Tiron Dissolution Method Used to Remove and Characterize Inorganic Components in Solls Hideomi Kodama* and G

1180 SOlL SCI. SOC. AM. I., VOL. 55, JULY-AUGUST 1991

Wada, K., T. Henmi, N. Yoshinaga, and S.H. Patterson. 1972. Im- Zealand. Aust. J. Soi! Res. 19:185-195. ogolite and allophane formed in saprolite ofbasalt on Maui, Ha- Shoji, S., and Y. Fujiwara. 1984. Active aluminum and iron in the waii. Clays Clay Miner. 20:375-380. humus horizons of Andosols from northeastern Japan: Their Wada, K., and T. Higashi. 1976. The categories ofaluminium- and forms, properties, and significance in clay weathering. Soil Sei. iron-humus complexes in ando soils determined by selective dis- 137:216-226. solution. J. Soil Sei. 27:357-368. Shoji, S., and J. Masui. 1971. Opaline silica ofrecent volcanic ash Wada, K., and Y. Tokashiki. 1972. Selective dissolution and dif- soils in Japan. J. Soil Sei. 22:101-108. ferential infrared spectroscopy in quantitative mineralogical Shoji, S., and J. Masui. 1972. Amorphous clay mineral of recent analysis of volcanie-ash soil clays. Geoderma 7:199-213. volcanic ash soils. Part 3. Mineral composition of fine clay frae- Wada, S.-I., A. Eto, and K. Wada. 1979. Synthetic allophane and tions. J. Sei. Soil Manure Jpn. 43:187-193. imogolite. J. Soil Sei. 30:347-355. Shoji, S., and M. Saigusa. 1978. Occurrence oflaminar opaline silica Wada, S.-I., and K. Wada. 1980. Formation, composition, and strue- in some Oregon Andosols. Soil Sei. Plant Nutr. 24:157-160. ture ofhydroxy-aluminosilicate ions. J. Soil Sei. 31:457-467. Tait, J.M., N. Yoshinaga, and B.D. Mitchell. 1978. The occurrence Wada, S.-I., and K. Wada. 1981. Reactions between aluminate ions ofimogolitein some Scottish soils. SoilSei. Plant Nutr. 24:145-151. and orthosilieic aeid in dilute alkaline to neutral solutions. Soil Tokashiki, Y., and K. Wada. 1975. Weathering implications ofthe Sei. 132:267-273. mineralogy of clay minerals oftwo ando soils, Kyushu. Geoderma Wang, c., and H. Kodama. 1986. POOogenic imogolite in sandy 14:47-62. Brunisols ofEastern Ontario. Can. J. Soil Sei. 66:135-142. Violante, P., and J.M. Tait. 1979. Identification ofimogolite in some Wang, C., JA McKeague, and H. Kodama. 1986. POOogenic im- volcanic soils from Italy. Oay Miner. 14:155-158. ogolite and soil environments: Case study ofSpodosols in Quebec, Wada, K. 1978. Allophane and imogolite. Dev. Sedimentol. Canada. Soil Sei. Soc. Am. I. 50:711-718. 26:147-187. Wells, N., C.W. Childs, and C.J. Downes. 1977. Silica Springs, Ton- Wada, K. 1980. Mineralogical characteristics of Andisols. p. 87- gariro National Park, New Zealand - analyses of the spring water 107. In B.K.G. Theng (00.) Soils with variable charge. New Zea- and characterization of the aluminosilicate deposit. Geochim. land Soc. Soil Sei., Lower Hutt. Cosmochim. Acta 41:1497-1506. Wada, K. 1989. Allophane and imogolite. p. 1051-1087. In J.B. White, J.L. 1971. Interpretation ofinfrared spectra ofsoil minerals. Dixon and S.B. Weed (00.) Minerals in soil environments. 2nd Soil Sei. 112:22-31. 00. SSSA, Madison, WI. Yoshinaga, N., and S. Aomine. 1962. Imogolite in some ando soils. Wada, K., and D.J. Greenland. 1970. Selective dissolution and dif- Soil Sei. Plant Nutr. (Tokvo) 8:22-29. ferential infrared spectroscopy for characterization of "amor- Young, A.W., A.S. Campbell, and T.W. Walker. 1980. Allophane phous" constituents in soils clays. Oay Miner. 8:241-254. isolated from a podzol developed on a non-vitric parent material. Wada, K., and M.E. Harward. 1974. Amorphous clay constituents Nature (London) 284:46-48. of soils. Adv. Agron. 26:211-260.

Tiron Dissolution Method Used to Remove and Characterize Inorganic Components in Solls Hideomi Kodama* and G. John Ross

