Article pubs.acs.org/Langmuir
Differential Etching of ZnO Native Planes under Basic Conditions Nathan Johann Nicholas,† William Ducker,‡ and George V. Franks*,† †
Chemical and Biomolecular Engineering, University of Melbourne, Parkville, Victoria, Australia Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24060, United States
‡
S Supporting Information *
ABSTRACT: The in situ dissolution of polished (0001), (101̅0), and (0001̅) surfaces of ZnO was studied using Atomic Force Microscopy under alkaline conditions. In aqueous NaOH solution the (0001) plane forms a stepped surface whereas the (0001)̅ plane converts into more stable {1011̅ }̅ planes. Dissolution of the (101̅0) plane leaves a combination of (0001) and (101̅1̅) planes. Dissolution in solutions containing both NaOH and Na3citrate causes the (0001) plane steps to increase in number and reduce in height, and cause an overall increase in the rate of dissolution in the ⟨101̅0⟩ directions. These observations are explained using a mechanism based on edgewise dissolution where the etching rate depends on the number of surface oxygen atoms per zinc atom. Large areas of single index faces (over 50 μm2) of (0001) and (0001̅), suitable for surface chemistry studies, were also generated by chemical dissolution.
1. INTRODUCTION In the past decade or so there has been renewed interest in ZnO due to its potential use in many applications,1 such as LEDs,2−4 solar cells,5 catalysis,6 gas sensors,7 and piezoelectric devices.8 Research has focused on tailoring the shape of ZnO particles and surfaces for each application, using the “bottom up” method of crystal growth. A great number of different crystal shapes have been formed,9 ranging from rods,10,11 platelets,12 wires,13,14 or tetrahedrons,14 to more complex shapes like “urchins”,15 “flowers”,16 and even “pine trees”.17 One way this control of the ZnO crystal shape can be achieved is by introducing small molecules into the growth process. Comparatively little attention has been given, however, to the so-called “top down” approach of ZnO etching. This is surprising as etching not only has the potential to create shapes that are difficult to achieve via growth methods (such as the formation of ZnO nanotubes18,19), but also can produce surfaces of ZnO that are not normally expressed during growth.20 Previous work on the wet etching of ZnO focused on how the different surfaces respond to acidic20−25 conditions. The majority of these studies used ex situ techniques, which give good insight into products of etching, but are limited in their ability to show the etching mechanism. One exception to this is the work by Valtiner et al.,26,27 who have performed in depth studies of the in situ etching of single crystal ZnO under both acidic27 and basic26 conditions. They showed that under alkaline conditions the (0001) plane is stabilized by hydroxide adsorption, and dissolution proceeds at crystal step edges. There have been only a few studies on the effect of small molecules on the ZnO dissolution process, and so far these have only been studied ex situ.20,24,28 © 2012 American Chemical Society
The ability of ZnO to form the many different shapes observed in the literature, and indeed its versatility in many different applications, is due, in large part, to its anisotropic crystal structure. While several polymorphs of ZnO are known to exist, most often ZnO conforms to the Wurtzite structure with alternating hexagonally close packed sheets of zinc and oxygen atoms such that every atom is tetrahedrally coordinated, as shown in Figure 1.
Figure 1. The idealized structure of Wurtzite ZnO showing the Zn terminated (0001), O terminated (0001̅), and the six equivalent faces {101̅0}. Light gray atoms represent zinc, red atoms represent oxygen.
Typically crystals will be comprised of the surfaces that have the lowest free energy. These low energy surfaces are usually low Miller index planes.9 For ZnO the most commonly observed surfaces (when grown in the absence of growth Received: November 30, 2011 Revised: February 29, 2012 Published: March 6, 2012 5633
dx.doi.org/10.1021/la2047273 | Langmuir 2012, 28, 5633−5641
Langmuir
Article
Figure 2. (a−d) Time series of 10 μm by 10 μm AFM scans showing the in situ dissolution of the Zn-terminated basal plane ZnO sample in 2 M NaOH. (e and f) 1 μm by 1 μm AFM images showing the morphology before and after etching. (g) Ex situ SEM confirming hexagonal stepped morphology and the presence of a (101̅1) plane (dotted outline). (h) Height trace of a step from part d (indicated by the dashed red line) showing the incline of the dissolution front. Note: The range of the z-axis for parts a−f varies between different AFM images, and is shown in the top left corner of each image. for fluid recirculation (see Figure S1 in Supporting Information). During an experiment the etching solution (either 2 M NaOH or 2 M NaOH with 100 mM Na3citrate) was circulated between the fluid cell (V ≈ 1.5 mL) and storage reservoir (V = 100 mL) via a peristaltic pump at a rate of 0.25 mL/min (4.2 μL/s) to keep the degree of undersaturation approximately constant during the experiment, while keeping the flow rate low enough so no erosion due to flow occurred. NaOH (2 M) was elected as the etchant concentration as it was observed to etch ZnO at a reasonable rate. At higher concentrations (3 M) the etching rate was too rapid on the length scales studied in this work, particularly for the (0001̅) surface. An Asylum Research MFP3D Atomic Force Microscope using Veeco NP cantilevers was used to capture in situ images of the etching behavior of ZnO. Images were obtained at a scan rate of 1 Hz with 512 scan lines and 512 scan points regardless of the scan size. During imaging the pump was turned off to prevent acoustic vibrations and pressure waves from producing noise in the scan. In the cases of the Zn-terminated basal plane ZnO and the prismatic plane ZnO samples, the etching experiments were conducted such that the samples were initially etched in 2 M NaOH for several hours, then the solution was changed to 2 M NaOH and 0.1 M Na3citrate. For the O-terminated basal plane ZnO, two separate experiments were conducted on virgin samples (for the etching with and without Na3citrate). Additional preliminary investigations into the effects of counterions (using LiOH, NaOH, and CsOH) were also performed. However, it was discovered that while the choice of counterion does affect the etching rate of the different ZnO surfaces (particularly the (0001̅) surface), it does not affect the morphology, and as this is the focus of this paper, these results have been omitted from this work. Ex situ images were obtained under high vacuum with use of an FEI Quanta Scanning Electron Microscope (ESEM) with an acceleron
directing molecules) are the following: the Zn-terminated basal plane, designated as (0001), the O-terminated basal plane at the opposite end of the crystal, designated as (0001̅), and the six chemically equivalent sides of a hexagonal prism, designated as (101̅0) (formally they are denoted as {101̅0}, as (101̅0) is only one of the six equivalent planes). In this work the in situ dissolution of three polished faces of ZnO, (0001), (0001̅), and (101̅0), is studied under basic conditions. The effect of citrate, a known growth modifier,10−12,30,34 on ZnO dissolution is also investigated. Using this differential approach a mechanism of dissolution is developed, including the influence of citrate.
