HA was synthesized by a precipitation reaction between calcium hydroxide and orthophosphoric acid. Both reagents were AnalaR grade. 0.3 M orthophosphoric acid was added drop-wise to a 0.5 M calcium hydroxide solution under continuous stirring at room temperature, while the pH was kept above 10.5 by the addition of ammonia solution. Stirring was maintained for further 16 hours after complete addition of the reactants. The precipitate obtained was aged for a further week and then washed with boiling water. To make the HA suspension suitable for electrospraying, ethanol was added as a liquid carrier until the 3% wt concentration achieved. To investigate the influence of nozzle tip angle to the stable cone-jet domain, such HA suspension was used as the spraying liquid during the process.

As shown in Figure 1(a), basically, the setup consists of a ground electrode and a nozzle, which is connected to a high voltage supply (Glassman Europe Ltd., Tadley, UK). The liquid is fed to the nozzle by an infusion pump (Harvard apparatus PHD 4400) at a controlled flow rate. The flowing medium is subjected to an electrical field, which is generated by the applied voltage between the nozzle and the ground electrode. Five different stainless steel nozzles were designed for the study. The nozzle tip angle varies from 15˚ to 90˚, as shown in Figure 1(a). The internal and the outer diameter of the nozzles are 800 μm and 1100 μm, respectively.

To investigate the influence of nozzle tip angle on the stability of the cone-jet mode electrospraying, the nozzles afore-mentioned were used. In these experiments, the nozzle tip to substrate distance was kept at 40 mm. HA suspension (3 wt%) was syringed to the nozzle, and the flow rate was varied from 5 μl·min^-1 to the maximal value for stable cone-jet mode spraying. The applied voltage between the nozzle and the ground electrode was varied up to 7 kV to establish the modes of electrospraying. A high speed digital camera (Weinberger AG, Dietikon, Switzerland) in conjunction with a computer was used to observe and characterize the cone-jet mode. At a chosen flow rate, the cone-jet mode was achieved within a range of applied voltage. Thus, the relationship between the applied voltage and flow rate for the suspension jetting in the cone-jet mode was established. To compare and confirm the results got from the HA suspension, pure ethanol, which had been proven to give a stable cone-jet mode spraying, was also used as the process liquid with the same procedures.

To investigate the influence of nozzle tip angle on the deposited relic size distribution, HA suspension was syringed into five different nozzles. The nozzle tip-ground electrode distance and liquid flow rate were set at 40 mm and 10 μl/min, respectively. Microscope glass slides were used as substrates to collect the deposited relics. The collecting time was controlled at 10 seconds. Three deposited glass slides were prepared for each nozzle. The relics collected on the glass slides were then examined using an optical microscope. 50 relics on each glass slide were stochastically chosen and measured.

Due to the importance of the relic size to TAEA patterning process, as reported previously [23], the nozzle which gives the smallest relics was used for TAEA patterning process (Figure 1(b)) to verify the findings. Pure

titanium was used as substrate. The nozzle tip-substrate distance and the liquid flow rate were controlled at 40 mm and 10 μl/min. A copper TEM grid and 3 wt% HA suspension were used as the template and the liquid. The spraying time was set at 10 minutes. The patterns prepared were examined and measured using the optical microscope and the scanning electron microscope (FE-SEM, JEOL JSM/6310F).

It is well-known that the flow liquid could form a conical shape meniscus when the electric driving force is controlled at specific range. Using a conventional nozzle with the tip angle of 90˚, the liquid cone formed symmetrically near the centre of the nozzle orifice (Figure 2(e)). However, it was observed an entirely different scenario while a nozzle with tilted outlet was used. As shown in Figures 2(a)-(d), the liquid cone and jet locate near the sharp tip rather than in the centre of the nozzle orifice. It was also observed that the diameter of the liquid cone dramatically reduced compared with the liquid cone formed when using the nozzle with 90˚ tip angle. This phenomenon demonstrates that the electric field, which drives the liquid to form the cone shape meniscus, may not be symmetry, and an electric field concentration may occur at the sharp tip of the tilted nozzle.

The relationship between the nozzle tip radius and the electric field was investigated by Jones and Thong in 1979 [24]. It was assumed that, irrespective of geometrical

considerations, the same field conditions are required for the onset of the cone-jet mode. The electrical potential V will be related to the electrical field E generated in the liquid surface situated at the capillary edge, by

where f(R, h) is a geometrical function of the nozzle tip radius, R, and the nozzle tip-grounded plane distance, h. An accepted approximation [12] for this function is given in Eq.2.

where A is an empirical constant ranging from 0.5 to 0.707. Variation of the f(R, h) value with respect to changes in R can be obtained from

when during jetting,. This means that while the nozzle tip radius reduces, the f(R, h) value will reduce, and the electric field at the nozzle tip will increase accordingly when the same electrical potential is applied.

