The introduction of silver ions into glass by diffusion from an external source is of interest in modifying the optical and electrical properties of glass for device applications. Silver is introduced to fabricate in-glass waveguides, while potassium ions are introduced to pattern the silver diffusion by locally impeding the diffusion. Electric fields assist the silver diffusion, leading to faster diffusion rates of ions and allowing greater concentrations of silver without a metallic cluster formation. However, ion depletion layers are formed in the glass as a result of the application of electric field. Here, we study the nature of these depletion layers using depth profiles of composition after electric field diffusion, using cross-sectional energy dispersive x-ray spectroscopy analysis and infrared spectroscopy (Fourier transform infrared attenuated total reflection spectroscopy). We confirm the accelerated diffusion of silver by an electric field and show that potassium impedes the diffusion of silver even when a field is present. We find that an ion depletion layer is formed below the glass surface by the applied electric field which leads to a thermal relaxation and depolarization process when the samples are reheated. Observation of this process enables activation energies and threshold temperatures for the diffusion of Ag, K, and Na to be evaluated and compared with values obtained by composition profiles. Potassium was found to increase the initially low temperature threshold for silver diffusion, increase the activation energy, and also impede sodium diffusion.

The modification of ordinary window glass, soda-lime glass, by the introduction of metal ions into its surface by ion exchange is useful for endowing it with special mechanical,1 optical, electrical, and biological properties. The most common ion exchange process is one in which the alkali cations in the glass, principally sodium, are displaced by ions diffusing from an external source. The introduction of potassium by thermal diffusion from a molten salt bath is a well-known chemical toughening process,2–6 while the introduction of silver is of interest for constructing optical waveguide structures, optical amplifiers, laser host materials, solar spectral converters,7–14 and for imparting an antimicrobial property to the surface.15–17 The importance of silver in creating optical structures derives from the increase in refractive index when silver ions replace sodium ions and the introduction of useful nonlinear properties by the presence of silver nanoparticles.10–12 One of the challenges in creating loss-free optical waveguide structures is that the diffusion of silver from a molten salt too readily creates metallic nanoparticles instead of causing exchange of individual ions.13–16 It has been found that the application of an electric field enhances the ion exchange of silver into glass,18–24 either when using electrodes in the molten salt25–27 or metallic silver electrodes on the surface of the glass.28,29 It has also been observed that there is an interaction between potassium and silver that inhibits the thermal indiffusion of silver.30,31 While the interaction between silver and potassium ions may be of nuisance value when attempting ion exchange of silver in glass that has been previously chemically toughened with potassium, it can be helpful when creating waveguide structures by ion exchange. By prediffusing potassium into defined areas, silver can be prevented from entering those areas of a waveguide design where it is not wanted.30 There are also reasons to believe that the electric field process is preferable to the thermal process if ion exchange rather than silver nanoparticles is the desired outcome, since glass laden with silver nanoparticles can be converted into ion exchange glass by an electric field “poling” process.32 The use of potassium to act as a diffusion mask would be useful to fabricate devices, but it is not known whether the potassium inhibition effect is also present when electric field induced ion exchange is used. In this study, we address these questions by comparing the effectiveness of ion exchange by thermal diffusion and electric field-assisted diffusion and examining the interactions between the potassium ion and the silver ion. We use cross-sectional analysis by energy dispersive analysis of x-rays to obtain diffusion profiles and examine the structure of the ion exchanged glass by infrared absorption, its optical properties by spectrophotometry. We are particularly interested in examining the nature of the ion depletion layer to determine whether there is a built-in polarization field remaining after field-assisted diffusion. The incorporation of potassium and silver ions together is of interest in combining toughening with other enhanced properties,1,33 so stress effects associated with the depletion layer are also studied.

Soda-lime silicate glass samples of 1.9 mm thickness were cut into squares of 15 × 15 mm2 and cleaned with detergent and then ethanol. Potassium nitrate, KNO3 (Sigma-Aldrich, Inc.), was placed on the samples and melted to cover the surface by heating in an oven (B & L Tetlow, Melbourne, Australia). Annealing temperature is 450 °C and the annealing time was 25 h. After annealing, KNO3 was removed from the glass surface by dissolving in water. After drying, Ag paste (RS Components, RS 186-3593, with 56% silver content and conductivity of the order 0.001 ohm cm) was applied to both surfaces of the samples and dried. Two different sample types of 1.9 mm thickness, one prediffused with potassium and the other without, were used in the setup shown in Fig. 1.

FIG. 1.

(a) The configuration used for applying DC voltage to the samples for ion diffusion from the electrodes in contact with the sample. The sample shown has a potassium prediffused layer at its surfaces and silver paste are the electrodes. (b) The configuration used for applying field using electrodes is not in direct contact with the sample surfaces. iP is the polarization current arising from ion diffusion. (c) The configuration used for thermal relaxation ion spectroscopy (TRIS) measurements. iD is the depolarization current arising from ion diffusion in the sample in response to the built-in field.

FIG. 1.

(a) The configuration used for applying DC voltage to the samples for ion diffusion from the electrodes in contact with the sample. The sample shown has a potassium prediffused layer at its surfaces and silver paste are the electrodes. (b) The configuration used for applying field using electrodes is not in direct contact with the sample surfaces. iP is the polarization current arising from ion diffusion. (c) The configuration used for thermal relaxation ion spectroscopy (TRIS) measurements. iD is the depolarization current arising from ion diffusion in the sample in response to the built-in field.

Close modal

The applied DC voltage was 1500 V, providing a field strength in the sample of 790 V/mm. Electrical resistance was measured using a TENMA® model 72-9405 resistance meter when applying the voltage. The resistance measurement range for 1500 V applied was found to be from 20 GΩ to 1 MΩ. The electrical conductivity of the glass samples undergoing diffusion at various temperatures was calculated from the applied voltage, the measured current, the sample thickness, and the area of the electrode which covered the entire face of the sample. The temperature dependence of the conductivity was found to show an activated behavior, enabling an activation energy to be calculated from the Arrhenius equation as follows:

k=AexpEakBT,
(1)

where k is the diffusion rate constant, A is a pre-exponential factor, Ea is the activation energy, kB is the Boltzmann constant, and T is the temperature. The activation energy Ea is dependent on the interaction energy of the positive ions with their adjacent negatively charged non-bridging oxygen (NBO) atoms.

