DETERMINATION OF PHYSICAL RESPONSE IN (Mo:AlN) SAW DEVICES

(1)

DETERMINATION OF PHYSICAL RESPONSE

IN (Mo/AlN) SAW DEVICES

J. C. CAICEDO*,||

_,

J. A. PÉREZ†,‡, H. H. CAICEDO§,¶and H. RIASCOS‡ *Tribology Polymers,

Powder Metallurgy and Processing of Solid Waste Research Group, Universidad del Valle, CaliColombia

†_{Synchrotron Radiation and Particle Accelerators,}

Universidad Autonoma Barcelona, Spain

‡_{Departamento de F}_{_sica,

Universidad Tecnologica de Pereira, Grupo Plasma, Laser y Aplicaciones A. A 097, Colombia

§_{Department of Bioengineering,}

University of Illinois at Chicago, IL 60612, USA

¶_{Department of Anatomy and Cell Biology,}

University of Illinois at Chicago, IL 60612, USA

||_{[email protected]}

Received 22 August 2012 Revised 16 January 2013 Accepted 18 January 2013

Published 28 March 2013

This paper describes the experimental conditions in surface acoustic wave (SAW) designed on alu-minum nitride (AlN) ¯lms grown on Si3N4substrates by using pulsed laser deposition. Moreover it was studied the dependency of optical properties with temperature of deposition. The thickness, measured by pro¯lometry technology, was 150 nm for all ¯lms. Moreover, SAW devices with a Mo/AlN/Si3N4

con¯guration were fabricated employing AlN bu®er and Mo Channel. The morphology and compo-sition of the ¯lms were studied using atomic force microscopy (AFM), scanning electron microscopy (SEM) and X-ray photoelectron spectroscopy analysis (XPS), respectively. The optical re°ectance spectra and color coordinates of the ¯lms were obtained by optical spectral re°ectometry technique in the range of 400900 cm1_{. In this work, a clear dependence in morphological properties, optical}

properties, frequency response and acoustic wave velocity as function of applied deposition temper-ature was found. It was also observed a reduction in re°ectance of about 10% and an increase of acoustic wave velocity of about 1.2% when the temperature was increased from 200C to 630C.

Keywords: Aluminum nitride; pulsed laser deposition; optical re°ectance; color purity frequency response; surface acoustic wave.

1. Introduction

Aluminum nitride (AlN), an electrical insulating ceramic with a wide bandgap of 6.3 eV, is a poten-tially useful dielectric material very important in

¯elds such as optoelectronics and microelectronics. For instance, it may serve as the dielectric gate in high voltage and high-power electronic devices; AlN is also being investigated as a substitute for the silicon

Surface Review and Letters, Vol. 20, No. 2 (2013) 1350017 (15 pages)

°c World Scienti¯c Publishing Company DOI:10.1142/S0218625X13500170

Surf. Rev. Lett. Downloaded from www.worldscientific.com

(2)

dioxide buried layer in silicon-on-insulator (SOI) substrates due to its low thermal expansion coe±-cient, high breakdown dielectric strength and high chemical and thermal stability.1 _{Moreover, metal}

nitrides based on AlN are of high interest because of their electrical, optical and acoustical properties.2

Their hardness and thermal coe±cients of expansion are comparable to that of Si. AlN is used for acoustic wave devices on Si, optical coatings for spacecraft components, heat sinks in electronic packaging applications, as well as electroluminescent devices in the wavelength range from 215 nm to the blue end of the optical spectrum.2_{Some optical properties near}

the fundamental bandgap of AlN have been reported by Yamashita et al.3_{since 1979. Several theoretical}

studies have been well documented in the past4,5_;

however, reliable experimental data2_{has been}

repor-ted just recently because both the sample quality and the spectroscopic techniques have only been sub-stantially improved in recent years. Precise knowl-edge of the optical constants is particularly important in view of the use of AlN thin ¯lms in optical ¯lters and light emitting laser diodes since they also show temperature dependence.6

Pulsed laser deposition (PLD) is largely applied for processing thin ¯lms and other structures. The highly no equilibrium nature of the PLD process is attractive for the synthesis of stoichiometric thin ¯lms of various metal nitrides, carbon nitrides, oxides and oxi-nitrides from the corresponding bulk targets. PLD appears to be a suitable method to transfer stoichiometric ally complex monolayer structures from AlN ceramic structures to substrates. Thus, from AlN stoichiometric ¯lms many infrared (IR) measurements presented in the literature were used to estimate the ¯lm quality. The IR spectra provided important information about the composition, ho-mogeneity, crystallinity and the residual stresses present in the ¯lms. Moreover, the ellipsometric and re°ectance measurements have been used to deter-mine the refractive index in many AlN compounds, since the ellipsometry and re°ectance are sensitive and nondestructive techniques used for studies of optical properties and microstructures of surfaces and thin ¯lms.2 _{Additionally, single-chip front end RF}

modules incorporating surface acoustic wave (SAW) ¯lters are a matter of intense research. At present these modules are fabricated by bonding single crystal piezoelectric SAW devices onto integrated circuits.

Thin ¯lms of polycrystalline AlN are promising materials for the integration of SAW devices on Si3N4 substrates due to their good piezoelectric properties and the possibility of deposition at low temperature compatible with the manufacturing of Si integrated circuits. AlN thin ¯lms of su±cient quality for SAW applications can be obtained by the PLD technique.7

The goal of this work is to study the e®ect of the applied deposition temperature on the chemical, morphological properties, optical properties, fre-quency response and acoustic wave velocity of binary AlN ¯lms deposited by PLD on Si3N4(100) for use in optical and electronic applications. Here, using nitrogen as working gas, we report results on AlN ¯lms deposited from Al targets, their characterization by X-ray photoelectron spectroscopy (XPS) and scanning electron microscopy (SEM) as well as re°ectance investigations.

