Surface Typochemistry of Hydrothermal Pyrite

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 13

ISSN 0016-7029, Geochemistry International, 2008, Vol. 46, No. 6, pp. 565–577. © Pleiades Publishing, Ltd., 2008.

Original Russian Text © V.L. Tauson, D.N. Babkin, E.E. Lustenberg, S.V. Lipko, I.Yu. Parkhomenko, 2008, published in Geokhimiya, 2008, No. 6, pp. 615–628.

Surface Typochemistry of Hydrothermal Pyrite:


Electron Spectroscopic and Scanning Probe Microscopic Data.
I. Synthetic Pyrite
V. L. Tauson, D. N. Babkin, E. E. Lustenberg, S. V. Lipko, and I. Yu. Parkhomenko
Vinogradov Institute of Geochemistry, Siberian Branch, Russian Academy of Sciences,
ul. Favorskogo 1a, Irkutsk, 664033 Russia
e-mail: vltauson@igc.irk.ru
Received February 19, 2007

Abstract—Techniques of X-ray photoelectron and Auger electron spectroscopy, scanning probe microscopy
were used to demonstrate that the natural surface of hydrothermally synthesized pyrite, as well as vacuum frac-
tures, contain a number of sulfide-sulfur species: disulfide, monosulfide, and, more rarely, polysulfide. The nat-
ural surface of hydrothermal pyrite is chemically modified compared to the inner volume into a nonautonomous
phase film up to ~500 nm thick, which has a variable composition resembling that of pyrrhotite but with broader
variations toward FeS2. Its principal distinctive feature is the presence of a peak at ~710 eV in the XPS Fe 2p3/2
spectrum, which is often higher than the main peak of bivalent low-spin Fe(II) in the pyrite structure (707 eV).
The “basic” structure of the nonautonomous phase is a layer of variable composition Fe2+[S, S2, Sn]2–, whose
S/S2 ratio varies from ~0.5 to ~2.0, averaging at ~1.1. This layer may include admixtures of minor elements, as
follows from the appearance of a nonautonomous phase in the presence of As, which does not, however, form
an individual phase. The polymerization of S at the surface is thereby more significant. The major oxisulfide
components of this phase may be the sulfite and thiosulfate ions at a subordinate concentration of sulfate
because of the instability of coexisting sulfate and disulfide ions, which results, in the presence of oxygen, in
sulfite (thiosulfate) and sulfide ions in the nonautonomous phase. In line with XPS, scanning probe microscopic
(SPM) data show that, at a high S activity in the “pure” system, the surface of the crystals contains practically
no nanometer-sized phases and is characterized by low roughness (14–17 nm). At a low S fugacity in equilib-
rium with pyrrhotite and sphalerite, the average roughness of the surface increases to 25–65 nm, with the max-
imum height of the surface features of ~100–500 nm. This is consistent with Auger spectroscopic data, obtained
after the etching (ion milling) of the surface with Ar+, on the thickness of the nonstoichiometric surface layer.
Comparison with analogous data on other sulfides shows that crystals growing in hydrothermal environments
have surface layers up to ~500 nm thick, which are different from the main volume of the crystal in chemistry,
stoichiometry, and, possibly, also structure. This is scale of the surface heterogeneity at which the typochemis-
try of mineral surfaces may be manifested. The typochemistry of pyrite stems from the ability of the nonauto-
nomous phase to “record” the growth conditions of crystals in terms of two major factors: the purity of the sys-
tem (the occurrence of other phases, including virtual ones, i.e., potentially possible phases of admixture ele-
ments) and S fugacity (which influences the S/S2 ratio at the surface). The geochemical role of the surface
nonautonomous phase in pyrite may be very significant, particularly when minor elements are captured that are
incompatible with the pyrite structure but can be easily accommodated in the less rigid structure of the nonau-
tonomous phase.
DOI: 10.1134/S0016702908060037

INTRODUCTION ical equipment of X-ray photoelectron spectroscopy


The most dynamically developing avenues in the (XPS) to the use of synchrotron radiation (XPS-SR),
chemistry of mineral surfaces are the investigation of which provides a higher surface sensitivity and spectral
pristine fracture surfaces, which are obtained in an inert resolution. This approach made it possible to examine
gas atmosphere or under vacuum in the intermediate in much detail atomic transformations in the surface
(transfer) chamber of a spectrometer, and the examina- layer (for example, with reference to pyrite, discussed
tion of the mechanisms of oxidation, leaching, and here) and to identify such phenomena as autofredox
other analogous processes proceeding at the surfaces of reactions and polymerization of the ligand [1, 2]. At the
minerals during the processing and concentration of same time, the geochemical and genetic aspects of the
ores. The progress achieved in these fields (particularly problem remained largely uncertain, because these
in the former) over the past years was related mostly to results did not immediately pertain to real (but not syn-
the transition from the application of traditional analyt- thetic, obtained under high vacuum, which is atypical

565
566 TAUSON et al.

