1. science
2. chemistry

best
@buchi
www.buchi.com
Information Bulletin
Number 58/2010
How to Achieve Low Detection and Quantification Limits
for the Nitrogen Determination with Kjeldahl
en
best@buchi 58/ 2010 en
concentration of the titration solution on the detection and quantification limits were investigat
best results were obtained by using 2 % boric acid with 3 g potassium chloride per liter. A titra
solution of 0.005 M HCl worked best. With these parameters, detection limits for distillation of
solutions as low as 0.008 mg nitrogen and quantification limits of 0.02 mg nitrogen can be ac
How to Achieve Low Detection and Quantification Limits for the Nitrogen
Introduction
Often the potential of the reference electrode Eref, wh
Determination with Kjeldahl
which can lead to measurement variations. The varia
Authors: Dr. Claudia Blum-Fretz,
Stephan Buschor, Jürgen Müller
Abstract
Kjeldahl is one of the most commonly
used techniques to determine the protein
content in food and feed samples. The
detection and quantification limits are
important characteristics of analytical
methods. The impact of the concentration of boric acid, the addition of
potassium chloride, and the concentration of the titration solution on the
detection and quantification limits were
investigated. The best results were
obtained by using 2 % boric acid with
3 g potassium chloride per liter. A
titration solution of 0.005 M HCl
worked best. With these parameters,
detection limits for distillation of
standard solutions as low as 0.008 mg
nitrogen and quantification limits of
0.02 mg nitrogen can be achieved.
Introduction
Kjeldahl
For almost 130 years, the determination
of nitrogen using the method developed
by the Danish chemist Johan Kjeldahl
(1849–1900) has been an internationally
accepted standard. The method, which
is named after its inventor, has since
found widespread application in life science and chemistry and has extended
its scope to the determination of nitrogen
and proteins in dairy products, meat
products, beer, cereals, and other food
materials [1].
The Kjeldahl procedure involves three
major steps:
In the digestion step, the organically
bonded nitrogen is converted into
ammonium ions by oxidation with
concentrated sulfuric acid.
In the distillation step, the sample is
alkalinized to convert the ammonium
ions to ammonia. The latter is then
distilled into a boric acid solution
(via steam distillation).
In the final titration step, the ammonia
is titrated and the nitrogen content can
be calculated.
2
is stirred. To demonstrate the stirring effect, a detaile
Kjeldahl
For almost 130 years, the determination of nitrogen using the method developed by the Dani
Theoretical
of pH has been an internationally accepted standard. The me
chemist Johanbackground
Kjeldahl (1849–1900)
measurements
and
which is named after
its boric
inventor,acid
has since found widespread application in life science and c
titration
and has extended its scope to the determination of nitrogen and proteins in dairy products, m
products,
beer,
cereals,
andlogarithm
other food materials [1].
The
pH value
is the
negative
The
Kjeldahl
procedure
involves
three
of the hydronium ion activity and
is major steps:
In the digestion step, the organically bonded nitrogen is converted into ammonium ions by ox
measured with an electro-chemical
with concentrated sulfuric acid.
sensor.
practice this
is athe
measurement
In theIndistillation
step,
sample is alkalinized to convert the ammonium ions to ammonia. T
of isathen
potential
a (via steam distillation).
distilleddifference
into a boricbetween
acid solution
mea- is titrated and the nitrogen content can be calculated.
reference
electrode
In the final
titration E
step,
thethe
ammonia
ref and
suring electrode E. The measured voltTheoretical
background
pH measurements and boric acid titration
age
U is the potential
difference of of
E and
ErefThe
. The
of negative
pH is performed
pHcalculation
value is the
logarithm of the hydronium ion activity and is measured with an
chemical to
sensor.
In practiceequations
this is a measurement of a potential difference between a refere
according
the following
measuring
(1 electrode
- 2), whichand
are the
derived
from theelectrode.
Nernst The measured voltage U is the potential difference of
Eref. The calculation of pH is performed according to the following equations (1 - 2), which are
equation [2 - 4].
from the Nernst equation [2 - 4].
pH = pH 0 −
E − E ref
slope
(1)
(1)
The
quotient
equation(2)(2)represents
represents the slope of the pH function and shows that the slope
The
quotient
in inequation
function
of
the
temperature.
Neue
Neue
Formeln
the slope
ofFormeln
the
pH function and
Figure 1: Schematic description of the pH Electro
shows that the slope is a function of
f ln−log * R * T
Figure 1: Schematic representation of
theslope
temperature.
=
(2)
the pH electrode
z*F
1
measuring electrode (e.g., Ag/AgCl)
2 ion activity
internal reference solution
pH
negative
of
the
hydronium
f ln −log
R
*
T
•
f lnlogarithm
R
*
T
•
− log
1
measuring
electrode
(e.g.,
Ag/AgCl)
(2)
3 the
pHwhen
sensitive
glass
membrane
(2)sensor
= at=zero point of pH sensor
(2)
slope
pH
(i.e.,
pH
the
signal
is 0 mV)
pH0 slope
z
•
z•F
4 internal
sample
solution
(e.g., boric acid as receiving
reference
solution
E
potential at F
measuring
electrode 2
5 pH
liquid
(e.g., ceramic diaphragm)
potential of the reference electrode
(should
be junction
constant)
Eref
3
sensitive
glass membrane
6
reference
electrolyte
(e.g.,logarithm
3 M KCl) (2.303)
conversion
factor
for
the
change
of
the
natural
(ln)
to
the
common
f
pHln-log
negative logarithm of the
4 sample solution
7
reference electrode (e.g., Ag/AgCl)
R
universal
gas
constant
(8.3145
J/(K*mol))
hydronium
ionx (activity
U (e.g., voltage
boric acid
as(5)
receiving solution)
(5)
xabsolute
measurement
) = )x (=LOD
) • k) • k
( LOQ
LOD
T x ( LOQ
temperature
[K]
zero point
of
pH
sensor
pHz0 pH atnumber
5
liquid
junction
of electrons transferred (for pH: 1)
variability
the potential is produced at the li
when constant
the sensor
The(e.g.,
ceramicofdiaphragm)
F (the pH
Faraday
(9.6485*104 C/mol)borate
and hydronium
ion, etc.).