ABSTRACT moving these components from crystalline components and determining their overaJl composition. Treatments using NaOH and Na]COJ have been widely used to In the past 80 yr, a number of chemicaJ dissolution remove poorly crystalline and noncrystalline aluminosUicates and methods have been proposed, modified, and im- hydrous oxides of Si and Al. However, NaOH treatment is orten too proved. Based on chemicaJ reagents used and chemical harsh and Na]COJ treatment is usually too mild and must be ce- reactions involved, these methods can be c1assified peated. Treatment using e)kaline Tiron (4,5-dihydroxy-l,3-benzene- into one or more of the following four categories: (i) disulfonic acid (disodium saltI, CJI.NR]OaSJ wasas effective as aIkaline, (ii) acid, (iii) complexing, and (iv) reducing. acid oxalate in the dissolution of allODbll1llll!..imOl!olite. and ooorly Table 1 shows some major methods c1assifiedin these qystalline bvdrous Fe oxides. In contrast to oxalate, Tiron effec- categories. The extraction capacity and selectivity of tively removed oDaline sUica. and the difference between extracted these dissolution methods are c10sely related to the Si could be used to measure the amount of opaline sUica. Unlike mineral components to be extracted (Kodama and oxalate, Tiron dissolved little mametite. As sJbbsite is partially dis- Jaakkimainen, 1982). For best efficiency, therefore, it solved bv Tiro~ the rates or Al release by gibbsite «214m) dissolution at different temperatures were used to estimate the gibbsite is essentiaJ to select the dissolution method(s) appro- contents of soil days containing this mineral. Generally, a single priate to the components to be extracted. treatment with Tiron appears to be sufficlent to mildly but effectively Poody crystalline and noncrystaJline inorganic com- clean up sampIes ror x-ray diffraction of crystalline minerals with ponents in soils are mainly hydrous oxide compounds amounts of extracted Si, Al, and Fe indicating the presence and of Si, Al, and Fe (Mn) and hydrous aJuminosilicates quantities or poorly crystalline and noncrystalline aluminosUicates such as allophane and imogolite. Table 2 lists these as weil as hydrous oxides of Fe, Al, and Si, including opaline silica. minerals with some oftheir characteristics. Among the many dissolution methods summarized in Table 1, NaOH, Na2C03, oxaJate, and dithionite-citrate are perhaps most frequently used to remove the hydrous oxide and aJuminosilicate minerals listed in Table 2. Hydroxylamine and Tiron are aJso applied, but less HE CHARACTERIZAnON of poorly crystaJline and T noncrystalline inorganic soil components is of frequently. Since the Tiron method was originally pro- considerable interest in view of their strong influence posed by Biermans and Baert (1977), it has been ex- on the physicaJ, chemical, and biological activities of tensively applied in our laboratory (Wang et al., 1981; soils. ChemicaJ extractions are prerequisite for re- Kodama and Jaakkimainen, 1982; McKeague et al., 1983; Wang et al., 1986; Kodama and Wang, 1989; Land Resource Research Centre, Research Branch, Agriculture Can- Kodama et aJ., 1989). This is because we have found ada, Ottawa, ON, KIA OC6 Canada. LRRC Contribution no. 90- 16. ReceivOO 27 Mar. 1990. .Corresponding author. Abbreviations: AA, atomic absorption; DCB, dithionite-citrate-bicarbonate; XRD, x-ray diffraction. Published in Soil Sci. Soc. Am. J. 55:1180-1187 (1991). 1181 KODAMA & ROSS: INORGANIC COMPONENTS CHARACfERIZED BY TIRON DISSOLUTION

application. The sources of these reference minerals and soil Table 1. Major chemical dissolution methods. sampies are listed in Table 3. The materials to be fraction- Chemical reagents ated were suspended in water and the day fractions (< 2#Lm) used for were separated by a sedimentation method without centrif- Category dissolution Selected references ugation (Jackson, 1968). The pretreatment with H202 to de- Alkaline NaOH Foster (1953) stroy organic matter was applied only to soil sampies, Hashimoto and Jackson (1960) allophane, and imogolite prior to suspension. Fractionated Na2CO. Mitchell and Farmer (1962) Follet et aI. (1965) clay fractions of five soil sampies were saturated with Mg2+ and freeze-dried. The day fractions of other sampies were Acid Ha Deb (1950) not saturated with Mg2+before freeze-drying. For consoli- Segalen (1968) Alkaline plus NaOHlHa dated sampies such as hematite, magnetite, goethite, gibb- acid site, lepidocrocite, opal, and obsidian, the mineral sampies Complexing U.C2O. Chester and Hugbes (1967) were first crushed to sand size and pure mineral grains were H2c.O. Ball and Beaumont (1972) selected by handpicking. (NH.)2C2OJH2c.O. Tamm (1922) After pulverizing by a Spex MixerJMill (Spex Industries, Schwertmann (1959) Metuchen, NJ), powdered sampies were passed through a Na2C2OJH2c.O. Heumi and Wada (1976) c.u.Na20.S. (Tiron) Biermans and Baert (1977) 50-#Lmnylon sieve. When further fractionation was required, Na.P20, McKeague (1967) as in the case of gibbsite, goethite, and hematite, the sampies NH2OH.HClIHa Chao and Zhou (1983) were suspended in water and fractionated using a sedimen- Reducing tation method without centrifugation. The fine-silt (2-5 #Lm) Reducing plus Na,CJisO,lNaHCO,J Mebra and Jackson (1960) and day «2 #Lm)fractions were freeze-dried. All other frac- complexin& Na2S.0. K2C2OJH2C2OJMg Jeffiies (1946) tionated sampies were dried at 60°C in an oven. The ground (NH.)2C2OJH2C2OJ Ducbaufour and Souchier (1966) silicate minerals were prepared by dry grinding for 96 to 180 Na.g.O. h using a mechanical grinder. This procedure was followed to obtain partially noncrystalline specimens of known compositions (Kodama and Jaakimainen, 1982). Before use, the distinctive merit in this method, inc1uding a high ef- ground sampies were dried overnight at 110°C in an oven. ficiency equivalent to that obtained by a combination Synthetic materials and two ferrihydrite sampies were used of other extraction methods. Therefore, several ex- as obtained or received, since they were already in a fine traction procedures may simply be replaced by a single powdery state. extraction with Tiron. The objective ofthis study was The Tiron method used in this investigation was modified to compare Tiron and other common extraction meth- from the method originally proposed by Biermans and Baert ods on a wide variety of materials. (1977). Detailed procedures are as folIows. First 31.42 g of Tiron (Sigma no. D-7389, Sigma Chemical Co., St. LQuis, MATERIALS AND METHODS MO) was dissolved in approximately 800 mL of H20 in a plastic beaker. The resulting solution was acidic (pH ~5.5) A wide variety of crystalline, poorly crystalline, and non- and light brownish yellow. A l00-mL Na2C03 solution con- crystalline minerals were selected to test the suitability of taining 5.3 g of anhydrous Na2C03 was prepared in a plastic Tiron as an extractant for Si, Al, and Fe. Selected soil sam- beaker and added to the Tiron solution while stirring. The pIes were also employed to provide examples of practical

