Integrated SPR Bio-Sensing

eme
bio-sensing

#1

Hi,
I am try to reproduce the results of this paper:
Hybrid metallic ion-exchanged waveguides for SPR biological sensing
I am still new to SPR (Surface Plasmon Resonance) based structures. Since the length of the structure reaches 1000 um, I decided to try this on EME solver. I got initial results similar to figure 9 (wavelength shift).
Figure 9 :
Generated figure:

  1. I need help understanding why this structure basically has resonance at this specific wavelength.
  2. I can’t find the first 2 modes at 700 nm as reported in the paper. I only find the third mode. (True this is the mode that has the largest overlap with the input but still I am curious to know where the other modes are).
  3. I am not an expert with EME simulations so are there any suggestions to improve the simulation, please?

Thanks in advance!


#4

Hi Aya,

Q1. I need help understanding why this structure basically has resonance at this specific wavelength.
A1. As is mentioned in the paper, the resonance here refers to a sharp dip in the transmission spectrum, which is a result of the waveguide mode better coupling to the plasmonic mode at that specific wavelength. If you look at the hybrid modes in the detection zone, you will see the modes having multiple peaks due to the coupling of waveguide mode to the surface plasmonic mode.

Q2. I can’t find the first 2 modes at 700 nm as reported in the paper. I only find the third mode. (True this is the mode that has the largest overlap with the input but still I am curious to know where the other modes are).
A2. It is difficult to tell why this is the case without looking at your simulation settings. But it would be worth checking whether you have fine enough meshes over the gold layer.

Q3. I am not an expert with EME simulations so are there any suggestions to improve the simulation, please?
A3. Again, it would be difficult to say how you can improve your simulation without seeing what your current simulation setting is. You might want to consider uploading your simulation file for us to better understand what your simulation settings are and how it compare with the settings mentioned in the reference.

Note that the paper does not provide much information about the index profile of the waveguide as a result of the ion-exchange. You might need to make sure that you are using the same index profile as in the paper to obtain a result reasonably comparable to that of the reference.

Thanks,

Shin-Sung


#5

Thanks a lot for your reply. It seems the file didn’t get attached in my first post. So would you please have a look on this file? SPR_sensor.lms (437.7 KB)


#6

Thanks for uploading the simulation file.
Sure, I will have a look.


#7

Hi Aya,

I had a look at your simulation setting and could notice a couple of issues with it:

1. Number of modes
The number of modes you are using for each group seems to be a bit small.

You might need to increase these numbers until the results converge. You might also want to have a look at the following post for further information:

2. Broadband mode source
The mode solver of the Mode source uses a frequency domain technique to calculate the modes of a structure. This technique is inherently single frequency. For broadband simulations, the mode solver calculates the modes at the center frequency of the source. This could introduce some error in your transmission spectrum, especially when the mode profiles away from the center frequency is considerably different from the one at the center frequency. For further information about broadband mode source, please have a look at the following KB page:
https://kb.lumerical.com/en/index.html?ref_sim_obj_mode_source_-_broadband.html

3. Choice of solver
The result in the reference is obtained by first using FDE solver for calculation of modal indices and the modal profiles, and then applying equations 4 -6 to calculate the transmission vs. wavelength.
If you are to compare the published result with your simulation result, I think it makes sense first to follow exactly the same steps mentioned in the reference.
I suspect that EME solver should give similar result but the cause of the discrepancies between your result and the published result would be rather difficult to identify. The transmission spectrum based on equations 4 - 6 should be taken an approximation. So if you are to compare the EME result with the published result, it would be difficult to discern whether the discrepancies in transmission comes from the approximation made in using those equations or from some simulation errors within the EME.
Moreover, you might need to use a very large number of modes in the EME setting to account for the highly confined plasmonic modes, making the simulation time very long.

Thanks,

Shin-Sung