In very broad terms: intensity may get lost either by absorption or by scattering.

I'd check for absorption with a spectral measurement (FTIR or the like), even an assignment of the vibrational bands may be possible if present.

As for scattering, I'd compare straightforward transmission and specular reflectivity measurements with angle integrating ones to get an idea (using an integrating shere set-up). Depending on spectral range there are different options in integrating sphere materials (special Teflon called spectralon (excellent afaik), BaSO4, gold). It is very important, though to have a set-up with adapted geometry, I have seen strange results over the years due to inappropriate apparatus design.

In very broad terms: intensity may get lost either by absorption or by scattering.

I'd check for absorption with a spectral measurement (FTIR or the like), even an assignment of the vibrational bands may be possible if present.

As for scattering, I'd compare straightforward transmission and specular reflectivity measurements with angle integrating ones to get an idea (using an integrating shere set-up). Depending on spectral range there are different options in integrating sphere materials (special Teflon called spectralon (excellent afaik), BaSO4, gold). It is very important, though to have a set-up with adapted geometry, I have seen strange results over the years due to inappropriate apparatus design.

Thanks for your suggestions!

At the moment I've only got access to an old PE 983 spectrophotometer with a specular reflectance accessory (i.e. I can only measure specular reflectance and transmittance at normal incidence). I'm scanning over 2.0-6.0µm waveband, but there are no obvious absorption bands. Possibly something would show up on an instrument with better resolution.

I have a few process/material related parameters I'm going to play around with to see if I can improve the transmittance.

I have an idea that it's possible I'm not actually depositing Al2O3, maybe the film is slightly metal-rich and the material is partially dissociating during the process. Possibly the temperature is too high.

I wouldn't say that neither the material is dissociating, nor the temperature was too high. The deposition rate was OK. If it is a part of a multilayer design, why do you think that Al2O3 is responsible for the low transmittance, which is significant?

Thanks for your answer Dr Zorc.

The reason I think Al2O3 is responsible for the lower than expected transmittance is that my design also includes germanium layers, and my other multilayer designs containing germanium (and also other materials) don't show the same effect.

It seems only to occur with designs containing Al2O3.

I have also noticed a sort of charring or blackening of the material at one side of the crucible after deposition - maybe this is normal for Al2O3, but again, I don't see that effect with the other materials I work with.

The reflectance of the coating is very close to that predicted theoretically, but the transmittance is somewhat lower.

I've also checked the substrate material (Si), and it transmits as expected.

If so, then you have to make a single, Al2O3 layer, characterize it and then fit the deposition conditions, oxygen pressure, etc. Any ion assist present?

Thanks, I think that will be the next step. There is an end-Hall type ion source available, but I haven't tried that yet.

It's quite strange, because I've previously made germanium / Al2O3 AR multilayers with very good transmittance!

It must be an effect of non-stoichiometry caused by lack of oxygen. Normally it should not happen but sometimes trivial causes could be responsible. We had once a similar "ghost effect" (you are in Scotland, aren't you?) when the needle valve did not open the flow properly. Careful with ion assist! Use oxygen only as the working gas, otherwise argon might increase the absorbance if not used correctly! To make sure, bake the bad sample for 30 minutes in air at 250C, that should decrease absorbance!

Yes, I'm in Scotland!

Thanks again for the advice, I'll try this and also try baking the sample next week.

I'll let you know how it goes.

You have not mentioned what was partial pressure of Oxygen. You must see for what oxygen partial pressure absorption is negligible. It must be characterized for a single layer of Al2O3. Have you seen effect of oxygen on absorption of single layer of al oxide?

Dr Shankar, thanks for your input. I'll be experimenting with different partial pressures of oxygen shortly. Previously, I have been able to achieve very good transmittance with Al2O3 without the use of additional oxygen, over the 3-5μm band. This latest coating design is aimed to maximise transmittance over 2.5-5.5μm.

Professor Irene, thanks for your suggestions too - I will also be experimenting with oxygen annealing soon. Also, thanks for your lectures at NanoElli09 in Belgrade, very interesting!

My experience with SiO2 and ZnO is that you realy need partial oxygen pressure to get stochimetric materials and I assume this also holds for Al2O3.

You might have oxygen deficiency in your Al2O3, presuming that your E-beam system is just a plain vacuum system. I normally deposit in a background of oxygen, but I don't know whether this can be done in your system. The true expert on E-beam evaporation of oxides is R.H. (Bob) Hammond at Stanford University (Geballe Laboratory for Advanced Materials).

Thanks again guys.

Some experiments with a partial oxygen pressure of 4.0x10^-2 Pa has at least shown an improvement in the appearance of the evaporant material. Previously, a charring or blackening of the Al2O3 surface was apparent post-deposition; bleeding oxygen into the chamber has removed this effect.

The reflectance curve has strayed quite a bit from the theoretical prediction, but I guess I'll have to recalibrate the QWOT figures for the material (with partial pressure of oxygen) and try again. I still intend to experiment with oxygen annealing of 'bad' samples too.