Spin triplet supercurrents in thin films of ferromagnetic CrO2
Anwar, M.S.
Citation
Anwar, M. S. (2011, October 19). Spin triplet supercurrents in thin films of ferromagnetic CrO2. Casimir PhD Series. Retrieved from https://hdl.handle.net/1887/17955
Version: Corrected Publisher’s Version
License: Licence agreement concerning inclusion of doctoral thesis in the Institutional Repository of the University of Leiden
Downloaded from: https://hdl.handle.net/1887/17955
Note: To cite this publication please use the final published version (if applicable).
grow epitaxially on TiO2 in the form of rectangular grains. On an untreated TiO2 substrate, the grains are aligned with their longer side parallel to the c-axis. They are parallel to the b-axis when grown on a pretreated substrate.
On sapphire the grains also are rectangular but they grow in six directions due to the hexagonal structure of the substrate. The magnetic properties of the films are strongly connected to the morphology, which is why they are presented together in this chapter. Basically, films grown on sapphire show a sixfold magnetic anisotropy corresponding to their morphology. Strained films on TiO2 show an easy magnetization along the b-axis, relaxed films along the c-axis. Intermediate situations also occur as function of film thickness and temperature. 1
1Parts of this chapter has been published in
M. S. Anwar and J. Aarts, Inducing supercurrents in thin films of ferromagnetic CrO2, Supercond. Sci. Technol. 24, 024016 (2011).
24
3.1. Introduction 25
3.1 Introduction
As discussed in Chapter 2, CrO2 is a half metallic ferromagnet which was utilized to investigate the existence of a long range proximity effect in super- conductor (S) / ferromagnet (F) hybrid structures. It was observed that a supercurrent can pass through CrO2, which is only possible for spin triplet supercurrents [11]. How the triplet pairs are generated was not understood, but it was noticed that the 100 nm thick CrO2 films used in the experiments exhibited biaxial magnetic anisotropy (presence of two independent magnetic easy axes) [35]. Intrinsically, CrO2 has uniaxial magnetic anisotropy with an easy axis along the crystallographic c-axis, with a magnetic moment of 2 µB/Cr atom, where µB is the Bohr magneton, corresponding to a saturation magnetization of 651 emu/cm3= 651 kA/m. In thin films, the magnetic prop- erties can be different from the bulk, in particular when the substrate induces strain in the film because of a lattice mismatch. It is therefore important for our experiments on induced superconductivity to control and understand both the growth and the magnetic properties of thin CrO2films.
So far, Chemical Vapor Deposition (CVD) is the only successful technique to deposit thin films of CrO2 [36, 37]. They are grown on isostructural TiO2
and hexagonal sapphire substrates. The lattice parameters of TiO2are closely matched with CrO2 and epitaxial growth is possible with small (although not negligible) effects of substrate induced strain. It has been reported that pretreatment of the TiO2substrates with Hydrofluoric acid (HF) can enhance the strain in the films [38–41]. Of course, this affects both the electronic and the magnetic properties. The growth situation for a sapphire substrate is rather complicated because of its hexagonal structure, which is close to Cr2O3. Growth on sapphire actually starts as Cr2O3and then changes to the required CrO2[37, 42].
Miao et al. [43] investigated the effects of such TiO2 pretreatment on the magnetic properties of CrO2 thin films and found that pretreatment induces more strain and affects, in particular, the magnetic anisotropy. In the strained films the magnetic easy axis lies along the b-axis for films of thickness less than 50 nm. Films grown on untreated substrates are strain free and have bulk-like properties even for very thin films. However, the effects are subtle. It was also reported that a 60 nm thick CrO2film had the c-axis as easy axis, even though the substrate was pretreated [44]. It shows that the substrate treatment is a critical factor in obtaining a certain anisotropy, be it biaxial or otherwise.
We grow CrO2 thin films on TiO2 and sapphire substrates. The morphol- ogy is studied with Atomic Force Microscopy (AFM), the crystal structure with X-ray diffraction (XRD) and magnetic properties are investigated in a Quantum Design SQUID magnetometer. This chapter is divided in the follow-
CrO2 thin films can be deposited on isostructural rutile TiO2 substrates at a temperature of 390 ◦C using CrO3 as a solid precursor heated at 260◦C.
