- Email: [email protected]
Water-soluble polyethylenimine as an efficient dispersant for gallium zinc oxide nanopowder in organic-based suspensions
Powder Technology 305 (2017) 226–231
Contents lists available at ScienceDirect
Powder Technology journal homepage: www.elsevier.com/locate/powtec
Water-soluble polyethylenimine as an efficient dispersant for gallium zinc oxide nanopowder in organic-based suspensions Chia-Chen Li a,⁎, Jia-Hao Jhang a, Hsin-Yi Tsai a, Yung-Pin Huang b a b
Institute of Materials Science and Engineering, Department of Materials & Mineral Resources Engineering, National Taipei University of Technology, Taipei 10608, Taiwan Material and Chemical Research Laboratories, Industrial Technology Research Institute, Hsinchu 30011, Taiwan
a r t i c l e
i n f o
Article history: Received 29 June 2016 Received in revised form 29 September 2016 Accepted 1 October 2016 Available online 4 October 2016 Keywords: Gallium zinc oxide Transparent conducting oxide Dispersion Polyethylenimine Tris(2-butoxyethyl) phosphate
a b s t r a c t Appropriate dispersants for the dispersion of gallium zinc oxide (GZO) nanopowder in the commonly used organic medium-dimethylacetamide (DMAC) are proposed by this study. The dispersion efficiencies and stabilization mechanisms of three dispersants—the organic-based dispersant tris(2-butoxyethyl) phosphate (OP3) and two water-based dispersant polyethylenimines (PEIs) with different molecular weights—are compared. Surprisingly, the water-based PEI polyelectrolytes exhibit greater efficiency than the organic-based OP3 for the dispersion of GZO nanopowder in the organic-based suspensions. This is because that the stabilization mechanism of GZO in the very polar DMAC is primarily related to steric repulsion rather than to electrostatic repulsion according to the theoretical and experimental analyses. © 2016 Published by Elsevier B.V.
1. Introduction Zinc oxide (ZnO)-based powders have attracted considerable interest as transparent conducting oxide (TCO) materials in recent years because of their competitive advantages of high natural abundance, low cost, nontoxicity, low resistance, high transparency, and good chemical and thermal stabilities [1–10]. The traditional TCO is indium tin oxide (ITO), which is a very expensive material that is unstable in hydrogen plasma [2,9,10]; thus, numerous doped ZnO compounds, including In2O3-doped ZnO, Al2O3-doped ZnO (AZO), and Ga2O3-doped ZnO (gallium zinc oxide, GZO), have been proposed as replacements for ITO [1– 10]. Among the doped ZnO powders, AZO is the most common; however, GZO is the most chemically stable because Ga is less reactive than Al toward environmental oxygen and because the ZnO lattice is less distorted by the substitution of Zn2+ with Ga3+ than by the substitution of Zn2+ with Al3+ [11–13]. TCO materials are generally and extensively manufactured as films for use in optical and electrical devices [2,14,15] such as solar cells, flat panels, liquid crystal displays, transistors, light-emitting diodes, and window coatings. Several approaches have been used for preparing TCO films [1,2,16–20], including various physical vapor deposition, sputtering, chemical vapor deposition, epitaxy, ion plating, spray pyrolysis, and solution growth processes. Among these techniques, the method of sol-gel spin coating, which is a solution growth process, has ⁎ Corresponding author. E-mail address: [email protected] (C.-C. Li).
http://dx.doi.org/10.1016/j.powtec.2016.10.006 0032-5910/© 2016 Published by Elsevier B.V.
attracted the interest of numerous research groups [21–24]. Interest in sol-gel spin coating stems from its distinct advantages of cost effectiveness, simplicity, and suitability for mass production because of its high deposition rate. Nevertheless, disadvantages and drawbacks of this method include the challenging conditions needed to control the homogeneity of the sol-gel reaction and the difficulty associated with simultaneously optimizing a chemical reaction and achieving a uniform coating. A new approach using a well-dispersed suspension has been proposed as a simple alternative method of depositing TCO films; this method also features the advantages of low cost and suitability for mass production [10]. Although the preparation of uniform films through the use of welldispersed powder suspensions is advantageous, most commercially available doped ZnO powders such as AZO and GZO are very fine, with particle sizes of b50 nm. Such fine particles undergo severe aggregation in suspension and are not easily deagglomerated and dispersed. The demand for doped ZnO powders has increased for a wide range of products, including paints and coatings that provide heat-insulating or antistatic functions [25] and various optical and electrical devices; consequently, techniques for preparing stable suspensions of these powders are reasonably expected to increase in importance and to attract increased attention in the near future. Thus far, few reports regarding the dispersion of these powders or the development of appropriate dispersants have been published. In this work, appropriate dispersants for the dispersion of doped ZnO in organic-based suspensions are proposed and explored. GZO powder and dimethylacetamide (DMAC), a common solvent in
C.-C. Li et al. / Powder Technology 305 (2017) 226–231
227
industrial processes, were adopted as a model for this investigation. The purpose is to identify an effective dispersant for GZO powder and to achieve a stable dispersion of GZO in an organic suspension with a high solid content. Moreover, the dominant stabilization mechanism for the dispersion of GZO in suspensions is discussed and clarified. 2. Materials and methods High-purity GZO nanopowder (54W-GZ10, ITRI, Taiwan) with a hexagonal wurtzite crystal structure was used. The content of gallium oxide (Ga2O3) in GZO powder was 1.3 wt%, as determined by inductively coupled plasma atomic emission spectroscopy (ICAP-9000, JarrellAsh, USA). The primary particle size of GZO was approximately 20 nm, as characterized by transmission electron microscopy (JEM-2100, JEOL, USA). The surface area of GZO was 47 m2/g. The dispersants used included tris(2-butoxyethyl) phosphate (OP3, 95%; TCI, Tokyo, Japan) and two types of polymeric polyethylenimines (PEIs) with molecular weights of 1800 g/mol (L-PEI, 99%; Alfa Aesar, USA) and 10,000 g/mol (H-PEI, 99%; Alfa Aesar, USA). The polar organic solvent dimethylacetamide (DMAC, 99%; J.T. Baker, USA) was used as the dispersion medium. The electrokinetics of GZO nanopowder was analyzed using the electroacoustic method (ZetaProbe, Colloidal Dynamics Inc., North Attleborough, MA, USA). For the zeta-potential measurements, GZO suspensions in DMAC with a solid content of 0.5 wt% were prepared with the addition of 5 wt% of different dispersants. The suspensions were mixed and deagglomerated by a ball mill (MUBM236, Yeong-Shin, Taiwan) at 300 rpm with the addition of 50% mill charge of 5-mm yttria-stabilized zirconia media for 24 h at room temperature. In the adsorption experiments, the suspensions were prepared with a solid content of 20 wt% (based on the weight of solvent) in the presence of various concentrations of dispersants (based on the weight of GZO powder). The powder suspensions were deagglomerated by ball milling for 24 h and centrifuged at 5500 rpm to separate GZO nanopowder from the supernatant. The separated nanopowder was dried in an oven at 80 °C, and the amount of dispersant adsorbed onto the nanopowder was determined by thermogravimetric analysis (Q50, TA Instruments Ltd., Crawley, UK). The rheology of GZO suspensions with a solid content of 20 wt% and with OP3, L-PEI, and H-PEI added at various concentrations was analyzed by using a concentric cylinder rheometer (AR1000, TA Instruments Ltd., UK). In addition, the thixotropic behavior was analyzed based on the suspensions with solid contents of 10–40 wt% in the presence of 10 wt% of L-PEI or H-PEI. Cone-plate geometry fixtures with a diameter of 20 mm and cone angle of 1° were chosen for the steady-shear rotation tests. All the rheological curves were fitted using the software (V 5.2.26, Rheology Advantage™, UK) attached to the rheometer. The reliability of the equation was evaluated by the coefficient of determination (R2). The R2 obtained from the fittings of all samples are within 0.95–1.0. 3. Results and discussion For the dispersion of nanopowders, the investigated dispersant candidates are generally small molecules or polymers with low molecular weights because of their reasonable size compared to the small nanoparticles. Three dispersants—OP3 and two types of PEI—were chosen, and their dispersion efficiencies for GZO nanopowder in the polar organic solvent—DMAC were explored. The chemical structures of the investigated dispersants are shown in Fig. 1. OP3 is a small molecular surfactant with a phosphate group, and L-PEI and H-PEI are both polyethylenimines (PEIs) with molecular weights of 1800 g/mol and 10,000 g/mol, respectively. The solubility parameters (δ) of the dispersants and solvent are 23.5 MPa1/2 for PEI [26], 20.5 MPa1/2 for OP3 [27], and 22.7 MPa1/2 for DMAC [28]. The δ values of OP3 and PEI are very close to that of DMAC, indicating that DMAC is good solvent for dissolving OP3 and PEI. On the other hand, the logarithm of acid dissociation constant (pKa) of the suspension components was also
Fig. 1. Chemical structures of (a) tris(2-butoxyethyl) phosphate (OP3) and (b) polyethylenimine (PEI).
