RATE PROCESSES

3rd exam coverage

Scope
reaction Rates application of reaction rates to nucleation crystal growth grain growth recrystallization precipitation sintering oxidation solid state reactions role of kinetics in the development of microstructure

REFERENCES
1. Chemical Reaction Engineering. O. Levenspiel. Wiley (1972)
2. An Introduction to Transport Phenomena in Materials Engineering. D.R. Gaskell. Macmillan (1992). 3. Transport and Chemical Rate Phenomena, N.J. Themelis. Gordon and Breach (1995). 4. Transport Phenomena. A. Byron Bird., W.E. Stewart, E. N. Lightfoot. Wiley (1994). 5. Physical Chemistry of High Temperature Technology. E.T. Turkdogan. Academic Press (1979). 6. Problems in Metallurgical Thermodynamics and Kinetics. G.S. Upadhyaya and R.K. Dube. Pergamon (1977). 7. Materials Science and Engineering An Introduction. William D. Callister, Jr. 5th edition (2000).

REACTION RATES

RATES OF REACTIONS
The rate of a reaction is the speed at which a reaction happens. If a reaction has a low rate, that means the molecules combine at a slower speed than a reaction with a high rate. Some reactions take hundreds, maybe even thousands of years while other can happen in less than one second. The rate of reaction depends on the type of molecules that are combining.

There is another big idea for rates of reaction called collision theory. The collision theory says that the more collisions in a system, the more likely combinations of molecules will happen. If there are a higher number of collisions in a system, more combinations of molecules will occur. The reaction will go faster, and the rate of that reaction will be higher.

Reactions happen, no matter what. Chemicals are always combining or breaking down. The reactions happen over and over but not always at the same speed.

A few things affect the overall speed of the reaction and the number of collisions that can occur.

Concentration:
If there is more of a substance in a system, there is a greater chance that molecules will collide and speed up the rate of the reaction. If there is less of something, there will be fewer collisions and the reaction will probably happen at a slower speed.

Temperature:
When you raise the temperature of a system, the molecules bounce around a lot more (because they have more energy).

When they bounce around more, they are more likely to collide. That fact means they are also more likely to combine. When you lower the temperature, the molecules are slower and collide less. That temperature drop lowers the rate of the reaction.

Pressure:
Pressure affects the rate of reaction, especially when you look at gases. When you increase the pressure, the molecules have less space in which they can move. That greater concentration of molecules increases the number of collisions. When you decrease the pressure, molecules don't hit each other as often.

The lower pressure decreases the rate of reaction.

APPLICATION OF REACTION RATES IN NUCLEATION

The Role of Nucleation and Growth in Phase Transformations

There are two distinct stages in a phase transformation. The first of these stages is referred to as nucleation and the second stage is referred to as growth. The overall rate of a phase transformation is determined by both the nucleation rate and the growth rate, since any given grain produced by a phase transformation goes through sequential nucleation and growth stages.

If two or more processes must occur simultaneously [i.e. in parallel] during a reaction, the overall rate of reaction will be determined by the slowest of these.

However, if the processes must occur one after the other [i.e. serially], then the overall rate of reaction will depend on the rate of all of the individual processes.

For example, if you are building a house, the foundation [and associated systems] must be laid first and then the walls framed. Thus, the time taken to the end of framing is the time taken to lay the foundations, plus the time taken to complete framing. However, if the house is to be veneered with brick, then bricks must be carried and mortar laid at the same time. If the bricklayer happens to be faster than the mortar carrier, then the bricklayer must wait for more mortar to be brought up.

Hence, the slowest member of the crew, in this case the mortar carrier, controls the rate at which the brick walls come up. Try waiting for a contractor to finish a house and you can observe this for yourself! There is nothing intrinsically slower or faster about serial and parallel processes, instead this depends on the nature of the processes themselves. For example, in computers, a parallel printer port is faster than a serial communications port, but is slower than serial USB or [especially] Firewire ports.

Nucleation
Consider the formation of a solid metal from its liquid. The production of a new grain begins when a solid nucleus forms in the liquid.

Do not confuse this nucleus with the atomic nucleus - in the present context a nucleus might contain up to a few hundred atoms. If a solid nucleus forms in the liquid then a new solid - liquid interface is produced.

