Difficult Limits: Examples and Techniques

Phil Petrocelli, mymathteacheristerrible.com

February 24, 2019

Contents
1 Introduction 1

2 Basics & reminders 1

2.1 Basic limit rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
2.1.1 2-sided and 1-sided limits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
2.1.2 Limit properties for arithmetic performed on functions: the foundations . . . . . . . 2
2.1.3 Limits involving compositions of functions . . . . . . . . . . . . . . . . . . . . . . . . 2

3 Difficult limits: Some issues 3

3.1 Limits at ±∞, substituting ∞ for x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
3.2 Indeterminate forms, or, ”L’ôpital’s Rule is coming, but isn’t here yet” . . . . . . . . . . . . 3
3.3 Rational functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

4 Difficult limits: The techniques 4

4.1 Rephrasing, un-simplifying, or non-obvious tricks that really help . . . . . . . . . . . . . . . 4
4.1.1 Double what you know! Read theorems and formulae forwards and backwards . . . 4
4.2 Rationals: factoring and cancelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
4.3 Rationals: Asymptotes, limits at ±∞, and the keys to dealing with them . . . . . . . . . . 6
4.4 Rationals: dealing with roots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
4.5 Rationals: rational expressions mean, ”divide,” so let’s get to it . . . . . . . . . . . . . . . . 7
4.6 Rationals: the conjugate trick . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
4.7 Roots: making a rational expression(!) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

1 Introduction
I began writing this handout with the intention of it being a quick and dirty approach to difficult limits
problems in introductory Calculus. However, as I began to write about the various subtleties of the topic,
I found the somewhat non-prescript character of the information here to need a bit more commentary than
I originally intended. This handout is intended to slow down for a second, remind the reader that there are
things they already know from 12-ish years of their mathematical career thus far, that it’s still relevant,
and consider what was in between the lines when your instructor covered this stuff in a brief hour or two.
This entire handout is about the intuitive process of doing tricky algebra. Sometimes a change for new
Calculus students, especially those fresh out of high school. Furthermore, the concept of limits is the
foundation upon which all of Calculus is built! Why not explore it a little bit? Please trust that
what I discuss here will be a useful 0th or first step in the solution to many problems in your first 2 years
of Calculus, and what makes this handout seem a little thick is the many lines of steps to solving example
problems.

1
2 Basics & reminders
Let’s begin with a cursory review of limits and limit rules. Generally, this is covered early in your Calculus
textbook. Since this handout is more about difficult limit calculations, I’ll just provide the end results of
limit law proofs, so that we may use them to solve difficult limits problems straight away.

2.1 Basic limit rules

2.1.1 2-sided and 1-sided limits
There are 3 basic ways in which we consider limits:

1. limits as x approaches a from x < a

x → a−

2. limits as x approaches a from x > a

x → a+

3. limits as x approaches a
x→a

Recall that
lim f (x) = L,
x→a

that is, L exists, iff (if, and only if)

lim f (x) = lim f (x) = L

x→a− x→a+

2.1.2 Limit properties for arithmetic performed on functions: the foundations

There are a handful of basic ways to deal with limits of functions that are built out of arithmetic on other
functions:

1.
lim f (x) ± g(x) = lim f (x) ± lim g(x) (1)
x→a x→a x→a

Read as, ”the limit of a sum is the sum of the limits,”

2.
lim f (x) · g(x) = lim f (x) · lim g(x) (2)
x→a x→a x→a

Read as, ”the limit of a product is the product of the limits,”

3.
f (x) limx→a f (x)
lim = (3)
x→a g(x) limx→a g(x)
Read as, ”the limit of a quotient is the quotient of the limits,” and

4. 󰁳 󰁴
n
lim f (x) = n lim f (x), (4)
x→a x→a

Read as, ”the limit of the nth root of a function is the nth root of the limit.”
For even values of n, please note that there is a domain restriction on the nth root, which means
there is ultimately a range restriction on f (x) and a range restriction on limx→a f (x).

2
2.1.3 Limits involving compositions of functions
It’s a good idea to note here that there is a special property of limits that is very useful, since most real-
world functions are not monomials related by basic arithmetic operations as above, but can be thought of
as composite functions, that is, some f (g(x)) or something more complex.

