(a) Production

Titanium dioxide is the principal white pigment used commercially, due to its high refractive index, its ease of dispersion into a variety of matrices, and its inertness towards those matrices during processing and throughout product life (Considine, 1974).

Two main processes exist for making titanium dioxide pigments: the sulfate process and the chloride process. The sulfate process, the older of the two, was first used in Europe and the USA around 1930 and was the primary process until the early 1950s, when the chloride process was developed. By 1981, the chloride process accounted for approximately 78% of US production and by 1982 for 68% of world titanium dioxide pigment production (Considine, 1974; Lynd & Lefond, 1983).

In the sulfate process, rutile or anatase titanium dioxide is produced by digesting ilmenite (iron titanate) or titanium slag with sulfuric acid. The major concern in selecting the starting material for use in the process is that it contain as little as possible of impurities such as chromium, vanadium, manganese, niobium and phosphorus, which impair pigment properties. Ilmenite containing as little as 40% titanium dioxide can be used to produce pigment-grade titanium dioxide. Titanium slag may also be used as the starting material. Slag is produced by smelting ilmenite in an electric furnace and typically contains about 70% titanium dioxide, although concentrations may reach 85% (Considine, 1974; Lowenheim & Moran, 1975; Lynd & Lefond, 1983). Some typical materials used for pigment production throughout the world are shown in Table 1.

Table 1

Composition of typical commercial ilmenite concentrates and titaniferous slag (weight percent).

The sulfate process is a batch process in which concentrated sulfuric acid is added to ground ilmenite or titanium slag in proportions of 1.5:1 (acid to ore). An organic flocculent or antimony oxide may be added to induce aggregation of suspended titanyl and iron sulfates into a solid porous cake. The cake is dissolved in a dilute acid solution to release the sulfate agglomeration into solution. If necessary, scrap iron is added to reduce iron [III] to iron [II]. Also during this step, small amounts of titanium [IV] are reduced to titanium [III] to prevent later oxidation of iron [II]. The solution is clarified by settling and filtration. The resulting mother liquor is concentrated and subjected to steam for 6 h. Seed crystals may be added to aid nucleation. About 95% of the titanium in the mother liquor is hydrolysed to titanium hydrate or metatitanic acid (H₂TiO₃), which is collected on a filter and washed. The final filter cake is calcined at 900–1000°C to form titanium dioxide. The product is ground, quenched and dispersed in water; the coarse particles are separated in the thickener, re-ground and filtered; and the cake is dried in a rotary steam dryer and pulverized. The resulting product is anatase titanium dioxide. The rutile form is made by seeding the mother liquor with rutile seed crystals and conditioning the precipitated pigment with phosphates, potassium, antimony, aluminium or zinc compounds prior to calcination. The recovery of pigment-grade titanium dioxide in sulfate process plants is approximately 80% (Lowenheim & Moran, 1975; Lynd & Lefond, 1983; Lynd, 1985).

The chloride process is a continuous process that requires ores with a high content of titanium dioxide or concentrates such as natural or synthetic rutile. Natural rutile contains approximately 95% titanium dioxide; synthetic rutile or ilmenite (iron titanate) concentrates must have a minimum titanium dioxide content of 60% to produce economical yields of pigment in this process. The titanium dioxide content in ilmenite may be increased by reducing iron to its elemental form, followed by chemical or physical separation. Another method of enrichment is reducing iron [III] to iron [II] and chemically leaching it out of the mineral. A third method involves prior selective chlorination to remove iron and other impurities (Considine, 1974; Lowenheim & Moran, 1975; Lynd & Lefond, 1983).

In the chloride process, ore is ground and mixed with coke in a fluidized or static bed reactor and chlorinated at temperatures of 850–1000°C. Titanium tetrachloride is produced, along with chlorides of impurities present in the starting material, which include chlorides of iron, vanadium and silicon; these are removed chemically and through fractional distillation. Hydrogen chloride and carbon dioxide are present after chlorination and are vented prior to fractional distillation. Conversion to titanium dioxide is accomplished by burning titanium tetrachloride with air or oxygen at temperatures of 1200–1370°C. The resulting fine-grained oxide is sometimes calcined at about 500–600°C to remove any residual chlorine or hydrogen chloride. These gases are separated, and the chlorine is collected and recirculated to the chlorinator. Approximately 90% of the chlorine may be recycled. Aluminium chloride is added to titanium tetrachloride to assure near-total conversion to rutile titanium dioxide. A typical yield from this process is 90% (Lowenheim & Moran, 1975; Lynd & Lefond, 1983; Lynd, 1985).

