Size Estimation of Biological Ink Particles Dispersed in Liquids Using Atomic Force Microscopy

Toshihiko Matsuura*, Takamine Kato, Makoto Horii, Shohei Todo, Ken-ichi Minato, and Takashi Ueno 1 Laboratory of Biotechnology, Hokkaido University of Education, Hakodate, Hokkaido 040-8567, Japan 2 Department of Electrical and Electronic Engineering, Hakodate National College of Technology, Hakodate, Hokkaido 042-8501, Japan 3 Department of Material and Environmental Engineering, Hakodate National College of Technology, Hakodate, Hokkaido 042-8501, Japan


Introduction
][3][4] They are utilized not only as food additives but also as pigments (sepia) all over the world. 5)Recently, Ueno and coworkers have developed a method for purifying size-controlled ink particles isolated from squid ink sacs. 6)This purification method enables us to prepare ink particles of different sizes in the monodisperse state.Electron spin resonance (ESR) spectra demonstrated that the size-controlled ink particles are highly pure and are of reliable quality compared with a commercial synthetic melanin. 7)The ink particles have a potential application as nonpoisonous black inks for inkjet printers. 8)Also, it has been reported that the mixture of size-controlled ink particles in the titania paste enhances porous structures in TiO 2 films for dye-sensitized solar cells (DSSCs). 9)When the TiO 2 film electrodes prepared using a paste mixture of titania and ink particles were sintered at 450°C, the color of the electrodes changed to white.This temperature was in good agreement with that of the thermal decomposition of ink particles.In addition, the atomic force microscopy (AFM) analysis suggested that the thermal decomposition of the ink particles produces many pores in the films.The ink particles in the paste consequently enhanced the film porosity.The effects of the paste composition on the film morphology and the performance of DSSCs using a standard dye were also revealed.To improve the DSSC performance of porous TiO 2 film electrodes utilizing ink particles, it is indispensable to precisely measure the diameter of the ink particles because it may affect the film porosity.
Dynamic light scattering (DLS) and scanning electron microscopy (SEM) are used widely to measure biological particle sizes.DLS is a conventional technique used to measure biological particles dispersed in liquids, but accurate results are obtained only for the particles having a narrow size distribution. 10)SEM provides an impression of the size, number, and form of biological particles, but the particles are subject to deformation caused by the high vacuum conditions and the coating processes for sample preparation. 11)FM is a powerful technique for measuring the size distribution of biological particles directly under an ambient atmosphere, because the particles can be imaged one by one without the high vacuum condition and the coating processes for sample preparation. 12,13)The structural conformation of biomolecules deposited on solid substrates generally differs from that dispersed in water. 14)anno et al. suggested a suitable method for examining the size distribution of vesicles on mica directly by AFM. 15)They succeeded in calculating the diameters of vesicles dispersed in aqueous suspensions from the surface areas of dome-shaped structures that collapsed as a result of vesicle adsorption on mica.
In the present study, AFM was used to determine the size distribution of the ink particles deposited on mica.The AFM observation under an ambient atmosphere can be conducted without the high vacuum condition and the coating processes, and is therefore a simple and easy approach.However, the deformation of the ink particles on mica was induced by the meniscus force due to the surface layer of water under an ambient atmosphere.From the heights and widths of their distorted structures on solid substrates, we will suggest a method for estimating the diameters of the ink particles dispersed in aqueous suspensions.The estimation method in the present study would be a valid technique for precisely measuring the size distribution of biological particles by AFM.

