<< Back to Issues and Controversy: Measurement of Crystalline Silica


Because of the potential health hazard of the crystalline forms of silica in the industrial environment, federal regulations require monitoring the presence of silica(*) forms in the workplace atmospheres and marketed products. From time-to-time, it is important to review and evaluate the regulations and methods for monitoring these crystalline silicas. This paper reviews and evaluates the use of X-ray powder diffraction for detecting, identifying and quantifying the crystalline and amorphous silicas in all types of samples from airborne dusts to bulk commercial products. This project has been sponsored by the Chemical Manufacturers Association.

Crystalline Silica in the Industrial Environment

The inhalation of dust as a primary cause of pulmonary disease has been a problem in the mining industry since antiquity. The problems were first termed pneumonokoniosis by Zenker (1867), but the generic term has since been shortened to pneumoconiosis. The word originally implied that the lung had been seriously damaged by dust, but the meaning has been broadened to include all pulmonary manifestation of dust inhalation (Goldberg and Jacobson, 1972). The result of pneumoconiosis is to harden the linings of the lung by creating fibrous growths that make the lung effectively inoperative. Cures are impossible, so prevention is the goal of modern industry.

The most important form of pneumoconiosis is silicosis which could affect workers in many industries as well as in the mines. The cause of silicosis is primarily inhaled particles of crystalline silicon dioxide, SiO2, most commonly quartz which is a ubiquitous mineral in nature. The main technique for prevention of silicosis is to clean up the atmosphere that workers might inhale. Early prevention procedures in "hard rock" mining were to shift from drilling dry to drilling with water flushing the cuttings away from the working face. The freshly broken quartz particles from the quartz in the rocks and veins were extremely reactive both because of their angular shapes with active edges and because of the fresh surfaces. It is now known that ageing of quartz particles diminishes the activity but does not eliminate it, so it is necessary to reduce the atmosphere particulate content anywhere workers are liable to encounter crystalline silica.

Modern practices in mines and industrial plants include directing the affected atmosphere away from the worker and into collector systems that remove the respirable particle component before the atmosphere is recirculated. The atmosphere and personnel are continuously monitored by sampling devices that can accumulate the particulate matter in a quantitative fashion for subsequent laboratory analysis. These samples present the analytical laboratory with several challenges: the detection of the silica minerals in the sample, the quantification of the amount of crystalline silica in the sample, and the need to accomplish these measurements rapidly and accurately on large numbers of samples on a routine basis. Three methods are commonly employed: infra-red analysis, chemical analysis on treated samples and X-ray powder diffraction (Gebhardt, 1975 and Hamilton et al., 1990). It is the purpose of this paper to review and evaluate the X-ray diffraction procedures in the light of modern equipment, sample handling and our understanding of the diffraction process.

Forms and sources of crystalline silica

Silicon dioxide, SiO2, may occur in many crystalline forms of which quartz is the most common. However there are many other polymorphs which may occur as natural minerals and/or are synthesized in the laboratory. Frondel (1962) presents a very comprehensive review of the mineralogy of SiO2 discussing all mineral forms. Drees et al. (1989) reviews the occurrences of silica minerals in soils. Table 1 lists most of the known forms whose powder diffraction patterns have been recorded in the Powder Diffraction File (PDF, 1991). A few other clathrasils and some doubtful phases have been mentioned in the literature. Figure 1, redrawn from Ostrovsky (1967), shows the phase relations of the stable forms as a function of pressure and temperature. Only six forms appear on this diagram, and there is considerable question whether tridymite is truly stable. Achieving stability is always difficult in SiO2 phase studies because of the sluggish nature of all the transitions. Once a major structure type is formed, it tends to continue to exist even outside its true stability field as a metastable phase. Figure 2, modified from Sosman (1955), shows diagrammatically the metastable behavior of the many polymorphs which can occur at atmospheric pressure. All of the major phases may be supercooled and superheated because of the sluggish nature of the transitions to the different structure types. Within each structure type, the transitions are fast because there are no bonds broken. The transitions are all caused by changing the angle of the Si-O-Si bonding as a function of temperature.


Figure 1. Pressure-temperature phase relations for SiO2.
Figure 1. Pressure-temperature phase relations for SiO2.

