3.1. Optical Characteristics of Silicon Based Zipper Photonic Crystal Cavity
The magnitude of the optical Q factor and different optical modes can affect the optomechanical coupling rate, thereby affecting the sensitivity of the accelerometer. The optical mechanical coupling rate between the fundamental mode of the one-dimensional photonic crystal with symmetrical field strength distribution and the in-plane vibration mode of the mechanical vibrator is the highest [
25], and the following parameters such as optical resonance frequency are all parameters of the fundamental mode.
The coupling mode of the zipper cavity was obtained through simulation, which means that the electric field energy is mainly limited in the slit.
Figure 4 shows the electric field distribution corresponding to the fundamental and second-order modes of the three cross-sections.
Figure 4a,b shows the mode distribution in the xy plane, and it can clearly be seen that the electric field is mainly confined in the air slit. When the mechanical vibrator is affected by acceleration and vibrates, half of the nanobeams in the zipper cavity will generate corresponding displacements. The electric field confined in the slit is sensitive to this change, which is conducive to optomechanical coupling.
Figure 4c,d shows the pattern distribution in the xz plane. It can be seen that each mode spot is spaced between two air holes in the x-direction.
Figure 4e,f shows the mode distribution in the yz plane. The second order mode slit has less electric field energy in this plane because the center position of the zipper cavity was selected when cutting the plane, while the electric field in the x direction of the second order mode is mainly distributed on both sides of the center. The characteristic frequencies corresponding to the fundamental and second-order modes are 195.48 THz and 188.02 THz, respectively, with corresponding wavelengths of 1.53 μm and 1.59 μm.
The Q factor of an optical cavity is an important factor reflecting its quality. The influence of different structural parameters on the Q factor and wavelength was simulated and analyzed. The blue curve corresponds to the left axis, which is the wavelength curve in nm, while the orange curve corresponds to the right axis, which is the Q factor curve.
Figure 5a,b respectively plots the curves of the Q factor and mode wavelength as a function of the air pore lattice period
am and nanobeam width w.
Figure 5 shows that the required mode wavelength is positively related to the lattice period and nanobeam width, but its sensitivity to the nanobeam width w is lower than that of the lattice period
am. When the range of w variation is 100 nm, the wavelength fluctuation is only less than 60 nm, while a change of 40 nm in
am can shift the fundamental mode wavelength close to 65 nm. The tolerance of the lattice period needs to be more strictly controlled during actual manufacturing. The fluctuation of the Q factor is not very obvious, and it remains at around 4 × 10
4 within the designed parameter variation range, with relatively large factors obtained at individual parameters. The Q factor showed a downward trend with the increase of nanobeam width, and the maximum value was about 8 × 10
4.
Figure 6a,b respectively plots the curves of the Q factor and mode wavelength as a function of the lateral length
hx and longitudinal length
hy of the air hole. The fundamental mode wavelength is negatively correlated with the size of the air hole, and the proportional coefficient is basically the same. However, due to the smaller nominal value of the lateral size, the same size fluctuation has a more significant impact on the wavelength and Q factor. The Q factor shows an overall downward trend with the increase of
hx, and has a maximum value near 420 nm when
hy is taken.
Figure 7a,b respectively plots the curves of the Q factor and mode wavelength as a function of slit width s and pore number N. The fundamental mode wavelength is negatively correlated with the size of the air hole, and the overall Q factor shows a decreasing trend with the increase of the slit width s, but the fluctuation is not significant. From
Figure 7b, it can be seen that the overall fundamental mode wavelength shows an upward trend with the increase of the number of holes N, but the fluctuation is not significant, at <5 nm, and the maximum Q factor is close to the order of 10
5.
3.2. Mechanical Properties
At the same time, the mechanical mode of the zipper cavity was analyzed, and two fixed blocks were added at both ends. Fixed constraint conditions were used at the distal boundary of the fixed block, and the solid mechanics module in structural mechanics was used for simulation.
Figure 8a,b respectively shows two xy in-plane mechanical modes.
Figure 8a shows the differential mode, where the two sides of the zipper cavity are displaced in two directions, which has a more significant impact on the optical performance of the cavity.
