THE NATURE OF GARNIERITES--I STRUCTURES, CHEMICAL ...clays.org/journal/archive/volume 21/21-1-27.pdf · THE NATURE OF GARNIERITES--I STRUCTURES, CHEMICAL COMPOSITIONS AND COLOR CHARACTERISTICS

Clays and Clay Minerals, 1973, Vol. 21, pp. 27-40. Pergamon Press. Printed in Great Britain

THE N A T U R E OF GARNIERITES--I

S T R U C T U R E S , C H E M I C A L C O M P O S I T I O N S A N D C O L O R C H A R A C T E R I S T I C S

G. W. BRINDLEY and PHAM THI HANG Materials Research Laboratory, and Department of Geosciences, The Pennsylvania State University,

University Park, Pennsylvania, 16802, U.S.A.

(Received 25 August 1972)

Abstract-X-ray diffraction patterns of garnierites indicate that most samples resemble serpentine- group minerals or a talc-like mineral, or a mixture of these forms, and give respectively 7 A and l0 A basal reflections. From a survey of some 40 garnierites, four of predominantly serpentine type and seven of predominantly talc-like type were selected for detailed study. The talc-like garnierites exhibit little variation of the l 0 A basal spacing with low-temperature heating or with immersion in liquids, though some may contain a small proportion of expandable layers. Chemical analyses show considerable deviations of octahedral/tetrahedral cation ratios from the values 3/2 and 3/4 for normal serpentine and talc minerals, and may be interpreted in terms of mixed 1 : I and 2 : 1 layer types, either as separate phases and/or as interstratifications, or as defect structures of various kinds. The H 2 0 + contents of the talc-like forms of garnierite are considerably greater than that of normal talc and point to a mineral of composition 3(Mg, Ni)O'4SiO2.2H20 or [(Mg, Ni)3Si4010(OH)2]'H20-a talc monohydrate formula. The green color of garnierites is related to the NiO weight per cent and a color index is derived based on the Munsell color charts.

INTRODUCTION THE long-standing problem of the nature of garnieri tes, used as a general name for hydrous magnes ium-n icke l silicates, is related to the fine- grained nature of these materials, their poor crystalline order, and especial ly their f requent occur rence as int imate mixtures of two or more components . It is general ly agreed that these components are related to the serpent ine group minerals and to some form of 2 : 1 type layer silicate such as talc, s teven- site, or a smecti te . The results of early studies gave rise to many mineral names including n icke l - gymnite , garnierite, genthite, noumei te , nepouite , pimelite, deweyl i te , etc. Faust (1966) rev iewed much of the early work on these materials and it is unncesssary to repeat his survey here.

Because of the poorly defined nature of these materials and their f requent occurrence as mixtures, Pecora , Hobbs and Mura ta ( 1 9 4 9 ) r e c o m m e n d e d that garnieri te be used as a general name, analogous to bauxi te and limonite, so that under c i rcumstances where a more detailed analysis is impract icable or impossible , a general ly unders tood term can be applied. Faus t (1966) regarded this suggestion as " exce l l en t " and proposed that the "garnier i te g r o u p . . , includes all hydrous n icke l -magnes ium si l icates".

Ex tens ive studies of hydrous n icke l -magnes ium silicates have been repor ted in the Russian litera-

ture (Ginzburg and Rukavishnikova , 1951; D ' y a k o n o v , 1963; Vi tovskaya and Berkhin, 1968, 1970; Vi tovskaya et al., 1969). This work has been surveyed in detail by Vi tovskaya and Berkhin (1968, 1970). In these studies, the t e rm 'kerol i te ' has been widely used as well as garnierite. Kero- lites also are commonly mixtures of serpentine- and talc-like minerals. The dist inction be tween them and garnierites appears to be primari ly composi- t ional, with the N i O contents of keroli tes ranging f rom zero to about 20 per cent and the M g O contents of garnieri tes f rom zero to about 15 per cent. The use of these names is said to be advan- tageous to indicate composi t ion i'anges and proper- ties, but also because the two materials "have fully defined features of genes is" (Vi tovskaya and Berkhin, 1970, see p. 39). The terms ct- and fl- keroli te have been used to indicate variet ies in which a serpentine form or a talc-like form respec- t ively predominates (Ginzburg and Rukavish- nikova, 1951 ; D ' y a k o n o v , 1963).

Mineral names are not usually given to mixtures, nor are different names justified on the basis of genesis. There fore in the present text , the recom- mendat ions of Pecora et al. (1949) and of Faust (1966) regarding the use of garnierite as a general name will be fol lowed.

There is abundant ev idence that many garnierites are indeed intimate mixtures of 1 : 1 and 2 : 1 type

27

28 G . W . BRINDLEY and PHAM THI HANG

layer silicates or, more explicitly, mixtures of serpentine-like and "talc"-like components, with "talc" written in this manner to indicate that it may not be strictly the normal form of talc, by composition or by structure. Many garnierites show both 7 and 10 ,~ basal spacings in X-ray powder diffraction patterns. However , the poor crystalline order of many samples is such that one or other of these characteristic spacings may not be obvious, and indeed may not be seen against the generally high background scattering from garnierites. As will be shown in Part I1 of this study, high magnification electron microscopy may reveal directly the 7 and 10 A spacings, but when this technique is not available evidence of an indirect nature must be considered very cautiously. Faust (1966) showed that the compositions of many garnierites, as regards the cations occupying tetrahedral and octahedral coordination sites, lie near a line joining nickel serpentine to pimelite; he regarded pimelite as a montmorillonite-group mineral and possibly a nickeloan stevensite. A similar plot of tetrahedral to octahedral cation values had been used earlier by Faust and Fahey (1962) for minerals related to serpentine and talc, or stevensite. Vitovskaya and Berkhin (1968, 1970) have emphasized that chemical analyses of garnierites (and kerolites) lead to non-stoichiometric formulae, which can be resolved into mixtures having stoichiometric serpentine- and talc-like formulae.

