Analyst, November, 1979, Vol. 104, $9. 1055-1061 1055 Determination of Iron(l1) in Rock, Soil and Clay L. Th. Begheijn Department of Soil Science and Geology, Agricultural Universz't-v Wageningen, P.O. Box 37, 6700 A A Wageningen, The Netherlands A rapid and direct method for the determination of iron(I1) in silicates is described. Redox processes frequently occurring during decomposition are suppressed satisfactorily by limiting the reaction time to 10 s while main- taining the temperature at 60-65 "C. Reproducible decomposition tempera- tures are achieved by mixing concentrated sulphuric and hydrofluoric acids (1 + 3 V / V ) in the reaction vessel. The coloured iron(I1) - 1,lO-phenanthroline complex is used in the spectrophotometric determination of the two valency states of iron, iron(I1) directly and iron(II1) by difference after subsequent reduction by hydroquinone.Mean results of duplicates of the USGS geo- logical standards G-2, AGV-1 and BCR-1 are within 0.1% and those for DTS-1 and PCC-1 within 0.3% of the quoted average values for iron(I1) oxide. Keywords : Iron(II) determination ; silicates ; hydrofluoric acid decomposition ; spectrophotometry The determination of iron(I1) in mineral material, including clay fractions of soils, suffers from many possible sources of error. Apart from incomplete decomposition, interferences frequently reported are due to the reduction of iron(III), mainly by organic matter and sulphides. The principal sources of errors in iron( 11) determinations are the mechanical dry grinding of samples in air and oxidation during decomposition.Pruden and Bloomfieldl reported the effect of organic matter and concluded that its presence vitiates the deter- mination of iron(I1). Mitsuchi and Oyama2 lowered the temperature of the decomposition to less than 80 "C in order to overcome reduction of iron(II1) by organic matter. Kiss3 showed that air oxidation may cause appreciable loss of iron(I1) oxide. Clemency and Hagner4 considered an inert atmosphere unnecessary; however, they used a system in which both reduction and oxidation may have occurred. Replacement of air by nitrogen or carbon dioxide is often advocated for eliminating air oxidation. French and Adams5 reported that the hydrogen fluoride vapour evolved by the pre-mixed equal volumes of concentrated hydrofluoric and sulphuric acids with silicates effectively prevents air oxidation.Oxidation during sample preparation or pre-treatment may also lower the apparent iron( 11) content. French and Adams5 found that grinding rock samples for 10 min with continuous moistening with acetone produced no detectable oxidation. Removal of organic matter from soil samples by hydrogen peroxide does not seem to affect the iron(I1) in the minerals directly but may cause the pH of the suspension to become very low, resulting in degradation, especially in the 2 : 1 clay minerals6 During this degrada- tion iron(I1) fixed in a non-exchangeable form within aluminium interlayers may be removed and ~xidised.~ These effects may be suppressed by the addition of a 1 M sodium acetate buffer (pH 5) as suggested by Jackson8 for the removal of carbonates before hydrogen peroxide treatment .6s7 This study, partly an extension of the work of Clemency and Hagr~er,~ was made in order to investigate whether interfering oxidation and reduction processes can be eliminated by limiting the time and temperature of digestion while, a t the same time, achieving complete decomposition.Experimental US Geological Survey StandardsgJo of powdered rock with a wide range of iron(I1) oxide contents were selected for the experiments. The decomposition procedure recommended by Clemency and Hagner4 was modified by limiting the time