Owens Corning, Science and Technology Center, Granville, Ohio, USA
Glastechnische Berichte 43, 46-55 (2000)
Copyright © 2000 Russell M. Potter
A new method for measuring in-vitro fiber dissolution rate in physiological saline solutions is based on direct optical measurement of fiber diameter decrease. Dissolution times measured by this new method for a variety of vitreous silicate fibers with a wide range of dissolution rates are in good agreement with results from SEM and from traditional mass loss and solution analysis techniques. The new method is conceptually and experimentally simple, and it can be applied directly to fibers pulled from the original fiber sample with no pre-treatment.
For about the last 20 years there has been interest in the dissolution rate of fibers in model physiological fluids thought to simulate the environment in the lung. This interest has intensified in recent years as the link between a fiber's durability and its association with lung disease has been clearly identified [1 to 4]. Throughout this time a number of different in-vitro laboratory techniques have been developed and used to quantify fiber dissolution [for example, 5 to 21]. These differ in the pH and chemical composition of the model physiological fluid, in the flow rate of fluid past the fibers, in the nature of the measurements used to assess fiber dissolution rate, and in the way in which the dissolution rate is calculated and expressed. There is general consensus that flow-through techniques which minimize alteration of the fluid by the dissolving fibers model most accurately the environment in vivo . The most widely used analytical method has been chemical analysis of the fluid for one or more of the dissolving components. The mass loss is calculated from these data, and a dissolution rate constant is derived from the mass loss. The dissolution rate is most commonly expressed as kdis, the mass loss in units of ng/cm2/hr. It is also expressed as a dissolution velocity in various units of length per time.
Despite the success of the various in-vitro methods in measuring fiber durability, there are a number of fundamental and practical problems with the available methods that have prevented a general acceptance of any one as a universal standard. Besides the overriding requirement of good correlation with in-vivo data, such a standard in-vitro fiber dissolution rate measurement technique would have the following attributes:
This paper describes an in-vitro technique based on direct optical measurement of fiber diameter decrease which has been developed with these desirable criteria in mind.
The essential difference between the optical technique and other currently-used flow-through methods for measurement of fiber dissolution rate is in how the dissolution process is measured, not the nature of the physiological fluid or the supporting equipment to maintain constant fluid flow and temperature. The optical method requires an appropriately-equipped optical microscope and suitably-designed sample holders. These holders can then be used with an existing flow-through system without essential change.
For the work reported here the physiological saline solution is a modified Gamble’s solution having a pH near 7.4. The sample holders were maintained at 37 oC, the sample flow rates were generally in the range 0.2-0.3 ml/min. Details of the flow-through system have been published by Mattson .
Figure 1: A cross-sectional view of the fibers (C) and mount (A) in place in the cell (B) and a planar view of the mount without fibers.
The sample holders for the optical technique consist of a cylindrical mount and cell (Figure 1). The fibers are attached to the mount with epoxy so that they span the slot. The mount is enclosed in the cell, which is held together by 4 bolts and sealed with O-rings at top and bottom. During a run, 1.5 cm long pieces of rubber tubing are attached to the lower bolt heads on the outflow side of the cell to tilt it so that bubbles are removed.
The measurement device is an optical microscope with a 40x, 8 mm focal length, 0.5 numerical aperture objective and a 16X digital filar ocular. This provides direct observation of the fiber during measurement, which is essential to obtain accurate, reproducible fiber diameters. Standard video measuring equipment gave diameter measurements which varied unreproducibly with brightness of the microscope field of view, video gain, and fiber characteristics.
Table 1 contains details of the fibers included in this study, which were produced, free of binder or other organic coatings, by a variety of commercial or laboratory processes. The chemical compositions were measured using standard techniques (X-ray fluorescence, inductively-coupled plasma, atomic absorption, and various wet chemical techniques) on either the bulk glass from which the fibers were made or on the fibers themselves. In the case of a few samples, the reported analysis is that of a different fiber sample formed from the same bulk glass. Several fiber compositions in Table 1, which are well-known in the fiber dissolution literature, are identified by name in the table. These are not necessarily samples of the identical material studied elsewhere.
