Article pubs.acs.org/cm
New Insights into the Mechanism of ZnO Formation from Aqueous Solutions of Zinc Acetate and Zinc Nitrate Mei-Keat Liang,‡ Marion J. Limo,‡ Anna Sola-Rabada,‡ Martin J. Roe,† and Carole C. Perry* Biomolecular and Materials Interface Research Group, Interdisciplinary Biomedical Research Centre, School of Science and Technology, Nottingham Trent University, Clifton Lane, Nottingham NG11 8NS, U.K. S Supporting Information *
ABSTRACT: The controlled synthesis of ZnO at the micro- and nanoscale has been the focus of significant research due to its importance in electrical and optoelectronic applications, and the potential of tuning its properties at the crystal formation stage. We present a detailed study of ZnO growth processes which supports and consolidates previous findings and gives a clearer understanding of the mechanism of ZnO formation. The influence of synthesis conditions on ZnO formation was investigated by comparison of two different growth routes (Zn(CH3COO)2−NH3 and Zn(NO3)2· 6H2O−HMTA) both known to result in the formation of wurtzite structured, twinned hexagonal rods of ZnO. The identities of the solid phases formed and supernatants were confirmed by data from SEM, XRD, FTIR, XPS, TGA, and ICP-OES analysis; giving insight into the involvement of multistep pathways. In both cases, reaction takes place via intermediates known as layered basic zinc salts (LBZs) which only later transform to the oxide phase. In the ZnAc2−NH3 system, crystal growth evolves as Zn(CH3COO)2 → LBZA (A: acetate) → ZnO through a dissolution/reprecipitation process, with the formation of an additional product identified as LBZAC (C: carbonate). In contrast, in the Zn(NO3)2· 6H2O−HMTA system, solid-phase transformation occurs as Zn(NO3)2·6H2O → LBZN (N: nitrate) → ZnO with no evidence of dissolution. Similar comprehensive studies can be applied to other solid-state processes to further advance functional materials design.
■
ogy.21−24 A wide range of 1D nanometer to micrometer ZnO structures have been formed using solution synthesis methods such as rods, plates, tubes, rings, tetrapods, prisms, pyramids, spheres, hollow structures, flowerlike, and multineedle-shaped crystals.18,20,25−30 The driving force for chemical reactions leading to nucleation and crystal growth from solution is the need to minimize the free energy of the whole system, and the direction of crystal growth is dictated by the surface activities of the growing facets under specific growth conditions.23,31,32 Observations from biomineralization studies suggest that nucleation may follow more complex routes than proposed by classical nucleation theory.33,34 Crystallization could proceed through a multistep pathway involving structural and compositional transformation of amorphous precursors which subsequently crystallize, forming intermediates that are transformed to give the final crystalline structure.33−36 The progress of phase transformation usually occurs through dissolution and renucleation processes in a phenomenon referred to as Ostwald’s Law of phases which depends on the free energy of activation and the solubility of the intermediates.33,37−39 Ultimately, the resulting crystal morphology is thought to be governed by contributions from
INTRODUCTION From as early as 1935, studies on characterization of ZnO had begun;1−3 however, with improvements in technology to produce ZnO using synthetic processes, the past decade has seen a significant rise in research interest in developing the uses of ZnO especially for its potential electrical and optoelectronic applications.1,4−6 Emerging technological and industrial applications include the making of field emitters, varistors, acoustic wave devices, piezoelectric devices, solar cells, photocatalysts, transparent conducting materials, and possibly even chemical/ biosensors.7−9 Its biocompatible, biodegradable, and antimicrobial nature also encourages further developments for biomedical applications.10−12 As the physical/chemical properties and functions of a given material are dictated by its structure and/or morphology, controlled synthesis of materials at the micro- and nanoscale has been of research interest but is still met by challenges with a rising need for more facile, reproducible, efficient, economic, and environmentally friendly fabrication routes.13−16 Solution synthesis methods using alkoxides or simple salts as precursors17−20 have become a favored bottom-up synthesis route for metals/metal oxides as they allow for the fine-tuning of growth parameters such as solution composition, chemistry, temperature, reactant concentration, reaction rate, and solubility which are intricately linked and together influence the parameters of crystal growth and hence crystal morphol© XXXX American Chemical Society
Received: March 27, 2014 Revised: June 20, 2014
A
dx.doi.org/10.1021/cm501096p | Chem. Mater. XXXX, XXX, XXX−XXX
Chemistry of Materials