Table 2. Poorly crystaUine and noncrystalline minerals occurring in solls. Cbaracteristics Reference ldealized cbemical Structural cbaracteristics Electron microscopic and parameters X-ray diffraction pattern Minera1 formula Mostly spherical particles with 1,2 A DUijorband with a maximum near Amorphous 0.1-0.85 pm diameter Opaline Silica SiO.'IIH2O 0.4 um. Mostly spherical particles with 3 Amorphous A DUijorband near 0.4 nm with OpaI-A OO2'"H'<> diffuse maxima at 0.2, 0.15 and 0.1-0.85 pm diameter 0.12 um. Similar to opaline silica 3 Amorphous plus poorly A DUijorband near 0.41 um with 0paI-CT Si02."H2O crystaI1ine cristobalite minor bands corresponding to andlor tridymite cristobalite andlor tridymite Spherical 4 Rbombohedra1 Broad peaks at 0.25,0.22,0.197, Ferrihydrite Fe.O,:2FeOOH 0.171 and 0.15 nm. + 2.6H2O IJ, - 0.508um c um - 0.94 Acicular, possibly thin rolled plates 4 Hexagonal Broad peaks at 0.25,0.22,0.169 and Feroxyhite FeOOH 0.225 um. (3'-AlOOH) /J - 0.295 um c - 0.453 um Platy, mostly irregularly out1ined 5 A broad hump at 0.97 um. Wad Mn02."H2O Amorphous particles 6,7 Hollow spberu1es or polyhedrons Amorphous Broad bands at 0.33 and 0.225 nm. Allopbane 2Si02'Al20""H,0 with diameter of 3.5-5 nm 8,9 Amorphous Broad bands at 0.44, 0.26 and 0.15 Hisingerite 2SiO,'Fe20.."H2O um. 10 Two-dimensionaltabuIar Major peaks (broad) at 1.6,0.79, Smooth and curved threads varying 11 Imogolite SiO,'Al2O,'2.5H,0 structure and 0.33 um with less pronounced in diameter from 10 to 30 nm and 12 peaks at 0.56, 0.44, 0.37, 0.25, extendingseveralmicrometersin b 0.51 nm length. - 0.21 and 0.14 um. c - 0.84um (1957); 6-Wada (1977); 7- (1980); 5-Hariya and Harada t l-Lanning et 81. (1958); 2-Jones et aI. (1964); 3-Levin and Ott (1933); 4-Carlson and Schwertmann Yoshinaga and Aomine (1962); 8-Sudo and Nakamura (1952); 9-Kohyama and Sudo (1975); 10- Yoshinaga (1968); ll-Cradwick et aI. (1972); 12-BayJiss (1987). 1182 SOlL SCI. SOC. AM. J., VOL. 55, JULY-AUGUST 1991 pH of the Tiron solution rose to about 10.2 and the color be required). The total volume of the Tiron solution was then made up to 1 L. After standing for "",0.5h, the solution ofthe solution tumed greenish. Increments ofO.1 to 0.2 mL changed its color from greenish to light brownish yellow. of 4 M NaOH were added until the pH reached 10.5 (ac- The pH was adjusted to pH 10.5 since it may drop during cording to OUfexperience, about 23.5 mL of 4 MNaOH may storage. The solution was stored in a polypropylene con- tainer and kept in a refrigerator, allowing it to be used for Table 3. Source of natural and synthetic reference minerals and up to about 3 mo without a noticeable decline of its extrac- selected son sampIes. tion capacity. For extraction, 25 mg of a dry specimen was Aluminum bydroxides weighed in a 50-mL polypropylene centrifuge tube; 30 mL Gibbsite Richmond, Massachusetts, USA of the Tiron solution was then added to the tube and Boehmite (synthetic) Alcan Co. weighed. (Larger sampies may be used in the same solid/ Pseudoboehmite (synthetic) Land Resource Research Centre (LRRC), solution ratio in larger containers.) The tube was covered Ottawa loosely with a polypropylene cap, and placed in a water bath NoncrystalIine "Al(OH);' LRRC, Ottawa at 80°C for 1h. During the course of extraction, the contents (synthetic) were occasionally stirred by shaking manually. At the end Iron oxides and hydroxides of the I-h extraction, the tube was cooled in a cold-water Hematite Iron ore mine, Quebec, Canada bath, the cap removed, the outside of the tube dried, and Magbemite, (synthetic) Energy, Mine and Resources Canada the contents weighed. Water was added according to weight Magnetite Ishpeming, MI loss due to evaporation. The contents were well mixed and Goethite Diwabik, MN min at 17 600 g. A 20-mL sampie of the LRRC, Ottawa centrifuged for 1° Akaganäte (synthetic) supematant was siphoned off for analysis of extracted Si, Al, Lepidocrocite Giessen, Germany Ferrihydrite (S5) Supplied by Dn. Schwertmann and and Fe by the AA spectrophotometric method. It has been Murad proven in our laboratory that a reliable analysis can be obtained with or without destruction of Tiron prior to AA Silicates for ground sampies analysis. If Si in the supematant is below the detection limit Kaolinite Macon, GA of the AA method, a colorimetric method must be applied Diotite BancroCt, Ontario, Canada Microcline pertb, Ontario, Canada after destroying Tiron by treating with 30% (w/w) H202 in Fe-clinocblore Chester, VT a water bath until the color disappears. In this study, an Mg-bastingsite Faraday, Ontario, Canada autoanalyzer method (Technicon Autoanalyzer 11)using mo- Poorly crystalline hydrous silica and aluminosilicates lybdate (Wang and Schuppli, 1986) was employed to deter- mine Si in the extracts after the destruction of Tiron. Opaline silica (synthetic) Silica gel, F"JSber's Chemical reagent For comparison, the following selective cheniical disso- Opal Arizona lution methods were also employed: NaOH (Hashimoto and Opal Hawaii Jackson, 1960), Na2C03(Follet etal., 1965), ammonium Obsidian Little Lake, CA oxalate (Schwertmann, 1959), sodium oxalate (Higashi and Allophane Kanuma, Tocbigi, Japan Ikeda, 1974), DCB (Mehra and Jackson, 1960),and hydroxylamine (Chao and Zhou, 1983; Ross et al., 1985). Imogolite Kurayosbi, Tottori, Japan X-ray diffraction analysis was done on oriented specimens using an automated Scintag PAD V diffractometer (Scintag, Selected soil samples Sunnyvale, CA) with Co radiation and a graphite monoch- Northem Qu&ec 82-116, E horizon, Typic Cryorthods pedont romator. Oriented specimens were prepared by drying a 1- Northern Quebec mL water suspension containing 30 mg clay on 30- by 25- 82-128, Ds2 borizon, Typic Cryorthods pedont mm glass slide. Northem Quebec 82-151, Dbs borizon, Typic Cryorthods pedont RESULTS AND DISCUSSION Eastem Ontario 84-2477A, E borizon, Typic Dystrocbrept pedon* Dissolution Selectivity 7850, C3 borizon, saprolite-underlain Humic Nova Scotia§ Haplorthod To distinguish between crystalline and noncrystalline substances, a method should selectively dissolve t Wang et al. (1986). Kodama and Wang (1989). the noncrystalline portion only. In addition, the chem- §* McKeague et al. (1983).

