## Abstract

A new phononic crystal with the graded supercell configuration is proposed to broaden the Bragg scattering band gaps. The graded structural design can merge adjacent multiple band gaps into an extremely broad one. The proposed phononic crystal is made up of a periodic arrangement of supercells, and the supercells are composed of unit cells with graded structural parameters. The mechanical model of the graded phononic crystals is established based on the transfer matrix method to investigate in-plane elastic waves propagating and band structures of the periodic system. Modal analysis shows that the mechanism for the broadening of band gap is that the graded supercell configuration breaks some symmetries of the phononic crystal, resulting in the opening of the Dirac cone and creation of new band gaps. The effects of the main structural parameters related to graded supercell design on band gap broadening are studied by simulation and verified by the experiment. The present study is beneficial to the design of new functional materials with broadband vibration isolation performance.

## 1 Introduction

Phononic crystals (PnCs), whose material properties or lattice parameters distribute in a periodic way, are engineered periodic composite materials or structures with band gaps [1], in which the elastic wave propagation is prohibited [2]. In recent years, PnCs have become a research hotspot in the field of elastic and acoustic waves due to adjustable band structures and strong manipulability in attenuating mechanical waves at various frequencies [3–6]. Nevertheless, it still exists many challenges for the industrial applications of phononic crystals, one of which is that traditional PnCs tend to manipulate elastic waves only in a limited frequency range, which limits their engineering applications to a great extent [7,8].

Phononic crystals include Bragg scattering PnCs [9,10] and locally resonant phononic crystals [11] based on the mechanism for the creation of band gaps. The wavelength of the elastic wave in the Bragg scattering band gap is about the same order of magnitude as the periodic lattice size, so the low-frequency band gap is usually realized at the cost of a very large size [12]. To solve this problem, many approaches have been proposed. For example, it is found that using negative stiffness in a unit cell leads to decreasing the frequency range of band gaps [13]. Besides, Bragg scattering phononic crystals have good broadband performance compared to locally resonant phononic crystals [14]. It is expected that the Bragg scattering phonon crystals can achieve the ultra-broadband vibration isolation and noise reduction for engineering applications.

Functionally graded materials (FGMs) represent a kind of new type high-performance composites with gradually varying material contents [15–20], which gives us insights into the design of broadband PnCs. The graded design is currently applied to the local resonant phononic crystals for the purpose of broadening the band gap [21–26], while less research has been done on Bragg scattering phononic. Kushwaha and Halevi [27] first achieved ultra-wideband filter of noise by placing five groups of periodic metal rods in series, so that the band gaps of each group of structures are superimposed on each other. Kushwaha and Halevi's design formed the prototype of a graded structure to broaden the band gap. Recently, studies on graded Bragg scattering phononic mainly focuses on the elastic wave focusing, especially acoustic wave focusing, which is demonstrated to dissipate more energy and improve the sound absorption performance in a wide range of frequencies [28–33]. The graded design on Bragg scattering phononic is used to change the width and the central frequency of the band gaps by introducing the graded profile of the interlayer with piezoelectric [34], decoupled double defects [35], or different FGM compositions and geometry parameters [36], so as to design the desired filters. The existing research on the graded design only operates on a single band gap, but has not realized the merging of adjacent multiple band gaps into an extremely broad one. Therefore, a new mechanism needs to be introduced to realize the merging of adjacent multiple band gaps.

According to previous studies, the simplification of some fluctuation modes at high symmetry points in the Brillouin region can be eliminated by reducing the structural symmetry, thus opening new band gaps, or shifting the band gaps toward the low-frequency region [37–39]. With graded supercell configuration, the Bragg scattering phononic crystal can retain most of the band gaps of the nongraded structures to broaden the band gaps well [40]. With a wider original bandwidth, it is expected that Bragg scattering phononic crystals can achieve an extreme broad band gap easier, so as to satisfy the demands of the broadband vibration reduction in engineering applications.

