See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/50248558

Assignment of the vibrational spectra of lithium hydroxide monohydrate,

LiOH·H2O

Article in The Journal of Chemical Physics · February 2011

DOI: 10.1063/1.3553812 · Source: PubMed

CITATIONS READS
32 2,324

4 authors, including:

Geoffrey Dent
GDAnalytical Consulting
51 PUBLICATIONS 2,866 CITATIONS

SEE PROFILE

All content following this page was uploaded by Geoffrey Dent on 11 March 2014.

The user has requested enhancement of the downloaded file.

 THE JOURNAL OF CHEMICAL PHYSICS 134, 084503 (2011)

Assignment of the vibrational spectra of lithium hydroxide monohydrate,

LiOH·H2 O
Stewart F. Parker,1,a) Keith Refson,2,b) Robert I. Bewley,1,c) and Geoffrey Dent 3,d)
1
ISIS Facility, STFC Rutherford Appleton Laboratory, Chilton, Didcot, Oxon OX11 0QX, United Kingdom
2
Computational Science and Engineering Department, STFC Rutherford Appleton Laboratory, Chilton,
Didcot, OX11 0QX, United Kingdom
3
Avecia ASG, PO Box 42, Hexagon House, Blackley, Manchester M9 8ZS, United Kingdom
(Received 19 November 2010; accepted 21 January 2011; published online 24 February 2011)

The assignment of the vibrational spectra of lithium hydroxide monohydrate, LiOH·H2 O, has been
controversial for more than half-a-century. Here we show that only the combination of all three forms
of vibrational spectroscopy: infrared, Raman and inelastic neutron scattering spectroscopies coupled
with periodic-density functional theory calculations is able to satisfactorily assign the spectra. All
previous work based on empirical criteria is, at least partially, incorrect. The librational modes of
water do not follow the expected rock > wag > twist order and the calculations indicate that complete
or partial deuterium substitution would not be useful in assigning the modes. © 2011 American
Institute of Physics. [doi:10.1063/1.3553812]

I. INTRODUCTION The aim of the present work is to give an unambigu-

ous assignment of the hydroxyl and water librations in
Hydrates occur in a wide range of compounds from phar-
LiOH·H2 O. To this end we have obtained new infrared and
maceuticals to minerals and concrete.1 The presence of water
Raman spectra of 6 LiOH·H2 O and 7 LiOH·H2 O. To observe
in the compounds markedly alters their properties and behav-
the missing modes we have used inelastic neutron scattering
ior. The strength of the hydrogen bonding present is reflected
(INS) spectroscopy.14 To test the assignments, we have car-
in the vibrational spectrum as changes in position, width, and
ried out a periodic density functional theory (periodic-DFT)
intensity of the bands. This is particularly the case for the wa-
calculation of the complete unit cell. The combination of tech-
ter librational modes and these have been extensively studied
niques has proven to be essential to resolve the ambiguities in
in a large number of cases.2, 3
the literature.
Lithium hydroxide monohydrate, LiOH·H2 O, provides a
