American
Mineralogist,
Volume 80, pages 491-501,
1995
Influence of chemistry on the pyroelectric effect in tourmaline
KATE D. HAWKINS.
Centre for Microscopy and Microanalysis, University of Queensland, Brisbane, Queensland 4072, Australia;
and School of Mechanical and Manufacturing Engineering, Queensland University of Technology,
P.O. Box 2434, Brisbane, Queensland 4001, Australia
IAN D. R. MACKINNON
Centre for Microscopy and Microanalysis,
Institut fUr Kristallographie
University
of Queensland,
HELMUT
SCHNEEBERGER..
und Mineralogie, Ludwig-Maximilians-Universitiit,
Brisbane, Queensland
4072, Australia
TheresienstraJ3e 41, Munich, Germany
ABSTRACT
Pyroelectric coefficients were measured from a series of natural tourmaline crystals between -170 and - 500 K to quantify the variation of the pyroelectric effect with chemical
composition. The amount of Fe in tourmaline has a prominent influence on the pyroelectricity. Fe content linearly decreases the pyroelectric coefficient in the composition range
between 0.01(1) and 14.6(2) wt% FeO. Thus, to a first approximation, tourmaline pyroelectric coefficients may be predicted directly from the chemical composition derived by
routine electron probe microanalysis.
The relationships between pyroelectricity and chemistry indicate that the pyroelectric
coefficient is influenced to different extents by the occupancies of the X, Y, and Z cation
sites in the tourmaline structure. The octahedral Y site occupancy strongly influences the
pyroelectric coefficient due to the preference of Fe for this site. This work further suggests
that the addition of Fe and Mg cations to the smaller Z octahedral site causes the pyroelectric coefficient to increase. However, because an extended suite of samples is not available in which the Z site contains ions other than AI, this proposed trend has not been
experimentally determined. The chemistry of the ninefold coordinated X site and the
population of this site do not influence the pyroelectric coefficients of tourmaline.
INTRODUCTION
Tourmalines are naturally occurring minerals with the
general formula XY3Z6 (B03)3Si6018 (O,OH,F)4' where
usually X = Na, Ca, or a vacancy, Y = AI, Li, Fe2+, FeH,
and many other cations, and Z = AI, Mg, and FeH. It
has been known for centuries (Dietrich, 1985) that tourmaline crystals are pyroelectric materials that develop an
electrostatic charge when heated or cooled. Careful measurements of the combined primary and secondary pyroelectric effects in these minerals were first reported by
Ackermann (1915). The results of that work showed that
tourmaline exhibits pyroelectricity over a wide temperature range. Tourmaline is stable to high temperatures
and may find application as an infrared detector where
ferroelectric materials such as LaTi03 fail (Hamid, 1980).
Importantly, Ackermann's (1915) measurements
show
that differently colored tourmalines exhibit different degrees of pyroelectricity. For example, black varieties are
the weakest whereas rose colored tourmalines are the most
* Present address: CSIRO Division of Minerals, P.O. Box
124, Port Melbourne, Victoria 3207, Australia.
** Present address: BahnhofstraJ3e lOa, 86551 Aichach, Germany.
0003-004X/95/0506-0491$02.00
strongly pyroelectric. Subsequent investigations support
these observations and show that tourmaline pyroelectric
coefficients (P'3)range between -1.8 and 5.4 ~C/(m2. K)
at 296 K (Hayashi, 1912; Rontgen, 1914; Ackermann,
1915; Gladkii and Zheludev, 1965; Gavrilova, 1965; Fabel and Henish, 1971; Drozhdin et aI., 1975; Gavrilova
et al. 1989; the present work). Related studies have shown
that piezoelectricity, the inducement of an electric charge
by an applied stress, is also stronger in translucent colored tourmalines than in certain opaque black varieties
(Cady, 1946).
None of these previous studies has addressed the variation in the pyroelectric coefficients with either the precise chemical composition or crystal chemistry of differently colored tourmalines. Tourmaline crystal chemistry
is complex because the size and anion coordination of
the X, Y, and Z cation sites are different. As a result
many different cations can be accommodated
in tour~
maline by coupled isomorphous substitution (e.g., Henry
and Guidotti, 1985; Foit et aI., 1989; Burt, 1989). An
indication of this complexity is given by the summary of
natural and synthetic tourmaline end-members as well as
by the references to their structural descriptions in Table 1.
491
-------..-....------
492
HAWKINS
TABLE
1.
Tourmaline
End-member
end-member
x
ET AL.: PYROELECTRIC
EFFECT IN TOURMALINE
compositions
y
z
First description
Refined
structure
Natural end-members
Buergerite
Chromdravite
Dravite
Elbaite
Feruvite
Uddicoatite
Povondraite*
Schorl
Tsilaisite"
Uvite
Na
Na
Na
Na
Ca
Ca
Na
Na
Na
Ca
Fe~+
Mg.
Mg.
(AI,U).
Fe~+
(AI,U).
F~+
F~+
(Mn,AI).
Mg.
AI.
(Cr.AI)
AI.
AI.
(AI,Mg).
AI.
F&.+
AI.
AI.
(AI.Mg)
(Mg,AI)
AI.
Mason et al. (1964)
Rumantseva (1983)
Tschermak (1885)
Vernadsky (1913)
Grice and Robinson (1989)
Dunn et al. (1977)
Walenta and Dunn (1979)
Mathesius (1564)
Schmetzer and Bank (1984)
Dunn (1977)
Barton (1969)
Nuber and Schmetzer (1979)
Buerger et al. (1962)
Donnay and Barton (1972)
n.d.
Nuber and Schmetzer (1981)
Grice et al. (1993)
Fortier and Donnay (1975)
Nuber and Schmetzer (1984)
Schmetzer et al. (1979)
Synthetic end-members
Alkali-defect dravite
"Sodium aluminum"
tourmalinet
Na
Rosenberg and Foit (1979)
Rosenberg
and Foit (1979)
Rosenberg
Rosenberg
et al. (1986)
et al. (1986)
Note: n.d. = not determined.
Povondraite was initially termed "ferridravite" (Walenta and Dunn, 1979) with Y. = Mg. before the formula was altered to Y.
*
= Fe'+
crystal
structure refinement (Grice et aI., 1993).
· following
** Tsilaisite = manganoan elbaite.
t Also called aluminobuergerite (Foit, 1975) and olenite (Burt, 1989).
Primary pyroelectricity describes the development of
an electric voltage at the poles of certain noncentrosymmetric crystals, such as tourmaline (space group R3m),
as temperature is changed. A permanent electric dipole
or spontaneous polarization is inherent along the (00./)
axis in tourmaline. As temperature is varied, the charge
distribution in the structure shifts to produce a voltage
along this axis. This voltage dissipates as atmospheric
molecules are adsorbed onto the surface, so the crystal
soon reverts to electrical neutrality.
The primary pyroelectric coefficient is a vector property, isolated when the external electric field, applied stress,
and applied strain on a crystal are constant or zero (Nye,
1957). However, thermal expansion in a crystal held under such conditions establishes a strain field. Thus, a
component of the measured pyroelectric coefficient is
caused by the piezoelectric effect. This component, known
as secondary pyroelectricity, is important for this work,
as it produces between 75 and 90% of the observed pyroelectric effect in tourmaline (Zheludev, 1971; Bhalla
and Cross, 1981; Lang, 1974). It is impractical to prevent
thermal expansion of a crystal by mechanical clamping.
Thus, pyroelectric coefficients of tourmaline are measured parallel to the (00./) directions and the d33 piezoelectric tensor from unclamped crystals at constant stress
(0). The measured pyroelectric coefficient (P3) is the sum
of these primary and secondary coefficients.