Our group has done some work on graded hierarchical phononic crystals in the previous study [40]. This article is an extension of the previous work. First, we only studied the ideal simple layered model in the previous study, and in this article, the focus is on the engineering application structures based on the folded beam structure. Compared to the ideal simple layered model, the structure proposed in this article owns lower and more suitable range of frequency for engineering applications. Second, the different theories of beam model (Timoshenko beam and Euler—Bernoulli beam) are considered for the calculation of the band gap, and the applicable range of two kinds of beam model in graded design is discussed. To apply on engineering better, different structural parameters are calculated for different engineering applications. Furthermore, the experiment is taken to verify the conclusions we obtain from the theoretical calculations and numerical simulations. The broadband vibration isolation in engineering applications of the proposed graded structure has got a fuller discussion.

In this article, to broaden the Bragg scattering band gaps extremely, a new graded phononic crystal composed of periodic arrangement of supercells with graded structural parameters is developed, and the mechanism for the broadening of band gap is revealed. This study is organized as follows: In Sec. 2, basic assumptions and the structural model are presented. Also, the theoretical model is established by the transfer matrix method. In Sec. 3, the band structures and the frequency response function of the graded phononic crystal are numerically compared with that of the nongraded structures. The mechanism for graded structures to broaden band gaps is revealed by the analysis of modes. Finally, the effect of structural parameters on band gap broadening is studied numerically. In Sec. 4, the broadband vibration isolation of the proposed graded structure is verified by an experiment. In Sec. 5, the main conclusions are summarized.

## 2 Models and Methods

Folded beam phononic crystals have been shown to have a low broadband gap [41]. In this article, we take a folded beam-type phononic crystal as an example and further broaden its band gap by introducing a graded structural configuration.

In Fig. 1, the graded folded beam phononic crystals model is shown. The two-dimensional graded structure is shown in Fig. 1(a). It consists of supercells arranged periodically in the *x*-direction and *z*-direction. The grades are introduced through the increment of the thickness *h*_{i} of beams in the *z*-direction. Different unit cells own different thicknesses of beams. The graded value is defined as follows: *dh* = *h*_{i+1} − *h*_{i}. The graded order *g* refers to the number of cells in a supercell. Note that the grades only work in the *z*-direction, so only the wave conduction in the *z*-direction is considered this article. A supercell consists of several unit cells. The unit cell model is shown in Fig. 1(b). Each unit cell is a symmetrical folding beam structure. As shown in Fig. 1(c), the axisymmetric model is taken to introduce the composition of each unit cell. The part marked by dotted lines is the axisymmetric model of the *n*th unit cell. It is composed of two beams I and II with the same thickness, two mass blocks M_{1} connecting the adjacent unit cells in the center, and two mass blocks M_{2} connecting the beams on the left and right. The distance between the central layers of the two beams is kept constant, which is denoted as *s*. The lattice constant of the unit cell is *a*_{0}. The thickness of the beam is *h*_{i}.

Before proceeding to the solution, we make the following assumptions:

The material used in the structure is homogeneous, isotropic, and linearly elastic.

The mechanic model is simplified as a two-dimensional plane strain problem.

The mass blocks M

_{1}and M_{2}are assumed to be rigid bodies.Because only the propagation of elastic wave in the

*z*-direction is concerned, the degree-of-freedom in the*z*-direction of M_{1}and M_{2}is considered, while the*x*directional and rotational degree-of-freedom are ignored.Considering the symmetry of the structure, beam I and II will only be excited in the symmetric mode, and its antisymmetric mode will be ignored.

The transfer matrix method is mostly used to calculate the dispersion relations and the frequency response function of the one-dimensional phononic crystal. In the process of applying the transfer matrix method, it is only necessary to assume the displacement field as a linear combination of trigonometric functions, hyperbolic trigonometric functions. Since the structure proposed in this article is left-right symmetric, only its left-right symmetric part is taken for calculation in the analysis.

