by Dr.

OMAR HAZEM MOHAMMED

NTU, Technical of Eng. College Mosul
HIGER − ORDER PARTIAL DERIVATIVES

We have seen that if y = f(x), then

df d2 f d df
y′ = and y ′′ = 2 = ( )
dx dx dx dx
That is ,the second derivative of f is the derivative of the first derivative of f. , if z = f (x, y ),
then we can differentiate each of the two ''first'' partial derivatives∂ fl ∂ x and∂ fl ∂y with
respect to both x and y to obtain four second partial derivatives as follows :

Definition 1: SECOND PARTIAL DERIVATIVES

(i) Differentiate twice with respect to x :

∂2 z 𝜕2𝑓 𝜕 𝜕ƒ
= = ƒ 𝑥𝑥 = ( ) (1)
∂x 2 𝜕𝑥 2 𝜕𝑥 𝜕𝑥

(ii ) Differentiate first with respect to x and then with respect to y :

∂2 z 𝜕2𝑓 𝜕 𝜕ƒ
= = ƒ𝑥𝑦 = ( ) (2)
∂y ∂x ∂y ∂x 𝜕𝑦 𝜕𝑥

(iii ) Differentiate first with respect to y and then with respect to x :

∂2 z 𝜕2𝑓 𝜕 𝜕ƒ
= = ƒ𝑦𝑥 = ( ) (3)
∂x ∂y ∂x ∂y 𝜕𝑥 𝜕𝑦

(iv ) Differentiate twice with respect to y :

∂2 z 𝜕2𝑓 𝜕 𝜕ƒ
2
= = ƒ𝑦𝑦 = ( ) (4)
∂y 𝜕𝑦 2 𝜕𝑦 𝜕𝑦

REMARK1.Thederivatives ∂2 ƒl∂ x ∂y and ∂2ƒ∂ y ∂x arecalled the mixed second partialsl.

REMARK2. It is much easier to denote the second partials byƒxx , ƒxx , ƒyx andƒyy ,

1
 by Dr. OMAR HAZEM MOHAMMED
NTU, Technical of Eng. College Mosul

We will there f are use this notation for the remainder of this section. Note that the symbol ƒxy
indicates that we differentiate first with respect to y .

EXAMPLE 1 Let z =ƒ(x,y) =x 3 y 2 − xy 5 .Calculate the four second partial derivatives .

Solution. We have fx = 3x2 y2 − y5 and fy =2x3 y − 5xy4 .

(a) f ∂
xx = (f ) = 6xy2
∂× x
(b) f ∂
yx = (f ) = 6x2 y− 5y2
∂y x

2
(c) f ∂ 6x y − 5y 2
xy = (f ) =
∂x y

(d) f ∂
yy = (f ) = 2x3 − 20xy3
∂y y

In Example 1 we saw thatfxy = fyx This result is no accident ,as we see by the following
theorem whose proof can be found in any intermediate calculus text .t

Theorem 1 : Suppose that f, fx , fy , fxy and fyx are all continuous at (x0 , y0 ) Then

𝑓𝑥𝑦 (𝑥0 , 𝑦0 ) = 𝑓𝑦𝑥 (𝑥0 , 𝑦0 ) (5)

This result is often referred to as the equality of mixed partials ≠0

The definition of second partial derivatives and the theorem on the equality of mixed partials are
easily extended to functions of three variables. If w =( x ,y , z ), then we have the nine second
partial derivatives ( assuming that they exist ) :

∂2 f ∂2 f ∂2 f
= fxx , = fxy , = fxz ,
∂x 2 ∂y ∂x ∂z ∂x

∂2 f ∂2 f ∂2 f
= fyx , 2 = fyy , = fyz
∂x ∂y ∂y ∂z ∂y

∂2 f ∂2 f ∂2 f
= fzx , = fzy , = fzz .
∂x ∂z ∂y ∂z ∂z 2

Theorem2 If f, fx , fy , fz and fyx and all six mixed partial are continuous at a point
(x0 , y0 , z0 )then at a point This theorem was first stated by Euler in a 1734 paper devoted to
a problem in hydrodynamics

𝑓𝑥𝑦 = 𝑓𝑦𝑥 , 𝑓𝑥𝑧 = 𝑓𝑧𝑥 , 𝑓𝑦𝑧 = 𝑓𝑧𝑦 .