Like any other interface, this has a certain

interfacial energy per unit area of interface (SL) and so this new interface adds energy to the system (this is represented mathematically by the sign of

SL being positive).

Since a sphere has the lowest possible surface area to volume ratio, the nucleus minimizes its total interfacial energy by adopting a spherical shape. Interfacial energies do depend on temperature, but unless the temperature change is very large,

SL can be treated as a constant.

The actual value of SL is really very small. For example around 250 mJ m-2 of solid to liquid interface could be regarded as typical of a pure metal. In comparison a single 100 W light bulb uses 100 J every second! Nonetheless, it will be seen shortly that interfacial energies are crucially important in nucleation.

Producing a new solid will change the energy of the bulk material by a certain amount (GV) per unit volume of solid formed.

If GS and GL are respectively the energy of a unit volume of solid and that of the volume of liquid that transformed to make a unit volume of solid, then GS is added and GL is removed for each unit volume of solid formed.
In other words, the value of GV is given by:

GV = GS - GL

What about the sign of GV?

Unlike SL (which is only weakly sensitive to temperature), GV depends crucially on temperature.

Heres how: At the equilibrium melting temperature (TM), by definition the liquid and solid phases must have the same energy. Thus GV is zero and there is no thermodynamic driving force for transformation (either melting or solidification).

At a temperature T > TM
The energy of the liquid is lower than that of the solid, so that GV is positive and there is a thermodynamic driving force for melting.

At temperatures T < TM
The energy of the solid is lower than that of the liquid, so that GV is negative and there is a thermodynamic driving force for solidification.

The undercooling (T) below the equilibrium transformation temperature (i.e. TM) is defined as:

T = TM T
Thus, as T increases, GV becomes increasingly negative, since the solid becomes more and more stable than the liquid the further one goes below the equilibrium transformation temperature.

Now lets put the surface and bulk energies together to determine the overall change in energy (GL-S) on production of a solid nucleus:

GL-S = SL x surface area of nucleus + GV x volume of nucleus

For a spherical nucleus of radius r, this becomes:

GL-S = 4r2 SL + (4/3) r3 GV

First consider a case (Figure 1) in which T > 0 so that GV is negative. Remember that the SL term is invariably positive for solidification and that mathematically r2 initially goes up more quickly than - (r3) goes down. Thus, the value of GL-S initially increases with increasing r and only decreases again when a critical radius r* is exceeded. Any nucleus that forms with a radius r less than r* is unstable and can reduce its energy by reverting to the liquid state.

In contrast, any nucleus with a radius r greater than r* is stable and can reduce its energy by continuing to grow. It will also be noted that there is an activation barrier of magnitude G * to assembling a nucleus of radius r*.

Figure 1: GL-S as a function of r

Figure 1. GL-S as a function of r

The values of G* and r* depend crucially on T. As has already been noted, when T = 0, the GV term = 0 and so there is nothing to offset the

SL

term and hence both G * and r* are infinite (see figure 2). Since it is clearly impossible to assemble a nucleus of infinite size, or to expend an infinite amount of energy doing so, solidification clearly can not take place at the equilibrium melting temperature TM.
In contrast, as G increases, the GV term becomes increasingly negative and so the values of both G * and r* fall.

Figure 2: The effect of T on G* and r*.

What determines the value of T at which stable nucleation actually occurs (TN)?
A vital factor is the cooling rate (dT/dt).

When the cooling rate is low, there is more time for assembly of a nucleus of a given size than when the cooling rate is high.

Consequently:
As dT/dt increases, the typical value of r decreases. To allow stable nucleation r > r* and so, as dT/dt increases, nucleation will only be able to occur if r* is decreased. A decrease in r* requires an increase in T.

As dT/dt increases, TN increases.

Thus, when the cooling rate is very high, nucleation will take place at extremely high undercooling. Of course, solidification usually involves the formation of many nuclei, not just one and so the rate of formation of nuclei is also important, as follows:

As soon as T exceeds TN stable nuclei will be formed. The greater the value of TN the larger the driving force for nucleation (since Gv becomes increasingly negative as T increases). The greater the driving force for nucleation, the greater the rate (dN/dt, if the number of nuclei is N) at which these will form. The higher dT/dt is the greater the value of TN and hence the higher dN/dt. Thus, when a very high cooling rate is used, many nuclei are formed.