A nonuniform strain field caused by a temperature gradient produces tertiary pyroelectricity (Nye, 1957). This
effect produces erratic data. Therefore, tertiary pyroelectricity must be avoided by maintaining uniform heat distribution in the material during experimental determination of pyroelectric coefficients.
For this study, relationships between pyroelectricity and
the chemistry of tourmaline are investigated for selected
samples from -170 to - 500 K. From these data, the
effects of chemical composition on the P3 variation in
tourmalines are determined. These data complement previous studies, which correlated the crystal chemistry of
tourmalines with other material properties such as color
and refractive indices (e.g., Rossman et aI., 1991; Deer
et al., 1992) and magnetic and mechanical properties (e.g.,
Donnay et aI., 1967; Tsang et aI., 1971; Dobrovol'skaya
and Kuz'min, 1975; Helme and King, 1978; Tatli, 1980,
1985; Tatli and Ozkan, 1987).
EXPERIMENTAL METHODS
Preparation
of single-crystal
plates
Pyroelectric coefficients were obtained from singlecrystal tourmaline plates with normals parallel to the
(00./) crystallographic directions. Crystals detailed in Table 2 were aligned by back-reflection Laue diffractographs. The diffractographs showed the error in alignment, f) < 0.5°. This tolerance produces a 0.004% error
between the measured pyroelectric coefficient (PO and
the true coefficient (P3) ofthe plate, by the following equation:
P1
-
P1
P1'
= (l - cos f).
(I)
This equation is derived from Figure I, which schematically illustrates the relationship between coefficient vec300 and 900 ILm thick
tors and a plate surface. Plates
were sliced from the crystal and then polished with flatness and parallelism within a :t 1O-lLm tolerance (Hawkins, 1993).
Translucent polished plates were examined with the
petrographic microscope to select crystal volumes suitable for pyroelectric measurement. Material with cracks
or color zoning was rejected. Regions with uniaxial conoscopic images were cut from these larger plates to form
test plates for pyroelectric measurements, since biaxial
fringes are evidence for internal strain or chemical zoning
-
HAWKINS
TABLE2.
ET AL.: PYROELECTRIC
EFFECT IN TOURMALINE
493
Tourmaline mineral samples
Sample
Color and «00./) axis
length, basal dial (em)
S1b
S3a
S8
S19
S21
S32
S33a
S35
S36
S37
S40
S41
S42
S43
S49
S50
dark green (3.5, 0.5)
black (3.0, 0.9)
green-blue (1.0, 0.7)
black (0.5, 2.0)
black (3.8, 1.4)
dark green (2.5, 1.0)
black (1.2, 0.9)
dark blue (1.5, 0.7)
pale aqua (1.5, 1.3)
dark blue (1.5, 1.0)
mauve (1.0, 2.5)
black (1.0, 3.5)
dark olive (0.5, 5.0)
pale green (2.0, 0.5)
black (3.0, 1 .0)
black (5.1, 1.5)
Provenance
Maine, U.S.A.
Harts Range, Australia
Minas Gerais, Brazil
Ratnapura, Sri Lanka
Pakistan
Brazil
Rollstone Hill, U.S.A.
Hindu Kush, Afghanistan
Plumbago Mine, U.S.A.
Usakos, Namibia
Plumbago Mine, U.S.A.
Tourmaline Queen, U.S.A.
Madagascar
Himalaya Mine, U.S.A.
Black Jack, Australia
unknown
Museum catalogue no.
S.K. Dobos
S.K. Dobos
Daddow's Rock Shop
Smithsonian B14657
QGRS
Harvard H#91844
Harvard (no cat. no.)
AMNH 41859
AMNH 40730
AMNH 24116
AMNH 40732
AMNH 46244
Gemological Inst
Gemologicallnst
Australian D14988
D.J. Henry
Note: tourmaline samples were loaned and donated from the following sources: S.K. Dobos, Department of Earth Sciences, University of Queensland,
Brisbane, Australia;
Daddow's Rock Shop, Stanley Street, Woolloongabba,
Brisbane, Queensland, Australia; Quality Gem Rough Supplies (QGRS),
P.O. Box 129, Kyneton, Victoria, Australia;
D.J. Henry, Department of Geology and Geophysics,
Louisiana State University, Baton Rouge, Louisiana
70803-4101,
U.S.A.; Harvard University Mineralogical
Museum, 24 Oxford Street, Cambridge, Massachusetts
02138, U.S.A.; American Museum of
Natural History (AMNH), Central Park West at Seventy-ninth
Street, New York, New York 10024-5192,
U.S.A.; National Museum of Natural History,
Smithsonian
Institution, Washington,
DC 20560, U.S.A.; Gemological Institute, 1660 Stewart Street, Santa Monica, California 90404, U.S.A.; and the
Australian Museum, 6-8 College Street, Sydney, New South Wales 2000, Australia.
(Foord and Mills, 1978). These flaws may produce tertiary pyroelectric effects. For plates containing no volumes
of uniaxial structure, volumes of low biaxial distortion
were selected. Test plates were cut from the opaque schod
without petrographic examination. These schod samples
(S3a, S21, and S49) have highly mosaic structures, as
shown by neutron-rocking curves across individual Bragg
reflections (Hawkins, 1993).
Surface areas of opaque and densely colored plates were
measured from images captured with a Panasonic WVBL600 video-rate CCD camera fitted with a Canon 50
mm macro lens and extender tube. The captured images
were analyzed with the MD30-plus Image Analysis System on an IBM-PC 386. Conversely, the S36 plate did
not provide sufficient contrast for image analysis. The
surface area for this plate was obtained from its density
measured by heavy liquids and its thickness and mass.
Pyroelectric
coefficient measurement
Pyroelectric coefficients were measured by an electronic circuit (Schneeberger, 1992) built after Byer and Roundy (1972). A Au film electrode was evaporated onto both
faces of each plate of surface area A and then mounted
on a heating element and cooled to -170 K. The temperature was increased at a constant rate (dT/dt) between
::t2 and ::t6 Klmin to 500 K and then decreased at a
constant rate. The varying pyroelectric current (I) measured at 100 points during each increasing and decreasing
temperature ramp was used to calculate the pyroelectric
coefficient at each temperature, P'3 (T), by the following
equation (Byer and Roundy, 1972):
I
pj (T) =
A(dT/dt)'
The measurements
(2)
were repeated for each plate with dif-
ferent heating rates. At least two plates from each tourmaline specimen were measured to check for reproducibility.
Electron probe microanalysis
and structural formulae
The tourmaline plates were coated with a 200 A C film.
Chemical analyses were collected using a JEOL JXA8800L automated electron microprobe with four wavelength dispersive spectrometers. Thirteen elements were
calibrated against the following primary standards: Si02,
spessartine garnet; Ti02, Ti02; Al203, spessartine garnet;
Cr203, chromite; FeO, hematite; MnO, spessartine garnet; NiO, synthetic Ni2SO.; ZnO, ZnS; MgO, MgO; CaO,
wollastonite; Na20, albite; K20, orthoclase; F, MgF2.
Secondary standards were analyzed prior to each session.
Line scans were made approximately normal to each
other across each plate with a 15-kV, 15-nA beam. Analyses were taken at 100 or 200 Jim intervals. These intervals are broad compared with the fine, 1-3 Jim zonation
features in some tourmalines (e.g., Grice and Robinson,
1989; Cavarretta and Puxeddu, 1990; Henry and Du-
Fig. I. Geometric relationship between the true (P,) and
measured (PJ') pyroelectric coefficients and the surface of a crystal plate.
--
HAWKINS
494
TABLE3.
