Using combinatorial and high-throughput materials science methods, we have studied thin-film libraries of Sn 1−x Co x ͑0 Ͻ x Ͻ 0.6͒ and ͓Sn 0.55 Co 0.45 ͔ 1−y C y ͑0 Ͻ y Ͻ 0.5͒ alloy negative electrode materials for Li-ion batteries. Over one hundred compositions have been studied carefully by X-ray diffraction and electrochemical methods. The Sn 1−x Co x system is found to be amorphous for 0.28 Ͻ x Ͻ 0.43. For 0.43 Ͻ x Ͻ 0.6, the amorphous phase coexists with electrochemically inactive crystalline Co 3 Sn 2 . Amorphous materials with x = 0.4 show a specific capacity of 650 mAh/g, but differential capacity, dQ/dV, vs potential is not stable vs cycling indicating irreversible atomic-scale changes in the alloy, most likely due to tin aggregation. Adding carbon to this system, for example in the ͓Sn 0.55 Co 0.45 ͔ 1−y C y ͑0 Ͻ y Ͻ 0.5͒ library, has a number of positive effects. First, all alloys with 0.05 Ͻ y Ͻ 0.5 are amorphous, with carbon directly incorporated within the amorphous phase. Second, the addition of carbon increases, not decreases, the specific capacity from about 670 ± 15 mAh/g for y = 0.05 to 700 ± 15 mAh/g for y = 0.4. Third, compositions with y Х 0.4 show differential capacity vs potential curves that do not change during charge-discharge cycling, indicating that such alloys are stable on the atomic scale and hence are extremely good candidates for long cycle life. Stability increases with carbon content up to y = 0.4.

Magnetron cosputter deposited ternary libraries of Sn 1−x−y M x C y ͑M = Ti, V and Co͒ ͑0 Ͻ x Ͻ 0.5 and 0 Ͻ y Ͻ 0.5͒ have been studied structurally and electrochemically using combinatorial and high-throughput methods. Each of the sputtered Sn 1−x M x binary systems shows an amorphous composition range where the specific capacity for lithium decreases with M content. Adding carbon to the amorphous binaries, to make ternaries, causes the precipitation of crystalline Sn in the cases when M = Ti or V, but not when M = Co. We believe this is because stable carbides of Ti and V exist but stable Co carbides do not. The sputtered Sn-Co-C system was found to have a large amorphous range and the initial amorphous atomic arrangement in certain compositions are stable over at least 27 charge-discharge cycles of 600 mAh/g. Crystalline Sn was found to precipitate in composition ranges having competitive specific capacity in the Sn-Ti-C and Sn-V-C libraries causing rearrangement of the atoms during cycling leading to poor capacity retention.

A survey of the structural and electrochemical properties of combinatorially sputter deposited Sn-transition metal alloys ͓Sn 1−x M x ͑0 Ͻ x Ͻ 0.7; M = Ti, V, Cr, Mn, Fe, Co, Ni, Cu͔͒ is reported. Over 512 compositions have been studied. Sputtered libraries of Sn 1−x M x with M = Mn, Fe, Ni, and Cu show no evidence of nanocrystalline or amorphous phases at any composition. By contrast, libraries of Sn 1−x M x with M = Ti, V, Cr, and Co show composition ranges where the films are highly nanostructured or amorphous, suggesting that these elemental combinations are better glass formers. The transition metal contents of the amorphous or nanostructured phase regions are 0.37 Ͻ x Ͻ 0.40 and x Ͼ 0.48 to at least x = 0.65 for M = Ti, x Ͼ 0.39 to at least x = 0.60 for M = V, 0.47 Ͻ x Ͻ 0.73 for M = Cr, and 0.28 Ͻ x Ͻ 0.43 for M = Co. Electrochemical tests using a 64 channel Li/Sn 1−x M x combinatorial electrochemical cell show that the specific capacity of the alloys drops with transition metal content, as expected. The Sn 1−x Co x system shows an amorphous phase with the largest specific capacity, primarily because the amorphous phase is reached at the lowest transition metal content for Sn 1−x Co x . Capacity retention vs cycle number is generally best for those compositions that are amorphous or highly nanostructured. Arguments are presented to suggest that amorphous Sn 1−x V x alloys are the best choice among Sn 1−x M x alloys. Comparison with literature results for samples prepared by mechanical alloying, electrodeposition, vacuum deposition, etc. is made.

