RESEARCH ARTICLE

MATERIALS ENGINEERING 2016 © The Authors, some rights reserved;

exclusive licensee American Association for

Subatomic deformation driven by vertical the Advancement of Science. Distributed

under a Creative Commons Attribution

piezoelectricity from CdS ultrathin films NonCommercial License 4.0 (CC BY-NC).
10.1126/sciadv.1600209

Xuewen Wang,1,2* Xuexia He,1* Hongfei Zhu,3 Linfeng Sun,4 Wei Fu,1 Xingli Wang,5 Lai Chee Hoong,1 Hong Wang,1
Qingsheng Zeng,1 Wu Zhao,1 Jun Wei,6 Zhong Jin,3 Zexiang Shen,4 Jie Liu,3,7 Ting Zhang,2† Zheng Liu5,8,9†

Driven by the development of high-performance piezoelectric materials, actuators become an important tool for
positioning objects with high accuracy down to nanometer scale, and have been used for a wide variety of equip-
ment, such as atomic force microscopy and scanning tunneling microscopy. However, positioning at the sub-
atomic scale is still a great challenge. Ultrathin piezoelectric materials may pave the way to positioning an
object with extreme precision. Using ultrathin CdS thin films, we demonstrate vertical piezoelectricity in atomic
scale (three to five space lattices). With an in situ scanning Kelvin force microscopy and single and dual ac
resonance tracking piezoelectric force microscopy, the vertical piezoelectric coefficient (d33) up to 33 pm·V−1
was determined for the CdS ultrathin films. These findings shed light on the design of next-generation sensors
and microelectromechanical devices.

INTRODUCTION

Downloaded from https://www.science.org on October 14, 2024

Piezoelectric material has proven to be a key to achieving high-resolution ate negative and positive charges simultaneously. These charges will
positioning, which is crucial for fundamental research and industrial distribute at the top and bottom surfaces under external bias (Fig. 1).
applications. Recent studies on the piezoelectricity of two-dimensional However, the synthesis of ultrathin piezoelectric materials is the main
(2D) materials provide opportunities for constructing thin electro- challenge to determining the vertical piezoelectricity (d33). Up to now, a
mechanical coupled devices, such as nanoscale mechanical transducers few studies reported the synthesis of atomic thin piezoelectric materials
and microenergy systems (1–3). Because of the break of centrosymmetry (such as ZnO, CdS, and CdSe nanoplatelets), by a wet chemical method
in monolayer crystals, only a few 2D layered materials (including graphene, (7–10). Although these achievements present valuable advances, many
h-BN, C3N4, and 2H-ACh2, where A = W, Mo; Ch = S, Se, Te) have been critical problems still remain and limit their applications. For instance,
predicted to have piezoelectricity (4, 5), which has been exemplified in a because of organic solutions that were used to synthesize CdS, the re-
few layered MoS2-based flexible and suspended devices (1, 2, 6). In these maining organic molecules at the surface of CdS may cast shadows
studies, in-plane strain was produced by an atomic force microscopy (AFM) on the piezoelectricity of CdS. Here, using chemical vapor deposition,
tip (or flexible substrate) on the films, which resulted in charge aggrega- we synthesized the CdS thin films down to 2 to 3 nm. We demonstrate
tion when a lateral bias was applied to the materials (Fig. 1). However, the vertical piezoelectricity d33 of CdS in atomic scale (three to five space
these methods can only be used to study in-plane piezoelectricity, such lattices) by single and dual ac resonance tracking piezoelectric force mi-
as d11 and d22, although the vertical piezoelectric coefficient (d33) is croscopy (DART-PFM). The surface potential and charge distribution
the key parameter to evaluate the performance of piezoelectric materials. were measured by scanning Kelvin force microscopy (SKFM). We ob-
For intrinsic piezoelectric materials, vertical symmetry breaking still served the vertical piezoelectric domains at the CdS thin films, and the
remains even for thin films down to a few layers of atoms. Therefore, measured piezoelectric coefficient (d33) was 32.8 pm·V−1. We found that
theoretically, the vertical piezoelectricity (d33) can be obtained from this value was approximately three times larger than that of the bulk
ultrathin piezoelectric materials such as ZnO, CdS, and CdSe nano- CdS. In addition, we found that the contact PFM process could largely
platelets. In this case, when the stress is applied to ultrathin piezoelectric change the surface potential of a CdS thin film by producing the charges
materials toward out of plane, the geometric deformation will gener- at its surface. This study sheds light on the vertical piezoelectricity of
ultrathin (lattice scale) nanomaterials and promotes their application
in actuators with subatomic resolution.
1
School of Materials Science and Engineering, Nanyang Technological University, 50
Nanyang Avenue, Singapore 639798, Singapore. 2International Laboratory for Adapt-
ive Bio-nanotechnology (i-Lab), Suzhou Institute of Nano-Tech and Nano-Bionics, Chi-
nese Academy of Sciences, Suzhou 215123, China. 3Key Laboratory of Mesoscopic RESULTS AND DISCUSSION
Chemistry of Ministry of Education and Collaborative Innovation Center of Chemistry
for Life Sciences, School of Chemistry and Chemical Engineering, Nanjing University, Growth of CdS thin films for vertical piezoelectricity
Nanjing 210093, China. 4School of Physical and Mathematical Sciences, Nanyang
Technological University, 21 Nanyang Link, Singapore 637371, Singapore. 5Centre We synthesized CdS thin films onto SiO2/Si substrate by chemical
for Micro-/Nano-electronics (NOVITAS), School of Electrical and Electronic Engineer- vapor deposition with sulfur powder and Cd grains as precursors
ing, Nanyang Technological University, Singapore 639798, Singapore. 6Singapore (see details in Materials and Methods and fig. S1A). Figure 2 shows three
Institute of Manufacturing Technology, 71 Nanyang Drive, Singapore 638075, Singa-
pore. 7Department of Chemistry, Duke University, Durham, NC 27708, USA. 8Center
dominant morphologies of CdS thin films (i) uniform disc-like struc-
for Programmable Materials, School of Materials Science and Engineering, Nanyang tures, (ii) Janus structure with two kinds of thickness at one thin film,
Technological University, 50 Nanyang Avenue, Singapore 639798, Singapore. 9CINTRA and (iii) rounded microparticle at the center of a thin film. The typical
CNRS/NTU/THALES, UMI 3288, Research Techno Plaza, 50 Nanyang Drive, Singapore optical microscopic images are shown in Fig. 2 (A, D, and G), and all
637553, Singapore.
*These authors contributed equally to this work. the samples are circular. The radius of the as-prepared thin film is between
†Corresponding author. Email: [email protected] (Z.L.); [email protected] (T.Z.) 10 and 20 mm. To identify the thickness of as-prepared CdS thin films,

