Piezoelectricity takes part in multiple important functions and processes in biomaterials often vital to the survival of organisms. Here, we investigate the piezoelectric properties of fish scales of green carp by directly examining their morphology at nanometer levels. Two types of regions are found to comprise the scales, a smooth one and a rough one. The smooth region is comprised of a ridge and trough pattern and the rough region characterized by a flat base with an elevated mosaic of crescents. Piezoelectricity is found on the ridges and base regions of the scales. From clear distinctions between the composition of the inner and outer surfaces of the scales, we identify the piezoelectricity to originate from the presence of hydroxyapatite which only exists on the surface of the fish scales. Our findings reveal a different mechanism of how green carp are sensitive to their surroundings and should be helpful to studies related to the electromechanical properties of marine life and the development of bio-inspired materials.
I. INTRODUCTION
Piezoelectricity in essence is electricity from pressure and is one of the direct consequences of when the structure of a crystalline material is non-centrosymmetric. Though the piezoelectricity of natural materials has been identified for quite a long time, detailed investigations of its properties remain rather limited.1 In fact, all kinds of biomaterials, such as peptides and proteins, are non-centrosymmetric and believed to possess piezoelectricity.1 The piezoelectricity found in biomaterials play an important role in biological processes and functions since pressure generated by the surrounding environment is converted to electrical signals. Pioneering studies on the piezoelectricity of biomaterials were first carried out by Fukuda on wood,2,3 and then on bones (collagen fiber).4 Subsequently, piezoelectricity and its related phenomena were also investigated in glycine and thymine,5,6 voltage-dependent ion channels,7 collagen,8 microtubules,9 pineal glands,10 human nails,11 seashells,12 and so on. Recently, the polarity of the ferroelectricity in aortic elastin13 was shown to be capable of being switched by an applied electric field. At the same time, glucose was shown to act as a suppressant in aortic elastin.13,14 Moreover, piezoelectricity is believed to play a major role in our nervous system and brain memory.15 Biological piezoelectricity can also be used to create bio-inspired materials, for instance, peptide nanotubes.16
The outer surface layers of a fish are covered with scales. These scales protect the body, reduces friction between the fish and the water and sense external stimulations from the environment by its nervous system. From such, the scales in theory should have some sort of piezoelectric functionality since neurons have a variety of voltage-sensitive ion channels that transmit signals. Though there are studies on the hierarchical structure, mechanical properties and chemical modifications of fish scales,17,18 not many investigations have been conducted on its piezoelectricity. Given the recent development of multiple modes in Piezoresponse Force Microscopy (PFM),19,20 it has now become possible to investigate biological piezoelectricity at the nanoscale level.21,22 With a unique method of obtaining hysteresis loops of biological systems,14 PFM is a powerful tool in studying the electromechanical properties of biological samples.23,24
In this work, the morphology and possible piezoelectric behaviors in fish scales were characterized directly using PFM. The possible origin and existence of piezoelectricity is discussed. In addition, an investigation on the Young’s modulus of the scales is also presented. Our results confirm the existence of piezoelectricity in some parts along the surface of the scales which is believed to be correlated to hydroxyapatite. The obtained results should spark more interest on further investigations of piezoelectricity in biomaterials.
II. METHODS
The scales employed in this study are from grass carp (Ctenopharyngodon idellus). The samples were rinsed with distilled water, dried in the air and cut into small pieces of approximately 30 mm2 in size.
The atomic force microscopy (AFM) images were captured with an Asylum Research unit model MFP-3D-SA AFM. For the in-contact PFM mode, probes (mode: ASYELEC-01) composed of silicon were used, in which the cantilevers and tips were coated with Ti/Ir (5/20). The spring constants were 2 N/m (nominal); with the free frequency in air set to 70 kHz (nominal); and the drive frequency set to 330 kHz. For harmonic responses of both 1st order and 2nd order, probes (mode:ARROW-contPt-50) composed of silicon were used, in which the cantilevers and tips were coated with Pt/Ir. The spring constants were 0.2N/m(nominal); with the free frequency in air set to 14kHz(nominal). The resonant amplitude is 93.45nm/V.