Quartz

Quartz is classified as a tetrahedral framework structure in which the Si is 4-coordinated to the O atoms which in turn bridge two Si tetrahedra to form a framework of tetrahedra with 6 and 12-membered loops. The bonding is estimated to be about 50% covalent and 50% ionic. There are two thermal polymorphs with the same framework topology; but the high-temperature form, stable above 573C, is unquenchable, and only the low-temperature form is encountered in atmospheric samples or in the analytical laboratory. The composition shows very little deviation from stoichiometry; however, some water may be incorporated by breaking one Si-O-Si link with the formation of Si-OH HO-Si clusters. Quartz is considered to be chemically inert, reactive only in hydrofluoric and phosphoric acids and strongly basic solutions, but it does react in the lung linings to initiate fibrous tissue growth for unknown reasons. Whether the quartz acts as an irritant, a nucleation site, or is involved in the chemical reactions is not clear. By federal regulations, the permissible exposure limit, PEL, of quartz in the working atmosphere is 5% of the respirable particulates or 0.100 mg•m-3 during an 8 hour workday, and any product containing more than 0.1 weight percent quartz, cristobalite or tridymite must be labeled as a potential hazard.

Cristobalite

Cristobalite is second most common crystalline form of SiO2 which is encountered at ambient conditions. It is metastable at these conditions, but once formed persists indefinitely. It is also a tetrahedral framework structure composed of 6-membered rings. There are high- and low-temperature forms with the same topology, but as for quartz only the low form is encountered in industrial atmospheres. In spite of the shorter loops of tetrahedra, cristobalite has a more open structure than quartz, and traces of alkali ions may be incorporated in the cages combined with aluminum substituting for the silicon in the framework. Chemically, it is also considered inert with reactivity similar to quartz. With respect to silicosis reactivity, cristobalite appears more toxic than quartz, Wright (1978), but because it is less common than quartz, it has not received the attention in toxicology studies. The permissible exposure limit is set as one half the limit of quartz, i.e. 0.050 mgm-3.
Figure 2. Metastability behavior of SiO2 at ambient pressure.
Figure 2. Metastability behavior of SiO2 at ambient pressure.

Table 1. Crystalline forms of silica represented in the Powder Diffraction File.
Pattern # Code Phase Designation Cell Parameters (, degrees)
PDF-46-1441 0M Lutecite [NR] a = 8.77 b = 4.879
= 90.08
c = 10.72
PDF-46-1242 0 (at 53 GPa)
PDF-46-1045 * H Quartz a = 4.91344 c = 5.40524
PDF-46-570 CO a = 5.01 b = 21.52 c = 11.13
PDF-45-1374 * T Stishovite a = 4.1791 c = 2.6659
PDF-45-131 CH Zeolite SSZ-24 a = 13.603 c = 8.277
PDF-45-130 CH Zeolite SSZ-24 a = 13.671 c = 8.328
PDF-45-112 * C a = 24.225
PDF-45-111 i H a = 17.17 c = 28.28
PDF-44-1394 CO + organic a = 14.24 b = 20.14 c = 8.40
PDF-44-696 i O a = 20.05 b = 20.0 c = 13.4
PDF-43-784 0
PDF-43-745 0
PDF-43-596 0
PDF-42-1401 i O Tridymite a = 17.0859 b = 9.9313 c = 16.3041
PDF-42-22 CT a = 20.067 c = 13.411
PDF-42-5 iM a = 9.91 b = 20.63
= 99.7
c= 9.80
PDF-40-1498 a = 10.2387 c = 34.3829
PDF-39-1425 0 Cristobalite, syn. a = 4.9732 c = 6.9236
PDF-38-651 iR Clathrasil a = 13.887 c = 40.989
PDF-38-360 M Moganite [NR] a = 4.934 b = 10.761
= 92.29
c = 8.533
PDF-38-197 iO Zeolite O1 a = 13.836 b = 17.415 c = 5.042
PDF-35-63 0O --------- a = 19.51 b = 13.98 c = 21.6
PDF-34-1382 0 Silica X
PDF-34-717 Silica X1
PDF-32-993 T ---------- a = 12.75 c = 4.72
PDF-31-1234 0T Silica X2 a = 19.4 c = 14.50
PDF-31-1233 0T Silica Y a = 15.50 c = 6.60
PDF-30-1127 0 ----------
PDF-29-85 ----------
PDF-27-605 iC Cristobalite, high a = 7.13
PDF-18-1170 iM Tridymite-1M a = 18.504 b = 5.0064
= 105.84
c = 23.845
PDF-18-1169 H Tridymite, high a = 5.046 c = 8.236
PDF-16. 380 0 ----------
PDF-16-331 iC Melanophlogite a = 13.402
PDF-14-654 M Coesite a = 7.17 b = 12.38
= 120
c = 7.17
PDF-14-260 iH Tridymite-20H a = 9.92 c = 81.5
PDF-13-26 T Silica K (Keatite) a = 7.46 c = 8.61
PDF-12-708 H Quartz, disord. a = 5.006 c = 5.459
PDF-11-252 0H Quartz, high a = 5.002 c = 5.454

Codes
*, i, (blank), and 0 are PDF quality marks
C, H, R, T, O and M indicate the crystal system.