This relationship was investigated based on an assumption that the nozzles used are of the regular shape, that the tip angle is 90˚. It seems to be not applicable in our case. However, if taking an in-sight analysis of the cone-jet mode formed using the tilted nozzles, a linkage between the nozzle tip angle and an equivalent nozzle orifice radius could be observed.

As shown in Figure 2, high speed camera images clearly indicate the changes in the jetting towards a grounded plate. The diameter of the liquid conical meniscus, highlighted in red, formed under the varied nozzles was found to increase significantly with the increase of the tip angle. Cones of the minimum and maximum diameter appear while the 15˚ and 90˚ nozzle were used. To study and compare the electric field distribution at the nozzle tip, an equivalent nozzle of 90˚ tip angle, as highlighted in yellow in Figure 3(a), is assigned to all the nozzles. The relationship between the equivalent nozzle diameter and the nozzle tip angle was investigated. As shown in Figure 3(b), the relationship follows

where A and B are 83 and 7, respectively. L is the equivalent nozzle diameter and α is the nozzle tip angle. Clearly, according to Eq.4, the equivalent nozzle diameter of the titled nozzle decreases with the tip angle, which means L_15˚ < L_30˚ < L_45˚ < L_60˚ < L_90˚. Due to the relationship of the conventional nozzle radius and the electric field as afore-introduced, the electric field at each nozzle tip can be compared as shown in Figure 3(c). The relationship between the electric field and nozzle tip angle is given by:

where A, B and C are 7, 14 and 8, respectively. Therefore, the electric field concentration occurs at the sharp nozzle tip. The relationship of the electric field near the nozzle tip follows E_15˚ > E_30˚ > E_45˚ > E_60˚ > E_90˚.

When using a 3% wt HA suspension as the liquid for electrospraying, the relationship between applied voltage and flow rate is shown in Figure 4. The cone-jet mode can only be achieved when the flow rate and the applied voltage are in a distinctive domain. The maximal liquid flow rates for achieving a stable cone-jet mode using the nozzle of 15˚, 30˚, 45˚, 60˚ tip angles exhibit much higher than that of 90˚ nozzle. The nozzle with 45˚ tip angle shows the highest flow rate up-limit which is approximately 68 μl/min. It is almost double of the maximal flow rate while using the nozzle with 90˚ tip angle. The voltage range (ΔV between maximal and minimal applied voltage) at a specific flow rate was also found to decrease with the decrease of the nozzle tip angle. To confirm this finding, pure ethanol, which is a conventional liquid for electrospraying research, was also used. The similar scenario was observed. The highest flow rate up-limit for stable cone-jet mode was also found while using the nozzle of 45˚ angle. The difference of flow rate up-limit between the highest and the lowest was found to be further enlarged.

The relic size distribution diagrams were built using 3% wt HA suspension (Figure 5). It was found that the main relic size dramatically decreased with the decreasing nozzle tip angle. The main size of the HA relics deposited varied between 2 to 3 μm using the nozzle of 15˚ tip angle, which was nearly half of the main relic size range when the nozzle with 90˚ tip angle was used. This phenomenon is mainly due to the electric field concentration as demonstrated. As the same flow rate was set during the HA relics collection, more electric charge presented at the liquid cone formed at the sharper nozzle tip. Subsequently, the liquid cone formed a jet and then broken up into droplets. Higher electric charge presented in the surface of the droplets, emitted from the cone-jet mode with a nozzle of smaller tip angle. Such phenomenon enhanced the possibility that the highly charged droplet to reach the Reileigh charge limit [11], as showed in the following formula.

.

where is the Rayleigh charge limit (C), is the surface tension, ε₀ is the vacuum dielectric constant and r_d is the droplet radius (m).

When the charge on a single droplet reaches the limit, droplet breaks up into two or more smaller droplets by the over-limited charge. Therefore, when using the nozzle with 15˚ tip angle, the droplet fission occurred, and smaller liquid droplets were therefore formed.

Template-assisted electrohydrodynamic atomization spraying (TAEA) was previously developed to create bioceramic patterns to up-regulate the bone cell behavior [23]. To improve the resolution (the minimal diameter of the pattern) of TAEA, the droplets with reduced dimensions are compulsory. Based on the findings above, the nozzle with 15˚ tip angle was chosen for HA patterning trials. As shown in Figures 6(a) and (b), the HA line pattern of < 10 μm in width and the fine square HA patterns of the width below 10 μm were successfully achieved. This is nearly one tenth in diameter of the patterns reported previously. The resolution of TAEA was dramatically improved, as expected. In addition, the inset micrograph of Figure 5 clearly confirms that dense surface topography assembled by the HA nanocrystals, which favors the cellular attachment.