This mode is selected as the mode of optomechanical coupling, while the change in the same direction mode in
Figure 8b is relatively small. The corresponding characteristic frequencies of the two modes are 4.82 MHz and 4.85 MHz, respectively, corresponding to an effective mass of 13.8 pg.
A further simulation analysis was conducted on the influence of different lengths on the resonant frequency of mechanical modes.
Table 3 shows the differential modes at N = 47, 51, 53, and 55, with resonant frequencies of 6.56 MHz, 5.59 MHz, 5.18 MHz, and 4.82 MHz, respectively. The resonant frequency gradually decreases as the number of air holes or length of the zipper cavity increases.
After adding a mechanical vibrator, the mechanical vibrator and the zipper cavity form the optical mechanical system. The mechanical mode of the mechanical vibrator affects half of the optical cavities on the mechanical vibrator, driving changes in optical characteristics. The optical radiation pressure of the optical mode in turn heats or cools the mechanical mode of the mechanical vibrator. The mechanical modes of the mechanical vibrator mass block were also simulated and analyzed.
Figure 9 shows four mechanical modes of the mass block, with corresponding frequencies of 3.66 KHz, 13.04 KHz, 13.55 KHz, and 18.82 KHz, among which
Figure 9a–c are all modes outside of the xy plane. The displacement of the mechanical vibrator is in the z direction, while the optical cavity is in the xy plane. These modes cannot achieve good optomechanical coupling. And
Figure 9d shows the xy in-plane mode of the vibrator at this time, which is an in-plane vibration mode that vibrates along the direction of the detection acceleration. This mode will be “filtered” out by the light radiation pressure of a specific optical mode field and heated up for vibration, indicating that this mode has the highest optomechanical coupling rate. The main consideration is the relevant characteristics of this mode, and the relevant parameters that affect the mechanical mode characteristics of the mechanical vibrator include the size, quantity, and mass block size of the vibrating beam. Considering the actual manufacturing factors and simplifying the optimization process, the simulation here mainly focuses on the simulation analysis of the size and quantity of the vibrating beam.
Figure 10,
Figure 11 and
Figure 12 respectively show the effects of different beam lengths
lx, widths
ly, and quantities
Nl on the in-plane mode resonance frequency and effective mass of the mechanical vibrator. The photomechanical coupling can be described by using a single acoustic mode of a mechanical resonator and its optomechanical coupling rate with light. At low frequencies, this effective response is equivalent to a harmonic response where the effective mass is less than the total mirror mass. The definition of effective mass
meff is as follows:
Among them, ρ is the density of the mechanical vibrator material, Q (r) is the displacement of each point of the mechanical vibrator, and max (Q (r)) is the maximum displacement of the mechanical vibrator in this mechanical mode.
Figure 10 shows the curve of the resonant frequency and effective modal mass as a function of the length of the vibrating beam. It can be seen that the relationship between the resonant frequency and the length of the vibrating beam is a first-class inverse proportional function, and decreases overall with the increase of
lx. However, the decrease is rapid when
lx is small, and as
lx continues to increase to a certain value, the decrease rate of the resonant frequency slows down.
Figure 11 shows the curve of the resonant frequency and effective modal mass as a function of the width of the vibrating beam, within the range of 0.15 μm to 1 μm, and the resonant frequency is positively correlated with
ly.
Figure 12 shows the curve of the resonant frequency and effective modal mass as a function of the width of the vibrating beam. The number of vibrating beams gradually increases from 2 to 21, and the resonant frequency is also positively correlated with the number of vibrating beams
Nl. From
Figure 10b,
Figure 11b, and
Figure 12b, it can be seen that the effective mass is basically positively correlated with the size and quantity of the vibrating beam, and the transformation trend is close to a linear change. The main reason is that the increase in the size and quantity of the vibrating beam increases the overall displacement volume without changing the mode corresponding to the displacement distribution of the vibrating beam. The maximum effective mass is 5 ng, and the resonance frequency variation range is 18.8–52.3 KHz, which can meet various needs.
