Energy and bond strength development during ultrasonic consolidation

Journal of Materials Processing Technology 214 (2014) 1665–1672

Contents lists available at ScienceDirect

Journal of Materials Processing Technology journal homepage: www.elsevier.com/locate/jmatprotec

Energy and bond strength development during ultrasonic consolidation Gregory S. Kelly ∗ , M. Scott Just Jr., Suresh G. Advani 1 , John W. Gillespie Jr. 2 Center for Composite Materials, University of Delaware, Newark, DE 19716, United States

a r t i c l e

i n f o

Article history: Received 2 October 2013 Received in revised form 14 February 2014 Accepted 6 March 2014 Available online 15 March 2014 Keywords: Ultrasonic consolidation Bond strength Aluminum Infrared thermography

a b s t r a c t A process model was developed that couples the effects of the three adjustable ultrasonic consolidation (UC) process parameters (amplitude, force and speed) into a single term – thermal weld energy due to frictional and volumetric heat generation. Infrared thermography was used to evaluate weld energy during UC and a relationship was established between weld energy and the peel strength of ultrasonically consolidated aluminum. An optimum processing widow was identified for bonding Al 1100-0 and Al 3003-H14 based on UC processing temperature and weld time. Bonding occurred well below the material melting temperatures, confirming that thermal softening in the bulk of the material or melting are not the bonding mechanisms and that bonding takes place in the solid-state. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Ultrasonic consolidation (UC) is a low temperature solid state bonding process which can bond multiple metal foils or a foil to a substrate under the influence of pressure and ultrasonic vibrations. Bond quality is controlled through three adjustable process parameters (Fig. 1): sonotrode clamping force (Fc ), sonotrode oscillation amplitude () at a given frequency (f), and sonotrode speed (S). The material type and geometric parameters (e.g. foil thickness, width and surface roughness) also influence the bond properties. UC bonding mechanisms should be separated into two categories as suggested by Kong et al. (2005): (i) volumetric bonding effects and (ii) surface bonding effects. Volumetric bonding effects include elastic and plastic deformation enhanced through reduced yield stress due to acoustic and thermal softening. Bakavos and Prangnell (2010) investigated the bond interface of UC welds through microscopy and proved that significant amounts of plastic deformation occur during the UC process. Yang et al. (2009) investigated the microstructure of UC bonds using a scanning electron microscope and concluded that volume effects are important in flattening out rough surfaces and bringing the base metal into intimate contact. Kelly et al. (2013) have shown that plastic

∗ Corresponding author. Tel.: +1 3028318514 E-mail address: [email protected] (G.S. Kelly). 1 George W. Laird, Professor of Mechanical Engineering. 2 Donald C. Phillips, Professor of Civil and Environmental Engineering, Department of Materials Science and Engineering, Department of Mechanical Engineering. http://dx.doi.org/10.1016/j.jmatprotec.2014.03.010 0924-0136/© 2014 Elsevier B.V. All rights reserved.

deformation in aluminum bonded with UC is significantly increased due to acoustic softening and to a lesser extent by bulk thermal softening. This increased plastic deformation facilitates the onset of bonding by bringing additional material into intimate contact. Surface bonding effects include interfacial friction and shearing that break up the oxide layers and bring more metal from both surfaces into intimate contact. Using a scanning electron microscope, Bakavos and Prangnell (2010) identified evidence of oxide “flakes” near the bond interface that were dispersed during UC processing. Fujii et al. (2011) investigated the bulk and interface microstructures of ultrasonic welds through an electron backscattered technique and found a significant microstructural change near the bond interface. Mueller et al. (2013) used a scanning electron microscope and X-ray energy dispersive spectroscopy to develop a method to estimate the interdiffusion coefficient of copper–aluminum UC bonds as a function of accelerating voltage and apparent diffusion distance. de Vries (2004) used a force sensor to measure the interface forces during UC and found that the coefficient of friction and shear force at the bond interface play an important role in bond formation. Yang et al. (2009) concluded that surface effects as well as volume effects are important to UC bonding: plastic deformation and progressive shearing facilitate bonding through the dispersal of surface contaminants and oxides and allow for intimate contact between the pure metal of each part. Zhang and Li (2008) used a thermo-mechanical FEA to investigate temperature distribution and shear stresses during UC. A thermocouple or infrared (IR) camera are used to measure UC processing temperatures. Yang et al. (2009) and Sriraman et al. (2011) have measured UC temperatures by sandwiching a

