Visual walkthrough — Viscosity — dynamic μ, kinematic ν = μ - ρ; Newtonian vs non-Newtonian

We are going to answer one question: when you drag a layer of liquid, how much sideways force per area do you feel, and why?

---

Step 1 — The picture we are allowed to start with: two plates

WHAT. Take a puddle of fluid and trap it between two flat plates. The bottom plate is bolted down. The top plate we can grab and slide sideways. The vertical distance between the plates is a height we will call $h$ (just "the gap").

WHY. We want the simplest possible situation where fluid layers slide past each other. No pipes, no curves, no gravity tricks — just a flat sandwich. Everything hard comes later; first we need a clean picture where only one thing varies.

PICTURE. Below, the bottom plate sits still. We push the top plate to the right with speed $U$ (a speed, measured in metres per second). The blue stuff between is the fluid.

---

Step 2 — What the fluid actually does: a ramp of speeds

WHAT. The fluid does not move as one block. The layer glued to the bottom plate does not move at all. The layer glued to the top plate moves at the full speed $U$. Every layer in between moves at some in-between speed. Draw an arrow from each layer: the arrows grow from zero at the bottom to $U$ at the top — a straight ramp.

WHY the bottom layer is stuck. Real fluid sticks to a solid surface it touches. This is called the no-slip condition: the fluid right at a wall matches the wall's speed. It is an experimental fact — dust never fully blows off a fan blade for the same reason. Without it we'd have no drag at all.

WHY the ramp is straight. In this simple gap, nothing tells one layer to speed up more than another, so the speed rises evenly — like the steps of a perfectly even staircase. (We'll trust this straight line for now; Step 4 will name exactly what "straightness" buys us.)

PICTURE. The little arrows form a triangle: zero-length at the floor, full-length $U$ at the ceiling.

---

Step 3 — Why layers pull on each other: molecules trade momentum

WHAT. Zoom in until you can "see" molecules. Molecules in the fast (upper) layer are jiggling and occasionally wander down into a slower layer, carrying their fast sideways motion with them. Molecules from the slow layer wander up carrying their sluggishness. When a fast molecule lands among slow ones, it shoves them along; a slow one arriving among fast ones drags them back.

WHY this makes a force. Sideways motion (momentum) is being handed down the ramp from fast layers to slow ones. Handing over momentum is a force — that's Newton's second law read backwards. So every layer feels a sideways tug from its neighbours trying to erase the speed difference between them. This tug, spread over an area, is what we'll call shear stress.

KEY consequence. The force depends on the difference in speed between neighbouring layers — not on how fast they move together. If the whole deck of layers moved at one speed, no molecule would arrive "out of step," and there'd be zero viscous force. Speed alone does nothing; speed difference is everything.

PICTURE. Two neighbouring layers; curvy arrows show molecules crossing between them, red arrows show the resulting tug that tries to equalise their speeds.

---

Step 4 — Measuring "how different the speeds are": the shear rate

WHAT. We need a number for "how much do speeds change as you climb." Take two nearby heights. The speed changed by a small amount, call it $du$, over a small rise in height, call it $dy$. The ratio $\frac{du}{dy}$ is the slope of the speed ramp — how many m/s you gain per metre you climb.

WHY a ratio (and why this is a slope / derivative). Step 3 told us only the difference matters. But "difference over what distance?" — a 1 m/s gap across a hair-thin layer is far more violent than the same gap across a metre. So we must divide the speed change by the height it happened over. That "small change divided by small distance" is exactly what a derivative is: the local steepness of a curve. We reach for $\frac{du}{dy}$ because it is the one tool that captures local steepness, and Step 3 proved that local steepness is what generates force.

