B O L E T I N D E L A S O C I E DA D E S PA Ñ O L A D E

Cerámica y Vidrio
A R T I C U L O

Scanning probe microscopy applied to the study of domains and

domain walls in a ferroelectric KNbO3 crystal
J. CANET-FERRER1, L. MARTÍN-CARRÓN1,
J. L. VALDÉS1 and J. MARTÍNEZ-PASTOR1
1
Instituto de Ciencia de los Materiales de la Universidad de Valencia, E-46071 Valencia, Spain.

A commercial Atomic Force Microscope (AFM) and a semi-home made Scanning Near-Field Optical Microscope (SNOM)
have been used to characterize electrically, topographically and optically the domain walls among natural ferroelectric
domains in a KNbO3 crystal. The AFM measurements have been performed with a metallic coated tip in order to detect
electrostatic forces between the polarization field at the ferroelectric surface and the tip. An external electric field has also
been applied between the sample surface and the tip to tune this electrostatic interaction over the atomic forces. In optical
transmission images, acquired under near field conditions, we observe a clear contrast of the signal at the domain walls
between 180º spontaneous polarization domains; while the images of the surface topography, obtained simultaneously, show
a reasonably flat surface of the crystal. The scanning probe microscopy techniques used in this work are valuable tools for the
investigation of ferroelectric materials and, in particular, to characterize the domain walls, without needing a either especial
preparation or damage of the sample surface.

Keywords: KNbO3, ferroelectric, domain-wall, SNOM and AFM

Microscopias de barrido aplicadas al estudio de los dominios y las paredes de dominio en un cristal ferroelectrico de
KNbO3.

Hemos utilizado un Microscopio de Fuerzas Atómicas (AFM) comercial y un Microscopio Óptico de Campo Cercano (SNOM)
semi-casero para caracterizar eléctrica, óptica y topográficamente las paredes de dominio presentes entre los dominios
ferroeléctricos naturales de un cristal de KNbO3. Las medidas de AFM las hemos realizado con una punta recubierta con
metal, para detectar las fuerzas electrostáticas entre los campos de polarización de la superficie ferroeléctrica y la punta.
Además, hemos aplicado campos eléctricos externos entre la superficie de la muestra y la punta, de manera que se pueda
variar la fuerza electrostática en relación a las fuerzas atómicas. En imágenes de transmisión óptica, bajo condiciones de
campo cercano, observamos un claro contraste de la señal en las fronteras entre los dominios ferroeléctricos con polarización
espontánea a 180º, mientras que las imágenes de topografía muestran una superficie prácticamente plana. Las técnicas de
microscopía de prueba usadas en este trabajo se revelan muy valiosas para la investigación de materiales ferroeléctricos y,
en particular, para la caracterización de sus dominios y paredes de dominio, puesto que no es necesario preparar ni dañar la
superficie de la muestra para ser investigada.

Palabras clave: KNbO3, ferroeléctrico, pared de dominio, SNOM y AFM

1. INTRODUCTION

Potassium niobate (KNbO3) belongs to the group of A variety of experimental techniques such as polarizing
perovskite-type ferroelectrics of the barium titanate family. optical microscopy, etching, colloidal decoration, the
Extensive theoretical and experimental studies (1-4) have anomalous dispersion of x-rays, Atomic Force Microscopy
been performed on this material since the discovery of its (AFM), scanning electron microscopy and transmission
ferroelectricity (5), because its out-standing electro-optical, electron microscopy, have been used to study static domains
nonlinear optical and photorefractive properties. In the last (6,9). Electrostatic Force Microscopy (EFM) is a suitable
decade, KNbO3 (KNO) single crystal has received much technique for characterization of ferroelectric nano-domains
attention (4, 6-8) due to the relation existent between the (10). These Scanning Probe Techniques (SPM), in combination
piezoelectric properties and the domain structures. However, with Near Field Optical Microscopy (SNOM), are unique for
many of these properties are not well understood at the characterization of domains and domain walls in ferroelectric
nanometer scale in bulk material and thin films. Some crystals, since they provide the possibility of studying
ferroelectric crystals, for example LiNbO3 and KNbO3, form samples without either a especial preparation of the sample
natural periodic and quasi-periodic domain structures. or damaging the surface, as demonstrated in this work.
Domain wall movement plays a key role in the macroscopic We have incorporated small modifications in a commercial
response. In particular, the pining-depining dynamics, caused AFM to be used as an EFM, in order to study a KNO single
by randomly distributed defects and impurities, is of great crystal sample, which exhibits natural periodic ferroelectric
interest (9). domains. We have also employed AFM and SNOM techniques