The force map measurements were conducted in contact mode. The probes (mode: Multi 75) used were composed of silicon with a spring constant of 2.8 N/m (nominal). The force map images were captured as 32 × 32 pixels where the scanning speed was 1 s/pixel. Other images were acquired as 512 × 512 pixels in ambient conditions. The scales were polished by sandpaper (granularity No: 800) since its inner layer cannot be measured directly. All the data were analyzed by the software package Igor Pro 6.32 (by Oxford Instruments).
The scanning Electron Microscopy (SEM) images were captured with a Tescan Vega3 Lmh SEM, in which energy dispersive X-ray analyses (EDAX) were attached. The surface of the scales were highly insulating so the samples were coated with a thin layer of gold prior to the SEM scans. The related data were analyzed by the software Tescan Vega TC.
III. MORPHOLOGY AND POSSIBLE PIEZOELECTRIC BEHAVIORS
Figure 1 shows the surface morphology of the fish scales. Two types of morphology are identified: one appearing smooth and another rough. Figure 1(c) shows an enlarged view of the smooth regions which is comprised of ridges and troughs. Figure 1(d) shows an enlarged view of the rougher regions which consists of a unique type of elevated crescents.
(a) Images of a Ctenopharyngodon idellus; (b) an enlarged view of one of its scales. Two types of morphology exist in the scales: (c) a smooth type characterized by ridges (I) and troughs (II) and (d) a rough type comprised of a mosaic of elevated crescents (III) with a flat base (IV).
(a) Images of a Ctenopharyngodon idellus; (b) an enlarged view of one of its scales. Two types of morphology exist in the scales: (c) a smooth type characterized by ridges (I) and troughs (II) and (d) a rough type comprised of a mosaic of elevated crescents (III) with a flat base (IV).
In order to investigate the morphology of the two areas at nanometer levels, SEM was utilized to obtain the overall morphology of the two areas as shown in Figure 2. The upper part of Figure 2 shows the details of the smooth areas. Domain I is shown as stripes. The averaged width of the ridges (designated as domain I from here on after) is 8.0 ± 0.6 μm. The averaged distance between two adjacent domain Is (i.e. the average width of a trough labeled as domain II) is 41.8 ± 1.5 μm. Similar patterns were reported in other type of fishes.25 The lower part of Figure 2 shows the morphology of the rough areas. The average width and length of a domain III is 95.0 ± 5.2 μm and 279.2 ± 30.4 μm, respectively. The width of the flat domains IV is 230.1 ± 28.2 μm. Other images were acquired with 512 × 512 pixels at ambient conditions. The typical topography, amplitude and phase images for domains I and II (smooth areas) are shown in Figure 3. The same scans were also conducted for the rough areas (see Figure 4). In order to confirm whether there is piezoelectricity in the fish scales, phase-voltage and amplitude-voltage loops were performed by PFM shown in the upper and lower panels of Figure 5 for the smooth and rough parts, respectively. In the smooth areas, a hysteresis loop and an amplitude butterfly curve was observed for domain I (blue loops) which verifies that domain I is piezoelectric. Similar hysteresis loops were also observed for domain IV (Figure 5(c) (d)) indicating that it is also piezoelectric. Interestingly, piezoelectricity is found in regions that are elevated (I) and non-elevated (IV) in distinct parts of the fish scale. To our knowledge, there is no existing literature showing the existence of piezoelectricity in fish scales.
SEM images of the four domains. (a) Magnified image of a typical ridge (I) region. (b) Magnified image of a trough (II) region. (c) Magnified image of a crescent (III). (d) Magnified image of the base (IV) belonging to the rough region.
SEM images of the four domains. (a) Magnified image of a typical ridge (I) region. (b) Magnified image of a trough (II) region. (c) Magnified image of a crescent (III). (d) Magnified image of the base (IV) belonging to the rough region.
Topography of the smooth areas taken by AFM imaging. (a) Detailed profile of a ridge (I) section with its (b) corresponding amplitude and (c) phase images. (d) Detailed profile of a trough (II) section (e) corresponding amplitude and (f) phase images.
