Weighing according to the European

Pharmacopoeia: general considerations

and new general chapter 2.1.7. Balances
for analytical purposes

K. Fritsch1, C. Saal2, B. Spieldenner3

ABSTRACT
Weighing is a key activity in every quality control laboratory as it is one of the first steps in the
preparation of samples and reagents for most analytical procedures. It is also critical because
weighing errors will add up and propagate throughout the whole analysis, affecting the accuracy
and precision of the reported results. A new general chapter, Balances for analytical purposes
(2.1.7), has recently been published in the European Pharmacopoeia (Ph. Eur.). This new text
sets out clear requirements for an instrument that is essential to every analytical procedure
described within the pages of the Ph. Eur. This article explains in detail these requirements and
generally reviews the other quantity-related requirements present in Ph. Eur. texts.

KEYWORDS
Balances, quantities, weighing, general chapter 2.1.7, calibration, accuracy, precision, European
Pharmacopoeia.

1. INTRODUCTION
The weighing of samples and reagents is one of the first steps in many analytical procedures.
In all laboratories – but particularly those dealing with quality control of medicines – accuracy
and precision are key parameters in every weighing operation, as even very small errors can
ultimately affect the reported results. Although the General Notices of the European Pharma-
copoeia (Ph. Eur.) have always contained information on requirements for quantities and their
interpretation, there has never been a chapter specifically dedicated to this critical activity.
The Ph. Eur. General Methods Working Party took on the task of addressing this gap, and the
newly elaborated general chapter 2.1.7. Balances for analytical purposes [1] was adopted at the
168th session of the Ph. Eur. Commission in November 2020. Aligned with existing guidelines
for the use and qualification of balances published elsewhere [2-5], this chapter provides users
with comprehensive guidance on weighing for analytical purposes as described in Ph. Eur.
monographs and general chapters. General chapter 2.1.7 was included in Supplement 10.6 of
the Ph. Eur., published in July 2021, and became applicable on 1 January 2022. It will become
legally binding on 1 April 2022 when the new version of the General Notices [6], in which it is
referenced, enters into force (Supplement 10.7). It is not the aim of this article to add to the
requirements given in Balances for analytical purposes; rather, the intention is to explain in
detail the approach taken in the newly published general chapter. This article provides advice
on how to put these requirements into practice to ensure consistently high-quality weighing

1 K. Fritsch, Ad hoc specialist “Balances” in the Ph. Eur. General Methods (MG) Working Party, Mettler-
Toledo, Greifensee, Switzerland.
2 C. Saal, expert in the Ph. Eur. General Methods (MG) Working Party, Merck KGaA, Darmstadt, Germany.
3 B. Spieldenner (corresponding author’s e-mail: bruno.spieldenner@edqm.eu), European Pharmacopoeia
Department, European Directorate for the Quality of Medicines & HealthCare (EDQM), Council of Europe,
Strasbourg, France.

1
Pharmeuropa Bio&SN | January 2022

results and compliance with the Ph. Eur. It starts with a review of the Ph. Eur. requirements
applying to quantities, in particular those given in the General Notices, taking due account of
the impact of the new general chapter. Next, the section in general chapter 2.1.7 on “Equipment
Performance”, which gives the actual weighing requirements, is extensively discussed. The
article is intended as a tool for users who will implement 2.1.7. Balances for analytical purposes
into their quality management system.

2. QUANTITIES IN THE PH. EUR.

In the Ph. Eur., some requirements for quantities are given in the General Notices [6].
While the General Notices do not specifically cover considerations for weighing instru-
ments, they do explain how to interpret quantities in monographs.
A. THE 10 PER CENT-RULE

“In tests with numerical limits and assays, the quantity stated to be taken for examination cor-
responds to the quantity used during development of the analytical procedure. The amount actually
used may deviate by not more than 10 per cent from the stated quantity. In any case, the amount
used is accurately measured and the result of the test is calculated from this exact quantity.”