2. EXPERIMENTAL SECTION All chemicals used in this work were AR grade (99%+ purity) and used as received with no further purification. The solvent for all solutions is Millipore purified water (resistivity 18.2 MΩ cm−1). ZnO single crystals (10 mm × 10 mm × 0.5 mm SPC Goodwill) were obtained cut and polished perpendicular to the [0001], [0001̅], or [101̅0] directions. In this work we refer to the sample nominally normal to the [0001] direction as the Zn-terminated basal plane ZnO. The sample nominally normal to the [0001̅] direction is referred to as the O-terminated basal plane ZnO and the sample nominally normal to the [101̅0] as the prismatic plane ZnO. The Miller Indices (0001), (0001̅), and (101̅0) will be used to refer to the actual crystalline planar surfaces. The crystals were then cut into approximately 3 × 3 mm sections and mounted on a gold-coated glass disk that was loaded into an Asylum Research Atomic Force Microscope Bioheater Fluid Cell setup 5634
dx.doi.org/10.1021/la2047273 | Langmuir 2012, 28, 5633−5641
Langmuir
Article
voltage 5 kV at a working distance of 9.7 mm and a spot size of 2.0 nm. Samples were mounted on conductive tape with no pretreatment.
artifact caused by tip convolution. As the leading face of the AFM tip is inclined at an angle of about 15° relative to the [0001] direction, the tip convolution will lead to images with the (101̅0) plane at an angle of 105° relative to the (0001) plane. However, Figure 2h shows that the angle is approximately 150°, making the inclined plane closer to a (101̅1) plane (which would have an ideal inclination of 122°). The presence of the (101̅1) plane is also confirmed by the ex situ SEM (region enclosed by dots in Figure 2g). The difference between the observed and theoretical incline can be attributed to the fact that the system is dissolving in real time and the AFM scans are not instantaneous. As the step is etching while the AFM scan is being taken it will cause the dissolution front to distort, appearing to be more spread out (i.e., closer to horizontal). This will further be exaggerated by the speed at which the step is dissolving and it was observed that steps that were lesser in height propagated faster and showed a degree of inclination even further removed from the ideal for a (101̅1) plane. 3.2. Etching the Zn-Terminated Basal Plane ZnO with 2 M NaOH and 0.1 M Na3citrate. During crystal growth, citrate promotes growth in the ⟨101̅0⟩ directions and therefore leads to formation of crystals containing large areas of (0001) and (0001̅) planes. Here we use in situ AFM imaging to examine the effect of citrate on the dissolution of ZnO. The sample was initially etched with 2 M NaOH solution (without citrate present) to remove the defect-rich surface layers that etch quickly and are particular to each sample. Figure 3 shows a
3. RESULTS 3.1. Etching the Zn-Terminated Basal Plane ZnO with 2 M NaOH. The (0001) polar plane is the most characterized native face of ZnO due to it being the primary growth direction for ZnO, especially during epitaxial growth. Figure 2 shows the time evolution of the nominally Zn-terminated basal plane ZnO during etching. AFM images on a 10-μm scale of the Znterminated basal plane ZnO before etching (Figure 2a) show that the surface is quite flat (2.8 nm rms roughness over 100 μm2) and is characterized by a series of thin scratches that arise from the manufacturer’s polishing process. At higher magnifications (Figure 2e) it is obvious that the surface does not consist of molecularly smooth planes, instead it is comprised of irregular features on the order of 3 nm high and 100 nm in diameter. For convenience, we refer to this sample as Zn-terminated basal plane ZnO, but it is clear that there are no large areas of (0001) on the sample. Once the Zn-terminated basal plane ZnO is exposed to 2 M NaOH two different types of dissolution occur. Initially dissolution pits form in the [0001̅] direction at certain locations on the plane. It has been previously reported that dislocations occur primarily in the [0001] direction for hydrothermally grown ZnO.29,30 As dissolution occurs preferentially at defect sites, it is concluded that a dislocation is present at the point of origin of a dissolution pit. This explains why dissolution in the [0001̅] direction only is seen to occur at specific points, and why these points often occur in lines corresponding to the deeper scratches initially present on the surface. As the etching in the [0001]̅ direction occurs, material is also removed along the ⟨101̅0⟩ directions parallel to the (0001) plane surface. The newly created surfaces propagate away from the dislocation sites forming stepped, hexagonal pits as seen in Figure 2g. These pits continue to increase in size, expanding away from the dislocation in the ⟨101̅0⟩ directions and are the main method by which the ZnO crystal dissolves. This stepped etching morphology was also observed by Valtiner et al.26 for etching of ZnO under 3 M NaOH. Thus, under highly alkaline conditions, it would appear that the etching morphology is not dependent upon NaOH concentration. As dissolution continues, the steps become less frequent but increase in height. The steps seen in the AFM images (Figure 2, parts a−f) are on one side of a hexagonal dissolution pit, and there are similar steps going in five other directions, as is evident in Figure 2g. Under etching conditions, the basal plane becomes rougher (i.e., there is greater variation in height across the sample) over larger areas (100 μm2) but locally, i.e., the area between the steps, is very smooth. The continuing presence of the (0001) plane during dissolution and the observation that dissolution into the Zn-terminated basal plane ZnO only occurs at dislocation sites implies that the (0001) plane itself is stable (i.e., has a low surface energy) in the highly alkaline solution. The orientation of the new surface formed by the appearance of the steps (i.e., the surface of the dissolution front) can be approximated by studying the AFM cross section presented in Figure 2h. The cross section shows that the orientation of the dissolution front is not perpendicular to the (0001) plane and so is not purely (101̅0) plane, but rather it is at an incline, making it a higher index combination of (0001) and (101̅0) planes. This inclination cannot be attributed to an imaging
Figure 3. Effect of citrate ions on the Zn-terminated basal plane ZnO. (a−c) Time series of 10 μm by 10 μm AFM scans in the presence of 2 M NaOH and 0.1 M Na3citrate. (d) Step profiles over 4 μm showing how citrate increases the step density and reduces the average step height. Note that the sample designated as t = 0 has previously been etched in 2 M NaOH (without citrate) for 4 h.