218 Bol. Soc. Esp. Ceram. V., 45 [3] 218-222 (2006)

Scanning probe microscopy applied to the study of domains and domain walls in a ferroelectric KNbO3 crystal

to complete the study of the ferroelectric domains at nano- standard probe is replaced by a tip shaped and metal coated
and micro-metric scale, related to some of their basic optical monomode optical fiber. This optical probe is mounted on a
properties and electrostatic behaviour. pitch-fork quartz, whose natural resonance arises at 32.7 kHz
(without fiber) for vibration parallel to the sample surface.
During the scan, the amplitude of such lateral vibration is kept
2. EXPERIMENTAL DETAILS constant by the feedback electronics (Nanotec). The feedback
correction consists in a displacement of the sample along the
At room temperature, KNO has an orthorhombic crystal z-axis to maintain constant the tip-sample distance, so called
structure with space group Amm2 and presents natural periodic gap-width. The surface features are extremely correlated with
ferroelectric domains with 180º spontaneous polarization (8). this z-motion, which is used to form the topographic images.
We have used a commercial AFM (Nanotec, Spain) and a non The sample surface was illuminated with a laser diode of 980
commercial SNOM to characterize electrical, topographically nm coupled to the free end of the optical fiber mounted on the
and optically, respectively, the domain walls in a KNO bulk quartz fork. The light detection was made by a Si-photodiode,
crystal. The SNOM microscope is based on a quartz tuning whose electrical signal is connected to the AD/DA card of our
fork sensor head and piezoelectric streper (and scanning) AFM-electronics after pre-amplified (106-108, depending on
motors from Attocube, adapted to an AFM electronics (also excitation) by a low-noise external current amplifier (Femto).
from Nanotec).
The AFM measurements have been performed in both
contact and non-contact modes, because the interaction 3. RESULTS AND DISCUSSION
depends on the tip-surface distance and also on the tip motion
(11). The Nanotec piezoelectric scanner, with a scan size of 3.1. Electrostatic imaging with AFM.
17.5x17.5x3 µm3, is the base of our sample holder. The electronic
control and digital processing of the signals are carried out by The electrostatic signal measured in a commercial EFM is
the Dulcinea-electronics (Nanotec), where capabilities for that associated to changes on the surface potential induced
Dynamic Force Control and Kelvin Probe Microscopy (EFM) by an AC modulation added to the DC bias applied between
are integrated. A PointProbePlus-cantilever (Nanosensors) the tip and the sample surface, and typically measured by
with appropriate elastic and geometric properties (k = 2.8 a lock-in amplifier. The electrostatic force (tip-sample) in a
N/m and tip radius of 15 nm) has been used to obtain both given point of the surface is proportional to the square of
contact and tapping AFM images of the same scanned area. the electric potential between the tip and the sample. This
We have employed a programmable DC Keithley 230 voltage potential consists in the sum of the bias applied to the tip and
source to externally apply a lateral electric field on the surface the difference of the potential emerging, as a result of moving
(about 100 V/mm). We can also add a DC bias between the the tip to a different potential area (12). In addition, the normal
tip and one of these contacts by means of the AFM electronics. force presents three components due to the electrostatic
The schematic drawn of this experimental setup is shown in interaction: a static force and two oscillating forces, whose
Figure 1. frequencies depend on the AC modulating signal (13). We have
used a more simple approach to measure local electrostatic
forces, in order to obtain some more qualitative conclusions.
In our experiment, a conducting probe (metal coated Si tip)
has been used to observe changes on the surface potential due
to the abrupt variation of the spontaneous polarization across
the domain walls. We have obtained AFM images applying
several combinations of voltages at the contacts. Two different
effects were observed in contact mode with a conducting

Fig. 1- Scheme of experimental setup performed to obtain electrostatic

information. It is based on a typical AFM set up, where two voltage
sources have been added. As can be seen in the figure, the DC voltage
is applied between the tip and the surface. A lateral field can be also
applied by an external voltage source. Both sources were grounded at
the sample holder.