*x*-axis, the direction perpendicular to

*x*-axis in the working plane is taken as

*z*-axis. The system of coupled differential equations for transverse vibration of the uniform Timoshenko beam without external force is given by

*I*,

*ρ*, and

*S*are, respectively, the second moment of the section, the density, and the cross-sectional area of the beam.

*Q*(

*x*,

*t*) =

*κGSγ*(

*x*,

*t*),

*κ*is the Timoshenko shear coefficient depending on the cross section of the beam [42,43], in rectangular section,

*κ*= 5/6, and

*γ*(

*x*,

*t*) is the shear angle.

*u*(

*x*,

*t*) and

*θ*(

*x*,

*t*) can be written as

*u*(

*x*,

*t*) =

*U*(

*x*)

*T*(

*t*),

*θ*(

*x*,

*t*) =

*Y*(

*x*)

*T*(

*t*). With such an assumption, for $\omega <\kappa GS/(I\rho )$, $Ujn(x)$ can be written as follows:

*j =*I, II, indicating different beams in the same unit cell,

*n*denotes the

*n*th unit cell. $\lambda 12=(\u2212d+\Delta )/2$, $\lambda 22=(d+\Delta )/2$,

*d*=

*ω*

^{2}

*ρI*(1 +

*E*/

*Gκ*)/EI, Δ =

*ω*

^{4}

*ρ*

^{2}

*I*

^{2}(1 −

*E*/

*Gκ*)

^{2}+ 4EI

*ω*

^{2}

*ρS*, $Ajn$, $Bjn$, $Cjn$, and $Djn$ are the coefficient to be determined.

*z*-axis direction, we can assume that the mass block and the connecting block are rigid bodies with only translational motion and without rotation; thus, the rotation angle of the beam at

*x*

_{1}and

*x*

_{2}is described as follows:

*z*-directional displacement of the intermediate mass block $uM1n(x1,t)=uIn(x1,t)$.

*z*-directional displacement of the mass blocks on both sides $uM2n(x2,t)=uIIn(x2,t)$.

*k*is the

*z*-directional Bloch wave vector, and then the standard matrix eigenvalue problem can be obtained as follows:

The dispersion relation between the wave vector *k* and the frequency *f* can be obtained by solving the eigenvalues of the matrix **T** [44].

## 3 Results and Discussion

### 3.1 The Numerical Results of Bad Gap.

All the structures in this article are made of polylactic acid (PLA), and the specific material parameters are presented in Table 1. The geometric parameters are presented in Table 2.