Regarding the substrate temperature, there is a very small window to grow pure CrO2 around 390◦C; Cr2O5impurities are present at 380◦C and a more stable phase Cr2O3 appears at 400◦C [36].
We modified the Ishibashi setup by using a separate heater for precursor and substrate along with the furnace to keep the entire glass tube at precursor temperature. Such an apparatus has been used before as well [37, 45, 46]. The precursor boat and substrate holder are placed inside a borosilicate glass tube of 55 cm in length. The substrate holder is inclined at 45◦ to the flow of the Oxygen. Solid CrO3 precursor is loaded in a stainless steel boat, 11 cm away from the substrate holder. Thin films are deposited at a substrate temperature of 390◦C whereas the the precursor and furnace are heated at 260◦C. The va- pors of the precursor are transported from the precursor zone to the substrate zone by O2 with a flow rate of 100 sccm. The schematic and a photograph of the deposition setup is shown in Fig. 3.1. It was found that the quality of the films is sensitive to the substrate temperature, while the precursor tempera- ture is not so sensitive. The rate of deposition also increases with the increase of the substrate temperature due to accelerated dissociation of the precursor [37]. Finally, the quality of the film depends on the quality of the substrate.
For epitaxial growth of CrO2, good examples of substrates with lattice parameters closely matching the a-axis of the tetragonal CrO2 (a = b = 0.4421 nm, c = 0.2916 nm) are TiO2(100) (tetragonal with a = b = 0.4594 nm, c = 0.2958 nm) or Al2O3(0001) (hexagonal with a = 0.4754 nm, c = 1.299 nm), for which the method was specifically demonstrated [42]. The lattice mismatch between film and substrate is given by
mismatch = (af ilm− asubstrate)/asubstrate% (3.1)
3.2. Deposition of CrO2thin films 27
Figure 3.1: (a) A schematic of the CVD setup, (b) a photograph of the furnace, with the insets showing the precursor boat (left) and the substrate holder (right).
The lattice mismatch of the a-axis and the c-axis between CrO2 film and TiO2 substrate are -3.79% and -1.48% respectively [47] which suggests that a higher lateral growth rate can take place along the c-axis than along the in-plane a- = b-axis. The sapphire substrate is isostructural with Cr2O3 with a lattice mismatch of the a-axis of 4% [42], while the a-axis mismatch between CrO2and sapphire is of the order of -7%. This makes growth of Cr2O3rather than CrO2 favourable on sapphire. Still, CrO2can form since three times the CrO2c-axis fits the internal diagonal axis of the hexagonal structure to within 6%. The high amount of oxygen in the gas flow then apparently still favors CrO2growth, but only after a layer of Cr2O3has grown first, which results in a rather complex growth on sapphire.
The cleaning of substrates before deposition is a critical issue. Both sap- phire and TiO2 substrates are cleaned with organic solvents such as acetone and isopropanol in an ultrasonic bath for 15 minutes (min) in each. Acetone
Figure 3.2: (a) AFM image of an untreated TiO2 substrate. (b) AMF image of a pretreated TiO2 substrate shows the trenches with 0.4 ˚A (unit cell size) parallel to the b-axis.
is used to eradicate organic dust from the substrate surface, while isopropanol is used to remove the residue of acetone. The substrates are always dried with nitrogen gas after cleaning with deionized (DI) water for short time in an ultrasonic bath. It is also noticed that sometimes there are particular dust particles which do not remove with organic solvents. After the cleaning, to get rid off such particles, reactive O2ions plasma cleaning is utilized for 5 min.
For films deposited on TiO2 substrates, the quality can be enhanced with HF (hydrofluoric acid) pretreatment of the substrate before deposition. For this purpose TiO2substrates are treated with 25% concentrated HF for 45 min after cleaning with organic solvents. We investigated the substrate under AFM after every five minutes treatment and found atomic trenches at the substrate surface starting to develop after 30 min HF pretreatment. Figure 3.2 shows AFM images of untreated and pretreated substrates. The untreated substrate is uniformly smooth while the pretreated substrate has the atomic trenches parallel to the b-axis with a height of 0.4 nm. The trenches appear to be anisotropic etching but in Ref.[43] atomic step like structures were observed.