time series of AFM images of a Zn-terminated basal plane ZnO sample as it dissolves in the presence of citrate. From Figure 3 it can be seen that the stepped structure is maintained, but the average step height decreases and there is a commensurate decrease in the spaces between the steps, as highlighted by the comparative cross sections presented in Figure 3d. 3.3. Etching of the O-Terminated Basal Plane ZnO with 2 M NaOH. The next ZnO face to be studied was the Oterminated basal plane ZnO, which is nominally normal to the [0001̅] direction. Figure 4 shows the time evolution of the O5635
dx.doi.org/10.1021/la2047273 | Langmuir 2012, 28, 5633−5641
Langmuir
Article
3.4. Etching of O-Terminated Basal Plane ZnO with 2 M NaOH and 0.1 M Na3citrate. Unlike the other samples the O-terminated basal plane ZnO sample was etched in a solution of 2 M NaOH and 0.1 M Na3citrate, without pre-etching in a purely 2 M NaOH solution. This was done for two reasons. First the rate of etching of (0001̅) ZnO was seen to occur more rapidly than for the other faces. Second the etching process causes the (0001̅) plane to transform into a different set of planes. Consequently, if the sample was pre-etched there would be very little (0001̅) plane present, with only a limited time frame to study the influence of citrate on (0001̅) plane dissolution. When the O-terminated basal plane ZnO sample was etched in 2 M NaOH and 0.1 M Na3citrate there was no observable difference in morphology or etching rate when compared to the sample etched in 2 M NaOH only. Therefore it is concluded that Na3citrate has no influence on the etching of (0001̅) ZnO. 3.5. Etching of Prismatic Plane ZnO with 2 M NaOH. Etching of the (0001) and (0001̅) faces occurs primarily in the ⟨101̅0⟩ directions. Thus more insight into the dissolution process can be obtained by directly observing the dissolution of the (1010̅ ) face. Another reason for interest in etching of this face is that {101̅0} surfaces are the primary surfaces produced during hydrothermal growth. Therefore understanding how the {101̅0} dissolve is key to understanding how ZnO crystals will dissolve in general. The as received prismatic plane ZnO appears similar to the as received (0001) plane shown in Figure 2a,e. The time progression of etching of the prismatic plane ZnO in 2 M NaOH is shown in Figure 5. From Figure 5a it is apparent that the prismatic plane ZnO primarily reconstructs into trenches. These trenches appear to be running in the [112̅0] direction with one side being the (0001) plane and the other the higher index plane observed during the dissolution of the (0001̅) surface. This is highlighted by the corrugated nature of what was formerly the (101̅0) plane (shown in greater detail in Figure 5g). The planes forming the sides of these trenches are stable across most of the surface (except where a dislocation is believed to be present) as no further dissolution in either the [0001] or [0001̅] directions is seen to occur (as highlighted by the invariance with time of the trench marked by the asterisk in Figure 5a and cross sections along the [0001] direction presented as Figure 5i). However, at dislocation sites (assumed to be parallel to the [0001] direction) dissolution does continue to occur, as evident by the large pits formed. At these sites it can be seen that the dissolution occurs initially in the [0001]̅ direction, with the newly exposed (0001) plane forming the stepped structure that also appears to propagate away from the dislocation in the ±[112̅0] directions. However, ex situ SEM imaging (Figure 5h) shows that the dissolution front on one side of the dislocation is comprised of (101̅1) and (011̅1) planes, marked by a white bar and black dot, respectively (the other side is comprised of (101̅1) and (01̅11) planes). Thus the apparent propagation in the ±[112̅0] directions is merely the resultant combination of two different {1011̅ } dissolution fronts on each side of the dislocation. This is further heightened by the fact that dissolution in the [1̅010] direction (into the crystal) is limited by the stable high index plane that is formed opposite the (0001) plane, marked by a white triangle in Figure 5h. The SEM images also show the presence of the inclined (101̅X) surfaces near the dislocation (one of which is marked by a black dot).