Article
Zn(X)2 ↔ Zn 2 + + X2 −(X = CH3COO or NO3)
thermodynamic and kinetic effects as well as the formation of defects which can alter the crystal structure.40,41 The anisotropic growth habit of ZnO from aqueous solution at ambient reaction conditions forming elongated hexagonal crystals with the thermodynamically stable wurtzite structure has been widely reported.20,23,42−46 Layered basic zinc salts (LBZs) have been identified as intermediate products formed during solution synthesis of ZnO and are in themselves a material of great interest.47−54 The structure of LBZs is thought to be brucite-like and is attractive for possible applications such as anion exchange, catalysis, absorption, intercalation, and drug delivery.55−58 LBZs can also be used as a confined twodimensional reaction space and as precursors for the formation of unusual and unfavored ZnO structures such as mesoporous two-dimensional structures.56,59 In this contribution we report a detailed study on the synthesis of ZnO in solution via intermediates, layered basic zinc salts (LBZs). An understanding of the influence of synthesis conditions on product formation, dissolution/ reprecipitation processes and phase transformation is demonstrated by a comparison of two different growth routes both known to result in the formation of wurtzite-structured, twinned hexagonal rods of ZnO. The outcomes of our investigation additionally adds insight into previous studies and gives a clearer understanding of the mechanism of ZnO formation.
■
(b) hydrolysis of base
ZnAc 2−NH3 system: NH3 + H 2O ↔ NH4 + + OH−
Zn(NO3)2 ·6H 2O−HMTA system: C6H12N4 + 10H 2O ↔ 6CH 2O + 4NH4 + + 4OH− (c) precipitation of Zn2+ ions
Zn 2 + + 2OH− → ZnO + H 2O During the course of the reactions, samples were taken from the reaction vessels at selected times (±2 min). Typically, aliquots of 1 mL were sampled for the determination of [Zn2+] using ICP-OES, PerkinElmer Optima 2100DV. Larger volumes were sampled for precipitate characterization. Samples were collected in triplicate and centrifuged at 13000 rpm for 3 min to separate precipitates from the supernatants. The supernatants obtained were recentrifuged using the same conditions. The precipitates were washed three times with ddH2O. Cleaned precipitates were lyophilized at 70 °C using a Virtis110 freeze-dryer. Characterization of Lyophilized Precipitates. The crystallinity of the precipitates obtained was characterized using XRD (PANalytical X’Pert PRO, Cu Kα radiation with wavelength of 1.54056 Å). Ground samples (if necessary) were packed into an aluminum sample holder and scanned from 5° to 90° of 2θ at an accelerating voltage of 45 kV, 40 mA filament current, using a scan speed of 0.02° s−1 at room temperature. Diffraction patterns were analyzed using X’PertHighScore Plus (Version 2.0a) program for diffractogram manipulation, background determination, and peak identification. The mean crystallite size (grain size) ratio of c- and a-axis i.e. (0001)/(101̅0) planes was estimated using the Scherrer equation: d = 0.9λ/(β·cos θ), where d, λ, β, and θ are the grain size, X-ray wavelength, the full-widthat-half-maximum (fwhm), and the Bragg angle, respectively. The morphology and size of selected precipitates were studied using SEM (JEOL JSM840A, 20 kV, gold sputter-coated samples). For ZnO crystal size analysis, more than 30 crystals were studied for each sample and the aspect ratio (L/D = length/diameter average) of each individual crystal was determined using the Java-based image processing program (ImageJ software). FTIR (PerkinElmer Spectrum 100 FTIR Spectrometer with Diamond/KRS5 crystal and Nicolet Magna IR750) was used to detect the functional groups present in the lyophilized precipitates. Spectra were averaged from 32 scans at 2 cm−1 resolution with air as background. The amount of non ZnO component in the precipitates was determined by TGA (Mettler Toledo TGA/SDTA 851e) where samples were heated at 10° min−1 from 30 to 900 °C in air to ensure complete combustion of all organic material. The chemical constitution of the precipitates was investigated by XPS using a VG Scientific ESCALab MkII Xray photoelectron spectrometer with Al Kα Xray source (hν = 1483.6 eV). Synthesized precipitates as well as control samples (purchased ZnO (99.999%) and ZnAc2 (99.99%)) were ground and then mounted on standard sample holders. Survey spectra were collected covering the full binding energy (BE) range from 0−1200 eV using a step size of 1 eV and pass energy (PE) of 50 eV. To compensate for surface charging effects in the insulating samples, all binding energies were corrected with reference to the saturated hydrocarbon C 1s peak at 285.0 eV. High-resolution core level spectra of the Zn 2p, C 1s, and O 1s peaks were collected using a PE of 20 eV and step size of 0.2 eV, which were subsequently deconvoluted and fitted using standard mixed Gaussian−Lorentzian components using CASAXPS software.