Wang et al. Sci. Adv. 2016; 2 : e1600209 1 July 2016 1 of 9

RESEARCH ARTICLE

Fig. 1. Schematic illustration of local characterization of in-plane piezoelectricity and vertical piezoelectricity. In-plane piezoelectricity (piezo)
(d11, d22) of ultrathin materials is the planar electromechanical couple behavior, where the applied stress and produced piezoelectric potential are
located at the in-plane of exposed lattice plane. Vertical piezoelectricity (d33) focus on electromechanical interaction occurred in the vertical axis, which
is perpendicular to the surface of materials. The high-precision deformation actuator can be implemented using accurate positioning of the materials
surface by vertical inverse piezoelectricity.

Downloaded from https://www.science.org on October 14, 2024

Fig. 2. CdS thin films with three different shapes. (A to E, G, and H) Typical optical images (A, D, and G) and AFM images (B, C, E, and H) of CdS
thin films with uniform disc-like structure, Janus structure, and the structure of rounded microparticle at its center. The AFM line scan in (C) indicates
that the thickness of the CdS ultrathin film is ~3.2 nm, corresponding to approximately five lattices of CdS. The AFM line scans in (E) and (H) show
that the Janus structure has two kinds of thickness (2.1 and 3.7 nm) at one thin film, and the structure of the microparticle is grown on the thin film
with a height (h) of ~1.4 mm and radius of ~5 mm. (F and I) Corresponding structural diagram.

we characterized the samples by AFM (Fig. 2, B, C, E, and H). The film (2.2 nm). The height and radius of the particles are ~1.4 and ~5 mm,
thickness of these thin films ranges from 2.1 to 3.7 nm, which is close respectively (Fig. 2I). We examined more than 100 CdS thin films to
to the thickness of three to five crystal lattices (2.02 nm for three lattices calculate their size distribution (fig. S2). The three types of thin films
and 3.37 nm for five lattices) of hexagonal CdS along the z orientation. present average sizes of 8.92, 48.33, and 27.08 mm, respectively, with
The Janus structure has different geometries at the thin film (Fig. 2F). the Gaussian distribution. The SDs of the sizes are 0.84, 17.65, and
For the third structure, circular microparticles were grown on the thin 7.75 mm, respectively. The average size of the Janus structure is larger

Wang et al. Sci. Adv. 2016; 2 : e1600209 1 July 2016 2 of 9

RESEARCH ARTICLE

than that of the uniform disc-like structures, which suggests that the sec- Similar near–band edge emissions from CdS quantum dots, nano-
ondary nucleation could be preferentially formed in a large thin film. wires, and nanoribbons were also observed in previous reports
(7, 18, 19). This phenomenon can be explained by quantum confinement
Spectroscopic characterization of CdS thin films effect in a 2D thin film. As the dimension of the materials reaches nano-
We collected Raman spectra from these materials and were able to iden- scale (close to the exciton Bohr radius), the electrons and holes will be
tify the wurtzite CdS by the Raman vibration modes (11, 12). The first- squeezed into a confining dimension, following the appearance of dis-
and second-order longitudinal optical phonons in CdS thin films were crete energy states. Therefore, the band gap increases and becomes
observed from Raman spectra (fig. S3), which matched the Raman size-dependent, which ultimately gives rise to a blue shift in band edge
spectra of nanowires (12–14). It is well known that the optical properties emission (15, 20). For the circular microparticle at the center of the
of nanomaterials show significant departures from their bulk counter- thin film, the PL spectrum displays a pronounced peak at 595 nm
parts, because the scale of confinement approaches their exciton Bohr other than band edge emission at 514 nm (Fig. 3B). This contributed
radius (15). Notably, the thickness of our CdS thin films is in the to the defect-related emission from trap or surface states, for example
range of 2 to 3 nm, which is close to the exciton Bohr radius of CdS the unreacted sulfur atoms at the interface and Cd vacancies.
(~3.1 nm) (16, 17). To understand the optical properties of the as- We also collected the PL spectra and PL mapping from different
prepared CdS thin film, we investigated their band gap structure and types of CdS thin films (Fig. 2, D to F, and fig. S4). Figure 3C shows the
photoluminescence (PL). Figure 3A represents the energy band struc- typical optical microscopic image of the rounded microparticles at the
ture of CdS in the vicinity of the G-point of the Brillouin zone, center of the thin film and the corresponding PL intensity mapping
showing the direct gap emission at G-valley (green arrow) as well as at a peak of 514 nm. Different scale bars were applied to Fig. 3 (D and E)
electron transitions between defect energy level and electron band (yel- to reflect the fine structure of the CdS thin films. The PL mapping at the