Figure 4. The dissolution of the O-terminated basal plane ZnO in 2 M NaOH as a function of time. (a) 5 μm by 5 μm image of a typical Oterminated basal plane ZnO surface before etching. (b−d) 20 μm by 20 μm scans showing the evolution of the stable semipolar planes. (e) 40 μm by 40 μm scan showing the (0001̅) plane has almost completely reconstructed after 3 h. The 5 μm height scale bar applies to images b−e. (f) Ex situ SEM showing the formation of hexagonal dissolution pits across the entire ZnO surface.
terminated basal plane ZnO as it etches in the presence of 2 M NaOH. The first thing to note is that the polished sample is locally rough and there are no large planes of (0001̅) so the term O-terminated basal plane is used to designate the sample, despite the lack of (0001̅) initially present. After a short etching time (less than 1 h) the sample is predominantly comprised of smooth (0001̅) planes, with the occasional small hexagonal pyramidal pit. As time progresses the small hexagonal pyramidal pits grow rapidly, thereby reducing the amount of (0001̅) plane present, making the previously (0001̅) surface very rough on the 10-μm scale after about 3 h. At this time most of the sample terminates in higher index planes which make up the sides of the pits (see Figure 4, parts e and f). The presence of these pits suggests that these higher index planes are more stable than the (0001̅) face in basic solution. The sides of the pits appear to be quite smooth, with small ridges appearing from the incorporation of the shallow dissolution fronts that rapidly spread across the surface in the early stages of dissolution. The Miller indices of these newly formed faces are not able to be determined at this crystal orientation because the linear dissolution rate of the pits is again similar to the scan speed. Each image takes about 8 min to capture, so the dimensions of the pits change during this time, making them appear distorted. 5636
dx.doi.org/10.1021/la2047273 | Langmuir 2012, 28, 5633−5641
Langmuir
Article
Figure 5. Time series of 50 μm by 50 μm AFM scans showing the in situ dissolution of the prismatic plane ZnO. (a−c) Dissolution in 2 M NaOH (a = 4 h, b = 5 h, c = 6 h). An example of the two most stable faces, (0001) and (101̅1̅), have been labeled on part b. (d−f) After 6 h the etching solution was changed to 2 M NaOH and 0.1 M Na3citrate (d = 1 h after addition of citrate, f = 2 h, g = 3 h). Dashed line on part a indicates the location of the height traces measured in the [0001] direction presented in part i. (g) 5 μm enlargement of the reconstructed (101̅0) surface (rotated 90°). (h) Ex situ SEM image showing the dissolution habit in the pit. The (101̅1̅) plane is marked by a white triangle, the (101̅1) by a white bar, and the (011̅1) by a black dot (i) Time series of cross sections through the dashed line in part a (curves for parts a and f have been labeled). Black lines show the cross section with 2 M NaOH only (a−c) and light gray show lines the cross sections for dissolution in 2 M NaOH and 0.1 M citrate (d− f).
The high index plane (the surface of the etch pit opposite the (0001) planes, indicated by the white triangle in Figure 5h) is not only stable but also stationary (i.e., no dissolution parallel to the plane is occurring). As a result the orientation of this plane can be determined from the height trace without interference from the surface etching during the image scan. From the AFM profile it was determined that the orientation of this high index plane is close to (101̅1̅), when errors introduced by the tip width, thermal drift, and sample tilt were accounted for. 3.6. Etching of the Prismatic Plane ZnO with 2 M NaOH and 0.1 M Na3citrate. When the prismatic plane ZnO (which has been pre-etched with 2 M NaOH) is etched in 2 M NaOH and 0.1 M Na3citrate the results are consistent with observations made during the (0001) and (0001̅) plane dissolution experiments. The (101̅1̅) plane remains unaffected by the presence of citrate and the (0001) plane steps become
smaller and more frequent. This results in a subtle change in the morphology of the etch pits, which will be discussed in the next section.
4. DISCUSSION 4.1. Estimating the Relative Dissolution Rates in the [0001̅] and [101̅0] Directions. The observations of the dissolution pits on the prismatic plane ZnO can be used to estimate the relative dissolution rates in the two principal dissolution directions (along the dislocation and away from the dislocation). The rate of dissolution in each of these directions can be quantified unambiguously for each pit as they can be related to a fixed position on the sample (the stationary (101̅1̅) plane). The relative dissolution rates can be determined from the angle formed by the dissolution front at the top surface of the crystal. This has the advantage of not being affected by the size of the dissolution pit or by the merging of different 5637
dx.doi.org/10.1021/la2047273 | Langmuir 2012, 28, 5633−5641
Langmuir
Article