A Phenom XL® (Phenom-World B.V.) scanning electron microscope (SEM) was used for imaging using a backscattered electron detector as well as for energy dispersive x-ray spectroscopy (EDS) elemental profiling as a function of depth. The operating voltage for all measurements was 15 kV. For depth profiling, the cross-sectional glass samples were fractured by propagating cracks from a scratch. EDS measurements were integrated for 0.5 s at each of the 512 locations. The EDS depth profiles for silver, sodium, and potassium near the positive electrode were measured on samples prepared at temperatures of 120, 145, 170, 195, and 220 °C for diffusion times in the range of 1–3 h.

Fourier transform infrared attenuated total reflection spectroscopy (FTIR-ATR) was carried out using a VERTEX 80v spectrometer fitted with a Ge crystal ATR accessory from Brucker, USA. FTIR-ATR is sensitive to absorptions at depths on the order of 10 μm and has been used for silicate glass surface studies. The spectra were recorded with a spectral resolution of 4 cm−1, using 64 scans. The spectral lines were fitted with Gaussian profiles using OPUS software.

Transmission measurements in the wavelength range of 200–800 nm were carried out using a Varian-Cary model 5E UV-Vis-NIR spectrophotometer using 1 nm wavelength steps. The total transmittance was calculated from the transmitted light intensity through the sample, after subtraction of the background with the beam blocked.

A new technique we have developed for this work undertakes a measurement of the current passing between two capacitively coupled copper electrodes placed in each side of the sample as the temperature of a pretreated sample is ramped up at a constant rate. The principle of this measurement is that the motion of ions in the sample is detected in the circuit connecting the two copper electrodes. The ions move in response to the built-in potential differences induced by the polarization of the specimen during the electric field-assisted diffusion process. When the sample temperature increases, the positive ion mobility increases and the ions start to move in response to the built-in field. The ion movement corresponds to a current, which is carried between the plates as a displacement current. The temperature dependence of the current in the circuit is characteristic of the ion and its environment in the glass host. The current and the voltage between the plates is a type of spectrum of the temperature dependence of the ions in the sample in the built-in field region. We term the method “thermal relaxation ion spectroscopy” (TRIS). Samples were prepared by applying a voltage of 1500 V for various times at a range of temperatures. Samples of soda-lime glass, Ag diffused soda-lime glass, and soda-lime glass prediffused with potassium were prepared. The DC conditions are listed in Table I. The displacement current was measured as voltage across a resistance as shown in Fig. 1. After samples were prepared, each one was again placed in the oven between the copper electrodes and the temperature increased at a controlled rate while voltage was measured at each temperature point.

TABLE I.

Sample preparation conditions for thermal relaxation ion spectroscopy (TRIS) measurements.

Electrodes used for sample preparationPotassium statusVoltage (V)DC time (min)DC temperature (°C)
Cu not in contact None 1500 10 200 
Cu not in contact Prediffused 1500 10 240 
Ag in contact None 1500 10 260 
Ag in contact Prediffused 1500 10 280 
Electrodes used for sample preparationPotassium statusVoltage (V)DC time (min)DC temperature (°C)
Cu not in contact None 1500 10 200 
Cu not in contact Prediffused 1500 10 240 
Ag in contact None 1500 10 260 
Ag in contact Prediffused 1500 10 280 

Figure 2(a) shows the EDS depth profile of a reference sample containing a potassium layer prepared by thermal diffusion without field assistance by oven baking at 450 °C for 25 h with a molten KNO3 layer. The atomic percent was calculated from the Na, K, Ag, Ca, Mg, Si, and O counts. The decrease in sodium toward the surface indicates the region where sodium was replaced by potassium which had been supplied from the molten KNO3. Figure 2(b) shows the profile for the glass as received and Fig. 2(c) shows the profile for a sample without potassium that was oven baked with a silver paste electrode on its surface at 550 °C, for 20 h, showing that very little silver diffuses into the glass at this temperature, even in the absence of potassium. The elemental composition of the as-received glass is shown in Table II, which shows a small detectable amount of Ag. However, Ag is not present in soda-lime glass and this detection indicates the noise level of these measurements.

FIG. 2.

(Left panels) atomic composition as a function of depth in EDS analysis where the indigo line indicates sodium, the red line potassium, and the yellow line silver. (a) After baking with a molten KNO3 layer at 450 °C for 25 h, without an applied field. (b) A reference soda-lime glass sample as received. (c) A sample with silver paste applied to its surface after oven baking at 550 °C for 20 h. (Right panels) SEM images in backscattered electrons of the corresponding sample cross sections. The dashed line indicates the line used for the cross-sectional analysis. The scale bar is 30 nm in (a) and (b) and 10 nm in (c).

FIG. 2.

(Left panels) atomic composition as a function of depth in EDS analysis where the indigo line indicates sodium, the red line potassium, and the yellow line silver. (a) After baking with a molten KNO3 layer at 450 °C for 25 h, without an applied field. (b) A reference soda-lime glass sample as received. (c) A sample with silver paste applied to its surface after oven baking at 550 °C for 20 h. (Right panels) SEM images in backscattered electrons of the corresponding sample cross sections. The dashed line indicates the line used for the cross-sectional analysis. The scale bar is 30 nm in (a) and (b) and 10 nm in (c).

Close modal
TABLE II.

Elemental composition of the as-received glass determined by EDS in atomic %.

ElementOSiNaCaMgAlKAg
Atomic % 63.5 20.1 9.3 2.4 2.5 1.2 0.5 0.4 
ElementOSiNaCaMgAlKAg
Atomic % 63.5 20.1 9.3 2.4 2.5 1.2 0.5 0.4 

Figure 3 shows the atomic composition as a function of depth near the positive silver electrode after the application of DC voltage of 1500 V for times of 1 h at 170 °C. Silver diffuses into the sample under these conditions without prediffused potassium. The silver content of the glass was also observed in backscattered electron SEM images as atomic number contrast and the depth of diffusion in the images appears consistent with the EDS compositional analysis. The results for the depth of silver field-assisted diffusion as observed by SEM as a function of temperature and time are summarized in Table III. The quoted depths of diffusion in this work are the results determined from EDS. The total amount of sodium and silver is slightly lower than for sodium in the bulk, showing that the surface region has a net negative charge arising from the deficit of positive ions. This charged region leads to the possibility of a depolarization current which we describe in Sec. III G.

FIG. 3.

(Left) The atomic composition as a function of depth in a sample without prediffused K after field-assisted diffusion of Ag at 170 °C, 1500 V for 1 h. Na is indigo, Ag is yellow, and light blue is the total of Na and Ag. The net deficit of positive ions in the surface region is demonstrated by the line for Na + Ag. (Right) The SEM image of the diffused region showing the silver diffused region as white. The dashed line corresponds to the depth profile in (a).