Name | Value | Description |
---|---|---|

E | 3.13E9 Pa | Young's modulus |

ρ | 1240 kg/m^{3} | Density |

ν | 0.18 | Poisson's ratio |

Name | Value | Description |
---|---|---|

E | 3.13E9 Pa | Young's modulus |

ρ | 1240 kg/m^{3} | Density |

ν | 0.18 | Poisson's ratio |

Name | Value (mm) | Description |
---|---|---|

a_{0} | 16 | Lattice constants |

l | 140 | Width |

s | 8 | Distance between the middle layers of the two crossbeams |

oh | 2 | Initial thickness of beam |

dh | 0 | Graded value |

h_{i} | oh + dh*i | Thickness of beam |

b_{1} | 5 | Width of M2 |

b_{2} | 10 | Width of M1 |

d_{i} | (a_{0}_{−}m_{−}h)/2_{i} | Half height of M1 |

g | 5 | Graded order |

N | 1 | Number of supercells in z-direction |

th | 15 | 3D model depth |

l_{1} | l/2 − b_{2}/2 − b_{1} | Length of beam |

ld | l_{1}/h_{i} | Length-to-diameter ratio |

Name | Value (mm) | Description |
---|---|---|

a_{0} | 16 | Lattice constants |

l | 140 | Width |

s | 8 | Distance between the middle layers of the two crossbeams |

oh | 2 | Initial thickness of beam |

dh | 0 | Graded value |

h_{i} | oh + dh*i | Thickness of beam |

b_{1} | 5 | Width of M2 |

b_{2} | 10 | Width of M1 |

d_{i} | (a_{0}_{−}m_{−}h)/2_{i} | Half height of M1 |

g | 5 | Graded order |

N | 1 | Number of supercells in z-direction |

th | 15 | 3D model depth |

l_{1} | l/2 − b_{2}/2 − b_{1} | Length of beam |

ld | l_{1}/h_{i} | Length-to-diameter ratio |

To verify the theoretical model and determine the applicable length-to-diameter ratio (*ld)* range for Euler—Bernoulli beam, the first ordered band gap of the nongraded PnCs of the following three models is calculated and compared: (i) the theoretical Euler—Bernoulli beam model; (ii) the theoretical Timoshenko model; and (iii) the finite element method (FEM) with the comsol multiphysics software, and the result is presented in Fig. 2. The result shows that the Timoshenko beam model fits well with the FEM method. When *ld* > 10, the Euler—Bernoulli beam can fit well, too. So in the case that *ld* < 10, it is better to use the Timoshenko beam model. Once *ld* > 10, the Euler—Bernoulli beam model can be used.

To obtain band gaps with low frequency, the slender beam is selected here to discuss the influence of grades (*ld* = 30). Thus, the Euler—Bernoulli beam is used to do the following calculation. The band structure of the graded structure was calculated, and the results are plotted in Fig. 3. In Fig. 3(a), as a Bragg scattering phononic crystal, the first-order bandgap onset frequency of the original structure reaches to 289 Hz, which is comparable to the lowest frequency that can be reached in the bandgap of some local resonant phononic crystals [46]. In Figs. 3(a)–3(c), there exists a wide passband between the first- and second-order bandgaps in the graded-free structure. Observing Fig. 3(d), it can be found that the graded structure can broaden the bandgap frequency range without increasing the starting frequency of the structural bandgap. The fifth-order graded structure has many flat bands in the frequency range of 225.5–3000 Hz, and in these flat bands, the elastic wave cannot propagate. So the band gap frequency range of the fifth-order graded structure is broadened to 225.5–3000 Hz, which reduces the band gap starting frequency while greatly broadening the bandgap frequency range of the structure without grades.

Further, the frequency response function of the finite-period graded structure is calculated. A fifth-order graded two-period structure is considered, and its vibration transmission loss is calculated. In acoustic and vibration engineering, the amplitude–frequency response function is generally defined as follows: TL = 20log _{10}*X*/*X*_{0}, where *X* and *X*_{0} are the displacements, velocities or accelerations at the output and excitation ends, respectively. The displacements at both ends are taken as *X* and *X*_{0}. Given the free boundary conditions, the broadening of the band gap by the graded structure is investigated while ensuring that the geometric and material parameters of the finite-period graded structure are the same as those of the infinite-period graded structure. The frequency response curves of the fifth-order graded structure with an initial thickness of 2 mm and a grade of 0.5 mm are shown in Fig. 4. The blue ranges indicate the band gaps. These bandgap ranges are in consistent with those in the band structure of Fig. 3(d). It can be observed from the figure that the input vibration cannot propagate in several frequency ranges, and the broadening effect of the graded structure on the band gap is confirmed.

### 3.2 Mechanism of Band Gaps Broadening.

To investigate the mechanism for the graded structure to broaden the band gap, we take the two-order graded structure with Euler—Bernoulli beam (*ld* = 30) as an example and performed calculations. The taken graded structure consists of two unit cells. They have the same material parameters but different beam thicknesses with lattice constants 2*a*_{0}, and other parameters are the same as presented in Table 2. Figure 5 shows the band structure corresponding to the supercell structures with *dh* = 0, 0.1, 0.5, and 1 mm. The blue area is the part of the band gap, where the band gaps of the graded structure overlap with the graded-free structure, and the yellow area is the new band gap generated by the graded structure. The comparison of Fig. 5(a) with Figs. 5(b)–5(d) reveals two features of the graded structure in the band gap broadening: (i) the retention of the band gap of the nongraded structure and (ii) the generation of new band gaps in the middle- and high-frequency ranges.

