Journal of ELECTRONIC MATERIALS

DOI: 10.1007/s11664-016-4666-3
Ó 2016 The Minerals, Metals & Materials Society

Effects of Lu and Ni Substitution on Thermoelectric Properties

of Ca3Co4O9+d

WENCHAO YANG,1 HAOJI QIAN,1 JINYU GAN,1 WEI WEI,1

ZHIHE WANG,2 and GUODONG TANG1,3

1.—School of Materials Science and Engineering, Nanjing University of Science and Technology,
Nanjing 210094, China. 2.—National Laboratory of Solid State Microstructures, Department of
Physics, Nanjing University, Nanjing 210093, China. 3.—e-mail: [email protected]

Effects of (Lu, Ni) co-doping on the thermoelectric properties of Ca3Co4O9+d

(CCO) have been systematically investigated from 20 K to 350 K. The elec-
trical resistivity and thermopower of (Lu, Ni) co-doped samples increase, while
their thermal conductivity is significantly depressed as compared to that of
pristine CCO. The figure of merit (ZT) of co-doped samples is higher than
those of Lu-doped samples and pristine CCO. A maximum ZT of 0.0185 is
achieved at 350 K for Ca2.9Lu0.1Co3.9Ni0.1O9+d. We demonstrate that the
simultaneous increase of spin entropy and phonon scattering induced by (Lu,
Ni) co-doping boosts ZT of CCO. This study indicates that (Lu, Ni) co-doping
may promise an effective way to improve thermoelectric properties of the CCO
system.

Key words: Thermoelectric materials, Ca3Co4O9+d, thermoelectric

properties, spin entropy, phonon scattering

INTRODUCTION subsystems provide a way to control the electronic

and thermal transport properties separately. How-
Thermoelectric materials, which involve a con-
ever, for practical application, the thermoelectric
version between thermal and electrical energy, are
performance needs to be further improved.8–11
expected to play increasingly important roles in
So far, partial substitution of different elements
meeting energy challenges in the future.1 The
at the Ca site12–23 or Co site24–28 is a method used to
conversion efficiency can be determined by the
try to improve the thermoelectric properties of CCO.
figure of merit ZT = S2T/qj, where T, q, S, and j
However, very few groups study the effects of
are the absolute temperature, the electrical resis-
simultaneous substitution of two different elements
tivity, the thermopower, and the thermal conduc-
at Ca and Co sites.29–39 Our group previously
tivity, respectively. Recently, the misfit-layered
demonstrated that (Ce, Ni) co-doping considerably
cobalt oxides have received much attention due to
enhanced the thermoelectric performance of CCO
their large thermopower, low thermal conductivity,
compared with those of pure and Ce single-doped
and good chemical stability.1,2 In this family, envi-
CCO.40 We also studied the thermoelectric proper-
ronment-friendly Ca3Co4O9+d (CCO) system exhibit-
ties of the Ca3xLuxCo4O9+d system and found that
ing excellent thermoelectric performance and high
partial Lu substitution of Ca is an effective
temperature stability has been considered to be a
approach for improving the thermoelectric proper-
promising candidate thermoelectric materials.3–7
ties of the CCO system.41 In the CCO system, the
CCO consists of two alternating subsystems, a rock
large thermopower values were found to originate
salt type Ca2CoO3 layer and a CdI2 type CoO2 sheet,
from the spin entropy.42 In layered cobalt oxides
each having the same a, c, and b parameters, but a
with strong electron–electron interaction, the ele-
different lattice parameter b.4 These two different
mentary charge-transport process is the hopping of
a hole from Co4+ to Co3+, and a large electron–
electron on-site repulsion excludes double occu-
(Received August 17, 2015; accepted May 12, 2016) pancy of a site by the holes. Because this process
 Yang, Qian, Gan, Wei, Wang, and Tang

