Evaluation the effects of humidity and other process parameters on TiO2 Nanofibers by RSM (CCD) and experimental

INTRODUCTION

One-dimensional (1D) nano-structures such as nanorods, nanowires, and nanotubes and generally the nanocomposites containing various oxide materials have been widely considered by researchers (1) due to their various potential applications. Many applications of different nano-structures can be found in the literature such as solar cells (2, 3), photo-degradation (4, 5) and various other applications (6-9)

Recently, potential applications of titanium Dioxide (TiO₂) (including environmental regeneration, electronics, sensors, solar cells, and other related fields) (1-3, 10, 11)motivates researchers. It has been found that surface area property (e.g. large specific surface area) and consequently the mesoporosity of TiO₂ improves light absorption. Likewise, the particle dispersion increases the absorbance capacity of the reactants followed by more photocatalytic reactions (12). Therefore, Titanium Dioxide is a broadband semiconductor with many interesting features such as transparency to transmit visible light, suitable thermal expansion coefficient, and low absorption coefficient. Moreover, it is known as an excellent photocatalyst for degrading organic materials (13). Titanium has two natural polymorphs: rutile and anatase. Generally, only anatase has significant photocatalytic activity (14). Among photocatalyst, anatase phase of TiO₂ has been extensively studied. Because this phase of TiO₂ has high photocatalytic activity, chemical stability, environmentally friendly and cost-effective (1, 14). Accordingly anatase phase of TiO₂ improves stability of the Dye Sensitized Solar Cells (DSSC) (15).

There are several methods to synthesize TiO₂ nanofibers, such as sol-gel, electrospinning, etc. Among these, electrospinning is an excellent method for producing synthetic fibers like carbon, organic and inorganic hybrid nanowires and other fibers in nanoscale (16) due to low price, variability, and simple process (17).

The electrospinning process involves a high voltage source connected to a needle and a metal collector, where the fibers are accumulated (18). The needle is attached to an injection pump, where a polymeric solution is delivered. The needle is connected to the positive electrode and the collector is connected to the negative electrode. As a result, a difference in potential between the collector and the needle of the electric field is generated which causes drops on the needle to be drawn and nanofibers is formed (19). When the applied electric field overcomes the surface tension of a droplet, the solution is drawn, heading to the negative electrode as a jet to form fibers on the surface of the collector (20). As these fibers are in nanoscale, they have a very large surface-to-volume ratio. The diameter and morphology of the fibers can be adjusted by changing the rheological properties of the solution and the process parameters of the electrospinning (18).

The results of research studies on fibers synthesized by electrospinning indicate that several factors such as solubility, tip-collector distance, temperature, humidity, etc. affect their structure and morphological properties. Most electrospinning processes are generally performed in an uncontrolled environmental condition with unknown and variable relative humidity. However, to understand the formation of fibers under varying relative humidity conditions, limited number of studies have been carried out (21, 22).

Several published works on synthesis of titanium dioxide nanofibers are appeared in the literature, to study the effect of three key parameters (injection rate (mL/h), potential difference (kV), and synthesis solution concentration) on fiber diameter (18, 23). In some reports, the so-called Response Level Method (RSM) as statistic model has been used to predict the diameter of titanium dioxide in the electrospinning process. In a study conducted by Li and Xia (2003), three electrospinning parameters including the voltage between tip and collector and tip-collector distance (TCD, the distance between the positive and negative electrodes) were investigated by the Box-Behnken Design (BDD) method during synthesis of nanofibers (23). In another study, six effective parameters (TCD, voltage, flow rate, amount of PVC, amount of TTIP and pH) were studied using two statistical analyses, Response Level Method (RSM) and Box–Wilson central composite design (CCD) techniques (24). The RSM is a statistical method for analyzing the effects of several independent variables for the response. This method uses multivariable data in experiments that are designed to solve multivariate equations simultaneously (25). Additionally, the main advantage of RSM is to reduce the number of empirical experiments needed to evaluate different variables and their interactions. Therefore, RSM makes simplicity and reduces the time and cost more than that of classical methods spend to optimize a process (26). There was neither enough number of effective electrospinning parameters nor statistical validity to support results out of previous studies on nanofibers fabricated from TiO₂

