We present a novel technique that can rapidly and accurately measure surface tension and viscosity by direct thermal processing of an optical fiber. We demonstrate the applicability of this technique for a variety of glass compositions from silica to soft glass fibers, and these results have been validated against results obtained with other techniques. In addition, this characterisation technique has been used to measure the surface tension and viscosity for previously unmeasured glass compositions. The techniques are ideal for acquiring critical parameters of relevance to the conditions for the controlled fabrication of new glass compositions into microstructured fibers.
©2012 Optical Society of America
The surface tension, γ, is a crucial factor for glass fabrication and processing. It controls the fining and homogenization of glass melts and fire-polishing of glasses [1,2]. In addition, surface tension is a particularly critical parameter for the controlled fabrication and tapering of microstructured optical fibers [3,4], because it dictates the degree of microstructure collapse or expansion and only with their knowledge can the specific pressure required to actively control the internal structure be known .
There are many different techniques for measuring the surface tension of glass melts, including the ring [6,7], bubble pressure [8–11], fiber elongation , dipping cylinder methods  and drop shape (pendant drop and sessile drop) . For the sessile drop method in particular, the contamination between the substrate and melt can be an issue, and for some glass compositions a suitable substrate may not be available at all. The drop and ring methods suffer from the exposure of a large volume of molten glass to air for an extended period of time which can lead to significant evaporation. Methods such as the maximum bubble pressure technique suffer from the problem of some melts corroding the capillary tube, and of an overestimate of the surface tension when bubbles are not formed slowly enough.
We have developed a method which is a rapid, accurate, non-contact, direct technique and we demonstrate that it can be used to measure surface tension and viscosity of an optical fiber’s glass composition. Since our technique does not need sample preparation or involve any numerical analysis, we can quickly and accurately take multiple measurements of the surface tension for any given optical fiber. In addition we demonstrate the measurement of viscosity in situ on the same fiber optic and for the same temperature as that at which the surface tension is measured.
The technique uses a CO2 laser operating at 10.6 µm to heat a small volume of glass. The 10.6 µm radiation is readily absorbed in most glass compositions, with an absorption length as low as 4 µm in silica at 1800°C , enabling fast measurement times compared to conventional techniques in which several hours can be required to heat and cool large volume samples. It also enables the measurement of a full range of glass compositions, from silica based glasses with high temperature softening point, to heavy metal oxide based glasses with lower softening temperatures.
The technique we have developed is based on same fundamental principle as the fiber elongation technique, which has previously been reported to be inaccurate  and limited in its application to viscosities from 108 to 1013 dPa.s [12,16,17]. However, we demonstrate that our extension to this technique allows for the accurate measurement of surface tension to within ± 0.004 N/m, providing valuable information for the processing of optical glass compositions. We also show that the technique can be extended to yield values of the viscosity of the glasses at the same conditions.
In section 2 of this article we present a theoretical background to our techniques for measuring surface tension and viscosity and in section 3 we detail our experimental methods. The surface tension results of a range of glasses that have been used for optical fiber fabrication including silica, heavy metal oxide glasses and fluoride-containing glasses are presented in section 4 along with an experimental validation of our viscosity measurement technique through a comparison of results for fused silica.
2. Theoretical background of surface tension and viscosity measurements
2.1 Surface tension
Surface tension can be directly measured using a solid single-material optical fiber of radius r by considering the balance between the downward force Fw = mg exerted on the mass m, hanging below a heated volume of optical fiber and the upward force due to surface tension Fs = 2πγr as shown in Fig. 1 .
When the downwards force is greater than the force due to the surface tension, the fiber will taper. When the downwards force is less, the fiber length below the heated zone will be pulled upwards to form a glass sphere.
When the two forces are in equilibrium the following expression can be derived , for the surface tension:
The viscosity measurement technique is based on the same physical principles as that used in a standard fiber elongation viscometer measurements [16,17], except we show that it can be extended to include contraction of the fiber length due to surface tension pulling material up into a ball, thus allowing fast, accurate measurements of very low viscosities directly on an optical fiber.
Following the approach described in , where an entire fiber is heated in a furnace or dilatometer, the dynamic viscosity η can be expressed as:16,17].
We rewrite Eq. (2) in terms of the parameters measured using our technique:Fig. 1. The length of fiber being heated is approximated equal to the beam width of the laser, LI = LH≈2w, cross-section area is given by A = πr2 and the rate of elongation of the heated length is rewritten in terms of the rate of contraction of the length below the heated zone, dL/dt = dLI/dt.