of natural terrestrial processes) mineral surfaces. were as large as 5 mm (along edges) and had cubic (in
Another methodology of the research was proposed in the pure system) and cuboctahedral and more compli-
[3, 4] and involved the distinguishing of the primary cated (in the systems with As) morphologies [9]. Some
(juvenile) surface and factors affecting it depending on of the experiments were carried out with the use of Ti
the conditions under which the minerals were pro- “traps” to gather the fluid for analysis and pH determi-
duced, as well as studying the chemistry of the surfaces nation.
and its variations with depth in the crystal for minerals
from ore deposits of various genetic types. Our research The surface of the synthesized crystals was analyzed
was centered on a detailed justification of this by X-ray photoelectron spectroscopy (XPS) and Auger
approach. In its first part (presented in this publication), electron spectroscopy (AES) on LAS-3000 (Riber).
we analyze data on pyrite crystals synthesized under The instrument was equipped with an OPX-150 hemi-
hydrothermal conditions. The second part of this spherical electron analyzer with a retarding potential
research (an oncoming paper) will be devoted to natural and Auger electron spectrometer with a cylindrical mir-
pyrite crystals from gold deposits of various genetic ror analyzer and energy resolution of no worse than
types. We will thereby pay attention to advances along 0.3%. Atoms at the sample surface were excited by the
the aforementioned “classic” lines of research concern- radiation from an Al anode (AlKα = 1486.6 eV) at an
ing the mechanisms of the reconstruction and stabiliza- emission current of 20 mA and a tube voltage of 10 kV.
tion of mineral surfaces, although, as could be a priori Vacuum in the chamber was 5 × 10–10 torr. The spectra
thought, the applicability of this information to real sit- were calibrated on the C1s line of carbon with a binding
uations is limited. We focused our study on pyrite energy of 285 eV. The spectra were standardized using
because the geochemistry, morphology, and structure the line Au 4f7/2at 84.0 eV. The spectra were processed
of this mineral widely spread at ore deposits, as well as by a specialized program with regard for the nonlinear
its typochemistry and typomorphism, continuously background and the mixed Lorentz-Gaussian shapes of
attract keen interest of many researchers in hope of the peaks. This made it possible to subtract overlapping
developing criteria for searches for and identification of peaks and to obtain exact energy values (within ±0.1 eV),
ore deposits and evaluating the physicochemical condi- FWHM parameter (bull peak width at its half-maxi-
tions under which ore mineralization was produced at mum height), and areas beneath the peaks, which were
them [5, 6]. At the same time, many aspects of the sur- used to calculate the atomic concentrations of elements
face properties of this mineral remain largely obscure. and their speciation (in the presence of chemical shifts)
accurate to ±10%. The surface of the samples was ion
milled with a beam of Ar+ ions with an energy of 2 keV
EXPERIMENTAL AND ANALYTICAL and an emission current of 20 mA. For this purpose, Ar
TECHNIQUES was pumped into the analytical chamber to a pressure
of 2 × 10–5 torr.
Pyrite crystals were synthesized (1) in sealed optical
quality quartz glass ampoules at T = 400°C and P = The Auger spectra were registered in the differential
0.5 kbar [7] and (2) by the conventional technique of form dN(E)/dE at a primary beam energy of 2.5 keV
hydrothermal thermal-gradient synthesis in Ti inserts at and a modulation voltage of 2.3 V. To run the analytical
T = 400 and 500°C and P = 0.5 and 1 kbar [8]. The profiles, i.e., to analyze the crystal composition for a
inserts or quartz glass ampoules were placed in stain- certain depth from its surface, we etched the crystal at
less steel autoclaves and held for 3 days at 400 or 450°ë a rate of approximately 0.1 nm/s by a beam of Ar ions.
to homogenize the material and then for 9 days at a tem- The Auger spectra were successively registered at spec-
perature gradient (15°ë along the outer wall of the ified time intervals of etching (5–10 min). We applied
autoclave). The mineralizers were either pure NH4Cl conventional techniques for the processing of spectra
solutions or those with added HCl or Na2S and (more in order to obtain qualitative and quantitative informa-
rarely) 20% solution of CaCl2. The experiments ended tion on the composition of the surface [10]; the atomic
with the quenching of the autoclaves in running cold concentrations of elements were calculated from the
water. Pyrite was synthesized in both “pure” systems, intensities of Auger peaks with the use of element sen-
whose starting materials contained only Fe and S, and sitivity coefficients. The peaks were localized accurate
multicomponent systems, which contained, along with to ±0.1 eV, and elemental compositions were analyzed
these elements, also ZnS or As. Pyrite crystallized accurate to ±10% rel.
together with sphalerite or sphalerite + pyrrhotite in the
former case, and no additional phases were formed in The surface topography of the crystals was exam-
the latter one, but the crystals contained As (a few hun- ined on an SMM-2000 (Russia) scanning probe in the
dredths of a weight percent in the bulk of the crystal and modes of scanning tunneling microscopy (STM) and
up to ~1 wt % in the surface layer; according to Auger atomic force microscopy (AFM) in contact mode. In
electron spectroscopy, of As L3M45M45). The single the absence of vacuum at solid conducting samples, the
pyrite crystals obtained in quartz ampoules were no resolution of STM amounted to a few tenths of a
larger than 1 mm, and we could examine them only by nanometer. The software of the microscope makes it
Auger spectroscopy. The crystals grown in Ti inserts possible to analyze the roughness and other character-

GEOCHEMISTRY INTERNATIONAL Vol. 46 No. 6 2008


SURFACE TYPOCHEMISTRY OF HYDROTHERMAL PYRITE 567

istics of surfaces and determine the height and shapes Table 1. Energy of the S 2p3/2 binding of S particles in the
of nanometer-sized features on them. structures of pyrite and some other compounds
Crystal matrix
S species Binding energy, eV
IDENTIFICATION OF S AND Fe SPECIATION or compound
ON THE SURFACE OF PYRITE CRYSTALS S2– FeS2 161.4–161.8
We used a traditional radiation source to excite XPS S2– (sur.)* FeS2 161.2
spectra and thus failed to identify fine features of the 2–
surface S and Fe speciation (compare with [1, 11]). S 2 (sur.)* FeS2 162.0
However, high-resolution data led us to propose a 2–
model for the convolution of the spectra in situations S 2 (bulk)* FeS2 162.7
allowed by the resolution of the peaks. For example, if 2–
S2 FeS2 162.3–162.5
the difference between the “bulk” and “surface” sulfide
and disulfide ions (of the order of a few tens of an eV) 2–
S n (n > 2) FeS2 163.2–163.7
is comparable with the resolution of the peaks of the
spin-orbital S 2p doublet, then the sulfide and disulfide S0 (S8) α-S 164.0–164.2
ions can be distinguished even if traditional methods 2– Na and K sulfites
are applied, because the chemical shift then amounts to SO 3 165.6–167.5
and hydrosulfites
1 eV, and modern software taking into account the 2– Na thiosulfate and thio- 161.7–162.9 (SII)**
actual shapes of the peaks and overlaps makes it possi- S2 O3
sulfate pentahydrite 167.6–168.6 (SVI)
ble to resolve them. Table 1 presents binding energy val-
ues for the component 3/2 of the S2p3/2–1/2 doublet 2–
S2 O y Na and K sulfoxides 163.8–169.0
(mostly based on data on pyrite; in the absence of infor- (y = 5÷8)
mation on this mineral, we utilized data on correspond-
2–
ing chemical compounds). Analogous data on Fe 2p3/2 SO 4 Fe2+ and Fe3+ sulfates 168.7–169.1
are reported in Table 2. In the analysis, we used materi-
als from [1, 2, 11–20] and sources referred to therein. If * (sur.) and (bulk) mean purely surface (atoms in the first layer) and
“bulk” (deeper layers) states, respectively (XPS-SR data);
the volume of the data was sufficient (n ≥ ~10), we ** S in two oxidation states (formally, S2– and S6+)
selected average values, otherwise intervals and expert
estimates were utilized. The data listed in Tables 1 and 2
provide a rough idea on the distribution of Fe and S on Table 2. Position of the Fe 2p3/2 peak for various Fe2+ and
the surface of pyrite, and this distribution may be char- Fe3+ compounds
acterized as follows:
Range of values, Most probable value,
(1) bivalent “pyrite” Fe (bulk, as it is referred to in Compound
eV eV
2–
[1]), i.e., that is bound to the S 2 anion in the same FeS2 706.5–707.6 707.0
manner as in the pyrite structure, makes the predomi- FeS 706.8–713.6 710.0
nant (in terms of intensity) contribution to the XPS
peak of Fe 2p3/2 and has a bond energy of 707 eV; this FeO 709.1–710.7 709.8
peak may be correlated with the S 2p doublet, which FeCO3 710.2–710.5 710.5
2– FeSO4 710.8–713.6 711.0
corresponds to the S 2 , dianion with a binding energy
Fe3O4 707.9–711.4 709.8
of the most intense peak 2p3/2  162.5 eV;
Fe2O3 710.4–711.8 711.0
(2) bivalent “nonpyrite” Fe, which can be related to
FeOOH 710.2–711.8 711.1
the presence of Fe2+ sulfide, oxide, or carbonate in the
absence of the sulfate ion on the surface; the position of Fe2(SO4)3 713.1–713.3 713.2
the Fe 2p3/2 peak is in this case at 710 eV, and its inten-
sity should correlate with either the line of oxygen (if
this is oxide or carbonate) or the sulfide sulfur peak (5) an intermediate S oxidation state, which is char-
(S 2p3/2  161.5 eV, if this is monosulfide); acterized by peaks S 2p3/2 within the range of 166
(sulfite) to 169 eV (salts of higher sulfoxide anions).
(3) trivalent Fe, which is characterized by the posi-
tion of the peak Fe 2p3/2 at 711 eV, if it is not sulfate
but oxide or oxi-hydroxide; the peak of the Fe–O com- RESULTS
ponents on the pyrite surface in contact with air has the
same position [20]; Auger Spectroscopy
(4) sulfate of trivalent Fe, which is characterized by Based on the analysis of differential Auger spectra
the greatest chemical shift for both Fe and S, the peak of pyrite crystals obtained at 450°ë, 1 kbar in NH4Cl
Fe 2p3/2 is displaced to 713 eV and the peak S 2p3/2 solutions, we have arrived at the conclusion [4] that the
shifts toward 169 eV. surface of hydrothermal and metasomatic pyrite is oxi-

