**5. Imaging procedure**

As in AFM scanning, the detector signal can be fed back to the scanner *z* actuator. This mode of operation is called constant signal mode, in contrast to the open-loop or constant distance mode. The constant signal mode is robust and allows an accurate tracking of the sample surface, but it also presents a few problems. For example, the magnetic signal can be positive or negative, while stable feedback is only possible when the interaction does not change sign. This makes it necessary to bias the signal: the application of a voltage between the sample and the tip introduces an additional (electrostatic) force. Another problem of this mode is that the magnetic and non-magnetic interactions are mixed. The mixing ratio depends on the tip-sample distance which itself depends on the magnetic interaction. This makes the contributions very difficult to separate. For operation in air, it is known that the interaction with the surface contaminant layer and the damping (in dynamic mode) have a stronger influence on the tip than the van der Waals interaction (Porthun et al., 1998).

Quantitative data about the sample stray field can only be derived from MFM images when topographic signal contributions are not included. This is especially important when the tip is brought very close to the sample (in order to improve resolution), since non-magnetic forces become increasingly stronger. The solution to this problem is to keep the topography influence constant by letting the tip follow the surface height profile (Porthun et al., 1998). This constant distance mode places higher demands on instrument stability, because it is sensitive to drift. In the Digital Instruments microscope (Nanoscope 3A Multimode), the specific method employed to separate signal contributions is called lift mode (Fig. 6). It involves measuring the topography on each scan line in a first scan (left panel), and the magnetic information in a second scan of the same line (right panel). The difference in height *h* between the two scans, the so-called lift height, is selected by the user. Topography is measured in dynamic AM mode and the data is recorded to one

Fig. 5. Scanning electron images of AFM probes like the ones used for MFM. The probes are

As in AFM scanning, the detector signal can be fed back to the scanner *z* actuator. This mode of operation is called constant signal mode, in contrast to the open-loop or constant distance mode. The constant signal mode is robust and allows an accurate tracking of the sample surface, but it also presents a few problems. For example, the magnetic signal can be positive or negative, while stable feedback is only possible when the interaction does not change sign. This makes it necessary to bias the signal: the application of a voltage between the sample and the tip introduces an additional (electrostatic) force. Another problem of this mode is that the magnetic and non-magnetic interactions are mixed. The mixing ratio depends on the tip-sample distance which itself depends on the magnetic interaction. This makes the contributions very difficult to separate. For operation in air, it is known that the interaction with the surface contaminant layer and the damping (in dynamic mode) have a stronger influence on the tip than the van der Waals interaction

Quantitative data about the sample stray field can only be derived from MFM images when topographic signal contributions are not included. This is especially important when the tip is brought very close to the sample (in order to improve resolution), since non-magnetic forces become increasingly stronger. The solution to this problem is to keep the topography influence constant by letting the tip follow the surface height profile (Porthun et al., 1998). This constant distance mode places higher demands on instrument stability, because it is sensitive to drift. In the Digital Instruments microscope (Nanoscope 3A Multimode), the specific method employed to separate signal contributions is called lift mode (Fig. 6). It involves measuring the topography on each scan line in a first scan (left panel), and the magnetic information in a second scan of the same line (right panel). The difference in height *h* between the two scans, the so-called lift height, is selected by the user. Topography is measured in dynamic AM mode and the data is recorded to one

coated with a magnetic thin film. Specifications are mentioned in the text (Bruker

Corporation, 2011)

**5. Imaging procedure** 

(Porthun et al., 1998).

image. This height data is also used to move the tip at a constant local distance above the surface during the second (magnetic) scan line, during which the feedback is turned off. In theory, topographic contributions should be eliminated in the second image.

Fig. 6. Outline of the lift mode principle. Magnetic information is recorded during the second pass (right panel). The constant height difference between the two scan lines is the lift height *h* (adapted from Hendrych et al., 2007)

Magnetic data can be recorded either as variations in amplitude, frequency, or phase of the cantilever oscillation. It is argued that phase detection and frequency modulation give the best results, with a higher signal-to-noise ratio (Porthun et al., 1998; Hartmann, 1999). However, these detection modes can require the addition of an electronics module to the microscope. In our MFM measurements we have used amplitude detection, which measures changes in the cantilever's amplitude of oscillation relative to the piezo drive. The signal depends on the force derivative in the following manner (Porthun et al., 1998):

$$f = f\_0 \sqrt{1 - \frac{\partial F \, / \partial z}{c}}\tag{6}$$

with 0*f* the free resonance frequency of the cantilever in the case of no tip sample interaction. In the amplitude detection, the cantilever is oscillated at a fixed frequency *ext* <sup>0</sup> *f f* , where in the case of / 0 *F z* the oscillation amplitude is already slightly below the maximum amplitude at 0*f* . When the resonance frequency changes this will result in a change in cantilever oscillation amplitude which can easily be detected. The disadvantage of this technique is that it is very slow for cantilevers with low damping and that a change in cantilever damping will be misinterpreted as change in resonance frequency.

