Why Coordinate Choice Matters
A three-dimensional region can often be described in more than one coordinate system. The region itself does not change, but the equations and bounds used to describe it can change substantially. A boundary that is a simple constant-coordinate surface in one system may require a variable bound or a complicated inequality in another.
The practical goal is to choose coordinates that fit the region’s shape and symmetry. A region organized around the \(z\)-axis often suits cylindrical coordinates. A region organized around a point, especially a sphere or cone centered at the origin, often suits spherical coordinates. Rectangular boxes and boundaries parallel to coordinate planes often suit Cartesian coordinates.
Coordinate Descriptions and Their Boundaries
In Cartesian coordinates, a point is described by \((x,y,z)\). In cylindrical coordinates, a point is described by \((r,\theta,z)\), where \(r\) is distance from the \(z\)-axis and \(\theta\) is the angle in the \(xy\)-plane. The conversion relations are \(x=r\cos\theta\), \(y=r\sin\theta\), and \(x^2+y^2=r^2\). In spherical coordinates, a point is described by \((\rho,\theta,\phi)\), where \(\rho\) is distance from the origin, \(\theta\) is the azimuthal angle, and \(\phi\) is measured down from the positive \(z\)-axis:
Here \(r\ge 0\), \(\rho\ge 0\), and a full turn in the azimuthal angle can be represented by \(0\le\theta<2\pi\). The polar angle satisfies \(0\le\phi\le\pi\). These conventions make the corresponding geometric boundaries especially recognizable.
| Coordinate system | Boundaries it often simplifies |
|---|---|
| Cartesian \((x,y,z)\) | Planes such as \(x=a\), \(y=b\), or \(z=c\); rectangular boxes |
| Cylindrical \((r,\theta,z)\) | Cylinders \(r=a\), vertical planes \(\theta=a\), and horizontal planes \(z=c\) |
| Spherical \((\rho,\theta,\phi)\) | Spheres \(\rho=a\), cones \(\phi=a\), and azimuthal half-planes \(\theta=a\) |
A coordinate description using bounds means that each coordinate is restricted to an interval, possibly with endpoints depending on coordinates chosen earlier in the description. A description is especially convenient when the intervals are simple and the dependence between bounds is limited. The best choice is not necessarily the one with the shortest-looking first inequality; it is the one that describes the whole region accurately with the least complicated restrictions.
Worked Example: A Ball in Three Coordinate Systems
Worked Example: A Ball in Three Coordinate Systems
Describe the ball of radius \(3\) centered at the origin in Cartesian, cylindrical, and spherical coordinates. Compare the resulting bounds.
The ball consists of points whose distance from the origin is at most \(3\). In Cartesian coordinates, the distance condition is:
One set of iterated Cartesian bounds is:
To check these bounds, the first two restrictions give \(x^2+y^2\le 9\), and the last gives \(z^2\le 9-x^2-y^2\). Therefore \(x^2+y^2+z^2\le 9\). Conversely, every point satisfying that inequality meets the displayed bounds in the chosen order.
In cylindrical coordinates, \(x^2+y^2=r^2\), so the ball condition becomes \(r^2+z^2\le 9\). Thus one description is:
For each allowed \(r\), the two bounds on \(z\) are exactly the solutions of \(r^2+z^2\le 9\). The angle makes a full turn, so this describes the entire ball, not only a slice of it.
In spherical coordinates, distance from the origin is \(\rho\), so the ball condition is simply \(0\le\rho\le 3\). A complete description is:
Every point in these bounds has distance \(\rho\le 3\). Conversely, every point in the ball has a distance \(\rho\) in that interval and has angles in the stated ranges. Here spherical coordinates give the most direct bounds because the boundary is a sphere centered at the origin. The Cartesian and cylindrical descriptions are still correct, but their bounds vary with the other coordinates.
Worked Example: A Region Inside a Sphere and a Cone
Worked Example: A Region Inside a Sphere and a Cone
Describe the region inside the sphere \(x^2+y^2+z^2\le 4\) and above the cone \(z=\sqrt{x^2+y^2}\). Include the cone’s interior on the positive \(z\)-axis.
The sphere gives \(\rho\le 2\). In spherical coordinates, the cone equation becomes \(\rho\cos\phi=\rho\sin\phi\). For \(\rho>0\), this is \(\phi=\pi/4\). The region above the cone is on the side containing the positive \(z\)-axis, so \(0\le\phi\le\pi/4\). The origin is included and is obtained by \(\rho=0\). Therefore the bounds are:
To check the description, \(\rho\le 2\) places every point inside the sphere. Also, for \(0\le\phi\le\pi/4\), \(\cos\phi\ge\sin\phi\), so \(z=\rho\cos\phi\ge\rho\sin\phi=r=\sqrt{x^2+y^2}\). Thus each described point is above the cone. Conversely, the sphere condition gives \(\rho\le 2\), and the cone inequality \(z\ge\sqrt{x^2+y^2}\) corresponds to \(\phi\le\pi/4\) in the upper half-space. The full range of \(\theta\) includes every direction around the \(z\)-axis.
For comparison, cylindrical coordinates give a less direct description. The sphere gives \(z\le\sqrt{4-r^2}\) in the upper half, while the cone requires \(z\ge r\). These bounds can both hold only if \(r\le\sqrt{4-r^2}\), so \(2r^2\le4\), or \(0\le r\le\sqrt{2}\). The cylindrical bounds are therefore:
At \(r=\sqrt{2}\), the lower and upper \(z\)-bounds both equal \(\sqrt{2}\), as expected where the cone meets the sphere. Spherical coordinates express the two central boundaries with constant bounds, so they are the more natural choice here.
