Making Optical Scattering Standards

Making Solid Light Scattering Standards

There are many tricky things that I work on.

Things become especially tricky when there are no long term standards to refer to.

In particular, I have been trying to build instrumentation to measure scattering and absorbance coefficients in solutions. I have been trying to measure levels that are around the same level as human tissue. It's pretty tricky, since the best standard that is used (as far as I have seen) is intralipid- a soy-oil micro-emulsion which looks a whole lot like milk.

Intralipid is sold in pharmacies for use as a parenteral nutrient. That means it's used for feeding people intravenously.

It's nice in the sense that it's relatively harmless to work with- I could probably drink it without getting sick. It's also relatively cheap, and readily available.

However, it's not an ideal standard for light scattering.

Ideally, it would be useful to have a standard which had a known, fixed particle size, with a known concentration. Intralipid is a lipid emulsion, composed of a bunch of tiny oil-bubbles bordered by an emulsifier. Unfortunately, since there is an excess of emulsifier present in solution, as the Intralipid is diluted, the particle size likely changes.

As particle size changes, both anisotropy and scattering coefficient changes. As anisotropy and scattering coefficient change, the scattering profile changes drastically.

It's tricky (if not impossible) to measure the particle sizes and anisotropy at high concentrations of intralipid because of practical limits on real optical measurements.

So it's necessary to rely on previous studies which indicate that the scattering coefficient is linearly proportional to intralipid concentration. It's not very feasible to verify the conclusions.

Furthermore, since the Intralipid expires, it's necessary to make fresh solutions for each experiment. So it's quite difficult to compare experiments and results of instrument performance.

I have wanted to make solid standards for quite some time now.

I ran a preliminary test Friday October 20th, 2006, trying to use Smooth-On 'Crystal Clear 204' (a slow-curing urethane polymer) and Silicycle 'Silia-P Flash Silica Gel' (40-63 micrometer irregularly shaped particles).


The idea was to include a solid scattering agent into a transparent solid block, so that I can run tests on permanent blocks.

I have seen some solid scattering standards when I visited the Institut national d'optique (INO) in Quebec City. Unfortunately, they aren't available to me. There is also a nice recipe listed at the Oregon Medical Laser Center web site. That's the web site for Steve Jacques and Scott Prahl among others, and it's a really good site for learning about biomedical spectroscopy overall.

This recipe is my own version of Delpy's 'Solid Epoxy Phantoms'. They suggest epoxy resin, and TiO2 to compose the scattering standard. The disadvantage is the necessity to suspend the TiO2 in ethanol. (Here is another page on making tissue-equivalent phantoms.)

I made several blends of Crystal-Clear 204 and the silica gel to make 3 different scattering levels.
Crystal clear and silica gel

One level was completely transparent, a second was 0.268g of silica for 11.076g of urethane, and a third was 1.075g of silica for 16.705g of urethane. That's 0, 2.36 and 6.05 mass % of silica to final mass.
Urethane + Silica gel

I had mixed the urethane in a few nalgene bottles, and was initially not sure how viscous the monomer solutions would be. The urethane comes in two solutions, from which the A and B solutions are to be added in a 100A:90B weight ratio. I weighed out 82.60g of A and 76.86g of B, which would have been about 100A:93B. Unfortunately, when I poured the container of B into the container with A, a significant portion of B remained in the B container. Hence I think that the final ratio was less than 100A:90B. Unfortunately, I didn't weigh out the B container to find out the amount of B remaining in the container. Pour slowly and carefully to avoid trapping air bubbles in the solution.

To mix the two monomer solutions, I used a melting point tube to slowly mix the two. I did this by stirring carefully to avoid mixing in air.

Next I weighed the silica gel into plastic weighing boats, and carefully poured in some of the pre-mixed urethane. Again I mixed the silica into the urethane by stirring slowly and carefully. I was surprised by how easily the silica was included into the polymer, with no difficulty in distributing the silica evenly throughout the urethane.

I then poured the urethane into a few molds I had prepared in advance. I tried several different containers, shapes and sizes. My initial intention was to make 1" diameter discs about 1" thick. I also wanted to try filling 1cm by 1cm plastic cuvettes, the plastic weighing boats, a nalgene bottle and a syringe.

The 1" diameter disc was cast in a ceramic support for use in kilns. The ceramic support was a porous material, and I first coated the inside surface with paraffin (by heating the ceramic up, and melting paraffin into the porous surface). Next I used paraffin and packing tape to seal the bottom edge of the ceramic, and then sprayed the inside of the mold with a mold-release agent (Mann Ease Release 300). I also sprayed the inside of one of the plastic cuvettes with the mold release agent. I didn't treat the plastic weighing boats, nalgene bottles or syringe.

I tried both injecting the urethane into the cuvettes with a syringe, however that caused a large number of air-bubbles to be formed. I then tried pouring the urethane from the plastic weighing boats into the various molds. Pouring was quite easy, and resulted in few bubbles.

The crystal clear 204 has a pot life of 120 minutes, which was more than enough for my manipulations. I began at 8:47pm, and ended at 9:30pm, with a fair bit of experimentation. Hence there shouldn't be any trouble in making a significant number of urethane standards in a single casting process.

After pouring the urethane into the molds, I placed the various castings into a vacuum chamber. We have a vacuum-desicator chamber, normally used to dry chemicals.
Degassing the urethane mixtures Degassing the urethane mixtures

According to specifications, the crystal clear should be fully cured after 48 hours. I found that 72 hours later, the urethane was not yet fully cured. When I first removed the castings from the molds, they were still quite gummy. 24 hours after removing them from the molds, they hardened fully. I was able to machine the urethane in my milling machine with no trouble (see my USB-key page).

While trying to de-mold the 1" diameter castings from the ceramic molds, I broke both of the cylindrical castings. The ceramic was originally quite porous, and can be broken by dropping it onto a hard surface. After the casting process, the ceramic was VERY hard. I finally smashed the ceramic molds apart by hitting them with the claw of a hammer, and caused the castings to fracture along the point of impact.
Final pieces

The castings in the nalgene, plastic weighing bottles, and syringe were easily pulled out of the molds (the syringe and nalgene bottles were sliced with a knife to pull out the casting).

Unfortunately, while de-moulding the cuvette castings, it was obvious that the urethane stuck to the plastic surfaces of the cuvettes. While prying the sides of the cuvettes away from the cast urethane, I pried chips off of the otherwise clean and smooth urethane faces.

Additionally, bubbles in the castings were confined to the sides and bottom of the molds. If I was to machine the castings, I could easily exclude the included bubbles. Furthermore, the hardened urethane was polished using fine sandpaper to a relatively clear surface. I think that I will be able to create a perfectly glossy and smooth surface using a buffing wheel and some polishing compound (though I have not yet tried).

Overall, I think this is an excellent method for creating solid scattering phantoms.

The difficulty of characterizing the silica particles for scattering properties is identical to the difficulties in characterizing TiO2 particles in epoxy, however no additional solvent is needed to disperse the silica, and so there will be no extra solvent pockets formed (I suspect that this might be an issue with ethanol-dispersed TiO2, though I have no experience with that system).

Good luck, and please let me know of any other methods you may know of.