Downloaded from https://www.science.org on October 14, 2024

low arrow), which results in photoemission with band gap fluores- outside area suggests a high uniformity and homogeneity of CdS thin
cence and defect-related luminescence (7). The PL properties for films. Meanwhile, strong and ring-like PL emission was observed at the
all three kinds of CdS thin film were obtained at room temperature middle range of the sample. The intensity of PL emission from the
using 457-nm laser as an excitation source (Fig. 3, B to F, and fig. S4). microparticle is much stronger than that from the thin film. Therefore,
Only a strong emission peak at ~506 nm was observed from the ul- it is difficult to recognize the small distinction of PL mappings between
trathin CdS thin film with a thickness of 2 to 3 nm, which is con- those two materials, despite the fact that there is ~8-nm disparity for
sidered to be band edge emission. Notably, the value is close to the band edge emissions at different positions. The intensity of PL mapping
band edge emission of bulk CdS (~514 nm) with a slight blue shift. at 595 nm confirms that the defect-related emission only existed at the

Fig. 3. Spectroscopic characterization of CdS thin film. (A) Energy (E) band structure in the vicinity of the G-point of the Brillouin zone, showing the
photon emission process. (B and C) PL spectrum of CdS thin film from points i and ii marked in (C) with plus signs, showing strong band edge
emission (506 nm) of CdS ultrathin film and defect-related emission (595 nm). (C) Optical image of CdS thin film with a rounded microparticle at its
center. a.u., arbitrary units. (D and E) PL mapping at an emission of 514 nm with a different scale bar, demonstrating high uniformity and homo-
geneity of CdS thin films at the outside area. (F) PL mapping at the 595-nm emission, indicating that the defect-related emission only occurs at the thicker
CdS microparticle.

Wang et al. Sci. Adv. 2016; 2 : e1600209 1 July 2016 3 of 9

RESEARCH ARTICLE

relatively thicker CdS rounded microparticle. Those ring-like and electrostatic interactions between the tip and the sample surface. Such
size-dependent band edge emissions, as well as strong defect-related surface potential is related to the work functions of scanned materials
emission, provide new possibilities for the construction of opto- and reflects the local charge distribution when a bias voltage was applied
electronic devices. to their surface. SKFM has been used to measure the surface potential
difference in 2D graphene-boron nitride heterostructures, as well as the
Surface potential and vertical piezoelectricity of CdS doping and charge carrier concentration in semiconducting nano-
thin films materials (21, 22). Figure 4C shows the band diagram change of the
To investigate the surface potential of CdS thin films, SKFM was AFM tip and sample in SKFM mode. The work function differences
carried out by noncontact probe scan mode. As shown in Fig. 4A, between the sample (ϕsample) and the probe (ϕtip) create a contact
SKFM can be used to explore the surface potential by tracking the potential difference when they are electrically contacted. Therefore,

Downloaded from https://www.science.org on October 14, 2024

Fig. 4. Noncontact SKFM and standard contact PFM investigation for CdS thin film. (A and B) Schematic illustration of SKFM (A) and PFM (B)
measurements. (C) Band diagram of tip and sample when they are electrically separated (top graph) and electrically contacted (bottom graph). d,
distance; VL, vacuum levels; q, electronic charge; Vc, contact potential difference. (D) Optical image of CdS thin films. (E and F) Topography (E) and
phase (F) images observed by SKFM mode for the single CdS thin film marked in (D). (G to I) Corresponding potential mappings with tip voltages of 3, 6, and 9 V,
respectively. Insets show histograms of the surface potential distributions. The CdS ultrathin film has a higher positive voltage (~0.9 V) than the substrate,
demonstrating that a large amount of charges are accumulated at a CdS thin film after contact PFM scanning. (J) Amplitude images observed by contact
PFM technology with tip voltages from 1 to 6 V, showing remarkable inverse piezoelectricity. (K) Average amplitude variations versus applied
voltages calculated from (J). Error bars indicate 1 SD. Scale bars, 2 mm (E to J). The linearly fitted line shows that the measured piezoelectric coefficient
deff is ~16.4 pm·V−1, whereas the vertical piezoelectric coefficient d33 is ~32.8 pm·V−1.

Wang et al. Sci. Adv. 2016; 2 : e1600209 1 July 2016 4 of 9

RESEARCH ARTICLE

tracking the contact potential difference by applying the equivalent and the contact area is negligible in this situation (33). The nonplanar
external voltage can determine the work function and local charge PFM tip and small deformation of rigid substrate result in a weak in-
distribution at the surface of scanned samples. We applied SKFM dentation limit and make the contact area with the sample less than
mode to the single CdS thin films with different tip voltages (3, 6, and the tip radius (~25 nm) (fig. S8). According to previous reports (32–34),
9 V). As seen from the topography and phase images in Fig. 4 (E and the value of d33 is approximately twice than measured deff. On this
F), the CdS thin film presents stronger dark and light contrasts than basis, we can estimate the piezoelectric coefficient d33 of the CdS thin
the substrate. It has a thickness of ~3.09 nm and phase variation of film to be 32.8 pm·V−1, which is three times larger than that of the
~3°. Figure 4 (G to I) shows the surface potential images with bulk CdS materials (9.71 pm·V−1) and ZnO materials, and even com-
increased tip voltages, which indicate that the CdS thin film has high- parable to some of the inorganic piezoelectric ceramics (see fig. S9 and
er positive voltage than the substrate. The surface potential of CdS table S1) (35). Previous studies have demonstrated that the piezo-
sample is quantitatively determined to be ~0.9 V higher than that electric coefficient of CdS strongly depended on its carrier concentra-
of the substrate, indicating that a large amount of charges are accu- tion. The d33 coefficient of CdS decreased by 60% when the carrier
mulated at the CdS thin film. We believe that this resulted from the pi- concentration increased from 109 to 1018 cm−3 (36, 37). To investigate
ezoelectricity of the CdS sample because the contact PFM was the electrical conductivity of CdS materials, we fabricated the Hall
perfomed several times before the noncontact SKFM measurement. device and field-effect transistor devices (figs. S10 and S11). The
Similar behaviors were also recorded from other CdS samples via Hall device presents poor conductivity, and picoampere-level current
SKFM. Before the CdS thin film was scanned by the PFM tip, it can be detected under higher source-drain voltage, making it difficult
showed ~200-mV surface potential, which is higher than the sub- for the Hall device to calculate the carrier concentration. We further
strate. After scanning, this value increased up to ~1 V (fig. S5). This fabricated the field-effect transistor using the single CdS thin film (see