FIG. 3.

(Left) The atomic composition as a function of depth in a sample without prediffused K after field-assisted diffusion of Ag at 170 °C, 1500 V for 1 h. Na is indigo, Ag is yellow, and light blue is the total of Na and Ag. The net deficit of positive ions in the surface region is demonstrated by the line for Na + Ag. (Right) The SEM image of the diffused region showing the silver diffused region as white. The dashed line corresponds to the depth profile in (a).

Close modal

Figure 4 shows the results of field-assisted diffusion of silver into potassium-containing samples. The apparent step changes of potassium concentration seen in Fig. 4 that are concurrent with sharp changes in the Ag level are artifacts arising from the interaction of the EDS peaks of Ag and K which artificially raises the apparent potassium level. Although correction for this overlap is not made explicitly here, care is taken in the interpretation of the EDS profiles to avoid errors. The electric field resulting from the application of 1500 V DC to silver electrodes does not cause the diffusion of silver into a potassium-containing sample held at 170 °C (see Fig. 4) as it did when potassium is absent as shown in Fig. 3. At higher temperatures, however, the field does cause the silver to move into glass prediffused with K, with potassium remaining near the surface. It appears that the field drives the silver deeply into the glass, bypassing potassium while displacing sodium to the negative electrode side. The evidence for this is the presence of a sharp concentration increase of sodium beyond the zone where silver has penetrated, while potassium remains on the surface.

FIG. 4.

The left panels show the atomic distribution as a function of depth in a sample prediffused with K after field-assisted diffusion of Ag. The analysis is along the lines shown in the cross-sectional SEM images in the right panels. Na is indigo, Ag is yellow, and K is red. (a) At 170 °C, after applying 1500 V for 1 h, the presence of K appears to prevent the inward diffusion of Ag. (b)–(d) show that at 220 °C, field-assisted diffusion of Ag takes place after 1, 2, and 3 h, respectively. The interaction between silver and Na ions causes Na to diffuse inward while potassium remains on the surface.

FIG. 4.

The left panels show the atomic distribution as a function of depth in a sample prediffused with K after field-assisted diffusion of Ag. The analysis is along the lines shown in the cross-sectional SEM images in the right panels. Na is indigo, Ag is yellow, and K is red. (a) At 170 °C, after applying 1500 V for 1 h, the presence of K appears to prevent the inward diffusion of Ag. (b)–(d) show that at 220 °C, field-assisted diffusion of Ag takes place after 1, 2, and 3 h, respectively. The interaction between silver and Na ions causes Na to diffuse inward while potassium remains on the surface.

Close modal

Figure 5 shows SEM images of a field-assisted diffusion sample held at 220 °C for 3 h with an applied voltage of 1500 V. Cracks are observed on the edge of the potassium diffused sample, suggesting that the introduction of Ag weakens the glass, which is possibly the result of stress changes arising from a deficit of positive ions.

FIG. 5.

SEM images showing the silver diffused region as bright in a sample held at 220 °C for 3 h with electric field assistance. The upper layers show vertical cracking consistent with tensile stress and a horizontal crack appears at the depth where the potassium-containing layer ends.

FIG. 5.

SEM images showing the silver diffused region as bright in a sample held at 220 °C for 3 h with electric field assistance. The upper layers show vertical cracking consistent with tensile stress and a horizontal crack appears at the depth where the potassium-containing layer ends.

Close modal
TABLE III.

A summary of the depth of field-assisted silver diffusion observed by SEM as a function of time and temperature, in samples with and without prediffused K. The applied voltage was 1500 V in all cases.

TemperatureTime (h)Potassium statusSilver depth ( μm)
220 With K 11 
220 With K 23 
220 With K 44 
195 With K 
170 No K 15 
170 With K No detection 
TemperatureTime (h)Potassium statusSilver depth ( μm)
220 With K 11 
220 With K 23 
220 With K 44 
195 With K 
170 No K 15 
170 With K No detection 

The color in transmitted light of the glass sample that was baked with Ag paste on its surface at 550 °C for 20 h without field-assisted diffusion is shown in Fig. 6(a). The color change from clear to yellow indicates that silver was present as particles. The depth profile of this sample shown in Fig. 2(c) is consistent with the presence of a small number of silver particles indicated by noise in the Ag trace. Figure 6(b) shows the sample that was diffused with silver using an applied voltage of 1500 V at 220 °C for 1 h. This color of the glass did not change even though silver was present in significant amounts on and below the surface.

FIG. 6.

Color of (a) a sample baked without applied voltage at 550 °C for 20 h and (b) a sample with DC voltage of 1500 V applied at 220 °C for 3 h.

FIG. 6.

Color of (a) a sample baked without applied voltage at 550 °C for 20 h and (b) a sample with DC voltage of 1500 V applied at 220 °C for 3 h.

Close modal

The visible light adsorption of both samples was measured as a function of wavelength in the UV and visible regions as shown in Fig. 7. The baked sample shows an absorption peak around 400 nm which is attributed to the presence of Ag particles.34,35 The field-assisted sample does not show specific absorption peaks in this region. However, the transmittance is lower than for untreated soda-lime glass, largely the result of sodium precipitation at the surface on the negative electrode side that caused an observed roughening of the surface on that side, scattering light and reducing the direct transmitted beam.

FIG. 7.

The light transmittance of various samples of field-assisted silver diffused glass as a function of wavelength in the UV and visible regions from 200 to 800 nm. Untreated soda-lime glass (red), field-assisted diffusion with a DC voltage of 1500 V applied at 220 °C for 1 h (gray), 2 h (yellow), and 3 h (blue). The transmittance of the sample baked at 550 °C for 20 h without field assistance is shown in green.

FIG. 7.

The light transmittance of various samples of field-assisted silver diffused glass as a function of wavelength in the UV and visible regions from 200 to 800 nm. Untreated soda-lime glass (red), field-assisted diffusion with a DC voltage of 1500 V applied at 220 °C for 1 h (gray), 2 h (yellow), and 3 h (blue). The transmittance of the sample baked at 550 °C for 20 h without field assistance is shown in green.