It should be noted that an attractive interaction ( / 0 *F z* ) leads to a negative amplitude change (dark contrast in the image), while a repulsive interaction ( / 0 *F z* ) gives a positive amplitude variation (bright contrast).

Finally, Fig. 7(b) shows a typical MFM image. In this case, the sample was a piece of metal evaporated tape: a standard sample that is used to check whether the microscope is correctly tuned to image magnetic materials (Koch, 2005). It is clear that no correlation exists between the topography data shown on the left, and the magnetic data on the right. Consequently, the separation of both contributions is successful.

Magnetic Force Microscopy: Basic Principles and Applications 49

Thin films of amorphous SiMn and GeMn were prepared by conventional radio frequency sputtering. The Mn concentration ([Mn]) in the samples was in the ~ 0.1–24 at.% concentration range. Additionally, thin films of amorphous SiCo and GeCo were also deposited by sputtering. The Co concentration ([Co]) in the samples stayed in the ~ 1.7–10.3 at.% range. Pure samples were also prepared following identical conditions. The films, typically 1700 nm thick, were deposited principally on c-quartz and c-Si substrates. After deposition the films were submitted to thermal annealing treatments in the range of 200−900 oC. The samples were characterized by a great variety of experimental techniques: (1) the composition of the films was determined mainly by energy dispersive x-ray spectrometry (EDS), (2) the atomic structure of the films was investigated by Raman scattering spectroscopy and x-ray diffraction (XRD) experiments, (3) the surface of the films was investigated by scanning electron microscopy (SEM) and AFM, (4) their optical properties were examined by means of transmission measurements, (5) the electrical resistivity of the films was measured using the standard van der Pauw technique, and (6) their magnetic properties were investigated by superconducting quantum interference device (SQUID) magnetometry and MFM. Except the SQUID measurements, all experimental characterizations were always carried out at room temperature. For further details, see Ferri

As confirmed by the Raman measurements, as the thermal annealing advances, the SiMn samples show crystallization signals that are accompanied by the growth of randomly dispersed sub-micrometre structures on the surface of the films. These structures are Mncontaining Si crystallites, surrounded by Si crystallites, amorphous Si and the MnSi1.7 silicide phase (Ferri et al., 2009a). It is worth mentioning that the MnSi1.7 is representative of a group of several Mn-silicides of the MnxSiy form, with y/x approximately equal to 1.7: Mn4Si7, Mn15Si26, Mn27Si47, etc. Therefore, in this work, the Mn-silicides are simply identified

The morphology and magnetic characteristics of the SiMn20% sample were investigated by means of AFM and MFM measurements (Fig. 8). Based on the AFM results the observed structures are typically ~ 750–1200 nm large and 300–400 nm high. Also, the image contrast present in Fig. 8(b) is a clear indication of the magnetic activity present in sample SiMn20%. At these dimensions, the contrast shown by the MFM images occurs because of force gradients between the FM tip and the magnetic activity present on the sample's surface. In this study, the MFM images were achieved after topography measurements (tapping mode) followed by sample surface scanning at a constant 200 nm height (lift mode). According to this procedure, no van der Waals forces are expected to be detected, and any change in the vibration amplitude of the cantilever is proportional to the gradient of magnetic fields perpendicular to the sample surface (Hartmann, 1999). It is worth noting that no MFM contrast was observed in the Mn-free film and SiMn20% sample as-deposited nor after

In addition to the presence of magnetic activity in the sample under study, it also produces a remarkable contrast in the MFM image of Fig. 8(b). The FM materials are known to form

**6.2 Experimental considerations** 

et al., 2009a, 2009b, 2010a, 2010b, 2011.

scanning the samples under the tapping mode.

**6.3 Results and discussion** 

by MnSi1.7.