Other Polymorphs

Other crystalline forms of SiO2 include tridymite, coesite, stishovite, and melanophlogite which are natural minerals and keatite and several clathrasil-type forms which have no natural counterpart. Although all members of this list may be potentially toxic, only tridymite is listed as a hazardous material primarily because of its similarity to cristobalite in structure and chemical properties and its reported stability in the SiO2 phase diagram. Wright (1978) states that tridymite has a similar toxicity to cristobalite and that stishovite seems to be unreactive. No evaluation of the other forms is given. The other minerals are very rarely encountered in nature or in industry. Tridymite is also very rare, and may actually not exist as pure SiO2. All natural tridymite contains small amounts of Na, and it is impossible to synthesize without the presence of an alkali. Its structure is composed of tetrahedra linked in 6-membered loops similar to but with a different topology than cristobalite. The sluggish nature of the transition precludes its formation directly from quartz or cristobalite by heat treatment without the appropriate mineralizers present.

Opal

Opal is a mineraloid of SiO2nH2O which is paracrystalline to amorphous. It is a surprisingly common phase in nature, occurring principally in sedimentary environments. It forms as concentrations of siliceous animal skeletons such as diatoms and radiolaria which may form thick sedimentary layers. Opal also forms when ground water leaches silica from highly siliceous local rocks and soils and redeposits it as a silica gel in interstices and crevices. Opal also occurs in volcanic regions as a secondary reaction of hot water with the siliceous volcanic rocks.

Opal forms at low temperatures, typically less than 100C, and is usually X-ray amorphous. As opal ages geologically, it dewaters and ultimately devitrifies to quartz as chalcedony, a cryptocrystalline variety of quartz. However, if the formation temperature is high, the opal may develop a structure very similar to cristobalite and/or tridymite. The effective crystallite size of this form is less than 500, and its X-ray diffraction pattern is considered distinct from the truly crystalline forms of silica. On the basis of studies started by Jones et al. (1964), opal is classified into three forms: opal-A (X-ray amorphous), opal-C (resembles crystalline cristobalite), and opal-CT (shows a disordered form with some tridymite character).

Toxicity studies have been made only on the diatomite form of opal. Because of its amorphous nature, it is not easy to detect and quantify opal by X-ray diffraction. A diffraction pattern for opal-C has been reported which is listed in the Powder Diffraction File as PDF-38-448, and traces from opal-CT have also been reported in the literature. Heating amorphous opal to 1100C will cause it to recrystallize as true cristobalite and allow its detection and quantification. However, other amorphous forms of silica also behave the same way. Table 2 lists all the hydrated forms of silica in the PDF. When monitoring for opal, it may prove necessary to monitor some of these forms as well.

Table 3. Low-angle diffraction patterns of the principal crystalline silica phases and opal
Quartz
PDF 46-1045
Cristobalite
PDF 39-1425
Tridymite
PDF 18-1170
Opal
PDF 38-448
d I hkl d I hkl d I hkl d I hkl

4.2550 18 100 4.040 100 101 4.328 90 040 4.08 100
3.3435 100 101 3.515 <1 110 4.236 2 131 3.14 9
2.4569 9 110 3.136 8 111 4.107 100 211 2.86 10
2.2815 8 102 2.841 9 102 3.867 20 221 2.51 30
2.2361 4 111 2.487 13 200 3.818 50 041 2.13 4
2.1277 6 200 2.467 4 112 3.672 2 022 2.03 4
1.9799 4 201 2.342 <1 201 3.642 4 112 1.937 5
1.8180 13 112 2.1179 2 211 3.461 2 050 1.878 5
1.8017 <1 003 2.0196 2 202 3.396 4 231
1.6717 4 202 1.9294 4 113 3.250 4 150
1.6592 2 103 1.8715 4 212 3.215 2 202
1.6083 <1 210 1.7591 <1 220 3.017 4 311
1.5419 9 211 1.7303 <1 004 2.975 25 042
1.6922 2 203 2.950 2 321
1.6349 <1 104 2.776 8 160
1.6122 3 301 2.609 2 023
1.6013 1 213 2.500 16 341
2.490 14 410
2.385 2 133
2.342 2 223
2.308 16 043


Identification of the Amorphous Forms of Silica

The direct identification of the several amorphous forms of silica is essentially impossible by X-ray diffraction unless the sample is a single homogeneous phase. There are undoubtedly several distinct structures for the amorphous states of silica. Silica glass is anhydrous and has a different type of tetrahedral linkage than is found in opal where the tetrahedra cluster into spherical agglomerates, and the water plays a role in the linkage. Other forms of amorphous silica may have different structures also as evidenced by the clathrasils formed from gel synthesis. The broad diffraction peak due to the structure of the particular state may shift in diffraction angle, but there is little else to distinguish the states.

