http://journa

ite

the
d in
in

he

HelpCommentsWelcome
Journal of

MATERIALS RESEARCH
Phase analysis of seeded and dopedPbMg1///3Nb2///3O3 prepared
by organic solution of citrates
J. C. Carvalho, C. O. Paiva-Santos, M. A. Zaghete, C. F. Oliveira, and J. A. Varela
I.Q.-UNESP. 14800-900, Araraquara, S. P., C.P. 355, Brazil

E. Longo
DQ-UFSCar. 13560-905, S. Carlos, S.P., C.P. 676, Brazil

(Received 15 November 1994; accepted 18 March 1996)

PbMg1/3Nb2/3O3 (PMN) prepared by organic solution of citrates was analyzed by
the Rietveld method to determine the influence of seeds and dopants on the perovsk
and pyrochlore phase formation. It was observed that pyrochlore phase formation
increases with an increase in calcination time when no additives are included during
preparation. It was also observed that a greater amount of perovskite phase appeare
doped or seeded samples. The fraction of perovskite phase increased from 88 mol %
pure sample to,95 mol % in doped and seeded samples calcined at 800±C for 1 h. It
is clear that the addition of dopants or seeds during PMN preparation can enhance t
formation of perovskite phase.
)

s

t
r

c
s

N

t
n
e

e
t
g

l
t

i

x-

e-
d
so-
red
i-
te

d

I. INTRODUCTION

PbMg1/3Nb2/3O3 (PMN) has been studied since
1950.1 It has perovskite (Pe) type structure ABO3

2

at room temperature and characteristics that make
useful as dielectric in multilayer ceramic capacitors.2–5

These characteristics are low sintering temperatu
(,1000±C), high dielectric constant (8000–15,000
and a high electrostriction coefficient6 (0.1%). These
properties depend on the precursors’ purity, the proce
ing method, as well as the reaction time and temperatu
Any change in these parameters can induce the forma
of pyrochlore phase (Py). This phase has a low dielect
constant (130–200)7 that causes the overall dielectri
constant of PMN to decrease. The PMN was fir
prepared by mechanical mixture of oxide precursors3,5

that promote the formation of pyrochlore phase. Aimin
to get PMN free of pyrochlore phase, several PM
preparation methods have been performed.

Swartz and Shrout5 prepared perovskite PMN a
800 ±C by using columbite precursor which was the
prepared at 1000±C; even so, 2% of pyrochlore phas
was present. Ravindranathanet al.2 prepared PMN by
the sol-gel method where Pe phase could be form
at 775±C. In this work, they were able to prepar
pure perovskite type PMN by seeding the material wi
1 wt. % of Pe PMN, also taking advantage of lowerin
the crystallization temperature by 75±C. Liou and Wu8

could enhance the dielectric properties of PMN b
adding PbTiO3 and excess of MgO and PbO. Among a
the methods described in the literature, chemical rou
have shown better performance to decrease the amo
of pyrochlore phase.

In this work PMN has been prepared by organ
solution of citrates9 using the concept of controlled
ls.cambridge.org Downloaded: 12 Mar 2014

J. Mater. Res., Vol. 11, No. 7, Jul 1996
it

re
,

s-
re.
ion
ic

t

g

ed

h

y
l
es
unt

c

nucleation10 through the introduction of BaTiO3 seed
particles or barium/titanium doping solution.

II. EXPERIMENTAL

Precursor reagents used were niobium ammoniac o
alate NH4H2[NbO(C2O4)] 20.5% in niobium (CBMM–
Companhia Brasileria de Minerac¸ão e Metalurgia),
basic magnesium carbonate (MgCO3)4Mg(OH)2 ? 5H2O
(Cinética Qu´ımica), lead acetate (CH3COO)2Pb? 3H2O
(Merck), citric acid C6H8O7 ? H2O (Merck), ethylene gly-
col HOCH2CH2OH (Synth), barium acetate C4H6BaO4

(Vetec), and titanium isopropoxide Ti(OC3H7)4 (Rulsag).
Standard solutions of precursor citrates were pr

pared (magnesium, niobium, barium, and titanium) an
gravimetric analyses were used to standardize these
lutions. Lead acetate aqueous solution was also prepa
and standardized by complexometry with ethylened
amine tetra-acetic acid (EDTA). Magnesium carbona
was also standardized by EDTA.