2
 by Dr. OMAR HAZEM MOHAMMED
NTU, Technical of Eng. College Mosul

EXAMPLE 2 : Let 𝑓(𝑥, 𝑦, 𝑧) = 𝑥𝑦 3 − 𝑧𝑥 5 + 𝑥 2 𝑦𝑧 be a function , Calculate all for nine

second partial derivatives and show that all three pairs of mixed partials are equal

Solution : We have

fx = y 3 − 5zx 4 + 2xyz ,

fy = 3xy 2 + x 2 z ,

and

fz = − x 5 + x 2 y

Then

fxx = −20zx 3 + 2yz , fyy = 6xy , fzz = 0 ,

∂ 3
fxy = (y − 5zx 4 + 2xyz ) = 3y 2 + 2xz ,
∂y

∂
fyx = (3xy 2 + x 2 z ) = 3y 2 + 2xz ,
∂x
∂ 3
fxz = (y − 5zx 4 + 2xyz ) = − 5x 4 + 2xy ,
∂z
∂
fzx = ( − x 5 + x 2 y ) = −5 x 4 + 2xy ,
∂x
∂
fyz = (3 xy 2 + x 2 z ) = x 2 ,
∂z
∂
fzy = ( − x5 + x2 y ) = x2
∂y

We conclude this section by pointing out that we can easily define partial derivatives of orders
higher than two . For example,

∂3 f ∂ ∂2 f ∂
fzyx = = ( ) = (f )
∂x ∂y ∂z ∂x ∂y ∂z ∂x zy

EXAMPLE 3 Calculate and for the function of Example 2 .

Solution We easily obtain the three third partial derivatives:

3
 by Dr. OMAR HAZEM MOHAMMED
NTU, Technical of Eng. College Mosul

∂ ∂
fxxx = (fxx ) = (20zx 3 + 2yz ) = −60zx 2
∂x ∂x
∂ ∂
fzzy = (fxz ) = (5x 4 + 2yz ) = 2x
∂y ∂y

∂ ∂
fyxz = (fyx ) = (3y 2 + 2xz ) = 2x
∂z ∂z
Note that fxzy = fyxz This again is no accident and follows from the generalization of Theorem 2
to mixed third partial derivatives. Finally, the fourth partial derivative fyxzx is given by

∂ ∂
fyxzx = (fyxz ) = ( 2x) = 2.
∂x ∂x
PROBLEMS

In problems 1-12 , calculate the four second partial derivatives and show that the mixed partials
are equal .

1 . f ( x , y ) = x 2 y.

. f ( x , y ) = xy 2 y.2

3 . f (x ,y ) = 3exy3

4. f ( x , y ) = sin (x 2 + y 3 )
4x
5 . f (x ,y ) =y5

6 . f ( x , y ) = ey tanx .

7 . f (x ,y ) =In ( x 3 y 5 − 2 )

8 . f ( x , y ) =√xy + 2y 3

9. f (x ,y ) = (x + 5y sin x )‫اكتبالمعادلةهنا‬.

10 . f ( x , y ) = sinh ( 2x - y )
x2 −y2
11 . f (x ,y ) = sin−1 (x2 + y2 )

12 . f ( x , y ) =sec x y

4
 by Dr. OMAR HAZEM MOHAMMED
NTU, Technical of Eng. College Mosul

In Problems 13 -21 , calculate the nine second partial derivatives and show that the three pairs of
mixed partials are equal

13 . f ( x , y , z ) = xyz

14 . f ( x , y , z ) = x 2 y 3 z 4
x+ y
15 . f ( x , y , z ) = z

16 . f ( x , y , z ) = sin ( x + 2y +z 2 )
xz
17. f ( x , y , z ) = tan−1
y

18 . f ( x , y , z ) = cos xyz

19 . f ( x , y , z ) =e3xy cosz

20 . f ( x , y , z ) = ln ( xy + z)

21 . f ( x , y , z ) = cosh√x + yz

22 . How many third partial derivatives are there for a function of (a) two variables; (b) three
variables?

