Higher-Order Partial Derivatives and Mixed Partial Derivatives

The Definition of Higher-Order Partial Derivatives

Higher-order partial derivative

Suppose the function z=f(x,y)z = f(x, y) has the partial derivatives fx(x,y)f_x(x,y) and fy(x,y)f_y(x,y) on a region DD. If the partial derivatives of these two partial derivatives also exist, they are called the second-order partial derivatives of f(x,y)f(x,y).

A function of two variables has four second-order partial derivatives:

∂∂x(∂z∂x)=∂2z∂x2=fxx\frac{\partial}{\partial x}\left(\frac{\partial z}{\partial x}\right) = \frac{\partial^2 z}{\partial x^2} = f_{xx}

∂∂y(∂z∂x)=∂2z∂x∂y=fxy\frac{\partial}{\partial y}\left(\frac{\partial z}{\partial x}\right) = \frac{\partial^2 z}{\partial x \partial y} = f_{xy}

∂∂x(∂z∂y)=∂2z∂y∂x=fyx\frac{\partial}{\partial x}\left(\frac{\partial z}{\partial y}\right) = \frac{\partial^2 z}{\partial y \partial x} = f_{yx}

∂∂y(∂z∂y)=∂2z∂y2=fyy\frac{\partial}{\partial y}\left(\frac{\partial z}{\partial y}\right) = \frac{\partial^2 z}{\partial y^2} = f_{yy}

where fxyf_{xy} and fyxf_{yx} are called the mixed partial derivatives.

Similarly, third-order, fourth-order and so on partial derivatives can be defined. The second-order and higher-order partial derivatives are collectively called the higher-order partial derivatives.

The Condition for the Mixed Partial Derivatives to be Equal

Theorem1

Clairaut’s theorem (the Schwarz theorem): if the two mixed partial derivatives fxyf_{xy} and fyxf_{yx} of the function z=f(x,y)z = f(x,y) are continuous on a region DD, then throughout that region

fxy=fyxf_{xy} = f_{yx}

that is, the mixed partial derivatives do not depend on the order of differentiation.

For elementary functions the mixed partial derivatives are usually continuous on their domain, so in general fxy=fyxf_{xy} = f_{yx} and you can choose whichever order of differentiation is more convenient.

The Higher-Order Partial Derivatives of a Function of n Variables

For a function of nn variables u=f(x1,x2,…,xn)u = f(x_1, x_2, \ldots, x_n) there are n2n^2 second-order partial derivatives, of which n(n−1)2\frac{n(n-1)}{2} pairs of mixed partial derivatives are equal under the continuity condition.

In particular, a function of three variables u=f(x,y,z)u = f(x,y,z) has nine second-order partial derivatives: fxx,fxy,fxz,fyx,fyy,fyz,fzx,fzy,fzzf_{xx}, f_{xy}, f_{xz}, f_{yx}, f_{yy}, f_{yz}, f_{zx}, f_{zy}, f_{zz}, and under the continuity condition fxy=fyxf_{xy}=f_{yx}, fxz=fzxf_{xz}=f_{zx}, fyz=fzyf_{yz}=f_{zy}.

Worked Examples

Example 1: Computing Second-Order Partial Derivatives

Find all the second-order partial derivatives of z=x3y2−3xy3+xy+1z = x^3 y^2 - 3xy^3 + xy + 1.

second-order partial derivativesmixed partial derivativesa polynomial function

Approach:

A second-order partial derivative is one more differentiation of a first-order partial derivative; all four of them have to be written out, and you should check whether the mixed partial derivatives agree.