TABLE I. Preparation conditions for lead magnesium niobate dope
with 1 wt. % of barium/titanium Pechini solution or seeded with
1 wt. % of BaTiO3, and the codes used for them in the text.

Temperatures±Cd Calcination time (h) Additive type Code

700 1.5 Seed 15S7
700 1.5 Dopant 15D7
700 1.5 None 15N7
700 3.0 Seed 30S7
700 3.0 Dopant 30D7
700 3.0 None 30N7
800 1.0 Seed 10S8
800 1.0 Dopant 10D8
800 1.0 None 10N8
IP address: 200.145.174.165

 1996 Materials Research Society 1795

http://www.mrs.org/publications/jmr/comments.html
http://journals.cambridge.org


J. C. Carvalho et al.: Phase analysis of seeded and doped PMN

u

a

n
i-
if-

ial
s

ar-
y.

d
ize
s

ic
r.

d
l
as
er
of
in
n
e

ed

ed
ite
t

s
-
t

-

FIG. 1. X-ray diffractograms of all samples.

Stoichiometric mixtures of respective cation sol
tions were stirred at 90±C until complete chelation.
At this point seeds or dopants were introduced in
the solution, as described in the next paragraph,
http://journals.cambridge.org Downloaded: 12 Mar 2014

1796 J. Mater. Res., Vol.
-

to
nd

then the temperature was increased to 140±C until the
complete polymerization that results from sterificatio
among metallic citrates and ethylene glycol in acid env
ronment. The so-formed resin was characterized by d
ferential thermal analysis and infrared spectroscopy.11,12

The polymer was calcined at 350–400±C in order to
eliminate the organic matter, and the resultant mater
was pulverized in an agate mortar. This powder wa
calcined at different temperatures and times, and ch
acterized by x-ray diffraction and infrared spectroscop

Barium and titanium citrate solubilized in ethylene
glycol with stoichiometry 1 : 1 were prepared to be use
as dopant. Gravimetric analysis was used to standard
this solution. Weighted amount of this solution wa
included to the PMN resin at 90±C under stirring, in
order to represent 1 wt. % of the entire ferroelectr
mass, after calcination and elimination of the polyme
The remaining solution was polymerized at 140±C and
heated at 400±C, pulverized, and calcined at 800±C for
1 h. The resulting powder was characterized by XRD an
IR, and then ground in a ball mill (with ethylene glyco
as medium) for 72 h. The resulting suspension w
then classified by centrifugation, and particles small
than 0.1 mm were selected as seeds. An amount
this suspension representing 1 wt. % of seed particles
relation to the entire ferroelectric mass, after calcinatio
and elimination of the polymer, was introduced into th
PMN resin at 90±C while stirring. In both cases the
PMN-based polymers were calcined and characteriz
by x-ray diffraction and infrared spectroscopy.

BaTiO3 composition was chosen as dopant and se
because it is a ferroelectric material and has a perovsk
type structure, with lattice dimension very close to tha
of PMN.

X-ray diffraction data was obtained in a Siemen
D-5000 model equipment with monochromatized cop
per radiation obtained by 40 kV and 30 mA filamen
current. The Rietveld method13 program used was the
DBWS9006-PC release 12.8.9114 and the pseudo-Voigt
function15 was applied to the refinements. The crys
tained in

2

0

TABLE II. Lattice dimensions, molar phase percentage, and agreement factor indexes. Enclosures are the standard deviations ob
the computations.

