# Overview

• Waveguides can be defined in a substrate using implanted ions
• The pattern for the implant can be defined using photolithography, with mask openings patterned in a resist layer.
• The ion implant feature in DEVICE CHARGE can be used to simulate mask-defined ion implant profiles in 2D and 3D
• An index perturbation can be applied to create a mapping between the implanted ion density and the variation in the material index

# DEVICE CHARGE Simulation

One of the features of the DEVICE CHARGE solver is the implant doping object. This object enables you to define an ion doping profile in a target substrate that is constrained by a mask opening (the mask opening could be patterned by photolithography and specified by a GDSII file). The implant profile is also specified by

• Implant angle
• Peak concentration
• Distribution function and parameters (e.g. range and straggle)
• Lateral scatter

These parameters can be extracted from measurements and/or process configuration. In this example, we use the first technique to extract the parameters from measured implant profiles reported in Ref. 1. The plot below shows the fit to the Pearson 4 distribution of the measured profile for AgO2 implanted at 6MeV into SiO2.

Once defined, the CHARGE solver can be used to generate a finite element mesh (2D or 3D) with the implanted ion density distribution, which will be used to calculate the index perturbation. Note that the substrate material defined for the electrical simulation is silicon rather than SiO2; this is done because doping implants into oxides are ignored in the CHARGE solver, but this will not affect the optical simulation.

# Index Perturbation

To determine the index perturbation, Huggin’s model is used [2,3]:

$n(\chi) = 1 + \frac{R(\chi)}{V(\chi)}=1+\frac{\sum_M a_M N_M(\chi)}{k + b_{Si} + \sum_M c_M N_M(\chi)}$

where $a_M$, $b_M$ and $c_M$ are material-dependent parameters; $N_M(\chi)$ is the number of moles of ion $M$ contributed by the oxide $M_mO_n$ per mole of oxygen ions contributed by all components of the glass; $k$ is a small process-dependent parameter; and $\chi$ is the fractional contribution of each oxide.

For a given density of implanted Ag ions (e.g. from an AgO2 source), $N_{Ag}$ will vary spatially, and $\Delta n$ can be determined relative to the pure silica case. In the example, this calculation is performed in the “Calculate Index Perturbation” step in the “Ion Implant Waveguide” workflow, which takes the material-dependent parameters and the implanted ion density (from the previous step) as inputs and generates a dataset with a spatially varying $\Delta n$. This result is available for visualization and it is loaded into the (n,k) material import object.

# DEVICE FEEM Simulation

The modal fields and properties (e.g. effective index) can be determined using the FEEM solver. The FEEM solver includes the (n,k) material import object, which is configured to cumulatively add its spatially-varying index perturbation to the material index of the underlying volume (as defined in the geometry setup). In this case, the optical substrate material is silica (SiO2), and the $\Delta n$ value calculated in the previous step will be added to that index.

The FEEM solver is configured to calculate the fields over a large volume to account for the weak confinement of the mode, which is due to the small change in index. The resulting fields and mode properties are available as a result from the solver, and the effective index of the fundamental mode is reported as a result of the workflow.

# Taking the Model Further

For optical simulation in FDTD Solutions or MODE Solutions, the same index perturbation can be calculated and exported in a format suitable for import as an (n,k) import object. In the “Calculate Index Perturbation” step of the workflow, there is a parameter “export_for_mode” that can be set to a non-zero value to enable data export.

When run, this step of the workflow will now create and save a rectilinear dataset, which can be loaded into the MODE Solutions or FDTD Solutions design environments

matlabload("mode_nkdata.mat");
importnk2(pinch(nkimportData.index),nkimportData.x,nkimportData.y,nkimportData.z);


This will create a structure with a spatially varying index. The background index for the unperturbed silica is also reported as a result of the workflow step. In the attached MODE Solutions project, the import object has been loaded and the simulation configured similarly to the example in DEVICE FEEM. The resulting waveguide modal fields and properties can now be calculated.

# Files

01_implant_wg.ldev (6.6 MB)
02_implant_wg.lms (535.7 KB)
Ag_implant_SiO2_6MeV.txt (265 Bytes – please change extension to .csv for this example)

# References

1. Marquez, H., et al., Opt. Mat. , 35 , 927 (2013)
2. Huggins, M. L., J. Opt. Soc. Am. , 30 , 420 (1940)
3. Fantone, S. Appl. Optics , 22 , 432 (1983)
3 Likes

Hi，the [01_implant_wg.ldev] file can’t work. The program broke down.
Any help will be grateful!

OS 版本: 6.3.9600.2.0.0.272.7

I just ran the file with 2019b-R1 and it works just fine.

What version are you using? If it is an old version (maybe 2017?), it might not support the feature as it was added recently.

Please share the error message that you get and I will be glad to take a look at it.

Thanks

The version is 8.19.1584. I’ll try the newest version. Thank you for attention!