Downloaded from https://www.science.org on October 14, 2024

phenomenon indicates that the contact PFM process can largely change fig. S11). The source-drain current was measured at the picoampere
the surface potential of CdS thin film by producing the charges at its level when a bias voltage (25 V) was applied, indicating the low
surface, and the vertical piezoelectricity of CdS is the key to the gen- carrier concentration in our CdS thin film. The carrier concentration
erated charges. (nc) can be calculated by nc = 1/reme, where r is the resistivity, e is
We used standard contact PFM technology to investigate the ver- the electronic charge, and me is the electron mobility (36). Electron
tical piezoelectricity of single CdS thin film. As shown in Fig. 4B, the mobility (me) can be estimated by me = Gm(L/WCgVds), where Gm
mechanical response of the sample was measured when the vertical is the peak transconductance (G = dIds/dVg), Cg is the gate capaci-
electrical field was applied to its surface through conducting the tance, and L and W are the channel length (7 mm) and width (10 mm),
PFM tip. A typical piezoelectric material is strained (or stressed) under respectively. We deduce that the mobility me is 0.0023 cm2 V−1 s−1,
a vertical electrical field, which is reflected in the amplitude change for and the carrier concentration nc is in the range of ~1012 to ~1015 cm−3.
PFM imaging. The vertical motion of PFM tip and the vertical electric Therefore, this robust piezoelectric coefficient of CdS thin film is presum-
field applied make PFM imaging a powerful tool to explore the ably due to the free boundary and low carrier concentration, which
electromechanical functionalities in a vertical direction, namely, quan- results from the low dimensional structure with large surface and defects.
tify the piezoelectric constant of d33 for piezoelectric materials. Up to now,
this technology has been extensively used to study piezoelectric/ DART-PFM technology for imaging vertical
ferroelectric ceramic films and ZnO nanomaterials, such as nanowires piezoelectric domain
and nanorods (23–28), as well as to image domain structures and do- The DART-PFM was used to understand the piezoelectricity of CdS
main evolution in ferroelectric thin films (29–32). Here, PFM imaging thin films by in-site characterization. Although PFM technology is con-
was first used for local characterization of ultrathin CdS thin films. Tip sidered to be a powerful tool to monitor the electromechanical response
voltages from 1 to 6 V were applied to study the vertical electromechanical of piezoelectric materials, it fails at rough surface or atomic thin films.
response of CdS thin films. The amplitude images under different tip For such samples, the noise floors of the surface height topography
voltages are shown in Fig. 4J, and the corresponding topography and usually reach tens of picometers, which is comparable to the electro-
phase images are displayed in fig. S6. Surprisingly, significant ampli- mechanical displacement in atomically thin nanomaterials at the vertical
tude change was observed in the amplitude images when different tip direction. As shown in fig. S12, DART-PFM uses dual ac resonance
voltages were applied, indicating strong inverse piezoelectricity. The tracking to measure the shift of resonance frequency while avoiding
average amplitude variations can be calculated from the selected areas crosstalk between changes in the contact resonance frequency and sur-
in Fig. 4J (details in fig. S7). From average amplitude variations versus face topography (38, 39). DART-PFM has proven to be a reliable method
applied voltages in Fig. 4K, the vertical piezoelectric constant of deff is to study the electromechanical elastic properties of inorganic ceramics
~16.4 pm·V−1, which is not equal to the real vertical piezoelectric and biological systems (38). Here, we used this technology to investigate
coefficient d33 because a nonuniform electric field is applied to the the local piezoelectricity of CdS thin films (Fig. 5, A to K). Figure 5A
sample during the piezoelectric measurements. Pt/Ir-coated tips were shows the height topography of single CdS thin film with a thickness of
used for PFM scanning, and the force applied is in the range of 100 to ~3.16 nm. The corresponding resonant frequency, phase, and amplitude
200 nN. Such external force will result in a weak indentation to the images are displayed in Fig. 5 (B to D). For the resonant frequency
sample because of the support of the rigid SiO2 substrate. It is re- image, the CdS thin film (dark) shows a remarkable contrast against
ported that an AFM tip can only penetrate ~3 nm at SiO2 (3). The the regions outside (bright) because the resonant frequency of the tip-
thin and soft nature of the CdS (~3 nm) ensures that the mechanical surface interaction changes from ~347 to 342 kHz when the tip is ap-
deformation of SiO2 is the dominant contribution, as shown in fig. S8. plied to the CdS thin film. The corresponding phase change is also
The stiff tip and surface and high dielectric constant lead to a signif- observed, as shown in Fig. 5C. This variation of resonant frequency
icant potential drop between the tip and surface, that is, Vs ≪ Vtip, probably results from the tip-surface spring constant, which originated

