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

net/publication/333167904

A revised jump-diffusion and rotation-diffusion model

Article  in  Chinese Physics B · May 2019

DOI: 10.1088/1674-1056/28/5/056105

CITATIONS READS

0 49

3 authors, including:

Hua Li
Jinan University (Guangzhou, China)
16 PUBLICATIONS   52 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Simulation of tidal current field View project

All content following this page was uploaded by Hua Li on 04 June 2019.

The user has requested enhancement of the downloaded file.

A revised jump-diffusion and rotation-diffusion model

Hua Li(李华), Yu-Hang Chen(陈昱沆), Bin-Ze Tang(唐宾泽)

Citation:Chin. Phys. B . 2019, 28(5): 056105. doi: 10.1088/1674-1056/28/5/056105
Journal homepage: http://cpb.iphy.ac.cn; http://iopscience.iop.org/cpb

What follows is a list of articles you may be interested in

Recent progress on magnetic-field studies on quantum-spin-liquid candidates

Zhen Ma(马祯), Kejing Ran(冉柯静), Jinghui Wang(王靖珲), Song Bao(鲍嵩), Zhengwei Cai(蔡正蔚), Shichao Li(李世超),
Jinsheng Wen(温锦生)
Chin. Phys. B . 2018, 27(10): 106101. doi: 10.1088/1674-1056/27/10/106101

Dynamic behaviors of water contained in calcium—silicate—hydrate gel at different

temperatures studied by quasi-elastic neutron scattering spectroscopy

Zhou Yi(易洲), Pei-Na Deng(邓沛娜), Li-Li Zhang(张丽丽), Hua Li(李华)

Chin. Phys. B . 2016, 25(10): 106401. doi: 10.1088/1674-1056/25/10/106401

--------------------------------------------------------------------------------------------------------------------
 Chin. Phys. B Vol. 28, No. 5 (2019) 056105

A revised jump-diffusion and rotation-diffusion model

Hua Li(李华)† , Yu-Hang Chen(陈昱沆), and Bin-Ze Tang(唐宾泽)
Department of Physics, Jinan University, Guangzhou 510632, China

(Received 18 December 2018; revised manuscript received 20 March 2019; published online 19 April 2019)

Quasi-elastic neutron scattering (QENS) has many applications that are directly related to the development of high-
performance functional materials and biological macromolecules, especially those containing some water. The analysis
method of QENS spectra data is important to obtain parameters that can explain the structure of materials and the dynamics
of water. In this paper, we present a revised jump-diffusion and rotation-diffusion model (rJRM) used for QENS spectra
data analysis. By the rJRM, the QENS spectra from a pure magnesium-silicate-hydrate (MSH) sample are fitted well for
the Q range from 0.3 Å−1 to 1.9 Å−1 and temperatures from 210 K up to 280 K. The fitted parameters can be divided
into two kinds. The first kind describes the structure of the MSH sample, including the ratio of immobile water (or bound
water) C and the confining radius of mobile water a0 . The second kind describes the dynamics of confined water in pores
contained in the MSH sample, including the translational diffusion coefficient Dt , the average translational residence time
τ0 , the rotational diffusion coefficient Dr , and the mean squared displacement (MSD) hu2 i. The rJRM is a new practical
method suitable to fit QENS spectra from porous materials, where hydrogen atoms appear in both solid and liquid phases.

