betaFeedback

WorksPhysics and Astronomy

Cladding effect on the mode index engineered tuned cavity

Abstract

Photonic integrated circuits require a cladding material on top to prevent any outside interaction with the photonic circuit elements and electromagnetic modes that could cause damage. Mohit et al. proposed selective tuning of the resonance of a cavity by using the mode-index engineering method. However, their work did not consider adding a cladding (upper cladding) material on top of the tuned cavity. In this study, I aim to build upon their work by investigating the impact of depositing cladding material on the frequency distribution of tuned cavities through analytical studies and numerical experiments. I identify crucial calculation parameters and discuss the optimum conditions for high-resolution tuning and large tuning range.

For the total length of cavity LL and the effective mode index nn, the resonance frequency is given by f=mcLnf = \frac{mc}{Ln}, where mm is the mode number and cc is the speed of light in vacuum. When the tuning process is performed by the mode-index engineering method [1], the resonance frequency is tuned to f′f'. The tuned resonance frequency is given by f′=mc(L−L′)n+L′(n+δn)f' = \frac{mc}{(L-L')n+L'(n+\delta n)}, where δn=n′−n\delta n = n' - n, L′L' is the partial length of cavity whose effective mode index nn is changed to n′n' and L−L′L-L' is the length of the cavity with unchanged effective mode index nn and LL is the total length of the cavity. The relative change in resonance frequency is evaluated by, δff=f′−ff=11+L′δnLn−1\frac{\delta f}{f} = \frac{f'-f}{f} = \frac{1}{1+\frac{L'\delta n}{Ln}}-1.

The addition of cladding on top of the tuned cavity changes the effective mode index and resonance shift from f′f' to fcladdingf_{\mathrm{cladding}}. So, the resonance frequency shifts in two steps,

f→f′→fcladdingf \to f' \to f_{\mathrm{cladding}}

where f′f' is tuned cavity resonance.

f=mcLn→f′=mc(L−L′)n+L′(n+δn)→fcladdingf=\frac{mc}{Ln}\to f'=\frac{mc}{(L-L')n+L'(n+\delta n)}\to f_{\mathrm{cladding}}

For any tuned cavity, (L−L′)n+L′(n+δn)=constant=K(L-L')n+L'(n+\delta n)=\text{constant}=K. After adding a cladding on top i.e. upper cladding,

effective mode index of L-L’ segment changes: n→n′n\to n'

effective mode index of L’ segment changes: n+δn→npn+\delta n\to n_p

For small values of (np−δn−n)(n_p-\delta n-n), and (n′−n)(n'-n), mode number mm remains the same. In Figure 1, two configurations are shown, A and B; in the case of A, δn\delta n is large and limits high tuning resolution, but in B, δn\delta n, δα\delta\alpha and δβ\delta\beta are small numbers, high-Q resonators are fine-tuned i.e. configuration B is more favorable for tuning procedure because of the small δn\delta n, δα\delta\alpha, and δβ\delta\beta. Therefore, I limit the numerical experiment to configuration B only. Now I evaluate, Δfcladd.′f′=fcladding−f′f′\frac{\Delta f'_{\mathrm{cladd.}}}{f'}=\frac{f_{\mathrm{cladding}}-f'}{f'},

Schematic diagram of the layer structure cross-sections of two configurations

Figure 1. The schematic diagram of the layer structure cross-sections of two configurations. First configuration (A): cavity core waveguide material Si3N4Si_3N_4 (yellow color) on top of SiO2SiO_2 cladding (grey color). A dielectric material layer (dark grey color) is deposited on the partial length of the cavity to tune the cavity’s resonance selectively. Finally, an upper cladding is added (blue color). Second configuration (B): cavity core waveguide material Si3N4Si_3N_4 (yellow color) on top of SiO2SiO_2 cladding (grey color) with a mid-cladding (green color) on top of Si3N4Si_3N_4. A dielectric material layer (dark grey color) is deposited on a partial length of the cavity to tune the cavity’s resonance selectively. Finally, an upper cladding is added (blue color).