GEOCHEMISTRY INTERNATIONAL Vol. 46 No. 6 2008


568 TAUSON et al.

Table 3. Data obtained on hydrothermally synthesized pyrite crystals with the use of Auger electron spectroscopy with the
ion milling of the surface with Ar+
Conditions of synthesis Depth of the Composition of surface
Exp. no.
T, °C P, kbar Solution composition layer, nm* (S/Fe)at O, at %
1 450 1 20% CaCl2 120 1.0–1.1 6–10
2** 450 1 20% NH4Cl 30–240 1.3–1.9 2–8
3 400 0.5 10% NH4Cl 90–240 1.0–1.5 0–6
4 400 0.5 same 60–240 1.0–1.7 5–7
5 400 0.5 " 60–360 1.1–1.9 0–8
6** 400 1 9% NH4Cl + 1% HCl 30–360 1.0–2.0 0–5
7*** 400 1 10% NH4Cl 120 0.5–0.6 7–15
Note: experiments 3–5 were conducted in quartz ampoules, and the others were carried out in Ti inserts;
* approximate value calculated from the average etching rate [3];
** in association with S0;
*** in association with pyrrhotite and sphalerite.

dized to a depth of approximately 0.1–0.5 µm. The general tendency is an increase in the S/Fe ratio with
original (juvenile) surface of a crystal growing under depth in the crystal, and the non-stoichiometry occurs
these conditions without adsorbed gases may contain within a layer ~300 nm thick. Our data indicate that the
about 10–20 at % oxygen in the oxide–sulfide layer. juvenile surface of hydrothermal pyrite can contain sul-
These measurements were conducted on samples that fide sulfur in two modes: as S2– and S– (the formal S
were held in air for a fairly long time after the experi- valence in the disulfide ion). This observation is impor-
ments, and this could modify the composition of their tant for the further analysis of the X-ray photoelectron
surface layers [19, 20]. Because of this, we undertook spectra.
an analogous study of fresh samples, which were ana-
lyzed within as brief as possible time span after their
synthesis and which were stored in an Ar atmosphere in XPS
sealed glass ampoules during this time. The results are Table 4 reports the conditions under which the crys-
presented in Table 3 and reveal a degree of oxidation of tals were synthesized and the results obtained by ana-
the fresh surface of hydrothermal pyrite (synthesized at lyzing their spectra. For brevity’s sake, the table does
400–450°ë in chloride solutions in quartz or Ti contain- not show some less important details, for example, the
ers) even lower than the degrees of oxidation reported characteristics of the less intense component 2p1/2 of the
in [4]. The oxygen concentrations in the surface layers S 2P doublet and the proportions of the Lorentz and
of the crystals are at a minimum (0–8 at %) in the pres- Gauss functions in the deconvolution of the peaks. Note
ence of free sulfur (S0) and are slightly higher (up to that the shapes of the peaks were approximated by a
15 at %) in association with pyrrhotite, i.e., at the low- mixed Lorentzian–Gaussian function in variable pro-
est possible sulfur activity for pyrite at a given temper- portions. The results reported in Table 4 will be dis-
ature. In general, we detected a decrease in the oxygen cussed below, and here we will consider only the most
concentration with depth in the crystals (in the course topical problems related to the peaks. In order to eluci-
of its etching). This decrease is not always monotonous, date the situation, we will compare all obtained data
most probably because of the heterogeneity of the sur- and the results of crystal etching. It should be, however,
face itself and because it is hard to successively etch and borne in mind that etching notably differs in XPS and
conduct analysis at exactly the same points [21]. The Auger spectroscopy. In the former case, Ar+ ions bom-
thickness of the oxygen-bearing layer of the crystal is bard the whole surface of the sample, which is fairly
~200–400 nm, according to the earlier data [4]. It is par- uneven and rough, because it consists of surfaces of dis-
ticularly interesting to examine the stoichiometry of the crete, randomly oriented crystals. Because of this, the
surface layer [22, 23]. Table 3 shows the S/Fe atomic etching rate in this situation is practically unknown:
proportions, which do not reach the pyrite stoichiome- randomly oriented crystals are ion-milled at different
try even with regard for (O, Cl) anionic admixtures in rates, and it is hard to expect any clear-cut correlations.
the surface layer but assume intermediate values In our experiments, etching was carried out for 0.5 and
between 1 and 2 (i.e., FeS and FeS2). Values close to 2 1 h, and its effect was pronounced the most conspicu-
are characteristic of crystals obtained in the presence of ously in a decrease of the concentrations in the surface
excess S, when S0 occurred as an individual phase. The of oxide sulfur species: they decreased to value compa-

GEOCHEMISTRY INTERNATIONAL Vol. 46 No. 6 2008


SURFACE TYPOCHEMISTRY OF HYDROTHERMAL PYRITE 569

Table 4. XPS data on synthetic pyrite


Conditions of synthesis XPS data***
Exp.
P, Solution com- Fe 2p3/2 S 2p3/2
no. T, °C Phases**
kbar position* E, eV F, eV Rel. peak at % E, eV F, eV Rel. peak at %
7 400 1 10% NH4Cl Py, Po, Sp 707.2 2.1 FeII–S2 3 161.0 6.9 S 51
(20.6; –8.0) 711.0 4.0 FeII–S (np) 68 163.0 4.8 S2 39
713.0 4.0 FeIII–SO4 29 168.0 5.0 SO4 10
8 400 1 same Py, Sp 708.0 2.5 FeII–S2 30 161.0 3.0 S 28
(1.9; –6.9) 710.0 3.0 FeII-S (np) 42 163.0 4.8 S2 51
712.5 2.5 FeIII–SO4 28 168.0 5.0 SO4 21
9 400 0.5 " Py, S0 706.0 4.0 FeII–S2 76 162.0 1.6 S (in S2O3) 34
(–2.4) 709.1 3.3 FeII–S (np) 18 163.5 3.1 S2 + S0(Sn?) 34
712.2 2.5 FeIII–S2O3 6 167.0 5.0 S2O3 32
etching for 1 h
706.0 4.0 FeII–S2 91 161.5 3.0 S 42
710.0 2.5 FeII–S (np) 9 163.0 4.0 S2 50
167.1 4.9 S2O3 8
10 400 0.5 " Py 706.0 4.0 FeII–S2 80 161.0 3.0 S 49
709.7 3.1 FeII–S (np) 16 162.5 2.7 S2 48
713.0 2.5 FeIII–SO4 4 167.5 2.0 SO4 3
etching for 1 h
706.0 5.5 FeII–S2 95 161.0 3.0 S 38
711.0 2.6 FeII–S (np) 5 163.0 4.0 S2 60
167.5 2.0 SO4 2
11 450 1 " (0.9) Py 706.8 4.0 FeII–S2 100 161.6 3.5 S 56
163.0 4.1 S2 35
166.0 3.0 SO3 9
etching for 0.5 h
706.0 4.0 FeII–S2 71 161.6 3.0 S 45
712.0 7.0 FeIII–SO3 29 163.4 3.1 Sn 37
166.0 4.0 SO3 18
12 450 1 9% NH4Cl Py (As) 706.7 5.2 FeII–S2 65 160.8 5.0 S 67
+ 1% Na2S 710.1 4.6 FeII–S (np) 35 162.3 3.3 S2 33
(8.2)
etching for 0.5 h
706.8 3.9 FeII–S2 56 162.0 3.1 S 55
710.4 6.0 FeII–S (np) 44 163.4 2.9 Sn 29
166.0 3.0 SO3 16
13 450 1 9% NH4Cl Py, S0 706.8 4.5 FeII–S2 100 162.5 2.5 S2 61
+ 1% Na2S 164.3 2.5 S0 26
(7.2) 166.0 3.0 SO3 9
168.0 2.0 SO4 4
etching for 0.5 h
706.0 4.0 FeII–S2 90 160.8 2.5 S 47
710.1 3.0 FeII–S (np) 10 162.3 3.2 S2 51
166.0 2.1 SO3 2
14 450 1 10% NH4Cl Py (As) 706.1 4.0 FeII–S2 28 161.5 2.8 S 40
(1.0) 709.3 5.0 FeII–S (np) 49 163.5 2.6 Sn 52
711.0 4.0 FeII–S2O3 23 167.0 2.0 S2O3 8
etching for 0.5 h
707.0 5.0 FeII–S2 48 161.0 2.1 S 38
710.0 5.0 FeII–S (np) 30 163.3 2.4 Sn 49
712.0 7.0 FeIII–S2O3 22 167.0 2.0 S2O3 13

