Phys Chem Minerals (2006) 33:553–557 DOI 10.1007/s00269-006-0102-1
ORIGINAL PAPER
Natural iron-containing blue and colorless euclase studied by electron paramagnetic resonance K. J. Guedes Æ K. Krambrock Æ M. V. B. Pinheiro Æ L. A. D. Menezes Filho
Received: 11 May 2006 / Accepted: 7 July 2006 / Published online: 8 August 2006 Springer-Verlag 2006
Abstract Natural blue and colorless rare-gem mineral specimens of euclase from Brazil are investigated by electron paramagnetic resonance (EPR). Angular dependences of Fe3+ EPR spectra in three mutually perpendicular crystal planes are analyzed revealing g and D tensors with significant low-symmetry effects, as for example, the high asymmetry parameter E/ D = 0.28. Fourth-order degree Stevens parameters are also included in analysis. The anisotropy of both g and D tensors is consistent with Fe3+ substituting for Al3+ ions in strongly distorted AlO5(OH) octahedra in ˚. which the oxygen distances range from 1.85 to 1.98 A Fe3+ is not responsible for the blue color because colorless and blue euclase show nearly the same Fe3+ concentration as measured by EPR. However, total iron content in blue sample is much higher than in the colorless one suggesting that the existing model that Fe2+–Fe3+ intervalence charge transfer transition may explain the blue color of euclase. Keywords
EPR Æ Fe3+ Æ Euclase Æ Low symmetry
Introduction Euclase is a rare gem mineral with chemical formula BeAlSiO4(OH) and monoclinic space group P21/a (C52h) with four molecular units (Z = 4) and lattice ˚ , b = 14.29 A ˚ and c = 4.618 A ˚, parameters a = 4.763 A with b = 11015¢ (Mrose and Appleman 1962). It belongs to the orthosilicate group. The structure of euclase can be described by zigzag chains of BeO3(OH) and SiO4 tetrahedra along crystallographic a axes interconnected by distorted AlO5(OH) octahedra (Hazen et al. 1986). Intrinsically colorless, it is believed that blue and green colors are related to iron and chromium impurities, respectively (Stocklmayer 1998). In this work we present a detailed analysis of Fe3+ related EPR spectra in single crystals of euclase from the Rio Grande do Norte state in Brazil. EPR rotational patterns are measured for three mutually perpendicular crystal planes. The local site symmetry, the g tensor, the electronic fine structure tensor D and the fourth-order degree Stevens operators are determined and interpreted in terms of the local environment of distorted octahedra symmetry for Fe3+ substituting for Al3+ ions.
Experimental K. J. Guedes Æ K. Krambrock (&) Æ M. V. B. Pinheiro Departamento de Fı´sica, ICEx, Universidade Federal de Minas Gerais, CP 702, 30.123-970 Belo Horizonte, MG, Brazil e-mail:
[email protected] L. A. D. Menezes Filho Luiz Menezes Minerais, Rua Esmeralda 534, Prado, 30.410-080 Belo Horizonte, MG, Brazil
Natural blue and colorless euclase specimens from the Alto Equador mine in the state of Rio Grande do Norte, Brazil, were chosen for our investigation and were characterized for their impurity contents with electron microprobe analysis (EMA) using an JEOL JXA8900R spectrometer. The EMA results in Table 1 are presented in weight percent (wt%) for
123
554
Phys Chem Minerals (2006) 33:553–557
Table 1 Electron microprobe analysis of blue and colorless euclase specimens (wt%) Sample
Color
Al2O3
SiO2
FeO
Cr2O3
Total
1 2
Colorless Blue
33.368 33.035
42.075 41.451
0.006 0.016
0.008 0.007
75.461 74.488
two representative samples, one colorless and one blue. In our work we used samples with homogeneously distributed blue color. Comparison between the blue and colorless samples shows that the iron content in the blue sample is a factor of about 2.5 higher than in the colorless sample. Both show approximately the same low chromium content. Total iron content was analyzed to be due to FeO. The total value for the composition of euclase is less than 100% because EMA did not permit for the determination of BeO nor of water content. For the EPR measurements samples were cut to pieces of about 3 · 3 · 3 mm3 using the perfect cleavage plane (0 1 0). EPR spectra were recorded with a homemade spectrometer which includes a 500 mW klystron (Varian), a commercial cylindrical resonance cavity (Bruker), an electro-magnet (Varian) with maximum field amplitudes of 800 mT and a He flux cryosystem (Oxford) for low-temperature measurements. For the angular variations, the sample holder was rotated with a goniometer, in steps of 5 with a precision of about 0.2. The microwave frequency was stabilized by an automatic frequency control and measured with a high-precision (six digits) frequency meter (PTS). For g-factor calibration, the diphenyl–picryl–hydrazyl standard in powder form (DPPH ALDRICH) was used (g = 2.0037). Spectra were recorded using 100 kHz field modulation and the lock-in technique (EG and G Princeton).