It seems generally agreed that the component giving a 7 A basal reflection resembles a serpentine group mineral, usually chrysotile or lizardite, but much uncertainty exists regarding the component giving a 10 .& basal spacing. The latter is frequently described as non-expanding in water or in ethylene glycol, and therefore is said to be "talc-like" rather than like a smectite group mineral. Pimelite has often been used as the name for this component, although Spangenberg (1938) used pimelite for a mineral with a variable layer spacing ranging from a maximum of 12.8 ,~, to 12-2 A when air-dry, and 9 .6A when fully dry. Slansky (1955) expressed doubt that pimelite belongs to the smectite group and did not find the variability of spacing reported by Spangenberg. Kato (1961), in a broad study of garnierite from New Caledonia, described the 10 ,~ layer mineral as "talc" because "the common occurrence of pimelite is rather doubtful", and stated that probably the non-swelling talc-like mineral "has hitherto been mistaken for pimelite". Maksimovic (1966) studied a series of minerals which he described as ranging from /3-kerolite to pimelite, respectively the magnesium and nickel end members of the series; he concluded that the available X-ray data "strongly suggest that a non-expanding 10 A layered pimelite predominates", and stated

that "only three out of 23 samples swelled partially, due to the presence of a small amount of an expanding material". Evidently more care is needed in the use of the name pimelite, for if it is indeed a swelling-type mineral, the name should not be used for non-swelling materials. The complex nature of the garnierite from the Oeyama mine, Kyoto prefecture, Japan, has been discussed by Muchi(1965, 1966); it appears to contain 7 A- and 10 A-type components and also a long-spacing component of 20-22 A, containing 12 A layers of "montmorillonite type" but which does not expand with ethylene glycol.

OBJECTIVES, MATERIALS, AND METHODS OF THE PRESENT STUDY

The present study was initiated in the hope that a broad survey of many garnierites would provide some essentially monomineralic samples suitable for detailed study. Some forty samples were available and from each sample, particles in a size range about 0-8-0.2 mm were separated by sieving, followed by careful hand-picking under a binocular microscope. Sufficient material was selected for X-ray diffractometer examination (filtered CuKa radiation, recording rate 1 ~ (20)/min). Materials showing dominantly 7 or 10 A basal reflections and an absence of crystalline impurities were then meticulously hand-picked in amounts sufficient for further study.

Diffraction patterns were recorded mainly from samples lightly packed in shallow cavities in glass slides, but when the available material was insufficient, samples "slurried" on glass slides were used. There appeared to be little tendency for enhance- ment of basal reflections.

Purification of materials by selective settling using dense liquids, or by acid treatment was avoid- ed because of possible alteration of the garnierite. Magnetic separation occasionally was useful. When samples appeared satisfactory apart from a small proportion of quartz, the quartz percentage was determined directly by X-ray intensity measure- ments using the so-called "method of known addi- tions", (Brindley 1961, p. 495).

Because many samples gave broad and somewhat asymmetric X-ray basal peak profiles, the usual simple tests for interstratified systems were applied, i.e. low-temperature heat-treatment and immersion in liquids that commonly cause swelling, such as water and ethylene glycol.

Electron micrographs of the selected materials were taken with a Philips electron microscope, model EM300, operated at 100 kV and magnifi- cations mainly in the range 3-104-1-10 a X. Subse- quently a collaborative research with Dr. Natsu Uyeda, of Kyoto University, was undertaken in

THE NATURE OF GARNIERITES-I 29

which micrographs at 106X magnification were made showing directly the 7 and 10A basal spacings. This aspect of the research is reported separately in the following joint paper.

Chemical analyses of the selected garnierites were made using a Perkin-Elmer model 303 atomic absorption spectrometer and about 40 mg samples, after equilibration at 110~ in air. The procedure followed closely that of Medlin, Suhr and Bodkin (1969).

Thermogravimetric analyses were made on about 5-10 mg samples using a recording Cahn micro- balance, with a nearly uniform heating rate of about 5 ~ C/min from room temperature to about 1000~ with the sample heated in air. The weight losses from 110~ to about 1000~ were taken as H 2 0 + values for the total chemical analyses.

Combined X-ray and thermal studies have been made on samples that were heated to progressively higher temperatures for periods of about 3 hr a n d then X-rayed after cooling to room temperature. Several investigators and particularly Vitovskaya and Berkhin (1968, 1970) have related the heating products to the initial mineralogy and it seemed worthwhile to examine this procedure in relation to the selected garnierites. These thermal studies are reported as Part I I I of the present study.

R E S U L T S A N D D I S C U S S I O N

X-ray diffraction data, 7 ,~-type garnierites Figure 1 shows diffraction patterns of selected

7 A-type garnierites (a . . . . . . f ) together with patterns, A, for lizardite from Kennack Cove, Lizard, England, and B, for a mixed clino- and orthochrysotile from Transvaal, S. Africa. The similarity of the patterns justifies these garnierites being described as "serpentine-like".

In the notation of Whittaker and Zussman (1956), an asterisk (*) indicates indices for a monoclinic lattice, and the symbol (?) indices for an orthorhombic lattice. Indices of types hkO and Okl are common to both systems; only hOl and hkl reflections differentiate monoclinic and orthorhombic forms of serpentine minerals. Reflections marked Q arise from quartz impurity. The indices used for lizardite, pattern A, relate to a two-layer, orthorhombic unit cell.

The best crystalline of the 7 A-type garnierites, sample MN6 from Morro do Niquel, M.G., Brazil, gives the diffraction pattern l a, with clear 20/? reflections with both odd and even numbered / indices, consistent with a two-layer regularity in an orthorhombic structure. The 020 (or simply 02) diffraction band is similar to that of chrysotile, and the background from 200 onwards also is similar to

] I I

i ! i i I/

P I '/L I

Fi~. 1. Diffractometer patterns of A, lizardite; B, mixed clino-orthochrysotile; a, . . . . f, 7 A-type garnierites; a, sample MN6; b, sample 3996; c, sample HMC5; d, Village Ba,

Serbia; e. Kambalda, Australia;f, New Caledonia. Ni-filtered, CuKa radiation.

i

I

I 1 I J I I

"\ I I I ~ W " I

L i L ~ 10 20 30 40 50 60

28,degrees

30 G . W . BRINDLEY and PHAM THI HANG

that of chrysotile; this background probably arises in part from a 20, 13 diffraction band. From these observations (supported later by electron micrographs), this sample appears to be a fairly well- crystallized orthochrysotile. A more detailed diffraction study will be undertaken later.

