Oil in and under Ice Detection using Nuclear Magnetic Resonance

The role of remote sensing during an oil spill is a vital component to the success of a response and often require multiple platforms and sensors during a response (API, 2013). Although several existing sensors are used operationally for marine oil spills, an area of continued research is the detection of oil in or under ice. Arctic ice-prone waters present unique challenges to existing remote sensing methods commonly used during marine oil spill responses (Dickins et al., 2008).

Current operational methods to determine if oil has become trapped in or under ice require deploying personnel onto the ice. This method is a challenge to both safety and time, with the ability to only cover a relatively small area in a single day. The benefits of remote sensing methods are that they do not require placing personnel on ice and they can cover large geographic areas in a short period of time. There are a number of ongoing research efforts to leverage the safe operations and areal coverage of remote sensing to detect oil in and under ice and efforts continue to determine the best methods to exploit the suite of surface and subsurface remote sensing tools for the Arctic.

The International Oil and Gas Producers (IOGP) is conducting the Arctic Oil Spill Response Technology joint industry program (JIP) to improve the technologies and methodologies to respond to oil spills in Arctic waters (www.arcticresponsetechnolgy.org). Platforms include both surface and subsurface remote detection technologies with sensors that range from visible to infrared and acoustics (Pegau et al., 2016).

Since 2006, ExxonMobil has been researching the use of Nuclear Magnetic Resonance (NMR) in the Earth’s magnetic field to detect oil in and under ice from an airborne platform. Early research advances included the detection of a crude oil surrogate in the presence of water and simulated ice with a 6 m NMR antenna in the Earth’s magnetic field, for the first time (Palandro et al., 2015). The final phase of research and development occurred in late 2016 in Newfoundland, Canada. A NMR antenna was operated from a helicopter to detect a crude oil surrogate under simulated ice to test the sensitivity of the ruggedized system.

The background on NMR as well as the application of detecting oil in and under ice has previously been documented elsewhere (e.g., Chavez et al., 2015). In brief, NMR is used to characterize matter that contains atomic nuclei with spin and angular momentum (e.g., carbon-13 and hydrogen). A common use of an applied-field NMR is in medicine as a Magnetic Resonance Imaging (MRI) device. For the application of detecting oil in and under ice in the Earth’s magnetic field the characterization of hydrogen is used and is accomplished by:

1. Placing a sample in a static magnetic field to align the magnetic moments (i.e., the Earth’s magnetic field)

2. Perturbing the magnetic moments of the hydrogen protons with one or more excitation pulses

3. Detecting the magnetic field generated by the nuclei as they return to equilibrium (precess)

The challenge of working in the Earth’s magnetic field is that, in comparison to a typical magnetic field for an MRI, the Earth’s magnetic field is weaker than the typical MRI field by ~20,000 times. To compensate for this large discrepancy, novel NMR techniques are required.

The discrimination of the proton NMR signals from oil and water exploits the difference in the time-dependent response of a hydrogen proton in water compared to a hydrogen proton in oil. The relaxation time, T₂, is the time required for the transverse magnetic polarization to drop to zero. The T₂ value of protons in oil is significantly shorter than that of protons in water and can therefore be exploited to distinguish the NMR signals. A key factor allowing the use of this technology is that ice, which is the limiting detection challenge for all other technologies, is in effect invisible to NMR. This is due to the rigid structure of ice, which has a T₂ relaxation time too short to be detected.

Previous testing of the ability of NMR to discriminate the hydrogen proton of oil from water has taken a stepwise approach by first ensuring that the T₂ relaxation times of various crude oils at 0° C was sufficiently different from seawater. The next step was to build several 1 m NMR coils to test wire configurations and type, as well as to design the required electronic components to enhance the signal to noise ratio (SNR). The project determined that a wire coil in a modified figure-8 pattern provided the greatest SNR, while still managing logistical considerations (e.g., weight and cost).

The 1 m antenna was scaled-up to a 6 m diameter version and underwent testing at the US Army Corps of Engineers Cold Regions Research and Engineering Laboratory (CRREL) in Hanover, New Hampshire, USA (Figure 1). This antenna was tested with actual ice to field confirm the ‘invisible’ nature of ice to the NMR system. Further, a crude oil surrogate was included in the testing to ensure the larger NMR antenna’s ability to discriminate oil in the presence of water. Crude oil surrogates have been used throughout development of the NMR system to better manage logistics and regulatory challenges. Crude oil surrogate (e.g., Crisco^® oil) NMR properties in the earth’s magnetic field were tested. In each case the measured properties were directly comparable to Alaska North Slope crude oil, and therefore a supported the justification of continued use during testing. CRREL testing also included the use of an adiabatic inversion to suppress the water signal. This step is required as the volume of water is orders of magnitude greater than the oil.

