Inﬂuence of solvent molecular geometry on the growth of nanostructures Journal of Colloid and Interface Science

(cid:1) Inﬂuence of water and ethanol molecular geometry on the growth mechanism of ZnO nanorods. (cid:1) Role of carbon chain length as a limiting factor for ZnO monomer stacking in wurtzite crystal system. (cid:1) New theory Formulation based on solvent molecular geometry to predict the aspect ratio of nanorods. (cid:1) Nitrogen doping Inﬂuence on the synthesis and stabilization of ZnO monomer for nanorods formation. of on a incorporated ( -GO) substrate, the ﬁrst showing the inﬂuence of solvent molecular geometry on the growth mechanism of The solvents such as ( N -GO-ZnO-W) allow a large number of functional atoms along a, b and c-axis to coordinate in all possible directions with the metal ions of wurt- zite hexagonal crystal system of ZnO and thus leads to lower aspect ratio nanorods. On the contrary, the unavailability of binding sites along a-axis for solvents such as ethanol ( N -GO-ZnO-E) provides a size- limiting effect and leads to preferred growth along b and c-axis, thus generating ZnO nanorods with a higher aspect ratio. The study shows that the number of interacting atoms, carbon chain length and the solvent molecular geometry inﬂuence the aspect ratio and therefore a solvent could be used to tune the nanostructures morphology and hence the performance of devices based on them.


Introduction
Nanostructures with various materials and morphologies such as nanowires [1], nanorods [2], nanobelts [3], nanosheet [4], nanocable [5], and nanocomb [6], etc. have attracted significant attention for application in electrochemical sensing, biosensing and energy storage etc [7][8][9][10][11][12]. Morphologies such as nanorods offer higher aspect ratios (length to width ratio) and hence the large surface-to-volume ratio, which is needed for enhanced performance of devices made from them. To this end, a wide variety of physical (chemical vapor deposition, molecular beam epitaxy, pulsed laser deposition, magnetron sputtering, and thermal evaporation) and chemical (chemical bath deposition, electrochemical deposition, hydrothermal, solvothermal, sol-gel, and precipitation) methods have been explored [13][14][15][16][17][18]. These methods involve controlling the process parameters such as temperature, deposition time, stirring speed, reducing agent, vacuum condition, catalysts, concentration and solvent, etc. to obtain nanostructures with desired aspect ratio. Additionally, the influence of ambient conditions such as solvent properties such as surface tension, dielectric constant, pH, and viscosity etc. has been studied for the growth mechanism of nanostructures [18,19]. Interestingly, none of these extensive nanoparticles (NPs)-based studies conducted over more than 150 years has considered the influence of solvent molecular geometry on the growth of nanostructures and their morphology. Herein, with ethanol and water solvents we demonstrate for the first time the influence of solvent molecular geometry on the aspect ratio of nanostructures such as ZnO nanorods. The ethanol (N-GO-ZnO-E) and water (N-GO-ZnO-W) are selected here due to the availability of different number of interacting atoms and molecular geometries. The ZnO used here presents an ideal testbed to gain insight into nanoparticles growth mechanism, as it carries 3d transition metal ions and offers a rich family of structures such as whiskers, wires, rods, tubes, belts, cages, rings, combs, prisms, etc. [20][21][22][23][24][25]. The ZnO nanorods were synthesized on a highly conductive nitrogen incorporated graphene oxide (N-GO) substrate (shown in Fig. 1), in water and ethanol solvents. The synthesis steps include the dual interaction of 1,4-Phenylenedimethanamine carrying two NH 2 groups on the opposite position of a benzyl ring. NH 2 groups are expected to covalently react with the hydroxyl, epoxy, and carboxyl groups of GO in a series of reduction and condensation reactions to produce pyridinic, pyrrolic, graphitic-N and pyridinic-NO species [26][27][28]. These nitrogen configurations have already been shown to alter the GO chemistry by the redistribution of electronic charges in the p electronic system of GO matrix [29,30]. Nitrogen doping intrinsically modifies the material chemistry by the generation of electrophilic and nucleophilic centers in the vicinity of heteroatom [27], which, consequently, plays a vital role in the concentration and growth kinetics of NPs [31]. The synthesized NPs were extensively studied through X-ray photoelectron spectroscopy (XPS), high resolution transmission electron microscopy (HR-TEM), X-ray diffraction (XRD), energy dispersive X-ray spectroscopy (EDX) and EDX mapping.  Reference Water Purification System. All the reagents used in this experiment were of analytical grade and were used without further purification.