2. Experimental

Experiments were carried out in standard PLD con-¯guration consisting of a laser system, a multiport stainless steel vacuum chamber equipped with a gas inlet, a rotating target and a heated substrate holder. A Nd:YAG laser that provides pulses at the wave-length of 1064 nm with 9 ns pulse duration and repe-tition rate 10 Hz was used. The laser beam was focused with anf ¼23 cm glass lens on the target at the angle of 45, with respect to the normal. During deposition, the substrate was rotated at a low speed of 2.2 rpm for enhancing the thickness and uniformity of deposited ¯lms. The distance between the target and the sub-strate was 6.5 cm. Before deposition, the vacuum chamber was evacuated down to 1105_{mbar by using} a turbomolecular pump backed with a rotary pump. The ¯lms were deposited in a nitrogen atmosphere as working gas; the target was an aluminum high purity (99.99%) target. The ¯lms were deposited with a laser °uence of 7 J/cm2for 10 min on silicon nitride (100) substrates (Si3N4). The working pressure was 9103_{mbar and the substrate temperature were} varied from 200C to 630C. In a previous work,8_we

showed that most of the piezoelectrically driven devices require AlN ¯lms containing a high volume of microcrystals oriented with the c-axis normal to the surface and with the same crystallographic polarity, as this ensures the highest piezoelectric response. The

(3)

thickness, measured by pro¯lometry technology, was 150 nm for all high-quality polycrystalline piezoelec-tric (0002) AlN thin ¯lms, as shown in reference.8_For

each ¯lms the deposition time was 10 min. XPS was used on AlN samples to determine the chemical com-position and the bonding of aluminum and nitrogen atoms using ESCA-PHI 5500 monochromatic AlK radiation and a passing energy of 0.1 eV. The surface sensitivity of this technique is so high that any con-tamination can produce deviations from the real chemical composition; therefore, the XPS analysis is typically performed under ultra high vacuum condi-tions with a sputter cleaning source to remove any undesired contaminants.

Morphologic characteristics of the coatings like grain size and roughness were obtained using an atomic force microscopy (AFM) from Asylum Research MFP-3Dr and calculated by a scanning probe image processor (SPIPr) which is the standard program for processing and presenting AFM data, therefore, this software has become the de-facto standard for image processing in nanoscale.

Optical re°ectance spectra and color coordinates of the samples were obtained by spectral re°ectome-try in the range 400900 cm1_{by means of an Ocean} Optics 2000 spectrophotometer. The coated samples received the white light from a halogen lamp illumi-nator through a bundle of six optical ¯bers, and the light re°ected on the samples was collected by a single optical ¯ber and analyzed in the spectrophotometer. The ¯ber was ¯xed in perpendicular direction to the sample surface. An aluminum mirror ¯lm freshly de-posited by rapid thermal evaporation in high vacuum was used as the reference sample, and the experi-mental spectra were normalized to 100% re°ectance of the reference sample. The morphology and micro-device structure (SAW) on AlN surface ¯lms was analyzed by SEM (Leika 360 Cambridge Instru-ments).

To determine the quality of the AlN piezoelectric SAW devices this one has been fabricated on thin ¯lms of AlN with di®erent deposition temperatures. The manufacture of the ¯lters is to de¯ne interdigi-tated electrodes of Mo of 0.1 mm thick over the layer of piezoelectric material by a photolithography pro-cess, from a Mo target using a RF source, as substrate was used ¯lms AlN with 20 oscillations/min. For electrodes of Mo a deposition time 10 min with a °ow of argon of 25 sccm and an incident power of 800 W

was used. The periodicity of the interdigital trans-ducer (IDT) and the speed of propagation of the acoustic wave in the substrate determines the fre-quency with which the acoustic wave can propagate. In these devices, the IDT switch 50 pairs f ¯nger with a width of 10m, so that the acoustic wavelength is 40m. The area of the electrodes is 1.45102_cm2 which agrees with the mask geometry used.

Taking into account the deposition conditions of AlN ¯lm, the frequency response was measured by a network analyzer (HP 8720C. From the frequency responses, the insertion loss at center frequency was monitored and correlated with deposition tempera-ture and surface roughness of AlN ¯lms. The propa-gation of the SAW was studied through the analysis of the scattering parameters on Mo/AlN/Si trans-mission con¯gurations.

3. Results and Discussion

3.1. XPS analysis

The XPS survey spectrum for AlN thin ¯lm grown at 300C is shown in Fig.1. The peaks at 534.4, 397.3, 124.4 and 73.9 eV correspond to O1s, N1s, Al2s and Al2p binding energies, respectively. Calculation of the peak areas from the deconvoluted spectrum (see Fig. 2) without O1s contribution gives an atomic ratio of Al:N¼0:392:0.588, which is similar to the stoichiometry of Al0:40N0:60.8 The core electronic

1000 800 600 400 200 0

0 10000 20000 30000 40000 50000 60000 70000

Intensity (c/s)

Binding energy (eV)

Al-N

O (KLL)

O(1s)

N(1s)

Al(2s)

Al(2p)

Fig. 1. XPS survey spectrum of AlN coatings deposited on Si3N4 at 300C, where di®erent elemental signals are

showed.

(4)

spectra carry the information of the chemical com-position and bonding characters of the AlN ¯lms. The integral of N1s, A2s and Al2p spectra corrected by relevant sensitive factors can evaluate the con-centrations of Al and N elements in the AlN ¯lms. The corresponding integral of the deconvoluted peaks can also be used to estimate the bond contents, which are described by the following formula9_:

C_i¼XðA_i=S_iÞ=XðA_j=S_jÞ; ð1Þ

whereS is the sensitivity factor,Ais the integral of deconvoluted peaks andCis the atomic content. The numerator is the sum of the integral of one sort of bond; the denominator is the sum of the integral of all types of bonds decomposed from the whole peak of N1s, A2s and Al2p in the sample. The atomic con-centrations of Al, N and O elements from XPS

analysis of the ¯lm deposited at 300C are listed in Table1. For all the AlN ¯lms obtained, the O con-centration is less than 3 at.%. We can speculate that the majority of the oxygen at the surface of the non-treated sample (which is smooth and almost perfect) is chemically bonded into alumina-dominated oxide-nitride material.