Close modal

FTIR-ATR spectroscopy gives the vibrational spectrum of the modified region near the surface of the samples. The vibrational density of states (VDOS) of soda-lime glass has signatures arising from the Si-O stretching of Si-centered siloxane linked tetrahedra as shown in Fig. 8. The bending of Si–O–Si bonds in CVD deposited amorphous silica shows three different FTIR adsorption peaks in the range of 750–1250 cm–1.36 In soda-lime glass, this region may show as many as seven vibrational modes relevant to the SiO4 network structure including modes affected by sodium or potassium interactions with non-bridging oxygen (NBO) atoms,37 One of these, centered at 970 cm−1, has been associated with vibrations of the sodium ion and the interacting NBO atom as shown in the peak fitting in Figs. 9(a)9(e), since its intensity depends on the sodium concentration.37–39 When sodium is replaced by the heavier potassium ion, the peak moves to around 940 cm–1 depending on the potassium concentration40 as shown in Figs. 9(a) and 9(b).

FIG. 8.

FTIR-ATR spectra of soda-lime glass samples: (a) As received (blue) and with potassium prediffused into the sample by oven baking (red). (b) Prediffused with potassium and then treated using silver electrodes with DC voltage applied at 1500 V at 220 °C for the times shown. The blue curve is without treatment, green is for 1 h treatment, light blue is for 2 h, and orange is for 3 h. The change in peak shape is caused by the displacement/removal of the sodium peak, leaving the potassium peak more obvious. (c) Oven baked with Ag paste and no field assistance at 550 °C for 20 h (black). The peak does not shift substantially upon silver diffusion. (d) After field-assisted diffusion using Ag paste electrodes at 220 °C for 1 h without prediffused potassium (purple); blue is for glass as received. The peak shifts upward (see text).

FIG. 8.

FTIR-ATR spectra of soda-lime glass samples: (a) As received (blue) and with potassium prediffused into the sample by oven baking (red). (b) Prediffused with potassium and then treated using silver electrodes with DC voltage applied at 1500 V at 220 °C for the times shown. The blue curve is without treatment, green is for 1 h treatment, light blue is for 2 h, and orange is for 3 h. The change in peak shape is caused by the displacement/removal of the sodium peak, leaving the potassium peak more obvious. (c) Oven baked with Ag paste and no field assistance at 550 °C for 20 h (black). The peak does not shift substantially upon silver diffusion. (d) After field-assisted diffusion using Ag paste electrodes at 220 °C for 1 h without prediffused potassium (purple); blue is for glass as received. The peak shifts upward (see text).

Close modal
FIG. 9.

FTIR-ATR spectra fitted with component peaks attributed to Si-O vibrations (SiO 1, 2, 3) and vibrations that involve the NBO atom and the alkali cation (Si-NBO-Na and Si-NBO-K). (a) Soda-lime glass. (b) Thermally prediffused potassium. (c) Potassium prediffused with DC applied at 220 °C for 1 h. (d) No potassium Ag baking. (e) No potassium with DC applied at 170 °C for 1 h.

FIG. 9.

FTIR-ATR spectra fitted with component peaks attributed to Si-O vibrations (SiO 1, 2, 3) and vibrations that involve the NBO atom and the alkali cation (Si-NBO-Na and Si-NBO-K). (a) Soda-lime glass. (b) Thermally prediffused potassium. (c) Potassium prediffused with DC applied at 220 °C for 1 h. (d) No potassium Ag baking. (e) No potassium with DC applied at 170 °C for 1 h.

Close modal

Figure 8(a) compares FTIR spectra for a soda-lime glass sample as received and a thermally diffused potassium-containing sample. The soda-lime glass shows a peak at 983.6 cm−1 while the K diffused sample shows a peak at 970.1 cm−1. This shift to lower wavenumbers arises from the lower vibrational frequencies associated with the higher mass of the potassium ion.41 The potassium displaces sodium from the surface layers of the sample and modifies the FTIR-ATR spectrum as shown in Fig. 9(b).

Figure 8(b) shows the FTIR spectra of the samples prediffused with potassium then diffused with silver using an applied DC voltage of 1500 V at 220 °C for times of 1, 2, and 3 h. All samples show peaks at 950 cm−1 where the shift to a lower wave number is expected from the presence of potassium and the reduction in the levels of sodium. The shift is greater than observed for the baked samples [Fig. 8(c)] indicating a greater contribution from potassium as shown in the peak analysis in Fig. 9(c), which would occur as silver displaces sodium but not potassium from the surface layers.

Figure 8(c) shows the FTIR spectrum for a sample without potassium and diffused with silver for 20 h at 550 °C using oven baking without an applied field. The peak does not change significantly, indicating that Ag diffused by oven baking does not affect the mode frequencies, a result that would be expected if silver aggregated into particles without ion exchange with sodium [Fig. 9(d)].

Figure 8(d) shows the effect of field-assisted diffusion of silver into a sample without potassium. There is an upward shift in the spectrum that is associated with a diffusion of sodium away from the surface caused by the electric field and the ingress of silver ions. The peak analysis [Fig. 9(e)] indicates that there is no potassium peak as expected, the sodium peak is reduced, and the peak is shifted to higher frequencies most likely as a result of the deficit of sodium near the surface. The displacement of sodium ions by the electric field may cause the formation of some vacancies and possibly some hydrogen substitution. Amma et al.42 have found that acid etching of soda-lime glass causes a similar upward shift in the peak wavenumbers in this region which they attribute to a reduction in sodium ion concentration in the surface caused by acid etching, causing an increase in hydrogen at the surface.42 

The peak positions for all the spectra are summarized in Table IV and a peak fitting analysis is shown in Fig. 9 and summarized in Table V.

TABLE IV.

FTIR peak center positions for the silver diffused samples prepared with and without applied DC voltage and with and without prediffused potassium.

Temperature, voltage, timeIon contentPeak position (cm−1)
No treatment No Ag, K 983.6 
No treatment With K/No Ag 970.1 
220 °C, 1500 V, 1 h With Ag, K 954.6 
170 °C, 1500 V, 1 h With Ag/No K 1006.7 
550 °C, 0 V, 20 h With Ag/No K 972.0 
Temperature, voltage, timeIon contentPeak position (cm−1)
No treatment No Ag, K 983.6 
No treatment With K/No Ag 970.1 
220 °C, 1500 V, 1 h With Ag, K 954.6 
170 °C, 1500 V, 1 h With Ag/No K 1006.7 
550 °C, 0 V, 20 h With Ag/No K 972.0 
TABLE V.

Peak height results obtained from the fitting FTIR-ATR spectra shown in Fig. 9. The peak positions and full widths at half maximum are fixed during the fitting process.