Pattern lb (sample 3996, from Morro do Cerisco, Liberdade, Brazil) and pattern lc (sample HMC5, from Riddle, Oregon) show resemblances to pattern la , but diagnostic reflections, such as 20l 's , are scarcely seen. The basal reflections, 002 and 004, are weak so that basal plane orientation seems not to occur. These characteristics identify only 7 A, serpentine-like minerals. Electron micrographs show mainly fluffy, ill-defined, probably aggre- gated particles. Pattern lb has a slight rise of background near 20 ~-9 ~ which may indicate the presence of some very poorly developed 10 A-type phase.

Patterns ld , le , lf, (samples respectively from Village Ba, Serbia; Kambalda, Australia; and New Caledonia) are tess well developed than patterns la , b, c. Patterns le , f s h o w a diffusion of the 002 basal reflection towards smaller 20, which may indicate an interstratification or an admixture of longer spacings. The diffraction features extending from 20 --- 34 ~ to higher angles probably represent the combined 20, 13 two-dimensional diffraction band. The 06 diffraction band also has considerable diffusion towards higher values of 20. Evidently the samples corresponding to patterns ld, e, and f contain considerable structural disorder.

When the 7 A-type garnierites are soaked in water, or in ethylene glycol, or are heated at 110~ generally very little change occurs in the diffraction patterns. However , the sample from New Cale- donia, pattern l f , was exceptional in that ethylene glycol treatment sharpened the 002 peak profile and moved the lattice spacing from about 7.6 A to about 7.3~ A. Possibly some of the diffused scattering comes from an expandable component which, after expansion, no longer contributes to the broad 7 A peak.

The main conclusion from these data is that the samples corresponding to patterns la , b, c, d are dominantly 7 A, serpentine-like garnierites, while the samples corresponding to le , f may contain material with longer spacings.

X-ray diffraction data, 10 ,~-type garnierites Figure 2 shows diffraction patterns of selected

10 A-type garnierites (a . . . . . . h) together with a pattern K of a so-called '/3-kerolite' from Goles Mountain, Yugoslavia, supplied by Dr. Z. Maksi- movic. The term '/3-kerolite' is retained here because" this was the name applied to the particular sample. The garnierite samples are from the fol-

lowing sources: (a) Sample NC, from Durham, N. Carolina; (b) Sample MN4A, from Morro do Niquel, Brazil; (c), (d), (e), Samples RO3B, RO24, RO3A, from Riddle, Oregon; ( f ) , (h), Samples 206, 201 from Serbia, Yugoslavia; and (g), Sample GUS, from an unknown source, but probably from New Caledonia.

The patterns in Fig. 2 are similar to each other, and to those given by Maksimovic (1966), from which he concluded that a series of minerals of similar structures exists from a magnesium end- member to a nickel end-member. These patterns justify the description "talc-l ike" because of the 10 A spacing and mainly non-swelling, non-shrink- ing characteristics. The names used by Maksi- movic for the end-members,/3-kerolite and pimelite, are open to question.

Several of the patterns in Fig. 2 show weak indi- cations of a reflection near 20---- 12 ~ which may correspond to a 7 A-typoe mineral. The first basal reflection from these 10 A-type garnierites is always broad and the peak profile has varying degrees of asymmetry, with a spreading towards smaller angles, i.e. longer spacings. Heating the samples in air at 110~ for about 3 days modifies the patterns to the extent shown in Fig. 3. The 10 A peak is sharpened and the spacing moves to slightly smaller values, from an average d(0o01 ) = 10.25---0-15A before heating to 10.10+0-I0 A after beating. The diffraction patterns as a whole become weaker, which suggests that the water released at 110~ is sufficiently organized to contribute to the original diffracted intensities.

Soaking the samples in water or in ethylene glycol has little effect on the 001 peaks of most of the samples; neither the spacings nor the profiles are much changed. Evidently these materials do not behave in the manner of smectites. However, the samples from N. Carolina, pattern 2a, and from an unknown source, pattern 2g, do show appreciable changes. Wetting with water increases the 001 peak intensity but has little effect on the spacing or the peak profile. Ethylene glycol spreads the diffracted intensity towards smaller angles and moves the maximum towards larger angles. The expansion appears to be irregular (there is no sharp reflection at smaller angles, as with smectites) and to affect only a part of the material.

These results are consistent with those of Maksimovic (1966).

Chemical analyses and mineral formulae The results of chemical analyses on materials

equilibrated at 110~ in air are recorded in Tables 1 and 2 respectively for the predominantly 7 A- and 10 A-type garnierites. The water contents recorded


(00~)

1 020

002

I I

003

1 o

"~ ~ I

I I

I I 200

310

.3k (002) ..'%. i �9

J ' ,~, ,ws" '~ -'~ \ L ) / J , %r ~ .........

/ "v'~,~, ~ ' ' '~

viva,It (oo2}

{OOl)

I r ~#V'~a?'*~V'~,~,'~, ~ ~,~ '~*'

I

I I I t I I 10 20 30 40 50

20,degrees

Fig. 2. Diffractometer patterns of K, sample of/3-kerolite, a . . . . . h, 10 A-type garnierites; a, from N. Carolina; b, sample MN4A; sample RO3B; d, sample RO24; e, sample RO3A; f, sample 206, Village Ba, Serbia; g, sample GUS; h, sample 201, Serbia.

Ni-filtered, CuKa radiation. Q = quartz reflection.

I

l 60

as HzO+ are taken as the weight loss between t 10 and 1000~ obtained from the thermogravimetric data discussed later.

The formulae recorded in Tables 1 and 2 under headings A and B have the following significance: Under A, the compositions are normalized with respect to 2 Si atoms in Table 1, corresponding to the serpentine formula Mg3Si2Os(OH)4, and to 4 Si atoms in Table 2, corresponding to the talc formula Mg3Si40,0(OH)2. Under B, the compositions are normalized with respect to a combined tetrahedral + octahedral cation valence of 14 in Table 1 and 22 in Table 2; this requires that the oxygen anions total 7 in Table 1 and 11 in Table 2. In both formulations, Mg, Ni, AI, and Fe 3+ are considered to occupy octahedral positions and E(octahedral cations) is taken as ~s 2+ + 3/2R3+). In discussing the chemical relations, Mg will be used to represent (Mg, Ni, A I . . . ) when no ambiguity arises between a magnesium end member and a nickel-containing garnierite.

Octahedral]tetrahedral, or O/T, cation ratios.