Keywords: revised jump-diffusion and rotation-diffusion model (rJRM), data analysis of quasi-elastic neutron
scattering (QENS) spectra, dynamics of water, magnesium-silicate-hydrate (MSH) samples
PACS: 61.05.fg, 64.70.qj, 66.30.jj DOI: 10.1088/1674-1056/28/5/056105

1. Introduction translational motion of water and taking into account the con-
Quasi-elastic neutron scattering (QENS) has undergone tribution of the neutron scattering lengths in Sears expansion
remarkable development since the 1970s. [1] A QENS exper- to deal with the rotational motion. The new model of rJRM
iment measures the energy transfer (E) and momentum ex- seems to give a better fitted line for a QENS spectrum from
change (Q) between a neutron and a sample target. Because of a pure magnesium-silicate-hydrate (MSH) sample at Q values
the very large incoherent scattering cross section of a hydro- 0.3–1.9 Å−1 and E from −120 µeV to 120 µeV. Two kinds
of parameters can be extracted for QENS spectra fitted by the
gen atom as compared to any other element present in the in-
rJRM. One depicts dynamics of water confined in the MSH
vestigated samples, QENS [2–4] is a powerful tool for studying
sample, and the other reflects the structure of the MSH sam-
the dynamics of water molecules in a confined environment.
ple.
The fitted results of QENS spectra can provide information
about both the long-time diffusive motions (on the time scale
of 0.1 ns) and short-time rotational motions of water molecules 2. Fitting methods of QEMS spectra
contained in porous samples. So the analysis method of QENS 2.1. The jump-diffusion and rotation-diffusion model
spectra is important to obtain parameters explaining the dy- The jump-diffusion and rotation-diffusion model has al-
namics of confined water. ready been used to fit QENS spectra by Bordallo et al. [2] in
As we know, there are mainly three types of theoretical 2006 and us in recent years. [11,12] We adopt this model, differ-
models [5] used to fit QENS data: the empirical diffusion model ent from Bordallo et al., by choosing the scattering function in
(EDM), [6–8] the relaxed cage model (RCM), [4,9,10] and the the form of S(Q, E) instead of S(Q, ω). The JRM can be stated
jump-diffusion and rotation-diffusion model (JRM). [2,11,12] as
Although all the three models have been used to fit QENS
S(Q, E)
spectra data, seldom of them can give well fitted QENS spec-
tra both for Q ≤ 1 Å−1 and Q > 1 Å−1 at the neutron energy = A[Cδ (E) + (1 −C)ST (Q, E) ⊗ SR (Q, E)] ⊗ R(Q, E), (1)
transfer larger than 50 µeV. where Q is the neutron scattering vector (momentum ex-
The JRM combines both the translational and rotational change), E the energy transfer of neutron, A the Debye–Waller
motions of water, so it can give more comprehensive and factor (DWF), δ (E) the Dirac delta function, C the factor pre-
related parameters describing the water contained in porous senting the part of elastic contribution to the total QENS spec-
samples, but there still exist some approximations. In this tra, (1 − C) the factor denoting the quasi-elastic component
paper, a revised JRM (rJRM) has been developed based on in the whole QENS spectra, ST (Q, E) and SR (Q, E) the scat-
JRM by adding the contribution to the elastic part from the tering functions contributed by the translational motion and
† Corresponding author. E-mail: [email protected]

© 2019 Chinese Physical Society and IOP Publishing Ltd 　http://cpb.iphy.ac.cn

http://iopscience.iop.org/cpb　

056105-1
 Chin. Phys. B Vol. 28, No. 5 (2019) 056105
rotational motion, respectively, ⊗ the convolution action, and l(l + 1)h̄Dr
× , (5)
R(Q, E) the experimental resolution function. [l(l + 1)h̄Dr ]2 + E 2
In Eq. (1), the DWF [13] can be given as
  where jl (l = 0, 1, 2, 3) is the spherical Bessel function, Dr the
1
2 2
A = exp − u Q , (2) rotational diffusion coefficient, and a the radius of gyration
3
taken as the O–H distance (0.98 Å) in a water molecule. [2]
where u2 is the mean square displacement (MSD).


