Opening the book…
What makes something rotate? Not just a force, but a force applied in the right place. Pushing a door near its hinges barely turns it; the same push at the handle swings it easily. The rotational effectiveness of a force is the torque. For a force \(F\) applied at a distance \(r\) from the axis, making an angle \(\theta\) with the line from the axis to the point of application,
Key Equation: Torque
\[\tau = rF\sin\theta = r_\perp F, \tag{10.12}\]
where \(r_\perp = r\sin\theta\) is the lever arm, the perpendicular distance from the axis to the line of the force (Figure 10.4a). Grouping the same two factors the other way gives an equally useful third form, \(\tau = rF_\perp\), in which \(F_\perp = F\sin\theta\) is the component of the force perpendicular to \(\vec r\) (Figure 10.4b). Read either way, the message is identical: only the perpendicular part of the force does any turning. A force aimed straight away from the axis (\(\theta = 0\)) or straight at it (\(\theta = 180^\circ\)) has no perpendicular component and produces no torque at all. The SI unit is the newton-meter (\(\,\mathrm{N\cdot m}\)). Torque has a sense (we will call counterclockwise positive), and in Section 10.4 we will give it a full vector definition with the right-hand rule.
Checkpoint 10.4
A stubborn bolt will not budge, so you slip a length of pipe over the wrench handle to double its length, then push just as hard as before and still perpendicular to the handle.
- By what factor does the torque on the bolt change?
- Starting over with the original wrench, suppose instead you keep the handle length the same but push at \(30^\circ\) to the handle rather than perpendicular to it. By what factor does the torque change now?
- Which of the two changes helps more? And why is a push \(30^\circ\) past perpendicular just as effective as one \(30^\circ\) short of it?
Torque as a vector
So far we have treated torque as a magnitude with a clockwise-or-counterclockwise sense. That is enough for motion in a plane, but torque is really a vector, and its full definition is the cross product of the position vector and the force, exactly the vector product introduced in Chapter 3:
Key Equation: Torque as a vector
\[\vec{\tau} = \vec{r}\times\vec{F}, \qquad |\vec{\tau}| = rF\sin\theta. \tag{10.13}\]
Here \(\vec r\) runs from the axis to the point where the force is applied, and \(\theta\) is the angle between \(\vec r\) and \(\vec F\) when the two are drawn tail to tail. The magnitude \(rF\sin\theta\) is the lever-arm result we already know (equation (3.17)). The direction is given by the right-hand rule: point the fingers of your right hand along \(\vec{r}\), curl them toward \(\vec{F}\), and your thumb points along \(\vec{\tau}\), perpendicular to the plane of \(\vec{r}\) and \(\vec{F}\) (Figure 10.5). For rotation in the \(xy\)-plane the torque points along \(\pm z\): out of the page for a counterclockwise torque, into the page for a clockwise one. This is the precise meaning behind the counterclockwise-positive sign convention we have been using.
Note: Two ways to go wrong with the cross product
First, order matters. The cross product is anticommutative, so \(\vec F\times\vec r = -\vec r\times\vec F\): writing the factors in the wrong order reverses the torque. The position vector always comes first. Second, \(\theta\) is the angle between \(\vec r\) and \(\vec F\) drawn tail to tail, not the angle you happen to see in the sketch between the force and some nearby edge or wall. If you push on a wrench handle at \(30^\circ\) past perpendicular, then \(\theta = 120^\circ\), not \(30^\circ\).
The same right-hand rule promotes angular velocity to a vector. Curl the fingers of your right hand in the direction the body is turning; your thumb then points along \(\vec\omega\), parallel to the rotation axis. A wheel spinning counterclockwise in the page has \(\vec\omega\) out of the page. With that convention, \(\vec\tau\) and \(\vec\omega\) point the same way when a body is speeding up and opposite ways when it is slowing down: the rotational echo of \(\vec F\) and \(\vec v\) for straight-line motion.
Example 10.3: Torque from the cross product
A force \(\vec{F} = (12\,\hat\imath + 3\,\hat\jmath)\,\mathrm{N}\) is applied at a point located at \(\vec{r} = (1.4\,\hat\imath + 0.50\,\hat\jmath)\,\mathrm{m}\), on an object pivoted about the \(z\)-axis. Find the torque about the axis.