sections of the same age is constant. Because the angle formed at the apex of the triangle is constant as the front advances along the dislocation (in the [0001̅] direction), the etching rate away from the dislocation (in the ±[112̅0] directions) also remains constant for any particular age. This means that as the crystal etches along the dislocation, the lateral dissolution rate (in the ±[112̅0] directions) at any specified distance near the dislocation is the same at any given time, but as it propagates away from the point of origin the etching rate slows. This is another advantage of using the angle formed at the dissolution front to determine the relative dissolution rates in the [0001̅] and ±[112̅0] directions, as it averages across steps of different ages, but maintains the same age range with time. As stated in Section 3.5, dissolution in the [1120̅ ] direction occurs by removal of atoms from two different [101̅0] directions. Thus the rate of dissolution in the ⟨101̅0⟩ directions is related to the dissolution rate in the ±[112̅0] directions. When citrate ions are added into the system, the average dissolution angle increases from 153 ± 6° to 170 ± 2°. This indicates either an overall increase in lateral dissolution (in the ±[112̅0] directions) or hindrance of dissolution along the dislocation (in the [0001]̅ direction). However, from the AFM cross sections (Figure 5i between the 18 and 35 μm marking on the x-axis scale) it is apparent that the etching rate along the dislocation remains constant before and after citrate addition. Therefore it must be concluded that the citrate ions increase the overall dissolution rate in the ±[112̅0] direction, and as such the ⟨101̅0⟩ directions. 4.2. The Influence of Surface Oxygens on the Dissolution of ZnO. Having studied the etching of the Znand O-terminated basal plane ZnO samples, as well as the prismatic ZnO surface, some general comments can be made. The (0001) and {101̅1̅} planes are stable as large areas of these faces are maintained throughout the etching process. The {1011̅ } planes are partially stable, forming the dissolution front, and the (0001̅) and (101̅0) planes are unstable, restructuring into other planes. The soluble species under basic conditions investigated in this work are [Zn(OH)4]2−,31 so the final state of the dissolved zinc atom must be tetrahedrally coordinated by four hydroxide ions, rather than by four −O−Zn bonds, as in the bulk. We attribute the different rates of dissolution to the fact that the different surface planes consist of Zn atoms that begin further along the path to full hydroxylation. The crystals we study are immersed in water, so the surface atoms have the opportunity to react with water and hydroxide. The exact surface species are not known (as the surface is likely a combination of oxide, hydroxide, and hydrate species); instead of counting the extent of hydroxylation, we prefer to count the number of bonds to surface oxygens. We define a surface oxygen as any oxygen that is bound to less than four zinc atoms. As the Zn atoms are tetrahedrally coordinated in the solid state, stability could also be considered in terms of how many Zn−O−Zn bonds must be broken to convert the Zn atom to an aqueous state. However, the surface oxygens on the (0001̅) and (101̅0) planes are also nominally bound to other Zn, i.e., they are ideally surface Zn− O−Zn bonds. And given the ambiguity of the ZnO surface structure in water this could lead to confusion. Thus the number of surface oxygens was chosen instead as they are easier to define (as even though the precise state of the surface oxygens is unknown, the number of surface oxygens on each face can be determined per Zn atom). These surface oxygens are labeled in deep red in Figure 7.
dissolution fronts. Two examples of this angle are shown in Figure 6. Table 1 shows the dissolution angle for three different
Figure 6. A 2D projection of Figure 5b reproduced in gray scale, illustrating the dissolution angle represented in Table 1. The dissolution angle gives an indication of how fast a dissolution front is advancing along a defect, relative to how fast it is propagating away.
Table 1. Change in Dissolution Front Angle with Time and Solution Composition etchant 2 M NaOH
2 M NaOH/100 mM citrate
etch pit no.
time (h)
age (h)
angle (deg)
1 1 1 2 2 2 3 3 3 3 3 3 3
2.32 2.97 3.05 3.88 4.93 6.17 3.88 4.93 6.17 7.42 8.47 9.52 10.32
1.15 1.80 1.89 3.03 4.08 5.31 1.15 2.20 3.44 4.69 5.74 6.79 7.59
154 152 150 146 146 154 156 161 162 170 168 172 169
etch pits as a function of time. Each of the etch pits started growing at different times and thus have different “ages”, which are also shown in the table. The age of the pit was obtained by plotting the distance the dissolution front has advanced along the dislocation (i.e., in the [0001̅] direction) with time and extrapolating back to zero distance. This gave the time of origin of a dissolution front which could be subtracted from the experimental time to give the age of a dissolution front. From Table 1 it can be seen that the angle near the apex of the triangle is approximately constant at 153 ± 6° for each etch pit and remains constant over time. Figure 5 shows that the dissolution rate in the ⟨1010̅ ⟩ directions slows as a section of the dissolution pit aged (i.e., as the section etched away from the dislocation over time). This can be attributed to the observation that the average step height increases as a section ages. It should be noted, however, that even though the etching rate decreases as a front ages, the dissolution rate for different 5638
dx.doi.org/10.1021/la2047273 | Langmuir 2012, 28, 5633−5641
Langmuir
Article
Figure 7. The atomic structure for the oxygen stabilized bulk terminated ZnO projected along the [112̅0] direction, showing the surface oxygens for each of the different low index faces described in this work. Gray represents zinc atoms, red oxygen, and dark red surface oxygen. Note as (1010̅ ) is equivalent to (1̅010), (1̅011̅) is equivalent to (101̅1̅).
