Development of Advanced Magnetic Force Microscopy
Tips for Magnetic Characterization
R.D. Gomez, Jon Orloffand Klaus Edinger
University of Maryland, College Park, MD
and
Sy-Hwang Liou
University of Nebraska, Lincoln, NE
PROJECT SUMMARY:
This research will emphasize novel means of the development of improved magnetic tips for magnetic force microscopy for characterizing the magnetic properties of thin film magnetic materials. The improved tips will be achieved by utilizing a variety of magnetic coatings which can be tailored to the properties of the materials under investigation, and by using state-of-the-art technologies like ion-beam milling to fabricate tips capable of much higher spatial resolution and enhanced sensitivity.
INTRODUCTION:
The study of the magnetic properties of matter on a nanometer scale is of interest both from the scientific and the technological points of view. From a scientific standpoint, our present understanding of magnetism at these length scales is in its infancy at best. It is only in recent years that experimental tools and sufficient computing power have become available to tackle the complicated magnetic behavior occurring the microscopic length scales. In comparison with bulk magnetic studies where the effects of interfaces, defects, morphological grains and so on are treated as perturbations and lumped into global parameters, micromagnetic models attempt to incorporate these as locally varying parameters in combination with the relative contributions of the Zeeman, magnetostatic, magnetoelastic and local anisotropies. The goal of predicting the magnetic domain structures is thus an inherently complex problem, and we are just beginning the development of models starting from well-known magnetic alloys having the simplest geometrical configurations. The so called “standard problem #1”, inititiated by the National Institues of Science and Technology several years ago, consists of a rectangular 1um x 2um strip of a 50 nm Permalloy thin film with no crystalline anistropy and perfect edges. From a technological point of view, the ability to measure and control the properties of magnetic domains is of great interest in the areas of data storage, micro electromechanical systems (MEMS), magnetic ramdom acess memories (MRAMS), magneto-electronics (spin transistors), quantum computing and even as magnetic sensors for the industrial and biomedical applications. Often, the novel features of these devices, such as the giant magneto-resistance in field sensing applications or the ultimate areal density for recording media, have its origins upon the nascent micromagnetic domain configurations.
Magnetic force microscopy is arguably the most important imaging tool for studying a wide variety of local magnetic phenomena. Since its development in 1987, it has emerged as a powerful micromagnetic tool and has revealed magnetic processes with unprecented clarity, resolution and ease. It allows the direct visualization of magnetic domains and provides the experimental basis for theoretical modeling. The technique measures change of the interaction force between a magnetized probe and the local stray magnetic field from the sample, point by point, as the probe is scanned across the surface. The probe is a silicon-based cantilever with a ferromagnetic tip on the free end and placed in close proximity (~25 nm) with sample surface. The inherent resolution depends upon the confinement of the interaction at the end of the probe and sensitivity depends upon the ratio of the cantilever spring constant and the magnetic moment. At present, the resolution of the commercial MFM probes is about 50-100 nm with a force constant of about 0.01 N/m, which is roughly equivalent to resolving the field gradients from a 10-12 emu source at a distance of 50 nm.
Despite the impressive performance and widespread use of the MFM, there are important probe-related limitations that need to be overcome to realize the full potential of this technique. First of these is the enhancement of resolution and sensitivity. As is well known from microscopy, in order to measure something at a given scale it is necessary to have a probe whose fundamental size is well below the size of the object to be measured. In the case of atomic force microscopy, for example, in order to characterize precisely and accurately a trench with a width of 0.1 micrometers and a depth of 0.2 micrometers, one needs a probe in the shape of a rod with a length of at least 0.2 micrometers and a well characterized diameter of less than 0.1 micrometers. In the case of magnetic force microscopy (MFM) the force between the probe and the sample is carried by the magnetic field. Obviously, the smaller the magnetically active area of the probe, the less it will be affected by areas from far away since the dipole nature of the field causes it to diminish rapidly with distance. Therefore, in order to make a high resolution MFM it would be necessary to create an extremely small magnetic probe. The smaller volume of the magnetic probe will result in a lower magnetic moment and a smaller interaction volume, and thus a weaker force. Hence, the lateral resolution of the MFM probe will also be limited by its sensitivity (e.g. the spring constant of the cantilever). The improvement in resolution would have to be complemented with an enhancement of the probe sensitivity.
