In the space of a mere 30 years, metaland ceramic injection mouldingtechnology, better known by theiracronyms MIM and CIM, have
come a long way, becoming the almost $1billion industry it is today and now pre-dicted to pass the $2 billion mark duringthis decade.
There are hundreds of MIM firmsworldwide, all producing, or trying toproduce more or less the same products.These are mainly parts for watches, medical applications, orthodontics appliances, computer disk drives, andparts for the automotive industry.Understandably some are a little betterat it than others.
While the globalisation of MIM willundoubtedly continue, it will inevitablylead to a plateau when supply of MIMparts overtakes demand. The ensuingbattle for market share will naturally heatup and - as some soothsayers predict -culminate in takeovers and a generalshakeout from which only the fittest willemerge.
But who will be those 'fittest'? Will wesee a repeat of the pattern followed bySilicon Valley's semiconductor industry inthe 1960s when, drawn by the lure ofcheap labour, entire manufacturing opera-tions were transplanted to S. E. Asia andother countries? In that case we can fore-see a shift in the centre of gravity of theMIM industry towards Eastern Europe,China, India or even Africa.
However, cheap labour, like a mirage,is a fleeting illusion and now increasinglyredundant with the availability of robot-ics and automation. As in evolution theo-ry, the survivors will be those who willadapt to the ever changing market condi-tions and advances in technology.
The majority of today's commercialMIM applications are what we could callimprovements over conventionally man-ufactured products. MIM has successfullydisplaced many investment cast parts suchas orthodontic appliances, parts requiringextensive or complex machining opera-tions like watch and disk drive compo-nents, or parts where the material is sim-ply too hard to machine, for exampleceramics or carbides.
MIM's full potential however has yetto be exploited. MIM can do so muchbetter than merely produce more cost-effective versions of existing products. Itis these unprecedented and, in some cases,impossible applications - by today'sstandards - that constitute MIM's nextfrontier. And the gate is wide open to theintrepid explorer.
But, like the great explorers of yore,to increase our chances of reaching thetreasures that lie hidden in strangenew worlds where no one has gonebefore, we would do well to first look atthe obstacles we will encounter on ourpath. Unlike the 12 labours of Herculesthere are only three, but they are ofstature. They are Fineness, Atmosphereand Gravity. Let's examine them one byone and see how we can control them,and perhaps even turn them into ourallies.
Fines have always been the bane of PMand in fact of metallurgy in general, per-haps even of life on Earth, causing allkinds of problems from dust-laden desertwinds to explosions, volcanic ash, atmos-pheric haze, turbidity in liquids, pollu-tion, contamination, abrasion, asthma,allergies, cancers, silicosis, etc.
Powders for PM are usually classifiedto remove the undesirable fines whichhave poor pressing characteristics. MIM'sgreatest contribution has been to givethose good-for-nothing fines the prop-erty of plasticity, thus allowing us toshape them into value-added products.
Tomorrow's technology-based manu-factured goods will have to be madesmaller. We already use our cellularphones to surf the web, buy shares on thestock market and take photographs.Before long we'll also use them to trans-mit and receive video signals, monitor ourbody functions, or to instruct the homerobot to cook fettuccine alla carbonaraand select a nice wine to go with it. Allthis amounts to squeezing more thingswith greater functionality into the limitedspace of a handheld gadget, drivingdesigners of electronic products into so-called system-on-a-chip (SOC), chip-scale-package (CSP) and multilevel pack-age design - and undoubtedly up the wallas well.
But miniaturisation can have a radicaldomino effect. Size reduction of integrat-ed systems implies that not just one butall components of the system be madesmaller, including the tools to fabricatethese. The situation is analogous to thatof the Swiss watch industry at the turnof the 18th Century when fabrication oftiny gear trains had to wait until micro-gear cutting machines were firstdesigned. Things may go fairly well until suddenly an impassable obstaclepops up.
As integrated circuit technology forgesahead towards higher integration, there isa corresponding increase in the I/O(input/output) count. That is the numberof interconnections between the siliconchip and the external circuitry, whichrequires the bond pitch - the distancebetween contiguous interconnections - todecrease proportionally. This in turnrequires that wirebonding tools, theminuscule ceramic tubelets with orificesas small as 25 micrometres (m), mustalso be made smaller. State-of-the-artwirebonding tool manufacturing technol-ogy is unable to oblige. This problemconstitutes today's biggest bottleneck inUltra Large Scale Integration.
