Coping with Significant Figures

by Patrick Duffy
Last revised: August 30, 1999

1. WHERE DO THEY COME FROM? 2

2. GETTING STARTED 3

2.1 Exact Numbers 3

2.2 Recognizing Significant Figures 3

2.3 Counting Significant Figures 4

3. CALCULATIONS 5

3.1 Rounding Off Numbers: When to do it 5

3.2 Basic Addition and Subtraction 6

3.3 Multiplication and Division 6

3.4 Square Roots 7

3.4.1 What are Square Roots? 7
3.4.2 How to Use Significant Figures With Square Roots 7

3.5 Logarithms 8
3.5.1 What Are Logarithms? 8
3.5.2 How to Use Significant Figures With Logarithms 9

3.6 Using Exact Numbers 9

4. AN EXAMPLE CALCULATION 10

5. CONCLUDING COMMENTS 13

6. PRACTICE PROBLEMS 14
 2

During the course of the chemistry lab this year, you will be introduced (or re-introduced) to the
idea of using significant figures in your calculations. Significant figures have caused some
trouble for students in the past and many marks have been lost as a result of the confusion about
how to deal with them. Hopefully this little primer will help you understand where they come
from, what they mean, and how to deal with them during lab write-ups. When laid out clearly
and applied consistently, they are relatively easy to use. In any event, if you have difficulties I
(or for that matter any member of the chemistry teaching staff at Kwantlen) will be only too
happy to help sort them out with you.

1. Where do they come from?

Nothing in life is perfect (though some things are very good), and so as a result any measurement
made or experiment performed will have some small error or uncertainty associated with it. For
instance, when you read a digital watch to tell the time, you can only see the time to the nearest
second. For that matter, the watch itself may be showing the wrong time. We say in either of
these two cases that only a certain number of meaningful digits can be obtained from the watch.
In the first case, if we cannot read the watch to any better than the nearest second, what
justification do we have in reporting more digits than this (for example, tenths of seconds)? In
the second case, if we have an inaccurate watch (it's out by a minute, say), it would be pointless
reporting the time read from this watch to the nearest second. After all, if the minutes are not
right (and so have little meaning), what reason would there be to justify reporting the seconds,
which would surely have even less meaning?

Similarly, consider the times when we read a thermometer. We could, in principle, take a
thermometer with markings for every degree and read it to many more decimal places than the
markings show. Why would we want to do this, though? If the thermometer showed a
temperature somewhere between 21o and 22o, we might be justified in reading it to 1 decimal
place (by making an estimate of the position of the column of alcohol or mercury between the
two marks), say 21.4o. We would probably not be justified (and very likely not correct, either) to
record the temperature to any more decimal places than that, simply because we cannot read the
thermometer that accurately (there aren't enough markings to help us guess more decimal places)
and very likely the thermometer is not that accurate to begin with, else why not put more
markings on in the first place?

These examples and others like them are the motivation for significant figures. By using them in
our quoted measurements and answers to calculations, we are saying that we only believe the
answer we have obtained up to a certain point. In the case of the watches above, we would only
really be able to quote the time as (say) 7:11:56 in the first case, and 7:12 in the second case. For
the thermometer, we could only report the temperature as 21.4o. To report any more digits for
the measurements above would not be meaningful, since we do not trust them any more than this.

Significant figures and their use in calculations are the very first step to carrying out full error
analysis during the course of the experiments you will perform. Some might suggest that full
error analysis be taught straight away, but I for one disagree. After all, there are enough new
concepts to cope with in the chemistry course and labs. In addition, proper error analysis
involves complex mathematical analysis for which most students do not yet have the tools;
 3

simply giving you the recipes to do it will not teach you any more about error analysis than a
calculator will teach you about how to multiply numbers. I believe that the important thing here
is to first teach students where the errors come in during the course of an experiment, then to
teach them how to propagate them properly once they have a feel for what the errors should be.
Significant figures are an ideal starting point for gaining this understanding.

2. Getting Started

2.1 Exact Numbers

Before we launch into all the finer points of dealing with significant figures, I should mention
that my statement above that nothing in life is perfect isn't quite true. There are numbers which
are known exactly, and these come into significant figure calculations quite often. As an
example, count your ears. Most people have two. How much do we trust this number? Well,
we know that we have exactly two ears, and that we are sure that we have that many and no more
or less (unless, of course, our parents spent too much time touring the uranium enrichment plant
before we were born, in which case there might be some doubt as to the exact number of ears we
do have...). We can therefore place infinite confidence in this number, and as a result we say that
this number is infinitely accurate. In fact, all such counting numbers are infinitely accurate.
This is an important point and will be dealt with again in a later section. Let's drop it for the
moment, though, and move on to numbers which are inexact.

2.2 Recognizing Significant Figures

If the numbers you're dealing with do not arise from counting something, then chances are they're
inexact (at least to some degree), and you'll have to recognize what parts of the number are
significant. This is what this section is all about. Let's start with the basics (as the section title
suggests we are), and give some basic rules for recognizing (in an inexact number) exactly what
parts of it are significant. Believe it or not, the rules are few and (hopefully) fairly simple to
follow:

1. Any zeros to the left of the first nonzero number do not count as significant figures. That is,
0.00000000000001 has as many significant figures as does 1, and 123 has more
significant figures than 0.000023. This is because zeros which appear to the left of the first
nonzero number are only acting as a marker for the decimal place. To see this, rewrite the
numbers in scientific notation: 0.000023 becomes 2.3 10-5, which we see now has
fewer significant figures than does 123. (Review your scientific notation if you don't
understand how I converted 0.000023".)

2. Any zeros embedded in the number are counted as significant figures. For instance, 103"
has the same number of significant figures as 123".

3. Zeros to the right of the lst nonzero number can be either significnt or insignificnt. There are
three possible cses:

Check to see if the number you are dealing with is known exactly. Numbers that come
from conversion of units (like, say 1000 mL in a litre nd 100 cm in a metre) are exmples
 4

of these kinds of numbers. If the number you are checking is one of these, then you
should read section 3.6 for advice about how to deal with it in calculations.

If the number is not exact, but does not have a decimal point somewhere in it, then the
zeros after the last nonzero number do not count as significant. For instance, the zeros in
113700 (assuming it is not an exact number) do not count as significant.

If the number does contain a decimal point in it somewhere, the zeros count as
significant. For instance, the zeros in 113700. count as significant, as opposed to the
last example where they did not. Note that the only difference between the two cases is
the decimal point in the second.

Let's look at an example to (try to) make things absolutely clear. Have a look at 1230 and
1230.. Neither number is exact (we'll assume this, anyway), and both represent the same
number (one thousand two hundred and thirty). Note, though, that one of the two numbers does
not have a decimal point. Because of this, the two numbers do not have the same number of
significant figures. The number 1230 (without the decimal point) has three significant figures,
while 1230. has four.

It is because of the problems that the zeros cause that many scientists do not use ordinary decimal
notation when writing numbers but instead use scientific notation. That is, they would write
1230. as 1.230 103, and 1230 as 1.23 103. In this way, all numbers written around
the decimal place count as significant figures, and it is immediately obvious just by looking that
1.230 103 has more significant figures than does 1.23 103.

2.3 Counting Significant Figures

Digesting the rules above is the hard part. Once you've done that, counting the number of
significant figures is easy. All you do is leave out anything that does not count according to the
rules above and count all the remaining numbers.

Let's do an example or two to make things absolutely clear. For starters, consider the number
0.0002340500. There are four zeros to the left of the first nonzero number. These will not
count. Of course the zero inside the number does count, and, by the third rule above, since the
number contains a decimal point, the last two zeros also count. Therefore the number has a total
of seven significant figures, five of which are from 23405, and two of which are from the last
two zeros.

Let's also try the number 762000. It's not an exact number (because I say it isn't, and besides, if
it were it wouldn't belong with these examples). It has no decimal point, so the three zeros after
the 2 don't count, and the number therefore has only three significant figures. By contrast, the
number 762000. has six significant figures, since the number contains a decimal point.

(You can see how important it is when communicating numbers to write exactly what you mean!)
 5

Now that we have the basic rules for counting them outlined, let's get on with how we use them
in calculations.

3. Calculations

3.1 Rounding Off Numbers: When to do it

It is important that the answer reported at the end of the calculation reflect the estimated accuracy
of the numbers used in obtaining it. This means, for instance, that when you only really believe
your answer to two decimal places, you should not report the 10 that your calculator may give
you, but only two, and correctly rounded.

There is a problem which creeps in here, though, and that is of round off error. What's that?
Well, round off error occurs when you round off the numbers during the intermediate steps of
your calculations and use only those rounded off numbers in later work. Think it doesn't make a
difference? Try the following example for yourself:

Evaluate the product 1.5 1.5 1.5 1.5 in two different ways. The first way, round off your
answer to two significant figures after each multiplication then do another multiplication using
the rounded answer. I get (using my hyperadvanced HP 15C) the following numbers:

1. 1.5 1.5 = 2.25. Rounded to two significant figures, this is 2.3.

2. 2.3 1.5 = 3.45. Rounded to two significant figures, this is 3.5.

3. 3.5 1.5 = 5.25. Rounded to two significant figures, this (the final answer) is 5.3.

Now let's try the same thing again, and this time we'll multiply all the numbers together and only
round off at the end:

1. 1.5 1.5 = 2.25.

2. 2.25 1.5 = 3.375.

3. 3.375 1.5 = 5.0625. Rounded to two significant figures, this (the final answer) is 5.1.

You can see then that keeping and using all the digits your calculator gives you during a
calculation does make a difference. The moral of the story is therefore that even if you are only
allowed to keep two significant figures for a final answer, you should never round off numbers in
the middle of a calculation.

The thing is, though, that we (the lab supervisors and instructors) expect you to demonstrate a
knowledge of how many significant figures you should be keeping in a calculation. What we (or
at least I, anyway) would like you to do, therefore, is to underline the last significant digit in any
number you report.
 6

Clear? No? Well, read on, and look over the examples at the end of the calculation section and
hopefully it'll be better.

3.2 Basic Addition and Subtraction

In addition and subtraction, your answer can be trusted no more than your least precise number.
That is, if you add 0.0032 (a number with four decimal places) to 5.1 (a number with one decimal
place), you really cant trust your answer any farther than one decimal place, since you dont
know one of the numbers in the sum any more accurately than that.

Have a look at the subtraction below:

82.54
- 23.3344
59.2056

The answer can be trusted no further than the second decimal place, since we do not know one of
the two numbers in the sum any more precisely than that. This means that the answer will
provide four significant figures to later calculations, as it has two significant digits before the
decimal and two after.

If you're confused, don't worry: all will become clear (hopefully) with time and a few more
examples.

Let's try another example. This time, let's add the following two numbers:

24.43
+ 26.92561
51.35561

Again in this case no more than two decimal places can be trusted, because the first number can
be trusted no more than this..

If the number were to be carried on to later calculations, again we would use all the digits in the
answer, although for significant figure counting, we would count only four (two before the
decimal and two after).

3.3 Multiplication and Division

The basic rule here is even simpler than for addition and subtraction. When multiplying or
dividing two numbers, the result of the operation can be trusted to no more significant figures
than the least precise number in the multiplication or division.

As an example, consider what you'd get when you multiply the numbers 1.10 (3 significant
figures) and 1.7 (2 significant figures). Your calculator would (rightly) show the answer 1.87.
However, because 1.7 only has two significant figures (whereas 1.10 has three), the answer
would provide only two significant figures to later calculations. Youd report the answer as 1.87.
 7

Similarly, if you were to divide 1.98 (3 significant figures) by 1.100 (4 significant figures), your
calculator would show the answer 1.8. 1.98 has three significant figures, and 1.100 has four, so
the answer would provide three significant figures to later calculations. In this case, youd have
to pad the right side of the result with zeros until the correct number of significant figures was
indicated. In this case, you'd show it as 1.80.

3.4 Square Roots

3.4.1 What are Square Roots?

Before you learn what to do with square roots, those of you who have not seen them before might
want to find out what they are. Put mathematically, if our number is b, and its square root is
a, then we can write that a a = b.

Hmmm... Let's do an example or two. The square root of 9 is 3, because 3 3 = 9. As another

example, the square root of 4 is 2, because 2 2 = 4. Square roots are (unfortunately) not all
whole numbers, too. For example, the square root of 2 is 1.41421... because 1.41421...
multiplied by itself is 2.

3.4.2 How to Use Significant Figures With Square Roots

Now that you know what they are, the rule for their application is fairly simple. When you take
the square root of a number with n significant figures, the answer you get will provide n+1
significant figures. For instance, if you take the square root of 2.00 (which you know by now has
3 significant figures), the actual square root will provide four, and we would report it as
1.41421...

The reason for this rule originates in calculus, but the answer can be seen more easily than that.
What we want to do when we find the square root of a number is report a number which, when
multiplied by itself and rounded off to the correct number of significant figures, will give back
the original number. For example, take the square root of 3.0. Your calculator will show
1.732050808 (or some number close to it). The answer as we would round it would be 1.73
(since 3.0 has two significant figures). To see that two decimal places (three significant figures
in this case) are necessary, try just squaring the number 1.7 and rounding that off to two
significant figures. Your calculator will show 2.89, and rounded off to two significant figures
this is 2.9. Now try squaring 1.73. Your calculator will show the answer 2.9929. Rounding off
to two significant figures gives back the original number, 3.0.

By extension (for the reasons explained in the previous paragraph), whenever you multiply a
number by itself (called squaring the number), you lose one significant figure in your answer.
For instance, if you squared 2.50, you would get 6.25, but (if it were a final answer), you would
report it as 6.25. Of course, if the number were to be used in later calculations, you would keep
all the digits.
 8

3.5 Logarithms
Logarithms and powers of 10 are probably the hardest things you'll have to worry about as far as
significant figures are concerned. Because of this, I'll go into a little detail for you about what
they are before dealing with them specifically. Those of you who have taken grade 11 and 12
math (or Math 112 at Kwantlen) and feel comfortable with logs can go on to section 3.5.2; the
rest of you read on, and I'll try to explain what they are for you.

3.5.1 What Are Logarithms?

Essentially, a logarithm is just part of another (different) way of writing a number. You're used
(probably) to writing numbers as fractions, for instance 31 , and as decimal numbers, for instance
0.3333333333... Logarithms employ a third and less common way, which consists of writing the
number as one number {the base) raised to the power of another number (the exponent). Instead
of writing 8, for instance, we could write 23, which means 2 multiplied by itself three times, or
2 2 2. In this case, 2 is the base and 3 is the exponent. We could also write 25 as 52, or
5 5. 5 is the base here, and 2 is the exponent.

Usually, when we talk about logs, we talk in terms of exponents with a specific base. There are
two bases that are commonly used in chemistry: e, and 10. e is a nonrepeating decimal
number, the first 11 or so digits of which are 2.718281828459. 10, of course, is 10. We can
write numbers as a combination of either one of these two bases raised to some power (which
will change with the choice of base and the number we wish to represent, of course).

How does all the above relate to logs? Well, when you take the log of a number, you are finding
the exponent to which either e or 10 must be raised in order to recreate the number of which
you took the log. You must be specific about the kind of log you are taking too. When you want
to use e as a base, you say that you are taking the natural logarithm of the number. When you
want to use 10 as a base, you say that you are taking the logarithm base 10 of that number.
These are accessed on your calculator via different buttons, too. Usually, to take the natural
logarithm of a number, you use the button marked lnx on your calculator. The button marked
logx is for when you want to use base 10 for your logarithms.

This, by now, is probably monstrously confusing, so let's take a look at a few examples to clear
things up. Suppose you wanted to take the natural log of 9 (for whatever reason). If you
punched it up on the calculator, you would hit the sequence:

9 [lnx].

You would then see the answer (which is 2.197225...). That means that e2.197225... = 9. If you
were to use base 10, you would use the keystrokes:

9 [logx]

and see the answer (which is 0.9542425...). That means that 100.9542425... = 9.
 9

Earlier on, in the introduction to this section, I used whole number bases and exponents. Don't
worry about how you would multiply anything by itself a fractional number of times; your
calculator will deal with that for you.

3.5.2 How to Use Significant Figures With Logarithms

The object with logarithms and exponents is, just like with squares and square roots, to keep
enough digits to recreate the number when you perform the inverse operation. For instance, if
you take the log of a number (whether natural or base 10), you want to keep enough of the digits
so that when you hit [10x] or [ex] on your calculator (to undo the base 10 log or natural log
respectively), you will get back the original number.

Suppose then that you want to take the log of a number and need to know how many digits are
significant in the resulting log. Let's consider the log (base 10) of 317.88235. Punching this up
on your calculator will show 2.502266415 (etc.). Now, take the log of 3.1788235. Your
calculator will show 0.502266415. Now take the log of 0.0031788235. Your calculator should
show -2.502266415. Hopefully you've noticed a pattern here, and that is that all these logs that
we've taken are the same, except for the numbers to the left of the decimal point. There's a
reason for this, and that is that the numbers to the left of the decimal point serve only to indicate
the position of the number in the decimal, and are not affected by the number of significant
figures in the number whose log we have taken. If we want to take the antilog of the number,
therefore, the only part of the exponent which contributes to the final form of the number is to the
right of the decimal point.

How many decimal places do we report when taking logs, then? Well, we report as many
decimal places in the log of the number as there are significant figures in the original number.

Hmph. Monstrous confusion. Let's do an example. The number 3.17 has three significant
figures. If we take its log (base ten), the calculator shows 0.501059262. How many decimal
places would we report in a final answer? We would report three, because the original number
had three significant figures, so we would write 0.501. (As a side note, if we were to use this
number in later calculations, we would count the significant figures by the usual rules.)

It's the same rules in reverse for taking antilogs of numbers. Suppose we have a number like
0.564789534 which we've figured has 5 significant figures. If we take the antilog of this number
the calculator shows 3.671043528. Now, because the exponent has 5 significant figures, all of
which are to the right of the decimal, the final answer will have 5 significant figures, and we
would report 3.6710 as a final answer.

3.6 Using Exact Numbers

Exact (counting) numbers are very simple to use in calculations which contain them. Why?
Because exact numbers can be treated as though they have an infinite number of significant
figures, and so the accuracy of an answer obtained using them will not be affected by them.

As an example, let's multiply the number 3.772 (an inexact number with four significant figures)
by 2 (which, for this example, is an exact number). The answer will be 7.544, and will have 4
 10

significant figures because 3.772 has four significant figures, which is a smaller number than
infinity.

Square roots are equally as easily dealt with. When taking the square root of a number, provided
that number is exact you may keep and use as many digits as you like or can read off your
calculator. For instance, the square root of 3 is (to 11 decimal places) 1.73205080758.

Sound good so far? Well, there's a catch. Suppose you have to multiply something (say the
number 4.275) by some fraction, like 43 , which is composed (in this example, anyway) of two
counting numbers. You first evaluate 43 on your calculator and record the answer as 1.33, since
you don't want to be bothered recording everything your calculator showed you, which was
probably something like 1.333333333. Anyway, you take this number (1.33) and multiply it by
the other number to get your answer, 5.68575. You figure that since the four and the three that
made up the fraction were counting numbers, they are therefore infinitely accurate, and so your
answer will provide as many significant figures as 4.275 has.

This is not correct, because the second you approximated 43 by a number with three significant
figures, you lost the accuracy of the numbers from which it came, and your once exact number
now only provides three significant figures.

The moral of this story (and a recurring theme throughout this magnum opus of modern
literature) is that you should never round off any number (no matter what kind) during the course
of a calculation.

4. An Example Calculation
In order to show how we'd like you to present calculations, I'll present here a big example
problem and the solution to go along with it, along with some comments. Those of you who
don't yet understand the chemistry involved should just watch what I do with the numbers.

The problem:
1.2000 g CuSO45H2O was dissolved in 100.0 mL of H2O to make solution A. 5.00 mL of this
solution was taken and diluted to 50.00 mL to form solution B. Calculate the concentration of
CuSO45H2O in solutions A and B.
 11

The solution:
Step 1: We need to calculate the number of moles of CuSO45H2O so that we can calculate
the concentration. 1 mole of CuSO45H2O consists of the following:

Element Atomic mass Quantity total

Cu 63.54 g/mol 1 63.54 g
S 32.066 g/mol 1 32.066 g
O 15.9994 g/mol 4 63.9976 g
H 1.00794 g/mol 10 10.0794 g
O 15.9994 g/mol 5 79.9970 g
249.6800 g

The underlined digits in each number are the last significant digits in that number. Note that
since the numbers under Quantity are counting numbers; they are exact and do not affect the
number of significant figures in the numbers under the total column. Also note that since
63.54 has only two decimal places, the total mass may be trusted to only two decimal places and
therefore only five significant figures.

Anyway, now that we know what the molar mass is, we can move on to

Step 2: Calculate the total number of moles of CuSO45H2O used.

The number of moles of CuSO45H2O can be found by taking the mass of CuSO45H2O used and
dividing it by the molar mass:

1 mol CuSO 4 5H 2 O
moles CuSO45H2O = 1.2000 g = 4.8061518743... 10-3
249.6800 grams

Again, the underlined digits in the numbers are the last significant digits in the number in
question. Since the least number of significant figures provided is five, the answer will have five
significant figures.

We now know the number of moles of CuSO45H2O used and can move on to

Step 3: Calculate the concentration of solution A.

The concentration of solution A is the number of moles put into it divided by its volume in litres.
We need, therefore, to convert 100.0 mL (the volume used for solution A) to litres.

1L
100.0 mL = 0.1000 L
1000 mL

The underlined numbers have their usual meaning. Note that neither the 1 L or the 1000 mL
have an underlined number. This is because these are exact numbers (exactly 1 L is defined to be
exactly 1000 mL). As a result, they do not affect the number of significant figures which the
volume will provide to the concentration calculation.
 12

At any rate, the concentration of solution A is:

4.8061518743 10-3 moles

= 4.8061518743 10-2 moles/L
0.1000 L

Note the drop in the number of significant figures in the answer. This is because the volume
contributes four significant figures, while the moles contribute five. The answer must have the
lesser of the two numbers of significant figures (in division), so it has four.

Now were ready for

Step 4: Calculate the number of moles used in solution B.

The number of moles of CuSO45H2O used in solution B is the concentration of solution A
multiplied by the volume of it taken. We start (as before) by converting the volume to litres.
5.00 mL of it was used, so:

1L
5.00 mL = 0.00500 L
1000 mL

As before, the unit conversion factors used here do not affect the number of significant figures in
the answer since they are exact.

Now, the number of moles used in solution B is:

0.00500 L 4.8061518743 10-2 moles/L = 2.40307593715 10-4 moles

Notice here again the drop in the number of significant figures, since only three are provided by
the volume.

We are, at last, ready for

Step 5: Calculate the concentration of solution B.

As before, we take the number of moles used and divide by the volume of solution used. We
first convert 50.00 mL to litres:

1L
50.00 mL = 0.05000 L
1000 mL

And, as before, the exact unit conversion does not affect the number of significant figures in the
volume.

The concentration is therefore:

2.40307593715 10-4 moles

= 4.8061518743 10-3 moles/L
0.05000 L
 13

Now, since this is the final answer, we could round it off to the correct number of significant
figures and report it as 4.81 10-3 moles/L

5. Concluding Comments
Don't panic. There's a lot to what I've written above, and a lot of it may be new. If you do have
questions, I'm more than happy to discuss them with you and try and help you solve your
problems. In the end and with time though, you'll get the hang of it with a bit of work. I've
provided some practice problems below for you to try. Good luck!
 14

6. Practice Problems
If you feel confident about your newfound (and profound) understanding of significant figures,
you might want to try these problems to test yourself. Answers will be provided on request.

1. How many significant figures do each of the following inexact numbers have?

a) 3.22
b) 1.042
c) 0.00032
d) 0.22340
e) 230

2. Complete the following expressions. Assume all answers are final nd may be rounded off to
the correct number of significant figures. (Assume that the number 2 which appears twice in
problem 2 (d) is a counting number.)

a) 3.22 + 5.272 =
b) (3.22 + 5.272) 10.00 =
c) 3.22 + 5.272 - 1.76553 =
d) 2 4.327 - 2 =

3. For advanced chemists only. A student used 24.28 mL of 0.1232 mol/L NaOH to neutralize
25.00 mL of HCl. Calculate the concentration of the HCl in g HCl/kg solution, given that the
density of the HCl solution was 0.952 g/mL, and that the atomic masses of H and Cl are
1.00794 g/mol and 35.453 g/mol respectively.