In order to show the JRM fitting, the QENS spectra from
The translational component ST (Q, E) is usually modeled
one sample previously studied [10] by using the new global
by a Lorentzian with a half-width at half-maximum (HWHM)
model are reanalyzed, which is the pure MSH sample mea-
ΓT (Q) as
sured at 210–280 K and Q values 0.3–1.9 Å−1 . The QENS
1 ΓT (Q)
ST (Q, E) = . (3) data were collected by Professor Sow-Hsin Chen’s group at
π ΓT2 (Q) + E 2
the high-resolution backscattering silicon spectrometer at the
According to the model of Singwi and Sjölander, [14] the Q- Spallation Neutron Source in Oak Ridge National Laboratory,
dependence of ΓT (Q) in the random jump diffusion process details of which are given in Ref. [10]. Figure 1 shows the
can be depicted as follows: MSH QENS spectra fitted by using JRM for 230 K and 280 K
h̄Dt Q2 at Q values of 0.3 Å−1 and 1.5 Å−1 , respectively. In Fig. 1,
ΓT (Q) = , (4)
1 + Dt Q2 τ 0 the fitted line of the measured QENS data can be stated as the
where h̄ is the Plank constant, Dt the self-diffusion coefficient sum of three parts of ENS, QENS1, and QENS2, which are
(translational diffusion coefficient), and τ0 the average trans- corresponding to the ACR(Q, E), A(1 − C) j02 (Qa)ST (Q, E) ⊗
lational residence time between two jumps. R(Q, E), and the rest of the expanded Eq. (1), respectively. The
The rotational component SR (Q, E) can be expressed by ENS is the signal from the immobile species, such as water or
the Sears expansion [2,11,15] as OH chemically bound in a sample. The QENS1 and QENS2
1 3 are the signals from mobile water confined in nanoscale pores
SR (Q, E) = j02 (Qa)δ (E) + ∑ (2l + 1) jl2 (Qa)
π l=1 in the sample.

QENS data fitted by JRM at T=230 K QENS data fitted by JRM at T=280 K
100 100
(a) QENS data (b) QENS data
Intensity/meV-1
Intensity/meV-1

fitted line fitted line

Q=0.3 A-1 ENS ENS
10-2 QENS1 10-2 Q=0.3 A-1 QENS1
QENS2 QENS2

10-4 10-4

10-6 10-6
100 100
Q=1.5 A-1 Q=1.5 A-1
Intensity/meV-1

Intensity/meV-1

10-2 10-2

10-4 10-4

10-6 10-6
-100 -50 0 50 100 -100 -50 0 50 100
E/meV E/meV

Fig. 1. MSH QENS spectra data fitted by JRM for temperatures (a) 230 K and (b) 280 K at Q values of 0.3 Å−1 and 1.5 Å−1 , respectively.

From Figs. 1(a) and 1(b), it can be seen that the QENS representing the immobile water in Eq. (1). But there should
spectra are fitted well only for 230 K at 0.3 Å−1 (shown in be an elastic signal coming from the mobile water even with-
Fig. 1(a)) and 280 K at 1.5 Å−1 (shown in Fig. 1(b)). Why? out the presence of chemical bound water, [10,13,16] which is not
Is there a miss of some parts in the QENS spectra fitting? Re- included in Eq. (3). Re-examining the rotational component in
examining the translational component in JRM, we find that JRM, the part of the QENS spectra due to localized rotational
the elastic fraction is accounted for only by a delta function motion taken from the Sears expansion [15,17] should be a func-
056105-2
 Chin. Phys. B Vol. 28, No. 5 (2019) 056105
tion of the neutron scattering lengths due to the interaction The acoh and ainc play the roles of effective coherent and inco-
of the incident neutrons with the confined water molecules, herent neutron scattering lengths, η = 0.95 represents the ef-
which is also not included in Eq. (5). By adding the contribu- fect of the nuclear statistics on the scattering cross section, [17]
tion to the elastic part from the mobile water and taking into x is the parameter to cancel units of b2l . The taken values of
account the contribution of the neutron scattering lengths in a2coh and a2inc are obtained approximately from those of H2 and
the Sears expansion, a revised JRM (rJRM) is developed in O2 and listed in Table 1.
the next section.

