Western Springs, IL). Crystals were identified as silicon; to- tal silicon concentration (measured as silica) in the submitted sample was 49 ppm. Concentration of ...
NOTES
Silicon Crystals in Water of Autoclaved Glass Bottles JEFFREY J. LOHMILLER1, DVM, MS AND NEIL S. LIPMAN1,2, VMD Abstract Crystals were detected in glass bottles containing autoclaved (standard cycle: 121 C, 17 lb/in2, for 20 min) tap water. Bottles were fitted with a sipper tube and stopper. Crystals were observed twice at approximately a 6-month interval. Analysis of crystals and crystal-laden water by transmission electron microscopy and inductively coupled plasma and energy X-ray analysis revealed that the crystals were silicon. Subsequently, procedures and processes involved with preparation of water bottles were analyzed to determine the source or factors involved with silicon contamination. Analyses of silicon concentrations were performed on samples of tap water (0.88 to 1.20 ppm), tap water autoclaved in polycarbonate (0.89 to 1.20 ppm) and glass water bottles (0.84 to 11.0 ppm), glass and polycarbonate bottles after purposeful contamination of tap water with various amounts of alkaline detergent (1.10 to 1.70 ppm), and on steam condensate from the autoclave (0.23 and 0.47 ppm) to ascertain the source of silicon. Water from autoclaved glass bottles had greater concentrations of silicon, compared with unautoclaved tap water. Siliconized rubber stoppers and stainless-steel sipper tubes were excluded as sources of silicon contamination. Autoclaving glass bottles with stopper pieces did result in crystal formation, despite only modest increases in silicon concentration. Detergent contamination did not influence silicon concentrations. Analysis of the data indicated that autoclaving glass water bottles can lead to increases in silicon concentration and, under some circumstances, crystals may form. During the fall of 1994, crystals were identified in a case of half-pint glass bottles filled with tap water that had been autoclaved and were to be used for barrier-maintained rodents. The water was discarded, and the bottles were reprocessed. Crystals were not detected again for approximately 6 months. In the second episode, only glass bottles in a case that contained autoclaved glass bottles and polycarbonate bottles developed crystals. The physical condition of affected bottles ranged from worn and scratched to new an unmarred. Similarly, affected bottles had rubber stoppers that ranged from worn and pitted to stoppers that appeared new. Crystals were angular, translucent, refracted light, and settled to the bottom of bottles, appearing as an opaque white precipitate. Light microscopic examination revealed crystals to be 200 to 350 m x 375 to 1,550 m. Within these flat, angular crystals, smaller (7.5- to 10- m diameter) cuboidal crystals were embedded (Figure 1). In contrast to the large crystals, the smaller crystals were birefringent in polarized light. Due to recurrence of crystals at our institution and reports of incidents at other institutions (1), we investigated the composition of the crystals. Initial qualitative analysis of a highly concentrated sample of crystal-laden water was performed by means of transmission electron microscopy and inductively coupled plasma (ICP) and energy X-ray analysis (Midwest Laboratories Inc., Western Springs, IL). Crystals were identified as silicon; total silicon concentration (measured as silica) in the submitted sample was 49 ppm. Concentration of silicon in the sample submitted was artificially high, because the sample contained concentrated crystals, which were dissolved prior to determination of the silicon concentration. Subsequently, we attempted to identify the source of the silicon that led to crystal formation. Potential sources included the water source and purification and distribution system, contamination during processing of the bottles, steam used for autoclaving, and various components of the bottles. The water supply came from Lake Michigan. Quarterly analysis of raw water from the intake cribs by the Bureau of Water 1
The Department of Pathology and the Committee on Comparative Medicine and Pathology, The University of Chicago, 5841 S. Maryland Ave., Chicago, IL. 2 Research Animal Resource Center, Memorial Sloan-Kettering Cancer Center and Cornell University Medical College, 1275 York Ave., New York, NY. 62