Fig. 2- Typical topography of KNO surface. Left: image obtained at

Our SNOM is capable of recording simultaneously the 100V in lateral field and cero bias in contact mode. Right: the bias field
topography and the optical signal (transmission or reflectivity) applied of 30V, 60V and -60V from top to bottom. These contrasts are
images. The operational principles of this SNOM are similar perfectly reproducible and more intense under higher electric fields,
to that of the AFM working in non-contact mode, but the but are clearly different for both polarizations.

Bol. Soc. Esp. Ceram. V., 45 [3] 218-222 (2006) 219

 J. CANET-FERRER, L. MARTÍN-CARRÓN, J. L. VALDÉS and J. MARTÍNEZ-PASTOR

tip: (1) some reproducible contrasts with different features 3.2. Near-field optical microscopy Characterization.
depending on the analyzed signal (topography, lateral force or
normal force), and (2) a clear electrostatic contrast double-line To complete the study of ferroelectric domains in KNO,
at the domain wall. As shown below, they have been observed we have performed SNOM measurements. The advantage of
under different imaging conditions: without electrostatic field, our Near-Field Microscope is the possibility of acquiring the
applying lateral field and adding a small bias (less than 10 V) SNOM images, containing the optical information, and the
between the tip and one of the contacts. topographical images, simultaneously. In this way, we can
A typical example of these contrasts is shown in Figure correlate the optical information of the sample (transmission
2. The left panel is the topographical image, which was or reflectivity) with the surface topography. Our optical (and
acquired without applying any bias to the tip, but applying topographic) probes reach a resolution better than 100 nm
100V between the two metallic contacts made on the sample along the lateral directions and around 1 nm in the axial one.
surface. The topographic resolution decreases when that When the tip is approached close to the surface, maintaining
lateral electric field is applied, which is attributed to an a constant height about 15 nm, the light arrives to the detector
increase of the normal force component (used as the feedback by tunnelling this gap and going across the sample. As a
signal), because of the presence of the electrostatic repulsion result, the light reaching the Si-photodetector from each point
force. When a bias is added to the tip, also a great number
of reproducible marks are observed. Their contrast increases
applying this bias and their shape depends on the sign of the
electric field (right side images in Figure 2).
All these measurements confirm that the measured double-
line signature is related to the domain wall and the other are
just electrostatic effects, which cannot be associated to the
domain structure. In our opinion, these contrasts could be
produced either by superficial voltages (like spontaneous
polarization), charges, inductive effects and/or tip-sample
capacitive effects. The last ones are interesting since they could
be used to characterize ferroelectric surfaces by analysing the
static component of the tip-surface interaction.
The domain walls seem also to be observed in topographic
images, even without applying external fields (not shown).
They were appearing like protuberances of the surface, which
could be differentiated from surface wrinkles by correlating
with normal force images. Given that the normal force signal
coming from the domain wall is very weak, we are able to
detect better domain walls by applying the lateral electric
field. In this way, in Figure 3 a “dark line“ (attractive force)
is observed close to the domain wall by scanning along the
backward direction, while a “bright line” (repulsive force)
appears by scanning along the forward direction. This fact
(different brightness of the two scanning directions) is the
main reason why we can infer that this is an electrostatic effect Fig. 4- (a) and (b) correspond to topography images obtained simulta-
as well. neously as the optical reflectivity (c) and transmission (d) ones under
near field conditions.

Fig. 3- Domain wall observed by electrostatic interaction. Left: topographic image of the domain wall, which appears without applying field.
Middle and right: backward and forward normal force signal, which exhibit attractive or repulsive contrasts at the domain wall depending on
the scanning direction.