Solution
Both vectors lie in the \(xy\)-plane (\(r_z = F_z = 0\)), so in the determinant of equation (3.19) every term that multiplies \(\hat\imath\) or \(\hat\jmath\) contains a \(z\)-component and vanishes. Only the \(z\)-component survives, \(\tau_z = r_x F_y - r_y F_x\): the \(x\) of the first vector times the \(y\) of the second, minus the \(y\) of the first times the \(x\) of the second.
\[\vec{\tau} = \vec{r}\times\vec{F} = \big[(1.4)(3) - (0.50)(12)\big]\hat{k}\,\mathrm{N\cdot m} = (4.2 - 6.0)\hat{k}\,\mathrm{N\cdot m} = \boxed{-1.8\,\hat{k}\ \,\mathrm{N\cdot m}}.\]
The order matters: \(\vec{r}\) comes first because torque is defined as \(\vec{r}\times\vec{F}\); reversing it would flip the sign.
The torque has magnitude \(1.8\,\mathrm{N\cdot m}\) and points in the \(-z\) direction (into the page), so it would tend to turn the object clockwise. Note that the two terms compete: the \(y\)-component of the force, acting at lever arm \(r_x = 1.4\,\mathrm{m}\), tries to turn the object counterclockwise (\(+4.2\,\mathrm{N\cdot m}\)), while the \(x\)-component, acting at lever arm \(r_y = 0.50\,\mathrm{m}\), tries to turn it clockwise (\(-6.0\,\mathrm{N\cdot m}\)). Here the second wins.
Checkpoint 10.5
A wheel lies flat in the plane of this page and is spinning clockwise.
- In which direction does \(\vec\omega\) point?
- A torque is applied that makes the wheel spin faster. In which direction does \(\vec\tau\) point?
- Now the wheel slows down instead. Which of your two answers changes, and which does not?
Newton’s second law for rotation
Torque is to rotation what force is to translation, and it enters an exactly analogous law. The reason is short enough to give here. Apply Newton’s second law to one particle of the body in the tangential direction, \(F_{t,i} = m_i a_{t,i}\), and multiply both sides by that particle’s distance \(r_i\) from the axis. The left side becomes the torque on the particle, and on the right \(a_{t,i} = r_i\alpha\) turns the expression into
\[\tau_i = r_i F_{t,i} = m_i r_i^{2}\,\alpha.\]
Every particle of a rigid body shares the same \(\alpha\), so summing over the whole body factors it out, and what is left of the sum is precisely the moment of inertia. The result is Newton’s second law for rotation:
Key Equation: Newton’s second law for rotation
\[\sum\tau = I\alpha. \tag{10.14}\]
Compare \(\sum F = ma\): torque replaces force, moment of inertia replaces mass, and angular acceleration replaces linear acceleration. One condition is built into the derivation and easy to overlook: the torques on the left and the moment of inertia on the right must be taken about the same fixed axis.
Example 10.4: A hanging mass turning a wheel
A wheel of radius \(R\) and moment of inertia \(I\) turns on a frictionless horizontal axle. A light cord is wrapped around its rim, and from the free end hangs an object of mass \(m\). When released, the object falls and the cord unwinds, spinning the wheel. Find the angular acceleration of the wheel, the linear acceleration of the object, and the cord tension.
