Advances of unilateral mobile NMR in nondestructive materials testing

Magnetic Resonance Imaging 23 (2005) 197 – 201

Advances of unilateral mobile NMR in nondestructive materials testing Bernhard BlqmichT, Federico Casanova, Juan Perlo, Sophia Anferova, Vladimir Anferov, Kai Kremer, Nicolae Goga, Klaus Kupferschl7ger, Michael Adams Institute of Technical Chemistry and Macromolecular Chemistry, RWTH Aachen University, D-52056 Aachen, Germany

Abstract Unilateral mobile NMR employs portable instrumentation with sensors, which are applied to the object from one side. Based on the principles of well-logging NMR, a hand-held sensor, the NMR-MOUSE (MObile Universal Surface Explorer) has been developed for nondestructive materials testing. In the following, a number of new applications of unilateral NMR in materials science are reviewed. They are the state assessment of polyethylene pipes, the characterization of wood, the in situ evaluation of stone conservation treatment, highresolution profiling of rubber tubes and 2-D imaging for defect analysis in rubber products. D 2005 Elsevier Inc. All rights reserved. Keywords: Mobile NMR; NMR-MOUSE; Unilateral NMR imaging; Polyethylene; Cultural heritage

1. Introduction NMR can be conducted in highly inhomogeneous fields by detecting echoes. The initial magnetization converted into magnetization detectable by one or more echoes can be modified by magnetization filters in the same way as contrast is introduced in NMR imaging [1]. By systematic variation of a filter parameter and evaluation of the transverse relaxation during detection of the echo train, 2-D data sets are obtained for correlation of parameters in a manner similar to multidimensional spectroscopy [2]. The use of this type of NMR is of great value in well logging [3] and it also offers interesting applications in materials science [1,4]. In solid materials, translational molecular diffusion is absent and higher field gradients can be tolerated. This implies that unilateral NMR can be conducted at higher fields (0.5 vs. 0.02 T) and higher field gradients (10 vs. 0.1 T/m) for materials investigations than for well logging. A compact unilateral sensor developed for nondestructive materials testing is the NMR-MOUSE (MObile Universal Surface Explorer, registered trademark of RWTH Aachen) [5]. The original NMR-MOUSE employs a U-shaped magnet with the radiofrequency coil situated in the magnet gap

T Corresponding author. Tel.: +49 241 8026420; fax: +49 241 8022185. E-mail address: [email protected] (B. Blqmich). 0730-725X/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.mri.2004.11.058

following earlier concepts of unilateral sensors with electromagnets for moisture detection in soil and road decks [6]. A simpler design is the bar-magnet NMRMOUSE that consists of a figure-of-8-type surface coil mounted on one of the pole faces of a bar magnet [7]. This bEasy NMR-MOUSEQ provides an excellent signal-to-noise ratio in a single-shot CPMG train from unfilled natural rubber (Fig. 1). By adjusting the magnet geometry of the original NMR-MOUSE, the bProfile NMR-MOUSEQ is obtained with a thin and flat sensitive volume that is suited for measuring depth profiles with a spatial resolution of better than 30 Am. All measurements reported below were conducted with CPMG sequences.

2. Applications of the original NMR-MOUSE The Easy NMR-MOUSE has a penetration depth of about 3 mm, whereas the maximum usable depth is 10 mm and more for the original NMR-MOUSE, depending on the physical dimensions of the components. Both devices can operate with dead-times less than 10 As, providing minimum echo times of 20 As, so that they are well suited for analysis not only of rubber but also of semicrystalline polymers like polyethylene (PE) and cellulose. The CPMG echo decay of PE can well be fitted with the sum of two exponential functions, where the fast relaxing component A shortexp{ t/T 2eff,short} is assigned to the

198

B. Blu¨mich et al. / Magnetic Resonance Imaging 23 (2005) 197–201

Fig. 1. Easy NMR-MOUSE (left) with a bar magnet and single shot CPMG echo decay from unfilled natural rubber (right).

crystalline regions and the slowly relaxing component A longexp{ t/T 2eff,long} to the amorphous regions. The fit parameters are interpreted in terms of the crystallinity A short/(A short+A long), the average size of the crystallites 1/T 2eff,short , and the molecular order in the amorphous domains 1/T 2eff,long. Although this interpretation still needs to be verified in further investigations, the data obtained on PE pipes can consistently be explained in this way. PE water pipes are squeezed tight to turn off the water flow for repair, and lifetime predictions are often based on measurements at 808C. For this reason, we have subjected

