American Mineralogist, Volume 86, pages 566–577, 2001

Displacive components of the low-temperature phase transitions in lawsonite

HINRICH-WILHELM MEYER,1,* STEFAN MARION,2 PETER SONDERGELD,3 MICHAEL A. CARPENTER,1

KEVIN S. KNIGHT,4 SIMON A.T. REDFERN,1 AND MARTIN T. DOVE1
1
Department of Earth Sciences, University of Cambridge, Downing Street, Cambridge CB2 3EQ, U.K.
2
University of Oslo, Center for Materials Science, N-0349 Oslo, Norway
3
Institut für Experimentalphysik, Universität Wien, Strudlhofgasse 4, A-1090 Wien, Austria
4
ISIS Facility, Rutherford Appleton Laboratory, Chilton, Didcot, Oxfordshire, OX11 OQX, U.K.

ABSTRACT
The phase transitions in deuterated lawsonite were investigated with high-resolution, time-of-
flight neutron diffraction between 2 and 500 K. From the analysis of spontaneous strain data, the
thermodynamics of the phase transition at 273 K are not changed by the deuteration process. Shifts
in atomic positions with temperature indicate continuous changes for a framework oxygen and for
one of the deuterium atoms, whereas for the other deuterium atom, a more discontinuous behavior
was observed in the average structure. Comparison of O···D and O···O bond lengths with IR data
from a non-deuterated lawsonite permits a detailed analysis of assignments of O-H stretching modes.

INTRODUCTION strument, operating at 20 kV and 3 nA. Elements with concen-

Lawsonite has attracted considerable interest over the last trations less than 2σ were excluded from further consideration,
few years both because of its importance as a marker of mod- and consequently only Ca, Al, Fe and Si were taken into ac-
erate P and low T conditions in nature (Pawley 1994; Pawley count. The average of four analyses, based on eight oxygen
et al. 1996; Schmidt 1995) and because it exhibits phase tran- atoms per atomic formula unit, is Ca 1.00, Al 1.95, Fe 0.05, Si
sitions arising from interactions between hydrogen bonding and 2.01, where the estimated uncertainties are ±0.01.
lattice relaxations (Libowitzky and Armbruster 1995; To provide a deuterated sample suitable for neutron diffrac-
Libowitzky and Rossman 1996; Lager et al. 1999; Meyer et al. tion, powdered lawsonite with an average grain size of ~77 µm
2000; Sondergeld et al. 2000; Martin-Olalla et al. 2001). was separated using heavy liquids. Approximately 1 g of this
Libowitzky and Armbruster (1995) reported two phase transi- powder was placed in a platinum sample holder, which was
tions, one at 273 ± 5 K, and the second below 155 K. They mounted in a closed, gas-tight furnace (details in Marion et al.,
described the room-temperature transition in terms of the de- in preparation). The deuteration was achieved by flowing dried
velopment of co-operative hydrogen bonds. The aim of the Ar gas at room temperature through a bottle of D2O and subse-
present powder neutron diffraction study was to address the quently through the hot zone of the furnace, where the exchange
following questions: (1) Does the exchange of deuterium for reaction took place. The experiments were performed at 425
hydrogen change the critical temperatures and the thermody- °C at a vapor pressure of approximately 3.2 kPa. The progress
namic character of the transitions? (2) How do the positions of of the deuteration reaction was followed via a thermo-balance.
the deuterium atoms evolve with temperature and how does From static proton NMR spectroscopy experiments the sample
their behavior correlate with the results from X-ray structure was 90.5 ± 2.5% deuterated.
determinations and IR experiments? (3) Is it possible to clarify Neutron powder diffraction data were collected using the
details of the mechanism of the transition at 273 K in terms of time-of-flight high-resolution diffractometer HRPD (Johnson
the contributions of proton ordering and framework displace- and David 1985, Ibberson et al. 1992) at the ISIS neutron spal-
ments? The data collected also allow some further analysis of lation source (Rutherford Appleton Laboratory). For the ex-
the relationships between bond lengths and vibrational frequen- periments, 3 grams of deuterated lawsonite were packed in an
cies for hydrogen bonding. aluminum can with thin vanadium front and back windows.
This and a 100 W cartridge heater were then mounted on a
METHODS candlestick sample holder. The complete sample set-up was
The lawsonite sample is from Valley Ford, Sonoma County, subsequently quenched in liquid nitrogen before being intro-
California (sample no. 120943 of the Harvard mineral collec- duced into a helium cryo-furnace, where it was cooled to 2 K.
tion). A sample from this batch was studied by Meyer et al. During cooling from 100 K the diffraction pattern was con-
(2000). The composition of the sample was determined by elec- stantly monitored for any signs of unwanted effects. None were
tron microprobe analysis (EMPA) on a Cameca Camebax in- observed. The data sets were focused to a Bragg angle of
168.329° and normalized to the incident monitor spectrum. The
background was subtracted and the data were corrected for
* E-mail: [email protected] detector efficiency using a diffraction pattern from vanadium.
0003-004X/01/0004–566$05.00 566
 MEYER ET AL.: PHASE CHANGES IN DEUTERATED LAWSONITE 567