Fig. 7. Topographic image (a) and magnetic force gradient image (b) of a metal evaporated tape (Koch, 2005)

## **6. Applications of MFM in the study of Si and Ge-based magnetic semiconductors**

#### **6.1 Motivation**

Driven by the promise of controlling charge and spin degrees of freedom, and its consequent technological impact through the realization of spintronic devices, many different ferromagnetic (FM) semiconductors have been investigated over the last few years. The potential advantages of this class of devices (in the form of ultra-dense nonvolatile semiconductor memories, spin transistors and light emitting devices with polarized output, etc.) are expected to be, in addition to the low energy required to flip a spin: higher speed, greater efficiency and better stability (Zutic et al., 2004). Thus far, most of the work on FM semiconductors has been focused on Mn-containing II–VI or III–V compounds in which manganese replaces a fraction of group II or III sub-lattices (Dietl & Ohno, 2006). For practical reasons, however, the interest in a specific FM semiconductor depends on the existence of magnetic activity near or above room temperature as well as its compatibility with the current micro-electronics industry. Mn-containing Si- or Gebased compounds partially fit these requirements since they possess a mature processing technology and because of some recent experimental work reporting Curie temperatures well above 300 K (Zhang et al., 2004; Kim et al., 2007). Furthermore, the low solubility of Mn in crystalline (c-)Si or c-Ge can be partially circumvented by using their amorphous counterparts, which also provide a more homogenous Mn distribution. Indeed, this is a particularly interesting feature since charge and spin states are sensitive mostly to the local environment so the magnetic activity existing in c-Si or c-Ge should also be observable in amorphous Si or Ge.

Based on these facts, this section reports on the MFM characterization of amorphous Si and Ge thin films containing different amounts of Mn and Co. Even though the amorphous character of the as-deposited films, thermal annealing at increasing temperatures induces their crystallization. Following this procedure, their magnetic properties have been systematically investigated as a function of the impurity concentration and atomic structure.

#### **6.2 Experimental considerations**

48 Atomic Force Microscopy – Imaging, Measuring and Manipulating Surfaces at the Atomic Scale

Fig. 7. Topographic image (a) and magnetic force gradient image (b) of a metal evaporated

Driven by the promise of controlling charge and spin degrees of freedom, and its consequent technological impact through the realization of spintronic devices, many different ferromagnetic (FM) semiconductors have been investigated over the last few years. The potential advantages of this class of devices (in the form of ultra-dense nonvolatile semiconductor memories, spin transistors and light emitting devices with polarized output, etc.) are expected to be, in addition to the low energy required to flip a spin: higher speed, greater efficiency and better stability (Zutic et al., 2004). Thus far, most of the work on FM semiconductors has been focused on Mn-containing II–VI or III–V compounds in which manganese replaces a fraction of group II or III sub-lattices (Dietl & Ohno, 2006). For practical reasons, however, the interest in a specific FM semiconductor depends on the existence of magnetic activity near or above room temperature as well as its compatibility with the current micro-electronics industry. Mn-containing Si- or Gebased compounds partially fit these requirements since they possess a mature processing technology and because of some recent experimental work reporting Curie temperatures well above 300 K (Zhang et al., 2004; Kim et al., 2007). Furthermore, the low solubility of Mn in crystalline (c-)Si or c-Ge can be partially circumvented by using their amorphous counterparts, which also provide a more homogenous Mn distribution. Indeed, this is a particularly interesting feature since charge and spin states are sensitive mostly to the local environment so the magnetic activity existing in c-Si or c-Ge should also be

Based on these facts, this section reports on the MFM characterization of amorphous Si and Ge thin films containing different amounts of Mn and Co. Even though the amorphous character of the as-deposited films, thermal annealing at increasing temperatures induces their crystallization. Following this procedure, their magnetic properties have been systematically investigated as a function of the impurity concentration and atomic structure.

**6. Applications of MFM in the study of Si and Ge-based magnetic** 

tape (Koch, 2005)

**semiconductors** 

observable in amorphous Si or Ge.

**6.1 Motivation** 

Thin films of amorphous SiMn and GeMn were prepared by conventional radio frequency sputtering. The Mn concentration ([Mn]) in the samples was in the ~ 0.1–24 at.% concentration range. Additionally, thin films of amorphous SiCo and GeCo were also deposited by sputtering. The Co concentration ([Co]) in the samples stayed in the ~ 1.7–10.3 at.% range. Pure samples were also prepared following identical conditions. The films, typically 1700 nm thick, were deposited principally on c-quartz and c-Si substrates. After deposition the films were submitted to thermal annealing treatments in the range of 200−900 oC. The samples were characterized by a great variety of experimental techniques: (1) the composition of the films was determined mainly by energy dispersive x-ray spectrometry (EDS), (2) the atomic structure of the films was investigated by Raman scattering spectroscopy and x-ray diffraction (XRD) experiments, (3) the surface of the films was investigated by scanning electron microscopy (SEM) and AFM, (4) their optical properties were examined by means of transmission measurements, (5) the electrical resistivity of the films was measured using the standard van der Pauw technique, and (6) their magnetic properties were investigated by superconducting quantum interference device (SQUID) magnetometry and MFM. Except the SQUID measurements, all experimental characterizations were always carried out at room temperature. For further details, see Ferri et al., 2009a, 2009b, 2010a, 2010b, 2011.