Current worldwide demand for titanium dioxide is about 2.8 million tonnes (Anon., 1988a,b,c). Production volumes in 1978–86 in several countries are given in Table 2.

Table 2

Titanium dioxide pigment production by country in 1978–86 (thousand tonnes).

(b) Use

The principal use of titanium dioxide is as a whitening and opacifying agent in paints, varnishes, lacquers, paper, plastics, ceramics, rubber and printing ink (Table 3). These industries accounted for 93–95% of the titanium dioxide pigment used in the USA during 1982–86 (Mannsville Chemical Products Corp., 1983; Lynd & Hough, 1986).

Table 3

Percent distribution of titanium dioxide pigment used within US industries, 1982–86.

Titanium dioxide is the most common white synthetic pigment used in the paint industry. Modern paint plants often handle titanium dioxide as a slurry for convenience and to avoid airborne particulates (Schurr, 1981). In the paper industry, both anatase and rutile titanium dioxides are used to improve opacity. The total amount of pigment varies with the grade of paper but may comprise 2–40% of the final sheet, of which as much as 25% may be titanium dioxide. The pigment is dispersed during the pulping process and retained in the sheet during formation with organic polymers. The optical efficiency of titanium pigment is improved by synthetic silicas and silicates (Baum et al., 1981). Paper is waterproofed by coating the sheets with polyethylene resins whitened with titanium dioxide (Locker, 1982).

Titanium oxide is incorporated into various plastic products to confer opacity and whiteness, and also because it resists degradation by ultraviolet light and is chemically inert (Lynd, 1985).

Commercial use of anatase titanium dioxide as an opacifier for ceramic and enamel products began in 1946. Rutile pigments may also be used for this purpose (Friedberg, 1980; Lynd, 1985).

Titanium dioxide is also incorporated into rubber tyres to make whitewall tyres (Lynd, 1985). In printing inks, titanium dioxide is mixed with coloured pigments to add opacity or lighten the hue. In a typical nitrocellulose ink formulation, titanium pigment is present at 35 wt%. An acrylic-based ink formulation contained 30 wt% titanium dioxide (Burachinsky et al., 1981).

Other uses for titanium dioxide are as a catalyst in the production of alcohol fuels (Klass, 1984), as a delusterant in a variety of synthetic fibres at up to 2 wt% (Davis & Hill, 1982), as a component of flame-retardant formulations for wood (Wegner et al., 1984), in cosmetics as a physical sunscreening agent (Isacoff, 1979), as a component of gums, resins and waxes used for making dental impressions (Paffenbarger & Rupp, 1979), and as a pigment in floor coverings, leather products and soaps (Mannsville Chemical Products Corp., 1983).

(c) Regulatory status and guidelines

Occupational exposure limits for titanium dioxide in 13 countries or regions are presented in Table 4.

Table 4

Occupational exposure limits for titanium dioxide.

Titanium dioxide may be used as a food colour additive with the following specifications limiting impurities: antimony compounds, < 100 mg/kg; zinc compounds, < 50 mg/kg; soluble barium compounds, < 5 mg/kg; and hydrochloric acid-soluble compounds, <3.4 g/kg (Commission of the European Communities, 1962).

(a) Natural occurrence

Titanium dioxide occurs naturally in three crystalline forms: anatase, rutile and brookite. Only anatase and rutile are of commercial importance. Rutile can be mined directly, but anatase is obtained through the processing of ilmenite, a natural iron titanate (Lynd, 1985).

Rutile is a widespread accessory mineral found in high-grade metamorphic gneisses and schists and in igneous rocks. It also occurs in black sand deposits in many parts of the world. Typically, the composition of rutile is 95% titanium dioxide, the remaining 5% being silicon, chromium, vanadium, aluminium and iron oxides (Lynd & Lefond, 1983).

Ilmenite is a common accessory grain in igneous rocks such as anorthosites, gabbros and basic lavas; it may also occur as an intergrowth in magnetite and haematite. Weathering and alterations along grain boundaries result in the removal of iron from the ilmenite lattice, producing leucoxene or pseudorutile. Ilmenite is also found in sand deposits (Lynd & Lefond, 1983).

