research papers
Seed layer formation by deposition of microcrystallites on a revolving substrate: modeling of the effective linear elastic, piezoelectric, and dielectric coefficients
aHolcombe Department of Electrical and Computer Engineering, Clemson University, Clemson, SC 29634, USA, and bDepartment of Materials Science and Engineering, Clemson University, Clemson, SC 29634, USA
*Correspondence e-mail: art.ballato@gmail.com, jballat@clemson.edu
Averaging of material coefficients of crystallites deposited at an angle to a rotating substrate is considered. A simple model is proposed, and applied to determine effective linear dielectric, piezoelectric, and elastic constants of all Laue groups. While these represent tensors of rank 2, 3, and 4, the method applies generally to tensors of any rank. Results are then particularized for 6mm crystals, and applied numerically to zinc oxide, ZnO. It is shown that, by means of the rotating substrate method, depositions may be achieved having the equivalent of hexagonal anisotropy, enabling the creation of `engineered' structures.
Keywords: deposition; sputtering; rotating substrate; crystallites; dielectric permittivity; piezoelectricity; elasticity; zinc oxide; engineered polycrystals; seed layers; texture.
1. Introduction
Deposition technology is critical in micro- and nano-electronic fabrication of sensors, transducers, filters, energy harvesters, resonators, etc. (Martin, 2010; Behrisch, 2013a, Behrisch, 2013b; Bundesmann & Neumann, 2018). These devices most often consist of multi-layer structures, the reliability and functionality of which depend on the ability to create uniform and stable layer interfaces. One modality for assuring interface compatibility is an initial deposition between working layers of a thin `seed' or buffer layer, whose material and composition are chosen to afford compatible adhesion and to provide suitable conditions for subsequent uniform working layer growth (Clement et al., 2012; Noh et al., 2014; Li et al., 2020). This is particularly important when forming structures comprised of single, or quasi-single crystal piezoelectric working layers. The axial orientation of these, with respect to the substrate normal, i.e. the texture (Bunge, 1982; Suwas & Gurao, 2008) allows their effective piezo-coupling and elastic values to be optimized (Bjurström et al., 2006; Yanagitani et al., 2007a; Du et al., 2009; Moreria et al., 2011).
It is found that often superior seed layers are achieved by deposition of microcrystallites at a single angle to a revolving substrate (Stedile et al., 1994; Auger et al., 2002; Mertin et al., 2018). The substrate rotation averages the crystallite properties, resulting in a polycrystalline aggregate film with material properties modified from the single crystal values of the seed material. After deposition of the seed layer, the substrate rotation is halted, and deposition of the textured film is commenced. This oriented film may be of the same material as that of the seed, or may, indeed, be of any material, deposited at the same angle as that of the seed microcrystals, or at any desired angle. This is arranged by suitable adjustment of the deposition apparatus parameters.
This work derives effective values for the elastic stiffnesses, piezoelectric stress coefficients, and dielectric permittivities of the seed layer using a simple model outlined below. The model may easily be expanded to accommodate a more realistic orientation distribution function characterizing the incident seed layer
rather than a fixed single deposition angle. It is noted that material values deviate from those of the bulk as crystallite size diminishes; this is usual when the micro/nanoboundary is crossed. The effective crystallite material coefficients are to be substituted for those of the bulk, with no changes in the averaging procedure described herein.If the seed layer is not constrained to be `thin', then other potential configurations emerge. It becomes possible to construct an assortment of tailored structures such as acoustic and optical Bragg reflectors consisting of alternating high- and low-impedance layers (Newell, 1964; Link et al., 2006; Dvoesherstov et al., 2013), as utilized, for example, in solidly mounted resonators (Lakin, 2003; Iborra et al., 2012; Nguyen et al., 2019). Other options are enabled as well, such as superlattices (Schuller, 1980; Akçakaya et al., 1990; Bykhovski et al., 1997), including Fibonacci and Moiré designs (MacDonald, 1988; Dean et al., 2013; Mouldi & Kanzari, 2013). A simple example is offered in Section 4 where ZnO microcrystals are considered to be deposited with (103) planes parallel to the revolving substrate plane. Substrate rotation, and consequent angular averaging (see Section 2), produces an engineered crystal, again with 6mm symmetry, albeit with modified material coefficients, which are computed.