Instrumentation
The structural characterization of the nanomaterials were carried out using X-ray photoelectron spectroscopy (XPS) on a Multi-Lab 2000 spectrometer (Thermo VG Scientific, Southend-On-Sea, Essex, UK) in an ultra-high vacuum chamber along with X-ray diffraction (XRD) spectrum carried out on a Rigaku D/max-2500, using filtered Cu Ka radiation at Center for Research Facilities (CCRF) of Chonnam National University (CNU). The highresolution transmission electron microscopy (HR-TEM) and energy dispersive X-ray spectroscopy (EDX) were studied using a JEM-2100F microscope at 200 kV in the Korea Basic Science Institute (KBSI) of CNU.

Synthesis of nanomaterials
GO was synthesized from natural graphite powder (~325 mesh, 99.999%) according to the improved Hummer's method [38]. The synthesis of N-GO-ZnO-E, N-GO-ZnO-W, and GO-ZnO-E nanomaterial is shown in Fig. 1. Initially, GO was dispersed in double distilled water (1 mgmL À1 ) and sonicated for 1 h. The resulting homogeneous suspension was mixed with an aqueous solution of 1,4-Phenylenedimethanamine (1 mgmL À1 ) in a round-bottom flask equipped with a magnetic stirrer bar. After 2 h of continuous stirring, a solution of ZnCl 2 in ethanol was added to the homogeneous solution and kept for 30-minute stirring. Thereafter, 100 mL of N 2 H 4 ÁH 2 O was added as a reducing agent followed by stirring and heating at 110°C for the next 10 h. The black solution obtained was filtered three times with double distilled water and was dried under vacuum at 50°C for overnight. The N-GO-ZnO-W was prepared in the same protocol, except ZnCl 2 solution was prepared in water rather than in ethanol. GO-ZnO-E was also prepared in the same protocol except with the addition of 1,4-Phenylenedimethanamine.

Structural analysis of N-GO-ZnO-W and N-GO-ZnO-E
The crystallinity of the N-GO-ZnO-W and N-GO-ZnO-E were evaluated from the XRD spectra. As can be seen in Fig. 2 [8,32]. The enhanced diffraction peaks of (1 0 0), (0 0 2), (1 0 1) and (1 0 3) crystal faces indicate a polycrystalline behavior, accommodated with preferential ZnO crystal orientation along the c-axis which provides hexagonal rod-like morphology. The high diffraction intensity of (0 0 2) in N-GO-ZnO-E represents growth of ZnO nanorods with a high aspect ratio which was also observed from HRTEM study. Further, the absence of any impure peak suggested the synthesis of pure hexagonal structured ZnO in both nanomaterials. Besides, a GO reduction reflection at 26.1°and honeycomb structure formation by the sp 2 hybridized carbon reflection at 42.9°with a 'd' spacing of 0.34 and 0.21 nm respectively [33] confirmed the XPS results regarding the successful integration of 1,4 Phenylenedimethanamine into GO network.
In order to better understand the obtained oxidation states and peak intensities of N-GO-ZnO-W and N-GO-ZnO-E nanomaterials by XPS, it is important to understand the expected chemical reactions between 1,4-Phenylenedimethanamine and GO. Following nitrogen configurations are expected to obtain from the reaction between GO oxygenated functionalities and NH 2 group of 1,4-Phenylenedimethanamine.
(i) Reduction of epoxy and OH group of GO by NH 2 to produce graphitic-N and pyrrolic-N respectively. (ii) A reaction between the -COOH and NH 2 to produce pyridinic-N.  [34,35]. Zn and O1s peaks of N-GO-ZnO-E exhibited comparatively stronger intensity then N-GO-ZnO-W, thus; suggesting the presence of a higher proportion of Zn and O atom in its electronic environment. The survey spectra of both nanomaterials were deconvoluted to figure out the mechanism of ZnO nanorods morphology and growth kinetics by understanding the individual oxidation state of each atom present in the system.  [38,39]. The higher intensity of graphitic-N and lower intensity of pyrrolic-N (Fig. 3c) is justifying the low intensity of CAO and high intensity of OH (Fig. 3d), and vice versa for the lower graphitic and higher  [32,35,40]. The difference in individual binding energies of both nanomaterials is due to the difference in charge transfer between Zn 2+ to O 2À , which are surrounded in a different chemical environment. Further, the spin-orbit splitting of 23.07 (Fig. 3e) and 23.06 eV (Fig. 3e 0 ) between Zn2p 3/2 and Zn2p 1/2 of N-GO-ZnO-E and N-GO-ZnO-W respectively stamped the presence of pure +2 oxidation state [40]. (N-GO-ZnO-W)). The only difference between two nanomaterials is the solvent, in which zinc precursor was dissolved and added. There are several important parameters in a solvent such as, dielectric constant, viscosity, surface tension, and pH, which play a crucial role in governing the production, dispersion, and morphology of a NP. The mechanism of ZnO nanorods formation in both nanomaterials (N-GO-ZnO-W and N-GO-ZnO-E) is divided into following four steps:

(I) monomer (ZnO) formation
In case of N-GO-ZnO-E, reaction between ZnCl 2 and C 2 H 5 OH resulted in the formation of Zn(ethoxide) 2 from the conjugate base of ethanol (AOC 2 H 5 ). The alkoxide group made metal alkoxides very susceptible to hydrolysis, condensation, and nucleophilic reactions which usually follow alcoxolation and/or oxolation reaction to produce metal oxide [41]. The thermodynamics of these reactions depends on the strength of the entering nucleophile, the electrophilicity of metal, and on the partial charge and stability of the leaving group [42]. Herein, for coordinately saturated Zn (OC 2 H 5 ) 2 in the absence of a catalyst, hydrolysis has occurred by the attack of a nucleophile, resulting in the formation of Zn(OH) 2 , which followed either alcoxolation or oxolation to initiate the formation of ZnO monomer by the removal of a proton. Zn On the other hand, in case of N-GO-ZnO-W, an aqua complex of Zn (hydroxy)chloride was formed by the reaction of ZnCl 2 and H 2 O. The aqua complex was later on hydrolyzed by reacting with second H 2 O molecule resulting in the formation of Zn(OH) 2 . Finally, the Zn (OH) 2 underwent either oxolation or simple condensation due to heating to produce ZnO [43].

(II) nucleation of ZnO
After the formation of initial ZnO monomer, the nucleation threshold started to build-up depending on the rate of supersaturation, which in turn is dependent on the solubility of the ZnO in the solvent. The solubility of any monomer in a chemical system is directly proportional to the dielectric constant of the solvent [44]. As we know, the dielectric constant of water (e = 80) is higher than ethanol (e = 25); therefore higher ZnO monomer solubility in N-GO-ZnO-W nanomaterial has led to its supersaturation reduction, and subsequent nucleation rate. Conversely, the decrease in surface energy due to a better ethanol-ZnO interaction, the relatively higher ZnO stability in N-GO-ZnO-E resulted in a massive supersaturation and increased nucleation rate. Herein, it is important to note that the monomer formation in both cases was strongly enhanced by the presence of different nitrogen configurations. This has altered the electronic environment of GO, enhanced the system reactivity which led to the generation of a higher number of monomer, accelerated nucleation rate, and finally produced smaller particles with narrower size distribution [30]. However, the eventual survival battle of a monomer is with the dielectric constant of its solvent which determines the final number of nuclei for growth. The mechanism of nucleation in both nanomaterials is shown in Fig. 5a.