Operations, Department of Water, City of Chicago were obtained. We examined records for the 2 years preceding initial detection of the crystals as well as the period throughout which crystals were detected. Silicon concentration (measured as silica) during this period ranged from 0.39 to 1.18 ppm. These amounts were not sufficiently high to account for the increased concentrations that would result in crystallization. Fluctuation in silicon content of lake water is associated with temperature inversions during the late fall and spring as well as consumption of silicon by diatoms, algae, and other plants during growth (2). Inversions result in mixing of deep water layers, which are in contact with lake sediment, with surface water. Inversions increase water silicon concentrations as vegetative break-down products and diatoms, which are high in silicon, are disturbed from the lake bottom. The water treatment and purification process was evaluated to identify introduction of agents that may have contained silicon. During the water treatment process, chlorine was added to kill certain microflora, and fluoride was added for dental prophylaxis. Alum was added as an adsorbent to remove particulates, and the resulting solids were removed as a flocculum prior to water entering the distribution system. Polymers and phosphates were added to the water prior to distribution to coat pipes and prevent lead absorption into the water from older pipes of the distribution system. None of these additives contained silicon. However, some municipalities may use silicon-containing products to coat old lead pipes to diminish leaching. Steam released into the autoclave also could have served as a source of silicon. The autoclave was supplied with steam that may have contained one or more additives from the central steam plant. Chemicals that were added to the boiler to prevent excessive scale buildup and to prevent foaming that may have contaminated the steam included lye, sulfuric acid, sodium sulfite, and a silicon-containing antifoaming agent. Cyclohexylamine, an alkaline product (pH 12.5) that was added to prevent pipe corrosion, was the only agent added directly to the steam, and it was added at a rate designed to maintain steam condensate at or near pH 8.2. Steam condensate returning from the autoclave contained 0.23 and 0.47 ppm silicon and had pH 8.03 and 6.51 in samples obtained at 2 sampling times. Steam did not appear to be the source of silicon contamination. After excluding tap water and steam as the source of silicon
bottles. An alkaline detergent (Clout®; Pharmacal Research Laboratories Inc.; Naugatuck, CT) was the sole chemical used in the tunnel washer. Baskets of bottles destined for use in barrier rodent facilities were placed in clear autoclave bags (Associated Sales and Bag Co.; Milwaukee, WI), closed with a metal tie, and autoclaved in a high-vacuum steam autoclave (American Sterilizer Company [AMSCO]; Erie, PA), using a standard autoclave cycle for liquids (121 C, 17 lb/in2, for 20 min). Detergent residue from improperly rinsed bottle components was examined as a potential source of silicon. Known volumes of tunnel washer sump fluid containing diluted alkaline detergent were placed in glass and polycarbonate water bottles, which were then autoclaved. Volumes of sump fluid (0, 0.25, 0.5 and 1 ml of sump fluid per 250 ml of water) were purposefully large to maximize the chances for detecting differences in silicon concentrations. The resulting silicon concentrations did not differ with varying amounts of sump fluid; however, silicon concentrations were higher in glass bottles, compared with polycarbonate bottles (Table 1). Silicon concentrations were determined in larger numbers of glass and polycarbonate bottles after autoclaving. Glass bottles consistently had increased silicon concentrations; however, the degree of increase varied considerably (Figure 2). The resulting silicon concentration in water from autoclaved glass bottles correlated with pH of the water after autoclaving (Figure 3). Silicon concentrations increased markedly when pH of water was > 8.5. Increases in pH of water in glass (mean SD, 0.71 0.26 pH 0.09 pH units) units) and polycarbonate (0.67 bottles were attributed to autoclaving, presumably because of alkalinity