In November 2016, a ruggedized NMR system was deployed from a helicopter at a pond outside of St. John’s, Newfoundland Labrador, Canada. The 6 m NMR coil consisted of multiple copper wire bundles in a figure-8 pattern consisting of two D-shaped halves. The NMR coil was accompanied by 6 m pre-polarization coil placed within the same fiberglass housing (Figure 2) consisting of multiple aluminum wire bundles. The total weight of the NMR coil assembly and fiberglass housing was 890 kg. The NMR system included onboard electronics housed in a standard aluminum equipment rack that measures approximately 2 m³ and secured to the internal rail system.

To accommodate the testing, a platform was built at the pond site. The platform was constructed of Geofoam (expanded polystyrene) and framed with wood (Figure 2). The Geofoam was used to simulate ice, specifically as it has a similar ‘invisible’ NMR signal and could be built to the detect oil at different distances from the NMR antenna. The NMR system also includes electronic components carried on the helicopter to control the transmit pulse and measure the receiving data to provide a real-time detection signal.

The distinction of oil from water was accomplished by using a combination of prepolarization and adiabatic inversion of the acquisition signal. This process allows for the derivation of a numerical term in arbitrary units, X score. In essence, the X score is the ratio of oil to water after the suppression of water using the pre-polarization and adiabatic inversion techniques.

Prior to the helicopter test, several measurements were made with different oil surrogate concentrations and simulated ice thicknesses (i.e., distance of oil surrogate from NMR antenna). During these tests, the oil surrogate was located in gas cans above the NMR antenna to easily change the volume and distance (the antenna is magnetically symmetrical so that there is no difference between positioning the sample above or below the antenna). Four oil volumes corresponding to four different oil thicknesses (0.0, 0.5, 1.0 and 2.0 cm) were measured at three distances from the antenna (20, 56 and 92 cm). Each test was run five times to derive a mean X score for each scenario per day. Seven days of X score testing results were then averaged for each test scenario. At a minimum, 0.0 oil thickness (no oil) was run at the start and close of each day’s testing runs.

The helicopter test used a Sikorsky S-92^® helicopter that slung the NMR antenna to the Geofoam platform, whereupon it place the antenna onto the platform to take measurements. An individual was inside the helicopter solely to operate, via a laptop computer, the NMR system and record the data from the tests. Four helicopter tests were made: no oil (three passes), 1.0 cm thick oil at a distance of 79 cm and 1.0 cm thick oil at a distance of 53 cm. For all cases, the water surface was 80 cm from the NMR antenna. Each test consisted of a preliminary scan to check for noise and detect strongest water signal (e.g., distance to water) as well as 16 pairs of actual data acquisition using the described methodology. Total time to derive the X score was ~ 3.5 minutes for each test.

Following the helicopter testing, more sensitivity testing was completed. This testing followed the same protocols outlined above, specifically for oil thicknesses of 0.5 and 1.0 cm at distances of 92 and 127 cm. This testing was done to test the operational limit of the NMR system as it is currently configured.

The tests completed prior to helicopter deployment provided the opportunity to modify measurement parameters, specifically preamplifier gain and transmitter output. These parameters became ‘locked in’ for the duration of the testing to allow for measurements made to be directly comparable across pre-, during and post- helicopter testing. Although the X score is the ratio of oil to water, this value never reached zero during any of our no oil tests. This is easily seen during the pre-helicopter testing when averaging no oil scenarios to those where oil is present, both in relation to oil thickness and distance of oil sample from NMR antenna (Figure 3). The values presented are an average and standard deviation of X scores taken over seven different days. It was determined that at the two highest oil thicknesses (1.0 and 2.0 cm) for all three distances (20, 56 and 92 cm) that the crude oil surrogate could be clearly differentiated from water alone. This is also true for the lowest oil thickness (0.5 cm) at distances of 20 and 56 cm. However, the same assertion cannot be made for an oil thickness of 0.5 cm at a distance of 92 cm, where mean X score values ranged from 0.0345 to 0.0392. These values intersect with the no oil mean X score values, which ranged from 0.0207 to 0.0389.