Experimental procedure
In accordance with the procedures described in the literature, 6) ink particles of approximately 300 nm in diameter were isolated from the ink sacs of the neon flying squid (Ommastrephes bartramii) by ultrafiltration (Millipore Labscale TFF System).The obtained ink particles were dispersed with ultrapure water.The concentration of the aqueous suspension was adjusted to a range of 5.0×10 -2 -1.0×10 2 g/l.The sample was never frozen during purification or in storage.A similar procedure was performed for preparing the ink particles isolated from the ink sacs of the common cuttlefish (Sepia officinalis).
Mica was used as a substrate for the visualization of the ink particles using AFM, because freshly cleaved mica has an atomically flat surface.Mica (Okenshoji #08-1101) was cut into pieces of 1 × 1 cm 2 using scissors.The substrate was glued to steel discs with a double-coated tape, and was cleaved with a mending tape immediately before use.Five microliters of the aqueous suspension was placed on a freshly cleaved piece of mica so as to cover the entire surface.The suspension was allowed to stand on the mica for approximately 1 min, and then was blown off with air.After sufficient evaporation of the solvent, the sample was observed by AFM in air.
Two silicon cantilevers with different spring constants were used for comparison.One was a standard stiffness cantilever (Bruker RTESP) with a nominal spring constant of ca.40 N/m and a curvature radius of ca. 8 nm.This spring constant is almost the same as that used in the previous studies. 16)The other was a medium stiffness cantilever (Olympus OMCL-AC200TN) with a nominal spring constant of ca. 9 N/m and a curvature radius of ca.7 nm.AFM observation was conducted under an ambient atmosphere (50 ± 10% relative humidity) using a scanning probe microscope (Shimadzu SPM-9500J3) in a dynamic mode.The scanning rate was between 0.3 and 1 Hz at 512 lines per frame.All the AFM measurements were performed at room temperature.The heights and widths of ink particles were recorded by analyzing cross-sections using the supplied software (Shimadzu SPM manager 2.11).

Dispersion of ink particles on mica
AFM images of the ink particles isolated from the ink sacs of squid were obtained using standard stiffness cantilevers.Figure 1 shows typical AFM images of the squid ink particles deposited on mica.The brighter region in the topographic images represents an elevated height.The white spherical protrusions correspond to single ink particles.These images indicate that the number density of the ink particles on mica is related to the concentration of the loading suspension.At 10 g/l, the ink particles aggregated laterally, as shown in Fig. 1(a).The two-dimensional aggregates were observed irrespective of whether or not an ultrasonic procedure was performed on the loading suspension.It is difficult to conclude that the aggregates were fabricated in suspension because ink particles were prepared in the monodisperse state.Therefore, it seems more reasonable to suppose that the evaporation of the solvent caused the formation of aggregates of ink particles. 17) submonolayer of coverage of the ink particles on mica was a limit in the sample concentration range used in the present study.When the concentration of the loading suspension was decreased, AFM images revealed that each particle was dispersed on mica.At 2 g/l, sparser coverage was achieved, with single ink particles isolated from one another as shown in Fig. 1(b).
In the case of commercial Sepia inks, the sample contains nearly spherical particles that range from 2 to 20 µm in diameter. 18)Multi-micrometer-sized aggregates comprised of spherical particles with an average diameter of approximately 150 nm were observed by SEM and AFM. 11,19)The 150 nm particles were liberated from the aggregates by mild sonication. 18)The formation of the multi-micrometer-sized aggregates probably depends on the method of sample preparation.In the present study, such huge aggregates were not observed.This was due to the fact that the ink particles prepared according to the previously described methods 6) were dispersed thoroughly in water.Therefore, the ink particles used in the present study would be suitable for precise measurements of the size distribution of ink particles.

Size distribution of the squid ink particles
The cross sections indicated by the solid white lines in the topographic images (Fig. 1) provide the height and width of the squid ink particles deposited on mica.We measured the heights of the squid ink particles (n = 478) selected randomly from AFM images obtained using standard stiffness cantilevers.A histogram was constructed to examine the height distribution of the observed ink particles, and then was fitted with a Gaussian distribution, as shown in Fig. 2(a).The distribution curve revealed that the mean height of the ink particles on mica was 176 nm, with a standard deviation of 38 nm.The AFM analysis was also performed on the apparent lateral size of the ink particles on mica.We assumed that the effect of the cantilever tip shape on the measurement was negligible because the size of the ink particles was remarkably larger than the curvature radius of the tip.However, the widths of the ink particles at half-maximum were measured to prevent the overestimation of width that may have occurred as a result of the tip-sample convolution effect. 20,21)igure 2(b) shows a histogram of the width of the ink particles at half-maximum (n = 519).The solid line corresponds to a Gaussian distribution.The mean width of the ink particles at half-maximum was estimated to be 340 nm, with a standard deviation of 76 nm.The height and width data demonstrate that the ink particles on mica were not complete spheres but rather were somewhat distorted.This deformation might have been caused mainly by the meniscus force due to the surface layer of water under an ambient atmosphere. 13,22)However, the size of the ink particles dispersed in aqueous suspensions could be estimated from the height and width of the distorted structure on mica, as we discuss below.To investigate the effects of the spring constant, topographic images of the squid ink particles deposited on mica were also acquired using medium stiffness cantilevers, as shown in Fig. 3.In the cross section, the height of the ink particles seemed to increase compared with the heights shown in Fig. 1. Figure 4(a) summarizes the heights of the squid ink particles (n = 300) selected randomly from AFM images obtained using medium stiffness cantilevers.The histogram is fitted with a Gaussian distribution.The distribution curve revealed that the mean height of the ink particles on mica was 196 nm, with a standard deviation of 30 nm.The width distribution of the ink particles measured by AFM is shown in Fig. 4(b).The solid line corresponds to a Gaussian distribution.The mean width of the ink particles at half-maximum was estimated to be 325 nm, with a standard deviation of 46 nm (n = 300).AFM analysis using medium stiffness cantilevers suggested that the height of the ink particles was 10% larger than that using standard stiffness cantilevers.On the other hand, the width of the ink particles was slightly smaller when using medium stiffness cantilevers.Thus, it was clarified that the spring constant of cantilevers affects the size measurement of biological particles.On the other hand, DLS analysis revealed that the mean diameter of the squid ink particles was 315 nm, with a standard deviation of 170 nm. 6) We therefore found discrepancies between the AFM and DLS data.