Topography of the smooth areas taken by AFM imaging. (a) Detailed profile of a ridge (I) section with its (b) corresponding amplitude and (c) phase images. (d) Detailed profile of a trough (II) section (e) corresponding amplitude and (f) phase images.
Topography of the rough areas taken by AFM imaging. (a) Detailed profile of a crescent plateau (III) with its (b) corresponding amplitude and (c) phase images of the same domain III. (d) Detailed profile of the base (IV) region with its (e) corresponding amplitude and (f) phase images of the same domain IV.
Topography of the rough areas taken by AFM imaging. (a) Detailed profile of a crescent plateau (III) with its (b) corresponding amplitude and (c) phase images of the same domain III. (d) Detailed profile of the base (IV) region with its (e) corresponding amplitude and (f) phase images of the same domain IV.
Phase-voltage and amplitude-voltage loops acquired from different parts of the scales. (a) Phase-voltage and (b) amplitude-voltage loops from domains I and II depicted by blue and red curves, respectively. (c) Phase-voltage and (d) amplitude-voltage loops from domains III and IV depicted by blue and red curves, respectively.
Phase-voltage and amplitude-voltage loops acquired from different parts of the scales. (a) Phase-voltage and (b) amplitude-voltage loops from domains I and II depicted by blue and red curves, respectively. (c) Phase-voltage and (d) amplitude-voltage loops from domains III and IV depicted by blue and red curves, respectively.
To ensure that the strain we have measured is not dominated by quadratic electrostrictive responses, we examine first and second harmonic strain responses of domain II and domain IV where we think the piezoelectricity exists. Since the first harmonic response is associated with linear effects and the second harmonics response is associated with quadratic effects,20 it will be clear whether the conclusion that this material’s piezoelectricity is reliable. As shown in Figure 6 (a) and (b), the quadratic response is much smaller than first-order responses. Our result matches the feature of typical piezoelectric materials well. Figure 6 (c) and (d) are the straight line fitting for the amplitude of first harmonic response of domain I and domain IV using a series of bias. The r-square of each line is 0.987 and 0.990 which means the linear result is convincible.
First and second harmonic responses of (a) domain I and (b) domain IV. The straight line fitting for the amplitude of first harmonic response of (c) domain I and (d) domain IV by applying a series of bias.
First and second harmonic responses of (a) domain I and (b) domain IV. The straight line fitting for the amplitude of first harmonic response of (c) domain I and (d) domain IV by applying a series of bias.
IV. ELEMENTAL COMPOSITION ANALYSIS
In an attempt to search for the origin of the piezoelectricity, elemental composition analysis on all four of the domains by EDAX was performed (see Table I for the results of the elemental composition analysis). As reported before, the main constituent of the outer surface of fish scales is hydroxyapatite18,25 which is known to be a piezoelectric molecule.26,27 In hydroxyapatite, the mass ratio of Calcium and Phosphorous (Ca/P) = 2.2. As listed in Table I, the Ca/P mass ratios of domains I and IV, which are piezoelectric, are equal or very close to the reported value for hydroxyapatite. On the contrary, the Ca/P mass ratios of domains II and III are quite different from that of hydroxyapatite, which suggests that there are no traces of hydroxyapatite in domains II and III, which explains why there is no piezoelectric response in these domains. The results of the elemental compositions analyses coincide very well with the PFM scans. Our findings indicate that hydroxyapatite is present on the outer surface of the scales, and hydroxyapatite is the main component that contributes to the piezoelectricity of the scales.
Elemental composition analysis of green carp fish scale.
area . | domain . | Ca . | P . | Ca/P . | piezoelectricity . | hydroxyapatite . |
---|---|---|---|---|---|---|
smooth | I | 30.9a | 14.0 | 2.2 | + | + |
II | 14.0 | 4.2 | 3.3 | - | - | |
rough | III | 21.8 | 11.5 | 1.9 | - | - |
IV | 22.3 | 9.7 | 2.3 | + | + |
area . | domain . | Ca . | P . | Ca/P . | piezoelectricity . | hydroxyapatite . |
---|---|---|---|---|---|---|
smooth | I | 30.9a | 14.0 | 2.2 | + | + |
II | 14.0 | 4.2 | 3.3 | - | - | |
rough | III | 21.8 | 11.5 | 1.9 | - | - |
IV | 22.3 | 9.7 | 2.3 | + | + |
Mass fraction (%); +/-: with/without. The results are the averaged data (n = 10).