Table 1: Source and chemical compositional data for the fiber samples
|oxide chemical composition|
|sample number||sample name||fiber source||pre processing||analysis source||SiO2||Al2O3||CaO||MgO||Na2O||K2O||B2O3||FeO||Fe2O3||TiO2||SO3T||SrO||BaO||F||P2O5||MnO||ZrO2|
|1||7753||laboratory continuous fiber||none||g||66.03||0.63||5.90||3.08||15.64||0.81||7.66||-||0.35||0.12||-||-||-||-||-||-||-|
|2||B-01||laboratory continuous fiber||none||g||61.33||0.11||16.28||3.44||15.12||0.41||3.00||-||0.23||0.11||-||-||-||-||-||-||-|
|3||MMVF11||laboratory continuous fiber||none||g||63.64||3.95||7.36||2.93||15.71||1.38||4.28||-||0.32||-||0.18||-||-||-||-||-||-|
|4||MMVF10||laboratory continuous fiber||none||g||57.15||5.43||7.82||4.25||15.11||1.07||8.63||-||0.10||-||-||-||-||0.76||-||-||-|
|5||MMVF10||production glass wool||respirable separate||f||57.50||5.10||7.50||4.13||14.95||1.06||8.75||-||0.07||0.01||0.12||-||-||0.83||-||-||-|
|6||MMVF10a||production glass wool||respirable separate||f||57.21||5.10||7.17||4.48||15.64||1.04||8.40||-||0.05||0.01||0.00||0.00||0.01||0.36||-||-||0.02|
|7||NK8340||laboratory continuous fiber||none||g||64.93||2.12||9.10||3.08||14.59||0.51||4.63||0.12||0.37||-||0.14||-||-||-||-||-||-|
|8||NK8340||production glass wool||blended||f||64.92||2.12||9.08||3.03||14.56||0.51||4.66||0.12||0.38||-||0.15||0.10||-||-||-||-||-|
|9||-||laboratory continuous fiber||none||g||66.49||0.60||8.43||2.82||15.33||0.12||5.48||0.17||0.22||-||0.15||-||-||-||-||-||-|
|10||-||production glass wool||blended||f||66.61||0.61||8.41||2.77||15.48||0.12||5.44||0.19||0.20||-||0.16||-||-||-||-||-||-|
|11||-||laboratory continuous fiber||none||s||66.51||0.66||8.95||2.87||14.94||0.32||5.36||-||0.13||-||-||-||0.11||-||-||-||-|
|12||-||laboratory glass wool||none||f||66.51||0.66||8.95||2.87||14.94||0.32||5.36||-||0.13||-||-||-||0.11||-||-||-||-|
|13||-||laboratory continuous fiber||none||s||62.78||0.88||8.69||2.57||14.33||0.40||9.94||-||0.17||-||-||-||-||-||-||-||-|
|14||-||laboratory continuous fiber||none||s||62.78||0.88||8.69||2.57||14.33||0.40||9.94||-||0.17||-||-||-||-||-||-||-||-|
|15||-||laboratory continuous fiber||none||s||62.78||0.88||8.69||2.57||14.33||0.40||9.94||-||0.17||-||-||-||-||-||-||-||-|
|16||-||laboratory continuous fiber||none||s||62.78||0.88||8.69||2.57||14.33||0.40||9.94||-||0.17||-||-||-||-||-||-||-||-|
|17||-||laboratory continuous fiber||none||s||62.78||0.88||8.69||2.57||14.33||0.40||9.94||-||0.17||-||-||-||-||-||-||-||-|
|18||-||laboratory glass wool||none||f||62.78||0.88||8.69||2.57||14.33||0.40||9.94||-||0.17||-||-||-||-||-||-||-||-|
|19||-||laboratory glass wool||none||f||62.60||0.93||8.71||2.58||14.27||0.40||9.97||-||0.13||-||-||-||-||-||-||-||-|
|20||-||laboratory continuous fiber||none||g||63.29||1.05||9.12||2.54||14.28||0.51||8.74||-||0.13||-||-||-||0.15||-||-||-||-|
|21||-||laboratory glass wool||none||f||63.49||1.03||9.12||2.59||14.09||0.49||8.47||-||0.22||-||0.10||-||0.14||-||-||-||-|