1666

G.S. Kelly et al. / Journal of Materials Processing Technology 214 (2014) 1665–1672

Fc

Fc

Sonotrode z

λ

y

S Substrate

lc

τxz

Foil

x w

Foil

Fig. 1. Diagram of ultrasonic consolidation and forces acting on the foil.

thermocouple between two layers of foil. Thermocouple measurements during UC have two disadvantages when compared to IR camera measurements: (i) the thermocouple is between the two foil layers and can interfere with bonding and (ii) thermal measurements can only be taken at one location per thermocouple. de Vries (2004) showed that an IR camera is a useful method to measure UC temperatures because the entire transient temperature distribution can be easily measured. Koellhoffer et al. (2011) used an IR camera to record UC temperatures and calculated friction at the bond interface using thermal FEA. Previous studies on UC bond quality can be broken into two categories: (i) microscopy to identify bonded area and (ii) mechanical bond strength measurements. Kong et al. (2005) used “linear weld density” (LWD) to evaluate the percent of the foil width that is bonded during UC. LWD was defined as the percentage of contact points across the width of the foil that show evidence of diffusion. Friel et al. (2010) found that linear weld density generally increased with increasing amplitude and force. Obielodan et al. (2010) used optical microscopy to identify gaps between adjacently bonded foils and optimized the foil spacing during processing to minimize these defects. Several methods have previously been used to test the strength of UC bonds. Obielodan et al. (2011) used lap-shear to test the strength of aluminum–titanium UC bonds. A drawback in using a lap-shear test to evaluate the strength of UC bonds is that tensile failure happens in the base material before failure at the bond interface unless the bond is very weak. Zhang et al. (2009) developed a “push-pin” test to measure the strength of UC bonds and also found that strength increases with increasing amplitude and deceasing speed. Kong et al. (2005) used a floating roller peel test to evaluate the strength of UC bonds. Friel et al. (2010) also used a peel test to evaluate UC bonds and relate the peel strength to the LWD. Peel testing is useful in investigating the change in bond quality throughout the length of a UC seam weld. During UC, weak bonds are created with relatively low values of UC process parameters because there is insufficient energy for bonding to occur. High UC process parameters can also result in poor bond quality due to excessive damage to the parts as bonds are formed and subsequently broken. A window of optimum UC process parameters was identified by Kong et al. (2003) for Al 6061 through peel testing. A similar study was also done by Kong et al. (2004) for Al 3003. Janaki Ram et al. (2006) investigated the effect of UC process parameters on linear weld density (LWD) and found that the highest quality bonds generally occur at high , high Fc and low S. Hopkins et al. (2012) tested the lap-shear and transverse tensile strength of UC bonds and confirmed that bond strength follows the same trends in LWD found by Janaki Ram et al. (2006). Despite the fact that trends between individual UC process parameters and bond strength have been established, there is a lack of literature on a comprehensive method to relate , Fc and S to bond strength for a given material.