A CO2 laser operating at 10.6 µm, beam quality M2 < 1.2 and maximum power of 10 watts is used to heat the fiber. The temperature is adjusted by changing the average power, measured using a thermal power meter. The temperature is measured using an optical pyrometer. In order to ensure minimal temperature fluctuations the laser is allowed to stabilize over 5 minutes at each required power level prior to use on the fiber.
All optical fibers measured are initially uncoated with the exception of the fused silica, which is mechanically stripped in addition to cleaning with high purity acetone and then isopropanol. The fiber is secured in a v-groove by a magnet, and suspended vertically. The CO2 laser beam is expanded by propagation over 6 meters, and then focused to a spot size, diameter 2·w, by a gold coated spherical mirror with radius of curvature 90 cm as shown in Fig. 2 . The beam diameter was measured to be 100 ± 5 µm using a razor blade mounted on a micrometer controlled stage, in both the x and y axis. Astigmatism is minimized by ensuring that the angle between the incident and reflected CO2 beams is less than 5 degrees.
A large diameter, long radius of curvature mirror was chosen to allow scanning the beam 10 cm up the length of the fiber while simultaneously maintaining the fiber within depth of focus of 1.5 mm, thus minimizing the change in temperature as the beam is scanned. An alternative approach would be to translate the stage across the beams path. Precise alignment onto the optical fiber is achieved using a co-propagating HeNe alignment laser.
The CO2 laser power incident on the fiber and the power measured behind the heated fiber is monitored during the experiments to ensure all power is absorbed, and that the beam is centered on the optical fiber. For all measurements of fibers with outer diameters of 125 µm and greater, the power behind the fiber was measured to be zero as expected for the small spot size of 100 µm and the high absorption at the CO2 wavelength of the glass compositions used.
3.1 Surface tension
To measure the surface tension, the CO2 laser beam is initially aligned to the bottom of the optical fiber, causing it to contract into a ball due to surface tension. The beam is scanned up the fiber until the heated zone tapers, causing the ball to drop off and the piece of glass that is detached is then collected and weighed with an electronic balance to within ± 0.01 mg. Care must be taken to select a suitable scanning speed for a given viscosity to ensure the taper is formed only at the point directly above the ball. This process is repeated about 10 times for each optical fiber sample and the mean and standard deviation of the masses are recorded to demonstrate the reproducibility and accuracy of the technique. The diameter of the fiber is also measured to ± 1 µm for each segment of fiber, using an optical microscope.
The entire process, from the laser scanning to collection and weighing of over 10 samples of the same glass takes under half an hour, making it a very rapid way of measuring the surface tension when the glass to be tested has already been made into an optical fiber.
To measure the viscosity, a travelling microscope is used to measure the initial fiber length L, and subsequently the change in length ΔL, after a time Δt, as illustrated in Fig. 1. These measurements are repeated for a range of temperatures achieved by varying the CO2 laser power for a fixed beam diameter of 100 ± 5 µm.
To measure the temperature of the optical fiber we built a disappearing filament optical pyrometer with telescopic imaging and red filter. As this is a comparative brightness technique it does not require the heated area to fill the full viewing area of the device, and is less sensitive to uncertainty in the emissivity of the object under examination. The uncertainty in our temperature measurement due to a 50% uncertainty in the emissivity of the optical fiber is 6%, for a temperature of 2500K and operating wavelength of 650nm. For each CO2 laser power, 6 measurements are taken with the optical pyrometer and the mean and standard deviation of the temperature recorded.
The values of the density used in Eq. (3) are from known sources at room temperature, since it refers only to the density ρ of glass suspended below the heated zone and the value of γ is used from that measured in the previous experiments.
3.3 Glass composition and fiber fabrication
The glasses used for the surface tension and viscosity measurements are listed in Table 1 . The composition of the commercial glasses from Schott Glass Co. (LLF1, F2, SF57, N-FK5, N-FK51A) were measured using energy dispersive x-ray spectroscopy and field emission scanning electron microscope from FIE Co. The compositions of the glasses fabricated in-house (Lead-Germanate, Tellurite, ZBLAN) are quoted as the nominal compositions of the raw material batches from which the glasses were prepared using the melt-quench technique.
For the Schott, Asahi and in-house glasses, the fiber preforms were prepared using the extrusion technique, and these preforms were drawn down to fiber using a soft glass drawing tower . For the silica glass, we used commercial SMF28 fiber from Corning.
4.1 Surface tension
The surface tension was measured using the above method for the glass compositions shown in Table 1 and the results are recorded in Table 2 . A comparison between our surface tension measurements and those measured independently for similar, or the same compositions are summarized in Fig. 3 . The surface tension was shown to be measurable to a precision of 0.004 N/m with our technique, with the statistical variations in the diameter of the fibers used making the largest contribution to the overall error.