To liberate the zinc atom, a number of oxygen−zinc bonds must be cleaved. We hypothesize that a lower number of bond cleavages requires less energy and provides more access to the zinc atom, and therefore that complete hydroxylation proceeds more rapidly when a zinc atom is bonded to a great number of surface oxygen atoms. Table 2 shows that stability is inversely
formed. If a zinc tetrahedra at the edge where the (0001) and (101̅0) planes meet (highlighted by the dotted triangle in Figure 7a) is removed, it results in the formation of a (101̅1) plane. As this edge dissolves, the {101̅1} planes would form in preference to the {101̅0} planes (having fewer surface oxygens), and are easier to form than the more stable {101̅1̅} planes (which would require the removal of a zinc tetrahedron from the base of a step, at the junction between the (0001) and (101̅0) planes, i.e., the removal of a zinc tetrahedra from an inner corner). As these {1011̅ } planes continue to etch rapidly in the ⟨101̅1̅⟩ directions, creating a new {101̅1}/(0001) edge, the original step can begin to etch from two sites (instead of one). This would then change the appearance of the plane to a higher index plane, effectively forming the observed {101̅X} planes. We now consider the observation that the step height on the Zn-terminated basal plane ZnO increases as it ages. There are only two ways for a step to increase in size during dissolution. Either etching occurs into the plane at the base of the step, or several smaller steps must combine to form a larger step. As dissolution into the plane was only observed at specif ic sites (attributed to the presence of dislocations) it must be concluded that the general increase in step height (on the (0001) plane) observed must be due to the coalescence of smaller steps. Assuming material is removed from the outer corner of a step (as outlined previously) primarily in the [101̅1̅] direction, it results in the movement of the step in the [101̅0] direction (which was the observed direction of step propagation). As smaller steps are initially formed and are seen to propagate faster, it follows then that the rate at which a step propagates must decrease with step height. If step propagation were independent of height, there would be no change in step height with time. If larger steps propagated faster there would be little chance of step coalescence, as the larger steps are seen to occur at the top of a dissolution pit on the (0001) surface and coalescence could only occur when the top step propagates slower than those below it. This agrees with the hypothesis that dissolution of the (0001) plane is governed by etching rate in the [101̅1̅] direction. As the dissolution front advances in the [1010̅ ] direction, each time a (1011̅ ) layer is removed, the area of (101̅1) (and by extension the step height) will govern the etching rate in the [101̅0] direction. For instance, consider two different steps given in Figure 8, one that is two zinc tetrahedra high (step a) and one that only a single
Table 2. The Number of Exposed Surface Oxygens Per Zinc Tetrahedra for the Different Faces of Oxygen-Stabilized ZnO Observed in This Work ZnO surface
surface O per zn Tetrahedron
O dangling bonds per nm2
stability in 2 M NaOH
(0001) (1011̅ )̅ (101̅1) (0001̅) (101̅0)
1 1.5a 2.5a,b 3 3
32.8 26.0 15.6 10.9 5.9
stable stable semistable unstable unstable
a The (101̅1) and (101̅1̅) planes have two different zinc tetrahedra configurations with a different number of surface oxygens, and so an averaged value is used. b(101̅1) was used as an example of the (101̅X) planes. The higher index surfaces have an average number of surface oxygens that approaches 3.
dependent upon the number of surface oxygen atoms per zinc atom in the outer layer. A similar trend was observed by Palacios-Lidón et al.32 for acid etching of high index planes. However, their analysis focused on the oxygen dangling bonds density as the reason for the observed trend, which does give good agreement with the results observed in this work (as shown in Table 2). While oxygen dangling bonds density is important for dissolution under acidic conditions (as they are likely sites for proton interaction) it cannot explain the same observed trend for dissolution under basic conditions because hydroxide ions are unlikely to bond to a surface oxygen. The availability of surface oxygens to be converted to hydroxide species does not fully explain the appearance of the {101̅X} planes during the dissolution process. As the crystal etches into the (0001) plane along a dislocation, a series of {101̅0} planes is exposed. These {101̅0} planes should reconstruct into (0001) and {1011̅ }̅ and cease propagating in the ⟨101̅0⟩ directions, as was observed for the prismatic plane studies. Therefore these high index {101̅X} planes should not form. If an edge-wise dissolution mechanism is considered, however, it can be seen how these higher index planes are 5639
dx.doi.org/10.1021/la2047273 | Langmuir 2012, 28, 5633−5641
Langmuir
Article
adsorbing specifically onto the (0001) plane making the formation of (0001)/(101̅1) edges more favorable. This increase in the number of steps then explains the increased rate of etching in the ⟨101̅0⟩ directions observed in Figure 5. By splitting a larger step into smaller ones, each smaller step is going to begin to etch independently (as the etching in the [101̅0] direction is governed by the presence of (0001)/(101̅1) edge) and so the overall etching rate in the ⟨101̅0⟩ directions would increase. At the dislocation site, however, it can be seen (Figure 5h) that there is no (0001) plane present, and so the citrate will have no affect on the etching rate near the dislocation. Due to the different spacing of the atoms, the citrate ions are not likely to bind to the {1011̅ }̅ planes in the same manner as they appear to do on the (0001). Thus they are unlikely to affect the dissolution of the {101̅1̅} planes, which is consistent with our observations.
5. CONCLUSIONS From the in situ dissolution studies of the Zn-terminated, Oterminated, and prismatic planes of ZnO the following observations were made: • The dissolution of ZnO begins by rapid etching in the [000 ± 1] directions at specific points on the (0001) and (0001̅) planes. • The (0001) and {101̅1̅} planes are stable under alkaline conditions, with the semistable, high index {101X ̅ } planes forming the main dissolution front. The {101̅0} planes reconstruct into the stable (0001) and {101̅1̅} planes, while the (0001̅) plane reconstructs into {101̅1̅} planes, forming many hexagonal pyramidal pits. • As the stable faces have the least number of surface oxygen atoms per zinc atom, it is concluded that surface stability is dependent upon the ability of the surface zinc atoms to be converted from the solid oxide network to the soluble tetrahydroxide complex. These observations are consistent with the hypothesis that etching proceeds in the {1010̅ } directions via the continual removal of edge atoms in the ⟨101̅1̅⟩ directions. This leads to the rate of etching of a step to be dependent upon the height of a step, with smaller steps etching faster. The addition of citrate molecules results in an increase in the number of steps on the (0001) plane, thereby increasing the number of edges and increasing the net rate of dissolution in the ⟨101̅0⟩ directions. The differential approach to single crystal etching utilized here could easily be used to understand the etching mechanism of other crystal systems.