Experimental results Figure 1 shows the angular variation of EPR spectra in steps of 5 as first-derivates measured at room temperature and microwave frequency of 9.39 GHz, with magnetic field B in the ab plane, and with the crystal mounted with its c* axis perpendicular to B. The c* axis is defined as being perpendicular to the ab plane. The field positions of the EPR lines varied strongly with orientation of the sample and ranged from about 5 mT to about 800 mT. The linewidths of the EPR lines depends strongly on orientation ranging from 1 to 10 mT. Figure 2 shows the peak positions of EPR lines for three mutually perpendicular rotation planes ab, ac* and bc*. The field positions of EPR lines are shown
123
as dots and the calculated angular variations as solid lines. The analysis is described below. The EPR lines belong to Fe3+ ions in the high spin state S = 5/2 with ground-state multiplet (3d5) 6S5/2. No EPR lines due to chromium impurities were detected. The angular dependences in the ab and bc* plane are due to two magnetically inequivalent Fe sites, whereas in the ac* plane, the plane with monoclinic symmetry, only one set of Fe-related spectra is observed. Figure 3 shows the projection of the crystal structure of euclase onto the bc plane, demonstrating that two of the four Al octahedra are equivalent in pairs. In each Al octahedron, the oxygen distance ranges from 1.85 to ˚. 1.98 A The high-field EPR lines vary strongly with orientation of the magnetic field with respect to sample orientation. In addition, in extreme high-field positions, the EPR lines broaden strongly and lose much of their intensities as can be observed in Fig. 1. The angular pattern of Fe3+ is rather complex in euclase, as is typical for the case of Fe3+ ions in intermediate ligand fields [see chapter 7 of Abragam and Bleaney (1986)].
Discussion The EPR angular dependencies are analyzed using the following spin Hamiltonian X m Bm ð1Þ H ¼ bSgB þ SDS þ 4 O4 m¼0;2;4
in monoclinic symmetry. The first term in Eq. 1 represents the electronic Zeeman interaction, the second the electronic fine structure interaction and the third term is due to fourth-order terms in Stevens notation. The parameters of Eq. 1 have their usual meaning [see chapter 3 of Abragam and Bleaney (1986)]. The parameters of the spin Hamiltonian were evaluated by fitting simultaneously all line positions in the three mutually perpendicular crystal planes (Fig. 2) using exact diagonalization of the spin Hamiltonian of Eq. 1 for monoclinic symmetry. In total, 592-line position were taken into account in the analysis for the three mutually perpendicular crystal planes. Also, small misalignments of the exact crystal mounting in the resonance cavity were taken into account. The final fitting parameters of the g tensor and also the electronic fine structure tensor D given in GHz are summarized in Table 1, together with the polar coordinates of their principal axes and errors. The |D| and |E|
Phys Chem Minerals (2006) 33:553–557
180 150
Angle (degree)
Fig. 1 Angular variation of Fe3+-related X-band EPR spectra as first-derivates in the ab plane with the rotation axis along c* axis, measured at room temperature
555
120 90 60 30 0 0
100
200
300
400
500
600
700
Magnetic Field (mT)
(c)
800 600
a c*
400 200 0 0
Magnetic Field (mT)
Fig. 2 Angular variation of the EPR line positions (shown as dots) in the (a) bc*, (b) ab and (c) ac* plane of euclase single crystal measured at room temperature at X-band. Solid lines correspond to the final fitted line positions (see text) considering also small misalignment of sample
30
60
800
90
120
150
b
180
