Gas Suppression via Copper Interlayers in Magnetron Sputtered Al

Subscriber access provided by UNIV OF ARIZONA

Article

Gas suppression via copper interlayers in magnetron sputtered Al:Cu2O multilayers Alex Hand Kinsey, Kyle Slusarski, Steven Sosa, and Timothy P. Weihs ACS Appl. Mater. Interfaces, Just Accepted Manuscript • Publication Date (Web): 07 Jun 2017 Downloaded from http://pubs.acs.org on June 7, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Applied Materials & Interfaces is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

Gas suppression via copper interlayers in magnetron sputtered Al:Cu2O multilayers Alex H. Kinsey*, Kyle Slusarski, Steven Sosa, Timothy P. Weihs Department of Materials Science and Engineering Johns Hopkins University 3400 N Charles St. Baltimore, MD 21218 *Corresponding author. [email protected] Keywords: Nanoenergetics, reactive nanolaminates, Al-Cu2O, Al-Cu, Gas Suppression

Abstract The use of thin foil, self-propagating thermite reactions to bond components successfully depends on the ability to suppress gas generation and avoid pore formation during the exothermic production of brazes. To study the mechanisms for vapor production in diluted thermites, thin film multilayer Al-Cu-Cu2O-Cu foils are produced via magnetron sputtering, where the Cu layer thickness is systematically increased from 0 nm to 100 nm in 25 nm increments. The excess Cu layers act as diffusion barriers, limiting the transport of oxygen from the oxide to the Al fuel, as determined by slow heating DSC experiments. Furthermore, by adding excess Cu to the system, the temperature of the self-propagating thermite reactions drop below the boiling point of Cu, eliminating the metal vapor production. It is determined that Cu vapor production can be eliminated by increasing the Cu interlayer thickness above 50 nm. However, the porous nature of the final products suggests that only metal vapor production is suppressed via dilution. Gas generation via oxygen release is still capable of producing a porous reaction product.

Introduction Thermites are oxidation-reduction reactions in which a metal fuel reduces a metal oxide and produces substantial amounts of heat, often enough to melt and/or vaporize the product phases. The original thermite formulation consisted of Al and Fe2O3 powders, and its molten Fe product was used to join steel railroad lines.1 1 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 31

More recently, thermites have been studied for their use in pyrotechnics, propellants, and other energetic applications due to their high temperatures, heat output, and gas formation. Thermites are typically fabricated using powder based methods, such as mixing loose powders,2 producing nanocomposite powders via arrested reactive milling,3,4 or using electrophoretic deposition of powders to create films.5,6 A more extensive list of such powder-based methods is available in recent reviews by Dreizin7 and Zhou et al.8 Another, less common method is to sputter deposit9–20 alternating layers of a metal and an oxide to form nanolaminate films or foils, otherwise known as reactive multilayers. These layered samples offer the advantage that their local and average stoichiometries, as well as their average spacing of the two reactants, are well defined. In powder-based methods, particles can agglomerate and packing inhomogenieties can arise, thereby producing non-uniform mixtures at the local level. Thus, researchers have turned to uniformly layered thermite samples to study reaction mechanisms such as the role of the terminal oxide,14 the role of interface layers on reaction kinetics,13,15 the effect of reactant spacing (bilayer period),11,12 and the effect of sample thickness (number of bilayers).21 Most of these thermite multilayer studies have been performed on the Al/CuO system due to its high thermal output and propensity to produce gas for applications such as micro initiators17 or exploding foil initiators.18,20 In this study we investigate the less energetic Al/Cu2O system for its potential use in joining applications. While Al/Ni multilayer foils have already been developed as local heat sources for joining components,22 thermite chemistries provide a new opportunity for joining because unlike the Al/Ni formation reaction, thermites produce molten metal which can act as a braze. The Al/Ni reaction produces little, if any, molten product and simply acts as a heat source to melt solder or braze layers. One significant limitation to using thermites in joining applications is that they produce gas, which can form via the vaporization of the metals or via decomposition of the oxide and the release of oxygen.23–25 This gas formation can be detrimental to bonding due to the formation of pores and material ejection.26,27 One way to 2 ACS Paragon Plus Environment

Page 3 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


reduce gas formation in thermites is to introduce excess metal to reduce the reaction temperature and thereby limit the boiling of products.27 Here we fabricate magnetron sputtered Al/Cu2O foils with Cu interlayers in order to suppress gas formation upon reaction by lowering the reaction temperature. The thickness of the Cu interlayers is varied systematically and the properties of the resulting reactions are characterized to identify a critical thickness of Cu that eliminates the production of Cu vapor.