Figure 8. How step height influences lateral etching rate. Step a is twice the height of step b and so propagates in the [101̅0] direction at half the speed. The numbers represent the order of removal of the Zn tetrahedra and (×) marks the original location of each step.
tetrahedra high (step b). After four units have been removed (in the numbered order), step a will have advanced two units in the [101̅0] direction, while step b will have advanced four units, thus the step that is smaller in height advances faster. This example assumes that the rate of removal of each zinc tetrahedra is the same for each step, which given they are chemically the same and are exposed to the same solution conditions, is not unreasonable. It is easy to see that if step heights were reversed (i.e., step a was 1 unit high and step b 2 units) step b would have merged with step a forming a new step that would be 3 units high. This mechanism is only valid if (1) the step etches at the (0001)/ (101̅1) edge and (2) the etching of a single step occurs primarily in the [101̅1̅] direction. If the first assumption was invalid, and dissolution occurred anywhere on the (1011̅ ) plane, then pitting would be observed on the (101̅1) plane (from Figure 2g and Figure 5h we can see that pitting does not occur) and there would be no dependence of step height on etching rate, which is contrary to the results shown by Figure 5, parts a−c and d−f. If the second assumption was invalid, and the etching of a step occurred primarily in the [101̅0] direction (i.e., in Figure 8 for step a the second and third tetrahedra to be removed were switched) then we would see a reduction in height of small steps with time, which was not observed. Furthermore, the etching of (0001) ZnO steps has also been previously26 shown to occur by shortening the length of a step, rather than decreasing the step height. 4.3. The Effects of Citrate on the Dissolution of ZnO. The results presented in this paper show that the citrate molecules only influence the dissolution of (0001) ZnO. It has been well established that citrate encourages the expression of (0001) ZnO during growth. It is thought that citrate selectively adsorbs to the (0001) plane thereby hindering growth in that direction.10−12,30,33,34 As seen in Figures 3 and 5d−f, citrate ions cause a reduction in average step size. As there is no growth possible in this system (as the solution is undersaturated with respect to ZnO) the reduction in step size must be due to the splitting of a larger step into smaller ones. Therefore we speculate that citrate is
■
ASSOCIATED CONTENT
* Supporting Information S
Further details on the experimental setup and procedure. This material is available free of charge via the Internet at http:// pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was funded in part by the Particulate Fluids and Processing Centre, a Special Research Centre of the Australian 5640
dx.doi.org/10.1021/la2047273 | Langmuir 2012, 28, 5633−5641
Langmuir
Article
Chemical Etching. Thin Solid Films WOE-11, 2004 2005, 486 (1−2), 42−45. (21) Zhu, J.; Emanetoglu, N. W.; Chen, Y.; Yakshinskiy, B. V.; Lu, Y. Wet-Chemical Etching of (11(2)0) ZnO Films. J. Electron. Mater. 2004, 33 (6), 556−559. (22) Jo, W.; Kim, S.-J.; Kim, D.-Y. Analysis of the Etching Behavior of Zno Ceramics. Acta Mater. 2005, 53 (15), 4185−4188. (23) Gerischer, H.; Sorg, N. Chemical Dissolution of Zinc-Oxide Crystals in Aqueous-Electrolytesan Analysis of the Kinetics. Electrochim. Acta 1992, 37 (5), 827−835. (24) Han, X. G.; Jiang, Y. Q.; Xie, S. F.; Kuang, Q.; Zhou, X.; Cai, D. P.; Xie, Z. X.; Zheng, L. S. Control of the Surface of ZnO Nanostructures by Selective Wet-Chemical Etching. J. Phys. Chem. C 2010, 114 (22), 10114−10118. (25) Maki, H.; Ikoma, T.; Sakaguchi, I.; Ohashi, N.; Haneda, H.; Tanaka, J.; Ichinose, N. Control of Surface Morphology of ZnO by Hydrochloric Acid Etching. Thin Solid Films 2002, 411 (1), 91−95. (26) Valtiner, M.; Torrelles, X.; Pareek, A.; Borodin, S.; Gies, H.; Grundmeier, G. In Situ Study of the Polar Zno(0001)-Zn Surface in Alkaline Electrolytes. J. Phys. Chem. C 2010, 114 (36), 15440−15447. (27) Valtiner, M.; Borodin, S.; Grundmeier, G. Stabilization and Acidic Dissolution Mechanism of Single-Crystalline ZnO (0001) Surfaces in Electrolytes Studied by in-Situ Afm Imaging and Ex-Situ Leed. Langmuir 2008, 24 (10), 5350−5358. (28) Wu, J. S.; Xue, D. F. Controlled Etching of Hexagonal ZnO Architectures in an Alcohol Thermal Process. Mater. Res. Bull. 2010, 45 (3), 295−299. (29) Dhanaraj, G.; Dudley, M.; Bliss, D.; Callahan, M.; Harris, M. Growth and Process Induced Dislocations in Zinc Oxide Crystals. J. Cryst. Growth 2006, 297 (1), 74−79. (30) Andeen, D.; Kim, J. H.; Lange, F. F.; Goh, G. K. L.; Tripathy, S. Lateral Epitaxial Overgrowth of ZnO in Water at 90 Deg. Adv. Funct. Mater. 2006, 16 (6), 799−804. (31) Richardson, J. J.; Lange, F. F. Controlling Low Temperature Aqueous Synthesis of ZnO. 1. Thermodynamic Analysis. Cryst. Growth Des. 2009, 9 (6), 2570−2575. (32) Palacios-Lidon, E.; Perez-Garcia, B.; Vennegues, P.; Colchero, J.; Munoz-Sanjose, V.; Zuniga-Perez, J. Anisotropic Chemical Etching of Semipolar {10(1̅)(1̅)}/{10(1)1} ZnO Crystallographic Planes: Polarity Versus Dangling Bonds. Nanotechnology 2009, 20 (6). (33) Kuo, C. L.; Kuo, T. J.; Huang, M. H. Hydrothermal Synthesis of Zno Microspheres and Hexagonal Microrods with Sheetlike and Platelike Nanostructures. J. Phys. Chem. B 2005, 109 (43), 20115− 20121. (34) Yahiro, J.; Kawano, T.; Imai, H. Nanometric Morphological Variation of Zinc Oxide Crystals Using Organic Molecules with Carboxy and Sulfonic Groups. J. Colloid Interface Sci. 2007, 310 (1), 302−311.