Temperature, voltage, timeIon content1150.0 cm−11070.0 cm−11025.0 cm−1970.5 cm−1942.2 cm−1
SiO 1SiO 2SiO 3NBO-NaNBO-K
No treatment No Ag, K 0.12 0.25 0.27 0.91 0.00 
No treatment With K/No Ag 0.09 0.24 0.19 0.88 0.09 
220 °C, 1500 V, 1 h With Ag, K 0.06 0.26 0.23 0.70 0.41 
550 °C, 20 h With Ag/No K 0.11 0.30 0.14 0.95 0.00 
170 °C, 1500 V, 1 h With Ag/No K 0.16 0.38 0.45 0.72 0.00 
Temperature, voltage, timeIon content1150.0 cm−11070.0 cm−11025.0 cm−1970.5 cm−1942.2 cm−1
SiO 1SiO 2SiO 3NBO-NaNBO-K
No treatment No Ag, K 0.12 0.25 0.27 0.91 0.00 
No treatment With K/No Ag 0.09 0.24 0.19 0.88 0.09 
220 °C, 1500 V, 1 h With Ag, K 0.06 0.26 0.23 0.70 0.41 
550 °C, 20 h With Ag/No K 0.11 0.30 0.14 0.95 0.00 
170 °C, 1500 V, 1 h With Ag/No K 0.16 0.38 0.45 0.72 0.00 

It is clear from our results that Ag drives Na out of the surface region. However, the location of the introduced Ag ions is not certain, since there is no obvious peak associated with silver in the range of 650–1300 cm−1. We do not observe the SiO-H peak expected at 3650 cm−1 (Ref. 42) for NBO-Na sites occupied by protons. There are two hypotheses, for further investigation:

  1. Many NBO-Na sites are unoccupied after silver diffusion, associated with an ion depletion region that contains vacancies at the NBO sites and a net negative charge. The charge would be balanced by a net positive charge in the sample at greater depth.

  2. The NBO-Na site is occupied by Ag ions and the spectral features associated with Ag are very weak or lie outside our measurement range.

An Arrhenius plot for the electrical conductivity of soda-lime glass undergoing field-assisted diffusion with a voltage of 1500 V applied to silver electrodes for a potassium prediffused sample is shown in Fig. 10, giving an activation energy of 1.07 eV, calculated from the slope using Eq. (1). For comparison, the activation energy calculated in the same way without potassium has a lower value of 0.77 eV (Fig. 10). The higher activation energy in the presence of potassium is consistent with other observations of the ability of potassium to impede the ion exchange process in silver. For comparison, the activation energy was also calculated from the atomic composition profiles from EDS analysis and the results are given along with literature studies in Table VI.

FIG. 10.

Arrhenius plot derived from the temperature dependence of the electrical conductivity of a glass sample containing prediffused potassium (blue), without potassium (orange), and placed between silver electrodes with an applied voltage of 1500 V. Temperature (T) is in Kelvin. The slope of this plot gives the activation energy for Ag with and without K present as shown in Table VI .

FIG. 10.

Arrhenius plot derived from the temperature dependence of the electrical conductivity of a glass sample containing prediffused potassium (blue), without potassium (orange), and placed between silver electrodes with an applied voltage of 1500 V. Temperature (T) is in Kelvin. The slope of this plot gives the activation energy for Ag with and without K present as shown in Table VI .

Close modal
TABLE VI.

Activation energies for electric field-assisted ion exchange in glass derived from the sample resistance during treatment temperature from TRIS measurements and from EDS depth profile measurements. Activation energies are calculated from Arrhenius plots as a function of 1/T for soda-lime glass samples with various compositions as shown.

Measurement methodElectrode used for field application/detectionGlass compositionBasis of activation energy calculationActivation energy (eV)Temperature range (°C)Probable activated ion
DC applied Ag Soda-lime Conductivity 0.77 120–220 Ag 
DC applied Ag Soda-lime with K Conductivity 1.07 120–220 Ag 
AC applied43  Graphite 20 wt. % Na2O 80 wt. % SiO2 Conductivity 0.74 ± 0.06 400–475 Na 
Voltage from sample Ag Soda-lime TRIS 1.05 140–240 Ag 
Voltage from sample Ag Soda-lime with K TRIS 1.37 160–260 Ag 
Voltage from sample Cu plate not in contact Soda-lime TRIS 1.12 280–360 Na 
Voltage from sample Cu plate not in contact Soda-lime with K TRIS 0.96 440–520 Na 
DC applied Ag Soda-lime with K Diffusion depth 1.11 195–220 Ag 
Thermal diffusion None Soda-lime with K Diffusion depth 0.64 400–550 
Measurement methodElectrode used for field application/detectionGlass compositionBasis of activation energy calculationActivation energy (eV)Temperature range (°C)Probable activated ion
DC applied Ag Soda-lime Conductivity 0.77 120–220 Ag 
DC applied Ag Soda-lime with K Conductivity 1.07 120–220 Ag 
AC applied43  Graphite 20 wt. % Na2O 80 wt. % SiO2 Conductivity 0.74 ± 0.06 400–475 Na 
Voltage from sample Ag Soda-lime TRIS 1.05 140–240 Ag 
Voltage from sample Ag Soda-lime with K TRIS 1.37 160–260 Ag 
Voltage from sample Cu plate not in contact Soda-lime TRIS 1.12 280–360 Na 
Voltage from sample Cu plate not in contact Soda-lime with K TRIS 0.96 440–520 Na 
DC applied Ag Soda-lime with K Diffusion depth 1.11 195–220 Ag 
Thermal diffusion None Soda-lime with K Diffusion depth 0.64 400–550 

The TRIS measurements’ result is shown in Fig. 11. TRIS measurements provide additional valuable insights into the ease of diffusion of ions in soda-lime glass and the effect of potassium. The activation energies derived from TRIS shown in Tables VI and VII are reasonable in agreement with those derived from current and diffusion depth. A more detailed summary of the TRIS peak energies is shown in Table VII and it shows that potassium impedes the electric field-assisted diffusion of silver. In addition, the TRIS measurements also show that the field-assisted diffusion of sodium is also impeded by the addition of potassium.

FIG. 11.

TRIS measurements for four different samples of soda-lime glass, plotted (left) as the voltage across the 10 MΩ internal resistance in series with the sample in the circuit of Fig 1(c) and on the right as an Arrhenius plot of the natural logarithm of the voltage as a function of inverse temperature in Kelvin across the resistor. The linear portions of the Arrhenius plots give the activation energies for ion motion presented in Tables VI and VII.

FIG. 11.