Table 4. Inßuence of the chemical composition of minerals on the extraction performance of various dissolution methods, as expressed by All Si and FelSi ratios of extracts. Original Hydroxylamine Tiron Oxalate atomic ratio NaOH NazCO, Ground minera1s AllSi FelSi Si Al AlISi Si Al AllSi Si Al Fe AllSi FelSi Si Al Fe AIISi FeISi Si Al Fe AllSi FelSi _gkg-I- _gq-'- -gq-'- _gkg-I- _gq-'-

Kaolinite 5 55 11.0 25 161 2 6.4 171 167 0.98 88 78 0.89 127 129 - 1.07 (144 b)t 0.99 -* Microc1ine 8838- 0.44 3 15 5.0 (160 h) 0.32 - 119 44 0.37 18 46 107 2.6 5.9 Diotite 84 41 93 0.49 1.11 5 17 38 3.4 7.6 (180 b) 0.33 0.91 67 24 0.36 59 20 0.34 Fe clinoch1ore 2 18 17 9.0 8.5 32 82 69 2.6 2.2 (132b) 0.62 0.51 54 39 0.72 48 19 0.40 113 79 62 0.70 0.55 Mg bastiogsite 19 48 0.26 0.66 3 14 29 4.7 9.7 (96 b) 0.29 0.61 27 7 0.33 73 t h -grindingbours. Not determined. * 1183 KODAMA & ROSS: INORGANIC COMPONENTS CHARACfERIZED BY TlRON DISSOLUTION

With the use of ammonium oxalate, there is the ica1 composition of noncrystaHine substances should possibility that NH~ might be fixed in expansible clay not influence the extraction performance of the meth- minerals, particularly in vermiculte, to cause struc- od. To test for such an influence on dissolution meth- tural contraction ofthese minerals. To avoid this pos- ods, Al/Si and Fe/Si ratios of extracts from the ground sibility, Higashi and Ikeda (1974) proposed the use of silicates were compared with those of the original ma- sodium oxalate in place of ammonium oxalate. Cod- terials (Table 4). The DCB method was not used in erre (1985, personal communication) illustrated the this comparison, because this method has a specific effect on vermiculite-bearing soil sampies by compar- selectivity for Fe and extracts noncrystalline and crys- ing their XRD patterns before and after treatments talline free Fe oxide compounds. As far as Al and Si with ammonium and sodium oxalate. Some of the proportions are concerned, the NaOH method worked ammonium oxalate treated sampies showed an in- weHfor the five mineral compositions. The results also crease in the 1.0-nm peak intensity at the expense of show that NaOH has a high extraction capacity for Si the l.4-nm peak intensity of the vermiculite compo- and Al from Si-Al minerals such as kaolinite and nent. Unlike the ammonium oxalate method, the Ti- microcline (Table 4). Recently, Kodama et al. (1989) ron method does not cause such structural alterations. suggested that the high amounts of Si and Al extracted from ground kaolinite might have been due to attack Dissolution 01Poorly Crystalline Silicate of NaOH on residual crystalline kaolinite. The ex- and Aluminosilicate Minerals traction by NazC03 showed a trend similar to that by NaOH, but the extraction capacity ofNazC03 was less The dissolution of poorly crystalline silicate min- than that of NaOH, although the NazC03 treatment erals, aHophane, and imogolite by Tiron was compared was repeated three times. Oxalate and hydroxylamine with the dissolution of these minerals by NaOH and showed a specificselectivity for Al and Fe; consequent- oxalate (Table 5). Tiron extracted nearly as much Si ly, Al/Si and Fe/Si ratios of extracts were much higher from opaline silica as did NaOH, whereas oxalate ex- than anticipated. The Tiron method showed balanced tracted almost no Si from it. Since the opaline silica capability in extraction capacity and selectivity for the contained about 180 g HzO kg-l, the amounts of Si mineral compositions tested (Table 4).From these data, extracted by the two methods corresponded closely to it is evident that an extraction methods except the Tiron the total amounts of Si from the opaline silica. Un- method generally require at least one more compli- doubtedly, the low pH (3.0-3.5) ofthe oxalate solution mentary extraction method to obtain reliable data for is the main reason for the low Si extraction. For the an overall composition of noncrystalline substances Arizona opal (opal-CT, Jones and Segnet, 1971), the commonly consisting of Si, Al, and Fe. Shortcomings of the methods other than the Tiron method may be 82-151 explained by the pH of the extraction media. No Fe is (<2j1111) extractable at the high pH of the NaOH and NazC03 solutions, whereas very little Si is extractable at the low pH of the oxalate and hydroxylamine. Cleaning Up Crystalline Minerals BEFORE TlRON TREATMENT There are always some less crystalline substances associated with clay-size crystalline minerals, and these substances are susceptible to selective chemica1 l~ dissolution. This raises the question of where to draw ~O)0.... ~ C') the line between weH and poorly crystalline sub- .... ci stances. The distinction may depend on the chemica1 dissolution method applied. In general, it is preferable to avoid attack on crystalline substances by the chem- ica1reagents used. ~ AFTER TlRON TREATMENT Figure 1gives an example ofhow the alkaline Tiron method effectively cleaned up soil clay sampies. The ~ ~ C') sampie, pretreated with HzOz and deferrated with di- .... -rci N C') thionite-citrate, contained 5.18% Si, 10.05% Al, and ci 0.87% Fe, which were extractable by Tiron. A 1.4-nm mineral, quartz, and albite were barely recognized in ~ci oe; the sampie before extraction with Tiron. After Tiron N extraction and subsequent saturation with Mgz+,an ci amphibole mineral and microcline as weH as quartz I~ ;::: and albite were definitely identified. With glycerol and ci heat treatment, the 1.4-nm mineral was identified as vermiculite. It is evident that Tiron removed coating 40 and cementing agents to disintegrate aggregates and 2 10' 20 30 expose more cleaned mineral particles. The drastic im- DEGREE29,CoK a provement in the XRD patterns can be attributed to better orientation of phyHosilicates and more exposure Fig. 1. X-ray diffraction patterns of a soll day (82-151) before and after Tiron treatment (CPS: counts per second). Spacings are in of other silicates, as weH as to the increased concen- nanometers. tration of crystalline components. 1184 SOlL SCI. SOC. AM. J., VaL. 55, JULY-AUGUST 1991