GEOCHEMISTRY INTERNATIONAL Vol. 46 No. 6 2008


570 TAUSON et al.

Table 4. (Contd.)
Conditions of synthesis XPS data***
Exp.
P, Solution com- Fe 2p3/2 S 2p3/2
no. T, °C Phases**
kbar position* E, eV F, eV Rel. peak at % E, eV F, eV Rel. peak at %
15 450 1 10% NH4Cl Py (As) 706.1 3.5 FeII–S2 32 161.3 2.3 S 22
(1.2) 709.0 3.5 FeII–S (np) 31 163.0 2.2 S2 32
712.0 7.0 FeIII–S2O3 37 167.5 5.0 S2O3 37
170.0 2.0 SO4 9
etching for 0.5 h
706.6 3.9 FeII–S2 44 161.0 2.5 S 49
710.5 4.0 FeII–S (np) 56 163.0 2.7 S2 41
167.0 2.0 S2O3 10
16 450 1 20% CaCl2 Py (As) Poor resolution of the spectrum 161.0 2.6 S 50
(0.9) 163.3 3.0 Sn 44
167.0 2.0 S2O3 6
etching for 0.5 h
706.5 3.8 FeII–S2 48 160.5 3.0 S 55
711.0 7.0 FeII–S (np) 52 162.5 3.0 S2 37
166.0 2.0 SO3 8
* numerals in parentheses show the pH of the trapped solution;
** synthesized phases: Py—pyrite; Py(As)—As-bearing pyrite (<~0.1 wt % As); Po—pyrrhotite; S0—elementary sulfur;
Sp—sphalerite (numerals in parentheses show mol % FeS in sphalerite and the corresponding ( log f S ), bar, value);
2

*** E—binding energy, F—bull width at half maximum of peak height, see text to the normalization of the peaks. In calculating at %, the
total of all species of the element was assumed to be equal to 100%.

rable with the AES data during 1-h etching for a layer these could actually be polysulfide ions. In this context
at a depth of ~100–200 nm (Table 4, experiments 9, 10). it is pertinent to mention experiment 14, in which a high
Etching for 0.5 h is less efficient and does not always concentration of the polysulfide ion Sn (n > 2) also
reveal a decrease in the concentrations of oxide sulfur remained after etching. It could have been explained by
species. Note that, according to the AES data, we the effect of As on the bond energy of S in the surface,
describe the spectrum of reduced sulfur with two 2p3/2–1/2 but, according to [2], the 2p3/2 peak in arsenopyrite is
doublets, which are provisionally ascribed to mono- shifted for approximately 0.2 eV toward lower (but not
and disulfide sulfur (polysulfide in some cases). This higher) energy values compared to the analogous S peak
approach seems to be somewhat artificial but is war- in the disulfide ion of pyrite, a phenomenon reportedly
ranted by the fact that the spectra (if described by a sin- explained by the slightly lower electron density at the
gle doublet with a 2p3/2 leading line in the vicinity of S atom (because of the differences in the bond lengths).
162 eV) occur to be shifted toward either S2– (~161) or Concerning the proportion of oxidized sulfur species, it
2–
S 2 (~163 eV). Because of this, the only way out was is pertinent to mention that it is fairly difficult to distin-
guish between the sulfite and thiosulfate and between
to assume that the surface always contained two sulfur the sulfate and thiosulfate ions based on the binding
species in variable proportions, which was also con- 2–
firmed by AES data (Table 3). However, their quantita- energy in the presence of S 2 and S2–, and this can be
tive proportions should be evaluated very cautiously, done only with regard for data on the state of Fe. We
although such data are presented in the last column of usually admitted the presence of sulfate when a Fe 2p3/2
Table 4. In principle, etching should decrease the S/S2 peak appeared near 713 eV (it corresponded to trivalent
ratio (as in Table 4, we do not specify here the charges Fe) (Table 2): this provides evidence for the maximum
of the species). This is actually the case (experiments degree of oxidation, which logically should correspond
10, 11, and 12), although the opposite examples were to the maximum oxidation of the S species, i.e., the
also encountered (experiments 15, 16). In some situa- occurrence of sulfate. It usually corresponds to the peak
tions, particularly in the presence of As, a peak appears S 2p3/2 at 168 eV, which is slightly lower than in Table 1.
whose binding energy is close to that of polysulfides It should be borne in mind that these data pertain to the
(Table 1). If S0 is present at the surface, this can be “bulk” material, and binding energy values are lower in
explained by the closeness of the corresponding peak to the surface (for example, by 0.5−0.7 eV for the disul-
164 eV (experiments 9, 13). However, in its absence, fide ion of pyrite; Table 1). We believe that the peak at

GEOCHEMISTRY INTERNATIONAL Vol. 46 No. 6 2008


SURFACE TYPOCHEMISTRY OF HYDROTHERMAL PYRITE 571

Table 5. Fe fraction and the total fraction of oxidized S species on the surface of pyrite crystals synthesized in various systems
Py in multiphase Py in the system Py in the “pure” “Pure” Py in association
Element
systems with As system with S0
Fraction of Fe(np) 0.42–0.68 0.3–0.49 0–0.16 0–0.18
Fraction of oxidized 0.1–0.21 0.06–0.46 0.03–0.18 0.13–0.32
S species