EXPERIMENTAL SECTION
Materials. Zinc acetate (Zn(CH3COO)2), zinc nitrate hexahydrate (Zn(NO3)2·6H2O), 1,3-hexamethylenetetramine (HMTA, C6H12N4), 1 M hydrochloride acid (HCl), 1 M potassium hydroxide volumetric standard (KOH), and zinc atomic absorption standard solution (1000 ppm) were purchased from Sigma-Aldrich; ammonia solution (NH3, 35%) was from Fisher Scientific. All chemicals were used without further treatment. Distilled−deionized water (ddH2O) with conductivity less than 1 μS cm−1 (25 °C) was used as the solvent for all ZnO synthesis reactions. ZnO Synthesis and Kinetics Studies. ZnO crystals were synthesized in aqueous solution using precursors Zn(CH3COO)2 and Zn(NO3)2·6H2O following previously described hydrothermal synthesis methods in which twinned hexagonal rods were formed. In the first method, similarly described by Masuda and co-workers,60 30 mM of Zn(CH3COO)2 (hereafter labeled as ZnAc2) in 50 °C water and 30 mM of NH3 were prepared as stock solutions. For accurate preparation of the required NH3 stock concentration, the concentration of a 5% NH3 solution used for stock solution preparation was determined by titration prior to each experiment. For all reactions, the 30 mM NH3 stock solution at room temperature was vigorously stirred into an equal volume of 30 mM ZnAc2 solution (at 50 °C). The pH of the reaction solutions was 7.0 ± 0.1. The mixtures were placed in a water-bath set at 50 °C for up to 168 h. ZnO crystals were also synthesized as previously described by Tomczak et al.20 and Liang et al.:61 100 mM stock solutions of precursor Zn(NO3)2·6H2O and HMTA were prepared and mixed in equal volumes in glass vials while stirring vigorously using a magnetic stirrer. The pH of each solution was 6.9 ± 0.1. Solutions were incubated at 20 °C for 24 h (±2 min) then transferred to a water-bath set at a temperature of 65 °C for an additional 48 h. The precursor and base concentration used in the ZnAc2−NH3 system (15 mM) was lower than the concentration used in the Zn(NO3)2·6H2O−HMTA system (50 mM); however, these concentrations were necessary to maintain neutral pH where the formation of elongated hexagonal ZnO rods occurs. Equal molar ratios between zinc ions and hydroxyl ions were achieved in both systems as shown in the following equations: (a) dissociation of precursor
■
RESULTS AND DISCUSSION ZnO Formation in the ZnAc2NH3 System via LBZA as the Intermediate Compound. The reaction between ZnAc2 and NH3 at pH 7.0 ± 0.1 and 50 °C produced a layered structure with irregular shapes at the early stages of the reaction (Figure 1a). When the reaction was prolonged, a mixture of the B
dx.doi.org/10.1021/cm501096p | Chem. Mater. XXXX, XXX, XXX−XXX
Chemistry of Materials
Article
Figure 1. SEM of precipitates collected at (a) immediately at the start of synthesis, (b) 7 h, (c) 24 h and (d) 48 h. Insets are highermagnification images.