Materials and Methods Fabrication Multilayer thermite foils were produced by DC magnetron sputtering using three 12 in. x 5 in. targets: an aluminum 1100 alloy, a Cu-diluted Cu2O (80Cu2O-20Cu, by weight), and copper (99.99% OFE). The oxide target was diluted with copper in order to increase the conductivity of the target; it was fabricated by Plasmaterials using pressed powders. The sputter chamber was evacuated below 5 x 10-6 Torr, and subsequently backfilled with Ar (99.999%) to a pressure of 1.7 mTorr for deposition. The dilution of the oxide target allowed the use of a DC power supply, as opposed to an RF supply, as other researchers have used,10,14 for the deposition of the oxide layers. Another benefit of adding Cu to the oxide target was that it produced a baseline dilution of the thermite foil by incorporating excess inert metal. Sputtering from a diluted oxide target also eliminated the need to purge the system of oxygen between the deposition of metal and the oxide. Such purging steps are required when using reactive deposition to deposit oxide layers, as the oxygen must be removed from the system so that it does not interact with the Al deposition, as done by others.11–13,15,17,19,28 The ability to deposit all species of the multilayers in a single system without purging between layers decreases overall deposition time, providing an opportunity to increase throughput as the process is scaled. To minimize the risk of oxygen contamination within the aluminum layer, the oxide and

3 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 31

aluminum targets were placed on opposite sides of the chamber and baffles were employed to limit cross contamination.

Figure 1. (a) A schematic representation of the deposition chamber showing the location of the sputter targets and carousel rotation. Depending on the initial position of the substrates on the carousel, two types of foils can be fabricated as shown in (b), bilayer and quad layer. Two types of foils were fabricated; a bilayer foil consisting of alternating layers of Al and Cu2O-Cu, and a quad layer, with Cu layers deposited between each Al and Cu2O-Cu layer. For deposition of the Al and Cu2O-Cu layers, the substrates were held stationary in front of both sputter targets for fixed amounts of time, in an alternating fashion. For deposition of the Cu layers, the substrate was passed slowly in front of the Cu target via a rotating carousel. The carousel was water cooled to avoid heating the substrates, although direct measurement of substrate temperature was not performed. When the substrates were rotating (depositing the Cu layers), the Al and Cu2O-Cu targets were shuttered. When the Al and Cu2O-Cu layers were being deposited (during the holding segments), the Cu target was shuttered. A shield was employed between the Cu and Al targets to minimize mixing of plasmas. In order to deposit Cu layers between both the Al and the Cu2O-Cu layers, the direction of rotation was switched after each Al layer was deposited. Thus, a bilayer foil was deposited at the same time as a


Page 5 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


quad layer foil, on a substrate that never rotated in front of the Cu target. A schematic of the deposition chamber and foil geometries is provided in Figure 1. Foils were deposited onto polished aluminum 6061 substrates cleaned with acetone and ethanol to remove any residue from the surface. A 250 nm adhesion layer of Al was deposited first to ease removal of the foils from their substrates. The Cu2O-Cu was deposited next, followed by alternating layers of Al and Cu2O-Cu in the bilayer sample. The final layer in these foils was Cu2O-Cu for a total of 11 Cu2O-Cu layers and 10 Al layers, not including the adhesion layer. For quad layer foils, another Al layer was deposited after the Al adhesion layer, followed by a Cu2O-Cu layer. The final layer in these foils was Al for a total of 11 Al layers (not including the adhesion layer) and 10 Cu2O-Cu layers. In all foils, the thermite bilayer (the thickness ratio of Al to Cu2O-Cu) was 1µm, with 255 nm and 745 nm for the Al and Cu2O-Cu layer thickness, respectively. Four quad layer foils were produced with Cu layer thicknesses of 25 nm, 50 nm, 75 nm and 100 nm. In order to change the Cu layer thickness for the different quad layers, the rotation rate and/or the power to the Cu target was varied. Deposition parameters are summarized in Table 1.

Target Power (W) Rotation Atomic Ratio Adiabatic Equivalence Rate Unoxidized Temperature Cu2ORatio Cu Al (Hz) Cu:Al (°C) Cu 900 0.5 1.30 1:1.54 2471 0 (Bilayer) 300 800 300 800 900 2 1.71 1:1.38 2200 25 300 800 900 1 1.71 1:1.05 2086 50 300 800 800 0.6 1.71 1:0.84 2054 75 300 800 900 0.5 1.71 1:0.71 2054 100 Table 1. The deposition parameters along with calculated equivalence ratios, atomic ratio of Cu to Al, and adiabatic reaction temperatures. The atomic ratio is calculated for the elemental copper in the as-deposited foils that is unoxidized, as in not bound to oxygen to form Cu 2 O. Foil (nm of Cu)