The formulae given under A and B in Tables 1 and 2 express primarily the ratios of atoms rather than their absolute numbers. In Table 1, the O/T ratios are distinctly less than the ideal ratio 3]2 for serpentine-group minerals, and in Table 2 are generally greater than 3/4 for a talc-like mineral. These deviations from the ideal ratios can be attributed to deficiencies of octahedral cations in 7 A-type garnierites, and of tetrahedral cations in 10 A-type garnierites.

An alternative approach, which is more or less in line with arguments presented by Faust (1966) and by Vitovskaya and Berkhin (1968, 1970), is to consider that the O/T ratios fall between 3/2 and 3/4 because of a mixing of the two kinds of layers. The mixing may be in various forms, such as mixed phases or interstratifications or even as defective layers. For example, a 2:1 talc-like layer that is deficient in tetrahedral cations and associated oxygen anions will appear chemically to consist partly o fa 1 : 1 type layer. Conversely, ifa 1 : 1 type layer is deficient in octahedral cations, so that parts

C C M - V o l . 21, No. 1 - C

32 G .W. BRINDLEY and PHAM THI HANG

I - - - 3 - - - " f I -1 I I J j I I

' 001 l ] l I [ I

~,,i 0201 I 1

V'W ?" I I

. / , ~ . J

~, ~..,..z \ j j l"

I _ _ , A _ I 1 _ _ 1 t l I I . , I 1 I I0 20 30 40 50 60

20, degrees

Fig. 3. Diffractorneter patterns of 10 ~.-type garnierites, after holding at I 10~ for 3 days.

of the octahedral sheet are missing, the 'free' silica will appear in a calculation as if it developed elements o f a 2 : 1 type structure.

From the foregoing considerations, one sees that an analysis of the chemical data in terms of a mixing of two kinds of ideal layers will indicate the extent of the deviation of a mineral from one or other of the regular end-member forms, but will not specify the nature of the deviation. Structural information can be obtained only by structure-sensitive methods of analysis. In putting forward the following analysis of the chemical data, the limitations of the method must not be overlooked.

Calculations of the effective proportions of 1 : 1 and 2:1 type layers in garnierites from their total compositions can be made as follows: In 7 ,&-type garnierites, let (1 - -x ) be the proportion of serpentine-like layers and x the proportion of talc-like layers, and in I 0 ,&-type garnierites let the converse arrangement hold. With the basic formulae written in the oxide forms, 3MgO.2SiO2-2HoO and 3MgO-4SiO2"H20, (where Mg includes Ni and other octahedral cations), then the composite

formulae are:

7 `&-type garnierites, 3 MgO. (2 + 2x) SiO2" (2 - x) H20, x = proportion of talc-like layers,

10 ,&-type garnierites, 3 MgO. (4 -- 2x) SiO2" (1 + x) H2 O, x = proportion of serpentine-like layers.

If [SiOJMgO] represents the mole ratio of these oxides, then

x = 1.5 [S iOJMgO] - 1 for 7 ,&-type garnierites

x = 2 - 1 . 5 [S iOJMgO] for 10 ,&-type garnierites.

Values of x are listed in Tables 1 and 2, and range up to 0.268 for 7 ,&-type minerals, and to 0.359 for 10 ,&-type minerals. Evidently the deviations from the normal regular arrangements can be considerable.

Water content, H204, of garnierites. Little atten- tion has been given previously to the water contents

THE NATURE OF G A RN IE RIT E S- I

Table 1. Chemical analyses and structural formulae of some 7 A-type garnierites

33

(1) #MN6 (2) #3996 Morro do Niquel, Morro do Cerisco (3) #HMC5

Brazil Brazil Riddle, Oregon (4) #New Cal

New Caledonia

Quartz 1.5 1-3 SiO2 (d) 43-5 43.7 AI20.~ 0.25 0.51 Fe203 2.20 6.10 MgO 35.1 30.4 NiO 4.40 5.50 CaO 0.20 0-05 Na20 0.04 0.13 K20 0.03 0.05 H~O+(a) 12'8 12"7 Total 100-0 100.4

A B A B A Si 2"00 2.09 2"00 2" 11 2.00 A1 0'01 0'01 0.03 0"03 0-01 Fe 3§ 0-08 0.08 0-21 0.22 0.09 Mg 2.41 2.51 2.07 2.19 1.57 Ni 0.16 0.17 0.20 0.21 0.70

Oct. (b) 2.70 2.81 2.63 2.77 2.42 H O+ 1.96 2.05 1.94 2.05 1.85 x(c) 0.110 0.141

1"2 43"2 0'24 2'74

22"7 18'6 0'12

11'9 100'7

B 2"18 0'01 0'10 1'71 0.76 2'62 2'01

0" 240

35.0 0-47 0-i0 2.70

49-3 < 0.05

0.14 0.08

11.4 99.2

A B 2.00 2' 14 0.03 0.03 0.00 0.00 0.23 0.25 2.27 2-43 2.54 2.70 2.17 2"31

0.180

(a) H20 + = weight loss from 110-1000~ (b) E Oct. = E(R2++3/2 R3+). (c) x = Proportion of 10 ~.-type layers, calculated from [Si[E Oct.]. (d) Total silica-quartz.

of these materials. In the present work, H 2 0 + is taken as the weight loss of material between 110 ~ and about 1000~ It is seen in Tables 1 and 2 that, with H 2 0 + included, the total analyses come near to 100 per cent. From the thermogravimetric curves, the weight loss is obtained with an uncertainty of the order of 2-3 per cent, which leads to an uncertainty of the order of 0.2-0.3 in the weight percent of H20+. The shapes of the thermogravimetric curves are considered in detail later.

Water content of lO.4-type garnierites. Figure 4 shows the H 2 0 + values given under A in Table 2 plotted against total octahedral cations, E(R2++ 3/2R3+), relative to Si = 4.00. A scale of x values calculated from the [MgO/SiO2] mole ratio, as described earlier, also is given. Over the range of compositions studied, x varies almost linearly with the octahedral cations, so that a plot of H 2 0 + versus x gives substantially the same kind of dia- gram. The vertical lines attached to each observed value correspond to an uncertainty of - - + 3 per cent.

The observed values fall near to a straight line, marked (c), which passes through the point H 2 0 + = 2, Mg = 3, S i = 4, and x = 0. The composition of this point, 3MgO.4SiO2.2H~O or [Mg3Si4010(OH)z]'H20 can be called talc monohydrate.

