ELSEVIER

l Short Communication

Magnetic Resonance Imaging, Vol. 14, Nos. 7/S, pp. 887-889, 1996 Copyright 0 1996 Elsevier Science Inc. Printed in the USA. All rights reserved

0730-725x/96 $15.00 + .OO

PI1 SO730-725X(96)00173-9

INVESTIGATION OF FLUOROCARBON BLOWING AGENTS IN INSULATING POLYMER FOAMS BY ‘!‘I? NMR IMAGING

C.A. FYFE, Z. MEI, AND H. GRONDEY

Department of Chemistry, University of British Columbia Vancouver, B.C., Canada, V6T lZ1

Currently, there is no reliable and readily accessible technique with which the distribution and diffusion of blowing agents in rigid insulating foams can be detected and monitored. In this paper, we demonstrate that 19F NMR microscopic imaging together with 19F solid-state MAS NMR spectroscopy is ideally suited for such measurements and yield quantitatively reliable information that will be critical to the development and fabrication of optimized insulating materials with alternative blowing agents. Polystyrene (PS) and polyurethane (PU) foam samples were investigated with the objective of determining quantitatively the amount of blowing agents in the gaseous phase and dissolved in the polymer phase, and to determhte and monitor the distribution of the blowing agents in aged foams as a function of time and temperature. The concentrations of the gaseous blowing agents in the cells and dissolved in the solid were simultaneously and quantitatively measured by ‘v MAS NMR spectroscopy. An unfaced 1-yr-old PS foam filled with CHBCFzCl has about 13% of total HCFCs dissolved in the solid; while there is about 24% of HCFCs in the solid of a faced 3-mos-old PU foam filled with CHJCClzF. The data from ‘w NMR imaging demonstrate that the distributions of the blowing agents in an aged foam are quite uniform around the center part (2 cm away from any edge) of a foam board; however, a gradient in blowing agent concentration was found as a function of distance from the initial factory cut edge. The effective diffusion coefficients of the blowing agents can be directly calculated from the imaging data. Quantitative diffusion constants and activation barriers were determined. Additionally, a foam treated with a second blowing agent was monitored with chemical shift selective imaging and the diffusion of the second gas into the foam and the out-diffusion of the original gas were determined. Copyright 0 1996 Elsevier Science Inc.

Keywords: Fluorocarbon blowing agents; Polymer foams; 19F NMR imaging.

INTRODUCTION

Chlorofluorocarbon (CFC) -filled closed-cell foams have the lowest thermal conductivity of any insulation material currently available. However, there has been considerable concern regarding the environmental impact of CFCs. The Montreal Protocol’ and subsequent international agreements have decided on a worldwide phase-out of CFCs by the year 2000. Hydrochlorofluorocarbons (HCFCs) and hydrofluorocarbons (HCFs) have been chosen as replacements for CFCs. Both of them have much less environmental impact; unfortunately, their thermal conductivities are higher than those of CFC’s.

However, the quality of an insulation material is not solely determined by its initial thermal conductiv- ity, but also by its long-term performance. When an

insulation foam ages, the thermal conductivity in- creases as air diffuses into the foam while blowing agents diffuse out. Thus, being able to easily measure and compare diffusion rates of blowing agents in dif- ferent polymer matrices is very important in order to optimize the product design. In this study, we have investigated the potential of 19F NMR imaging and solid state ‘9 MAS NMR spectroscopy for the investi- gation of diffusion processes in insulation foams.

EXPERIMENTAL

For this study, two types of laminate foam boards were received from Dow Chemical Co.: polystyrene (PS) blown with CH,CF,Cl, and polyurethane (PU) blown with CH,CCl,F. The boards were about 5 cm thick and

Address correspondence to C.A. Fyfe, Department of B.C., Canada, V6T 1Zl. Chemistry, University of British Columbia, Vancouver,

887

888 Magnetic Resonance Imaging l Volume 14, Numbers 7/8, 1996

were cut into 30- x 30 cm pieces for easier handling. Several foam boards were exposed to air at defined tem- peratures (between 303 K and 363 K) for defined times (2 1 days to 5 mo) . For the NMR experiments, cylinders with diameters of 8.7 mm, and 3.4 mm were cut from the board using sharpened cork borers.

In other experiments, cylinders of 8.7 mm diameter were cut from fresh blocks; and for a defined period of time (5 days to 2 mo), each of them was exposed to CFC& gas in a pressured container at 342 K.

All NMR measurements were carried out on a Bruker MSL-400 spectrometer equipped with microimaging ac- cessory. For the imaging experiments, a Bruker single- frequency ‘H imaging probe (400.13 MHz) was modified to tune to 19F (376.5 MHz), using a home-built 10 mm horizontal coil. ‘H and ‘% images were obtained with spinecho pulse sequences. For those samples that con- tained mixtures of two different blowing gases, a soft 500 ps Gaussian pulse replaced the hard 90” pulse in the spin echo sequence. After measuring all the relevant relaxation constants, the imaging parameters were chosen to avoid T1 and T2 weighting (TE = 2 ms, and TR = 100 ms) . An inplane resolution of 270 pm was obtained from 6.8 G/cm gradients.