2.2. The revised jump-diffusion and rotation-diffusion Table 1. Effective neutron scattering lengths used in the rJRM fitting.
model Taken values H2 [17] O2 [17]
In the rJRM, the elastic part is added to ST (Q, E) (instead a2coh /barn 0.307 0.142 0.33
of Eq. (3)) as a2inc /barn 6.34 6.34 0.02

1 ΓT (Q)
ST (Q, E) = p(Q)δ (E) + (1 − p(Q)) ,
π ΓT2 (Q) + E 2 After the above two revisions based on the JRM, the ex-
 2 panded Eq. (1) can be shown as follows after inserting Eqs. (4),
3 j1 (Qa0 )
p(Q) = , (6)
Qa0 (6), and (7):

where p(Q) is the elastic component coming from the mo-

S(Q, E)
bile water confined in pores, [8,10,13] j1 the first order spherical
Bessel function, and a0 the confining radius. = ACR(Q, E) + A(1 −C)p(Q)xb20 j02 (Qa)R(Q, E)
In the rJRM, SR (Q, E) includes the contribution of neu- 1 ΓT (Q)
+ A(1 −C)(1 − p(Q))xb20 j02 (Qa) ⊗ R(Q, E)
tron scattering lengths [15,17] (instead of Eq. (5)) as π ΓT2 (Q) + E 2
1 ΓT (Q)
1 3 + A(1 −C)(1 − p(Q))x

SR (Q, E) = x b20 j02 (Qa)δ (E) + ∑ (2l + 1)b2l jl2 (Qa) π ΓT2 (Q) + E 2
π l=1
1 3 l(l + 1)h̄Dr
∑ (2l + 1)b2l jl2 (Qa) [l(l + 1)h̄Dr ]2 + E 2 ⊗ R(Q, E)

l(l + 1)h̄Dr ⊗
× , (7) π l=1
[l(l + 1)h̄Dr ]2 + E 2
1 3
where + A(1 −C)p(Q)x ∑ (2l + 1)b2l jl2 (Qa)
π l=1
(
4a2coh + 2(1 − η)a2inc , l is even, l(l + 1)h̄Dr
b2l = × ⊗ R(Q, E). (8)
2(1 + η)a2inc , l is odd. [l(l + 1)h̄Dr ]2 + E 2

QENS data fitted by rJRM at T=230 K QENS data fitted by rJRM at T=280 K
100 100
(a) QENS data (b) QENS data
fitted line
Intensity/meV-1

Intensity/meV-1

fitted line
Q=0.3 A-1 ENS Q=0.3 A-1 ENS
QENS1 QENS1
10-2 QENS2 10-2 QENS2

10-4 10-4

10-6 10-60
100 10
Q=1.5 A-1 Q=1.5 A-1
Intensity/meV-1

Intensity/meV-1

10-2 10-2

10-4 10-4

10-6 10-6
-100 -50 0 50 100 -100 -50 0 50 100
E/meV E/meV

Fig. 2. MSH QENS spectra data fitted by rJRM for temperatures (a) 230 K and (b) 280 K at Q values of 0.3 Å−1 and 1.5 Å−1 , respectively.