#### **6.3 Results and discussion**

As confirmed by the Raman measurements, as the thermal annealing advances, the SiMn samples show crystallization signals that are accompanied by the growth of randomly dispersed sub-micrometre structures on the surface of the films. These structures are Mncontaining Si crystallites, surrounded by Si crystallites, amorphous Si and the MnSi1.7 silicide phase (Ferri et al., 2009a). It is worth mentioning that the MnSi1.7 is representative of a group of several Mn-silicides of the MnxSiy form, with y/x approximately equal to 1.7: Mn4Si7, Mn15Si26, Mn27Si47, etc. Therefore, in this work, the Mn-silicides are simply identified by MnSi1.7.

The morphology and magnetic characteristics of the SiMn20% sample were investigated by means of AFM and MFM measurements (Fig. 8). Based on the AFM results the observed structures are typically ~ 750–1200 nm large and 300–400 nm high. Also, the image contrast present in Fig. 8(b) is a clear indication of the magnetic activity present in sample SiMn20%. At these dimensions, the contrast shown by the MFM images occurs because of force gradients between the FM tip and the magnetic activity present on the sample's surface. In this study, the MFM images were achieved after topography measurements (tapping mode) followed by sample surface scanning at a constant 200 nm height (lift mode). According to this procedure, no van der Waals forces are expected to be detected, and any change in the vibration amplitude of the cantilever is proportional to the gradient of magnetic fields perpendicular to the sample surface (Hartmann, 1999). It is worth noting that no MFM contrast was observed in the Mn-free film and SiMn20% sample as-deposited nor after scanning the samples under the tapping mode.

In addition to the presence of magnetic activity in the sample under study, it also produces a remarkable contrast in the MFM image of Fig. 8(b). The FM materials are known to form

Magnetic Force Microscopy: Basic Principles and Applications 51

In this case, basically, the observed magnetic contrast occurs because of variations in the magnetization orientation along the sub-micrometre structures [Fig. 8(b)]. In other words, the presence of these Mn-based structures (probably Mn dimmers, in combination with the MnSi1.7 phase) can lead to the appearance of magnetic activity (Bernardini et al., 2004; Affouda et al., 2006) whose main characteristics are highly influenced by the size and shape of the structures. Fig. 10 shows the surface topography in connection with the measured magnetic contrast of a single sub-micrometre structure. The figure also displays the height profile and MFM voltage achieved under horizontal [Fig. 10(b)], vertical [Fig. 10(c)] and

Fig. 10. (a) Magnetic force microscopy image of an isolated sub-micrometre structure present in the SiMn20% film after thermal annealing at 600 oC. Its height profile (as obtained by AFM) and corresponding MFM voltage along the horizontal, vertical and diagonal dashed lines drawn in (a) are represented, respectively, in (b), (c) and (d). Note the MFM voltage pattern due to the presence of magnetic vortices in the structure (Ferri et al., 2009a)

is non-uniform (and/or highly influenced by the presence of MnSi1.7) around it.

It is interesting to observe the quite different topographic (AFM profile) and magnetic (MFM voltage) patterns achieved from the very same structure exclusively due to the presence of magnetic activity. The effect of manganese on the formation of these magnetic vortices is also remarkable suggesting that, once the structure is formed, the Mn distribution

The Mn-free, GeMn3.7% and GeMn24% films deposited under crystalline quartz substrates were also investigated through similar MFM measurements (Ferri et al., 2010a). Since these samples showed a flat surface, the magnetic activity of these three films was evaluated by scanning the MFM tip along a ~ 20 μm line across the crystalline quartz substrate partially

diagonal [Fig. 10(d)] scans along the structure.

Fig. 8. (a) AFM and (b) MFM images of the sputter-deposited SiMn20% film after thermal annealing at 600 oC. The AFM scanning was performed in the tapping mode, whereas MFM in the lift mode by means of a Co/Cr coated tip magnetized just before scanning. The measurements were carried out under room conditions (temperature and atmosphere) from a 1.7 µm thick film deposited on crystalline silicon (Ferri et al., 2009a)

domain structures to reduce their magnetostatic energy that, at very small dimensions such those experienced by a (sub-)micrometre dot, for example, adopts the configuration of a curling spin or magnetization vortex (Shinjo et al., 2000). When the dot thickness becomes much smaller than the dot diameter, all spins tend to align in-plane. In the curling configuration, the spin directions change gradually in-plane in order to maintain the exchange energy and to cancel the total dipole energy (Fig. 9). The development of these magnetic vortices is well documented in the literature and its comprehensive description can be found in many works (Zhu et al., 2002; Soares et al., 2008).