(III) ZnO nanorods formation
After sufficient nucleation, ZnO nuclei undergo agglomeration to form particles. These ZnO NPs get crystallize into hexagonal discs either by NPs diffusion, by the surface reaction or by both mechanisms. The NPs self-assembly depends on the inter-particle distance, particle interaction with solvent, and chemically or physically adsorbed organic/inorganic ions. We believe that the presence of different nitrogen configurations in both nanomaterials is crucial for a surface-integration controlled growth and stabilization of the nanocrystal. A 2 nm HRTEM images of N-GO-ZnO-E (Fig. 4h), and N-GO-ZnO-W (Fig. 4h 0 ) showed a lattice line of 0.12, 0.13, 0.14, 0.19, 0.25, 0.28 nm, depicting a Wurtzite-type structure of ZnO nanorods [45], which also confirmed the XRD findings (Fig. 2). Wurtzite crystal structure present hexagonal polar repeated units due to the difference in surface energy of its crystal faces [46]. The top plane of the crystal structure is positively terminated Zn 2+ (0 0 0 1) while the bottom is surrounded by negatively terminated O 2À (0 0 0 1 À ) atoms. These repeated polar units are perpendicular to c-axis (0 0 0 2), which is a layer of side atoms to enclose the crystal. The sequential layering of Zn 2+ and O 2À atoms along the c axis gives rise to intrinsic polarity. The difference in surface energy between (0 0 0 1) and side surfaces favours anisotropic crystal growth along c-axis, thus; resulting in a rod-like morphology. The inherited instability of the (0 0 0 1) and (0 0 0 1 À ) faces requires additional respective complementary charges to stabilize the system [47]. The difference in chemisorption of different molecules, atoms onto ZnO, directs NP morphology differently which subsequently defines chemical properties of the nanomaterial. Similarly, herein, the interaction of different nitrogen configurations (pyrrolic, pyridinic, graphitic-N, and pyridinic-N + O À ) and solvent (ethanol and water) with ZnO will determine the Debye length, surface potential barrier height, surface charge and its layer thickness.
(i) Following types of interactions are possible between pyrolic, pyridinic, graphitic-N, pyridinic-N + O À , ZnO, and ethanol/ water (Fig. 5b). The XPS analysis of N-GO-ZnO-W (Fig. 3c 0 ) and N-GO-ZnO-E (Fig. 3c) has revealed the presence of pyridinic, pyrrolic, graphitic-N along with an additional pyridinic-N + O À peak in latter. The lone pair of Pyridinic-N has interacted with O of ZnO, which resulted in the generation of pyridinic-N + O À in N-GO-ZnO-E, but in case of N-GO-ZnO-W, the higher polarity of water has increased the shielding effect of pyridinic-N [48]; thus, justifying the absence of pyridinic-N + O À peak in its XPS spectra. (ii) Pyrrolic-N is a H bond donor due to the availability of acidic-H. The electronegative O atoms of ZnO, and C 2 H 5 -OAH/HAOAH are H-bond acceptor for the acidic-H of Pyrrolic-N. (iii) Graphitic-N is a H bond acceptor due to the absence of acidic-H and will accept a H from Pyrrolic-N, and HAOAH/HAOC 2 H 5 . (iv) Pyridinic-N has a lone pair so it will be a H-bond acceptor for the HAOC 2 H 5 and pyrrolic-N. (v) The O À of pyridinic-N + O À will accept a H-bond from H-OC 2 H 5 and pyrrolic-N. (vi) The inter-molecular H-bonding between HAOAH and H-OC 2 H 5 is likely to occur in N-GO-ZnO-W and N-GO-ZnO-E respectively.
As Zn and O atoms of ZnO are forming consecutive layered structures [48], therefore these interactions among a growing crystal surface will act as a substrate for the next layer of ZnO. The extent of interactions along c-axis will determine the length of the ZnO nanorods, while the interactions along m-plane will define the width of ZnO nanorods which we believe are dependent on the geometry of the solvent molecules (ethanol, water). If we have a look at the structure of water and ethanol, H 2 O has a bent structure and 3 atoms to propagate the surface interactions in three directions. Both H and O atom of H 2 O will propagate interactions with ZnO either along caxis or m-plane. Contrarily, in ethanol, there are only two interacting atoms; a single O and a single H which will further the interactions either along c-axis or m-plane. m-plane interactions on another side of H-OC 2 H 5 are stopped due to the attachment of C 2 H 5 at the end of the molecule. C 2 H 5 is acting as a dead-end to advance the interactions along m-plane which resulted in the formation of ZnO nanorods with a high aspect ratio (high length and short width) in N-GO-ZnO-E. H 2 O has more atoms to propagate interaction in mplane than H-OC 2 H 5 , therefore, ZnO nanorods with a low aspect ratio (short height and high width) were obtained in N-GO-ZnO-W. Besides the absence of pyridinic-N interactions with ZnO and HAOAH (absence of pyridinic-N + O À shown by XPS) has also significantly reduced the ZnO stacking along c-axis.