Perovskite Pyrochlore Refinement indexes

Samples a (Å) mol % a (Å) mol % Rwp S

15N7 4.0487(2) 85.35(2) 10.6056(7) 14.65(2) 10.66 1.55
15D7 4.0474(3) 91.77(3) 10.6051(8) 8.23(2) 11.12 1.59
15S7 4.0485(2) 86.61(2) 10.6157(6) 13.39(2) 11.16 1.5
30N7 4.0475(2) 80.84(2) 10.6015(7) 19.16(3) 9.37 1.51
30D7 4.0455(3) 93.18(3) 10.599(1) 6.82(2) 11.61 1.84
30S7 4.0477(3) 89.30(3) 10.6073(9) 10.70(3) 10.32 1.6
10N8 4.0467(2) 88.04(2) 10.5962(6) 11.96(2) 10.02 1.71
10D8 4.0452(2) 94.55(3) 10.599(1) 5.45(2) 12.67 1.77
10S8 4.0459(2) 95.01(2) 10.605(1) 4.99(2) 10.83 1.75
IP address: 200.145.174.165

11, No. 7, Jul 1996

http://journals.cambridge.org


J. C. Carvalho et al.: Phase analysis of seeded and doped PMN

/

a

of

y

r-
s
re

d
e
e

e
he
er
at
re

by

in
y.
in
e
al
y

r
tors
FIG. 2. Final Rietveld plot of PMN doped with 1 wt. % of barium
titanium in Pechini solution, calcined at 800±C for 1 h: (1) observed,
y0; (—) calculated,yc ; (— ) residual,y0 2 yc; (j) Bragg positions.

tal structure parameters used for perovskite (Pe) ph
were16 space groupPm3m, a ­ 4.0441 Å, with the
atomic positions Pb (0.027, 0.027, 0.0697), MgyNb
(0.523, 1y2, 1y2), and O (0.540, 1y2, 0). The crystal
structure parameters used for pyrochlore (Py) pha
were17 space groupFd3m, a ­ 10.6029Å, with the

FIG. 3. Final Rietveld plot of pure PMN, calcined at 700±C for 1.5 h:
(1) observed,y0; (—) calculated,yc ; (—) residual,y0 2 yc ; (j)
Bragg positions.
http://journals.cambridge.org Downloaded: 12 Mar 2014

J. Mater. Res., Vol. 1
se

se

FIG. 4. Perovskite molar percentage obtained by Rietveld method
all PMN samples.

atomic positions Pb (1y2, 1y2, 1y2), MgyNb (0, 0, 0), O
(0.3175, 1y8, 1y8), and O0 (3y8, 3y8, 3y8). Chemical
composition of the pyrochlore phase, determined b
Wakiya et al.,17 is Pb1.86Mg0.24Nb1.76O6.5.

Infrared spectrums were obtained in Perkin-Elme
567 and Nicolet FTIR spectrometers. All condition
considered to prepare powders used in this work a
listed in Table I.

III. RESULTS AND DISCUSSIONS

All samples considered in Table I were analyze
using x-ray diffraction. Figure 1 shows that the relativ
intensity of (012) perovskite peak to the (044) pyrochlor
peak (around 2u ­ 50± indicated in the figure) changes
accordingly to the kind of heat treatment applied to th
sample. In pure samples (15N7, 3ON7, and 10N8) t
relative peak intensity of pyrochlore phase is great
than that of other samples. This is an indication th
the proportion of pyrochlore phase is greater in pu
samples.

A precise quantitative phase analysis was done
using the Hill and Howard18 method. This method is
based on the fact that the scale factor obtained
Rietveld method is proportional to the pattern intensit
Also, the relative intensity of peaks for the phases
the pattern is proportional to their relative amount in th
sample, so the relative mass fraction is also proportion
to the relation of scale factor for each phase, given b
Eq. (1).

Wi ­ sSVZMdi

.X
j

sSVZMdj , (1)

whereS is the scale factor,V is the unit cell volume,
and ZM is the weight of the unit cell in atomic units
(number of formula units,Z, per cell times the atomic
weight, M, of the formula unit).18,19

Listed in Table II are the cell parameters, mola
percentage of each phase, and the agreement fac
IP address: 200.145.174.165

1, No. 7, Jul 1996 1797

http://journals.cambridge.org


J. C. Carvalho et al.: Phase analysis of seeded and doped PMN

e

i

f

%

e

e
r
e

h

i

d

g

e
d
er

at
t
r

at

n
of

d
d

as defined by Scott20 (Rwp, S). Figures 2 and 3 show
the results of refinement of two different samples (r
spectively 15N7 and 10S8). The calculated (yc), the
observed (y0), and the residual (y0 2 yc) diffractograms
indicate that the refinement was very satisfactory. The
diffractograms show that there is no other phase bes
the perovskite and pyrochlore. In fact, the diffractogram
for all samples considered in this study show (Fig. 1) th
only these two phases are present.