This rule applies only in those cases where the amount of sample actually weighed enters
into the calculation of the final result, e.g. quantification of impurities or content determination.
When a testing procedure contains such a quantitative calculation, some flexibility is allowed
for the actually weighed amount, which may deviate by up to 10 per cent from the quantity
prescribed in the Ph. Eur. (i.e. in monographs or general chapters). Performing an analytical
procedure with a quantity ± 10 per cent would be considered to be acceptable since it would still
be covered by the original validation. In such cases, it is very important to calculate the result of
the test from the actual quantity employed and this quantity must be accurately weighed.
The 10 per cent-rule only applies to the case described above. It does not apply to reagents,
which must be used in the prescribed amounts, taking due account of the rules for significant
figures detailed below under section B. Similarly, it does not apply in tests where the limit is not
numerical, but is defined by comparison with a reference substance prepared under the same
conditions instead.

B. NUMBER OF SIGNIFICANT FIGURES FOR MASSES

The General Notices also provide more general information on how quantities are described in
Ph. Eur. texts, in particular the number of significant figures given for a quantity. The number
of significant figures for a quantity implies specific requirements for the quantities (e.g. masses
and volumes) to be measured. For example, in a related substances test by liquid chromatogra-
phy, the amount prescribed for a chemical reference substance (CRS) used for peak identifica-
tion (i.e. qualitative purposes) is given without a decimal (5 mg), while the same amount for an
impurity CRS, used as external standard in a quantitative or limit test, is stated with one extra
decimal place (5.0 mg). This difference illustrates how the intended use of a quantity affects the
need for more or less accuracy in the weighing operation.
A recent major revision of the General Notices provided an opportunity to review and clarify the
section on “Quantities” and to add a cross-reference to the new general chapter on balances
(2.1.7). The new version of the General Notices was published in Suppl. 10.7, available since
October 2021, and contains the following additional paragraph:
“For masses, requirements for balances for analytical purposes are provided in general chapter
2.1.7 and are applicable to all texts. In addition, during weighing, the indication of a balance must
match the target mass value given in the text when mathematically rounded to the same number of
significant figures. For example, if the target mass value in the text is 50.0 mg, the ‘minimum weight’
(mmin) of the balance used must be smaller and the weighing is done within ± 5 subunits after the
last figure of the stated mass value (for example, 50.0 mg is to be interpreted as 49.95 mg to 50.04
mg or 49.950 mg to 50.049 mg, depending on the readability of the balance).”

2
 Weighing according to Ph. Eur.: general considerations and new general chapter 2.1.7

In addition to these requirements, if the described quantity is used in the calculation of a quan-
titative result, the 10 per cent-rule (see section A) also applies. This translates into the need to
have an amount in the range of ± 10 % that is weighed according to the number of significant
figures provided: e.g. a weighing value of 52.54 mg (rounded to 52.5 mg) would satisfy the
conditions of the example given in the extract above.

C. SPECIAL CASE: LIMITS EXPRESSED AS A RESIDUE MASS

Some tests described in monographs, such as residue on evaporation or non-volatile matter,

are used to verify that any impurity residue from manufacturing is smaller than a given limit.
This limit is often given in two formats, relative to the amount of substance to be examined (m/m
or percentage) and expressed as a maximum residue mass:
Extract from the monograph on Acetone (04/2019:0872): “Residue on evaporation: maximum
50 ppm. Evaporate 20.0 g to dryness on a water-bath and dry at 100-105 °C. The residue weighs a
maximum of 1 mg.”

The test description refers to a sample treatment and the determination of the mass of a
leftover residue; however, the requirements of general chapter 2.1.7 do not strictly apply to
the residue mass. Depending on the quality of the sample, the residue obtained in such tests
can be very small, for example, from maximum 1 mg as in the extract above to a few hundred
micrograms or even less, and it is technically impossible to satisfy the requirements of general
chapter 2.1.7 by weighing such small amounts. However, based on how the sample is handled
in the process, the residue is not weighed alone but always in a container and the value of the
residue is obtained by calculation. The actual weighing of the residue in its container yields
results in the milligram, or even in the gram range, where compliance with the requirements of
general chapter 2.1.7 is easily achievable. The amount of residue is then calculated by subtract-
ing the mass of the tared vessel.