In summary, conventional machiningtechnology is unable to produce presentday's microprecision and tomorrow'snanoscale parts. Moulding these items
22 MPR September 2003 0026-0657/03 2003 Elsevier Ltd. All rights reserved.
The shape ofthe future?MIM experts R L Billiet and H T Nguyen pay tribute to the remarkable progress the industryhas made in 30 years while glancing forward tothe next frontier, its potential problems andways to circumvent them...
using MIM technology however is sensi-ble, cost-effective and, in some cases, theonly viable alternative. But attempting tofit a 10 m powder particle, a typical con-stituent in today's MIM powders, into a12 m concave mold cavity feature wouldbe like asking a blindfolded player toscore in basketball. Thus we have nochoice but to use sub-micrometre powdersor nanoparticulates as they are nowincreasingly called.
While fineness is key to achievingmicro- and nano- design features,nanoparticulates bring along their ownproblems. Producing them is difficultenough and always costly. The real and asyet unsolved problem is their handling.Imagine, we are talking about particlesthe size of those in cigarette smoke, i.e.typically 0.01-1.0 m.
As a powder's particle size goes down,its specific surface area and consequentlyits surface activity go up with a concomi-tant increase in the tendency to formstrongly bonded agglomerates resulting indifficulty to achieve a high volume load-ing (also called packing density) in thefeedstock. Nanoparticulates also displaylower sintering temperatures, faster sin-tering kinetics with associated increasedgrain growth.
A particulate material's average parti-cle diameter allows us to estimate its spe-cific surface area, usually expressed inm2/g. For a population of uniform spheri-cal particles we can use the formula
where A is the specific surface area, in m2/g is the density, in g/cm3, and d is the particle diameter, in micro-metres (m)
Equation (1) shows us that, at con-stant density, the product of the specificsurface area times the average particlediameter is a constant. Thus, if we wereto comminute a powder consisting ofuniform spherical particles of diameter30 m into 0.03 m diameter nanos-pheres, the specific surface wouldincrease thousand fold.
State-of-the-art MIM is unable to han-dle the high surface activity associatedwith nanoparticulate materials. When
making a MIM feedstock, the polymericbinder has to stick to or wet the fillerpowder in order to get a high volume load-ing. The higher the filler's surface activitythe more difficult it becomes to wet it.
A particulate material that doesn'twet will not disperse in the binder butinstead forms strong agglomerates as theaffinity between particle surfaces isstronger than between particle surfaceand binder molecules,
MIM can overcome the surface activi-ty problems associated with nanopartic-
ulate materials by the use of surfaceactive agents or surfactants. By coatingthe surface of the particulates with amolecular monolayer of a suitable sur-factant, the surface activity can be dras-tically reduced so that the polymericbinder will now wet the thus surfactant-coated particulates.
The amount of surfactant needed isonly an infinitesimal fraction of the totalmass of the binder. Also, in most cases itis unnecessary to coat the entire surfacearea of the particulates.
When as little as 25 per cent of theirsurface is coated, their wettability will
metal-powder.net September 2003 MPR 23
Figure 1: The Shuttles takeoff is assisted by more than 350 000 pounds of fine aluminiumpowder, all burned in less than two minutes. Photograph: Courtesy NASA.
already have greatly improved while coat-ing more than 50 per cent habitually doesnot carry additional benefits.
If aliens from some distant planetwere to visit us, their first questionwould probably be: How can you guyssurvive in this poisonous gas? Our scien-tists tell us things haven't always beenthat bad and that there were the goodold days, at least for the anaerobic lifeforms from which we evolved, when therewas hardly any gaseous oxygen on ourplanet. But right now there's plenty of itand it burns - sometimes slowly, some-times fast - our cars, our ships, ourbridges, our forests, the Eiffel tower, andeven ourselves.
Because of atmospheric oxygen, finemetal powders are pyrophoric, hencetheir use in fireworks and rockets. TheSpace Shuttle is put into orbit by burningabout 352,000 pounds of fine aluminiumpowder, all of it in less than two minutes.Most nanopowders are so reactive theyneed to be constantly kept under a blanket of inert gas. This makes fabrica-tion and processing of nanoparticulatematerials complicated and extremelycostly.