Fig. 9. Drawing of the magnetic moment configuration for ferromagnetic tri-dimensional sub-micrometre structures (Soares et al., 2008). At these very small dimensions, the magnetization adopts the pattern of a curling spin or magnetization vortex. In this curling arrangement, the spin directions change gradually in-plane in order to maintain the exchange energy and to cancel the total dipole energy

Fig. 8. (a) AFM and (b) MFM images of the sputter-deposited SiMn20% film after thermal annealing at 600 oC. The AFM scanning was performed in the tapping mode, whereas MFM in the lift mode by means of a Co/Cr coated tip magnetized just before scanning. The measurements were carried out under room conditions (temperature and atmosphere) from

domain structures to reduce their magnetostatic energy that, at very small dimensions such those experienced by a (sub-)micrometre dot, for example, adopts the configuration of a curling spin or magnetization vortex (Shinjo et al., 2000). When the dot thickness becomes much smaller than the dot diameter, all spins tend to align in-plane. In the curling configuration, the spin directions change gradually in-plane in order to maintain the exchange energy and to cancel the total dipole energy (Fig. 9). The development of these magnetic vortices is well documented in the literature and its comprehensive description

Fig. 9. Drawing of the magnetic moment configuration for ferromagnetic tri-dimensional sub-micrometre structures (Soares et al., 2008). At these very small dimensions, the magnetization adopts the pattern of a curling spin or magnetization vortex. In this curling arrangement, the spin directions change gradually in-plane in order to maintain the

a 1.7 µm thick film deposited on crystalline silicon (Ferri et al., 2009a)

can be found in many works (Zhu et al., 2002; Soares et al., 2008).

exchange energy and to cancel the total dipole energy

In this case, basically, the observed magnetic contrast occurs because of variations in the magnetization orientation along the sub-micrometre structures [Fig. 8(b)]. In other words, the presence of these Mn-based structures (probably Mn dimmers, in combination with the MnSi1.7 phase) can lead to the appearance of magnetic activity (Bernardini et al., 2004; Affouda et al., 2006) whose main characteristics are highly influenced by the size and shape of the structures. Fig. 10 shows the surface topography in connection with the measured magnetic contrast of a single sub-micrometre structure. The figure also displays the height profile and MFM voltage achieved under horizontal [Fig. 10(b)], vertical [Fig. 10(c)] and diagonal [Fig. 10(d)] scans along the structure.

Fig. 10. (a) Magnetic force microscopy image of an isolated sub-micrometre structure present in the SiMn20% film after thermal annealing at 600 oC. Its height profile (as obtained by AFM) and corresponding MFM voltage along the horizontal, vertical and diagonal dashed lines drawn in (a) are represented, respectively, in (b), (c) and (d). Note the MFM voltage pattern due to the presence of magnetic vortices in the structure (Ferri et al., 2009a)

It is interesting to observe the quite different topographic (AFM profile) and magnetic (MFM voltage) patterns achieved from the very same structure exclusively due to the presence of magnetic activity. The effect of manganese on the formation of these magnetic vortices is also remarkable suggesting that, once the structure is formed, the Mn distribution is non-uniform (and/or highly influenced by the presence of MnSi1.7) around it.

The Mn-free, GeMn3.7% and GeMn24% films deposited under crystalline quartz substrates were also investigated through similar MFM measurements (Ferri et al., 2010a). Since these samples showed a flat surface, the magnetic activity of these three films was evaluated by scanning the MFM tip along a ~ 20 μm line across the crystalline quartz substrate partially

Magnetic Force Microscopy: Basic Principles and Applications 53

as absolute magnetic data are available (such as those given by SQUID magnetometry, for example) the adopted experimental procedure can provide a convenient method to analyze the magnetic properties of microsized (or sub-microsized) isolated systems. As a final point, it is important to mention that the room temperature magnetic activity observed in the present GeMn samples (Fig. 11), occurs, basically, because of the presence of the Mn5Ge3

For the magnetic characterization of the SiCo and GeCo films (deposited on crystalline quartz) the MFM technique was used similarly to the GeMn samples, since these samples also showed a flat surface (Ferri et al., 2010b). The main results of these MFM measurements are shown in Fig. 12, which illustrates results obtained in some SiCo and GeCo samples without annealing and after thermal treatment up to the crystallization temperature. In these samples, after crystallization, the non-magnetic CoSi2 silicide and CoGe2 germanide phases were found, as confirmed by XRD measurements (not shown). Therefore, we must to keep in mind that the only phase that can cause ferromagnetism at room temperature (or higher) for the samples in question, is the metallic Co, which has a Curie temperature of ~