Homogeneous versus Heterogeneous Nucleation

In the example considered in the last section, solid nuclei were formed at random locations in the liquid. This is called homogeneous nucleation.
In contrast, what would happen if nucleation occurred preferentially at the solid mold walls of a casting instead of homogeneously. Nucleation at preferred sites is referred to as heterogeneous nucleation.

What is gained by nucleating heterogeneously, instead of homogeneously?

Homogeneous nucleation: A new solid - liquid interface is created and this is not offset by the removal of any existing interfacial area. The result is a large increase in interfacial energy.

Heterogeneous nucleation: By nucleating, for example, on the walls of the mold, an area of interface between the mold and the liquid is removed. This is then replaced by a new solid to liquid interface and a new solid to mold wall interface. Removing some of the existing mold to liquid interface helps to offset the interfacial energy added by the new interfaces. The result can be a much smaller increase in interfacial energy than for homogeneous nucleation.

How can the effectiveness of heterogeneous nucleation be determined?

The best place to start is to consider the Young equation for the balance of interfacial tensions for the geometry shown in figure 3.

Figure 3: Derivation of the Young equation.

Figure 3: Derivation of the Young equation.

By resolving forces in the horizontal direction it can be seen that:

ML = SM + SL cos
Where:

is known as the contact angle and the

terms represent the interfacial tensions
of the mold to liquid (ML), solid to mold (SM) and solid to liquid (SL) interfaces.

Note: Interfacial tensions have units of N m-1, whereas the units of interfacial energy are J m-2. So long as consistent units are used [e.g. do not mix SI and non-SI units], the interfacial tension and interfacial energy are numerically equal and the same symbol g is used for both.

However, within a single calculation, interfacial tensions and energies should not be mixed. For example, one can describe the contents of a bottle of soda as still 20 percent full [the optimists view!] or 80 percent empty [the pessimists view!], but mixing these up when trying to work out how many drinks are left will give nonsense.

The lower the value of the contact angle, the more effectively the solid nucleus wets the mold and the more effective a heterogeneous nucleation site the mold will be. Remember that cos (0)= 1 and cos (90) = 0 so for

best wetting (i.e. q tending to zero), cos be as high as possible. In other words, the higher the value of

should

lower the value of SM , the more effective the mold will be as a heterogeneous nucleation site.

ML and the

Does using the Young equation to explain heterogeneous nucleation make sense? Yes, because to be effective as a heterogeneous nucleation site: the amount of interfacial energy added by the presence of the nucleus should be as small as possible; the amount of interfacial energy removed by the presence of the nucleus should be as large as possible; so that, on balance, the least possible amount of total interfacial energy is added.

Thus, the most potent heterogeneous nucleation sites for new solid are those that have a high interfacial energy with the liquid and a low interfacial energy with the solid nucleus.

Looking at the Young equation, these same conditions will result in efficient wetting of the heterogeneous nucleation site by the new solid. (The reason that the terms wet and wetting are used in the context of a solid nucleus forming on a solid surface is that contact angles were first considered for liquid droplets wetting solid surfaces and nucleation theory is much more recent.)

The effect of heterogeneous nucleation on the activation barrier to nucleation is represented by a function called the shape factor, which has the symbol S(). The exact derivation of S() depends on the shape of the nucleus, but as q decreases so does S().
When is low, S() <<1. The shape factor does not change the value of r*.

However:

G*HET = G*HOM S()

where HET and HOM refer to heterogeneous and homogeneous nucleation. The availability of potent heterogeneous nucleation sites, i.e. sites that result in a low S(), greatly reduces the size of the activation barrier to nucleation. Thus, the value of TN for heterogeneous nucleation is often much smaller than for homogeneous nucleation.

Hence,
in most cases, nucleation occurs heterogeneously, rather than homogeneously. The example given here has focused on solidification. However, the same basic theory is also applicable to solid-state phase transformations.

In the case of solid-state phase transformations, nucleation usually occurs heterogeneously, often on grain boundaries (although other heterogeneous nucleation sites are also important). Homogeneous nucleation is only observed in cases where the interfacial energy between a second phase and the matrix in which this precipitates is very small. For this to happen, the precipitate and matrix need to have similar crystal structures, lattice parameters and suitable chemistry. In cases where the interfacial energy between the precipitate and matrix is very small, homogeneous nucleation is possible because G*HOM is low, even at relatively small T.