All data sets were collected from 32000 µs to 230000 µs in TABLE 1. Refinement details for all temperatures
time channel bins of ∆t/t = 1 × 10–4, the time window corre- T (K) space parameters χ2 wRp Rp
sponding to a d-spacing of 0.66 to 2.49 Å in the back-scatter- group refined
2 P21cn 84 3.232 0.0556 0.0540
ing detector bank. The following data collection times were 50 P21cn 84 3.688 0.0388 0.0365
employed: 200 µAh for the data set at 2 K, 180 µAh for the 100 P21cn 84 3.435 0.0585 0.0552
data sets between 50 K and 285 K, and 100 µAh for data sets 130 Pmcn 61 3.475 0.0626 0.0588
175 Pmcn 61 4.348 0.0664 0.0619
collected at 300, 350 and 400 K. The data set at 500 K was 225 Pmcn 61 4.278 0.0656 0.0604
collected for about 45 minutes before the end of the allocated 268 Pmcn 61 5.981 0.0780 0.0690
experiment time. All data collections included a 15 minute in- 285 Cmcm 42 8.468 0.0941 0.0796
300 Cmcm 42 4.787 0.0971 0.0845
terval at each temperature prior to data collection to allow for 350 Cmcm 42 4.074 0.0948 0.0833
thermal equilibration. 400 Cmcm 42 3.848 0.0910 0.0797
500 Cmcm 42 1.457 0.1619 0.1438
Rietveld refinements were carried out using the program
GSAS (Larson and Von Dreele 1994). Starting values for lat-
tice parameters and atomic positions were taken from ure 2 with spontaneous strain data from X-ray powder diffrac-
Libowitzky and Armbruster (1995). The data set at 500 K was tion and dilatometry experiments [taken from Meyer et al.
sufficient for the refinement of lattice parameters only, but at (2000)]. In all three crystallographic directions, the strains from
all other temperatures, full structure refinements were per- neutron diffraction qualitatively describe the same effects as
formed. The data sets were refined with isotropic atomic dis- the strains from other experiments. The strains are small and
placement factors and convergence was rapidly achieved. differences in the data relate to both the limited resolution of
Refinement results of the atomic displacement factors for the the X-ray equipment and to uncertainties in defining the baseline
Al positions were non-positive and so estimates of the occu- in each experiment. The experimental uncertainty in the neu-
pancies for this site were derived from bond-length calcula- tron data propagated from the estimated standard deviation
tions. A small amount of Fe3+ (3% of the site occupancy of Al) values for the lattice parameters is ~0.01%, but the real uncer-
was introduced and gave reasonable results for the displace- tainty is probably somewhat greater.
ment parameters of the octahedral site. This amount is in good The square of the strains, ei2 in Figure 3, show a linear depen-
agreement with the electron microprobe analysis. Refinement dency below the transition temperature of 273 K (T1), which
with anisotropic atomic displacement parameters and split-sites indicates that the transition is being tricritical in character
for the deuterium positions did not result in crystallographi- (Q4 ∝ |Tc – T|). A linear fit to the data gives a critical temperature
cally meaningful values. Details of the refinement parameters of 273 K for e1, which is in excellent agreement with the result
are in Table 1, the refined lattice parameters are in Table 2, and from heat capacity measurements (Martin-Olalla et al. 2001).
refined atomic positions and isotropic atomic displacement The heat capacity measurements also indicated a tricritical be-
parameters are in Table 3. The variations of lattice parameters havior for the transition at T1 (Martin-Olalla et al. 2001). There
with temperature are shown in Figure 1. are insufficient data for a precise determination of the critical
temperature or a strain analysis for the second transition at T2
(<155 K).
RESULTS
Strain analysis
Atomic positions
To check if the H-D exchange had actually changed the
In Figure 4 Fourier maps, based on observed structure fac-
thermodynamic character of the transitions, spontaneous strains
tors, are displayed for sections parallel to (100) at tempera-
were calculated from the refined lattice parameters. These were
tures between 2 and 400 K. All maps show an equivalent part
compared to spontaneous strain data from non-deuterated
of the structure which contains a sequence of D2O-OD-OD-
samples (from Meyer et al. 2000). The spontaneous strains were
D2O highly affected by the temperature changes. An equiva-
calculated in the manner described by Carpenter et al. (1998)
lent part of the structure is shown as a polyhedral drawing in
using the following formulae for lattice constants and the unit
Figure 5. Figure 4, lower right, shows the high-symmetry struc-
cell volume:
ture at 400 K. Mirror planes, which relate the two deuterium
atoms of the D2O molecule to each other, are shown by dashed
a − a0 b − b0 c − c0 V −V
e1 = , e2 = , e3 = , and eVol = V 0 (1) lines. On cooling through the first phase transition at T1 = 273
a0 b0 c0 0
K this mirror-plane is lost and the D2O molecule starts to rotate
The baseline used to calculate the values of a0, b0, c0, and V0 in the plane of the map. This effect increases on further cool-
was taken from Meyer et al. (2000). It is of the form y = yc+yl · ing (center of Fig. 4). On further cooling, a second mirror-plane
θs · coth (θs/T) and describes plateau-like features in the lattice perpendicular to the first is lost and the Dwa atom leaves its
constants and the unit-cell volume (Fig. 1) at low temperatures place and moves out of the plane to give rise to the previously
and a linear behavior at high temperatures. θs is a material con- observed ferroelectric pattern (Sondergeld et al. 2000). The
stant and is related to the temperature at which the temperature maps in Figure 6, which are sections perpendicular to the maps
dependence changes from constant to linear (Salje et al. 1991a, in Figure 4 through the D2O-OD-OD-D2O chains, show the vio-
1991b). lations of the second mirror plane at 2 K and 100 K. The orien-
The results of the strain calculations are compared in Fig- tation of these sections is perpendicular to (100) and along the
568 MEYER ET AL.: PHASE CHANGES IN DEUTERATED LAWSONITE