Solution
Step 1: One equation for each object. Treat the two objects separately. For the falling block (taking down as positive), the weight points down and the tension up:
\[mg - F_T = ma.\]
For the wheel, the cord’s tension acts at the rim with lever arm \(R\), producing a torque \(F_T R\), and Newton’s second law for rotation gives
\[F_T R = I\alpha.\]
Step 2: The link between them. That is two equations in three unknowns (\(a\), \(\alpha\), \(F_T\)), so we need one more relation. The cord does not slip, so the block’s acceleration matches the rim’s tangential acceleration: \(a = R\alpha\), i.e. \(\alpha = a/R\). Substituting into the wheel equation and dividing both sides by \(R\):
\[F_T R = I\,\frac{a}{R} \;\Longrightarrow\; F_T = \frac{Ia}{R^2}.\]
Step 3: Solve for \(a\). Substitute this \(F_T\) into the block equation. Then gather both terms containing \(a\) on one side and factor it out:
\[mg - \frac{Ia}{R^2} = ma \;\Longrightarrow\; mg = ma + \frac{Ia}{R^2} = a\left(m + \frac{I}{R^2}\right).\]
Divide both sides by the parenthesis, then multiply numerator and denominator by \(R^2\) to clear the fraction-within-a-fraction:
\[a = \frac{mg}{m + I/R^2} = \boxed{\frac{mgR^{2}}{I + mR^{2}}}.\]
Step 4: The other two. Each follows by substituting this \(a\):
\[\alpha = \frac{a}{R} = \boxed{\frac{mgR}{I + mR^{2}}}, \qquad F_T = \frac{Ia}{R^{2}} = \frac{I}{R^2}\cdot\frac{mgR^2}{I + mR^2} = \boxed{\frac{mgI}{I + mR^{2}}}.\]
In the tension, the \(R^2\) in the numerator of \(a\) cancels the \(R^2\) in the denominator of \(I/R^2\).
Checks. If \(I \to 0\) the wheel offers no resistance and \(a \to mgR^2/(mR^2) = g\), \(F_T \to 0\): the block is in free fall. A massive wheel (large \(I\)) makes \(a\) small. The units work too: \(I\) and \(mR^2\) are both \(\,\mathrm{kg\cdot m^2}\), so they can be added, and \(a\) comes out in the units of \(g\).
Notice also that \(F_T < mg\) for any finite \(I\), since \(I/(I + mR^2) < 1\). It has to be: the block is accelerating downward, so the upward pull of the cord cannot balance its weight. Setting \(F_T = mg\) is the most common mistake in problems of this kind, and it would leave the block hanging motionless.
Rolling down a ramp
With \(\sum\tau = I\alpha\) in hand we can return to the rolling wheel of Section 10.2. Set a wheel on an incline and it accelerates down the slope, spinning faster as it goes. Something must be exerting the torque that spins it up, and gravity cannot: gravity acts at the center of mass, so its lever arm about the center is zero. The normal force acts along a line through the center too. The only remaining candidate is static friction at the contact point, and it is what makes rolling on a ramp possible at all. On a perfectly frictionless ramp a wheel would slide down without ever beginning to turn.
Take a body of mass \(m\), radius \(R\), and moment of inertia written in the form \(I_{\rm cm} = \beta mR^2\), where \(\beta\) is a pure number fixed by the shape: \(\beta = 1\) for a hoop, \(\tfrac{1}{2}\) for a solid disk or cylinder, \(\tfrac{2}{5}\) for a solid sphere (Table 10.1). On an incline of angle \(\theta\), Newton’s second law along the slope and about the center of mass gives two equations,
\[mg\sin\theta - F_f = ma_{\rm cm}, \qquad F_f R = I_{\rm cm}\alpha = \beta mR^2\,\frac{a_{\rm cm}}{R},\]
where the rolling condition \(a_{\rm cm} = R\alpha\) has already been used in the second. The second equation says \(F_f = \beta m a_{\rm cm}\); substituting into the first and solving,
Key Equation: Rolling down an incline
\[a_{\rm cm} = \frac{g\sin\theta}{1+\beta}, \qquad F_f = \frac{\beta}{1+\beta}\,mg\sin\theta, \qquad \mu_s \ge \frac{\beta\tan\theta}{1+\beta}. \tag{10.15}\]
The last condition comes from requiring \(F_f \le \mu_s F_N\) with \(F_N = mg\cos\theta\): if the surface cannot supply that much friction, the body slips and the rolling condition fails. Notice what is absent from all three results. Neither \(m\) nor \(R\) appears, only the shape factor \(\beta\). Two spheres of wildly different size and mass roll down the same ramp side by side, and a hoop (\(\beta = 1\)) always loses to a solid sphere (\(\beta = \tfrac{2}{5}\)). More of the hoop’s mass sits far from its center, so friction must take a larger share of the downhill pull just to spin it up, leaving less to speed it down the slope. Section 11.1 tells the same story in terms of energy.