compared. Although the pKa of GZO has not been reported in the literature, the content of Ga in GZO is extremely low and Ga exhibits a higher cationic potential than does Zn, as a consequence of its smaller ionic radius and higher charge valence [29], we assumed that GZO should be more acidic than ZnO and the pKa of GZO is slightly lower than that of ZnO. The pKa of ZnO has been reported to be 9.3 [30]. From the electrokinetic measurements, it is found that GZO exhibits a negative zeta potential at − 1.5 mV in DMAC. This indicates that DMAC should be like an alkaline medium for the dispersion of GZO, although the exact pKa is not clear due to its very different reports in the literature, ranging from −0.5 to 30 [31–33]. Since DMAC is a high polar solvent, the dispersed GZO is possible to carry charge via the ionization mechanism based on the electron affinity differences between itself and the dispersion medium, parallel to the case occurred in an aqueous suspension [34]. When H-PEI or L-PEI was added, the GZO powder changed to exhibit a positive charge in DMAC. The pKa of PEI has been reported to be ~11 [35,36]. The possibility for the varied zeta potential of GZO is that both PEIs are positively charged and have adsorbed on GZO. The low positive value of zeta (ζ) potentials due to the addition of PEI may be caused by two reasons. One is that PEI does not have a good dissociation in DMAC; the other is the adsorbed amount of PEI on GZO is low. And the combined reason is also probable. The dispersant OP3 has a reported pKa ranged from − 3 to − 9 [37], lower than any possible pKa value reported for DMAC. Therefore, OP3 would preferentially dissociate to carry a negative charge in DMAC. As shown in Table 1, the resulting zeta potential of GZO with the addition of OP3 is −4.0 mV, just evidencing the specific adsorption of OP3 onto GZO. The adsorption isotherms of OP3, L-PEI, and H-PEI on GZO nanopowder are compared in Fig. 2(a). Their significant adsorptions correspond well to the behaviors predicted on the basis of the zeta-potential measurements. The adsorption amounts of the three dispersants increased with increasing dispersant concentration and finally reached saturation, demonstrating typical chemisorption behavior [38]. The adsorption saturations for OP3, L-PEI, and H-PEI were approximately 20, 74, and 112 mg/g GZO, respectively. We further analyzed the exact adsorption behavior of the dispersants using the Langmuir monolayer adsorption equation (Eq. (1)) [39]:
Ce Ce K ¼ þ ; As Cm Cm
ð1Þ
Table 1 Zeta potentials (ζ) and adsorption layer thicknesses (t) of GZO nanopowders in DMAC upon the addition of various dispersants. Dispersant
ζ (mV)
t (nm)
None OP3 L-PEI H-PEI
−1.5 −4.0 9.3 15.0
– 3.1 3.1 5.1
228
C.-C. Li et al. / Powder Technology 305 (2017) 226–231
Fig. 3. Apparent viscosity of 20 wt% GZO suspensions obtained at a shear rate of 300 s−1 plotted as a function of dispersant concentration. The inset shows the static rheological curves of 20 wt% GZO suspensions as functions of various concentrations of H-PEI.
Fig. 2. (a) Adsorption amounts of three dispersants on GZO as functions of the equilibrium concentration of dispersants. (b) Replot of the data in (a) on the basis of the Langmuir equation. Note that the solid lines are the fitting results.
where Ce is the equilibrium concentration of the dispersant in suspension, As is the adsorbance of the dispersant, Cm is the saturated/monolayer adsorbance of the dispersant, and K is a constant. When Ce/As was plotted as a function of Ce, straight lines were obtained for these dispersants, as shown in Fig. 2(b). The results in Fig. 2(b) indicate that all isotherms in Fig. 2(a) correspond to Langmuir-type adsorption (i.e., the dispersants are adsorbed chemically). Generally, a powder can be better stabilized if more dispersant molecules adsorb onto the particles because greater amounts of adsorbed molecules might result in a thicker adsorption layer on the particles, which, in turn, results in a greater repulsive energy among the particles, thereby preventing agglomeration. Hence, the fact that H-PEI exhibited the largest adsorption amount on GZO in Fig. 2(a) suggests that it might be the most efficient dispersant. To determine the relative dispersion efficiencies of dispersants, we compared the rheologies of powder suspensions with a solid content of 20 wt% stabilized by the addition of different dispersants at various concentrations. The inset of Fig. 3 reveals the viscosity of suspensions upon the addition of H-PEI. When the concentration of added H-PEI was b5 wt%, the rheological curve showed that viscosity decreased with increasing shear rate, demonstrating typical pseudoplastic flow behavior. That is, the suspension still contained powder agglomerates when the added concentration of H-PEI ([H-PEI]) was too low. When the [H-PEI] was increased to 5 wt%, a substantial change in the
rheological behavior of the suspensions was observed. The viscosities of the suspensions remained constant over the whole investigated range of shear rates, demonstrating typical Newtonian flow behavior (i.e., no powdered agglomerates existed in the suspensions). The viscosities of suspensions obtained at a shear rate of 300 s−1 as a function of [H-PEI] are replotted in Fig. 3; the viscosities that resulted from the additions of OP3 and L-PEI are compared in the figure. When L-PEI or H-PEI was used as the dispersant, the lowest concentration at which a stable viscosity was obtained was 5 wt%, whereas it was 10 wt% when the dispersant was OP3. In addition, the lowest stable viscosity of suspensions to which OP3 was added was higher than that of suspensions to which L-PEI or H-PEI was added. Obviously, PEI is a better dispersant than OP3, and L-PEI and H-PEI exhibit good efficiency for the dispersion of GZO in DMAC. To further compare the effect of the molecular weight of PEI on its ability to stabilize GZO in suspensions, the non-equilibrium shear stress obtained upon increasing and decreasing the shear rate for suspensions with solid contents of 10, 20, 30, and 40 wt% in DMAC was analyzed. Fig. 4 shows the resulting thixotropic hysteresis loops for these suspensions with 10 wt% added PEI. By fitting the increasing section of the thixotropic curve with the Herschel–Bulkley equation (Eq. (2)) [40–42], we
Fig. 4. Thixotropic curves of GZO suspensions with solid contents of (a) 10, (b) 20, (c) 30, and (d) 40 wt% in the presence of 10 wt% of L-PEI and H-PEI, respectively. The inset shows the viscosity obtained from the increscent section at 300 s−1 as a function of the solid content of suspensions.
C.-C. Li et al. / Powder Technology 305 (2017) 226–231
obtained detailed information pertaining to the flow behavior of suspensions: σ ¼ σ y þ ηγn ;
3 2 2 h −4a2 A123 4 2a2 2a2 5; þ 2 þ ln VA ¼ − 2 2 6 h h h −4a2
terms, sterically repulsive energy (VRs) and electrical double layer repulsive energy (VRe), are generally included, in which (VRe) is given by
ð2Þ
where σ is the shear stress, σy is the yield stress that can break apart the agglomerates in suspensions, η is the viscosity, γ is the shear rate, and n is the flow index. The flow will be Newtonian if n = 1 and shear thinning if n b 1 [39–41]. For the suspensions with added H-PEI at solid contents of 10–40 wt%, the n values are within 0.98–1.00, indicating that these suspensions exhibit well-dispersed Newtonian flow behavior. Note that the good dispersion at 40 wt%, which is an extremely high solid content for a suspension loaded with a nanosized powder, affirms that H-PEI is a highly efficient dispersant for nano-GZO in the polar organic solvent. Upon the addition of L-PEI, suspensions with various solid contents exhibited different values of n. At the lower solid content of 10 wt%, the suspension exhibited good dispersion, with n = 0.91. At solid contents of 20–30 wt%, suspensions became shear thinning, with 0.81 ≤ n ≤ 0.83, demonstrating that soft agglomerates formed in the suspensions. At the higher solid content of 40 wt%, the suspension showed more significant shear thinning, with n = 0.73, indicating more severe agglomeration of the powder. In addition, the calculated σy for suspensions with added H-PEI are all small, showing an insignificant increase from 0.09 Pa at 10 wt% to 0.12 Pa at 20 wt%, to 0.15 Pa at 30 wt%, and to 0.16 Pa at 40 wt%. For suspensions containing added L-PEI, σy increased significantly with increasing solid content, from 0.11 Pa at 10 wt% to 0.48 Pa at 20 wt%, to 0.52 Pa at 30 wt%, and to 0.63 Pa at 40 wt%. A greater σy value indicates a greater degree of powder agglomeration; therefore, the σy data show good agreement with the results for n. Moreover, the suspensions with H-PEI added at various concentrations show insignificant hysteresis loops (i.e., the viscosity is time-independently stable). In contrast, when L-PEI was added at a solid content of 40 wt%, the suspension exhibited a significant hysteresis loop. The inset of Fig. 4 compares the viscosity of suspensions with the presence of L-PEI and H-PEI, which is obtained at a shear rate of 300 s−1. Obviously, at any solid content of GZO, the addition with H-PEI results in a lower viscosity. Besides, the increase in the viscosity with increasing solid content for the suspension with the addition of H-PEI is not as drastic as that with L-PEI addition. According to the aforementioned comparisons, HPEI is significantly a more efficient dispersant than L-PEI although the latter may have better solubility in DMAC due to lower molecular weight. We further clarified the dispersion mechanisms of PEI and OP3 by using the extended Derjaguin–Landau–Verwey–Overbeek (DLVO) theory to calculate the stabilization force on the basis of their adsorptions [38,39,43]. The attractive potential between particles was assumed to be primarily due to van der Waals force (VA):
229
V Re ¼
2πaεr ε0 ψ20 ln 1 þ e−κh ; kT
ð5Þ
where εr and ε0 are the dielectric constants of solvent and vacuum, which are 37.8 [49] and 8.85 × 10−12 F/m [29], respectively; k is the Boltzmann constant (1.38 × 10−23 J/K) [29], T is the temperature (298 K), κ is the reciprocal of double layer thickness, and ψ0 is the surface potential of the particles. Since suspensions are organic-based in this investigation, it was assumed that κ is equal to zero because of the extremely low concentration of ions. Also, the surface potential ψ0 was reasonably approximated by the electrokinetic zeta potential (ζ) (Table 1). The resulted VRe and VA of GZO in the presence of various dispersants, as calculated using Eq. (5), are shown in Fig. 5(a). From the low values of all resulting VRe, it is clear that the contribution of VRe to the stabilization of GZO in DMAC is very small, which should be insufficient for the dispersion of the powder [50]. VRs comprises contributions of mixing free energy (Vmix Rs ) and elastic free energy (Vela Rs ), as described by Eqs. (6) and (7), respectively:
V mix Rs ¼
2 4πkTC 2 1 h h −χ t− 3a þ 2t þ 2 2 3V sol ρ2dis 2
ð6Þ
ð3Þ
where h is the center-to-center distance between two particles, a is the average radius (1 × 10−8 m) of the powder, and A123 is the Hamaker constant that is determined from the Hamaker constants of powder (A11), solvent (A22), and dispersant (A33). A123 ¼
pffiffiffiffiffiffiffi pffiffiffiffiffiffiffipffiffiffiffiffiffiffi pffiffiffiffiffiffiffi A22 − A33 : A11 − A33
ð4Þ
Because the Hamaker constant was not found for GZO and GZO contains only 1.3 wt% of Ga2O3, the A11 of ZnO, which is 9.21 × 10−20 J [44], was substituted for that of GZO for above calculations. A22 is 6.80 × 10−20 J corresponding to DMAC [45,46]; A33 is 8.82 × 10−20 J corresponding to PEI [45,47] and 7.18 × 10−20 J corresponding to OP3 [45,48]. On considering the repulsive energy for the stabilization of particles, two
Fig. 5. Theoretical potential energies: (a) VA and VRe and (b) VT as a function of the surfaceto-surface distance between GZO particles in the presence of various dispersants in DMAC.