Keep in mind that the reason we even bother to talk about combining functions in these ways at all in
this context is that whenever the argument of a limit is anything other than a monomial (i.e., one term),
the limit laws help us break up the argument into multiple limits that we can solve term-by-term, when
needed. (I say, ”when needed,” because, sometimes, breaking things up makes the problem worse; this will
be covered in this handout.) When we say limx→5 [x + 1] is 6, we are actually breaking up the argument
using the sum rule, above. Specifically, limx→5 [x + 1] = limx→5 [x] + limx→5 [1] = 5 + 1 = 6. Of course, the
shorthand here is that we can use the original form, substitute x = 5 into the expression, then evaluate.
We will often creatively, using tricky algebra, break up the functions that compose the argument of a
limit in order to ”rephrase” the argument in ways that allow us to use algebra and the limit laws to calculate
the limit. Many times in Calculus, there is a simpler problem or two ”hiding out” in what’s presented to
us that allows us to turn a terrible-looking problem into tractable, easily-solvable subproblems based on
first principles. I cannot stress this enough. I find that Calculus, in large part, is a practice of intuition
and inference, more than it is a practice of actual calculations. Tricky limits problems is usually our first
introduction to this kind of idea.

3 Difficult limits: Some issues

What happens when positing limx→a f (x) = f (a) results in something undefined or difficult to deal with?

3.1 Limits at ±∞, substituting ∞ for x

For this topic, we must get very specific about +∞ and −∞ and the roles they play. +∞ and −∞ are not
numbers. There are no calculations we can do on them. They are the upper and lower bounds of R (the
set of all real numbers), placeholders to show that there is no smallest element of nor largest element of R.
The roles of +∞ and −∞ can simply be recalled in the definition of any x being a real number,
as for all x 󰂃 R, x 󰂃 (−∞, +∞). Please recall that when using interval notation, parentheses are used
to show that the endpoints are not included in the interval. Using inequality notation, we can also say
−∞ < x < +∞, for all x 󰂃 R. If you find yourself saying, ”I’ll just substitute ∞ for x, ...,” please take
pause, for it’s not an allowable operation.
Limits as x → ∞ are, however, asking an important question that sometimes Calculus teachers fail
to cover, and that is, ”Does this function have an asymptote?” Asymptotes pop up everywhere, including
rational functions, exponentials, inverse trig functions, etc. When you see a limit as x → ±∞, ask yourself
about asymptotes as you proceed.

3.2 Indeterminate forms, or, ”L’ôpital’s Rule is coming, but isn’t here yet”
Just what are indeterminate forms? What do they look like? Here is a typical list:
0 ∞
, , ∞ ± ∞, 0 × ∞, 00 , 1∞ , ∞0
0 ∞
Please note that these forms are pretty specific. Here are some things that are not indeterminate forms:

1. Things like 1/0 or any n/0 (n ∕= 0), f (x)/0, (f (x) ∕= 0), etc. are not indeterminate.

2. Limits that go to infinity. This still means that our function approaches infinity as x → a (i.e., grows
without bound) and that’s fine. This is the usefulness of the concept of infinity, above.

3
 3. Limits that do not exist (DNE) are not indeterminate forms, either. Please refer to 1.1.1, above.
These are usually sources of confusion for new Calculus students. Please resist categorizing lots of things
as indeterminate forms; I’d say they are kind of rare. More specifically, feeling like you’ve found an indeter-
minate form in your calculations should first prompt a question about your assumptions and calculations,
like, ”Did I do enough algebra here?” Additionally, you should always look at the graph of the function
in question. If the graph appears to do something unexpected according to the statement of the problem,
that’s a strong hint that more algebra can be done to clarify just what is going on.
Here is a simple example. At first blush, using straight substitution, limx→0 xx appears to produce 0/0.
But does it, really? Of course not. Well, mostly not. Please recall some closely-related points about limits.
1. Implicit in all of our problem-solving approaches is consideration of the function only on its
domain. This is one reason why learning how to specify domains for given functions is an important
part of pre-Calculus courses.
2. A corollary here is that we are only ever concerned with the behavior of a function super
super super super super close to the value that x is approaching (sometimes called a, δ-
neighborhood) in the limit we’re evaluating, not at that value. This mostly helps us with the,
”Did I do enough algebra?” question, as we’ll see.
3. Finally, a limit is about where a function is going, not whether it ever gets there. Fur-
thermore, the value that x approaches in a limit need not necessarily be in the domain of f (x). In
the case of, say, polynomials, f (x) does both - it goes to a particular value, and is also able to be
evaluated there, i.e., it gets there. Not every function behaves as straightforwardly as polynomials
do.
In light of these points, let us reconsider limx→0 xx . Recall from basic algebra and considerations involving
rational functions, that a fraction is defined so long as the denominator is not zero. Furthermore, if we can
guarantee that the denominator will never be zero, we can proceed with dividing the numerator by the
denominator. Since we are not interested in the value of xx at zero, we can safely divide the x terms, leaving
us with limx→0 1 = 1. Now we can say we’ve done enough algebra. Please confirm with your graphing
calculator or by hand that the graph of y = x/x is indeed a horizontal line at y = 1 with a hole at (0, 1).

3.3 Rational functions

I would venture to say that lots of the difficult limits problems we see in Calc I involve rational functions.
There are simply too many things to mind, and numerators and denominators always seem to interact in
unexpected ways. Throw a few trig functions in there, and things really begin to get weird sometimes.
This is why a lot of PreCalc courses spend a lot of time on rational functions; they are notoriously tricky
and sometimes things are hiding out in their definitions. With some algebra tricks, however, we can make
lots of improvements and, step by step, deal with complicated-looking functions.
It is worth mentioning here that if we’ve worked with rationals before, we know a couple of things —
limit-related things — about rational functions:
1. Vertical asymptotes are where the denominator of a rational function is zero. Specifically, limx→a f (x),
where a is a vertical asymptote is either ±∞ (function blows up in absolute value in the same direc-
tion on either side of the asymptote) or DNE (function blows up in different directions on either side
of the asymptote). The trick of setting the denominator equal to zero and finding those values of x
that make that true is still the technique here.
2. Horizontal asymptotes are simply limx→±∞ f (x). Techniques like the leading coefficient test for
rationals composed of polynomials, the conjugate trick, polynomial division, factoring and canceling,
as well as rationalizing the denominator all still work here.
In the limits game, we must still take extra care with rationals.

4
4 Difficult limits: The techniques
4.1 Rephrasing, un-simplifying, or non-obvious tricks that really help
The main issue with this entire topic, and with a lot of Calculus (and possibly Mathematics, in general),
is that there is not really a prescriptive technique for calculation. I find that new Calculus students often
have an expectation that there are 1 : 1 or 1 : n, for small n techniques for a particular pattern. This is
unfortunate, and is likely a result of training, thus far, that has focused only on techniques, where the goal
of solving problems on exams is focused on one or two techniques, for ease of understanding the problem,
testing understanding, and, even more likely, easy grading. Calculus is a very different practice.

4.1.1 Double what you know! Read theorems and formulae forwards and backwards
I cannot stress this one enough. We are so used to seeing equations and theorems in only one direction.
Consider a well-known expression: √ √ √
a·b= a· b (5)
No doubt, you’ve used this many times, like:
√ √ √
8= 2· 4
√
= 2·2
√
=2· 2
√
=2 2

√ √
What happens when we want to simplify something like, 2 2 + 24? Can we? Let’s do it, using the
product rule for radicals:
√ √ √ √ √
2 2 + 24 = 2 2 + 2 12
√ √
= 2(2 + 12)
√ √ √
= 2(2 + 4 3)
√ √
= 2(2 + 2 3)
√ √
= 2 2(1 + 3)

Not too bad. But if we read the product rule backwards, we can proceed a little quicker, or just differently,
based on what we might see. Here, if we see 24 as 8 · 3, we might proceed:
√ √ √ √ √ √
2 2 + 24 = 2 2 + 8 3 (factor 24)
√ √ √ √
= ( 2)2 2 + 8 3 (rewrite 2, cleverly)
√ √ √ √ √
= 2 2 2+ 8 3 (expand)
√ √ √
= 8+ 8 3 (use (11), backwards)
√ √
= 8(1 + 3)
√ √
= 2 2(1 + 3)

Of course, this problem is nearly nonsensical. But it does raise a simple point, and that is, algebra
techniques are the core of dealing with tricky problems, and I believe this is one of the cores of doing
calculations many times in Calculus. Another classic is, sin2 x + cos2 x = 1. Of course, whenever we see
sin2 x + cos2 x in an expression, we replace it with 1. But, you may remember from trig proofs, that
sometimes, we replace 1 with sin2 x + cos2 x!

5
4.2 Rationals: factoring and cancelling
x2 − 4
lim
x→2 x − 2

If we take this limit and substitute 2 for x, it would appear that we get an indeterminate form. This should
prompt us to ask if we’ve done enough algebra. There is an additional signal, since this is a rational function
with polynomials: see if we can factor the top and bottom, then see if anything cancels. Additionally, if
possible, graph the function to see if it looks like we expect it to, if it’s weird, or if it nudges us toward a
technique we should use. Begin by factoring the top of the fraction. Of course, x2 − 4 = (x + 2)(x − 2),
via the difference of squares pattern. Rewriting, we have:

x2 − 4 (x + 2)(x − 2)
lim = lim (factored form)
x→2 x − 2 x→2 x−2
= lim (x + 2) (cancelled form)
x→2
= lim x + lim 2 (by (6))
x→2 x→2
=2+2
=4

Quite a bit different than 0/0, yes? Canceling and not worrying about holes, etc. is totally valid here,
keeping in mind that calculating a limit is about where a function is going, rather than if it ever gets there.

4.3 Rationals: Asymptotes, limits at ±∞, and the keys to dealing with them
Earlier in this handout, I drew the parallel between limits at infinity and asymptotes. Here is the one limit
that underlies the way we deal with most limits of rational functions, x → ∞. Recall from algebra, that,
for all f (x) = x1n , where n is a positive integer, f (x) has an asymptote of 0. In terms of limits:
1
lim = 0, n 󰂃 Z, n > 0 (6)
x→∞ xn
This fact, along with the other techniques mentioned in this section, will allow us to handle tough-looking
limits with ease, as we break things down into this form. In fact, when dealing with rationals, we will do
all manner of algebraic manipulations in order to have components of our limits at infinity take on the
form 1/xn . As a thought exercise, think about ways in which we might combine this idea with, say, (9),
above.

4.4 Rationals: dealing with roots

Consider:
x2
lim √
x→∞ x4 + 1
This one is really interesting. The thing that makes it difficult to deal with is that we cannot do anything
with the terms under the radical in the denominator. There simply is no operation we can do in its current
state; a rephrase is needed.
We know this rule for working with fractions and radicals:
󰁵 √
a a
= √ (7)
b b
In words, we can say, ”the square root of a quotient is the quotient of the square roots.” This says that we
have two choices when dealing with that expression: we can divide a by b and take the square root, or we

6
can ”distribute” the radical to the numerator and denominator of the fraction, take the root of each, and
then divide. Reading right to left,
√ 󰁵
a a
√ = (8)
b b
will help us if we are clever about rephrasing the numerator. We need a way to pull the entire fraction under
the radical. If we think about how x2 is the result of having taken a root, as in this classic un-simplifying,
󰁳
x2 = (x2 )2
√
= x4

Then, we can rewrite and work:

√
x2 x4
lim √ = lim √
x→∞ x4 + 1 x→∞ x4 + 1

󰁵
x4
= lim by (8)
x→∞ x4 + 1

󰁵
x4
= lim by (4)
x→∞ x4 + 1

√
= 1 (by the leading coefficient test)

=1

4.5 Rationals: rational expressions mean, ”divide,” so let’s get to it

Consider:
x5 + 3x4 + 1
lim
x→∞ x3 − 1
This one is actually pretty straightforward. I recommend that if you can divide, you should at least try it.
The least it will do is help visualize the graph of the function. One thing I feel like teachers don’t emphasize
P (x)
enough is, for a rational expression f (x) where f (x) = Q(x) , where P (x) and Q(x) are polynomials, if we
actually do the division, f (x) is of the form f (x) = A(x) + R(x), where A(x) is the asymptote as x → ±∞,
and R(x) is the remainder. R(x) is a rational term whose denominator is Q(x). I feel like, for many,
prior to Calculus or more generalized math studies, asymptotes were positioned as lines. Lines are but one
possibility.
After we do the division, our limit becomes:
x5 + 3x4 + 1 2 x2 + 3x + 1
lim = lim [x + 3x + ]
x→∞ x3 − 1 x→∞ x3 − 1

x2 + 3x + 1
= lim [x2 + 3x] + lim
x→∞ x→∞ x3 − 1

=∞+0 (by the leading coeff test)

=∞

7
4.6 Rationals: the conjugate trick
The essence of any use of the conjugate trick in algebra is the difference of squares pattern. Two conjugate
binomials multiply to give the difference of their squares:

(a + b)(a − b) = a2 − b2 (9)

Consider:
√
2− 3+x
lim
x→1 x−1
Straight substitution leads to 0/0, but did we do enough algebra? We really do not want the denominator,
which evaluates to 0 with straight
√ substitution, to sit there
√ alone. We can multiply top and bottom of the
fraction by the conjugate of 2 − 3 + x, which is, 2 + 3 + x:
√ √ √
2− 3+x 2− 3+x 2+ 3+x
lim = lim · √
x→1 x−1 x→1 x−1 2+ 3+x

4 − (3 + x)
= lim √
x→1 (x − 1)(2 + 3 + x)

−(x − 1)
= lim √ (expand and factor the numerator)
x→1 (x − 1)(2 + 3 + x)

−1
= lim √ (cancel terms)
x→1 2 + 3 + x

1
= − lim √
x→1 2+ 3+x

1
=− (evaluate by direct substitution)
4
Quite a bit different than 0/0, yes? Draw the graph of the original function to see that this limit is indeed
correct.
When dealing with expressions that have both polynomial terms and not, in particular, with radicals,
some form of the conjugate trick might help, so definitely try it! It’s hard to see these things hiding out in
some expressions! I feel like I am surprised over and over.

4.7 Roots: making a rational expression(!)

Consider: 󰁳
lim 4x2 + 3x + 2x
x→−∞

This one is super tough. I recommend having a look at the graph, if you can. Here it is, for reference.

8
 5 y

4.5

3.5

2.5

1.5

0.5
x
−10 −9 −8 −7 −6 −5 −4 −3 −2 −1 1
−0.5

−1

−1.5

−2

From the picture, we get the idea that perhaps there is an asymptote √ in the range −1 < y < 0, as
x → −∞. Clear from the picture, but not very obvious from the expression 4x2 + 3x + 2x. Furthermore,
a picture doesn’t help much with algebra. Thinking algebraically, the square root presents a real problem
here. And it’s not obvious how much the +2x term contributes to where the asymptote is. Here is the
same graph, but without the +2x term.

9
 5 y

4.5

3.5

2.5

1.5

0.5
x
−10 −9 −8 −7 −6 −5 −4 −3 −2 −1 1 2 3
−0.5

−1

−1.5

−2

It’s probably easier to reason about the shape of the graph for x > 0, but for x → −∞, it’s tough. In
this case, viewing our limit expression as having potential for the conjugate trick (4.6) helps here. In yet
another un-simplifying move, we create a fraction where there was not one to begin with.
√
󰁳 4x2 + 3x + 2x
lim 2
4x + 3x + 2x = lim
x→−∞ x→−∞ 1

√ √
4x2 + 3x + 2x 4x2 + 3x − 2x
= lim ·√
x→−∞ 1 4x2 + 3x − 2x

4x2 + 3x − 4x2
= lim √
x→−∞ 4x2 + 3x − 2x

x
= 3 · lim √
x→−∞ 4x2 + 3x − 2x

10
Probably doesn’t look much better, I agree, but we could use the technique from ??:
x 1
3 · lim √ = 3 · lim 󰁴
x→−∞ 4x2 + 3x − 2x x→−∞ 2
4x +3x
−2
x2

1
= 3 · lim 󰁴
x→−∞ 3
4+ x −2

1
=3·
0
Problematic. This one may take a moment to see, but there are a couple of non-obvious hints here. The
first is that x → −∞, and the other one is the x2 term. Specifically, when we take the square root at the
very end, the square root has to evaluate to −2, because of the domain restriction of x → −∞, leaving us
with
1 1
3 · lim 󰁴 =3· (by (6), then taking the proper root)
x→−∞
4+ 3
−2 −2 − 2
x
1
=3·
−4
3
=− ,
4
which seems to align with our graph. The radical in the denominator says to evaluate to the positive
square root, but the x → −∞ says that we must evaluate it as negative. Not so much a calculation as it
is an insight. In addition it is the creation of the fraction from an expression that ultimately helped here.

11
Reporting errors and giving feedback
I am so pleased that you have downloaded this study guide and have considered the techniques herein. To
that end, I am the only writer and the only editor of these things, so if you find an error in the text or
calculations, please email me and tell me about it! I am committed to prompt changes when something is
inaccurate. I also really appreciate it when someone takes a moment to tell me how I’m doing with these
sorts of things, so please do so, if you feel inclined.

My email address is: [email protected].

Please visit https://mymathteacheristerrible.com for other study guides. Please tell others about it.

Please donate
I write these study guides with interest in good outcomes for math students and to be a part of the solution.
If you would consider donating a few dollars to me so that these can remain free to everyone who wants
them, please visit my PayPal and pay what you feel this is worth to you. Every little bit helps.

My PayPal URL is: https://paypal.me/philpetrocelli.

Thank you!

12