23 . How many fourth partial derivatives are there for a function of (a) two variables; (b) three
variables?

24 . How many nth partial derivatives are there for a function of (a) two variables; (b) three
variables ?

In Problems 25 -30 , calculate the given partial derivative

25 . f ( x, y ) = x 2 y 3 + 2y: fxyx

26 . f ( x, y ) = sin (2 xy 4 ) ; fxyt

27 . f ( x, y ) = In (3x − 2y ); fyxy

28 . f ( x, y, z ) = x 2 y + y 2 z − 3 √xz ; fxyz

29. f ( x, y, z ) = cos(x + 2y + 3z ) ; fzzx

30 f ( x, y, z ) = exy sin z ; fzxyx .

5
 by Dr. OMAR HAZEM MOHAMMED
NTU, Technical of Eng. College Mosul

THE CHAIN RULE

In this section we derive the chain rule for function s of two and three variables Let us recall
the chain rule for the composition of two f unctions ofone variable ∶

Let y = f(u)and u = g (x)and assme that f and g are differentiable. Then

dy dy du
= = f ′ (g(x))g ′ (x)(1)
dx du dx
If z =f (x, y) is a function of two variables , then there are two versions of the chain rule

Theorem 1 CHAIN RULE Let z =f (x, y) be differentiable and suppose that x =x (t) and y=
y(t). Assume further thatdx/dt anddy/dt exist and are continuous Then z can be written as a
function of the parameter t, and

𝑑𝑧 𝜕𝑧 𝑑𝑥 𝜕𝑧 𝑑𝑦 𝑑𝑥 𝑑𝑦
= + = 𝑓𝑥 + (2)
𝑑𝑡 𝜕𝑥 𝑑𝑡 𝜕𝑦 𝑑𝑡 𝑑𝑡 𝑑𝑡

We can also write this result using our gradient If g(t) =x (t) I +y(t)j, then g'(t)= (dx/dt)i +
(dy/dt)j, and (2)can be written as
𝑑 Theorem 2 ′ (𝑡) = [𝑓(𝑔(𝑡))]′ = ∇𝑓 ∙ 𝑔′ (𝑡)
𝑓(𝑥(𝑡), 𝑦(𝑡)) = (𝑓 °𝑔) (3)
𝑑𝑡

Theorem 2 CHAIN RULE Let z =f (x, y) be differentiable and suppose that x and y are
function of the two variables r and s That is , x= x(r, s) and y =(r, s)

Suppose further that ∂x/ ∂r , ∂x/ ∂s , ∂y/ ∂r and ∂y/ ∂s all exist and are continuous . Then
z can be written as a function of r and s, and

𝜕𝑧 𝜕𝑧 𝜕𝑥 𝜕𝑧 𝜕𝑦
= + (4)
𝜕𝑟 𝜕𝑥 𝜕𝑟 𝜕𝑦 𝜕𝑟

𝜕𝑧 𝜕𝑧 𝜕𝑥 𝜕𝑧 𝜕𝑦
= + (5)
𝜕𝑠 𝜕𝑥 𝜕𝑠 𝜕𝑦 𝜕𝑠

We will leave the proofs of these theorems until the end this section .

EXAMPLE 1∶ Let z = f(x, y) = xy 2 . Let x = cos t and y = sin t Calculate dz/dt

6
 by Dr. OMAR HAZEM MOHAMMED
NTU, Technical of Eng. College Mosul

Solution.

dz ∂z dx ∂z dy
= + = y 2 (− sin t) + 2xy(cos t)
dt ∂x dt ∂y dt

= (sin 2 t)(− sin t) + 2(cos t (sin t) (cos t)

= 2 sin t cos2 t − sin3 t

We can calculate this result another way . Since z = xy 2 we have z = (cos t)(sin2 t) Then

dz
= (cos t) 2(sin t (cos t) + (sin2 t)(− sin t)
dt
= 2 sin t cos2 t − sin3 t
r
EXAMPLE 2 Let z = f (x, y) = sin xy 2 Suppose that x = s and y = er−s Calculate∂z/
∂r and ∂z/ ∂s