The measured surface tension of silica (0.30 N/m) agrees with previously measured values using the fiber elongation and pendant drop methods [2,12]. The lower surface tension values (0.24-0.26 N/m) of the PbO- and Bi2O3-containing glasses (LLF1, F2, SF57, Bismuth, Lead-Germanate) compared with silica is consistent with previous results showing decreasing surface tension with increasing PbO and Bi2O3 content in silicate glasses [6,7,14,22].
The surface tension values of SF57 and F2 agree within the measurement error with the values measured using the fiber elongation method . The higher surface tension of SF57 compared with F2 and LLF1, despite higher PbO, we attribute to the lower content of potassium ions in SF57. Potassium ions were found to decrease the surface tension of silicate glasses significantly .
The tellurite glass shows significantly lower surface tension (0.16 N/m) than the heavy metal oxide glasses containing PbO and Bi2O3 (0.24-0.26 N/m). We attribute this to the higher content of the heavy metal oxide TeO2 and the absence of the lighter glass network former oxides SiO2 and GeO2.
The ZBLAN fluoride glass shows the lowest surface tension (0.09 N/m) of all glasses investigated. This result is consistent with the low surface tension value (0.17 N/m) measured for a ZBL fluoride glass .
The surface tension of the fluoride-phosphate glass N-FK51A (0.22 N/m) is similar to the surface tension values of alkaline earth phosphate glasses (0.22-0.25 N/m)  and the PbO- and Bi2O3 containing glasses of this work.
We attribute the lower surface tension of the fluoroborosilicate glass N-FK5 (0.24 N/m) compared with silica glass to presence of potassium, boron and fluoride ions in the glass. These three ions were found to decrease the surface tension of glass melts [1,2,22,23].
The significantly lower surface tension of tellurite and fluoride glasses compared with all the other glasses investigated is consistent with our observation that tellurite and fluoride glasses show significantly lower degree of hole closure during microstructured fiber drawing. For a ZBLAN fluoride glass preform with 7 holes, the hole size (relative to the outer fiber diameter) even increased during fiber drawing, which we attribute to the low surface tension of the ZBLAN glass.
In order to demonstrate the validity of our viscosity measurement technique, whereby we are measuring the contraction of the fiber rather than the elongation, we measure the viscosity of corning SMF28 optical fiber and compare it measured values in literature for fused silica .
The results in Fig. 4 demonstrate that our method is consistent with other measurement techniques for the viscosity of fused silica, down to a viscosity of ~103 dPa.s (1 dPa.s = 1 poise).
For a CO2 laser intensity greater than 1000 Wmm−2, the temperature approaches 3000 K and the vaporization rate of silica from the fiber increases substantially . Because our viscosity measurement utilizes the surface tension force, viscosities less than 103 dPa.s cannot be accurately measured for this glass composition as the surface is affected significantly by vaporization.
In order to determine the contribution to the uncertainty of our viscosity measurement from the increase in radius of the heated section of fiber, we measured Δt for 18 increments of ΔL = 1cm in a single trial on a SMF28e optical fiber. The results show that there is an 8% fluctuation in Δt, but no trend as the radius increases. This indicates that the uncertainty in the technique is not dominated by the change in radius, but primarily limited for our experimental setup by the 10% fluctuation of our CO2 laser power, which could be improved by the implementation of a feedback loop.
As viscosity is strongly dependent on temperature, the fundamental limitation to this technique is the inaccuracy of the temperature measurement due to the difficulties associated with accurately determining the emissivity of the fiber.
We have developed and tested an elegant technique utilizing a scanning CO2 laser that rapidly and accurately measures surface tension directly on an optical fiber, with precision of up to 0.004 N/m. We demonstrate the versatility and validity of this technique by measuring a range of glass compositions from silica to soft glass fibers. We confirm results obtained with other techniques and give values of surface tension for previously unreported compositions including ZBLAN, tellurite and lead germanate. In addition, we demonstrate the validity of our method for measuring viscosity of an optical fiber at identical conditions to those of our surface tension measurements.
Our technique is ideal for acquiring critical parameters at the exact conditions for the fabrication of new glass compositions in microstructured fibers. It provides valuable information for those involved with the drawing of microstructured optical fibers using new glass compositions, as well as those involved in post fabrication processes such as tapering.
We would like to thank Asahi Glass Co. for the supply of bismuth glass and wish to thank the members of the Institute for Photonics & Advanced Sensing at the University of Adelaide for their work in the fabrication of the optical fibres tested. T. Monro acknowledges the support of an Australian Research Council Federation Fellowship.