converts the Co4+ (s = 1/2) to a Co3+ (s = 0) and vice properties (thermopower, thermal conductivity, elec-
versa, a spin of 1/2 along with the hole is also trical resistivity) were measured using the physical
transferred, which implies a transfer of the spin property measurement system (PPMS, PPMS-9T EC-
entropy r = kBln 2. The spin degrees are predicted II) from 20 K to 350 K.
to produce a large contribution of the Heikes
formula43: RESULTS AND DISCUSSION
l r Figure 1 shows XRD patterns of Ca3xLux-
Q¼ ¼ ; ð1Þ
eT e Co4yNiyO9+d samples. All diffraction peaks match
to the standard JCDPDS card (no. 21-0139) of CCO.
where l is the chemical potential, and r (the entropy We show the enlarged (002) diffraction peak and
per electron) equals kBln(gsgc) with gs and gc the (004) peak in the inset of Fig. 1. It is shown that the
spin and configuration degeneracies, respectively. A enlarged (002) and (004) diffraction peaks of co-
previous study suggested that Ni doping can doped samples shift to higher 2h values in compar-
provide a new hopping model for transporting the ison to the pristine CCO because Lu3+ ion (0.98 Å)
spin entropy, and as a result the thermopower has a similar ionic radius to Ca2+ (1.00 Å) ion.46
increases.44 Thus, simultaneous substitution of Lu XRD analysis indicates that Lu and Ni are success-
and Ni at Ca and Co sites may increase ther- fully doped into the lattices of CCO. The morphology
mopower due to their combined effects on the spin of Ca3xLuxCo4yNiyO9+d is characterized by SEM
entropy enhancement. Meanwhile, substitutions of (Fig. 2). It is found that the samples are uniform
Ni at Co site and heavy metal element Lu at Ca site and small pores are present in the matrixes. The
cause effective phonon scattering, causing the ther- particles are combined closely through sintering
mal conductivity to decrease.45 Therefore, it is of and twice-annealing. The excellent crystallinity of
great interest to investigate the effects of (Lu, Ni) specimens and the clean surfaces are also identified
co-doping on the thermoelectric properties of CCO. by the SEM images. Grain sizes of our samples by
In this work, we investigate systematically the sol–gel method are smaller than those of other
effects of (Lu, Ni) co-doping on the thermoelectric materials prepared by solid state reaction.47 The
properties of CCO. It is found that (Lu, Ni) co- densities of our samples are 3.8383 g/cm3 (x = 0.1,
doping is more effective than single rare earth y = 0.1), 3.8260 g/cm3 (x = 0.1, y = 0.2), 3.8568 g/
elements doping in improving thermoelectric per- cm3 (x = 0.2, y = 0.1), and 3.8093 g/cm3 (x = 0.2,
formance of CCO. The simultaneous increase of spin y = 0.2).
entropy and phonon scattering induced by (Lu, Ni) Figure 3 shows the temperature dependence of
co-doping boosts ZT of CCO. This work indicates resistivity (q) for Ca3xLuxCo4yNiyO9+d. It demon-
that (Lu, Ni) co-doping promises an effective way for strates that (Lu, Ni) co-doping leads the q to
improving thermoelectric properties of CCO system. increase. A reentrant is observed for all samples,
indicating a typical semiconductor behavior. The
EXPERIMENTAL transition temperature (Tmin) is indicated by
arrows, shown in Fig. 3. As (Lu, Ni) doping level
Ca3xLuxCo4yNiyO9+d samples were prepared by increases, Tmin shifts to higher temperature. For
a sol–gel chemical solution route as follows: the semiconductor, q is in inverse proportion to the
stoichiometric CaCO3, Co(NO3)2Æ6H2O, Lu(NO3)3Æ
6H2O, Ni(NO3)2Æ6H2O and citric acid monohydrate
were dissolved in distilled water. The intensive
mixed solution was dehydrated at 335 K for 12 h.
The resulting gel was heated at 473 K for 3 h to
remove the moisture. The obtained carbonaceous
xerogel was crushed and calcined at 823 K for
another 3 h in air. The obtained powders were
cooled down to room temperature and sintered at
1173 K for 12 h under O2 atmosphere in corundum
crucibles (fused porous alumina). Finally, the pow-
ders were pressed into pellets at 15 MPa and
annealed at 1173 K for another 36 h under O2
atmosphere.
X-ray diffraction (XRD, Bruker D8) with Cu Ka
radiation was carried out for phase identification.
Scanning electron microscope (SEM) investigations
were conducted with a FEI Quanta 250F. X-ray
photoemission spectroscopy (XPS, PHI 5000 Ver-
saProbe) were performed with a surface sciences Fig. 1. XRD patterns of Ca3xLuxCo4yNiyO9+d. The left inset part
instruments spectrometer equipped with a monochro- shows the enlarged (002) diffraction peak and the right shows the
matized Al Ka radiation. All the thermoelectric enlarged (004) peak.
Effects of Lu and Ni Substitution on Thermoelectric Properties of Ca3Co4O9+d

Fig. 2. SEM images showing the fractured cross sections of Ca3xLuxCo4yNiyO9+d: (a) x = 0.1, y = 0.1; (b) x = 0.1, y = 0.2; (c) x = 0.2, y = 0.1;
(d) x = 0.2, y = 0.2.

