Skip to main content

Section 6.2 Line Integrals

References.

Introduction.

It is often useful to “sum” (integrate) a quantity along a curve. One example is computating the total mass along a curve from a variable density. Another physical example is the summation of increments of work done by or against a force to get the total work done as an object moves along a curve.
These mass and work examples are different in an important way: the first involves integrating a scalar quantity with respect to increments in arc length, \(ds\text{,}\) whereas the second involves a vector quantity, the force, acting in relation to increments in spatial coordinates like \(dx\text{.}\)

Topics.

Subsection 6.2.1 Scalar Line Integrals: Integrating With Respect to Arc Length Along a Curve

Subsubsection 6.2.1.1 Scalar Line Integrals in The Plane

If a plane curve \(C\) is given parametrically as
\begin{equation} x=x(t), \; y=y(t), \quad a \leq t \leq b\tag{6.2.1} \end{equation}
or in vector form \(\vec{r}(t) = x(t)\veci + y(t)\vecj\text{,}\) then one can start by considering sums of the form
\begin{equation*} \sum_{i=1}^n f(x_i^*,y_i^*) \Delta s_i \end{equation*}
where the interval \([a,b]\) of \(t\) values is divided into \(n\) subintervals \([t_{i-1},t_i]\text{,}\) \(1 \leq i \leq n\text{,}\) each point \((x_i^*,y_i^*)=(x(t_i^*),y(t_i^*))\) is given by some \(t_i^*\) in \([t_{i-1},t_i]\) and so lies on the \(i\)-the sub-arc, and \(\Delta s_i\) is the length of that sub-arc.
A now familiar limit process turns these approximations into an integral along the curve:
Definition 6.2.1.
If \(f\) is defined at each point of a smooth curve \(C\) given by Equation (6.2.1) then the line integral of \(f\) along \(C\) is
\begin{equation} \int_C f(x,y) \, ds = \lim_{n \to \infty} \sum_{i=1}^n f(x_i^*,y_i^*) \Delta s_i\tag{6.2.2} \end{equation}
This is also called the line integral of \(f\) along \(C\) with respect to arc-length.
Next, we can express this in terms of an integral in variable \(t\) over the interval \([a,b]\text{.}\)
For \(f(x,y)=1\text{,}\) the limit in Equation (6.2.2) is the definition of arc-length as in Section 3.3 and it was seen there that in effect,
\begin{align} ds \amp= \sqrt{dx^2+dy^2}\notag\\ \amp= \sqrt{\left( \frac{dx}{dt} \right)^2 + \left( \frac{dy}{dt} \right)^2}\ dt,\tag{6.2.3}\\ \amp= \left\| \frac{d\vec{r}}{dt} \right\|\ dt, \quad\text{with } \vec{r}=\vector{x,y}\tag{6.2.4} \end{align}
The same argument here gives:
\begin{equation} \int_C f(x,y) \, ds = \int_{t=a}^b f(x(t),y(t)) \sqrt{\left(\frac{dx}{dt}\right)^2 + \left(\frac{dy}{dt}\right)^2} \; dt\tag{6.2.5} \end{equation}