230
V ela Rs
C.-C. Li et al. / Powder Technology 305 (2017) 226–231
5=2 ðh−2aÞ ¼ 0:75G t− ða þ t Þ1=2 ; 2
ð7Þ
where C represents the saturated adsorption concentration of the dispersant (g/ml) that can be obtained from Fig. 2(a), t is the adsorption thickness (Table 1), Vsol is the molecular volume of the solvent (92.8 cm3/mol) [51], G is the elastic modulus of the adsorbed layer (105 dyne/cm2) [39], and ρdis is the density of the dispersant, which are 1.02 g/cm3 and 1.03 g/cm3 for OP3 [52] and PEI [53], respectively. χ is the Flory–Huggins constant, which includes two parts relating to entropy (χS) and enthalpy (χH), as described in Eq. (8). When two particles approach a distance less than double the value of t, Vmix Rs will increase due to the increase in osmotic pressure. When the distance between particles is less than t, Vela Rs will increase due to the increase of energy for the compression of the adsorbed layer. χ ¼ χS þ χH ¼ χS þ
V sol ðδ −δdis Þ2 kT sol
ð8Þ
δsol and δdis represent the solubility parameters of the solvent and the dispersant, respectively; they are 23.5 MPa1/2 for PEI [26], 19.5 MPa1/2 for OP3 [27], and 22.7 MPa1/2 for DMAC [28]. According to thermodynamics, the change in entropy can be ignored when systems are liquid or solid [39]; therefore, χH was the only term to be considered. The t values required for the calculations of Eqs. (6) and (7) were obtained by the analyses of relative viscosity (ηr) of suspensions, using viscometric method [39]. ηr is relevant to the apparent hydrodynamic volume fraction of the dispersed phase (ϕ) that contains particles and the corresponding adsorbed layers, and it can be approximated by the modified Einstein equation (Eq. (9)) [39]: ηr ¼ 1 þ 3ϕ þ 23ϕ2 :
ð9Þ
The average t of the adsorbed dispersant can then be determined by Eq. (10): t ¼ Δϕ=SA ;
ð10Þ
where Δϕ is the effective volume of the adsorbed layer calculated from the difference between the apparent volume (resulted ϕ from Eq. (9)) and the actual volume (ϕ0) of the dispersed phase and SA is the total surface area of particles in the suspension. For 20 wt% GZO suspensions (ϕ0 = 3.97%), the values of ηr are 5.34, 1.42, and 7.34 when 10 wt% of OP3, H-PEI, and L-PEI are added, respectively. The t values upon these conditions were calculated and listed in Table 1. Using t, the resulting ela total potential energies (VT; VT = VA + VRe + Vmix Rs + VRs ) of GZO as functions of surface-to-surface distance (h-2a) between particles are shown in Fig. 5(b). Comparing the total potential energy VT and the VRe in Fig. 5(a), it is obvious that the contribution of steric repulsion ela (Vmix Rs + VRs ) is much greater than that of VRe and will dominate the stabilization of the powder. Among the three dispersants, H-PEIadsorbed-GZO owned the highest steric hindrance, which extremely high energy barrier will repel particles apart effectively when they are too close. The L-PEI-adsorbed-GZO showed a maximum energy barrier of ~ 22 kT, which just exceeds the thermal energy of particles (~ 15 kT) [50]. Hence, L-PEI is theoretically efficient, but its dispersion efficiency is obviously poorer than that of H-PEI because of its lower molecular weight. For the adsorption of OP3, a maximum energy barrier of ~ 8 kT was obtained. Therefore, based on the theoretical calculation and experimental rheology, OP3 is clearly not an efficient dispersant for stabilizing GZO in the organic suspensions. 4. Conclusions Although PEI is a common water-based dispersant and OP3 is an organic-based surfactant, they are miscible with the polar dispersion
medium DMAC, which qualified them as dispersant candidates in this investigation. The dispersion efficiencies and stabilization mechanisms of GZO nanopowder in the polar organic solvent—DMAC with the additions of dispersants, OP3 and two types of PEI with molecular weights of 1800 g/mol and 10,000 g/mol, were studied. The interactions between GZO and dispersants were analyzed by zeta potentials and adsorption isotherms. The dispersion stabilities of GZO in DMAC with the additions of dispersants were analyzed by measurements of suspension rheologies and calculations of potential energies based on the extended DLVO theory. The rheological results indicate that the water-based dispersants L-PEI and H-PEI are instead to be more efficient than the organic-based dispersant OP3 for the dispersion of GZO in the organic medium DMAC. Although the polyelectrolyte L-PEI is probable to be more soluble in DMAC than H-PEI due to its lower molecular weight, which solubility is usually a considerable issue for the selection for dispersants, L-PEI is not as efficient as that of H-PEI. With 10 wt% added dispersant, a good dispersion of GZO suspension with a solid content of 40 wt% was obtained when the dispersant is H-PEI, whereas a less stable dispersion was obtained when the dispersant is L-PEI. According to the results of the DLVO calculations, the steric repulsive energy contributed more than the electrostatic repulsion and is the primary mechanism for the stabilization of GZO nanopowder in the polar organic medium DMAC. Acknowledgments The authors are grateful for the financial support by the Material and Chemical Research Laboratories of Industrial Technology Research Institute. References [1] S. Nagarani, M. Jayachandran, C. Sanjeeviraja, Review on gallium zinc oxide films: material properties and preparation techniques, Mater. Sci. Forum 671 (2011) 47–68. [2] P.K. Nayak, J. Yang, J. Kim, S. Chung, J. Jeong, C. Lee, Y. Hong, Spin-coated Ga-doped ZnO transparent conducting thin films for organic light-emitting diodes, J. Phys. D. Appl. Phys. 42 (2009) 035102. [3] C.M. Hsu, W.C. Tzou, C.F. Yang, Y.J. Liou, Investigation of the high mobility IGZO thin films by using co-sputtering method, Material 8 (5) (2015) 2769–2781. [4] C.G. Kuo, I.C. Lin, C.I. Chuang, C.F. Yang, H.Y. Yang, C.C. Wu, Deposition of ZnO-In2O3Ga2O3 (IGZO) thin films by using spray coating method, Communication and Engineering, 3rd ed.CRC Press/Balkema, Communication and Engineering, China 2014, pp. 11–14. [5] Y.H. Yang, S.S. Yang, C.Y. Kao, K.S. Chou, Chemical and electrical properties of lowtemperature solution-processed InGaZn-O thin-film transistors, IEEE Electron Device Lett. 31 (4) (2010) 329–331. [6] R.M. Pasquarelli, D.S. Ginley, R. O'Hayre, Solution processing of transparent conductors: from flask to film, Chem. Soc. Rev. 40 (2011) 5406–5441. [7] S.Y. Park, B.J. Kim, K. Kim, M.S. Kang, K.H. Lim, T.I. Lee, J.M. Myoung, H.K. Baik, J.H. Cho, Y.S. Kim, Low-temperature, solution-processed and alkali metal doped ZnO for high-performance thin-film transistors, Adv. Mater. 24 (6) (2012) 834–838. [8] J.S. Park, J.K. Jeong, H.J. Chung, Y.G. Mo, H.D. Kim, Electronic transport properties of amorphous indium-gallium-zinc oxide semiconductor upon exposure to water, Appl. Phys. Lett. 92 (7) (2008) 072104-1–072104-3. [9] F.H. Wang, H.P. Chang, C.C. Tseng, C.C. Huang, H.W. Liu, Influence of hydrogen plasma treatment on Al-doped ZnO thin films for amorphous silicon thin film solar cells, Curr. Appl. Phys. 11 (1) (2011) S12–S16. [10] A. Alkahlout, A comparative study of spin coated transparent conducting thin films of gallium and aluminum doped ZnO nanoparticles, Phys. Res. Int. 2015 (2015) 1–8. [11] K. Yim, H.W. Kim, C. Lee, Effects of annealing on structure, resistivity and transmittance of Ga doped ZnO films, Mater. Sci. Technol. 23 (1) (2007) 108–112. [12] T. Minami, H. Sato, H. Nanto, S. Takata, Group III impurity doped zinc oxide thin films prepared by RF magnetron sputtering, Jpn. J. Appl. Phys. 24 (1985) L781–L784. [13] T. Minami, Transparent conducting oxide semiconductors for transparent electrodes, Semicond. Sci. Technol. 20 (2005) S35–S44. [14] Z. Shen, P.E. Burrows, V. Bulović, S.R. Forrest, M.E. Thompson, Three-color, tunable, organic light-emitting devices, Science 276 (1997) 2009–2011. [15] H.L. Hartnagel, A.L. Dawar, A.K. Jain, C. Jagadish, Semiconducting Transparent Thin Films, CRC Press, Florida, 1995. [16] Q.B. Ma, Z.Z. Ye, H.P. He, S.H. Hu, J.R. Wang, L.P. Zhu, Y.Z. Zhang, B.H. Zhao, Structural, electrical, and optical properties of transparent conductive ZnO:Ga films prepared by DC reactive magnetron sputtering, J. Cryst. Growth 304 (1) (2007) 64–68. [17] K.T. Ramakrishna Reddy, T.B.S. Reddy, L. Forbes, R.W. Miles, Highly oriented and conducting ZnO:Ga layers grown by chemical spray pyrolysis, Surf. Coat. Technol. 151–152 (2002) 110–113.