Figure 3. Diffraction patterns of the principal 
crystalline silica minerals. A. Quartz. B. Cirstobalite. C. Tridymite. D. Opal-C.
Figure 3. Diffraction patterns of the principal crystalline silica minerals. A. Quartz. B. Cirstobalite. C. Tridymite. D. Opal-C.

Heating an amorphous silica usually causes it to recrystallize as cristobalite. Most samples will convert at temperatures as low as 1100C, but further heating to 1450C may be required to convert quantitatively the material. The thermal behavior of the sample would be more sensitive to the starting material than the final cristobalite. However, if the conversion is complete, the methods for quantifying cristobalite may be used to quantify the amorphous silica. Whether the conversion is affected by the presence of other phases in the sample needs to be studied.

Literature Review

There is considerable literature on the detection and quantification of the silica minerals by X-ray diffraction. An extensive review of all aspects of the silica problem was conducted by Babyak and Kamchak (1975, 1978), which contains many of the early references. The present list of references contain all of the papers published since 1948 that were located in a recent literature search. This date represents the real beginning of quantitative X-ray powder diffraction. No attempt was made to cover the large number of papers on quantification in general, although a few have been included among the references for specific purposes. Very detailed discussions of quantitative analysis may be found in Klug and Alexander (1974) and Brindley (1980). A very recent review of quantitative X-ray diffraction analysis published by Snyder and Bish (1989) covers most of the other pertinent literature. Recent reviews of silica determinations have been presented by Nenadic et al. (1971), Anderson (1975) and Hamilton et al. (1990).

The papers covered in this literature review may be classified as concentrating on laboratory procedures specifically for the determination of silica, evaluations of these procedures and inter-laboratory tests, analytical problems such as particle size and amorphous states, and discussions of theoretical and technical aspects of the quantification. These categories are indicated in the reference list.

Standardized Procedures

As is necessary in any good routine type analysis, especially where the results may have legal implications, it is necessary to establish standard procedures for collecting and processing the samples. The procedures for the determination of respirable crystalline silica are primarily designed by governmental safety and health agencies to meet the requirements of federal health regulations. They may be modified by industrial analytical laboratories to meet specific needs and take advantage of local equipment. In the United States, the principal organizations are the Occupational Safety and Health Administration, OSHA, and the National Institute for Occupational Safety and Health, NIOSH. In the United Kingdom, it is the Health and Safety Executive, HSE; and in Canada, it is the Ministry of Labour. The procedures are listed in the reports section of the references. In addition to the reports, a few agency papers are in the open literature such as Abell et al. (1981b) and Anderson et al. (1976). Because of the importance of determining silica levels in mine atmospheres, the U. S. Bureau of Mines has also devoted considerable effort to silica analysis, and their reports are also listed.

Although the agencies have established the procedures, most of the analyses performed are done by service laboratories or by companies monitoring their own processes. Many of these companies have established their own procedures as dictated by their needs, but the results are usually compared with test samples processed by both the developed procedure and the standard procedures. A review of silica procedures by Anderson (1975) has covered the earlier literature. Newer papers include Donovan et al. (1972), Allen et al. (1974), Altree-Williams (1977), Altree-Williams et al. (1977), Bumsted (1973), Davis and Johnson (1982a and b), Henslee and Guerra (1977), Machacek (19__), Malik and Viswanathan (19__), and Tossavainen (1979).

In addition to the procedures for determining quartz and cristobalite in respirable samples, there are several procedures for quantifying the levels of these minerals in bulk samples. These procedures are necessary because of the requirement that hazard labels be applied to any product that contains 0.1 weight percent crystalline silica. The more recent papers include silica analysis in clays (Murray and Merkl, 19__; Carter et al.,1987); paint (Kamarzchik, 1980); coal, (Davis et al, 1986); dolostone (Emig and Smith, 1989); and rocks (Pawlowski, 1985). A general review of the theory and methods used for bulk analyses is presented by Snyder and Bish (1989).

Evaluations of Standard Procedures In order to establish confidence in the use of standard procedures, it is necessary to test the methods in the working laboratories. Several tests are described in the literature. Several of the tests were monitored by NIOSH personnel or contractors including Peters (1976), IIT (1983), Giles and Cee (19__), and A. D. Little, Inc. (1976). Other reports appear in the open literature including Donovan et al. (1972), Edwards et al. (1955), Chung (1982), Nagelschmidt (1956), and Pickard et al. (1985). All these tests indicate that reproducibility and accuracy of the results were not within the desirable tolerance limits, which is why there has been so much effort applied to improving the analysis methods.


1. In this review the term "silica" will be used to imply a substance with the chemical composition SiO2 or SiO2.xH2O. "Crystalline" and "amorphous" will indicate the X-ray diffraction response to the material, where amorphous means no diffraction pattern is observed. The term "free" has no meaning when applied to crystalline silica and will not be used.