Table S. Amounts of major elements extracted from opaline silica, Table 7. Amounts of AI extracted from hydrous AI oxide minerals opal, allophane, and imogoIite by three different chemical disso- by four different chemicaI dissolution methods. lution methods. Amount of AI extracted NaOH Tiron Oxalate Mineral Origin Tiron Oxalate NaOH DCBt Si AI Fe Si AI Si AI Fe g kg-' gkg-' AIuminum Syntbetic 221 215 -* Opaline silica 397 379 1 hydroxide Opal, Arizona 278 trt 38 tr 0 0 (noncrystaIline) Obsidian, 73 13 53 10 8 Pseudoboehmite Synthetic 301 40 CaIüornia Boehmite Syntbetic 2 1 Richmond, 138 12 282 6 Opal, Hawaii 43 tr 50 0 tr 0 Gibbsite <2 "m AIlopbane, 114 165 SO 115 163 53 MA Richmond, 51 8 275 Kanuma, Japan Gibbsite 2-5 "m Imogolite, 73 103 12 68 96 14 MA Richmond, 23 6 230 Kurayoshi, Japan Gibbsite 5-20 "m MA Richmond, 5 149 tr t Trace. Gibbsite 20-44 "m MA Table 6. Amounts of Si, AI, and Fe extracted from sudace soU sam- t Dithionite-citrate-bicarbonate. pies by the Tiron and oxalate methods. *Not deterinined. Tiron Oxalate Sampie (horizon) Si Al Fe Si AI Fe ASit Aluminum Hydroxide and Oxyhydroxide Minerals 82-116 (E) 1050 28 26 tr 13 25 1050 Tiron and oxalate extracted nearly equal amounts 84-2477A (E) 1025 133 290 tr 89 294 1025 of Al from noncrystalline synthetic hydrous Al oxide (Table 7). However, Tiron extracted far more Al from t ASi - Tiron-extracted Si minus oxalate-extracted Si. pseudoboehmite than did oxalate. Neither method at- amount ofSi extracted by Tiron was substantially low- tacked a well-crystalline boehmite. With fractionated er than that by NaOH. Besides these minerals, this gibbsite sampIes, there was a general effect of surface opal sampIe contained pyroxene as an impurity. Only area on the amount of Al extracted, as amounts of a fraction of Si from volcanic glassy substances such extractable Al increased with a decrease in particle as obsidian and Hawaiian opal was extractable with size. Compared with oxalate and DCB treatments, Ti- NaOH and Tiron. Obsidians are essentially unhy- ron dissolved considerably more Al from gibbsite, par- drated silicate glasses that have a three-dimensional ticularly clay-size gibbsite. This indicated that Tiron disordered framework structure. Therefore, very lim- partially dissolves gibbsite, confirming the finding of ited dissolution of the obsidian by alkaline solutions Biermans and Baert (1977). Therefore, if any soil sam- was anticipated. However, the results forthe Hawaiian pIes containing clay-size gibbsite requires chemical opal were not fully understood, since its XRD pattern dissolution treatment for cleaning up, the oxalate with a single hump near 0.4 nm was typical of opal. method should be applied. If oxalate treatment does One explanation is that the opal is composed of sub- not improve XRD patterns, the Tiron method may microscopically crystalline silica substances that are still be usable with the correction formula K(m2 - still amorphous to x-rays because of extremely small ml)/a2 - al), where K is the conversion factor, 2.89, crystallite size. from Al to Al(OH)3' ml and m2 are the amounts of Al Dissolution of allophane and imogolite by Tiron extracted by Tiron from a sampIe after a I-h reaction was comparable to that by oxalate. Assuming that this at 60 ° and 80°C, respectively,and al and a2 are the allophane has a chemical composition similar to the I-h reaction rates of clay-size gibbsite at 60 ° and one reported for an allophane from the same locality 80°C, respectively. The al and a2 rates were prede- (Sudo, 1968), the amounts of Si and Al extracted in- termined from separate solubility experiments on clay- dicate that both treatments dissolved almost all of the size gibbsite. The solubility of gibbsite increased ex- mineral. Similarly, Tiron and oxalate were also equally ponentially with increase in temperature and time of effective extractants for imoglite, which contained reaction with Tiron (Fig. 2), and these solubility curves some quartz and feldspar. could be explained by a combination offirst-order (up Since Si is extracted from allophane and imogolite to 45 min) and two-dimensional diffusion-controlled but not from opaline silica by oxalate (Sio)and Si is (>45 min) reactions (Fig. 3). Since a I-h reaction time extracted from all three by Tiron (Sh), the difference at 80°C was used as a routine extraction procedure, (ßSi = SiT - Sio)may be used as an indicator of the 60°C was considered to be the appropriate choice for amount of opaline silica. This idea was tested on the another temperature. Thus, from Fig. 2, al (60°C) = clay fractions of surface soil sampIes. Table 6 gives 0.28 and a2 (80°C) = 0.58.The correctionformula two representative cases showing high ßSi values, alcan then be simplified to 9.63 (m2 - ml)' A working though amounts of Al and Fe extracted were com- example is illustrated in Fig.4. A saprolite from Nova parable between the two dissolution methods. Scotia required Tiron treatment to obtain improved Differential XRD patterns (not shown) indicated the XRD patterns from which crystalline components presence of a broad hump with a maximum near 0.39 could be identified. The pattern before treatment nm, which is a characteristic feature of standard silica showed only one peak at 0.483 nm, which correspond- gel (Kodama and Wang, 1989). Therefore, this pro- ed to the strongest peak of gibbsite. After the treat- cedure appears to be a useful indicator of amounts of ment, several minerals including gibbsite were opaline silica. identifiable (Fig. 4). Tiron extracted 16.8 p.gAl mg-I KODAMA & ROSS: INORGANIC COMPONENTS CHARAcrERIZED BY TIRON DISSOLUTION 1185