Research Council (ARC), and by the ARC Discovery Scheme DP0985970.
■
REFERENCES
(1) Triboulet, R.; Perriere, J. Epitaxial Growth of ZnO Films. Prog. Cryst. Growth Charact. Mater. 2003, 47 (2−3), 65−138. (2) Sim, A. Y. L.; Goh, G. K. L.; Tripathy, S.; Andeen, D.; Lange, F. F. Photoluminescence of Hydrothermally Epitaxied ZnO Films. Electrochim. Acta Electrochim. Acta 2007, 52 (8), 2933−2937. (3) Xiao, Z. Y.; Liu, Y. C.; Dong, L.; Shao, C. L.; Zhang, J. Y.; Lu, Y. M.; Zhen, D. Z.; Fan, X. W. The Effect of Surface Properties on Visible Luminescence of Nanosized Colloidal Zno Membranes. J. Colloid Interface Sci. 2005, 282 (2), 403−407. (4) Thompson, D. B.; Richardson, J. J.; DenBaars, S. P.; Lange, F. F. Light Emitting Diodes with ZnO Current Spreading Layers Deposited from a Low Temperature Aqueous Solution. Appl. Phys. Express 2009, 2 (4), 042101/1−042101/3. (5) Yan, F. P.; Huang, L. H.; Zheng, J. S.; Huang, J.; Lin, Z.; Huang, F.; Wei, M. D. Effect of Surface Etching on the Efficiency of ZnOBased Dye-Sensitized Solar Cells. Langmuir 2010, 26 (10), 7153− 7156. (6) Li, D.; Haneda, H. Morphologies of Zinc Oxide Particles and Their Effects on Photocatalysis. Chemosphere 2003, 51 (2), 129−137. (7) Perez-Lopez, O. W.; Farias, A. C.; Marcilio, N. R.; Bueno, J. M. C. The Catalytic Behavior of Zinc Oxide Prepared from Various Precursors and by Different Methods. Mater. Res. Bull. 2005, 40 (12), 2089−2099. (8) Wang, X.; Song, J.; Liu, J.; Wang, Z. L. Direct-Current Nanogenerator Driven by Ultrasonic Waves. Science 2007, 316 (5821), 102−105. (9) Wang, Z. L. Zinc Oxide Nanostructures: Growth, Properties and Applications. J. Phys.: Condens. Matter 2004, 16 (25), R829−R858. (10) Garcia, S. P.; Semancik, S. Controlling the Morphology of Zinc Oxide Nanorods Crystallized from Aqueous Solutions: The Effect of Crystal Growth Modifiers on Aspect Ratio. Chem. Mater. 2007, 19, 4016−4022. (11) Kim, J. H.; Andeen, D.; Lange, F. F. Hydrothermal Growth of Periodic, Single-Crystal Zno Microrods and Microtunnels. Adv. Mater. (Weinheim, Ger.) 2006, 18 (18), 2453−2457. (12) Cao, X. L.; Zeng, H. B.; Wang, M.; Xu, X. J.; Fang, M.; Ji, S. L.; Zhang, L. D. Large Scale Fabrication of Quasi-Aligned ZnO Stacking Nanoplates. J. Phys. Chem. C 2008, 112 (14), 5267−5270. (13) Sugunan, A.; Warad, H.; Boman, M.; Dutta, J. Zinc Oxide Nanowires in Chemical Bath on Seeded Substrates: Role of Hexamine. J. Sol-Gel Sci. Technol. 2006, 39 (1), 49−56. (14) Tian, Z. R. R.; Voigt, J. A.; Liu, J.; McKenzie, B.; McDermott, M. J.; Rodriguez, M. A.; Konishi, H.; Xu, H. F. Complex and Oriented ZnO Nanostructures. Nat. Mater. 2003, 2 (12), 821−826. (15) Wang, D. B.; Song, C. X.; Hue, Z. S. Shape-Controlled Synthesis of Zno Architectures. Cryst. Res. Technol. 2008, 43 (1), 55−60. (16) Shang, T.-M.; Sun, J.-H.; Zhou, Q.-F.; Guan, M.-Y. Controlled Synthesis of Various Morphologies of Nanostructured Zinc Oxide: Flower, Nanoplate, and Urchin. Cryst. Res. Technol. 2007, 42 (10), 1002−1006. (17) Zhao, F. H.; Li, X. Y.; Zheng, J. G.; Yang, X. F.; Zhao, F. L.; Wong, K. S.; Wang, J.; Lin, W. J.; Wu, M. M.; Su, Q. Zno PineNanotree Arrays Grown from Facile Metal Chemical Corrosion and Oxidation. Chem. Mater. 2008, 20 (4), 1197−1199. (18) Xu, L. F.; Liao, Q.; Zhang, J. P.; Ai, X. C.; Xu, D. S. SingleCrystalline ZnO Nanotube Arrays on Conductive Glass Substrates by Selective Disolution of Electrodeposited ZnO Nanorods. J. Phys. Chem. C 2007, 111 (12), 4549−4552. (19) Wang, H. Q.; Li, M.; Jia, L. C.; Li, L.; Wang, G. Z.; Zhang, Y. X.; Li, G. H. Surfactant-Assisted in Situ Chemical Etching for the General Synthesis of ZnO Nanotubes Array. Nanoscale Res. Lett. 2010, 5 (7), 1102−1106. (20) Takahashi, K.; Funakubo, H.; Ohashi, N.; Haneda, H. MicroPatterning of ZnO Single Crystal Surface by Anisotropic Wet5641
dx.doi.org/10.1021/la2047273 | Langmuir 2012, 28, 5633−5641