T ABLE 2. Refined lattice parameters of lawsonite between TABLE 3b. Atomic positional parameters of lawsonite in space group
2 and 500 K Pmcn at 130, 175, 225 and 268 K
T (K) a (Å) b (Å) c (Å) V (Å3) Atom T (K) x y z Uiso*100 (Å2)
2 5.86465(2) 8.7652(3) 13.1002(5) 673.413(3) Ca 130 –1/4 0.0812(3) 0.2496(4) 0.59(6)
50 5.8639(2) 8.76442(4) 13.099(5) 673.208(3) 175 –1/4 0.0810(4) 0.2493(4) 0.65(7)
100 5.86332(2) 8.76659(4) 13.1021(5) 673.464(3) 225 –1/4 0.0812(4) 0.2492(5) 0.83(7)
130 5.86243(2) 8.76981(4) 13.1063(5) 673.825(3) 268 –1/4 0.0803(5) 0.2471(7) 0.92(9)
175 5.86092(2) 8.77586(4) 13.1132(6) 674.473(3)
Al 130 0 0 0 0.10(5)
225 5.85942(2) 8.78269(4) 13.1209(6) 675.223(4)
268 5.85539(3) 8.79152(5) 13.1326(7) 676.036(4) 175 0 0 0 0.16(6)
285 5.85424(3) 8.79471(6) 13.1365(8) 676.350(5) 225 0 0 0 0.26(6)
300 5.8543(4) 8.79679(7) 13.1387(9) 676.628(6) 268 0 0 0 0.40(8)
350 5.85603(3) 8.80228(7) 13.1442(9) 677.533(6) Ala 130 0 0 1/2 0.10(5)
400 5.85851(3) 8.80727(7) 13.149(8) 678.457(5) 175 0 0 1/2 0.16(6)
500 5.8642(5) 8.817(1) 13.159(1) 680.415(9) 225 0 0 1/2 0.26(6)
268 0 0 1/2 0.40(8)
Si 130 –1/4 0.7298(6) 0.1319(4) 0.29(5)
TABLE 3a. Atomic positional parameters of lawsonite in space group 175 –1/4 0.7291(7) 0.1324(4) 0.38(5)
P21cn at 2, 50, and 100 K 225 –1/4 0.7298(7) 0.1323(4) 0.39(6)
268 –1/4 0.731(1) 0.1312(6) 0.37(7)
Atom T (K) x y z Uiso*100 (Å2)
Sia 130 –1/4 0.7285(6) 0.3659(4) 0.29(5)
Ca 2 –0.266(3) 0.0812(3) 0.2511(3) 0.35(8)
175 –1/4 0.7292(7) 0.3664(4) 0.38(5)
50 –0.272(3) 0.0811(3) 0.2506(3) 0.26(8)
100 –0.269(3) 0.0813(3) 0.2503(4) 0.33(9) 225 –1/4 0.7291(7) 0.3660(5) 0.39(6)
Al 2 0 –0.000(2) –0.002(1) 0.15(5) 268 –1/4 0.729(1) 0.3664(6) 0.37(7)
50 0 0.000(2) –0.004(1) 0.13(6) O1 130 –1/4 -0.2019(3) 0.2487(3) 0.42(5)
100 0 0.000(2) –0.004(1) 0.17(5) 175 –1/4 -0.2020(3) 0.2488(4) 0.57(6)
Ala 2 0.002(3) –0.002(2) 0.498(1) 0.15(5) 225 –1/4 -0.2022(3) 0.2492(4) 0.61(6)
50 –0.006(4) –0.002(2) 0.497(1) 0.13(6) 268 –1/4 -0.2034(4) 0.2510(6) 0.68(8)
100 0.005(4) –0.002(2) 0.505(1) 0.17(5) O2 130 0.0243(5) 0.1226(3) 0.1207(2) 0.40(3)
Si 2 –0.248(3) 0.731(5) 0.133(3) 0.21(4) 175 0.0235(6) 0.1231(3) 0.1202(2) 0.45(3)
50 –0.253(4) 0.730(5) 0.132(4) 0.11(5)
225 0.0224(7) 0.1242(4) 0.1195(2) 0.49(3)
100 –0.252(4) 0.730(5) 0.132(4) 0.18(5)
Sia 2 –0.250(3) 0.727(5) 0.366(3) 0.21(4) 268 0.021(1) 0.1270(6) 0.1179(3) 0.50(4)
50 –0.254(4) 0.727(5) 0.366(4) 0.11(5) O2a 130 0.0230(5) 0.1230(3) 0.3848(2) 0.40(3)
100 –0.249(4) 0.727(5) 0.366(4) 0.18(5) 175 0.0230(6) 0.1305(3) 0.3847(2) 0.45(3)
O1 2 –0.239(2) –0.2016(2) 0.2485(3) 0.30(5) 225 0.0234(7) 0.1304(4) 0.3845(2) 0.49(3)
50 –0.247(3) –0.2016(3) 0.2485(3) 0.44(6) 268 0.024(1) 0.1289(5) 0.3842(3) 0.50(4)
100 –0.244(3) –0.20172(3) 0.2485(3) 0.46(6) O3 130 –1/4 -0.1139(4) 0.0628(3) 0.23(4)
O2 2 0.024(2) 0.1185(8) 0.1250(5) 0.27(3) 175 –1/4 -0.1137(5) 0.0636(3) 0.27(4)
50 0.017(3) 0.1185(8) 0.1248(5) 0.28(3) 225 –1/4 -0.1138(5) 0.0638(3) 0.39(4)
100 0.018(3) 0.1184(9) 0.1218(6) 0.37(3)
268 –1/4 -0.1135(8) 0.0661(4) 0.34(5)
O2a 2 0.021(3) 0.1261(8) 0.3819(5) 0.27(3)
50 0.013(3) 0.1255(9) 0.3819(5) 0.28(3) O3a 130 –1/4 -0.1136(4) 0.4333(2) 0.23(4)
100 0.015(3) 0.127(1) 0.3862(6) 0.37(3) 175 –1/4 -0.1137(5) 0.4332(3) 0.27(4)
O2b 2 0.474(3) 0.3674(8) 0.8870(5) 0.27(3) 225 –1/4 -0.1129(5) 0.4330(3) 0.39(4)
50 0.466(3) 0.3664(8) 0.8869(5) 0.28(3) 268 –1/4 -0.1116(8) 0.4333(4) 0.34(5)
100 0.467(3) 0.3675(9) 0.8824(6) 0.37(3) O4 130 –1/4 0.3925(4) 0.0514(2) 0.29(4)
O2c 2 0.475(2) 0.3740(8) 0.6170(5) 0.27(3) 175 –1/4 0.3934(5) 0.0512(3) 0.44(5)
50 0.468(2) 0.3741(9) 0.6170(5) 0.28(3) 225 –1/4 0.3943(5) 0.0515(3) 0.49(5)
100 0.470(3) 0.374(1) 0.6197(6) 0.37(3) 268 –1/4 0.3936(8) 0.0516(5) 0.76(6)
O3 2 –0.253(3) –0.1130(4) 0.0636(2) 0.15(3)
O4a 130 –1/4 0.3812(4) 0.4612(3) 0.29(4)
50 –0.261(3) –0.1132(4) 0.0635(3) 0.14(4)
100 –0.257(3) –0.1135(4) 0.0637(3) 0.16(4) 175 –1/4 0.3822(5) 0.4601(3) 0.44(5)
O3a 2 –0.252(3) –0.1155(4) 0.4339(2) 0.15(3) 225 –1/4 0.3835(5) 0.4594(3) 0.49(5)
50 –0.257(3) –0.1154(4) 0.4337(2) 0.14(4) 268 –1/4 0.3868(8) 0.4577(4) 0.76(6)
100 –0.256(3) –0.1147(4) 0.4341(2) 0.16(4) O5 130 –1/4 0.3575(4) 0.2482(4) 1.21(6)
O4 2 –0.258(3) 0.3916(4) 0.0513(2) 0.16(4) 175 –1/4 0.3580(4) 0.2484(5) 1.42(7)
50 –0.264(3) 0.3912(4) 0.0512(2) 0.22(4) 225 –1/4 0.3586(4) 0.2488(5) 1.83(8)
100 –0.263(3) 0.3919(4) 0.0513(2) 0.26(4) 268 –1/4 0.3599(6) 0.249(1) 2.2(1)
O4a 2 –0.247(3) 0.3806(4) 0.4613(2) 0.16(4) Dw 130 –1/4 0.3841(5) 0.1723(3) 2.80(9)
50 –0.255(3) 0.3806(4) 0.4609(2) 0.22(4)
175 –1/4 0.3875(6) 0.1734(4) 3.2(1)
100 –0.255(3) 0.3806(4) 0.4609(2) 0.26(4)
O5 2 –0.234(2) 0.3589(3) 0.2479(3) 0.41(7) 225 –1/4 0.3912(6) 0.1746(4) 3.8(1)
50 –0.239(3) 0.3592(3) 0.2482(4) 0.39(7) 268 –1/4 0.407(1) 0.1805(8) 5.7(3)
100 –0.241(3) 0.3590(4) 0.2485(4) 0.72(7) Dwa 130 –1/4 0.4379(6) 0.2878(4) 5.9(1)
Dw 2 –0.250(3) 0.3826(4) 0.1721(3) 2.26(7) 175 –1/4 0.4371(7) 0.2902(4) 5.8(2)
50 –0.259(3) 0.3827(5) 0.1726(3) 2.26(8) 225 –1/4 0.4337(7) 0.2925(4) 6.0(2)
100 –0.260(3) 0.3842(5) 0.1726(3) 2.43(9) 268 –1/4 0.420(1) 0.3015(8) 6.4(3)
Dwa 2 0.699(2) 0.4402(5) 0.2841(3) 3.5(1) Dd 130 –1/4 0.2797(4) 0.0313(3) 2.13(8)
50 0.694(3) 0.4408(5) 0.2843(3) 3.2(1) 175 –1/4 0.2796(5) 0.0329(3) 2.38(9)
100 0.701(3) 0.4410(5) 0.2851(3) 3.6(1)
225 –1/4 0.2794(5) 0.0350(3) 2.6(1)
Dh 2 –0.250(3) 0.2797(4) 0.0290(2) 1.93(7)
50 –0.256(3) 0.2792(4) 0.0288(3) 2.17(9) 268 –1/4 0.2781(7) 0.0383(5) 3.6(2)
100 –0.254(3) 0.2797(4) 0.0295(3) 2.16(8) Dda 130 –1/4 0.3189(5) 0.4042(3) 3.1(1)
Dha 2 –0.234(3) 0.3227(4) 0.4014(3) 2.8(1) 175 –1/4 0.3158(6) 0.4063(4) 3.8(1)
50 –0.243(3) 0.3225(4) 0.4017(3) 2.9(1) 225 –1/4 0.3132(6) 0.4092(4) 4.3(1)
100 –0.240(3) 0.3218(4) 0.4024(3) 3.0(1) 268 –1/4 0.307(1) 0.4178(5) 5.2(2)
 MEYER ET AL.: PHASE CHANGES IN DEUTERATED LAWSONITE 569