of reaction and avoiding external1056 BEGHEI J N : DETERMINATION OF Analyst, Vol. I04 heating. A temperature of 60-65 "C is reached spontaneously if 0.10 g of silicate sample is mixed successively with 1.0 ml of sulphuric acid (sp.gr. 1.84) and 3.0 ml of hydrofluoric acid (sp. gr. 1.19). The subsequent 2-min boiling step in the presence of boric acid was examined for the effect of complete dissolution of iron after disintegration of the silicate structure (Table I). Separate samples of some of the standard rocks were each decomposed for different periods of time. After optimum conditions of decomposition were evaluated the effect of organic matter was investigated by treating the selected geochemical standards in the presence of 25 mg of humic acid (Fluka, practicum grade, relative molecular mass 500-1 000; Table 11). The colour development of iron(I1) with 1 ,lo-phenanthroline was tested under experi- mental conditions.Separate solutions of ammonium iron(I1) sulphate and ammonium iron(II1) sulphate (analytical-reagent grade) and a mixture of both were transferred into silica beakers containing 10 ml of boric acid (4% m/V), 1.0 ml of hydrochloric acid (sp. gr. 1.16), 1.0ml of sulphuric acid (sp. gr. 1.84) and 3.0ml of hydrofluoric acid (sp. gr. 1.19), and boiled for 2 min. The contents of the silica beakers were then transferred into 100-ml polypropylene calibrated flasks, containing 40 ml of boric acid, diluted to volume and filtered. Aliquots of 4.00 ml were pipetted into 100-ml calibrated flasks that contained 20 ml of buffer solution and 8.0 ml of 1,lO-phenanthroline solution (see Reagents), diluted with water to volume and mixed. After the addition of 1 ,lo-phenanthroline solution absorbances of these solutions at 510nm were measured after periods of 0, 3, 5 , 10, 15, 23 and 30min.About 25mg of hydroquinone were then added to the remainder of each solution (about 50ml were left). The rate of colour development, after reduction of iron(II1) by the hydroquinone, was studied by measuring the absorbances at 5-min intervals from 0 to 30min after the hydroquinone addition. Apparatus Platinum - rhodium (95 + 5 ) crucibles of capacity 30 ml. Fwne cupboard. Silica beakers, 150 ml. PolyProPylene calibrated Jlasks, 100 ml. Vitatron Universal Photometer UFD with 1-cm cell. Reagents water was used to prepare all solutions. All reagents, except for hydroquinone, were of analytical-reagent grade, and de-ionised HydroJluoric acid, s#.gr. 1.19. Sulphuric acid, sp. gr. 1.84. Hydrochloric acid, sp. gr. 1.16. Boric acid solution, 4% m/V (saturated). Iron stock solution, 1.000 g 1-1 of iron. Iron standard solution, 0.010 g 1-1 of irorc(l1I) oxide. Potassium hydrogen phthalate bufer solution, 0.5 M, pH 4.1. 1 ,lO-Phenanthroline hydrochloride (monohydrate) solution, 0.25y0 m/V. Hydroquin one. Dissolve 40 g of boric acid in 1 1 of water. Dissolve 1 .OOO g of iron powder (Merck) in 25 ml Dilute 7.00 ml of the iron stock Dissolve 100 g of potassium Dissolve 0.25 g of Prepare freshly each day (solutions become coloured of 4 M hydrochloric acid. Dilute with water to a volume of 1 1. solution with water to a volume of 1 1. hydrogen phthalate in 1 1 of water and heat gently. C,,H,N,.HCl.H,O in 100 ml of water.with age). Procedure mill (TEMA). The standard USGS rock samples passed a 150-pm sieve without previous powdering. Powder soil samples by grinding for 30 s in a mechanical tungsten carbide ball (or ring) A finely powdered sample is obtained of which 80% is less than 50pm. Freeze-dried samples of clay fractions may be used after light grinding with a pestle andNovember, 1979 IRON(II) IN ROCK, SOIL AND CLAY 1057 mortar. Samples used for separation of clay fractions should be