Size estimation of ink particles in aqueous suspension
The heights of the squid ink particles determined by AFM were significantly smaller than their diameters dispersed in aqueous suspensions determined by DLS. 6) In contrast, their widths determined by AFM were wider than the values reported in the DLS studies.The height and width distributions obtained from AFM analysis demonstrated that the ink particles on mica were hemisphere-like rather than spherical.We therefore attempted to estimate the diameters of the ink particles in aqueous suspensions (D 0 ) from the heights (h) and widths (D) of the hemisphere-like structure on mica in air.Given that the ink particles dispersed in aqueous suspension are spherical (Fig. 5(a)), the surface area of a particle (A 0 ) is defined as If the ink particles deposited on mica are assumed to be partial spheres (Fig. 5(b)), the surface area of the particles (A) is expressed as Here, h and D are the height and diameter of a partial sphere, respectively.c ( = �ℎ( − ℎ) ) is the radius of the bottom circle.The surface area of the ink particles on mica should be consistent with that dispersed in aqueous suspensions.Hence, By introducing Eqs. ( 1) and ( 2) into Eq.( 3), we obtain the following equation: This equation is an expansion of the model that calculates the diameters of spherical particles in suspension from the surface area of the dome-shaped structure that collapsed by adsorption of the particles on mica. 15)This equation enables us to easily estimate the diameter, although the equation derived from the sameness of volumes requires us to solve the cube root to estimate the diameter.Table I summarizes the heights and widths of the squid ink particles on mica, and the diameters estimated from Eq. ( 4).Regardless of the spring constants of AFM cantilevers, the diameters of the ink particles in aqueous suspensions were estimated to be approximately 300 nm, close to the value reported in the DLS studies (315 ± 170 nm). 6)As compared with the DLS studies, the standard deviations estimated from the AFM analysis were decreased, meaning that a precise estimation was achieved.We succeeded in precisely estimating the sizes of the ink particles dispersed in aqueous suspensions from the height and width of the distorted structure on mica in air.However, our estimation may have yielded diameters that are slightly small, because the widths of the ink particles at half-maximum were measured to prevent overestimation.Our estimation technique should be applicable for measuring the size distribution of other biological particles by AFM.

