touch such material with scissors and needles; so askillful and artistic young woman employee from Toledo was sent t o New York to do what the ordinary seamstress could not, for a specialist is needed to cut and sew glass which differs from other cloths in breaking and wickedly sticking into the hands.
FIBER GLASS Mechanical Development J. H. PLUMMER Owens-Illinois Glass Company, Newark, Ohio
F
This description is perfectly accurate, for this fabric (Figure 2) had been made by using bundles of coarse glass fibers for the warp and weaving them together with a filling of silk. Additional silk warp threads were added between each bundle of glass to provide the n e c e s s a r y strength. The fibers used were about 0.001 inch in diameter, over five times as large as those in use today, and the fabric could not be folded or creased without breaking. To prevent irritation to the wearer, the entire dress was lined with heavy silk.
T H E recent history of glass fibers dates from the beginning of the World War when the Germans were unable to import asbestos for insulating purposes, and it was necessary for them to develop other materials. The hand process of producing glass fibers was used a t first, but later the winding drums were increased in size and driven mechanically. Furthermore, the glass was pulled from orifices in the furnace itself, instead of being first made into rods and spun from them, The resultant fibers were very coarse by present standards and were harsh and brittle. The interior of the Gossler plant a t Dusseldorf in which glass wool was made by this process is shown in Figure 3 This view was r e p r o d u c e d from a photograph made in G e r m a n y in 1915, and now this factory is producing wool and glass textiles by processes described in this paper. An i m p o r t a n t d e velopment was made in 1931 when a blast of steam replaced the drum as a means of attenuating the glass s t r e a m . This method permitted the continuous production of flexible fibers of any d i a m e t e r desired w i t h i n t h e l i m i t s of 0.001 inch and greater. It was no longer neces-
IBER glass is a mineral fiber, the only artificial textile fiber to come from nature without passing through an animal or vegetable form. It is resistant to all acids except hydrofluoric; it will not mildew, rot, or support combustion, and it will stand much higher temperatures than any other textile fiber. Because of these properties many attempts have been made throughout past ages to produce a textile of glass. As far as is known, the first time that wearing apparel was made from spun glass was in 1893. Georgia Cayven, a prominent actress, visited the Columbian Exposition. There she saw an exhibit prepared by Michael Owens, then with the Libbey Glass Company, in which workmen were spinning glass by drawing the fibers from a heated rod by means of a drum rotated by a foot pedal. Small b u n d l e s of t h e s e fibers were woven into glass lamp shades. The appearance of these shades excited Miss Cayven’s interest, and she asked the company to make for her the 12 yards needed for a dress. This was done a t a cost of 25 dollars a yard. This was a great deal of money to invest in an experiment, but Miss Cayven also secured the exclusive right to wear glass cloth on the stage, and during one of her appearances the Spanish Princess Eulalia saw it. During her visit to the exposition she was shown a counterpart of the first dress, and, before sailing to Spain, she too was fitted for an Ameriean glass gown. Figure 1 shows a model wearing one of these early dresses. The parasol is made of the same fabric, and both the dress and the parasol are now in the possession of the Toledo Museum of Art. Another of these dresses is in the Madrid Museum, and the third has been lost. No others were made for the reason best expressed in the following contemporary account of the fitting of Miss Cayven’s dress:
0
FIGURE 1. GLASSDRESS PARASOL MADE BY MICHAELJ. OWENS IN 1893
AND
HOWto make it up was the next question, for Madame La Modiste vowed she wouldn’t 726
JULY, 1938
IKDUSTRIAL AND ENGEVEBRING CHEMISTRY
sary to stop the process every few minutes to remove the fibers that had collected. Furthermore, the process was easily controlled and only a few seconds were required to change from the production of fiber of one diameter to that of another. The first fibers made in this way were utilized in the manufacture of an impingement-type air filter which provided for the first time an easily replaceable filter unit. T o make this filter, two mats, one composed of fibers about 0.010 inch in diameter and the other from fibers about one-fifth of that size, were first bonded into semirigid mats by means of a latex binder. The fibers in the two mats were then coated with an adhesive, and the coated packs were placed in a container with grilled faces. By placing the pack made from the heavy fibers on the intake side of the filter, a progressive pack was obtained which removed more than 95 per cent of the dust from the air which passed through it. When the unit became dirty, it couId be removed, discarded, and replaced with a fresh one. During the development of the steam blown process, it was found that the diameters of the fibers produced were intimately related to the viscosity of the glass, the size of the orifices in the tank, and the rate a t which the fibers were drawn. All of these factors were under control except the size of the orifices. In the early work these orifices consisted of holes in a refractory block, and, after a few hours of operation, the holes had enlarged so as to require replacement of the bushing. Further progress could be made only after suitable materials had been found which would withstand operating temperatures and the solvent action of the molten glass.