Sample
SiO,
TiO,
AI,O,
Cr,O,
FeO
MnO
NiO
ZnO
MgO
CaO
Na,O
K,O
F
-F= 0
Total
Si
B
Na
Ca
K
~Xsite
Ti
AI
Cr
Fe
Mn
Ni
Zn
Mg
Li*
~Ysite
AI
Mg
Fe
~Zsite
F
OH
Average
ET AL.: PYROELECTRIC
EFFECT
of the pyroelectric
IN TOURMALINE
compositions
and structural
S1b
S3a
S8
S19
S21
S32
S33a
S35
37.32(50)
0.18(5)
36.32(34)
0.01(2)
3.59(24)
0.98(8)
0.01(2)
0.04(4)
0.19(3)
2.11(3)
1.73(14)
0.02(1)
1.29(29)
-0.54
83.25(70)
35.04(41)
0.63(7)
31.06(26)
0.01(2)
14.29(15)
0.59(4)
0.02(3)
0.08(5)
1.98(12)
0.46(5)
2.28(7)
0.07(1)
0.59(5)
-0.25
86.86(58)
37.04(42)
0.01(2)
37.47(29)
0.01(2)
2.40(7)
1.41(5)
0.01(2)
1.88(7)
0.00(1)
0.04(2)
2.08(11)
0.01(1)
0.51(15)
-0.21
82.64(65)
36.12(58)
1.18(20)
27.13(35)
0.08(4)
0.83(12)
0.01(2)
0.01(3)
0.03(4)
13.76(30)
3.94(50)
0.85(8)
0.01(1)
0.70(10)
-0.29
84.34(96)
35.16(44)
0.60(13)
32.67(16)
0.01(1)
11.07(70)
0.45(3)
0.01(2)
0.41(7)
2.23(56)
0.11(2)
2.04(8)
0.04(1)
0.44(13)
-0.19
85.31 (77)
37.04(55)
0.27(4)
35.74(29)
0.01(1)
5.98(13)
0.34(4)
0.01(1)
0.05(5)
1.10(6)
0.11(2)
2.58(8)
0.03(1)
1.33(24)
-0.21
84.4(1.1)
35.12(47)
0.80(21 )
32.88(60)
0.01(1)
11.41(19)
0.15(2)
0.01(2)
0.07(4)
3.30(27)
0.52(16)
1.98(12)
0.05(1)
0.46(14)
-0.19
86.58(88)
37.74(84)
0.02(2)
36.61 (58)
0.01(1)
6.74(16)
0.52(4)
0.01(1)
0.09(4)
0.10(1)
0.12(2)
2.69(11)
0.03(1 )
1.15(8)
-0.48
85.3(1.3)
6.00
3.00
0.54
0.36
0.00
0.90
0.02
0.88
0.00
0.48
0.13
0.00
0.00
0.05
(1.44)
1.56
6.00
0.00
0.00
6.00
0.66
3.11
6.00
3.00
0.76
0.08
0.02
0.86
0.08
0.37
0.00
1.95
0.09
0.00
0.01
0.51
6.00
3.00
0.65
0.01
0.00
0.66
0.00
1.15
0.00
0.33
0.19
0.00
0.22
0.00
(1.11)
1.89
6.00
0.00
0.00
6.00
0.26
3.45
6.00
3.00
0.67
0.02
0.01
0.70
0.08
0.57
0.00
1.58
0.07
0.00
0.05
0.57
(0.08)
2.92
6.00
0.00
0.00
6.00
0.24
3.62
6.00
3.00
0.81
0.02
0.01
0.84
0.03
0.82
0.00
0.81
0.05
0.00
0.01
0.27
(1.01)
1.99
6.00
0.00
0.00
6.00
0.68
3.32
6.00
3.00
0.66
0.10
0.01
0.77
0.10
0.62
0.00
1.63
0.02
0.00
0.01
0.84
6.00
3.00
0.83
0.02
0.01
0.85
0.00
0.86
0.00
0.90
0.07
0.00
0.01
0.02
(1.14)
1.86
6.00
0.00
0.00
6.00
0.58
3.50
3.01
5.90
0.00
0.10
6.00
0.32
3.48
formulae
--
Ions per formula unit
6.00
3.00
0.27
0.70
0.00
0.97
0.15
0.00
0.01
0.12
0.00
0.00
0.00
2.72
3.00
5.31
0.69
0.00
6.00
0.37
3.38
test plates
3.22
6.00
0.00
0.00
6.00
0.25
3.50
Note: n.a. = not analyzed. Values in parentheses are 1 esd of the microprobe traverse analyses.
* Li ions pfu are the difference from full Y site occupancy.
trow, 1992). The intervals used in this work allowed each
plate to be investigated within the constraints of instrument time, yet gave sufficient data to obtain an average
composition
with estimated standard deviations
(esd)
representative of chemical variation across the plates.
Structural formulae were normalized to six Si atoms
per formula unit. The assumptions for cation site assignments are similar to those summarized by Rosenberg
and Foit (1979). B is assumed to be present in stoichiometric amounts. Na, Ca, and K are assigned to the [91X
site. Multivalent cations are assumed to be in the oxidation states Fe2+, Mn2+, and Ti4+. Al is assigned to the
[6JY
smaller [6JZsite and excess Al is assigned to the larger
site. All remaining cations are assigned to the Y site. If
there is a deficiency of Al, the smaller cations such as Mg
and excess Fe2+ (in lieu of FeH) are assigned to the Z
site, using crystal ion radii (Shannon, 1976) as a guide.
No attempt is made to partition the same ion between
both octahedral sites. An exception is the S3a schod in
which Fe is partitioned between the Y and Z sites, using
the ratio derived from M6ssbauer studies and refined
during single-crystal
X-ray structure determination
(Hawkins, 1993).
Some Y site vacancies are present when these criteria
are followed. However, structural refinement has shown
that Y site vacancies are unlikely in tourmaline (Donnay
and Barton, 1972; Foit and Rosenberg, 1971; Rosenberg
and Foit, 1979; Hawkins, 1993). Therefore, Y site vacancies are filled by Li ifthe mineral has a lithia-rich provenance, i.e., if it contains elbaite or liddicoatite components. The number of H ions is estimated from bond
valence sums (Brese and O'Keeffe, 1991).
RESULTS
Test plate quality
Chemical variation within the tourmaline test plates may
affect their suitability for electrical measurements. Thus,
two statistical tests were applied to the sets of traverse
analyses to assess (1) whether there is chemical zonation
within the plates and (2) whether two or more plates from
the same crystal have the same average chemistry.
First, the variation in concentrations along a traverse
was compared with the counting statistics for each element. For this study, microprobe traverses do not show
zoning at a 95% confidence level if the relative variation
HAWKINS
ET AL.: PYROELECTRIC
EFFECT
495
IN TOURMALINE
TABLE 3.-Continued
8ample
8i02
Ti02
AI203
Cr203
FeO
MnO
NiO
ZnO
MgO
CaO
Na20
K20
F
-F=O
Total
8i
B
Na
Ca
K
2: X site
Ti
AI
Cr
Fe
Mn
Ni
Zn
Mg
U*
1: Y site
AI
Mg
Fe
2: Z site
F
OH
836
837
840
841
842
843
849
850
38.83(26)
0.01(1)
38.32(23)
0.01(1)
2.13(21)
0.57(5)
0.01(1)
0.13(3)
0.02(1)
0.52(8)
2.18(5)
0.02(1)
0.92(6)
-0.39
83.29(47)
37.68(33)
0.01(2)
36.98(38)
0.01(2)
4.84(22)
0.80(5)
0.01(2)
0.27(10)
0.03(2)
0.32(3)
2.59(7)
0.02(1)
1.13(7)
-0.47
84.21(60)
39.45(76)
0.01(1)
40.39(26)
0.01(2)
0.01(1)
0.18(4)
0.01(2)
0.02(4)
0.00(1)
0.49(5)
1.65(11)
0.01(1)
n.a.