Checkpoint 10.6
A solid sphere (\(\beta = \tfrac{2}{5}\)) and a thin spherical shell (\(\beta = \tfrac{2}{3}\)) have the same mass and radius and are placed on the same ramp. Using equation (10.15) and without computing anything numerically, decide which one requires the larger coefficient of static friction to roll without slipping, and explain in words why the answer had to come out that way. What happens to both if the ramp is made steeper and steeper?
Example 10.5: An unfurling roll of paper
A roll of paper is held by its free end and released, unfurling as it falls (Figure 10.6). The roll has outer radius \(R = 6.0\,\mathrm{cm}\), inner radius \(r = 1.8\,\mathrm{cm}\), and mass \(m = 200\,\mathrm{g} = 0.200\,\mathrm{kg}\), and it falls \(s = 3.0\,\mathrm{m}\). Treating it as a uniform thick-walled (annular) cylinder, for which \(I = \tfrac{1}{2}m(R^2 + r^2)\), and assuming \(R\) and \(m\) stay essentially constant, find (a) its translational acceleration, (b) its angular acceleration, (c) its final translational speed, (d) its final angular speed, and (e) the tension in the sheet. Take \(g = 9.8\,\mathrm{m/s^2}\).
Solution
Step 1: The physical picture. The sheet plays exactly the part the ground plays for a rolling wheel: the point where the paper leaves the roll is instantaneously at rest, so the roll “rolls” down the sheet. The tension therefore acts at radius \(R\), and the unwinding constraint is \(a = R\alpha\).
Step 2: Two equations. Newton’s second law for the falling center of mass (down positive), and for rotation about the center:
\[mg - F_T = ma, \qquad F_T R = I\alpha = I\,\frac{a}{R}.\]
Dividing the second equation by \(R\) gives \(F_T = Ia/R^2\), the same structure as the hanging mass and wheel earlier in this section.
Step 3: Solve for \(a\). Substitute this \(F_T\) into the first equation, gather the \(a\) terms, and factor:
\[mg - \frac{Ia}{R^2} = ma \;\Longrightarrow\; mg = a\left(m + \frac{I}{R^2}\right).\]
Now divide both sides by \(m\) (so that \(m/m = 1\) and \(I/R^2 \div m = I/(mR^2)\)), and then by the parenthesis:
\[a = \frac{g}{1 + I/(mR^2)}.\]
Writing it this way isolates the dimensionless ratio \(I/(mR^2)\), which depends only on the shape of the roll. With \(I = \tfrac{1}{2}m(R^2 + r^2)\), the \(m\) cancels, and dividing each term in the parenthesis by \(R^2\) gives
\[\frac{I}{mR^2} = \frac{\tfrac{1}{2}m(R^2 + r^2)}{mR^2} = \frac{1}{2}\left(1 + \frac{r^2}{R^2}\right) = \frac{1}{2}\left(1 + \left(\frac{1.8}{6.0}\right)^{2}\right) = \frac{1}{2}(1 + 0.09) = 0.545.\]
The two radii appear only as a ratio, so there is no need to convert centimeters to meters for this step. Then:
(a) Translational acceleration. \[a = \frac{9.8\,\mathrm{m/s^2}}{1 + 0.545} = \frac{9.8}{1.545}\,\mathrm{m/s^2} \approx \boxed{6.3\,\mathrm{m/s^2}}.\]
(b) Angular acceleration. From \(a = R\alpha\), divide by \(R\), which now must be in meters (\(R = 0.060\,\mathrm{m}\)):
\[\alpha = \frac{a}{R} = \frac{6.34\,\mathrm{m/s^2}}{0.060\,\mathrm{m}} \approx \boxed{106\,\mathrm{rad/s^2}}.\]
(c) Final translational speed. The acceleration is constant and the roll starts from rest, so \(v^2 = 0 + 2as\):
\[v = \sqrt{2(6.34\,\mathrm{m/s^2})(3.0\,\mathrm{m})} = \sqrt{38.0\,\mathrm{m^2/s^2}} \approx \boxed{6.2\,\mathrm{m/s}}.\]
(d) Final angular speed. The unwinding constraint holds for speeds too, \(v = R\omega\):
\[\omega = \frac{v}{R} = \frac{6.17\,\mathrm{m/s}}{0.060\,\mathrm{m}} \approx \boxed{103\,\mathrm{rad/s}}.\]