Due to the complex relationship between the UC process parameters and bond quality, several previous researchers have suggested relating energy to UC bond strength. Bakavos and Prangnell (2010) investigated the relationship between electrical energy and tensile shear strength using an ultrasonic spot welder and found that strength increases as energy increases; however, this work assumes that all electrical energy is converted to mechanical vibrations at the bond interface. As discussed by de Vries (2004), electrical energy input into the UC system is the upper limit of energy available for bonding as there will be system losses and suggested that measurements taken at the bond interface would be a better indicator. A method to calculate weld energy based on IR camera thermal measurements is presented and the relationship between weld energy and bond strength is investigated in this work. 2. Model development Ultrasonic vibrations during UC are responsible for heat generation and for stick-slip motion at the bond interface. The temperature increase during UC reflects the total thermal weld energy (E) due to friction heat generation (Efr ) and volumetric heat generation (Ev ) due to plastic deformation. E = Efr + Ev

(1)

E is an indirect indication of shear stress and acoustic softening in the material. At low levels of shear stress, displacement across the foil thickness is near zero and there is maximum displacement for frictional heating. At high levels of shear stress, friction and volumetric heating occur along with increased strain energy and acoustic softening in the material. E is developed during the weld time (tw ): time that the sonotrode is in contact with a given area of the foil. tw is related lc and S according to Eq. (2). tw =

lc S

(2)

Throughout tw the mechanics of the process transition from frictional sliding motion before bonding to stick-type motion after bonding. Frictional heat generation (qfr ) is related to the shear stress ( xz ) due to sonotrode oscillations, the average speed of the foil–substrate interface during a given oscillation (sfr ) and the contact area (Ac ) between the foil and the substrate according to Eq. (3). qfr (t) = Ac xz (t)sfr (t)

(3)

 xz and sfr vary throughout tw as the process transitions from slip to stick dominated motion. Ac is equal to the product of the foil width (w) and the contact length (lc ) between the sonotrode and the foil in the y-direction (shown in Fig. 1) according to Eq. (4). Ac = wlc

(4)

G.S. Kelly et al. / Journal of Materials Processing Technology 214 (2014) 1665–1672

λ/2

Eq. (10) shows the final expression for E including the frictional and volumetric contributions.



E = 2f Fc

Displacement



tw

␮(t)slip (t) dt + Vˇ 0

1/f

-λ/2

Time Fig. 2. Sonotrode oscillation over one time period.

 xz is related to the coefficient of friction () at the foil–substrate interface, Fc and Ac according to Eq. (5). (t)Fc Ac

(5)

 can vary throughout tw as surface conditions change and subsequent bonding occurs. The amplitude of frictional sliding (slip ) during a given oscillation is related to the difference between  and portion of the amplitude that contributes to foil deformation (def ) according to Eq. (6). slip (t) =  − def (t)

(6)

The values of slip and def will change throughout tw as bonding occurs. sfr (Fig. 2) is related to slip and f according to Eq. (7). sfr (t) = 2fslip (t)

(7)

The expression for Efr as a function of the UC process parameters and  is shown in Eq. (8).





tw

Efr =

tw

qfr (t) dt = 2f Fc 0

(t)slip (t) dt

(8)

0

Ev due to plastic deformation is related to the material’s yield ˙ the material volume (V), a stress ( y ), the plastic strain rate (ε), material dependent heat conversion factor (ˇ) and tw according to Eq. (9).



Ev =



tw

qv (t) dt = Vˇ 0

tw

y : ε˙ p dt 0

tw

y : ε˙ p dt

(10)

0

In general, the solution to Eq. (10) is not straightforward; however, it can be solved when plastic deformation is negligible and  is a constant. In this case Eq. (10) simplifies to Eq. (11). E = 2fFc tw = 2fFc

xz (t) =

1667

(9)

lc S

(11)

E is proportional to , Fc , tw and inversely proportional to S; however, previous work by Koellhoffer et al. (2011) has shown that  can vary significantly depending on the specific material and the specific UC process parameters, so calculating E according to Eqs. (10) and (11) is not always straightforward. Using an infrared (IR) camera it is possible to capture the temperature change (T) during UC and to use T to calculate an average value of  through thermal FEA when plastic deformation is negligible. A more straightforward method for determining E during UC without making restrictive assumptions is to establish a relationship between E, T and tw . A 2D thermal finite element model in Abaqus 6.11-1 is used to calculate this relationship. E = E(T, tw )