|22||-||laboratory continuous fiber||none||f||47.35||3.19||6.66||2.60||19.06||0.01||19.18||-||0.03||0.01||0.02||0.02||-||3.09||-||-||-|
|23||-||laboratory continuous fiber||none||f||53.17||5.09||8.82||4.99||20.62||0.57||6.07||-||0.18||0.07||0.26||0.01||-||-||-||-||-|
|24||-||laboratory continuous fiber||none||g||50.04||8.08||11.25||1.52||4.94||0.07||23.35||-||-||-||-||0.49||-||-||-||-||-|
|25||-||laboratory continuous fiber||none||g||56.64||1.10||7.61||2.72||14.62||0.82||16.25||0.04||0.25||0.13||-||-||-||-||-||-||-|
|26||-||laboratory continuous fiber||none||g||56.56||1.07||7.40||2.65||14.22||0.78||12.91||0.06||0.19||0.13||-||-||4.15||-||-||-||-|
|27||-||laboratory continuous fiber||none||g||56.47||1.08||7.15||2.55||13.92||0.74||9.54||0.06||0.20||0.12||-||-||8.10||-||-||-||-|
|28||-||laboratory continuous fiber||none||f||47.31||7.40||12.25||2.09||5.91||0.80||23.04||0.11||0.10||-||-||0.54||-||-||-||-||-|
|29||-||laboratory rock wool||none||f||40.94||21.74||16.41||11.62||0.83||0.91||-||4.59||1.06||1.71||-||-||-||-||-||-||-|
|30||607||production rock wool||respirable separate||f||58.30||1.25||38.70||0.40||0.30||0.10||-||-||0.10||0.05||-||-||-||-||0.05||-||-|
|31||MMVF34||production rock wool||shot removed||f||38.54||22.95||15.18||10.22||1.69||0.78||-||5.66||0.97||1.95||-||-||-||-||0.37||0.31||-|
|32||MMVF34||production rock wool||respirable separate||f||38.84||22.99||15.13||10.41||1.68||0.81||-||5.28||1.60||1.96||0.08||0.054||0.04||-||0.41||0.32||0.059|
|33||FI980889||production rock wool||none||s||45.59||14.92||14.28||11.10||2.01||1.03||-||7.07||0.58||1.59||0.35||-||-||-||0.40||0.18||-|
|34||FI980890||production rock wool||none||s||39.88||20.61||13.54||11.49||1.57||0.83||-||6.64||0.59||1.63||0.29||-||-||-||0.34||0.17||-|
|35||QFHA19||production rock wool||shot removed||f||40.14||18.91||19.12||11.57||1.77||0.45||-||4.62||0.81||1.58||-||-||-||-||-||0.16||-|
|36||QFHA22||production rock wool||blended||f||41.62||21.62||15.56||9.83||1.72||0.43||-||5.54||1.15||1.65||-||-||-||-||-||0.16||-|
|37||QFHA23||production rock wool||blended||f||39.53||21.92||16.76||11.01||1.94||0.43||-||4.95||1.36||1.64||-||-||-||-||-||0.17||-|
|38||QFHA25||production rock wool||blended||f||38.26||22.73||25.19||6.58||0.74||0.39||-||3.38||0.64||1.17||0.60||-||-||-||-||0.17||-|
|39||MMVF22||production slag wool||none||f||38.11||10.97||37.86||10.41||0.40||0.47||-||-||0.37||0.45||1.25||-||-||-||-||0.68||-|
|40||MMVF22||production slag wool||respirable separate||f||38.35||10.55||37.50||9.90||0.38||0.45||-||-||0.30||0.45||1.81||0.05||0.04||-||-||0.70||0.06|
|41||1260C||production high-temp wool||none||f||76.43||1.59||0.40||20.39||0.16||0.22||0.08||0.28||0.12||-||-||-||-||-||-||-||-|
|42||-||production high-temp wool||none||f||63.99||0.56||16.39||13.45||0.08||0.08||-||-||0.16||-||-||-||-||-||-||-||5.14|
Chemical analysis source:
f - on the fiber sample;
g - on the bulk glass from which fiber was formed;
s - on a different fiber sample formed from the same bulk glass
Total sulfur is reported as SO2T
If no value is reported for FeO, total iron is reported as Fe2O3.