2. The model
Microcrystallites are deposited on a revolving substrate, all with a specified facet lying in the substrate plane. The x3 axis of the crystallite is thus inclined with respect to the substrate normal (X3) by a facet angle θ; see Fig. 1. The substrate revolution, in angle φ, creates a distribution of crystallite normals ([002] in the diagram) that form a cone about the substrate normal. The model assumes a uniform areal distribution of crystallites on the substrate. A suitable angular averaging then yields effective values 〈q〉 = , where q is the quantity to be averaged. In the following sections this averaging is carried out schematically on the point-group matrices characterizing the linear elastic, piezo, and dielectric crystal properties. These three families of phenomenological coefficients are much utilized in electro-acoustic applications and are used here to illustrate the results applied to tensors of ranks 2, 3, and 4 in their matrix form. The model may be extended to material tensors of arbitrary rank.
Axial conventions, and associated nomenclature, for the point groups are given, e.g. in Institute of Radio Engineers Standards Committee (1949); symmetry relations relating the tensorial material coefficients are succinctly discussed by Bhagavantam & Suryanarayana (1949), Fumi (1952a), Fumi (1952b), Fumi (1952c), Fieschi (1957), and in engineering format by Cady (1946), Cady (1964) Mason (1966), Nye (1985). Tensor coefficients describing the elastic, piezo, and dielectric effects, are transformed according to standard rules (Bond, 1943; Juretschke, 1952; Bechmann, 1960; Hearmon, 1957). Initially, the crystallite x3 axis is normal to the substrate reference plane, and a first rotation about its x1 or x2 axis carries it to the configuration shown in Fig. 1. The elastic, piezoelectric, and dielectric matrices resulting from this first rotation are denoted [c′], [e′], and [ɛ′], respectively. Subsequent rotation of the crystallite sitting on the substrate plane is described by a second rotation about its x3′ axis; the averaging is carried out on the resulting transformed quantities, with final averaged results denoted as 〈c″〉, 〈e″〉, and 〈ɛ″〉.
3. Model results for all Laue groups
In this section are given the averaged components, 〈″〉, of the dielectric, piezoelectric, and elastic tensors for all Laue groups (Bechmann & Hearmon, 1969; Brendel, 1979).
3.1. Linear dielectric permittivities
Five distinct, symmetric, permittivity matrices, [ɛ], corresponding to Laue groups I, II, III, IV–VI, and VII, characterize the linear dielectric behavior of crystals (Cady, 1946, 1964; Mason, 1966; Nye, 1985). The geometry of Fig. 1 is realized by a rotation about an in-plane x1 or x2 axis by facet angle θ; this transforms these to five distinct matrices [ɛ′] in standard fashion. Anticipating the final result, these [ɛ′] matrices are not given explicitly. A further rotation about the resulting x3′ axis (normal to the substrate plane), and subsequent averaging by the revolving substrate produces two distinct 〈ɛ″〉 matrices, the components of which are given in terms of the components of [ɛ′] in Table 1, with 〈ɛ22″〉 = 〈ɛ11″〉, and zero off-diagonal elements. In terms of the components of the original [ɛ] matrices, the final result is given in Table 2.