Growth
Once stable nucleation has been achieved, the growth stage of the phase transformation begins.
As the name of this stage implies, during growth the stable nucleus proceeds to grow into a grain.

For growth to occur, two things must happen: The interface between the product and parent phases (the phase boundary) must be mobile (i.e. able to move), otherwise the new phase cant grow.

 If, according to the phase diagram, there is a difference in composition between the two phases, then it will be necessary to transfer solute from the parent to the product phase (or vice-versa, depending on the compositions of the two phases). For example, if the carbon-rich phase Fe3C is to be transferred from the -Fe to the Fe3C.

precipitate in a matrix of -Fe, then carbon must

The mechanism by which the composition of the parent and product phases is adjusted involves the diffusion of solute.

Since motion of the phase boundary and diffusion must occur at the same time, whichever of these two processes is slowest will control the overall rate of growth. Cases where mobility of the interface is the rate limiting step are referred to as interface controlled, whilst those in which diffusion is the rate limiting step are known as diffusion controlled. Most, but not all, phase transformations are diffusion controlled.

As an analogy, think about what controls the rate of someone walking across a floor. Normally, the rate controlling step is the rate at which one can walk - think of this as being like diffusion control. If, however, something very sticky is split on the floor, then the rate controlling step becomes detaching ones feet from the floor - think of this as being like interface control.

The Nature of Diffusion

Diffusion involves the random-walk of atoms. In this type of motion (which is approximated, but not usually achieved, by persons whom have consumed an excess of certain ethanol-bearing beverages) each atom moves by a series of random jumps, independently of other atoms. The effect of this random walk is that any gradient in concentration is gradually removed.

In other words, diffusion tends to homogenize. To verify this yourself, try experimenting with putting drops of food coloring into water and watching what happens to the color over time (also try changing the temperature of the water and observe the effect this has on the rate of homogenization).

How is homogenization possible if diffusion is a random process?

Consider a case when NL atoms are present on the left hand side of a specimen and NR atoms are present on the right hand side.

In a random walk process, the probability of any given atom jumping from left to right (PLR) is the same as that from right to left (PRL).

The behavior of any individual atom can not be predicted, but the overall number of jumps from left to right (NLR) and right to left (NRL) in time t can be predicted since: NLR = NL PLR NRL = NR PRL The net motion of atoms from left to right (NNET) would then be: NNET = NLR - NRL = NL PLR - NR PRL = P (NL - NR) where P = PLR = PRL.

Thus, so long as there is a gradient in concentration, the net flow of atoms will be from the concentrated to the less concentrated side (i.e. diffusion tends to homogenize).
In other words, diffusion occurs down a concentration gradient.

How can this be represented mathematically?

Ficks first law of diffusion states: J = -D dc/dx

where:
J = flux (i.e. the number of atoms passing through 1 m2 of a reference plane per second), c = concentration and

x = distance (so that dc/dx is the concentration gradient).

The D term is known as the diffusion coefficient or diffusivity and has units of m2 s-1.

The flux is a measure of how fast material is actually transported by diffusion. The (dc/dx) term and the minus sign reflect diffusion taking place down a concentration gradient. The diffusion coefficient is a measure of how rapidly a species is able to diffuse. Consider an analogy. Feeling fed up with the crowds in the city and wanting to get away from people provides the motivation to visit the countryside where there are fewer people (the concentration gradient).

Having a car provides a faster means of getting there than walking (the diffusivity).

Thus, if a lot of fed-up city people have cars then there will be many people moving through the city limits in a given period of time (the flux).

What controls the value of D?

The answer to this is two things, firstly the solute and solvent involved (i.e. D is material specific) and secondly D depends on temperature (i.e. diffusion is thermally activated). Mathematically, the effect of temperature and material on diffusion is represented by the Arrhenius relationship:

D = D0 exp [-Q / (RT)]

where:
D0 is a material-specific constant,

R is the gas constant and T the temperature.

Q represents an activation energy. t is not surprising to realize that increasing the temperature makes diffusion faster, since increasing the temperature would be expected to increase the frequency with which atoms jump. However, to really understand diffusion, it is necessary to understand the role that the material plays in this process.