useful test case for the assignment of water librational modes.
The structure has been determined by both single crystal II. EXPERIMENTAL
x-ray and neutron diffraction.4 The possibility of isotopic 6
LiOH·H2 O and 7 LiOH·H2 O were obtained from Aldrich
substitution on all three elements makes LiOH·H2 O a partic- and used as received. The inelastic neutron scattering mea-
ularly attractive subject for vibrational spectroscopy conse- surement was performed using the spectrometers TOSCA
quently, it has been studied by both infrared5–8 and Raman (Ref. 15) and HET (Ref. 16) at the ISIS pulsed spallation neu-
spectroscopy.8–10 However, despite efforts spanning half-a- tron source at the Rutherford Appleton Laboratory, Chilton,
century the interpretation of the spectra is controversial with UK. TOSCA offers high resolution, ∼1.25% E/E between
no clear consensus as to the assignments and not all of the 16 and 4000 cm−1 . On TOSCA there is a fixed relationship
modes have been observed. An infrared and Raman study11 between momentum transfer, Q (Å−1 ), and energy transfer, E
suggested that the structure consisted of OH− · · ·H+ · · ·OH− (cm−1 ) given by Q2 ≈ E/16, thus it follows a fixed trajectory
units rather than hydroxyl ions and water molecules linked through (Q,E) space. Combination modes between the exter-
by hydrogen bonds. This interpretation was based on the ab- nal modes and the internal modes (phonon wings) are allowed
sence of modes due to O–H stretch and H–O–H bending in transitions in INS spectroscopy. They are unusual in the sense
the Raman spectrum. However, this is not uncommon; e.g., that the total intensity (fundamental + wing) distribution be-
the Raman spectrum of K2 [FeCl5 (H2 O)] (Ref. 12) shows no tween the fundamental and the wing depends on Q2 , thus at
evidence for the presence of water. The proposed structure is large energy (momentum) transfer most of the intensity is in
also clearly at variance with the crystal structure determined the wing and the fundamentals are not observable. This ef-
by single crystal neutron diffraction4 which shows discrete fect is particularly severe in the O–H stretch region. The HET
OH− and H2 O entities, as was noted in a Comment.13 spectrometer used a fixed incident energy (thus is analogous
to a Raman spectrometer) and was designed to access low
a) Author to whom correspondence should be addressed. Electronic mail: momentum transfers at large energy transfer, albeit at lower
[email protected]. FAX: 00-44-1235-44-5720. resolution. At energy transfers approaching the incident en-
b) Electronic mail: [email protected].
c) Electronic mail: [email protected]. ergy there is a steeply rising background resulting from mul-
d) Present address: GDAnalytical Consulting, 53 Nudger Green, Dobcross, tiple scattering. To first order, this background is Q indepen-
Oldham OL3 5AW, UK. Electronic mail: [email protected]. dent. HET had detectors at small scattering angles, 3–30◦ , the

