Lecture on

Thermofluid Engineering and Microsystems Design

• There are many microscaled devices that involve “heat” or/and “fluid”
flows.

• Examples such as thermal force-actuated devices and valves and pumps

in micro fluidic systems.

• Thermofluid principles are used in the design of these devices for

both “performance” (i.e. functions) and “strength” (e.g. fluid-induced
forces).

• Thermofluid principles are also used in the design of

microfabrication processes such as chemical vapor deposition,
oxidation, etc.
Characteristics of Moving Fluids

● Fluids have volume but no shape.

● Compressible fluids (gases)

● Incompressible fluids (liquids)

● Fluids cannot withstand “normal” stresses, other than “hydrostatic pressures”

● Shear stress is responsible for fluid flow.

● Shear stress is directly proportional to the velocity gradient in moving fluid:
y Velocity profile, u(y)
uo The shear stress:    du( y) (5.2)
dy
 where µ = dynamic viscosity of the fluid
u(y)

LV
● Many fluid flow cases are characterized by Reynolds number: Re  (5.3)

in which ρ= mass density; V = velocity; L = characteristic length
● Laminar fluid flows occur at Re < 10-100 for compressible fluids,
and Re< 1000 for incompressible fluids.
 The Continuity Equation

It is often used to compute the volumetric flow, Q and the velocity, V of a moving
fluid through conduits with variable cross-sectional areas.

In the situation such as illustrated below:

1 2
Q = V1A1 = V2A2 m3/s (5.6)
V1 Reducer V2

To micro fluidic

1 2

Diameter
d1 = 1000 m V1 V2 d2 = 20 m
 The Momentum Equation

This equation is derived on the basis of conservation of momentum. It is used to

compute the fluid flow-induced forces on the interfacing solids.

It is used in assessing the strength of microvalves and pumps in a design process.

2

C’
1
C

A’ D’
A

V1 D

B B’

The force required to drive the fluid from 1-1 to 2-2, or the flow-induced forces to be:
(5.7)
 This equation relates the
Laminar Fluid Flow in Circular Conduits
volumetric flow, Q and the
- The Hagen-Poiseuille Equation corresponding pressure drop, ∆P.
w
r
Velocity profile: Shear stress a
profile: (r) x
Vr(r)

x2 - x1 = L
x1 x2

Q
 a 4
 d (P  gy) (5.16)
8  dx 
where y = elevation of the tube from a reference plane.
The pressure drop in the fluid over the tube length, L is:
8LQ
P  (5.17)
 a4
1 meaning a reduction in half in the radius→24=16
NOTE: The pressure drop, P 
a 4 times increase in pressure drop (pumping power)!!
The equivalent head loss in relation to Q is:
128LQ
h f ,𝑙  (5.18)
g d 4
Laminar Fluid Flow in Circular Conduits For conduits with non-circular
- The Hagen-Poiseuille Equation cross-sections.

In such cases, hydraulic diameter, dh is used in the Hagen-Poiseuille equations.

This diameter is defined as:

4A
dh 
p (5.19)
where A = cross-sectional area of fluid flow
p = wet perimeter.
w

h h1

Rectangular conduit filled Rectangular conduit filled

with fluid with fluid up to h1
4wh1
4A  4(wh)  2wh dh 
dh  w  2h1
p 2(w  h) w  h
Incompressible Fluid Flow in
Microconduits
Observation: Droplets of water on flat surfaces exhibit “spherical topography”
and such phenomenon is possible only with “small” droplets.

Reason: It is the “surface tension” of the water that produces such spherical
surface of droplets of liquids.
Surface Tension in Liquids

● It is the cohesion forces of molecules that exist in all liquids.

● When a liquid is in contact with air or a solid, the inter- molecular forces
in the liquid bind the liquid molecules beneath the contacting surface,
whereas no such force exist at the contacting surface.
● Consequently, when the liquid is in contact with air, the inter-molecular forces
of the liquid tend to bond the liquid molecules together.
● Since there is no force at the liquid/air contacting surface, the shape of the
liquid at the interface becomes spherical.
● In the case of larger sized droplets, the “weight” of the liquid droplet itself
exceed the inherited surface tension, and no droplet of spherical shape is
possible.