Worked Example: A Vertical Cylinder with a Flat Top and Bottom
Worked Example: A Vertical Cylinder with a Flat Top and Bottom
Describe the solid inside the cylinder \(x^2+y^2\le 4\), between the horizontal planes \(z=0\) and \(z=3\).
In Cartesian coordinates, the base condition is \(x^2+y^2\le4\), which is a disk rather than a rectangular range. One possible set of bounds is \(-2\le x\le2\), \(-\sqrt{4-x^2}\le y\le\sqrt{4-x^2}\), and \(0\le z\le3\).
In cylindrical coordinates, the cylinder is \(r\le2\), and the two horizontal planes are the constant bounds \(z=0\) and \(z=3\). The angle makes a complete turn. Thus:
Checking the bounds, \(0\le r\le2\) is equivalent to \(x^2+y^2\le4\), while \(0\le z\le3\) imposes exactly the stated vertical restrictions. The full angle range covers the whole circular base.
Spherical coordinates do not make this region as simple. The cylinder condition becomes \(\rho\sin\phi\le2\), and the top plane becomes \(\rho\cos\phi=3\). Consequently, the radial bound depends on the angle or the angular bounds depend on the radius. Cylindrical coordinates are preferable because both the curved side and the flat ends are standard cylindrical coordinate surfaces.
A Practical Selection Method
The choice of coordinates can be made systematically. First identify the boundaries that define the region, then ask which coordinate system turns the largest number of them into constant-coordinate surfaces. Finally, check the ranges and intersections: a collection of plausible-looking bounds can still describe too much, omit part of the region, or include the wrong side of a boundary.
Look for rectangular faces, circular cylinders, spheres, cones, and planes. Note whether the region is centered on the origin or aligned with the \(z\)-axis.
Use \(x,y,z\) for coordinate-plane boundaries, \(r,\theta,z\) for axial cylinders and horizontal planes, and \(\rho,\theta,\phi\) for spheres and cones centered at the origin.
Use a full azimuthal range only when the region goes all the way around the axis. Restrict \(\theta\) for a wedge, and restrict \(\phi\) when the region lies within a cone or between cones.
Translate the bounds back into inequalities for \(x,y,z\), or test the boundary surfaces and representative points. Confirm that every described point is in the region and every point in the region is described.
When Boundaries Are Centered Elsewhere
Coordinate systems are most convenient when their centers and axes match the region. A sphere centered at \((0,0,1)\), for example, is not described by a constant \(\rho\) in the spherical coordinates defined from the origin. If its radius is \(2\), its Cartesian equation is \(x^2+y^2+(z-1)^2\le4\). Cylindrical coordinates about the \(z\)-axis give \(r^2+(z-1)^2\le4\), and hence the bounds \(0\le r\le2\), \(0\le\theta<2\pi\), and \(1-\sqrt{4-r^2}\le z\le1+\sqrt{4-r^2}\). The sphere is not reduced to a constant \(z\)-bound, but the cylindrical description uses its symmetry about the \(z\)-axis.
This example also illustrates why “use spherical coordinates for a sphere” is not a sufficient rule. The center of that sphere matters. A sphere centered at the origin has a constant \(\rho\)-bound; a sphere centered elsewhere generally does not in this coordinate system. The most useful choice is the one that simplifies the actual boundaries of the region, not merely one boundary’s name.
Common Pitfall: Keeping the Wrong Angular Range
Suppose the region inside the sphere \(\rho\le2\) is restricted to the upper half-space \(z\ge0\). It is tempting to use the full spherical angle range \(0\le\phi\le\pi\), because that is the standard range for all directions. But \(z=\rho\cos\phi\), and \(\cos\phi\) is negative when \(\pi/2<\phi\le\pi\). The proposed bounds therefore include points with \(z<0\), which are not in the upper half-space.
The correction is to impose \(\cos\phi\ge0\), which gives \(0\le\phi\le\pi/2\). The complete corrected description is:
Substitution verifies the correction: \(z=\rho\cos\phi\ge0\) throughout these bounds. Conversely, every point of the ball in the upper half-space has \(\rho\le2\) and a polar angle in \([0,\pi/2]\). The mistake was not in the radial bound; it was in using an angular range that covered more directions than the region.
Why the Coordinate Match Works
The method is effective because each coordinate system builds certain shapes directly into its coordinates. In cylindrical coordinates, fixing \(r\) gives a vertical circular cylinder, and fixing \(z\) gives a horizontal plane. In spherical coordinates, fixing \(\rho\) gives a sphere centered at the origin, while fixing \(\phi\) gives a cone around the \(z\)-axis. Cartesian coordinates describe coordinate planes by fixing one coordinate.
When a boundary is one of these coordinate surfaces, its equation becomes a constant bound. This is exactly why the ball in the first example has the simple spherical bound \(\rho\le3\), and why the vertical cylinder in the third example has the simple cylindrical bound \(r\le2\). A mismatch does not make a description impossible; it usually makes one or more bounds depend on other coordinates.
Check Your Understanding
For each question, consider both which coordinates simplify the boundaries and how you would verify the proposed bounds.
- Which coordinate system gives the simplest bounds for the ball \(x^2+y^2+z^2\le16\), and what are those bounds?
- Why are cylindrical coordinates a natural choice for a vertical cylinder between two horizontal planes?
- In spherical coordinates, what condition on \(\phi\) describes the upper half-space \(z\ge0\)?
- For the region inside \(\rho\le3\) and within the cone \(0\le\phi\le\pi/6\), which bounds vary and which remain constant?
- What should be checked before concluding that a set of coordinate bounds describes exactly the intended region?