MIM can overcome the pyrophoricityproblem associated with nanoparticulatematerials by coating the surface of thealready surfactant-coated particles with apolymeric binder that will effectivelyshield them from contact with atmospher-ic oxygen. Consequently, a properly pre-pared feedstock can be handled andmoulded without having to place theentire operation under inert gas (with
moulding operators in scuba diving out-fits). As dewaxing and sintering are cus-tomarily performed in an oxygen-freeatmosphere, the issue of pyrophoricitybecomes immaterial once the green partshave been moulded. Clearly the use ofwater as a solvent to extract water-solublebinders from green nanostructuresbecomes questionable.
With a few rare exceptions, most of usspend our entire lives forcibly stuck to theEarth's surface. We are so complacentabout living in a gravitational field thatwe hardly ever realize that, each time westep on our bathroom scale, we measureits effect on our body. MIM part produc-ers on the other hand, are constantlyreminded of gravity.
During binder removal, whatever contri-bution the binder was making to the greenpart's tensile strength evanesces. At the out-set of sintering, interparticulate bond for-mation gradually builds up the tensilestrength again. Between these two events,the compact's tensile strength goes througha minimum that is often insufficient tocounter the gravitational pull. As a resultthe part will droop or sag. This is a majorproblem in MIM and at present there areonly partial solutions to alleviate it.
The magnitude of gravitational sagdepends on a number of factors. Factorsrelated to the filler are its particle size,shape, surface morphology and density.Large dense particles are subjected to agreater force than small lightweight onesfollowing Newton's second law of motion.Smooth spherical particles will sag morethan spiky ones which tend to mechanical-ly interlock. Factors related to the greenpart are its geometry and density.
A green part in the shape of a pyramidwill be more resistant to sagging than a part with a long cantilever feature. A highly loaded part, i.e. moulded from a feedstock with high packing density,will deform less. Factors related to theprocessing environment include the sinter-ing atmosphere, the rate of temperaturerise, the support on which the part isplaced, mechanical vibration transmittedfrom circulating fans and vacuum pumps,among others. Although the problem ofgravitational sag has not been entirelyovercome, many proprietary "tricks" existto mitigate its effects.
So now that we are at least aware ofthe obstacles, let's boldly go and see whatlies ahead in MIM.
Designer materials on demand
In the early days of MIM the onlyfine metal powders available were car-bonyl iron and nickel. Almost anticipat-ed for MIM, these powders, with particlesizes in the 3-8 m range, were essential-ly spherical, relatively inexpensive andeasy to process as they could be sinteredto near full density at temperatures ofonly 1200C (2200F) in forming gas, anon-flammable mixture of 10 per centhydrogen in argon, thus obviating theneed for sophisticated sintering equipment.
MIM's first large-scale commercialproduction of nickel-iron parts was sin-tered in cheap ceramic hobbyist kilnsplaced in a steel tank through whichforming gas was made to flow. Part buyersin need of stainless steel had to contentthemselves with nickel-iron alloys con-taining sufficient nickel to make themcorrosion resistant. As things stand, manycontemporary MIM part producers haveyet to emerge from MIM's nickel-ironage.
With the erection of the world's largestinert gas atomiser in the late 1970s byAvesta (now Carpenter Powder Products),fine spherical prealloyed 316L stainlesssteel powder became routinely available,soon followed by Pfizer's (now AmetekSpecialty Metal Products) MIM 17-4PHstainless steel, developed for a militaryapplication.
These days, although many powdersuppliers will - for a price - producealmost any alloy composition in gradessuitable for MIM, the iron and nickel car-bonyls, 316L and 17-4PH, together with
24 MPR September 2003 metal-powder.net
Figure 2. Getting smaller! Miniaturisation is an area where MIM can go, but others cant follow.
tungsten carbide powders and Alcoa'ssuperground alumina remain the main-stay of today's MIM industry. So, after 30years, we have just half a dozen powdersthat are used to produce probably wellover 90 per cent of all contemporaryMIM applications.
It is a common scenario for a partbuyer to approach a MIM firm in thehope that his machining-intensive appli-cation, say a brass watchcase, can be pro-duced more economically by MIM. Theclient's material specification calls forbrass only because this material is avail-able, cheap, and easy to machine. In a sit-uation like this the MIM part producerwill, almost invariably, suggest to pro-duce the watchcase in 316L. If, for someoutlandish reason - fear of material sub-stitution is a very common one - theclient would insist on brass, it would cer-tainly be possible to produce a MIMbrass feedstock but its cost would over-shadow that of 316L...