FIGURE 2. CONSTRUCTION OF SPUN GLASSFABRICOF 1893 ( X 5.5)
Many refractories were investigated in the search for materials from which these bushings could be made. However, none were satisfactory and it was not until bushings were made of precious metal alloys that a constant orifice diameter could be maintained. These alloys permitted the use of smaller orifices and led to the development of a form of fiber glass well adapted for service as heat and sound insulation in all types of building construction. By heating the bushings electrically, it was possible to reduce the orifices to very small diameters and still maintain a satisfactory flow of glass. Furthermore, the flow could be accurately regulated by control of the bushing temperatures.
I N ORDER to reduce production costs and to produce fibers finer than 0.001inch in diameter, experiments were conducted in an effort to increase the velocities a t which fibers were
FIGURE 3. INTERIOR VIEW
727
OF
GOSSLERPLANTIK 1915
drawn. The use of high-pressure superheated steam made it possible to attenuate the streams of molten glass a t speeds exceeding the velocity of sound. These high speeds produced much smaller fibers and permitted more rapid production. Almost any type of fiber glass, long or short, coarse or fine, could now be produced by a choice of orifice diameters, working temperatures, and steam pressures. The fibers could be fabricated into many forms such as bats and granules for insulation of houses, blankets of many sizes for industrial purposes, bonded materials for rigid and semirigid insulation, and plastic cements. Fiber glass products have been used as insulation in such extremes of temperature as those encountered in containers for solid carbon dioxide and the regenerators of glass furnaces. To obtain satisfactory thermal insulation it is necessary to maintain close control of all phases of the operation. A satisfactory bat or blanket can be assured only by the use of long fibered wool or by a suitable combination of long and short fibers Long fibers are necessary in order to obtain structural strength, and short fibers are required for a high degree of interfelting. If the fibers are too fine, the mat lacks resilience. Best results are obtained by holding the average fiber diameter a t about 0.0004 inch. The form of fiber glass then produced had the appearance of natural wool, and this resemblance was probably responsible for the idea of developing fiber glass for textile purposes. The first experiments consisted of an attempt to make yarn by passing insulating wool through machines used for producing yarn from natural wool. This process proved to be unsatisfactory although short pieces of sliver were made which, when hand-twisted, gave definite hope of developing a glass thread. About this time a very fine, long-fibered, silky type of fiber glass was produced for experimental work on battery separator plates. This material was made in the form of a laminated mat which, when split into thin layers, could readily be
728
INDUSTRIAL AND ENGINEERING CHEMISTRY
VOL. 30, NO. 7
1
JULY, 1938
=
INDUSTRIAL AND ENGINEERING CHEMISTRY
no longer satisfactory. Production was required to be constant in terms of grains per second.
