Ma · Mathematics

How to Find a Tangent Plane and Normal Vector

Published By the 9to5Science editorial team

At a glance

  1. 1. Identify the surface

    Write the surface as an implicit equation F(x,y,z)=0 or as a parametrized surface.

  2. 2. Evaluate derivatives

    Compute the gradient, or take the cross product of the parametrization’s tangent vectors.

  3. 3. Find the normal

    Use the resulting vector as a normal vector at the specified point.

  4. 4. Write the plane

    Substitute the point and normal vector into the point-normal plane equation.

A tangent plane is the plane that best approximates a smooth surface near a specified point. Its normal vector is perpendicular to every direction that lies in that plane.

The method depends on how the surface is given. For an implicit equation, use the gradient. For a parametrized surface, use a cross product of tangent vectors.

The point-normal equation of a plane

If a plane contains the point (x0,y0,z0)(x_0,y_0,z_0) and has normal vector n=(a,b,c)\mathbf{n}=(a,b,c), its equation is

a(x−x0)+b(y−y0)+c(z−z0)=0.a(x-x_0)+b(y-y_0)+c(z-z_0)=0.

Thus, once you know a point on the surface and one normal vector there, the tangent-plane calculation is immediate.

A normal vector is not unique: every nonzero scalar multiple of a normal vector is perpendicular to the same plane and defines the same plane. For example, (1,2,3)(1,2,3) and (−2,−4,−6)(-2,-4,-6) define the same plane.

Implicit surfaces: use the gradient

Suppose the surface is defined by

F(x,y,z)=0.F(x,y,z)=0.

At a smooth point P=(x0,y0,z0)P=(x_0,y_0,z_0), the gradient is

∇F(x0,y0,z0)=(Fx(x0,y0,z0),Fy(x0,y0,z0),Fz(x0,y0,z0)).\nabla F(x_0,y_0,z_0)=\left(F_x(x_0,y_0,z_0),F_y(x_0,y_0,z_0),F_z(x_0,y_0,z_0)\right).

If this gradient is nonzero, it is normal to the surface at PP. Therefore the tangent plane is

Fx(x0,y0,z0)(x−x0)+Fy(x0,y0,z0)(y−y0)+Fz(x0,y0,z0)(z−z0)=0.F_x(x_0,y_0,z_0)(x-x_0)+F_y(x_0,y_0,z_0)(y-y_0)+F_z(x_0,y_0,z_0)(z-z_0)=0.

The condition ∇F(P)≠0\nabla F(P)\ne \mathbf{0} matters. If the gradient is zero, the standard gradient formula does not directly determine a unique tangent plane; the point may be singular or require a separate analysis.

Worked example: a sphere

Find the tangent plane and a normal vector to the sphere

x2+y2+z2=14x^2+y^2+z^2=14

at P=(1,2,3)P=(1,2,3).

First write the equation as F(x,y,z)=0F(x,y,z)=0:

F(x,y,z)=x2+y2+z2−14.F(x,y,z)=x^2+y^2+z^2-14.

Compute the gradient:

∇F=(2x,2y,2z).\nabla F=(2x,2y,2z).

At PP,

∇F(1,2,3)=(2,4,6).\nabla F(1,2,3)=(2,4,6).

So (2,4,6)(2,4,6) is a normal vector. You may simplify it to (1,2,3)(1,2,3), since both vectors have the same direction.

Using (1,2,3)(1,2,3) in the point-normal equation gives

(x−1)+2(y−2)+3(z−3)=0.(x-1)+2(y-2)+3(z-3)=0.

Expanding,

x+2y+3z=14.x+2y+3z=14.

Therefore, the tangent plane is

x+2y+3z=14\boxed{x+2y+3z=14}

and one normal vector is

n=(1,2,3).\boxed{\mathbf{n}=(1,2,3)}.

The point really lies on the sphere because 12+22+32=141^2+2^2+3^2=14, which is a useful check before beginning.

Graphs of functions: a convenient formula

If the surface is given as a graph

z=f(x,y),z=f(x,y),

rewrite it as

F(x,y,z)=f(x,y)−z=0.F(x,y,z)=f(x,y)-z=0.