34.7(1.0)
0.03(4)
35.77(25)
0.01(2)
8.84(18)
1.72(8)
0.01(3)
0.89(15)
0.05(2)
0.03(2)
1.93(9)
0.03(2)
0.49(16)
-0.21
84.3(1.0)
37.87(61)
0.07(4)
38.36(63)
0.01(2)
2.29(14)
0.34(6)
0.01(3)
0.06(7)
0.03(2)
0.44(13)
2.13(10)
0.02(2)
0.87(24)
-0.37
82.1(1.0)
38.18(15)
0.03(2)
38.61(24)
0.01(1)
0.79(6)
0.77(4)
0.01(2)
0.05(4)
0.01(1)
1.48(8)
1.69(4)
0.01(1)
1.10(6)
-0.46
82.27(32)
35.39(80)
0.41 (23)
33.2(1.0)
0.01(2)
14.04(94)
0.16(5)
0.01(2)
0.07(8)
0.63(11 )
0.10(6)
1.77(20)
0.04(2)
0.46(20)
-0.43
86.1(1.4)
36.91(48)
0.02(2)
35.49(56)
0.01(3)
8.32(22)
0.31(4)
0.01(2)
0.20(6)
0.11(2)
0.06(2)
2.66(9)
0.03(1)
1.60(28)
-0.67
85.05(73)
6.00
3.00
0.65
0.09
0.00
0.74
0.00
0.98
0.00
0.28
0.07
0.00
0.01
0.00
(1.66)
1.34
6.00
0.00
0.00
6.00
0.45
3.40
6.00
3.00
0.80
0.05
0.00
0.85
0.00
0.94
0.00
0.64
0.11
0.00
0.03
0.01
(1.28)
1.72
6.00
0.00
0.00
6.00
0.57
3.53
6.00
3.00
0.51
0.25
0.00
0.76
0.00
1.15
0.00
0.10
0.10
0.00
0.01
0.00
(1.64)
1.36
6.00
0.00
0.00
6.00
0.55
3.31
6.00
3.00
0.58
0.02
0.01
0.61
0.05
0.64
0.00
1.99
0.02
0.00
0.01
0.16
(0.13)
2.87
6.00
0.00
0.00
6.00
0.25
3.62
6.00
3.00
0.84
0.01
0.01
0.86
0.00
0.80
0.00
1.13
0.04
0.00
0.02
0.03
(0.98)
2.02
6.00
0.00
0.00
6.00
0.82
3.22
82.2(1.2)
6.00
3.00
0.49
0.08
0.00
0.57
0.00
1.24
0.00
0.00
0.02
0.00
0.00
0.00
(1.74)
1.26
6.00
0.00
0.00
6.00
n.a.
3.88-F
Ions per formula unit
6.00
6.00
3.00
3.00
0.65
0.65
0.01
0.07
0.01
0.00
0.67
0.72
0.00
0.01
1.28
1.16
0.00
0.00
1.28
0.30
0.25
0.05
0.00
0.00
0.11
0.01
0.01
0.01
(0.08)
(1.46)
2.92
1.54
6.00
6.00
0.00
0.00
0.00
0.00
6.00
6.00
0.27
0.44
3.62
3.38
in concentration for each element does not exceed twice
the relative variation in counting statistics. By this criterion, all plates except S40 and S43 showed zoning in
some elements.
,
Second, differences between the average chemical compositions of plates cut from the same tourmaline specimen were determined for each element within a 95% confidence interval by a Student t-test (Walpole and Myers,
1978). In most cases, the compositions of all elements
were consistent between plates. Thus, the analyses shown
in Table 3 are the averages of microprobe traverses within these confidence limits for each sample.
Pyroelectricity vs. temperaturecurves
In this temperature range, the pyroelectric response
for tourmaline is a convex curve, as illustrated in Figure
2. Curves for the repeated temperature cycles and for
several plates from the same crystal were fitted with a
quadratic function. The error in P'3shown in Table 4 is
calculated from the spread of the data from the fitted
function (Schneeberger, 1992). In most cases, plates from
the same specimen gave the same curve within these error margins. Thus, the dimensional tolerances, chemical
zonation, and flaws in these tourmaline plates were ac-
------
ceptable for reproducible measurement of P'3.Repeat data
collected from the same crystal were added to form the
quadratic functions for P'3vs. T given in Table 4. The Sib
and S41 samples are exceptions to the reproducibility
generally observed within samples. The two Sib samples
gave consistently different intersecting curves, as shown
by the equations in Table 4. An explanation for these
differences is not obvious because both plates are of similar quality. Thus, SIb results are omitted from further
analysis. Similarly, one S41 plate was omitted from the
average curve for this sample because it produced pyro10% higher
electric coefficients that were consistently
than those for three other S41 plates. This difference may
be attributed to inaccurate surface-area measurement of
one plate.
The low-temperature form of the pyroelectric response
curve is shown in Figure 2. The lowest temperature obtained in this work was limited by the liquid N2 coolant,
and so the polarity inversion in tourmaline P'3 at <20 K
(e.g., Drozhdin et ai., 1975; Gavrilova et ai., 1989) was
not detected.
The high-temperature
region of the curve adopts one
of two forms. First, there is an upward inflection of the
curve shown in Figure 2. This behavior has been noted
-
~-
496
HAWKINS
ET AL.: PYROELECTRIC
10
Second, a smooth decrease in the pyroelectric coefficient was observed in the schor! S33a plate above ",430
K. This behavior, which is probably caused by the onset
of electrical conductivity, has not been previously reported in tourmaline. Such conductivity is different from
the electric current generated in the external circuit from
the pyroelectric voltage. This tourmaline is omitted from
further analysis because it produced an anomalously low
Pj. This problem arose from a difficulty in the accurate
determination of surface area because of chipped edges.
Further upper temperature limits were encountered with
S21, S32, and S41 plates, as shown in Table 4. The signals
from these plates contained high levels of electronic noise
at higher temperatures.
8
Q'
~6
5
u
3-
4
tJ ..,
~2
0
0
200
400
Temperature (K)
600
800
Fig. 2. The general form of the pyroelectric response with
temperature change for tourmaline. The approximate temperature range for the present measurements is between T, and T2.
A low-temperature polarity inversion is shown at <20 K (after
Gavrilova et ai., 1989).
for only a pale rose sample above 450 K (Ackermann,
1915; Gavrilova et aI., 1983, 1989). However, this type
of behavior was observed during one experiment of the
Sl b/5 plate above -420 K and for both S49 plates above
400 K. A similar inflection has been predicted for the
thermal expansion vs. temperature relationship in other
minerals. For example, strong anharmonic thermal vibration above 1300 K in forsterite produces an upward
inflection of the calculated thermal expansion coefficient
curve (Reynard et aI., 1992). Thermal expansion causes
the secondary pyroelectric effect, which is a strong effect
in tourmaline. Thus, this region of the curve in Figure
2 may reflect the effects of anharmonic vibration on pyroelectricity. These results indicate that the onset of pyroelectricity that is caused by anharmonic vibration of
ions in the structure and high thermal expansion may
occur at a lower temperature in Slb/5 and S49 than in
other tourmalines.
Empirical equations
for the temperature
Pyroelectric equations
*
-0.2379
4.379
+ 1.732 x 10-2TX 10-2
coefficient variation with bulk chemistry
Clearly, tourmaline bulk chemistry determines the pyroelectric response of this mineral. To investigate trends
between the pyroelectric coefficients and chemistry, pyroelectric coefficients were calculated from equations for
14 samples in Table 4 at two temperatures. Low-temperature behavior is represented by data at 193 K. Although
500 K, some plates
most samples were measured to
gave erratic responses above -400 K. Therefore, higher
temperatures are represented by data at 383 K, so all 14
samples could be included in the data set. The variations
in tourmaline pyroelectric coefficients with bulk chemistry are shown by the graphs in Figure 3.