PE pipes to such treatment and analyzed the NMR signal at eight points symmetrically placed on the circumference in the deformation zone (Fig. 2). The data scatter from point to point is due to local strain induced by cooling and crystallization during the production. Upon deformation, the crystallinity decreases, and upon annealing 408C below the melting temperature of the ideal PE crystal, it decreases further. During deformation macromolecular chains are drawn out of the small crystallites. At the same time, local order is induced in the amorphous domains by the deformation. This order is reduced upon annealing when

0,7

0,6

0,5

0,4

0,3

Fig. 2. Nondestructive morphology characterization of PE pipes. Top left: deformation and measurement positions. Top right: Relative amplitude ratio corresponding to the degree of crystallinity. Bottom left: Relaxation rate of the rigid component interpreted as the mean size of the crystallites. Bottom right: Relaxation rate of the soft component interpreted as the molecular order in the amorphous domains.

B. Blu¨mich et al. / Magnetic Resonance Imaging 23 (2005) 197–201 Table 1 Fit parameters for CPMG decays of wood Type of wood

T 2eff,short (ms)

T 2eff,long (ms)

A long/(A short+A long)

Pine Pine weakly degraded Pine strongly degraded Birch Weathered oak Hardwood floor

0.11F0.02 0.13F0.03

0.81F0.08 0.77F0.09

0.30 0.35

0.10F0.01

0.55F0.12

0.24

0.13F0.01 0.17F0.01 0.11F0.01

1.18F0.10 0.85F0.02 0.54F0.08

0.27 0.37 0.30

the macromolecular chains assume a more random-coil-like conformation to lower their free energy [8]. These first results on PE pipes conducted with the original NMRMOUSE and an echo time of 30 As demonstrate that unilateral NMR is a suitable tool not only to study rubber but also semicrystalline polymers. Similar results are obtained with the Easy NMR-MOUSE. A systematic investigation of unilateral NMR for the characterization of PE pipes is in progress. Another hard material is wood. It consists of cellulose, lignin and bound water. Again, the CPMG decay can be fitted with a biexponential function. Here, the long component characterizes the bound water and the short component is assigned to the cellulose and the lignin [9,10]. Following a first study of different types of wood dried for at least 2 years in sheltered spaces, the relaxation time T 2eff,long of the slowly relaxing component appears to be indicative for the type and the state of the wood (Table 1). These results measured with an echo time of 60 As are in agreement with studies of paper degradation by the NMRMOUSE [11]. Portable unilateral NMR is useful for quality assessment of stone conservation treatments. To reduce the water ingress of stone in monuments and statues exposed to weather and pollution and to strengthen degraded stone, coatings from organic polymers or silica are applied to the surface. This modifies the surface wetting properties and reduces the pore size in the upper few millimeters of the

199

stone. For measurement of an NMR signal, the stone needs to be wet [12–17]. A signal can readily be obtained with the NMR-MOUSE, but contrary to the signal of fully water saturated stone measured in homogeneous fields, the NMR relaxation time distribution determined in strongly inhomogeneous fields from partially wetted stone cannot readily be interpreted in terms of a pore size distribution. The wetting properties of the stone and the treatment, as well as the effect of the strong magnetic field gradient on signal attenuation by translational diffusion need to be taken into account. Nevertheless, significant differences between the water signals from the wetted sandstone of window frames, treated and untreated, were observed in a first field study (Fig. 3). The relaxation time distributions of the CPMG decays lend themselves to the interpretation, that signal from large pores is lost and appears in the treated stone in smaller pores due to a pore-size reduction by the stone treatment. A laboratory study is on its way to gain a detailed understanding of relaxation time distributions measured for partially wetted stone with the strongly inhomogeneous fields of the NMR-MOUSE. Yet this first field study of stone conservation efforts already demonstrates the feasibility of the approach. 3. Profiling The sensitive volume of the original NMR-MOUSE is shaped like a soup dish. Depending on the NMR frequency and with it on the measurement depth, this shape changes [19]. By suitably modifying the profiles of the magnetic fields B 0 and B 1, the sensitive volume can be shaped planar and thin, so that only signal in planar slices less than 30 Am thick is excited. The slice profiles of the original NMRMOUSE and the Profile NMR-MOUSE are compared in Fig. 4. The dramatic improvement in spatial resolution achieved with the Profile NMR-MOUSE is obvious. By accurately shifting the position of the slice through the object, depth profiles are obtained with excellent spatial resolution and simple hardware.

untreated

frequency

30

treated

20 10 0 0.01

0.1

1

10

1 00

T2eff [ms]

Fig. 3. Characterization of stone treatment for stone preservation. Left: Paffendorf Castle in Germany with the scaffolding to facilitate the measurements of the sandstone window frames. Middle: Experimental setup showing the NMR-MOUSE mounted on a tripod and a low-field spectrometer in a pilot’s case. Right: Distributions of CPMG relaxation times from water in the partially saturated sandstone window frames without and with stone conservation treatment. The experimental CPMG decays were inverted by the UPEN algorithm [18].