|
|
3.2. Linear piezoelectric stress moduli
Sixteen distinct matrices [e], characterize the linear piezoelectric behavior of the 20 piezoelectric classes (Cady, 1946, 1964; Mason, 1966; Nye, 1985). Each of the Laue groups is represented. The pairs (4, 6), (422, 622), (4mm, 6mm), and (23, 43m) share the same matrices. The geometry of Fig. 1 is realized by a rotation about an in-plane x1 or x2 axis by facet angle θ; this transforms each of these sixteen to one of five distinct matrices [e′] in standard fashion. The Laue group I results for both x1 and x2 rotations are identical, so there are but nine distinct [e′] matrices. Again, anticipating the final result, these [e′] matrices are not given explicitly. A further rotation about the resulting x3′ axis (normal to the substrate plane), and subsequent averaging by the revolving substrate produces three distinct 〈e″〉 matrices, that we denote as 〈A〉, 〈B〉, and 〈C〉. The structure of matrix 〈A〉 is given in Table 3, where the 〈e″〉 components are expressed in terms of the [e′] components; it is identical in form to the matrix of the unrotated classes (4, 6). Matrix 〈B〉 is that of matrix 〈A〉 with e14″ = e25″ = 0 and has the same structure as the matrix of the unrotated classes (4mm, 6mm). Finally, matrix 〈C〉 is that of matrix 〈A〉 with e15″ = e24″ = e31″ = e32″ = e33″ = 0 and has the same structure as the matrix of the unrotated classes (422, 622). The averaged results for the twenty piezoelectric classes are given in Table 4.
|
|
The pertinent [e′] matrix components are given explicitly in terms of the components of the original [e] matrices in Table 5 and Table 6 for Laue group I. These relations are easily specialized for the remaining Laue groups of higher symmetry.
|
|
3.3. Linear elastic stiffnesses
Nine distinct, symmetric, stiffness matrices, [c], corresponding to Laue groups I, II, III, IVa, IVb, Va, Vb, VI, and VII, characterize the linear elastic behavior of crystals (Cady, 1946; 1964; Mason, 1966; Nye, 1985). The geometry of Fig. 1 is realized by a rotation about an in-plane x1 or x2 axis by facet angle θ; this transforms each of these nine to matrices [c′] in standard fashion. A further rotation about the resulting x3′ axis, (normal to the substrate plane), and subsequent averaging by the revolving substrate produces, for all Laue groups, a single 〈c″〉 matrix with hexagonal symmetry. The components of 〈c″〉 are given in terms of the components of [c′] in Table 7. The pertinent [c′] matrix components are given explicitly in terms of the components of the original [c] matrices in Tables 8 and 9 for Laue group I. These relations are easily specialized for the remaining Laue groups of higher symmetry.
|
|
|
4. Example: zinc oxide, Laue group VIb, class 6mm
Zinc oxide, in its 6mm polytype, is similar in many ways to the III-nitrides, and to polarized ceramics. Its high piezoelectric coupling has led to its use in a wide variety of acousto-electronic devices. Provided here are the effective material coefficients, and piezoelectric coupling factors, of deposited ZnO crystallites subjected to the averaging procedure described in Section 2. The 6mm symmetry requires isotropy in the basal plane, hence all rotations (tilts) about any axis, (e.g. x1 or x2), in the (001) plane will yield identical results. We take the first rotation about x1 by a facet angle θ corresponding to the (103) plane (Kushibiki et al., 2009); in the coordinate system of the substrate, this results in the equivalent of monoclinic symmetry. The averaging process then yields 〈″〉 quantities but again with 6mm symmetry. The numerical input values are taken from Jaffe & Berlincourt (1965) and Pearton et al. (2005).
4.1. Dielectric permittivities
A first rotation, about x1, results in [ɛ′] quantities as follows: ɛ11′ = ɛ11; ɛ12′ = ɛ13′ = 0; ɛ22′ = ɛ11C2 + ɛ33S2; ; and ɛ33′ = ɛ33C2 + ɛ11S2. The resulting 〈ɛ″〉 components, taken from Table 2, and numerical values for ZnO referred to the facet plane (103), are given in Table 10.