Fig. 12. MFM signal (as obtained from the voltage difference at the bare substrate-film edge region−see sketch of Fig. 11) as a function of the tip-to-sample distance, for as-deposited (AD) and thermally annealed (a) SiCo and (b) GeCo films (pure and containing different amounts of Co) deposited on c-quartz. The Co contents and the annealing temperatures are indicated in

The MFM measurements for the Co-free Si and Ge films (both amorphous and annealed up to the crystallization temperature) suggest the absence of magnetic activity. This experimental result is expected [since it was also observed in the set of Mn-free Si and Ge samples (Ferri et al., 2009a, 2010)], and is in accord with the literature (Bolduc et al., 2005; Cho et al., 2002). When annealed at high temperatures, the XRD results indicate the presence of non-magnetic phases in the films containing Co. In addition, it is known that Co is less efficient than Mn in promoting ferromagnetic alignment, and a high magnetic moment, for the case of Ge (Continenza et al., 2006). Therefore, we expect a similar magnetic behaviour from the Co for the Si matrix. Taking these considerations into account, and remembering the fact that the MFM experiments were performed at room temperature, it is expected for

the figure. The lines joining the experimental data points are just guides to the eye

ferromagnetic germanide phase (Ferri et al., 2009b, 2010a).

1382 K (Ko et al., 2006).

covered by the desired Ge film (see sketch in Fig. 11). By adopting this procedure, at the bare substrate-film edge, the MFM tip will experience a signal difference which is proportional to the magnetic response of the probed region. Considering that crystalline quartz gives no magnetic contrast in the MFM measurements, the observed MFM signal is exclusively due to the GeMn films. In fact, and in accord with the literature (Cho et al, 2002) and our SQUID results, no MFM signal has been observed from both the amorphous and crystallized Mn-free Ge films. Also, and in order to confirm that the MFM signal is mainly of magnetic nature (Porthun et al., 1998), the measurements were carried out at a fixed tip-tosample (substrate + film) distance *d* in the 100−2500 nm range. The main results of these MFM measurements, in conjunction with the SQUID data, are shown in Fig. 11. Here it is important to point out that similar results were obtained for the SiMn samples according this procedure (not shown).

Fig. 11. MFM signal (as obtained from the voltage difference at the bare substrate-film edge region−see sketch) as a function of the magnetization of saturation (as obtained from the SQUID measurements at *T* 300 K). The MFM data correspond to three different MFM tipto-sample distances ( *d* 200, 1000, and 2000 nm). The measurements were carried out on the GeMn3.7% and GeMn24% films, deposited on crystalline quartz: both amorphous (AD−asdeposited) and after crystallization at the temperatures indicated in the figure. The lines joining the experimental data points are just guides to the eye (Ferri et al., 2010a)

The experimental data of Fig. 11 indicates that the MFM signal decreases with the distance *d* : demonstrating the magnetic character behind the interaction between the MFM tip and the sample. Except for minor deviations in the MFM signals obtained with the lowest *d* values, which were clearly affected by the experimental conditions (temperature, film thickness, and instrumental resolution, for example), the MFM signal scales with the magnetization of saturation, as obtained from the SQUID measurements. Indeed, the MFM signal increases with [Mn] and after the crystallization of the GeMn films. Therefore, as far

covered by the desired Ge film (see sketch in Fig. 11). By adopting this procedure, at the bare substrate-film edge, the MFM tip will experience a signal difference which is proportional to the magnetic response of the probed region. Considering that crystalline quartz gives no magnetic contrast in the MFM measurements, the observed MFM signal is exclusively due to the GeMn films. In fact, and in accord with the literature (Cho et al, 2002) and our SQUID results, no MFM signal has been observed from both the amorphous and crystallized Mn-free Ge films. Also, and in order to confirm that the MFM signal is mainly of magnetic nature (Porthun et al., 1998), the measurements were carried out at a fixed tip-tosample (substrate + film) distance *d* in the 100−2500 nm range. The main results of these MFM measurements, in conjunction with the SQUID data, are shown in Fig. 11. Here it is important to point out that similar results were obtained for the SiMn samples according