TABLE 3c. Atomic positional parameters of lawsonite in space group 5.865

Cmcm at 285, 300, 350, and 400 K
Atom T (K) x y z Uiso*100 (Å2) 5.860
Ca 285 –1/4 0.0804(6) 1/4 1.1(1)

a (Å)
300 –1/4 0.0807 (6) 1/4 1.2(1) 5.855
350 –1/4 0.0805(6) 1/4 1.4(1)
400 –1/4 0.0803(6) 1/4 1.4(1) 5.850
Al 285 0 1/4 0 0.5(1)
300 0 1/4 0 0.6(1) 5.845
350 0 1/4 0 0.7(1)
400 0 1/4 0 0.7(1) 8.815
Si 285 –1/4 0.7311(4) 0.1330(3) 0.42(8) 8.805
300 –1/4 0. 7311(4) 0.1331(3) 0.45(8)
350 –1/4 0. 7311(4) 0.1331(3) 0.49(8) 8.795

b (Å)
400 –1/4 0. 7309(4) 0.1330(3) 0.52(8)
8.785
O1 285 –1/4 –0.2028(4) 1/4 0.64(9)
300 –1/4 –0.2030(5) 1/4 0.66(9) 8.775
350 –1/4 –0.2033(5) 1/4 0.71(9)
400 –1/4 –0.2036(5) 1/4 0.89(9) 8.765
O2 285 0.0225(3) 0.1283(2) 0.1166(1) 0.61(4)
300 0.0226(3) 0.1281(2) 0.1167(1) 0.63(5) 13.16
350 0.0223(3) 0.1290(2) 0.1164(1) 0.74(5)
400 0.0225(3) 0.1288(2) 0.1165(1) 0.75(5) 13.14

c (Å)
O3 285 –1/4 –0.1122(3) 0.0675(2) 0.09(6)
300 –1/4 –0.1123(3) 0.0675(2) 0.13(6)
350 –1/4 –0.1125(3) 0.0677(2) 0.20(6) 13.12
400 –1/4 –0.1123(3) 0.0678(2) 0.27(6)
O4 285 –1/4 0.3910(3) 0.0480(2) 1.28(7) 13.10
300 –1/4 0. 3913(3) 0.0482(2) 1.28(8)
350 –1/4 0. 3915(3) 0.0482(2) 1.42(8)
400 –1/4 0. 3916(3) 0.0481(2) 1.37(8) 681.0
O5 285 –1/4 0.3606(6) 1/4 1.7(1)
300 –1/4 0. 3602(7) 1/4 1.7(1)
678.0
V (Å)

350 –1/4 0. 3615(7) 1/4 2.1(1)

400 –1/4 0. 3607(7) 1/4 2.2(1)
Dw 285 –1/4 0.4153(6) 0.1886(3) 5.9(1) 675.0
300 –1/4 0. 4160(6) 0.1886(4) 6.0(1)
350 –1/4 0. 4160(6) 0.1886(4) 6.2(1)
400 –1/4 0. 4161(6) 0.1889(4) 6.5(1) 672.0
Dd 285 –1/4 0.2890(6) 0.0551(4) 7.3(1) 0 100 200 300 400 500
300 –1/4 0. 2889(6) 0.0554(4) 7.2(1)
350 –1/4 0. 2896(6) 0.0556(4) 7.6(1) T (K)
400 –1/4 0. 2893(6) 0.0557(4) 7.5(1)

FIGURE 1. Lattice constants of lawsonite between 2 and 500 K

from neutron diffraction. The coth-function used as a baseline in
direction indicated by the straight line in the 285 K plot in Fig- spontaneous strain calculations is also shown in the figures. Estimated
ure 4. They do not quite cut through the central positions of the standard deviations are smaller than the symbol size.
D atoms, but they do show the shifts below T2. Above T2 Dwa
and Dw oscillate in a direction perpendicular to the chain, while
the deuterium atoms which form OD bonds oscillate in the plane
of the chain. Above T1 these oscillations become stronger (bot-
tom of the Fig. 6).
and almost doubles its magnitude. (δO2)4 is shown in the lower
A displacive order parameter? part of Figure 7 and the linear relationship implies that δO2 is
To quantify the shifts in atomic positions with tempera- scaling linearly with the order parameter for the temperature
ture, a quantity δ was calculated as the modulus of the vector range between T1 and T2. The displacements of other oxygen
difference between atomic positions at two temperatures: atoms are all much smaller. The variation of δDd (Fig. 8) has
an apparent discontinuity between 268 K and 285 K, but shows
no influence by the transition at T2. The magnitude of δDda is
. (2) almost twice that of δDd, whereas (δDd)4 and (δDda)4 both
vary approximately linearly with temperature (lower part of
The index 400 refers to the 400 K data set, which was taken as Fig. 8). The thermal evolution of δDwa (Fig. 9) is quite similar
reference state and the index T to the data at any temperature. to that of δO2, with a large displacement at T2, whereas the
Thus, δ is a measure of the shift of each atom from its average variation δDw is more like that of δDd and δDda.
position in the parent Cmcm structure. The δ atoms O2 (Fig.
7), Dd and Dda (Fig. 8) and Dw and Dwa (Fig. 9) show the O-D, O···D, O-O distances
most pronounced effects. The deviation from zero for the O2 Figure 10 schematically shows the four different D posi-
atom, δO2, is only small from high temperatures down to 285 tions with coordinating oxygen atoms in the P21cn structure of
K, when it starts to increase smoothly. At ~T2 it shows a jump lawsonite. Within a cut-off distance of 2.6 Å, four oxygen at-
570 MEYER ET AL.: PHASE CHANGES IN DEUTERATED LAWSONITE