(b)
600 400
a
200 0
0
800
30
60
90
120
150
180
(a)
b
600
c*
400 200 0 0
30
60
90
120
150
180
Angle (degree)
values are 15.8 and 4.4 GHz, respectively. The inclusion of the fourth-degree Stevens parameters decrease only very little the total error in the adjusted line positions. Their values are: B04 = –0.004(1) GHz, B24 = –0.003 (1) GHz and B44 = –0.023(6) GHz. The absolute sign in the electronic fine structure parameters D and E were determined from the temperature dependence of the relative intensities of lowand high-field EPR lines [see chapter 3 of Abragam
and Bleaney (1986)]. The temperature dependence of the EPR lines is shown in Fig. 4. At low temperature the population of the levels at highest energy is reduced according to the Boltzmann distribution and therefore, the intensity of the corresponding lines is decreased in relation to the levels with lowest energy. The energy level corresponding to the ms = –5/2 state is lower than that of the ms = 5/2 state. It can be observed that the intensities of the low-field EPR lines
123
556
Phys Chem Minerals (2006) 33:553–557
Fig. 3 Projection of the euclase structure onto the bc plane, showing the Al octahedral, oxygen atoms (open circles), hydrogen atoms (small black circles), silicon atoms (large gray circles) and beryllium atoms (small gray circles )
a
c
6K
30 16 K
EPR (arb. units)
Fig. 4 Temperature dependence of EPR spectra measured from 6 K up to 300 K for the magnetic field along B || b
b
50 K
20 80 K 100 K
10
150 K 300 K
0
0
200
400
600
Magnetic Field (mT)
increase by lowering the temperature, whereas the intensities of the high-field lines decrease. This means that both D (= 3B02) and E (= B22) have negative signs. In addition, at low temperatures, forbidden transitions of type Dms = ± 2 were observed in the low-field region. The EPR analysis for Fe3+ in euclase is consistent for Fe3+ ions substituting for Al3+ ions. We may argue that Fe3+ has the same valence state as Al3+ and both ions have nearly identical ionic radii for octahedral coordination. Further, for rotation of sample in the ab and bc planes, two magnetically inequivalent sites are observed, whereas in the ac plane only one site. The same observation was obtained from nuclear magnetic resonance spectra of the Al nuclei in euclase (Eades 1955). No correlation between the directions of the principal tensor axis and the site symmetry of the distorted octahedron could be established what is not surprisingly because the local point symmetry of the Al site is Ci. A wide range of literature exists for Fe3+ related EPR spectra in different low-symmetry minerals. In most cases the g tensor is nearly isotropic and approximately that of the free spin value. In addition, most |E/D| ratios are in the range of 0.10–0.20
123
(Abragam and Bleaney 1986). Using the symmetry considerations on the EPR parameters for triclinic point symmetry (McGavin 1987) and diagonalization of the D tensor (Table 2) results in |D| = 3 B02 = 15.8 GHz and |E| = B22 = 4.4 GHz. This leads to a ratio of |E/D| for Fe3+ in euclase of 0.28 which is very large; falling, however, within the range of standardized EPR parameters with ratios of up to 0.33 (Rudowicz 1986). From the temperature dependence of the EPR spectra we found that both E and D have negative sign. Note, that the zero-field splitting is rather large with |D| = 15.8 GHz compared to Fe3+ in monoclinic spodumene with |D| = 4.0 GHz (Manoogian et al. 1965), however, less than in orthorhombic andalusite with |D| and |E| of 42.0 and 4.7 GHz, respectively (Holuj et al. 1966). However, again the |E/D| ratio 0.11 is less here than in euclase. It is still not clear what effects are responsible for the contribution to the large zero-field splitting of S state ions. In addition to the crystal field, spin-spin and spin-orbit interactions, it appears that covalent binding effects play a dominant role. Blue euclase, the highly prized gem form of euclase, presents three absorption bands centered at about 670, 860 and 1,250 nm. The first band has been attributed to