C.-C. Li et al. / Powder Technology 305 (2017) 226–231 [18] A.R. Kaul, O.Y. Gorbenko, A.N. Botev, L.I. Burova, MOCVD of pure and Ga-doped epitaxial ZnO, Superlattice. Microst. 38 (4–6) (2005) 272–282. [19] S.M. Park, T. Ikegami, K. Ebihara, Effects of substrate temperature on the properties of Ga-doped ZnO by pulsed laser deposition, Thin Solid Films 513 (1–2) (2006) 90–94. [20] V. Fathollahi, M.M. Amini, Sol-gel preparation of highly oriented gallium-doped zinc oxide thin films, Mater. Lett. 50 (4) (2001) 235–239. [21] C.Y. Tsay, H.C. Cheng, Y.T. Tung, W.H. Tuan, C.K. Lin, Microstructural and optical properties of Ga-doped ZnO semiconductor thin films prepared by sol–gel process, Thin Solid Films 519 (5) (2008) 1516–1520. [22] C.Y. Tsay, K.S. Fan, C.M. Lei, Synthesis and characterization of sol–gel derived gallium-doped zinc oxide thin films, J. Alloys Compd. 512 (1) (2012) 216–222. [23] K.K. Banger, Y. Yamashita, K. Mori, R.L. Peterson, T. Leedham, J. Rickard, H. Sirringhaus, Low-temperature, high-performance solution-processed metal oxide thin-film transistors formed by a ‘sol–gel on chip’ process, Nat. Mater. 10 (1) (2011) 45–50. [24] J.S. Park, W.J. Maeng, H.S. Kim, J.S. Park, Review of recent developments in amorphous oxide semiconductor thin-film transistor devices, Thin Solid Films 520 (6) (2012) 1679–1693. [25] M. Lee, J. Dho, Controlling the electrical and the optical properties of amorphous IGZO films prepared by using pulsed laser deposition, J. Korean Phys. Soc. 58 (3) (2011) 492–497. [26] G. Bakeri, A.F. Ismail, M. Rahimnejad, T. Matsuura, Analysis of polyetherimide/Nmethyl-2-pyrrolidone/nonsolvent phase separation behavior, J. Polym. Res. 21 (2014) 386-1–386-7. [27] S. Mae, (Meth)acrylic polymer particle, method for producing the same, plastisol composition made from the polymer particle, and article made from the plastisol composition, US Patent No. 8,039,541 B2, 2011. [28] S.P. Deshmukh, K. Li, Effect of ethanol composition in water coagulation bath on morphology of PVDF hollow fibre membranes, J. Membr. Sci. 150 (1) (1998) 75–85. [29] M.W. Barsoum, Fundamentals of Ceramics, CRC Press, Florida, 2003. [30] Z. Deng, Y. Tian, Z. Yin, Q. Rui, H. Liu, Y. Luo, Physical vapor deposited zinc oxide nanoparticles for direct electron transfer of superoxide dismutase, Electrochem. Commun. 10 (5) (2008) 818–820. [31] A.C.L. Su, J.W. Collette, Ligand effect on the stereoselectivity and catalyst activity in a rhodium-catalyzed synthesis of 1,4-hexadiene from ethylene and butadiene, J. Organomet. Chem. 90 (1975) 227–234. [32] J.P. Richard, G. Williams, A.C. O'Donoghue, T.L. Amyes, Formation and stability of enolates of acetamide and acetate anion: an eigen plot for proton transfer at α-carbonyl carbon, J. Am. Chem. Soc. 124 (12) (2002) 2957–2968. [33] F.Z. Dörwald, Side Reactions in Organic Synthesis: A Guide to Successful Synthesis Design, John Wiley & Sons, 2006. [34] A. Kitahara, Zeta potential in nonaqueous media and its effect on dispersion stability, Prog. Org. Coat. 2 (1973/74) 81–98.
231
[35] R.R. Burgess, Use of polyethyleneimine in purification of DNA-binding proteins, Methods Enzymol. 208 (1991) 3–10. [36] I.S. Blagbrough, A.A. Metwally, A.J. Geall, Measurement of polyamine pKa values, Methods Mol. Biol. 720 (2011) 493–503. [37] A. Jakimska, B. Huerta, Z. Barganska, A. Kot-Wasika, S. Rodríguez-Mozaz, D. Barceló, Development of a liquid chromatography-tandem mass spectrometry procedure for determination of endocrine disrupting compounds in fish from Mediterranean rivers, J. Chromatogr. A 1306 (2013) 44–58. [38] M.S. Kuo, S.J. Chang, P.H. Hsieh, Y.C. Huang, C.C. Li, Efficient dispersants for TiO2 nanopowder organic suspensions, J. Am. Ceram. Soc. 99 (2) (2016) 445–451. [39] T. Sato, R. Ruch, Stabilization of Colloidal Dispersions by Polymer Adsorption, CRC Press, Florida, 1980. [40] S.J. Chang, W.S. Liao, C.J. Ciou, J.T. Lee, C.C. Li, An efficient approach to derive hydroxyl groups on the surface of barium titanate nanoparticles to improve its chemical modification ability, J. Colloid Interface Sci. 329 (2009) 300–305. [41] E. Kissa, Dispersions, Characterization, Testing, and Measurement, CRC Press, Florida, 1999. [42] Z. Astolfi-Filho, V.R.N. Telis, E.B. de Oliveira, J.S. dos Reis Coimbra, J. Telis-Romero, Rheology and fluid dynamics properties of sugarcane juice, Biochem. Eng. J. 53 (2011) 260–265. [43] D. Brown, P.R. Jaffé, Effects of nonionic surfactants on the cell surface hydrophobicity and apparent hamaker constant of a Sphingomonas sp. Environ. Sci. Technol. 40 (1) (2006) 195–201. [44] M. Rahaman, M.N. Rahaman, Ceramic Processing, CRC Press, Florida, 2006. [45] P.C. Hiemenz, R. Rajagopalan, Principles of Colloid and Surface Chemistry, Marcel Dekker, New York, 1997. [46] F.K. Ko, Y. Wan, Introduction to Nanofiber Materials, Cambridge University Press, Cambridge, 2014. [47] H.R. Kang, Waterfast Ink Jet Compositions and Process, US Patent No. 4,659,382, 1987. [48] Y. Liu, L. Huang, S.M. Li, T. Harner, J. Liggio, OH-initiated heterogeneous oxidation of tris-2-butoxyethyl phosphate: implications for its fate in the atmosphere, Atmos. Chem. Phys. 14 (2014) 12195–12207. [49] V.H. Agreda, Acetic Acid and Its Derivatives, CRC Press, Florida, 1992. [50] L.L. Schramm, Emulsions, Foams, and Suspensions, Wiley-VCH, Weinheim, 2005. [51] E.A. Grulke, Solubility parameter values, in: J. Brandrup, E.H. Immergut, E.A. Grulke (Eds.), Polymer Handbook, 4th ed., vol. 7, Wiley, New York 1999, pp. 675–714. [52] J. Crawford, J. Doyle, D.G. Jones, J. Lippe, F. Llados, G.D. Todd, Toxicological Profile for Phosphate Ester Flame Retardants, U.S. Department of Health and Human Services, Washington, D.C., 2012 [53] R. Jordan, Surface-Initiated Polymerization II, Springer, New York, 2006.