layered structures and hexagonal rods was observed (Figure 1b). With aging, the hexagonal rods became the prominent structure (Figure 1c,d) and grew anisotropically as demonstrated by the increase of aspect ratio (L/D) from ∼4 at 24 h to ∼7 at 72 h. The amount of Zn2+ consumed with time was studied using ICP-OES and the percentage of Zn2+ remaining in solution (with respect to the initial amount) was calculated. As implied by the profile shown in Figure 2, % of Zn2+ was reduced to a minimum of ∼35% after 2 h of reaction followed by a gradual increase of % of Zn2+ to >40% (after 16 h), eventually stabilizing at ∼43% after 24 h.
Figure 3. (a) XRD characterization of precipitates collected at different reaction times; the four-digit and three-digit Miller indices correspond to the crystal planes of ZnO and LBZA, respectively, while the diffraction peak from the aluminum sample holder is marked with an asterisk, (b) a schematic representation of the crystal planes of ZnO hexagonal rods with the wurtzite structure and (c) the region from 32° to 35° of 2θ has been highlighted where peaks found confirm the formation of intermediates. Purchased ZnO is designated with a hash.
diffraction peaks of hexagonal wurtzite structure of ZnO (JCPDS card No. 36-1451). The characteristic diffraction peaks of ZnO at 31.98°, 34.63°, and 36.48° were assigned to the (101̅0), (0002), and (101̅1) planes, respectively.23,50,64 A broad peak at ∼33.2° of 2θ was also observed in the 1-h precipitate (Figure 3c), which corresponds to the (100) diffraction in LBZA.56,65 These results confirmed that ZnO formation in this ZnAc2−NH3 reaction occurred via the formation of an intermediate compound as follows: Zn(Ac)2 → LBZA → ZnO. Interestingly as the reaction progressed, there was a shift toward low angle as a peak was observed at ∼32.9° of 2θ in the 24- and 48-h precipitates (Figure 3c). This new peak could be attributed to the formation of an additional intermediate phase with intercalated carbonate anions54,66 besides the acetate anions and/or Zn(OH)2 species which can be present at the later stages of ZnO formation.53,54,66−68 FTIR analysis gave further evidence corroborating the formation of a zinc hydroxyl double salt and is discussed in detail below. FTIR peaks identified from precipitates are shown in Table 1. In agreement with XRD results, a band arising from the ZnO stretching mode in the 430−530 cm−1 region69−71 was observed in FTIR spectra for all precipitates collected (Figure 4) but only became prominent for precipitates collected at 16 h and beyond. The presence of acetate (COO−) from LBZA was confirmed (Figure 4i, iii, iv,and vi) by peaks arising from the asymmetric (νas) and symmetric stretching (νs) of COO−
Figure 2. Percentage of Zn2+ in solution with respect to the initial concentration of Zn2+ (15 mM) as a function of reaction time.
