F17_NMR

NMR Study of Slow-Curing Epoxies

Jarod White

Introduction

In modern NMR spectroscopy, samples are placed in a strong external magnetic field and then subjected to radiofrequency electromagnetic pulses to manipulate the magnetization of the sample. The decay of the sample’s magnetization back to equilibrium can reveal key insights into what nuclides make up the sample and how the nuclei are structured into molecules.

A common application of NMR spectroscopy involves obtaining the chemical and structural characteristics of hydrogen in a sample. We will use NMR spectroscopy to study the curing process of slow-curing epoxy which includes many hydrogen nuclei in molecular configurations that change over the curing process. The characteristic relaxation times measured in NMR spectroscopy, spin-lattice relaxation (T1), and spin-spin relaxation (T2) can be used to infer things about the sample including the viscosity of the solution and how free the molecules are to move about and rotate. By measuring T1 and T2over the whole curing time, we will be able to see how epoxy transitions from its initial liquid state to its final solid state. Cured epoxy is a solid polymer network which means that it consists of repeated connected chains of base monomer segments. Epoxy develops from an initial liquid mixture of two types of monomer molecules, an epoxy resin and a hardener, and hardens as the monomers connect together to form a large, rigid, polymer network.

Theory

Nuclei with a non-zero spin like Hydrogen exhibit nuclear magnetic resonance, an effect where the nucleus absorbs electromagnetic radiation of correct frequency to change spin state in a magnetic field. This property is used in pulsed NMR experiments, where a sample is subjected to an RF electromagnetic pulse of the Larmor precession frequency to rotate the sample’s magnetization vector away from its equilibrium value.

The magnetization rotates more the longer the pulse is applied. A 90 degree pulse rotates the magnetization 90 degrees into the XY-plane and a 180 degree pulse rotates it from +z to –z. The apparatus measures the perturbed magnetization in the XY-plane with a solenoid. Nuclear magnetic resonance depends on the nuclei’s gyromagnetic ratio as well as the local magnetic field.

Modern pulse NMR experiments measure the decay of the sample magnetization back to equilibrium following the application of an RF pulse. A 90 degree pulse causes a portion of the sample to magnetize in the XY-plane and makes their Larmor precessions coherent. After the pulse, the magnetization decays back to equilibrium (pointing in the Z direction). The XY magnetization is measured with a solenoid and the resulting signal is called the free-induction decay (FID). The FID decay of the Z-magnetization back to equilibrium occurs with time constant T1, called the spin-lattice relaxation.

The decay of the coherence between the Larmor precessions occurs with time constant T2, called the Spin-Spin relaxation.

Epoxy adhesives are made by mixing an epoxy resin and a curing agent (“hardener”) which then crosslink with each other to form a large, rigid polymer network. The epoxy resin, hardener, and cured polymer all have hydrogen nuclei which will be measured in our NMR experiment. As the epoxy cures, the relative populations of these hydrogens will change and affect the T2 values. As well, T1 will increase as the epoxy cures and becomes more rigid because there will be less rotational and translational degrees of freedom for spin realignment.

Experimental Setup

We use the TeachSpin PS2 pulsed NMR apparatus for this experiment. It consists of a permanent magnet (0.5T) enclosure, RF pulse generator, receiver, and magnet temperature controller (top).

T1 and T2 are measured for the sample over 80 minutes of curing time. Each T1 point is measured by using the inversion recovery method. This means measuring the echo height of several 180°-90° pulse trains with varying 𝜏. It was determined that varying 𝜏 between 0 and 500 ms captured most of the T1 decay for the epoxy. 10 data points (𝜏, echo height measurements) are taken by hand for each T1 measurement. Because each 𝜏 value must be set by hand, it takes about a minute to complete each T1 measurement.

For each epoxy sample, the instrument is recalibrated to find the Larmor frequency and the ideal 90-degree and 180-degree pulse lengths and the calibration parameters are not adjusted over the course of the epoxy curing. The instrument outputs a FID (shown in blue) and an envelope of the FID (shown in yellow). When the instrument is tuned perfectly to resonance, there should be no frequency component left in the FID (below image is slightly off-resonance). The actual measurements are performed on the FID envelope for consistency throughout the experiment.

Results

T1 and T2 were measured over 80 minutes of curing time for four separate samples of the Parbond 5011 epoxy. Each sample showed an apparently linear increase in T1 over the curing time with peaks located around the 40 minute mark.

One strange result from the T1 measurements is the presence of a peak at around the 43 minute mark. The previous semester's group appeared to observe this as well but no literature results show similar behavior.

The T2 measurements also seem to show a peak at around the 37 minute mark. Previous literature results do show T2 peaks for sub components of T2 [2] but not for the bulk T2 value that we are measuring in this experiment.

One possible explanation for the T1 peak is the gel effect. The autocatalytic gel effect is known to occur in curing epoxies and is characterized by local viscosity sharply increasing for the larger polymerizing chains [3]. This could result in a peak in T1 as the motion of the macromolecules becomes much more constrained and thus makes it more difficult for the macromolecules to realign with the external magnetic field. The consumption of monomers by the polymerizing chains also becomes much larger at the point of the gel effect [3] so the T2 peak might correspond to a sharp increase in the conversion rate between the monomers to some intermediary which has a larger T2.

It was also noticed that the FID signal intensity from the initial 90° decreased over the course of the experiment.

By the 80 minute mark, FID amplitudes were reduced to 6.5 V with T2 decays of <0.2 ms resulting in only three or four echoes being observed before the echo became indistinguishable from background. Our instrument was also unable to detect any FID signal for fully-cured epoxy (left to cure overnight).

Conclusion & Further Work

T1 increased from as the epoxy cured over 85 minutes as predicted but no leveling off of T1 was observed. Future experiments could attempt to increase the range of T1 measurments and measure the curing epoxy over very long time scales (several hours) to try to see T1 level off as it is expected to do [2].

T2 decreased as predicted and appeared to decrease in a roughly exponential fashion. Future experiments could attempt to measure T2 at a much faster rate than was performed in this experiment to get very fine resolution of the peak and also confirm the exponential trend.

A major experimental issue was that FID signal intensity decreased over the curing time and ultimately is undetectable in the fully-cured epoxy. The excat reason for this could be studied by a future experiment but might need different NMR equipment to improve rangibility.

An interesting peak at ~40 minutes into the curing time was observed for both T1 and T2. This could be due to the gel effect in the epoxy pre-polymers. A future experiment could study the frequency response of the FID to try to determine the relative contributions to T1 and T2 and see if these rapidly change near the peaks, which could indicate the presence of the gel effect.

References

[1] Pulsed/CW NMR Spectrometer. PS2-A/B/C Instructor’s Manual. Rev 1.4. 9/13. TeachSpin Inc.

[2] H. Kimoto, Pulsed NMR Study of the Curing Process of Epoxy Resin. Analytical Sciences, 2008 Jul;24(7):915-20

[3] S. Kaufman, J. Polym. Sci., Part A: Polym. Chem., 1982, 14, 149.