The high-resolution spectra of Al2p and N1s, were recorded from AlN ¯lms, as shown in Fig. 2. From Fig. 2(a), the Al2p peak is composed of a shoulder separated by 1.7 eV with intense peak. The XPS spectrum of Al2p can be ¯tted well by two Gaussian functions. The value of binding energy obtained for the Al2p peak was 73.9 eV and the higher value for Al2p was 75.6 eV, respectively. According to the lit-erature1012_{for the Al2p peak, the ¯rst one (73.9 eV)}

and the second one (75.6 eV) can be assigned to AlN and AlO bonds.

82 80 78 76 74 72 70 68 66 64 0

200 400 600 800 1000 1200 1400 1600 1800

Intensity (c/s)

Binding energy (eV) Al2p Al-N

73.9 eV

Al-O 75.6 eV

(a)

408 406 404 402 400 398 396 394 392 390 0

200 400 600 800 1000 1200 1400 1600 1800

Intensity (c/s)

Binding energy (eV) N1s

N-N 400.2 eV

N-AL 397.3 eV

(b)

Fig. 2. High-resolution XPS spectrum of: (a) Al2p and (b) N1s of AlN ¯lm is around 300C.

Table 1. XPS analysis-derived atomic concentrations of Al and N elements in the AlN ¯lm synthesized by PLD with 7 mbar and substrate deposition temperatures as function of atomic percentage (at.%).

Substrate temperature (C) Pressure (103_mbar) _{Atomic composition (at.%)}

AlN ¯lms

Al N O

200 9 39.0 58.6 2.4

300 9 39.4 58.8 2.0

400 9 39.6 59.0 1.8

500 9 39.9 59.3 1.5

600 9 40.1 59.5 1.3

630 9 40.3 59.7 1.1

(5)

The appearance of the peak at 73.9 eV clearly shows that Al has reacted with N; therefore, it can be assigned to AlN.11,12 _{In Fig.} _2(b)_{, N1s peak is}

composed of spin doublets, each separated by 2.9 eV. The XPS spectrum of N1s, well ¯tted by two Gaussian functions, depicts the N1s spectrum with values at 397.3 and 400.2 eV characteristic for NN and AlN bonds, respectively.13

XPS results demonstrated that Al atoms bonded to N in the form of nitride; taking into account that ele-mental concentration of the AlN ¯lm was obtained by adjusting the laser incidence on Al target, and the N2was the working gas. In this research, it was found that amounts of Al(N) in AlN ¯lm were maximum at the current establishment of process conditions and the ratio of Al and N in the ¯lm was about 2:1. In general, formative Al(N) phase indicates that the aluminum and nitrogen activity and activation energy provided by the present deposition conditions are enough for the formation of a AlN thin ¯lm. Although the surface temperature of the substrate during deposition of AlN ¯lm is around 300C, the substrate lies in a high-density plasma region and a high ion-to-atom ratio of aluminum and nitrogen, that behavior can promote the formation of AlN phase at low temperatures below 300C. Moreover, in this work it a correlation was determined between temperature deposition and chemical properties. Due to the reduc-tion of oxygen concentrareduc-tion with the increase of deposition temperature, this behavior was associated to high surface energy activity generated by high sur-face temperatures that induce more nitrogen reaction. In order to see the e®ects of increased substrate temperature during thin ¯lms growth, the con-centrations in these depth pro¯les were obtained by curve-¯tting of high-resolution XPS spectra and measurements of the XPS peak areas. Thus, the AlN concentration is in accordance with the stoichiometry necessary to act like an AlN bu®er ¯lm in the MoSAW device. The dependence of aluminum and nitrogen in the deposition temperature is shown in Fig. 3, which represents various chemical states as described in Table1 and Fig.2.

3.2. Surface analysis of AlN ¯lms via

AFM

The scanning probe image processor (SPIPr) can be used for di®erent purposes including semiconductor

inspections, physics, chemistry and nanotechnology. In this work, the SPIPr was used to determine morphologic characteristics of the coatings. Grain size and roughness were obtained using an AFM from Asylum Research MFP-3Dr. Moreover, SPIP soft-ware packages are considered the de facto standard for processing 3D visualization of images taken by scanning probe microscopes. The software can import data through an algorithm that is able to reconstruct images from ASCII ¯les, bitmap and jpeg images and other general formats. Furthermore, SPIP was used on grain size analysis and surface roughness analysis. Nevertheless, most of the energy on the propa-gating SAWs is contained in the region within one wavelength below the surface, so they are very sen-sitive to the surface state and are a®ected by any change, such as residual stresses and the level of roughness. Moreover, authors such as Zhuang et al.

have shown that at the same frequency and in the small roughness limits (RMSwavelength), the SAW velocity increases according to the decrease in the RMS roughness.10 _Figure _4(a)_4(c) _{shows a}

5m5m image of AlN ¯lms deposited with dif-ferent deposition temperatures. The general trend in the AlN ¯lms with the decreasing temperature exhi-bits a decrease of the grain size and roughness. Figure4 also shows microdrops on AlN surface gen-erated by PLD energy deposition corresponding to high laser °uency. As this ¯gure was obtained by AFM, observed drops are characteristic of the PLD technique and are caused by the release of particu-lates from the Al target upon impact with the

200 300 400 500 600 700 38,8

39,0 39,2 39,4 39,6 39,8 40,0 40,2 40,4

Al N

Aluminium Concentration (at.%)

C

o

m

p

o

s

itio

n

(a

t.

%

)

58,6 58,8 59,0 59,2 59,4 59,6 59,8

C

o

m

pos

it

io

n (

a

t.

%

)

Fig. 3. Temperature dependence for AlN ¯lms at the concentrations of Al and N.