Figure 4 and Table II show the evolution o
perovskite phase with calcination conditions. For
pure sample treated at 700±C, the amount of Pe phase
decreases with the calcination time, from 85.35(2) mol
when treated for 1.5 h to 80.84(2) when treate
for 3 h. However, for the PMN powder calcined a
800 ±C for 1 h, the amount of Pe phase increas
from 88.04(2) mol % for pure sample to 94.55(3
and 95.01(2) mol % for doped or seeded sampl
respectively. These are significant increases, conside
the estimated standard deviation obtained in the Rietv
refinements, and justify the use of seeds or dopan
Figure 4 also shows that the effect of seeds on t
formation of perovskite phase is more significant fo
higher calcination temperature (800±C).

Figure 5 shows the variation in 2u of full width at
half maximum (FWHM) to the phases of all sample
studied. This figure indicates that perovskite phase h
sharper peaks than pyrochlore phase for all samp
considered, indicating greater crystalline size for pe
ovskite phase. It is also observed that the crystalline s
for perovskite phase increases from 700±C to 800±C.

FIG. 5. Full width at half maximum (degrees) of perovskite an
pyrochlore phases.
http://journals.cambridge.org Downloaded: 12 Mar 2014

1798 J. Mater. Res., Vol. 1
-

se
de
s

at

a

d
t
s

)
s,
ing
ld
ts.
e

r

s
as
les
r-
ze

The FWHM21 for perchlorate phase presents a strikin
variation with 2u at 700±C. This can indicate the
presence of inhomogeneity or defect in its lattice. Th
compositional fluctuation, if present, can be explaine
by the presence of carbonates at temperatures low
than 800±C, as indicated in Fig. 6. This figure shows
the infrared spectroscopy of pure sample calcined
750 ±C for 3 h. The observed band in this figure a
1400 cm21 is characteristic of carbonate stretching. Fo
samples calcined at 800±C for 1 h, the infrared spectrum
showed no characteristic band at 1400 cm21, indicating
that the carbonates have vanished after calcination
this temperature (Fig. 7).

The formation of carbonates during calcinatio
of powders prepared by using the decomposition
polyester has been observed by several authors.10,22–25

During the polymer decomposition, intermediates lea
titanium and lead zirconium carbonates are forme

FIG. 6. Infrared spectrum of pure PMN calcined at 750±C for 3 h.

FIG. 7. Infrared spectrum of calcined powders at 800±C for 1 h.
IP address: 200.145.174.165

1, No. 7, Jul 1996

http://journals.cambridge.org


J. C. Carvalho et al.: Phase analysis of seeded and doped PMN

m

s
M
c

r

n
o

t

c
l

m

c.

c.

y

s.

.

s,
of

s

nd

m.

c.

s,
before the nucleation of PbZrxTi12xO3 (PZT) phase.10,22

These two carbonates have different reactivities that le
to formation of nonhomogeneous PZT phase. In the sa
way, the formation of PMN phase by decompositio
of the polyester containing Pb, Mg, and Nb can al
lead to the formation of PM and PN carbonates. P
carbonate is more stable than PN carbonates which
lead to preferential decomposition of the latter, favorin
the formation of P3N pyrochlore phase, especially fo
long time at low temperature calcination (700±C). This
is in agreement with the results of this study (Fig. 4).

The use of seed BaTiO3 particles as well as stoichio-
metric BayTi solution seems to favor the decompositio
of these mixed carbonates leading to the maximizati
of the perovskite phase formation.26

IV. CONCLUSIONS

From results of this study it is concluded that bo
BaTiO3 seeds and stoichiometric BayTi solutions favor
the formation of PMN perovskite phase. The existen
of PN and PM carbonate phases for temperatures be
800 ±C seems to limit the perovskite phase formatio
for pure PMN composition.

ACKNOWLEDGMENTS

The authors acknowledge FAPESP, FINEPyPADCT,
and CNPq for granting this research.