|
4.2. Piezoelectric stress coefficients
A first rotation, about x1, results in [e′] quantities as follows: e15′ = e15C; e16′ = e15S; e21′ = e31S; e22′ = [e33S2 + (e31 + 2e15)C2]S; e23′ = [e31S2 + (e33 − 2e15)C2]S; e24′ = [e33S2 − e31S2 + e15(C2 − S2)]C; e31′ = e31C; e32′ = [e31C2 − 2e15S2 + e33S2]C; e33′ = [e33C2 + 2e15S2 + e31S2]C; and e34′ = [e33C2 − e31C2 − e15(C2 − S2)]S. The subsequent averaging, from Table 4, results in matrix 〈B〉 symmetry, with components as follows: 〈e15″〉 = 〈e24″〉 = (1/2)(e15′ + e24′) = (1/2)[2e15C2 + (e33 − e31)S2]C; 〈e31″〉 = 〈e32″〉 = (1/2)(e31′ + e32′) = (1/2)[e31(1 + C2) + (e33 − 2e15)S2]C; 〈e33″〉 = [e33C2 + (e33 + 2e15)S2]C; numerical results are given in Table 11.
|
4.3. Elastic stiffnesses components
A first rotation, about x1, results in the [c′] quantities given in Table 12. When these are put into the 〈c″〉 matrix of Table 7, the entries of Table 13 result; numerical values for ZnO referred to the facet plane (103), are also given.
|
|
4.4. Computation of piezo-coupling values
Piezo-coupling coefficients determine efficacy of piezoelectric transduction. As measures thereof, `they appear in considerations of bandwidth and insertion loss in transducers and signal processing devices, in the location and spacing of critical frequencies of resonators, and in electrical/mechanical energy conversion efficiency in actuators and energy harvesters' (Ballato & Ballato, 2023). Calculation of these quantities for the one-dimensional case of thickness vibrations of plates and films requires knowledge only of the effective dielectric, piezoelectric, and elastic tensors referred to the plate/film coordinates; these are the 〈″〉 quantities in our situation. An eigenvalue equation is solved, yielding three modal elastic stiffnesses, cm, three modal piezoelectric coefficients, em, and the effective permittivity in the thickness direction, ɛ3. Then the coupling coefficients are given in the generic form |km| = |em|/(cmɛ3)1/2 (Foster et al., 1968; Wittstruck et al., 2005; Ballato & Ballato, 2023).
In Table 14 are given the relevant data for averaged ZnO with zero facet angle. Corresponding wave velocities are Vm = (cm/ρ)1/2. The of bulk crystalline ZnO is ρ ≃ 5.606 × 103 (kg m−3) (Pearton et al., 2005). Table 15 provides this data for averaged ZnO with facet angle of the (103) plane. The |km| values are plotted in Fig. 2, as a function of tilt angle γ, for averaged ZnO with facet angle of the (103) plane.
|
5. Conclusions
The rotating substrate method of crystallite deposition is modeled, allowing computation of effective material coefficients of the layers resulting from the averaging. Modeling of the averaging applies generally to tensors of any rank, but is applied here to dielectric, piezoelectric, and elastic coefficients. A worked numerical example particularized to 6mm ZnO is provided. Layers comprised of crystallites deposited with zero facet angle are contrasted with those deposited on facet plane (103). It is proposed that the method enables creation of `engineered' layered structures.
Acknowledgements
JB acknowledges support from the J. E. Sirrine Foundation.
Funding information
Funding for this research was provided by: J. E. Sirrine Textile Foundation.