of the steam. To identify the component of glass water bottles that contributed to the increased silicon concentration after autoclaving, water bottles containing pieces of rubber stopper or sipper tube were autoclaved. Black, siliconized, rubber stoppers (new and badly worn) were cut into 8 equal- sized pieces. Stopper pieces were FIG. 1. Photomicrograph of crystals. Notice the flat angular nature of the large added to glass and polycarbonate water bottles such crystal (open arrow) with the smaller cuboidal crystal (solid arrow) embedded in that stopper weight-to-water volume ratios were the the crystal matrix. Bar = 5 m. same for each pair of glass and polycarbonate bottles. A stainless-steel sipper tube alone was placed inside contamination, controlled quantitative experiments were con- glass and polycarbonate bottles filled with tap water. To maxiducted to evaluate the role of detergent residue from the tunnel mize the silicon contamination from the bottle components, washer, water pH, bottle composition (glass or polycarbonate), tap water was made alkaline (pH 8.5) by the addition of a 5M rubber stoppers, and sipper tubes on silicon concentration. Sili- solution of NaOH prior to autoclaving. All bottles were covcon concentrations were determined by means of ICP analysis ered with aluminum foil prior to autoclaving to prevent by the Bureau of Water Operations, Department of Water, City excessive evaporative loss. New or worn stopper pieces did not of Chicago; Chicago, IL. The detection limit of ICP for silicon increase the silicon concentration in either bottle type when was 0.1 ppm. Water samples were collected directly from the tap compared with its control sample (Table 2). Although the conor from autoclaved bottles after the water cooled to ambient centration of silicon in the glass bottle with sipper tube was temperature. Samples were stored in sterile plastic vials (Becton increased with respect to the autoclaved control glass bottle, Dickinson Labware, Franklin Lake, NJ) until analyzed. The pH we would have needed to evaluate additional samples to deterof all water samples was measured before and after autoclaving, mine the importance of this finding. The increase in silicon using a pH probe (Cole-Parmer Instrument Co., Chicago, IL). concentration observed in water from this bottle may have reAutoclaved water bottles had rubber stoppers and stainless-steel flected the variability of silicon observed after autoclaving glass bottles (Figure 3). We did observe a small number of crystals in sipper tubes in place, except where indicated. Standard procedures for processing water bottles at our insti- glass bottles that contained stopper pieces, but crystals were tution were performed. Polycarbonate and glass water bottles not evident in polycarbonate bottles with stopper pieces. These were placed in stainless-steel baskets and were sanitized in a tun- crystals resembled those observed during the initial 2 episodes, nel washer. Black, siliconized rubber stoppers containing as determined on the basis of light microscopy. We systematically examined potential factors involved in the stainless-steel sipper tubes were bulk washed and sanitized in a mesh basket in the tunnel washer. Stoppers with sipper tubes development of silicon crystals in autoclaved water bottles. Wawere allowed to drain prior to being inserted into water-filled ter source and processing were not contributory. However, it 63
Table 1. Effect of sump fluid from a tunnel washer on silicon concentration in autoclaved glass and polycarbonate water bottles.
Bottle type
Amount of sump fluid (ml of sump fluid per 250 mL of tap water)
Glass Glass Glass Glass Polycarbonate Polycarbonate Polycarbonate Polycarbonate
Silicon concentration* (ppm)
pH
1.5 1.5 1.7 1.6 1.2 1.1 1.1 1.2
NA 7.89 7.89 7.91 NA 8.04 8.19 8.25
0 0.25 0.5 1.0 0 0.25 0.5 1.0
*Known volumes of tunnel washer sump fluid containing diluted alkaline detergent were placed in glass and polycarbonate water bottles and autoclaved (121 C, 17 lb/in2, for 20 min). Silicon concentrations were measured as silica by means of inductively coupled plasma analysis. Data represents one sample per bottle condition. NA = data not available.
Table 2. Effect of siliconized rubber stoppers and stainless-steel sipper tubes on silicon concentration in autoclaved water bottles.
