UV-vis Spectrophotometer Measurements on NPN

ByGreg Madejski

We’d like to get verification of the absorption spectra of our npAu/npMgF2 SERS structure. The SERS effect should occur where the absorbance is maximized.

Total Light = Absorbance + Reflection||Scattering + Transmittance

Last Friday, I had the opportunity to work with Jennifer Kruschwitz who possesses a Konica-Minolta integrating sphere spectrophotometer from 380-740 nm (10 nm increments). It’s primarily made for examining the reflectance and transmittance of paint swatches/filter glass, but it has a 3x5mm aperture which is reasonably sized for our chips. One concern with my SERS chips is that they have only a small active area (9x 0.5×0.5 um windows = 0.0225 mm^2). I have been using chips for these experiments on the order of 1.5 mm^2 to get decent signal.

I metallized chips using the Denton Sputterer (15mA current, 100 mTorr, 0.1 nm/sec) for varying amounts of time to produce Au layers of different thicknesses. The chips became more opaque as I added more Au, by eye.

I made a few preliminary measurements of reflectance on partially etched NPN, after cribbing together a reflectance mount for the system, using carbon tape and a black aluminum plate.

2238 NPN, 80s etched (2mmx0.7mm window). It doesn’t change very much with wavelength.

This week, I’ve made a few measurements on the transmittance side of things. 1148X NPN (labeled as 1148 in a 4″ cassette, need to verify with TEM) is some 5 slot NPN, with total area = 1.5 mm^2. I created an aperture for the chip to sit on with an index card and carbon tape.

Cut a hole behind the chip (3x3mm), and used carbon tape to surround hole. 1″ line used to align chip to tool.

 

2238 NPN (2mmx0.7mm) Transmittance. The fact that the broken window is not 100% transmissive is concerning.

 

1148X NPN Transmittance Error Bars. n=3. Remounted the chip on the aperture 3 times.

 

1148X NPN Transmittance, without error bars

 

Contribution of Au layers to decrease in Transmittance. Peaks in this plot can mean reflection or absorption, where they seem to appear around 500nm and 730nm. If the reflection contributions are constant across this wavelength range, then I would expect the peaks are absorptive. It is curious that the 35nm Au thickness blocks as much light as the 50nm film, but the error bars are substantial enough that this is probably measurement error.

I can probably decrease the error of my measurements by improving aperture mount (if light leaks in underneath it impacts the transmittance). More work to come.

 

 

 

Similar Posts