Second, is the development of specialized probes whose properties are optimized for a given specimen and free from instrument-induced distortions. Since the MFM relies on a mutual interaction, it is inherently invasive. Thus, it is easy to imagine that measurement process itself could cause irreversible changes to the system and the measured image may not reflect the intrinsic state of the sample. Conversely, the probe’s moment itself may change as it moves in varying fields, which would cause a nonlinear instrument response. This would render the interpretation of images to be quite complicated and equivocal. To overcome this problem, it is necessary to use probes whose properties are compatible with the sample at hand. For instance in one extreme, a sample with very low coercivity such as garnets or Permalloy, would require the use of high sensitivity, low moment probe. The low probe moment ensures that fringing field is much lower that the sample coercivity, and increased sensitivity is to compensate the reduction in the interaction forces. On the other extreme, for the case of permanent magnet samples, the probe would have to be relatively stiffer and its coercivity must be higher than the stray field generated at the sample surface. In cases where no known materials exist that could withstand the strength of the specimen field, it may be judicious to use superparamagnetic probes and interpret the images accordingly. Most interesting problems are likely to have requirements that are midway between these extreme cases, while certain applications, such as understanding the magnetic evolution of soft magnetic materials at high fields, might require a low moment probes with high coercivity.
Finally, the MFM requires the fundamental understanding of the magnetic characteristics of the probes themselves, which can be incorporated into theoretical models of image interpretation. At present the generally accepted model for MFM assumes a point dipole at the tip apex. This picture is adequate in qualitative descriptions which treat the images as representations of the distribution of magnetic charges from the divergence of the volume magnetization or the normal component of the surface magnetization. Several sophisticated theoretical descriptions for image representation have been proposed in the literature which take into consideration the finite volume of the tip. Unfortunately, because of the absence of direct experimental evidence of the probe’s magnetization distribution, the models simply provide possible explanations of observed contrast formation rather than offer precise magnetization reconstruction.
The goal of this work is address the aforementioned limitations of conventional MFM probes, by improving the resolution and sensitivity, by developing processes to tailor probes with predefined moment and coercivity, and by developing characterization and calibration methods for incorporation into theoretical models of image reconstruction. In this work, we will use our combined resources in thin film preparation, micromachining using focused ion beam (FIB) facility and expertise in magnetic force microscopy. By using this collaborative approach, the composition, probe size, cantilever mechanical sensitivity will be systematically studied and tailored over a wide range of parameters. The main feature that differentiates this project from other probe development efforts is the nature of the collaboration. This effort spans a very broad based development --- exploring the interrelationships of film chemistry, micromachining, MFM implementation and theory.
II. STATE OF THE ART IN MFM IMAGING
The efficacy of the MFM as a micromagnetic tool and the need for advanced tips is best illustrated by considering a case study of a micron-sized NiFe (Permalloy) island. Permalloy is one of the most important and well known magnetic alloys, owing to its unique characterisitics that make it ideally suited for field generation and detection. In fact, nearly all magnetic recording systems use Permalloy in one or more components and it is likely to be a major ingredient in the next generation of magnetic devices. Nevertheless, despite the enormous volume of work in studying this material, it is only now that we are beginning to understand the dependence of the magnetic properites with the material shape, edge irregularities, defects and pinning centers, and magnetization history.
The set of MFM images in figure 1 shows the micromagnetic evolution of a NiFe square element with a 3 um lateral dimension and about 26 nm thick. (some discussion on the MFM). Image is sensitive to change in M rather than M itself. Remanence image shows 7 domain with crosstie structure, evolving showing the pinning effects of the xtie. It progressive expands until a transtion at 60Oe is observed. This is interesting BUT: contrast is weak. Apart from the domain walls and xtie contrast, the interior regions show little or weak contrast. The resolution is limited, unable to measure the width of the 90 and 180 degree walls. Perturbation due to probe field evident in image at 60 Oe, which limits the increments that the field that could be applied in the evolution studies. Most importantly, the internal magnetization of the probe is unknown which makes it very difficult to recontruct the magnetization of the sample. If we have an ideal probe, what can we do? Can measure the domain walls, can derive the ratio anisotropy versus exchange field, can determine the precise switching and reversal behavior, i.e., nucleation at edges or at some interior regions. Can determine the micromagnetic everywhere and not only near the edges.
PROGRAM OBJECTIVES
The questions that will be addressed are:
What are the factors that determine the experimental resolution limit of magnetic force microscopy and how much can it be improved through the use of advanced MFM tips?
What are the ultimate limits on magnetic grain sizes, and what are the factors that influence them?
We propose to address these questions by systematically using a variety of coating chemistry, thin film fabrication techniques, and focused ion beam (FIB) technology combined with theoretical modeling to:
(i) understand the micromagnetics of coated thins on Si and SiN3 AFM probes, both as function of chemistry and geometry,
(ii) develop new models and experimental methods to characterize the MFM tips, and to quantify their interaction (and perturbation) with the sample,
(iii) elucidate the relationship between probe size and MFM resolution,
(iv) understand and control the mechanical properties of micro-machined AFM probes.
We will then use the MFM to investigate, in detail, the evolution of magnetic domain structures in nanostructured thin films with a spatial resolution previously unobtainable. We expect that the proposed studies will play a central role in our continuing efforts to elucidate magnetization reversal mechanisms and to develop predictive models of the reversal process. The proposed studies will also guide our efforts to develop new thin film media for high-density information storage and for magnetic sensor applications.