It should be noted that if each of the seven days of data are analyzed independently, oil can be clearly discriminated from water alone as the X score is greater than one standard deviation, with a single exception. As stated, the values discussed above are averages over all of the testing days. For example in the case of 0.5 cm oil thickness at a 92 cm distance between NMR antenna and oil sample on 2 August 2016, the X score mean was 0.0345 versus a mean no oil X score of 0.0219, which can be discriminated. However, the no oil high X score for 3 August was 0.0297, which, at one standard deviation for both scores, cannot be discriminated (average X score was 0.0392 for 0.5 cm at 92 cm on 3 August). There is a single case where the X score for 0.5 cm oil thickness at 92 cm distance could not be discriminated from no oil in the same day.

Figure 4 shows the NMR antenna being deployed from a long line under the helicopter. The use of a helicopter introduces several new logistical challenges, including increased safety risk. The tests were completed in a safe manner and the tethered NMR antenna performed well during the flights, both in flight characteristics and data acquisition. During the tests a load master located near the platform was used to ensure that the 6 m diameter NMR antenna rested securely on the 6.2 m wide platform. The limited results align with the pre-helicopter testing. No oil X scores were 0.0272, 0.0282 and 0.0258. Two measurements were taken for 1.0 cm oil thickness at a distance of 53 cm, values were 0.0415 and 0.0401. The single value derived for 1.0 cm oil thickness at 79 cm was 0.0403. Although these X score values are not directly comparable to previously recorded values as the distances are different, they are lower than expected values. The average X score for 1.0 cm oil thickness at 56 cm is 0.0710. The helicopter test results showed a clear distinction of oil from water, but could not distinguish the distance between the two oil samples (53 and 79 cm).

Sensitivity testing to characterize the system performance as a function of operating parameters followed directly after the helicopter tests. These tests extended the detection results and discriminated a 1.0 cm oil thickness layer at a distance of 92 cm, but could not clearly discriminate the same oil thickness for 127 cm. Extrapolating the results, it was concluded that the NMR system could discriminate 1.0 cm oil thickness at a distance of ~ 110 cm. An oil thickness layer of 0.5 cm was also tested, with extrapolated results showing that it can be discriminated from oil at a distance of ~ 90 cm. It is expected that these limits could be extended by acquiring data for a longer period of time (i.e., > 3.5 minutes) to increase signal strength.

The results from testing the helicopter ruggedized NMR system support the initial project operational criteria objective of the detection of ~ 1 cm thick oil layer below ~ 1 m of ice without placing personnel on the ice and in a reasonable amount of time. The motivation for the initial criteria was selected based on oil exploration where first year ice would be present. The testing completed prior to helicopter flights provided the primary source of data through a matrix of different oil thicknesses and distances separating sensor from sample. These results show a clear discrimination of oil from water for 1.0 and 2.0 cm oil surrogate thicknesses at all distances (i.e., ice surrogate thicknesses). However, there is overlap in X score values for 0.5 cm thickness at 92 cm and no oil values. The root cause for this is that the water suppression method is not fully suppressing the water signal, and therefore the X score for no oil is not zero. This further complicated direct comparison of data from the helicopter and sensitivity tests. Although the signal from water is suppressed and allows for data to be assessed from each day of testing, the comparison across multiple days is challenged by the no oil X score benchmark values changing over time. In an operational setting, this challenge can be alleviated by ensuring that reference data points are taken each day that the sensor is deployed over ice where no oil is present.

The NMR system, comprised of the antenna and onboard electronics, has been designed with operational considerations. The weight of the NMR antenna was considered during design and can safely be deployed from helicopters used in offshore operations. The shape and form of the NMR antenna also considers in-flight aerodynamics. Further, the design of the NMR antenna air frame allows for it to be folded and stored in a standard-width shipping container box for storage when not in use (Figure 5). The onboard NMR electronics can be operated using helicopter or battery power and can be easily transferred from one helicopter to another, as the components are mounted in a standard rack. A beta version of software has been developed to allow untrained personnel the ability to operate the NMR system via the onboard laptop.

The NMR system as designed is capable of being operationally deployed and detect oil in an ice under ice. There remain potential improvements to the system that require continued testing and minor modifications. Primarily, improvements and validation to the water suppression methodology would allow for higher confidence when attempting to detect lower oil concentrations and discriminating from water. This, in turn, could lead to improved overall sensitivity of the NMR system.

The authors would like to thank Fugro for operational support throughout the construction and testing of the ruggedized NMR system. Thanks are also extended to Lake Central Air Services for avionic and aeronautical design support.