Wang et al. Sci. Adv. 2016; 2 : e1600209 1 July 2016 5 of 9

RESEARCH ARTICLE

Downloaded from https://www.science.org on October 14, 2024

Fig. 5. DART-PFM technology for characterization of CdS thin film. (A to D) Topography (A), resonance frequency (B), phase (C), and amplitude (D)
images for a single CdS thin film, showing remarkable resonance frequency variations. (E to H) Top-view 3D topography (E), resonance frequency (F),
phase (G), and amplitude (H) images for marked area in (A). (I) Overlap image of topography and resonance frequency. Marked areas present the
notable change of resonance frequency. (J and K) Height profiles measured by line scans along lines 1 and 2 in (E), and shadow areas (I, III, and V)
corresponding to marked areas in (I).

from the electromechanical reaction and piezoelectric constant of the half ball as the PFM tip, which has a diameter of 50 nm (Fig. 6A), and
materials being scanned. To understand the vertical piezoelectric of the sur- one CdS sample with a thickness of 3 nm. In this simulation setup, the PFM
face of CdS thin film locally, the imaging area was zoomed to 400 × 400 nm, tip contacts the CdS sample, to which a 5-V dc voltage was applied.
as marked in Fig. 5A. The top view of 3D topography, resonance frequency, The stress distributions on the CdS sample were investigated, as shown
phase, and amplitude images are shown in Fig. 5 (E to H, respectively). in Fig. 6 (C and D). The region surrounding the contact point ex-
The height topography shows that the CdS thin film is composed of periences the highest stress. The stress then decays gradually toward
ultrathin nanosheets (roughness, 0.67 nm). The resonance frequency the edges of the CdS sample. The strong stress is contributed by the large
of a small nanosheet varies from ~346 to ~343 kHz at different loca- deformation produced by vertical piezoelectricity of the CdS sample.
tions (Fig. 5F). The contrast is observed from the marked areas (I, II, This unprecedented high vertical piezoelectricity (d33 = 32.8 pm·V−1)
III, IV, and V) in the overlap images of topography and frequency in of CdS thin film allows extremely precise positioning at the subatomic
Fig. 5I and the corresponding height profiles in Fig. 5 (J and K). The level. We built a 3D FE model to demonstrate this. CdS thin film (thick-
dark areas represent low resonance frequency, implying the strong ness of 3 nm and radius of 15 nm) was used as the actuating material,
piezoelectric domains. Similar piezoelectric geometries were also found at and −1- to −5-V dc voltages were applied to its surface as driven voltages.
the boundary between the CdS thin film and substrate, as shown in fig. Applying a negative tip bias to the CdS surface leads to a positive electrical
S13. As mentioned above, the frequency response in such scanning field along the vertical direction and results in the expansion of
model is dominated by the tip-sample electromechanical interac- materials. Figure 6 (E to G) shows the simulation results of a strained
tion, which mainly relies on the piezoelectricity and the crystallinity CdS thin film with different voltages applied to its surface. By increasing
of samples. Therefore, it enables us to study the electromechanical prop- the negative driven voltage from −1 to −5 V, the deformation of the CdS
erties locally on the rough surface of the atomic nanosheets. sample was linearly increased from ~30 to ~150 pm (Fig. 6, E and F). This
result is in accordance with experimental data. Although the roughness of
Simulation on vertical piezoelectricity and actuator with the CdS thin film is increased slightly under large potential (Fig. 6E), the
subatomic resolution deformation magnitude remains linearly dependent on the driven
To better explain the experimental results of vertical piezoelectricity and voltage at the picometer level, which is already close to the lattice constant
estimate the deformation of CdS thin film under applied voltage, we of CdS. Our simulation demonstrates high vertical piezoelectricity in
developed a 3D finite element (FE) model with COMSOL Multiphysics, CdS thin film, which can be applied as an active material for subatomic
as schematically shown in Fig. 6. This model consists of a cone with a deformation actuators.

Wang et al. Sci. Adv. 2016; 2 : e1600209 1 July 2016 6 of 9

RESEARCH ARTICLE

Downloaded from https://www.science.org on October 14, 2024

Fig. 6. Simulation of vertical piezoelectricity and subatomic deformation actuator. (A) Three-dimensional image of potential drop on CdS film.
(B) Scanning electron microscopy image of a conductive tip for PFM characterization. (C and D) Bottom and side views of stress distribution on CdS
film. (E to G) Simulation for subatomic deformation actuator. Different potentials were applied to surface deformation curves (E), mappings (G), and
vertical deformation (F) of CdS thin films.

CONCLUSIONS the Cd grain and at the center of the furnace to collect the products.
In summary, we prepared ultrathin CdS thin films and demonstrated Then, the system was flushed with argon flow (100 sccm) for more than
their strong vertical piezoelectricity at the atomic scale (three to five three cycles and heated to the targeted temperature of 600°C, with a rate
space lattices) by using noncontact SKFM, contact PFM, and DART- of 20°C/min. When the temperature reached and kept in targeted
PFM. We resolved the locally vertical piezoelectricity of CdS thin films temperature for 30 min, the furnace was instantly switched off and
using DART-PFM and determined the vertical piezoelectric domains cooled down to room temperature naturally. Finally, the CdS thin films
at the CdS thin films. Their vertical piezoelectric coefficient (d33) was were grown on the Si/SiO2 substrate.
~32.8 pm·V−1, which was three times higher than the CdS bulk. In
addition, vertical piezoelectricity in the CdS thin film and its applica- Characterization and 3D FE simulation
tions in subatomic deformation actuators were demonstrated by 3D Both Raman and PL data were collected at room temperature using
FE models. Strong vertical piezoelectricity makes this material 457-nm laser as an excitation source (WITec alpha300 RAS Raman
promising for the construction of atomically thin piezoelectric devices. system). Sample size, thickness, morphologies, surface potential, and
vertical piezoelectricity were evaluated with optical microscopy
(Olympus BX51), AFM, SKFM, PFM, and DART-PFM (Cypher S,
MATERIALS AND METHODS Asylum Research). The FE simulation of the vertical piezoelectricity
and subatomic deformation actuator was performed with FE modeling
Growth of CdS thin films software (COMSOL Multiphysics 5.0). Details can be found in the Sup-
The synthesis of ultrathin CdS nanoplatelets was carried out in a con- plementary Materials.
ventional tube furnace (ThermoFisher, Lindberg/Blue MTF55035C-1)
equipped with a 1-inch quartz tube and connected to Ar gas inlets (fig. S1).
The starting materials were sulfur (1 g) and Cd grain (60 mg), which SUPPLEMENTARY MATERIALS
were placed in different porcelain boats inside the quartz tube. The Supplementary material for this article is available at http://advances.sciencemag.org/cgi/
Si/SiO2 substrates (230-nm SiO2 on Si) were placed downstream of content/full/2/7/e1600209/DC1