To explain the mechanism for the broadening of band gap in the graded structure, we first analyze the mechanism for the formation of band gap in the original, nongraded structure. In the analysis of the dispersion curves, the propagation of the vertically incident elastic wave in the graded structure corresponds to its wave propagation along the z-direction of the first Brillouin zone. In Figs. 5(e)–5(h), each mode corresponds to each ordinary bandgap boundary point. Looking at the mode diagram, it can be found that the lower boundary points of the band gap in Figs. 5(f) and 5(h) own an antisymmetric mode of beam I and II, while the upper boundary points of the band gap in Figs. 5(e) and 5(g) own an symmetric mode of beam I and II. In the vibration process, the blocks M1 and M2 vibrate as rigid bodies, which lead to the interference of the scattered transverse waves by the phase extinction of the structure in the process of elastic wave propagation. In simple terms, the blocks M1 and M2 have infinite bending stiffness compared to the beams. This also leads to a sudden change in bending stiffness experienced periodically during elastic wave propagation, resulting in a strong wave reflection that cancels out the incident wave, thus inhibiting wave propagation and forming a band gap.

The graded structure retains the band gap of the nongraded structure to a large extent. Observing the vibrational modes of the overlapping part of the nongraded structure and the band gap in Figs. 5(e)–5(h), we can see that the vibrational modes corresponding to the onset frequency and the cutoff frequency of the band gap are the same. So it can be inferred that the band gap of the graded structure, whose frequency is overlapped with the nongraded structure, is a superposition of the Bragg bandgap in the graded-free structure.

Another feature of the graded structure is the generation of new band gaps in the middle- and high-frequency range. When *dh* = 0, the structure has mirror symmetry, so the supercell structure can degenerate into a unit cell phononic crystal with a lattice constant *a*_{0}. At the boundary of the first Brillouin zone, *k* = π/*a*_{0} in Fig. 5, as shown in *S*_{1} and *S*_{2}, there will be double degeneracy due to the symmetry of the structure, and the Dirac cone of the dispersion is created at the same time. This symmetry of the structure is reflected in the mode as local modal symmetry. Due to the periodicity of PnCs, there is no sequential distinction between the two unit cells in the modal diagram. Taking the modes in Figs. 5(k) and 5(l) as an example, the upper unit cell's mode of Fig. 5(k) and the lower cell's mode of (l) are symmetric modes, while the remaining other layer is antisymmetric.

To broaden the band gap as much as possible, many scholars have proposed methods to remove the simplicity in recent years, such as introducing material anisotropy [47] and changing the structure to reduce the symmetry of the structure [47,48]. When *dh* ≠ 0, as shown in Fig. 5(d), *dh* = 1 mm, the mirror symmetry of the structure is broken, the double degeneracy disappears, and the Dirac cone opens, resulting in a new bandgap. To some extent, the frequency of the original bandgap is also affected. For example, when *dh* increases from 0 to 0.1 mm, the frequency range of the second band gap increases from 1221.9–2385.6 Hz to 1259.3–2426.4 Hz. The original double degeneracy, such as the dispersion curve *S*_{1} and *S*_{2}, gradually opens into a flat band, forming a new band gap in the passband range of Fig. 5(a). Thus, the introduction of the graded structure breaks the symmetry and generates a new band gap, broadening the band gap. At the same time, some low-order dispersion curves are shifted to the high-frequency region, thus reducing the bandgap frequency to a certain extent.

The aforementioned phenomena are similar to those observed in one-dimensional layered composites with hierarchical structures [49] and in two-dimensional solid–liquid systems with fractal structures [50]. The difference is that the graded structure proposed in this article is assembled by a simple variation of structural dimensions in one direction, whereas the hierarchical and fractal structures have differences in internal and external gradations and shapes.

