1H, 23Na and 35Cl Imaging in Cementitious Materials with NMR
- 734 Downloads
A set-up especially designed for semi-simultaneous measurements of 1H, 23Na and 35Cl in ordinary cementitious materials using nuclear magnetic resonance was built. This setup makes use of the main field of a whole body magnetic resonance imaging system (Philips Intera), which has allowed us to combine two measurement setups into one, i.e., a 23Na/35Cl and a 1H insert. This 1.5 T field was chosen as a compromise between the signal-to-noise ratio of the spin-echo signal, which increases at higher frequencies, and the line broadening due to the presence of magnetic impurities of these materials, which leads to a decrease of the resolution at higher magnetic fields. The preliminary experiments show that this setup can be used to the study the interaction of different types of ions with cementitious materials. One-dimensional profiles of the moisture and dissolved ions can be measured with a spatial resolution of about 2 mm for 1H, 6 mm for 23Na and 9 mm for 35Cl.
KeywordsNuclear Magnetic Resonance Phase Lock Loop Nuclear Magnetic Resonance Signal Voltage Control Oscillator Programmable Gain Amplifier
It is widely acknowledged that chloride-induced corrosion is one of the main degradation mechanisms in civil structures based on reinforced concrete. The corrosion starts as soon as the chloride comes in contact with reinforcement steel bars . The source of chloride can be natural, i.e., sea water, or from de-icing salts. In general the chloride will enter a concrete by advection with moisture or diffusion within the moisture present in concrete. Gaining insight into these transport phenomena can not only improve the assessment of durability aspects of existing structures, but might lead to improved design for new reinforced concrete structures that are to be used in aggressive environments.
At the moment there is lack of experimental data on this topic that can be used to validate or discard the wide range of models available in the literature. Whereas such models might be correct, erroneous input is bound to lead to false predictions (i.e., the garbage in, garbage out principle). Modelling with faulty input parameters might have catastrophic consequences.
There are various methods available to measure ionic chloride content in porous building materials. The most common method is to drill a specimen out of a concrete structure and analyse it chemically in the laboratory (see, e.g., ). Traditionally this is done by pulverizing the sample and extracting the soluble chloride content using nitric acid solution. The solution is then analysed for chloride ion concentration using wet chemistry. The most obvious drawback of this method is its destructiveness, but there are more shortcomings, such as its poor spatial resolution and irreproducibility.
Nowadays there is a wide range of non-destructive techniques available [3, 4, 5]. While most of these methods are able to detect chloride content non-destructively, none of them is able to measure moisture and ion transport simultaneously. For scientific research a method that can measure combined moisture and ion profiles with high accuracy and high temporal and spatial resolution would be desired as to verify transport models. Moreover, one would like to measure multi ion transport and also to study the interaction of these ions with the materials, where ion exchange can be present.
Various properties of the nuclei studied by NMR
Natural abundance (%)
These studies show that it is possible to quantitatively study chloride in cementitious materials using NMR imaging techniques. However, the reported studies on Cl transport are limited often to white cements, whereas ordinary cements always contain magnetic impurities (e.g., Fe), which can influence the relaxation behaviour (see, e.g., ). Since the time scale of our experiments covers the region from a few seconds to a few days, our original aim was to build an insert designed for semi-simultaneous measurements of 1H, 23Na and 35Cl in ordinary cementitious materials. As the gyromagnetic ratios and hence the resonance frequencies of the selected nuclei are too far apart to be covered by one insert without seriously compromising the sensitivity, we have chosen to use the main field of a 1.5 T whole body MRI system (Philips Intera). The large space within this system allows us to combine two measurement setups into one, i.e., a 23Na/35Cl and a 1H insert. This 1.5 T field was chosen as a compromise between the signal-to-noise ratio of the spin-echo signal, which increases at higher frequencies, and the line broadening due to the presence of magnetic impurities in these materials, which leads to a decrease of the resolution at higher magnetic fields. We will first discuss the home-built NMR setup, i.e., the multi-nuclei RF electronics and the multi-nuclei insert designed for this purpose. To test the set-up we have looked at the hydration of ordinary cement, i.e., Portland CEM I and blast furnace CEM III, with a 4 m salt solution, where we have focussed on the relation between the 23Na and 35Cl content/concentration as a function of time, which could not be studied up to now.
2 The NMR-Setup
2.1 Multi-Nuclei RF Electronics
For NMR frequencies below 30 MHz, the RF output signal of the Radioprocessor board is fed into a programmable gain amplifier (PGA), which drives the power amplifier that generates the RF field used to excite the nuclei of interest in the sample. The NMR signal picked up from the sample is amplified by a PGA, before it is supplied to the RF input of the Radioprocessor board. A duplexer isolates the RF signals to and from the sample from each other. The PGA’s are controlled by digital signals via a digital input/output module (DIO) connected to the PC used for experiment control and data acquisition.
For NMR frequencies above 30 MHz, the NMR frequency is shifted to a frequency around 30 MHz, which can be handled by the Radioprocessor board. For this frequency range a variety of RF circuits, such as mixers, is commercially available. The frequency shifting is implemented by mixing the input and output signals of the Radioprocessor board with a signal at a frequency that is a multiple of 2 MHz. The latter signal is created by a voltage controlled oscillator (VCO), which is part of a phase locked loop (PLL). This PLL is synchronized to the 10 MHz reference signal of the Radioprocessor board, which ensures that the phases of all NMR signals are mutually synchronized also.