0021-9606/2011/134(8)/084503/7/$30.00 134, 084503-1 © 2011 American Institute of Physics

084503-2 Parker et al. J. Chem. Phys. 134, 084503 (2011)

low-Q bank, and at large scattering angles 110–138◦ , the tations of the point group and assign IUPAC symmetry la-
high-Q bank. The HET spectra presented are difference spec- bels. DFPT was also used to compute the dielectric response
tra where the smoothed spectra of the high-Q bank have been and the Born effective charges, and from these the mode os-
subtracted from the data from the low-Q bank. The two spec- cillator strength tensor and infrared absorptivity were calcu-
trometers are highly complementary. The samples were held lated. The Raman activity tensors were calculated using a hy-
in aluminium sachets, cooled to 15 K and the spectra were brid finite displacement/DFPT method.23 In addition to the di-
recorded for either six hours (TOSCA) or two hours (HET). rect evaluation of frequencies and intensities at zero wavevec-
6
Li has a very large absorption cross section for slow neu- tor, phonon dispersion was also calculated along high sym-
trons, thus the INS spectrum is essentially obtained from a metry directions throughout the Brillouin zone. For this pur-
very thin sample and is consequently of much lower quality pose, dynamical matrices were computed on a regular grid of
than for 7 LiOH·H2 O and is not shown. wavevectors throughout the Brillouin zone and Fourier inter-
FT-Raman spectra with 1064 nm excitation over the polation was used to extend the computed grid to the desired
range 100–3500 cm−1 , were recorded using a Perkin Elmer fine set of points along the high-symmetry paths.24 Transition
Series 2000 spectrometer with a Raman attachment. The spec- energies for isotopic species were calculated from the dynam-
tra were recorded at 4 cm−1 resolution. ical matrix that is stored in the CASTEP checkpoint file us-
Mid-infrared, 200–4000 cm−1 , spectra were recorded at ing the PHONONS utility.25 The INS spectrum was generated
4 cm−1 resolution using a Perkin Elmer 1725 Fourier trans- from the eigenvectors by the program ACLIMAX.26 The calcu-
form infrared spectrometer with a room temperature deuter- lated INS spectra include overtones and combinations up to n
ated triglycine sulphate detector. Spectra were recorded at = 10 (n = 1 are fundamentals, n = 2 are first overtones and
room temperature by ATR using a Golden Gate accessory and binary combinations, etc,. . . ).
also from CsI discs.
Periodic-DFT calculations of the crystalline structure III. RESULTS
were carried out using the plane-wave pseudopotential
method implemented in the CASTEP code.17, 18 Exchange At room temperature LiOH·H2 O crystallizes in the mon-
and correlation were approximated using the Perdew-Burke- oclinic space group C2/m ≡ C2h 3
(number 12) with four for-
Ernzerhof (PBE) functional.19 Norm-conserving pseudopo- mula units in the conventional centered cell4 and two in the
tentials were generated using the kinetic-energy optimized primitive cell, see Fig. 1. The lithium ions and the water
method20–22 with a plane wave cutoff energy of 700 eV. Bril- molecules lie on C2 sites and the hydroxyl ions lie in the
louin zone sampling of electronic states was performed on mirror plane. As shown in Fig. 1 each lithium ion is tetrahe-
a 3 × 3 × 5 Monkhorst–Pack grid. The equilibrium struc- drally coordinated by four oxygen atoms, two from hydroxyls
ture, an essential prerequisite for lattice dynamics calculations and two from waters, and the water molecules bridge adja-
was obtained by Broyden-Fletcher-Goldfarb-Shanno (BFGS) cent lithium ions. The water molecules are hydrogen bonded
geometry optimization after which the residual forces were to the hydroxyls, HO–H. . . OH− , while the hydrogen from
converged to zero within 0.0012 eV/A. Phonon frequencies the hydroxyl ion is not involved in hydrogen-bonding. Using
were obtained by diagonalization of dynamical matrices com- the correlation method,27 the vibrations can be classified as
puted using density-functional perturbation theory18 (DFPT). shown in Table I. The presence of two independent formula
An analysis of the resulting eigenvectors was used to map the units in the unit cell results in 18 Raman active in-phase (Ag
computed modes to the corresponding irreducible represen- and Bg ) and 15 infrared active out-of-phase (Au and Bu ) modes
with no coincidences. The three acoustic modes are inactive in
both forms of spectroscopy because they have zero frequency
at zero wavevector. All modes are allowed in the INS spec-
trum since this is not restricted to observation of modes at
zero wavevector as are infrared and Raman spectroscopy.

TABLE I. Classification of vibrations of LiOH·H2 O.

Vibration

O–H stretch
OH− Ag + Bu
H2 O Ag + Bg + Au + Bu
Bend
H–O–H Ag + Au
Librations
OH− Ag + Bg + Au + Bu
H2 O Ag + 2Bg + Au + 2Bu
Translations
FIG. 1. Two adjacent primitive cells of the conventional monoclinic cell of Li+ Ag + 2Bg + Au + 2Bu
LiOH·H2 O. Each lithium ion (purple) is tetrahedrally coordinated by two H2 O + OH− 3Ag + 3Bg + Au + 2Bu
oxygens (red) from waters and two from hydroxyls. The water molecules Acoustic translations Au + 2Bu
bridge the lithium ions. (Hydrogen is shown in white).
084503-3 Lithium hydroxide monohydrate J. Chem. Phys. 134, 084503 (2011)