056105-3
 Chin. Phys. B Vol. 28, No. 5 (2019) 056105
A QENS spectrum can be fitted by the rJRM by summing three sample, respectively. The other is parameters including DWF
parts representing the ENS, QENS1, and QENS2, which are or MSD, Dt , Dr , and τ0 , which all describe the dynamics of
given respectively by the sum of the first and second items, mobile water confined in pores. Figure 3 shows the seven fitted
the third item, and the rest items in Eq. (8). To test the rJRM parameters, with their fitted standard errors, and the MSD cal-
fitting, the QENS spectra analyzed are the same as those fit- culated by Eq. (2) as a function of the Q value (0.3–1.9 Å−1 )
ted by JRM in Subsection 2.1 for the whole Q values, which at 230 K and 280 K.
are from the pure MSH sample measured at 210–280 K for From Fig. 3(a), it can be seen that the DWF is nearly equal
Q of 0.3–1.9 Å−1 . Figure 2 shows the rJRM fitting for the to 1 for Q from 0.3 Å−1 to 1.9 Å−1 at 230 K, but it is smaller
MSH QENS spectra. From Figs. 2(a) and 2(b), it can be seen than 1 and increases slightly with increasing Q at 280 K. This
that the QENS spectra data are fitted well for E ranges from makes all the calculated MSD (in Fig. 3(b)) < 1.6 Å2 for all the
−120 µeV to 120 µeV both for high and low Q. Comparing Q at 230 K, whereas the MSD decreases from 11.5 Å2 down
with Figs. 1(a) and 1(b), it is found that the rJRM fitting is to near zero with increasing Q at 280 K. This is the first new
better for 230 K at 1.5 Å−1 (shown in Fig. 2(a)) and 280 K at result compared with the QENS spectrum fitted by RCM [4,10]
0.3 Å−1 (shown in Fig. 2(b)), because of the obvious higher or EDM, [6–8] where the MSD cannot be directly obtained be-
contribution of the QENS2 due to accounting for the contribu- cause of the DWF taken approximately as unity in the data
tion of neutron scattering lengths in Eq. (7).
fitting process. From Fig. 3(c), it can be seen that the ratio
of immobile water (mainly bound water) C decreases with in-
3. The extracted parameters creasing Q at both 230 K and 280 K, but it is larger at the lower
3.1. The fitted parameters temperature. Figure 3(d) shows that the x values all are close
There are seven parameters extracted by using rJRM to to 1 for both 230 K and 280 K. From Fig. 3(d), it can also
fit a QENS spectrum, which are A, C, Dt , Dr , τ0 , a0 , and x. be seen that the x values are slightly larger at the lower tem-
These fitted parameters can be divided into two types. One is perature 230 K for Q > 1 Å−1 . This may be attributed to the
structure parameters including C and a0 , which describe the more active rotational motion for water molecules confined in
immobile water fraction and the radius of pores existing in the smaller pores (larger Q) and at the lower temperature 230 K.