166 eV is related to the sulfite ion and explain the inter- to pyrrhotite but with broader compositional variations
mediate cases by the presence of thiosulfate. Recall that toward FeS2. According to the data of Table 5, the frac-
S is contained in the thiosulfate ion in two chemical tion of Fe in the nonautonomous phase in the Fe–Zn–S
states: with formal valences of +6 and –2. Because of system is 42–68%, i.e., is almost as high as in the sys-
this, if the presence of thiosulfate is admitted, its contri- tems with As (30–49%), although no individual As
bution to the intensity of the peak of S2– should be taken phases appear in them; finally, this fraction in the
into account proceeding from the approximately equal “pure” systems is insignificant (0–18%). It shows no
intensities of peaks for both valence modes of S in thio- pervasive correlation with the fraction of oxidized S
sulfate [24] (see also Table 6, experiment 9). species in the surface layer (Table 5) and with possible
The interpretation of Fe peaks is a fairly compli- crystallization in acidic (pH 0.9–1.2) or weakly alka-
cated task, and here we adhere in this context to the ide- line (pH 7.2–8.2) solutions. Neither shows it an unam-
ology described in [20], whose authors approximated biguous relation to the S/S2 ratio, although (at equal
the Fe 2p3/2 spectrum by a small number of explicit other experimental conditions) both of these values
peaks. Another approach was employed in [1], in which markedly decrease with increasing S activity in the sys-
the high-energy “tail” of the 2p3/2 peak was described tem, as can be clearly seen in Fig. 1 and Table 6. Com-
by a series of virtual, i.e., not resolved, multiplet peaks paring Figs. 1a with 1c and 1b with 1d, one can see how
related to various spin states of the Fe2+ and Fe3+ ions. much are the differences between the S 2p doublets and
The major peak at 707 eV (~70% of the whole signal) the corresponding Fe 2p3/2 spectra in our experiments
was attributed to Fe2+ of “bulk” (i.e., completely coor- with different S activity. The differences in fs2 are about
dinated, as in the crystal volume) sites of Fe atoms in a one order of magnitude at the complete variation for the
low-spin state in pyrite. Our data indicate that the inter- pyrite stability field at this temperature encompassing
pretation [1] should be invalidated for naturally occur- almost six orders of magnitude. However, the contribu-
ring surfaces of pyrite crystals (but not vacuum frac- tions of monosulfide S and Fe of the nonautonomous
tures), because all data indicate that the multiplets have phase (which is denoted FeII-S(np) in Table 4 to
low intensities, whereas the additional peaks appearing emphasize its relation to the “pyrrhotite” mode of Fe
in our samples (particularly at ~710 eV) have intensi- occurrence) to the corresponding spectra are much
ties either comparable with or (often) even strongly higher at a low S fugacity, when pyrrhotite appears in
exceeding that of the peak at 707 eV. Note that an anal- the system. This is fully consistent with our idea that
ogous peak at 709.6 eV was described in [20] as peak 2 the nonautonomous phase on the pyrite surface resem-
for the fresh surface of pyrite and was interpreted fairly bles pyrrhotite. Conceivably, the composition of the
uncertainly: as corresponding to some surface defects nonautonomous phase at a high S activity enriches in
of unknown structure. However, its intensity accounted sulfoxide anions (sulfite and thiosulfate ions) and,
for as little as 19% of the major peak at 707 eV, whereas sometimes, in polysulfide ions (Tables 4 and 6). Etch-
this peak often prevails in our spectra (experiments 7, 8, ing materials indicate that the S/S2 ratios may thereby
14). Moreover, it often appears in the absence of any remain high enough (experiments 9, 13). Recall how-
traces of oxide species and oxygen on the surface (par-
ticularly during ion milling), i.e., obviously is not
related to FeO (Table 2). Because the fraction of Fe cor- Table 6. Fraction of nonautonomous Fe phase and the mono-
responding to this peak is at a minimum in pyrite in the sulfide/disulfide ion ratio on the surface depending on S fugac-
“pure” system and at a maximum in the multiphase sys- ity in the experiments at 400°C in 10% NH4Cl solution
tem (Table 5), it is reasonable to think that the peak at
~710 eV belongs to Fe2+ of a nonautonomous surface Exp. no. Fe(np)/ ∑ Fe S/S2 – log f S2 , bar
phase. The possibility of the origin of nonautonomous
phases and their composition were discussed in [3] with 7 0.68 1.31 8.0
reference to natural pyrite. In fact, the occurrence of a 8 0.42 0.55 6.9
nonstoichiometric layer on the surface of pyrite crystals
9 0.18 0.02* 2.4
with two modes of S occurrence (mono- and disulfide)
definitely points to the existence of a certain nonauton- * the monosulfide concentration was determined with the deduc-
omous phase of variable composition, which is similar tion of S2– from the composition of the thiosulfate ion.

GEOCHEMISTRY INTERNATIONAL Vol. 46 No. 6 2008


572 TAUSON et al.

Intensity (normalized)
1.1
1.0 experiment (b)
0.9 sum of components (‡)
0.8 deconvolution
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
1.1
1.0 (c) (d)
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
175 170 165 160 155 720 715 710 705 700
Binding energy, eV

Fig. 1. XPS S 2p and Fe 2p3/2 spectra for pyrite crystals in experiments (a, b) 8 and (c, d) 7 (Table 4). As the sulfur fugacity decreases
and pyrrhotite appears, the S/S2 ratio increases and the FeII-S(np) peak at ~710 eV becomes predominant in the Fe 2p3/2 spectrum.

ever that the procedure used to obtain these values was SPM
not faultless, and this part of the problem calls for fur- Scanning probe microscopy was applied to visualize
ther investigation on high-resolution XPS-SR equip- the nonautonomous phases and examine their morphol-
ment. Another possible reason for the absence of a ogy and size characteristics. The surface of the exam-
clearly pronounced dependence between the Fe con- ined crystals was almost not treated: we only carefully
centration of the nonautonomous phase and the S/S2 purified it on a paper filter soaked in ethanol. We exam-
ratio is the higher surface sensitivity of the Fe 2p spec- ined three types of samples: “pure” pyrite (Table 4,
trum compared to that of S 2p [2]. In the former case, experiment 11), pyrite from the multiphase system
photoelectrons come from an almost two times lower whose surface was almost completely covered with a
number of surface layers. The appearance of a rela- nonautonomous phase (Figs. 1c, 1d, experiment 7), and
tively low concentration of a nonautonomous Fe phase pyrite synthesized in the system with a single solid
phase but in the presence of As (experiment 14). We
in experiment 13 as a result of etching was associated obtained five scans for each of these samples various
with the nearly complete removal of oxidized S species magnifications. Figure 4 shows three-dimensional
(SO3 and SO4) from the surface. In another “pure” sys- AFM images in the contact mode, which display amaz-
tem without excess S (experiment 11), etching resulted ingly significant differences between the morphologies of
in the appearance of Fe3+, which was likely accompa- the surfaces. While the surface of “pure” pyrite contains
nied with the polymerization of S and an increase in the only single submicrometer-sized inclusions (Fig. 4a), a
sulfite ion concentration (Fig. 2). Finally, the results of well-developed nanometer-sized structure of the non-
experiment 12 demonstrate that the concentration of autonomous phase was observed in the Fe–Zn–S sys-
FeII–S(np) in the presence of As can be high even in the tem (Fig. 4b), and an analogous structure, but at a
absolute absence of sulfoxide anions, and etching did smaller scale, is also pronounced in the system with As
(Fig. 4c; note that the scale of the scan is 2.5 times
not decrease the concentration of this mode of Fe lower than that in Fig. 4b). Moreover, in the system
occurrence and resulted (as in the previous case) only with As, the surface included areas not occupied by the
in S polymerization and the appearance of the sulfite nonautonomous phase. The boundary of one of these
ion (Table 4, Fig. 3). Hence, etching of the pure surface domains is shown in Fig. 5 in tunnel mode. The images
can bring about the formation of trivalent Fe, which were analyzed for a number of parameters, and the
does not change its valence in the presence of As and is most indicative of these were the maximum height of
accommodated in the nonautonomous phase. the “topographic” features and the roughness of the sur-

GEOCHEMISTRY INTERNATIONAL Vol. 46 No. 6 2008


SURFACE TYPOCHEMISTRY OF HYDROTHERMAL PYRITE 573

Intensity (normalized)
1.1 (b)
1.0 experiment
(‡)
0.9 sum of components
0.8 deconvolution
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
1.1
1.0 (c) (d)
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
175 170 165 160 155 720 715 710 705 700
Binding energy, eV

Fig. 2. XPS S 2p and Fe 2p3/2 spectra for pyrite crystals in experiment 11: (a, b) original crystals, (c, d) after ion milling with Ar+
for 0.5 h. It is worth noting the absence of the FeII-S(np) peak at ~710 eV for this “pure” system. Ion milling resulted in the appear-
2– 2–
ance of a wide FeIII peak, an increase in SO 3 , and the appearance of S n .