The question can be asked whether line (c) can be reproduced by mixtures of talc-like and serpentine-like components. With normal talc and normal serpentine compositions, line (a) is obtained, and with talc monohydrate and normal serpentine com- positionS, line (b) is obtained. To reproduce line (c), a more hydrous composition is required for the serpentine component, and in fact a combination of talc monohydrate and serpentine dihydrate repro- duces line (c).

An alternative procedure can be used, however, based on the slope of line (c), which gives an increment of 2 H 2 0 + per unit increment of octahedral cations, so that points on line (c) correspond to

[Mg3Si40,0(OH)2].H20 +p [M gO ' 2H 20] talc monohydrate

where p = octahedral cations in excess of 3.0 when referred to Si = 4.0. The additional term p[MgO. 2H20] can be written p[Mg(OH)2H~O] or p/2 [Mgz[S]'(OH)42H20], where Mg also includes Ni. These formulations suggest a hydrated brucite-like layer, or a cation-defective structure related to Mg3(OH)6, with one Mg 2+ replaced by 2H +. Ginzburg and Rukavishnikova (1951) also attributed the excess of octahedral cations to the possible presence of a free hydroxide.

34 G . W . B R 1 N D L E Y and P H A M T H I H A N G

o<

"6 .=

3 k; e,i

~ . .~o

o H ~

r ~ ~ I= I J O0~o

<

. ~ 0 ~

~ ~ .~

< ~ ~

m~-- ~ ~ A ~ A ~ ~~

. ~ o

~011 o


Of Q2 0.3 0.4 I x ~ I s+ /~

/ 1

H ~O + . (c)

2

~N ~s h drate ral~: mono y

r/5 1 ~ ~ , ~ , [ , ,

3.0 35 Octahedrd cations

Fig. 4. H20+ values of 10 .A-type garnierites versus total octahedral cations relative to Si= 4.0. The scale of x values, proportion of serpentine-type layers, is derived from [MgO/SiO~] mole ratios. Calculated lines (a), (b),

(c), - see text.

The concept of a talc (or nickel talc) monohydrate can be compared with the formula of a mica such as phlogopite:

Phlogopite: K Mg~ (SisAl) O10 (OH)z Talc monohydrate: H20 (Mg, Ni . . . . . )~ Si4 O10

(OH)2.

The comparison suggests that the additional H~O molecule occupies the position of the K + ion in phlogopite, i.e. in the holes of the quasi-hexagonal oxygen network. One recalls that H~O and K + have very similar packing radii, about 1"35 A. The suggested arrangement is compatible with the observed basal spacing, about 10 ~ as in micas, and not 9-4 as in normal talc. It is also compatible with the fact that the 10 ~ spacing does not readily diminish with mild heat treatment, and also with the fact that the additional H20 is not readily removed in thermogravimetric experiments.

The possible replacement of K + in a mica structure by H30 + has been discussed by Brown and Norrish (1952) to explain the low K~O, high H~O+ contents ofillites and Rosenqvist (1963) considered that HsO replaces K + in interlayer positions, with H + ions appearing elsewhere in the structure as

O H - in place of O '~-. Evidently it is not revolution- ary to suggest that the additional H20 in the talc monohydrate forrfiula may occupy interlayer positions in a 10 ,~-type layer structure.

Possible structure models of dominantly 10 ,'l-type garnierites. Various structure models can be suggested consistent with the chemical relations already discussed. Two such models are illustrated and compared in Fig. 5. Figure 5a represents a hydrated 2 : 1 talc-like layer with a defective silica sheet in which water molecules are placed where silica is missing. The defective part of the layer will appear in chemical analysis as a hydrated 1:1 serpentine-like layer. The overall basal spacing will be around l 0 A. Figure 5b is obtained by removing part of the lower tetrahedral sheet of Fig. 5a and placing it at the upper tetrahedral level. The result corresponds to a hydrated 2 : 1 talc-like layer, with an excess hydrated hydroxide sheet. The chemical equivalence of the two models is evident. Addition- ally, interstratifications and segregations of the 2:1 and 1 : 1 layers may exist. Such a complex system cannot be unraveled in detail from X-ray powder diffraction patterns.

Water content o f 7A-type garnierites. Figure 6 shows H 2 0 + values taken from Table 1, columns A, plotted against octahedral cations for Si = 2.0, together with one point P, taken from the data of Faust et al. (1969) for pecoraite, a nickel chrysotile. Four of the observed points lie on a straight line which extrapolates to about 2.1 H 2 0 + when the octahedral cations equal 3.0 and this is close to the value for normal serpentines. The remaining samples, apart from No. 4, indicate a decrease of Hz O+ with decreasing octahedral cations, which is

lO 0 0 0 0 O 0 0 0 0

1o oooooooo a

0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 b

Fig. 5. Schematic diagrams of defective talc-like layers. Shaded bands represent octahedral sheets; unshaded bands, Si-O tetrahedral sheets; open circles, water molecules. (a) shows a defective tetrahedral silica sheet replaced by water molecules. (b) shows part of the lower tetrahedral sheet of (a) moved to the upper level to complete, as far as possible, the 2:1 layer, and leaving an

excess of the octahedral sheet.

36 G . W . B R I N D L E Y and P H A M T H I H A N G

0.2 O1 , , , • / 2.5

-I [ 7A-type garnierite [ ]

1 ~ 2.0

t " ( b ~ " l ~ Serpentine

] 2.5 30

Odahedral cations

Fig. 6. H.,O+ values of 7 A-type garnierites versus total octahedral cations relative to S i = 2.0. The scale of x values, proportion of talc-like layers, is derived from [MgO/SiO2] mole ratios. Calculated lines (a), (b),-see

text.

generally what would be expected with admixed talc-like layers. With normal talc and normal serpentine compositions, one derives line (a) and with talc monohydrate and normal serpentine compositions, one derives line (b), which has nearly the same slope as that of the line through the experi-

mental points. Without elaborating the discussion, it appears that a serpentine composition modified by varying proportions of a talc monohydrate composition comes near to the observed data.

Thermogravimetric ( TG ) analyses Figure 7 gives the thermogravimetric curves

from which the values of HzO+ have been obtained; the fractional weight loss with respect to the sample weight at 110~ expressed as a percentage value, is plotted against temperature in ~ The curves were recorded as a variation of weight against time, with a superposed temperature-time calibration. The intervals of the temperature scale diminish with increasing temperature. The four curves in Fig. 7a and the seven curves in Fig. 7b correspond to the similarly numbered analyses for 7 A-type and 10 A-type garnierites in Tables 1 and 2 respectively.