TRIS measurements for four different samples of soda-lime glass, plotted (left) as the voltage across the 10 MΩ internal resistance in series with the sample in the circuit of Fig 1(c) and on the right as an Arrhenius plot of the natural logarithm of the voltage as a function of inverse temperature in Kelvin across the resistor. The linear portions of the Arrhenius plots give the activation energies for ion motion presented in Tables VI and VII.

Close modal
TABLE VII.

The temperature and voltage across the resistor and corresponding current at the peak voltage and the activation energy for ion diffusion in four sample types are shown for TRIS measurements presented in Fig. 11. Potassium appears to impede both sodium and silver ion diffusion as it drives the temperature at peak voltage to higher values.

ConditionPeak position (°C)Peak height (mV)Peak current ( μA)Activation energy (eV)
Soda-lime glass 550 270 0.027 1.12 
Soda-lime with K 580 360 0.036 0.96 
Ag 230 720 0.072 1.05 
Ag with K 310 550 0.055 1.37 
ConditionPeak position (°C)Peak height (mV)Peak current ( μA)Activation energy (eV)
Soda-lime glass 550 270 0.027 1.12 
Soda-lime with K 580 360 0.036 0.96 
Ag 230 720 0.072 1.05 
Ag with K 310 550 0.055 1.37 

The activation energies of Ag show an increase when potassium is present when compared at the same temperature range. The activation energies decrease in the high temperature range.

The results of this study are now interpreted in terms of relevant properties of the three ions. Silver has a smaller ionic radius43 (0.113 nm) than potassium (0.133 nm) or sodium (0.098 nm) and is likely to have a weaker binding energy in the sites interacting with NBO atoms because of its higher electronegativity (1.89)44 compared to sodium (0.93) or potassium (0.89). These properties mean the mobility of silver in diffusion, either thermal or field assisted, is likely to be significantly higher than for either sodium or potassium. This greater mobility has been observed previously by Guldiren et al.45 Of the three ions, potassium is likely to have the lowest mobility because of its large ionic radius and because of its lower electronegativity which will give it the highest binding energy of the three ions in the sites interacting with NBO atoms. Note that the mobility of ions in soda-lime glass is electric field dependent, increasing with field strength.46 

Being the highest electronegative element, silver is the most likely to show a preference for forming neutral atoms from positive ions, resulting in the tendency to form neutral aggregates in the form of silver nanoparticles. While this occurs readily in soda-lime glass, it does not appear to happen in aluminosilicate glass which does not have NBO sites for ions.6 Once the metallic state is formed, the inward migration of individual ions is likely to be impeded, as is observed with thermal diffusion of Ag. However, the presence of an electric field encourages the migration of individual ions, throwing the balance in favor of dispersing the ions in the glass.

The weak binding energy of a silver ion in a site that interacts with an NBO atom may allow it to diffuse in other locations from the other ions, although the strong binding of the large potassium ion will still impede the diffusion of silver. The factors also allow potassium to impede the diffusion of sodium when both ions are present in the same vicinity, in agreement with our results.

We show that DC voltage application at an elevated temperature creates a positive ion depletion region47–49 which has a greater number of negative NBO atoms compared to positive ions. The depletion region is associated with a built-in electric field and a charge double layer that is revealed in the TRIS measurement. This phenomenon is reminiscent of the charge double layers that form in the solution when ions are displaced by an electric field applied to electrodes and are revealed when the applied voltage is removed (see, for example, Guo and McKenzie50). It was confirmed that if there is no field applied to the glass, then there is no observed voltage in TRIS.

The peaks in the TRIS measurement reveal the temperature at which the largest ion movement of ions of a particular species occurs. This temperature is characteristic of the mobility of the ions as a function of temperature. In the case of Ag, the peak occurs at 340 °C, when the dominant mobile species is Ag. Sodium starts to move from 360 °C and starts to fall from 540 °C.51 This indicates that when Ca becomes active, it disturbs the movement of Na. This disturbance of ion movement may be considered an aspect of the “mixed-alkali effect.” This effect is the change in the mobility of one type of ion when it is gradually replaced by a second type of ion.52,53

The main findings of this experimental study of electric field-assisted ion migration in soda-lime glass are as follows:

The application of an electric field in glass at elevated temperature polarizes the glass and creates a negatively charged region depleted of metal ions in the surface near the anode and a positively charged region deeper inside the glass. This “built-in” charge creates an electric field that drives the ions backward in the process we have described as thermal relaxation ion spectroscopy, TRIS, where the temperature at which peak current occurs is characteristic of the ion that is moving at that temperature. TRIS allows the evaluation of the temperature at which an ion species becomes mobile and also enables the activation energies for field-assisted diffusion to be evaluated. Potassium has the lowest mobility of the three ions K, Na, and Ag in glass because of its high electronegativity and large ionic radius. Consequently, interactions between ions resulting from the presence of potassium impede the diffusion of both of the other ions, as confirmed in TRIS. This allows potassium to be used as a mask for defining the areas to be ion exchanged in a field-assisted process. Silver has the highest mobility in soda-lime glass because of its small ionic radius and high electronegativity.

A potential problem with the use of potassium as a mask for defining regions of silver diffusion for optical waveguide devices is the weakening of the glass at the surface, owing to the presence of stress gradients which may initiate cracking. One puzzle remains; we have no evidence that silver occupies the same sites in the glass as the other ions, only that silver is present in the surface region after the field-assisted ion exchange process as individual ions with displacement of the sodium ions.

The authors acknowledge the assistance of Dr. Tino Kausmann of Chemical Engineering, the University of Sydney, with scanning electron microscopy and Dr. Joonsup Lee of Sydney Analytical (Vibrational Spectroscopy), the University of Sydney, with FTIR-ATR. The Australian Research Council is thanked for financial support through the award of Grant Number LP160101322.