HF with 25% concentration is a strong acid. We also etched substrates with the considerably lower concentration of 4% HF for 5 min followed by the cleaning with organic solvents as described above. We then also find atomic trenches but only after annealing the substrate at 400 ◦C for about fifteen minutes. Although the differences were not investigated exhaustively, both types of pretreatment appear to yield films with the same characteristics.
3.3. Crystallography 29 It is obvious that some of the decomposed precursor deposits at the inner walls of the process glass tube. It was noticed that with progressive depositions of films, some solid micro particles fall down on the surface of the substrate, and make the film rough. Not only substrate cleaning is important but cleaning of the process glass tube is therefore important to grow high quality CrO2films via CVD.
The growth rate on a TiO2 substrate is higher than that on a sapphire substrate. On TiO2 it takes about 10 min for a total thickness of 100 nm, but on sapphire about 1 h. The thickness of the films can be measured with X-ray reflectometery (XRR) or Rutherford Backscattering (RBS). XRR mea- surements were not successful for films on TiO2. For sapphire substrates, the deposition of an initial layer of Cr2O3poses another problem in measuring the thickness of only the CrO2part. In this situation RBS is a powerful technique to measure the accurate thickness. The difficulty lies in the fact that the de- position rate is very sensitive (exponentially) to the substrate temperature, so that the deposition time cannot be used reliably to estimate the thickness of the film after a single calibration run with RBS. Because of this difficulty, we generally used the magnetic properties of the films to measure their thickness.
Cr2O3is an antiferromagnetic insulator, not contributing to the magnetization at low fields, while CrO2has a magnetic moment of 2.0 µB/Cr-atom. We cal- culated the magnetic moment for our films after measuring thickness by RBS and found 1.76 µB/Cr-atom, which is close (about 10%) to the theoretical value and indicates that the thickness can be estimated to within 10%.
3.3 Crystallography
After the successful growth of CrO2 on TiO2 and sapphire substrates, we investigated the crystallography and phase purity by XRD (using a Siemens D5005, and Cu K-alpha radiation).
3.3.1 Deposition on TiO2
We deposited CrO2 films on two different kind of TiO2 substrates, untreated and pretreated. Figure 3.3 shows the typical XRD patterns for both types.
It reveals that the films are single phase CrO2, single crystal and a-axis ori- ented without other visible phases of Chromium oxide. Of course, there may be low concentrations Cr2O3 or Cr2O5 present under the threshold of XRD and induced more with the passage of time. For untreated films, the (200) peak corresponding to CrO2 appeared at 40.78◦ and stays at this angle for a thickness range 50 - 200 nm. For pretreated films the angle of the (200) peak is increasing with the increase in the thickness of the films (see Fig. 3.3b).
Figure 3.3: Typical XRD patterns (logarithmic intensity) of CrO2 thin films of various thicknesses deposited on (a) untreated, (b) pretreated, TiO2 sub- strates.
Figure 3.4 shows the relative change of the out-of-plane a-axis of the various films with respect to bulk CrO2. Untreated films show some contraction at 100 nm thickness which has almost disappeared at 200 nm. The films are therefore weakly or not strained. The pretreated films show a much larger contraction (0.6%) at least up to 150 nm before it disappears, while film of 400 nm actually still showed such contraction. It could be that that substrate had an unusually favorable step structure, or that the substrate of the film of 220 nm was not pretreated correctly. Generally, the data show that films on pretreated substrates are strained.
3.3.2 Deposition on sapphire
Figure 3.5 shows the basic XRD patterns between 39◦ and 42.25◦ (2θ)◦ for three CrO2thin films deposited on sapphire substrates with different thickness.
A strong peak of (006) at 41.7ocomes from the sapphire (0001) substrate, the peak at 39.80◦is due to the presence of Cr2O3, and a peak at 40.78◦represents the (200) reflection of a-axis oriented CrO2. It is clear from the Figure that a 20 nm thick film shows no signature of the CrO2 phase, whereas the Cr2O3
phase is significantly present. The peak of (200) corresponding to the CrO2
appears for a 120 nm thick film and is strongly increased for the 400 nm thick film, while the (006) peak from Cr2O3 is present for both films. The results
3.4. Morphology 31
Figure 3.4: Relative change in out-of-plan axis (a-axis) of CrO2films compared to bulk CrO2as a function of thickness of the film, for films grown on untreated TiO2, on pretreated TiO2, and on sapphire.
show that Cr2O3 leads CrO2 in growth and deposits as an initial layer, which is reasonable in the context of lattice mismatch. Rabe et al. [42] found the same results, they also reported the presence of an initial layer of 40 nm Cr2O3
after which CrO2 grows in columnar fashion.