Wang et al. Sci. Adv. 2016; 2 : e1600209 1 July 2016 7 of 9

RESEARCH ARTICLE

Supplementary Materials and Methods 22. C. Maragliano, S. Lilliu, M. S. Dahlem, M. Chiesa, T. Souier, M. Stefancich, Quantifying
fig. S1. Growth of CdS thin films. charge carrier concentration in ZnO thin films by Scanning Kelvin Probe Microscopy.
fig. S2. Statistical analysis for the size of CdS thin films. Sci. Rep. 4, 4203 (2014).
fig. S3. Raman spectra of as-obtained CdS sample. 23. E. Broitman, M. Y. Soomro, J. Lu, M. Willander, L. Hultman, Nanoscale piezoelectric re-
fig. S4. PL spectra and PL mapping from the CdS thin films. sponse of ZnO nanowires measured using a nanoindentation technique. Phys. Chem.
fig. S5. SKFM characterization of CdS thin film. Chem. Phys. 15, 11113–11118 (2013).
fig. S6. Topography and phase images of PFM characterization of CdS thin film when different 24. Y.-F. Lin, J. Song, Y. Ding, S.-Y. Lu, Z. L. Wang, Piezoelectric nanogenerator using CdS
tip voltages (1 to 6 V) were applied. nanowires. Appl. Phys. Lett. 92, 022105 (2008).
fig. S7. The average amplitude change can be established from eight different areas of CdS sample. 25. C. Harnagea, M. Alexe, D. Hesse, A. Pignolet, Contact resonances in voltage-modulated
fig. S8. Illustration of weak indentation and strong indentation. force microscopy. Appl. Phys. Lett. 83, 338 (2003).
fig. S9. Thickness versus piezoelectric coefficient distribution from previous literature results. 26. N. V. Andreeva, M. Tyunia, A. V. Filimonov, A. I. Rudskoy, N. A. Pertsev, S. B. Vakhrushev,
fig. S10. Hall device of CdS thin film. Low-temperature evolution of local polarization properties of PbZr 0.65 Ti 0.35O 3 thin
fig. S11. Electrical characterization of CdS thin film. films probed by piezoresponse force microscopy. Appl. Phys. Lett. 104, 112905 (2014).
fig. S12. Schematic illustration of experimental setup for DART-PFM. 27. L. Aggarwal, J. S. Sekhon, S. N. Guin, A. Arora, D. S. Negi, R. Datta, K. Biswas, G. Sheet, Direct
fig. S13. DART-PFM characterization for the boundary of CdS thin film. evidence of strong local ferroelectric ordering in a thermoelectric semiconductor. Appl.
table S1. Summary of piezoelectric coefficient from different materials. Phys. Lett. 105, 113903 (2014).
References (40–50) 28. M. Hussain, A. Khan, O. Nur, M. Willander, E. Broitman, The effect of oxygen-plasma treat-
ment on the mechanical and piezoelectrical properties of ZnO nanorods. Chem. Phys. Lett.
608, 235–238 (2014).
29. M. Trapatseli, D. Carta, A. Regoutz, A. Khiat, A. Serb, I. Gupta, T. Prodromakis, Conductive
REFERENCES AND NOTES atomic force microscopy investigation of switching thresholds in titanium dioxide thin
films. J. Phys. Chem. C 119, 11958–11964 (2015).
1. H. Zhu, Y. Wang, J. Xiao, M. Liu, S. Xiong, Z. J. Wong, Z. Ye, Y. Ye, X. Yin, X. Zhang, Observation
30. C. Li, L. Huang, T. Li, W. Lü, X. Qiu, Z. Huang, Z. Liu, S. Zeng, R. Guo, Y. Zhao, K. Zeng,
of piezoelectricity in free-standing monolayer MoS2. Nat. Nanotechnol. 10, 151–155 (2015).
M. Coey, J. Chen, Ariando, T. Venkatesan, Ultrathin BaTiO3-based ferroelectric tunnel junc-