1.
G. Y.
Odo
,
L. N.
Nogueira
, and
C. M.
Lepienski
, “
Ionic migration effects on the mechanical properties of glass surfaces
,”
J. Non-Cryst. Solids
247
,
232
236
(
1999
)
2.
M. E.
Nordberg
,
E. L.
Mochel
,
H. M.
Garfinkel
, and
J. S.
Olcott
, “
Strengthening by ion exchange
,”
J. Am. Ceram. Soc.
47
,
215
219
(
1964
).
3.
B.
Sugarman
, “
Strength of glass (a review)
,”
J. Mater. Sci.
2
,
275
283
(
1967
).
4.
G. J.
Copley
, “
New developments in glasses
,”
Mater. Sci. Eng.
8
,
1
31
(
1971
).
5.
W. M.
Steyn
and
W. J.
de Wet
, “
Ion-exchange parameters of glass/molten salt systems
,”
J. Am. Ceram. Soc.
58
,
315
316
(
1975
).
6.
A.
Tervonen
,
S. K.
Honkanen
, and
B. R.
West
, “
Ion-exchanged glass waveguide technology: A review
,”
Opt. Eng.
50
,
071107
(
2011
).
7.
M. M.
Morsi
,
S.
El-Konsol
, and
M. I.
El-Shahawy
, “
Optical spectra of borate glasses containing Ti and Co in relation to their structure
,”
J. Non-Cryst. Solids
83
,
241
250
(
1986
).
8.
S.
Honkanen
,
P.
Pöyhönen
,
A.
Tervonen
, and
S. I.
Najafi
, “
Waveguide coupler for potassium- and silver-ion-exchanged waveguides in glass
,”
Appl. Opt.
32
,
2109
2111
(
1993
).
9.
J.
Saarinen
,
J.
Huttunen
,
S.
Honkanen
,
S. I.
Najafi
, and
J.
Turunen
, “
Computer-generated waveguide holograms by double-ion exchange process in glass
,”
Electron. Lett.
28
,
876
878
(
1992
).
10.
M. J.
Li
,
S.
Honkanen
,
W. J.
Wang
,
R.
Leonelli
,
J.
Albert
, and
S. I.
Najafi
, “
Potassium and silver ion-exchanged dual-core glass waveguides with gratings
,”
Appl. Phys. Lett.
58
,
2607
(
1991
).
11.
E.
Borsella
,
G.
Battaglin
,
M. A.
Garcìa
,
F.
Gonella
,
P.
Mazzoldi
,
R.
Polloni
, and
A.
Quaranta
, “
Structural incorporation of silver in soda-lime glass by the ion-exchange process: A photoluminescence spectroscopy study
,”
Appl. Phys. A
71
,
125
(
2000
).
12.
M. A.
Garcí
, “
Photoluminescence of silver in glassy matrices
,”
J. Appl. Phys.
96
,
3737
(
2004
).
13.
M. C. Ö.
Yanık
,
M.
Sarıgüzel
,
Y.
Öztürk
,
T.
Yıldız
,
E. D.
Kaçar
,
A.
İyiel
,
İzmirlioğlu
, and
E.
Günay
, “
Influence of different process conditions on mechanical, optical and surface properties of silver ion exchanged soda-lime silicate glass
,”
J. Non-Cryst. Solids
493
,
1
10
(
2018
).
14.
J. P.
Blondeau
and
O.
Véron
, “
Graded index waveguide elaboration by silver–sodium or silver–lithium ionic exchange for various glass compositions
,”
Opt. Mater.
34
,
278
286
(
2011
).
15.
D.
Guldiren
and
S.
Aydın
, “
Antimicrobial property of silver, silver-zinc and silver-copper incorporated soda lime glass prepared by ion exchange
,”
Mater. Sci. Eng. C
78
,
826
832
(
2017
).
16.
A. R.
Moussa
,
A. M.
El-kady
,
A. N.
Elboraey
,
D. Y.
Zaki
,
H. M.
Elgamily
,
M. I.
Ramzy
, and
G. T.
El-Bassyouni
, “
Antimicrobial properties of tissue conditioner containing silver doped bioactive glass nanoparticles: In vitro study
,”
Adv. Nat. Sci. Nanosci. Nanotechnol.
9
,
035003
(
2018
).
17.
I.
Sondi
and
B.
Salopek-Sondi
, “
Silver nanoparticles as antimicrobial agent: A case study on E. coli as a model for gram-negative bacteria
,”
J. Colloid Interface Sci.
275
,
177
182
(
2004
).
18.
J. M.
White
and
P. F.
Heidrich
, “
Optical waveguide refractive index profiles determined from measurement of mode indices: A simple analysis
,”
Appl. Opt.
15
,
151
(
1976
).
19.
K.
Chiang
, “
Construction of refractive-index profiles of planar dielectric waveguides from the distribution of effective indexes
,”
J. Lightwave Technol.
3
,
385
391
(
1985
).
20.
A.
Podlipensky
,
A.
Abdolvand
,
G.
Seifert
,
H.
Graener
,
O.
Deparis
, and
P. G.
Kazansky
, “
Dissolution of silver nanoparticles in glass through an intense dc electric field
,”
J. Phys. Chem. B
108
,
17699
17702
(
2004
).
21.
M. M.
Martins
,
L. R. P.
Kassab
,
D. M.
da Silva
, and
C. B.
de Araújo
, “
Tm3+ doped Bi2O3-GeO2 glasses with silver nanoparticles for optical amplifiers in the short-wave-infrared-region
,”
J. Alloys Compd.
772
,
58
63
(
2019
).
22.
G. V.
Rao
and
H. D.
Shashikala
, “
Optical, dielectric and mechanical properties of silver nanoparticle embedded calcium phosphate glass
,”
J. Non-Cryst. Solids
402
,
204
209
(
2014
).
23.
G. V.
Rao
and
H. D.
Shashikala
, “
Effect of heat treatment on optical, dielectric and mechanical properties of silver nanoparticle embedded CaO-CaF2-P2O5 glass
,”
J. Alloys Compd.
622
,
108
114
(
2015
).
24.
A.
Sonal Sharma
and
S.
Aggarwal
, “
Optical investigation of soda lime glass with buried silver nanoparticles synthesised by ion implantation
,”
J. Non-Cryst. Solids
485
,
57
65
(
2018
).
25.
T.
Izawa
and
H.
Nakagome
, “
Optical waveguide formed by electrically induced migration of ions in glass plates
,”
Appl. Phys. Lett.
21
,
584
586
(
1972
).
26.
C. R.
Lavers
,
B. J.
Ault
, and
J. S.
Wilkinson
, “
Characterization of secondary silver ion exchange in potassium-ion-exchanged glass waveguides
,”
J. Phys. D Appl. Phys.
27
,
235
(
1994
).
27.
N. H.
Ray
and
M. H.
Stacey
, “
Increasing the strength of glass by etching and ion-exchange
,”
J. Mater. Sci.
4
,
73
79
(
1969
).
28.
J.
Viljanen
