No abstract

Low-frequency noise was measured in the frequency range from 0.1Hzto10kHz on a variety of commercially available magnetic sensors. The types of sensors investigated include anisotropic magnetoresistance (AMR), giant magnetoresistance (GMR), and tunnel magnetoresistance (TMR) effect devices. The 1∕f noise components of electronic and magnetic origin are identified by measuring sensor noise and sensitivity at various applied magnetic fields. Commercial magnetometers typically consist of four elements in a Wheatstone bridge configuration and are biased with either a constant voltage or current. Voltage fluctuations at the sensor output are amplified by a pair of battery powered low-noise preamplifiers and input to a spectrum analyzer. A two-channel cross-correlation technique is used when the performance of a single preamplifier is not sufficient. For the AMR and GMR sensors investigated, both electronic and magnetic components contribute to the overall sensor noise. Maximum noise occurs at the bias field which gives maximum sensitivity. The noise of TMR based sensors is primarily due to resistance fluctuations in the tunnel barrier, having little to no field dependence. The best low-field detectivity of the sensors that have been measured is on the order of 100pT∕Hz0.5 at 1Hz.

Thin structures of alternating magnetic and nonmagnetic layers with a total thickness of a few hundred nanometers exhibit a phenomenon known as giant magnetoresistance. The resistance of microfabricated giant magnetoresistors (GMRs) is dependent on the strength of an external magnetic field. This paper examines magnetic labeling methodologies and surface derivatization approaches based on protein-protein binding that are aimed at forming a general set of protocols to move GMR concepts into the bioanalytical arena. As such, GMRs have been used to observe and quantify the immunological interaction between surface-bound mouse IgG and alpha-mouse IgG coated on superparamagnetic particles. Results show the response of a GMR network connected together as a set of two sense GMRs and two reference GMRs in a Wheatstone bridge as a means to compensate for temperature effects. The response can be readily correlated to the amount of the magnetically labeled alpha-mouse IgG that is captured by an immobilized layer of mouse IgG, the presence of which is confirmed with X-ray photoelectron spectroscopy and atomic force microscopy. These results, along with a detailed description of the experimental testing platform, are described in terms of sensitivity, detection limits, and potential for multiplexing.

Commercially available superparamagnetic nanospheres are commonly used in a wide range of biological applications, particularly in magnetically assisted separations. A new and potentially significant technology involves the use of these particles as labels in magnetoresistive assay applications. In these assays, magnetic bead labels are used like fluorescent labels except that the beads are excited and detected with magnetic fields rather than with photons. A major advantage of this technique is that the means for excitation and detection are easily integrable on a silicon circuit. A preliminary study of this technique demonstrated its basic feasibility, and projected a sensitivity of better than 10−12 molar [Baselt et al., Biosensors Bioelectronic 13, 731 (1998)]. In this article we examine the theoretical signal to noise ratio of this type of assay for the special case of a single magnetic bead being detected by a single giant magnetoresistive (GMR) detector. Assuming experimentally observed and reasonable parameters for the magnetic label and the sensitivity of the GMR detector, the signal to noise ratio is calculated to be greater than 5000:1 for detection of a single 1 μm diameter magnetic microsphere immobilized on the surface of a 1 μm×1 μm GMR sensor. Based on this large signal to noise ratio, the detection format should be applicable to more complicated assays where linear quantification is required or to assays requiring significantly smaller beads. Detection of microsphere labels approaching 10 nm may be possible upon further technological advances.

Pinned spin-dependent tunneling devices were fabricated and tested in a mode suited for low-field sensing. The basic structure of the devices was NiFeCo125/Al2O325/CoFe70/Ru9/CoFe70/ FeMn125 (in Å). This structure had a tunneling resistivity of 110 MΩ μm2 and exhibited a 20% magnetoresistance when a field was swept along the easy direction of the soft electrode. High sensitivity, low hysteresis operation was achieved by applying a bias field orthogonal to the easy axis. A sensitivity of 3%/Oe with negligible hysteresis was observed using this mode of operation. A sensor using this type of material was designed to achieve a minimum resolvable field in the picotesla range. The sensor consists of a bridge with four elements, each having 16 tunnel junctions in series. A signal-to-noise ratio of 1:1 at 1 pT (10−8 Oe) is possible assuming achievable values for the tunneling resistivity, device size, bias level, and sensitivity.