCryst. Solids 235237, 761 1998.
24
S. Sen and J. F. Stebbins, Phys. Rev. B 55, 3512 1997.
25
R. J. Elliott, L. Perondi, and R. A. Barrio, J. Non-Cryst. Solids 168, 167
1994.
26
S. Voss, . W. Imre, and H. Mehrer, Phys. Chem. Chem. Phys. 6, 3669
2004.
27
C. Cramer, S. Brckner, Y. Gao, and K. Funke, Phys. Chem. Chem. Phys.
4, 3214 2002.
28
. W. Imre, S. Voss, and H. Mehrer, Phys. Chem. Chem. Phys. 4, 3219
1
2
3

J. Chem. Phys. 133, 154509 2010

Smedskjaer et al.

2002.
. W. Imre, S. Voss, and H. Mehrer, J. Non-Cryst. Solids 333, 231
2004.
30
Y. Gao, J. Solid State Chem. 178, 3376 2005.
31
D. Wilmer, T. Kantimm, O. Lamberty, K. Funke, M. D. Ingram, and A.
Bunde, Solid State Ionics 7071, 323 1994.
32
Y. Gao and C. Cramer, Solid State Ionics 176, 921 2005.
33
A. H. Verhoef and H. W. den Hartog, J. Non-Cryst. Solids 182, 235
1995.
34
C.-P. E. Varsamis, A. Vegiri, and E. I. Kamitsos, Phys. Rev. B 65,
104203 2002.
35
C.-P. E. Varsamis, A. Vegiri, and E. I. Kamitsos, J. Non-Cryst. Solids
307310, 956 2002.
36
A. Vegiri, C.-P. E. Varsamis, and E. I. Kamitsos, Phys. Rev. B 80,
184202 2009.
37
M. M. Smedskjaer, J. C. Mauro, S. Sen, and Y. Z. Yue, Chem. Mater. 22,
5358 2010.
38
M. M. Smedskjaer and Y. Z. Yue, J. Non-Cryst. Solids 355, 908 2009.
39
M. M. Smedskjaer, J. Deubener, and Y. Z. Yue, Chem. Mater. 21, 1242
2009.
40
R. L. A. Everman and R. F. Cooper, J. Am. Ceram. Soc. 86, 487 2003.
41
A. Y. Liu and M. L. Cohen, Science 245, 841 1989.
42
S. Vepek, J. Vac. Sci. Technol. A 17, 2401 1999.
43
F. M. Gao, J. L. He, E. D. Wu, S. M. Liu, D. L. Yu, D. C. Li, S. Y. Zhang,
and Y. J. Tian, Phys. Rev. Lett. 91, 015502 2003.
44
X. G. Luo, X.-F. Zhou, Z. Y. Liu, J. L. He, B. Xu, D. L. Yu, H.-T. Wang,
and Y. J. Tian, J. Phys. Chem. C 112, 9516 2008.
45
L. F. Kirkegaard, M. Korsgaard, Y. Z. Yue, and S. Mrup, Glass Sci.
Technol. Amsterdam, Neth. 78, 1 2005.
46
T. Rouxel, J. Am. Ceram. Soc. 90, 3019 2007.
47
R. D. Shannon, Acta Crystallogr., Sect. A: Cryst. Phys., Diffr., Theor.
Gen. Crystallogr. 32, 751 1976.
48
L. M. Donohoe and J. E. Shelby, Phys. Chem. Glasses: Eur. J. Glass Sci.
Technol. B 47, 16 2006.
49
H. Doweidar, J. Mater. Sci. 25, 253 1990.
50
H. Doweidar, G. M. El-Damrawi, Y. M. Moustafa, and R. M. Ramadan,
Physica B 362, 123 2005.
51
N. P. Lower, J. L. McRae, H. A. Feller, A. R. Betzen, S. Kapoor, M.
Affatigato, and S. A. Feller, J. Non-Cryst. Solids 293295, 669 2001.
52
S. Kapoor, H. B. George, A. Betzen, M. Affatigato, and S. Feller, J.
Non-Cryst. Solids 270, 215 2000.
53
M. M. Smedskjaer, J. C. Mauro, and Y. Z. Yue, J. Chem. Phys. 131,
244514 2009.
54
D. I. Novita, P. Boolchand, M. Malki, and M. Micoualut, Phys. Rev. Lett.
98, 195501 2007.
55
N. Tsakiris, P. Argyrakis, and I. Avramov, Phys. Rev. E 81, 022101
2010.
56
J. A. Duffy and M. D. Ingram, J. Am. Chem. Soc. 93, 6448 1971.
57
J. A. Duffy, B. Harris, E. I. Kamitsos, G. D. Chryssikos, and Y. D.
Yiannopoulos, J. Phys. Chem. B 101, 4188 1997.
58
J. A. Duffy, E. I. Kamitsos, A. P. Patsis, and G. D. Chryssikos, Phys.
Chem. Glasses 34, 152 1993.
59
E. I. Kamitsos, G. D. Chryssikos, A. P. Patsis, and J. A. Duffy, J. NonCryst. Solids 196, 249 1996.
60
M. M. Smedskjaer, J. C. Mauro, and Y. Z. Yue, Phys. Rev. Lett. 105,
115503 2010.
61
P. K. Gupta and J. C. Mauro, J. Chem. Phys. 130, 094503 2009.
62
J. C. Mauro, P. K. Gupta, and R. J. Loucks, J. Chem. Phys. 130, 234503
2009.
29

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
141.214.17.222 On: Sun, 08 Mar 2015 11:24:36