In diluted solution
signal is 0 mV)
6
reference
electrolyte (e.g.,
3 M KCl)
solutions.
If the
solution
is not
stirred, a cloud of pota
E potential at measuring electrode
7
reference
electrode
(e.g.,
Ag/AgCl)
and 3) is created at the exterior of the liquid junction
potential
of the reference
Eref Zusätzlicher
U
voltage the
measurement
Satz
aufauf
Seite
4:
Zusätzlicher
Satz
Seite
4:
is stirred,
cloud of potassium and chloride ions i
electrode (should be constant)
increases and the measured pH value decreases.
The variability of the potential is
fln-log conversion factor for the change
In this
study,
thethe
LOD
andand
LOQ
were
always
calculated
according
In this
study,
LOD
LOQ
were
always
calculated
accordint
of the natural (ln) to the common produced at the liquid junction (zeta
thethe
indirect
method
to be
able
to compare
thethe
findings.
ForFor
Kjeldahl
indirect
method
to be
able
to compare
findings.
Kjeld
potential, different mobility of borate
logarithm (2.303)
method
is
well
suited,
because
the
matrix
is
completely
destroyed
b
method
is
well
suited,
because
the
matrix
is
completely
destroye
and hydronium ion, etc.). In diluted
R universal gas constant
sulfuric
acid.
with
sulfuric
acid.
•mol))
solutions, the variability is higher
with
(8.3145
J/(K
than in concentrated solutions. If the
T absolute temperature [K]
solution is not stirred, a cloud of
z number of electrons transferred
potassium
and chloride ions (black
Zusätzlicher
(forZusätzlicher
pH: 1)
Satz
aufauf
Seite
5:
Satz
Seite
5:
dots in Figures 2 and 3) is created at
F Faraday constant
the exterior of the liquid junction and
(9.6485•10 4 C/mol)
The
experiment
was
setset
up up
in
the
following
way:
1) optimization
of to
The
experiment
was
in the
following
way:
1) optimization
reduces the surface potential. If the
concentration
2)
optimization
of
the
KCl
addition
and
finally
3)
optim
concentration
2)
optimization
of
the
KCl
addition
and
finally
3)
op
Often the potential of the reference
solution is stirred, the cloud of potastitrant
concentration.
titrant
concentration.
sium and chloride ions is removed
electrode
Eref, which
should be constant,
shows a small variability which can
from the surface so that the potential
increases and the measured pH value
lead to measurement variations. The
variability of the potential is largest when
decreases.
the solution is stirred. To demonstrate the
stirring effect, a detailed view of the pH
sensor is shown in Fig. 1.
best@buchi 58/ 2010 en
1. The pH increase due to dilution of the receiving solution by distillate is less
important in low concentrated boric acid. The variability of the amount of distilled
water has less impact on the pH value and will therefore lead to less variability of
the blank values. The pH change related to dilution is shown in Figure 4.
pH change by dilution of boric acid
5.9
5.7
pH
5.5
5.3
5.1
4.9
4.7
Figure 2:
without
stirring
Figure
2:Liquid
Liquidjunction
junction
without
stirring
4.5
0
Figure
3: Liquid
junction60with stirring
20
40
80
Added H2O [ml]
100
120
4%
The stirring effect can be minimized by adding potassium chloride to low concentrated (< 4%) boric
acid to ensure that a sufficient amount
of potassium chloride is always at the surface of the liquid
Figure 4: pH change when diluting 60 ml receiving solution at different
junction.
concentrations of boric acid
140
2%
1%
with stirring
The use of diluted boric acid is beneficial for the determination of low nitrogen amounts for the
following three main reasons:
2. The blank values in less concentrated boric acid are smaller for the same
1. The pH increase due to dilution of the
receiving
solution
distillate isimportant
less important
in low
reason
as above.
Thisby
is particularly
because
usually low concentrated
concentrated boric acid. The variability titration
of the amount
of
distilled
water
has
less
impact
on
the
pH value
solutions are used for the determination of low
nitrogen
amounts. For
and will therefore lead to less variabilitythe
of determination
the blank values.
The
pH
change
related
to
dilution
is have smaller blank
of low nitrogen contents it is advantageous to
shown in Figure 4.
values, because the difference in titration volumes between the blanks and the
samples gets larger.
pH change by dilution of boric acid
5.9
3. The pH change caused by the distilled nitrogen is more important the lower
the concentration of the receiving solution is. Small amounts of nitrogen cause
a considerable increase in pH, thus making the titration more accurate.
5.7
5.5
5.3
pH
g
without stirring
Detection limit and quantification limit
The so-called detection limit (limit of detection LOD) and quantification limit (limit of
quantification LOQ) are important characteristics of analytical methods. They
have to be determined for each method, analyte, and matrix.
Figure5.1
3:
with
stirring
Figure
3:Liquid
Liquidjunction
junction
with
stirring
The DIN 32 645 standard defines the two terms and describes the procedure used to
values based on analytical results [5]. In this best@buchi, the definitions
by adding potassium
chloride to low concentratedcalculate
(< 4%)these
boric
4.9
of
the
aforementioned
standard are used (the terminology used in other standards may
unt of potassium chloride is always at the surface of the liquid
The stirring effect can be minimized
be slightly different).
4.7
by adding potassium chloride to low
concentrated (<of
4%)
boric
acid toamounts
ensure for
Definitions
cial for the determination
low
nitrogen
the
4.5
that a sufficient amount of potassium
Detection limit: The smallest content of the analyte that is significantly different 0
20
40
60
80
100
120
140
from
the blank value.
chloride
is
always
at
the
surface
of
the
he receiving solution by distillate is less important in low
Added
H2O [mL]
4%
2%
1%
liquidof
junction.
y of the amount
distilled water has less impact on the
pH value
ty of the blank values. The pH change related to dilution
is
Quantification
limit: The smallest content of the analyte that can be determined Figure
change
when
receiving solutionquantitatively.
with different concentrations of boric acid
The use4:ofpH
diluted
boric
acid diluting
is benefi-60 ml
cial for the determination of low nitrogen
2.