FeO. The linear decrease in pyroelectric coefficient with
increase in Fe content (as wt% FeO), shown in Figure 3a,
is the most important result. The data at 193 and 383 K
are strongly linear, as the linear least-squares correlation
coefficient r is -0.9 in both cases. This high correlation
is statistically significant, since the random probability
for 14 data points with a correlation coefficient between
0.85 and 0.90 is <0.1% (Taylor, 1982). The higher temperature data are more strongly linear, and the slope of this
line is higher than for the 193 K data. Thus, an increase
in Fe content has a stronger influence on tourmaline py-
of p~
p; ["C/(m2.K)]
8ample
81b/5*
81b/6*
83a
88
819
821
832
833a*
835
836
837
840
841
842
843
849
850
dependence
Pyroelectric
-
-
TABLE4.
EFFECT IN TOURMALINE
9.54 x 10-oP
+ 1.178 x 10-2T - 8.133 x 10-oP
-0.3686 + 1.112 x 10-q - 4.259 x 10-oP
-0.979 + 2.194 x 10-2T-1.873
x 10-'P
-0.8659 + 1.97 x 10-2T1.652 x 10-'P
-2.661 + 2.849 x 10-2T - 3.170 x 10-'P
-1.588 + 2.257 x 10-Q1.976 x 10-'P
-0.6975 + 1.056 x 10-Q6.903 x 10-oP
-1.88 + 2.308 x 10-2T - 1.976 x 10-'P
-1.19 + 2.33 x 10-2T1.987 x 10-'P
-1.310 + 2.024 x 10-2T1.578 x 10-'P
-0.4096 + 1.816 x 10-Q1.226 x 10-'P
-0.8918 + 1.950 x 10-2T1.620 x 10-'P
-0.9894 + 1.979 x 10-2T1.465 x 10-'P
-0.7167 + 2.246 x 10-2T - 1.883 x 10-'P
-0.9472 + 1.500 x 10-2T - 9.123 x 10-oP
-1.585 + 1.954 x 10-2T1.566 x 10-'P
Error
0.10
0.10
0.17
0.07
0.20
0.15
0.22
0.15
0.07
0.11
0.10
0.10
0.16
0.07
0.14
0.10
0.06
from these samples are not included in the anlaysis of p; variation with chemistry.
T range (K)
170-423
170-500
170-500
170-500
170-500
170-463
170-483
170-433
170-500
170-500
170-500
170-500
170-443
170-500
170-500
170-403
170-500
HAWKINS
ET AL.: PYROELECTRIC
roelectricity at higher temperature. This effect could be
caused by an increase in electrical conduction of Fe-rich
tourmaline at elevated temperatures and is consistent with
the proposed onset of electrical conductivity in schor!.
The pj may approach a plateau for the Fe-rich tourmalines S3a, S21, and S49, as indicated by the term 'schor!'
in Figure 3a. This small invariant region is more evident
at low temperature.
SiOz. The pj variation with SiOz content is linear, as
shown in Figure 3b. However, this variation results from
the heavy element content such as Fe, since there is only
minor substitution for Si in tourmaline (e.g., Walenta and
Dunn, 1979; Henry and Guidotti, 1985; Gallagher, 1988).
An inverse proportionality between weight percent SiOz
and weight percent FeO is seen by comparison of Figure
3a and 3b.
MnOz. A weak linear relationship between pj and weight
percent MnOz is shown in Figure 3c.
Alz03. There are two regions denoted by 1 and 2 in
circles in Figure 3d. In the AI-rich region (1), pj increases
with Alz03 concentration. This region is detailed in Figure 3e. A minimum pj value corresponding to the schor!
samples S3a, S21, and S49 identified in the FeO relationship in Figure 3a occurs around 31-33 wt% A1Z03. The
data in this minimum region fit the linear relationship
shown in Figure 3e without degrading the correlation coefficient. Therefore, the data points at this minimum are
included in region 1. However, r may appear artificially
high because the data set is small. With further depletion
of Al from tourmaline, the pyroelectric coefficient increases again (Fig. 3d, region 2). The data point shown
near 27 wt% Alz03 represents the Sl9 uvite-dravite, supporting the data in Figure 3f and indicating that incorporation oflarge amounts ofMg in the tourmaline lattice
increases pj.
MgO. There is a discontinuity in data between
3 wt%
MgO and the S 19 uvite-dravite at 14 wt% MgO in Figure
3f. At MgO concentrations below the sharp inflection
around 0.10-0.50 wt% MgO, the relationship between
pj and weight percent MgO is strongly linear, despite the
large composition error bars (Fig. 3g). However, a paucity
of suitable dravite or uvite tourmalines prevented detailed investigation of the effect of Mg incorporation on
the pyroelectric coefficients.
ZnO, NazO, CaO, and F. Zn, F, and the X-site cations
do not significantly affect the pyroelectric coefficients. Zno
content vs. pj displays a very weak inflection around 0.10.5 wt% ZnO in a trend similar to that of MgO vs. Pj.
There is a weak correlation between NazO and pj, as
shown in Figure 3h.
-
DISCUSSION
Pyroelectric
coefficients of Fe-containing
497
EFFECT IN TOURMALINE
b) SiO,
a) FeO
6.00
~
5.00
~.f...+,
;
i? 4.00
I
~ 3.00
1J)(f)
i;P-~
h...!..
.+
M.g
".:'
2.00
[illKJ
2
4 6 8
WIll.FeO
~..~
$-
_
o
10 12 14 16 33
I
5.00
~
b
+
1
...
~
I
3.00
+t
+
+
:f..
37
wt% Si02
41
39
$
e
.:'
+
+......-
t
[§KJ
e...~
~CD ~++
G)
+
e
([).:~~..E!j
2.00
35
-@- e~
-:--r'
d) AI,O,
c) MnO
6.00
Mi?4.00
E
*-"-
~
GJ
1.00
0.00
,
GJ$0.....
e '. '$
+
..-~
-+-~I"
~-
e
e.$. e
$41
41B- -e-
1.00
0.00
6.00
0.0
e)
0.5
1.0
wt%MnO
¥ 4.00
'"
ME
wt% A1203
f) MaO
I§]
t
I I
+....t...:-::..-
2.00
e
..~~...-$
([) ..~...1:(ffr
E!
l$~
[!§
30
32
34
36
WIll.
38
40
42
2
0
A'zDJ
6.00
g) MgO low concentration
5.00 §]-t--
br£'
ill
-r
"
1.00
0.00
28 30 32 34 36 38 40 42
__ +
.
-t"~_...'+
....+~."'+-
~3.00
§]
".::
2.0 26
AI"O'1 high concentration
I383K I
5.00
1.5
-+--+'-'1
~
2.00
'1"
h)
+
-+'"
+-:c
3-
--+- + +
~
-$-
i!...
10
12
14
Na,O
+..
.....__
4
6
8
WIll. MgD
.$.h..d
-
*"-
$-~
-<17--6)- -6)- '€r
1.00
0.00
0.00
0.05
wIll.MgD
0.10
0.15 0.5
1.0
1.5
2.0
wIll.N'zD
2.5
3.0
Fig. 3. Relationships between P'3 and the bulk chemistry of
tourmaline. Error bars are :t 1 esd for chemical composition (Table 3) and P'3 (Table 4). Equations have the form p, = A + Bx,
where x is the weight percent oxide; and r is the correlation
coefficient of the linear least-squares fit of the line to the data.