treated with hydrogen peroxide buffered at pH 5 to remove most of the organic matter8 and kept between pH 5 and 7.5 during dispersion and separation. Weigh accurately, to within 0.1 mg, about 100 mg of sample into a platinum crucible. Add 1.0 ml of sulphuric acid and homogenise.Add, from a 10-ml polypropylene measuring cylinder, 3.0 ml of hydrofluoric acid. Caution-Appropriate safety measures should be taken. Swirl gently for 10 s to allow the temperature to rise to 60-65 "C. Copious vapours now blanket the mixture. Transfer quickly into a silica (see Note 1) beaker that contains 10 ml of boric acid solution (which inactivates part of the hydrofluoric acid) and 1.0 ml of hydro- chloric acid. (Wear gloves throughout this last part of the procedure and work in a fume cupboard). Cover the beaker with a polypropylene cover, heat and boil gently for 2 min in order to dissolve the decomposition products. Cool, transfer into a 100-ml polypropylene (see Note 1) calibrated flask, which contains another 40 ml of boric acidsolution (see Note 2).Dilute to volume with water, homogenise and filter. Prepare a similar blank solution for each batch of analyses. NOTES- 1. 2. Standard laboratory glassware (Jena glass) may release appreciable amounts of metals (sodium, This additional boric acid is added in order to obtain an excess over the hydrofluoric acid. calcium, barium and aluminium). This would make the filtrate unusable for further elemental analysis. Determination of iron(II) Pipette 2.00 ml of the filtrate into a 50-ml calibrated flask. Caution-Do not pipette orally. Add 10 ml of the buffer solution and 4.0 ml of 1,lO-phenanthroline solution. If highest accuracy (1% relative or better) is desired, the solution can be made 0.01 M in nitrilotriacetic acid (NTA) as suggested by Fadrus and Malq.ll Immediately (or within 10 min) measure the absorbance at 510 nm against water as a reference.Keep the remainder of the solution for the determination of total iron (see below). If the filtrate is coloured by organic matter, a duplicate solution without phenanthroline should be used as a blank (check that the absorbance of phenanthroline is near to zero). Dilute with water to a volume of about 30 ml. Dilute to volume and mix. Determination of total iron 40ml) and homogenise. absorbance as for the iron(I1) determination. With a spatula add about 20 mg of hydroquinone to the remainder of the solution (about Leave for a period between 10 and 30min and measure the Calibration graph solution into separate 50-ml calibrated flasks. 1,lO-phenanthroline solution and 25 mg of hydroquinone.and homogenise. 20 mg of hydroquinone. concentration. make readings of absorbance less accurate. Pipette 0.0, 10.0, 20.0 and 30.0ml of the standard iron solution and 2.0ml of blank Add 10ml of buffer solution, 4.0ml of Dilute to volume with water Read the absorbance at 510 nm after 30 min against water as a reference. The absorbance of the blank solution should be read before and after the addition of Construct a calibration graph of absorbance versus iron(II1) oxide More concentrated solutions The graph is linear up to 5 mg 1-1 of iron. Results and Discussion The effect of the time of decomposition on the recoveries of iron(I1) and total iron [iron(II) As can be seen, the decomposition is complete within Extended tests with the geological standard AGV-1 show complete recovery of As shown in Table I, boiling is clearly essential for the Fig.2(a) shows that in this procedure, plus iron(III)] is shown in Fig. 1. 