Qualitatively, the introduction of the graded structure brings dimensional diversity in the super cell, and this diversity remains most of the bandgap corresponding to the unit cell nongraded structure without increasing the bandgap onset frequency, and so the overall bandwidth is broadened.

### 3.3 Analysis of Structural Parameters on Band Gap Broadening.

The effects of the main structural factors related to graded design on band gap broadening are studied in this section, which includes the length-to-diameter ratio *ld* and the design parameters of graded structure: the graded order *g* and the graded value *dh*. The graded supercell configuration with Euler—Bernoulli beam is taken as the example to analyze the influence of the aforementioned parameters on bandgap broadening.

#### 3.3.1 Structural Parameter ld.

Figure 6 shows the change of band gap with different parameters *ld* in the unit cell. The black part in the figure is the band gap, and the solid line and dotted line, respectively, represent the upper and lower boundaries of the band gap. Here, the band gap between the second and third band is defined as the first band gap, and the band gap between the fourth and fifth band is defined as the second band gap. With the increase of *ld*, the upper and lower boundaries of the band gap decrease gradually, the frequency range of the band gap decreases, and the width of the band gap also decreases. Because as *ld* increases, the stiffness of the beam decreases and the band gap frequency decreases.

For graded structures, *ld* increasing in the arithmetic sequence has a great influence on the band gap broadening effect. Next, the influence of graded order *g* on the band gap broadening effect is discussed.

#### 3.3.2 Graded Order g.

The graded order *g* is a significant factor on band gap broadening. Figure 7 displays the band structure of graded PnCs with graded order *g* = 2,3,4,5. Note that to reduce the influence of the gradient value, a small graded value *dh* = 0.3 mm is adopted here. The gray area is the band gap range. It can be seen that when the graded order changes from 2 to 3, the band gap is greatly broadened, but there are still some narrow passbands. However, when the gradient order rises to 4 or 5, the band gap widening effect does not change significantly, but the narrow passbands gradually disappear, and the emergence of the two band gaps is completely realized, resulting in the band gap broadening. This is because that due to the big graded order, there exists a great difference in stiffness within a supercell, resulting in a great difference in band gap frequency. Finally, a wider frequency range after the superposition of original band gaps appears.

Therefore, it can be concluded that the graded order has a step change in the band gap broadening, and the larger the order is, the better the band gap broadens.

#### 3.3.3 Graded Value dh.

Figure 8 shows the effect of the graded value on the bandgap broadening. To control the variables, the graded level *g* = 2 with the same order, the same initial thickness, and the same structural and geometric parameters except for the graded value is chosen here to calculate their band gaps. In Fig. 8, the black part is the bandgap frequency range, and the white is the passband range. It is easy to find that when the graded value is small, *dh* < 0.5 mm, the graded structure cannot broaden the bandgap well. When the graded value gradually increases, the bandgap can be broadened better, and the passband range of the white part gradually decreases. When 1 < *dh* < 2 mm, the graded broadening effect reaches the best among the calculated graded values. With the graded value increasing to more than 2 mm, the passband range tends to become wider again and the graded broadening effect weakens.

As mentioned earlier, one of the mechanisms for broadening the bandgap of the graded structure is to retain most of the bandgap of the nongraded structure. The larger the graded value is, the larger the difference in stiffness between adjacent cells is, which leads to a great difference in bandgap frequencies generated at different orders. When the graded value is small, the corresponding bandgap frequency ranges of adjacent units are close to each other, and the original passbands cannot be effectively eliminated; while when the graded value is too high, the corresponding bandgap frequency ranges of adjacent units are different, and the higher-order passbands generated by units with smaller stiffness and the lower-order passbands generated by units with larger stiffness are likely to have similar frequencies, and the coupling effect in this case cannot broaden the bandgap well.