Downloaded from https://www.science.org on October 14, 2024

2. W. Wu, L. Wang, Y. Li, F. Zhang, L. Lin, S. Niu, D. Chenet, X. Zhang, Y. Hao, T. F. Heinz,
tions through interface engineering. Nano Lett. 15, 2568–2573 (2015).
J. Hone, Z. L. Wang, Piezoelectricity of single-atomic-layer MoS2 for energy conversion
31. A. Belianinov, Q. He, A. Dziaugys, P. Maksymovych, E. Eliseev, A. Borisevich, A. Morozovska,
and piezotronics. Nature 514, 470–474 (2014).
J. Banys, Y. Vysochanskii, S. V. Kalinin, CuInP2S6 room temperature layered ferroelectric.
3. J. Qi, Y.-W. Lan, A. Z. Stieg, J.-H. Chen, Y.-L. Zhong, L.-J. Li, C.-D. Chen, Y. Zhang, K. L. Wang,
Nano Lett. 15, 3808–3814 (2015).
Piezoelectric effect in chemical vapour deposition-grown atomic-monolayer triangular
32. B. Y. Lee, J. Zhang, C. Zueger, W.-J. Chung, S. Y. Yoo, E. Wang, J. Meyer, R. Ramesh, S.-W. Lee,
molybdenum disulfide piezotronics. Nat. Commun. 6, 7430 (2015).
Virus-based piezoelectric energy generation. Nat. Nanotechnol. 7, 351–356 (2012).
4. K.-A. N. Duerloo, M. T. Ong, E. J. Reed, Intrinsic piezoelectricity in two-dimensional
33. S. V. Kalinin, D. A. Bonnell, Imaging mechanism of piezoresponse force microscopy of
materials. J. Phys. Chem. Lett. 3, 2871–2876 (2012).
ferroelectric surfaces. Phys. Rev. B 65, 125408 (2002).
5. M. Zelisko, Y. Hanlumyuang, S. Yang, Y. Liu, C. Lei, J. Li, P. M. Ajayan, P. Sharma, Anomalous
34. F. Johann, Á. Hoffmann, E. Soergel, Impact of electrostatic forces in contact-mode
piezoelectricity in two-dimensional graphene nitride nanosheets. Nat. Commun. 5, 4284 (2014).
scanning force microscopy. Phys. Rev. B 81, 094109 (2010).
6. S. Manzeli, A. Allain, A. Ghadimi, A. Kis, Piezoresistivity and strain-induced band gap tuning
35. H. D. Espinosa, R. A. Bernal, M. Minary-Jolandan, A review of mechanical and electro-
in atomically thin MoS2. Nano Lett. 15, 5330–5335 (2015).
mechanical properties of piezoelectric nanowires. Adv. Mater. 24, 4656–4675 (2012).
7. S. Ithurria, M. D. Tessier, B. Mahler, R. P. S. M. Lobo, B. Dubertret, A. L. Efros, Colloidal nano-
36. D. A. Scrymgeour, J. W. P. Hsu, Correlated piezoelectric and electrical properties in individual
platelets with two-dimensional electronic structure. Nat. Mater. 10, 936–941 (2011).
ZnO nanorods. Nano Lett. 8, 2204–2209 (2008).
8. X.-D. Wen, R. Hoffmann, N. W. Ashcroft, Two-dimensional CdSe nanosheets and their in-
37. T. Ogawa, H. Oikawa, A. Kojima, Decrement caused by screening effect of conduction
teraction with stabilizing ligands. Adv. Mater. 25, 261–266 (2013).
electrons on the effective charge of CdS crystals. Jap. J. Appl. Phys. 10, 593–599
9. C. Bouet, B. Mahler, B. Nadal, B. Abecassis, M. D. Tessier, S. Ithurria, X. Xu, B. Dubertret, Two-
(1971).
dimensional growth of CdSe nanocrystals, from nanoplatelets to nanosheets. Chem. Mater.
38. B. J. Rodriguez, C. Callahan, S. V. Kalinin, R. Proksch, Dual-frequency resonance-tracking
25, 639–645 (2013).
atomic force microscopy. Nanotechnology 18, 475504 (2007).
10. S. Pedetti, S. Ithurria, H. Heuclin, G. Patriarche, B. Dubertret, Type-II CdSe/CdTe core/crown
39. A. Gannepalli, D. G. Yablon, A. H. Tsou, R. Proksch, Mapping nanoscale elasticity and
semiconductor nanoplatelets. J. Am. Chem. Soc. 136, 16430–16438 (2014).
11. V. Sivasubramanian, A. K. Arora, M. Premila, C. S. Sundar, V. S. Sastry, Optical properties of dissipation using dual frequency contact resonance AFM. Nanotechnology 22, 355705
CdS nanoparticles upon annealing. Physica E 31, 93–98 (2006). (2011).
12. C. Hu, X. Zeng, J. Cui, H. Chen, J. Lu, Size effects of Raman and photoluminescence spectra 40. D. F. Crisler, J. J. Cupal, A. R. Moore, Dielectric, piezoelectric, and electromechanical coupling
of CdS nanobelts. J. Phys. Chem. C 117, 20998–21005 (2013). constants of zinc oxide crystals, Proc. IEEE 56, 225–226 (1968).
13. H. M. Fan, Z. H. Ni, Y. P. Feng, X. F. Fan, J. L. Kuo, Z. X. Shen, B. S. Zou, Anisotropy of 41. D. A. Scrymgeour, T. L. Sounart, N. C. Sinnons, J. W. P. Hsu, Polarity and piezoelectric
electron–phonon coupling in single wurtzite CdS nanowires. Appl. Phys. Lett. 91, response of solution grown zinc oxide nanocrystals on silver. J. Appl. Phys. 101, 014316
171911 (2007). (2007).
14. S. Kar, B. Satpati, P. V. Satyam, S. Chaudhuri, Synthesis and optical properties of CdS nano- 42. M.-H. Zhao, Z.-L. Wang, S. X. Mao, Piezoelectric characterization of individual zinc oxide
ribbons. J. Phys. Chem. B 109, 19134–19138 (2005). nanobelt probed by piezoresponse force microscope. Nano Lett. 4, 587–590 (2004).
15. A. P. Alivisatos, Semiconductor clusters, nanocrystals, and quantum dots. Science 271, 43. H. J. Fan, W. Lee, R. Hauschild, M. Alexe, G. Le Rhun, R. Scholz, A. Dadgar, K. Nielsch, H. Kalt,
933–937 (1996). A. Krost, M. Zacharias, U. Gösele, Template-assisted large-scale ordered arrays of ZnO
16. J. Tian, G. Cao, Semiconductor quantum dot-sensitized solar cells. Nano Rev. 4, 10.3402/nano. pillars for optical and piezoelectric applications. Small 2, 561–568 (2006).
v4i0.22578 (2013). 44. T.-Y. Ke, H.-A. Chen, H.-S. Sheu, J.-W. Yeh, H.-N. Lin, C.-Y. Lee, H.-T. Chiu, Sodium niobate
17. B. B. Srivastava, S. Jana, N. Pradhan, Doping Cu in semiconductor nanocrystals: Some old nanowire and its piezoelectricity. J. Phys. Chem. C 112, 8827–8831 (2008).
and some new physical insights. J. Am. Chem. Soc. 133, 1007–1015 (2011). 45. J. Wang, C. Stampfer, C. Roman, W. H. Ma, N. Setter, C. Hierold, Piezoresponse force mi-
18. N. N. Hewa-Kasakarage, M. Kirsanova, A. Nemchinov, N. Schmall, P. Z. El-Khoury, croscopy on doubly clamped KNbO3 nanowires. Appl. Phys. Lett. 93, 223101 (2008).
A. N. Tarnovsky, M. Zamkov, Radiative recombination of spatially extended excitons in 46. Z. Wang, A. P. Suryavanshi, M.-F. Yu, Ferroelectric and piezoelectric behaviors of individual
(ZnSe/CdS)/CdS heterostructured nanorods. J. Am. Chem. Soc. 131, 1328–1334 (2009). single crystalline BaTiO3 nanowire under direct axial electric biasing. Appl. Phys. Lett. 89,
19. R. Banerjee, R. Jayakrishnan, R. Banerjee, P. Ayyub, Effect of the size-induced structural 082903 (2006).
transformation on the band gap in CdS nanoparticles. J. Phys. Condens. Matter 12, 47. M. Minary-Jolandan, R. A. Bernal, I. Kuljanishvili, V. Parpoil, H. D. Espinosa, Individual GaN
10647–10654 (2000). nanowires exhibit strong piezoelectricity in 3D. Nano Lett. 12, 970–976 (2012).
20. D. A. B. Miller, D. S. Chemla, T. C. Damen, A. C. Gossard, W. Wiegmann, T. H. Wood, 48. Y. Luo, I. Szafraniak, N. D. Zakharov, V. Nagarajan, M. Steinhart, R. B. Wehrspohn,
C. A. Burrus, Band-edge electroabsorption in quantum well structures: The quantum- J. H. Wendorff, R. Ramesh, M. Alexe, Nanoshell tubes of ferroelectric lead zirconate titanate
confined Stark effect. Phys. Rev. Lett. 53, 2173–2176 (1984). and barium titanate. Appl. Phys. Lett. 83, 440 (2003).
21. T. Gao, X. Song, H. Du, Y. Nie, Y. Chen, Q. Ji, J. Sun, Y. Yang, Y. Zhang, Z. Liu, Temperature- 49. T. D. Nguyen, J. M. Nagarah, Y. Qi, S. S. Nonnenmann, A. V. Morozov, S. Li, C. B. Arnold,
triggered chemical switching growth of in-plane and vertically stacked graphene-boron M. C. McAlpine, Wafer-scale nanopatterning and translation into high-performance piezo-
nitride heterostructures. Nat. Commun. 6, 6835 (2015). electric nanowires. Nano Lett. 10, 4595–4599 (2010).