FOR OUR STRAIGHT RAMP the slope is the same everywhere, so we can compute it with whole quantities instead of tiny ones: $\underset{\text{shear rate}}{\underbrace{\dot{\gamma }}}=\frac{du}{dy}=\underset{\text{(top speed)}÷\text{(gap)}}{\underbrace{\frac{U}{h}}}$

Term by term: $\dot{\gamma }$ (read "gamma-dot") is our new name for the slope; $U$ is the full speed change from floor to ceiling; $h$ is the height it happened over; $U/h$ is that change per metre. Units: $\frac{\text{m/s}}{\text{m}}={\text{s}}^{-1}$ — a per-second, because it's a rate of shearing, not a distance.

PICTURE. The ramp with a right triangle drawn under it: horizontal leg $=U$, vertical leg $=h$, and the slope $U/h$ is the "steepness" of the ramp's edge.

---

Step 5 — Newton's experimental guess, drawn as knobs

WHAT. Now measure the sideways force per area, $\tau$ ("tau"), needed to keep the top plate gliding. Newton reasoned about two knobs you could turn:

• Push the top plate faster ($U$ up) → you feel more force. So $\tau$ grows with $U$.
• Make the gap wider ($h$ up) → the speed difference is smeared over more layers, each barely out of step → you feel less force. So $\tau$ shrinks with $h$.

WHY combine them as $U/h$. Both knobs point to the same buried quantity: force tracks speed-per-gap, i.e. $\tau \propto U/h=\dot{\gamma }$. That's not a coincidence — Step 3 already told us force comes from local steepness, and Step 4 named that steepness $\dot{\gamma }$. The two knobs are just two ways of turning the one dial that matters.

PICTURE. Two side-by-side boards: left, faster plate ⇒ thicker force arrow; right, wider gap ⇒ thinner force arrow. The dial in the middle reads $\dot{\gamma }=U/h$.

---

Step 6 — Naming the constant: dynamic viscosity μ, and the law

WHAT. A proportionality is not yet an equation. To turn $\propto$ into $=$ we insert one number that depends only on which fluid it is. Call that number $\mu$ ("mu"), the dynamic viscosity.

WHY define it this way. Honey and water sit in identical plates with identical $U$ and $h$ — same $\dot{\gamma }$ — yet honey needs far more force. The only difference is the fluid itself, so all of "how sticky this fluid is" gets packed into a single multiplier $\mu$. It is, by definition, whatever number makes the force come out right: $\boxed{\tau =\mu \dot{\gamma }=\mu \frac{du}{dy}}$

Reading the box left to right:

• $\tau$ — the sideways stress you must apply (Pa),
• $\mu$ — the fluid's stickiness dial (Pa·s); big for honey, tiny for air,
• $\frac{du}{dy}$ — the shear rate from Step 4 (s⁻¹); how out-of-step the layers are.

Multiply "how sticky" by "how out-of-step" and you get "how much force per area." Notice the shape is the simplest honest one: if $\frac{du}{dy}=0$ (uniform flow) then $\tau =0$ automatically — exactly the fact Step 3 demanded — and doubling the gradient doubles the stress.

Units, earned not memorised: $[\mu ]=\frac{[\tau ]}{[\dot{\gamma }]}=\frac{\text{Pa}}{{\text{s}}^{-1}}=\text{Pa⋅s}=\frac{\text{kg}}{\text{m⋅s}}$

PICTURE. The finished law as a straight line: horizontal axis $\dot{\gamma }$, vertical axis $\tau$, a line through the origin whose slope is $\mu$. Steeper line = stickier fluid.

---

Step 7 — Edge cases the line must survive

We drew one straight line. A law is only trustworthy if it behaves at the extremes. Walk each corner:

Case A — zero shear ($\dot{\gamma }=0$). All layers move alike (or nothing moves). $\tau =\mu \cdot 0=0$. Picture: the line passes exactly through the origin. Uniform motion costs no viscous force — the misconception from Step 3, now guaranteed by the formula.

Case B — reverse the drag ($\dot{\gamma }<0$). Slide the top plate left instead of right. The gradient flips sign, so $\tau$ flips sign: the fluid now tugs the other way. Picture: the line continues straight into the third quadrant. Viscous stress always opposes the relative sliding — the sign takes care of itself.