Δfcladd.′f′=(L−L′)n+L′(n+δn)(L−L′)n′+L′np−1=K(L−L′)n′+L′np−1\frac{\Delta f'_{\mathrm{cladd.}}}{f'}=\frac{(L-L')n+L'(n+\delta n)}{(L-L')n'+L'n_p}-1=\frac{K}{(L-L')n'+L'n_p}-1

I introduce two new parameters δα\delta\alpha and δβ\delta\beta,

np=n+δn+δαn_p=n+\delta n+\delta\alpha
n′=n+δβn'=n+\delta\beta

where δα\delta\alpha is the change in the effective mode index of the L’ segment, and δβ\delta\beta is the change in the effective mode index of the L-L’ segment after adding a cladding on the tuned cavity.

Δfcladd.′f′=(L−L′)n+L′(n+δn)(L−L′)(n+δβ)+L′(n+δn+δα)−1=K(L−L′)(n+δβ)+L′(n+δn+δα)−1\frac{\Delta f'_{\mathrm{cladd.}}}{f'}=\frac{(L-L')n+L'(n+\delta n)}{(L-L')(n+\delta\beta)+L'(n+\delta n+\delta\alpha)}-1=\frac{K}{(L-L')(n+\delta\beta)+L'(n+\delta n+\delta\alpha)}-1
Δfcladd.′f′=KK+(L−L′)(δβ)+L′(δα)−1\frac{\Delta f'_{\mathrm{cladd.}}}{f'}=\frac{K}{K+(L-L')(\delta\beta)+L'(\delta\alpha)}-1

Ideally, I want Δfcladd.′f′\frac{\Delta f'_{\mathrm{cladd.}}}{f'} to be independent of L’ so that adding an upper cladding doesn’t introduce a worsening of frequency distribution, though the resonance will shift. Only two factors, δα\delta\alpha and δβ\delta\beta, contribute to the shift of resonance and worsening of frequency distribution due to the addition of cladding on top of tuned resonators (KK is constant for all tuned cavities). The critical quantity is, ΔM\Delta M which is defined as,

ΔM=(Δfcladd.′f′)L′=L−(Δfcladd.′f′)L′=0\Delta M=\left(\frac{\Delta f'_{\mathrm{cladd.}}}{f'}\right)_{L'=L}-\left(\frac{\Delta f'_{\mathrm{cladd.}}}{f'}\right)_{L'=0}

This evaluates the maximum change in frequency distribution due to addition of an upper cladding.

ΔM=1(1+δαn+δn)−1(1+δβn)\Delta M=\frac{1}{\left(1+\frac{\delta\alpha}{n+\delta n}\right)}-\frac{1}{\left(1+\frac{\delta\beta}{n}\right)}
ΔM∼δβ−δαn+(δβ+δn+δα)∼δβ−δαn\Delta M\sim\frac{\delta\beta-\delta\alpha}{n+(\delta\beta+\delta n+\delta\alpha)}\sim\frac{\delta\beta-\delta\alpha}{n}

Only the magnitude of ΔM\Delta M is relevant. Ideally, I want this quantity to be below 1/Q1/Q or tuning resolution. These numbers depend on the cavity’s material configuration and should be evaluated using numerical methods such as finite-difference time-domain (FDTD) analysis. For an example of a numerical experiment, I consider a 100 nm thick 4000 nm wide silicon nitride (Si3N4Si_3N_4) core waveguide on silicon dioxide (SiO2SiO_2) cladding, which I call lower cladding. Another SiO2SiO_2 cladding is on top of Si3N4Si_3N_4 material, which I call mid cladding, and Al2O3Al_2O_3 layer is used for tuning step, and the top cladding is called as upper cladding. The refractive index of Si3N4Si_3N_4, SiO2SiO_2 and Al2O3Al_2O_3 are assumed as 2.000, 1.457, and 1.76 respectively at 772 nm. The simulation results are estimated at operating wavelength, 772 nm, for fundamental quasi-transverse electric (TE0_0) mode. Figure 1 shows two configurations, A and B. When the thickness of mid-cladding (tmct_{mc}) decreases, the mode overlaps with tuning dielectric and upper cladding material increases, and δn\delta n, δα\delta\alpha, and δβ\delta\beta increase; configuration B becomes A when the thickness of mid-cladding is zero. Figure 2(a) and (Figure 2b) shows the dependence of δn\delta n, δα\delta\alpha on tmct_{mc} and thickness of tuning dielectric material Al2O3Al_2O_3, respectively. Note that values of δα\delta\alpha (see Figure 2(b)) are similar in order of δn\delta n when an SiO2SiO_2 upper cladding is added. But when the tmct_{mc} is large the variation of δα\delta\alpha is small with thickness of Al2O3Al_2O_3 or δn\delta n. This is an important observation; a relatively larger value of the mid-cladding and tuning dielectric material thicknesses smaller values are found for δα\delta\alpha and δβ\delta\beta (see Figure 2(b), 3(a) and (b)). Figure 3(a) and (b) shows the δα\delta\alpha and δα−δβ\delta\alpha-\delta\beta dependence on the thickness of tuning dielectric material Al2O3Al_2O_3 and upper cladding refractive index (nucn_{uc}) for tmc=500t_{mc}=500 nm. Note that as the upper cladding refractive index decreases, δα−δβ\delta\alpha-\delta\beta decreases significantly, favoring the mode index engineering method for tuning applications. In the discussed example, ΔM\Delta M is ∼4×10−7−10−6\sim4 \times10^{-7}-10^{-6} for δn∼2×10−4\delta n \sim2 \times10^{-4} and nuc=1.3−1.35n_{uc}=1.3-1.35. Therefore, the method discussed by Mohit et al. [1] attains high-resolution tuning and a wide tuning range for wafer-scale manufactured microring cavities with loaded-Q factor in the order of 1 million. In principle, even higher Q resonators can be tuned with appropriate optimizations of materials and use of low-index upper cladding materials.