Contents lists available at ScienceDirect
Powder Technology journal homepage: www.elsevier.com/locate/powtec
Water-soluble polyethylenimine as an efficient dispersant for gallium zinc oxide nanopowder in organic-based suspensions Chia-Chen Li a,⁎, Jia-Hao Jhang a, Hsin-Yi Tsai a, Yung-Pin Huang b a b
Institute of Materials Science and Engineering, Department of Materials & Mineral Resources Engineering, National Taipei University of Technology, Taipei 10608, Taiwan Material and Chemical Research Laboratories, Industrial Technology Research Institute, Hsinchu 30011, Taiwan
a r t i c l e
i n f o
Article history: Received 29 June 2016 Received in revised form 29 September 2016 Accepted 1 October 2016 Available online 4 October 2016 Keywords: Gallium zinc oxide Transparent conducting oxide Dispersion Polyethylenimine Tris(2-butoxyethyl) phosphate
a b s t r a c t Appropriate dispersants for the dispersion of gallium zinc oxide (GZO) nanopowder in the commonly used organic medium-dimethylacetamide (DMAC) are proposed by this study. The dispersion efficiencies and stabilization mechanisms of three dispersants—the organic-based dispersant tris(2-butoxyethyl) phosphate (OP3) and two water-based dispersant polyethylenimines (PEIs) with different molecular weights—are compared. Surprisingly, the water-based PEI polyelectrolytes exhibit greater efficiency than the organic-based OP3 for the dispersion of GZO nanopowder in the organic-based suspensions. This is because that the stabilization mechanism of GZO in the very polar DMAC is primarily related to steric repulsion rather than to electrostatic repulsion according to the theoretical and experimental analyses. © 2016 Published by Elsevier B.V.
1. Introduction Zinc oxide (ZnO)-based powders have attracted considerable interest as transparent conducting oxide (TCO) materials in recent years because of their competitive advantages of high natural abundance, low cost, nontoxicity, low resistance, high transparency, and good chemical and thermal stabilities [1–10]. The traditional TCO is indium tin oxide (ITO), which is a very expensive material that is unstable in hydrogen plasma [2,9,10]; thus, numerous doped ZnO compounds, including In2O3-doped ZnO, Al2O3-doped ZnO (AZO), and Ga2O3-doped ZnO (gallium zinc oxide, GZO), have been proposed as replacements for ITO [1– 10]. Among the doped ZnO powders, AZO is the most common; however, GZO is the most chemically stable because Ga is less reactive than Al toward environmental oxygen and because the ZnO lattice is less distorted by the substitution of Zn2+ with Ga3+ than by the substitution of Zn2+ with Al3+ [11–13]. TCO materials are generally and extensively manufactured as films for use in optical and electrical devices [2,14,15] such as solar cells, flat panels, liquid crystal displays, transistors, light-emitting diodes, and window coatings. Several approaches have been used for preparing TCO films [1,2,16–20], including various physical vapor deposition, sputtering, chemical vapor deposition, epitaxy, ion plating, spray pyrolysis, and solution growth processes. Among these techniques, the method of sol-gel spin coating, which is a solution growth process, has ⁎ Corresponding author. E-mail address: [email protected] (C.-C. Li).
http://dx.doi.org/10.1016/j.powtec.2016.10.006 0032-5910/© 2016 Published by Elsevier B.V.
attracted the interest of numerous research groups [21–24]. Interest in sol-gel spin coating stems from its distinct advantages of cost effectiveness, simplicity, and suitability for mass production because of its high deposition rate. Nevertheless, disadvantages and drawbacks of this method include the challenging conditions needed to control the homogeneity of the sol-gel reaction and the difficulty associated with simultaneously optimizing a chemical reaction and achieving a uniform coating. A new approach using a well-dispersed suspension has been proposed as a simple alternative method of depositing TCO films; this method also features the advantages of low cost and suitability for mass production [10]. Although the preparation of uniform films through the use of welldispersed powder suspensions is advantageous, most commercially available doped ZnO powders such as AZO and GZO are very fine, with particle sizes of b50 nm. Such fine particles undergo severe aggregation in suspension and are not easily deagglomerated and dispersed. The demand for doped ZnO powders has increased for a wide range of products, including paints and coatings that provide heat-insulating or antistatic functions [25] and various optical and electrical devices; consequently, techniques for preparing stable suspensions of these powders are reasonably expected to increase in importance and to attract increased attention in the near future. Thus far, few reports regarding the dispersion of these powders or the development of appropriate dispersants have been published. In this work, appropriate dispersants for the dispersion of doped ZnO in organic-based suspensions are proposed and explored. GZO powder and dimethylacetamide (DMAC), a common solvent in
C.-C. Li et al. / Powder Technology 305 (2017) 226–231
227
industrial processes, were adopted as a model for this investigation. The purpose is to identify an effective dispersant for GZO powder and to achieve a stable dispersion of GZO in an organic suspension with a high solid content. Moreover, the dominant stabilization mechanism for the dispersion of GZO in suspensions is discussed and clarified. 2. Materials and methods High-purity GZO nanopowder (54W-GZ10, ITRI, Taiwan) with a hexagonal wurtzite crystal structure was used. The content of gallium oxide (Ga2O3) in GZO powder was 1.3 wt%, as determined by inductively coupled plasma atomic emission spectroscopy (ICAP-9000, JarrellAsh, USA). The primary particle size of GZO was approximately 20 nm, as characterized by transmission electron microscopy (JEM-2100, JEOL, USA). The surface area of GZO was 47 m2/g. The dispersants used included tris(2-butoxyethyl) phosphate (OP3, 95%; TCI, Tokyo, Japan) and two types of polymeric polyethylenimines (PEIs) with molecular weights of 1800 g/mol (L-PEI, 99%; Alfa Aesar, USA) and 10,000 g/mol (H-PEI, 99%; Alfa Aesar, USA). The polar organic solvent dimethylacetamide (DMAC, 99%; J.T. Baker, USA) was used as the dispersion medium. The electrokinetics of GZO nanopowder was analyzed using the electroacoustic method (ZetaProbe, Colloidal Dynamics Inc., North Attleborough, MA, USA). For the zeta-potential measurements, GZO suspensions in DMAC with a solid content of 0.5 wt% were prepared with the addition of 5 wt% of different dispersants. The suspensions were mixed and deagglomerated by a ball mill (MUBM236, Yeong-Shin, Taiwan) at 300 rpm with the addition of 50% mill charge of 5-mm yttria-stabilized zirconia media for 24 h at room temperature. In the adsorption experiments, the suspensions were prepared with a solid content of 20 wt% (based on the weight of solvent) in the presence of various concentrations of dispersants (based on the weight of GZO powder). The powder suspensions were deagglomerated by ball milling for 24 h and centrifuged at 5500 rpm to separate GZO nanopowder from the supernatant. The separated nanopowder was dried in an oven at 80 °C, and the amount of dispersant adsorbed onto the nanopowder was determined by thermogravimetric analysis (Q50, TA Instruments Ltd., Crawley, UK). The rheology of GZO suspensions with a solid content of 20 wt% and with OP3, L-PEI, and H-PEI added at various concentrations was analyzed by using a concentric cylinder rheometer (AR1000, TA Instruments Ltd., UK). In addition, the thixotropic behavior was analyzed based on the suspensions with solid contents of 10–40 wt% in the presence of 10 wt% of L-PEI or H-PEI. Cone-plate geometry fixtures with a diameter of 20 mm and cone angle of 1° were chosen for the steady-shear rotation tests. All the rheological curves were fitted using the software (V 5.2.26, Rheology Advantage™, UK) attached to the rheometer. The reliability of the equation was evaluated by the coefficient of determination (R2). The R2 obtained from the fittings of all samples are within 0.95–1.0. 3. Results and discussion For the dispersion of nanopowders, the investigated dispersant candidates are generally small molecules or polymers with low molecular weights because of their reasonable size compared to the small nanoparticles. Three dispersants—OP3 and two types of PEI—were chosen, and their dispersion efficiencies for GZO nanopowder in the polar organic solvent—DMAC were explored. The chemical structures of the investigated dispersants are shown in Fig. 1. OP3 is a small molecular surfactant with a phosphate group, and L-PEI and H-PEI are both polyethylenimines (PEIs) with molecular weights of 1800 g/mol and 10,000 g/mol, respectively. The solubility parameters (δ) of the dispersants and solvent are 23.5 MPa1/2 for PEI [26], 20.5 MPa1/2 for OP3 [27], and 22.7 MPa1/2 for DMAC [28]. The δ values of OP3 and PEI are very close to that of DMAC, indicating that DMAC is good solvent for dissolving OP3 and PEI. On the other hand, the logarithm of acid dissociation constant (pKa) of the suspension components was also
Fig. 1. Chemical structures of (a) tris(2-butoxyethyl) phosphate (OP3) and (b) polyethylenimine (PEI).