Ni) lead to an increase in impurity scattering.40

Meanwhile, the substitution of Lu3+ for Ca2+
increases m*.41 Consequently, the l decreases.
Meanwhile, Lu3+ doped into Ca2+ sites lead n to
reduce. As a result, the increase of q is observed in
(Lu, Ni) co-doped samples.
The temperature dependence of thermopower (S)
from 20 K to 350 K for Ca3xLuxCo4yNiyO9+d is
shown in Fig. 4. S of all samples increases with the
increasing temperature. S of co-doped samples is
significantly larger than that of the pristine CCO. In
particular, S of Ca2.8Lu0.2Co3.8Ni0.2O9+d reaches up
to 137.8 lV/K at 350 K. The results indicate that
(Lu, Ni) co-doping can effectively enhance S of CCO.
The increase in S can be mainly attributed to the
Fig. 3. Temperature dependence of resistivity for Ca3xLux-
Co4yNiyO9+d.
enhanced spin entropy.49 In the high temperature
limit, the spin entropy contribution to the ther-
mopower in the strong correlation system can be
expressed by the Heikes formula,50
carrier concentration (n) and the carrier mobility
  
(l).48 l is given by: kB g3 c 
  S ¼  ln ; ð3Þ
4 5 es0 ðkTÞr e g4 1  c
l ¼ 1=2 C r þ ; ð2Þ where g3 and g4 are the spin orbital degeneracies for
3p 2 m
Co3+ and Co4+ ions, respectively, c is Co4+ concen-
where m* and s0 are carrier effective mass and tration, kB is the Boltzmann constant and e is the
relaxation time, and e, k, T, C, and r are electron electron charge. The spin orbital degeneracies
charge, Boltzmann’s constant, temperature, gamma g3 = 1 and g4 = 6 are determined according to the
function, and scattering law parameter, respec- low-spin electronic configurations of Co3+ and Co4+
tively. Thus, m* and scattering probability deter- ions.50–55 Co 2p XPS spectra of Ca3xLux
mine the l. It is considered that impurity ions (Lu, Co4yNiyO9+d are shown in Fig. 5. The main peaks
 Yang, Qian, Gan, Wei, Wang, and Tang

Fig. 4. Temperature dependence of thermopower for Ca3xLux-

Co4yNiyO9+d. Fig. 6. Temperature dependence of the thermal conductivity for
Ca3xLuxCo4yNiyO9+d and the inset shows the temperature
dependence of the lattice thermal conductivity for these samples.