The blank
values
in less
concentrated
boric acid
smaller for
the
reason
as above.
is limit [5].
amounts
for the
following
three
main
In general,
the are
quantification
limit
is same
three times
higher
than the This
detection
particularly
important
because
usually
low
concentrated
titration
solutions
are
used
for
the
nge by dilution
of boric acid
There are two ways to calculate these limits. The results achieved from these two
reasons:
determination of low nitrogen amounts.
For the determination of low nitrogen contents it is
methods are not equal but are equivalent:
advantageous to have smaller blank values, because the difference in titration volumes between the
blanks and the samples is more important.
3
best@buchi 58/ 2010 en
3. The pH change caused by the distilled nitrogen is more important the lower the concentration of the
receiving solution is. Small amounts of nitrogen cause a considerable increase in pH, thus making the
titration more accurate.
Detection limit and quantification limit
The so-called detection limit (LOD) and quantification limit (LOQ) are important characteristics of
analytical
methods.
Theymethod”)
have to be determined
each
method,the
analyte,
and
matrix.
Direct
method
(“Blank
Experimental
k
factorfor
used
to calculate
With the determination of a large number
x (LOQ) based on x (LOD);
The
DIN 32
standard
defines
terms
and
describes
used to calculate
of
blanks
(n ≥645
10),[5]the
detection
and the
Equipment
two the
factor
is usually
k=3 the
[5]. procedure
these values limit
based
analytical
results. In this paper, the definitions of the aforementioned
AutoKjeldahl Unitstandard
K-370 with Kjeldahl
quantification
canonbe
calculated
are used
(the
terminology
used
standards
may be
slightly different).
based
on the
standard
deviations
of in
theother
Indirect
method
(“Calibration
line
Sampler K-371; Schott Titronic Uniblank measurements and the slope of the
method”)
versal, dosage instrument (Buchi P/N
Definitions:
calibration
line. For Kjeldahl the slope
043596); Analytical balance, reading
A calibration line (in the range of the limit
would be the linearity between the nitroprecision +/- 0.1 mg; Statist24cp, Verof quantification) is established (range 0
Detection limit: The smallest content oftothe
analyte that is significantly different
from the blank value.
sion 2.0., statistical program for method
gen content and the consumption of the
10 times x LOD). Based on the slope
titration solution. The calibration line of
of this line, the detection and quantivalidation for analytical laboratories,
Quantification limit: The smallest content of the analyte that can be determined quantitatively.
the entire working range is used.
©2000-2005, Georg Schmitt, Michael
fication limit can be calculated. In this
This method can only be used if a suitHerbold, Arvecon GmbH, Walldorf,
case, the uncertainty of the blank is
In general, the quantification limit is three times higher than the detection limit [1]. There are two ways
able blank is available. A blank should
Germany.
estimated by extrapolation of the
to calculate these limits. The results achieved from these two methods are not
equal but are
have
identical
properties
to
those
of
the
calibration
data.
This
method
is
more
equivalent:
Chemicals
actual sample, but without any analyte.
laborious and needs more statistical
This
is
rarely
the
case,
as
most
analyses
know-how
than
the
direct
method,
but
Ammonium dihydrogen phosphate
Direct method (“Blank method”)
are
done
in
complex
matrices
such
as
is
often
necessary
due
to
the
reasons
99.99and
% (Merck,
1.01440), dried; boric
With the determination of a large number of blanks (n ≥ 10), the limit of detection
quantification
mentioned
above.
acid
(Brenntag,
80948-155);
food
or
environmental
samples,
which
can be calculated based on the standard deviations of the blank measurements and the slope of the potassium
cannot
be imitated
easily.
calculations
are explained
in detail
chloride
(Merck,
104936); 0.05 M
calibration
line. For
Kjeldahl the slope The
would
be the linearity
between
the nitrogen
content
and the
The
calculations
detection
and The
in the
DIN 32 645
standard.
Several
acid (Riedel de Haën,
consumption
of of
thethe
titration
solution.
calibration
line of
the entire
workinghydrochloric
range is used.
quantification
limits
statistical
can Abe
usedshould
to
35320),
the titration
solutions were
This method can
onlyare
be performed
used if a suitable
blankprograms
is available.
blank
have identical
properties
according
to the
equations
- 5).
calculate
the detection
by diluting
to those of
actual(3sample,
but without
any analyte.
Thisand
is quantification
rarely the case,prepared
as most analyses
arethis standard
Neue Formeln
using the indirect
solution.
done in complex matrices such as foodlimit
or environmental
samples,method
which cannot
be imitated easily.
The calculations of the detection and quantification
limits
are performed according to equations (3 - 5).
according to DIN
32 645.
s
Samples
In this study, the LOD and LOQ were
x ( LOD ) = Φ n;α • L
(3)
(3)
always
calculated
according
to
the
direct
Solutions of ammonium dihydrogen
b
and the indirect method to be able to
phosphate were diluted to obtain an
f ln −log • R * T
slope =
compare the(2)
findings. For Kjeldahl, the
absolute nitrogen amount per sample
z • F1 1
(4)
Φ n;α = t f ;α •
+
(4)
direct method is well suited, because the
between 0.005 mg and 0.5 mg. Each
m n
matrix is completely destroyed by the
sample was determined in triplicate.
digestion with sulfuric acid.