(6)

substrate; consequently, imperfections on the surface of the ¯lms are generated. In a previous work,14 _we

found a high aluminum content of roughly 60% in this area, thus corroborating the formation of droplets. Moreover, the correlation between roughness and grain size with deposition temperature is shown in Fig. 5. The quantitative values were extracted from the AFM images by means of statistical analysis (SPIPr). Each data that point in the plots represents an average over three AFM images. The corre-sponding error bar was obtained by a standard deviation of the values.

In Fig.5, it is possible to compare the grain size and roughness values for the AlN ¯lms, thus, in this

paper a decreasing tendency in surface measurements was observed when the deposition temperature was increased. The lower grain size and roughness values obtained in this study by the AlN ¯lms deposited on silicon substrate represented a decrease of grain size around 42% and roughness 31%, compared to that obtained for the AlN ¯lms deposited at 200C. This fact is relevant since the continuity of the surface is important for optical and electrical SAW devices.

The variation of grain size and roughness (see Fig. 5) obtained in the AlN ¯lms suggests that thermal activation generated by the increase in the deposition temperature on surface substrates modi¯es the super¯cial morphology by increasing the energy associated with atoms on the substrate surface and/ or the growing ¯lms surface which produces a relative reduction of grain size and roughness.

3.3. Dominant wavelength and color

purity analysis

The re°ectivity measure is the fractional amplitude of the re°ected electromagnetic ¯eld, while re°ectance refers to the fraction of incident electromagnetic power that is re°ected at an interface. The re°ectance is thus the square of the magnitude of the re°ectivity. The re°ectivity can be expressed as a complex num-ber as determined by the Fresnel equations for a single layer, whereas the re°ectance is always a pos-itive real number.15

In certain ¯elds, re°ectivity is distinguished from re°ectance by the fact that re°ectivity is a value that applies to thick re°ecting objects. When re°ection 200 300 400 500 600 700

35 40 45 50 55 60 65 70 75

70 80 90 100 110 120 130 140

Gr

ai

n

si

z

e

(

n

m

)

Temperature (°C)

AlN Grain size AlN Roughness

R

o

ugh

n

e

s

(n

m)

Fig. 5. Surface measurements obtained via AFM and SPIPr analysis for AlN ¯lms with a total thickness of 150 nm: Correlation between grain size and roughness with deposition temperature.

(a) (b) (c)

Fig. 4. AFM images for AlN ¯lms grown onto Si3N4substrate with di®erent deposition temperatures: (a) 200C, (b) 400C

and (c) 630C.

(7)

occurs from thin layers of material, internal re°ection e®ects can cause the re°ectance to vary with surface thickness. Re°ectivity is the limit value of re°ectance as the surface becomes thick; it is the intrinsic re-°ectance of the surface, hence irrespective of other parameters such as the re°ectance of the rear surface. On the other hand, the dominant wavelength (of a color stimulus) is de¯ned as: \the wavelength of the monochromatic stimulus that, when additively mixed in suitable proportions with the speci¯ed achromatic stimulus, matches the color stimulus considered".15

In this work, the AlN ¯lms has been compared with pure aluminum because the mirror ¯nish aluminum has the highest re°ectance of any metal in the 200500 nm and the 300010000 nm (far IR) regions, while in the 500700 nm visible range it is slightly outdone by AlN and silver, but in the 7003000 nm range (near IR) it is slightly outdone by, gold and copper materials.16

The re°ectance spectrum or spectral re°ectance curve is the plot of the re°ectivity as a function of wavelength as can be seen in Fig.6. Therefore, all the deposited ¯lms appear highly re°ective with a white-and aluminum-colored appearance. Figure6(a)shows the optical re°ectance spectra of the AlN single layers

obtained at di®erent deposition temperatures. The re°ectance of the aluminum and the eye sensibility are shown for comparison. The spectra of the samples show high re°ectance for long wavelengths, near to 33% for the AlN ¯lms deposited with 200C and close to 23% for AlN ¯lms deposited with 630C. These values of re°ectance at these wave numbers are in good agreement with previous reports for AlN ¯lms.17 A clear decrease in re°ectivity for short wavelengths is seen, characteristic of a system with high free-electron density18 _{with a re°ectance edge below}

550 nm due to a screened plasma resonance. The white and aluminum colors of the AlN ¯lms are a result of the steep plasma re°ection edge that occurs in the visible region where the re°ectivity minimum is around 560 nm.

The color purity refers to the degree of purity present at a color; therefore in a pure color do not exist any achromatic e®ect within its composition, which can vary its capacity for re°ection. One of the three dimensions of color refers to the color purity that a surface can re°ect. When the tone is full and is not mixed with an achromatic e®ect (brightness), this presents its maximum saturation or purity. On the other hand, if a particular color is combined with an

400 450 500 550 600 650 700 750 800 0

10 20 30 40 50 60 70 80 90 100

R

e

fle

c

ta

n

c

e

(%

)

Wavelength (nm)

Eye sensitivity Aluminium 200ºC 300ºC 400ºC 500ºC 600ºC 630ºC AlN films

(a)

0,0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,0

0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9

AlN 630 °C

AlN 200 °C 460 400

470 480 485

490 495 500 505 510

515 520 530

540

550

560

570

580

590 600

610 620

650

Y A

xi

s

X Axis

700 E

Green

Orange

Violet Blue Cyan

Yellow

Red Achromatic point

Line o f Pur

ples

Aluminum

(b)

Fig. 6. Dominant wavelength and color purity results: (a) Optical re°ectance of AlN ¯lms deposited onto Si3N4 (100)

substrates at di®erent deposition temperature, also aluminum optical re°ectance and eye sensibility were also plotted as references. (b) Chromatic diagram, in thex, y coordinates, of the re°ectivity for AlN ¯lms. White coordinates of achromatic point are located at (1/3, 1/3) (color online).

(8)

achromatic e®ect, or with its complement (luminosi-ty), it su®ers a loss of purity or neutralization. In this study, color purity changes with the temperature of deposition, at ranges from 0.63 to 0.66 for AlN ¯lms, away from color purity of pure aluminum (0.90), hence con¯rming that the ¯lms are less white com-pared with the aluminum. Color coordinates of AlN single layer ¯lms are shown in Fig.6(b).