compared. Although the pKa of GZO has not been reported in the literature, the content of Ga in GZO is extremely low and Ga exhibits a higher cationic potential than does Zn, as a consequence of its smaller ionic radius and higher charge valence [29], we assumed that GZO should be more acidic than ZnO and the pKa of GZO is slightly lower than that of ZnO. The pKa of ZnO has been reported to be 9.3 [30]. From the electrokinetic measurements, it is found that GZO exhibits a negative zeta potential at − 1.5 mV in DMAC. This indicates that DMAC should be like an alkaline medium for the dispersion of GZO, although the exact pKa is not clear due to its very different reports in the literature, ranging from −0.5 to 30 [31–33]. Since DMAC is a high polar solvent, the dispersed GZO is possible to carry charge via the ionization mechanism based on the electron affinity differences between itself and the dispersion medium, parallel to the case occurred in an aqueous suspension [34]. When H-PEI or L-PEI was added, the GZO powder changed to exhibit a positive charge in DMAC. The pKa of PEI has been reported to be ~11 [35,36]. The possibility for the varied zeta potential of GZO is that both PEIs are positively charged and have adsorbed on GZO. The low positive value of zeta (ζ) potentials due to the addition of PEI may be caused by two reasons. One is that PEI does not have a good dissociation in DMAC; the other is the adsorbed amount of PEI on GZO is low. And the combined reason is also probable. The dispersant OP3 has a reported pKa ranged from − 3 to − 9 [37], lower than any possible pKa value reported for DMAC. Therefore, OP3 would preferentially dissociate to carry a negative charge in DMAC. As shown in Table 1, the resulting zeta potential of GZO with the addition of OP3 is −4.0 mV, just evidencing the specific adsorption of OP3 onto GZO. The adsorption isotherms of OP3, L-PEI, and H-PEI on GZO nanopowder are compared in Fig. 2(a). Their significant adsorptions correspond well to the behaviors predicted on the basis of the zeta-potential measurements. The adsorption amounts of the three dispersants increased with increasing dispersant concentration and finally reached saturation, demonstrating typical chemisorption behavior [38]. The adsorption saturations for OP3, L-PEI, and H-PEI were approximately 20, 74, and 112 mg/g GZO, respectively. We further analyzed the exact adsorption behavior of the dispersants using the Langmuir monolayer adsorption equation (Eq. (1)) [39]:
Ce Ce K ¼ þ ; As Cm Cm
ð1Þ
Table 1 Zeta potentials (ζ) and adsorption layer thicknesses (t) of GZO nanopowders in DMAC upon the addition of various dispersants. Dispersant
ζ (mV)
t (nm)
None OP3 L-PEI H-PEI
−1.5 −4.0 9.3 15.0
– 3.1 3.1 5.1
228
C.-C. Li et al. / Powder Technology 305 (2017) 226–231
Fig. 3. Apparent viscosity of 20 wt% GZO suspensions obtained at a shear rate of 300 s−1 plotted as a function of dispersant concentration. The inset shows the static rheological curves of 20 wt% GZO suspensions as functions of various concentrations of H-PEI.
Fig. 2. (a) Adsorption amounts of three dispersants on GZO as functions of the equilibrium concentration of dispersants. (b) Replot of the data in (a) on the basis of the Langmuir equation. Note that the solid lines are the fitting results.
where Ce is the equilibrium concentration of the dispersant in suspension, As is the adsorbance of the dispersant, Cm is the saturated/monolayer adsorbance of the dispersant, and K is a constant. When Ce/As was plotted as a function of Ce, straight lines were obtained for these dispersants, as shown in Fig. 2(b). The results in Fig. 2(b) indicate that all isotherms in Fig. 2(a) correspond to Langmuir-type adsorption (i.e., the dispersants are adsorbed chemically). Generally, a powder can be better stabilized if more dispersant molecules adsorb onto the particles because greater amounts of adsorbed molecules might result in a thicker adsorption layer on the particles, which, in turn, results in a greater repulsive energy among the particles, thereby preventing agglomeration. Hence, the fact that H-PEI exhibited the largest adsorption amount on GZO in Fig. 2(a) suggests that it might be the most efficient dispersant. To determine the relative dispersion efficiencies of dispersants, we compared the rheologies of powder suspensions with a solid content of 20 wt% stabilized by the addition of different dispersants at various concentrations. The inset of Fig. 3 reveals the viscosity of suspensions upon the addition of H-PEI. When the concentration of added H-PEI was b5 wt%, the rheological curve showed that viscosity decreased with increasing shear rate, demonstrating typical pseudoplastic flow behavior. That is, the suspension still contained powder agglomerates when the added concentration of H-PEI ([H-PEI]) was too low. When the [H-PEI] was increased to 5 wt%, a substantial change in the
rheological behavior of the suspensions was observed. The viscosities of the suspensions remained constant over the whole investigated range of shear rates, demonstrating typical Newtonian flow behavior (i.e., no powdered agglomerates existed in the suspensions). The viscosities of suspensions obtained at a shear rate of 300 s−1 as a function of [H-PEI] are replotted in Fig. 3; the viscosities that resulted from the additions of OP3 and L-PEI are compared in the figure. When L-PEI or H-PEI was used as the dispersant, the lowest concentration at which a stable viscosity was obtained was 5 wt%, whereas it was 10 wt% when the dispersant was OP3. In addition, the lowest stable viscosity of suspensions to which OP3 was added was higher than that of suspensions to which L-PEI or H-PEI was added. Obviously, PEI is a better dispersant than OP3, and L-PEI and H-PEI exhibit good efficiency for the dispersion of GZO in DMAC. To further compare the effect of the molecular weight of PEI on its ability to stabilize GZO in suspensions, the non-equilibrium shear stress obtained upon increasing and decreasing the shear rate for suspensions with solid contents of 10, 20, 30, and 40 wt% in DMAC was analyzed. Fig. 4 shows the resulting thixotropic hysteresis loops for these suspensions with 10 wt% added PEI. By fitting the increasing section of the thixotropic curve with the Herschel–Bulkley equation (Eq. (2)) [40–42], we
Fig. 4. Thixotropic curves of GZO suspensions with solid contents of (a) 10, (b) 20, (c) 30, and (d) 40 wt% in the presence of 10 wt% of L-PEI and H-PEI, respectively. The inset shows the viscosity obtained from the increscent section at 300 s−1 as a function of the solid content of suspensions.
C.-C. Li et al. / Powder Technology 305 (2017) 226–231
obtained detailed information pertaining to the flow behavior of suspensions: σ ¼ σ y þ ηγn ;
3 2 2 h −4a2 A123 4 2a2 2a2 5; þ 2 þ ln VA ¼ − 2 2 6 h h h −4a2
terms, sterically repulsive energy (VRs) and electrical double layer repulsive energy (VRe), are generally included, in which (VRe) is given by
ð2Þ
where σ is the shear stress, σy is the yield stress that can break apart the agglomerates in suspensions, η is the viscosity, γ is the shear rate, and n is the flow index. The flow will be Newtonian if n = 1 and shear thinning if n b 1 [39–41]. For the suspensions with added H-PEI at solid contents of 10–40 wt%, the n values are within 0.98–1.00, indicating that these suspensions exhibit well-dispersed Newtonian flow behavior. Note that the good dispersion at 40 wt%, which is an extremely high solid content for a suspension loaded with a nanosized powder, affirms that H-PEI is a highly efficient dispersant for nano-GZO in the polar organic solvent. Upon the addition of L-PEI, suspensions with various solid contents exhibited different values of n. At the lower solid content of 10 wt%, the suspension exhibited good dispersion, with n = 0.91. At solid contents of 20–30 wt%, suspensions became shear thinning, with 0.81 ≤ n ≤ 0.83, demonstrating that soft agglomerates formed in the suspensions. At the higher solid content of 40 wt%, the suspension showed more significant shear thinning, with n = 0.73, indicating more severe agglomeration of the powder. In addition, the calculated σy for suspensions with added H-PEI are all small, showing an insignificant increase from 0.09 Pa at 10 wt% to 0.12 Pa at 20 wt%, to 0.15 Pa at 30 wt%, and to 0.16 Pa at 40 wt%. For suspensions containing added L-PEI, σy increased significantly with increasing solid content, from 0.11 Pa at 10 wt% to 0.48 Pa at 20 wt%, to 0.52 Pa at 30 wt%, and to 0.63 Pa at 40 wt%. A greater σy value indicates a greater degree of powder agglomeration; therefore, the σy data show good agreement with the results for n. Moreover, the suspensions with H-PEI added at various concentrations show insignificant hysteresis loops (i.e., the viscosity is time-independently stable). In contrast, when L-PEI was added at a solid content of 40 wt%, the suspension exhibited a significant hysteresis loop. The inset of Fig. 4 compares the viscosity of suspensions with the presence of L-PEI and H-PEI, which is obtained at a shear rate of 300 s−1. Obviously, at any solid content of GZO, the addition with H-PEI results in a lower viscosity. Besides, the increase in the viscosity with increasing solid content for the suspension with the addition of H-PEI is not as drastic as that with L-PEI addition. According to the aforementioned comparisons, HPEI is significantly a more efficient dispersant than L-PEI although the latter may have better solubility in DMAC due to lower molecular weight. We further clarified the dispersion mechanisms of PEI and OP3 by using the extended Derjaguin–Landau–Verwey–Overbeek (DLVO) theory to calculate the stabilization force on the basis of their adsorptions [38,39,43]. The attractive potential between particles was assumed to be primarily due to van der Waals force (VA):
229
V Re ¼
2πaεr ε0 ψ20 ln 1 þ e−κh ; kT
ð5Þ
where εr and ε0 are the dielectric constants of solvent and vacuum, which are 37.8 [49] and 8.85 × 10−12 F/m [29], respectively; k is the Boltzmann constant (1.38 × 10−23 J/K) [29], T is the temperature (298 K), κ is the reciprocal of double layer thickness, and ψ0 is the surface potential of the particles. Since suspensions are organic-based in this investigation, it was assumed that κ is equal to zero because of the extremely low concentration of ions. Also, the surface potential ψ0 was reasonably approximated by the electrokinetic zeta potential (ζ) (Table 1). The resulted VRe and VA of GZO in the presence of various dispersants, as calculated using Eq. (5), are shown in Fig. 5(a). From the low values of all resulting VRe, it is clear that the contribution of VRe to the stabilization of GZO in DMAC is very small, which should be insufficient for the dispersion of the powder [50]. VRs comprises contributions of mixing free energy (Vmix Rs ) and elastic free energy (Vela Rs ), as described by Eqs. (6) and (7), respectively:
V mix Rs ¼
2 4πkTC 2 1 h h −χ t− 3a þ 2t þ 2 2 3V sol ρ2dis 2
ð6Þ
ð3Þ
where h is the center-to-center distance between two particles, a is the average radius (1 × 10−8 m) of the powder, and A123 is the Hamaker constant that is determined from the Hamaker constants of powder (A11), solvent (A22), and dispersant (A33). A123 ¼
pffiffiffiffiffiffiffi pffiffiffiffiffiffiffipffiffiffiffiffiffiffi pffiffiffiffiffiffiffi A22 − A33 : A11 − A33
ð4Þ
Because the Hamaker constant was not found for GZO and GZO contains only 1.3 wt% of Ga2O3, the A11 of ZnO, which is 9.21 × 10−20 J [44], was substituted for that of GZO for above calculations. A22 is 6.80 × 10−20 J corresponding to DMAC [45,46]; A33 is 8.82 × 10−20 J corresponding to PEI [45,47] and 7.18 × 10−20 J corresponding to OP3 [45,48]. On considering the repulsive energy for the stabilization of particles, two
Fig. 5. Theoretical potential energies: (a) VA and VRe and (b) VT as a function of the surfaceto-surface distance between GZO particles in the presence of various dispersants in DMAC.