The RF signal to excite the nuclei is created as follows. The output signal of the Radioprocessor board is mixed with the output signal of the VCO by an image reject mixer (IRM), which significantly attenuates frequencies other than the desired NMR frequency. The output signal of this IRM is amplified by a PGA and filtered to further suppress unwanted frequencies, before it is fed into the RF power amplifier. The NMR signal from the sample is first amplified by a PGA, and subsequently mixed with the output signal of the VCO in an IRM. The output signal of this IRM is filtered by a band pass filter and further amplified by a PGA, before it is fed into the RF input of the Radioprocessor board.
The output frequency of the VCO can be adjusted by changing the divisor of the programmable divider, which is controlled via a DIO connected to the PC. The routing of the signals for NMR frequencies below or above 30 MHz, respectively, is adapted by switching the coaxial relays R, which are also controlled by the DIO module. If needed, some coaxial relays are added to facilitate the use of different RF power amplifiers for different nuclei or different duplexers. The control of all modules in the system depicted in Fig. 1, as well as the data analysis and presentation, is performed by MATLAB® routines.
2.2 NMR Insert
Since our aim is to perform quantitative NMR measurements, special attention was given to the impedance matching of the NMR probe. To reduce the effects of variations of the dielectric permittivity by a changing 1H, 23Na or 35Cl content in a sample, a Faraday shield  was added between the LC circuit of the probe head, i.e., the RF coil, and the sample (see Figs. 2, 3). This shield consists of 0.6 mm copper wires running parallel to the axial direction of the coil. The wires are interconnected and grounded at the lower side of the shield, well outside the RF coil. A small slit in this part of the shield prevents the generation of Eddy currents and consequent RF power losses. Care has to be taken in choosing the appropriate wire thickness to prevent acoustic resonance of these wires, which can be picked up as ghost signal. Additionally, the quality factor Q of the LC circuit was reduced to about 40 by adding a small series resistor to the RF circuit (see Fig. 3).
To generate a magnetic field gradient 3 pairs of coils are used, which can be combined to Anderson gradient coil sets, either for the upper or lower RF coil. These coils were made from a single plate of 3 mm copper, from which they were cut using a water-jet. The coils have a typical resistance lower than 0.05 Ω. To produce a gradient of 0.15 Tm−1 about 50 A are needed and therefore these coils are water cooled. No attempts were made to switch off the field gradients during the individual NMR measurements, i.e., during individual spin-echo sequences. With this gradient of 0.15 Tm−1 typically a one-dimensional resolution in the order of 2 mm is obtained for 1H, whereas the resolution for 23Na and 35Cl is respectively in the order of 6 and 9 mm.
The spin-echo experiments are performed at a fixed frequency, corresponding to the centre of one of the RF coils, i.e., corresponding to the centre of the appropriate gradient coil set. The vertical position of the sample can be controlled by a step motor, and hence first a measurement can be performed on the Na/Cl contents, after which the sample is moved over to the position corresponding to the RF coil for H. Again this step motor is placed well outside the whole body scanner and a Bowden cable is used. The spin-echo signal was excited by straightforward (90 x −τ−180 y ) Hahn pulse sequences. Using the LC circuit described above and a 1.5 KW Tomco® wide-band RF power amplifier (0.5–150 MHz), a 90 degree flip angle of the spins could be achieved with pulses having a duration of 25 μs for 1H and 35 μs for both 23Na and 35Cl.
To test the performance of the setup we have looked at the interaction of 23Na and 35Cl with ordinary cement during hydration, i.e., CEM I Portland cement and CEM III Blastfurnace cement. We will first shortly discuss the signal-to-noise ratio of the setup and the measured nuclear relaxation in micro concrete samples. Finally we will discuss the hydration measurements of cement paste with 4 molal (m) NaCl solution.
3.1 Signal-to-Noise Ratio
3.3 Relaxation Behaviour of Ions in Cement
3.4 Hydration of Cement with a NaCl Solution
Using a specially designed NMR setup, the 1H, 23Na and 35Cl content in cementitious materials can be measured quasi-simultaneously. The sensitivity of the current setup is sufficient to study the intrusion of seawater in cementitious materials. The relaxation of 1H, 23Na and 35Cl can be used to obtain pore-size information, and thereby information on the pore-ion concentration distribution Moreover, the setup has shown that by multi nuclei measurements more insight can be gained on interactions with the cement matrix. It is observed that the Na/Cl ratio changes during the hydration, indication a chemical/physical interaction with the cement matrix.
In the near future we want to improve the signal-to-noise ratio for 35Cl by removing any material causing a background signal. We also want to extend the RF-coil for 1H as to be able to measure also 7Li. This will make the setup also suitable to study alkali silica reactions (ASR), which form also a very important damage mechanism for concrete structures. Moreover, by increasing the sample diameter from 20 to 26 mm, i.e., the maximum diameter allowed within this setup, a 1.7 times increase in the signal-to-noise ratio and thereby a 2.8 times decrease in measurement time can be obtained.
This research is supported by the Dutch Technology Foundation STW, which is part of the Netherlands Organisation for Scientific Research (NWO), and which is partly funded by the Ministry of Economic Affairs.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.