3.4 Morphology
CrO2(100) thin films with various thicknesses were deposited on TiO2 and sapphire substrates. The morphology is investigated under atomic force mi- croscope (Veeco/Bruker Multimode AFM, OLYMPUS Cantilever, with a tip of 7 nm in diameter and 11 µm in height) utilizing tapping mode at room temperature.
3.4.1 Deposition on TiO2
High quality CrO2 thin films can be deposited on isostructural TiO2 sub- strates, but it was mentioned above, the properties of the films sensitively depend on the pretreatment of the substrate surface. It was shown by Miao et al. [43] that this is observable in the morphology of the films.
Figure 3.5: XRD patterns of CrO2thin films deposited on sapphire substrate with three different thickness values, 20 nm, 120 nm and 400 nm. There is no trace of CrO2in the 20 nm thick film. Relatively weak signatures of CrO2are present for 120 nm, which become clear for 400 nm thickness.
Figure 3.6a shows the morphology for a 70 nm thick CrO2 film deposited on an untreated TiO2substrate. The grains mostly have a rectangular shape, while some are close to square, with aspect ratio’s between 1 and 3 and rms surface roughness of the order of 4 nm. The rectangular grains have their longer side along the substrate c-axis, apparently because of the lower lattice mismatch along this direction. For a 100 nm thick film deposited on pre- treated TiO2 substrate, almost all the grains are rectangular (see Fig. 3.6b), less wide than for films deposited on untreated substrates, and more uniform, with an rms roughness less than 2 nm. An important difference is that the rectangular grains are aligned along the b-axis rather than the c-axis. It shows that HF pretreatment strongly affects the lateral growth rate and makes it higher along the b-axis, apparently because of the atomic trenches (≈ 0.4 nm) on the substrate surface, which lie along the b-axis (see Fig. 3.2). Our results on the morphology are similar to the observation of Miao et al. [43], except that in their case the films on untreated substrates were square rather than rectangular along the b-axis2. The homogeneity and flatness of the films is also
2There are some small mistakes in Ref. [43] which may lead to confusion. In quoting the
3.4. Morphology 33
Figure 3.6: (a-b) AFM images of films deposited on TiO2(100) substrates:
(a) untreated (b) pretreated. The substrate orientation is the same for both images and given by the arrow in (b). (c) A TEM cross section image of CrO2
on TiO2(100). The top dark band is a 5 nm thick a-MoGe layer. (Courtesy of M. Porcu from Delft University.)
shown in a Transmission Electron Microscope (TEM) image (cross-section) of a 50 nm thick CrO2 thin film deposited on TiO2, given in Fig. 3.6c.
3.4.2 Deposition on sapphire
Growth on sapphire (Al2O2)(0001)(a = 4.754 ˚A; c = 12.990 ˚A; spacegroup:
R¯3c) proceeds in a very different way, as shown by Rabe et al. [42]. The hexag- onal sapphire lattice does not provide a good fit to the rectangular surface net
relative lattice mismatch, the values along the b-axis and the c-axis have been interchanged, as have been the substrate b-axis and the c-axis in Fig. 2 of Ref. [43].
Figure 3.7: (a) AFM image of CrO2deposited on sapphire. The film grows in the form of rectangular grains oriented 60◦ with respect to each other. The inset shows a zoomed-in image with three needle-like grains visible. (b) A TEM image revealing a layer of Cr2O3 and the columnar growth of CrO2. The upper dark layer is the amorphous MoGe film deposited after cleaning the surface. (Courtesy of M. Porcu from Delft University.) (c) A model representing growth mechanism for CrO2rectangular crystallites with six fold rotational symmetry and the c-axis as longer side over three atomic layers of Cr2O3. (taken from Ref.[42])
of TiO2, and the growth actually starts as Cr2O3, which also has a hexagonal structure. After an initial layer of around 40 nm, columnar growth of CrO2
starts, with the c-axes along all of the three equivalent in-plane axes of the Cr2O3. The morphology then consists of rectangular and rather needle-shaped grains oriented at 60◦ to each other, with often star-like structures. This is clearly seen in the morphology measured by AFM and shown in Fig. 3.7a, with the inset showing a close-up of three needles with 60◦ orientation. The films are relatively rough with rms values of the order of 12 nm.