intense bands are seen at 999, 853, 678, and 633 cm−1 . In
the Raman spectrum a medium band is seen at 838 cm−1 and
a weak band at 693 cm−1 . All of these bands are√unaffected
by 6 Li/7 Li substitution but give the expected 1/ 2 shift on
deuterium substitution6, 7, 9 confirming the assignment to the
water and hydroxyl librational modes that are the main fo-
cus of this study. We assign the Raman bands at 838 and
693 cm−1 as the centrosymmetric counterparts of the infrared
bands at 853 and 678 cm−1 , respectively. In the INS spectrum
of 7 LiOH·H2 O five bands are clearly seen at 1038/988, 872,
796, 657, and 624 cm−1 , where the 796 cm−1 is the “missing”
band.
Two assignments of the hydroxyl librational modes
FIG. 2. INS (recorded on TOSCA) (a), FT-Raman (c) and infrared spectra
have been proposed: that they are the lowest energy (near
(e) of 7 LiOH·H2 O and FT-Raman (b) and infrared spectra (d) of 6 LiOH·H2 O.
The infrared and FT-Raman spectra were obtained at room temperature, the 650 cm−1 ) modes in this group6, 8 and that they are the high-
INS spectrum on TOSCA at 15 K. est energy modes7 (near 1000 cm−1 ). We note that the libra-
tional modes in LiOH are at low energy (419, 620 cm−1 in
7
LiOH),28–30 favoring the first assignment. Further, since the
The INS, FT-Raman, and infrared spectra of 7 LiOH·H2 O INS spectral intensity depends on the amplitude of vibration,
in the fingerprint region are given in Fig. 2 together with the to first order it would be expected that the two hydroxyl li-
FT-Raman and infrared spectra of 6 LiOH·H2 O. The spectra of brational modes would have similar intensity as would the
7
LiOH·H2 O in the O–H stretch region are shown in Fig. 3. three water librational modes, although the relative intensities
The hydroxyl O-H stretch gives sharp bands in the in- of the two groups would be different. This is exactly what is
frared and Raman9 spectra, while the water O-H stretch observed in the INS spectrum and strongly favors the assign-
modes give broad, intense bands in the infrared spectrum. ment of the hydroxyl librations to the modes near 650 cm−1 .
These observations are consistent with the hydroxyls being The HET INS data confirms this assignment because the
only weakly hydrogen-bonded while the water molecules are mode at 4285 cm−1 can be assigned to the combination (OH−
strongly hydrogen-bonded. In the INS spectrum both bands stretch + OH− libration: 3566 + 640 (average value) = 4200
are observed, in addition a third band at 4250 cm−1 is also ≈ 4250 cm−1 observed). The same combination was also seen
present. The water bending mode gives a strong band at for the hydroxyl ion in anhydrous hydroxyapatite.31
1579 cm−1 in the infrared spectrum, only a very weak feature In agreement with previous assignments,6, 9 the signif-
at 1719 cm−1 is observed in the Raman spectrum, even though icant lithium isotope shifts of the bands at 540, 412, and
it is an allowed transition. In the HET spectrum recorded with 385 cm−1 in the Raman spectrum and 485 and 434 cm−1 in
2420 cm−1 incident energy [Fig. 3(b)] two bands at 1610 and the infrared spectrum identify these as lithium translational
1704 cm−1 are clearly seen confirming the assignments. modes.
For the librational modes five bands are expected in both
the infrared and Raman spectra. In the infrared spectra four
IV. DISCUSSION
The assignment of the water librations to individual mo-
tions is often problematical. The usual order is rock > wag >
twist,2, 12 which is the expected order based on the moments
of inertia. The twist is frequently not observed in the infrared
or Raman spectra and suggests that the 796 cm−1 INS band
can be assigned to this mode. A simplistic assignment would
be that the 1038/988 cm−1 bands are the rock, the 872 cm−1
band is the wag, and the 796 cm−1 band is the twist. This has
the virtue of agreeing with previous assignments and of being
consistent with the spectra.
A major advantage of INS spectroscopy is that it is possi-
ble to calculate the intensities quantitatively. This can be done
using a classical “balls and springs” force field approach or
by using the atomic displacements in each mode generated
by a vibrational analysis from an ab initio calculation. Initial
efforts using the classical approach and a model consisting
of two formula units gave reasonable agreement with the ob-
served INS spectrum. However, all three of the water libra-
FIG. 3. Spectra of 7 LiOH·H2 O in the water bend and O–H stretch regions.
INS spectra recorded on HET with (a) 6857 cm−1 and (b) 2420 cm−1 incident
tional modes have similar INS intensity (since the amplitudes
energy (c) FT-Raman spectrum [×8.5 ordinate expansion relative to Fig. 1(c)] of vibration are similar), thus it was not possible to arrive at
and 1(d) infrared spectrum [×2 ordinate expansion relative to Fig. 1(e)]. an unambiguous assignment.
084503-4 Parker et al. J. Chem. Phys. 134, 084503 (2011)

TABLE II. Comparison of experimental [fom neutron diffraction (Ref. 4)] and calculated lattice parameters, bond distances, and angles in
LiOH·H2 O.