3. EQUIPMENT PERFORMANCE
A. REQUIREMENTS OVERVIEW

The requirements for balances in general chapter 2.1.7 are divided into three major activities:
• Calibration of balances
• Performance check to assess the precision of the balance
• Performance check to assess the accuracy of the balance
Before these three requirements are analysed in detail, it is important to understand the
terminology used. In the context of general chapter 2.1.7, “accuracy” is used according to the
definition given in ICH Q2 (R1) [7]:
“The accuracy […] expresses the closeness of agreement between the value which is accepted
either as a conventional true value or an accepted reference value and the value found. This is
sometimes termed trueness.”

It should be noted that in all other fields of metrology the concept referred to in the definition
above is called “trueness” and a different definition is given for accuracy, see the International
Vocabulary of Metrology (VIM) [8].
Precision, on the other hand, is defined in a similar way in both the VIM and ICH Q2 (R1),
namely:
“The precision […] expresses the closeness of agreement (degree of scatter) between a series
of measurements obtained from multiple sampling of the same homogeneous sample under the
prescribed conditions […].”

3
Pharmeuropa Bio&SN | January 2022

Figure 1 – Distribution of results

a) Results which are accurate (true) and

precise (repeatable).
b) Results which are accurate but not
precise; they are widely scattered
around the central point.
c) Results which are precise but not
accurate; they are tightly clustered but
not around the central point (biased).
d) Results which are neither accurate,
nor precise.

In other words, the accuracy requirement described in general chapter 2.1.7 corresponds to an
assessment of the systematic error, i.e. the bias, whereas the precision requirement assesses
the random error. By defining specific acceptance criteria for both tests, the general chapter
ensures that both the random error and the systematic error of the balance are kept at low
levels. As will be detailed later, both acceptance criteria are expressed as relative limit values,
i.e. at 0.10 per cent. From a practical perspective, these requirements are quite stringent and
are meant to ensure that the weighing error is usually small, if not negligible, relative to other
errors generated in subsequent steps of the analytical procedure (such as dilution and HPLC
analysis). It is worth noting that these performance requirements are in line with the approach
described in the USP [3, 4].

B. CALIBRATION

As defined in the VIM [8], calibration is a set of activities carried out on a measuring instrument
to establish a relationship between known values (from measurement standards) and the asso-
ciated measured values (indications) given by the instrument being calibrated. This relationship
consists of deviations and their associated uncertainties.
It is evident that the relationship between the known and the measured values can only be
established if the associated measurement uncertainties are determined. Measurement uncer-
tainty quantifies how far away from the true value a measurement result might reasonably be
with a certain probability.
The calibration requirement stipulated in general chapter 2.1.7 builds on this concept by stating:
“Calibration is part of balance qualification and is performed by the user or by a suitable competent
body. Its aim is to establish traceability of measurement results to the SI units (metrological trace-
ability). The calibration results include measurement uncertainty and are documented in a calibra-
tion certificate.”
The calibration guideline EURAMET cg-18 [2] is a widely used guideline for the calibration of
balances and it is the most widespread reference document in the field of metrology of bal-
ances [9, 10].

4
 Weighing according to Ph. Eur.: general considerations and new general chapter 2.1.7

In this context, the Ph. Eur. chapter also tries to clear up a misconception about calibration –
that is, by explaining the difference between calibration and adjustment. Adjustment is defined
in the VIM as a:
“Set of operations carried out on a measuring system so that it provides prescribed indications
corresponding to given values of a quantity to be measured.”

In other words, after adjusting an instrument, its indications will correspond – as closely as
possible – to the quantity values of the measurement standards used. Unfortunately, the words
calibration and adjustment are often used interchangeably, or worse, randomly. The VIM also
points this out:
“Adjustment of a measuring system should not be confused with calibration, which is a prerequisite
for adjustment. After an adjustment of a measuring system, the measuring system must usually be
recalibrated.”