In the current study, the synthesized TiO₂ nanofibers were characterized and analyzed under various processing conditions each in five levels. These parameters were flow rate of the synthesis solution from 0.1 to 0.9 mL/h, the voltage between the needle tip and the collector ranged from10 to 30 kV, the distance between the needle tip and the collector (TCD) ranged from10 to 30 cm, the percent of PVP in the polymeric solution (5 to 13%) and the ambient relative humidity under controlled conditions at five levels (ranged from 24 to 42 %).

By comparing the methodology of the current research with those of previous research works, its novelty is well appeared. While in a previous work (27) sampling was completely random, in the current research all parameters are designed and optimized using the Design Expert software, Response Surface Model (RSM) and Composite Central Design (CCD). Hence, the synthesis process is performed accurately, but at a lower cost and reproducibility is also improved. The outputs of the software are three-dimensional diagrams representing the effect of each parameter on diameter of the synthesized nanofibers were investigated. Moreover, in the previous studies the parameters were mostly investigated individually, but in the current study all of the parameters are examined in an integrated manner with more emphasis on humidity of the synthesizing space. Therefore, the current study has the excellence and innovation in examining titanium dioxide nanofibers as it shows the following attributes: 1) High test accuracy (low error), 2) Examine the process parameters at acceptable levels and 3) The study explored the impact of ambient relative humidity as a significant parameter on diameter of the synthesized nanofibers for the first time and with great accuracy of morphology and diameter.

MATERIAL AND METHODS

Solution preparation

In order to synthesis TiO₂ nanofibers by electrospinning technique, PVP (Mw 1,300,000 g mol^-1 and TTIP (Ti(OiPr)₄, both from Aldrich) were used as precursor materials. The preparation procedure of electrospinning solutions was performed in three steps. In the first step 2 mL of TIPP, 2 mL of absolute ethanol and 2 mL of glacial acetic acid were mixed for about 10 min. In the second step 0.1, 0.3, 0.5, 0.7 and 0.9 mL of PVP and 6 mL of absolute ethanol were mixed separately for about 10 min. In the last step, the solution containing the TTIP was added to the polymer solution dropwise and stirred for about 30 min. Finally, to make a viscous electrospinning solution, it was heated at 50 ºC for about 10 min.

Electrospinning

In the electrospinning process, the PVP/TIPP solution was loaded in a syringe with a 21 gauge stainless steel needle. The syringe was placed in the syringe pump and the flow rate levels were set as 0.1, 0.3, 0.5, 0.7 and 0.9 mL/h. The distance between the needle tip and collector was altered from 10 to 30 cm. To collect the synthesized nanofibers easily, the surface of collector was covered with an aluminum foil. One of the key parameters in electrospinning process is the voltage between the needle tip and collector (10) that was set between 10-30 kV. A typical electrospinning setup system is shown in Fig. 1. Each experimental run of the synthesizing nanofibers samples (Fig. 2) lasted for about 20 min.

The collected nanofibers were kept in laboratory for about two hours to remove moisture and then calcinated at 300, 400 and 500 ºC. During the calcination procedure, the diameter of the fibers was decreased as organic components such as PVP were burnt out. Therefore, calcination formed and changed the phase of fiber from anatase to rutile.

To control the relative humidity of the synthesizing space and provide various levels of this parameter, a closed chamber made of glass and a steam generator were provided to isolate the electrospinning space. The ambient relative humidity was set at five levels as 24, 28.5, 33, 37.5 and 42 %. The synthesized nanofibers can be used in many applications such as solar cells and photocatalyst activity.