Size distribution of the cuttlefish ink particles
The size estimation from AFM was applied to the size distribution measurement of the ink particles isolated from the ink sacs of cuttlefish.AFM images of the cuttlefish ink particles were acquired using standard stiffness cantilevers.Figure 6 shows a typical AFM image of the cuttlefish ink particles deposited on mica.The white spherical protrusions with an average diameter of approximately 110 nm correspond to individual ink particles.We measured the height and width of the cuttlefish ink particles selected randomly from AFM images obtained using standard stiffness cantilevers.The height and width distributions of the cuttlefish ink particles measured by AFM are shown in Fig. 7.The solid lines correspond to a Gaussian distribution.The distribution curve shown in Fig. 7(a) revealed that the mean height of the ink particles on mica was 111 nm, with a standard deviation of 35 nm (n = 762).On the other hand, the mean width of the ink particles at half-maximum was estimated to be 204 nm, with a standard deviation of 27 nm (n = 382) from Fig. 7(b).Table II summarizes the height and width of the cuttlefish ink particles on mica, and the diameter estimated from Eq. ( 4).The diameter of the ink particles in aqueous suspension was estimated to be approximately 180 nm, close to the value reported in the DLS studies (176 ± 81 nm). 6)As compared with our AFM studies, previous SEM studies suggested that the cuttlefish ink particles have smaller diameters (ca.150 nm). 18)It is well known that the process of sample preparation for SEM often induces the deformation of biological samples and causes image artifacts. 11)In SEM imaging, biological samples are completely dry, since the specimen chamber is under a high vacuum.It is also required that the biological samples are electrically conductive and electrically grounded to prevent the accumulation of electrostatic charge at the surface during electron irradiation.In the conventional SEM measurement, the ink particles are coated with an ultrathin layer of an electrically conductive material such as gold/palladium alloy. 18)It is quite likely that the severe conditions in the sample preparation enhanced the deformation of the ink particles.The AFM observation under an ambient atmosphere, however, can be conducted without the high vacuum condition and without the coating processes, and is therefore a simpler, easier approach.The deformation results from the meniscus force, but the size estimation method suggested in the present study enables us to precisely estimate the diameters of size-controlled ink particles isolated from the ink sacs of both squid and cuttlefish.Thus, this estimation method would be a valid technique for measuring the size distribution of biological particles by AFM.
Table II.The height and width at half-maximum of the cuttlefish ink particles deposited on mica, and the diameter of the ink particles dispersed in aqueous suspensions estimated from Eq. (4).

Conclusion
The size distributions of size-controlled ink particles isolated from the ink sacs of squid and cuttlefish were determined by AFM.Topological images acquired using standard stiffness (ca.40 N/m) cantilevers suggested that the squid ink particles deposited on mica were not spherical but instead were hemisphere-like.The AFM analysis revealed that the heights and widths were 176 ± 38 nm and 340 ± 76 nm, respectively.In the AFM analysis performed using medium stiffness (ca. 9 N/m) cantilevers, the heights and widths of the squid ink particles deposited on mica were 196 ± 30 nm and 325 ± 46 nm, respectively.We found discrepancies among these values obtained using cantilevers with different spring constants.From the height and width data, however, the diameters of the ink particles in aqueous suspension were estimated to be approximately 300 nm.The estimated values were almost the same regardless of the spring constant of cantilevers, and were in good agreement with that obtained from the DLS analysis.As compared with the DLS studies, the standard deviations estimated from the AFM analysis were decreased, meaning that a precise estimation was achieved.A precise estimation was also achieved for the cuttlefish ink particles.Therefore, the estimation method suggested in the present study would be useful for precisely measuring the size distribution of biological particles by AFM.

Fig. 1 .
Fig. 1. (Color online) Typical AFM images of the squid ink particles deposited on mica.These were obtained using standard stiffness cantilevers.The cross sections indicated by the solid white lines in the images are shown below each image.(a) The ink particles aggregated at a loading suspension of 10 g/l.(b) At 2 g/l, sparser coverage was achieved, with single ink particles isolated from one another.

Fig. 2 .
Fig. 2. (Color online) Histograms of (a) height and (b) width of the squid ink particles selected randomly from AFM images obtained using standard stiffness cantilevers.The solid lines correspond to a Gaussian distribution.

Fig. 3 .
Fig. 3. (Color online) A typical AFM image of the squid ink particles acquired using medium stiffness cantilevers.

Fig. 4 .
Fig. 4. (Color online) Histograms of (a) height and (b) width of the squid ink particles selected randomly from AFM images obtained using medium stiffness cantilevers.The solid lines correspond to a Gaussian distribution.

Fig. 5 .
Fig. 5. (a) Spherical surface of a sphere with diameter D 0 .(b) Spherical surface of a partial sphere with the diameter D and height h.c is the radius of the bottom circle.

Fig. 6 .
Fig. 6. (Color online) A typical AFM image of the cuttlefish ink particles acquired using standard stiffness cantilevers.

Fig. 7 .
Fig. 7. (Color online) Histograms of (a) height and (b) width of the cuttlefish ink particles selected randomly from AFM images obtained using standard stiffness cantilevers.The solid lines correspond to a Gaussian distribution.

Table I .
The heights and widths at half-maximum of the squid ink particles deposited on mica, and the diameters of the ink particles dispersed in aqueous suspensions estimated from Eq. (4).