Reference coordinates of pure aluminum are shown for comparison in that ¯gure. The results of color measurement indicate that all ¯lms re°ect a hue slightly shifted far to white as compared to aluminum re°ectivity and with similar color purity. In this work, it was observed the high dependence on re°ectance percentages when the deposition temperature is var-ied from 200C to 630C. Figure 7(a) exhibits one constant region for wavelengths ranged from 760 to 800 nm; in those regions it is possible to appreciate the e®ect of deposition temperature on changes in the re°ectance values. Figure 7(b) shows the decreasing of re°ectance when temperature is increased, which indicate that temperature promotes the absorbance in the AlN deposited via PLD. The last behavior changes in optical properties (re°ectance) can be re-lated with changes generated by the reduction of grain size and reduction of surface roughness when

the depositing temperature is increased producing an activation surface energy observed form AFM results (see Figs.5and6).

From the re°ectance spectra, a weak but clear e®ect of deposition temperature on the optical prop-erties is seen. As the temperature is decreased, the re°ectance of the ¯lms tends to be higher in the near IR region. So, dominant wavelength and color purity have been calculated from the re°ectance spectra of all ¯lms by using the javaoptics software (see Table2). The dominant wavelength was varied for all the samples, from 562 to 570 nm, and a little close to that of pure aluminum reference (574 nm).

In Fig. 8(a), the in°uence of deposition tempera-ture on color purity can be observed. This graph shows the increase of purity values toward color gray purity. Figure 8(b) shows the di®erences in the dominant wavelength for all AlN ¯lms deposited with di®erent deposition temperatures.

The wavelength is an important optical charac-teristic for di®erent materials. So, when the wave-length is changed it is possible to observe that natural color is changed. This modi¯cation on optical prop-erties is associated to changes generated by the re-duction of grain size and rere-duction of surface roughness observed form AFM results (see Fig. 6),

760 765 770 775 780 785 790 795 800 20

22 24 26 28 30 32 34

200ºC 300ºC 400ºC 500ºC 600ºC 630ºC

R

e

fl

ect

an

ce (

%

)

W avelength (nm ) AlN film s

(a)

200 250 300 350 400 450 500 550 600 650 20

22 24 26 28 30 32 34 36

Wavelength 760 - 800 (nm)

R

e

fl

ect

ance (

%

)

Deposition temperature (°C) AlN films

(b)

Fig. 7. Temperature-Re°ectance dependence: (a) Re°ectances of AlN ¯lms deposited onto Si3N4(100) substrates at

dif-ferent deposition temperature from 760 to 800 nm wavelength, (b) re°ectance as function of deposition temperature (color online).

(9)

moreover, the changes in chemical composition of AlN ¯lms observed form XPS results (see Fig. 2) shows that oxygen concentration is reduced when the temperature deposition is increased, so, the low oxy-gen concentration induces changes in the energy bands generating changes at optical properties (color purity and dominant wavelength). Finally the purity color dependence in AlN ¯lms with temperature obtained in this work demonstrates the possibility of some purity color control.

3.4. SEM analysis

SEM was carried out to quantitatively study the surface morphology of our samples in relation to an increase in deposition temperature in AlN ¯lms grown onto Si3N4 (100). Figure 9 shows SEM micrographs for AlN ¯lms with random distribution of microparticles or microdrops that were analyzed

on these surfaces in agreement with AFM results (see Fig. 5). Accordingly, the deposition temperature clearly a®ects the reduction of microdrops; this can be possible due to great surface mobility when the temperature was varied from 200C to 630C. This surface mobility reduces the possibility that the microdrops are anchored on the surface when they arrive with high energy on AlN ¯lm. Other possible reason can be associated with high super¯cial tem-perature that produces surface di®usion of micro-particles or microdrops, which can decrease the overall number of particles. A typical image is given in Fig. 9(a). The magni¯cation image showed in

Fig. 10(b) exhibits that Mo lines are continuous,

without any interruption or short circuit. After electrical measurements, it was possible to deduce an insulator behavior for our AlN structures, a good prerequisite for using them as SAW (Mo/AlN/SAW) devices.

200 250 300 350 400 450 500 550 600 650 0,625

0,630 0,635 0,640 0,645 0,650 0,655 0,660 0,665

Pure alum inium (0.90)

C

o

lo

r p

u

ri

ty

Deposition tem perature (°C)

AlN film s

(a)

200 250 300 350 400 450 500 550 600 650 561

562 563 564 565 566 567 568 569 570 571

Pure aluminum 574 (nm)

D

o

m

in

a

nt

w

avel

engt

h (

n

m

)

Deposition temperature (°C)

AlN film s

(b)

Fig. 8. Temperature dependence at color results for AlN ¯lms: (a) color purity and (b) dominant wavelength as function of deposition temperature.

Table 2. Optical properties for AlN ¯lms deposited onto Si3N4substrate.

Deposition temperature (C) Dominant wavelength (nm) Color purity Axis X Axis Y

200 562 0.63 0.234 0.121

300 567 0.63 0.242 0.123

400 569 0.63 0.240 0.132

500 569 0.64 0.246 0.137

600 569 0.65 0.262 0.158

630 570 0.66 0.302 0.270

Reference Aluminum 574 0.90 0.358 0.371

(10)

3.5. Frequency response of Mo/AlN

SAW devices

Hadjoub et al.19 _{reported an improvement in the}

Rayleigh Velocity (VR) by decreasing the thickness of the ¯lms and substrates using Si3N4with a velocity of 5694 m/s value very next to that of VR ðSiCÞ ¼ 6810 m/s. These values are referred to substrates for fast VRðAl₂O₃Þ ¼5680 m/s, VRðSiÞ ¼4710 m/s, VR ðSiO2Þ ¼3410 m/s, VR ðTiÞ ¼2960 m/s and VR ðWÞ ¼2670 m/s, reported in the literature. It clearly shows the importance of using substrates with good

speed of Rayleigh and compatible with existing silicon technologies so that in our case we used Si₃N₄because VR is 20% higher than that of Si.