Nanofiber Characterization

The diameter and morphology of the electrospun TiO₂ nanofibers were investigated by Scanning Emission Microscope (SEM) (Leo 1450VP, Germany) and Field Emission Scanning Electron Microscope (FESEM) (MIRA3 TE SCAN, Czech Republic). Morphological and structural analysis were studied by Transmission Electron Microscope (TEM) (Leo 912 AB, Germany). Crystallization measurements were performed by X-Ray Diffraction (XRD) (EXPLORER, Italy) with CuKα line, from 20 to 80^°. FTIR (AVATAR 370 FT-IR, USA) was used to find out the functional groups of the TiO₂ nanostructures.

Experimental Design

Using the Design Expert software version 10, a five-factor, five-level Central Composite Design (CCD) was established to verify the impact of process variables, namely PVP concentration (X₁), voltage (X₂), distance (X₃), humidity (X₄), and flow rate (X₅) on diameter of the TiO₂ nanofibers. Screening experiments were carried out to determine the range of variables. The parameters and their levels are shown in (Table 1). A total of 50 experimental runs were performed in triplicate, including 8 replicates of the center point for each variable.

Statistical analysis

In order to obtain the empirical models that predict the response, different regressions were fitted to the data. These models were statistically compared to select the appropriate model. It is noteworthy to say that a statistical model is suitable when “lack of fit” is not significant and it has the highest R², “R² adjusted” and “R² predicted”. To select the appropriate model for predicting the desired response, the lack of fit test was performed. The significance of a lack of fit test for a model indicates that the model does not have the ability to predict the results for new observations. Therefore, the insignificance of the lack of fit test means that the model is able to be fit well with the empirical data. The “F” test was also performed to compare and analyze the variance (ANOVA) of the models and parameters.

The analysis of variance was applied for the response. The model has been statistically significant about the response (P < 0.05). To show the significance of the variables on the response, “F- test” and (p< 0.05) were used. Non-significant variables were eliminated by the backward elimination algorithm, which reduced the number of terms in the polynomial models. The results of statistical analysis of the fitted model show that this model has optimum response to a high-performance and high-power response considering R², R² adjusted and R² predicted. All experiments were carried out randomly in triplicate.

The effect of independent variables on the response

In RSM, for each dependent variable there is a model that expresses the principal and reciprocal effects of factors on each separate variable. The multivariate model is written as:

Y = β₀ + β ₁X₁ + β ₂X₂ + β ₃X₃ + β ₄X₄ + β ₅X₅ + β ₁₂X₁X₂ + β ₁₃X₁X₃ + β ₁₄X₁X₄ + β ₁₅X₁X₅ + β ₂₃X₂X₃ + β ₂₄X₂X₄ + β ₂₅X₂X₅ + β ₃₄X₃X₄ + β ₃₅X₃X₅+ β ₄₅X₄X₅ + β ₁₁X₁² + β ₂₂X₂² + β ₃₃X₃² + β ₄₄X₄² + β ₅₅X₅² (1)

where, Y is the predicted response (diameter of nanofibers), X₁, X₂, X₃, X₄ and X₅ are the affecting parameters, β ₀ is the interception coefficient, β₁, β₂, β₃, β₄ and β₅ are constant coefficients. In other words, this equation expresses the relation between TiO₂ nanofibers’ diameter and all the affecting parameters taken into account in the current study.

Response surface

Surface plots provide the relationship between responses and the experimental levels of each parameter (variable), as well as the type of interactions between two test variables, and hence determine the optimal level of each factor. In each plot, the response surface level was plotted using the Design Expert software version 10. Moreover, empirical data were optimized, the experimental factors were examined in triplicate and theoretical and actual values were compared. Eventually, the effect of humidity of synthesis environment on TiO₂ nanofibers were investigated. In this section the main objective was to study the variation of diameter and morphology of nanofibers under varying ambient humidity.