CASTEP
Expt. Expt. lattice parameter Optimized lattice parameter

a (Å) 7.4153 7.5528

b (Å) 8.3054 8.2764
c (Å) 3.1950 3.2467
β (◦ ) 110.107 111.8945
Volume of primitive cell (Å3) 92.389 94.157
d(HO–H) (Å) 1.002 1.014 1.014
 H–O–H (◦ ) 104.3 103.8 103.8
d(O–H) (Å) 0.951 0.971 0.971
d(H2 O · · · · HO) (Å) 1.684 1.652 1.655
d(H2 O · · · · Li) (Å) 1.982 1.975 1.986
d(HO · · · · Li) (Å) 1.965 1.980 1.985
d(HO · · · · HO) (Å) along c axis 2.250 2.238 2.285
d(HO · · · · OH) (Å) within cell 3.009 3.041 3.050
d(H2 O · · · · OH2 ) (Å) in ab plane 4.872 4.858 4.832
d(H2 O · · · · OH2 ) (Å) along c axis 3.195 3.195 3.247
 Li · · · · O(H2 ) · · · · Li (◦ ) 107.4 108.0 109.6
 HO–H · · · · OH (◦ ) 174.8 174.2 174.4
 Li · · · · O(H) · · · · Li (◦ ) 80.0 79.6 79.6
 H–O · · · · HOH (◦ ) 96.4 95.9 95.5
 H–O · · · · OH (◦ ) 167.6 168.2

To overcome this difficulty we have carried out ab initio lattice parameters are almost identical (see Table III) and are
calculations of the complete unit cell using the periodic-DFT not shown. The calculated spectra have been scaled by 0.95
code, CASTEP.17, 18 Table II compares geometric parameters so as to bring the water and hydroxyl librational modes into
for the structure optimized at the experimental lattice param- agreement. It can be seen that both transition energies and
eters and also for where both geometry and lattice optimiza- relative intensities are generally well-reproduced and clearly
tion have been carried out. The lattice optimization results in demonstrates the capability of state-of-the-art ab initio calcu-
a small increase of the cell parameters and a correspondingly lations for spectra prediction. We also note that this successful
small, ∼2%, increase in cell volume. The intramolecular bond reproduction of the infrared and Raman spectra completely
distances and angles are the same in the two calculations and invalidates the argument of Tyutyunnik11 for a formulation of
agree well with experiment (<0.02 Å for distances and 0.5◦ the structure as OH− . . . H+ . . . OH− units.
for the angle). The intermolecular distances and angles are
also very close, those of the optimized lattice calculation are
slightly larger because of the increased cell size.
Figure 4 compares the observed and calculated infrared
and Raman spectra for 7 LiOH·H2 O. The spectra shown (and
all subsequent ones) are for the calculation at the experimental
lattice parameters, those for the calculation at the optimized

FIG. 5. Comparison of the observed (middle) and scaled (by 0.95) CASTEP
FIG. 4. Comparison of observed [(a) infrared and (c) Raman] and scaled calculated INS spectra of 7 LiOH·H2 O: (lower) -point only and (upper) in-
(by 0.95) CASTEP calculated [(b) infrared and (d) Raman] spectra of cluding dispersion across the Brillouin zone. The calculated spectra include
7 LiOH·H O.
2 overtones and combinations up to n = 10.
084503-5 Lithium hydroxide monohydrate J. Chem. Phys. 134, 084503 (2011)

TABLE III. Transition energies and assignments for LiOH·H2 O.