References
Akçakaya, E., Farnell, G. W. & Adler, E. L. (1990). J. Appl. Phys. 68, 1009–1012. Google Scholar
Auger, M. A., Gago, R., Fernández, M., Sánchez, O. & Albella, J. M. (2002). Surf. Coat. Technol. 157, 26–33. CrossRef CAS Google Scholar
Ballato, A. & Ballato, J. (2023). Int. J. Ceram. Eng. Sci. 5, e10182. CrossRef Google Scholar
Bechmann, R. (1960). Acta Cryst. 13, 110–113. CrossRef IUCr Journals Google Scholar
Bechmann, R. & Hearmon, R. F. S. (1969). The Third Order Elastic Constants. In Landolt-Börnstein – Numerical Data and Functional Relationships. Group III, Vol. 2. Berlin: Springer. Google Scholar
Behrisch, R. (2013a). Sputtering by Particle Bombardment I: Physical Sputtering of Single-Element Solids. Berlin, Heidelberg: Springer. Google Scholar
Behrisch, R. (2013b). Sputtering by Particle Bombardment II: Sputtering of Alloys and compounds, Electron and Neutron Sputtering, Surface Topography. Berlin, Heidelberg: Springer. Google Scholar
Bhagavantam, S. & Suryanarayana, D. (1949). Acta Cryst. 2, 21–26. CrossRef IUCr Journals Web of Science Google Scholar
Bjurström, J., Wingqvist, G. & Katardjiev, I. (2006). IEEE Trans. Ultrason. Ferroelect. Freq. Contr. 53, 2095–2100. Google Scholar
Bond, W. L. (1943). Bell Syst. Tech. J. 22, 1–72. CrossRef Google Scholar
Brendel, R. (1979). Acta Cryst. A35, 525–533. CrossRef CAS IUCr Journals Web of Science Google Scholar
Bundesmann, C. & Neumann, H. (2018). J. Appl. Phys. 124, 231102. CrossRef Google Scholar
Bunge, H.-J. (1982). Texture Analysis in Materials Science: Mathematical Methods. London: Butterworths. Google Scholar
Bykhovski, A. D., Gelmont, B. L. & Shur, M. S. (1997). J. Appl. Phys. 81, 6332–6338. CrossRef CAS Google Scholar
Cady, W. G. (1946). Piezoelectricity: An Introduction to the Theory, Applications of Electromechanical Phenomena in Crystals. New York: McGraw-Hill. Google Scholar
Cady, W. G. (1964). Piezoelectricity: An Introduction to the Theory, Applications of Electromechanical Phenomena in Crystals. New York: Dover. Google Scholar
Clement, M., Olivares, J., Capilla, J., Sangrador, J. & Iborra, E. (2012). IEEE Trans. Ultrason. Ferroelect. Freq. Contr. 59, 128–134. CrossRef Google Scholar
Dean, C. R., Wang, L., Maher, P., Forsythe, C., Ghahari, F., Gao, Y., Katoch, J., Ishigami, M., Moon, P., Koshino, M., Taniguchi, T., Watanabe, K., Shepard, K. L., Hone, J. & Kim, P. (2013). Nature, 497, 598–602. CrossRef CAS PubMed Google Scholar
Du, J., Xian, K., Wang, J. & Yang, J. (2009). Ultrasonics, 49, 149–152. CrossRef PubMed CAS Google Scholar
Dvoesherstov, M. Yu., Cherednik, V. I., Bosov, S. I., Orlov, I. Ya. & Rudenko, O. V. (2013). Acoust. Phys. 59, 513–520. CrossRef CAS Google Scholar
Fieschi, R. (1957). Physica, 23, 972–976. CrossRef Google Scholar
Foster, N. F., Coquin, G. A., Rozgonyi, G. A. & Vannatta, F. A. (1968). IEEE Trans. Sonics Ultrason. 15, 28–40. CrossRef CAS Google Scholar
Fumi, F. G. (1952a). Acta Cryst. 5, 44–48. CrossRef CAS IUCr Journals Web of Science Google Scholar
Fumi, F. G. (1952b). Acta Cryst. 5, 691–694. CrossRef IUCr Journals Web of Science Google Scholar
Fumi, F. G. (1952c). Nuovo Cimento, 9, 739–756. CrossRef Google Scholar