(e) Tension in the sheet. From the first equation of Step 2:
\[F_T = mg - ma = m(g - a) = (0.200\,\mathrm{kg})(9.8 - 6.34)\,\mathrm{m/s^2} \approx \boxed{0.69\,\mathrm{N}}.\]
Checks. The tension came out well below the roll’s weight of \(1.96\,\mathrm{N}\), as it must for a roll that is accelerating downward. Part (e) can also be checked from the rotation equation: \(F_T = Ia/R^2 = m(0.545)a = (0.200)(0.545)(6.34) = 0.69\,\mathrm{N}\). And the moment of inertia behaves sensibly at the extremes: as \(r \to 0\) the formula gives \(\tfrac{1}{2}mR^2\), the solid cylinder, while as \(r \to R\) it gives \(mR^2\), the thin hoop.
Note: Taking torques about an accelerating center of mass
The example above applied \(\sum\tau = I_{\rm cm}\alpha\) about the center of the roll: a point that is itself accelerating downward, and therefore not the origin of an inertial frame. That step is nonetheless legitimate, and it is worth knowing why: for any rigid body, torques and moments of inertia taken about an axis through the center of mass obey \(\sum\tau_{\rm cm} = I_{\rm cm}\alpha\) even when the center of mass accelerates. The center of mass is the one accelerating point for which this works. About any other accelerating point, extra terms appear and the simple law fails.
Answers to Checkpoints
Checkpoint 10.4. (a) Torque is \(rF\sin\theta\). Doubling \(r\) with the same perpendicular push (\(\sin 90^\circ = 1\)) doubles the torque, a factor of \(2\). (b) With the original handle, pushing at \(30^\circ\) to it multiplies the torque by \(\sin 30^\circ = 0.5\), only half the torque of a perpendicular push. (c) Lengthening the handle helps more: it doubles the torque, while the angled push halves it. A push \(30^\circ\) past perpendicular makes \(120^\circ\) with the handle, and one \(30^\circ\) short makes \(60^\circ\); since \(\sin 120^\circ = \sin 60^\circ = 0.866\), both give the same torque. Only the component of the force perpendicular to the handle turns the bolt, and it is the same in both cases.
Checkpoint 10.5. (a) Curl the fingers of your right hand in the clockwise sense of rotation: your thumb points into the page, so \(\vec\omega\) points into the page. (b) Speeding up means \(\vec\omega\) grows in the direction it already points, so \(\vec\tau\) (like \(\vec\alpha\)) points into the page too. (c) If the wheel slows, \(\vec\tau\) reverses and points out of the page, while \(\vec\omega\) still points into the page. The direction of \(\vec\omega\) is set by the sense of rotation; the direction of \(\vec\tau\) is set by whether the rotation is speeding up or slowing down.
Checkpoint 10.6. The shell. The required coefficient is \(\mu_s \ge \beta\tan\theta/(1+\beta)\), and the factor \(\beta/(1+\beta)\) increases with \(\beta\): it is \(2/7 \approx 0.29\) for the solid sphere and \(2/5 = 0.40\) for the shell. In words, the shell carries its mass farther from its center, so a larger torque, and therefore a larger friction force, is needed to spin it up in step with its acceleration down the ramp. As the ramp steepens, \(\tan\theta\) grows without limit, so eventually both must slip; the shell reaches that point first. (At \(\theta = 90^\circ\) there is no normal force at all, so no friction is possible.)
This page titled 10.4: Torque and the Right-Hand Rule was last modified on Thu, 12 Nov 2026 04:26:58 GMT and is shared under a CC BY 4.0 license and was authored, remixed, and/or curated by Michael Ragsdale.
The book could not be loaded just now.