Bottle type Control (tap water) Glass Polycarbonate Glass Polycarbonate Glass Polycarbonate Glass Polycarbonate
Stopper
Sipper tube
Silicon concentration* (ppm)
+ (Worn) + (Worn) + (New) + (New) -
+ +
0.98 4.18 1.02 4.16 1.03 3.83 1.03 6.01 1.05
* Bottle components were added to water (pH 8.5) filled glass and polycarbonate bottles which were then autoclaved (121 C, 17 lb/in2, for 20 min). Silicon concentrations were measured as silica by means of inductively coupled plasma analysis. Data represents one sample per bottle condition. (-) = indicated components was not in bottle during autoclaving. (+) = indicated component was in bottle during autoclaving.
should be noted that major differences in silicon content of water may exist as a result of geologic factors including soil type and condition (3). Water from certain wells, especially those in the Southwestern United States, may contain high concentrations of silicon (4) as a result of geologic factors. Detergent residues were also excluded as a potential silicon source. Tap water from autoclaved glass bottles consistently yielded high concentrations of silicon, compared with unautoclaved tap water and tap water autoclaved in polycarbonate bottles. This result was consistent with that of another study (5) in which the investigator examined several types of glass bottles and effects of the amount of silica on parenteral solutions after autoclaving. In that study, it was concluded that autoclaving glass bottles increased silica concentration in the fluid and that repeated autoclaving had a significant influence on the release of silica into the fluid. The chemical nature of silicon is complex. It exists in several forms, such as monionic, colloidal (polymerized silicic acid), and polymeric (4). Colloidal and dissolved silicon are in equilibrium with each other, and the predominant form is pH-dependent. At pH > 8, silicic acid dissociates into silicate anion (SiO32-) increasing the solubility of silica (6). If the solubility of silicate anion is exceeded, it may precipitate as silicon dioxide (SiO2). If other multivalent cations such as calcium, magnesium, iron, or aluminum are present, insoluble silicate 64
Control
Polycarbonate
Glass
Bottle Composition FIG. 2. Silicon concentration in autoclaved polycarbonate and glass water bottles. Bottles were fitted with new black rubber stoppers and stainless-steel sipper tubes and were autoclaved (121 C, 17 lb/in2, for 20 minutes). Silicon concentrations, measured as silica, were determined by inductively coupled plasma analysis. Control samples were unautoclaved tap water.
Glass bottles Polycarbonate bottles
pH FIG. 3. Correlation of silicon concentration and final pH of water in autoclaved polycarbonate and glass water bottles. Bottles were fitted with new black rubber stoppers and stainless-steel sipper tubes and were autoclaved (121 C, 17 lb/in2, for 20 minutes). Silicon concentrations, measured as silica, were determined by inductively couple plasma analysis.
salts may precipitate (6). Silicon solubility is greatly dependent on temperature (6). This was reflected in the increased silicon concentration observed in autoclaved glass bottles. Silicon solubility is also influenced by pH. At pH < 7.0, the solubility of silicon moderately increases as the pH decreases. However, the solubility of silicon is dramatically increased at pH > 8 (6). At pH < 8, the predominant precipitate is SiO2, and at pH > 8.0, silica salts precipitate when metallic cations are present. The magnitude of increased silicon concentration, pH of water, and concentrations of other solutes determine whether the solubility product for silicon is exceeded, favoring crystallization. We were not able to ascertain the exact composition of the glass because of several factors, which included manufacturer reluctance to release the information and the fact that a subset of the bottles was extremely old. Glass can be composed of many oxides, such as silicon dioxide, sodium oxide, calcium oxide, aluminum oxide, potassium oxide, and magnesium oxide. It is likely that the composition of glass bottles used within our facility varied due to manufacturer and age of some of the bottles. It should be noted that this study measured only changes in sili-