The solution was dosed into the
(5)
(5)
x ( LOQ ) = x ( LOD ) *• k
(5)
Kjeldahl flasks using the Titronic
Universal dosage instrument.
x(LOD)
detection limit
limit
The determination was carried out
xx(LOD)
quantification
limit
(LOQ) detection
quantification
limit
xΦ
with the (m),
AutoKjeldahl
unit K-370
factor,
depending
on
number
of
blank
measurements
(n),
sample
replicates
and
(LOQ)
n,α
Zusätzlicher
Satz
auf
Seite
4:
Φ
factor,
depending
on
number
with
Kjeldahl
Sampler
K-371
using
significance
level
(
α
)
n,α
the parameters given in Table 1.
of blank
measurements
(n),of blank measurements
standard
deviation
sL
b this
slope
of the
calibration
line; for
Kjeldahl
the relation
between
titration solution
consumption
sample
replicates
(m),
and
In
study,
the
LOD
and LOQ
were
always
calculated
according
to the direct
and
and
the
nitrogen
content
(example:
14.28
ml
of
0.005
M
HCl
corresponds
to
1
mg
Nitrogen,
α
)
significance
level
(
the indirect method to be able to compare the findings. For Kjeldahl, the direct
sL
b=well
14.28).
Tablematrix
1: Parameters
for the Kjeldahl
sampler system
K-370/K-371
standard
deviation
of blank
method
is
suited,
because the
is completely
destroyed
by the
digestion
quantile of the Student t-distribution, depending on degree of freedom f (f = n-1) and
tf;α
measurements
with sulfuric
acid. level α
b
slopesignificance
of the calibration
line; for Distillation
Titration
n
number
of blank
measurements
Kjeldahl
the relation
between
m
number
of consumption
sample replicates Water
50 ml
Type
Boric acid
titration
solution
k
factor
usedcontent
to calculate
Zusätzlicher
Satz
auf
Seite 5:the x(LOQ) based on x(LOD); the factor is usually k=3 [1].
and the
nitrogen
NaOH
90 ml
Titration solution
HCl 0.005 M
(example: 14.28 ml of 0.005 M Indirect
(“Calibration
method”)
HClmethod
corresponds
to 1 set
mg upline
The
experiment
was
in the
following
way: 1)
optimization
of the
boric
acid
A
calibration
line
(in
the
range
of
the
limit
of quantification)
is5established
(range
0 toreceiving
10 times
Reaction
time
s
Volume
sol.x LOD).
60 ml
nitrogen, b= 2)
14.28).
concentration
optimization
of
the
KCl
addition
and
finally
3)
optimization
of
thecase,
Based
on
the
slope
of
this
line,
the
limit
of
detection
and
quantification
can
be
calculated.
In
this
tf;α quantile of the Student
titrant
concentration.
the uncertainty
of the blank is estimatedDistillation
by extrapolation
of the
data.
Thismode
method is more
time
240 calibration
s
Titration
Standard
t-distribution, depending on
degree of freedom f (f = n-1)
Steam power
100 %
End-point pH
4.65
and significance level α
n
number of blank measurements
Algorithm
1
m number of sample replicates
4
best@buchi 58/ 2010 en
The experiment was set up in the following way: 1) optimization of the boric acid
concentration 2) optimization of the KCl
addition and finally 3) optimization of the
titrant concentration.
of 0.01 M corresponds to the addition
of 0.75 g / liter boric acid. The pH of the
boric acid was adjusted to 4.65.
both cases, a significance level of 99 %
was used.
Results and Discussion
Impact of the titration solution
The following concentrations were
used to investigate the impact of the
titration solution on the detection
and quantification limits: 0.0025 M HCl,
0.005 M HCl, 0.01 M HCl, and 0.05 M
HCl. The analyses were carried out
using 2 % boric acid with 0.04 M KCl.
Impact of concentration of
boric acid
In Table 2, the mean values of the
blanks and their relative standard
deviation (rsd) are given using different
concentrations of boric acid. The
results in Table 2 show that the higher the
boric acid concentration, the higher the
Calculations
blank value. As shown in Figure 4, the
All calculations of the detection and
pH increase due to the dilution of the
boric acid by the distillate becomes more
quantification limits according to the
Impact of KCl addition
important the higher the concentration.
direct method (blank method) were
Different amounts
potassium chloride
Therefore, more titration solvent is
performed using the equations
Results
and of
Discussion
needed to get back to the endpoint pH
were added to 2 % boric acid. The final
(3- 5). The calculations according
Impact
of concentration
boric
acid to the indirect method (calibration of 4.65. The relative standard deviations
concentrations
of KCl inofthe
boric
Inacid
Tablewere
2, the0.01
meanM,
values
and their
relative were
standard
deviation
are given
are higher the lower the boric acid
0.02of the
M, blanks
0.04 M,
method)
carried
outrsdusing
the using
different
concentrations
acid.
concentration.
0.06 M,
and 0.1 M.of Aboric
concentration
statistical program “Statist24cp.” For
Impact of boric acid concentration
Boric acid solutions of 4 %, 2 %, and
1 % and pure water (0 % boric acid)
were compared. The pH of the boric acid
was adjusted to 4.65. To compare the
impact of the concentration of boric
acid, 10 blanks and a sample series of
5 samples with different nitrogen
contents were analyzed.
Table 2: Mean values of blank analyses with different boric acid concentrations (titration solution 0.005
M HCl, n=10)
Table 2: Mean values of blank analyses with different boric acid concentrations (titration solution 0.005 M HCl, n=10)
4 % boric acid
2 % boric acid
1 % boric acid
0 % boric acid
mean value [ml]
9.291
0.503
0.366
4 % 0.995
2%
1%
0%
sd
0.216
0.033
0.026
0.037
boric
acid
boric
acid
boric
acid
boric
acid
rsd [%]
2.3
3.3
5.1
9.5
mean value [ml]
9.291
0.995
0.503
The results in Table 2 show that the higher the boric acid concentration, the higher the blank value. As
shown
to the dilution of the boric acid
by the distillate becomes0.026
more
sd in Figure 4, the pH increase due 0.216
0.033
important the higher the concentration. Therefore, more titration solvent is needed to get back to the
rsd [%]
2.3
endpoint
pH of 4.65. The relative standard
deviations also are higher3.3
the lower the boric acid 5.1
concentration.
0.366
0.037
9.5
In Figure 5, the mean values of the recoveries of the sample series with different boric acid
In Figure 5, the mean values of the recoveries of the sample series with different boric acid concentrations are shown.
concentrations are shown.