In the actual detection of accelerometers, mechanical sensitivity
χ(
ω) is an important standard for measuring the performance of an accelerometer, defined as the displacement of the mechanical vibrator caused by unit acceleration. The greater the mechanical sensitivity, the greater the displacement of the mechanical vibrator caused by unit acceleration. When other parameters are the same, the overall sensitivity of the accelerometer is higher. χ(ω) is given by the following equation [
25]:
Among them,
ωm is the resonant angular frequency of the mechanical vibrator,
ω is the angular frequency of acceleration,
i = √−1, and
Q is the mechanical Q factor of the mechanical vibrator.
Figure 13 simulates and analyzes the effects of different resonant frequencies and Q factors on mechanical sensitivity when other parameters are simultaneously present. From
Figure 13a, it can be seen that, when the resonant frequency decreases, the overall mechanical sensitivity increases, but the flat frequency response bandwidth decreases. When
ω <<
ωm, the mechanical vibrator has a flat response curve to AC acceleration, and when
ω >>
ωm, the mechanical sensitivity gradually decreases. Therefore, a compromised design must be made between bandwidth and sensitivity based on actual needs. And when
ω =
ωm, the mechanical sensitivity is the highest and positively correlates with the Q factor, so it is also necessary to increase the mechanical Q factor of the mechanical vibrator as much as possible to enhance mechanical sensitivity. At the same time, the displacement of the mechanical vibrator will cause changes in optical resonance frequency, and the optical mechanical coupling rate is defined as:
Among them,
ωc is the angular frequency of the optical mode, while
x is the displacement of the mechanical vibrator. In cavity optomechanical interaction, the optomechanical coupling index can be mainly divided into two parts. Firstly, the photoelastic effect caused by mechanical motion can change the refractive index of the medium, thereby changing the resonant frequency of the optical resonant cavity. This may be the dominant factor in some systems, such as micro disk structures. Secondly, the equivalent refractive index change caused by the moving dielectric boundary caused by mechanical motion leads to a change in the optical resonant frequency. In the F–P cavity, the second factor dominates. In the case of the commonly studied Fabry–Perot cavity, GOM =
ωc/
Lc, where
Lc is approximately the physical length of the cavity. A similar relationship applies to echo wall structures, where the optical mechanical coupling is proportional to the reciprocal of the cavity radius (R),
gom =
ωc/R. Both devices utilize the radiation pressure or scattering force of light. In contrast, the zipper cavity uses gradient force operation, and its optomechanical coupling length can range from the optical wavelength
Lom to λ on the scale of c, where the length of optical mechanical coupling exponentially scales with the air slit
s,
Lom ~
λceαs. Among them α is proportional to the refractive index comparison between the nanobeam forming the zipper cavity, and the surrounding cladding is directly proportional [
26], which means that the optical mechanical coupling index
gom of the zipper cavity can reach the GHz/nm level, achieving more effective optical mechanical coupling.
The previous analysis results were based on ideal situations, but in actual situations, due to the effect of gravity, there will be a certain height difference between the two ends of the photonic crystal air slots located on the mass block and the fixed block, which will affect the resonant frequency and optical Q factor of the photonic crystal. Therefore, it is necessary to consider the impact of height differences on the fundamental mode optical Q factor and conduct simulation analysis.
Figure 14 shows the influence of the height difference ∆
h at both ends of the photonic crystal zipper cavity on Q and the influence of different beam parameters on height difference, respectively.
Figure 14a shows the variation of the Q factor with height difference ∆
h. It can be seen that, when the height difference is within the range of 4 nm, the overall Q factor remains close to 10
5, and after exceeding 15 nm, the Q factor has already decreased by an order of magnitude. So, it is necessary to maintain a height difference of less than 4 nm.
Figure 14b–d shows the relationship between the height difference ∆
h and the number of beams
Nl, beam width
lx, and beam length
ly, respectively. It can be seen that the height difference is negatively correlated with these parameters. That is, as the number of beams increases and the width of the beam increases, the traction force on the mass block increases, and the displacement of the vibrator affected by gravity decreases, resulting in a smaller height difference. As the beam becomes longer, the traction force on the central mass block becomes weaker and the height difference becomes greater. From the figure, it can be seen that the height difference can be controlled within 4 nm by increasing the number of beams, increasing the width of beams, and reducing the length of beams. Due to the fact that beam parameters can also affect resonant frequency, it is not only necessary to consider the height difference during actual manufacturing, but also to obtain appropriate parameters based on factors such as actual needs.