1 min. total iron with 15 s of swirling. complete dissolution of the decomposition products.1058 BEGHEI JN : DETERMINATION OF Analyst, Vol. 104 formation of the iron(I1) - 1 ,lo-phenanthroline complex is practically instantaneous; subse- quent addition of hydroquinone has no further effect. A negligible effect of iron(II1) at very low iron(I1) concentration is indicated in Fig. 2 ( b ) . The maximum absorbance after reduction of iron(II1) (after addition of hydroquinone) is reached in 5 min (Fig. 2 ( b ) ] . An effect of iron(II1) is seen in Fig. 2(c) in the presence of a moderate iron(I1) concentration. Seconds swirling Minutes on hot-plate Time of decomposition Fig.1. Measured iron(I1) and total iron contents as a function of time of decomposition. Solid lines : contents according to USGS (1972); A, iron(I1) plus iron(II1); and 0, iron(I1). Positive errors in the iron(I1) values may increase by up to 5% relative after 30 min standing, or 2% relative after 10min. According to Fadrus and Mal$ll this could be eliminated by addition of NTA. Results for iron(I1) and total iron (I1 plus III), as compared with USGS standards, are shown in Table 11. Recoveries of iron(I1) oxide are good for the felsic, intermediate and mafic standard rocks (G-2, AGV-1 and BCR-1, respectively), but low (97-98y0) for two ultramafic rocks (DTS-1 and PCC-1). Still lower recoveries (92-96y0) of total iron were found in the ultramafic samples indicating that iron(II1) especially, in the samples, is only partially liberated by the proposed procedure. Solid residues, after hydrofluoric acid - sulphuric acid treatments, amounting to 2 and 10% of the original mass of DTS-1 and PCC-1, respectively, were observed. Longer decomposition times (up to 1 min) or longer boiling (5 min) hardly increased the measured iron(II1) oxide content and had no effect on iron(I1) oxide.Enstatite and a nickel - chromium - iron spinel were identified by X-ray diffraction in the solid residues of PCC-1 and DTS-1, respectively. This indicates that this method is less suitable for measuring total iron in ultramafic rocks. In soil and clay samples with low iron(I1) oxide contents (0.1-0.4~0) analysed in this laboratory, the standard deviation is about 0.05-0.06% .7 TABLE I EFFECT OF BOILING ON RECOVERY OF IRON(II) OXIDE AND IRON(III) OXIDE Sample 100 mg of rock, 10 s decomposition in hydrofluoric acid - sulphuric acid. Iron(I1) oxide found, yo Iron(II1) oxide found, y-, -- With Rock boiling boiling boiling boiling BCR-1 .. . . 6.59 8.94 3.76 3.68 AGV-1 . . . . 1.55 2.09 4.34 4.35 (3-2 .. . . 1.22 1.42 1.09 0.96November, 1979 IRON(II) IN ROCK, SOIL AND CLAY 1059 The results of these experiments demonstrate that the problems with the usual recom- mended longer decomposition periods, with external heating, are avoidable. The disintegra- tion of most silicates is virtually instantaneous. This also applies to a wide range of powdered clay fractions examined in this laboratory, as checked against the determination of total iron by X-ray fluorescence spectrometry.The minimum period needed for adding reagents and transferring products into the silica beaker is about 15 s. A short time of reaction appears essential for suppressing interfering. factors. (Fig. 1 shows that oxidation starts after 10 s of decomposition, possibly due to oxidising constituents in the sample.) Although the period of mixing (swirling) should be as short as possible, it should still be reproducible and should ensure adequate mixing. Thc 10-s period of swirling appears to satisfy these con& tions. hydroqu inone hydroquinone I I I I I c 0 20 40 60 0 20 40 Timehin I Without I With - hydroquinone I Fig.2. Formation of iron(I1) - l,l0-phenanthroline complex in the presence of iron(II1) before and after the addition of hydroquinone : (a), concentration of iron(I1) 1.7 mg l-l, iron(II1) nil; (b), concentration of iron(I1) nil, iron(II1) 1.4 mg 1-l; and (c), concentration of iron(I1) 1.7 mg l-l, iron(II1) 1.4 mg 1-I. The sequence of addition of the acids may also influence oxidation. Additions of the acids in the order sulphuric and then hydrofluoric acid to a mixture of iron(I1) and iron(II1) salts resulted in almost complete recovery (96.4y0), whereas after addition in the reverse order (hydrofluoric then sulphuric acid), only 89.4% of iron(I1) was recovered (Table 111). The sequence of addition did not cause significant differences in results for rock BCR-1 or for silicate material.Apparently, oxidation takes place only after decomposition, whereas the iron salts tested are dissolved immediately. TABLE I1 RESULTS FOR IRON(I1) OXIDE AND TOTAL IRON [AS IRON(II1) OXIDE] IN INTERNATIONAL GEOLOGICAL STANDARDS Sample 100 mg of rock, 10 s decomposition in hydrofluoric acid - sulphuric acid, 2 min boiling in boric acid - hydrochloric acid. Iron(I1) oxide, yo Total iron as iron(II1) oxide, yo USGS Rock (1969) BCR- 1 . . 8.91 AGV- 1 . . 2.04 G-2 . . 1.44 DTS-1 . . 6.79 PCC- 1 . . 4.94 USGS* (1972) Found S.D.t Recovery 8.80 8.90 0.09 101.1 2.05 2.06 0.09 100.5 1.45 1.48 0.11 102.1 7.23 7.07 0.12 97.8 5.24 5.06 0.03 96.6 bSGS USGS* (1969) (1972) Found S.D.7 Recovery 13.50 13.40 13.67 0.12 102.0 6.80 6.76 6.73 0.14 99.6 2.76 2.65 2.58 0.13 97.4 8.85 8.64 8.31 0.11 96.2 8.53 8.35 7.66 0.03 91.7 m * USGS (1972) used for reference.t S.D. = standard deviation (six determinations per sample).1060 BEGHEI J N : DETERMINATION OF Analyst, Vol. 104 The effects on measured iron(I1) contents of even large amounts of organic matter (10 and 25 mg) appear to be small (Table IV); there is a slight rising trend, but none of the differences are significant at P = 0.05. At lower organic matter contents, as reported by Brinkman,' the measured iron(I1) oxide content in clay after removal of organic matter was within 0.06% of the value. TABLE I11 SPECTROPHOTOMETRIC DETERMINATION OF IRON(II) IN THE PRESENCE OF IRON(III) FROM PURE IRON SOURCES 'WITH 1 ,lo-PHENANTHROLINE Absorbances are read within 10 min after adding 1,lO-phenanthroline.Mass of iron(I1) Mass of iron(I1) Recovery, No.* Test takenlmg foundlmg % 1 30.1 mg of Fe(NH,),(SO4),.6H,O 4.29 4.27 99.5 2 30.9 mg of Fe(NH,)(S04),.12H,0 0 0.07 - 3 4 6 30.1 mg of Fe(NH4),(SO4)2.6HaO + 28.4 mg of Fe(NH,),(S04),.6H,0 + 27.3 mg of Fe(NH4),(SO4),.6H2O + 30.9 mg of Fe(NH4)(S0,)a.12H,0 4.29 4.44 103.5 22.4 mg of Fe(NH,)(SO4),.12H,O 4.04 3.61 89.4 23.5 mg of Fe(NH,)(S0,),.12H20 3.89 3.75 96.4 * 1, 2 and 3. fluoric acid and 1.0 ml of sulphuric acid, boil for 2 min. that contain 40 ml of boric acid. 4. In a platinum crucible. swirl and 2-min boil. 5. In a platinum crucible. and 2-min boil. Dissolve in 10 ml of boric acid plus 1.0 ml of hydrochloric acid, 3.0 ml of hydro- Add to 100-ml polypropylene flasks Add 3.0 ml of hydrofluoric acid and 1:O ml of sulphuric acid; 10-s Add 1.0 ml of sulphuric acid and 3.0 ml of hydrofluoric acid; 10-s swirl Dilute to volume.The widely applied and improved spectrophotometric iron(I1) - 1 ,lo-phenanthroline method, as described by Sandell,12 was slightly modified. The tedious procedure of buffering by citrate was replaced by the addition of a fixed (excess) volume of biphthalate solution of pH 4.1. This brings the pH of the final solution (including foregoing acid treatments) up to 3.1. Interferences of iron(II1) are limited, possibly due to the presence of fluoride (see Note 3), which masks iron(II1) and hence inhibits or slows down the reaction Fe3+ + 3(phen) + e- -+ Fe(phen),2+ NOTE- 3. Complexation by boric acid forms hydrofluoroboric acid (HBF,) givins ions H+ and BF,-; a The latter process proably leaves sufficient F- in solution to little BF4- dissociates