0.95 scaled CASTEP of 7 LiOH·H2 O

Expt. lattice parameter
Optimized lattice Infrared Raman Expt.
parameter intensity/ intensity INS Infrared Raman
Mode (cm−1 ) (cm−1 ) (D/Å)2 (amu) (Å4 ) Symmetry 7 Li(cm−1 ) 6 Li(cm−1 ) 7 Li(cm−1 ) 6 Li(cm−1 ) 7 Li(cm−1 )

H2 O + OH b 94 92 0.053 Bg 90s,brb
transa
H2 O + Li c 119 115 0.191 Bg 130w
trans
OH c trans 139 151 1.031 Ag
H2 O b trans 200 194 0.894 Ag 193w 214vs 212vs
+ a OH trans
H2 O + OH b 214 218 0.018 Bu 249w
trans
H2 O + OH a 255 244 1.770 Bg 246s 246s
trans
H2 O + OH a 302 299 8.543 Bu
trans
H2 O + OH + 325 320 6.061 Au 322w 336w 336w
Li b trans
Li b trans 353 357 1.906 Bg 369w 352w 352w
H2 O + OH + 367 368 0.201 Ag 385s 366s
Li b trans
Li c trans 389 401 0.875 Bg 388vw 412s 391s
Li c trans 395 408 5.195 Bu 418w 433s 413s
Li a trans 425 427 10.822 Bu 471w 485vs,br 459vs,br
Li a trans 432 442 8.846 Au 521sh 489sh
Li b trans 491 501 6.407 Ag
OH lib 582 590 5.542 Au 624vs 633vs 632vs
OH lib 591 600 0.776 Bg
OH lib 645 651 7.624 Bu 657vs 682vs 678vs
OH lib 666 669 1.205 Ag 696w 693w
H2 O rock 837 842 6.663 Bg 796s 839m 838m
H2 O rock 858 864 5.555 Bu 853vs 853vs
H2 O twist 969 970 0.004 Au 872s
H2 O wag 984 982 2.769 Bg 955m
H2 O twist 1022 1027 1.897 Ag 988s
H2 O wag 1028 1027 9.698 Bu 1038s 999s,br 999s,br
H2 O bend 1541 1542 5.657 Au 1594m,br 1581s,br 1579s,br
H2 O bend 1649 1650 7.838 Ag 1703m 1717vw
H2 O 2584 2626 591.256 Bg
asymmetry
stretch
H2 O 2707 2756 72.620 Au
symmetry
stretch
H2 O 2752 2797 123.323 Bu 2879m,br 3000s,br
asymmetry
stretch
H2 O 2763 2813 640.811 Ag 2828
symmetry
stretch
OH− stretch 3425 3455 814.064 Ag 3608m,br
OH− stretch 3426 3455 8.261 Bu 3566mc 3566mc
a
Cell directions refer to the primitive cell.
b
s = strong, m = medium, w = weak, sh = shoulder, br = broad, v = very.
c
Reference 9.
084503-6 Parker et al. J. Chem. Phys. 134, 084503 (2011)