Intensity (normalized)
1.1 (b)
1.0 experiment (‡)
0.9 sum of components
0.8 deconvolution
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
1.1
1.0 (c) (d)
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
175 170 165 160 155 720 715 710 705 700
Binding energy, eV

Fig. 3. XPS S 2p and Fe 2p3/2 spectra for pyrite crystals in experiment 12. The peak FeII-S(np) at ~710 eV appears in the presence
of As in the system and the absence of sulfoxide anions. The ion milling of the surface induced the appearance of the sulfite ion and
polymerization of sulfur.

GEOCHEMISTRY INTERNATIONAL Vol. 46 No. 6 2008


574 TAUSON et al.

(a) (‡)

Z, µm
1.031
0
3.838
0

Y, µ µm
m X,
3.838 0

(b)
Z, nm
456.4 (b) Z, nm
0 95.52
5.098 0
0 0

Y, µ µm
m X,

m
X, µ
5.098 0

0
Y, µm 2.559
2.559
(c)
Fig. 5. (a) Two- and (b) three-dimensional scanning tunnel-
Z, nm ing microscopic images of an area of the nonautonomous
phase on the surface of a crystal from experiment 14.
195.4
0
1.912 DISCUSSION
0
The two major mechanisms commonly discussed in
Y, µm various variants for the relaxation and reconstruction of
µm X, the surface of pyrite in application to the ultrapure sur-
1.912 0 face of a vacuum fracture are as follows: an auto-redox
reaction and the polymerization of the ligand. Both of
these variants have to explain the experimentally estab-
Fig. 4. Three-dimensional atomic force microscopic images lished fact of the presence of three modes of S occur-
of the surface of pyrite crystals: (a) experiment 11 (Table 4), 2–
“pure” pyrite; (b) experiment 7, pyrite from the multiphase rence on the pyrite surface: the S 2 , disulfide ion, the
system; (c) experiment 14, pyrite with an As admixture. The 2–
scans in Figs. 4b and 4c show the nanometer-sized granular S2– monosulfides ion, and the S n (n > 2) polysulfide
structure of the nonautonomous phase with various grain
sizes. ion. Their proportions in the vacuum fracture of a pyrite
crystal are 85 : 15 : 5 [25]. It is thought that the latter
two are produced due to the instability of the disulfide
face, which was applied in the format of the Mean Area ion because of a mechanical effect (fracturing and
Roughness over the scan field. The latter was 14–17 nm crushing [26], etc.), which resulted in the breaking of
for pure pyrite and ranged from 14 to 50 nm for the sys- the S–S bond and the origin of the S– ion-radical. The
tem with As and from 25 to 65 nm for pyrite in the mul- further scenarios may be different. One of the transfor-
tiphase system. The maximum height of the features on mation means does not affect Fe and implies the com-
the surface with nonautonomous phases ranged from bination of two radicals S– + S– = S2– + S0 [1, 27] or their
2– 2–
~100 to ~500 nm, which is consistent with the afore- interaction with the disulfide ion 2S– + S 2 = S2– + S 3
mentioned AES data on the thickness of the nonstoichi- [2]. The formation of elementary S should likely be
ometric layer that consisted of the nonautonomous viewed as the limiting case of polymerization. Another
phase. scenario assumes the oxidation of bivalent Fe: S– + Fe2+ =

GEOCHEMISTRY INTERNATIONAL Vol. 46 No. 6 2008


SURFACE TYPOCHEMISTRY OF HYDROTHERMAL PYRITE 575

S2– + Fe3+ [2]. The second mechanism implies the insta- minor elements, which, in turn, affect its structure. This
bility of the disulfide ion itself, which is “dispropor- follows from the appearance of a nonautonomous phase
2– 2– in the presence of As, although this element does not
tionated” according to the reaction 2S 2 = S 3 + S2– produce any individual phases. The polymerization of S
[25]. It was demonstrated in [2] that the surface of is pronounced more clearly at the surface, according to
pyrite crystals is stabilized by redox reactions, and the aforementioned data [2]. These data confirm that
polymerization on this surface is at a minimum or is there may be another mechanism forming a nonauton-
absent, whereas the surfaces of other Fe dichalco- omous phase: a polymorphic transition in the surface
genides (marcasite, arsenopyrite, and loellingite) of the layer [30]. During its crystallization, pyrite may con-
orthorhombic crystal system is stabilized according to tain a thin surface film of a nonautonomous phase (mar-
both mechanisms. Polymerization is at a maximum for casite) that is unstable in volume, and this leads to the
loellingite, whose surface does not contain either enrichment of the crystal surface in As and, for exam-
monomers or dimmers. The proportions of these mech- ple, Au because of the structural similarity between
anisms are controlled by the “bulk” crystal structure, marcasite and arsenopyrite.
the distributions of atom spacings, and the electronega- 2–
tivities of the atoms [2]. Various mechanisms were pro- An increase in the SO 3 concentration during etch-
posed to account for the oxidation of the surface of ing (Table 4, experiments 11, 12) demonstrates that the
pyrite [19, 20, 25, 28]. Here we discuss them only cur- layer underlying the nonautonomous phase may con-
sorily, to the extent necessary for understanding the tain oxidized S species. The main oxisulfide compo-
chemical features of crystal surfaces. nents of this phase are the sulfite and thiosulfite ions at
a subordinate amount of sulfate. The latter may, how-
The data of our research do not contradict currently ever, “overlap” the nonautonomous phase as a result of
adopted concepts that the natural surface of pyrite, as later oxidation processes. Thus, the fine structure of the
well as its vacuum fractures, contain more than one surface layer can be more complicated. We believe that
mode of sulfide S occurrence: disulfide, monosulfides, the reason for this is the instability of coexisting sulfate
and, more rarely, polysulfide (Table 4). This is, how- and disulfide ions, which results (in the presence of
ever, the only analogy between them. It could be oxygen) in the appearance of the sulfite (thiosulfate)
expected that the bombardment of the surface with Ar+ and sulfide ions in the nonautonomous phase according
should have produced Fe3+ and S polymers according to to the formal reaction
the mechanisms discussed above, with the participation
of S–. This was the case with experiments 11 and 12 FeSO4 + FeS2 + 1/2O2 = [FeSO3 + FeS](np) + SO2.
(Table 4), but generally no unambiguous relations can The formation of sulfide was documented during Cd
be traced. According to our data, the natural surface of 2–
hydrothermal pyrite is chemically modified, compared adsorption on pyrite as a result of S 2 disproportion-
to the volume of the crystal, into a surface (up to ation [31]. Sulfite– and thiosulfate–sulfide complexes
~300 nm thick) pyrrhotite-like nonautonomous phase. with “thiosulfate” bonds like O3S–S may be regarded as
Its principally important distinctive feature is the occur- thermodynamic compromises needed for the coexist-
rence of a peak at ~710 eV in the Fe 2p3/2 spectrum, ence of oxidized and reduced S species. It is interesting
with this peak often having a height greater than that of that such instability of coexisting polysulfide and sul-
the peak of bivalent low-spin Fe(II) in the pyrite struc- fate ions is also manifested in lazurite, an S-bearing
ture (707 eV). As was shown by the spectroscopic aluminosilicate mineral, during its heating in air, which
results and those of scanning probe microscopy, the results in the origin of the sulfite ion [32]. Comparison
nonautonomous phase is highly sensitive to the purity with data on pyrrhotite and galena [7, 21, 30] shows
of the system. Since this phase is transitional (in the that sulfide crystals growing in hydrothermal environ-
chemical sense) between pyrite and pyrrhotite, it is sen- ments have surface layers up to ~500 nm thick, whose
sitive to the S activity: the closer to the pyrrhotite field, chemical composition, stoichiometry, and possibly,
the greater its contribution to the surface modes of Fe also structure differ from those in the crystal volume.
and S. This means that, according to the classification This is the scale of the surface heterogeneity at which
[29], this phase can be attributed to a nonautonomous the typochemistry of mineral surfaces can manifest
“prephase,” i.e., a surface phase object formed during itself. This range of sizes is inaccessible for electron-
the crystallization under conditions close to the stabil- microprobe or ion-microprobe analytical techniques
ity field of the neighboring “bulk” phase. In this situa- and can be efficiently examined by methods of surface
tion, even a small excess of the chemical potential of spectroscopy, scanning probe microscopy [30], and
the component (because of, for example, a temperature high-resolution scanning electron microscopy.
gradient) can significantly affect the composition of the
surface layer. The “basic” structure of the nonautono-
mous phase may contain no oxidized modes of S spe- CONCLUSIONS
cies and be a layer of variable composition Fe2+[S, S2, Models for the reconstruction of fresh pyrite surface
Sn]2– with the S/S2 ratio varying from ~0.5 to ~2.0, produced by vacuum fracture are only partly applicable
averaging at ~1.1 (Table 4). This layer may include to the situation of the real surface of hydrothermally