● Thus, surface tension is a dominant factor in “small” volume of liquids.

The surface tension of the small volume of fluids at the contacting
surface of the conduits, and the friction at the interface result in radically
different flow phenomena in microconduits .
Surface tension in small volume fluids presents obstacle to the flow,
and extra pressure is required in pumping such flows – Capillary flow.

Magnitude of Surface tension in a liquid

Surface tension Wet perimeter Coefficient of
Fs S surface tension, 

The coefficient of surface tension,  with a unit of N/m is a measure of the

magnitude of the surface tension.

The  - value for water can be obtained by the following empirical formula:

(T) = 0.07615 – 1.692 x 10-4T (5.23)

where T is the temperature in oC and γhas units of N/m

 Pressure change due to surface tension across liquid volumes
a2P
2aLP
2 a

P 
 2
(5.24a) P  (5.24b)
a a
Combining the above two cases for a volume in a microconduit:
Tube wall

3
P  Fluid volume Radius, a Tube diameter

a Radius, a d

∆P is the minimum pressure to be overcome for pumping this volume of liquid.

Numerical
Determine the pressure required overcoming
the surface tension of water in a small tube of
0.5 mm inside diameter. Assume that the
water is at 20oC
 Example

Determine the pressure required overcoming the surface tension of water in a

small tube of 0.5 mm inside diameter. Assume that the water is at 20oC.

Solution:

We first determine the surface tension coefficient of water at 20oC from Eq. (5-23)
to be  = 0.073 N/m.

The tube has a radius of a = 250 m = 250x10-6 m.

Following the expressions in Eqs. (5-24a,b), we have the pressure required to

overcome the surface tension to be:

3 3x0.073
P   6  876 N/m2 or 876 Pa
a 250x10
Overview of Heat Conduction in
Micro Structures
To assess temperature distribution (i.e. variations), T(r,t) in a micro structure
subject heat flow, in which r = position vector, t = time.

The computed temperature distribution T(r,t) is used to assess the induced

thermal stresses and strains (and displacements) in the structure such as.

Fourier Law of Heat Conduction

Total amount of heat flow through the slab, Q

during time period, t is:
Amount of
Q A(T a  T b )t
heat flow, Q
Area, A Qk (5.26)
d
where k = thermal conductivity of the solid with
d
a unit Btu/in-s-oF or W/m-oC

Thermal conductivity, k is a material property, which represents a material’s

ability to conduct heat. It is normally a constant in normal range of temperature.
Heat flux, q, which is equal to:
Q (  )
q  k Ta Tb (5.27)
At d
is a more meaningful quantity in heat transfer analysis. It represents the “intensity”
of heat flow. It has a unit of Btu/in2-s or W/m2.

The above is the Fourier law of heat conduction in simple one-dimensional

case. For a more general case, it is expressed as:
 
q r ,t   k T r ,t   (5.28)
→
In which r  position vector : (x, y, z) in Caartesian coordinate system
q(x, y, z, t)  q x  q y  q z
2 2 2
qz (5.29)
z q(r,t)
qx T ( x, y, z, t)
where qx   k x (5.30a)
qy x
T (x, y, z, t)
x qy   k y (5.30b)
y
y Position vector:
T (x, y, z, t)
r: (x,y,z) qz   k z (5.30c)
z
The Heat Conduction Equation

 2T (r↼
Q 1 T (r ,t)
,t)   (5.31)
k  t
where the Laplacian is defined as:

 2  2  2
2
in Cartesian coordinate system, and
 x2  y2  z2

 2  1   1  2   2 in cylindrical polar coordinate system

2
 r 2 r r r 2   2  z 2

In the heat conduction equation , Eq. (5.31), the term Q = Q(r,t) is the heat
generated by the solid material.
 In MEMS and microsystems, electric resistance heating is commonplace.
In such case, this amount of heat generation is equal to:

2
Power P
= Current, I χ Resistance, R
watts (W) amperes (A) Ohms (Ω)

The power in the above expression has a unit of watt, which is equivalent to
1 Joule/sec. It is also equivalent to 1 N-m/sec in the SI units.