Recently, some reports determined the relation between theh=andVSAW by experiment or theory,

wherehis AlN thickness,is wavelength andV_SAWis SAW velocity. The relationship between the thick-ness and the wavelength is h=¼0:004 the value according to the literature.1820 _{The frequency}

re-sponse characteristics of Mo/AlN SAW devices fab-ricated using the di®erent deposition conditions, such

(a) (b)

Fig. 9. SEM micrographs where is observed the microparticles or microdrops: (a) AlN ¯lms deposited with 200C (b) AlN ¯lms deposited with 630C.

(a) (b)

Fig. 10. SEM micrographs showing the interdigital structure of Mo coating on AlN ¯lms obtained by PLD: (a) general interdigital structure (b) magni¯cation images where it is possible to observe the width and distribution of lines that conform the interdigital device.

(11)

as, 200C, 400C and 630C, have been measured and results are shown in Figs.11(a)11(d). By using the AlN bu®er layers, as can be seen by comparing the wide-frequency scan spectra of those three devices [which are plotted in the inserts of Fig. 11(c)], the center frequency was shifted from approximately 0.618 GHz [see Fig. 11(a)] to higher frequencies of 0.621 GHz [see Fig.11(b)] and 0.625 GHz [see Fig.11 (c)]. This indicates that the AlN bu®er layer could be used as high phase-velocity substrate in the fabrica-tion of high-frequency SAW devices, as with Mo.

In this research using di®erent deposition tem-peratures of AlN bu®er the insertion loss was reduced. When using the AlN bu®er layers, the center fre-quency was shifted from approximately 0.618 [see

Fig. 11(a)] to 0.621 GHz [see Fig. 11(b)] and

0.625 GHz [see Fig. 11(c)]. This indicates that the AlN bu®er layer could be used as a high phase velocity substrate in the fabrication of high-frequency SAW devices. Additionally, Fig. 11(c) shows that using AlN bu®er reduced the insertion loss. To ex-amine in more detail the dependence of insertion loss

0,50 0,55 0,60 0,65 0,70

-40 -35 -30 -25 -20 -15 -10 -5

Insertion Loss (dB)

Frequency (GHz) 200 °C

-19.04 (dB)

(a)

0,50 0,55 0,60 0,65 0,70

-40 -35 -30 -25 -20 -15 -10 -5

Frequency (GHz) 400 °C

-12.07 (dB)

(b)

0,50 0,55 0,60 0,65 0,70

-40 -35 -30 -25 -20 -15 -10 -5

Frequency (GHz)

630 °C -5.16 (dB)

(c) (d)

Fig. 11. Frequency response characteristics of fabricated Mo/AlN SAW devices with di®erent deposition temperatures (a) 200C (b) 400C (c) 630C and (d) the insertion loss by the use di®erent deposition temperatures of AlN bu®er layer.

(12)

on the deposition condition of AlN bu®er, the inser-tion loss was measured for all fabricated SAW devices as was shown for the frequency response superposi-tion in Fig.11(d).

The results are plotted in Fig.12as a function of temperature deposition of AlN ¯lms and the esti-mated quantitative values are also listed in Fig. 11. As can be seen in Fig.12, for SAW devices fabricated using the sputtered-AlN bu®er with Mo channel the insertion loss was exactly correlated with the corre-sponding temperature value for all AlN ¯lms.

This can be identi¯ed by comparing the results obtained from the device with AlN ¯lms deposited on silicon nitride substrate. For the aforementioned devices, the temperature deposition was increasing systematically, as already shown in Fig. 12. More-over, the device with AlN deposited to 630C revealed the relatively large insertion loss in relation to frequency value (0.625 GHz). This may be attrib-uted to the relatively low surface roughness (90 nm) AlN ¯lm used in Mo/AlN device considering that the surface roughness of AlN ¯lm may also be a®ected by the surface morphology and deposition temperature of AlN bu®er layer,21_{the results shown in Fig.}_11(d)

con¯rm that frequency response characteristics (in-sertion loss) can be a®ected by reduction of oxygen concentration, reduction of grain size, roughness and decreasing with re°ectance of AlN ¯lms together with increase of color purity and dominant wavelength. In this sense, the property of AlN bu®er may strongly

govern the characteristic of SAW device using the Mo/AlN system.

Benedic et al.22 _{reported the direct relationship}

between increased frequency and speed in SAW devices. All this is related to the morphological qualities of roughness and grain size favor in in-creased rate of velocity. Therefore, the resonance frequency of a SAW is determined by the equation f¼V=k, where f is the resonance frequency, V is the phase velocity of the acoustic wave andKis the acoustic wavelength, which in turn is de¯ned by the electrode pitch in the IDT. The re°ected signal of the Rayleigh mode in AlN ¯lms deposited to highest temperatures (630C) is observed to be much stronger than that of the AlN ¯lms deposited to lowest temperatures (200C) and the phase velocity of AlN ¯lms deposited to (630C) is much higher. For The Mo/AlN SAW devices, the measured reso-nant frequencies of the Rayleigh mode waves ranges from 0.618 to 0.625 GHz corresponding to phase velocities of 24720 to 25000 m/s, indicating good re-peatability and reproducibility.2325Thus, the lower deposition temperature of the substrate generates a perturbation of the acoustic velocity that is related to a change in frequency of the oscillator circuit. Figure 13 shows improvement of velocity with in-creased temperature of the ¯lms deposited during the growth of AlN and was used as substrates for SAW devices building Mo metal.

Taking into account the last behavior, it is possi-ble to observe that the quasi-static electric ¯eld

200 300 400 500 600 700

-22 -20 -18 -16 -14 -12 -10 -8 -6 -4

0.624 (GHz) 0.623 (GHz)

0.620 (GHz)

0.625 (GHz)

0.621 (GHz)

Inse

rt

io

n L

o

ss

(

d

B

)

Temperature (°C) Mo/AlN SAW

0.618 (GHz)

Fig. 12. The measured insertion losses of fabricated Mo/ AlN SAW devices, as a function of the temperature values from the deposited AlN ¯lms.