GEOCHEMISTRY INTERNATIONAL Vol. 46 No. 6 2008


576 TAUSON et al.
2–
synthesized pyrite. Although the same species S 2 , S2−, ACKNOWLEDGMENTS
and
2–
Sn (n > 2) were detected under different condi- The authors thank Yu.V. Shchegol’kov for assis-
tance in the spectroscopic research.
tions and in proportions determined by these condi-
tions, the appearance of these species is controlled not This study was supported by the Russian Founda-
2– tion for Basic Research (project no. 06-05-64171) and
by auto-redox reactions or S 2 , disproportionation but the Siberian Branch of the Russian Academy of Sciences
by the development of a pyrrhotite-like nonautonomous (integration project 96). The equipment for the AFM and
phase in the surface layer of the crystal (~300 nm STM measurements was provided for us by Rosnauka
thick). This phase contains variable proportions of FeII, in 2004 within the scope of a program for the support
2–
S2–, S 2 , and (more rarely) the polysulfide ion. The oxi- of equipment availability for research organizations.
dized S species of the phase are dominated by the
sulfite and thiosulfate ions. The chemistry of the nonau- REFERENCES
tonomous phase is likely determined by the instability 1. H. W. Nesbitt, M. Scaini, H. Hochst, et al., “Synchrotron
of association of the sulfate and polysulfide (disulfide) XPS Evidence for Fe2+–S and Fe3+–S Surface Species
ions, for which a more suitable “thermodynamic com- on Pyrite Fracture–Surface, and Their 3D Electronic
promise” is a sulfide–sulfite (thiosulfate) ion associa- States,” Am. Mineral. 85, 850–857 (2000).
tion. The typochemistry of the pyrite surface is based 2. S. L. Harmer and H. W. Nesbitt, “Stabilization of Pyrite
on the ability of the nonautonomous phase to “record” (FeS2), Marcasite (FeS2), Arsenopyrite (FeAsS) and
the growth conditions of the crystals in terms of two Loellingite (FeAs2) Surfaces by Polymerization and
major factors: the purity of the system (the occurrence Auto-Redox Reactions,” Surf. Sci. 564, 38–52 (2004).
of additional phases, including virtual ones, i.e., poten- 3. V. L. Tauson, R. G. Kravtsova, and V. I. Grebenshchik-
tially possible phase of minor elements) and S fugacity. ova, " Chemical Typomorphism of the Surface of Pyrite
SPM data indicate that the surface of the crystals at a Crystals of Gold Ore Deposits," Dokl. Akad. Nauk 399
high S fugacity contains almost no nonautonomous (5), 673–677 (2004) [Dokl. Earth Sci. 399A, 1291–1295
phases and is characterized by low roughness (14–17 nm). (2004)].
At a low S fugacity in equilibrium with pyrrhotite and 4. V. L. Tauson and R. G. Kravtsova, “Typochemistry of
Fe-rich sphalerite, the average roughness of the surface Mineral Surface: Features of Surface Composition (Evi-
increases to 25–65 nm, and the maximum height of the dence from Gold-Bearing Pyrite from an Epithermal
Deposit),” Geol. Geofiz. 45 (2), 222–227 (2004).
“topographic” features reaches ~100–500 nm. This is
consistent with the data of Auger spectroscopy with the 5. S. D. Sher and A. V. Demchenko, “On the Significance
of Studying the Morphology of Pyrite Metacrysts in
ion milling of the surface by Ar+ ions on the average Searching for Gold Deposits in the Lena Region,” Geol.
thickness of the surface layer of approximately 300 nm. Rudn. Mestorozhd., No. 4, 84–96 (1962).
The most typical feature of the nonautonomous phase 6. T. T. Lyakhovich and I. V. Vedyaeva, “Typochemistry of
on pyrite is its XPS peak of FeII 2p2/3 at ~710 eV, with Pyrites—Criteria for Searching for and Identification of
the intensity of this peak often exceeding that of the Ore Deposits,” Razved. Okhr. Nedr, No. 8, 31–34
peak of low-spin FeII in the pyrite structure (707 eV). (2002).
The nonautonomous phase on the pyrite crystal surface 7. V. L. Tauson, I. Yu. Parkhomenko, D. N. Babkin, et al.,
can be regarded as a precursor phase according to the “Cadmium and Mercury Uptake by Galena Crystals
classification [29]; this phase can be formed in the pres- under Hydrothermal Growth: A Spectroscopic and Ele-
ence of As by a process similar to a polymorphic tran- ment Thermo-Release Atomic Absorption Study,” Eur. J.
sition (pyrite–marcasite). In this case it shows charac- Mineral. 17, 599–610 (2005).
teristics typical of the surface of orthorhombic Fe 8. V. L. Tauson and V. V. Akimov, “Effect of Crystallite
dichalcogenides: the formation of the polysulfite ion by Size on Solid State Miscibility: Applications to the
the reaction of S polymerization [2]. The geochemical Pyrite–Cattierite System,” Geochim. Cosmochim. Acta
55 (10), 2851–2859 (1991).
role of the surface nonautonomous phase on pyrite may
also be significant, particularly during the uptake of 9. V. L. Tauson and R. G. Kravtsova, “Estimates of Gold
Admixture in the Structure of Pyrite from Epithermal
minor elements that are incompatible with the pyrite Gold–Silver Deposits, Northeastern Russia,” Zap. Vse-
structure but can be readily accommodated in the less ross. Mineral. O-va 131 (4). 1–11 (2002).
rigid structure of the nonautonomous phase. Because of 10. Practical Surface Analysis: Auger and X-Ray Photoelec-
the abundance of S species and the absence of hard con- tron Spectroscopy, Ed. by D. Briggs and M. P. Seah
straints in terms of space and bond length (because of (Wiley, Chichester, 1990; Mir, Moscow, 1987) [in Rus-
the variations in the sulfide/disulfide ratio), many of sian].
them can be incorporated into the surface layer even 11. A. G. Schaufuss, H. W. Nesbitt, I. Kartio, et al., “Reac-
being absolutely incompatible with the crystal matrix. tivity of Surface Chemical States on Fractured Pyrite,”
This important geochemical conclusion drawn from Surf. Sci. 411, 321–328 (1998).
our results will be considered in detail in our oncoming 12. M. Descostes, F. Mercier, N. Thromat, et al., “Use of
paper and illustrated by the example of pyrite from var- XPS in the Determination of Chemical Environment and
ious gold deposits. Oxidation State of Iron and Sulfur Samples: Constitution