(12)

Due to the high thermal conductivity of aluminum and the thin foil, thermal gradients through the foil thickness are approximately 5 degrees Celsius (Koellhoffer et al., 2011). Thicker foils with a lower thermal conductivity will have more significant thermal gradients. The small thermal gradient allows a single value of E to be used in the thermal model in order to approximate the individual contributions from Efr and Ev . E in the thermal model – due to friction heating (Koellhoffer et al., 2011) and localized plastic deformation at the bond interface (Bakavos and Prangnell, 2010) – was assumed to be distributed evenly over the entire contact area between the foil and the substrate. The model used a value of lc = 5.4 mm (Kelly et al., 2013). The entire cross section (x–z plane) of the foil and the substrate was modeled, while only a portion of the sonotrode was modeled since the sonotrode was very large in comparison to the foil and the substrate. A sufficiently large portion of the sonotrode was modeled such that there was no heat flux through the boundary ensuring that the size of the modeled sonotrode did not influence the results. Surfaces exposed to air were modeled using a convection boundary condition with a convective heat transfer coefficient

Fig. 3. Thermal FEA boundary conditions.

1668

G.S. Kelly et al. / Journal of Materials Processing Technology 214 (2014) 1665–1672

Table 1 Thermal FEA material properties.

Table 2 UC materials.

Material

k (W/mC)

 (kg/m3 )

c (J/kgC)

Component

Material

Al 1100-0 Al 3003-H14 Ti-6Al-4V

237 159 17

2700 2730 4500

898 893 528

Width (w) (mm)

Length (l) (mm)

Thickness (h) (mm)

Foil Substrate

Al 1100-0 Al 3003-H14

9.5 25.4

305 152

1.0 4.8

3. Experimental procedure Input: E, tw

UC Parameters: λ, Fc, tw

Materials: Substrate & Foil

Ultrasonic Consolidation

Thermal FEA

Output: ΔT

Measure: ΔT

Peel Test Specimen

Create Plot of E = E(ΔT, tw )

Calculate: E(ΔT, tw )

Peel Test

Evaluate: P = P(E)

Measure: Peel Strength (P)

Fig. 4. Flow chart summarizing the approach to evaluate the influence of UC process parameters on weld energy and peel strength.

of h = 5 W/m2 C and a T∞ = 20 C.The value of h was varied in the normal range of natural convection (from 2 to 15 W/m2 C) and it was found that it did not influence the temperature distribution in the foil due to the short time-scale of UC. The sonotrode-foil and substrate-foil interfaces were assumed to have perfect thermal conductance. Fig. 3 shows the thermal model’s boundary conditions. Table 1 summarizes the material properties used in the thermal FEA. The three material properties are the thermal conductance (k), density () and specific heat (c). The substrate was modeled using Al 3003-H14 properties, the sonotrode was modeled using Ti–6Al–4V properties and the foil was modeled using Al 1100-0 properties. A flow chart summarizing the modeling process is shown in Fig. 4.