The initial vapid weight loss, corresponding to loss of hygroscopic water, is largely complete by about 110~ which is taken as the lower temperature limit for obtaining values of H20+. The rapid losses at higher temperatures, 550-700~ in Fig. 7a and 800-900~ in Fig. 7b, arise mainly from

=--~. - - ' N . "

1. MN6

" ~ - . ~ 2, 3996

�9 " ~ ' - . 3, HMC5

4 New Cal . . . . ' " " .. ' - 5

o

�9 ~:-~'-'~..~:~..: ~ . ~ ~ . . -._< ~_. - ~ . ~ _ '"'- ~ ' : ' ~ Z ~ - ' ~ - . . ~ - - ~

1 /~ -ker "" "'" ~ " ~ " ~ ~

2, N C0~o~ - - ""~.~ "~... ~'~.. ~'q~ \~

5, ~o2~ ........... ........ " ......... ~ \ : ~ ' " - - RO3A "" """ " " . \ "~ '~- '~ ~10 - - 6 , " ...... ~ - -10

7, GUS

i I I I I I _ I I _I ....... J 100 200 400 600 800 1000

T e m p e r a t u r e ~

Fig. 7. Thermogravimetric curves of (a), 7 A-type garnierites, (b) 10 A-type garnierites. Sample numbers 1-4 in (a), and 1-7 in (b) correspond to samples listed respectively in

Tables 1 and 2.


dehydroxylation and these temperatures are normal for serpentine-group and talc minerals respectively. The weight becomes constant by about 1000~ which is taken as the upper limit for determining H20+. It was impracticable to extend the measure- ments beyond 1000~

The curves of Fig. 7b are of particular interest because of the high H~O+ values recorded for the 10A-type garnierites and the proposed monohydrate formula. However, no clear separation is found between the loss of water by dehydroxylation and the loss of 'monohydrate water'. The rapid loss between about 800 and 900~ where much of the dehydroxylation reaction occurs, does not account for all the dehydroxylation. As in other clay mineral reactions, the dehydroxylation process begins before and continues after the range of most rapid reaction. This is seen in the curves of Fig. 7a where most of the total weight loss is due to dehydroxylation. In Fig. 7b, the 'monohydrate water' is released in the temperature range from 110 to about 700~ throughout which there is a slowly accelerating loss of weight; there is little evidence for a retardation in the weight loss which might be attributed to the completion of the removal of the monohydrate water.

Two curves in Fig. 7b are anomalous. Curve 5, for sample RO24, shows a weight increase around 200~ which may be an oxidation effect. However, the Fe203 content of this sample (see Table 2) is not high compared with the values for other samples, and the NiO content is about the same as for samples RO3A and RO3B (all three samples are from Riddle, Oregon). No explanation has been found for this unusual behavior. Curve 7 in Fig. 7b

shows two inflections, with the first near 600 ~ which could correspond to a 7 A-type component; Table 2 shows that sample 7 has a high value of x, 0.274, the estimated proportion of serpentine-type layers.

Color characteristics and N i O content It is an interesting question whether color can be

related to the NiO content. The colors of the samples were matched against those in the exten- sive color charts given by Munsell (1929), in which colors are specified with respect to three variables, as follows: The hue, denoted here by u, is described qualitatively in zones, such as Y (yellow), GY (green-yellow), G (green) . . . . . and also by numbers with ten steps per color zone. Thus hue numbers 20-30 span the range of yellows, 30-40 green- yellows, 40-50 greens, 50-60 blue-greens, and so on. Color intensity or chroma is denoted by r and lightness by z, with 10 steps 0-10 for each variable. A Munsell color is denoted by u/z/r.

If the NiO content determines the "green-ness" of a sample, then the departure of the hue number from that of the lowest yellow, namely 20, is a useful parameter to consider. As regards the other two variables, it seems likely that NiO content will increase with r and decrease with z. Therefore the simplest combined function to consider is (u -29) r /z The results, set out in Table 3 and Fig. 8 show an almost proportional relation between NiO content and (u- -20)r / z . This result must be partly for- tuitous because the u scales in the Munsell color charts are given in steps of 2.5, and r and z in steps of unity. Evidently r, chroma, which varies between 1 and 6 for the samples studied, has a big influence on (u -- 20)r/z and an error of one unit in estimating

Table 3. Relation between green color of garnierite samples and nickel content

Type of garnierite

Munsell Mineral color* sample u/z/r

(u - 20)r NiO, wt % z analysed estimated

7 A type

10 A type

(1) MN6 32.5/8/2 3.1 4-4 (2) 3996 30/6/3 5-0 5-5 (3) HMC5 37.5/8/4 8.7 18.6 (4) New Cal 42.5/6/6 32-5 49.3

Serbia 2 42.5/7/4 12.9 Kambalda 40/6/6 20.0

(2) NC 45/8/1 3.1 5.5 (3) MN4A 45/8/2 6.3 16-8 (4) RO3B 40/8/4 10.0 18-2 (5) RO24 42'5/8/4 11.2 19.6 (6) RO3A 45/8/3 9-4 20.8 (7) GUS 42'5/8/6 16.9 31.5

Serbia 206 45/8/4 12-5 Serbia 201 45/8/6 18.8

26-27 40-43

25-27 37-39

*u = hue, z = lightness, r = color intensity or chroma.

38 G.W. BRINDLEY and PHAM THI HANG

2O

I0

r , i ' i ' i

~ '4 Q

0 2 ~

I , I , I t ' 10 20 / 0 40 50

N i O , wei~h l percent

Fig. 8. Color index (u--20)r/z versus NiO, weight percent. Symbols E], 1-4, correspond to 7 A-garnierites;

O, 1-7, correspond to 10 A-garnierites.

r may have a large effect. For example, the value of (u--20)r/z for sample MN4A, with r = 2, is 6"3, which is rather far from the mean line in Fig. 8; if r is given the value 3, the function becomes 9.4 and is then nearly on the mean line,

DISCUSSION AND CONCLUSIONS

X-ray diffraction data show that garnierites are commonly minerals with basal spacings near 7.2- 7.3 A and near 10 A, and only a minority of the samples examined appear to be dominantly of the one, or the other type. The diffraction characteristics of the 7 A-type resemble those of the serpentine group minerals, chrysotile and lizardite, but often are insufficient for specific identification. The diffraction characteristics of the 10 A-type suggest a talc-like mineral with a high degree of layer stacking disorder, and in some cases possibly with some interstratification of layers that exhibit a swelling behavior, but for the most part they are nonswelling minerals.