220 Bol. Soc. Esp. Ceram. V., 45 [3] 218-222 (2006)

Scanning probe microscopy applied to the study of domains and domain walls in a ferroelectric KNbO3 crystal

of the sample surface will depend on the tip-surface gap and

on the sample refractive index, since the evanescent fields are
transformed into propagating waves by interaction between
the probe and the sample in the near-field range (14). This is
the reason why the transmission signal usually contains the
optical information as well as the topographic one (15).
Figure 4 shows the SNOM images (transmission and
reflectivity) of the same surface region of the sample. We
have also included the topography belonging to these images
to check that we are scanning the same region (top images
in Figure 4). At first sight, we can see that the reflectivity
image mostly reproduces the topography of the sample
(plus a high number of dust particles), which is not the case
for the transmission image. In this image we observe clear
contrasts, which are attributed to the change of the refractive
index produced at the domain walls (16). The correlation
between optical and mechanical information is considered
one of the key issues in SNOM microscopy, necessary to
interpret correctly optical images. The surface roughness is
usually reproduced in SNOM images and optical information
sometimes becomes a (less intense) second order effect. In
addition, topographic artefacts can be sometimes observed, due
to inappropriate feedback response, sample flatness and other
effects (14-15). This is not the case in the SNOM transmission Fig. 5- Profiles (a) to (d) are extracted from the corresponding images
images measured in KNO, which exhibit a clear contrast in Fig. 4.
associated to the domain walls between 180º spontaneous
polarization domains, and signal changes induced by surface
protuberances (dust particles) become negligible. 4. CONCLUSIONS
In Figure 5 we present some profiles extracted from the
images in Figure 4. We measure changes in the transmitted In contact mode AFM measurements, a coulombian
light larger than 30 mV [see Fig. 5(b)] over an average value repulsion to the tip can be produced by applying a lateral
of around 2 V, that is, the contrast in the transmittance, ∆T/T, electric field. On the one hand, this repulsion explains the
is larger than 1 %, which corresponds to a contrast in the lost of resolution in spite of the enhancement of the contrasts
refractive index, ∆n/n, larger than 2%. The contrast for n has in the topographic images. On the other hand, some domain
been estimated considering a simple expression that relates walls were observed without applying field, which appeared
the optical transmission with the refractive index for normal like a protrusion of the surface. They could be also observed
incidence conditions, as contrasts in normal force images under electric field,
indicating that they are not merely topographic effects. In
the SNOM transmission images we observe a clear contrast
thus it should be considered as an approximation. These of the signal at the domain walls between 180º spontaneous
contrasts appears without having topographic signal, that is, polarization domains, which is of the order of 1 % in the optical
the topographic images obtained simultaneously as the optical transmittance (there are more and less intense throughout
ones show a reasonably flat surface [see Figs. 5(a) and (c)], the surface) and corresponds to a contrast in the refractive
except sharp peaks associated to the dust particles. On the index of around a 2%. These contrasts appear without having
contrary, the topography of the sample is mainly determining topographic signal. We also determined the average separation
the reflectivity image shown in Figure 4 (including similar between domain walls to be around 1500 nm and their typical
sharp peaks as in topography, associated to the dust particles). width 700 nm, approximately. These SPM techniques are very
Furthermore, when we analyse in more detail the reflectivity useful for characterization of domains and domain walls in
profiles [like that shown in Fig. 5(d)], we observe an extra ferroelectric materials in the nanometre scale. Both of them
modulation associated to an optical contrast, of course weaker provide the possibility of studying samples without a special
than that observed in transmission. This optical contrast can preparation of the surface (chemical selective etching, for
be separated from that corresponding to the topographic example), as done to observe periodic domain structures by
features. The fact that this modulation is practically inverse traditional optical microscopy.
to that observed in the transmission profiles (maxima in
transmission correspond to minima in reflectivity) confirm
that the optical contrast in the domain-wall is produced by ACKNOWLEDGEMENTS
a change in the refractive index. Finally, we can extract from
the profiles that the average separation between domain We want to acknowledge C. Zaldo for providing the
walls is around 1500 nm and their typical width 700 nm, sample.
approximately, which can be the origin of the strong second We acknowledge the “Ministerio de Educación y Ciencia”
harmonic generation typically observed in this kind of samples for financial support under contract MAT2002-04603-C05-
at room temperature. 02 and L. Martín-Carrón wants also to acknowledge the
“Ministerio de Educación y Ciencia” for financial support
under contract program “Juan de la Cierva”.