4.0 1.2

3.0
0.9
e1 (x 10 )
-3

2.0

e1 ( x 10 )
-5
0.6
1.0

2
0.0 0.3

1.0
0.0

0.0
e2 (x 10 )

6.0
-3

-1.0
4.0

e2 ( x 10 )
-6
-2.0

2.0

2
0.0
e3 (x 10 )
-3

-1.0 0.0

-2.0
4.0
-3.0
e3 ( x 10 )
-6

2
2.0
eVol (x 10 )
-3

1
2

-1 0.0

-2
0 100 200 300 400 500
0 100 200 300 400 500
T (K)
T (K)

F IGURE 2. Spontaneous strain for lawsonite from neutron FIGURE 3. Thermal evolution of the square of the spontaneous
diffraction (squares), X-ray diffraction (plus signs and crosses) and strains, ei2, indicating tricritical behavior below T1 (=273 K).
dilatometry data (dots). The error bar in each plot is the estimated
standard deviation of the X-ray experiments. Estimated standard
deviations for the neutron data are smaller than the symbol size.

oms are nearest neighbors and nine oxygen atoms are next- whereas both the other atoms deviate quite clearly from ideal
nearest-neighbors of the deuterium atoms. D-O distances with a separation of 2.217(8) Å and 2.400(5) Å.
Two oxygen atoms coordinate Dw: O5 is at a distance of Figure 11 and Table 4 show the temperature variations of
1.019(4) Å away and O4 is located 1.586(4) Å away. Both dis- the D-O distances of Figure 10. Above T1 the D atoms are con-
tances are near the ideal bond lengths for hydrogen bonds strained by symmetry to two independent positions (Dw and
(Libowitzky 1999). The Dd atom is shown with four oxygen Dd), whereas in the two low temperature space groups each of
neighbors, the O4 atom having a distance of 1.024(4) Å from these splits into two (Dw, Dwa, Dd, Dda). Three of the four
Dd. Of the other three oxygen atoms, O4a is at an almost typi- short O-D distances (Dw-O5, Dd-O4 and Dda-O4a) seem not
cal distance of 1.662(4) Å, while O2 and O2c occupy positions to be affected by the symmetry change at T2, which is in con-
2.484(9) Å and 2.398(9) Å, respectively, away from Dd. For trast to the behavior of the Dwa-O5 distance (Table 4a).
Dda four neighboring oxygen atoms are observed; O4a, at a The remaining nine D-O distances show different tempera-
distance of 0.939(4) Å, is the nearest neighbor. The distance to ture dependencies (Table 4b). The Dw-O4 distance at about
the next-nearest atom (O5) is 2.035(6) Å and the atoms O2a 1.6 Å increases slowly over a large temperature range and seems
and O2b are observed at comparatively large distances of to significantly change this behavior on heating only in the vi-
2.294(9) and 2.398(9) Å, respectively. Around the Dwa atom, cinity of T1. A similar temperature dependency is shown by
three coordinating oxygen atoms are observed within a dis- Dd-O4a, but with slightly higher values for the bond length.
tance of 2.6 Å. O5 forms the nearest neighbor at 0.942(5) Å, The other two Dd-O distances, Dd-O2 and Dd-O2c at ~2.4 Å
 MEYER ET AL.: PHASE CHANGES IN DEUTERATED LAWSONITE 571

2K Hwa
100 K
Ca
O5
Hw
Si Si
Hh O4

O4a
Hha

Hwa
O5 Ca
Hw

175 K 268 K

285 K 400 K

FIGURE 5. Polyhedral drawing of a part of the P21cn-lawsonite

structure in the same orientation as in Figure 4. The arrangement of
hydrogen atoms can be followed from two OH at the bottom of the
plot via the H2O molecule to the next two OH molecules (reproduced
with kind permission of E. Libowitzky). The numbers refer to the
deviation of the atoms from the lost (100) mirror-plane (x=0) in
FIGURE 4. Fourier maps (Fobs) of the lawsonite structure containing
Libowitzky’s structure model.
a D2O-OD-OD-D2O chain. Dashed lines equal mirror planes in the
high symmetry space group. The dashed inclined line in the 2 K and
285 K plots indicates the direction of the second cut through the nuclear
density shown in Figure 6. The c-axis points upwards, b to the right.
Contour intervals are 0.5 fm/Å3 between 0.50 and 6.

Å, respectively. The longest separation is the O-O distance O5-

merge above T2 and subsequently decrease slightly. Dda-O5 is Dw···O2 at 3.2 Å.
at ~2.0 Å at low temperatures and increases to ~2.2 Å at T1,
where it increases abruptly to 2.6 Å. Two other Dda-O dis- DISCUSSION
tances below 2.6 Å, between Dda and O2b and O2a, merge One of our objectives was to determine how far the deute-
above T2 and exhibit a negligible variation over the tempera- rium-hydrogen exchange affected the critical temperatures and
ture range. Dwa-O2 and Dwa-O4a show contrasting behavior the thermodynamic character of the transitions. Sondergeld et
to each other. Whereas Dwa-O2 increases in steps from 2.2 to al. (2000); Martin-Olalla et al. (2001); and Meyer et al. (2000)
2.5 Å, Dwa-O4a decreases with temperature. showed that the transition at T1=273 K in a non-deuterated
The O-D···O distances (Fig. 12) were calculated for the D- sample can be described by a tricritical model. From Figure 2,
O environments in Figure 10. They cover a range between 2.6 the spontaneous strains of a deuterated sample are qualitatively
and 3.2 Å. The two shortest distances (at low temperatures) are the same as those of non-deuterated samples. That the deutera-
O5-Dw···O4 and O4-Dda···O4a. The next longest distance is tion process does not change the thermodynamic character is
shown for O4a-Dda···O5 and O5-Dwa···O4a, which plot on top demonstrated in Figure 3, which shows that the squares of the
of each other. Above T1, O5-Dwa···O4a becomes O5-Dw···O4, spontaneous strains, ei2, vary linearly with temperature. The
and is therefore not plotted above this temperature. transition at T1 in both deuterated and non-deuterated lawsonite
Around 2.9 Å, the O4a-Dda···O2a and O4a-Dda···O2b dis- is thus tricritical within experimental uncertainty. The critical
tances merge above T2. O4-Dd···O2 and O4-Dd···O2c are also temperature of T1=273 K determined by a linear fit to the data
separate below T2 only. They were calculated at 2.95 and 3.1 (Fig. 3) is in excellent agreement with the transition tempera-
572 MEYER ET AL.: PHASE CHANGES IN DEUTERATED LAWSONITE

2K 100 K
O5 O5

Hw Hh Hha Hw
Hwa

175 K 268 K

285 K 400 K

FIGURE 6. The nuclear density maps for the second orientation indicated on Figure 4 by a dashed inclined line. Temperature-induced
increases of atomic displacements are shown. The sequence indicates the development of the ferroelectric pattern below the second transition,
(the top row).
ture determined by specific heat measurements on a non-deu-
14 terated sample (Martin-Olalla et al. 2001). Thus, the transition
mechanism is not determined by the dynamics of the H or D
12 motions. This contrasts with, for example, TlH 2PO 4 and
δO2
10 δO2new TlD2PO4, in which the transition temperature changes from 230
K to 353 K due to the substitution of D for H (Rios et al. 1998).
δO2 (x 10 )
-3

