Antarctic NMR

NMR equipment

Members of the group have made seven trips to Antarctica to study the content and structure of sea ice in the polar ice cap, in the area around Scott Base. The trips took place from 1994–2009. The people involved are listed below:

  • 1994 (November): Paul Callaghan, Craig Eccles
  • 1995 (October): Joe Seymour, Paul Callaghan, Craig Eccles
  • 1997 (October): Robin Dykstra, Joe Seymour, Paul Callaghan, Craig Eccles
  • 1999: Suzanne Furkert, Robin Dykstra, Paul Callaghan
  • 2002: Mark Hunter, Ocean Mercier, Paul Callaghan
  • 2006: Paul Callaghan, Miang Lim, Robin Dykstra, Mark Hunter
  • 2009: Robin Dykstra, Meghan Halse, Achim Gadke, Evan McCarney

This page illustrates some of results from the first trip in 1994, along with the experience of carrying out NMR research in such a remote environment.

About Antarctica

The pack ice surrounding Antarctica has a significant impact on global climate, as it has a surface area of almost 20 million square km and an albedo (reflecting power) of 95%. Consisting of frozen sea water, pack ice, or sea ice, is formed gradually over several months during the Antarctic winter. Sea ice has a complex structure and is far from being a homogenous material.

As the sea water freezes, concentrated salt water accumulates in small cavities called brine pockets. Although they are generally less than 0.5 mm in diameter they can extend for many millimetres vertically. During the winter, gravity causes much of the trapped brine to move down through the ice, and results in a non-uniform salinity profile.

The presence of the brine pockets gives the sea ice unusual mechanical properties. An understanding of these properties and their implications for annual sea ice formation and break up can be gained by macroscopic stress measurements, or by inference from microscopic analysis of sea ice morphology. Until now, microscopic studies have been practically non-existence, apart from some laboratory studies on home-grown sea ice. However given the unusual conditions of Antarctic sea ice formation, such as variable temperature gradients, long growth periods and wave action, an on-site analysis would seem to be preferable.

NMR as a probe of sea ice

NMR provides a non-invasive technique for studying sea ice morphology using imaging or bulk measurement techniques. Not surprisingly, previous work has been done with high field instruments, however the possibility of taking such a facility to the Antarctic is remote.

Given the low level of electromagnetic interference in Antarctica, NMR studies which use the Earth's magnetic field would seem to be a good alternative. Although an imaging system based on the Earth's field has been demonstrated, the resolution is too coarse to allow detailed studies of the microscopic sea ice structure.

The alternative is to use NMR to provide bulk measurements of liquid water content and relaxation times. More ambitious experiments using pulsed-gradient spin-echo NMR can be used to obtain information on restricted self diffusion, from which average brine pocket dimensions can be inferred.

NMR in Antarctica

The pulse sequence used to observe a free induction decay in the Earth's magnetic field.

In the spring of 1994 we spent several days on sea ice, 25 km from Scott Base, in Antarctica's McMurdo Sound. An Earth's field NMR apparatus was assembled in a polar tent and measurements were made on ice cores extracted from the pack ice.

Each sample was placed inside the NMR probe which was housed in a snow cave to reduce the effects of generator and computer interference.

The probe consisted of a large solenoidal polarizing coil, capable of producing a field of approximately 0.03 Tesla, an excitation coil and two receiver coils. The polarizing coil was isolated from the sample by a glass dewer so as to minimise heating effects.

The orientation of the various coils in the NMR probehead with resepct to the Earth's field.

Free induction decays were acquired as shown in the diagram below:

The polarizing coil was energised for approximately 5 seconds to produce a bulk magnetization in the sample. The polarizing current pulse was then switched off, but sufficiently slowly so that the magnetization had time to redirect itself along the Earth's magnetic field - almost vertical and of strength 65 µT in this location.

A 2 ms, 90° pulse at the Larmour frequency of 2.76 kHz, was then applied to the sample via the excitation coil to tip the magnetization back into the horizontal plane where it could be subsequently detected as an induced emf in one of the receiver coils. T2 relaxation data was obtained using a standard spin-echo pulse sequence.


All the necessary switching pulses, as well as the audio-frequency pulse applied to the excitation coil, were originally generated by a Tecmag-Aries system and Macintosh IIci computer.