200 300 400 500 600 700

24700 24750 24800 24850 24900 24950 25000 25050

V

e

lo

c

it

y

(m

/s)

Temperature (°C)

0.624 (GHz) 0.623 (GHz)

0.620 (GHz)

0.625 (GHz)

0.621 (GHz)

Mo/AlN SAW

0.618 (GHz)

Fig. 13. Acoustic wave velocity as a function of deposition temperature in Mo/AlN/Si3N4system.

(13)

associated with the SAW interacts with the charge carriers of the ¯lm. The SAW velocity is, therefore, dependent on the AlN thin ¯lm conductivity changes. These velocity changes can easily be interpreted as a shift in the resonant frequency, when the SAW delay line is incorporated in an oscillator con¯guration. The e®ect of temperature is one of the easiest parameters that can be sensed because almost all physical and chemical e®ects depend on temperature.26

In certain applications, the e®ect of temperature should be minimized in order to avoid any interfer-ence with other e®ects. For acoustic wave sensors, the temperature e®ect induces dimensional dilatations and small changes in the mechanical and electrical properties. Two kinds of interaction can thus be measured: (i) shift of the acoustic wave velocity due to intrinsic changes of material properties (elasticity, density, etc) this is the case for almost all SAW devices using metal-nitrides materials with low thickness27_{and (ii) for acoustic devices for which the}

resonant frequency or the velocity is dependent on the dimensional values (Lamb wave devices, thickness shear mode, etc.), temperature can a®ect changes of both material properties and boundary conditions.28

3.6. Correlation between optical

properties and acoustic wave velocity

in Mo/AlN SAW devices

Figure 14(a) shows the relationship between optical properties (re°ectance), acoustic wave velocity and deposition temperature in AlN ¯lms. In this research, it is clearly shown that, when the deposition tem-perature decreases, reduction of grain size (see Fig.5), reduction of re°ectance [see Fig. 7(b)] and the in-crease of the color purity [see Fig. 8(a)] is associated with an increase of frequency response (see Fig. 13) due to the increase of acoustic wave velocity (see Fig. 13) produced by energy activation at higher de-position temperatures. From this correlation, it is possible to determine that the two-merit index29

associates to an acceptable re°ectance with accept-able acoustic wave velocity at the same temperature deposition, and it also associates to an acceptable roughness with acceptable acoustic wave velocity at the same temperature deposition.

Therefore, the AlN ¯lms deposited between 400 and 500Co®er the best synergy for optical and acoustic properties with good re°ectance and an acceptable

acoustic wave velocity. Moreover, AFM results in Fig.14(b)show a direct relationship between the de-crease in roughness of the AlN thin ¯lms and the in-crease of the deposition temperature.

In this research, is shown a clear correlation is shown between the roughness and the Rayleigh ve-locity. At lower roughness the highest velocity is obtained for the samples grown at a higher temper-ature of 630C. Therefore, the AlN ¯lms deposited around 450C o®er the best synergy for optical and acoustic properties with good re°ectance and an ac-ceptable acoustic wave velocity. Furthermore, the nanometric grain size and in thickness of such thin ¯lms AlN/Si3N4a low surface roughness (120 nm)

200 250 300 350 400 450 500 550 600 650 700 20 22 24 26 28 30 32 34 36 24650 24700 24750 24800 24850 24900 24950 25000 25050 Re(%)

Wavelength 760 - 800 (nm)

R e fl ect a n ce ( % )

Deposition temperature (°C) Mo/AlN/Si SAW V(m/s) V e lo c it y ( m /s ) (a)

200 250 300 350 400 450 500 550 600 650 700 70 80 90 100 110 120 130 140 Velocity Roughness

Deposition temperature (°C)

R oughness ( n m ) 24700 24750 24800 24850 24900 24950 25000 25050 25100 V e lo c ity (m /s ) (b)

Fig. 14. Correlation between optical, morphological and acoustic properties with the deposition temperature: (a) correlation between optical properties (Re°ectance) and acoustic wave velocity in Mo/AlN SAW devices and (b) correlation between roughness and Rayleigh velocity as a function of temperature deposit AlN thin ¯lms.

(14)

and the nucleation side is characterized by a rela-tively small grain size (70 nm), which can be de-creased with increased substrate temperature improving processes nucleation, hence enhancing the propagation of elastic waves as shown in Table3.

The morphological properties of AlN are highly dependent on their polarity; the nitrogen-¯nished ¯lms typically have a greater roughness than the ¯nished aluminum, but instead, its crystal quality is better than conventional AlN obtained by other de-position techniques. Thus, initial reservoir conditions greatly a®ect the polarity of the material obtained.30 In principle, it might be thought that the whole ma-terial has the same polarity; however, it is possible that there are areas in which the crystals have dif-ferent polarities. This may explain why the lower roughness (due to increased substrate temperature during the AlN deposition) on AlN ¯lms has favored an increase in the velocity of the waves in the SAW devices.

4. Summary

In this paper, we have investigated the role of deposition temperature in determining the optical properties of AlN ¯lms and the frequency response characteristics of Mo/AlN/Si3N4 SAW devices. The temperature deposition was varied from 200 to 630C to analyze their e®ect on the optical and morphological properties of the ¯lms. XPS con¯rmed the formation of the binary ¯lms. The study revealed that temperature deposition has a marked in°uence on the chemical and physical nature of all the AlN thin ¯lms (oxygen concentration grain size and roughness).

For piezoelectric ¯lms with smooth surface, a decrease in the RMS roughness can lead to an enhancement in its SAW velocity. Therefore, the physical e®ect is most appreciated with changes of the roughness and grain size when the substrate tem-perature is increased. It was found a decrease in the re°ectance of 10%, a increase of color purity of 4.5% and dominant wavelength around 1.5% with temperature deposition between 200 and 630C. Furthermore, in this work it was evidenced that the deposition temperature clearly a®ects the reduction of microdrops; this can be possible due to great sur-face mobility when the temperature was varied from 200 to 630C.