The UC equipment used in this work was a seam welder custom built by AmTech. The 147 mm diameter titanium sonotrode was textured using electric discharge machining (EDM). Bond properties were controlled by adjusting the following UC process parameters: Fc ranging from 300 to 6000 N,  ranging from 7 to 44 ␮m and S ranging from 0 to 300 mm/s. The system operated at a constant frequency (f) of 20 kHz. A vise was used to hold the substrate (Fig. 1) in a fixed position during processing. The geometry of the foils and substrates used during UC are listed in Table 2. The 12 sets of UC process parameters listed in Table 3 were chosen in order to create a wide range of UC bond strengths. Within this array of UC process parameters,  was varied from 27 to 44 ␮m, Fc from 1000 to 4000 N and tw from 0.07 to 0.17 s. Three specimens were created at each set of process parameters for a total of 36 specimens by bonding the foil to the substrate in preparation for peel testing. The bonded length (y-direction) of each specimen is 80 mm. UC thermal development was monitored using an FLIR SC6100 infrared (IR) camera. Values of T reported in this work were measured at the center of the top surface of the foil. During processing, the center of the IR camera lens was placed 56 cm horizontally and 5 cm vertically away from the specimen as shown in Fig. 5. The IR camera captured data during UC at a rate of 10 Hz and at a resolution of 640 × 512 pixels. Kelly et al. (2013) have shown that an IR camera frequency in the range of 10 Hz accurately captures UC temperature due to the following: (i) temperatures are recorded at the nip-point between the sonotrode and the foil and by definition this nip-point travels at the same speed as the sonotrode (unlike a thermocouple which remains fixed at a single point) and (ii) temperatures measured at the nip-point are the result of the thermal buildup of approximately 1400–3400 oscillations that occur at each foil location during tw , so high frequency measurements are not required. High frequency measurements would be required to capture the heating that occurs during a single oscillation. Steady state temperatures were generally measured after the sonotrode had traveled approximately 30 mm along the length.

Fig. 5. Schematic of the ultrasonic seam welder and the relative position of the IR camera.

G.S. Kelly et al. / Journal of Materials Processing Technology 214 (2014) 1665–1672

1669

Table 3 Array of UC process parameters for conducting the experiments. UC process parameter set

#1

#2

#3

#4

#5

#6

#7

#8

#9

#10

#11

#12

 (␮m) Fc (kN) tw (s)

27 4.0 0.17

27 4.0 0.13

32 2.0 0.15

32 2.0 0.13

36 1.0 0.13

36 1.8 0.17

36 1.8 0.13

36 1.8 0.07

36 3.0 0.13

36 4.0 0.13

40 1.8 0.11

44 1.0 0.13

Table 4 Average steady-state temperature (Tss ) measured via IR camera and average weld energy (E) calculated according to Eq. (13) for each of the 12 UC process parameter sets. UC process parameters set

#1

#2

#3

#4

#5

#6

#7

#8

#9

#10

#11

#12

Tss (C) stdv Tss (C) E (J) stdv E (J)

60 3.7 35.9 2.2

50 10.9 26.4 5.7

122 6.4 69.6 3.6

103 18.5 53.7 9.7

170 2.7 88.9 1.4

226 30.4 136.1 18.3

175 44.6 91.6 23.4

134 3.4 51.8 1.3

218 14.0 114.1 7.3

192 28.8 100.7 15.1

237 24.1 115.7 11.8

252 9.5 132.0 5.0

Based on the work by Kong et al. (2005), the strength of UC bonds was measured through peel testing using a floating roller peel test fixture. Details on the floating roller peel test fixture can be found in ASTM D3167-03a (ASTM, 2004). An Instron 5565 with a 500 N load cell was used to conduct the peel tests at a crosshead rate of 154 mm/min. Each specimen was peel tested and the maximum peel strength along the 80 mm bond length was recorded. Only data from the steady-state temperature region was used in the analysis (the first 30 mm of peel strength data is excluded).

Temperature (C)

300 250 200 150 100 50 0 0

10

4. Results and discussion

Thermal measurements were taken with the IR camera throughout the entire weld duration of the 12 sets of UC process parameters shown in Table 3. Fig. 6 shows a typical IR camera image recorded during UC. A series of IR camera images was captured for each weld as the sonotrode traveled down the length of the weld in the y-direction and the temperature was recorded at the center of the foil’s top surface from each IR camera image. Fig. 7 shows a typical plot of T versus position along the weld length. At the beginning of the weld a thermally-transient region was seen as the force and amplitude were applied to the specimen causing temperature to increase. After approximately 30 mm there was a steady-state temperature region (Tss ). The analysis was simplified by focusing on regions of the weld in the steady-state temperature region. The average steady-state temperature for each of the 12 UC process parameter sets is shown in Table 4. T was calculated as the difference between the temperature at the center of the foil’s top surface and the ambient temperature. At the temperatures measured in this

30

40

50

60

70

Fig. 7. Temperature variation along the length of a weld:  = 36 ␮m, Fc = 3.0 kN, tw = 0.13 s.