Mean values of nitrogen recovery
130
120
110
100
90
Recovery [%]
80
70
0.01 mg N
60
0.05 mg N
50
0.2 mg N
0.1 mg N
0.5 mg N
40
30
20
10
Figure 5: Mean values (n=3,
except for 0.2 mg in 4 %
boric acid, n=1) of nitrogen
recovery in samples with their
standard deviation when
using 0 % - 4 % boric acid as
receiving solution
0
-10
4 % boric acid
2 % boric acid
1 % boric acid
0 % boric acid
Figure 5: Mean values (n=3, except for 0.2 mg in 4 % boric acid, n=1) of nitrogen recovery in samples
with their standard deviation when using 0 % - 4 % boric acid as receiving solution
5
best@buchi 58/ 2010 en
In Table 3 and Figure 6, the detection and quantification limits calculated based on the data of the blank analyses and the
sample series according to both methods proposed by DIN 32645 are shown.
Table 3: Detection limit and quantification limit calculated using the direct and indirect method.
Direct method (blank method) 1
4%
boric acid
2%
boric acid
1%
boric acid
0%
boric acid
Detection limit [mg N]
0.053
0.008
0.006
0.009
Quantification limit [mg N]
0.159
0.024
0.019
0.027
0.010
0.012
0.019
Table 3: Detection limit and quantification limit calculated using the direct and indirect method.
Direct method (blank method) 1
Indirect method (calibration method)
4 % boric acid
2 % boric acid
1 % boric acid
0 % boric acid
Detection limit [mg N]
0.053
0.008
0.006
0.009
4%
2%
1%
0%
boric acid
boric
acid
boric
Quantification limit [mg N] boric acid0.159
0.024
0.019
0.027acid
Detection
[mg N] (calibration method)
0.071
Indirectlimit
method
4 % boric acid
2 0.032
% boric acid
1 %0.041
boric acid
0 % boric
acid
0.239
0.063
Detection
limit
[mg
N]
0.071
0.010
0.012
0.019
A factor Φn,α of 3.5 was used, which is valid for 4 blanks and a triplicate determination of the samples, which are typical conditions of Kjeldahl determination;
Quantification
limitof [mg
N] the factor Φn,α0.239
0.032
0.041
0.063
would be 3.9.
for
duplicate determination
the samples,
Quantification limit [mg N]
1
Detection limit and quantification limit
0.300
Nitrogen [mg]
0.250
0.200
xLOD (direct)
xLOQ (direct)
0.150
xLOD (indirect)
xLOQ (indirect)
0.100
0.050
0.000
4 % boric acid
2 % boric acid
1 % boric acid
0 % boric acid
Figure
6: C6:
alculated
detectiondetection
and quantification
limits according
to the
direct and
method
on method
the data of the
Figure
Calculated
and quantification
limits
according
to indirect
the direct
and based
indirect
sample based
series using
0
%
4
%
boric
acid
as
receiving
solution
on the data of the sample series using 0 % - 4 % boric acid as receiving solution
Taking
into into
account
the blank
in Table 2, in
theTable
mean 2,
values
the recovery
5), and
the calculated
Taking
account
the values
blank presented
values presented
the of
mean
values rates
of the(Figure
recovery
rates
detection
and
quantification
limits
(Table
3
and
Figure
6),
it
is
evident
that
the
best
results
are
obtained
by
using
2
% boric
(Figure 5), and the calculated detection and quantification limits (Table 3 and Figure 6), it is clear
that acid.
The
blank
value
around
1
ml
is
in
a
good
range,
as
well
as
its
relative
standard
deviation
of
approx.
3
%.
The
recovery
rates and
the best results are obtained by using 2 % boric acid. The blank value around 1 ml is in a good range,
their
standard
deviations
are
better
than
those
obtained
with
the
other
concentrations
of
boric
acid.
The
calculated
detection
and
as well as its relative standard deviation of approx. 3 %. The recovery rates and their standard
quantification
limits
are
lowest,
but
comparable
to
the
ones
with
1
%
boric
acid.
The
subsequent
analyses
were
therefore
carried
deviations are better than those obtained with the other concentrations of boric acid. The calculated
outdetection
using 2 % and
boricquantification
acid.
limits are lowest, but comparable to the ones with 1 % boric acid. The
6
subsequent analyses were therefore carried out using 2 % boric acid.
Impact of KCl addition
best@buchi 58/ 2010 en
Impact of KCl addition
The measured pH shift caused by the stirring effect (using different concentrations of KCl in 2 % boric acid) is presented in Table 4.
Table. 4: pH shift caused by stirring using different KCl concentrations in 2 % boric acid, with and without dilution by distillate
not stirred
stirred
Δ pH
No KCl
4.66
4.48
-0.18
0.01 M KCl
4.66
4.60
-0.06
0.02 M KCl
4.66
4.62
-0.04
0.03 M KCl
4.66
4.64
-0.02
0.04 M KCl
4.66
4.64
-0.02
0.06 M KCl
4.66
4.64
-0.02
0.1 M KCl
4.65
4.63
-0.02
No KCl + 150 ml H2O
5.31
5.03
-0.28
0.04 M KCl + 150 ml H2O
5.38
5.34
-0.04
0.06 M KCl + 150 ml H2O
5.36
5.37
+0.01
0.1 M KCl + 150 ml H2O
5.25
5.27
+0.02
As predicted by theory (see chapter “Theoretical background of pH measurements and boric acid titration”), the addition of
potassium chloride (KCl) minimizes the stirring effect. In 2 % boric acid, a concentration of 0.03 M KCl is sufficient to decrease
the pH shift to 0.02. If the boric acid is diluted with 150 ml water, which corresponds to the approx. amount of distillate after
4 min distillation time, 0.06 M KCl is needed to minimize the stirring effect.
The amount of added KCl also influences the titration speed and the blank values (see Tables 5 and 6).