Figure 3.7b shows a TEM image illustrating a cross-section of a 100 nm
3.5. Magnetic properties 35 thick CrO2 film deposited on sapphire. An initial layer of Cr2O3 less than 40 nm thick is clearly visible and preceding a columnar-like growth of CrO2
grains. These results are similar to the previous results of the Ref. [42]. The top dark layer is a capping layer of a-MoGe as will be used in subsequent experiments.
The growth mechanism of CrO2 on Al2O3 can be understood using the model presented by Rabe et al., [42], along with XRD and TEM analysis. An isostructural initial layer of Cr2O3 precedes the deposition of a-axis oriented CrO2 (see Fig. 3.5), the b-axis and the c-axis are in-plane. The analysis of lattice mismatch between CrO2c-axis and Cr2O3reveals that the c-axes of the grains are oriented along the internal diagonal axis with the three equivalent in-plane axes of Cr2O3(0001), which is also a preferred growth direction of the crystallites. It leads to a rectangular grain morphology of the CrO2crystallites oriented with the six fold rotational symmetry of the hexagonal structure of the Cr2O3 initial layer or the Al2O3 substrate.
3.5 Magnetic properties
Magnetization (M ) as a function of temperature (T ) and field (H) were in- vestigated in a SQUID magnetometer (MPMS-5S, Quantum design). M (T ) and M (H) are measured in the rage of 5 K to 400 K at different thickness of the films for films deposited on both TiO2 (untreated and pretreated) and sapphire. The magnetization data presented in this chapter are corrected for diamagnetic signal coming from the underlying substrate.
3.5.1 Films on TiO2
With respect to the magnetization, the situation for films grown on TiO2 is somewhat complicated. Bulk CrO2has uniaxial magnetic anisotropy with the crystallographic c-axis as an intrinsic easy axis at all temperatures. For films it was shown that, in a thickness range below roughly 30 nm, the substrate strain can lead to the b-axis being the easy axis of magnetization [48]. For slightly thicker films, this can change as function of temperature: the b-axis is the easy axis at low temperatures but becomes the hard axis at room temper- ature. Moreover, analysis of magnetoresistance and planar Hall effect showed that at intermediate thickness (100 nm), films can develop biaxial magnetic anisotropy, in which two magnetic easy axes occur, one in between the c-axis and the b-axis, and one mirrored around the c-axis to lie in between the c-axis and the b-axis [35]. It should be clear that such magnetic characteristics of the films very much depend on the strain state, and therefore on the surface pre- treatment of substrates. Here, we show magnetization data for our untreated
Figure 3.8: Magnetization as a function of applied field along two different directions for CrO2 thin films deposited on untreated TiO2 substrates for 110 nm and 20 nm thick films measured at 10 K, 100 K, and 300 K. The horizontal dotted lines illustrate the saturation magnetization (651 kA/m).
Figure 3.8 shows the magnetization as a function of applied field along the c-axis and the b-axis for two untreated films. For both the 20 nm film and the 110 nm film, the c-axis clearly is the easy axis at all temperatures (10 K, 100 K and 300 K) with a coercive field Hc of about 30 mT. This is similar to bulk behavior and can be expected on basis of the morphology, which shows little effects of strain. The XRD investigation also shows strain free deposition of the films on an untreated substrate. Figure 3.9 shows the magnetization behavior for three pretreated films. The 120 nm film has the c-axis as easy axis and the b-axis as hard axis for all temperatures. For the 70 nm and 50 nm films both the c-axis and the b-axis appear as easy axes at low temperatures (10 K and 100 K), and the b-axis is becoming hard at room temperature. There is no indication of a full reversal, meaning that the c-axis has become a hard axis, and we cannot draw a clear conclusion with respect to the possibility of biaxial anisotropy, although the fact that neither the c-axis nor the b-axis is a real hard axis at low temperatures points to the possibility that more easy
3.5. Magnetic properties 37
Figure 3.9: Magnetization loops for CrO2 thin films deposited on pretreated TiO2substrates for 120 nm, 70 nm and 50 nm thick films. They were measured with the field along the b-axis and the c-axis, at temperatures 10 K, 100 K and 300 K. The horizontal dotted lines illustrate the saturation magnetization (651 kA/m).
directions are present.
