Integral of secant function

BigGlenntheHeavy

Senior Member
Joined
Mar 8, 2009
Messages
1,577
Speaking of the secant, sec(x)dx = lnsec(x)+tan(x)+C, Prove it.\displaystyle Speaking \ of \ the \ secant, \ \int sec(x)dx \ = \ ln|sec(x)+tan(x)|+C, \ Prove \ it.
 
sec(x)dx\displaystyle \int sec(x)dx

sec(x)(sec(x)+tan(x)sec(x)+tan(x))dx\displaystyle \int sec(x)\cdot\left(\frac{sec(x)+tan(x)}{sec(x)+tan(x)}\right)dx

=sec2(x)+sec(x)tan(x)sec(x)+tan(x)dx\displaystyle =\int\frac{sec^{2}(x)+sec(x)tan(x)}{sec(x)+tan(x)}dx

Let u=sec(x)+tan(x),   du=(sec2(x)+sec(x)tan(x))dx\displaystyle u=sec(x)+tan(x), \;\ du=(sec^{2}(x)+sec(x)tan(x))dx

Then, we get:

1udu\displaystyle \int\frac{1}{u}du

=ln(u)+C\displaystyle =ln(u)+C

Resub:

ln(sec(x)+tan(x))+C\displaystyle \boxed{ln(sec(x)+tan(x))+C}

Of course, for the purists, we can throw in the absolute value. :D
 
BigGlenntheHeavy said:
Speaking of the secant, sec(x)dx = lnsec(x)+tan(x)+C, Prove it.\displaystyle Speaking \ of \ the \ secant, \ \int sec(x)dx \ = \ ln|sec(x)+tan(x)|+C, \ Prove \ it.

It is very difficult (I would say impossible) to deduce that we should multiply by [sec(x)+tan(x)]/[sec(x)+tan(x)] without knowing the expected final result. Of course, in this case you have given the final result - you wanted to prove an identity. So we can attack the RHS and go:

d/dx [ln{|sec(x)+tan(x)}] = 1/[sec(x)+tan(x)] * [sec(x)*tan(x) + sec[sup:1q83zvwe]2[/sup:1q83zvwe](x)] = sec(x) ..... Q.E.D.
 
The previous method is the one usually shown in calc texts. It just uses a clever substitution to hammer it into something we can work with.

Here is another way I just thought up

sec(x)dx\displaystyle \int sec(x)dx

1cos(x)dx\displaystyle \int\frac{1}{cos(x)}dx

Multiply top and bottom by cos(x):

=cos(x)cos2(x)dx\displaystyle =\int\frac{cos(x)}{cos^{2}(x)}dx

=cos(x)1sin2(x)dx\displaystyle =\int\frac{cos(x)}{1-sin^{2}(x)}dx

Now, let u=sin(x),   du=cos(x)dx\displaystyle u=sin(x),\;\ du=cos(x)dx

11u2du\displaystyle \int\frac{1}{1-u^{2}}du

12[11+u+11u]du\displaystyle \frac{1}{2}\left[\int\frac{1}{1+u}+\int\frac{1}{1-u}\right]du

12[1u+11u1]du\displaystyle \frac{1}{2}\left[\int\frac{1}{u+1}-\int\frac{1}{u-1}\right]du

=12ln(u+1u1)\displaystyle =\frac{1}{2}ln(\frac{u+1}{u-1})

12ln(sin(x)+1sin(x)1)\displaystyle \frac{1}{2}ln(\frac{sin(x)+1}{sin(x)-1})

Here, I thought I was stuck. But hammering away at the identity, it does indeed work.

Continuing:

sin(x)+1sin(x)1sin(x)+1sin(x)+1=(sin(x)+1)2sin2(x)1=(sin(x)+1)2cos2(x)\displaystyle \frac{sin(x)+1}{sin(x)-1}\cdot\frac{sin(x)+1}{sin(x)+1}=\frac{(sin(x)+1)^{2}}{sin^{2}(x)-1}=\frac{(sin(x)+1)^{2}}{-cos^{2}(x)}

Because of the absolute value, we can shed the negative in the denominator in cos2(x)\displaystyle -cos^{2}(x) and write:

12ln((sin(x)+1)2cos2(x))\displaystyle \frac{1}{2}ln\left(\frac{(sin(x)+1)^{2}}{cos^{2}(x)}\right)

Property of logs:

=ln(sin(x)+1cos(x))=ln(tan(x)+sec(x))\displaystyle =ln\left(\frac{sin(x)+1}{cos(x)}\right)=ln(tan(x)+sec(x))

Is that acceptable, SK?. :D

I was thinkin', perhaps another way would be to note that sec(x)=2eix+eix\displaystyle sec(x)=\frac{2}{e^{ix}+e^{-ix}} and use complex integration.
 
Subhotosh Khan, I'm curious, what does this proof have to do with the Oxford English Dictionary (O.E.D.)?
 
BigGlenntheHeavy said:
Subhotosh Khan, I'm curious, what does this proof have to do with the Oxford English Dictionary (OED)?

Actually it is Q[the letter that follows P].E.D.

which comes from (Latin) quod erat demonstrandum - which translates to "which was to be demonstrated".

Curiously, translated to Greek - Q.E.D. - does become OED (???? ???? ?????? ), according to my Greek friend.
 
Thank you Subhotosh Khan, I figured it probably was a typo, but then one never knows unless one asks.

By the way galactus, good show on your second proof - never saw that one before.
 
Top