Steps:

  1. The first-order partial derivatives: fx=3x2y2−3y3+yf_x = 3x^2 y^2 - 3y^3 + y fy=2x3y−9xy2+xf_y = 2x^3 y - 9xy^2 + x

  2. The second-order partial derivatives: fxx=6xy2f_{xx} = 6xy^2 fxy=6x2y−9y2+1f_{xy} = 6x^2 y - 9y^2 + 1 fyx=6x2y−9y2+1f_{yx} = 6x^2 y - 9y^2 + 1 fyy=2x3−18xyf_{yy} = 2x^3 - 18xy

  3. Checking: fxy=fyxf_{xy} = f_{yx} (a polynomial function, so the mixed partial derivatives are continuous)

Answer: fxx=6xy2f_{xx}=6xy^2, fxy=fyx=6x2y−9y2+1f_{xy}=f_{yx}=6x^2 y-9y^2+1, and fyy=2x3−18xyf_{yy}=2x^3-18xy

Example 2: Second-Order Partial Derivatives of a Composite Function

Let z=f(x2+y2)z = f(x^2 + y^2), where ff has continuous second derivatives. Find ∂2z∂x2\frac{\partial^2 z}{\partial x^2} and ∂2z∂x∂y\frac{\partial^2 z}{\partial x \partial y}.

second-order partial derivativesa composite functionan abstract functionthe chain rule

Approach:

The outer function is composed once, so the second derivative has to keep both the chain rule term of the outer function and the second-order term of the inner function.

Steps:

  1. Put u=x2+y2u = x^2 + y^2, so z=f(u)z = f(u)
  2. The first-order partial derivative: ∂z∂x=f′(u)⋅2x=2xf′(u)\frac{\partial z}{\partial x} = f'(u) \cdot 2x = 2x f'(u)
  3. The second-order partial derivative ∂2z∂x2\frac{\partial^2 z}{\partial x^2}: ∂∂x[2xf′(u)]=2f′(u)+2x⋅f′′(u)⋅2x=2f′(u)+4x2f′′(u)\frac{\partial}{\partial x}[2x f'(u)] = 2f'(u) + 2x \cdot f''(u) \cdot 2x = 2f'(u) + 4x^2 f''(u)
  4. The mixed partial derivative ∂2z∂x∂y\frac{\partial^2 z}{\partial x \partial y}: ∂∂y[2xf′(u)]=2x⋅f′′(u)⋅2y=4xyf′′(u)\frac{\partial}{\partial y}[2x f'(u)] = 2x \cdot f''(u) \cdot 2y = 4xy f''(u)

Answer: ∂2z∂x2=2f′(x2+y2)+4x2f′′(x2+y2)\frac{\partial^2 z}{\partial x^2} = 2f'(x^2+y^2) + 4x^2 f''(x^2+y^2) and ∂2z∂x∂y=4xyf′′(x2+y2)\frac{\partial^2 z}{\partial x \partial y} = 4xy f''(x^2+y^2)

Example 3: Verifying the Laplace Equation

Verify that u=1x2+y2+z2u = \frac{1}{\sqrt{x^2+y^2+z^2}} satisfies the Laplace equation uxx+uyy+uzz=0u_{xx} + u_{yy} + u_{zz} = 0 (for r≠0r \neq 0).

the Laplace equationhigher-order partial derivativesa harmonic functiona function of three variables

Approach:

Compute the second-order partial derivatives one by one and add them; the key is not to miss the chain rule for rr.

Steps:

  1. Put r=x2+y2+z2r = \sqrt{x^2+y^2+z^2}, so u=1ru = \frac{1}{r}
  2. ux=−1r2⋅xr=−xr3u_x = -\frac{1}{r^2} \cdot \frac{x}{r} = -\frac{x}{r^3}
  3. uxx=−1r3−x⋅(−3)r−4⋅xr=−1r3+3x2r5u_{xx} = -\frac{1}{r^3} - x \cdot (-3)r^{-4} \cdot \frac{x}{r} = -\frac{1}{r^3} + \frac{3x^2}{r^5}
  4. By symmetry, uyy=−1r3+3y2r5u_{yy} = -\frac{1}{r^3} + \frac{3y^2}{r^5} and uzz=−1r3+3z2r5u_{zz} = -\frac{1}{r^3} + \frac{3z^2}{r^5}
  5. uxx+uyy+uzz=−3r3+3(x2+y2+z2)r5=−3r3+3r2r5=−3r3+3r3=0u_{xx}+u_{yy}+u_{zz} = -\frac{3}{r^3} + \frac{3(x^2+y^2+z^2)}{r^5} = -\frac{3}{r^3} + \frac{3r^2}{r^5} = -\frac{3}{r^3} + \frac{3}{r^3} = 0

Answer: u=1ru = \frac{1}{r} satisfies the Laplace equation for r≠0r \neq 0; it is a harmonic function (the fundamental solution in three dimensions).