This requirement for recalibration is also highlighted in the Ph. Eur. general chapter:
“…it is recommended to perform calibration before any maintenance operation is carried out on
the balance that significantly alters its measurement performance. ‘Significant’ operations include
repairs, transfer of the balance to another location or mechanical adjustment of one or more weigh-
ing parameters. The balance must be re-calibrated after significant operations.”

Here, between the lines, the concept of “as found” and “as left” calibration is inferred to ensure
the traceability of weighing results during routine use and over the lifetime of the balance.

C. PERFORMANCE CHECKS

After setting the requirement for calibration, general chapter 2.1.7 expresses further require-
ments for the two performance checks that must be carried out between calibrations with the
frequency defined in the user’s quality management system.

a. Repeatability requirement and determination of minimum weight

This section includes a precision requirement, i.e. a repeatability test, with an acceptance crite-
rion of 0.10 per cent that the balance must meet to be considered appropriate. This requirement
is directly linked to the smallest net weight, i.e. the smallest quantity projected to be weighed
on the balance. It is important to emphasise that the smallest net weight is defined by the user,
depending on the intended applications and should not be confused with the minimum weight,
which will be described later on in this text.
The test weight for the repeatability test does not need to be chosen to match the lowest
working point of the balance. General chapter 2.1.7 recommends choosing a test load that is
not more than 5 per cent of the maximum balance capacity but also not lower than 100 mg (see
Figure 2).
A test load that is sufficiently large will be easier to handle and will avoid the introduction of han-
dling errors that could negatively influence the resulting standard deviation and thus the perfor-
mance of the balance. It is not appropriate to select a test load close to the lowest working point
of the balance when measuring repeatability. As an example, on a semi-micro balance with a
nominal capacity of 200 g and a minimum weight of about 20 mg, it is not recommended to use
a test weight of 20 mg for the repeatability test. General chapter 2.1.7 suggests using a test load
of at least 100 mg and not more than 5 per cent of the nominal capacity of the balance, e.g. 10 g
in this case. For the repeatability test, the test load does not need to be a calibrated standard
weight; any test load can be used provided its mass does not change during the test.
A very important corollary to the repeatability test is the determination of the minimum weight
(which should not be confused with the previously mentioned smallest net weight). As stated in
general chapter 2.1.7:
“The “minimum weight” is the smallest net sample mass that can be weighed on the balance, whilst
continuing to comply with the repeatability test criterion.”

5
Pharmeuropa Bio&SN | January 2022

Figure 2 – Test weight selection for repeatability testing. This selection is consistent with
other recommendations in the field [4]

Based on the expression of the acceptance criterion and using the results of the repeatability
test, the smallest net sample mass that will satisfy the repeatability test criterion can be deter-
mined. This mass is called the minimum weight and is related to s, the standard deviation of the
indicated values in the repeatability test, according to the following equation:

General chapter 2.1.7 defines a lower value of “0.41 × d” for the standard deviation, where d is
the scale interval (or readability) of the balance; this value is independent of the obtained re-
peatability test results. It stems from the rounding error associated with all weighing operations.
Indeed, a weighing operation consists of two readings (tare and net sample weight), which have
associated rounding errors. For a single reading, the rounding error is calculated as “0.29 × d”
and the propagation of the two errors by a quadratic sum gives “0.41 × d”.
This technical limit for the standard deviation also impacts the minimum weight, which cannot
be smaller than “2000 × 0.41 × d = 820 × d”. For example, on a semi-micro balance with a read-
ability of 0.01 mg the smallest possible minimum weight would be 8.2 mg. In practice, however,
the installation location and the condition of the instrument itself will affect the standard devia-
tion, resulting in a minimum weight higher than this ideal theoretical threshold. If the smallest
net weight (that the user wants to weigh) is greater than or equal to the minimum weight (as
calculated from the repeatability of the balance), the repeatability criterion expressed in general
chapter 2.1.7 is satisfied.