(a) FeO 193 K: A = 2.59, B = -0.08, r
= -0.86; 383 K: A =
4.79, B = -0.10, r = -0.90. (b) SiOz 193 K: A = -4.51, B =
0.18, r = 0.61; 383 K: A = -5.91, B = 0.27, r = 0.73. (c) MnO
193 K: A = 1.99, B = 0.25, r = 0.27; 383 K: A = 4.02, B =
0.28, r = 0.24. (d) Alz03. (e) Alz03 193 K: A = - 2.99, B = 0.14,
r = 0.83; 383 K: A = -2.94, B = 0.20, r = 0.90. (f) MgO. (g)
MgO 193 K: A = 2.65, B = -9.14, r = -0.88; 383 K: A = 4.81,
B = -9.92, r = -0.86. (h) NazO 193 K: A = 2.94, B = -0.39,
r = -0.44; 383 K: A = 4.84, B = -0.31, r = -0.28.
tourmalines
These data form a basis for the prediction of tourmaline pyroelectric coefficients directly from electron microprobe FeO analyses. All tourmaline compositions studied
conform to the linear relationship between weight percent
FeO and pj shown in Figure 3a. Fe-rich tourmalines are
often shown in this study to be black or darkly colored.
Thus, this relationship supports the general observation
(e.g., Dietrich, 1985) that black tourmalines exhibit weaker
pyroelectric behavior than pale colored tourmalines.
However, the S19 uvite-dravite is Fe-poor and has a cor-
------.-
498
HAWKINS
TABLE5.
ET AL.: PYROELECTRIC
EFFECT IN TOURMALINE
Predicted pyroelectric coefficients PO [ILC/(m2.K)] and
errors for selected tourmaline end-member compositions
p~ at
p~
193 K at 193 K
(Fig. 4) (Eq. 3)
Endmember
Tourmaline (FeO = 0%)
Schorl (FeO = 20.46%)
Uvite
Elbaite
Aluminobuergerite
2.59(0.50)
1.02(0.56)
p~
at 296 K
(Eq.4)
p~
at 383 K
(Eq.5)
3.95(0.51 )
2.03(0.57)
4.79(0.52)
2.68(0.58)
6.00
5.00
Q
'"'5 4.00
.,
~3.00
I:>
Fe-free tourmalines
2.45
3.08
5.75
elbaile
Na(All.SLi1.5)AI6
uvite
CaMg3(AlsMg)
o
519
"2.00
o
aluminobuergerite
NaAl3 Al6
1.00
4.5
5.0
5.5
6.0
6.5
7.0
7.5
8.0
8.5
9.0
AI ions per formula unit
respondingly high P'3 value, despite the dark brown or
black color. Therefore, these results confirm that color,
which has been used previously to characterize tourmalines, is not as reliable as chemistry for the prediction of
the pyroelectric coefficients.
Predictive equations between the P'3 and FeO content
at three temperatures were devised by simple regression
analysis, with weight percent oxide as the independent
variable and P'3 as the dependent variable. These equations include the following:
at 193 K
pj = 2.594
-
O.077(wt% FeO)
ILC/(m2. K)
at 296 K
pj = 3.949
-
0.094(wt% FeO)
1LC/(m2.K)
at 383 K
pj
(3)
(4)
= 4.791 - 0.103(wt% FeO)
ILC/(m2. K).
(5)
Pyroelectric coefficients and errors calculated at 193,
296, and 383 K for an ideal schod end-member containing 20.46% FeO (Dietrich, 1985) and a tourmaline with
FeO = 0% are listed in Table 5. Pyroelectric coefficients
predicted for tourmalines at 296 K are in good agreement
with values for similar samples reported at ambient temperature. For example the P'3 value of 2.03(57) ILC/(m2.
K) at 296 K for the schod agrees extremely well with the
value of 2.2 1LC/(m2.K) for a black tourmaline (Gladkii
and Zheludev, 1965). This black tourmaline is probably
close to a schod in composition. Further, the P'3 of 4.4
ILC/(m2. K) at ambient temperature reported for pink
tourmalines (Hayashi, 1912; Ackermann, 1915) is within
the error of the value 3.95(51) 1LC/(m2'K) at 296 K predicted for Fe-free tourmaline. Pink tourmaline is usually
an elbaite with minor Mn substitution.
The errors in P'3predicted from Equations 3-5 are calculated as a 90% confidence interval based on a Student
t-test (Walpole and Myers, 1978). These intervals are large
percentage errors in P'3. Such errors may be reduced only
by increasing the number of samples in the data set. However, it is very difficult to obtain tourmaline crystals of
the quality required for this measurement technique. For
Fig. 4. The number
ys. p, at 193 K.
of AI ions per tourmaline
formula
unit
example, the 14 samples used in this analysis represent
only 27% of the tourmaline crystals collected for this work
(Hawkins, 1993).
It is surprising that the correlation between weight percent FeO and P'3 is so high, despite the assumption that
Fe is divalent for all tourmaline and that the effects of Fe
partitioning between the Y and Z sites have not been
taken into account.
The strong effect of FeO concentration may obscure
the effects of other ions on P'3. Thus, an investigation of
the complex crystal chemical substitutions in tourmaline
is used to show the dependence of the pyroelectric response on such ions.
Cation substitution
effects on the pyroelectric coefficient
Y site. The relationship between the number of Al ions
per formula unit (pfu) andp'3 is shown in Figure 4. Above
6.5 Al ions in region 1, P'3 increases with Fe removal by
the schod-elbaite substitution on the Y crystallographic
site: 2Fe~+
Li~ + AR+. This substitution corresponds
""
to the dependence
of P'3 on weight percent FeO. Therefore, the schod-elbaite substitution is probably the dominant mechanism for Fe removal to produce high P'3coefficients in AI-rich tourmalines. In these tourmalines, Z
is usually fully occupied by AI.
A similar linear increase in P'3 with increased Mn suggests that Fe is also removed by the Y site cation substitution: Fe?+
Mn~+. However, this cation exchange is
""
not important for all tourmaline compositions studied,
since the amount ofMn incorporated by this substitution
is small, <2 wt% MnO. Further, the relationship is weak,
as the linear-fit correlation coefficients for data at 193 and
383 K are -0.25, as shown in Figure 3c. At high Mg
Mg~+ substitutions deplete Fe
concentrations the Fe?+
""
then increases, as shown
from the structure. The P'3value
in Figure 3f.
Around the minimum inflection of the weight percent
AI203 vs. P'3 relationship in Figure 4, Fe removal frorr
the Y site and substitutions for Al on the Z site occur
HAWKINS
ET AL.: PYROELECTRIC
The effect of Al content on pj is probably a result of the
substitution of AI for Y and Z site cations, particularly
Fe. However, there may be properties of AI, Fe, and other
ions that affect polarity in the structure and pyroelectric behavior.
Z and X sites. The regions in Figure 4 that are low and
intermediate in Al content, shown as region 2, indicate
that a further set of cation substitutions may affect the
pj values in these tourmalines. A line cannot be drawn
from these data because there are insufficient data points.
Thus, the data from samples S19, S3a, and S21 (labeled)
are discussed individually. The low Al content of these
tourmalines suggests that there are other cations on the
Z site. For example, the S 19 uvite-dravite structural formula (given in Table 3) is depleted in AI. Thus, 0.69 Mg
ions and 5.31 Al ions are incorporated in the Z site by
the uvite substitution Nait + AH+-= Ca~++ Mgi+.Powder electron-spin resonance data collected at 130 K
499
EFFECT IN TOURMALINE
uvite
CaMg3(AlsMg)
1.00
0.0
0.5
1.0 1.5 2.0 2.5
Mg ions per fonnula unit
3.0
Fig. 5. The number of Mg ions per tourmaline
YS.p, at 193 K.
3.5
4.0
formula unit
~
(Hawkins, 1993) indicate that there is also minor Fe3+ on
the Z site of S19. In addition, the S3a schorl has been
shown by M6ssbauer studies and population parameter
refinement during least-squares structure analysis (Hawkins, 1993) to contain 0.10 Fe ions per six Z sites. This
analysis suggests that a 13% substitution of Mg for Al
and a 1.7% substitution of Fe for Al on the Z site in the
S 19 uvite-dravite
and S3a schorl, respectively, may
strongly increase the pyroelectric coefficient. Other ions,
for example Mn (Johnston and Duncan, 1975), partition
to the tourmaline Z site. Thus, there are many possible
Z site substitutions for Al that may affect Pj.