8 For lawsonite the behavior of the spontaneous strain below

T1 can be described with a Landau model, which implies that a
6 description of the phase transition on an atomic scale must in-
volve displacive components as opposed to a model that de-
4
pends only on the configurational effects of atomic ordering.
2 This issue has been discussed in detail (Carpenter 1992; Car-
penter et al. 1994; Martin-Olalla et al. 2001). Shifts of atomic
0 positions of the type shown in Figures 7, 8, 9 can be used to
65
4
(δO2) off scale T < 130 K
look for such components. δO2, δDw, and δDwa seem to vary
continuously at T1, whereas the thermal evolution of δO2 and
55 (δO2)
4
δDwa at T2 changes more abruptly. More abrupt changes exist
(δO2new)
4
in δDd and δDda (Fig. 8) at T1, but the shifts still roughly scale
)
-10

45 as δ4~(T – Tc). These shifts of atomic positions describe the

(δO2) (x 10

35
4

25
▼

15
FIGURE 7. (Top) The atomic shift of the O2 atom for refinements in
5 the highest possible space group for each temperature interval (δO2, full
symbol), and in the reduced-symmetry space group (δO2new, open symbol).
0 100 200 300 400 500 Both show a continuous behavior below T1 and a discontinuity at T2.
T (K) (Bottom) (δO2)4 (left axis) and (δO2new)4 (right axis) both display linear
temperature dependency below T1. The line is a guide to the eye.
 MEYER ET AL.: PHASE CHANGES IN DEUTERATED LAWSONITE 573

60 70

60
50
δ Dd δ Dw
δ Ddnew 50
δ Dwnew

δ Dw(a) (x 10 )
40
δ Dd(a) (x 10 )

-3
-3

δ Dda δ Dwa
δ Ddanew 40
δ Dwanew
30
30
20
20

10
10

0 0
10 (δ Dd)
4 4
7 (δ Dwanew) off scale T < 130 K
4
(δ Ddnew) (δ Dw)
4

8 4 4
(δ Dda) (δ Dwnew)

(δ Dda) (x 10 )

(δ Dw(a)) (x 10 )
4
(δ Dd) (x 10 )

-6
-6

(δ Ddanew) 5 4
(δ Dwa)
6 (δ Dwanew)
4
4

4
4

4 3
-7

2
1

0
0 100 200 300 400 500 0
0 100 200 300 400 500
T (K)
T (K)

FIGURE 8. Atomic shifts of the Dd deuterium atoms as a function FIGURE 9. (Top) Τhe thermal evolution of the atomic shift of the
of temperature. (Top) The atomic shift of the Dd deuterium atoms as a Dw deuterium atoms gives evidence of continuous changes for the
function of temperature shows a discontinuity at T1 for both Dd and temperature range down to T2. There the data for Dwa and Dwanew are
Dda, but a more continuous temperature dependence as a result of the both highly affected by the transition, Dw and Dwnew seem rather
second refinement process (δDdnew, δDdanew). (Bottom) The behavior unimpressed. (Bottom) (δDw)4 and (δDwa)4 both display a temperature
of (δDda)4 maybe interpreted as continuous contrast to (δDd)4, which dependence close to linear. The lines are drawn in as guides to the eye.
clearly is non-continuous. The changes in (δDdnew)4 and (δDdanew)4
are indeed too small for any statement. Left axis: (δDd)4, right axis: all
other symbols.

O2c
O2b
2.40 2.40 FIGURE 10. Schematic drawing of
the deuterium positions and some
1.66 2.04 O5 neighboring oxygen atoms as observed
Dd O4a Dda
O4 in the P21cn structure of lawsonite at 2
O4a
1.02 0.94 K. The cut-off distance for the graph was
chosen at 2.6 Å.
2.48 2.29

O2 O2a

1.59 O4 2.22 O2
Dw Dwa
O5 O5
0.94
1.02

2.40
O4a
574 MEYER ET AL.: PHASE CHANGES IN DEUTERATED LAWSONITE

1.4
d(Dwa···O4a) d(Dw···O4)
2.8 d(Dwa···O2) d(Dd···O4a)
d(Dd···O2) d(Dda···O5)
2.6 d(Dd···O2c) d(Dda···O2b)
d(Dda···O2a) 1.3
2.4

2.2
1.2
distance D···O (Å)

distance D-O (Å)

2.0

1.8
1.1

1.6

d(Dw-O5)
1.4 1.0
d(Dd-O4)
d(Dda-O4a)
1.2 d(Dwa-O5)

1.0 0.9

0.8
0 50 100 150 200 250 300 350 400

T (K)
FIGURE 11. D-O distances for the D-O pairs of Figure 10. The right axis pertains to the four O-D distances with a length of ~1 Å at the
bottom of the figure. The other bond lengths (left axis) are next-nearest-neighbor distances calculated for the deuterium-oxygen pairs of Figure
10 separated more than 1.6 Å from each other.

TABLE 4a. Thermal evolution of the D-O distances from Figure 11 TABLE 4b. Thermal evolution of the D-O distances from Figure 11
T (K) Dw-O5 (Å) Dwa-O5 (Å) Dd-O4 (Å) Dda-O4a (Å) T (K) Dwa-O4a (Å) Dw-O4 (Å) Dd-O4a (Å) Dd-O2 (Å)
2 1.019(4) 0.942(5) 1.024(4) 0.939(4) 2 2.400(5) 1.586(4) 1.662(4) 2.484(9)
50 1.018(4) 0.944(5) 1.025(5) 0.930(4) 50 2.392(5) 1.592(5) 1.660(5) 2.48(1)
100 1.025(5) 0.928(5) 1.026(5) 0.929(4) 100 2.378(5) 1.591(5) 1.668(5) 2.45(1)
130 1.021(5) 0.875(5) 1.024(5) 0.925(4) 130 2.326(5) 1.587(5) 1.684(5) 2.420(4)
175 1.017(5) 0.884(5) 1.027(5) 0.915(5) 175 2.280(6) 1.603(6) 1.712(6) 2.401(4)
225 1.014(6) 0.874(6) 1.032(6) 0.903(5) 225 2.234(6) 1.616(6) 1.741(6) 2.374(4)
268 0.983(9) 0.874(9) 1.030(9) 0.875(8) 268 2.07(1) 1.70(1) 1.794(8) 2.320(7)
285 0.939(4) 0.939(4) 0.902(5) 0.902(5) 285 1.860(5) 1.860(5) 2.083(6) 2.280(5)
300 0.944(5) 0.944(5) 0.905(6) 0.905(6) 300 1.858(6) 1.858(6) 2.089(6) 2.279(4)
350 0.939(5) 0.939(5) 0.902(6) 0.902(6) 350 1.858(5) 1.858(5) 2.098(6) 2.276(5)
400 0.940(5) 0.940(5) 0.907(6) 0.907(6) 400 1.863(5) 1.863(5) 2.098(6) 2.278(5)