RESULT AND DISCUSSION

As previously mentioned Design Expert software used and a five-factor, five-level Central Composite Design (CCD) was established to verify the impact of process variables, namely concentration (X₁), voltage (X₂), distance (X₃), humidity (X₄), and flow rate (X₅) on the diameter of TiO₂ nanofibers. The CCD matrix of the predicted and observed results of the electrospinning of TiO₂ nanofibers are presented in Table 2.

In order to obtain a suitable experimental model for predicting the response of Nanofibers diameter, various regressions were fitted to the data. These models were then analyzed to select the appropriate model. As illustrated in
Table 3, the best model to predict the responses was quadratic model.

From statistical point of view, a desired model is the one in which the goodness of fit test (p (lack of fit) > 0.05) is not significant, and also has the highest value of 𝑅² 𝑎𝑑𝑗𝑢𝑠𝑡𝑒𝑑 and 𝑅² 𝑝𝑟𝑒𝑑𝑖𝑐𝑡𝑒𝑑. The details of statistical analysis of the models fitting for responses are presented in (Table 4). According to this Table, the goodness of fit test of the fitted model on the responses was not significant. Therefore, the selected model and the test have a very high degree of certainty.

According to the ANOVA analysis for the responses (see Table 3 and Table 4), the quadratic model did not agree with the statistical tests reference values. It is noteworthy that after fitting, the model was subjected to an algorithm that eliminates feedbacks and hence the number of model parameters was decreased. Also, the fitness test for the responses was not significant. The significance of values of the test show that the studied model does not have the ability to predict the values of the function. In other words, the significance of the test does not mean that the model is able to be fit well in accordance to the test data. Based on the results presented inTable 3, in terms of statistical analysis of the fitted model, high-power and high-performance coefficients (𝑅², 𝑅² Adjusted and 𝑅² Predicted) for optimization are met.

As previously stated (Eq.1), RSM scheme is defined for each independent variable of the model indicating main and mutually exclusive effects of parameters for each variable. According to the analysis of variance (Table 4), it was determined that the main effects (X₁, X₂, X_3, X₄, X₅) and interactions (X₁X₂, X₁X₄, X₁X₅, X₂X₃, X₂X₄, X₄X₅₎ and quadratic effect (X₃²) on nanofiber diameter response were significant at the 99% (P < 0.0001) confidence level. Similarly, the effect of X₁X₃, X₃X₅, X₂², X₄² and X₅² was significant at 95% level (P < 0.05). However, other parameters did not have notable effect on response (diameter) at 95% level. The highest mean diameter was related to the sample 42 (3115 nm) and lowest value was recorded for the sample 2 (18 nm). Eventually, the predicting relationship was provided as follows:

Y (Diameter) = +324.14 + 430.11 X₁ - 205.33 X₂ - 219.48 X₃ - 287.58 X₄ + 154.69 X₅ - 213.58 X₁X₂ - 55.90 X₁X₃ - 123.53 X₁X₄ + 178.26 X₁X₅ + 96.08 X₂X₃ + 152.21 X₂X₄ - 31.38 X₂X₅ +19.53 X₃X₄ - 44.26 X₃X₅ - 144.14 X₄X₅ + 55.12 X₁² + 54.44 X₂² + 80.44 X₃² + 49.19 X₄² +49.44 X₅² (2)

The surface plots of representing interaction effects of the independent variables are depicted in Fig. 3a, the effect of polymer content (ranged from 7 to 11 percent), on nanofibers’ diameter, is very clear, however the effect of voltage (from 15 to 25 kilovolts), on diameter reduction is very low and not impressive. In general, the interaction of these two variables increases the diameter of nanofibers. Hence, it can be concluded that with a constant voltage, increasing the concentration of polymer, significantly increases the diameter of nanofibers, which agrees with a study led by Mali et al. (2015). They conducted similar study on polymer concentration for nanofiber synthesis. In this experiment, they concluded that the more polymer concentration, the bigger diameter of fibers (18).