0.65

R2 = 0.998

0.6 0.55 0.5

E/Δ

4.1. Thermal measurements of UC

20

Weld Length (mm)

0.45 0.4 0.35 0.3 0.06

0.08

0.1

0.12

0.14

0.16

0.18

tw (s) Fig. 8. Ratio of E at the bond interface to T at the center of the foil’s top surface as a function of tw calculated via thermal FEA.

work, thermal softening in the bulk of the material is present, but very small in comparison to acoustic softening that explains the measured bulk plastic deformation (Kelly et al., 2013). This does not preclude a localized region of plastic deformation at the interface such as that reported by Bakavos and Prangnell (2010). 4.2. Weld energy (E) The unknown relationship shown in Eq. (12) between E, T and tw was determined using the thermal FE model. The results of the thermal model are presented in Fig. 8 in terms of the energy required to increase the temperature per degree Celsius (E/T) as a function of tw between tw = 0.07 s and tw = 0.17 s. The relationship Table 5 Table of constants used in Eqs. (13)–(15).

Fig. 6. IR camera image of temperature distribution across the bond interface (◦ C).

Constant

c1 (W/C)

c2 (J/C)

c3 (1/m2 )

c4 (N/m)

Value

2.09

0.245

270.0

−3730

1670

G.S. Kelly et al. / Journal of Materials Processing Technology 214 (2014) 1665–1672

Fig. 9. Contour plot of E as a function of T and tw calculated according to Eq. (13).

between E/T and tw was modeled using Eq. (13) and the constants in Table 5 and the fit has an R2 value of 0.998. The maximum thermal gradient though the foil thickness was found to be 12 degrees Celsius resulting in a negligible thermal softening gradient through the foil’s thickness.

3.0E+04

P (N/m)

2.5E+04 2.0E+04 1.5E+04 1.0E+04 5.0E+03

E(T, tw ) = T (c1 tw + c2 )

(13)

0.0E+00 0

Using Eq. (13), energy in the steady-state temperature region was calculated as a function of Tss and tw . E is shown in Table 4 for the 12 UC process parameters sets. Fig. 9 shows a contour plot of E as a function of T and tw . E is greatest at high T and longer values of tw .

10

20

30

40

27μm, 4.0kN, 0.17s 32μm, 2.0kN, 0.15s 36μm, 1.0kN, 0.13s 36μm, 1.8kN, 0.13s 36μm, 3.0kN, 0.13s 40μm, 1.8kN, 0.11s

4.0E+04

Pmax (N/m)

3.5E+04

27μm, 4.0kN, 0.13s 32μm, 2.0kN, 0.13s 36μm, 1.8kN, 0.17s 36μm, 1.8kN, 0.07s 36μm, 4.0kN, 0.13s 44μm, 1.0kN, 0.13s

3.0E+04 2.5E+04 2.0E+04 1.5E+04

Pmax = 270.0 (E) – 3730 R2 = 0.79

1.0E+04 5.0E+03 0.0E+00 0

20

40

60

60

70

Fig. 10. Plot of peel strength (P) as a function of position along the length of the weld in the y-direction:  = 36 ␮m, Fc = 3.0 kN, tw = 0.13 s.

5.0E+04 4.5E+04

50

Weld Length (mm)

80

100

120

140

160

180

E (J) Fig. 11. Plot of maximum peel strength (Pmax ) v. weld energy (E) for each of the 12 UC process parameter sets.