TABLE 4C. Thermal evolution of the D-O distances from Figure 11

T (K) Dwa-O2 (Å) Dda-O5 (Å) Dda-O2a (Å) Dda-O2b (Å) Dd-O2c (Å)
2 2.217(8) 2.035(6) 2.294(9) 2.398(9) 2.398(9)
50 2.219(9) 2.036(7) 2.30(1) 2.39(1) 2.40(1)
100 2.25(1) 2.042(7) 2.28(1) 2.40(1) 2.41(1)
130 2.411(5) 2.073(7) 2.318(4) 2.318(4) 2.420(4)
175 2.410(6) 2.102(8) 2.299(5) 2.299(5) 2.401(4)
225 2.431(6) 2.143(9) 2.291(5) 2.291(5) 2.374(4)
268 2.49(1) 2.27(2) 2.286(9) 2.286(9) 2.320(7)
285 2.485(5) 2.637(6) 2.280(5) 2.280(5) 2.280(5)
300 2.480(5) 2.632(6) 2.279(4) 2.279(4) 2.279(4)
350 2.488(5) 2.632(6) 2.276(5) 2.276(5) 2.276(5)
400 2.488(6) 2.631(6) 2.278(5) 2.278(5) 2.278(5)
 MEYER ET AL.: PHASE CHANGES IN DEUTERATED LAWSONITE 575

3.3

3.2 d(O5-Dwa···O2)
d(O5-Dwa···O4a)
d(O5-Dw···O4)
d(O4a-Dda···O2a)
3.1 d(O4a-Dda···O2b)
d(O4a-Dda···O5)
d(O4-Dd···O2)
3.0 d(O4-Dd···O2c)
distance O-O (Å)

d(O4-Dd···O4a)

2.9

2.8

2.7

2.6

2.5
0 50 100 150 200 250 300 350 400

T (K)
FIGURE 12. O-D···O distances vs. temperature, calculated from O···O pairs from Figure 11. The O5-Dwa···Ox bonds do not correlate to the
temperature dependence of the relevant IR mode.

average structure only, however, and do not discriminate be- The new structural data can be used to check details of the
tween displacive and order-disorder models. This is because assignments given by Libowitzky and Rossman (1996) for the
the average positions of atoms partially disordered between four O-H stretching modes observed in infrared spectra in the range
split sites can appear to be somewhere in between the actual ~2800–3600 cm–1 by comparing the O-H frequencies with D-O
sites. The data above T1 were also refined in space group Pmcn and O…O distances. In Figure 13 data for the four O-H stretching
to test whether the sites in the Cmcm structure might be split. modes are reproduced from Meyer et al. (2000) at ~2800 cm–1
In the range between T1 and T2 it was unsuccessfully attempted (bottom), 3200 cm–1 (center), 3550 and 3620 cm–1 (top). At tem-
to refine the structures in space group P21cn. The results of peratures above 260 K the high frequency peaks become indis-
these refinements in Pmcn (Figs. 7 to 9) essentially confirm tinguishable. Frequencies for three of these peaks (at 2800,
the previous findings for O2, Dw and Dwa. The choice of re- 3200, and 3550 cm–1) increase with increasing temperature,
finement procedure does affect whether or not a discontinuity whereas the fourth peak at 3620 cm–1 shows the opposite trend.
for Dd and Dda is observed. These refinements may not eluci- As reviewed by Libowitzky (1999), O-H stretching frequen-
date the transition mechanism but they do confirm that the cies correlate with O···H and O···O distances. The D-O, D···O
framework participates significantly in the phase transitions. It and O···O distances are again only for the average structure but
is also possible that the changes in atomic positions at T2 are they are much greater than the shifts of the atomic positions, δ,
more abrupt than those at T1, but this remains a highly tenta- and are not significantly affected by considerations of the av-
tive conclusion. From their infrared study, Libowitzky and erage versus split positions of the D atoms.
Rossman (1996) concluded that the protons in the Cmcm struc- First, there is an approximate correlation between the O-H
ture are dynamically disordered between split positions and stretching frequency and the O-D distances in that a short O-D
then order at the phase transitions. An equivalent method for distance corresponds with a high O-H frequency, and vice versa.
investigating local structural behavior is needed to test whether Below T1 the O-D distances (Fig. 11) are grouped in two pairs
the O2 atoms, in particular, of the framework behave in the while the O-H frequencies are spread into four discrete ranges
same way. (Fig. 13), however, which suggests that other factors must also
576 MEYER ET AL.: PHASE CHANGES IN DEUTERATED LAWSONITE

has a bond length of 2.03 Å at 2 K and increases to ~2.2 Å at

T2. Again, the increasing D···O distances correlates with in-
3620 creasing O-H stretching frequency
The remaining mode at 3620 cm–1 is assigned to O5-Hwa
stretching and the corresponding D···O distance in O5-Dwa···O2
ω (cm )

is 2.21 Å at 2 K. There is also a D···O distance of 2.4 Å at 2 K

-1

3580 in O5-Dwa···O4a, however. With increasing temperature these

two D···O distances change in such a way that at T > T2 the
Dwa···O4a distance is smaller than the Dwa···O2 distance. In
other words the hydrogen-bonding configuration changes at
3540
the Pmcn ↔ P21cn transition point. The trend of decreasing
D···O distance with increasing temperature then correlates with
3280 the trend of decreasing O-H stretching frequency.
In Figure 14 the frequencies of the O-H stretching modes
3240
are plotted versus the four next-nearest neighbor D···O dis-
ω (cm )
-1

tances, as discussed above. The function derived by Libowitzky

(1999) from data for many minerals
3200
ν(cm–1) = 3632 – 1.79 × 10 6*exp[–d(H···O)/0.2146 (cm–1)]
(3)
3160
is shown for comparison. The overall agreement between
2980 distances and frequencies is consistent with these detailed as-
signments, allowing that O-D stretching frequencies are system-
atically lower than O-H stretching frequencies. The high
ω (cm )
-1

frequency stretching mode data for 300 K and above, plot sig-
2900
nificantly off the expected trend and reinforce the conclusion of
Libowitzky and Rossman (1996) that the protons are disordered
between discrete sites in the Cmcm structure rather than being
2820 located at their high symmetry positions. Libowitzky (1999) also
found an equivalent correlation between O···O distances and O-
0 100 200 300 400 500 H frequencies, but this is not observed for the O···O distances
T (K) found here. This presumably reflects the relatively complex dis-
FIGURE 13. Temperature dependence of the peak positions of the tribution of oxygens around the Dwa and Dda atoms in lawsonite.
four OH stretching modes in lawsonite as observed in IR experiments Whereas the mechanism for the Cmcm ↔ Pmcn transition
(Meyer et al. 2000). The two low frequency peaks at ~2800 cm–1 and
3200 cm–1 could not be detected above 290 K and 270 K, respectively.