0.8 0.8 1.0 D,(a) D2(a) 90' C r !

0.6 0.8 0.6

:? 'Ö N I es 11 - 0.6 0.4 0.4 :s: -es "8' - I cx 70'CQ -II SO'C ~ 0.4 0.2 0.2 ~ 50' C ~ SO' . C .- R 0.2 --. 0 15 30 45 45 60 90 TIME (MIN.) Fig. 3. Plots of first-order (D.) and two-dimensional diffusion (D:z) reactions as a function of reaction rates (a) against time for the dis- 00 20 40 60 80 100 solution of <2-pm gIöbsite by Tiron at four different temperatures. TIME (MIN.) CPS F"Ig. 2. llfOn dissolution curves represented by the &action reacted, a, agajnst time for a clay-size gibbsite at four different temperatures. BEFORE TlRON TREATMENT

sampie under the conditions for the routine procedure, 60 whereas 5.8 ILgAl was extracted after the I-h reaction at 60 oe. Therefore, the results indicate that the original sampie contained about 105g gibbbsite kg-l, This 30 estimate was comparable with the amount determined by thermogravimetric and x-ray methods (McKeague 0 et al., 1983). 90 AFTER TlRON TREATMENT M M Ch(+K) Iran Oxide and O:xyhydroxide Minerals 60 The dissolution capacity of Tiron for ferrihydrite is Ch similar to that of oxalate, although Tiron extracted somewhat less Fe than did oxalate (Table 8). For crys- 30 talline Fe oxyhydroxides such as goethite and lepidocrocite, the data of both treatments were in good 0 40' agreement. With maghemite and hematite, the 20' amounts of Fe extracted by oxalate, although small, DEGREE 28. CoK a were consistently much higher than those extracted by Fig. 4. X-ray diffraction patterns of a soil day (7850) containing Tiron, perhaps due to the acidity of the oxalate so- gibbsite (G) before and after Tiron treatment. Ch: chlorite, M: lution. The most striking observation was that Tiron mica, K: kaolinite, F: feldspars. reacted very little with magnetite, whereas oxalate dissolved nearly 70%ofmagnetite, Chao and Zhou (1983) oretically containing 629 g Fe kg-l). Silt-size hematite proposed an alternative method using hydroxylamine, (699 g Fe kg-l) required > 10 repeated extractions and which does not attack magnetite and other crystalline only about 60% of its coarser silt-size hematite was Fe oxides. Ross et al. (1985) and Wang et al. (1987) dissolved after 19 extractions. These facts should be found that the hydroxylamine method gave a good borne in mind when extraction data are analyzed in approximation of oxalate-extractable Fe and Al of soil detail to estimate crystalline and noncrystalline Fe ox- sampies, and they proposed that the hydroxylamine ide components. method should be used for soils containing magnetite. As additional information, it was found in our lab- Considering the results discussed above, Tiron also oratory that both Tiron and oxalate extracted sub- proved to be a more suitable reagent than oxalate, stantial amounts ofFe from siderite (FeC03, <44ILm). particularly for magnetite-bearing sampies. Although it is not an Fe oxide mineral, it is occasion- The DCB treatment is known to extract total free ally found in soils and sediments. Tiron and oxalate Fe. The etfectiveness of dissolution, however, is greatly extracted 157g Fe kg-l and 142g Fekg-l, respectively, atfected by the particle size of the crystalline Fe oxide which corresponds to about 300 g FeC03 kg-l. In con- minerals such as goethite and hematite (Table 8). The trast, dithionite-citrate extraction at room tempera- data clearly indicate that a single DCB treatment was ture for 16 h liberated only 40 g Fe kg-1 from this not sufficient to dissolve even clay-size goethite (the- mineral. 1186 SOlL SCI. Soc. AM. J., VOL. 55, JULY-AUGUST 1991

