Shefford P. Baker!Cornell University!Department of Materials Science and Engineering!Ithaca, New York, 14853!Deposition of Thin Metal Films "(on Polymer Substrates)!MS&E 5420 Flexible Electronics, 18 September, 2008!How to make a thin layer Suppose you want to make a thin (< 1 µm) metal film on a substrate—e.g. to provide electrical connections to a device on a flexible substrate. How could you do it?!bonding!• difficult to handle nm-scale layer!• trap dirt and voids at interface!painting!• easy!• poor control of geometry!• trap dirt and voids!atom by atom!• can be clean!• good geometry control!• expensive!Atom by atom deposition processes plasma!vapor!liquid!solid!physical!chemical!MECHANISM!PHASE!sputtering!evaporation!hot dip coating!segregation!reactive!sputtering!reactive!evaporation!plasma assisted"chemical vapor deposition!chemical vapor deposition!plating, anodization!Oxidation,!case hardening!Overview Introduction!Physical Vapor Deposition!Vacuum!Evaporation!Sputtering!Chemical Vapor Deposition!Deposition of Metals onto Polymers!Physical Vapor Deposition (PVD) The basic idea: evaporate atoms at a source, condense them on a substrate!Source atoms: may be vaporized by heating or bombardment with ions and/or electrons !Substrate: may be heated, biased, rotating, translating… etc!substrate!source!Vacuum: why and what Why vacuum? Need to remove “other” atoms and molecules (here generically called “particles”) from the deposition chamber!•# ensure that only film material deposited!•# deposit at an adequate rate (! long enough)!" Have vacuum!Units of pressure!•# Pascal (Pa) = N/m2 (official SI units)!•# torr (Torr) = 1 mm Hg = 133.3 Pa (conventional unit in the US)!•# Bar = 100,000 Pa = 750 Torr (used by some) !Properties of vacuum!•# Density of particles low compared with atmosphere!•# Particles travel long distances between collisions (average distance is mean free path)!•# Particles bombard surfaces at a high rate!•# Particles travel at about the speed of sound (in dry air, approx 330 m/s)!Levels of vacuum!•# Low vacuum: to about 10-2 Torr!•# High vacuum: 10-2–10-8 Torr!•# Ultra high vacuum (UHV): less than 10-8 Torr!Vacuum: important quantities ! n = 9.67 "1018PTparticles /cm3! "=kT#P= 2.33 $10%20TPd2cmWhere n = number of particles/cm3, P = pressure in Torr, T = temperature in K, ! = mean free path in cm, " = d2 collision cross section, G = bombardment flux of particles on surfaces inside the chamber, M = molecular weight of the particles.!Using ideal gas approximation:!! " = nkT2#= 3.51$1022PMTparticles /secparticle density:!mean free path:!bombardment flux:!Important quantities N2 at 25˚C!CO at 25˚C!Vacuum: numbers and flow pressure!(Torr)!particles/cm3!! H20!(cm)!flux"ptcls/cm2/s!time to "form monolayer!(1015 ptcles/cm2)!! He!(cm)!1!10-3!10-6!10-9!1016!1013!1010!107!0.0033!3.3!3300!33 # 106!0.011!11!11000!11 # 107!1021!1018!1015!1012!1 µs!1 ms!1 s!1000 s!Purity of film depends on rate of impingement of film atoms compared with other particles!Behavior of particles depends Knudsen number, Kn $ !/l where l is dimension of vacuum chamber and ! is mean free path!•# Kn >> 1 " laminar viscous flow: particle-particle interactions dominate, particles move in coherent manner along streamlines (turbulent flow only during initial stages of pumping)!•# Kn >> 1 " molecular flow: particle-particle interactions rare, flow controlled by wall collisions, particles rest on surface briefly and are reemitted (in random direction)!Vacuum: how to get one (spend $!) Vacuum level depends on!•# Pumping speed!•# Outgassing rate!•# Leak rate!Pumping speed depends on!•# Pump speed!•# Conductance (flow rate/%#)! Note: in molecular flow regime, conductance depends on extent to which pump can “see” all internal surfaces!Vacuum pumps Vane pump: fast, only good at low vacuum, oil can backstream!Sorption pump: condensation/"adsorption (depending on M) onto cold zeolite, fast, good to high vacuum, no back-streaming, M dependence, limited capacity!Diffusion pump: works by entrainment of gas by supersonic jet of pumping fluid vapor, only good at high vacuum, oil can backstream, needs forepump (very common, nothing to do with diffusion)!Cryo pump: condensation/"adsorption (depending on M) onto stage cooled by liquid He, fast, good to UHV except for H and He, no back-streaming, capacity better than sorption pump but still limited!Vacuum pumps Turbomolecular pump: a high speed (up to 100,000 rpm) turbine that works by capturing atoms on leading edges of blades, where they are re-emitted further into the pump (blade speed must be faster than gas particle speed!), only works in molecular flow regime, good to UHV!Sputter ion pump: ionizes gas, implants it in Ti layer and buries in with more Ti. Traps essentially all molecules. The best for UHV, capacity limited, need forepump!Typical strategy for UHV is turbopumps, backed by vane pumps (vane pump gets system to molecular flow regime, turbopump brings it to UHV), with ion pumps to get the light fast species that escape the turbopumps. Every separate chamber must be pumped. All components must be baked to >100˚C regularly. Pumps are typically more than half the cost of any evaporation system, and between a quarter and half the cost of sputter deposition systems.!Sublimation pump: evaporates Ti on walls, traps by chemisorption all gasses except inert gasses and other certain molecules, good to UHV, capacity limited, need forepump!Vacuum: composition changes during pumping Inside the vacuum chamber - start pumping, see change in composition of residual gasses!Initially pumping gas from volume of chamber " fast!!At about 10-4 Torr, outgassing from chamber walls dominates (Note H20 can be eliminated by heating chamber walls to 100˚C + during pumping!)!At lower pressure, diffusion of gas through seals or walls dominates. !Vacuum system key features Base pressure!Composition of base atmosphere!Pumping speed (cycle time)!Depends on!•# Pumps!•# Seals!•# Bakeout capability!Better vacuum always means!•# More cost!•# Lower throughput!Evaporation Oldest, best established PVD method (Edison patent 1884, earliest use Al reflective coatings for telescope mirrors)!Simple: !•# Raise T at source to evaporate material!•# If vacuum good, particles travel in straight line!•# Stick to any surface that$s cool enough, incl substrate!Complicated: film nucleation and growth!Need pressure less