Bol. Soc. Esp. Ceram. V., 45 [3] 218-222 (2006) 221

 J. CANET-FERRER, L. MARTÍN-CARRÓN, J. L. VALDÉS and J. MARTÍNEZ-PASTOR

REFERENCES 9. T. J. Yang et al. “Direct observation of pining and bowing of a single

ferroelectric domain wall”. Phys. Rev. Lett. 82, 4106-4109 (1999); and
1. M. Zgonik, R. Schlesser, I. Biaggio, E. Voit, J. Tscherry and P. Gunter. references therein.
“Materials constants of KNbO3 relevant for electro-and acousto-optics”. 10. M. Labardi, V. Likodimos and M. Allegrini. “Force-microscopy contrast
J.Appl.Phys. 74, 1287-1297 (1993). mechanisms in ferroelectric domain imaging”. Phys. Rev. B, 61, 14390-14398
2. A.V. Postnikov, T.Neumann, G. Borstel and M. Methfessel. “Ferroelectric (2000)
structure of KNbO3 and KTaO3 from first-principles calculations”. Phys. Rev. 11. P. Shivprasad et al. “Study of the electrostatic force between a conducting
B. 48, 5910-5918 (1993). tip in proximity with a metallic surface: theory and experiment”. J. of Appl.
3. R.D. King-Smith, D. Vanderbilt. “First-principles investigation of Phys. 88, 6940-6942 (2000)
ferroelectricity in perovskite compounds”. Phys. Rev. B. 49, 5828-5844 12. U. Zerweck, C. Loppacher, T. Otto, S. Granfström and L.M. Eng. “Accuracy
(1994) and resolution limits of Kelvin probe force microscopy”. Phys. Rev. B. 71,
4. C. Duan, W.N. Mei, J Liu and J.R. Hardy. “First-principles study on the 125424 (2005).
optical properties of KNbO3”. J. Phys.: Condens. Matter. 13, 8189-8195 13. R. Rousier, P. Vairac and B. Cretin. “Measurement of the contact potential
(2001). difference with an electrostatic force microscope”. Eur. J. Phys. 22, 657-662
5. B.T. Matthias. “New ferroelectric crystals”. Phys. Rev. 75, 1771 (1949) (2001)
6. R. Lüthi, H. Haefke, K.P. Meyer, E. Meyer, L. Howald and H.J. Güntherodt. 14. S. Wang and J.M. Siqueiros. “Influence of the sample-probe coupling on
“Surface and domain structures of ferroelectric crystals studied with the resolution of transmitted near-field optical image”. Opt. Lasers Eng.
scanning force microscopy”. J. Appl. Phys. 74, 7461-7471 (1993). 31, 517- (1999).
7. H. Bluhm, U.D. Schwarz and R. Wiesendanger. “Origin of the ferroelectric 15. B. Hecht, H. Bielefeldt, Y. Inoute, D. W. Pohl and L. Novotny. “Facts and
domain contrast observed in lateral force microscopy” Phys. Rev. B. 57, artefacts in near-field optical microscopy”. J. Appl. Phys. 81, 2492-2498
161-169 (1998). (1997).
8. V.Yu Topolov “Domain wall displacements and piezoelectric activity of 16. H. Chaib, T. Otto, and L.M. Eng. “Theoretical study of ferroelectric and
KNbO3 single crystals” J. Phys.: Condens. Matter 15, 561 (2003) optical properties in the 180º ferroelectric domain wall of tetragonal
BaTiO3”. Phys. Stat. Sol. 233, 250-262 (2002).

Recibido: 15.07.05
Aceptado: 31.01.06

222 Bol. Soc. Esp. Ceram. V., 45 [3] 218-222 (2006)