Subsubsection 6.2.1.2 Scalar Line Integrals in Space

The above concepts extend easily to space curves \(C\text{,}\) given parametrically as
\begin{equation} x=x(t), \; y=y(t), \; z=z(t), \quad a \leq t \leq b\tag{6.2.6} \end{equation}
or in vector form, \(\vec{r}(t) = x(t)\veci + y(t)\vecj + z(t) \veck\text{.}\)
We again define the line integral of \(f\) along \(C\) with respect to arc-length as
\begin{equation*} \int_C f(x,y,z) \, ds = \lim_{n \to \infty} \sum_{i=1}^n f(x_i^*,y_i^*,z_i^*) \Delta s_i \end{equation*}
where in the sum, the interval \([a,b]\) is divided into \(n\) (equally long) subintervals with \(\Delta s_i\) the distance between the endpoints of the corresponding part of the curve, and \((x_i^*,y_i^*,z_i^*)\) some point in that part of the curve.
Also as above, we use the Fundamental theorem of Calculus to turn this limit of a sum into a definite integral over interval \([a,b]\text{:}\)
\begin{align} \int_C f(x,y,z) ds \amp= \int_{t=a}^b f(x(t),y(t),z(t)) \sqrt{\left(\frac{dx}{dt}\right)^2 + \left(\frac{dy}{dt}\right)^2 + \left(\frac{dz}{dt}\right)^2} dt\notag\\ \amp= \int_{t=a}^b f(\vec{r}(t)) \left\|\frac{d \vec{r}}{dt}\right\|\ dt.\tag{6.2.7} \end{align}
The last vector form is convenient as it covers both the 2D ad 3D versions.
The simple and important special case \(f=1\) gives a compact formula for the arc length of the curve,
\begin{equation*} \int_C ds = \int_{t=a}^b \left\|\frac{d \vec{r}}{dt}\right\| \, dt. \end{equation*}

Subsubsection 6.2.1.3 Integrals Along Paths: Piecewise Smooth Curves

A path is another name for a piecewise smooth curve \(C\text{:}\) a collection of smooth curves \(C_1, C_2, \dots C_m\) that join end to end.
This is sometimes denoted \(C=C_1+ C_2 + \cdots + C_m\text{.}\)
The path integral along a path \(C\) is simply the sum of the line integrals along each smooth piece.

Subsection 6.2.2 Vector Line Integrals: Integrating With Respect to Position Coordinates

Subsubsection 6.2.2.1 Line Integrals in the Plane with Respect to the Coordinates, \(x\) and \(y\)

In some situations, the quantity to be summed is \(f(x,y) \Delta x\) or \(f(x,y) \Delta y\text{:}\) for example if \(f\) is a force acting on an object in the \(x\) direction, \(f(x,y) \Delta x\) is the work done by that force when the object moves distance \(\Delta x\) in that direction (no work is associated with movement perpendicular to the force).
Limits of sums give the work done by such a force in the \(x\)-direction as an object moves along curve \(C\) to be
\begin{equation} \int_C f(x,y) \, dx = \lim_{n \to \infty} \sum_{i=1}^n f(x_i^*,y_i^*) \Delta x_i\tag{6.2.8} \end{equation}
called the line integral of \(f\) along \(C\) with respect to \(x\). Likewise the line integral of \(f\) along \(C\) with respect to \(y\) is
\begin{equation} \int_C f(x,y) \, dy = \lim_{n \to \infty} \sum_{i=1}^n f(x_i^*,y_i^*) \Delta y_i\tag{6.2.9} \end{equation}
Much as with Equation (6.2.5), these line integrals can be computed with
\begin{align} \int_C f(x,y) \ dx \amp= \int_{t=a}^b f(x(t),y(t)) \frac{dx}{dt} dt\tag{6.2.10}\\ \int_C f(x,y) \ dy \amp= \int_{t=a}^b f(x(t),y(t)) \frac{dy}{dt} dt\tag{6.2.11} \end{align}
However, the absence of the square root term can make these new integrals easier to work with. It also allows the increments \(\Delta x\) and \(\Delta y\) to be negative depending on the direction of motion along the curve (and so intuitively, also the differentials \(dx\) and \(dy\)), which leads to an important difference in the behavior of these integrals.

Subsubsection 6.2.2.2 Line Integrals in Space Coordinates

Line integrals with respect to each of the space coordinate variables \(x\text{,}\) \(y\) and \(z\) can be defined similarly, so that for example
\begin{equation} \int_C f(x,y,z) dz = \lim_{n \to \infty} \sum_{i=1}^n f(x_i^*,y_i^*,z_i^*) \Delta z_i = \int_a^b f(x(t),y(t),z(t)) \frac{dz}{dt}\ dt\tag{6.2.12} \end{equation}

Subsubsection 6.2.2.3 Paired (and Tripled) Line Integrals and Vector Line Integrals

