Modified Viscometer for Employment in Torque Magnetometry

This invention relates to a method suitable for measuring the magnetic properties of a workpiece.

Methods suitable for measuring the magnetic properties of a workpiece are preferably simple to employ, efficient, inexpensive, and provide a desired measurement accuracy. One important magnetic property of the workpiece that we are especially interested in applying a preferred method, is that of measuring the magnetic anisotropy (Ku) of the workpiece, for example, the magnetic anisotropy of magnetic thin films.

There exist several methods for measuring the magnetic anisotropy of magnetic thin films. The most direct and dependable known methods are based on torque magnetometry. Torque magnetometry comprises a method of measuring the torque exerted on the magnetic thin film by an applied magnetic field, as it slowly rotates around the thin film, while maintaining a quasi-static equilibrium. The measured torque parameter, in turn, may be used to determine the required magnetic anisotropy (Ku) of the thin film.

Torque magnetometry, in turn, requires a torque magnetometer. Much attention has been given to the design of the torque magnetometer. See, for example, R. F. Penoyer, Rev. Sci. Instrum. 30, 711 (1959); A. A. Aldenkamp, C. P. Marks, and H. Zijlstra, Rev. Sci. Instrum. 31, 544 (1960; F. B. Humphrey and A. R. Johnston, Rev. Sci. Instrum. 34, 348 (1963); D. M. Paige and B. K. Tanner, J. Phys. E 15, 128 (1982); M. Tejedor, A. Fernandez, B. Hernando, and J. Carrizo, Rev. Sci. Instrum. 56, 2160 (1985).

We have found that the various designs of torque magnetometers may typically require a sensitive and delicate setup consisting of fine torsion fibers or linear springs, and a mechanism to measure the torque by either measuring the angular displacement or by applying a known counter torque to achieve zero net displacement, i.e., a null-type torque magnetometer. We have found, moreover, that these torque magnetometers, requiring as they do such a sensitive and delicate setup, may work against the cost and efficiency objectives of a preferred method, noted above, for employing torque magnetometry to measure the anisotropy of the thin film.

The present invention, in contrast, provides a novel alternative to the employment of the known torque magnetometers--an alternative that can preserve all the virtues of torque magnetometry, yet realize the preferred method objectives of efficiency, low cost and accurate measurement.

The present invention, accordingly, comprises a method for measuring magnetic properties of a workpiece, comprising:

(1) providing an assembly, the assembly comprising

(a) a viscometer, the viscometer comprising

(i) an external spindle, and

(ii) means for attaching the workpiece to the spindle; and

(b) a magnet which can develop a magnetic field enveloping the workpiece;

(2) positioning the workpiece so that it is attached to the spindle and enveloped by the magnetic field;

(3) actuating the viscometer, thereby imparting movement to the workpiece; and

(4) measuring a resultant torque thereby imparted to the workpiece.

The present invention, as defined, has an advantage that it may be employed for measuring the magnetization (M) of the workpiece, or for measuring particle orientation in various coated particle media, as well as for measuring the workpiece anisotropy (Ku). Further, the workpiece may be any material that has some appreciable magnetic capability, whether in the form of magnetic thin films, tapes, e.g., floppy disks, high temperature superconductor materials, or magneto-optic media, such as Co--Cr, Co--Ni, rare-earth-transition metals, etc.

The invention is illustrated in the accompanying drawing, in which:

FIG. 1 is a schematic diagram of an assembly of the invention;

FIG. 2 shows details of a viscometer used in the FIG. 1 assembly;

FIG. 3 is a diagram of further features of the FIG. 2 viscometer, as modified in accordance with the present invention;

FIG. 4 is a torque curve generated in accordance with the present invention; and

FIG. 5 is a plot of output voltage at different torsion angles, for determining the spring constant of the FIG. 2 viscometer.

Attention is now directed to the FIGS. 1-3, which show an assembly 10 (and details) of the present invention. The structure of the assembly 10 is first disclosed, followed by its operation, i.e., the method of the invention.

The FIG. 1 assembly 10 comprises a rotational viscometer 12, the viscometer 12, in turn, comprising an external spindle 14, and means 16 for attaching a magnetic workpiece 18 to the spindle 14. The workpiece 18 may be positioned so that it is enveloped by a magnetic field generated by an electromagnet 20. The assembly 10 further comprises a Hewlett Packard 310 series computer 22, a Hewlett Packard model 7090 measurement plotting system 24, and an external trigger switch 26.

Details of the rotational viscometer 12 are provided in FIG. 2. The viscometer 12 preferably is a Brookfield Instruments model LVTD viscometer, having a torque resolution of better than 0.6 dyne cm. The viscometer 12 is of the rotational type, as stated, and may be conventionally employed to measure the torque required to rotate the spindle 14 immersed in a given fluid. For our method, in sharp contrast, the viscometer 12 measures the torque experienced by the workpiece 18, attached to the spindle 14, and rotating in the externally applied magnetic field generated by the electromagnet 20.