III.
Preliminary Results
A magnetic probe can be fabricated by coating an AFM tip made of Si3N4 or Si with an appropriate material of high permeability. Such a probe can be successfully used for MFM. But since an AFM probe consists of a structure with dimensions of many micrometers having a sharp point, the inherent resolution for MFM is not high the magnetic material covers too large an area. If such a tip were fabricated and then machined using a high resolution ( < 0.05 micrometer) FIB, it is possible to remove the magnetic material everywhere except at the apex of the probe, as shown in Fig.1. This has to be done with great care to avoid damaging the magnetic material at the tip apex, and we believe it can be done much better by developing a coating the Si3N4 or Si probe with a "stoplayer" of material having a thickness of only a few tens of nm, so that the tip can be imaged at high resolution with the FIB prior to the micromachining step. The stoplayer can subsequently be removed by chemical means after the micromachining step.
The figure below shows an example a sharp probe we created using FIB. The original probe was pyramidal with facets (having half angles of 25o front, 17o side and 10o front) which tapered to a point roughly 30 nm. The probes are batch-fabricated using selective etching techniques, so that the resulting facets are more or less imposed by the crystallographic property of the material. It is quite difficult to prepare probes that have arbitrary geometry, e.g., very high aspect ratios. However, as shown in the figure, a very sharp protrusion can be prepared by ultra high precision FIB milling of the commercial probe. In this example, the nominal radius of the tip is about 300 nm with a length of about 2 microns and a diameter of about 100 nm.
Left: FIB micromachined MFM tip. Right. Schematic of the geometry of untrimmed commercial MFM tip (Digital Instrument spec).
One would suspect that a probe were coated with a magnetic material, then it is conceivable that a small amount of magnetic material would be left at the protrusion, and thereby satisfy our requirement of small magnetic volume, i.e., an ultra high resolution probe. However, things are not so simple. In some cases, as our initial results suggest, the MFM contrast are very different from what one would expect by merely reducing the effective volume size while keeping the magnetization unchanged. A number of important questions emerged from our initial work. First, we suspect that the process of material removal changes the magnetostatic energy of the magnetic film, which causes it to remagnetize in some, thus far, unknown manner. Secondly, we suppose that a reasonable alternative is to micromachine first and later deposit the magnetic layer. We have successfully implemented this approach and the results are very promising. In both cases, the nascent issue is how to control the films under those conditions. The exact nature of the remagnetization process of micromachined magnetic films and the micromagnetics of magnetic thin films deposited on protrusions can be understood and potentially controlled, only through a careful systematic study. Finally, there is the issue of ion implantation. The process of FIB causes a certain dose of gallium (or other atoms) to be implanted on the probe. While the use of FIB micromachining on a coarser (1 micrometer) scale is already being used commercially for the finishing step in the production of readwrite heads for magnetic disk drives, the effects on the magnetic properties on probes at the submicron length scales are still unexplored. Therefore, while a major portion of this study will be devoted to determining the relationship between probe tip morphology and the field distribution associated with the tip, substantial emphasis will be placed on extending the limits of FIB techniques on magnetic thin films.
Expected Outcome of This Research :
The purpose of the proposed research is to develop improved methods for understanding and characterizing the magnetic properties of nanostructured materials. This research should have significant implications for a variety of advanced technologies, including high-temperature permanent magnets, extremely high-density information storage applications and magnetic sensor applications. In each of these cases, the atomic-scale structure of the materials plays a dominant role in the macroscopic magnetic behavior, especially as regards the thermal and magnetic stability. The magnetic properties of nanostructured materials are usually controlled by the behavior of the fundamental “magnetic building blocks” - magnetic grains or clusters - and their interactions. The typical magnetic grain-size of these technology important materials is in the range of 5-10 nm. To have a better understanding these magnetic materials, it will be necessary to control and characterize the structure of the materials on the nanoscale level. We expect that the proposed activities will also have impacts as follows: The advanced MFM tip can serve as a small and sensitive magnetic sensor or a local magnetic field source for a variety of applications. It may be possible to use a similar approach for the improvement of other long-range scanning-probe force microscopy applications, e. g., to obtain images of ferroelectric domains.
(1) "Magnetic coating materials
The control of the size and shape of magnetic materials"
A. Fabrication and Characterization of Nanostructured Magnetic Clusters
The goals of this research are to synthesize and understand magnetic single domain particles and thin films with nanoscale magnetic particles that suitable for MFM applications. The desired properties include a very large magnetic anisotropy or the superparamagnetic instability inherent in extremely small magnetic particles.
1. Fabrication of Nanostructured Films
There are many ways to fabricate nanostructured magnetic films. We describe here a few of the methods that we have developed.