con concentration after autoclaving, but the concentration of other cations from oxides used to manufacture the glass might have increased after autoclaving (5). Cations released into solution after autoclaving may contribute to crystal formation. Siliconized rubber stoppers were not a source of silicon in our limited evaluation. Rubber stoppers are siliconized to prevent generation of rubber particles during repeated use, to reduce friction, and to reduce the natural tackiness of rubber. The appearance of crystals within bottles containing stopper pieces was surprising. Release of minute (< 32 m in diameter) rubber stopper particles with autoclaving has been reported (7, 8). Repeatedly autoclaved rubber stoppers become worn and pitted. The stoppers did not influence silicon concentration, but may have contributed to initiation of crystallization by increasing particulates in water, which can serve as a nidus for crystal formation. Manufacturers recommend neoprene stoppers for use when autoclaving liquids. Black, siliconized rubber stoppers should not be repetitively autoclaved. However, in the authors’ experience, it is commonplace for animal holding facilities to repeatedly autoclave siliconized black rubber stoppers. Safe amounts of silicon in drinking water have not been established by the United States Environmental Protection Agency or World Health Organization (9). Information is limited regarding toxicosis associated with excessive oral intake of silica. Soluble and amorphous silicon-containing compounds are used widely as stabilizers during processing of food, beverage, and pharmaceutical products. Chronic ingestion of silicon has been identified as an important risk factor for developing Balkan nephropathy in human beings (10). Increased concentrations of silica in well water have been reported in regions of Eastern Europe where this degenerative renal disease is reported. Experimental administration of silica-containing compounds to guinea pigs did induce renal lesions (11). It is unlikely that increased concentrations of silicon induced by autoclaving glass water bottles will lead to adverse effects in research animals. Although crystals or high concentrations of dissolved silicon have not been clearly implicated in health problems, the variation in silicon content from bottle to bottle could introduce unwanted variables in research animals. There also is the potential for crystals to occlude sipper tubes, and the possibility, albeit small, that crystals that have dripped onto the bedding could be inhaled by rodents during rooting. The definitive factor leading to crystallization in these instances was not clear. Glass water bottles appeared to contribute to increased silicon concentrations after autoclaving. Increased pH of water and degenerating stoppers may also have favored crystal formation. Analysis of data presented here indicated that
autoclaving glass water bottles in some circumstances leads to increased silicon concentrations and may cause conditions that favor crystallization. Veterinarians, scientists, and managers should be aware of the potential for crystal formation and should take precautions to eliminate this problem.
Acknowledgments The authors thank Ms. Ellen Wojcieszak of the Chicago Bureau of Water Operations for analysis of water samples.
References 1. Lipman, N. S. 1996. Personal communication. 2. Stauffer, R. E. 1986. Linkage between the phosphorus and silica cycles in Lake Mendota, Wisconsin, USA. Water Research, 20:597–610. 3. Carlisle, E. M., J. A. McKeague, R. Siever, et al. 1977. Silicon. p. 54–72. In Geochemistry and the environment, Vol. II. The relation of other selected trace elements to health and disease. National Academy of Sciences, Washington, D.C. 4. Meltzer, T. H. 1993. High-purity water preparation for the semiconductor, pharmaceutical, and power industries, pp. 261–265. Tall Oaks Publishing, Littleton, CO. 5. Passl, W. J. 1979. Interactions of parenteral solutions with sulphur-treated glass bottles. J. Pharm. Pharmacol. 31:721– 725. 6. Byrne, W. 1995. Reverse osmosis: a practical guide for industrial users. Tall Oaks Publishing, Littleton, CO. 7. Mannermaa, J. P., E. Muttonen, J. Yliruusi, et al. 1992. Comparison of different rubber stoppers; the effect of sterilization on the number of particles released. J. Parenteral Sci. Technol. 46:73–77. 8. Li, L. C., J. Parasrampuria, Y. Tian. 1993. Particulate contamination from siliconized rubber stoppers — a statistical evaluation. J. Parenteral Sci. Technol. 47:270–273. 9. DeZuane, J. 1990. Handbook of drinking water quality: standards and controls, pp. 135–136. Van Nostrand Reinhold, New York, NY. 10. Radovanovic, Z., L. Markovic-Denic, J. Marinkovic, et al. 1991. Well water characteristics and the Balkan nephropathy. Nephron. 57:52–54. 11. Dobbie, J. W., M. J. B. Smith. 1982. Silicate nephrotoxicity in the experimental animal: the missing factor in analgesic nephropathy. Scottish Med. J. 27:10–16.
65