The constant  in Eq. (5-38) is called thermal diffusivity of the material with a unit
of m2/sec. It has an important physical meaning of being a measure of how fast
heat can conduct in solids (thermal inertia). Mathematically, it is equal to:
k
 (5.32)
C
in which  and C are the respective mass density and specific heat of the solid.
The units for  is g/cm3, and the unit for C is J/g-oC.

Refer to Table 7.3 for the thermal physical properties of some common MEMS materials
 Newton’s Cooling Law For heat flow in fluids

Heat flow from Point A to Point B is expressed as

Ta q q-the heat flux (w/m2 or J/m2-s) in the expression:
A Tb
Fluid B
Ta > Tb Fluid
q = h (Ta –T b ) (5.33)

where h = heat transfer coefficient, W/m2-oC

The magnitude of h depends on the properties of the fluid, but the dominating
parameter is the velocity of the fluid in motion (forced convection).

Heat convection also occur in fluid under no influence of external force, ie

“natural (free) convection”. The h-value in forced convection is greater than
that in natural convection.
Numerical values of h are determined by the values of the Nusselt number (Nu)
from “dimensional analyses” in the following forms. The Nusselt number has an
expression of Nu = hL/k, in which L = characteristic length and k = thermal
conductivity of the fluid.
For forced convection:
Nu =  (Re)(Pr)
For Natural convection:

Nu =  (Re)(Pr) (Gr)
where α,β,γand δare constants determined by dimensional analyses with experiments.
LV
Reynolds number: Re  (5.3)

Cp
Prandtl number Pr  (5.34a)
k
3 2 g
Gr  L
Grashoff number (5.34b)
 2 ( t)
in which Cp is the specific of heat of fluids under constant pressure,  is the
volumetric coefficient of thermal expansion, t is the duration, and g is the
gravitational acceleration.
 Solid-Fluid Interaction
Modes of heat transfer:

Conduction in solids governed by Fourier law in Eq. (5.28)

Convection in fluids governed by Newton’s cooling law in Eq. (5.33)

There are MEMS structures, e.g. thermally actuated beams with their surfaces
being in contact with surrounding fluids.

At these interfaces the two modes of heat transfer take place with either:
conduction to convection, or convection to conduction.

The situation is further complicated with the building of a “boundary layer” at

the interface on the fluid side. Such boundary layer adds resistance to heat flow.
Consequently, the temperature of the solid at the interface is not equal to that
of the contacting fluids.

Because of both heat conduction and convection take place at the interface of
the solid structure and the surrounding fluid, the thermal boundary condition
at the interface needs to be specifically defined.
Boundary condition at solid-fluid interface

Boundary
Layer FLUID:
SOLID: Tf
T(r,t)
Boundary layer film
qs resistance, 1/h

qf
Normal line to
Boundary surface the surface, n
position: rs

T(r,t)
-k 
→
rs  h[T (r→s ,t)  T f ] (5.35)
n

The thickness of the boundary layer relates to the velocity of the surrounding fluid.
Thicker layers are produced with slow moving fluid, with extreme values in
natural convection cases, which is common in microsystems.
Heat Conduction in Multilayered Thin Films

Many MEMS devices are made of layers of dissimilar materials. Heat flow through
these layers of dissimilar materials require special formulations.
The governing DE for a multi-layer solid is:
Boundary conditions
 2Ti (x,t) 1 Ti (x,t)
T1(x,t): K1, 1 X = X1  (5.40)
T2(x,t): K2, 2
X = X2
x 2 i t
X = X3
in which the layer designation, i = 1,2,3,….
with xi  x  xi1 and t > 0, satisfying the
following conditions:

Prescribed initial conditions in xi  x  xi1

at t = 0, and
Prescribed boundary conditions
at x = 0 and x = xi+1 for t > 0.
These conditions are:
X = Xi
Ti(x,t): Ki, i Ti(xi+1,t) = Ti+1(xi+1,t) for i = 1,2,3,………. , and
X = Xi+1
Boundary conditions
 T i ( xi1,t)  T ( x ,t)
ki  k i1 i1 i1 for i = 1,2,3,……….
x x
x
 Example 5.9  The structure of a thermal actuator is made of a compound beam
involving silicon and SiO2 as illustrated below. A thin copper film is deposited
on the top of the SiO2 layer as the resistant heater.