The SAW propagation velocity is strongly de-pendent on deposition conditions. The Rayleigh mode and can be obtained withh=ratio larger than 0.004. The re°ection signal amplitude and phase velocity of Rayleigh wave have been increased with increasing of deposition temperature. Finally, it was possible to design an interdigital structure with Mo circuit on AlN ¯lms like an insulator, which can be used as SAW devices.

Acknowledgments

This work was supported by \El patrimonio Autó n-omo Fondo Nacional de Financiamiento para la Ciencia, la Tecnología y la Innovación Francisco Jose de Caldas" under contract RC-No. 275-2011 with Center of Excellence for Novel Materials (CENM). One of the authors, H. R, wishes to express his grati-tude for the kind hospitality o®ered to him fromGrupo de optica aplicadaat Universitat de Barcelona, Cata-lunya, Spain. Moreover, the authors acknowledge the

Table 3. In°uence of the substrate temperature of AlN thin ¯lms on the morphological propertiesroughness and the velocity of sound on the SAW.

Temperature substrate (C)

Insertion loss (dB) Frequency (Ghz) Grain size (nm) Roughness (nm) SAW velocity (m/s)

200 19.04 0.618 62 130 24720

300 16.45 0.620 56 117 24788

400 12.07 0.621 50 109 24840

500 8.1 0.623 42 102 24932

600 6.09 0.624 38 93 24964

630 5.16 0.625 36 90 25000

(15)

Serveis Cientí¯co-Tecnics of the Universitat de Bar-celona for and XPS, SEM and optical analysis.

References

1. C. Men and C. Lin,Mater. Sci. Eng. B133(2006) 124. 2. C. Ristoscu, C. Ducu, G. Socol, F. Craciunoiu and I. N.

Mihailescu,Appl. Surf. Sci.248(2005) 411.

3. H. Yamashita, K. Fukui, S. Misawa and S. Yoshida,

J. Appl. Phys.50(1979) 896.

4. N. E. Christensen and I. Gorczyca,Phys. Rev. B 50

(1994) 4397.

5. J. L. P. Hughes, Y. Wang and J. E. Sipe,Phys. Rev. B

55(1997) 13630.

6. L. F. Jiang, W. Z. Shen, H. Ogawa and Q. X. Guo,

J. Appl. Phys.94(2003) 5704.

7. M. Clement, L. Vergara, J. Sangrador, E. Iborra and A. Sanz-Hervas,Ultrasonics42(2004) 403.

8. J. A. Perez, H. Riascos, J. C. Caicedo, G. Cabrera and L. Yate,J. Phys.: Conf. Ser.274 (2011) 012119. 9. L. Rosenberger, R. Baird, E. McCullen, G. Auner and

G. Shreve,Surf. Interface Anal.40(2008) 12541261. 10. C. Zhuang, J. Zhao, F. Jia, C. Guan, Z. Wu, Y. Bai

and X. Jiang,Surf. Coat Technol.204(2009) 713. 11. P. W. Wang, S. Sui, W. Wang and W. Durrer,Thin

Solid Films295(1997) 142.

12. V. Brien and P. Pigeat,J. Cryst. Growth299(2007) 189.

13. A. M. Mahmood, R. Machorro, J. Heiras, F. F. Catrillon, M. H. Farias and E. Andrade,Diam. Relat. Mater.12 (2003) 1315.

14. J. A. Perez et al.,Neogranadina Sci. Eng.20 (2010) 107.

15. P. M. Raole, S. Mukherjee and P. I. John,Diam. Relat. Mater.14(2005), 482485.

16. M. Bass, E. W. Van Stryland, (eds.) Handbook of Optics(2nd ed.) Vol. 2, (McGraw-Hill, 1994). 17. H. A. Macleod, Thin-¯lm Optical Filters, 3rd Edn.

(CRC Press, 2001), pp. 158159.

18. N. S. Zayats, V. G. Boĭko, P. A. Gentsar, O. S. Litvin, V. P. Papusha and N. V. Sopinskiĭ,Semiconductors

42-2(2008) 195.

19. I. Touati, Z. Hadjoub and A. Doghmane,Phys. Chem. News27(2006) 26.

20. S. Niyomsoan, W. Grant, D. L. Olson and B. Mishra,

Thin Solid Films415(2002) 187.

21. J.-P. Jung, J.-B. Lee, J.-S. Kim and J.-S. Park,Thin Solid Films447

{

448(2004) 605.

22. F. Benedic, M. B. Assouar, F. Mohasseb, O. Elmazria, P. Alnot and A. Gicquel, Diam. Relat. Mater. 13

(2004) 347.

23. K.-Y. Hashimoto and M. Yamaguchi, IEEE Trans. Ultrason. Ferroelect. Freq. Control48(2001) 1181. 24. J. Kondoh and S. Shiokawa,Commun. Japan II 76

(1993) 69.

25. J. Kondoh and S. Shiokawa Liquid identi¯cation usingSH-SAW sensors Transducers'95 The 8th Int. Conf. Solid-State Sensors and Actuators Eurosensors IX.Stockholm, 1995, pp. 716719.

26. M. Hoummad, A. Campitelliand and W. Wlodarski,

Smart Mater. Struct.6(1997) 647.

27. R. Bechmann, A. D. Ballato and T. J. Lukaszek,Proc. IRE50(1962) 1812.

28. M. Hoummady and D. Hauden,Sens. Actuators A44

(1994) 177.

29. J. C. Caicedo, G. Zambrano, W. Aperador, L. Escobar-Alarcon and E. Camps,Appl. Surf. Sci. 258 (2011) 312.

30. Q. S. Paduano, D. W. Weyburne, J. Jasinski and Z. Liliental-Weber,J. Cryst. Growth261(2004) 259.