Magnetostrictive Fe 100−x Ga x alloys over the composition range 0 < x < 36 have been prepared by combinatorial sputtering methods. These films were investigated by x-ray diffraction and 57 Fe Mössbauer effect spectroscopy techniques that have been adapted for the efficient study of combinatorial samples. X-ray diffraction measurements show the presence of a disordered bcc phase with a lattice parameter that increases with increasing Ga content. Mössbauer effect measurements show that even for low Ga content the disordered alloys are not truly random but show some degree of short range Ga clustering. For x > 20, the Mössbauer effect results suggest the presence of short range D0 3 type order. This persists up to the maximum Ga content studied here (x = 36), although there is no evidence to support the formation of long range D0 3 structural order in any of the films prepared in this study.

normalPt1−xnormalMx (M=Ru,Mo,Co,Ta,Au,Sn) random alloy samples, covering most of the binary composition range, have been prepared via magnetron sputtering. The alloys were deposited through shadow masks onto 3M nanostructured thin-film catalyst support for testing in a 64-electrode polymer electron membrane fuel cell (PEMFC). CO stripping voltammograms and hydrogen oxidation polarization curves with pure hydrogen and with reformate containing up to 50ppm CO were measured on all the samples. In agreement with reports in the literature, Ru, Mo, and Sn were found to improve the CO tolerance of Pt, although the intrinsic hydrogen oxidation activity of Pt decreased significantly as the Sn content increased. The addition of Co to Pt had no impact on CO tolerance, possibly because of loss of surface Co through dissolution in the fuel cell. The addition of Au to Pt led to an increase in hydrogen oxidation overpotential when CO was present. Small amounts of Ta gave a small reduction in hydrogen oxidation overpotential in the presence of CO, but the overpotentials were still too high for practical application in a reformate-fed fuel cell.

A review of recent literature on Si:C composite and nanocomposite electrode materials is first presented emphasizing that most authors do not compare the experimental specific capacity of the composite with that expected based on the phases present. We provide such a comparison and suggest that much of the apparent confusion in the literature, when taken as a whole, can be understood if nanocomposites prepared by "aggressive" methods like high energy milling and high temperature heat-treatment contain significant amounts of amorphous or nanocrystalline SiC. In order to help resolve the confusion, samples of Si 1−x C x were prepared by high-energy mechanical milling for 0.25 Ͻ x Ͻ 0.5 and by combinatorial co-sputtering for 0 Ͻ x Ͻ 0.8. X-ray diffraction shows the mechanically milled samples to be a mixture of nanocrystalline SiC and Si. Electrochemical studies of the mechanically milled samples show that the attained specific capacity can be described accurately assuming that the Si is active and can reversibly react with 3.75 Li atoms per Si atom ͑Li 15 Si 4 ͒, while the SiC is inactive. The co-sputtered samples are amorphous or extremely nanostructured for all x. For 0 Ͻ x Ͻ 0.5, the specific capacity decreases with increasing x, from about 3580 mAh/g at x = 0, to about 1000 mAh/g at x = 0.5, presumably due to the formation of inactive regions of a-SiC. The capacity of the co-sputtered samples does not reach small values at x = 0.5, unlike the ballmilled samples, because there are presumably some regions of a-Si and a-C among the inactive a-SiC regions due to the high quenching rate of the sputtering process. Commercially relevant compositions are identified.

In situ atomic force microscopy measurements of patterned amorphous Sn-Co-C sputtered films reacting with Li in an electrochemical cell have been made. Prismatic-shaped patches of Sn0.34Co0.19normalC0.47 were found to undergo reversible volume expansion of 175±5% [(Vfinal−Vinitial)∕Vinitial] without fracture. The a-Sn0.34Co0.19normalC0.47 was found to have a reversible specific capacity of 700±10mAh∕g when cycled vs a Li metal negative electrode.

A modified version of the effective heat of formation ͑EHF͒ model of Pretorius et al. is presented to predict phase formation in codeposited silicon-transition metal films. The EHF model predicts that the first Si-M ͑transition metal͒ phase to form in thin-film diffusion couples is the phase with the most negative EHF at the composition of the lowest temperature eutectic ͑i.e., at the growth interface͒. Combinatorial thin-film libraries of codeposited Si-M produced using the modified composition spread method of Dahn et al. consist of intimately mixed Si and M atoms. We propose the entire film can be considered as the growth interface and that the Si-M phase with the most negative EHF at a given composition is present. Film nanostructure as a function of composition can then be directly determined from EHF diagrams ͑assuming regions of Si are amorphous͒. This information, when combined with the assumption that all Si-M phases are inactive, can be used to predict the specific capacity of all Si-͑transition or rare-earth metal͒ systems in Li/Si 1−x M x cells as a function of composition.