According to the plot of Fig. 3b, the two factors of polymer content and distance of electrodes have different effects on the diameter of nanofibers. With decrease of electrodes distance and increase of polymer content, the fiber diameter grows. In other words, with a constant distance, with increase of polymer concentration, nanofiber diameter also increases. In contrast, with a constant polymer concentration, with increasing electrode distance, the diameter of the Nanofibers decreases. However, according to the interaction graph (Fig. 3b), the interaction of these two factors has an increasing effect on diameter of nanofibers. This is mainly due to the much greater impact of polymer concentration on the diameter of the nanofibers than the electrode distance. This finding is in agreement with Sarlak et al. (24) who stated that increasing polymer concentration had the greatest effect on the electrospun nanofiber diameter.

Increasing the relative humidity of the synthesis environment noticeably decreases the diameter of nanofibers which is seen in Fig. 3c. A fairly linear reduction in diameter is seen when relative humidity increases from 24 to 37.5 percent. In general, the interaction of polymer concentration and humidity increases the diameter of the nanofibers, marking greater effect of polymer concentration on the fiber diameter than the effect of relative humidity. The Injection rate of synthesis solution assessment indicates that an increase in flow rate led to an increase in diameter of the nanofibers (Fig. 3d). However considering the detailed data it is seen that, this parameter has the least effect on diameter of the nanofibers.

According to Fig. 3e an instantaneous reduction in voltage and electrode distance increases the diameter of nanofibers, however, the effect of the former is less important than the latter. Hence, the interaction effect of these two parameters also increases the diameter of nanofibers. Comparing the graphs of Fig. 3b and Fig. 3e it can be seen that the effects of these two factors are fairly less than the effect of polymer concentration on the diameter of nanofibers. Considering the effect of voltage and the relative humidity of environment (Fig. 3f), it is evident that both parameters reduce the diameter of nanofibers, nonetheless, the effect of humidity is greater than the voltage. In other words, the effect of relative humidity on reduction of nanofibers’ diameter is more noticeable than the voltage. Fig. 3g shows that the effect of increasing flow rate on the increase in diameter of nanofibers is in contrast to the effect of increasing voltage. However, the nanofibers’ diameter is more affected by voltage variation than flow rate Based on the diagram of Fig. 3h, the interaction between the electrode distance and the relative humidity of environment has an increasing effect on the diameter of nanofibers with their increase. However, the relative humidity is fairly more effective than the distance of electrodes. The effect of increasing electrodes distance on reduction of nanofibers diameter is somehow higher than the decrease of flow rate (Fig. 3i), therefor their interaction decreases the diameter of nanofibers. It is also seen that the increase of humidity and decrease of flow rate have a decreasing effect on the diameter of nanofibers (Fig. 3j). But the effect of the humidity of laboratory space (synthesis environment) is far greater than the flow rate.

This can be implied from the inclination of the graph i.e. the humidity has a steeper slope than the other. It can be concluded that the relative humidity of the space of nanofiber synthesis laboratory is a key parameter affecting the geometrical properties of the nanofibers (more specifically their diameter) which should be controlled during nanofibers synthesis.

Optimization procedure

The effect of five process parameters namely voltage between electrodes, needle tip-electrode distance, injection rate (flow rate), concentration of synthesis solution (polymer content) and relative humidity of synthesis space (laboratory) on the diameter of titanium dioxide nanofibers were investigated, in five levels using the RSM method. These results are presented in Table 5 and Table 6.