THE necessity for constant production rates brought about some changes in the technic of melting glass, for glass of almost optical quality was required. There could be no seeds or bubbles in the molten mass as these were likely to interrupt the flow by obstructing the orifices. If the flow from two or three holes ceased even momentarily, the sliver produced was so far off-weight that it could not be used and production was halted until the flow could be resumed. The proper quality of glass was finally obtained by melting it in a large furnace and, in order to facilitate inspection, by forming it into small spheres which are automatically fed one a t a time into the molten mass in the textiIe furnace. This was a major improvement in that very high standards were immediately instituted in the inspection of these glass spheres. No piece was permitted to pass if it contained even the slightest flaw. As the quality of the glass improved, it became possible to reduce the diameter of the fibers further, and now uniform fibers only 0.0002 inch in diameter are consistently formed on a production basis. One bank of staple fiber machines used to form this type of sliver is shown in Figure 4. From this operation the sliver passes to the spinning room, a portion of which is shown in Figure 5. I n the extreme background are the spinners which draft and twist the sliver into yarn. I n the foreground of Figure 5 several spinners are shown twisting another form of fiber glass into yarns as fine or finer than silk. The development of this material paralleled that of the staple fiber operation for, during the early development of the latter process, a very long fibered form of fiber glass was produced, and the resultant thread possessed all the luster of silk. I n reducing the average fiber length so that the slivers could be processed on standard textile machinery, this resemblance to silk was lost, and the product was then similar in appearance to cotton thread. Since, with glass yarns, the resemblance to silk or cotton depends entirely on the length of the individual fibers, a separate development was started to produce a sliver containing only continuous filaments. The process that proved to be most feasible for producing such a sliver was one in which the fibers were drawn from the orifices by mechanical means. The old spinning drum process produced just the type of fiber desired, but the fiber diameter was too large for the sliver to be used as a textile material. By combining the newly developed electrically heated bushing with its fine orifices and the spinning drum, it was found that continuous fibers with a diameter of only 0.0002 inch were readily formed. With the proper type of fiber assured, the remaining problems consisted of the development of a means of gathering the fibers into a strand and winding this strand on a spindle so that it could be easily unwound.
MANY types of traverse were tried to wind the strand on the spindle in such a way that i t could be easily unwound and a t the same time eliminate or a t least reduce the jerkiness caused by guiding the strand first to one end of the spindle and then to the otlier. High traverse speeds set up considerable vibration, and it is necessary to use a continuously rotating traverse mechanism. Reece rolls were tried and discarded because they introduced too much friction which tended to break the strand. The problem was finally solved by setting the spindle so that the strand was forced to travel to one end, and by mounting a rotating disk with projecting pins behind the spindle in such a way that as the disk rotated the pins forced the strand to the opposite end. This requires a careful adjustment of angles, rotational speeds, and length of pins, but when the proper adjustment has been secured, the first pin will pick up the strand and carry it to the far end of the
729
spindle where it drops off. The strand then travels back unaided to the near end of the spindle where it is picked u p by a second pin. This form of traverse operates smoothly and produces a truly universal wind, in that no layer is ever wound exactly above any of the preceding ones. This type of winding is necessary; otherwise there is a tendency for the
FIGURE 7. FIBERGLASS FABRIC WOVEN JACQUARD LOOM
ON A
filaments of the strand in one layer to become entangled with other layers and cause the strand to break during unwinding. A smoothly operating traverse mechanism is one of the principal requirements in high operating speeds. Vibration must be eliminated as the least jerk in the drawing is likely to snap the filaments. The breaking of fibers near the bushing is serious, not because of the loss of one or two filaments but because in breaking they are likely to snap back and break more. Since it may require half an hour for an operator to start all of the fibers properly, a truly efficient operation is one in which no breaks occur. These difficulties have been overcome largely through the use of glass of optical quality, and one hundred and two filaments are pulled a t a time a t the rate of 6000 feet per minute. Experience indicates that this figure may soon be greatly exceeded, although the strands produced by this process are composed of one hundred and two filaments, each of which is over 12 miles long; the strand itself is little larger than a human hair. Continuous filament strands are produced on the machines shown in Figure 6. The strand which is drawn frqm above by the rotating spindles does not show in this view, as the traverse action causes it to move rapidly from side to side. This view also shows the Bakelite spools on which the strand is wound. From this operation the spools pass to the twisters (shown in the foreground of Figure 5 ) where the strands are unwound and twisted together into a yarn. Further processing may be accomplished with standard textile equipment, and the sheerest of fabrics may be produced. Such afabric, woven on a Jacquard loom is shown in Figure 7. R~CEIVE February D 15, 1938.