wavevectors across the Brillouin zone, whereas the infrared

and Raman spectra only sample the -point.
Table III compares the observed and calculated fre-
quencies and assignments. The symmetry assignments from
CASTEP are in complete agreement with those found
experimentally.8
The results clearly show that the expectation of rock
> wag > twist is incorrect; the twist and wag are at sim-
ilar frequencies with a small factor group splitting. The
rock is at lower energy and with a large factor group split-
ting. These assignments differ from those based on empirical
arguments6, 7, 10 of the bonding and demonstrate the difficulty
in assigning the modes. Deuteration and partial deuteration
have been used to assign the water librational modes in other
systems,2, 3 thus it is of interest to see if the same techniques
would discriminate among the librational modes in this case.
Access to the dynamical matrix means that the isotopic calcu-
FIG. 6. Dispersion curves of 7 LiOH·H2 O in the 0–1200 cm−1 region as cal- lations are simple and also allows calculation of isotope distri-
culated by CASTEP. butions that could not be achieved in reality. Table IV shows
the results for a variety of isotope distributions between the
two water molecules in the primitive cell. It can be seen that
Figure 5 compares the observed (middle) and scaled cal-
the isotope shifts do not follow any clear correlation with the
culated INS spectra for 7 LiOH·H2 O. The calculation at the
type of motion for either the full or partially deuterated cases.
-point, only, gives only modest agreement, in contrast, the
This is probably due to the symmetry reduction on partial sub-
calculation across the complete Brillouin zone gives excellent
stitution that changes the form of the normal modes, which is
agreement. Thus the hydrogen-bonding present in this system
why there is no shift in the highest energy wag mode in HDO:
results in significant dispersion in the librational and trans-
it is now a localized motion of the proton only. For systems
lational modes. This is more clearly seen in the dispersion
where the intramolecular coupling is strong, and particularly
curves shown in Fig. 6. The intramolecular water O–H stretch
if there is more than one water molecule in the primitive cell,
modes are also strongly dispersed whereas the hydroxyl O–H
it is likely that isotopic substitution will not be helpful and
stretch and the water bend are essentially dispersionless (see
ab initio calculations are essential to correctly assign the
Fig. 7). The calculation also confirms the assignment of the
modes.
second water bending mode ∼100 cm−1 higher than the in-
The hydroxide librations are confirmed as the modes
frared active mode. The difference in transition energies be-
around 600 cm−1 .
tween the infrared/Raman data and the INS data is partly due
For the translational modes, visualization shows that
to the different temperatures that the spectra were recorded at:
there is strong coupling between all the species present.
room temperature and 20 K, respectively, but is much more
This is indicated by the continuous intensity between 90 and
the result of the strong dispersion present in this material,
500 cm−1 in the experimental INS spectrum, indicating that
which the INS spectrum is sensitive to because it samples all
the modes are dispersed. The isotopic spectra7, 9 also show a
marked dependence on both 6 Li/7 Li and H/D substitution in-
dicating that coupling is present, in contradiction to the con-
clusions of Di Petro et al.8 This is also demonstrated by set-
ting the cross section of all the atoms in the unit cell to zero
except for the species of interest, thus only motions involving
these entities will contribute to the INS spectrum so gener-
ated. Figure 8 shows the experimental spectrum 8(a), and the

TABLE IV. Computed isotopic ratiosa for water isotopomers of

7 LiOH·H O.
2

2HDO 2D2 O H2 O + D2 O H2 O + HDO D2 O + HDO

1.22 1.33 1.28 1.19 1.29 rock

1.22 1.36 1.12 1.11 1.24 rock
1.28 1.31 1.25 1.13 1.29 twist
1.22 1.30 1.14 1.00 1.26 wag
1.05 1.26 1.02 1.02 1.27 twist
1.00 1.28 1.02 1.00 1.02 wag
FIG. 7. Dispersion curves of 7 LiOH·H2 O in the water bend and O–H stretch
region (1400–3600 cm−1 ) as calculated by CASTEP. a
Isotopic ratio is defined as [(100% H H2 O) / (substituted H2 O)].
 084503-7 Lithium hydroxide monohydrate J. Chem. Phys. 134, 084503 (2011)

(time on the SCARF compute cluster for the CASTEP cal-

culations) was provided by STFC’s e-Science facility.
1 H. G. Brittain, J. Pharm. Sci. 99, 3648 (2010); Y. Ye, R. A. Schwering, and
J. R. Smyth, Am. Mineral. 94, 899 (2009); R. J. Hand, Br. Ceram. Trans.
96, 116 (1997).
2 V. P. Tayal, B. K. Srivastava, and D. P. Khandelwal, Appl. Spectrosc. Rev.

16, 43 (1980).
3 H. D. Lutz, Struct. Bonding (Berlin) 69, 97 (1988).
4 K. Hermansson and J. O. Thomas, Acta Crystllogr. B 38, 2555 (1982).
5 L. H. Jones, J. Chem. Phys. 22, 217 (1954).
6 I. Gennick and K. M. Harmon, Inorg. Chem. 14, 2214 (1975).
7 Y. Hase, Inorg. Nucl. Chem. Lett. 16, 159 (1980).
8 E Di Petro, M. Pagliai, G. Cardini, and V. Schettino, J. Phys. Chem. B 110,

13539 (2006).
9 D. Krishnamurti, Proc. -Indian Acad. Sci., Chem. Sci. A 50, 223 (1959).
10 Y. Hase, Monatsch Chem. 112, 73 (1981).
11 V. I. Tyutyunnik, J. Raman Spec. 31, 559 (2000).
12 S. F. Parker, K. Shankland, J. C. Sprunt, and U. A. Jayasooriya, Spec-

trochim. Acta A 53, 2245 (1997).