The induced signal (~10uV) was amplified by a home built, low-noise, differential amplifier and bandpass filter. External, common-mode interference was minimised through the use of a second receiver coil. The resultant signal was digitized and accumulated by the Tecmag system in the usual way.

Note that unlike a conventional NMR system the uniformity of the polarizing field is not especially critical since the FID is acquired in the Earth's field, which in isolated areas such as this, is uniform to better than 1 part in 10^6. This and the fact that we were working at such low frequencies made for a compact and relatively inexpensive spectrometer.


An example of the quality of the NMR data which can be obtained with this equipment can be seen in figure 9 which shows free induction decays from a 500 ml water sample and a typical ice core.

In this initial set of T2 experiments 8 sea ice samples were taken from depths ranging from 60 to 1810 mm (just above sea level). For each sample a new core was extracted to minimise temperature changes. Even so, a temperature change of 1-2 °C was typical during each 20 minute experiment. A reference measurement using (warm) water was made before each T2 experiment to check on the stability of the apparatus and to provide a reference signal from which the water content in each sample could be determined.

In addition the conductivity of each sample was measured so as to determine its salinity, as shown below.

Although the water content in the ice sample was low, the FIDs obtained were of sufficient quality to determine transverse relaxation times. This data is summarised below, in plots of temperature and T2 as a function of depth.


Camp on the sea ice
Camp on the sea ice.

It appears that the water fraction and salinity measurements are in qualitative agreement, although the reason for the drop at middle depths is currently not clear.

As expected, temperature increases with depth, this just reflects the temperature gradient between the air (-10 °C to -15 °C) and the sea water (-1.7 °C).

T2 will depend on a number of factors including brine content, temperature and pore size. One might expect T2 to increase with depth since temperature does as well, but a downward trend is observed. This could be due to chemical exchange at the solid-liquid interface in the brine-pockets. This would explain the observed trend in T2 since the exchange rate will increase with temperature.

Further experiments

In 1995, 1997, 1999, 2002, 2006 and 2009 we returned to the Antarctic with improved equipment. As well as repeating the experiments discussed above, we made restricted diffusion measurements to determine the dimensions of the brine pockets at various depths.

Our latest apparatus has a probe that can be inserted directly into the sea ice, and avoids the need to remove core samples. In 2002 the apparatus used was the prototype of the magritek Terranova Earth's Field spectrometer with a specialised probe incorporating the gradient coils necessary for measurement of brine diffusion. In 2006 a fully commercial Terranova probe was used and permanent magnet systems were trialled. In 2009 further permanent magnet based systems were used.

Details of the work can be found in the publications listed below.


T.G. Haskell and W.H. Robinson. A sensitive and robust strain-meter for ice studies, Cold Regions Science and Technology, 23, p99 (1994).

P. Langhorn. Alignment of crystals in sea ice due to fluid motion, Cold Regions Science and Technology, 12, p197 (1986).

C. Richardson and E. Keller. The brine content of sea ice measured using an NMR spectrometer, Journal of Glaciology, 6, p89 (1966).

G. Planinsic, J. Stepisnik and M. Kos. Relaxation-time measurement and imaging in the Earth's magnetic field, JMR Series A, 110, p170 (1994).

P.T. Callaghan. Principles of NMR microscopy, Clarendon Press, Oxford (1991), Tecmag inc., 6006 Bellaire Blvd, Houston, TX 77081 USA.

P.T. Callaghan and C.D. Eccles. NMR studies of Antarctic Sea Ice, Bulletin of Magnetic Resonance 18 62-64 (1996)

P.T. Callaghan, C.D. Eccles and J.D. Seymour. An Earth's field NMR apparatus suitable for Pulsed Gradient Spin Echo measurements of self-diffusion under Antarctic conditions, Rev.Sci Instr. 68, 4263-4270 (1997)

P.T. Callaghan, C.D. Eccles, T.G. Haskell, P.J. Langhorne, and J.D. Seymour. Earth's field NMR in Antarctica: A Pulsed Gradient Spin Echo NMR study of restricted diffusion in Sea Ice, J Magn .Reson 133, 148-154 (1998)