Chemical analyses show significant deviations from the compositions of serpentine group minerals and of talc as regards both the octahedral/tetrahedral cation ratios and the water, H20+, contents. The O/T cation ratios can be related composition- ally to mixtures of stoichiometric serpentine- and talc-like components, but unique structural models cannot be derived from such considerations. These results are broadly in agreement with those of Vitovskaya and Berkhin (1968, 1970).

The H 2 0 + contents of the 10 A-type garnierites are considerably in excess of that of talc and are shown to increase almost linearly with the number of octahedral cations. In particular, seven analysed samples give H 2 0 + values consistent with a line passing through the composition point 3MgO. 4SIO2"2H20 or [Mg3Si4010(OH)z]'H20. Compari- son of this formula with that of a phlogopite suggests that the additional "monohydrate water" is anal-

ogous to the K ion in the mica structure and may occupy a similar interlayer position.

The increase of H 2 0 + as the O/T ratio increases above 3 can be described by the composition [Mg~Si4010(OH)2]'H20 +p[MgO'2H20] where Mg also includes Ni. The second term can also be written in the form p[Mg(OH)z'H20]. However, it does not produce a chloritic type of mineral, because the basal spacing remains near 10 A. It is most likely that the material represented by p[Mg(OH)e'H20] arises from the defective nature of the 2 : 1 layer structure.

The impact of the present study on questions of nomenclature cannot be overlooked. The terms 7 A-type and 10 A-type garnierite or serpentine- like and "talc-like" garnierite are useful labels but evade the problems of selecting names. The present results provide no justification for using separate names for low Ni- and high Ni-containing minerals, and the present authors concur with Pecora, Hobbs and Murata, and with Faust, that garnierite is a suitable general name. The use of pimelite at present adds confusion rather than clarification because there seems to be no agreement on whether it is, or is not, a swelling-type mineral.

Acknowledgments-Thanks are extended to the U.S. Agency for International Development (U.S.A.I.D.) for a scholarship to Pham Thi Hang, and to the Georgia Kaolin Company for a grant supporting the research program. We are greatly endebted to many donors of mineral samples whom we thank collectively here, and to Dr. Michael Fleischer and Miss Jaroslava Nigrin for trans- lations of Russian articles.

REFERENCES Brindley, G. W. (1961) Quantitative analysis of clay mix-

tures. In X-ray Identification and Crystal Structures of Clay Minerah., (Edited by G. Brown), Chap. XIV, 2nd Edn. The Mineralogical Society, London.

Brown, G. and Norrish, K. (1952) Hydrous micas: Mineral. Mag. 29,929-932.

D'yakonov, Yu S. (1963) Results of an X-ray study of kerolite: Doklady Akad. Nauk SSSR 148, 909-911. (English Trans.) Earth Sci. Series 148 107-109.

Faust, G. T. (1966) The hydrous nickel-magnesium silicates-the garnierite group: Amer. Mineral. 51, 279- 298.

Faust, G. T. and Fahey, J. J. (1962) The serpentine-group minerals: GeoL Survey Professional Paper 384-A, pp. 1-92.

Faust, G. T., Fahey, J. J., Mason, Brian and Dwornik, E. J. (1969) Pecoraite, Ni6Si4Ox0(OH)8, the nickel analog of clinochrysotile, formed in the Wolf Creek meteorite: Science 165, 59-60.

Ginzburg, I. I. and Rukavishnikova, I. A. (1951) Minerals from the ancient crust of weathering of the Urals: lzdat. A kad. Nauk. SSSR, 714 pp. [CA 11267, 1954].

Kato, Toshio (1961) A study of the co-called garnierite

THE NATURE OF G A R N I E R I T E S - I 39

from New Caledonia: Mineralogical J. (Japan)3, 107-121.

Maksimovic, Z. (1966) /3-kerolite-pimelite series from Goles Mountain, Yugoslavia. Proc. Intern. Clay Conf., Jerusalem. 1, 97-105.

Medlin, J. H., Suhr, N. H. and Bodkin, J. B. (1969) Atomic absorption analysis of silicates employing LiBO2 fusion: AtomicAbs. Newsletter, 8, 25-29.

Muchi, M. (1965) Electron microscopic observation on some fibrous minerals altered from serpentine minerals in the 0eyama District: Bull. Fukuoka Gakugei Univ. 15, 81-92.

Muchi, M. (1965) 10 ~ garnierite associated with nickelif- erous sepiolite from the 0eyama nickel mine, Kyoto Prefecture: Bull. Fukuoka Gakugei 16, 153-169.

Pecora, W. T., Hobbs, S. W. and Murata, K. J. (1949) Vari- ations in garnierite from the nickel deposit near Riddle, Oregon: Econ. Geol. 44, 13-23.

Rosenqvist, 1. Th. (1963) Studies in position and mobility

of the H atoms in hydrous micas: Clays and Clay Minerals 11, 117-135.

Slansky, E. (1955) The Ni-hydrosilicates from Kremze in South Bohemia. (Czech text, English abstract). Uni- versitas Carolina, Geologica 1, 1-28.

Spangenberg, K. (1938) Die wasserhaltigen Nickelsili- kate: Zentr. Mineral. Geol. A. 360-364.

Vitovskaya, I. V. and Berkhin, S. I. (1968) The problem of the nature of kerolite: Kora Vyvetrivaniya 10, 134-159.

Vitovskaya, I. V. and Berkhin, S. I. (1970) Nature of garnierite: Kora Vyvetrivaniya 11, 26-39.

Vitovskaya, I. V., Berkhin, S. I. and Yashina, R. S. (1969) The serpentine component of nickel silicates: Doklady Akad. Nauk SSSR 189, 1092-1094. (English Trans.) Earth Sci. Series 189,160-162.

Whittaker, E. J. W. and Zussman, J. (1956) Characteriza- tion of serpentine minerals by X-ray diffraction: Miner- alog. Mag. 31,107-126.