Analysis on no. 30 from reference 4.
Binder or other organic coatings such as oil can, if desired, be removed by low-temperature ashing in oxygen plasma. Removal by heat-treatment may change the annealing state of the fibers and should be avoided.
Fibers can generally be pulled by hand directly from the original sample and mounted individually. Typically these fibers will be larger in diameter than the average of the original sample since the larger diameter fibers tend to be longer and easier to see and to work with. We have not tried to pull fibers representative of the average fiber diameter, although this could be done. The fibers are mounted by applying a layer of epoxy resin to the top surface of the mount on either side of the slot and placing 8 to 10 individual fibers at about equal intervals along the slot. Each fiber should span the slot and be immersed in the epoxy on each side.
For samples which have been processed for shot removal, to isolate a respirable fraction, or for some other reason, it may not be possible to handle fibers individually due to their short length or small diameter. In this case about 2 mg of fibers are dispersed in about 5 ml of deionized water by ultrasonic treatment. The fibers are filtered onto a 45 mm diameter membrane filter and allowed to dry thoroughly. A very thin layer of epoxy resin is applied to the top surface of the mount on either side of the slot and allowed to cure to a tacky stage. The mount is turned upside down and pressed gently onto the fibers on the filter. The filter is then gently peeled off and discarded. This will leave many fibers adhered to the epoxy and sticking out over the slot in the mount. After the epoxy has cured, a binocular microscope is used to identify 10-12 fibers which are good candidates for measurement due to their smooth, uniform diameter, good adherence to the mount, and clear visibility in the slot. These fibers are marked by a small spot of colored nail polish as close as possible to them on the mount.
The mount is loaded into the cell as shown in Figure 1 and filled by syringe with simulated lung fluid at room temperature. The initial diameter of each fiber is measured using the diameter measurement procedure given below. The flow lines are connected to the cell, the tubing feet are placed on the bolt heads on the cell bottom at the outflow side of the cell, and the cell is placed into the constant temperature bath.
Typically, a production fiber can change diameter significantly even within a single microscope field of view (170 µm). It is therefore essential that successive diameter measurements be made at the same place on any given fiber. This is most easily accomplished by measuring the diameter a set distance (1/2 the field of view, for example) from the edge of the slot. We have generally found no reduction in dissolution rate at the slot edge even though a stagnant layer might be expected there, but it is good practice to check for this during a run.
The ideal image for measurement depends to some degree on the fiber diameter and whether the dissolution is congruent or incongruent. It probably also depends on the particular microscope design. Our experience indicates the following guidelines:
Congruent dissolution: When the focal plane is just below the optimum position for measurement, the fiber appears dark at the edges, grading to light in the center. As the focal plane is raised, the lighter, central portion of the fiber expands, and two dark bands appear along the edges yielding the image shown in Figure 2A. Often, particularly with smaller diameter fibers, there is a blue tinge outside of these bands. As the focal plane is raised further, the two bands lighten and become tinged with yellow. For fibers greater than about 4 µm in diameter, the image diameter is the same throughout this change in focal plane, but for fibers less that about 4 µm in diameter, the image diameter changes. The true fiber diameter is always the distance between the outer edges of the dark bands as shown in Figures 2A and 2B just before they become tinged with yellow. When the fiber diameter is smaller than about 2 µm, it may yield a sharply defined image like those of Figures 2A and 2B, but often the dark bands are absent, and the best image is that shown in Figure 2C, in which the fiber is uniformly dark.