For another pretreated film, with a thickness of 30 nm, the magnetization was measured with various angles θ between the c-axis and the applied field.
Figure 3.10 shows that the most square loop is actually found at the directions
± 30oaway from the c-axis, while the b-axis still shows the signature of a hard axis. The inset shows the ratio between the magnetization MS and the rema- nent magnetization Mr as a function of angle, which is a more quantitative way to find the easy axis. Note that easy axes at ± 30o are very reminiscent of the biaxial anisotropy observed before for 100 nm thick films [35].
The investigations of magnetization loops for pretreated and untreated CrO2thin films reveals that the substrate induced strain is strongly affecting
Figure 3.10: Magnetization loops at 10 K for a 30 nm thick CrO2 thin film deposited on pretreated TiO2 substrate at different angles (θ) of the field with respect to the c-axis. The inset shows the ratio between saturation to remanence magnetization as function of angle. It clearly illustrates an easy axis at 30◦ from the c-axis.
the magnetic anisotropy. The strain favors the b-axis as the magnetic easy axis for thinner films (less than 50 nm) and the c-axis is the easy axis with increase in thickness above 100 nm. Biaxial anisotropy can be found on pre- treated TiO2 substrates for intermediate thicknesses. Still, there appears to be no clear simple recipe to grow a film with biaxial anisotropy. The full treat- ment procedure (cleaning, HF concentration, time) determines the outcome for strain and anisotropy.
The results on the 30 nm thick film, which appear to show biaxial behavior, invite another look at the data on the 70 nm and 50 nm films of Fig. 3.9. In the occurrence of a true biaxial anisotropy the b-axis is not an easy axis, while in the data of M (H) for 70 nm and 50 nm thick films show that both c-and b- axis are equally easy axes. This points to a different picture, in which actually a part of the film close to the interface between substrate and film has the b-axis as an easy axis and the rest of the film can become relaxed with the c-axis as an easy axis. The thickness of first part might be small (30 nm - 50 nm). For such a film there are two easy axes present along both the c-axis
3.5. Magnetic properties 39 and the b-axis. So, in the intermediate thickness range we do not necessarily produce real biaxial anisotropy on pretreated substrates. It can also be a state of mixed uniaxial anisotropy.
3.5.2 Films on sapphire
Figure 3.11: (a) Magnetization loops at three different angles of an externally applied field in the plane of a CrO2 film deposited on a sapphire substrate.
The larger loops (circles and dotted line) are for 30◦ and 90◦, the smaller loop (dashed line) is for 0◦ from one side of the substrate. (b) FMR data for 0◦ (along first easy axis), 30◦, 60◦ and 90◦. (Courtesy of F. Czeschka from Munich University.)
Figure 3.11a shows magnetization loops measured at 10 K for a 80 nm thick CrO2 film deposited on sapphire. Magnetization loops were measured along three different orientations of an externally applied field with respect to one side of the substrate, namely 0◦, 30◦, and 90◦. It is clear that the magnetization loops along 30◦ and 90◦ are identical, while the measurement along 0◦ shows a small reduction in the coercive field. It demonstrates that the crystalline anisotropy induces a six-fold magnetic anisotropy in the film, which is connected with the six fold rotational symmetry of the rectangular grains (see Fig. 3.7). The magnetic easy axes are oriented at 60◦to each other.
This is confirmed by Ferromagnetic Resonance (FMR) data, where a strong signal appears at 0◦ and 60◦. A relatively smaller signal is observed at 30◦ and 90◦ as shown in Fig. 3.11b. The close correlation between morphology and magnetization indicates that the longer side of the grains is the easy axis.
Intrinsically, the c-axis is the easy axis of magnetization for bulk CrO2 and XRD analysis says that the films on sapphire are strain free. It confirms that
Figure 3.12: (a) Magnetization as function of temperature M (T ) normalized to M (10 K) for a 70 nm thick CrO2 thin film deposited on a TiO2 substrate measured along both in plane axes (c-axis and b-axis). (b) First order deriva- tive Ms(T ) shows a peak appeared around 50 K only along the c-axis and the dip at higher temperatures is corresponding to the Curie temperature. (c) Linearized for Bloch’s T3/2 law of magnetization along c-axis at temperature range of 5 - 105 K. (d) Along b-axis, a curved dip is clearly present with the deviations from Bloch’s Law at around 50 - 70 K. The solid lines are the linear fits.