(IV) final dispersion of ZnO nanorods
ZnO nanorods in both nanomaterials showed severe aggregation due to high surface energy of individual nanorods. The growth mechanism of ZnO nanorods in N-GO-ZnO-W and N-GO-ZnO-E is presented in Fig. 5b. To confirm the influence of nitrogen doping in the production and stabilization of initial ZnO monomers, ZnO nanorods were synthesized exclusively on the GO surface (Fig. 4a 00 ). Fig. 4(b 00 -d 00 ), shows a decent population of intermingled ZnO NPs in the form of sharp, pointy, and slightly bent nanorods onto GO sheets. The width of the nanorods was very less in comparison to N-GO-ZnO-E and N-GO-ZnO-W nanomaterials. This suggests that after initial adsorption of ZnO NPs on electrophilic centers of GO, ethanol solvent facilitated ZnO NPs stacking along c-axis. The adsorption of ZnO NPs on neighboring substrate surfaces was low due to the absence of ample electrophilic centers. Thus; the width of the nanorods was not only hampered by the structural configuration of ethanol but also due to the absence of nitrogen configurations (Fig. 4c 00 -e 00 ). Further, the lattice line of N-GO-ZnO-W (Fig. 4h 0 ), N-GO-ZnO-E (Fig. 4h), GO-ZnO-E (Fig. 4h 00 ) showed a very interesting behavior by assembling ZnO NPs in an oriented (OA) and non-oriented attachment (NOA) fashion for the latter two nanomaterials respectively, which differs in terms of the orientation of their crystal lattice at the grain boundary [8]. N-GO-ZnO-E and GO-ZnO-E showed no specific preference in attachment, while in N-GO-ZnO-W, a particular crystallographic alignment has occurred by rotation until the perfect crystal phase of two NPs or a twin matched in a minimum energy configuration to allow for continuous crystallographic planes. The high dielectric constant of H 2 O has induced strong attractive forces in interatomic interactions of the attaching NPs and other attaching surfaces. This attractive force is the actual driving forces to overcome the energy barrier for oriented particle-particle contact [8]. The successful synthesis of N-GO-ZnO-E, N-GO-ZnO-W, and GO-ZnO-E was ensured from elemental mapping as shown in Fig. 4i, i 0 and i 00 respectively.
This detailed spectroscopic analysis shows that ZnO nanorods obtained in the presence of ethanol showed a higher aspect ratio, while it was lower in the presence of water. Besides, the population of ZnO nanorods in both cases was attributed to the dielectric constant of water and ethanol as explained above in detail and shown in Fig. 5a. The obtained results conclude that the number of functional atoms in water are present in three dimensions (a, b and caxis) and therefore they can coordinate with the metal ions in all possible directions of wurtzite hexagonal crystal system of ZnO, leading to lower aspect ratio nanorods. However, in the case of ethanol, the unavailability of binding sites along a-axis provides a size-limiting effect for the ZnO NPs stacking along a-axis, thus leading to preferred growth along b and c-axis and eventually generating ZnO nanorods with a higher aspect ratio as shown in Fig. 5b. By following this theory, Fig. 6 has been made from the ZnO nanorods reported in the literature with methanol, ethanol, propanol, isopropanol, butanol, hexanol, and water to evaluate the applicability of presented theory with other solvent systems [49][50][51][52][53][54][55][56][57][58][59][60][61]. This analysis suggests the successful validation of the proposed theory. For example, the ZnO nanorods reported by Samulski et al. with methanol have a higher aspect ratio in comparison to the one obtained in our study using water [51]. The unavailability of binding site along a-axis in methanol result into nanorods with reduced width. In the case of water, the availability of binding site along 3-dimensions result in ZnO nanorods with a comparatively lower aspect ratio. Similar pattern can be noted from other reports as well. For example, the aspect ratio of ZnO nanorods linearly increase with the increase in carbon chain from methanol to hexanol. The reason behind this behavior is the steric hindrance caused by long-chain carbon atoms which are inactive and prevent further stacking of ZnO NPs in the direction of the carbon chain, thus escalating ZnO nanorod growth only along b and c-axis (as shown in Fig. 6) and eventually leading to a high aspect ratio ZnO nanorods. Even, nanorods obtained with hexanol showed a needle-like structure [58]. These observations reflect that the optimum conditions to obtain the desired aspect ratio of nanorods is not only a perfect selection of solvent molecular geometry but also the length of the carbon chain and steric hindrance is of equal and utmost importance.

Conclusion
The study concludes an interesting new finding regarding the influence of solvent (water and ethanol) molecular geometry on the controlled growth of ZnO nanorods. Herein we provided a detailed discussion about a series of individual factors such as surface tension, dielectric constant, pH, and viscosity of a solvent that are involved right from the synthesis of initial monomer to the final growth into a particular morphology. Further, the paper describes how the geometry of water and ethanol determines the aspect ratio of the ZnO nanorods and how replacing ethanol with water (ZnO precursor dissolving solvent) resulted in a steep decrease in ZnO nanorods population by~80%. The possible synthesis and dispersion route of ZnO nanorods was explored on a GO and a nitrogen-doped GO surface to understand the involvement of nitrogen in the production of the ZnO monomer. We unveiled a new way to predict and design the morphology of nanostructures by selecting a solvent of a particular molecular geometry. The study also reveals that the optimum conditions to obtain the desired aspect ratio of nanorods is not only about a perfect selection of solvent molecular geometry but also the length of the carbon chain and steric hindrance is of equal and utmost importance. The proposed growth theory was also validated with previously reported ZnO nanorods systems and showed a satisfactory validation which presents a substantial improvement in terms of predicting and designing different NP morphologies simply by selecting a solvent with a specific molecular geometry. In this regard, the present study could be further extended in the future for other metal oxide nanorods and other morphologies such as nanocubes, nanospindle, nanocones, nanocombs, and nanoflowers.

Declaration of Competing Interest
There are no conflicts to declare.