R~sum~- Les diagrammes de diffraction X des garni6rites indiquent que la plupart des 6chantillons ressemblent ~t des min6raux du groupe serpentine, ou hun min6ral du type talc, ou hun m~lange de ces deux formes, et donnent respectivement des r6flexions basales ~t 7 et 10 A. A partir d'une 6tude pr6alable de 40 garni6rites environ, on a s61ectionn6 pour une 6tude d6taill6e quatre 6chantillons du type serpentine pr6dominant et sept ~chantillons du type talc pr6dominant. Les garni6rites du type talc montrent peu de variation dans I'espacement basal ?~ 10 A Iors du chauffage h des temp6ratures peu 61ev6es ou Iors de l'immersion dans les liquides, en d6pit du fait que certaines d'entre elles contiennent une petite proportion de min6raux gonflants. Les analyses chimiques montrent des d6viations consid6r- ables dans les rapports des cations octa6driques/t6tra6driques par rapport aux valeurs ~ et ] trouv6es dans les min6raux normaux serpentine et talc; ceci peut 6tre interpr6t6 en termes du m61ange de couches 1 : 1 et 2 : 1, en phases s6par6es et/ou en interstratification, ou sur la base de structures con- tenant des d6fauts de diff6rentes natures. Les teneurs en HzO + des formes de garni6rite du type talc sont consid6rablement plus 61ev6es que celle du talc normal et indiquent l'existence d'un min6ral dont la composition, 3(Mg, Ni)O.4SiOz.2H~O ou (Mg, Ni)3Si~O,~(OH)z.HzO est celle d'un monohydrate tu talc. La couleur verte des garni6rites est reli6e ?~ la teneur en NiO en pour cent, et un index de couleur fond6 sur les tables de couleur de Munsell est ~tabli.

Kurzreferat-- R6ntgendiffraktionsbilder von Garnieriten deuten darauf hin, dab die meisten Proben Gesteinen der Serpentingruppe oder einem talkartigenoGestein bzw. einer Mischung dieser beiden Formen gleichen. Sie ergeben Basalreflexionen von 7 A bzw. 10 A. Von insgesamt 40 Garnieriten wurden vier von vorwiegend serpentiner Beschaffenheit und sieben von vorwiegend talkartiger Beschaffenheit fiir genaue Untersuchung ausgew~ihlt. Bei den talkartigen Garnieriten variiert die 10 A Basalstruktur bei geringer Erw~irmung oder Untertauchen in Fliissigkeiten wenig, obgleich einige davon einen kleinen Anteil an expansionsf~ihigen Schichten enthalten m6gen. Chemische Analysen erweisen erhebliche Abweichungen der Oktahedral-/Tetrahedral-Kationenverh~iltnisse von den Werten 3[2 und 3/4 ffir normale Serpentin- und Talkgesteine, und sie kSnnen als gemischte 1 : 1 und 2 : 1 Schichtarten (separate Phasen und/oder Einlagerungen) oder als defekte Strukturen verschiedener Arten interpretiert werden. Die HzO+ Gehalte der talkartigen Garnieritformem sind erheblich h6her als die von normalem Talk und weisen auf ein Gestein mit der Zusammensetzung 3(Mg, Ni)0.4SiO~. 2H zO oder [(Mg, Ni)zSi40 ~o(OH)2].H ~O hin-also auf eine Talkmonohydratformel. Die griine Farbe der Garnieriten wird mit dem NiO Gewichtsprozentsatz in Beziehung gebracht und ein Farbindex, der auf den Munsell Farbtabellen basiert, wird abgeleitet.

P e a m M e - PeHTreuorpaMMbI rapHHepHTOB yKa3bIBarOT, qTO 6OYlbmHHCTBO o6pa3I~OB HanOMHHa~OT MnHepanbI nH60 cepneHTrtnoBo~ rpyrtnbi, 3I]46o Ta~bKOO6pa3HbIe, ZIn60 CMeCb 3THX qbOpM H ~a~OT OCHOBHBIe oTpa~eHH~l 7 /~ a 10 /~ COOTBeTCTBeHHO. I,'I3 40 BHJIOB rapHnepnToa ~;~ ~eTa~bnoro aHanr~3a B~,i6pana qeTblpe npeI~MymecVBeHHO Tana cepnenTr~Ha rl CeMb npenMyttleCTBeHHO Tanr, Ko- o6pa3HblX. HamnH, qTO npn He3HatlHTe~bnOM narpeBaHHa ann i~pi~ ~orpy~eaHH B mnaKOCTb OCHOBHa~I npocnoflia xanbKoo6pa3nr, lx rapHr~epnToB 10 A ne3na~InTehbHO MeH~eTC~, XOTn, lfeI(OTO- pl, Ie n3 HHX MOryT co~epmaT~, pacmHpnrott~necn c~on. XnMI, IqecKI4~[ arla.rlrI3 noia3t, maeT 3na,tHTe.rlb- H~,~e OTrnoHenan OTHOmeHH~ KaTnOHOB OKTa3~p/xeTpaa)Rp OT 3Ha~eHn,~ 3/2 r~ 3/4 Ran aopMa.ababrX ceprlenTI, IHOB H Ta~bKOBbIX MHaepanoB u OH~I MOFyT 6btTb pacCMOTpeHbt Ka~ THI~bI CO CMeRIaI-IHBIMI, I cnoaMn 1:1 a 2:1, ann ~r pa3nr~qHb~e ~eqbe~T~,r B cTpy~Type. Tan~Koo6pa3m, ie rapnr~epHT~,t

40 G. W. B R I N D L E Y and P H A M T H I H A N G

co~ep~aT 60~b~ae H 2 0 , + qeM HOpMaYlbHbI~I Ta3IbK H 3TO yKa3blB~teT Ha M~aepasI coe~;aHea~eM 3(Mg, N i ) O . 4 S i O 2 - 2 H z O n a n (Mg, Ni)3Si4Olo(OH)2- H z O 2- - - qbopMyaa o)1HoocnoBnoro T a ~ x a . 3 e ~ e H ~ ~SeT rap~rteprIxoB 3aBHCHT OT c o ~ e p ~ a H I ~ B I/pOlleHTaX 170 Becy N i O H FIH~eKC LIBeTa Ran n o T a 6 n a u e LIBeTOB M y a c e ~ a .

Documents

THE NATURE OF GARNIERITES--I STRUCTURES, CHEMICAL ...clays.org/journal/archive/volume 21/21-1-27.pdf · THE NATURE OF GARNIERITES--I STRUCTURES, CHEMICAL COMPOSITIONS AND COLOR CHARACTERISTICS