Quaternary marine sediment in the Gulf of Maine basins contains 0.7 to 1.0 wt percent TiO₂ (determined by X-ray fluorescence spectrometry). Most of this TiO₂ exists in the form of silt-size rutile crystals that are visible by using the petrographic microscope with transmitted light (Valentine and Commeau, 1990). The identification of rutile was confirmed by using a scanning electron microscope (SEM) equipped with an energy-dispersive X-ray spectrometer (EDS) system. To quantify the amount of TiO₂ in the sediment contributed by rutile and its polymorphs, anatase and brookite, it was necessary to eliminate as many of the other minerals as possible, especially titanium-bearing minerals such as ilmenite, ilmenomagnetite, biotite, hornblende, oyroxene, and sphene. We accomplished this by developing a method of chemical dissolution that removed the bulk of the raw material and left the TiO₂ minerals intact.

Many methods using acids and bases have been developed over the years to dissolve rocks, minerals, and sediments for chemical analysis or to concentrate specific minerals (Dolcater et al., 1970; Church, 1971; Campbell, 1973, and the references cited therein). The method developed by Raman and Jackson (1965) to concentrate rutile and anatase in soils and sediments requires digesting the sample in concentrated hydrofluoric acid (HF) for 24 hours. However, Campbell (1973) found that digestion in HF for more than 2 hours results in a loss of anatase. The method of Dolcater et al. (1970) for concentrating titanium as a free oxide requires the use of hydrofluotitanic acid (H₂TiF₆), which is difficult to find on the commercial market. The acid can be prepared by the reaction of concentrated HF (48%) with an excess of TiO₂, but the procedure requires 36 hours to complete and should be attempted with caution because it is highly exothermic. Jackson (1979) provides a detailed method for digesting soils, but many of the recommended pretreatment steps employ sodium compounds such as sodium bicarbonate (NaHCO₃), sodium citrate (Na₃C₆H₅O₇-2H₂O), and sodium dithionate (Na₂S₂O₄), which are used for the removal of calcium carbonate, iron oxides, and phosphates. In combining the methods of Dolcater and Jackson, sodium compounds must be thoroughly washed from the sample because they form sodium fluotitanate (Na₂TiF₆) in the presence of hydrofluotitanic acid (Fig. 1K). French and Adams (1973) described an inexpensive method for decomposing silicates by HF digestion in polypropylene containers. Their technique was effective for a wide variety of rock types. However, it did not address the problem caused by the precipitation of insoluble fluorides, nor did it outline a procedure to concentrate any residue that remained.

As no one method gave the results we required, we modified procedures described in the literature and developed a process that removes 96 to 98 wt percent of the raw sample material. The residue is composed of rutile and minor amounts of micro- and cryptocrystalline TiO₂ (Fig. 1A-J), barite (Fig. 1L), elemental carbon (coal), and insoluble fluorides (Fig. 1J). The fluorides precipitate during the decomposition of siliceous material in hydrofluoric acid (e.g., MgF₂ and MgAlF₅-2.7H₂O; Lanmyhr and Kringstad, 1966).

Most of the sample analyzed by the method described were marine muds collected from the Gulf of Maine (Valentine and Commeau, 1990). The silt and clay fraction (up to 99 wt% of the sediment) is composed of clay minerals (chiefly illite-mica and chlorite), silt-size quartz and feldspar, and small crystals (2-12 um) of rutile and hematite. The bulk sediment samples contained an average of 2 to 3 wt percent CaCO₃. Tiher samples analyzed include red and gray Carboniferous and Triassic sandstones and siltstones exposed around the Bay of Fundy region and Paleozoic sandstones, siltstones, and shales from northern Maine and New Brunswick. These rocks are probable sources for the fine-grained rutile found in the Gulf of Maine.