References and links
1. S. Fujino, C. Hwang, and K. Morinaga, “Surface tension of PbO-B2O3 and Bi2O3-B2O3 glass melts,” J. Mater. Sci. 40(9-10), 2207–2212 (2005). [CrossRef]
2. W. D. Kingery, “Surface tension of some liquid oxides and their temperature coefficients,” J. Am. Ceram. Soc. 42(1), 6–10 (1959). [CrossRef]
3. C. J. Voyce, A. D. Fitt, and T. M. Monro, “Mathematical model of the spinning of microstructured fibres,” Opt. Express 12(23), 5810–5820 (2004), http://www.opticsinfobase.org/oe/abstract.cfm?URI=oe-12-23-5810. [CrossRef] [PubMed]
4. C. J. Voyce, A. D. Fitt, J. R. Hayes, and T. M. Monro, “Mathematical modeling of the self-pressurizing mechanism for microstructured fiber drawing,” J. Lightwave Technol. 27(7), 871–878 (2009). [CrossRef]
6. C. Hwang, B. K. Ryu, and S. Fujino, “Surface tension of bismuth borosilicate melts,” Thermochim. Acta 531, 70–74 (2012). [CrossRef]
7. S. Fujino, C. Hwang, and K. Morinaga, “Density, surface tension and viscosity of PbO/B2O3-SiO2 glass melts,” J. Am. Ceram. Soc. 87(1), 10–16 (2004). [CrossRef]
8. M. Yamashita, M. Suzuki, and H. Yamanaka, “Surface tension measurement of glass melts by maximum bubble pressure method,” Glastech. Ber. 73, 337–343 (2000).
9. A. E. Badger, C. W. Parmelee, and A. E. Williams, “Surface tension of various molten glasses,” J. Am. Ceram. Soc. 20(1-12), 325–329 (1937). [CrossRef]
10. C. A. Bradley, “Measurement of surface tension of viscous liquids,” J. Am. Ceram. Soc. 21(10), 339–344 (1938). [CrossRef]
11. S. Akhtar and M. Cable, “Some effects of atmosphere and minor constituents on the surface tension of glass melts,” Glass. Technol. 9, 145–151 (1968).
12. N. M. Parikh, “Effect of atmosphere on surface tension of glass,” J. Am. Ceram. Soc. 41(1), 18–22 (1958). [CrossRef]
13. L. Shartsis, S. Spinner, and A. W. Smock, “Surface tension of compositions in the systems PbO-B2O3 and PbO-SiO2,” J. Am. Ceram. Soc. 31(1), 23–27 (1948). [CrossRef]
14. L. D. Pye, A. Montenero, and I. Joseph, Properties of Glass-Forming Melts (CRC Press, 2005), Chap. 5.
16. H. R. Lillie, “Viscosity of glass between the strain point and melting temperature,” J. Am. Ceram. Soc. 14(7), 502–512 (1931). [CrossRef]
17. E. L. Bourhis, Glass (Wiley-VCH, 2008), Chap. 6.
18. C. A. G. Kalnins, H. Ebendorff-Heidepriem, N. A. Spooner, and T. M. Monro, “Radiation dosimetry using optically stimulated luminescence in fluoride phosphate optical fibres,” Opt. Mater. Express 2(1), 62–70 (2012). [CrossRef]
19. J. Stoetzel, “Fabrication of optical glass fibres by extrusion,” internship report (Otto Schott Institute at the University of Jena (Germany) and Institute for Photonics & Advanced Sensing at the University of Adelaide, 2011).
20. L. Shartsis and A. W. Smock, “Surface tension of some optical glasses,” J. Am. Ceram. Soc. 30(4), 130–136 (1947). [CrossRef]
21. S. Toyoda, S. Fujino, and K. Morinaga, “Density, viscosity and surface tension of 50RO–50P2O5 (R: Mg, Ca, Sr, Ba, and Zn) glass melts,” J. Non-Cryst. Solids 321(3), 169–174 (2003). [CrossRef]
22. N. P. Bansal and R. H. Doremus, Handbook of Glass Properties (Academic Press, 1986), Chap. 5.
23. N. P. Bansal and R. H. Doremus, “Surface tension of ZrF4-BaF2-LaF3 glass,” J. Am. Ceram. Soc. 67(10), C-197 (1984). [CrossRef]
24. G. Urbain, Y. Bottinga, and P. Richet, “Viscosity of liquid silica, silicates and alumino-silicates,” Geochim. Cosmochim. Acta 46(6), 1061–1072 (1982). [CrossRef]
25. H. L. Schick, “A thermodynamic analysis of the high-temperature vaporization properties of silica,” Chem. Rev. 60(4), 331–362 (1960). [CrossRef]