Case C — bolt both plates ($U=0$). Then $\dot{\gamma }=0/h=0$, stress $0$. A still fluid feels no viscous stress no matter how thick. Stickiness is dormant until something shears.

Case D — the gap shrinks toward zero ($h\to 0$). With $U$ fixed, $\dot{\gamma }=U/h\to \infty$, so $\tau \to \infty$. Picture: squeeze the plates together and the force to keep sliding blows up — exactly why a thin oil film between two nearly-touching surfaces (a bearing) can carry enormous load. This is the seed of Boundary Layer and Poiseuille Flow, where thin high-gradient regions dominate.

Recall Check yourself on the edge cases

If I move both plates rightward at the same speed, what is the viscous stress? ::: Zero — same speed means $\dot{\gamma }=0$, so $\tau =\mu \cdot 0=0$. Why does pushing the plates closer together (smaller $h$) increase the force? ::: Because $\dot{\gamma }=U/h$ grows as $h$ shrinks; same speed change is packed into a thinner gap, so the layers are more out-of-step.

---

Step 8 — When the line refuses to stay straight (non-Newtonian)

WHAT. For water, air, glycerin, $\mu$ is one fixed number — the graph is a single straight line. These are Newtonian fluids. But ketchup, cornstarch-water, and toothpaste bend the graph: their slope changes as you shear harder.

WHY the slope can change. In Step 6 we assumed the fluid's "stickiness" is a constant. But some fluids have internal structure — tangled chains, packed grains — that rearranges under shear. So the effective slope is no longer fixed; we call it the apparent viscosity ${\mu }_{app}=\tau /\dot{\gamma }$, and it depends on $\dot{\gamma }$.

The four boards to know:

• Newtonian — straight through origin, slope $\mu$ fixed. (water, air)
• Shear-thinning — curve bends down; slope drops as you shear faster. (ketchup, blood, paint — chains align and slip) → see Reynolds Number contexts where this matters.
• Shear-thickening — curve bends up; slope rises with shear. (oobleck — grains jam)
• Bingham plastic — flat refusal until a yield stress ${\tau }_0$, then a line: $\tau ={\tau }_0+{\mu }_p\dot{\gamma }$. (toothpaste)

A compact model for the bending ones: $\tau =K{\dot{\gamma }}^n$, with $n=1$ Newtonian, $n<1$ thinning, $n>1$ thickening.

PICTURE. All four curves on one $\tau$–$\dot{\gamma }$ board so the shapes contrast at a glance.

---

The one-picture summary

Everything above, on one board: the plates (Step 1), the speed ramp with no-slip (Step 2), the triangle giving $\dot{\gamma }=U/h$ (Step 4), and the $\tau$–$\dot{\gamma }$ graph whose slope is $\mu$ (Step 6), with the Newtonian line plus the bending non-Newtonian curves (Step 8).

Recall Feynman retelling — the walkthrough in plain words

Trap some liquid between two flat plates and slide the top one. The bottom liquid sticks to the bottom plate and doesn't move; the top liquid rides along with the top plate. In between, the liquid moves in a smooth ramp of speeds — slow at the bottom, fast at the top. Why does anyone feel any drag? Because molecules keep hopping between layers, carrying their speed with them, and a fast molecule shoving into a slow layer speeds it up while dragging itself down. That trading of motion is a force. Crucially, it only happens when the layers move at different speeds — if the whole slab drifts together, nobody's out of step and there's zero drag. So we measure "how different the speeds are per metre of height" — that's the slope of the ramp, $U$ over the gap $h$. Newton found the force per area is just that slope times one number for the fluid: sticky honey has a big number, slippery water a small one. That number is the viscosity $\mu$, and the law is force-per-area = stickiness × steepness. Finally, most everyday liquids keep the same stickiness no matter how hard you shear — a nice straight-line graph. But a few (ketchup, oobleck, toothpaste) rearrange inside when pushed, so their line bends: they get thinner or thicker or stubbornly refuse to flow until you shove hard enough.

---