The principal producing countries for rutile and ilmenite are Australia, Brazil, Canada, China, Finland, India, Malaysia, Mexico, Norway, Sierra Leone, South Africa, Sri Lanka and the USA (Lynd & Lefond, 1983).

(b) Occupational exposure

On the basis of a US National Occupational Exposure Survey, the National Institute for Occupational Safety and Health (1983) estimated that 1 270 000 workers were potentially exposed to titanium dioxide in the USA in 1981–83.

Although it is apparent that occupational exposure to titanium dioxide is extensive, there are few data on levels and sources of exposure (Santodonato et al., 1985), and the data available in the literature are reported as total dust or nuisance dust and not as titanium dioxide. Concentrations ranged from 10 to 400 mg/m³ during the grinding of titanium dioxide pigment, but documentation of these levels was not provided (Elo et al., 1972). Long-term exposures to titanium dioxide dust in a titanium pigment production factory occasionally exceeded 10 mg/m³, and exposures greater than 10 mg/m³ were common during the repair of production machinery (Rode et al., 1981).

(a) Oral administration

Mouse: Groups of 50 male and 50 female B6C3F1 mice, five weeks old, were fed diets containing 0, 2.5% or 5.0% titanium dioxide (anatase; purity, ≥98%) daily for 103 weeks. Mice were killed at 109 weeks of age, one week after exposure was stopped. At terminal sacrifice, there was no significant difference in survival between treated and control males: 32, 40 and 40 controls, low-dose and high-dose males were still alive at that time. In female mice, a dose-related trend for decreased survival was significant in treated groups (p = 0.001, Tarone test); survival at terminal sacrifice was 45, 39 and 33 among control, low-dose and high-dose females, respectively. No difference in body weights between treated and control groups and no significant increase in the incidence of tumours was observed in treated mice of either sex (National Cancer Institute, 1979).

Rat: Groups of 50 male and 50 female Fischer 344 rats, nine weeks old, were fed diets containing 0, 2.5% or 5.0% titanium dioxide (anatase; purity, ≥98%) daily for 103 weeks. Rats were killed at 113 weeks of age, one week after exposure was stopped. There was no significant difference in survival between treated and control groups of either sex; survival of males at terminal sacrifice was 31, 37 and 36 among control, low-dose and high-dose groups, respectively, and that of females was 36, 36 and 34 in control, low-dose and high-dose groups, respectively. No difference in body weights between treated and control groups and no significant increase in the incidence of tumours was observed in treated mice of either sex (National Cancer Institute, 1979).

(b) Inhalation/intratracheal administration

Rat: Groups of 50 male and 50 female Sprague-Dawley rats, eight weeks of age, were exposed by inhalation to 0 or 15.95 mg/m³ titanium dioxide [purity unspecified] for 6 h per day on five days per week for 12 weeks. After 140 weeks (128 weeks after the end of exposure), all surviving rats were sacrificed. Average survival was 116 and 113 weeks for control and treated males and 114 and 120 weeks for control and treated females, respectively. At terminal sacrifice, 39 control males, 44 treated males, 45 control females and 45 treated females were still alive. No difference in body weights between treated and control groups and no significant increase in the incidence of tumours was observed in treated mice of either sex (Thyssen et al., 1978). [The Working Group noted the short duration of exposure and the relatively low exposure level.]

Groups of 100 male and 100 female CD rats, five weeks of age, were exposed by inhalation to 0, 10, 50 or 250 mg/m³ titanium dioxide (rutile; 99% pure; 84% of dust particles of respirable size) for 6 h per day on five days per week for two years. At three, six and 12 months, five, five and ten rats of each sex at each dose, respectively, were removed for interim kills. No difference in mortality, body weight or clinical signs was observed. Nasal cavities were examined histologically, and no tumour was observed. Lung tumours were observed primarily in high-dose rats of each sex. The incidences of lung adenomas were: males - control, 2/79; low-dose, 1/71; mid-dose, 1/75; and high-dose, 12/77; females - control, 0/77; low-dose, 0/75; mid-dose, 0/74; and high-dose, 13/74. The incidences of cystic keratinizing squamous-cell carcinomas were: males — 0/79, 0/71, 0/75 and 1/77; females - 0/77, 1/75, 0/74 and 13/74. One anaplastic carcinoma occurred in a low-dose male. The lung tumours occurred in the bronchioloalveolar region, and no evidence of metastasis was observed. The authors noted difficulty in distinguishing between the squamous-cell carcinomas and keratinizing squamous metaplasia (Lee et al., 1985a,b, 1986).