230
V ela Rs
C.-C. Li et al. / Powder Technology 305 (2017) 226–231
5=2 ðh−2aÞ ¼ 0:75G t− ða þ t Þ1=2 ; 2
ð7Þ
where C represents the saturated adsorption concentration of the dispersant (g/ml) that can be obtained from Fig. 2(a), t is the adsorption thickness (Table 1), Vsol is the molecular volume of the solvent (92.8 cm3/mol) [51], G is the elastic modulus of the adsorbed layer (105 dyne/cm2) [39], and ρdis is the density of the dispersant, which are 1.02 g/cm3 and 1.03 g/cm3 for OP3 [52] and PEI [53], respectively. χ is the Flory–Huggins constant, which includes two parts relating to entropy (χS) and enthalpy (χH), as described in Eq. (8). When two particles approach a distance less than double the value of t, Vmix Rs will increase due to the increase in osmotic pressure. When the distance between particles is less than t, Vela Rs will increase due to the increase of energy for the compression of the adsorbed layer. χ ¼ χS þ χH ¼ χS þ
V sol ðδ −δdis Þ2 kT sol
ð8Þ
δsol and δdis represent the solubility parameters of the solvent and the dispersant, respectively; they are 23.5 MPa1/2 for PEI [26], 19.5 MPa1/2 for OP3 [27], and 22.7 MPa1/2 for DMAC [28]. According to thermodynamics, the change in entropy can be ignored when systems are liquid or solid [39]; therefore, χH was the only term to be considered. The t values required for the calculations of Eqs. (6) and (7) were obtained by the analyses of relative viscosity (ηr) of suspensions, using viscometric method [39]. ηr is relevant to the apparent hydrodynamic volume fraction of the dispersed phase (ϕ) that contains particles and the corresponding adsorbed layers, and it can be approximated by the modified Einstein equation (Eq. (9)) [39]: ηr ¼ 1 þ 3ϕ þ 23ϕ2 :
ð9Þ
The average t of the adsorbed dispersant can then be determined by Eq. (10): t ¼ Δϕ=SA ;
ð10Þ
where Δϕ is the effective volume of the adsorbed layer calculated from the difference between the apparent volume (resulted ϕ from Eq. (9)) and the actual volume (ϕ0) of the dispersed phase and SA is the total surface area of particles in the suspension. For 20 wt% GZO suspensions (ϕ0 = 3.97%), the values of ηr are 5.34, 1.42, and 7.34 when 10 wt% of OP3, H-PEI, and L-PEI are added, respectively. The t values upon these conditions were calculated and listed in Table 1. Using t, the resulting ela total potential energies (VT; VT = VA + VRe + Vmix Rs + VRs ) of GZO as functions of surface-to-surface distance (h-2a) between particles are shown in Fig. 5(b). Comparing the total potential energy VT and the VRe in Fig. 5(a), it is obvious that the contribution of steric repulsion ela (Vmix Rs + VRs ) is much greater than that of VRe and will dominate the stabilization of the powder. Among the three dispersants, H-PEIadsorbed-GZO owned the highest steric hindrance, which extremely high energy barrier will repel particles apart effectively when they are too close. The L-PEI-adsorbed-GZO showed a maximum energy barrier of ~ 22 kT, which just exceeds the thermal energy of particles (~ 15 kT) [50]. Hence, L-PEI is theoretically efficient, but its dispersion efficiency is obviously poorer than that of H-PEI because of its lower molecular weight. For the adsorption of OP3, a maximum energy barrier of ~ 8 kT was obtained. Therefore, based on the theoretical calculation and experimental rheology, OP3 is clearly not an efficient dispersant for stabilizing GZO in the organic suspensions. 4. Conclusions Although PEI is a common water-based dispersant and OP3 is an organic-based surfactant, they are miscible with the polar dispersion
medium DMAC, which qualified them as dispersant candidates in this investigation. The dispersion efficiencies and stabilization mechanisms of GZO nanopowder in the polar organic solvent—DMAC with the additions of dispersants, OP3 and two types of PEI with molecular weights of 1800 g/mol and 10,000 g/mol, were studied. The interactions between GZO and dispersants were analyzed by zeta potentials and adsorption isotherms. The dispersion stabilities of GZO in DMAC with the additions of dispersants were analyzed by measurements of suspension rheologies and calculations of potential energies based on the extended DLVO theory. The rheological results indicate that the water-based dispersants L-PEI and H-PEI are instead to be more efficient than the organic-based dispersant OP3 for the dispersion of GZO in the organic medium DMAC. Although the polyelectrolyte L-PEI is probable to be more soluble in DMAC than H-PEI due to its lower molecular weight, which solubility is usually a considerable issue for the selection for dispersants, L-PEI is not as efficient as that of H-PEI. With 10 wt% added dispersant, a good dispersion of GZO suspension with a solid content of 40 wt% was obtained when the dispersant is H-PEI, whereas a less stable dispersion was obtained when the dispersant is L-PEI. According to the results of the DLVO calculations, the steric repulsive energy contributed more than the electrostatic repulsion and is the primary mechanism for the stabilization of GZO nanopowder in the polar organic medium DMAC. Acknowledgments The authors are grateful for the financial support by the Material and Chemical Research Laboratories of Industrial Technology Research Institute. References [1] S. Nagarani, M. Jayachandran, C. Sanjeeviraja, Review on gallium zinc oxide films: material properties and preparation techniques, Mater. Sci. Forum 671 (2011) 47–68. [2] P.K. Nayak, J. Yang, J. Kim, S. Chung, J. Jeong, C. Lee, Y. Hong, Spin-coated Ga-doped ZnO transparent conducting thin films for organic light-emitting diodes, J. Phys. D. Appl. Phys. 42 (2009) 035102. [3] C.M. Hsu, W.C. Tzou, C.F. Yang, Y.J. Liou, Investigation of the high mobility IGZO thin films by using co-sputtering method, Material 8 (5) (2015) 2769–2781. [4] C.G. Kuo, I.C. Lin, C.I. Chuang, C.F. Yang, H.Y. Yang, C.C. Wu, Deposition of ZnO-In2O3Ga2O3 (IGZO) thin films by using spray coating method, Communication and Engineering, 3rd ed.CRC Press/Balkema, Communication and Engineering, China 2014, pp. 11–14. [5] Y.H. Yang, S.S. Yang, C.Y. Kao, K.S. Chou, Chemical and electrical properties of lowtemperature solution-processed InGaZn-O thin-film transistors, IEEE Electron Device Lett. 31 (4) (2010) 329–331. [6] R.M. Pasquarelli, D.S. Ginley, R. O'Hayre, Solution processing of transparent conductors: from flask to film, Chem. Soc. Rev. 40 (2011) 5406–5441. [7] S.Y. Park, B.J. Kim, K. Kim, M.S. Kang, K.H. Lim, T.I. Lee, J.M. Myoung, H.K. Baik, J.H. Cho, Y.S. Kim, Low-temperature, solution-processed and alkali metal doped ZnO for high-performance thin-film transistors, Adv. Mater. 24 (6) (2012) 834–838. [8] J.S. Park, J.K. Jeong, H.J. Chung, Y.G. Mo, H.D. Kim, Electronic transport properties of amorphous indium-gallium-zinc oxide semiconductor upon exposure to water, Appl. Phys. Lett. 92 (7) (2008) 072104-1–072104-3. [9] F.H. Wang, H.P. Chang, C.C. Tseng, C.C. Huang, H.W. Liu, Influence of hydrogen plasma treatment on Al-doped ZnO thin films for amorphous silicon thin film solar cells, Curr. Appl. Phys. 11 (1) (2011) S12–S16. [10] A. Alkahlout, A comparative study of spin coated transparent conducting thin films of gallium and aluminum doped ZnO nanoparticles, Phys. Res. Int. 2015 (2015) 1–8. [11] K. Yim, H.W. Kim, C. Lee, Effects of annealing on structure, resistivity and transmittance of Ga doped ZnO films, Mater. Sci. Technol. 23 (1) (2007) 108–112. [12] T. Minami, H. Sato, H. Nanto, S. Takata, Group III impurity doped zinc oxide thin films prepared by RF magnetron sputtering, Jpn. J. Appl. Phys. 24 (1985) L781–L784. [13] T. Minami, Transparent conducting oxide semiconductors for transparent electrodes, Semicond. Sci. Technol. 20 (2005) S35–S44. [14] Z. Shen, P.E. Burrows, V. Bulović, S.R. Forrest, M.E. Thompson, Three-color, tunable, organic light-emitting devices, Science 276 (1997) 2009–2011. [15] H.L. Hartnagel, A.L. Dawar, A.K. Jain, C. Jagadish, Semiconducting Transparent Thin Films, CRC Press, Florida, 1995. [16] Q.B. Ma, Z.Z. Ye, H.P. He, S.H. Hu, J.R. Wang, L.P. Zhu, Y.Z. Zhang, B.H. Zhao, Structural, electrical, and optical properties of transparent conductive ZnO:Ga films prepared by DC reactive magnetron sputtering, J. Cryst. Growth 304 (1) (2007) 64–68. [17] K.T. Ramakrishna Reddy, T.B.S. Reddy, L. Forbes, R.W. Miles, Highly oriented and conducting ZnO:Ga layers grown by chemical spray pyrolysis, Surf. Coat. Technol. 151–152 (2002) 110–113.