Apart from the behavior of the magnetization with respect to anisotropy, we also investigated the behavior of the saturation magnetization as function of temperature, MS(T ). For this, the films were cooled down to 5 K in a field of 100 mT (strong enough to saturate the samples). Ms(T ) is investigated along both in-plane orientations [001] (easy axis) and [010] (hard axis) in the
3.5. Magnetic properties 41 temperature range of 5 - 400 K and the data are presented in Fig. 3.12a for a 70 nm thick CrO2 film deposited on an untreated TiO2 substrate which is a relaxed film. The Curie temperature (Tc) is found to be about 390 K independent of the thickness of the films [48].
Figure 3.12a shows that MS(T ) along the c-axis behaves normally, by decreasing slowly with increasing temperature, but along the b-axis, MSrather shows a small dip-peak structure below 50 K, which is clearly present in Fig.
3.12b. A similar effect is found in a pretreated 110 nm thick film, as shown in the derivative of the normalized magnetization at low temperature (see Fig.
3.12c).
The decrease in MS(T ) is usually due to the excitation of spin waves, which is described by Bloch’s law,
Ms(T ) = Ms(0)(1 − BT3/2) (3.2) where B is a spin wave parameter connected to the spin wave stiffness D according to
B = 3.612( gµBkB
4πMSD) (3.3)
In turn, D is connected to the exchange energy J by,
D = JzSa2o
3 (3.4)
where z is the number of nearest neighbors, aois the lattice constant and S the spin quantum number [49].
Table 3.1: The values of the spin wave constant (B) and the spin wave stiffness (D) for different films.
Samples B(10−5K2/3)/D(meV ˚A2) B(10−5K2/3)/D(meV ˚A2)
T < 50K T > 50K
TiO2 (c-axis) 6.4/70 3.9/98
(b-axis) 6.5/69 2.6/127
sapphire 2.8/121 3.5/105
As can be seen in Fig. 3.12d, along the c-axis the Bloch law holds but with two different slopes with a cross-over around 70 K. The spin wave constant
Figure 3.13: Temperature dependent magnetization probed at 100 mT for a 200 nm thick CrO2 film deposited on a sapphire substrate. The inset shows the linearized magnetization according to Bloch’s T3/2 law of magnetization that shows the deviation around 70 K with two different linear fits.
Figure 3.13 presents the data on Ms(T ) for a 200 nm thick CrO2 film deposited on a sapphire substrate. The film was cooled down in a field of 100 mT and kept there during the measurements. The Curie temperature is about 390 K. It is again observed that Ms(T ) deviates from Bloch’s law around 70 K. From the Table 3.1 it is clear that the behavior for sapphire is opposite to the TiO2case in terms of spin wave stiffness constant, it is reduced at higher temperature.
3.6. Conclusions 43
3.6 Conclusions
CrO2(100) thin films were grown on both TiO2(100) (pretreated and un- treated) and sapphire(1000) substrates using the CVD technique. Pretreat- ment enhances the substrate induced strain, while the films deposited on un- treated substrates are strain free. The films grow epitaxially on TiO2 in the form of rectangular grains with two-fold rotational symmetry. On untreated TiO2 substrates, the grains are aligned along the c-axis. This changes to the b-axis when growing on a pretreated substrate. Films deposited on sapphire have rectangular grains in a six fold rotational pattern, coming from the hexag- onal structure of the substrate. Magnetically, unstructured films on TiO2have their easy axis along the c-axis. Thin strained films (typically below 50 nm) have the easy axis along the film b-axis. In the intermediate regime (typically between 50 nm and 150 nm) true biaxial anisotropy can occur (the easy axis lies in between b-axis and the c-axis), but we also find indications of mixed anisotropy (the film is a mixutre of both b-axis easy axis and c-axis easy axis).
On sapphire, the easy axis shows the six fold anisotropy corresponding to the morphology. From the temperature dependence of the saturation magnetiza- tion it appears that there is a change in behavior in the regime 70 K - 100 K, with change in the spin wave stiffness.