Hamster: Groups of 24 male and 24 female Syrian golden hamsters, six to seven weeks old, received intratracheal administrations of 0 or 3 mg titanium dioxide ([purity unspecified] particle size, 97% < 5 µm) in 0.2 ml saline once a week for 15 weeks. Animals were observed until spontaneous death and all control and treated hamsters died by 120 and 80 weeks, respectively, after the beginning of treatment. The respiratory tract and other organs with gross lesions were examined microscopically; no respiratory tract tumour occurred in treated hamsters, but two tracheal papillomas were found in untreated controls (Stenbäck et al., 1976).

(c) Subcutaneous injection

Rat: Groups of 20 male and 20 female Sprague-Dawley rats, 13 weeks old, received a single subcutaneous injection of 1 ml saline or 30 mg of one of three preparations of titanium dioxide (≥99%, ≥95% or ≥85% pure) in 1 ml saline into the flank. All rats were observed until spontaneous death, which occurred as late as 136, 126, 146 and 133 weeks in the control and titanium dioxide-treated groups, respectively. No tumour was observed at the site of the injection in any group (Maltoni et al., 1982). [The Working Group noted the inadequate reporting of the study.]

(d) Intraperitoneal injection

Mouse: Groups of 30 or 32 male Marsh-Buffalo mice, five to six months old, received a single intraperitoneal injection of 0 or 25 mg titanium dioxide (purity, ≥98%; manually ground) in 0.25 ml saline. All survivors (ten control and 13 treated mice) were killed 18 months after treatment. No difference in the incidence of local or distant tumours was observed between treated and control animals (Bischoff & Bryson, 1982). [The Working Group noted the small number of animals used.]

Rat: As part of a large study on various dusts, three groups of female Wistar rats, nine, four and five weeks of age, received intraperitoneal injections of granular titanium dioxide [purity unspecified] in 2 ml 0.9% sodium chloride solution. The first group received a total dose of 90 mg per animal by five weekly injections; the second group received a single injection of 5 mg per animal; and the third group received three weekly injections of 2, 4 and 4 mg per animal. One concurrent group of 32 five-week-old Wistar rats received a single injection of saline alone. Average lifespans were 120, 102, 130 and 120 weeks. No intra-abdominal tumour was reported in 47 and 32 rats from the second and third groups that were examined; six of 113 rats examined from the first group had sarcomas, mesotheliomas or carcinomas of the abdominal cavity [numbers unspecified]. Two controls had abdominal tumours. In a similar experiment with female Sprague-Dawley rats given single intraperitoneal injections of 5 mg per animal titanium dioxide, 2/52 rats developed abdominal tumours (average lifespan, 99 weeks). Controls were not available for comparison (Pott et al., 1987).

(e) Administration with known carcinogens

Hamster: Groups of 24 male and 24 female Syrian golden hamsters, six to seven weeks old, received intratracheal administrations of 3 mg titanium dioxide ([purity unspecified] particle size, 97% < 5 µm) plus benzo[α]pyrene in 0.2 ml saline, or benzo[α]pyrene in saline (controls) once a week for 15 weeks. Animals were observed until spontaneous death; all control and treated hamsters had died by 100 and 70 weeks, respectively. In the 48 hamsters [number of tumours per sex unspecified] treated with titanium dioxide plus benzo[α]pyrene, tumours occurred in the larynx (11 papillomas, five squamous-cell carcinomas), trachea (three papillomas, 14 squamous-cell carcinomas, one adenocarcinoma) and lung (one adenoma, one adenocarcinoma, 15 squamous-cell carcinomas, one anaplastic carcinoma). Two papillomas occurred in the larynx of benzo[α]pyrene-treated controls. In the same study, ferric oxide and benzo[α]pyrene induced a similar spectrum of tumours as that induced by the combination with titanium dioxide; aluminium oxide and benzo[α]pyrene produced no increase in tumour incidence compared with benzo[α]pyrene-treated controls (Stenbäck et al., 1976).

(a) Experimental systems

The toxicology of titanium dioxide has been reviewed (World Health Organization 1982).

(b) Humans

Overall evaluation

Titanium dioxide is not classifiable as to its carcinogenicity to humans (Group 3).