Table 8. AnlOunts of Fe extracted from Fe oxide and hydrous oxide CONCLUSIONS minerals by three different chemical dissolution methods. 1. Tiron treatment appeared to be more effective Mineral Tiron Oxalate DCBt than NaOH treatment in removing the noncrys- g kg-I talline components from aluminosilicates with- Ferrihydrite-Nl63 417 474 out significantly attacking the crystalline Ferrihydrite-S5 363 403 components. Goethite <2 12 15 582(1)* 2. Tiron treatment was as effective as oxalate treat- "m 5 8 574(1) Goethite 2-5 "m 2 5 545(1) 590(2) ment in the dissolution of allophane, imogolite, Goethite 5-20 "m Goethite 20-44 0.5 tr 380(1) 582(6) and short-range-ordered hydrous oxides (e.g., fer- Akaganeite "m 32 17 rihydrite). 9 10 Lepidocrocite <44 "m Maghemite 6 17 3. In contrast to oxalate treatment, Tiron treatment 0.4 37 204(1) 655(10) effectively removed opaline silica. The diff'erence Hematite 5-20 "m Hemlltite 20-44 ,.m 0.1 29 37(1) 395(19) between the amounts of Si extracted by the two Mqnetite <44 3 537 "m treatments could be used to measure the amount t DiQ1ionite-citntte-bicarbonate. of opaline silica. *Number in bracket indicates number oe treatments. 4. Unlike oxalate treatment, Tiron treatment dissolved little or no magnetite. 12 5. Generally, a single treatment with Tiron appears .." to be sufficient to clean up sampies for XRD ! analysis of crystalline minerals. The amounts of c w 10 AI extracted Si,Al, and Fe indicate the amounts of I-CJ short-range-ordered aluminosilicates and hy- er: IX: drous oxides of Fe, Al, and Si. l- X w 8 CI) REFERENCES I- Ball, D.F., and P. Beaumont. 1972. Vertical distribution ofextraet- ifi able iron and aluminum insoH profiles from a brown earth-peaty :!E . w podzol association. J. SoHSei. 23:298-308. .... 6 W Bayliss, P. 1987. Mineral nomenclature: Imogolite. Mineral. Mag. IL 51:327. 0 Biermans, V., and L. Baert..1977.. SeI<:ctiveexffiicti()A()ftl1eam~r- CI) phous Al, Fe, and Si oxides using an alkaline Tiron'solution.Clay ZI- Miner. 12:127-135. ::I Carlson; L., and U. SehwertInann.1980.Nat\J1'al?QCCU1'1'ence offer- 0 :!E Fe oxyhite W-FeOOH). aays aay Miner.. ~8:.~72-280. ,er: Chao, T.T., and L. zhou. 1983. Extraction techniqti~sfor selective w dissolution of amorphous iron oxides fromsoilsand sediments. > Soil Sei. Soc. Am. J. 47:225-232. Chester, R., and M.J. Hughes. 1967. A chemiciü technique for the ~.... separation of ferro-manganese minerals, carbOnate minerals and ::I adsorbed trace elements from pelagie sediments. Chem. Geo!. ~ 2:249-262. CJ Cradwiek, P.D.G., V.C. Farmer, J.D. Russen, C.R. Masson, K. Wada, 0 2 3 4 N. Yoshinaga. 1972. Imogolite, a hydrated aluminum silicate of tabular structure. Nature (London) Phys. sq, 240:187-189: Deb, B.C. 1950. The estimation offree ironoxides in soilsand elays NUMBER OF EXTRACTIONS and their removal. Soil Sei. 1:212-220. Fig. S. Dissolution by Tiron of Si, Al, and Fe from a soil day as Duehaufour, P., and B. Souehier. 1966. Not.esur une methode d'ex- expres~ i!l cumulative amounts (percent by weight) for four suc- traction combinee de l'aluminium et du fer libre dans les sols. cessive treatments (* indicates projection of a straight line through Sol 1:17-29. the dJl.tapoints from thesecond, third, and fourth extractions on Follet, E.A.C., W.J. MeHardy, B.D. MitchelI, and B.F.L. Smith. 1965. Chemical dissolution techniques in the study ofSoil clays: the Yaxis). Part I. ClayMiner. 6:23-34. Foster, M.D. 1953. Geochemical studies ofc~y minerais.m.T4e Ejfectiveness 01 the TiTon M ethod determination of free silica and free alumina in montmorillonites. Geochim. Cosmochim. Acta. 3:143-154. Hariya, Y., and Z. Harada. 1957~ On manganese-wads from som,e The cumulative amounts ofFe, Al, and Si extracted localities in Hokkaido: Studies in manganese oxide minerals (Rep. from a clay-size soil sampie (82-128) by four repeated 1). (In Japanese.) Mineral. Soc. Jpn. 3:300-313. Tiron treatments indicateq that nearly all the. extrac- Hashimoto, 1., and M.L. Jackson. 1960. Rapid dissolution of allo- tion occurred in the first treatment (Hg. 5). Extrapo- phane and kaolinite-halloysite after dehydration. Clays aay Min- lation ofa straight line through the data points from er, 7:102-113. Henmi,. T., and K. Wada. 1976. Morphology and composition of the second, third, and fourth extractions to zero ex- allophane. Am, Miner. 61:379-390. traction on the Yaxis (marked by in Fig. 5) shows Higashi, T., and H. Ikeda. 1974. Dissolution of allophane by aeid * oxalate solution. aay Sei. 4:205-211. that the extrapolated values are close to the amounts Jaekson, M.L. 1968. Soil ehemical analysis - Advanced course. extracted in the first treatment. These and other un- M.L. Jaekson, Madison, WI. published data (1988) on a number of clay-size soil Jeffries, C.D. 1946. A rapid method for the removal of free iron oxides in soils prior to petrographie analysis. Soil Sei. Soc. Am. sampies in Canada suggest that a single Tiron treat- Proe. 11:211-212. ment is sufficient to dissolve Si, Al, and Fe or poody Jones, J.B., I.V. Sanders, and E.R. Segnit. 1964. Strueture of opal. crystalline and nOQcrystallinecomponents in most soil Nature (London) 204:990-991. clays, Iones, J.B., and E.R. Segnit. 1971. The nature of opa!. I. Nomen- clature and eonstituent phases. J. Geo!. Soc. Aust. 18:57-68. KODAMA & ROSS: INORGANIC COMPONENTS CHARACTERIZED BY TIRON DISSOLUTION 1187