The probe for the solid-state ‘9 MAS NMR experi- ments was home built, incorporating a Doty 5 mm Super- sonic spinner system with a solenoid coil. Spinning speeds ranged from about 7.7 kHz to 11 kHz. A typical 90” pulse was 3 ps long, and a repetition time of 5 s was chosen to take into account the long T1 values of the HCFCs dissolved in the solid phase.

RESULTS AND DISCUSSION

By cutting samples orthogonal to the faces from the center of a board, cylinders were obtained where the

CENTER CUT EDGE

+-f---LA Fig. 1. (a) “F NMR images of PS foam cylinders. The foam had been aged for 200 days at 328 K. (b) The corresponding projections give the same information.

Table 1. Diffusion coefficient of CH,CF,Cl in PS as a function of temperature

Temperature (K) Diffusion coefficient (Dr) (cm/s)

300 1.88 x 1o-9 319 3.30 x 1o-9 328 5.60 x 1O-9 342 8.80 x 1o-9 346 1.05 x 1o-8 352 1.19 x 1o-8 358 1.71 x 10-a

concentration of blowing agent was constant over the radius.

A comparison of ‘H and 19F images shows that both exhibit the same gas distribution within the foam ma- trix; however, the total experimental times differ con- siderably due to the significantly shorter ‘9 relaxation times: A 64 scan, 128 x 128 ‘H imaging experiment took 13 hs, while the corresponding 19F imaging exper- iment needed only 14 min.

To quantify the spatial distribution of the blowing agent, projections of the intensity along the cylinder axis were acquired (Fig. 1). They show that in the center of the foam the gas is quite uniformly distrib- uted, while towards the edges that have been exposed to air, its concentration decreases.

From these projections measured as a function of aging time, it is possible to determine the diffusion coefficients. A mathematical model is available’ that describes this system very well.

Table 1 shows the DT values of CH,CF,Cl in a PS matrix at different temperatures. The activation energy was calculated to be 33.8 KJ/mol.

The diffusion of a gas into a foam matrix can also be observed. After a foam cylinder had been exposed to a second CFC gas, NMR experiments showed that the foam matrix contained both gasses. ‘9F chemical shift selective imaging experiments were then per- formed to study each blowing gas separately.

Figure 2 shows how the CH3CF2C1 and CFC& distri- butions change with the time the foam sample was exposed to the second blowing gas. Clearly, two diffu- sion processes occur: the diffusion of CH$ZF,Cl out, and the diffusion of CFC& into the foam matrix. A modification of the mathematical model was necessary to calculate the diffusion coefficients for both of the blowing agents. Here, taking into account the fact that the gas distribution is now a function of the radius, an expansion into Bessel functions is required to simulate the intensity distribution. The diffusion coefficient for CH,CF,Cl determined from this experiment agreed well with the value found for PS foam boards aged in air (DT at 342 K).

Investigation of fluorocarbon blowing agents 0 C.A. FYFE ET AL.

Repeating the same experiments on PU samples, it is found that the diffusion processes cannot be fitted satisfactorily with the same mathematical models. The projections show an overall rise in gas concentration after long exposure to elevated temperatures. The rea- son for this becomes clear when fast spinning 19F MAS NMR spectra are measured. These spectra make it pos- sible to observe the amount of gas dissolved in the solid polymer cell walls (Fig. 3). In all previous studies, it had always been assumed that the amount of dissolved gas was negligible. We found this to be somewhat valid for the PS systems, but not at all for the PU systems we investigated, where the solubility of blow- ing agent in the polymer matrix is considerable. The dissolved and the gaseous blowing agents seem to equilibrate fast compared to the diffusion process. A mathematical model that takes this additional process into account is currently being developed.

19F MAS NMR spectra also showed that during the posttreatment procedure, CFC13 gas had dissolved in

15 days

. .

Fig. 2. 19F NMR chemical shift selective imaging (projec- tions) of (a) CH, CF,Cl diffusing out of, and (b) CFC13 diffusing into a PS cylinder.

b) Jc c

l ) k d

d

889

Fig. 3. (a) 19F MAS NMR spectrum of CH$C12F in PU, freshly cut, and (b) after aging at 363 K for 4 mos. (c) 19F MAS NMR spectrum of CH,CF,Cl in PS, freshly cut, and (d) after aging at 363 K for 4 mos.

the solid polymer matrices, with the PU systems again showing a higher solubility.

CONCLUSION

It has been demonstrated that the combination of 19F NMR imaging and fast spinning 19F MAS spectros- copy is the ideal tool to determine the spatial distribu- tion of blowing agents within a foam sample, to mea- sure diffusion coefficients, and to monitor the amount of blowing agents dissolved in the solid matrix.

REFERENCES

1. Montreal Protocol on Substances that Deplete the Ozone Layer, Final Act, United Nations Environmental Pro- grams, Sept. 16, 1987 Montreal, Canada.

2. Jacobs, M.H. Diffusion Processes. New York: Springer Verlag; 1967.