According to Table 5, the high and low values of the effective factors on the response are in particular suffered to examine the correctness of the optimization and the success of the model. As it is seen the optimization should be achieved with a polymer content of 8.850 %, a voltage of 17.097 kV, an electrodes distance of 21.282 nm, a relative humidity of 36.879 %, and an injection rate of 0.659 mL/h in a test of 131.26 nm in diameter, which is predicted by the software. Accordingly, experiments were performed experimentally with the software prognostic standard, with a diameter of 113 nm, which is very close to the predicted value. The confidence interval and predication interval were used to prove this and the appropriateness of the experimental value. Based on the 95% confidence that the software is predicting, if the values obtained are empirically between 109.824 and 120.126 (confidence interval) and between 129.72 and 137.270 (Prediction interval), it can be claimed that the model in prediction was successful. This is quite true given the range and the experimental value obtained. Therefore, it can be safely said that a successful model, using the conditions in each experiment, can be approached similarly.

Characterization of titanium dioxide nanofibers

As stated previously, different process parameters have different impact on the diameter and morphology of the nanofibers. In this section, a statistical method is employed to predict the diameter of the fibers under different relative humidity and examine the effect of relevant parameters. It is noteworthy that relative humidity of synthesis environment by electrospinning method has not yet been considered as a key variable in such a process. In this part of the current study this parameter was thoroughly examined by considering five different levels (24, 28.5, 33, 37.5 and 42%) and the results revealed an interesting trend. The TiO₂ nanofibers were synthesized by electrospinning method with constant parameters including voltage (20 kV), polymer content (5.0 g per 5 mL absolute ethanol), titanium isopropoxide (1 mL titanium isopropoxide in 4 mL solution with 1:1 absolute ethanol and Glacial acetic acid), the flow rate (0.3 mL/h) and the distance between the electrodes (15 cm).

To remove the impurities, the synthesized nanofibers were then heated at 530 °C. This thermal treatment also created a crystal structure in the nanofibers. The morphological characteristics of the Nanofibers were then analyzed by FESEM, SEM, TEM, XRD, and FTIR.

The SEM and FESEM analyses were used to investigate the appearance of the nanofibers, including diameter and morphology. The diameter and morphology of the electrospun nanofibers are shown in Fig. 4 and Fig. 5 obtained from a sol-gel polymer solution (2 mL TIPP, 2 mL absolute ethanol and 2 mL acetic acid glacial in titanium solution and 0.3 PVP content in the polymer solution). The fibers were subjected to calcination at 530 °C using SEM and FESEM analyses. In the SEM images, four photos of the nanofibers (Fig. 4 and Fig. 5) were provided at a voltage of 20 kV, flow rate of 0.3 mL/h, electrodes distance of 20 cm, and constant concentration with a polymer content of 9% while the relative humidity varied from 24% to 42%. According to Fig. 4a, it is observed that the nanofibers with the lowest ambient humidity of 24%, have diameter of 525 nm and low morphology (exhibiting some additional outgrowth appendages on the surface). At relative humidity of 28.5%, the diameter reduced to 308 nm and their morphology was also improved. Further reductions in diameter and improvements in morphology are seen (166 nanometer at 33%, and 67 nm at 37.5%) as the ambient humidity increased. However, when ambient humidity rose to 42%, the nanofibers failed to form and hence sprayed onto the aluminum foil in form of some amorphous spots (Fig. 5). In other words, increasing humidity from a certain level prevents the formation of nanofibers and instead the titanium dioxide particles are sprayed droplet wise on the aluminum foil.

In the spectra of the calcined TiO₂ nanofibers at 500 ºC, and according to the previous studies (e.g. Li and Xia (23) and Dubey et al. (28)) at low frequencies of less than 1000 cm^-1 are reflects of the stretching vibration of Ti-O-Ti bonds (28, 29). In Fig.6, the FTIR analysis of the TiO₂ nanofibers shows two absorption peaks in the regions of 471 cm^-1 and 670 cm^-1 which is related to titanium dioxide bond. In Fig.6, in addition to the Ti-O-Ti bond peaks, the peak of O-H was also observed in the reign of 3433 cm^-1. The peak appears on 2339 cm^-1 probably indicates the potential for impurities.