3600

T > 300 K
O-H stretching frequency ω (cm )
-1

3400
influence the stretching frequencies. The mode with the lowest
frequency is at ~2838 cm–1, therefore this must correlate to the
shortest D···O distance shown on the graph. This is the Dw···O4 3200
ω3620
hydrogen bond (1.6 Å at 2 K). The corresponding O5···O4 dis- ω3550
ω3200
tance in O5-Dw···O4 of 2.603 Å (Fig. 12) is also close to the ω2800
3000
ideal distance of 2.5 Å observed in hydrogen bonds. The sec- Libowitzky's function

ond shortest D-O distance of 1.67 Å, between atoms Dd and

O4a, is assigned to the stretching mode with the next-highest 2800

frequency at 3200 cm–1. The corresponding distance between

O4 and O4a in O4-Dd···O4a is 2.66 Å, as shown in Figure 12. 1.6 1.8 2.0 2.2 2.4 2.6

D-O distances (Å)

For both these modes, increasing the D···O length correlates
with increasing frequency for O-H stretching.
FIGURE 14. Correlation of O-H stretching frequencies from Figure
The remaining two O-H stretching frequencies are observed
13 with observed O···D bonds. Also shown is a function predicting the
at ~3550 cm–1 and ~3620 cm–1 at low temperatures. The sepa- frequency-distance relationship as derived by Libowitzky (1999). In
ration between the two modes is small at low temperatures, general, the agreement between the data pairs observed frequency/D-
whereas at higher temperatures the two modes become indis- O bond is good for low temperature data. The arrow indicates data
tinguishable. Libowitzky and Rossman (1996) assigned these points of the high frequency stretching mode above ~300 K which plot
to O4a-Hha and O5-Hha. The Dda···O5 bond in O4a-Dda···O5 significantly away from the general trend.
 MEYER ET AL.: PHASE CHANGES IN DEUTERATED LAWSONITE 577

in lawsonite can be understood in terms of proton ordering, Los AlamosLaboratory Report LAUR 86-748.
Libowitzky, E. (1999) Correlation of O-H stretching frequencies and O-H···O hy-
there is also a significant contribution from framework dis- drogen bond lengths in minerals. Monatshefte für Chemie, 130, 1047–1059.
placements. The same may also be true of the Pmcn ↔ P21cn Libowitzky, E. and Armbruster, T. (1995) Low-temperature phase transitions and
transition but the thermodynamic character of this transition the role of hydrogen in lawsonite. American Mineralogist, 80, 1277–1285.
Libowitzky, E. and Rossman, G. R. (1996) FTIR spectroscopy of lawsonite be-
has not been so well defined. tween 82 and 325 K. American Mineralogist, 81, 1080–1091.
Martin-Olalla, J.–M., Hayward, S.A., Meyer, H.-W., Ramos, S., del Cerro, J., and
ACKNOWLEDGMENTS Carpenter, M.A. (2001) Phase transitions in lawsonite: a calorimetric study.
European Journal of Mineralogy, in press.
We thank C. Francis from Harvard University, USA, for supplying the Meyer, H.–W., Carpenter, M.A., Graeme-Barber, A., Sondergeld, P., and Schranz,
sample used in this study. Thanks also go to B.C. Schmidt (Bayreuth, Germany) W. (2000) Local and macroscopic order parameter variations associated with
for carrying out the NMR experiments and to Rutherford Appleton Laboratory low temperature phase transitions in lawsonite, CaAl2Si2O7(OH)2·H2O. Euro-
for providing beam time at ISIS. This work was performed within and finan- pean Journal of Mineralogy, 12, 1139–1150.
cially supported by the European Union TMR-Network on Mineral Transfor- Pawley, A.R. (1994) The pressure and temperature stability limits of lawsonite:
mations (contract no. ERB-FMRX-CT-97-0108). implications for H2O recycling in subduction zones. Contributions to Mineral-
ogy and Petrology, 118, 99–108.
Pawley, A.R., Redfern, S.A.T., and Holland, T.J.B. (1996) Volume behavior of hy-
REFERENCES CITED drous minerals at high pressures and temperatures: I. Thermal expansion of
Carpenter, M.A. (1992) Thermodynamics of phase transitions in minerals: A mac- lawsonite, zoisite, clinozoisite, and diaspore. American Mineralogist, 81, 335–
roscopic approach. In G.D. Price & N.L. Ross, eds., In The stability of miner- 340.
als, 172–215 p. Chapman and Hall, London. Rios, S., Paulus, W., Cousson, A., Quilichini, M., and Heger, G. (1998) Isotope
Carpenter, M.A., Powell, R., and Salje, E.K.H. (1994) Thermodynamics of effect in TlH2PO4 and TlD2PO4. Acta Crystallographica, B54, 790–797.
nonconvergent ordering in minerals: I. An alternative approach. American Min- Salje, E.K.H., Wruck, B., and Marais, S. (1991a) Order parameter saturation at low
eralogist, 79, 1053–1067. temperatures – numerical results for displacive and o/d systems. Ferroelectrics,
Carpenter, M.A., Salje, E.K.H., and Graeme-Barber, A. (1998) Spontaneous strain 124, 185–188.
as a determinant of thermodynamic properties for phase transitions in minerals. Salje, E.K.H., Wruck, B., and Thomas, H (1991b) Order-parameter saturation and
European Journal of Mineralogy, 10, 621–691. low-temperature extension of Landau theory. Zeitschrift für Physik B, 82, 399–
Ibberson, R.M., David, W.I.F., and Knight, K.S. (1992) The high resolution powder 404.
diffractometer (HRPD) at ISIS – a user guide. Rutherford Appleton Laboratory Schmidt, M.W. (1995) Lawsonite: Upper pressure stability and formation of higher
Report RAL-92–031. density hydrous phases. American Mineralogist, 80, 1286–1292.
Johnson, M.W. and David, W.I.F. (1985) HRPD: The high resolution powder Sondergeld, P., Schranz, W., Tröster, A. Carpenter, M.A., Libowitzky, E., and Kityk,
diffractometer at the SNS. Rutherford Appleton Laboratory Report RAL-85– A. V. (2000) Optical, elastic and dielectric studies of the phase transitions in
112. lawsonite. Physical Review B, 62, 6143–6147.
Lager, G.A., Libowitzky, E., and Schultz, A.J. (1999) Neutron diffraction study of
the low-temperature phase transitions in lawsonite. Proceedings of the 17th Gen- MANUSCRIPT RECEIVED AUGUST 2, 2000
eral Meeting of the International Mineralogical Association, Toronto, A99. MANUSCRIPT ACCEPTED NOVEMBER 28, 2000
Larson, A.C. and Von Dreele, R.B. (1994) General Structure AnalysisSystem (GSAS), MANUSCRIPT HANDLED BY BRYAN CHAKOUMAKOS