Pyroelectric
coefficients for Fe-free tourmaline
There are several Fe-free end-members shown in Table
1. Pyroelectric coefficients of the elbaite [XY3Z6 =
Na(AII.5Lil.5)AI6)' uvite [CaMg3(AI5Mg»), and aluminobuergerite (NaAI3AI6) end-members are shown by extrapolation in Figures 4 and 5. There are sufficient data for the
schorl-elbaite substitution series in Figure 4 to extrapolate a value of pj for elbaite, which is included in the
summary of pj values in Table 5. Similarly, the pj for
uvite is extrapolated in region 2 of Figure 5, although
there are fewer data points than in region 1. It may be
difficult to obtain samples in this composition range with
further work because there is a miscibility gap reported
in the AI-Mg (elbaite-dravite) phase field for natural tourmalines (e.g., Donnay and Barton, 1972; Taylor and Slack,
1984; Henry and Guidotti, 1985; Cavarretta and Puxeddu, 1990; Kassoli-Foumaraki,
1990).
The prediction of aluminobuergerite
pyroelectric behavior from Figure 4 is justified only if the same linear
trend continues from the schorl-elbaite region to higher
AI compositions. It may be difficult to explore this relationship further using the present pyroelectric apparatus
because large crystals are required. Aluminobuergerite is
a synthetic end-member (Foit, 1975) produced when AI
content on the Y site is increased by the liddicoatite, aluminobuergerite,
and olenite substitutions (Henry and
Guidotti, 1985; Burt, 1989). Only crystals < 1 mm in size
---...--
are produced by hydrothermal synthesis (e.g., Taylor and
Terrell, 1967; Foit, 1975; Vorbach, 1989).
ACKNOWLEDGMENTS
This study could not have been completed without many beautiful tourmaline specimens that were made available by colleagues and friends.
Steve Dobos and Darrell Henry kindly donated tourmalines from their
own mineral collections. The curators of the Harvard University Mineralogical Museum, American Museum of Natural History, National Museum of Natural History (Smithsonian Institution), the Gemological Institute, and the Australian Museum kindly loaned crystals of line quality
for this work. We (KD.H. and H.S.) thank L. Bohaty of the LudwigMaximilians-Universitiit
(now at Universitat KOln, Germany), and S.
Haussiihl ofUniversitiit KOln for many helpful discussions. Thanks to H.
Schmid, Universite de Geneve, Switzerland, for helpful discussions that
alerted us (KD.H.) to potential problems with biaxial crystals for pyroelectric measurement. The Australian Institute for Nuclear Science and
Engineering (AINSE) supported KD.H. by means of a postgraduate studentship.
REFERENCES
CITED
Ackermann, W. (1915) Beobachtungen iiber Pyroelektrizitiit in ihrer Abhangigkeit von der Temperatur. Annalen der Physik, 46, 197-220.
Barton, R., Jr. (1969) Relinement of the crystal structure of buergerite
and the absolute orientation of tourmalines. Acta Crystallographica,
B25, 1524-1533.
Bhalla, A.S., and Cross, L.E. (1981) Primary and secondary pyroelectricity in proper and improper ferroelectrics. Ferroelectrics, 38, 935-938.
Brese, N.E., and O'Keeffe, M. (1991) Bond-valence parameters for solids.
Acta Crystallographica, B4 7, 192-197.
Buerger, M.J., Burnham, C.W., and Peacor, D.R. (1962) Assessment of
the several structures proposed for tourmaline. Acta Crystallographica,
15, 583-590.
Burt, D.M. (1989) Vector representation of tourmaline compositions.
American Mineralogist, 74, 826-839.
Byer, R.L., and Roundy, C.B. (1972) Pyroelectric coefficient direct measurement technique and application to a NSEC response time detector.
Ferroelectrics, 3, 333-338.
Cady, W.G. (1946) Piezoelectricity, 806 p. McGraw-Hili, New York.
Cavarretta, G.c., and Puxeddu, M. (1990) Schorl-dravite-ferridravite
tourmalines deposited by hydrothermal magmatic fluids during early
evolution of the Larderello geothermallield,
Italy. Economic Geology,
85, 1236-1251.
Deer, W.A., Howie, R.A., and Zussman, J. (1992) An introduction to the
rock-forming minerals (2nd edition), 528 p. Longman, London.
Dietrich, R.V. (1985) The tourmaline group, 300 p. Van Nostrand Reinhold, New York.
500
HAWKINS
ET AL.: PYROELECTRIC
Dobrovol'skaya,
N.V., and Kuz'min, V.I. (1975) Magnetic properties of
tourmaline. Konstitutsiya i Svoistva Mineralov, 9, 124-130 (in Rus.
sian).
Donnay, G. and Barton, R., Jr. (1972) Refinement ofthe crystal structure
of elbaite and the mechanism of tourmaline solid solution. Tschermaks
Mineralogische-Petrographische
Mitteilungen, 18, 273-286.
Donnay, G., Senftle, F.E., Thorpe, A., and White, S. (1967) Magnetic
properties oftourmaline. Carnegie Institution of Washington Year Book,
65, 295-299.
Drozhdin, S.N., Novik, V.K., Koptsik, VA., and Kobyakov, LR (1975)
Pyroelectric properties of tourmaline and cancrinite crystals in a wide
range of temperatures. Soviet I;'hysics, Solid State, 16,2122-2123.
Dunn, P.J. (1977) Uvite, a newly classified gem tourmaline. Journal of
Gemmology, 15,300-307.
Dunn, P.J., Appleman, D.E., and Nelen, J.E. (1977) Liddicoatite, a new
calcium end-member of the tourmaline group. American Mineralogist,
62, 1121-1124.
Fabel, G.W., and Henisch, H.K (1971) Technique for the measurement
of pyroelectric coefficients. Solid-State Electronics, 14, 1281-1283.
Foit, F.F., Jr. (1975) Aluminobuergerite, Na'_xA1,AI.B,S;.0270,_..(OH)1+~
a new end-member of the tourmaline group. Eos, 56, 461.
Foit, F.F., Jr., and Rosenberg, P.E. (1977) Coupled substitutions in the
tourmaline group. Contributions to Mineralogy and Petrology, 62, 109127.
Foit, F.F., Jr., Fuchs, Y., and Myers, P.E. (1989) Chemistry of alkalideficient schor!s from two tourmaline-dumortierite
deposits. American
Mineralogist,74,1317-1324.
Foord, E.E., and Mills, RA. (1978) Biaxiality in 'isometric' and 'dimetric'
crystals. American Mineralogist, 63, 316-325.
Fortier, S., and Donnay, G. (1975) Schor! refinement showing composition dependence of the tourmaline structure. Canadian Mineralogist,
13, 173-177.
Gallagher, V. (1988) Coupled substitutions in schorl-dravite tourmaline:
New evidence from SE Ireland. Mineralogical Magazine, 52, 637-650.
Gavrilova, N.D. (1965) Temperature dependence of pyroelectric coefficients as studied by a static method. Kritallografiya, 10, 346-350.
Gavrilova, N.D., Drozhdin, S.N., Novik, V.K., and Maksimov, E.G. (1983)
Relationship between the pyroelectric coefficient and the lattice dynamics of the pyroelectrics. Solid State Communications,
48, 129-133.
Gavrilova, N.D., Maksimov, E.G., Novik, V.K, and Drozhdin, S.N. (1989)
The low-temperature
behavior of the pyroelectric coefficient. Ferroelectrics, 100, 223-240.