Table 5: Titration time of a blank value, 2 % boric acid with different concentrations of KCl
Titration time
No KCl
105 s
0.02 M
52 s
0.06 M
62 s
0.1 M
59 s
Table 6: Mean values of blanks (n=10) in 2 % boric acid with different concentrations of KCl.
No KCl
0.01 M
0.02 M
0.04 M
0.06 M
0.1 M
mean value
0.881
1.174
1.272
1.268
1.489
1.518
s
0.074
0.060
0.017
0.025
0.039
0.052
rsd
8.438
5.076
1.298
1.967
2.643
3.437
The more KCl is added to the boric acid, the higher the blank values. This phenomenon is related solely to the stirring effect.
If no KCl is added, the measured pH of the boric acid at the end of the distillation and start of titration is approx. 5.03 instead
of 5.34 (0.04 M KCl, see Table 4). In this case, less titration solution is needed to reach the endpoint of 4.65. Although
the blank values are higher with larger amounts of added KCl, the titration is faster due to a more stable pH measurement
(see Table 5). For the determination of low nitrogen contents it is advantageous to have smaller blank values, because
the difference in titration volumes between the blanks and the samples becomes more important. The ideal concentration
of KCl in the boric acid is a compromise between the desirable (stable pH measurement) and the undesirable (increase of
blank value).
7
best@buchi 58/ 2010 en
Figure
7 shows
the mean
valuesvalues
of the recoveries
of nitrogenofinnitrogen
the sample
Figure
7 shows
the mean
of the recoveries
in series.
the sample series.
Mean values of nitrogen recovery
160
140
Recovery [%]
120
100
0.01 mg N
80
0.05 mg N
0.1 mg N
60
0.2 mg N
0.5 mg N
40
20
0
no KCl
0.01 M KCl 0.02 M KCl 0.04 M KCl 0.06 M KCl 0.1 M KCl
Figure
7: Mean
valuesvalues
of samples
(n=3) with (n=3)
their standard
deviation
when using
0 - 0.1when
M KCl using
in 2 % boric
acid as
receiving
solution
Figure
7: Mean
of samples
with their
standard
deviation
0 - 0.1M
KCl
in 2 %
boric acid as receiving solution
The best results are obtained using 0.02 – 0.06 M KCl in 2 % boric acid. Based on the data of the
blank
analyses
and sample
series,
theM detection
limit acid.
and Based
the quantification
limit
were
calculated
(seeseries,
The
best results
are obtained
using 0.02
– 0.06
KCl in 2 % boric
on the data of the
blank
analyses
and sample
7 and
theTable
detection
limitFigure
and the 8).
quantification limit were calculated (see Table 7 and Figure 8).
Table 7: Detection limit and quantification limit calculated according to the direct and indirect method
using boric acid with different KCl concentrations
Table
7: Detection
and quantification
Direct
methodlimit
(blank
method) 2limit calculated according to the direct and indirect method using boric acid with
different KCl concentrations
No KCl
0.01 M
0.02 M
0.04 M
0.06 M
0.1 M
Detection
limit
[mgmethod)
N]
0.018
0.015
0.004
0.006
0.010
0.013
Direct
method
(blank
Quantification limit [mg N]
0.055
0.044
0.012
0.018
0.029
0.038
No KCl
0.01 M
0.02 M
0.04 M
0.06 M
0.1 M
Indirect method (calibration method)
Detection limit [mg N]
0.018
0.004
0.006M
0.010
No KCl 0.015
0.01 M
0.02 M
0.04
0.06
M
0.10.013
M
Detection limit
0.003
0.004
0.005
0.005
0.007
Quantification
limit[mg
[mg N]
N]
0.0550.014
0.044
0.012
0.018
0.029
0.038
Quantification limit [mg N]
0.046
0.010
0.014
0.016
0.018
0.025
Indirect method (calibration method)
8
No KCl
0.01 M
0.02 M
0.04 M
0.06 M
0.1 M
Detection limit [mg N]
0.014
0.003
0.004
0.005
0.005
0.007
Quantification limit [mg N]
0.046
0.010
0.014
0.016
0.018
0.025
best@buchi 58/ 2010 en
Detection limit and quantification limit
0.100
Nitrogen [mg]
0.080
0.060
xLOD (direct)
xLOQ (direct)
xLOD (indirect)
0.040
xLOQ (indirect)
0.020
0.000
no KCl
0.01 M KCl 0.02 M KCl 0.04 M KCl 0.06 M KCl 0.1 M KCl
Figure
8:C
Calculated
detection
and quantification
limitstoaccording
to the
direct
andusing
indirect
Figure 8:
alculated
detection
and quantification
limits according
the direct and
indirect
method
0 M -method
0.1 M KCl
0M
- 0.1
M KCl in
2 % boric acid as receiving solution
in using
2 % boric
acid
as receiving
solution
Taking into account the blank values presented in Table 6, the mean values of the recovery rates
(Figure 3), the minimized pH shift (Table 4), the titration time (Table 5), and the calculated detection
Taking
into account the
blank
values7presented
in Table
theclear
meanthat
values
the recovery
(Figure 7), the
and
quantification
limits
(Table
and Figure
4), 6,
it is
theofbest
resultsrates
are obtained
by minimized
using
pH
shift
(Table
4),
the
titration
time
(Table
5),
and
the
calculated
detection
and
quantification
limits
(Table
7
and
0.04 M KCl in 2 % boric acid. Although the detection and quantification limits are slightly lower Figure
with 8),
it
is
clear
that
the
best
results
are
obtained
by
using
0.04
M
KCl
in
2
%
boric
acid.
Although
the
detection
0.02 M KCl and the blank values are comparable, the mean values of the recoveries as well as the and
pH
quantification
are slightly
0.02 The
M KCl
and the blank
values are
comparable,
mean values
of the
shift
are morelimits
promising
withlower
0.04with
M KCl.
subsequent
analyses
were
thereforethecarried
out with
2
recoveries
as well
pH shift are more
promising
with 0.04 M KCl. The subsequent analyses were therefore carried
%
boric acid,
withasathe
concentration
of 0.04
M KCl.
out with 2 % boric acid, with a concentration of 0.04 M KCl.