The spindle 14 may be driven by a synchronous motor 28, which can be set to rotate at various speeds ranging from 0.3 rpm to 60 rpm. To achieve the quasi-equilibrium conditions required in the torque magnetometric measurements, however, a lower motor speed such as 0.3 rpm or 0.6 rpm, is preferred. A shaft 30 of the motor 28 may be connected to the spindle 14 through a calibrated beryllium-copper spring 32. The deflection of the spring 32, proportional to the amount of torque, may be detected by a rotary variable displacement transducer (RVDT). The transducer, in turn, provides an analog output signal, typically in the range of 0-10 mV or 0-1 V full scale. Knowing the spring 32 calibration constant (set forth below in an Appendix), the output voltage signal can be readily converted into dyne cm for torque. Other major mechanical components of the viscometer 12, for example, the housing, gear train, pivot cup, jewelled bearing and clutch, are so labelled in FIG. 2.

Further details on the viscometer 12 are provided in FIG. 3. In particular, FIG. 3 shows the spindle 14, and the means for attaching the magnetic workpiece 18 to the spindle 14, where the means is identified in FIG. 3 as a workpiece-holder 34. The spindle 14 and the workpiece-holder 34, taken together, are constructed in accordance with a novel design, discovered by us, for the purposes of modifying a conventional viscometer, to the end of the present invention. Accordingly, we may reference this modified viscometer by the generic expression, "torquemeter".

A portion of the modified viscometer 12 is shown in FIG. 3, and comprises a three-part assembly viz., a stainless steel spindle 36, a ceramic extension tube 38, and the thin film workpiece-holder 34 made of Plexiglas. The stainless steel spindle 36 is a cylindrical rod, preferably about 7.6 cm in length and 3.18 mm in diameter. At one end of the spindle 36, there is a cup with its outer diameter preferably about twice the diameter of the rod. This cup may be threaded from the inside, allowing the spindle 36 to screw directly onto the rotating assembly of the viscometer 12.

A straight ceramic tube 38 (4.75 mm o.d., 3.18 mm i.d., and having a camber of less than 1.5 mm over its 30.5 cm length) may be used as an extension between the stainless steel spindle 36 and the workpiece-holder 34. This extension tube 38 allows the workpiece 18 to reach the magnetic field generated by the electromagnet 20. The long and slender part of the stainless steel spindle 36 up to the neck of the cup may be snugly inserted in the ceramic tube 38 and glued with epoxy, keeping the alignment of the assembly straight. A preferred ceramic tube used for the extension part is made by McDanel Refractory Porcelain Co., PA, using material called Mulite (MV-20) consisting of magnetic or metallic impurities at the minimum level. Note that it is important that no metallic material be used for the parts of this assembly that are close to the magnetic field. Otherwise, large errors due to the eddy currents generated during the rotation of the assembly may result.

The workpiece-holder 34 preferably is made of Plexiglas. It may be made out of a rod 12.7 mm and 10.2 cm long. A 3.8 cm deep hole may be drilled at the center from the top to allow the ceramic extension tube 38 to insert snugly into it. A nylon screw can hold it in place. At the other end of the workpiece-holder rod, a diametrical slit about 0.76 mm wide and 7.6 mm deep may be cut vertically into the rod. The workpiece 18 may be placed in this slit, and may be held in place by another nylon screw coming from the side. Workpieces with flexible substrates can be sandwiched between a pair of thin microscope cover glasses, before attaching to the workpiece holder 34.

As identified above in FIG. 1, the assembly 10 comprises an electromagnet 20. The electromagnet 20 preferably comprises a Varian CFT-20 electromagnet with water-cooled coils. This electromagnet has 17.8 cm diam poles tapered down to a 10.2 cm diameter. When mounted, this forms a gap of 2.0 cm between the poles. In this configuration, the electromagnet 20 provides a dc field of about 16 kOe in the gap.

The structure of the assembly 10 of the present invention has now been disclosed, and we turn to its operation, i.e., the method of the invention. In accordance with one aspect of the method summarized above, a workpiece 18 comprising a C.sub.o --N.sub.i thin film is attached to the workpiece-holder 34. The workpiece 18 is then rotated in the magnetic field developed by the electromagnet 20.