Kodama, H., and M. Jaakkimainen. 1982. A comparative study of num from soi1s. Soi! Sei. Soc. Am. J. 49:783-785. selective chemical dissolution methods for separating noncrys- Schwertmann, U. 1959. Die fraktionierte Extraktion der fteien.Ei- talline components produced by grinding of silicates. p. 399-410. setiOxide in BOden". ihre mineralogischen Formen und ihre En- In H. van Olphen and F. Vemale (ed.) Int. Clay Cont: 7th, Bo- tsteliungsweisen. Z. PtIiinzenemähr. Düng. Bodenkd. 84:194-204. logna, ltaly. 6-1~ Sept. 1981. Elsevier Sci. Publ., New York. Segalen, P. 1968, Note sur une methode de determination des pro- Kooama, H., L.S. Kotlyar, and JA Ripmeester. 1989. Quantifica- dilits mineraux..al11orphes dans certains sols a hydroxydes tropi- tion of crystalline and noncrystalline material in ground kaolinite caux. Cah:'ORSTÖMSer. Pedol. 6:105-126. by x-ray powder diffraction, mfrared, solid-state nuc1ear magnetic Sudo, T.. 1968. Appendix I. Analysis (6). In Nendokobutsu (Clay tesonance, and chemical-dissolution analyses. Ciays Clay Miner. minerals). lwanami Publ., Tok)'o. 37:36+-370. Sudo, T., and T...Nakamura. 1952. Itisingerite ftom Japan. Am. Kodama, H., and C. Wang. 1989. Distribution and characterization Miner, 37:618-621. of noncrystalline inorganic components in Spodosols and Spo- Tamm,. O. 1922., Um bestämning av deoorganiska komponentenia dosol-like soils. Soil Sei. Soc. Am. J. 53:526-533. i markens gelkomplex.Medd. Statens Skogsfdersoeksanstalt Kohyama, N., and T. Sudo. 1975. Hisingerite occurring as a weath- 19:385-404. ermg product ofiron-rich saponite. Clays Ciay Miner. 23:215-218. Wada, K. 1977. Allophane and imogolite. p. 603-638. In J.B. Dixon Lanning, F.C., B.W.X. Ponnaiya, and C.F. Crumpton. 1958. The and S.ß. Weed (ed.) Minerals in soil environments. SSSA, Mad- chemical nature of silica in plants. Plant Physiol. 33:339-343. ison, WI: Levin, 1., and E. Ott. 1933. X-ray study of opals, silica glass and Wang, c., tI. Kodama, and N.M. Miles. 1981. Effect of various silica gel. Z. Kristallogr. 85:305-318. . pretreatments ort x-ray diffraction patterns o( clay ftactions of McKeague, JA 1967. An evaluation ofO.l M pyrophosphate and podzolic B horizons. Can. J. Soil Sei. 61:311-316. pyrophosphate-dithionite in comparison with oxalate as extrac- Wang, c., J.A. McKeague, and H. Kodama: 1986. Pedogenic im- tants of the accumulation products in podzols and some other ogolite and soil enviromnents: Case study ofSpodosols in Quebec, soils. Can. J. Soil Sei. 47:95-99. Canada. Soil Sei. Soc. Am. J. 50:711-718. McKeague, JA, D.R. Grant, H. Kodama, G.J. Beke, and C. Wang. Wang, c., and P.A. Sc.huppli. 1986. Determining ammonium oxa- 1983. Properties and genesis of a soil and the underiying gibbsite- late-extractable Si in soils. Can. J. Soil Sei. 66:751-755. bearing saprolite, Cape Breton Island, Canada. Can. J. Earth Sei. Wang, c., P.A. Sehuppli, and G.J. Ross. 1987. A comparison of 20:37-48. hydroxylamine and ammonium oxalate solutions as extractants Mehra, Q.P., and M.L. Jackson. 1960. Iran oxide removal ftom for Al, Fe and Si ftom Spodosols and Spodosol-like soils in Can- soils and days by a dithionite-citrate system with sodium bicar- ada. Geoderma 40:345-355. bonate buffer. Clays Clay Miner. 7:317-327. Yoshinaga, N. 1968. Identification ofimogolite in the filmy gel ma- MitchelI, ß.D., and V.C. Farmer. 1962. Amorphous day minerals terials in the Imaichi and Shichihonzakura pumice beds. Soil Sei. in some Scottish soil profiles. Clay Miner. Bull. 5:128-144. Plant Nutr. 14:238-246. Ross, G.J., C. Wang, and PA Schuppli. 1985. Hydroxylamine and Yoshinaga, N., and S. Aomine. 1962. Imogolite in some ando soils. ammonium oxalate solutions as extractants for iron and alumi- Soil Sei. Plant Nutr. 8:22-29.