Solid-State Diffusion In a liquid or gas, the atoms (or molecules, or ions as appropriate) are clearly free to move and this allows diffusion to occur.

In the solid-state, however, it is less clear that diffusion can occur and a mechanism is needed that allows atoms to jump.

At any temperature above absolute zero (0 K, i.e. - 273 C) some of the atomic sites in any crystal will be vacant at equilibrium and the number of vacant sites (vacancies) increases with temperature by an Arhennius relationship. This raises the possibility of diffusion occurring by atoms swapping places with vacant sites (see figure 4). In this way, both the solvent atoms (self diffusion) and substitutional solutes can diffuse.

Figure 4: Vacancy - atom interchange

The rate at which diffusion can occur will then depend on the availability and mobility of vacancies.
Hence the activation energy Q term in the Arhennius relationship actually consists of:

where

Q = QF + QM

QF and QM respectively represent the activation energies for the creation and migration of vacancies.

Q is a material (i.e. solute and solvent) specific constant with units of kJ mol-1.

Both the formation and migration of vacancies become easier as the temperature is increased because: Vacancies are created by atoms being shaken loose at interfaces such as grain boundaries, leaving a vacant site (see figure 5). The shaking comes from thermally-induced vibration of the atoms making up the lattice of the crystal. Hence, on-heating, the number of vacancies increases. Conversely, on-cooling, atoms present at boundaries tend to drop back into the vacant sites and the number of vacancies decreases.

Figure 5: Creation of vacancies.

 When a vacancy and an atom swap places, the atom is temporarily out of its equilibrium position and has to squeeze past neighboring atoms (figure 6). This becomes easier as the temperature increases, because thermally-induced lattice vibration dilates the lattice (i.e. lattice parameters increase with temperature).

Figure 6 : Activation barrier to vacancy atom interchange.

Consequently, whereas a given metal (say copper) might have only one vacancy atom interchange expected in billions of years at room temperature, millions of these interchanges will take place per second just below the metals melting temperature (1,083 C for copper).
Hence, there is a very strong positive effect of the temperature (T) term on the value of the diffusivity D.

As has already been noted, the activation energy (Q) is a material specific constant and the value of this reflects the relative ease with which vacancies can be created and move. The more strongly bound the material, the harder will be both of these two processes.

Hence, Q tends to increase with the melting temperature of the material, since a high melting temperature is indicative of strong bonding.

Thus, at a given temperature, tungsten will have a lower self diffusivity than say chromium (even though these are both BCC metals). Interestingly enough, at a temperature just below their respective melting point, different metals with the same crystal structure has about the same diffusivity.

Thus far, we have considered self diffusion and the diffusion of substitutional solutes, but not the diffusion of interstitials.
In the case of interstitial solutes, most interstitial sites are empty and so vacancies do not need to be created. Thus, the activation energy for interstitial diffusion is governed by the strain imposed on the lattice of the solvent as the solute atoms jump from interstitial site to interstitial site.

If the interstitial is very small, say atomic hydrogen (i.e. H and not molecular hydrogen H2) in iron then this strain will be quite small and the activation energy for diffusion will be very low.

Even with larger interstitial solutes, like carbon in iron, the activation energy for interstitial diffusion will be much lower than for self diffusion. Consequently, at any given temperature, the diffusivity of interstitial species will be much higher than for substitutional solutes, or self diffusion.

The Effect of Nucleation and Growth on Microstructural Development

The nature of the nucleation and growth process has a profound effect on microstructural development. For now, two key effects of nucleation and growth on microstructural development will be discussed:

When are diffusional phase transformations fastest?

In other words, how much undercooling will give the fastest rate of transformation?

As TN increases, nucleation occurs faster

However, an increase in TN implies that the actual temperature is lower and hence diffusion controlled growth becomes slower. Balancing these two factors results in fastest transformation at intermediate values of TN.

How can the grain size be controlled? Increasing the cooling rate will increase TN. This will make nucleation of grains faster and growth of grains slower. Consequently as dT/dt increases, the mean grain size decreases.

In summary: Phase transformations in ceramic materials are fairly similar to those in metals. The behavior of polymers is quite different to that of metals, but some of the basic concepts discussed here are still applicable.