Fig. 11. MFM signal (as obtained from the voltage difference at the bare substrate-film edge region−see sketch) as a function of the magnetization of saturation (as obtained from the SQUID measurements at *T* 300 K). The MFM data correspond to three different MFM tipto-sample distances ( *d* 200, 1000, and 2000 nm). The measurements were carried out on the GeMn3.7% and GeMn24% films, deposited on crystalline quartz: both amorphous (AD−asdeposited) and after crystallization at the temperatures indicated in the figure. The lines joining the experimental data points are just guides to the eye (Ferri et al., 2010a)

The experimental data of Fig. 11 indicates that the MFM signal decreases with the distance *d* : demonstrating the magnetic character behind the interaction between the MFM tip and the sample. Except for minor deviations in the MFM signals obtained with the lowest *d* values, which were clearly affected by the experimental conditions (temperature, film thickness, and instrumental resolution, for example), the MFM signal scales with the magnetization of saturation, as obtained from the SQUID measurements. Indeed, the MFM signal increases with [Mn] and after the crystallization of the GeMn films. Therefore, as far

this procedure (not shown).

as absolute magnetic data are available (such as those given by SQUID magnetometry, for example) the adopted experimental procedure can provide a convenient method to analyze the magnetic properties of microsized (or sub-microsized) isolated systems. As a final point, it is important to mention that the room temperature magnetic activity observed in the present GeMn samples (Fig. 11), occurs, basically, because of the presence of the Mn5Ge3 ferromagnetic germanide phase (Ferri et al., 2009b, 2010a).

For the magnetic characterization of the SiCo and GeCo films (deposited on crystalline quartz) the MFM technique was used similarly to the GeMn samples, since these samples also showed a flat surface (Ferri et al., 2010b). The main results of these MFM measurements are shown in Fig. 12, which illustrates results obtained in some SiCo and GeCo samples without annealing and after thermal treatment up to the crystallization temperature. In these samples, after crystallization, the non-magnetic CoSi2 silicide and CoGe2 germanide phases were found, as confirmed by XRD measurements (not shown). Therefore, we must to keep in mind that the only phase that can cause ferromagnetism at room temperature (or higher) for the samples in question, is the metallic Co, which has a Curie temperature of ~ 1382 K (Ko et al., 2006).

Fig. 12. MFM signal (as obtained from the voltage difference at the bare substrate-film edge region−see sketch of Fig. 11) as a function of the tip-to-sample distance, for as-deposited (AD) and thermally annealed (a) SiCo and (b) GeCo films (pure and containing different amounts of Co) deposited on c-quartz. The Co contents and the annealing temperatures are indicated in the figure. The lines joining the experimental data points are just guides to the eye

The MFM measurements for the Co-free Si and Ge films (both amorphous and annealed up to the crystallization temperature) suggest the absence of magnetic activity. This experimental result is expected [since it was also observed in the set of Mn-free Si and Ge samples (Ferri et al., 2009a, 2010)], and is in accord with the literature (Bolduc et al., 2005; Cho et al., 2002). When annealed at high temperatures, the XRD results indicate the presence of non-magnetic phases in the films containing Co. In addition, it is known that Co is less efficient than Mn in promoting ferromagnetic alignment, and a high magnetic moment, for the case of Ge (Continenza et al., 2006). Therefore, we expect a similar magnetic behaviour from the Co for the Si matrix. Taking these considerations into account, and remembering the fact that the MFM experiments were performed at room temperature, it is expected for

Magnetic Force Microscopy: Basic Principles and Applications 55

MFM and SQUID techniques can be very convenient to probe the magnetic properties of

The authors are indebted to Professor Antonio Ricardo Zanatta (Instituto de Física de São Carlos, Universidade de São Paulo, Brazil) for the support with the deposition and characterization of the Si and Ge samples. This work was financially supported by the

Affouda C. A., Bolduc M., Huang M. B., Ramos F., Dunn K. A., Thiel B., Agnello G., &

Bernardini F., Picozzi S., & Continenza A. (2004). Energetic stability and magnetic properties of Mn dimers in silicon, *Applied Physics Letters,* Vol. 84, No. 13, pp. 2289-2291. Bolduc M., Affouda C. A., Stollenwerk A., Huang M. B., Ramos F. G., Agnello G., & Labella

Bruker Corporation. (2011). MESP tips, In: *Bruker AFM Probes*, 23.09.2011, Available from: http://www.brukerafmprobes.com/Product.aspx?ProductID=3309. Cho S., Choi S., Hong S. C., Kim Y., Ketterson J., Kim B. J., Kim Y. C., & Hung J. H. (2002).

Continenza A., Profeta G., & Picozzi S. (2006). Transition metal impurities in Ge: chemical

Dietl T. & Ohno H. (2006). Engineering magnetism in semiconductors, *Materials Today,* Vol.