200

B. Blu¨mich et al. / Magnetic Resonance Imaging 23 (2005) 197–201

Fig. 4. Profiling of the sensitive volume. Left: cross-section through the sensitive volume of the U-shaped, original NMR-MOUSE. The signal is collected from a depth between 1 and nearly 3 mm. Right: By suitably profiling the magnetic fields, the slice thickness is reduced to less than 30 Am.

30 20

a defect in the rubber density corresponding to a hole (bottom, middle). 4. Imaging Unilateral NMR devices can be employed for imaging in a manner similar to the use of a magnifying glass, but with the capability to look beneath the surface. By fitting the NMR-MOUSE with surface gradient coils, an open tomograph is obtained with unrestricted access. Because of the inhomogeneous fields of the device, phaseencoding methods need to be employed. Feasibly short measurement times are obtained by suitable addition of echoes [20,21]. Depending on the sample and the resolution, the acquisition times for one image range between 1 min and 2 h. The image quality is quite acceptable, as demonstrated in Fig. 6 by an image of the defect in the rubber hose of Fig. 5 (bottom). The shape of the defect and the positions of the textile fiber reinforce-

spot 2 spot 1

10 0

relaxation weight [a.u.]

amplitude [a.u.]

An application of the profile NMR-MOUSE by no means obvious at first thought is to acquire depth profiles through round multilayer objects like textile-reinforced rubber tubes (Fig. 5, left). From the profile, the thickness of the individual layers can be determined with the accuracy of the slice thickness and the positioning of the slice. The layer of the textile fibers is readily identified by an abrupt change of the amplitude of the CPMG decay (Fig. 5, middle). A relaxation contrast profile is conveniently obtained from the CPMG decays by plotting the integral of the echo amplitudes in a certain time window normalized to the initial amplitude (Fig. 5, right). This parameter corresponds to the transverse relaxation weight in Fourier imaging and does not require fitting of the echo decay. The measurements on rubber tubes reported in Fig. 5 reveal two types of defects. A difference in the transverse relaxation detected for two spots in the inner lining of the tube reveals a problem with the vulcanization conditions (top, right). A difference in the signal amplitude of the inner lining reveals

250 200

100

failure

100 50 0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 depth [mm]

0.0 0.5 1.0 1.5 2.0 2.5 3.0 depth [mm]

relaxation weight [a.u.]

amplitude [a.u.]

150

reference

spot 1

150

0.0 0.5 1.0 1.5 2.0 2.5 3.0 depth [mm]

200

spot 2

350

reference failure

300 250 200 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 depth [mm]

Fig. 5. Profiling of textile-reinforced rubber hoses. Left: Photos of the objects. Middle: Signal amplitudes. Right: Relaxation weights. The different regions of the tube wall are marked by different gray shades of the background in the graphs. They correspond to the outer tube wall, the textile layer, and the inner tube wall. Top: Pair of rubber tubes with identical proton densities but a difference in the formulation or the curing. Bottom: Profiles at 2 locations marked as reference and failure. The failure was a hole in the inner tube wall.

B. Blu¨mich et al. / Magnetic Resonance Imaging 23 (2005) 197–201

Fig. 6. Imaging of the defect in the rubber hose of Fig. 5 (bottom left). Left: unilateral tomograph with the rubber tube in place. Right: NMR image showing the defect and the reinforcement fibers in the rubber matrix. The field of view is 44 cm, the lateral spatial resolution is (0.8 mm)2, the slice thickness is 0.66 mm, and the acquisition time was 120 min.

ments are clearly visible. 3-D images are obtained by shifting the sensitive slice through the object. 5. Summary Unilateral NMR has already found applications in nondestructive testing of rubber products. With the availability of new unilateral sensors like the Profile NMRMOUSE and the open tomograph, new applications are explored. Clearly, the development of ex situ spectroscopy [22] is a further intriguing perspective for unilateral NMR devices in materials testing and well logging. Acknowledgments The in situ measurements at Paffendorf Castle were done in cooperation with Prof. M. Raupach, Institut fqr Bauforschung, RWTH Aachen. The unilateral imaging work was conducted in the framework of a center of excellence bSurface NMR of Elastomers and Biological MaterialsQ supported by Deutsche Forschungsgemeinschaft. References [1] Blqmich B. NMR imaging of materials. Oxford7 Clarendon Press; 2000. [2] Hqrlimann M, Venkataramanan L. Quantitative measurement of two dimensional distributions functions of diffusion and relaxation in grossly inhomogeneous fields. J Magn Reson 2002;157:3 – 42. [3] Freedman R, Heaton N. Fluid characterization using nuclear magnetic resonance logging. Petrophysics 2004;45:241 – 50.