1.5 30
(a) (e) T=230 K
20 T=280 K
a0/A

1.0
A

10

0.5 0
0 0.5 1.0 1.5 2.0 0 0.5 1.0 1.5 2.0

15 0.10
(b) (f)
MSD/A2

10
τ0/ns

0.05
5

0 0
0 0.5 1.0 1.5 2.0 0 0.5 1.0 1.5 2.0

1.0 150
(c) (g)
Dt/A2Sns-1

0.8
100
C

0.6
50
0.4
0.2 0
0 0.5 1.0 1.5 2.0 0 0.5 1.0 1.5 2.0

3
8000
(d) (h)
x/barn-1

Dr/ns-1

2 6000
4000
1
2000
0 0
0 0.5 1.0 1.5 2.0 0 0.5 1.0 1.5 2.0
Q/A-1 Q/A-1

Fig. 3. (a)–(h) Seven fitted parameters with their fitted standard errors and the calculated MSD versus scattering vector at two temperatures of
230 K and 280 K.
056105-4
 Chin. Phys. B Vol. 28, No. 5 (2019) 056105
Figure 3(e) shows that the fitted a0 formed in the MSH gives the fitted Dr as a function of Q and T . From Fig. 4(b),
sample decreases with increasing Q, and its value is almost it can be seen that Dr ranges from 40 ns−1 to 8000 ns−1 and
the same at 230 K and 280 K for the same Q. This can be increases with the increasing Q at a given T , and it reaches
explained from the fact that the larger radius is probed at the peak values at 230 K. This shows distinctly that the rotational
smaller Q, and the radius of a pore formed in the sample is dynamics of confined water exhibits different behaviors below
unrelated to the measured temperature. This is the second and above 230 K. The 230 K is a crossover temperature for ro-
new result that the structure parameter a0 different with var- tational dynamics of water confined in pores of 2.0–19 Å. The
ious Q can be extracted by the rJRM fitting, although its fit- result is new compared with the same MSH QENS data ana-
ted standard error is larger for larger Q at lower temperature. lyzed over rotational dynamics of confined water in Ref. [10]
This result will be discussed more in Subsection 3.2. From by the new global model, where the crossover at 230 K was not
Fig. 3(f), it can be seen that τ0 is smaller than 0.1 ns at 230 K observed because the lowest temperature was taken as 230 K.
and 280 K, and its fitted standard error is larger for the larger
Q at the lower temperature. Figures 3(g) and 3(h) show the 20
Q=0.3 A-1 (a)
fitted parameters of Dt and Dr , which range 1–100 Å2 /ns and Q=0.5 A-1
Q=0.7 A-1
40–8000 ns−1 , respectively. From Figs. 3(g) and 3(h), we can 15
also see that Dt and Dr vary with increasing Q and different
Q=0.9 A-1 Q=1.5 A-1
temperatures. The former decreases with increasing Q at a Q=1.1 A-1 Q=1.7 A-1

a0/A
10 Q=1.3 A-1 Q=1.9 A-1
measured temperature, but the latter increases. The former
values are smaller at 230 K than those at 280 K, but vice versa
for the latter. Their fitted standard errors all are small except 5
for Dt at lower temperature 230 K for larger Q. The decrease
of parameter Dt with Q indicates the inactive translational mo- 0
200 220 240 260 280 300
tion of water molecules confined in smaller pores detected at
T/K
larger Q. So do the lower values of Dt for water at lower tem-
perature. Similarly, the behavior of parameter Dr in Q and (b)
8000
temperature can be explained as more active rotational motion Q=0.3 A-1
Q=0.5 A-1
of water molecules in smaller pores and at lower temperature. Q=0.7 A-1
6000 Q=0.9 A-1
The information about Dr is the third new result obtained by Q=1.1 A-1
Dr/ns-1

Q=1.3 A-1
the rJRM fitting. More discussion about the Dr will be given Q=1.5 A-1
4000 Q=1.7 A-1
in Subsection 3.2.
Q=1.9 A-1