Then

∇F=(fx,fy,−1).\nabla F=(f_x,f_y,-1).

At (x0,y0,z0)(x_0,y_0,z_0), where z0=f(x0,y0)z_0=f(x_0,y_0), a normal vector is (fx(x0,y0),fy(x0,y0),−1)(f_x(x_0,y_0),f_y(x_0,y_0),-1). The tangent plane can also be written directly as

z−z0=fx(x0,y0)(x−x0)+fy(x0,y0)(y−y0).z-z_0=f_x(x_0,y_0)(x-x_0)+f_y(x_0,y_0)(y-y_0).

Worked example: a paraboloid

Let

z=x2+xy+y2,z=x^2+xy+y^2,

and find the tangent plane at (1,1,3)(1,1,3). The point is on the surface because 12+1⋅1+12=31^2+1\cdot1+1^2=3.

Compute the partial derivatives:

fx=2x+y,fy=x+2y.f_x=2x+y,\qquad f_y=x+2y.

At (1,1)(1,1),

fx(1,1)=3,fy(1,1)=3.f_x(1,1)=3,\qquad f_y(1,1)=3.

The tangent-plane formula gives

z−3=3(x−1)+3(y−1).z-3=3(x-1)+3(y-1).

After simplifying,

z=3x+3y−3.\boxed{z=3x+3y-3}.

A corresponding normal vector is (3,3,−1)(3,3,-1). Multiplying it by −1-1 gives the equally valid normal vector (−3,−3,1)(-3,-3,1).

Parametrized surfaces: cross two tangent vectors

Suppose a surface is parametrized by

r(u,v)=(x(u,v),y(u,v),z(u,v)).\mathbf{r}(u,v)=\bigl(x(u,v),y(u,v),z(u,v)\bigr).

The vectors

ru=∂r∂u,rv=∂r∂v\mathbf{r}_u=\frac{\partial \mathbf{r}}{\partial u},\qquad \mathbf{r}_v=\frac{\partial \mathbf{r}}{\partial v}

lie in the tangent plane. Their cross product is perpendicular to both, so

n=ru×rv\mathbf{n}=\mathbf{r}_u\times\mathbf{r}_v

is a normal vector, provided the cross product is nonzero.

For example, consider

r(u,v)=(u,v,u2+v2).\mathbf{r}(u,v)=(u,v,u^2+v^2).

At (u,v)=(1,2)(u,v)=(1,2), the surface point is P=(1,2,5)P=(1,2,5). The tangent vectors are

ru=(1,0,2u),rv=(0,1,2v).\mathbf{r}_u=(1,0,2u),\qquad \mathbf{r}_v=(0,1,2v).

At (1,2)(1,2),

ru=(1,0,2),rv=(0,1,4).\mathbf{r}_u=(1,0,2),\qquad \mathbf{r}_v=(0,1,4).

Their cross product is

ru×rv=(−2,−4,1).\mathbf{r}_u\times\mathbf{r}_v=(-2,-4,1).

Hence the tangent plane is

−2(x−1)−4(y−2)+(z−5)=0,-2(x-1)-4(y-2)+(z-5)=0,

or

z=2x+4y−5.\boxed{z=2x+4y-5}.

Common mistakes

  • Using the point as the normal vector automatically. This works for a sphere centered at the origin, but not for a general surface.
  • Forgetting to move everything to one side. The gradient method requires F(x,y,z)=0F(x,y,z)=0.
  • Evaluating derivatives at the wrong coordinates. For z=f(x,y)z=f(x,y), evaluate partial derivatives at (x0,y0)(x_0,y_0), not at z0z_0.
  • Using a tangent vector as a normal vector. A tangent vector lies in the tangent plane; a normal vector is perpendicular to it.
  • Skipping the point check. Verify that the given point satisfies the surface equation before calculating the plane.
  • Accepting a zero normal vector. A normal vector must be nonzero. If a gradient or cross product vanishes, the usual formula does not provide a regular tangent plane.

References

  1. Calculus: Early Transcendentals, James Stewart
  2. Calculus, Michael Spivak

Spotted an error? Send a correction. Read how we write and check articles in our editorial policy.