entropy due to spin entropy competition mecha-

nism. Co and Ni make joint contribution to the
enhanced spin entropy. A new hopping model
provided by magnetic Ni ions further contributes
to the increase of the total spin entropy. At low
temperature, the undoped Ca3Co4O9+d has a higher
S than doped samples with x = 0.1 and y = 0.2. The
reason needs to be studied further.
Figure 6 shows the temperature dependence of
the thermal conductivity (j) for Ca3xLux-
Co4yNiyO9+d. j for all samples increases with
increasing temperature. It is found that j of (Lu,
Ni) co-doped samples is much lower than that of
pristine CCO. j of Ca2.9Lu0.1Co3.9Ni0.1O9+d sample
is 1.077 W/km at 350 K. In general, the total j
consists of the electronic thermal conductivity (je)
and lattice thermal conductivity (jL). The value of je
can be calculated using equation6:
Fig. 5. Co 2p XPS spectra of Ca3xLuxCo4yNiyO9+d. je ¼ L0 Tr; ð5Þ
where the Lorentz number (L0) is equal to
(A) with the binding energy of ca. 779.5 eV, 2.44 9 108 V2/K2, r and T are the electrical con-
788.9 eV, 794.6 eV can be identified as Co3+, while ductivity and the absolute temperature, respec-
the other peaks (B) at ca. 780.5 eV, 782.5 eV, 796 eV tively. jL of all samples have been calculated by
are assigned to Co4+.51 We calculate the ratio of subtracting je from j, which is shown in the inset of
fitting peak area of Co4+ to the total Co sites and Fig. 6. It indicates that jL is the main source of j.
find that c is reduced by (Lu, Ni) co-doping. Accord- Therefore, the decrease of j mainly originates from
ing to Eq. 3, the spin entropy enhancement from Co the reduction of phonon contribution. As shown in
ions can be expected. It is reported that Ni2+ and Fig. 2, the size of the particles decreases with
Ni3+ coexist in the Ni doped CCO materials.44 The increasing doping level. Smaller grains cause more
coexistence of Ni2+ and Ni3+ ions provides a new effective phonon scattering. Meanwhile, Ni and Lu
hopping model for transporting the spin entropy substitutions contribute the increase of impurity
and induces S to increase, as given by42: scattering. Therefore, both phenomena are respon-
    sible for the observed low thermal conductivity.
kB gðCo3þ Þ h c i kB gðNi2þ Þ 1  d The thermoelectric figure of merit (ZT) as a
S¼  ;
e gðCo4þ Þ 1  c e gðNi3þ Þ d function of temperature for Ca3xLuxCo4yNiyO9+d
ð4Þ samples is presented in Fig. 7. ZT of all samples
increases with increasing temperature. ZT of (Lu,
where g(Ni2+) and g(Ni3+) are the spin orbital Ni) co-doped samples is improved as compared with
degeneracy for Ni2+ and Ni3+ ions, d is Ni2+ concen- the pristine CCO. The previous data41 of Lu doped
tration. Ni contributes to the increase of the spin samples is given in Fig. 7 for comparison. ZT of (Lu,
Effects of Lu and Ni Substitution on Thermoelectric Properties of Ca3Co4O9+d