The resulting output from the RVDT of the viscometer 12 is stored into a buffer memory of the HP 7090 measurement plotting system, along with the corresponding time for each data point. On activating the trigger switch 26, the data storage begins. This switch may be manually triggered when, during the rotation of the workpiece 18, the film plane lies normal to the applied field (i.e., at 0.degree.), as determined by a careful visual inspection. For accuracy, one can employ a photo-switch, an LED, and a mirror arrangement for automatic triggering. At the end of one revolution, all the data are transferred to the HP 310 computer. From each pair of data consisting of the value of RVDT output in volts and the corresponding time in seconds, the torque in dyne cm and the position of the workpiece with respect to the applied fields in degrees, may be calculated. This is done through a program written in BASIC which incorporates the calibration constant and the rotational speed of the workpiece. The calibration procedure is described in the Appendix below, and an exemplary BASIC program is also listed below.

The magnitude of the intrinsic anisotropy K.sub.u of a magnetic thin film and its direction at an angle .delta. with respect to the film normal can be obtained from the torque curves using Eqs. (1) and (2), respectively:

Here, the positive sign is taken when .pi./4.ltoreq..alpha..sub.0 .ltoreq..pi./2 and the negative sign is taken when 0.ltoreq..alpha..sub.0 .ltoreq..pi./4.

For this, one needs to know the peak torque T.sub.p and the zero-torque position .alpha..sub.0 (i.e., the angle between the normal of the film and the applied field at which torque is zero) from the torque curves and the total shape anisotropy K.sub.s of the film. However, knowing the saturation magnetization M.sub.s and the volume V of the film, its K.sub.s is given by 2.pi.M.sub.s.sup.2 V.

Since this torque magnetometer is not a null-type, a correction for the angular position of the workpiece 18 with respect to the applied field direction, may be required. This correction is dependent upon the amount of torque, and is related through the spring constant k of the torquemeter, as shown in Eq. (3):

Here, at a given time t, .alpha..sub.observed =.omega.t, .omega. being the rotational speed of the workpiece, and T.sub.0 is the torque at t=0. Knowing k, .alpha..sub.corrected at all times can be easily calculated from Eq. (3). We compute this correction through a subroutine included in the BASIC program referred to earlier. The spring constant k can be experimentally determined as discussed in the Appendix.

In FIG. 4, we show the torque curves of a Co-Ni thin-film workpiece approximately 90-nm thick deposited at an oblique angle on a Kapton substrate by thermal evaporation. The workpiece used for torque measurements is in the form of a 16.5-mm-diam disk. The solid curve in the Figure represents torque vs .alpha..sub.observed curve, as obtained from the torque magnetometer, and the dashed curve is a plot of torque vs .alpha..sub.corrected.

When the direction of the intrinsic anisotropy in the film is known a priori, or in special cases where the anisotropy is either in-plane or perpendicular to the film, the value of T.sub.p alone from the torque curve curves is sufficient in determining K.sub.u of the film. In such cases, one does not need to correct for .alpha..

A small piece of a uniform wire made of pure nickel or iron of known radius r and length 1, may be used as a calibration standard. It is sandwiched between two thin glass slides, and mounted in the workpiece-holder 34 of the torquemeter, such that the length of the wire always lies in the plane of rotation. The total anisotropy K.sub.cal of this calibration workpiece is simply its shape anisotropy given by

where M.sub.s now refers to the saturation magnetization of either pure nickel (485 emu/cc) or pure iron (1671 emu/cc), whichever is the case. This calculated value of K.sub.cal is equivalent to the zero-to-peak value in volts of the output tracing from the torquemeter. A calibration standard made of pure nickel wire 0.10 mm in diameter and 10.0 mm in length produces a peak value of about 35.5 mV in the output. Using Eq. (4), the anisotropy K.sub.cal of this calibration sample may be calculated to be about 58.4 dyne cm. From this, we obtain a calibration constant of 0.608 dyne cm/mV.

The spring constant k of the torquemeter is determined by torsioning the spring of the torquemeter by a known amount (in degrees) and noting the corresponding output from the torquemeter (in mV). A protractor placed directly underneath the workpiece-holder 34 of the torquemeter, and a pointer temporarily attached to the workpiece-holder 34, indicates the torsion angle. With the motor 28 switched off, the spindle 14 of the torquemeter is allowed to reach equilibrium indicated by zero volt output. The protractor is now aligned with the pointer to indicate 0.degree.. The pointer is then manually moved to some new position, say 5.degree., and is held at that position for a few seconds. The output of the torquemeter connected to an X-Y recorder indicates a jump in the voltage, corresponding to this torsion of 5.degree..

In FIG. 5, we show a plot of output voltages corresponding to torsion angles ranging from +30.degree. to -30.degree., in steps of 5.degree.. From this plot, we obtain a spring constant of about 12.4 mV/deg or 7.54 dyne cm/deg using the calibration constant mentioned above. It is also noticed in this experiment that the torquemeter output is approximately linear up to +90.degree. torsion angle for a clockwise (cw) rotation, and up to -40.degree. for a counterclockwise (ccw) rotation. This corresponds to a torque limit of approximately 680 dyne cm for cw rotation, and 300 dyne cm for ccw rotation. ##SPC1##