Figure 2: Micrographs typical of fibers for diameter measurement:
A - congruently-dissolving fibers about 10 µm in diameter;
B - congruently-dissolving fibers about 4 µm in diameter;
C - congruently-dissolving fibers about 2 µm in diameter;
D - incongruently-dissolving fibers showing the unaltered core surrounded by a leached layer.
Incongruent dissolution: We have been able to detect and measure a leached-layer on an unaltered fiber core only if the layer is thicker than about 1 µm. At some positions of the focal plane most fibers show structures which appear to be leached layers but are not. It is good practice, when fibers are first immersed in the solution, to observe several as the focal plan is raised and lowered to identify structures which could be mistaken for a leached layer. The appearance of leached layers varies, depending on how much material has been leached from the layer, from barely distinguishable from the fluid to barely distinguishable from the unaltered core. The image in Figure 2D is of a highly leached layer which tends toward the former. The outer leached diameter is measured in a way similar to that of congruent fibers - by raising the focal plane through that level which makes the outer part of the fiber as dark as possible until it just begins to lighten. The true diameter of the leached-layer is then the distance between the outer edges of the darker area. The best image for measuring the core is when it first presents a clearly defined boundary with the leached-layer as the focal plane is raised. If the leached-layer is highly leached, it may never be dark enough to obscure the core. This is the case in Figure 2D, which shows a good image for measuring both the leached-layer and the core. If the leached-layer is only slightly leached, the core may not appear until the focal plane is raised further after measuring the leached-layer. The lower relief of the leached material, particularly after complete leaching, is good confirmation that a leached layer was measured rather than some other structure.
Typically, it is not necessary to plan a set measurement schedule, but rather to start with daily measurements and then lengthen the time interval as necessary until there is a 0.5 to 1.0 µm difference in diameter between subsequent measurements (or less if the initial diameters of the fibers are near 2 µm). This interval is then continued until the fibers are completely dissolved, too many fibers are broken or unmeasurable in some way, or the fibers have been measured at enough times to define the dissolution rate with confidence. The data do not define a clear, consistent point (such as a diameter decrease of 2 µm, for example) at which such confidence is generally possible, so it is advisable to follow the dissolution as far as time constraints allow.
The cooling of the cell during the measuring process sets an upper limit to the frequency at which diameter measurements can be made without introducing significant error into the calculated dissolution rate due to time at reduced temperature. Cell temperature measurements during the fiber diameter measurement and subsequent reheating of the cell in the water bath together with published data on the effect of temperature on fiber dissolution rate  allow an estimate of this error. For a typical measurement time of 20 minutes, 3% error in the dissolution rate corresponds to about three measurements per day, which is not a serious limitation even for the most rapidly dissolving fibers.
At the end of a run, if SEM data are desired to confirm the optical measurements or to define the thickness of a leached layer, the cell should be taken apart while completely immersed in water, and the mount lifted vertically through the water surface to minimize stress on the fibers. A few drops of ethanol put on the water surface near the fibers just before they break the surface will also help prevent fiber breakage by reducing surface tension. For SEM examination the mount should be embedded in epoxy and a polished cross-section made as nearly as possible at the point the fibers were measured optically.
In this paper congruent fiber dissolution rate is expressed as dissolution time, tdis, the time in days to dissolve a 1 µm diameter fiber. This parameter has the advantage of expressing directly the biologically relevant aspect of dissolution . Furthermore, tdis can be calculated from the optical data by a simple equation which depends on no other properties of the fiber or fiber sample:
The slope, m, is obtained by least-squares' analysis of the change of d with time, and tdis is calculated from the slope. The least squares' analysis does not force the line to pass through the origin since the initial measurement is not necessarily more accurate than any subsequent one.