3. K. Koumoto, I. Terasaki, and R. Funahashi, MRS Bull. 31,

206 (2006).
4. A.C. Masset, C. Michel, A. Maignan, M. Hervieu, O.
Toulemonde, F. Studer, B. Raveau, and J. Hejtmanek,
Phys. Rev. B 62, 166 (2000).
5. Y. Wang, Y. Sui, J.G. Cheng, X.J. Wang, J.P. Miao, Z.G. Liu,
Z.N. Qian, and W.H. Su, J. Alloys Compd. 448, 1 (2008).
6. N. Prasoetsopha, S. Pinitsoontorn, A. Bootchanont, P.
Kidkhunthod, P. Screpusharawoot, T. Kamwanna, V.
Amornkibamrung, K. Kurosaki, and S. Yamanaka, J. Solid
State Chem. 204, 257 (2013).
7. M. Shikano and R. Funahashi, Appl. Phys. Lett. 82, 1851
(2003).
8. R. Tian, R. Donelson, C.D. Ling, P.E.R. Blanchard, T.
Zhang, D.W. Chu, T.T. Tan, and S. Li, J. Phys. Chem. C
117, 13382 (2013).
9. Y. Masuda, D. Nagahama, H. Itahara, T. Tani, W.S. Seo,
Fig. 7. The thermoelectric figure of merit (ZT) as a function of tem- and K. Koumoto, J. Mater. Chem. 5, 1094 (2003).
perature for Ca3xLuxCo4yNiyO9+d. 10. M. Mikami, N. Ando, and R. Funahashi, J. Solid State
Chem. 178, 2186 (2005).
11. T.F. Yin, D.W. Liu, Y. Ou, F.Y. Ma, S.H. Xie, J.F. Li, and
J.Y. Li, J. Phys. Chem. C 114, 10061 (2010).
Ni) co-doped samples is higher than Lu doped 12. Y. Wang, L.X. Xu, Y. Sui, X.J. Wang, and J.G. Cheng, W.H
samples. ZT of Ca2.9Lu0.1Co3.9Ni0.1O9+d reaches up Su. Appl. Phys. Lett. 97, 062114 (2010).
to 0.0185 at 350 K. These results reveal that (Lu, 13. G.D. Tang, Z.H. Wang, X.N. Xu, L.K. Qiu, and Y.W. Du, J.
Ni) co-doping is more effective than single rare Appl. Phys. 107, 053715 (2010).
14. Y. Song and C.W. Nan, Physica B 406, 2919 (2011).
earth elements doping in improving thermoelectric 15. F.P. Zhang, X. Zhang, Q.M. Lu, J.X. Zhang, Y.Q. Liu, and
performance of CCO.2 The simultaneous increase of G.Z. Zhang, Solid State Sci. 13, 1443 (2011).
spin entropy and phonon scattering induced by (Lu, 16. G. Constantinescu, Sh Rasekh, M.A. Torres, J.C. Diez, M.A.
Ni) co-doping boosts ZT of CCO. Madre, and A. Sotelo, J. Alloys Compd. 577, 511 (2013).
17. S.W. Li, R. Funahashi, I. Matsubara, K. Ueno, S. Sodeoka,
and H. Yamada, Chem. Mater. 12, 2424 (2000).
CONCLUSIONS 18. S.W. Li, R. Funahashi, I. Matsubara, H. Yamada, K. Ueno,
We have investigated the effects of (Lu, Ni) co- and S. Sodeoka, Ceram. Int. 27, 321 (2001).
19. I. Matsubara, R. Funahashi, T. Takeuchi, and Y. Zhou,
doping on the thermoelectric properties of CCO. The International Conference of TE, Vol. 172 (2001).
thermopower and resistivity of (Lu, Ni) co-doped 20. I. Matsubara, R. Funahashi, T. Takeuchi, and S. Sodeoka,
samples increase, while their thermal conductivity J. Appl. Phys. 90, 462 (2001).
decreases significantly as compared to that of the 21. J. Nan, J. Wu, Y. Deng, and C.W. Nan, Solid State Com-
mun. 124, 243 (2002).
pristine CCO. The enhanced thermopower can be 22. G. Xu, R. Funahashi, M. Shikano, I. Matsubara, and Y.
attributed to the increase of the spin entropy. Lu Zhou, Appl. Phys. Lett. 80, 3760 (2002).
and Ni co-doping allows effective phonon scattering, 23. H.Y. Xu, H. Wang, Y.Q. Meng, and H. Yan, Solid State
resulting in the decrease of thermal conductivity. Commun. 130, 465 (2004).
The simultaneous increase of the spin entropy and 24. Y. Wang, Y. Sui, X.J. Wang, W. Su, and X. Liu, J. Appl.
Phys. 107, 033708 (2010).
phonon scattering induced by (Lu, Ni) co-doping 25. Q. Yao, D.L. Wang, L.D. Chen, X. Shi, and M. Zhou, J.
boosts ZT of CCO. ZT  0.0185 is achieved at 350 K Appl. Phys. 97, 103905 (2005).
for Ca2.9Lu0.1Co3.9Ni0.1O9+d samples, which is much 26. B.C. Zhao, Y.P. Sun, and W.H. Song, J. Appl. Phys. 99,
larger than that of non-doped and Lu doped CCO. 073906 (2006).
27. Y. Wang, Y. Sui, P. Ren, L. Wang, X.J. Wang, W.H. Su, and
This work indicates that co-doping method is a H.J. Fan, Chem. Mater. 22, 1155 (2010).
promising way to improve thermoelectric properties 28. S. Pinitsoontorn, N. Lerssongkram, N. Keawprak, and V.
of CCO. Amornkitbamrung, J. Mater. Sci. Mater. Electron. 23, 1050
(2012).
ACKNOWLEDGEMENTS 29. Y. Ou, J. Peng, F. Li, Z.X. Yu, F.Y. Ma, S.H. Xie, J.F. Li,
and J.Y. Li, J. Alloys Compd. 526, 139 (2012).
This work was supported by the National Natural 30. S. Butt, Y.C. Liu, J.L. Lan, K. Shehzad, B. Zhn, Y.H. Lin,
Science Foundation of China (Nos. 11204134, and C.W. Nan, J. Alloys Compd. 588, 277 (2014).
11164002, and 11364004), Natural Science Founda- 31. Y.Q. Zhou, I. Matsubara, S. Horii, T. Takeuchi, R. Funa-
tion of Jiangsu Province (No. BK2012404), Zijin hashi, and M. Shikano, J. Appl. Phys. 93, 2653 (2003).
32. H. Hao, L.M. Zhao, X. Hu, and Y. Liu, J. Mater. Sci.
Intelligent Program of Nanjing University of Science Technol. 25, 105 (2009).
and Technology, China Postdoctoral Science special 33. M. Senthilkumar and R. Vijayaraghavan, Adv. Mater. Res.
Foundation (No. 2013T60540), and China Postdoc- 584, 162 (2012).
toral Science Foundation (No. 2012M511279). 34. A. Bhaskar, R.Z. Lin, and C.J. Liu, Energy Convers. Man-
ag. 76, 63 (2013).
REFERENCES 35. A. Bhaskar, Y.C. Huang, and C.J. Liu, J. Electron. Mater.
43, 535 (2014).
1. I. Terasaki, Y. Sasago, and K. Uchinokura, Phys. Rev. B 56, 36. A. Bhaskar, Z.R. Lin, and C.J. Liu, J. Electroceram. 32, 269
R12685 (1997). (2014).
2. N.V. Nong, N. Pryds, S. Linderoth, and M. Ohtaki, Adv. 37. A. Bhaskar, Y.C. Huang, and C.J. Liu, J. Mater. Sci. Mater.
Mater. 23, 2484 (2011). Electron. 25, 249 (2014).
 Yang, Qian, Gan, Wei, Wang, and Tang