Plots of $\delta n$ and $\delta \alpha$ versus thickness of $Al_2O_3$

Figure 2. a) Plot of δn\delta n vs. thickness of Al2O3Al_2O_3 layer on SiO2SiO_2 cladding without an upper cladding. tmct_{mc} is the mid SiO2SiO_2 cladding thickness. b) Plot of δα\delta\alpha vs. thickness of Al2O3Al_2O_3 when an upper cladding SiO2SiO_2 is added on top. δα\delta\alpha is evaluated by estimating the effective mode index (with upper cladding) and subtracting the effective mode index (without upper cladding). Note that, at zero thickness of Al2O3Al_2O_3 i.e. δn=0\delta n=0, δβ=δα\delta\beta=\delta\alpha. For both (a) and (b), three different thicknesses of mid SiO2SiO_2 cladding (tmct_{mc}) are considered, i.e., 500 nm, 600 nm, and 700 nm. For the larger value of tmct_{mc}, smaller values of δn\delta n and δα\delta\alpha are found which is obvious due to less mode overlap with Al2O3Al_2O_3 and upper cladding as tmct_{mc} increases. But the critical parameter is ΔM\Delta M, which is dependent on δα−δβ\delta\alpha-\delta\beta.

Plots of $\delta \alpha$ and $\delta \alpha-\delta \beta$

Figure 3. Upper cladding refractive index (nucn_{uc}) dependent δα\delta\alpha and δα−δβ\delta\alpha-\delta\beta for tmc=500t_{mc}=500 nm, a) Plot of δα\delta\alpha vs. thickness of Al2O3Al_2O_3 on 500 nm SiO2SiO_2 mid cladding and b) Plot of δα−δβ\delta\alpha-\delta\beta vs. nucn_{uc} for 40 nm, 50 nm, 60 nm, and 70 nm thick Al2O3Al_2O_3 on 500 nm SiO2SiO_2 mid cladding; δβ\delta\beta is calculated at zero thickness of Al2O3Al_2O_3 from (a).

Now, I work out the ΔM\Delta M quantity for adiabatic design (see Ref. [1]). For adiabatic tapered regions (a and b; a = ad.1 at one end and b=ad.2b = \mathrm{ad}.2 at the second end),

f′=mc(L−L′−2l)n+L′n′+∫ad.1na(x) dx+∫ad.2nb(x) dxf' = \frac{mc}{(L-L'-2l)n+L'n'+\int_{\mathrm{ad}.1}n_a(x)\,dx+\int_{\mathrm{ad}.2}n_b(x)\,dx}

where L−L′−2lL-L'-2l region has effective mode index nn, L′L' region has effective mode index n′n' and two regions (a and b) of length ll have effective mode index as a function of x, i.e., na(x)n_a(x) and nb(x)n_b(x). Assuming the tapering region is linear of each adiabatic region of two ends (ad.1 and ad.2),