For incongruently-dissolving fibers it is not yet known whether biopersistence depends on leaching rate only, on total dissolution rate only, or on some combination of both. For this reason it is useful to define a leaching time, tl, the time in days for a 1 µm diameter fiber to be completely leached, and a total dissolution time, tn,, the time in days for a 1 µm diameter to be completely dissolved. They are given by equations similar to Equation 1, where d is the diameter of the unaltered core for tl, and d is the outer diameter of the leached layer for tn. They are evaluated in a manner analogous to tdis.
For congruently dissolving fibers, kdis in ng/cm2/hr is approximately equal to 5000/tdis . The exact relationships between kdis and the dissolution times are:
for congruently-dissolving fibers. And:
for incongruently-dissolving fibers. The factor 2 in the denominator of Equations 2 and 3 reflects the fact that the decrease in fiber diameter is twice the dissolution velocity. I have defined the mass fraction leached, Xl, equal to 0 for congruent dissolution. By this convention congruent dissolution is a special case of incongruent dissolution with tdis equated to tn. Generally, for incongruent dissolution, Xl will not be known, and kdis can be only estimated from the diameter decrease. For many compositions the mass fraction of SiO2+Al2O3 is a rough upper limit to Xl.
Table 2: Dissolution time data for the fiber samples
||-||-||-||-||optical method||mass loss method||solution analysis method||SEM method|
|sample number||sample name||initial fiber diameter||density||tdis||tn||tl||tdis||tn||tl||tdis||tn||tl||tdis||tn||tl|
Table 2 summarizes the results of tdis measurement by the optical method and by the methods of mass loss, solution analysis, and SEM fiber diameter measurement. The later three methods have been described elsewhere [10,13]. For the mass loss and solution analysis methods, values for tn and tl are calculated only for those samples for which the rate of mass loss with time (measured directly or calculated from solution analysis) show a clear break attributable to complete leaching of the fibers. Only when this break is apparent in the data, is it possible to determine unambiguously both the leaching and total dissolution. Thus, for example, a single, congruent dissolution rate is reported from the mass loss and solution analysis methods for sample no.4 even though SEM analysis shows the dissolution to be incongruent. The initial fiber diameter data in Table 2 are for the mounted fibers only and do not necessarily reflect the average fiber diameter of the sample.
The data in Table 2 indicate that the standard error associated with a linear fit to the data for a single measurement of tdis by the optical method is typically less than about 10% of the measured value. Most of the instances of larger standard error are the result of small initial fiber diameter, which limits the number of measured points and the accuracy of each measurement. For incongruently dissolving fibers, the standard error in both tn and tl is typically less than about 15%, which reflects the greater complexity of the optical image. Thus, the model of constant diameter decrease with time is in general a good one for both congruent and incongruent fiber dissolution.
Figure 3: Typical optical data of fiber diameter decrease (d) with time. The three samples shown span the range of dissolution times measured.
Although the dissolution process is essentially a linear diameter decrease with time, there are some departures from strict linearity which may be significant to details of the dissolution process. Figure 3 shows diameter decrease data for a range of tdis values together with the linear fits to the data by least-squares' analysis. The dissolution of samples no. 2 and no. 30 are strictly linear, which is the case for the majority of fibers. Some fiber samples, such as no. 31, show an initial period of slower dissolution. This may relate to the presence of a thin surface layer on the fibers from which the more volatile components have been partially removed during the fiberization process as discussed by Bauer  and by Mattson . Typically, this is a small part of the dissolution process both in terms of time and in terms of the amount of fiber dissolved. Some fibers, such as no. 8, show an initial period of more rapid dissolution. We have found this primarily with production mineral wool samples. This initial rapid dissolution can last for weeks through a diameter decrease of several micrometers, so it is unlikely to be related to a surface effect. It may relate to leached layers which are not recognizable optically.