One often gets a sum of line integrals in the plane in both \(x\) and \(y\text{,}\) and these can be abbreviated
\begin{equation} \int_C P(x,y) \, dx + \int_C Q(x,y) \, dy = \int_C P(x,y) \, dx + Q(x,y) \, dy\tag{6.2.13} \end{equation}
This combination has a concise vector form, called a Vector Line Integral:
\begin{align} \int_C P(x,y) \, dx + Q(x,y)\ dy \amp= \int_C \vector{P(x,y), Q(x,y)} \cdot \vector{dx,dy}\tag{6.2.14}\\ \amp= \int_C \vec{F} \cdot d\vec{r}\tag{6.2.15} \end{align}
where \(\vec{F}= \vector{P(x,y), Q(x,y)}\) and for the position vector \(\vec{r} = \vector{x,y}\text{,}\) we use the short-hand
\begin{equation*} d\vec{r} = d\vector{x,y} = \vector{dx,dy}\text{.} \end{equation*}
When \(P\) and \(Q\) are the components in the \(x\) and \(y\) directions of a force vector \(\vec{F}\text{,}\) this gives the total work done by the force as the object traverses curve \(C\text{.}\)
Likewise, integrals of the form
\begin{equation} \int_C P(x,y,z)\ dx + \int_C Q(x,y,z)\ dy + \int_C R(x,y,z)\ dz = \int_C P\ dx + Q\ dy + R\ dz\tag{6.2.16} \end{equation}
arise—in particular for the example of the work done by a force—and they have the same vector form (6.2.15), now with \(\vec{F} = \vector{P,Q,R}\text{.}\)
For a more completely vector notation form, describe the path \(C\) as \(\vector{r}(t),\ a \le t \le b\) and express each line integral above in terms of this: The integral in Equation (6.2.16) has the form
\begin{equation} \int_C P\ dx + Q\ dy + R\ dz = \int_a^b \left[ P(\vec{r}(t)) \frac{dx}{dt} + Q(\vec{r}(t)) \frac{dy}{dt} + R(\vec{r}(t)) \frac{dz}{dt} \right] dt = \int_a^b \vec{F}(\vec{r}(t)) \cdot \frac{d\vec{r}}{dt}\ dt\tag{6.2.17} \end{equation}
and in practice evaluation of \(\int_C \vec{F} \cdot d\vec{r}\) is done via this form, with the formal substutution \(\ds d\vec{r} = \frac{d\vec{r}}{dt}\ dt\text{:}\) this is sometimes called the line integral of \(\vec{F}\) along \(C\).
This integral \(\ds \int_C \vec{F} \cdot d\vec{r}\) can instead be defined rigorously by the familiar process of approximation by a sum of terms \(\vec{F}(\vec{r}(t_i^*)) \Delta \vec{r}_i\) and taking a suitable limit.

Subsubsection 6.2.2.4 Vector Line Integrals in Terms of the Arc-length Differential \(ds\text{:}\) Circulation and Flux