38. R.M. Tian, R. Donelson, C.D. Ling, P.E.R.B. Blanchard, 47. S. Demirel, A. Aksan, and S. Altin, J. Mater. Sci. Mater.
T.S. Zhang, D.W. Chu, T.T. Tan, and S. Li, J. Alloys Electron. 23, 2251 (2012).
Compd. 615, 311 (2014). 48. H.J. Goldsmid, Introduction to Thermoelectricity, ed. R.
39. A. Bhaskar, Z.R. Yang, and C.J. Liu, Ceram. Int. 41, 10456 Hull (Berlin: Springer, 2009), pp. 31–32.
(2015). 49. G.D. Tang, W.C. Yang, Y.Y. Jiang, Z.C. Wu, and Z.H.
40. G.D. Tang, W.C. Yang, Y. He, and Z.H. Wang, Ceram. Int. Wang, J. Mater. Sci. 50, 1746 (2015).
41, 7115 (2015). 50. W. Koshibae, K. Tsutsui, and S. Maekawa, Phys. Rev. B 62,
41. G.D. Tang, Z.H. Wang, X.N. Xu, L. Qiu, L. Xing, and Y.W. 6869 (2000).
Du, J. Mater. Sci. 45, 3969 (2010). 51. G.D. Tang, F. Xu, D.W. Zhang, and Z.H. Wang, Ceram. Int.
42. G.D. Tang, F. Xu, Y. He, L.Y. Wang, L. Qiu, and Z.H. 39, 1341 (2013).
Wang, Phys. State Solidi B 250, 1327 (2013). 52. G.D. Tang, X.N. Xu, C.P. Tang, Z.H. Wang, Y. He, L. Qiu,
43. Y.Y. Wang, N.S. Rogado, R.J. Cava, and N.P. Ong, Nature L.Y. Lv, L. Xing, and Y.W. Du, Europhys. Lett. 91, 17002
423, 425 (2003). (2010).
44. G.D. Tang, T. Yang, X.N. Xu, C.P. Tang, L. Qiu, Z.D. 53. G.D. Tang, C.P. Tang, X.N. Xu, Y. He, L. Qiu, L.Y. Lv, Z.H.
Zhang, L.Y. Lv, Z.H. Wang, and Y.W. Du, Appl. Phys. Lett. Wang, and Y.W. Du, Solid State Commun. 150, 1706
97, 032108 (2010). (2010).
45. S. Pinitsoontorn, N. Lerssongkram, A. Harnwunggoung, K. 54. S. Butt, W. Xu, W.Q. He, Q. Tan, G.K. Ren, and Y.H. Lin, J.
Kurosaki, and S. Yamanaka, J. Alloys Compd. 503, 431 Mater. Chem. A 2, 19479 (2014).
(2010). 55. E. Meza, J. Ortiz, D. Ruiz-Leon, J.F. Marco, and J.L.
46. R.D. Shannon, Acta Crystallogr. A 32, 751 (1976). Gautier, Mater. Lett. 70, 189 (2012).