into BF, and F-.13 reduce Fe3+ activity to a very low value.The use of hydroquinone may be a problem in the presence of titanium, which may cause an interfering brown colour.12 In this procedure and at the expected concentration of TABLE IV EFFECT OF ORGANIC MATTER ON RECOVERY OF IRON(II) OXIDE AND TOTAL IRON AS IRON(III) OXIDE Sample 100 mg of rock, 10 s decomposition in hydrofluoric acid - sulphuric acid, 2 min boiling in Blank determinations were carried out on 0, 10 and 25 mg of pure humic acid treated in the same way. boric acid - hydrofluoric acid. Iron(I1) oxide found, % Total iron as iron(II1) oxide found, % With 10 mg With 25 mg With 10 mg With 25 mg A I \ r A \ Without of humic of humic Without of humic of humic Rock humic acid acid acid humic acid acid acid BCR-1 .. . . 8.90 8.84 8.95 13.67 13.53 13.73 AGV-1 . . . . 2.06 2.01 2.21 6.73 6.67 6.63 G-2 .. . . 1.48 1.52 1.61 2.58 2.56 2.79November, 1979 IRON(II) IN ROCK, SOIL AND CLAY 1061 titanium (less than 1 mg 1-1 of titanium oxide) this interference appears to be of no practical importance. During the analysis of pyrite-containing soils (acid sulphate soils), iron(I1) disulphide (pyrite) is not included in the measured iron(I1) value because it is resistant to the applied concentrated acids. In fact, the latter residue is available for analysis of pyrite iron(I1) by a subsequent nitric acid di~solution.~~ An oxidimetric titration, as used by French and ad am^,^ for standard rock samples would obviate the need for filtration, but does not seem to be suitable for soils and clay fractions, because of possible interference by organic matter.Conclusion Analysis of iron(I1) in rocks, soils and clay fractions contairling some organic matter is subject to many interfering factors, which are somewhat unresolved in the literature. This paper evaluates recent published work and recommends a highly reproducible, rapid and simple method that does not require specialised apparatus. The proposed method is widely applicable to soils and clays and to most rocks; some reservation must be made, however, for the group of the ultrabasic crystalline rocks. The author thanks J. D. J. van Doesburg for X-ray diffraction analyses and N. van Breemen and R. Brinkman for their critical comments on a draft of this paper. A. Breeuwsma and R. A. Koning (Soil Survey Institute, Wageningen) are thanked for supplying the rock samples and for the useful discussions on a modified method designed for materials with high organic matter contents. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 4. References Pruden, G., and Bloomfield, C., Analyst, 1969, 94, 688. Mitsuchi, M., and Oyama, M., J . Sci. Soil Manure, Japan, 1963, 34, 23. Kiss, E., Analytica Chim. Acta, 1977, 89, 303. Clemency, C. V., and Hagner, A. F., Analyt. Chem., 1961, 33, 888. French, W. J., and Adams, S. J., Analyst, 1972, 97, 828. Douglas, L. A., and Fiessinger, F., Clays Clay Miner., 1971, 19, 67. Brinkman, R., Geoderma, 1977, 17, 111. Jackson, M. L., “Soil Chemical Analysis, Advanced Course,” published by the author, Department Flanagan, F. J., Geochim. Cosmochim. Acta, 1969, 33, 81. Flanagan, F. J., Geochim. Cosmochim. Acta, 1972, 37, 1189. Fadrus, H., and Malg, J., Analyst, 1975, 100, 549. Sandell, E. B., “Colorimetric Determination of Traces of Metals,” Third Edition, Interscience, New Mellor, J , W., “A Comprehensive Treatise on Inorganic and Theoretical Chemistry,” Volume V, Begheijn, L. Th., van Breemen, N., and Velthorst, E. J., Commun. Soil Sci. Plant Afial., 1978, 9, of Soil Science, University of Wisconsin, Madison, Wisc., 1956 (Fifth Printing 1969). York, 1959. 1946, p. 125. 873. Received April 17th, 1978 Amended June 4th, 1979 Accepted June 18th, 1978