V. FORCE CURVES AND YOUNG’S MODULUS
To further verify the origins of piezoelectricity, the cross sectional areas and inner surfaces of the fish scales were analyzed (see Figure 9). No traces of Calcium and Phosphorous elements were found in these areas which are in accordance with a former study concluding that only proteins reside in the inner layers of fish scales.25
Force curves were conducted on every pixel in the images (32 × 32 pixels) to determine the elasticity morphology of the scales. Figures 7 and 8 show the smooth and rough areas, respectively. The Young’s modulus can be calculated by the Derjaguin-Muller-Toporov (DMT) model. In the smooth areas, the Young’s modulus of domains I and II (Figure 7) are found to be 7.4 ± 2.2 GPa and 22.1 ± 1.0 GPa, respectively. In the rough areas, the Young’s modulus of domains III and IV (Figure 8) are 11.1 ± 1.2 GPa and 5.9 ± 2.0 GPa, respectively. These results are consistent with the data from other types of fish scales.25 In the piezoelectric domains (I and IV), the Young’s modulus are less than 10 GPa. This further substantiates that there is a mineral layer on the outer surface of the scale, from which the main component is hydroxyapatite. As a comparison, the Young’s modulus of demineralized inner scales is only about 36 MPa,18 which is 2∼3 orders of magnitude smaller than that of the mineralized regions of the scales.
(a) The topography and its (b) corresponding force map and (c) histograms of the Young’s modulus for a domain I sample. (d) The topography and its (e) corresponding force map and (f) histograms of the Young’s modulus for a domain II sample.
(a) The topography and its (b) corresponding force map and (c) histograms of the Young’s modulus for a domain I sample. (d) The topography and its (e) corresponding force map and (f) histograms of the Young’s modulus for a domain II sample.
(a) The topography and its (b) corresponding force map and (c) histograms of the Young’s modulus for a domain III sample. (d) The topography and its (e) corresponding force map and (f) histograms of the Young’s modulus for a domain IV sample.
(a) The topography and its (b) corresponding force map and (c) histograms of the Young’s modulus for a domain III sample. (d) The topography and its (e) corresponding force map and (f) histograms of the Young’s modulus for a domain IV sample.
SEM images taken along the perpendicular direction of the cross-sectional section of the smooth and rough areas. (a) The morphology of the smooth area and its (b) magnified image. (c) The morphology of the rough area and its (d) magnified image.
SEM images taken along the perpendicular direction of the cross-sectional section of the smooth and rough areas. (a) The morphology of the smooth area and its (b) magnified image. (c) The morphology of the rough area and its (d) magnified image.
VI. CONCLUSION
The morphology and piezoelectricity of fish scales of Ctenopharyngodon idellus were investigated by single-molecular approaches (mainly by SEM, AFM and PFM). Our results show that: (1) two distinct types of regions exist on the surface: a smooth type that possesses ridges (domain I) and troughs (domain II); and a rough type that is comprised of an elevated mosaic of crescents (domain III) above a flat surface (domain IV). (2) It is found that only domains I and IV are piezoelectric, and its origin is attributed to the presence of hydroxyapatite. (3) Studies on the cross sectional areas of the scales reveal that hydroxyapatite is found only on the surfaces and is not present below. As easily accessible natural polymers, fish scales can be employed as highly sensitive piezoelectric materials in high sensitive and high speed devices as well as be exploited for invasive diagnostics and other biomedical implications.
ACKNOWLEDGMENTS
This work was supported by the National Natural Science Foundation of China (Grant Nos. 11474146, 11504161, and U1532142), the Hong Kong, Macao and Taiwan Science & Technology Cooperation Program of China (Grant No. 2015DFH10200), the Science and Technology Research Items of Shenzhen (Grant No. KC2015ZDYF0003A) and a Nanshan District Key Lab funding. We also thank the help and discussions with Dr. Fu and his coworkers from Shenzhen Key Laboratory of Nanobiomechanics, Shenzhen Institutes of Advanced Technology.