Wang et al. Sci. Adv. 2016; 2 : e1600209 1 July 2016 8 of 9

RESEARCH ARTICLE

50. R. Blachnik, J. Chu, R. R. Galązka, J. Geurts, J. Gutowski, B. Hönerlage, D. Hofmann, J. Kossut, Wang, Z.L., and T.Z. cowrote the paper. All authors contributed to the result analysis and dis-
R. Lévy, P. Michler, U. Neukirch, D. Strauch, T. Story, A. Waag, II-VI and I-VII compounds; semi- cussions. Competing interests: Z.L. and Xuewen Wang are inventors of the technology which
magnetic compounds, in Landolt-Börnstein: Numerical Data and Functional Relationships in is patent pending (Application 10201604725P). Data and materials availability: All data
Science and Technology, O. Madelung, U. Rossler, M. Schulz, Eds. (Springer, Berlin, 1999) vol. 41B. needed to evaluate the conclusions in the paper are present in the paper and/or the Supple-
mentary Materials. Additional data related to this paper may be requested from the authors.
Acknowledgments
Funding: This work is supported by the Singapore National Research Foundation (NRF) (NRF Submitted 2 February 2016
RF Award No. NRF-RF2013-08), National Program on Key Basic Research Project (973 Program, Accepted 12 May 2016
Grand No. 2015CB351901), the National Natural Science Foundation of China (61574163), and Published 1 July 2016
a start-up funding from Nanyang Technological University (M4081137.070). Author contribu- 10.1126/sciadv.1600209
tions: Z.L., T.Z., and Xuewen Wang conceived and designed the experiments; Xuewen Wang
and X.H. performed the growth of materials and SKFM and PFM characterizations. L.S. and Citation: X. Wang, X. He, H. Zhu, L. Sun, W. Fu, X. Wang, L. C. Hoong, H. Wang, Q. Zeng,
Xingli Wang conducted Raman and PL measurements. Xuewen Wang and L.C.H. carried out W. Zhao, J. Wei, Z. Jin, Z. Shen, J. Liu, T. Zhang, Z. Liu, Subatomic deformation driven by
the DART-PFM measurement. H.Z. performed finite element analysis simulation. Xuewen vertical piezoelectricity from CdS ultrathin films. Sci. Adv. 2, e1600209 (2016).

Downloaded from https://www.science.org on October 14, 2024

Wang et al. Sci. Adv. 2016; 2 : e1600209 1 July 2016 9 of 9