It can be concluded that a moderate value of grades can extremely broaden the bandgap, while when the graded value is too large or too small, it can be difficult for the graded structure to broaden the bandgap.

## 4 Experimental Verification

The band gap broadening of the graded PnCs has been verified by the computational simulation of the band structure and the frequency response function mentioned earlier. In this section, the band gap broadening of the graded PnCs will be further verified by the experimental test. In Fig. 9, the schematic diagram of the experiment is shown. The following experimental equipment was used: lms test.lab 17 software, ASUS VM520U laptop (signal generator), SIEMENS SCM202 signal processor, Brüel & Kjær 2716C power amplifier, Brüel & Kjær 4825 shaker, two acceleration transducers, and an impedance head. The excitation signal is given from the PC, and it passed through the signal processor and power amplifier, and transmitted to the shaker, where the shaker impedance head is placed against one end of the specimen to give the initial excitation. The signal processor collects the signals from the two sensors and feeds them back to the terminal. The signal analysis software LMS is used to process the signals and finally obtain the frequency response function of the specimen at each initial excitation frequency. The sample is placed on a glass plate with a very small coefficient of friction, which realizes the side restraint to restrain the rotational mode of the mass block. The experimental specimen is made of the same PLA material as in the simulation, and the material parameters are presented in Table 1. The geometric parameters are the same as those in the simulation, as shown in Table 2.

Figure 10 shows the transmission comparison of experiments and simulations. The results obtained from the simulation show the existence of band gaps (blue part) in the frequency ranges of 282–868 Hz, 1212–2370 Hz, and 2831–3000 Hz. As a comparison, the experiment takes the sample with the same material and geometric parameters, and the result is shown in Fig. 10(b). The black part is the overlap of the experimental and simulation band gap. Comparing the experimental and simulation results, the band gap frequency range is not much different and the values are within the acceptable range. The passband obtained from the experimental test is narrower and the attenuation is smaller. This is because of the difference between the simulation and experimental test conditions, i.e., the simulation calculation does not consider damping at all, while the experimental test has unavoidable damp. Second, there exists uncertainty in the experiment itself and the incomplete balance of the experimental setup.

Figure 11 shows the transmission comparison of experimentally measured graded structure and graded-free structure. It is not difficult to find that the graded-free structure has several band gaps in the frequency range of 147–968 Hz, 1010–2555 Hz, and 2625–3000 Hz, while the graded structure merges adjacent multiple band gaps into an extremely broad one in the frequency range of 214–3000 Hz. This once again demonstrates that the graded design can have a great broadening effect on the band gap, merging adjacent multiple band gaps into an extremely broad one.

## 5 Conclusion

A new graded phononic crystal to broaden the Bragg scattering band gaps is proposed. The ability of graded PnCs to broaden the bandgap is demonstrated by theory, simulation, and experiment. The effects of the main structural factors related to graded design on band gap broadening are studied. The main conclusions are as follows:

The graded design can merge adjacent multiple band gaps into an extremely broad one, broaden the band gaps of Bragg scattering phononic crystals extremely.

The graded PnCs achieve the broadening of bandgaps mainly by (i) preserving the bandgap of the graded-free structure and (ii) breaking the mirror symmetry of the structure by introducing a graded change in the size of the structure, thus opening the Dirac cone and forming a new bandgap.

Structural parameters related to graded design have a significant influence on band gap broadening: (i) the larger the graded order is, the better the band gap broadens; (ii) a moderate graded value of an extremely broaden the bandgap. While the graded value is too large or too small, it can be difficult to broaden the bandgap.

## Funding Data

National Natural Science Foundation of China (Grant Nos. 51775201 and 52175095).

The Young Top-notch Talent Cultivation Program of Hubei Province, China.

## Conflict of Interest

There are no conflicts of interest.

## Data Availability Statement

The datasets generated and supporting the findings of this article are obtainable from the corresponding author upon reasonable request.