201

[4] Blqmich B, Anferova S, Kremer K, Sharma S, Herrmann V, Segre A. Unilateral NMR for quality control: the NMR-MOUSER. Spectroscopy 2003;18:18 – 32. [5] Eidmann G, Savelsberg R, Blqmler P, Blqmich B. The NMR MOUSE: a mobile universal surface explorer. J Magn Reson A 1996;122:104 – 9. [6] Hogan BJ. One-sided NMR sensor system measures solid concrete moisture. Design News May 5, 1986. [7] Blqmich B, Anferov V, Anferova S, Klein M, Fechete R, Adams M, et al. A simple NMR-MOUSE with a bar magnet. Magn Reson Eng 2002;15:255 – 61. [8] Strobl G. The physics of polymers. Berlin7 Springer; 1996. [9] Nabbe´ N, De Je´so B, Lartigue JC, Daude´ G, Pe´traud M, Ratier M. Moisture content and extractive materials in maritime pine wood by low field 1H NMR. Holzforschung 2002;56:25 – 31. [10] Araujo CD, MacKay AL, Haily JRT, Whitthall KP, Lee H. Proton magnetic resonance techniques for characterization of water in wood: application to white spruce. Wood Sci Technol 1992;26:101 – 13. [11] Blqmich B, Anferova S, Sharma S, Segre A, Federici C. Degradation of historical paper: nondestructive analysis by the NMR-MOUSE. J Magn Reson 2003;161:204 – 9. [12] Borgia GC, Camaiti M, Cerri F, Fantazzini P, Piacenti F. Study of water penetration inside rock materials by nuclear magnetic resonance tomography: hydrophobic treatment effects. J Cult Herit 2000;1: 127 – 32. [13] Appolonia L, Borgia GC, Borolotti V, Brown RJS, Fantazzini P, Rezzaro G. Effects of hydrophobic treatments of stone on pore water studied by continuous distribution analysis of NMR relaxation times. Magn Reson Imaging 2001;19:509 – 12. [14] Borgia GC, Borolotti V, Camaiti M, Cerri F, Fantazzini P, Piacenti F. Performance evolution of hydrophobic treatments for stone conservation investigated by MRI. Magn Reson Imaging 2001;19:513 – 6. [15] Borgia GC, Camaiti M, Cerri F, Fantazzini P, Piacenti F. Hydrophobic treatments for stone conservation: influence of the application method on penetration, distribution and efficiency. Stud Conserv 2003;48: 217 – 26. [16] Sharma S, Casanova F, Wache W, Segre A, Blqmich B. Analysis of historical porous building materials by the NMR-MOUSE. Magn Reson Imaging 2003;21:245 – 55. [17] Alesiani M, Baglioni P, Capuani S, Giorgi R, Maraviglia B. Effects induced on marbles by water repellent compounds: The NMR contribution. Appl Magn Reson 2002;23:63 – 73. [18] Borgia GC, Brown RJS, Fantazzini P. Uniform-penalty inversion of multiexponential decay data II: data spacing, T 2 data, systematic data errors, and diagnostics. J Magn Reson A 2000;147:273 – 85. [19] Anferova S, Anferov V, Adams M, Fechete R, Schroeder G, Blu¨mich B. Thermo-oxidative aging of elastomers: a temperature control unit for operation with the NMR-MOUSER. Appl Magn Reson 2004;27: 362 – 70. [20] Casanova F, Perlo J, Blqmich B, Kremer K. Multi-echo imaging in highly inhomogeneous magnetic fields. J Magn Reson 2004;166: 76 – 81. [21] Perlo J, Casanova F, Blqmich B. 3D imaging with a single-sided sensor: an open tomography. J Magn Reson 2004;166:228 – 35. [22] Meriles CA, Sakellariou D, Heise H, Moule´ AJ, Pines A. Approach to high-resolution ex situ NMR spectroscopy. Science 2001;293: 82 – 5.