Table 2 contains data for replicate runs for a number of samples and the standard deviation of the dissolution times of these replicate runs expressed as a percent of the mean value. This standard deviation is typically less than about 20% for both congruent and incongruent dissolution. Higher values are the result of too few diameter measurements because the dissolution process was not followed far enough (a total diameter decrease of 0.4 µm for the second run of sample no. 33 and 0.8 µm for the second run of no. 29) or was not measured frequently enough (only one measurement of unaltered core diameter for each of the runs of sample no. 22).
Figure 4: The correlation between optical and SEM diameter measurements of the same fibers. The line is a least squares' fit.
In order to check the accuracy of the optical fiber diameter measurements, we measured the final fiber diameters by SEM for a number of completed runs. In all cases the fibers are from a continuous laboratory bushing so that a small difference in the measurement position along the length of the fiber is not significant. Figure 4 shows a good correlation between the two diameter measurement techniques. The slope of the least squares' fit line is 1.00 with a slight offset indicating that the optical technique tends to yield a slightly larger diameter. This offset is small and, in any case, can have no effect on the dissolution rate calculated by the optical technique since the rate is based on diameter change.
Sample no. 25 was used to examine the effect of flow rate since its very rapid incongruent dissolution should make it particularly sensitive to flow rate changes. The dissolution rate is significantly reduced under stagnant conditions but is independent of flow in the range studied and reported here, 0.06 to 0.40ml/min. For convenience, the standard flow rate was set at 0.25 ml/min; it could certainly be reduced.
4.5 Effect of fiber diameter
The optical technique yields more accurate and reproducible dissolution rates when the initial fiber diameters are greater than about 2 µm since the measured fiber diameter changes can be larger and more points can be measured. This is also true of the mass loss and solution analysis methods due to poor dispersion of fine fibers in the mats [10,13]. It is therefore of general concern that measurements on larger diameter fibers be applicable to the dissolution of respirable fibers, which have diameters typically less than about 2 µm.
In most of the production fiber samples, the individual fibers measured by the optical technique span a large range of fiber diameters. In over half of these cases the dissolution times calculated from the individual fibers show a rough trend with dissolution time increasing with decreasing initial fiber diameter, particularly at diameters near 2 µm and below. The increase in dissolution time can be as much as a factor of 3. In addition, an increase in measured dissolution times sometimes occurs, regardless of initial fiber diameter, as individual fibers dissolve to diameters near 2 µm and below. This is true for both laboratory continuous fibers and for wool fibers. Fibers that have dissolved to diameters near 2 µm sometimes remain visible significantly beyond the time at which their earlier diameter change would predict complete congruent dissolution. The only evidence for a decrease in dissolution time with fiber diameter decrease is the larger dissolution time of sample no. 4 relative to sample nos. 5 and 6.
In general, one might expect that if there were any diameter effect, dissolution times would decrease rather than increase with fiber diameter since finer fibers would be cooled more rapidly and have, in consequence, more open structures. It is particularly hard to concieve of a process that would cause the small central portion of a large, congruently-dissolving fiber to dissolve more slowly than the rest of the fiber. The apparent correlation of dissolution rate to fiber diameter cannot be the result of systematic innacuracies of the optical diameter measurement below about 2 µm. Figure 4 shows good agreement between optical and SEM diameter measurements. In addition, fibers sometimes remain visible long beyond the time of complete dissolution predicted from their earlier rate of diameter reduction.
To further examine the effect of diameter, continuous fibers samples (nos.13-17) were produced at various diameters between 2.7 and 10.3 µm from a composition for which laboratory wool samples (no. 18 and no. 19) show a large effect of fiber diameter. The dissolution times measured by the optical method suggest little, if any, increase in dissolution rate for the smaller diameter fiber samples. Potential effects of fiber diameter on dissolution time remains an important, unresolved issue and the subject of continued investigation using the optical technique and other in-vitro methods.