It is noted that some level of intermixing is to be expected between the alternating layers. Marin et al. found that there is some 12-18 nm of intermixing between Al and Cu with Al deposited at higher power density levels of 800 W and Cu deposited at 400 W in a stationary fashion, and that the intermixing distance is independent of 5 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 31

the Cu layer thickness.15 Blobaum et al. found that a 10 nm thick intermixed layer of aluminum oxide layer formed when copper oxide layers were RF deposited on top of aluminum layers.10 Intermixing between layers is likely occurring in these foils as well, but the thickness of such intermixing will be smaller due to the lower power densities and the lack of RF sputtering in this investigation. More importantly, because the individual layer thicknesses are relatively large (255 nm for Al and 745 nm for Cu2O-Cu) the volume fraction of the intermixed regions will be less than 5%, and the intermixing is expected to have limited impact on the heats of reaction, reaction temperatures, and reaction velocities.29,30 Many authors have found that reaction velocities are higher in slightly fuel-rich formulations5,12,31,32 so we chose to deposit fuel-rich formulations as well, compared to the stoichiometric 2Al:3Cu2O thermite reaction. To assess how fuel rich, we refer to the equivalence ratio which is defined as, =

( / )

( / )

(1)

where is the mass of the fuel or oxide, respectively. The mass ratio of the Al to Cu2O-Cu remains constant in the four quad layer samples and produces an equivalence ratio of 1.71. The bilayer sample, though, has an extra Cu2O-Cu layer and fewer Al layers, which yields a lower equivalence ratio of 1.30. Using temperature dependent heat capacities from the NIST Chemistry workbook,33 adiabatic reaction temperatures were calculated for each sample as follows: '

Δ + () ∑

![#] %&

=0

(2)

and results are presented in Table 1. The calculations assume that all oxygen from the Cu2O reacts with the Al to produce Al2O3, and that the only other products are Cu and Al as shown in Equation 3. Al-Cu intermetallics and any possible reaction with the environment (such as oxygen in the air) are ignored. (2+x)Al + 3Cu2O + yCu → Al2O3 + xAl + (6+y)Cu 6 ACS Paragon Plus Environment

(3)

Page 7 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


To confirm that the foils were fabricated as desired, foil cross sections were imaged using a backscatter electron detector in a Tescan Mira SEM, and images for all five foil systems are presented in Figure 2.

Figure 2. Cross sections of PVD foils showing the Cu layers deposited between the Al (dark), and the Cu 2 O-Cu (light gray) layers in the quad layer sample. (See supplemental information for false colored version of figure).

Sample Characterization Foils were analyzed in a Perkin Elmer power compensated differential scanning calorimeter (DSC) with 5 to 7 mg of sample for each test. This required the inclusion of multiple fragments of foil and the use of copper DSC pans instead of alumina to prevent curling of the fragments during heating. To minimize sample reaction with the copper pans, a piece of alumina was placed at the bottom of the pans, below the samples. Both the sample and 7 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 31

reference pans had the same sized alumina separator. Samples were heated from 50 °C to 725 °C at 40°C/min with 40 mL/min flowing argon. Two scans were performed for each test with the second scan acting as the baseline that is subtracted from the first scan to measure the heat release from the samples. For quenching tests, foils were heated from 50 °C to 350 °C or 550 °C at 40°C/min with 40 mL/min flowing Argon using alumina pans. Crystalline phases were evaluated using a Phillips X’Pert X-ray diffractometer, operating in the symmetric θ-2θ condition, with Cu Kα radiation as the source. Diffraction scans were performed on all samples after deposition, and after quenching from 350 °C and 550 °C in the DSC. Scans were also performed on the 100 nm sample after a self-propagating reaction. Two conditions were used to characterize the reacted 100 nm sample: one in which the sample was reacted in ambient conditions, the same as the velocity and spectroscopy tests described below. In the other condition, the second sample was reacted in an inert environment. A small vacuum chamber was purged to below 1 x 10-2 Torr then backfilled with Ar (99.999%) to 105 Torr twice before reacting the sample. To characterize the properties of self-propagating reactions, 12-20 mm long samples were clamped between two glass slides at one end and ignited electrically with a power supply at the other, unclamped end. This allowed the thermite reactions to self-propagate in free-standing foils. Reaction velocities were determined by capturing the propagations with a NAC Memrecame HX-6 high speed camera recording at 8,000 frames per second. The velocities of the wavefronts were tracked in multiple locations for each video. At least three videos of separate foil samples were used to determine a given propagation velocity. Gaseous species and flame temperatures were determined by emission spectroscopy from the propagating foil. An AvaSpec-3648 spectrometer (Avantes Inc.) with a 25 µm slit and a grating with 300 lines/mm was connected via a fiber optic to a 1” collimating optic (LA 1951 ThorLabs Inc.). The collimating optic was placed approximately 30 mm from the foils being tested. The reactions were ignited electronically as done for the velocity


Page 9 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


measurements. Using a 100 µs integration time the spectrometer captured light with wavelengths from 300 to 1100 nm and with a quoted resolution of 1.1-1.3 nm for the given slit and grating. To determine the temperature of propagation, emitted radiation was characterized using a multiwavelength pyrometer technique described by Ng and Fralick.34 The technique assumes that emissivity is wavelength independent and manipulates Plank’s law of black body radiation to show: / ,-. /1 3 0 20

45 /6

7

6

= ' − 4 ln ;6

Gas Suppression via Copper Interlayers in Magnetron Sputtered Al

Recommend Documents