To determine the exact detection and quantification limits according to the indirect method, the highest
To determine the exact detection and quantification limits according to the indirect method, the highest value in the
value
in the calibration line shall not exceed 10 times the detection limit. If the value turns out to
calibration line shall not exceed ten times the detection limit. If the value turns out to exceed this limit afterwards, a new
exceed this limit afterwards, a new calibration line needs to be established. In the data shown in Table
calibration line needs to be established [5]. In the data shown in Table 7, the limit of detection is 0.005 mg and 0.006 mg,
7,respectively.
the limit ofTherefore,
detectionthe
is 0.005
mg and 0.006 mg, respectively. Therefore, the highest value in the
highest value in the calibration line should not exceed 0.06 mg. Consequently, new
calibration
line
should
not
exceed
0.06 mg. Consequently, new calibration lines needed to be
calibration lines needed to be established with lower nitrogen concentrations. The final calibration line is presented in
established
with
lower
nitrogen
concentrations.
calibration
line is presented in Figure 9, the
Figure 9, the corresponding detection and quantificationThe
limitsfinal
in Table
8.
corresponding detection and quantification limits in Table 8.
9
best@buchi 58/ 2010 en
Calibration line
Determined nitrogen content [mg]
0.03
0.025
0.02
0.015
0.01
0.005
0
0.000
0.005
0.010
0.015
0.020
0.025
0.030
Nitrogen content [mg]
FigureFigure
9:Calibration
line with concentrations
between 0.0025
and 0.025
mg nitrogen.
The mean
values of The
triplicate
9: Calibration
line with concentrations
between
0.0025
and 0.025
mg nitrogen.
mean
determinations
andoftheir
standard
deviations are and
shown.
values
triplicate
determinations
their standard deviations are shown.
Table 8:
Detection limit and quantification limit calculated according to the direct and indirect
method using the data of the final calibration line.
Table 8: D
etection limit and quantification limit calculated according to the direct and indirect method using the data of the
Indirect method
Direct method
final calibration line.
(blank method)1 (calibration method)
Direct method (blank method)
Indirect method (calibration method)
Detection limit [mg N]
0.008
0.004
Detection
limit [mg N]
0.008
0.004
Quantification
limit [mg N]
0.023
0.012
Quantification limit [mg N]
0.023
0.012
Based on the calibration line data (calibration method), the detection and quantification limits are
approx.
of the line
values
bymethod),
the blankthe
method.
Considering
the recoveries
the standard
Based
on the half
calibration
dataobtained
(calibration
detection
and quantification
limits areand
approx.
half of the values
deviations of the measured concentrations of the calibration line, it is obvious that accurate
obtained by the blank method. Considering the recoveries and the standard deviations of the measured concentrations of
quantifications cannot be performed for concentrations below 0.02 mg. Only at concentrations >
the calibration line, it is obvious that accurate quantifications cannot be performed for concentrations below 0.02 mg. Only
0.0225 mg, the recoveries are around 100 % with rsds lower than 5 %.
at concentrations > 0.0225 mg, the recoveries are around 100 % with rsds lower than 5 %. In this case, the direct method
gave more realistic detection and quantification limits.
Impact of the titration solution
Impact of the titration solution
In Table 9, the mean values of the blanks and their relative standard deviation rsd are given. Figure 10
shows
recovery
rates
of the and
sample
different
concentrations
titration
solution.
In Table
9, thethe
mean
values of
the blanks
their series
relative using
standard
deviation
(rsd) are given. of
Figure
10 shows
the recovery rates
of the sample series using different concentrations of titration solution.
Table 9: Mean values of blank analyses with different titration solutions (n=10)
0.05 M with
HCldifferent0.01
M HCl
0.005 M HCl
0.0025 M HCl
Table 9: Mean values of blank analyses
titration
solutions (n=10)
mean value [ml]
0.134
0.687
1.396
2.744
0.05 M HCl
0.01 M HCl
0.005 M HCl
0.0025 M HCl
sd
0.002
0.014
0.021
0.053
mean
1.396 1.9
2.744
rsdvalue
[%] [ml]
1.80.134
2.1 0.687
1.5
sd
rsd [%]
10
0.002
0.014
0.021
0.053
1.8
2.1
1.5
1.9
best@buchi 58/ 2010 en
Mean value of nitrogen recovery
200
180
160
Recovery [%]
140
0.005 mg N
120
0.01 mg N
0.15 mg N
100
0.02 mg N
80
0.025 mg N
0.03 mg N
60
40
20
0
0.05 M HCl
0.01 M HCl
0.005 M HCl
0.0025 M HCl
Figure
10: 10:
Mean
valuesvalues
of samples
(n=3) with(n=3)
their standard
deviation
whendeviation
using 0.0025
M – using
0.05 M0.0025
HCl as titration
solution
Figure
Mean
of samples
with their
standard
when
M – 0.05
M
HCl as titration solution
The standard deviations are larger when using a higher concentrated titration solution, due the fact that very small differences
inThe
titration
volume deviations
cause large differences
the calculated
nitrogenconcentrated
content. The accuracy
the titration
consumption
standard
are larger in
when
using a higher
titrationofsolution,
due
the fact using
the
integrated
titrator
with
a
20
ml
burette
in
the
AutoKjeldahl
Unit
K-370
is
limited
to
three
digits
(e.g.,
0.001 ml). For
that very small differences in titration volume cause large differences in the calculated nitrogen
concentrations
higher
than
0.01
M
HCl,
a
higher
accuracy
would
be
necessary
to
obtain
satisfying
results.
The
disadvantage
content. The accuracy of the titration consumption using the integrated titrator in the AutoKjeldahl
Unit
ofK-370
highly diluted
titration
solutions
(0.0025
M
HCl)
is
that
large
volumes
are
titrated
(higher
costs
per
sample)
and that the
is limited to three digits (e.g., 0.001 ml). For concentrations higher than 0.01 M HCl, a higher
titration
volumes
of samples
with lowtonitrogen
(0.005
mg N The
and disadvantage
0.01 mg N) are within
the statistical
spread of the
accuracy
would
be necessary
obtain content
satisfying
results.
of highly
diluted titration
high
blank
values.
solutions (0.0025 M HCl) is that large volumes are titrated (higher costs per sample) and that the
titration volumes of samples with low nitrogen content (0.005 mg N and 0.01 mg N) are within the
The
most promising
werevalues.