Ferri F. A. & Zanatta A. R. (2009). Structural, optical and morphological characterization of

Ferri F. A. (2010). *Synthesis and characterization of Si and Ge based films doped with magnetic* 

Ferri F. A., Pereira-da-Silva M. A., & Zanatta A. R. (2009). Evidence of magnetic vortices

*Journal of Physics D: Applied Physics,* Vol. 42, No. 13, pp. 132002-1−132002-5. Ferri F. A., Pereira-da-Silva M. A., & Zanatta A. R. (2011). Development of the MnSi1.7 phase

Ferri F. A., Pereira-da-Silva M. A., Zanatta A. R., Varella A. L. S., & de Oliveira A. J. A.

LaBella V. P. (2006). Observation of crystallite formation in ferromagnetic Mnimplanted Si, *The Journal of Vacuum Science and Technology A,* Vol. 24, No. 4, pp.

V. P. (2005). Above room temperature ferromagnetism in Mn-ion implanted Si,

Ferromagnetism in Mn-doped Ge, *Physical Review B,* Vol. 66, No. 3, pp. 033303-

trends and codoping studied by electronic structure calculations, *Physical Review B,* 

amorphous Ge100−xMnx films deposited by sputtering, *Journal of Physics D: Applied* 

*species*, PhD Thesis, Instituto de Física de São Carlos, Universidade de São Paulo,

formation in Mn-based sub-micrometre structures embedded in Si–Mn films,

in Mn-containing Si films, *Materials Chemistry and Physics,* Vol. 129, No. 1-2, pp. 148-

(2010). Effect of Mn concentration and atomic structure on the magnetic properties of Ge thin films, *Journal of Applied Physics,* Vol. 108, No. 11, pp. 113922-1−113922-5.

Brazilian agencies FAPESP and CNPq under CEPOF/INOF and INEO.

*Physical Review B,* Vol. 71, No. 3, pp. 033302-1−033302-4.

Vol. 73, No. 3, pp. 035212-1−035212-10.

*Physics,* Vol. 42, No. 3, pp. 035005-1−035005-6.

microsized (or sub-microsized) isolated structures.

**8. Acknowledgments** 

1644-1647.

1−033303-3.

9, No. 11, pp. 18-26.

São Carlos, Brazil.

153.

**9. References** 

the present samples a very weak or at least less intense magnetic signal than in the case of the Mn-containing films. Therefore, the results of Fig. 12 are in agreement with the initial expectations. Unlike observed in samples with Mn it is possible to identify only a slight decrease in the MFM signal with the tip-sample separation, due to the comparatively lower signal intensity. Here it is important to notice that the present procedure adopted in the MFM measurements is unique in the literature. Consequently, similar results obtained from others, for quantitative comparison purposes, are non-existent.

For the GeCo samples, we observed that the MFM signal intensity increased with increasing Co concentration [see Fig 12(b)]. The thermal treatment for samples with the same [Co], in principle, didn't intensify the magnetic signal. As an example of increasing MFM signal intensity with [Co], at a tip-sample separation of 500 nm, we observed that the Ge film with [Co] ~ 1.7 at.% showed a MFM signal of ~ 2 mV , and the Ge film with [Co] ~ 7.6 at.% exhibited a MFM signal of ~ 8 mV, both annealed at 500 oC. Still, as can be seen in Fig. 12(b), even the as-deposited Ge samples show magnetic signal, probably due to the existence of magnetically active Co atoms randomly distributed in the amorphous network. For the annealed samples, due to the diffusion of Co and the structural rearrangement of the network, it is expected that the number of magnetically active Co atoms increase (Ko et al., 2006). However, its magnetic activity does not exceed that of the amorphous films due to the formation of CoGe2. Finally, the increasing in the magnetic signal with increasing [Co] is expected since the number of magnetically active Co atoms probably also increases.

For the SiCo films, the situation seems somewhat different, and not systematic. At first, as shown in Fig. 12(a), the sample with [Co] ~ 2.8 at.% without annealing shows a relatively high value of magnetic signal due to the magnetically active Co. After annealing at 900 oC, its value is diminished, probably due to the formation of CoSi2. In contrast, the as-deposited Si film with the highest [Co] (~ 10.3 at.%) presents an extremely low MFM signal, probably due to the large number of magnetically inactive Co atoms, which may be associated with its highly disordered structure. After annealing at 900 oC, its magnetic activity is significantly increased due to the diffusion and consequent magnetic activation of the Co atoms. However, the magnetic activity is now limited by the existence of CoSi2, and, therefore, its magnetism is less intense than the as-deposited film with [Co] ~ 2.8 at.%, that, in principle, doesn't have the silicide phase.
