Hello Elie,

I attached a figure with a simple configuration for the reflectance measurement.  I have experience with reflectance measurement using a sphere as an accessory to a commercial spectrophotometer.  The configuration is for a single beam setup where you measure a reference and then measure your sample.  A few comments on top of that: 

• a large sphere will give more accurate measurements than a small sphere;
• try to get a parallel beam from the monochromator if possible using a concave mirror, or a least try to reduce the divergence of the beam;
• make sure that the light source is sufficiently stable;
• for the reference, it helps if you choose one with a reflectance similar to what you want to measure (same range of reflectance values, same roughness)
• you may want to use a second sample with relatively known reflectance (ex: silicon wafer) to check that your setup is working.

Hi,

regarding the reference measurement, there are two types of reference standards, specular reflectors like aluminum mirrors, and diffusive reflectors, for instance dielectric powders (BaSO4 is very common). It depends on your sample what is the right choice for you. In principle, it is desirable to have a standard whose reflection properties are similar to those from your sample. For instance, if you have a sample with a highly specular reflectance such as a thin metal film or a silicon wafer, a specular reflectance standard should be used.

Reference standards can be commercially purchased, but are quite expensive when they come along with a calibration certificate. A "low cost" solution is to use a sample with known optical constants as standard, for example a silicon wafer. The optical constants can be determined by an independent measurement such as spectroscopic ellipsometry. From the optical constants, the reflectance to be expected can be calculated using the Fresnel equations.

Among these two options, you can also use a Spectralon standard (labsphere, 99% of reflectance in the VIS NIR spectral region), or directly buy a sample of PTFE white powder as a reference measurement. Be sure you want to do diffuse reflectance measurements (specular light not included -you need a 'window' or an absorber to avoid the specular contribution) or just reflectance measurements (specular light included). These two documents could be of your interest: http://www.perkinelmer.com/pdfs/downloads/app_reflectancemeasurementsofmaterialsusedinthesolarindustry.pdf; http://www.iro.umontreal.ca/~poulin/fournier/theses/Ke.msc.pdf

Thanks you very much to all of you for all this very useful and clear information.

We plan to measure different types of sample but most of them are supposed to present a low reflectance in the visible range. Moreover, we want to measure the total reflectance. As a reference sample, we have a commercial thin film with known properties that we could use. Are there commercial references standards that have a low reflectivity in the visible range? 

Yes, there are. Here you can find some examples: http://www.labsphere.com/products/reflectance-standards-and-targets/reflectance-targets/spectralon-targets.aspx

Thank you !

Hi again,

I would have an additional question regarding the study of small samples. Indeed, the diameter of the aperture of our integrating sphere where we put our sample is 1 inch (See attached file).

In order to measure smaller samples (typically pieces of coated silicon or glass wafers), we have shutters ( see second attached file) that can be place at the sample aperture to reduce its diameter. There are black (I don't know their optical properties).

Is there a specific protocol when using this type of shutters to study small samples? Should we use the same protocol than the one described by Daniel Poitras (R=Rmeasured* (R_ref_theo/R_ref_measured) but whith the shutters at every step? Is it still possible to perform a reliable measurement using these shutters?

Thanks,

Regards,

Just a tip: Sometimes it is useful to stick PTFE white film (usually used by plumbers) to this kind of black (absorbing) surfaces. As you can measure, PTFE white film has a high reflectance (more than 90% in the VIS) so that you can both measure and calibrate with the shutters covered by white PTFE. I know labs where the old Integrating spheres have been re-coated with white PTFE to use them again.

Ok, thanks Juan Marcos, I will consider this possibility.

Thank you for the tip, Juan Marcos.  It is particularly good when the size of the light beam is still smaller than the sample and cover it well.  I will remember that.

I had the same problem as you Elie concerning the size of the samples, and had to use apertures (shutters) as you are planning to do.  In my case, the light beam was larger than the sample and hitting the aperture; one problem that I noticed was that the black material used for the aperture was not black enough in the near infrared (where we wanted to measure), and introduced some errors in the reflectance estimation.  I think that there are some black materials for IR on the market.

Hi, Elie

I used to measure very low light flux well below 10^-5 cd/m2 on the wall inside one of double integrating spheres. I believe you want to measure the absolute specular reflectance of nano_material sample. Based on the integraing sphere equation, for ideal sphere, ratio of two measurements should result in absolute reflectance of specular sample. But I advise you it is very difficult to eliminate stray light into real sphere, unless the integrating system is used in a perfect dark room. For example, port seal are not perfect for sample ports and detector ports in the integrating spheres, because they are intended to measure primarily total flux of very bright light sources. So, you should take great care on stray light elimination, which would introduce significant measurement error.

Data An integrating sphere system to realize very-low-luminance r...

Hi Shigeo,

Thanks a lot for these precious advices and for the article. Indeed I want to characterize nano-composite thin films. For now, I'm would like to measure the total reflectance of my samples, but indeed, it might be difficult to extract the nanostructures contribution in the signal, especially if it is noisy due to strait light effects. I didn't consider this problem up to now but i will definitely do something about it.

Best regards,

Elie

Elie,

     Good to know you have become aware of the stray light.  To check if your system is stray light free, linearity check of the photodetector attached to the integrating sphere is highly recommended if output power of your Xenon lamp can be varied.   Through linearity of the double logarithmic plot for the photodetector signal against the input power, you can check if offset in the photodetector signal.  The offset is mainly due to the stray light, if dark current is well suppressed by cooling down the photodetector.

   Attached please find data sheet for my favorite  Si-PIN photodetector.

Shigeo

Dear Shigeo,

Thank you for your advice again.

We are thinking about adding a chopper and a lock in amplifier on our r

system, do you think this would be enough to suppress the noise from straight light?

Regards,

Elie