13 H. D. Lutz, J. Raman Spec. 32, 301 (2001).
14 P. C. H. Mitchell, S. F. Parker, A. J. Ramirez-Cuesta, and J. Tomkinson,
FIG. 8. INS spectrum of 7 LiOH·H2 O, (a) experimental, and as calculated
by CASTEP for (b) water motions only, (c) hydroxide motions only and (d) Vibrational Spectroscopy with Neutrons (World Scientific, Singapore,
lithium motions only (×10 scale expansion). Only fundamentals (n = 1) are 2005).
15 D. Colognesi, M. Celli, F. Cilloco, R. J. Newport, S. F. Parker, V. Rossi-
shown in (b), (c), and (d).
Albertini, F. Sacchetti, J. Tomkinson, and M. Zoppi, Appl. Phys. A 74, S64
(2002).
16 A. D. Taylor, B. C. Boland, Z. A. Bowden, and T. J. L. Jones, Rutherford
calculated INS spectra for the hydroxyls only 8(b), the wa-
Appleton Laboratory Report RAL-87–012, 1987.
ter molecules only 8(c) and the lithium ions only 8(d). The 17 S. J. Clark, M. D. Segall, C. J. Pickard, P. J. Hasnip, M. J. Probert, K.
considerable overlap in the region below 300 cm−1 for Refson, and M. C. Payne, Z. Krist. 220, 567 (2005).
18 K. Refson, S. J. Clark, and P. R. Tulip, Phys. Rev. B 73, 155114 (2006).
all three species demonstrates significant coupling between
19 J. Perdew, K. Burke, and M. Ernzerhof, Phys. Rev. Lett. 77, 3865 (1996).
them. 20 A. M. Rappe, K. M. Rabe, E. Kaxiras, and J. D. Joannopoulos, Phy. Rev. B

41, 1227 (1990).

21 M. H. Lee, PhD thesis, Cambridge 1995; See http://boson4
V. CONCLUSIONS
.phys.tku.edu.tw/qc/my_thesis/index.htm.
22 See http://psi-k.dl.ac.uk/newsletters/News_67/Highlight_67.pdf
The assignment of the vibrational spectra of lithium hy- 23 V. Milman, A. Perlov, K. Refson, S. J Clark, J. Gavartin, and B. Winkler,
droxide monohydrate has been controversial for a long time. J. Phys.: Condens. Matter 21, 485404 (2009).
Here we show that only the combination of all three forms of 24 X. Gonze, J.-C. Charlier, and M. P. Teter, Phys. Rev. B 50, 13035 (1994).
25 K.
vibrational spectroscopy: infrared, Raman, and INS coupled Refson, Phonons and Related Calculations in CASTEP,
with periodic-DFT calculations is able to satisfactorily assign http://www.castep.org/
26 A. J. Ramirez-Cuesta, Comput. Phys. Commun. 157, 226 (2004).
the spectra. All previous work based on empirical criteria is, 27 W. G. Fateley, F. R. Dollish, N. T. McDevitt, and F. F. Bentley, Infrared and
at least partially, incorrect. Raman Selection Rules for Molecular and Lattice Vibrations: The Correla-
tion Method (Wiley-Interscience, New York, 1972).
28 F. Harbach and F. Fischer, J. Phys. Chem. Solids 36, 601 (1975).
ACKNOWLEDGMENTS 29 Y. Hase and I. V. P. Yoshida, Chem. Phys. Lett. 65, 46 (1979).
30 I. V. P. Yoshida and Y. Hase, Spectrosc. Lett. 12, 409 (1979).
The STFC Rutherford Appleton Laboratory is thanked 31 M. G. Taylor, S. F. Parker, and P. C. H. Mitchell, J. Mol. Struct. 251–253,
for access to neutron beam facilities. Computing resources 123 (2003).

View publication stats