Phys Chem Minerals (2006) 33:553–557
557
Table 2 Components of the g and D tensors of Fe3+ in high spin state S = 5/2 in euclase single crystals for room temperature; units of D tensor in GHz Direction angles
gxx
gyy
gzz
Dxx
Dyy
Dzz
Q F
1.993(5) 54(1) 106(1)
2.070(5) 131(1) 156(1)
2.127(5) 118(1) 38(1)
–0.86(4) 107(1) 294(1)
–9.63(4) 53(1) 10(1)
10.49(4) 42(1) 224(1)
The angles describe the orientation of the principal axes of the g and D tensors in relation to the crystal axes: Q polar angle and F azimuthal angle
intervalence charge transfer transitions (IVCT) of Fe2+ and Fe3+ pairs and the last two to crystal-field-enhanced transitions of Fe2+ (Mattson and Rossman 1987). Our investigation is not in contradiction to that model; however, a definite proof is still missing. EPR measurements at very low temperatures in our samples were not successful to detect Fe2+ which is not surprising. Comparison of the colorless and blue samples by EMA analysis show higher total iron content in blue samples, whereas EPR spectra show nearly the same concentration of Fe3+ ions in both types of euclase samples, leaving a possible higher concentration of Fe2+ ions in blue euclase.
Summary The rare-gem mineral euclase was investigated for Fe3+-related EPR spectra. It was shown that Fe3+ substitutes for Al3+ ions in the euclase structure with strongly distorted octahedral symmetry (actual point symmetry Ci). Rotation patterns of EPR spectra were analyzed using a triclinic spin Hamiltonian, and revealed that the zero-field splitting of Fe3+ ions is rather large with |D| of 15.8 GHz with pronounced low-symmetry effect E/D = 0.28 which is also evident from the low symmetry of the g tensor.
are grateful to W.T. Soares of the Physics Department of the Universidade Federal de Minas Gerais for the electron microprobe analysis.
References Abragam A, Bleaney B (1986) Electron Paramagnetic Resonance of Transition Ions, Dover Eades RG (1955) An investigation of the nuclear resonance absorption spectrum of Al27 in a single crystal of euclase. Can J Phys 33:286–297 Hazen RM, Au AY, Finger LW (1986) High-pressure crystal chemistry of beryl (Be3Al2Si6O18) and euclase (BeAlSiO4OH). Am Mineral 71:977–984 Holuj F, Thyler JR, Hedgecock NE (1966) ESR spectra of Fe+3 in single crystals of andalusite. Can J Phys 44:509–523 Manoogian A, Holuj F, Carswell JW (1965) The electron spin resonance of Fe+3 in single crystals of spodumene. Can J Phys 43:2262–2275 Mattson SM, Rossman GR (1987) Identifying characteristics of charge transfer transitions in minerals. Phys Chem Miner 14:94 McGavin DG (1987) Symmetry constrains on EPR spin-Hamiltonian parameters. J Magn Reson 74:19–55 Mrose ME, Appleman DE (1962) The crystal structures and crystal chemistry of va¨yrynenite (Mn,Fe)Be(PO4)(OH), and euclase, AlBe(SiO4)(OH) Zeitschrift fuer Kristallographie 117:16–36 Rudowicz C (1986) On standardization and algebraic symmetry of the ligand field Hamiltonian for rare earth ions at monoclinic sites. J Chem Phys 86:5045–5058 Stocklmayer S (1998) Blue euclase from Zimbabwe—a review. J Gemmol 26:209–218
Acknowledgments We acknowledge financial support from the Brazilian agencies FINEP, FAPEMIG, CNPq and CAPES. We
123