Exercises

Exercise 1

Find all the second-order partial derivatives of z=exyz = e^{xy}.

second-order partial derivativesmixed partial derivativesthe exponential function

Approach:

For the exponential function, differentiate term by term and check that the mixed partial derivatives agree.

Steps:

  1. zx=yexyz_x = ye^{xy} and zy=xexyz_y = xe^{xy}
  2. zxx=y2exyz_{xx} = y^2 e^{xy}
  3. zxy=exy+xyexy=(1+xy)exyz_{xy} = e^{xy} + xye^{xy} = (1+xy)e^{xy}
  4. zyx=exy+xyexy=(1+xy)exyz_{yx} = e^{xy} + xye^{xy} = (1+xy)e^{xy}
  5. zyy=x2exyz_{yy} = x^2 e^{xy}

Answer: zxx=y2exyz_{xx}=y^2e^{xy}, zxy=zyx=(1+xy)exyz_{xy}=z_{yx}=(1+xy)e^{xy}, and zyy=x2exyz_{yy}=x^2e^{xy}

Exercise 2

Let u=f(x,xy)u = f(x, \frac{x}{y}), where ff has continuous second-order partial derivatives. Find ∂2u∂x2\frac{\partial^2 u}{\partial x^2}.

second-order partial derivativesa composite functionan abstract function

Approach:

The inner function has two independent variables, so the second derivative has to expand all the product terms of the chain rule.

Steps:

  1. Put v=xv = x and w=xyw = \frac{x}{y}, so u=f(v,w)u = f(v, w)
  2. Write f1=∂f∂vf_1 = \frac{\partial f}{\partial v}, f2=∂f∂wf_2 = \frac{\partial f}{\partial w}, f11=∂2f∂v2f_{11} = \frac{\partial^2 f}{\partial v^2}, f12=∂2f∂v∂wf_{12} = \frac{\partial^2 f}{\partial v \partial w} and f22=∂2f∂w2f_{22} = \frac{\partial^2 f}{\partial w^2}
  3. ∂u∂x=f1⋅1+f2⋅1y=f1+1yf2\frac{\partial u}{\partial x} = f_1 \cdot 1 + f_2 \cdot \frac{1}{y} = f_1 + \frac{1}{y}f_2
  4. ∂2u∂x2=(f11⋅1+f12⋅1y)+1y(f21⋅1+f22⋅1y)=f11+2yf12+1y2f22\frac{\partial^2 u}{\partial x^2} = (f_{11} \cdot 1 + f_{12} \cdot \frac{1}{y}) + \frac{1}{y}(f_{21} \cdot 1 + f_{22} \cdot \frac{1}{y}) = f_{11} + \frac{2}{y}f_{12} + \frac{1}{y^2}f_{22} (using f12=f21f_{12}=f_{21})

Answer: ∂2u∂x2=f11+2yf12+1y2f22\frac{\partial^2 u}{\partial x^2} = f_{11} + \frac{2}{y}f_{12} + \frac{1}{y^2}f_{22}


Summary

Symbols Used in This Article

SymbolTypePronunciation/ExplanationMeaning in This Article
fxxf_{xx}Second-order partial derivativef double xDifferentiating twice with respect to xx
fxyf_{xy}Mixed partial derivativef x yDifferentiating first with respect to xx, then with respect to yy
fyxf_{yx}Mixed partial derivativef y xDifferentiating first with respect to yy, then with respect to xx
fyyf_{yy}Second-order partial derivativef double yDifferentiating twice with respect to yy
∇2u\nabla^2 uLaplacianLaplacianuxx+uyy+uzzu_{xx}+u_{yy}+u_{zz}

Found an issue with this content?

Tell us what could be improved.