0.01 M and 0.005 M HCl. With these solutions, more sample series needed to be
statistical
spreadtitration
of thesolutions
high blank
analyzed to establish a calibration line to calculate the limit of detection and quantification (data not shown). The calculated
detection
andpromising
quantification
limits are
presented
in Table
The most
titration
solutions
were
0.0110.
M and 0.005 M HCl. With these solutions, more
sample series needed to be analyzed to establish a calibration line to calculate the limit of detection
Table
Detection limit and
quantification
limit calculated
accordingdetection
to the directand
andquantification
indirect method using
titration solutions
and10:
quantification
(data
not shown).
The calculated
limitsdifferent
are presented
in
Table 10.
Direct method (blank method)
Table 10:
Detection limit and quantification
limit
according to the direct
andMindirect
0.01
M calculated
HCl
0.005
HCl
method
using
different
titration
solutions
Detection limit [mg N]
0.007
0.005
Direct method
(blank
Quantification
limit [mg
N] method)
2
Detection
limit(calibration
[mg N] method)
Indirect
method
Quantification limit [mg N]
Indirect method (calibration method)
Detection limit [mg N]
Quantification
limit [mg
[mg N]
Detection limit
N]
Quantification limit [mg N]
0.022
0.01 M HCl
0.007
0.022
0.01 M HCl
0.005
0.01 M HCl
0.013
0.005
0.013
0.005 M HCl
0.005
0.015
0.005 M HCl
0.003
0.009
0.015
0.005 M HCl
0.003
0.009
The detection and quantification limits are in the same order of magnitude as previous values (see Table 7 and Table 8). There
and
quantification
areMinHCl
theand
same
order
oftitration
magnitude
as previous values (see
isThe
also detection
no significant
difference
betweenlimits
the 0.01
0.005
M HCl
solutions.
Table 7 and Table 8). There is also no significant difference between the 0.01 M HCl and 0.005 M HCl
titration solutions.
11
best@buchi 58/ 2010 en
Conclusions
References
The detection limit and the quantification
limit are as low as approx. 0.008 mg nitrogen and 0.02 mg nitrogen when 2 %
boric acid with 3 g of KCl (0.04 M) is used
as receiving solution. Titration solutions
of 0.005 M HCl provide good results;
however, the detection and quantification
limits are not significantly influenced by
the choice of the titration solution.
The above-mentioned parameters are
suitable for low nitrogen concentrations.
For nitrogen concentrations usually found
in food samples, the standard application
using 4 % boric acid, without addition of
KCl, is recommended.
[1] Kjeldahl Guide, Buchi Labortechnik AG, 2008
[2] pH-Messung in der Praxis, Hamilton Bonaduz AG, 2007
[3]
[4] Handbook of Electrode Technology, Orion Research Incorporated, 1982
DIN 32645:2008-11 Chemical
analysis-Decision limit, detection limit and determination limit under repeatability conditions- Terms, methods, evaluation
BÜCHI Labortechnik AG
CH – 9230 Flawil 1
T +41 71 394 63 63
F +41 71 394 65 65
[email protected]
www.buchi.com
BUCHI UK Ltd.
GB – Oldham OL9 9QL
T +44 161 633 1000
F +44 161 633 1007
[email protected]
www.buchi.co.uk
BUCHI Hong Kong Ltd.
HK – Central
T +852 2389 2772
F +852 2389 2774
[email protected]
www.buchi.com.cn
Nihon BUCHI K.K.
J – Tokyo 110-0008
T +81 3 3821 4777
F +81 3 3821 4555
[email protected]
www.nihon-buchi.jp
BUCHI Canada Ltd.
CDN – Thornhill, Ontario L4J 6Z2
T +1 416 277 7407
F +1 905 764 5139
[email protected]
www.mybuchi.com
BÜCHI Labortechnik GmbH
D – 45127 Essen
Freecall 0800 414 0 414
T +49 201 747 490
F +49 201 237 082
[email protected]
www.buechigmbh.de
BÜCHI Labortechnik GmbH
NL – 3342 GT
Hendrik-Ido-Ambacht
T +31 78 684 94 29
F +31 78 684 94 30
[email protected]
www.buchi.nl
BUCHI Shanghai
RC – 500052 Shanghai
T +86 21 6280 3366
F +86 21 5230 8821
[email protected]
www.buchi.com.cn
BUCHI India Private Ltd.
IND – Mumbai 400 055
T +91 22 667 18983 / 84 / 85
F +91 22 667 18986
[email protected]
www.buchi.in
BUCHI Corporation
USA – New Castle,
Delaware 19720
Toll Free: +1 877 692 8244
T +1 302 652 3000
F +1 302 652 8777
[email protected]
www.mybuchi.com
BUCHI Sarl
F – 94656 Rungis Cedex
T +33 1 56 70 62 50
F +33 1 46 86 00 31
[email protected]
www.buchi.fr
BÜCHI Italia s.r.l.
I – 20090 Assago (MI)
T +39 02 824 50 11
F +39 02 57 51 28 55
[email protected]
www.buchi.it
BUCHI (Thailand) Ltd.
T – Bangkok 10600
T +66 2 862 08 51
F +66 2 862 08 54
[email protected]
www.buchi.com
We are represented by more than 100 distribution
partners worldwide. Find your local representative at
www.buchi.com
11592335 en 1008 / Technical data are subject to change without notice/ Quality Systems ISO 9001
The English version is the original language version and serves as basis for all translations into other languages.
[5]
12
Galster, Helmuth. pH-Messung–
Grundlagen, Methoden,
Anwendungen, Geräte. VCH
Verlagsgesellschaft GmbH,
Weinheim, 1990