Table 1 contains a number of sets of fibers which are essentialy the same composition but differ in the forming process or sample pre-processing (nos. 4-21). These samples provide some idea of how such differences can affect the dissolution time of a given composition. For the optical technique, the variations within a compositional set are not significantly greater than the typical reproducibility of the measurement except for cases which include samples with an average initial fiber diameter of 2-3 µm. The large variation in these cases may relate, in part at least, to the limited number of measurements possible for the small diameter samples and the large uncertainty of each measurement. The mass loss and solution analysis data are more limited than the optical data, but the average variation for all the sets of similar compositions is about the same.
For congruently-dissolving fibers, Figure 5 compares the dissolution times measured by the optical method with those measured by mass loss, solution analysis, and SEM. The agreement is good throughout the range. As expected from the SEM-optical comparison, dissolution times measured by the optical technique are within experimental error the same as those measured by SEM, falling close to the line indicating perfect agreement between the techniques. In almost all cases the dissolution times from the optical technique are less than or equal to those measured by mass loss or solution analysis. This probably relates to the isolated dissolution environment, which is characteristic of the optical and SEM methods, but which is difficult or impossible to achieve with the mass loss or solution analysis techniques.
Figure 5: Comparison of dissolution times for congruently dissolving fibers measured by the optical method to those measured by traditional methods. The line indicates exact agreement between the methods.
Figure 6: Comparison of dissolution times for incongruently dissolving fibers measured by the optical method to those measured by traditional methods. The line indicates exact agreement between the methods.
For incongruently-dissolving fibers, Figure 6 compares the total dissolution and leaching times measured by the optical method with those measured by mass loss, solution analysis, and SEM. The agreement is reasonable throughout the range of dissolution times. The greater scatter relative to Figure 5 is partly the result of the scale of the graph, but it also reflects the greater complexity of the optical image and the assumptions involved in extracting tn and tl from the mass loss and solution analysis data. The data do not indicate any tendency for the optical and SEM dissolution times to be shorter than those derived from mass loss and solution analysis.
The optical method presented here provides a direct measure of the time it takes a fiber to dissolve in model physiological solutions. The method is conceptually and experimentally simple and applicable to all fiber types. It does not depend on other measurements such as fiber diameter distribution or surface area, which are known to introduce error and poor reproducibility in the standard mass loss and solution analysis methods. Nor does it rely on expensive instrumentation or extensive laboratory expertise such as that needed for chemical analysis. Because individual, isolated fibers are measured, a single flow rate gives for all fiber compositions the high surface area to flow conditions thought to exist in the lung. Although the technique produces the most accurate and reproducible results when applied to fibers in the 4-10 µm diameter range, it can be used to measure the dissolution rate of fibers as small as about 2 µm in diameter. The time for analysis is roughly proportional to the fiber dissolution time. For tdis near 50 days/µm, the optical technique takes about 2 months.
d = fiber diameter at time t during dissolution (µm)
d0 = initial fiber diameter (µm)
doptical = fiber diameter measured by the optical technique (µm)
dSEM = fiber diameter measured by SEM (µm)d = fiber diameter decrease: d0-d (µm)
kdis = fiber dissolution rate (ng/cm2/hr)
m = the slope in a plot of t vs. d= fiber density (g/cm3)
t = time (days)
tdis = time to completely dissolve, congruently, a 1 µm diameter fiber (days/µm)
tl = time to completely leach a 1 µm diameter fiber (days/µm)
tn = time to completely dissolve, incongruently, a 1 µm diameter fiber (days/µm)
Xl = fraction of the initial mass leached during incongruent dissolution
Jon Bauer, Paul Boymel, Thomas Hanton, Martin Moore, and Thomas Steenberg provided fiber samples. Claire Hunkins and Cheryl Smith did most of the laboratory work which forms the basis of this paper. Walter Eastes and John Hadley provided guidance and critical review throughout the work. I am grateful to all these colleagues for their support.