A Grains-of-Rice Magnetic Gradiometer, 14-R9624

Printer Friendly Version

Principal Investigators
Jerome A. Helffrich
Todd H. Goyen
Michael A. Dooley

Inclusive Dates:  04/01/06 – 04/01/08

Background - The physics and engineering involved in observing the quantum level transitions in rubidium vapor were investigated using a laser-based optical absorption technique and scaled down to a handheld portable unit. The objective was to develop a magnetometer based on measurement of the frequencies separating these quantum levels and to use this magnetometer as a portable mineral and power line detection device.

Approach - The objective of this project was to have a field-testable, magnetic gradiometer smaller than current state of the art. This was achieved by designing and building a sensor front end to support a miniaturized magnetometer that implements a pump-probe interrogation scheme. Part of the work was to decide which interrogation method was most suitable for a portable sensor. The team selected what is known as the MX technique, in which a laser passes through a bulb filled with rubidium vapor and is incident on a photo diode. Two excitation coils wrapped around the bulb impart an oscillating magnetic field, which causes the laser absorption by the rubidium to vary. This setup allows for the electronics to be split into an ultra stable laser driver and a separate alternating current drive for the excitation coils with relaxed noise requirements because it will no longer affect the laser drive frequency.

Accomplishments - Researchers successfully demonstrated a rubidium (Rb) vapor total-field magnetometer that pushes existing size and weight standards down to new levels. This device, which is based on a millimeter-sized bulb of Rb vapor and microprocessor-driven control circuits, is exceptionally well suited to use in portable applications such as unmanned aerial vehicle based mine detection, surveillance and remote monitoring. It is portable and sensitive enough to be used in walkover surveys of buried power line and telecommunications infrastructure to locate cable routes and to infer depth of burial from signals emitted by them.

The magnetometer comes in two parts — a sensor and an electronics package — allowing the sensors to be remotely located, away from sources of electronic noise. The control electronics package is approximately 0.75 x 4.0 x 2.5 in.; the sensor head is approximately 0.75-in. diameter x 2.25-in. long, roughly the size of an AA battery, and weighs about 120 grams (Figures 1 and 2). The pre-prototype system shows a direct current sensitivity of 300 picoTesla, or at least an order of magnitude better than commercial fluxgate magnetometers, with a designed bandwidth of 1 kilohertz and a slew rate of 285 nanoTesla per millisecond. The signal-to-noise ratio is now 600 volts per volt, a drastic improvement since testing began. A plot of the in-phase and quadrature signals shows a gain of more than 120 million volts per watt of laser power. The system consumes three watts at steady state, most of which is used to heat the vapor cell to 70 degrees C. Despite the obstacle presented by this power requirement, the current system can run for more than six hours on the included lithium batteries, realizing the development of a truly field-portable system.

Figure 1. Engineering drawing of the optical system contained in the sensor head.

Figure 2. Internal view of the sensor head assembly. From left to right, the vertical cavity surface emitting laser diode light passes through a linear polarizer and then a quarter waveplate before entering the 87Rb cell, which is wrapped in fiberglass tape. Light passing through the cell is collected by the photodiode.

An experiment to test the hyperfine absorption was done by sweeping the laser frequency through the two ground state transitions (Figure 3). The separation between these peaks is 6.8 gigaHertz. The large peak width is due to Doppler broadening as well as the pressure of buffer gases within the Rb cell. The peak heights are due to the combination of laser power and cell temperature. To maximize the output signal, one must balance the alkali atom number, dependent on cell temperature, with laser power such that there is one atom for every incident photon. In the MX magnetometer, the laser is tuned to the center of the F = 2 peak; the cell temperature is adjusted to move the amplitude of the received signal to half the maximum input voltage. This ensures that the resonance signal will have the maximum amount of headroom and will not clip.

Figure 3. Absorption curve showing the F = 1 and F = 2 hyperfine transitions in 87Rb. In the MX magnetometer, the laser operates on the peak of the F = 2 transition.

The magnetometer was setup in a zero gauss chamber under a direct current magnetic field and the MX frequency was swept to record a quadrature (dispersion) curve for the system, resulting in quadrature and in-phase signals (Figure 4). The sensitivity of the system is defined as the full width at half maximum (FWHM) of the peak divided by the analog-to-digital converter resolution, or the signal to noise, whichever is lower. In this test, the width is approximately 5 kiloHertz and the signal to noise is approximately 70 Volts/Volt. This gives a sensitivity of 10 nanoTesla over a bandwidth of 1 kiloHertz, or 300 picoTesla/(Hertz)1/2.

Figure 4. Dispersion curve for the optically-pumped magnetometer system with a static direct current bias field. Note that the MX phase is not exactly zeroed and the In-phase and Quadrature channels are not correctly separated.

The measurement above was repeated for a number of known direct current magnetic fields, recording the zero crossing in the quadrature curve each time. This experiment allowed for the determination of the 87Rb gyromagnetic ratio (Figure 5). The ratio was measured to be 701.6 ± 18.0 kHz per gauss, which compares well with the actual value of 698.997 kiloHertz/gauss. The large error bars are due to the limited accuracy of 0.75 percent of the commercial magnetometer that was used to calibrate the field.

The work outlined here has led to the construction of a functional scalar optically pumped magnetometer. While sensitivities approaching the theoretical limit have not yet been demonstrated, the system is operating at a level acceptable for future use. Early testing has produced measurements that agree well with the expected values. The research conducted has bootstrapped SwRI into the field of atomic vapor magnetometry. While only a pre-prototype, our system should allow for future funding opportunities, allowing SwRI to capitalize on the invaluable knowledge that has been gained to further refine and miniaturize the magnetometer.

Figure 5. Results from the measurement of the 87Rb gyromagnetic ratio. The ratio was measured to be 701.6 ± 18.0 kiloHertz/gauss, which compares well with the actual value of 698.997 kiloHertz/gauss.

2008 Program Home