Circulation Form.
Recall from Equation (3.3.4) in Section 3.3 that the unit tangent vector to \(C\) is \(\vec{T} = d\vec{r}/ds\) and thus
\begin{equation} \int_C \vec{F} \cdot d\vec{r} = \int_a^b \vec{F}(\vec{r}) \cdot \frac{d\vec{r}}{ds}\ ds = \int_C \vec{F} \cdot \T\ ds\tag{6.2.18} \end{equation}
getting back to a line integral with respect to arc length as defined in Equation (6.2.2).
For the example of \(\vec{F}\) being the velocity in a fluid, the dot product in this integral is the rate of flow along the curve. Then for a closed curve, this integral measures the net flow around the curve: hence it is called the circulation around the curve.
Flux Form.
In the plane there is another useful version: \(\T = \vector{dx/ds, dy/ds}\text{,}\) so we can define a unit normal vector
\begin{equation} \N = \vector{dy/ds, -dx/ds}\text{,}\tag{6.2.19} \end{equation}
which is the rotation of \(\T\) a quarter turn clockwise. Thinking of \(\T\) as the “forward direction”, this choice of unit normal points to the right of the curve.
(Note that this is not necessarily the same as the the principle unit normal vector defined in Equation (3.3.8) in Section 3.3, which can be to the left or the right depending on which way the curve is turning.)
In the plane, this new choice has the advantage of always being defined, even at an "inflection point" where \(d\T/dt = \vec{0}\text{,}\) and always continuous so long as \(\T\) is.
The integral of \(\vec{F} \cdot \N\) along a curve is the integral of the component of \(\vec{F}\) normal to the curve; for example if \(\vec{F}\) is the velocity in a fluid, this measures the rate of flow across the curve: it is thus known as the flux of \(\vec{F}\) across \(C\).
This equals the circulation of \(\vec{G} = \vector{-Q, P}\text{,}\) which is \(\vec{F}\) rotated a quarter turn anti-clockwise:
\begin{align} \int_C \vec{F} \cdot \N\ ds \amp= \int_C \vector{P,Q} \cdot \vector{\frac{dy}{ds}, -\frac{dx}{ds}} ds\notag\\ \amp= \int_C \left( P \frac{dy}{ds} - Q \frac{dx}{ds} \right) ds\notag\\ \amp= \int_C -Q\ dx + P\ dy\notag\\ \amp= \int_C \vector{-Q, P} \cdot \T ds\tag{6.2.20} \end{align}
This connection will be useful in Section 6.4, where both the circulation and flux around a closed curve in the plane are related to a double integral over the region “inside” the curve, and also in later sections where these ideas are extended to integrals over surfaces and regions in space.

Subsection 6.2.3 Reversing the Orientation of a Curve

A curve \(C\) described by a parameterization of a curve \(x=f(t)\text{,}\) \(y=g(t)\text{,}\) \(a \leq t \leq b\) has an orientation, meaning a direction of motion from initial point \((x(a),y(a))\) to final point \((x(b),x(b))\text{.}\)
If we reverse the orientation, for example with new parameterization \(u=-t\) so that \(x=f(-u)\text{,}\) \(y=g(-u)\text{,}\) \(-b \leq u \leq -a\text{,}\) the new curve is denoted \(-C\text{.}\)
This reverses the order of limits on integration so that
\begin{equation} \int_{-C} f(x,y)\ dx = -\int_C f(x,y)\ dx, %\; %\int_{-C} f(x,y)\ dy = - \int_C f(x,y)\ dy\tag{6.2.21} \end{equation}
and so on, and for a vector line integral
\begin{equation} \int_{-C} \vec{F} \cdot d\vec{r} = - \int_C \vec{F} \cdot d\vec{r}\tag{6.2.22} \end{equation}
which of course is equally valid in the 3D case.
For the physical example where this integral gives the work done by force \(\vec{F}\) on an object as it moves along the curve \(C\text{,}\) this says naturally that going in the opposite direction negates the amount of work done.
However, scalar line integrals are not changed by this reversal:
\begin{equation} \int_{-C} f(x,y) \, ds = \int_C f(x,y) \, ds.\tag{6.2.23} \end{equation}
Compare this to Equation (6.2.18) for a line integral: there the integrand changes sign, due to \(\T\) negating when the orientation is reversed.
This is because \(\Delta s_i\) remains positive in Equation (6.2.2) whereas \(\Delta x_i\) and \(\Delta y_i\) in Equations (6.2.8) and (6.2.9) change sign.

Subsection 6.2.4 Properties of Vector Line Integrals

Vector line integrals have many properties in common with definite integrals \(\int_a^b f(x)\ dx\) in one dimension:
In comparison, scalar line integrals behave more like integrals over domains in two or three dimensions seen in Chapter 5, and like the one dimensional versions \(\int\limits_{[a,b]} f(x)\ dx\) introduced in Subsection 5.7.1.

Study Guide.

Study Section 6.2 of Calculus Volume 3
 8 
openstax.org/books/calculus-volume-3/pages/6-2-line-integrals
; in particular:
  • The Definitions and Theorems.
  • Examples 14–21, and according to your scientific interests, some of the application examples 22, 23 and 26 (and as usual, the Checkpoints following each).
  • The true/false exercises 39–43.
  • The following exercises (for ranges, do at least one in each): 49–53, 54, 55–58, 65–69.