GEOCHEMISTRY INTERNATIONAL Vol. 46 No. 6 2008


SURFACE TYPOCHEMISTRY OF HYDROTHERMAL PYRITE 577

of a Data Basis in Binding Energies for Fe and S Refer- 22. Yu. L. Mikhlin, A. V. Kuklinskiy, N. I. Pavlenko, et al.,
ence Compounds and Applications to the Evidence of “Spectroscopic and XRD Studies of the Air Degradation
Surface Species of an Oxidized Pyrite in a Carbonate of Acid-Reacted Pyrrhotites,” Geochim. Cosmochim.
Medium,” Appl. Surf. Sci. 165, 288–302 (2000). Acta 66 (23), 4057–4067 (2002).
13. J. R. Mycroft, G. M. Bancroft, N. S. McIntyre, et al., 23. Yu. L. Mikhlin, Ye. V. Tomashevich, G. L. Pashkov,
“Detection of Sulphur and Polysulphides on Electro- et al., “Electronic Structure of Non-Equilibrium Iron-
chemically Oxidized Pyrite Surfaces by X-Ray Photo- Deficient Layer at Hexagonal Pyrrhotite,” Appl. Surf.
electron Spectroscopy and Raman Spectroscopy,” J. Sci. 125, 73–84 (1998).
Electroanal. Chem. 292, 139–152 (1990). 24. A. S. Manocha and R. L. Park, “Flotation Related ESCA
14. A. M. Widler and T. M. Seward, “The Adsorption of Studies on PbS Surfaces,” Appl. Surf. Sci. 1, 129–141
Gold (I) Hydrosulphide Complexes by Iron Sulfide Sur- (1977).
faces,” Geochim. Cosmochim. Acta 66 (3), 383–402 25. H. W. Nesbitt and I. J. Muir, “X-Ray Photoelectron
(2002). Spectroscopic Study of a Pristine Pyrite Surface Reacted
15. A. R. Pratt, I. J. Muir, and H. W. Nesbitt, “X-Ray Photo- with Water Vapour and Air,” Geochim. Cosmochim.
electron and Auger Electron Spectroscopic Studies of Acta 58 (21), 4667–4679 (1994).
Pyrrhotite and Mechanism of Air Oxidation,” Geochim.
Cosmochim. Acta 58 (2), 827–841 (1994). 26. K. Sasaki, “Effect of Grinding on the Rate of Oxidation
of Pyrite Oxygen in Acid Solutions,” Geochim. Cosmo-
16. J. E. Thomas, W. M. Skinner, and R. St. C. Smart, “A chim. Acta 58 (21), 4649–4655 (1994).
Comparison of the Dissolution Behavior of Troilite with
Other Iron (II) Sulfides: Implications of Structure,” 27. H. W. Nesbitt, G. M. Bancroft, A. R. Pratt, and M. J. Scaini,
Geochim. Cosmochim. Acta 67 (5), 831–843 (2003). “Sulfur and Iron Surface States on Fractured Pyrite Sur-
face,” Am. Mineral. 83, 1067–1076 (1998).
17. J. E. Maulder, W. F. Stickle, P. E. Sobol, and K. D. Bom-
ben, “Handbook of X-Ray Spectroscopy. (A Reference 28. S. Karthe, R. Szargan, and E. Suoninen, “Oxidation of
Book of Standard Spectra for Identification and Interpre- Pyrite Surfaces: A Photoelectron Spectroscopic Study,”
tation of XPS Data),” (Perkin-Elmer Corp., Norwalk, Appl. Surf. Sci. 72, 157–170 (1993).
1990). 29. V. L. Tauson, “Systematics of Processes of Trace Ele-
18. C. D. Wagner, A. V. Naumkin, A. Kraut-Vass, et al., NIST ment Uptake by Real Mineral Crystals,” Geokhimiya,
X-ray Photoelectron Database www.nist.gov/srd/ No. 2, 213–219 (2005) [Geochem. Int. 43, 184–190
online.htm. (2005)].
19. A. G. Schaufuss, H. W. Nesbitt, I. Kartio, et al., “Incipi- 30. V. L. Tauson, B. A. Loginov, V. V. Akimov, and
ent Oxidation of Fractured Pyrite Surfaces in Air,” J. S. V. Lipko, “Nonautonomous Phases as Potential
Electron. Spec. Relat. Phenom 96, 69–82 (1998). Sources of Incompatible Elements,” Dokl. Akad. Nauk
20. C. M. Eggleston, J.-J. Ehrhardt, and W. Stumm, “Surface 406 (6), 806–809 (2006) [Dokl. Earth Sci. 407, 280–283
Structural Controls on Pyrite Oxidation Kinetics: An (2006)].
XPS-UPS, STM, and Modeling Study,” Am. Mineral. 31. B. C. Bostick, S. Fendorf, and M. Fendorf, “Disulfide
81, 1036–1056 (1996). Disproportionation and CdS Formation upon Cadmium
21. V. L. Tauson and N. V. Smagunov, “Composition of the Sorption on FeS2,” Geochim. Cosmochim. Acta 64, 247–
Surface of Pyrrhotite (Fe1-xS) Crystals Synthesized in 255 (2000).
Association with Greenockite (α-(Cd,Fe)S) under 32. V. L. Tauson and A. N. Sapozhnikov, “Stability of the
Hydrothermal Conditions: Introduction into the Modulated Structure of Baikal Lazurite and Its Recrys-
Geochemistry of Nonautonomous Phases,” Geokhimiya, tallization at a Temperature of 600°C over a Wide Range
No. 4, 448–454 (2004) [Geochem. Int. 42, 377–382 of Sulfur Dioxide Fugacities,” Crystallogr. Rept. 50
(2004)]. (Suppl. 1), 1–9 (2005).

GEOCHEMISTRY INTERNATIONAL Vol. 46 No. 6 2008

You might also like