 This heater will provide a maximum temperature of 50oC at the

top surface of the SiO2 layer. Determine the time required for the entire silicon
beam to reach the input temperature surface temperature 50oC.
 1400 m
 1000 m
 50 m

Material 1  Cu film heater

Silicon dioxide 2 m
Material 2 Silicon

40 m
Support

Given material properties are:

Thermal conductivities: k1 = 1.4 w/m-oC for SiO2 and k2 = 157 w/m-oC for silicon.
Thermal diffusivities: 1 = 0.62x10-6 m2/sec for SiO2 and 2 = 97.52x10-6 m2/sec for silicon.
 Solution:
Since heat will predominantly flow through the thickness of the compound beam
due to the short distance of the passage, a one-dimensional heat conduction analysis
along the thickness direction is justified.
Let T1(x,t) = temperature in SiO2
SiO2 T2(x,t) = temperature in Si
Si
From Eq. (5.47), we have the following DEs:

Thermally insulated boundary,

Surface temperature, Ts = 50oC

T1(x,t)  2 T 1 (x, t) 1  T 1 (x, t) 0 xa

For SiO2  (5.41a)
 x2 1 t
q=0
Heat Flow
T2(x,t)  2 T 2 (x,t)  1  T 2 (x,t)
For Si a xb (5.41b)
 x2 2 t
The initial conditions:
x
T 1 (x, t) t 0  F 1 (x)  20 C
o T 2 (x,t) t 0  F 2 (x)  20 oC
X = a = 2 m X = b = 42 m
The boundary conditions:
X=0  T 2 (x,t)
T 1 (x, t)  50oC 0
x0 x xb42  m
The compatibility conditions:
 T 1 (x,t)  T 2 (x,t)
T 1 (x, t) xa2 m  T 2 (x, t) xa2 m k1  k2
x xa2  m x xa2 m
The solution of this set of DEs and the associated conditions was carried out
by using MathCad, a commercial software package, with graphical output:

t = 600 µs
50
Temperature, oC

40 t = 100 µs
t = 50 µs
30
t = 1 µs
20 Time, t = 0
10
0
0 2 4 6 8 10 12 14 16 18 20 30
SiO2 Si Depth of the Beam, x (µm)

The temperature variations in both layers at selected instances are plotted as shown
in the graph above, from which we determined the time required for the silicon layer
to reach the input temperature of 50oC is 600 micro seconds. This information will
enable the design engineer to assess the sensitivity of the thermally actuated device.
 SUMMARY
● Thermofluids engineering principles are used in the design of MEMS
microsystems such as micro valves and micro fluididcs. Many of these
devices and systems are thermally actuated
● Another major application of thermofluid engineering principle is in
microfabrication such as chemical vapor deposition of thin films.
● Fluid-induced forces must be accounted for in the design of micro
valves and pumps. Fluids also affect thermal behavior of matters.
● Thermal analysis in MEMS and microsystems involve conduction and
convection heat transfer.
● Fourier law governs heat conduction in solids, whereas Newton’s cooling
law is used in convective heat transfer.
● Heat conduction equation, with or without convective boundary conditions,
is used to determine the temperature field (distribution) in the MEMS
structure. This temperature field is used to assess the induced thermal
stresses, strains and displacements. These thermally induced mechanical
behavior is critical in the design of MEMS and microsystems.
● Thermofluids engineering principles for sub-mcrometer scale are radically
different from those in macro-scale. Significant modifications of these
principles and formulations are necessary.
 End of
Chapter