XRD was used to characterize the crystallinity of the precipitates. As shown in Figure 3a, phase transformation was observed over time. During the early stages (up to 7 h), the main peak at 6.864 ± 0.025° was attributed to the (001) plane of layered basic zinc acetate (LBZA),48,50,58,62 with formula Zn5(OH)8(CH3COO)2·2H2O and an interlayer spacing of 12.87 ± 0.05 Å which is in agreement with reported values from ∼12 to ∼15 Å.50,58,62,63 ZnO eventually became the major component at 16 h and beyond as evidenced by the dominant C
dx.doi.org/10.1021/cm501096p | Chem. Mater. XXXX, XXX, XXX−XXX
Chemistry of Materials
Article
characteristic of the carbonate group (CO32−) were observed in the samples (Figure 4 ii, v, and vii), corresponding to the ν3 (CO 3 2−) asymmetric stretching (ν as) and ν 1 (CO 3 2−) symmetric stretching (νs) of CO32− (1507 and 1045 cm−1 respectively)56,66 and out-of-plane deformation of CO32− (831 cm−1).56,66 The ν3 (CO32−) is split into two peaks at 1507 and 1391 cm−1. The CO32− species in the samples are believed to originate from the dissolved carbon dioxide (CO2) present in the solution66 and are incorporated in the crystalline mineral forming a layered basic zinc hydroxyl double salt containing acetate and carbonate hereafter referred to as layered basic zinc acetate-carbonate (LBZAC) with formula Zn5(OH)x(CH3COO)y(CO3)z·nH2O (x+y+z = 10).54 The increase of the CO32− peak was directly proportional to the decrease of the COO− peak, clearly observed in the spectrum of the 48-h precipitate (Figure 4b). It is however not clear whether LBZAC was an intermediate product in the formation of ZnO or only a byproduct formed as a consequence of the reaction conditions. Nevertheless, XRD (Figure 3c) and FT-IR analysis (Figure 4b) suggests that LBZAC was present when dissolution of some of the solid phase initially formed was observed using ICP-OES analysis. The remaining undissolved solid phase (LBZA and LBZAC) could undergo phase transformation into ZnO. An additional peak at 1620 cm−1 was also observed in aged precipitates (i.e., 48-h precipitate) which was attributed to the O−H vibrational mode of the hydroxide slab in LBZA53,62 and/or O−H bending mode of water.56 The peak at 831 cm−1 previously attributed to CO32− species can also be associated with Zn(OH)2.73 However, this peak became prominent with the appearance of the other CO32− peaks, suggesting that it is more likely to be associated with CO32− species rather than Zn(OH)2. Thus, crystal growth evolves as Zn(Ac)2 → LBZA → ZnO with LBZAC being a plausible additional intermediate. From the FTIR results, the coordination mode of COO− ions to Zn2+ ions was deduced from the difference in frequency between νas(COO−) and νs(COO−), Δνas‑s.72,74 The Δνas‑s value in precipitates collected at up to 24 h was ∼163 cm−1 which implied that the COO− present in the precipitates has coordinated to Zn2+ either forming an ionic metal complex72 or as a bridging ligand.74 Therefore, the COO− detected by FTIR did not arise from free ions, corroborating the claim that LBZA (contains zinc hydroxide layers held together by acetate ions and intercalated water molecules) is the intermediate compound for ZnO formation in this reaction. The coordination mode of CO32− ions to Zn2+ ions was determined from the difference between ν3(CO32−) peaks, Δν3.66 The Δν3 value in precipitates with carbonate present (>4 h) was ∼115 cm−1 which is associated with unidentate coordination to Zn2+ ions.66 XPS results confirm the incorporation of Zn, O, and C in all the precipitates studied by detailed analysis of the Zn 2p3/2, O 1s, and C 1s peaks, at 1021.6 ± 0.1 eV, 531.6 ± 0.1 eV, 285.0 ± 0.0 eV, respectively; cf. the survey scan of the 48-h precipitate (Figure 5a). The full width half-maximum (fwhm) and peak position values are shown in Table S2 of SI. A corresponding reduction in atomic percentage of C with time was observed in the precipitates (Figure 5b). This reduction was due to the progressive transformation of intermediate compounds during the formation of ZnO. These results also include traces of the ubiquitous carbon contamination (C 1s peak), the so-called “adventitious carbon” present even on the cleanest and purest of samples. Practical detection limits for most elements are of
Table 1. FTIR Absorption Peaks Identified in Precipitates Collected at Different Reaction Times peak
wavenumber
i ii iii
∼1550 cm−1 ∼1505 cm−1 ∼1390 cm−1
iv v vi vii
∼1340 cm−1 ∼1045 cm−1 ∼1020 cm−1 ∼830 cm−1
viii ix
600−800 cm−1 430−530 cm−1
assignment antisymmetric stretching (νas) of CO2− antisymmetric stretching (νas) of CO32− symmetric stretching (νs) of CO2− and/or ν3 (CO32−) in-plane bending (δs) or deformation of CH3 symmetric stretching (νs) of CO32− CH3 rocking (ρr) out-of-plane deformation of CO32− and/or Zn(OH)2 vibration mode out-of-plane bending (π) of CH or COO O−Zn−O
Figure 4. (a) FTIR of precipitates collected at different reaction times and ZnO purchased from Sigma-Aldrich and (b) the region from 1800 to 1300 cm−1 has been highlighted where peaks found confirm the formation of intermediates. The identification of the peaks (i−ix) is given in Table 1. Purchased ZnO is designated with a hash (#).