∫ad.1na(x) dx=∫ad.2nb(x) dx=(n+n′)l/2\int_{\mathrm{ad}.1}n_a(x)\,dx=\int_{\mathrm{ad}.2}n_b(x)\,dx=(n+n')l/2

and after the addition of cladding, na(x)n_a(x) and nb(x)n_b(x) becomes na′(x)n'_a(x) and nb′(x)n'_b(x) respectively, therefore,

∫ad.1na′(x) dx=∫ad.2nb′(x) dx=(n′+np)l/2\int_{\mathrm{ad}.1}n'_a(x)\,dx=\int_{\mathrm{ad}.2}n'_b(x)\,dx=(n'+n_p)l/2

Therefore,

Δfcladd.′f′=(L−L′−2l)n+L′(n+δn)+(2n+δn)l(L−L′−2l)n′+L′np+(n′+np)l−1\frac{\Delta f'_{\mathrm{cladd.}}}{f'}=\frac{(L-L'-2l)n+L'(n+\delta n)+(2n+\delta n)l}{(L-L'-2l)n'+L'n_p+(n'+n_p)l}-1

The quantity ΔM\Delta M is estimated as,

ΔM=(Δfcladd.′f′)L′+2l=L−(Δfcladd.′f′)L′=0,l=0\Delta M=\left(\frac{\Delta f'_{\mathrm{cladd.}}}{f'}\right)_{L'+2l=L}-\left(\frac{\Delta f'_{\mathrm{cladd.}}}{f'}\right)_{L'=0,l=0}
ΔM∼δβ−δαn\Delta M\sim\frac{\delta\beta-\delta\alpha}{n}

which is similar to the ΔM\Delta M for non-adiabatic design. This is because only the difference of δβ\delta\beta and δα\delta\alpha contribute significantly to ΔM\Delta M, and ll is significantly smaller compared to LL.

Conclusion

Cladding on top of photonic-integrated devices reduces signal loss and crosstalk. It also offers protection against environmental damage, ensuring device reliability and durability. Therefore, it is essential to investigate the effect of thick cladding on top of tuned microring cavity. In this work, I have extended the work reported by Mohit et al. [1] by investigating the impact of upper cladding on tuned cavity. I introduced new parameters and quantities to estimate the worsening of frequency distribution due to cladding deposition on top of tuned cavities. It is found that the mode index-engineering method for passive tuning applications allows a scalable single-step tuning process for high-QQ microring resonators. In the discussed numerical experiment, resonators with loaded QQ-factor up to 10610^6 can be tuned with resolution below 1/Q1/Q and tuning range up to 103/Q10^3/Q without any significant impact on the frequency distribution after the addition of low index upper cladding. Low index materials for upper cladding are favorable to minimize the ΔM\Delta M; this is evident due to the high index contrast between the mid-cladding and upper-cladding material. Therefore, it is essential to review low-index materials that are optically transparent, low-cost, and compatible with complementary metal-oxide-semiconductor (CMOS) technology. The high-quality growth of such low-index materials on top of materials used in the mid-cladding and tuning step is a crucial area for future research. I envision that the work presented here will be expanded in further research, including different materials such as silicon, lithium niobate, silicon carbide, and experimental demonstrations.

Funding

I want to thank the Robert A. Welch Foundation (grants A-1261 and A-1547), the DARPA PHENOM program, the Air Force Office of Scientific Research (Award No. FA9550-20-10366), and the National Science Foundation (Grant No. PHY-2013771). This material is also based upon work supported by the U.S. Department of Energy, Office of Science, Office of Biological and Environmental Research under Award Number DE-SC-0023103, DE-AC36-08GO28308.

Acknowledgment

I want to thank Prof. Marlan O. Scully, who provided continued support and funds to conduct this study.

Disclosures

The author declares that he has no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data Availability

Data underlying the results presented in this paper are not publicly available at this time but can be obtained from the author upon reasonable request.

References

  1. [1]M. Khurana, S. Delfan, and M. O. Scully, “Selective passive tuning of cavity resonance by mode index engineering of the partial length of a cavity,” (2024).arxiv.org/abs/2409.04422

Paper details

vlnc:W_g9784576
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents