Professional Documents
Culture Documents
2
Sample
prepara,on-‐con$nued
Week
2
1
Preparing
solu2ons
How
to
prepare
a
solu,on?
Expression
of
concentra,on
Concentra,on
is
the
quan,ty
of
solute
in
a
known
amount
of
volume
or
mass
of
solu,on
or
solvent
Concentra,on
=
Amount
of
solute
Amount
of
solu,on
Week
2
2
Concentra,on
in
terms
of
percent
composi,on
The concentration of substance in commercial aqueous reagents, organic
solvents and commercial household products are usually expressed in
percent composition
Example: 37% is labeled on a HCl reagent bottle. This means that it contains 37 g
HCl per 100 g solution
Week
2
3
•
Parts
per
million
(ppm)
=
g
solute
106
g
solu2on
Or
=
µg/g
As
the
density
of
aqueous
solu,on
oPen
very
close
to
1.00
g/L,
we
usually
correlates
1
g
water
with
1
mL
water
(approxima,on)
Therefore,
1
ppm
=
1
µg/g can also be converted to 1 µg/mL
=
1
mg/L
Week
2
4
For
trace
analysis
in
liquid
sample
~Analysis
of
very
very
very
liLle
amount
of
analytes
Week
2
5
Conver,ng
percent
composi,on
to
molar
concentra,on
-‐conver$ng
‘commercial
term’
to
‘easier
laboratory
term’-‐
• Example:
A
commercial
aqueous
ammonia
(NH3
)
boYle
is
labeled
as
28.0%
w/w,
and
has
a
density
of
0.899
g/mL
and
MW
of
17.03
g/mol.
Let
say
you
want
to
convert
the
concentra,on
into
molarity
for
easier
calcula,on
in
your
lab
book,
how
can
you
do
that?
density
=
28
g/100
g
x
0.899
g/mL
=
0.2517
g/mL
=
0.2517
g/mL
/
17.031
g/mol
MW
=
0.01478
mol/mL
x
1000
mL/1L
=
14.8
mol/L
or
14.8
M
6
Week
2
6
Prepara,on
of
solu,on
Example:
How
to
prepare
250
mL
of
0.100
M
NaOH
from
NaOH
solid?
[MW:
NaOH
=
40
g/mol]
1.
Calculate
the
weight
(g)
of
NaOH
pellet
using
MW
to
produce
0.1
M
solu,on
Calcula$ons
????
2.
Weigh
???
g
of
solid
(generally
±
0.1
mg,
ie
up
to
4
decimal
places
in
grams)
3.
Dissolve
in
water,
transfer
(quan,ta,vely
with
rinsing)
to
a
250
mL
volumetric
flask,
and
dilute
to
the
mark
Dissolve
and
Calculate
Weigh
transfer
Week
2
7
Prepara,on
of
solu,on
Example:
How
to
prepare
25
mL
of
0.01
M
glutamic
acid
solu,on
from
glutamic
acid
powder?
[MW:
Glutamic
acid
=
147.13
g/mol]
0.03678 g
Week
2
8
Dilu,on
of
solu,on
The
moles
solute
in
concentrated
(1)
solu2on
equals
the
moles
in
dilute
(2)
solu2on
∴ M1 × V1 = M2 × V2
Example:
Describe
the
prepara,on
of
50
mL
of
0.100
M
NaOH
solu,on
from
a
0.5
M
solu,on
Calcula,ons…
How
to
???…
?
Glassware
requirement:
??
mL
pipet
and
??
mL
volumetric
flask
Week
2
9
Week
2+3
CHAPTER
2
Data
analysis
and
sta,s,cs
Week
2
10
Statistics in Analytical Chemistry
Week
2
11
ie. measurement of agreement between experimental
mean and true value (which may not be known!)
Measures of accuracy:
- Absolute error: E = xi - µ
(where µ = true or accepted value)
xi − µ
- Relative error: E R = | | × 100%
µ
ACCURACY
Relative error is more useful in practice
• degree
of
agreement
between
measured
value
• Absolute
error
=
measured
value
–
true
value
and
the
true
value
(which
may
not
be
known!)
E.g.
if
a
2.72
g
sample
is
analysed
to
be
2.62g,
• Therefore,
it
is
the
degree
of
agreement
the
absolute
error
is
-‐0.10g
between
measured
value
and
the
accepted
true
value
• Rela,ve
error
is
the
absolute
error
expressed
as
%
of
the
true
value
PRECISION
E.g.
(-‐0.10g/2.72
g)
x
100%
=
-‐3.7%
• Degree
of
agreement
between
replicate
measurements
of
the
same
quan,ty;
repeatability
of
a
result.
• Expressed
by
standard
devia$on,
the
coefficient
of
varia$on,
the
range
of
the
data
or
as
confidence
interval
(e.g.
95%)
about
the
mean
value
• How
similar
are
values
obtained
in
exactly
the
same
way?
• Useful
for
measuring
devia,on
from
the
mean.
di = x i − x
Week
2
12
High
Precision
High
accuracy
High
precision
Low
accuracy
Low
precision
Low
accuracy
Low
precision
High
accuracy
x
Week
2
13
Types
of
error
in
chemical
analysis
A.
Systema2c
(determinate)
Error
Operator/Instrument
error/Method
error
• Can
be
constant
(e.g
uncalibrated
weight
used
in
all
weighing)
or
variable
(e.g.buret
whose
volume
readings
are
in
error
by
different
amount
at
different
volumes)
• Readings
all
too
high
or
too
low
that
can
affect
accuracy.
How
to
detect?àblank
determina,ons,
analysis
of
standard
samples,
independent
analyses
by
alterna,ve/dissimilar
methods
B.
Random/accidental
(indeterminate)
Error
• Represents
experimental
uncertainty
occurred
in
any
measurement
• Random
and
cannot
be
avoided.
Affects
precision,
can
only
be
controlled
• Revealed
by
small
differences
in
successive
measurements
made
by
the
same
analyst
under
iden,cal
condi,onsà
can’t
be
predicted/
es,mated
but
can
be
dealt
with
sta*s*cs
• Data
scaYered
approximately
symmetrically
about
a
mean
value.
C.
Gross
Errors
• Serious
but
very
seldom
occur
in
analysis.
• Usually
obvious
-‐
give
outlier
readings.
• Detectable
by
carrying
out
sufficient
replicate
measurements.
• Experiments
must
be
repeated.
e.g.
Instrument
faulty,
contaminated
reagent
Week
2
14
SOURCES
OF
SYSTEMATIC
ERROR
Week
2
15
Significant
figures
(SF)
• The
SF
of
a
number
are
those
digits
that
carry
meaning
contribu,ng
to
its
precision.
•
Zero
is
significant
only
when:
A. It
occurs
in
the
middle
of
a
number.
Examples:
a) 401
-‐
3
SF
b) 6.0015
-‐
5
SF
c) 3.00
-‐
3
SF
d) 6.00
×
102
-‐
3
SF
Week
2
16
Basic Statistical Concepts/Definitions
• True value - value that remains unknown except
when a standard sample is analyzed
(value estimated from results of varying precision
depending on the method used)
• Accuracy - nearness of a measurement or result
to the true value (expressed in terms of error
error))
• Precision - variability of a measurement
(Standard deviations are precision indicators)
• Spread-- difference between the highest and
Spread
lowest results in a set (spread is a measure of
precision)
• Mean - average of a replicate set of results
• Median - middle value of a replicate set of results
Week
2
17
• Degree of Freedom - number of results in a set
(each time another quantity is derived from the
set, the degrees of freedom are reduced by 1)
• Range - difference between the highest and
lowest value of the results
• Standard Deviation (s or σ) - difference, with
respect to sign, between an individual result
and the mean or median of the set
• Relative Standard Deviation (RSD) - Also known
as the coefficient of variation, often used in
comparing precisions
• Variance (V)
(V) - square of the value of standard
deviation (σ2 or s2)
Week
2
18
Mean/Average
Median
N
• Arranged
in
accending
order,
if
data
in
the
middle
in
an
odd
number
record
it
as
the
∑ x i it
If
is
for
large
data
set,
called
popula,on
median.
x = i =1 mean
(µ)
N
Xi
=
individual
values
of
x
• Arranged
in
ascending
order,
if
two
middle
N
=
number
of
replicate
measurements
data
are
even
numbers
then
average
the
two
numbers
Range
The
difference
between
the
highest
and
lowest
result.
Week
2
19
Standard
Devia,on
(SD)
–
a
very
important
precision
indicator
Measure
of
spread
about
the
mean
Es,mate
the
variability
of
individual
measurement
(The
standard
devia,on
is
beYer
es,mated
by
the
pooling
of
results
from
more
than
one
set)
Can
be
calculated
using
Excel
spreadsheet
Small
sample
size
(N=20)
Popula2on
(N=
>20)
s =
(
∑ xi − x )
σ=
∑ xi − µ ( )2
N−1 N
Xi
=
individual
values
of
x
Xi
=
individual
values
of
x
x
=
mean
x
=
popula,on
mean
N
=
number
of
replicate
measurements
N
=
number
of
replicate
measurements
Week
2
20
! RELATIVE STANDARD DEVIATION (RSD)/
COEFFICIENT OF VARIATION (CV)
Standard deviation divided by mean
(depends on the units used)
Week
2
21
Sample Se (mg/g) (xi - mean)
Example: 1 0.07 4.9x10-5
Determine the 2 0.07 4.9x10-5
content of Se in 3 0.08 9.0x10-6
a batch of 4 0.07 4.9x10-5
brown rice 5 0.07 4.9x10-5
6 0.08 9.0x10-6
7 0.08 9.0x10-6
Mean = Σxi/N 8 0.09 1.69x10-4
= 0.077 9 0.08 9.0x10-6
Σ(x
xi-mean)2 = 4.01x10-4
∑ i
( x − x ) 2
S.D. = s= i = 0.007
N −1
Content of Se = 0.077 ± 0.007 mg/g
What does this result mean?
Week
2
22
10.00 10.07 10.07 10.03 10.13
10.00 10.08 10.00
10.24 10.05 10.24
10.08 10.22 10.05 10.22
• Calcula,on
of
SD
for
a
set
of
data
provides
indica,on
of
the
precision
inherent
in
par,cular
procedure.
CONFIDENCE LIMIT & CONFIDENCE INTERVAL
• For
large
• Confidence data
set,
Interval (CI) it
disoesn't
the range give
of
any
info
about
how
close
the
experimentally
values
determined
surrounding themmean ean
(,
within
mean, )to
the
which
true
m ean
value
(µ).
the
population mean, CONFIDENCE
µ is expected to LIMITlie with& CONFIDENCE
a INTERVAL
certain degree of probability
• The boundries • Confidence
of the range are Interval
called the (CI) is the range of values
Confidence Limits surrounding the mean mean,, within which the
• Confidence Level (CL) is themean,
population probability µ isthat the
expected to lie with a
true mean lies within a certain interval
(expressed ascertain %) degree of probability
Example: • The boundries of the range are called the
It is 99% probable Confidence
that µ forLimits a set of
measurement is 7.25mg ± 0.15.
• Confidence Level Thus, theis
(CL) meanthe probability that the
should lie in the interval from 7.10mg to 7.40mg
with 99% probability true mean lies within a certain interval
(expressed as %)
Example:
It is 99% probable that µ for a set of
measurement is 7.25mg ± 0.15. Thus, the mean
should lie in the interval from 7.10mg to 7.40mg
with 99% probability
Values of z for determining confidence limits
Week
2
23
Confidence level (%) z
Confidence
interval
for
small
data
set
(N
=
20)
ts •
Values
of
t
depend
on
degree
of
freedom,
v
(N
-‐
1)
and
CI ( µ ) = x ± confidence
level
(from
Table).
N
•
t
also
known
as
‘student’s
t’
and
will
be
used
in
N=Number
of
measurements/
replicates
hypothesis
test.
Week
2
24
Example:
Data
for
the
analysis
of
calcium
in
rock
are
given
by
14.35%,
14.41%,
14.40%,
14.32%
and
14.37%.
Within
what
range
are
you
95%
confident
that
the
true
value
lies??
ts
CI ( µ ) = x ±
Solu,on:
N
Mean,
x
=
14.37
SD,
s
=
0.037
At
different
confidence
level,
From
the
table,
at
95
%
confidence
level,
N
-‐
1
=
4,
t
=
2.78.
Confidence Level Confidence Interval
2.78 × 0.037 90% µ = 14.37% ± 0.04
µ=x±
Therefore,
CI
is
:
5 95% µ = 14.37% ± 0.05
= 14.37% ± 0.05% 99% µ = 14.37% ± 0.08
Summary:
So
you
are
95%
confidence
that
(in
the
If
the
confidence
level
increased,
the
absence
of
determinate
error),
confidence
interval
(CI)
also
increased.
The
the
true
value
falls
within
14.32%
to
14.42%
probability
of
the
true
mean
value
(µ)
appeared
in
the
interval
will
increase
Week
2
25
The
concentra,on
of
an
addi,ve
in
a
standard
sample
of
gasoline
was
measured
5
,mes
with
the
following
results:
0.13,
0.11,
0.12,
0.20,
and
0.14
%
by
mass.
(a) Calculate
the
sample
mean
and
standard
devia,on
of
the
data
Mean:
0.14,
s:
0.03536
(b)
Calculate
the
95%
confidence
interval
and
explain
what
does
the
data
means?
ts
CI ( µ ) = x ±
N
2.78 × 0.03536
µ=x±
5 there
is
a
95%
probability
that
the
concentra,on
will
fall
between
0.10%
to
0.18%
= 0.14% ± 0.04%
Week
2
26
Confidence
interval
for
large
data
set
(N
>20)
essions of Confidence
Interval
• CI
when
σ
is
Values of z for determining confidence limits
rge no. of data (>30)
µ = x ± z kσnown
(popula,on),
own population std
N Confidence level (%) z
n, σ zσ
(≤30)
mall no. of data (≤ µ=x±
µ=x± ts
50 0.67
knowing σ (know s) N N 68
80
1.0
1.29
90 1.64
N
=
Number
of
measurements/replicates
mber of measurements
95 1.96
ues from normal distribution curve
96 2.00
z-table)z=
the
value
from
the
standard
normal
om the z-
99 2.58
ues from normal distribution curve but
distribu,on
on the degree of freedom (N- (Nfor
-1)t(Read
he
selected
confidence
level
99.7 3.00
99.9 3.29
e t-
t-table) At 90% confidence level, z = 1.64,
known as the ‘student’s t’, generally
hypothesis tests σ
µ = x ± 1.64
N
Week
2
28
A.
To
determine
numer
of
replicate
Exa
gav
Example 1:
1: Calculate the number of replicates A so
needed to change the confidence interval by whe
1.5 µg/mL at 95% confidence level. con
Given, s = 2.4 µg/mL
At 9
Week
2
29
Sample Question
(Confidence Limit when s is known)
B. To
determine
systema2c
error
1)
Calcula2on
method
1
Example
A
standard
solu,on
gave
an
absorp,on
reading
of
0.470
at
a
par,cular
wavelength.
Ten
measurements
were
done
on
a
sample
and
the
mean
gave
a
value
of
0.461,
with
standard
devia,on
(s)
of
0.003.
Show
whether
systema,c
error
exists
in
the
measurements
at
95%
confidence
level.
Values of t for various levels of probability
Solu,on
Degrees of Freedom 80% 90% 95% 99%
At
95%
confidence
level,
N
=
10,
t
=
2.26,
(N
(N--1)
= 0.461 ± 2.26
(0.003) 9
19
1.38
1.33
1.83
1.73
2.26
2.10
3.25
2.88
10 59
∞
1.30
1.29
1.67
1.64
2.00
1.96
2.66
2.58
= 0.461 ± 0.002
The
calcula,on
gives
confidence
limit
of,
• Does
the
true
mean
0.470
0.459
<
µ
<
0.463
belong
to
the
interval?
Sample Question
• D(Confidence
oes
systema*c
error
Limit when s is p resent?
known)
AAS analysis of Cu in aircraft engine oil gave a mean value of
8.53 mg Cu/mL
Cu/mL.. Pooled results of many analyses showed
that s → σ = 0.32 mg Cu/mL
Cu/mL.. Calculate the confidence
Week
2
30
intervals (CI) at 90% & 99% confidence levels based on
(a) 1 (b) 4 (c) 16 measurements
(a) Confidence limit (CL) = µ = x ± t s
B. To
determine
systema2c
error
2)
Calcula2on
method
2
Example
A
standard
solu,on
gave
an
absorp,on
reading
of
0.470
at
a
par,cular
wavelength.
Ten
measurements
were
done
on
a
sample
and
the
mean
gave
a
value
of
0.461,
with
standard
devia,on
(s)
of
0.003.
Show
whether
systema,c
error
exists
in
the
measurements
at
95%
confidence
level.
Solu,on
At
95%
confidence
level,
N
=
10,
t
=
2.26,
N The
tcalc
>ttable
± t = (x − µ )
s Does
systema*c
error
10 present?
= (0.461 − 0.470)
0.003
t = 9.49
Week
2
31
0.463 N 10
Therefore, the
the reading 0.470 is
= 0.461 ± 0.002 NOT in the
range, and systematic
This means,error
0.459EXISTS
< µ < 0.463, ie 95% of the
time, the true value lies between 0.459 to
0.463
Therefore, the
the reading 0.470 is NOT in the
range, and systematic error EXISTS
dN/N
dN/N
mean that:- 80% of the time the true mean will lie
dN/N
dN/N
between
- 80% of the time the±1.29s
true of the measurements
mean will lie made
- or in other words 20% of the time the true
υ υ υ between ±1.29smeanof the
will NOTmeasurements
lie between ±1.29s made
υ
-3s -2s -1s 0 1s 2s 3s 4s -4s -3s -2s -1s 0 1s 2s 3s 4s -4s -3s -2s -1s 0 1s 2s 3s 4s - or in other words 20% of the time the true
υ -4s
mean is indicated by υ
mean will NOT lie between ±1.29s
4s -4s -3s -2s -1s 0 1s 2s 3s 4s -4s -3s -2s -1s 0 1s 2s 3s 4s
mean is indicated by υ
Week
2
33
- or in other words 20% of the time the true
mean will NOT lie between ±1.29s
t-‐test
The ‘t Test’ (Student t-
t-test)
(2) Comparison of means ( ) of two samples
Has two uses:
- eg Compare mean of new method with a
(1) Comparison of true value, µ and mean, reference (or standard) method
to detect if difference is significant
- Accept Null hypothesis (Ho) if NO significant
- Used to detect the existence of systematic difference between methods ie the results are
error or bias the same, or =0
- Calculate t (generally for 95% confidence level) - Calculate t, if tcalc < ttable, accept Ho to show that
- If value of tcalculate < tcritical (ie tcalc < ttable), there is NO significant difference in results
ACCEPT the null hypothesis, thus Ho: µ = - Use pooled estimate of std dev,
- Accepting Ho means that there is NO s2={(n1-1)s12+ (n2-1)s22} / (n1+n2-2),
significant difference (or no systematic
error) at the 95% confidence level, but or
there is 5% probability that there is a
sgnificant difference
Week
2
35
EXAMPLE 2: Determine if results differ significantly
Week
2
36
F-‐test
àComparing
the
precision
of
two
measurements
• Compare
standard
devia,ons
(ie
random
errors
of
2
data
sets)
F-TABLE
• Is
there
any
significant
The difference
‘F Test’between
both
methods?
• Compares Std dev (ie random errors of 2 data sets)
- One tailed test: test: test whether method A is more
precise than method B (assumes A is always precise)
- Two tailed test: test: test whether methods A and B
differ in their precision (ie any method can be precise)
- F is ratio of two 2
nominator
sample variances: ν1 s
F= = 1
ν2 s 22 denominator
Ho: Population variances are equal (or ≅ 1)
[F is always >1, thus the smaller ν ie the
more precise is always the denominator]
If Fcalc < Ftable (Accept Ho) which means that there
is NO significant difference in precision between
the two methods
Week
2
37
Example Question: ONE ONE--TAILED F TEST
A proposed method for COD of wastewater was
compared with a standardized method
The results are given as follows:
Standardized method (8 (8 determinations):
determinations):
mean =72 mg/L, s = 3.31 mg/L
Proposed method (9 (9 determinations):
determinations):
mean = 72 mg/L, s = 1.51 mg/L (↓)
Is the proposed method significantly more precise
than the standardized method?
F = (SStd)2/(SProp)2 = (3.31)2/(1.51)2 = 4.8
Data values: 8 for Std & 9 for proposed, thus from
the F-
F-table degrees of freedom (N (N--1) = 7numerator
and 8denominator, Fcrit = 3.50
Since Fcalc >Ftable , reject Ho. Thus there is a
significant difference bet the methods and the
proposed method is significantly more precise
Set as denominator
Week
2
38
Example: Determination of CO using a ‘Standard
Procedure’ gave an s value of 0.21 ppm. The
method was modified twice giving s1 of 0.15 and s2
of 0.12 (both 9 degrees of freedom).
freedom). Are the modified
methods significantly more precise than the std?
Ho : s1 = sstd Ho: s2 = sstd
s2std 0.212 s 2std 0.21 2
F1 = 2 = 2
= 1.96 F2 = 2 = 2
= 3.06
s1 0.15 s2 0.12
In standard methods the # of data is large, thus
s→σ,
→σ, & degrees of freedom becomes infinity, ∞
From F- num=∞, den=9; Fcrit = 2.71
F-table, num=∞
F1< Ftable : accept Ho but F2>Ftable : reject Ho
Only the 2nd modified method is is significantly
more precise than the standard method
Week
2
39
F-‐table
Week
2
40
The Q TEST or DIXON’S TEST
(Detection of gross errors)
‘The Q-
Q-Test is used for detecting outlier (suspected
unreasonable data) which statistically does not
belong to the set’
Example:: 10.05, 10.10, 10.15, 10.05, 10.45, 10.10
Example
• By inspection,
inspection, 10.45 seems to be out of the data
normal range (More easily observed when numbers are
arranged in a decreasing or increasing order)
10.05, 10.05, 10.10, 10.10, 10.15, 10.45
• Can/should this data be eliminated … ? (The mean
will change from the original value if changed!)
Contd…
Week
2
41
Use this
Equation:
Week
2
42
Q TABLE
No. of Observations Confidence Level
90% 95% 99%
3 0.941 0.970 0.994
4 0.765 0.829 0.926
5 0.642 0.710 0.821
6 0.560 0.625 0.740
7 0.507 0.568 0.680
8 0.468 0.526 0.634
9 0.437 0.493 0.599
10 0.412 0.466 0.568
Week
2
43
EXAMPLE QUESTION: Q-
Q-TEST
Week
2
44
Example:
An
analysis
on
calcite
gave
the
following
percentage
of
CaO:
55.45,
56.04,
56.23,
56.00,
55.08
Q:
Is
there
any
that
data
should
be
rejected
at
95%
confidence
level?
Solu2on:
• Arrange
data
55.08,
55.45,
56.00,
56.04,
56.08,
56.23
• Suspected
data:
55.45
OR
56.23
Qtable
from
5
determina2ons,
95%
=
0.710
56.23 - 56.08
Q calc =
56.23 - 55.45 Qcalc<Qtable.
Data
cannot
be
rejected.
= 0.19
55.45 - 56.00
Q calc = Qcalc
=
Qtable.
Data
cannot
be
rejected.
56.23 - 55.45
= 0.71
Week 2 45