REFERENCES

1. N. Kim, D. A. McHenry, S-J. Jang, T. R. Shrout, J. Am. Ceram
Soc. 73 (7), 923–928 (1990).

2. P. Ravidranathan, S. Komarmeni, and R. Roy, J. Am. Cera
Soc. 73 (4), 1024–1025 (1990).

3. H. S. Horowitz, J. Am. Ceram. Soc.71 (5), 250–251 (1988).
4. H. V. Anderson, M. J. Pennel, and J. P. Guha, Adv. Ceram.21,

91 –98 (1987).
5. S. L. Swartz and T. R. Shrout, Mater. Res. Bull.17, 1245 –1250

(1982).
http://journals.cambridge.org Downloaded: 12 Mar 2014

J. Mater. Res., Vol. 1
ad
e

n
o

an
g

n

h

e
ow
n

.

.

6. K. Uchino, Ceram. Bull.65 (4), 647–652 (1986).
7. J. Chen and J. P. Harmer, J. Am. Ceram. Soc.73 (1), 68–73

(1990).
8. Y-C. Liou and L. Wu, J. Am. Ceram. Soc.77 (12), 3255 –3258

(1994).
9. M. P. Pechini, U.S. Patent 3 330 697 (1967).

10. M. A. Zaghete, PhD Thesis, Federal University of São Carlos,
Brazil (1993).

11. J. C. Carvalho, M. A. Zaghete, J. A. Varela, and E. Longo, Pro
35± Congresso Brasilerio de Cerâmica e III Iberoamericano de
Cerâmica, Vidro e Refrat´arios, 144–148 (1991).

12. J. C. Carvalho, M. A. Zaghete, J. A. Varela, and E. Longo, Pro
36± Congresso Brasilerio de Cerˆamica, 157–164 (1992).

13. H. M. Rietveld, J. Appl. Cystallogr.2, 65 –71 (1969).
14. A. Sakthivel and R. A. Young,User’s Guide to Programs

DBWS9006 and DBWS9006-PC for Rietveld Analysis of x-ra
and Neutron Diffraction Patterns,School of Physics, Georgia
Institute of Technology, Atlanta, GA (1992).

15. R. A. Young and D. B. Wiles, J. Appl. Crystallogr.15, 430–438
(1982).

16. A. Verbaere, Y. Piffard, Z. G. Ye, and E. Husson, Mater. Re
Bull. 27 (10), 127–134 (1992).

17. N. Wakiya, A. Saiki, N. Shinozaki, and N. Mizutani, Mater. Res
Bull. 28 (2), 137–143 (1993).

18. R. J. Hill and C. J. Howard, J. Appl. Crystallogr.20, 467–474
(1987).

19. R. A. Young, A. Sakthivel, T. S. Moss, and C. O. Paiva-Santo
User’s Guide to Programs DBWS-9411 for Rietveld Analysis
x-ray and Neutron Diffraction Patterns,School of Physics, Geor-
gia Institute of Technology, Atlanta, GA (1995), pp. 15–16.

20. H. G. Scott, J. Appl. Crystallogr.16, 159–163 (1983).
21. G. Cagliot, A. Paoleti, and F. P. Ricci, Nucl. Instrum. Method

35, 223–228 (1958).
22. M. A. Zaghette, C. O. Paiva-Santos, J. A. Varela, E. Longo, a

Y. P. Mascarenhas, J. Am. Ceram. Soc.75 (8), 2088 –2093
(1992).

23. E. R. Leite, C. M. G. Sousa, E. Longo, and J. A. Varela, Cera
Int. 21 (3), 143–152 (1995).

24. P. P. Phule and S. H. Risbud, J. Mater. Sci.25 (2B), 1169 –1183
(1990).

25. S. Kumar, G. L. Messing, and W. B. White, J. Am. Ceram. So
76 (3), 617–624 (1993).

26. J. C. Carvalho, J. A. Varela, M. A. Zaghete, M. Cilense, W. C. La
and C. O. Paiva-Santos, unpublished.
IP address: 200.145.174.165

1, No. 7, Jul 1996 1799

http://journals.cambridge.org