C.-C. Li et al. / Powder Technology 305 (2017) 226–231 [18] A.R. Kaul, O.Y. Gorbenko, A.N. Botev, L.I. Burova, MOCVD of pure and Ga-doped epitaxial ZnO, Superlattice. Microst. 38 (4–6) (2005) 272–282. [19] S.M. Park, T. Ikegami, K. Ebihara, Effects of substrate temperature on the properties of Ga-doped ZnO by pulsed laser deposition, Thin Solid Films 513 (1–2) (2006) 90–94. [20] V. Fathollahi, M.M. Amini, Sol-gel preparation of highly oriented gallium-doped zinc oxide thin films, Mater. Lett. 50 (4) (2001) 235–239. [21] C.Y. Tsay, H.C. Cheng, Y.T. Tung, W.H. Tuan, C.K. Lin, Microstructural and optical properties of Ga-doped ZnO semiconductor thin films prepared by sol–gel process, Thin Solid Films 519 (5) (2008) 1516–1520. [22] C.Y. Tsay, K.S. Fan, C.M. Lei, Synthesis and characterization of sol–gel derived gallium-doped zinc oxide thin films, J. Alloys Compd. 512 (1) (2012) 216–222. [23] K.K. Banger, Y. Yamashita, K. Mori, R.L. Peterson, T. Leedham, J. Rickard, H. Sirringhaus, Low-temperature, high-performance solution-processed metal oxide thin-film transistors formed by a ‘sol–gel on chip’ process, Nat. Mater. 10 (1) (2011) 45–50. [24] J.S. Park, W.J. Maeng, H.S. Kim, J.S. Park, Review of recent developments in amorphous oxide semiconductor thin-film transistor devices, Thin Solid Films 520 (6) (2012) 1679–1693. [25] M. Lee, J. Dho, Controlling the electrical and the optical properties of amorphous IGZO films prepared by using pulsed laser deposition, J. Korean Phys. Soc. 58 (3) (2011) 492–497. [26] G. Bakeri, A.F. Ismail, M. Rahimnejad, T. Matsuura, Analysis of polyetherimide/Nmethyl-2-pyrrolidone/nonsolvent phase separation behavior, J. Polym. Res. 21 (2014) 386-1–386-7. [27] S. Mae, (Meth)acrylic polymer particle, method for producing the same, plastisol composition made from the polymer particle, and article made from the plastisol composition, US Patent No. 8,039,541 B2, 2011. [28] S.P. Deshmukh, K. Li, Effect of ethanol composition in water coagulation bath on morphology of PVDF hollow fibre membranes, J. Membr. Sci. 150 (1) (1998) 75–85. [29] M.W. Barsoum, Fundamentals of Ceramics, CRC Press, Florida, 2003. [30] Z. Deng, Y. Tian, Z. Yin, Q. Rui, H. Liu, Y. Luo, Physical vapor deposited zinc oxide nanoparticles for direct electron transfer of superoxide dismutase, Electrochem. Commun. 10 (5) (2008) 818–820. [31] A.C.L. Su, J.W. Collette, Ligand effect on the stereoselectivity and catalyst activity in a rhodium-catalyzed synthesis of 1,4-hexadiene from ethylene and butadiene, J. Organomet. Chem. 90 (1975) 227–234. [32] J.P. Richard, G. Williams, A.C. O'Donoghue, T.L. Amyes, Formation and stability of enolates of acetamide and acetate anion: an eigen plot for proton transfer at α-carbonyl carbon, J. Am. Chem. Soc. 124 (12) (2002) 2957–2968. [33] F.Z. Dörwald, Side Reactions in Organic Synthesis: A Guide to Successful Synthesis Design, John Wiley & Sons, 2006. [34] A. Kitahara, Zeta potential in nonaqueous media and its effect on dispersion stability, Prog. Org. Coat. 2 (1973/74) 81–98.
231
[35] R.R. Burgess, Use of polyethyleneimine in purification of DNA-binding proteins, Methods Enzymol. 208 (1991) 3–10. [36] I.S. Blagbrough, A.A. Metwally, A.J. Geall, Measurement of polyamine pKa values, Methods Mol. Biol. 720 (2011) 493–503. [37] A. Jakimska, B. Huerta, Z. Barganska, A. Kot-Wasika, S. Rodríguez-Mozaz, D. Barceló, Development of a liquid chromatography-tandem mass spectrometry procedure for determination of endocrine disrupting compounds in fish from Mediterranean rivers, J. Chromatogr. A 1306 (2013) 44–58. [38] M.S. Kuo, S.J. Chang, P.H. Hsieh, Y.C. Huang, C.C. Li, Efficient dispersants for TiO2 nanopowder organic suspensions, J. Am. Ceram. Soc. 99 (2) (2016) 445–451. [39] T. Sato, R. Ruch, Stabilization of Colloidal Dispersions by Polymer Adsorption, CRC Press, Florida, 1980. [40] S.J. Chang, W.S. Liao, C.J. Ciou, J.T. Lee, C.C. Li, An efficient approach to derive hydroxyl groups on the surface of barium titanate nanoparticles to improve its chemical modification ability, J. Colloid Interface Sci. 329 (2009) 300–305. [41] E. Kissa, Dispersions, Characterization, Testing, and Measurement, CRC Press, Florida, 1999. [42] Z. Astolfi-Filho, V.R.N. Telis, E.B. de Oliveira, J.S. dos Reis Coimbra, J. Telis-Romero, Rheology and fluid dynamics properties of sugarcane juice, Biochem. Eng. J. 53 (2011) 260–265. [43] D. Brown, P.R. Jaffé, Effects of nonionic surfactants on the cell surface hydrophobicity and apparent hamaker constant of a Sphingomonas sp. Environ. Sci. Technol. 40 (1) (2006) 195–201. [44] M. Rahaman, M.N. Rahaman, Ceramic Processing, CRC Press, Florida, 2006. [45] P.C. Hiemenz, R. Rajagopalan, Principles of Colloid and Surface Chemistry, Marcel Dekker, New York, 1997. [46] F.K. Ko, Y. Wan, Introduction to Nanofiber Materials, Cambridge University Press, Cambridge, 2014. [47] H.R. Kang, Waterfast Ink Jet Compositions and Process, US Patent No. 4,659,382, 1987. [48] Y. Liu, L. Huang, S.M. Li, T. Harner, J. Liggio, OH-initiated heterogeneous oxidation of tris-2-butoxyethyl phosphate: implications for its fate in the atmosphere, Atmos. Chem. Phys. 14 (2014) 12195–12207. [49] V.H. Agreda, Acetic Acid and Its Derivatives, CRC Press, Florida, 1992. [50] L.L. Schramm, Emulsions, Foams, and Suspensions, Wiley-VCH, Weinheim, 2005. [51] E.A. Grulke, Solubility parameter values, in: J. Brandrup, E.H. Immergut, E.A. Grulke (Eds.), Polymer Handbook, 4th ed., vol. 7, Wiley, New York 1999, pp. 675–714. [52] J. Crawford, J. Doyle, D.G. Jones, J. Lippe, F. Llados, G.D. Todd, Toxicological Profile for Phosphate Ester Flame Retardants, U.S. Department of Health and Human Services, Washington, D.C., 2012 [53] R. Jordan, Surface-Initiated Polymerization II, Springer, New York, 2006.