Solution

∂z ∂z ∂x ∂z ∂y 1
= + = (y 2 cos xy 2 ) + (2xy cos xy 2 )er−s
∂r ∂x ∂r ∂y ∂r s
r
e2(r−s) cos[(s)e2(r−s) ] 2r r
= + {cos[ e2(r−s) ] } e2(r−s)
s s s
And

∂z ∂z ∂x ∂z ∂y −r
= + = (y 2 cos xy 2 ) 2 + (2xy cos xy 2 )(−er−s )
∂s ∂x ∂s ∂y ∂s s
r
−re2(r−s) cos[(s)e2(r−s) ] 2r r
= + {cos[ e2(r−s) ] } e2(r−s)
s2 s s
The chain rules given in Theorem 1 and Theorem 2 can easily be extended to functions of three
or more variables .

Theorem 1' Let w= f (x, y , z) be a differentiable function If x = x(t) , y = y (t), z = z (t),

and ifdx/dt, dy/dt, and dz/dtexist and are continuous, then

𝑑𝑤 𝜕𝑤 𝑑𝑥 𝜕𝑤 𝑑𝑦 𝑑𝑤 𝑑𝑧
= + + (6)
𝑑𝑡 𝜕𝑥 𝑑𝑡 𝜕𝑦 𝑑𝑡 𝑑𝑧 𝑑𝑡

7
 by Dr. OMAR HAZEM MOHAMMED
NTU, Technical of Eng. College Mosul

Theorem 2' Let w = f (x, y , z) be a differentiable function and let x= x (r, s), y=y (r, s), and
z=z (r, s) Then if all indicated partial derivatives exist and are continuous, we have

𝜕𝑤 𝜕𝑤 𝜕𝑥 𝜕𝑤 𝜕𝑦 𝜕𝑤 𝜕𝑧
= + + (7)
𝜕𝑥 𝜕𝑥 𝜕𝑟 𝜕𝑦 𝜕𝑟 𝜕𝑧 𝜕𝑟

And

𝜕𝑤 𝜕𝑤 𝜕𝑥 𝜕𝑤 𝜕𝑦 𝜕𝑤 𝜕𝑧
= + + (8)
𝜕𝑠 𝜕𝑥 𝜕𝑠 𝜕𝑦 𝜕𝑠 𝜕𝑧 𝜕𝑠

Theorem 3' Let w = f (x, y , z) be a differentiable function and let x= x (r, s, t), y=y (r, s),
and z=z (r, s, t ) Then if all indicated partial derivatives exist and are continuous, we have

𝜕𝑤 𝜕𝑤 𝜕𝑥 𝜕𝑤 𝜕𝑦 𝜕𝑤 𝜕𝑧
= + +
𝜕𝑥 𝜕𝑥 𝜕𝑟 𝜕𝑦 𝜕𝑟 𝜕𝑧 𝜕𝑟

𝜕𝑤 𝜕𝑤 𝜕𝑥 𝜕𝑤 𝜕𝑦 𝜕𝑤 𝜕𝑧
= + + (9)
𝜕𝑠 𝜕𝑥 𝜕𝑠 𝜕𝑦 𝜕𝑠 𝜕𝑧 𝜕𝑠

𝜕𝑤 𝜕𝑤 𝜕𝑥 𝜕𝑤 𝜕𝑦 𝜕𝑤 𝜕𝑧
= + +
𝜕𝑡 𝜕𝑥 𝜕𝑡 𝜕𝑦 𝜕𝑡 𝜕𝑧 𝜕𝑡

PROBLEMS

In problems 1-11, use the chain rule to calculate dz/dt Check your answer by first writing z or w
as a function of t and then differentiating.

1. z = xy, x = et , y = e2t 2. z = x 2 + y 2 , x = cos t, y = sin t

y 3
3. z = , x = t 2 , y = t 3 4. z = ex sin y, x = √t, y = √t
x
y 5t 6. z = sinh(x − 2y), x = 2t 2 , y = t 2 + 1
5. z = tan−1 , x = cos 3t, y = sin
x

7. w = x 2 + y 2 , z 2 , x = cos t, y = sin t , z = t

8. w = xy − yz + zx, x ≠ et , y = e2t , z = e3t

x+y
9. w = , x = t, y = t 2 , z = t 3
z