Gladkii, V.V., and Zheludev, LS. (1965) Method and results of studying
the pyroelectric properties of certain single crystals. Soviet PhysicsCrystallography, 10, 50-53.
Grice, J.D., and Robinson, G.W. (1989) Feruvite, a new member of the
tourmaline group and its crystal structure. Canadian Mineralogist, 27,
199-203.
Grice, J.D., Ercit, T.S., and Hawthorne, F.e. (1993) Povondraite, a redefinition of the tourmaline ferridravite. American Mineralogist, 78,
433-436.
Hamid, S.A. (1980) Tourmaline as a pyroelectric infra-red radiation detector. Zeitschrift fUr Kristallographie, 151,67-75.
Hawkins, KD. (1993) Crystal chemical and structural analysis of pyroelectricity in tourmaline, 292 p. Ph.D. thesis, University of Queensland,
Brisbane, Queensland, Australia.
Hayashi, F. (1912) Observations ofpyroelectricity,
48 p. Thesis, Gottingen, Germany.
Helme, RG., and King, P.J. (1978) The elastic constants of iron tourmaline (schOr!). Journal of Material Science, 13, 1487-1489.
Henry, D.J., and Dutrow, RL. (1992) Tourmaline in a low grade clastic
metasedimentary
rock: An example of the petrographic potential of
tourmaline. Contributions to Mineralogy and Petrology, 112, 203-218.
Henry, DJ., and Guidotti, e.V. (1985) Tourmaline as a petrogenetic indicator mineral: An example from the staurolite-grade metapelites of
NW Maine. American Mineralogist, 70, 1-15.
Johnston, J.H., and Duncan, J.F. (1975) Manganese ion site distribution
studies in tourmaline by anomalous X-ray scattering methods. Journal
of Applied Crystallography, 8, 469-472.
Kassoli-Fournaraki,
A. (1990) Chemical variations in tourmalines from
pegmatite occurrences in Chalkidiki Peninsula, Northern Greece.
EFFECT IN TOURMALINE
Schweizerische Mineralogische und Petrographische Mitteilungen, 70,
55-65.
Lang, S.R (1974) Sourcebook ofpyroelectricity, 375 p. Gordon and Breach,
London.
Mason, R, Donnay, G., and Hardie, L.A. (\ 964) Ferric tourmaline from
Mexico. Science, 144,71-73.
Mathesius, J. (1564) Sarepta oder Bergpostill: Sampt der Jochimstalischen
Kurtschen Chroniken. Reprint, Navroni Technicke Muzeum Praha,
Prague, 1975.
Nuber, R, and Schmetzer, K (\ 979) Die Gitterposition des CrH im Turmalin: Strukturverfeinerung
eines Cr-reichen Mg-Al- Turmalins. Neues
Jahrbuch fUr Mineralogie Abhandlungen, 137, 184-197.
-(1981)
Strukturverfeinerung
yon Liddicoatit. Neues Jahrbuch fliT
Mineralogie Monatshefte, 5, 215-219.
-(1984)
Structural refinement oftsilaisite (manganese tourmaline).
Neues Jahrbuch fUr Mineralogie Monatshefte, 7, 301-304.
Nye, J.F. (1957) Physical properties of crystals: Their representation by
tensors and matrices, 322 p. Clarendon, Oxford, U.K
Reynard, R, Price, G.D., and Gillet, P. (1992) Thermodynamic
and anharmonic properties of forsterite, a-Mg2SiO,: Computer modeling versus high-pressure and high4emperature
measurements. Journal ofGeophysical Research, 97, 19791-19802.
Rontgen, W.e. (1914) Pyro- and piezoelectric investigations. Annalen der
Physik, 45, 737-800.
Rosenberg, P.E., and Foit, F.F., Jr. (1979) Synthesis and characterization
of alkali-free tourmaline. American Mineralogist, 64, 180-186.
Rosenberg, P.E., Foit, F.F., Jr., and Ekambaram, V. (\ 986) Synthesis and
characterization of tourmaline in the system Na20-Al20,-SiO,-B,O,H20. American Mineralogist, 71, 971-976.
Rossman, G.R., Fritsch, E., and Shigley, J.E. (1991) Origin of color in
cuprian elbaite from Siio Jose de Batalha, Paraiba, Brazil. American
Mineralogist, 76, 1479-1484.
Rumantseva, E.V. (1983) Chromdravite, a new mineral. Zapiski Vsesoyuznogo Mineralogicheskogo Obshchestva, 112,222-226.
Schmetzer, K, and Bank, H. (1984) Crystal chemistry oftsilaisite (manganese tourmaline) from Zambia. Neues Jahrbuch fliT Mineralogie
Monatshefte, 2, 61-69.
Schmetzer, K, Nuber, R, and Abraham, K (1979) Zur Kristallchemie
Magnesium-reicher
Turmaline. Neues Jahrbuch fUr Mineralogie Abhandlungen, 136,93-112.
Schneeberger, H. (1992) Pyroelektrische Eigenschaften nicht-ferroelektrischer Kristalle, 246 p. Ph.D. thesis, Ludwig-Maximilians-Universitat, Munich, Germany.
Shannon, R.D. (1976) Revised effective ionic radii and systematic studies
of interatomic distances in halides and chalcogenides. Acta Crystallographica, A32, 751-767.
Tatli, A. (1980) Ultrasonic investigation of the variation of the elastic
properties oftourmalines with structure, composition and temperature,
93 p. Ph.D. thesis, Middle East Technical University, Ankara, Turkey.
(1985) Zero-field elastic constants of uvite. Journal of Physics and
Chemistry of Solids, 46, 1015-1018.
Tatli, A., and Ozkan, H. (1987) Variation of the elastic constants of tourmaline with chemical composition. Physics and Chemistry of Minerals,
14,172-176.
Taylor, A.M., and Terrell, B.e. (1967) Synthetic tourmalines containing
elements of the first transition series. Journal of Crystal Growth, I,
238-244.
Taylor, RE., and Slack, J.F. (1984) Tourmalines from Appalachian-Caldeonian massive sulfide deposits: Textural, chemical and isotopic relationships. Economic Geology, 79, 1703-1726.
Taylor, J.R. (1982) An introduction to error analysis, 270 p. University
Science Books, Mill Valley, California.
Tsang, T., Thorpe, A.N., Donnay, G., and Senftle, F.E. (1971) Magnetic
susceptibility and triangular exchange coupling in the tourmaline minerai group. Journal of Physics and Chemistry of Solids, 32,1441-1448.
Tschermak, G. (1885) Lehrbuch der Mineralogie, 682 p. HOlder, Wien,
Austria.
Vernadsky, W. (1913) Uber die chemische Formel der Turmaline. Zeitschrift fliT Kristallographie, Kristallogeometri Kristallphysik, Kristall.
chemie, 53, 273-288.
Vorbach, A. (1989) Experimental examinations on the stability of syn
thetic tourmalines in temperatures from 250°C to 750 °C and pres
HAWKINS
ET AL.: PYROELECTRIC
sures up to 4kb. Neues Jahrbuch fUr Mineralogie Abhandlungen, 161,
69-83.
Walenta, K, and Dunn, PJ. (1979) Ferridravite, a new mineral of the
tourmaline group from Bolivia. American Mineralogist, 64, 945-948.
Walpole, R.E., and Myers, R.H. (1978) Probability and statistics for engineers and scientists (2nd edition), 580 p. Macmillan, New York.
EFFECT IN TOURMALINE
501
Zheludev, I.S. (1971) Physics of crystalline dielectrics. In Crystallography
and spontaneous polarization, vol. 1,336 p. Plenum, New York.
MANuSCRIYf
MANuSCRlYf
RECEfVED FEBRUARY 18, 1994
ACCEPTED JANUARY 12, 1995