G.S. Kelly et al. / Journal of Materials Processing Technology 214 (2014) 1665–1672

1671

2.8E+04

140

Peel strength testing was conducted on the three specimens from each of the 12 UC process parameter sets. Peel strength data is shown for only 32 of the 36 specimens that failed at the bond interface between the foil and the substrate. The remaining four specimens failed in tension. Fig. 10 shows a typical plot of the variation in peel strength along the length of the weld in the y-direction. The peel strength of each specimen was evaluated in terms of the maximum peel strength (Pmax ) within the region of steady-state temperature during UC processing (Fig. 7). The final foil widths after UC processing were used in the calculation of Pmax . A plot of Pmax versus E – calculated according to Eq. (13) – for all 32 specimens is shown in Fig. 11. The relationship between E and Pmax was approximated with a linear fit having an R2 equal to 0.79 according to Eq. (14) and the constant values shown in Table 5.

2.4E+04

120

2.0E+04

100

1.6E+04

80

1.2E+04

60

Pmax = c3 E + c4

(14)

The relationship between Pmax , T and tw was found by combining Eqs. (13) and (14) and is shown in Eq. (15). Pmax (T, tw ) = c3 T (c1 tw + c2 ) + c4

(15)

Fig. 12 shows a contour plot of Pmax as a function of T and tw calculated according to Eq. (15). Fig. 12 shows that Pmax increases with increasing T and increasing tw . While both T and tw are both important, T is the primary indicator of bond strength. Fig. 12 shows that bonding occurs at temperatures well below the melting temperature of aluminum. This indicates that significant thermal softening in the bulk of the material or melting is not responsible for bonding during UC. Temperature rise due to frictional and volumetric heat generation is an indirect measure of the strain energy in the material that is responsible for bringing the material into intimate contact through acoustic material softening, dispersion of oxide layers and ultimately solid-state bonding. Fig. 13 compares E – calculated according to Eq. (13) and the temperatures shown in Fig. 7 – to the peel strength shown in Fig. 10

P (N/m)

4.3. Bond strength

8.0E+03

E (J)

Fig. 12. Contour plot of Pmax as a function of T and tw calculated according to Eq. (15).

40

Peel Strength 4.0E+03

20

Weld Energy

0.0E+00

0 0

10

20

30

40

50

60

70

Weld Length (mm) Fig. 13. Plot of peel strength (P) and weld energy (E) as a function of position along the length of the weld in the y-direction:  = 36 ␮m, Fc = 3.0 kN, tw = 0.13 s.

at each position along the length of the weld. There was a clear relationship between the calculated weld energy and peel strength in both the steady-state and transient temperature regions of the weld. While this work focused on the steady-state weld regions, future work should focus on the transient temperature region and the potential to use an IR camera for monitoring and for real-time feedback control of UC bond strength. 5. Conclusions • Peel testing was confirmed to be an effective method for evaluating UC bonds. • A model was developed that allowed weld energy at the bond interface to be calculated based on UC processing temperature and weld time. • A linear trend was identified between the weld energy at the bond interface and the peel strength of UC bonds.

1672

G.S. Kelly et al. / Journal of Materials Processing Technology 214 (2014) 1665–1672

• A correlation between weld energy and peel strength of UC bonds was found in the thermally transient and steady-state weld regions. • Low processing temperatures were measured indicating that thermal softening in the bulk of the material or melting is not responsible for bonding during UC. Acknowledgements Research was sponsored by the Army Research Laboratory and was accomplished under Cooperative Agreement Number W911NF-06-2-011. The views and conclusions contained in this document are those of the authors and should not be interpreted as representing the official policies, either expressed or implied, of the Army Research Laboratory or the U.S. Government. The U.S. Government is authorized to reproduce and distribute reprints for Government purposes notwithstanding any copyright notation heron. References ASTM Standard D3167-03a, 2004. Standard Test Method for Floating Roller Peel Resistance of Adhesives. ASTM International, West Conshohocken, PA. Bakavos, D., Prangnell, P.B., 2010. Mechanisms of joint and microstructure formation in high power ultrasonic spot welding 6111 aluminum automotive sheet. Mater. Sci. Eng. A 527, 6320–6334. de Vries, E., (Dissertation) 2004. Mechanics and Mechanisms of Ultrasonic Metal Welding. The Ohio State University, Columbus, Ohio. Friel, R.J., Johnson, K.E., Dickens, P.M., Harris, R.A., 2010. The effect of interface topography for ultrasonic consolidation of aluminum. Mater. Sci. Eng. A 527, 4474–4483. Fujii, H.T., Sriraman, M.R., Babu, S.S., 2011. Quantitative evaluation of bulk and interface microstructures in Al-3003 alloy builds made by very high power ultrasonic additive manufacturing. Metall. Mater. Trans. A 42A, 4045–4055.