Hearmon, R. F. S. (1957). Acta Cryst. 10, 121–124. CrossRef IUCr Journals Google Scholar
Iborra, E., Clement, M., Capilla, J., Olivares, J. & Felmetsger, V. (2012). Thin Solid Films, 520, 3060–3063. CrossRef CAS Google Scholar
Institute of Radio Engineers Standards Committee (1949). Proc. IRE 37, 1378–1395. Google Scholar
Jaffe, H. & Berlincourt, D. A. (1965). Proc. IEEE, 53, 1372–1386. CrossRef Google Scholar
Juretschke, H. J. (1952). Acta Cryst. 5, 148–149. CrossRef IUCr Journals Google Scholar
Kushibiki, J.-I., Ohashi, Y., Arakawa, M. & Tanaka, T. (2009). J. Appl. Phys. 105, 114913. CrossRef Google Scholar
Lakin, K. M. (2003). IEEE Microw. 4, 61–67. CrossRef Google Scholar
Li, Y., Liu, X., Wen, D., Lv, K., Zhou, G., Zhao, Y., Xu, C. & Wang, J. (2020). Acta Cryst. B76, 233–240. CrossRef IUCr Journals Google Scholar
Link, M., Schreiter, M., Weber, J., Primig, R., Pitzer, D. & Gabl, R. (2006). IEEE Trans. Ultrason. Ferroelect. Freq. Contr. 53, 492–496. CrossRef Google Scholar
MacDonald, A. H. (1988). Fibonacci Superlattices, edited by C. R. Leavens & R. Taylor, pp. 347–380. In Interfaces, Quantum Wells, and Superlattices, NATO ASI Series, Vol. 179. Boston, MA: Springer. Google Scholar
Martin, P. M. (2010). Handbook of Deposition Technologies for Films and Coatings, edited by P. M. Martin, pp. 1–31. Amsterdam: Elsevier. Google Scholar
Mason, W. P. (1966). Crystal Physics of Interaction Processes. New York: Academic. Google Scholar
Mertin, S., Heinz, B., Rattunde, O., Christmann, G., Dubois, M.-A., Nicolay, S. & Muralt, P. (2018). Surf. Coat. Technol. 343, 2–6. CrossRef CAS Google Scholar
Moreria, M., Bjurström, J., Kubart, T., Kuzavas, B. & Katardjiev, I. (2011). IEEE Ultrason. Symp. Proc. pp. 1238–1241. Google Scholar
Mouldi, A. & Kanzari, M. (2013). PIER M, 32, 169–180. CrossRef Google Scholar
Newell, W. E. (1964). Proc. IEEE, 52, 1603–1608. CrossRef Google Scholar
Nguyen, N., Johannessen, A., Rooth, S. & Hanke, U. (2019). Ultrasonics, 94, 92–101. CrossRef CAS PubMed Google Scholar
Noh, Y.-K., Park, C.-H., Oh, J.-E., Lee, S.-T. & Kim, M.-D. (2014). J. Korean Phys. Soc. 64, 1577–1580. CrossRef CAS Google Scholar
Nye, J. F. (1985). Physical Properties of Crystals: Their Representation by Tensors and Matrices. Oxford University Press. Google Scholar
Pearton, S. J., Norton, D. P., Ip, K., Heo, Y. W. & Steiner, T. (2005). Prog. Mater. Sci. 50, 293–340. Web of Science CrossRef CAS Google Scholar
Schuller, I. K. (1980). Phys. Rev. Lett. 44, 1597–1600. CrossRef CAS Web of Science Google Scholar
Stedile, F. C., Freire, F. L. Jr, Schreiner, W. H. & Baumvol, I. J. R. (1994). Vacuum, 45, 441–446. CrossRef CAS Google Scholar
Suwas, S. & Gurao, N. P. (2008). J. Indian Inst. Sci. 88, 151–177. CAS Google Scholar
Wittstruck, R. H., Emanetoglu, N. W., Lu, Y., Laffey, S. & Ballato, A. (2005). J. Acoust. Soc. Am. 118, 1414–1423. CrossRef CAS Google Scholar
Yanagitani, T., Kiuchi, M., Matsukawa, M. & Watanabe, Y. (2007a). J. Appl. Phys. 102, 024110. CrossRef Google Scholar
This is an open-access article distributed under the terms of the Creative Commons Attribution (CC-BY) Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original authors and source are cited.