and
M.
Leppihalme
, “
Fabrication of optical strip waveguides with nearly circular cross section by silver ion migration technique
,”
J. Appl. Phys.
51
,
3563
(
1980
).
29.
K.
Forrest
,
S.
Pagano
, and
W.
Viehmann
, “
Channel waveguides in glass via silver-sodium field-assisted ion exchange
,”
J. Lightwave Technol.
4
,
140
150
(
1986
).
30.
A.
Belkhir
, “
Detailed study of silver metallic film diffusion in a soda-lime glass substrate for optical waveguide fabrication
,”
Appl. Opt.
41
,
2888
2893
(
2002
).
31.
G. H.
Chartier
,
P. J. R.
Laybourn
, and
A.
Girodt
, “
Masking process for double-ion-exchanged glass optical waveguides
,”
Electron. Lett.
22
,
925
926
(
1986
).
32.
O.
Deparis
, “
Poling-assisted bleaching of soda-lime float glasses containing silver nanoparticles with a decreasing filling factor across the depth
,”
J. Appl. Phys.
100
,
044318
(
2006
).
33.
R. G.
Walker
and
C. D. W.
Wilkinson
, “
Integrated optical waveguiding structures made by silver ion-exchange in glass. 2: Directional coupler and bends
,”
Appl. Opt.
22
(
12
),
1929
1936
(
1983
).
34.
G.
Molina
,
S.
Murcia
,
J.
Molera
,
C.
Roldan
,
D.
Crespo
, and
T.
Pradell
, “
Color and dichroism of silver-stained glasses
,”
J. Nanopart. Res.
15
,
1932
(
2013
).
35.
A.
Chorfa
,
N.
Belkhir
,
A.
Belbali
,
F.
Rubio
, and
J.
Rubio
, “
Coloration and structure behavior after silver and copper nanoparticles formation in soda lime glass
,”
Mater. Sci. Eng. Technol.
48
,
1166
1172
(
2017
).
36.
Y.
Shioya
,
T.
Ohdaira
,
R.
Suzuki
,
Y.
Seino
, and
K.
Omoteb
, “
Effect of UV anneal on plasma CVD low-k film
,”
J. Non-Cryst. Solids
354
,
2973
2982
(
2008
).
37.
O.
Véron
,
J.-P.
Blondeau
,
D. D.
Meneses
, and
C. A.
Vignolle
, “
Characterization of silver or copper nanoparticles embedded in soda-lime glass after a staining process
,”
Surf. Coat. Technol.
227
,
48
57
(
2013
).
38.
M. D.
Ingram
,
J. E.
Davidson
,
A. M.
Coats
, and
E. I.
Kamitsos
, “
Origins of anomalous mixed-cation effects in ion-exchanged glasses
,”
Glass Sci. Technol.
73
(4), 89–104 (
2000
).
39.
M. L.
Braunger
,
C. A.
Escanhoela
, and
E. C.
Ziemath
, Jr.
, “
Electrical conductivity of Ag–Na ion exchanged soda-lime glass
,”
Solid State Ion.
265
,
55
60
(
2014
).
40.
L.
De Ferri
,
D.
Bersani
,
A.
Lorenzi
,
P. P.
Lottici
,
G.
Vezzalini
, and
G.
Simon
, “
Structural and vibrational characterization of medieval like glass samples
,”
J. Non-Cryst. Solids
358
,
814
819
(
2012
).
41.
S.
Amma
,
J.
Luo
,
C. G.
Pantano
, and
S. H.
Kim
, “
Specular reflectance (SR) and attenuated total reflectance (ATR) infrared (IR) spectroscopy of transparent flat glass surfaces: A case study for soda lime float glass
,”
J. Non-Cryst. Solids
428
,
189
196
(
2015
).
42.
D. E.
Carlson
, “
Ion depletion of glass at a blocking anode: I, theory and experimental results for alkali silicate glasses
,”
J. Am. Ceram. Soc.
57
,
291
294
(
1974
).
43.
J. F.
Shackelford
,
Y.-H.
Han
,
S.
Kim
, and
S.-H.
Kwon
,
Materials Science and Engineering Handbook
, 4th ed. (
Taylor & Francis Group, LLC
,
2015
), p.
15
.
44.
A. L.
Allred
, “
Electronegativity values from thermochemical data
,”
J. Inorg. Nucl. Chem.
17
,
215
221
(
1961
).
45.
D.
Guldiren
,
İ.
Erdem
, and
S.
Aydin
, “
Influence of silver and potassium ion exchange on physical and mechanical properties of soda lime glass
,”
J. Non-Cryst. Solids
441
,
1
9
(
2016
).
46.
A.
Talimian
,
G.
Mariotto
, and
V. M.
Sglavo
, “
Electric field-assisted ion exchange strengthening of borosilicate and soda lime silicate glass
,”
Int. J. Appl. Glass Sci.
8
,
291
300
(
2017
).
47.
J.
Zhao
,
J.
Lin
, and
W.
Zhang
, “
Effect of electric field-assisted diffusion on surface layer structure and mechanical properties of soda-lime glasses
,”
J. Wuhan University of Technol.y Mater. Sci. Ed.
30
,
452
457
(
2015
).
48.
J.
Luo
,
H.
He
,
N. J.
Podraza
,
L.
Qian
,
C. G.
Pantano
, and
S. H.
Kim
, “
Thermal poling of soda-lime silica glass with nonblocking electrodes—Part 1: Effects of sodium Ion migration and water ingress on glass surface structure
,”
Am. Ceram. Soc.
99
,
1221
1230
(
2016
).
49.
A.
Tandia
,
M.
Reveil
,
K. D.
Vargheese
,
J.
Luo
,
J. C.
Mauro
, and
P.
Clancy
, “
Modeling the thermal poling of glasses using molecular dynamics. Part 1: Effects on glass structure
,”
J. Non-Cryst. Solids
461
,
98
103
(
2017
).
50.
E.
Guo
and
D. R.
McKenzie
, “
A post Gurney quantum mechanical perspective on the electrolysis of water: Ion neutralization in solution
,”
Proc. R. Soc. A
473
,
2207
(
2017
).
51.
N.
Takamure
,
D. R.
McKenzie
,
M. M. M.
Bilek
, and
A.
Kondyurin
, “
Chemical toughening of glass by potassium diffusion: How non-bridging oxygen and a surface calcium barrier limit the process
,”
Ceram. Soc. Jpn.
127
,
98
104
(
2019
).
52.
J. C.
Dyre
,
P.
Maass
,
B.
Roling
, and
D. L.
Sidebottom
, “
Fundamental questions relating to ion conduction in disordered solids
,”
Rep. Progr. Phys.
72
,
046501
(
2009
).
53.
P.
Maass
,
A.
Bunde
, and
M. D.
Ingram
, “
Ion transport anomalies in glasses
,”
Phys. Rev. Lett.
68,
3064
(
1992
).