StudyDeck

Chain Rule

Exam code: 4037
Written by: Ashika|Reviewed by: Caroline Carroll|Updated 2 July 2026

Chain Rulevideo

Chain Rule

Chain rule

What is the chain rule?

  •  The chain rule states if y{"language":"en","fontFamily":"Times New Roman","fontSize":"18"} is a function of u{"language":"en","fontFamily":"Times New Roman","fontSize":"18"} and u{"language":"en","fontFamily":"Times New Roman","fontSize":"18"} is a function of x{"language":"en","fontFamily":"Times New Roman","fontSize":"18"} then

 y=f(u(x)){"language":"en","fontFamily":"Times New Roman","fontSize":"18","autoformat":true}

 dydx=dydu×dudx{"language":"en","fontFamily":"Times New Roman","fontSize":"18"}

  • In function notation this could be written

 y=f(g(x)){"language":"en","fontFamily":"Times New Roman","fontSize":"18","autoformat":true}

 dydx=f'(g(x))g'(x){"language":"en","fontFamily":"Times New Roman","fontSize":"18","autoformat":true}

How do I know when to use the chain rule?

  •  The chain rule is used when we are trying to differentiate composite functions

    • “function of a function”

    • these can be identified as the variable (usually x{"language":"en","fontFamily":"Times New Roman","fontSize":"18"}) does not ‘appear alone’

      •  sin x{"language":"en","fontFamily":"Times New Roman","fontSize":"18"} – not a composite function, x{"language":"en","fontFamily":"Times New Roman","fontSize":"18"} ‘appears alone’

      • sin(3x+2){"language":"en","fontFamily":"Times New Roman","fontSize":"18"} is a composite function; x{"language":"en","fontFamily":"Times New Roman","fontSize":"18"} is tripled and has 2 added to it before the sine function is applied

How do I use the chain rule?

 STEP 1

 Identify the two functions

 Rewrite y{"language":"en","fontFamily":"Times New Roman","fontSize":"18"} as a function of u{"language":"en","fontFamily":"Times New Roman","fontSize":"18"};  y=f(u){"language":"en","fontFamily":"Times New Roman","fontSize":"18","autoformat":true}

 Write u{"language":"en","fontFamily":"Times New Roman","fontSize":"18"} as a function of x{"language":"en","fontFamily":"Times New Roman","fontSize":"18"};  u=g(x){"language":"en","fontFamily":"Times New Roman","fontSize":"18","autoformat":true}

 STEP 2

Differentiate y{"language":"en","fontFamily":"Times New Roman","fontSize":"18"} with respect to u{"language":"en","fontFamily":"Times New Roman","fontSize":"18"} to get dydu{"language":"en","fontFamily":"Times New Roman","fontSize":"18"} Differentiate u{"language":"en","fontFamily":"Times New Roman","fontSize":"18"} with respect to x{"language":"en","fontFamily":"Times New Roman","fontSize":"18"} to get dudx{"language":"en","fontFamily":"Times New Roman","fontSize":"18"}

 STEP 3

Obtain dydx{"language":"en","fontFamily":"Times New Roman","fontSize":"18"} by applying the formula dydx=dydu×dudx{"language":"en","fontFamily":"Times New Roman","fontSize":"18"} and substitute u{"language":"en","fontFamily":"Times New Roman","fontSize":"18"} back in for g(x){"language":"en","fontFamily":"Times New Roman","fontSize":"18","autoformat":true}

 

  • In trickier problems chain rule may have to be applied more than once

How do I differentiate (ax + b)n?

  • For n = 2 you will most likely expand the brackets and differentiate each term separately

  • If n > 2 this becomes time-consuming and if n is not a positive integer we need a different method completely

  • The chain rule allows us to use substitution to differentiate any function in the form y = (ax + b)n

    • Let u = ax + b, then y = un

    • Differentiate both parts separately

      • dudx=a{"language":"en","fontFamily":"Times New Roman","fontSize":"18"} and dydu=nun-1{"language":"en","fontFamily":"Times New Roman","fontSize":"18"}

    • Put both parts into the chain rule

      • dydx= dydu × dudx=a × nun-1 = anun-1 {"language":"en","fontFamily":"Times New Roman","fontSize":"18"}

    • Substitute u = ax + b back into your answer

      • dydx=an(ax+ b)n-1{"language":"en","fontFamily":"Times New Roman","fontSize":"18"}

How do I differentiate √(ax+b)?

  • The chain rule allows us to use substitution to differentiate any function in the form y=ax+b{"language":"en","fontFamily":"Times New Roman","fontSize":"18"}

  • Rewrite ax+b=(ax+b)12{"language":"en","fontFamily":"Times New Roman","fontSize":"18"} 

    • Let u = ax + b, then y = u½

    • Differentiate both parts separately

      • dudx=a{"language":"en","fontFamily":"Times New Roman","fontSize":"18"} and dydu=12u-12{"language":"en","fontFamily":"Times New Roman","fontSize":"18"}

    • Put both parts into the chain rule

      • dydx= dydu × dudx=a × 12u-12 = a2u-12 {"language":"en","fontFamily":"Times New Roman","fontSize":"18"}

    • Substitute u = ax + b back into your answer

      • dydx= a2ax+b-12 =a2ax+b{"language":"en","fontFamily":"Times New Roman","fontSize":"18"}

  • This method can be used for any fractional power of any linear or non-linear expression

    • Provided you know how to differentiate the non-linear expression

Are there any standard results for using chain rule?

  • The following general results are particularly useful

    • If y=(fx)n{"language":"en","fontFamily":"Times New Roman","fontSize":"18","autoformat":true} then dydx=nf'(x)f(x)n-1{"language":"en","fontFamily":"Times New Roman","fontSize":"18"}

  • If y=e f(x){"language":"en","fontFamily":"Times New Roman","fontSize":"18","autoformat":true} then dydx=f'(x)e f(x){"language":"en","fontFamily":"Times New Roman","fontSize":"18","autoformat":true}

    • If y=ln(fx){"language":"en","fontFamily":"Times New Roman","fontSize":"18","autoformat":true} then dydx=f'(x)f(x){"language":"en","fontFamily":"Times New Roman","fontSize":"18","autoformat":true}

    • If y=sin(fx){"language":"en","fontFamily":"Times New Roman","fontSize":"18","autoformat":true} then dydx=f'(x)cos(fx){"language":"en","fontFamily":"Times New Roman","fontSize":"18","autoformat":true}

    • If y=cos(fx){"language":"en","fontFamily":"Times New Roman","fontSize":"18","autoformat":true} then dydx=-f'(x)sin(fx){"language":"en","fontFamily":"Times New Roman","fontSize":"18","autoformat":true}

    • If y=tan(fx){"language":"en","fontFamily":"Times New Roman","fontSize":"18","autoformat":true} then dydx=f'(x)sec2(fx){"language":"en","fontFamily":"Times New Roman","fontSize":"18","autoformat":true}