X-Ray Diffraction (XRD) using CuKa radiation (λ= 1.54 A˚) at 40 kV and 30 mA at a size step of 0.040˚ and time step of 1 s (EXPLORER, GNR, Italy) was used to determine the form of amorphous or crystalline nanofibers. In determination of TiO₂ nanofiber phases, two phases of the Anatase and Rutile are important. At 200 °C, both polymer and solvents were removed from the nanofibers and the pure titanium dioxide nanofibers were obtained, possessing less diameter. Fig.7 shows the XRD analysis of the nanofibers synthesized with PVP / TiO₂ solution at ambient humidity of 37.5% and were then heated at 530 °C.

According to Fig.7a, nanofibers had a pure Anatase phase and no rutile phase was observed (30). The same analyses were performed for the other nanofibers at different levels of humidity. It should be noted that at temperatures above 600 °C the rutile phase began to form and at 800 °C the pure rutile phase was obtained which is in agreement with the previous studies (31). The peaks at 25.5º, 38.8º, 48.08º, 54.94º, 62.79º, 69.31º, 70.11º and 75.58º which are correspond to the planes of (101), (004), (200), (105), (211), (204), (116), (220) and (215) show the Anatase phase of the TiO₂ nanofibers (32). To verify the TiO₂ nanofibers, EDX analysis was also performed on the obtained structures (Fig.7 (b)). As seen in its corresponding EDX data from the area analyses of the nanofiber, the presence of Ti from this image clearly corresponds to the occurrence of pristine TiO₂ nanofibers.

Fig.8 shows the TEM images of different nanofibers formed from a solution containing 2 mL of titanium isopropoxide, 2 mL of ethanol and 2 mL of acetic acid in titanium solution and 6 mL of ethanol and 0.5 g of polyvinylpyrrolidone in the polymer solution. The nanofibers were calcinated at 500 °C and the effect of ambient relative humidity changing from 24 to 42% was investigated. It is clearly seen the effect of relative humidity on building blocks of the nanofibers. The relative humidity of 24% provided the nanofibers with fairly smooth and porous surface and a larger diameter (Fig.8a). At 28% humidity, it is seen that porosity increased, while the nanofiber diameter reduced (Fig.8b). This trend is also observed for further increase of relative humidity. The growing rate of porosity and reducing rate of diameter with increase of humidity has continued up to 39%, marking the effect of relative humidity on diameter and morphology of the nanofibers.

The nanofiber does not form higher than the aforementioned percent of relative humidity, which is vividly shown in the SEM image (Fig. 5). Therefore, it can be concluded that having very damp laboratory space, prevents the formation of fibers. Moreover, as illustrated in the TEM images, it is shown that different levels of humidity have a significant effect on the diameter and morphology of electrospun titanium dioxide.

CONCLUSION

Response surface methodology was employed to verify the effect of polymer concentration, distance of needle tip to electrodes, voltage between electrodes, flow rate and relative humidity of the synthesis process on diameter of TiO₂ nanofibers with electrospinning method. The quadratic model was chosen as the most efficient model in predicting nanofiber diameter and the polymeric content of nanofibers synthesis solution has the greatest effect on the physical structure of the nanofibers. Some interactions were found to significantly influence the response. Therefore, when synthesizing these nanofibers, the effective factors including polymer content, voltage and ambient humidity should be closely controlled. According to the results, with increasing the humidity of synthesizer atmosphere to a threshold level (39% in this study), the nanofiber diameter decreases to a minimum and its morphological characteristics are also improved (more porous fibers). However, further investigation should be carried out with varying relative humidity of the synthesizing environment from 35 to 45%, with a smaller increment, to find out a more accurate relative humidity threshold. Corresponding results of this study showed potential ability to produce various diameters for nanofibers in controlled situation.

ACKNOWLEDGMENTS

The authors acknowledge financial and laboratory supports received from Research deputy, Ferdowsi University of Mashhad, Iran. Project No: 2/46542

CONFLICT OF INTEREST

The authors declare that there are no conflicts of interest regarding the publication of this manuscript.