(1555 and 1392 cm−1 respectively),51,62,66,72 −CH3 bending (1338 cm−1)66 and CH3 rocking (1019 cm−1).66 A broad absorption peak in the 3000−3600 cm−1 region assignable to the O−H group of hydroxide or intercalated water in LBZA was also observed (Figure S1 in Supporting Information [SI]). Although LBZA was not detected by XRD from precipitates collected at 16 h and beyond, FTIR has shown the presence of trace amounts of COO− in these precipitates though the major component was ZnO. With progression of the reaction, additional products were identified. A series of peaks D
dx.doi.org/10.1021/cm501096p | Chem. Mater. XXXX, XXX, XXX−XXX
Chemistry of Materials
Article
Figure 5. (a) XPS survey spectrum recorded from an as-deposited 48h precipitate and (b) from XPS data: atomic % of carbon and zinc in the precursor (ZnAc2), 1- and 48-h precipitates.
the order of 0.5−1.0 atomic %. Hence, a sample of purchased 99.999% ZnO was analyzed by XPS in order to show that the C 1s peak detected in the 48-h precipitate was not just from trace contamination during the analysis but was also from LBZs in the sample (Table 2). The relative ratio of Zn/O in the samples was calculated, showing a decrease when LBZs was present in the samples (Table 2). Table 2. Atomic Percentage of C 1s, Zn 2p3/2, and O 1s Peaks Obtained from XPS and the Relative Ratio of Zn/O sample
%C
% Zn
%O
Zn/O
precursor (ZnAc2) 1-h precipitate 48-h precipitate ZnO#
43.9 21.7 18.9 13.0
17.7 32.0 41.0 44.5
38.4 46.3 47.6 42.4
0.5 0.7 0.9 1.0
The measured Zn 2p3/2 region of the 1-h precipitate (Figure 6(i)) was fitted with three components, with BE values of 1021.5 eV, 1022.6, and 1019.9 eV whereas for the 48-h precipitate only two peaks were required, with BE values of 1021.8 and 1022.5 eV (Figure 6(ii)). The BE value of 1021.7 ± 0.2 eV is very close to the corresponding reference value of 1021.9 ± 0.1 eV for crystalline ZnO75 and the value found for the purchased ZnO (99.999%) as shown in Figure S3 of SI). The different local chemical states of the Zn ions in the precipitates indicate the presence of Zn ions in a mixed environment, as is the case for LBZA/LBZAC. The presence of a higher binding energy component, such as the BE value of 1022.6 ± 0.1 eV, can be attributed either to acetate Zn groups (Zn−OCOCH3) or hydrated Zn (Zn−OH).76 The study of the Zn 2p3/2 region for ZnAc2 (Figure S4 of SI) indicates this higher BE value corresponds to carboxyl groups present in
Figure 6. Core-level spectra of as-deposited precipitates collected at 1 and 48 h of reaction for (a) Zn 2p3/2, (b) O 1s and (c) C 1s.
LBZA/LBZAC. The other lower binding energy component at 1019.9 eV for the 1-h precipitate is in accordance with the corresponding BE values (1019.7 ± 0.6 eV) of LBZs with a general formula [(Znocta)3(Zntetra)2(OH)8]2+·2(A−)·nH2O.47 The absence of the peak at 1019.7 ± 0.6 eV in the 48-h precipitate indicates that LBZs were consumed and most of the E
dx.doi.org/10.1021/cm501096p | Chem. Mater. XXXX, XXX, XXX−XXX
Chemistry of Materials
Article
the 48-h precipitate had the major weight loss within the second region (168 °C−250 °C). Total percentage weight loss curves of precipitates collected at different reaction times are shown in Figure S5 of SI. The weight loss in precipitates formed and collected immediately after mixing reagents (