Hopkins, C.D., Wolcott, P.J., Dapino, M.J., Truog, A.G., Babu, S.S., Fernandez, S.A., 2012. Optimizing ultrasonic additive manufactured Al 3003 properties with statistical modeling. J. Eng. Mater. Technol. 134, 1–10. Janaki Ram, G.D., Yang, Y., Stucker, B.E., 2006. Effect of process parameters on bond formation during ultrasonic consolidation of aluminum alloy 3003. J. Manuf. Syst. 25 (3), 221–238. Kelly, G.S., Advani, S.G., Gillespie, J.W., Bogetti, T.A., 2013. A model to characterize acoustic softening during ultrasonic consolidation. J. Mater. Process. Technol. 213, 1835–1845. Koellhoffer, S., Gillespie, J.W., Advani, S.G., Bogetti, T.A., 2011. Role of friction on the thermal development in ultrasonically consolidated aluminum foils and composites. J. Mater. Process. Technol. 211, 1864–1877. Kong, C., Soar, R., Dickens, P., 2005. A model for weld strength in ultrasonically consolidated components. J. Mech. Eng. Sci. 219, 83–91. Kong, C., Soar, R., Dickens, P., 2004. Optimum process parameters for ultrasonic consolidation of 3003 aluminum. J. Mater. Process. Technol. 146, 181–187. Kong, C., Soar, R., Dickens, P., 2003. Characterization of aluminum alloy 6061 for the ultrasonic consolidation process. Mater. Sci. Eng. A 363, 99–106. Mueller, J.E., Gillespie, J.W., Advani, S.G., 2013. Effects of interaction volume on Xray line-scans across an ultrasonically consolidated aluminum/copper interface. Scanning 35 (5), 327–335, http://dx.doi.org/10.1002/sca.21071. Obielodan, J.O., Janaki Ram, G.D., Stucker, B.E., Taggart, D.G., 2010. Minimizing defects between adjacent foils in ultrasonically consolidated parts. J. Eng. Mater. Technol. 132, 1–8. Obielodan, J.O., Stucker, B.E., Martinez, J.L., Hernandez, D.H., Ramirez, D.A., Murr, L.E., 2011. Optimization of the shear strengths of ultrasonically consolidated Ti/Al 3003 dual-materials structures. J. Mater. Process. Technol. 211 (6), 988–995. Sriraman, M., Gonser, M., Fujii, H., Babu, S., Bloss, M., 2011. Thermal transients during processing of materials by very high power ultrasonic additive manufacturing. J. Mater. Process. Technol. 211, 1650–1657. Yang, Y., Janaki Ram, G.D., Stucker, B.E., 2009. Bond formation and fiber embedment during ultrasonic consolidation. J. Mater. Process. Technol. 209 (10), 4915–4924. Zhang, C., Deceuster, A., Li, L., 2009. A method for bond strength evaluation for laminated structures with application to ultrasonic consolidation. J. Mater. Eng. Perform. 18 (8), 1124–1132. Zhang, C., Li, L., 2008. A friction-based finite element analysis of ultrasonic consolidation. Weld. J. 87, 187–194.