3.2. Discussion 2000

Among the parameters describing the structure of the

sample, we focus on the discussion about a0 . Figure 4(a) 0
200 220 240 260 280 300
shows a0 as a function of Q and T . It can be seen that the T/K
fitted a0 ranges from 19 Å to 2.0 Å, which are smaller than the Fig. 4. The fitted parameters of (a) a0 and (b) Dr versus scattering vector
Q and measured temperature T .
values of the neutron detected space 2π/Q. From Fig. 4(a), it
can be seen that a0 is almost of the same value for the same
Q at different T except for T < 230 K. It can also be seen that
4. Conclusion
the larger the Q value, the smaller the a0 value. The result in- We propose a rJRM model to fit QENS spectra data based
dicates that the radius of pores contained in the MSH sample on JRM model used before by adding the contribution to the
ranges from 2.0 Å to 19 Å for the QENS experimental detec- elastic part from the translational motion of water and taking
tion. This is consistent with the result that the MSH and CSH into account the contribution of neutron scattering lengths in
globules are packed together in the similar way, [18] and there Sears expansion to deal with the rotational motion of water.
are pores (radius < 3 nm) existing in the MSH sample. The fit- The rJRM seems to be better compared to JRM for fitting
ted parameter a0 shows more information compared with the MSH QENS spectra measured at Q values 0.3–1.9 Å−1 and
same MSH QENS data analyzed in Ref. [10] by the new global temperatures from 210 K to 280 K. Seven parameters (A, C,
model, where the fitted a0 is only the average radius of pores Dt , Dr , τ0 , a0 , and x) are extracted, which include not only
in the MSH sample. those referring to the dynamics of confined water such as A,
Among the parameters describing the dynamics of con- Dt , Dr , and τ0 , but also those revealing the structure of the
fined water, we focus on the discussion about Dr . Figure 4(b) sample such as C and a0 .
056105-5
 Chin. Phys. B Vol. 28, No. 5 (2019) 056105
The rJRM model can give well fitted line for the MSH References
QENS spectra, not only for the whole Q range of 0.3–1.9 Å−1 , [1] Bee M 2003 Chem. Phys. 292 121
but also for the whole E range from −120 µeV to 120 µeV. [2] Bordallo H N, Aldridge L P and Desmedt A 2006 J. Phys. Chem. B 110
17966
Comparing with other models used to fit QENS spectra, es- [3] Mamontov E, Wesolowski D J, Vlcek L, Cummings P T, Rosenqvist J,
pecially the new global model, [10] the rJRM can obtain three Wang W and Cole D R 2008 J. Phys. Chem. C 112 12334
[4] Li H, Fratini E, Chiang W S, Baglioni P, Mamontov E and Chen S H
new results for the extracted parameters. The first is the DWF 2012 Phys. Rev. E 86 061505
or MSD that can be fitted directly. The second is nine val- [5] Eckold G, Schober H and Nagler S E 2010 Studying Kinetics with
ues of a0 indicating the radius 2.0–1.9 Å of pores existing in Neutrons — Prospects for Time-Resolved Neutron Scattering (London:
Springer) pp. 19–75
the MSH sample. The third is that the fitted Dr increases with [6] Thomas J J, FitzGerald S A, Neumann D A and Livingston R A 2004
the increasing Q at a measured temperature and reaches peak J. Am. Ceram. Soc. 84 1881
[7] Yi Z, Deng P N, Zhang L L and Li H 2016 Chin. Phys. B 25 106401
values at 230 K, showing that the 230 K is a crossover tem-
[8] Viani A, Zbiri M, Bordallo H N, Gualtieri A F and Macova P 2017 J.
perature for dynamics of confined water based on rotational Phys. Chem. C 121 11355
motion. Although the rJRM is used to fit only one pure MSH [9] Chen S H, Liao C, Sciortino F, Gallo P and Tartaglia P 1999 Phys. Rev.
E 59 6708
sample here, in principle, it is expected useful also to analyze [10] Le P S, Fratini E, Zhang L L, Ito K, Mamontov E, Baglioni P and Chen
QENS spectra from porous materials, where hydrogen atoms S H 2017 J. Phys. Chem. C 121 12826
appear in both solid and liquid phases. Further research on [11] Li H, Zhang L L, Yi Z, Fratini E, Baglioni P and Chen S H 2015 J.
Coll. Interf. Sci. 452 2
new sample is recommended. [12] Deng P N, Yi Z, Zhang L L and Li H 2016 Acta Phys. Sin. 65 106101
(in Chinese)
[13] Bellissent-Funel M C, Chen S H and Zanotti J M 1995 Phys. Rev. E 51
Acknowledgments 4558
[14] Singwi K S and Sjölander A 1960 Phys. Rev. 119 863
We are particularly grateful to Professor Sow-Hsin Chen [15] Sears V F 1966 Can. J. Phys. 44 1299
from Department of Nuclear Science and Engineering in Mas- [16] Wang Z, Le P S, Ito K, LeaTo J B, Tyagi M and Chen S H 2015 J.
sachusetts Institute of Technology for providing help on the Chem. Phys. 143 114508
[17] Sears V F 1966 Can. J. Phys. 44 1279
study of QENS spectra and Peisi Le for providing the MSH [18] Chiang W S, Ferraro G, Fratini E, Ridi F, Yeh Y Q, Jeng U S, Chen S
QENS spectra data fitted already by other method. H and Baglioni P 2014 J. Mater. Chem. A 2 12991

056105-6

View publication stats