B. ACCURACY REQUIREMENT AND SENSITIVITY TEST

A third requirement and second routine performance check in general chapter 2.1.7 is to assess
the accuracy of the balance, which should satisfy the same 0.10 per cent acceptance criterion.

6
 Weighing according to Ph. Eur.: general considerations and new general chapter 2.1.7

However, in practice, for routine use, the general chapter proposes to test only one of the
main contributions to the accuracy error, i.e. sensitivity, with a modified acceptance criterion of
0.05 per cent. The explanation for the simplified testing strategy and the modified acceptance
criterion is given in general chapter 2.1.7:
“Accuracy is also impacted by two other parameters: eccentricity and linearity. A quadratic addi-
tion of the errors of these individual parameters, rather than a more conservative linear addition,
provides a more realistic approach to the assessment of the accuracy of the balance because the
three individual parameters are known to be largely independent of each other, and it is considered
unlikely that they will occur simultaneously and have the same algebraic sign. Therefore, the
acceptance criterion for each individual parameter can be set at 0.05 per cent, i.e. half the overall
accuracy tolerance of 0.10 per cent. While accuracy is impacted by all three parameters, the impact
of eccentricity and linearity is typically less than that of sensitivity. Hence, during a performance
check on accuracy, it can be considered sufficient to investigate sensitivity only (at 0.05 per cent),
all the more since the three individual parameters impacting accuracy are thoroughly evaluated
during calibration.”

Sensitivity characterises the change of indication of the balance divided by the change in load,
e.g. an ideal balance without a sensitivity deviation would indicate exactly the weighing value
corresponding to the applied test load. However, as a balance usually shows some sensitivity
deviation, the indication given corresponds to a larger or a smaller weighing value than the
applied test load [11].

Figure 3 – Test weight selection for accuracy testing. The test weight must be between
5 per cent and 100 per cent of the balance capacity

General chapter 2.1.7 stipulates that the mass of the test load must be at least 5 per cent of the
maximum capacity of the balance. Indeed, at the lower end of an analytical balance's measure-
ment range, the repeatability error is the most dominant contribution to the overall error of the
balance. Therefore, it is not meaningful to assess accuracy with small test weights, as any
potential systematic error would be masked entirely by the limited precision of the instrument in
that range.

7
Pharmeuropa Bio&SN | January 2022

General chapter 2.1.7 also expresses some requirements for the selection of the test loads used
for the test:
“It is generally sufficient to use the nominal weight as the test load value for the assessment, as
long as the relative maximum permissible error of the test load (i.e. the maximum permissible
error of the test load divided by the nominal weight) is not more than one third of the sensitivity test
specification (0.05 per cent). If this ratio cannot be achieved, the conventional mass value of the
test load (stated on the calibration certificate of the weight) must be used for the assessment. In this
case, the user must ensure that the weight uncertainty divided by the nominal weight is not greater
than one third of 0.05 per cent.”

This requirement is not very stringent, and in many cases weights of an OIML accuracy class
F1 or F2 (see [12], Table 1) or alternatively of an ASTM accuracy class 3 or 4, are sufficient.
For example: a microbalance with a capacity of 20 g is tested with a 20 g test weight, thereby
using its nominal mass value. The relative maximum permissible error (MPE) of the chosen
test weight must not exceed 1/3 of the sensitivity specification (i.e. about 0.017 %, resulting in
3.33 mg for a test load of 20 g). Considering a 20 g test weight of an OIML accuracy class F2
has a MPE of ± 0.8 mg (0.04 mg relative MPE), it is more than sufficient to satisfy the weight
accuracy requirement of general chapter 2.1.7.

D. USE OF BUILT-IN ADJUSTMENT WEIGHTS

In addition to verifying the performance of balances using external weights, it is accepted

practice to adjust the instruments by means of built-in reference weights so that the sensitivity
test does not have to be performed with external reference weights as frequently [4, 13-15].
At this point, it is appropriate to comment on a very frequent misconception about the perfor-
mance of balances that has been prevalent in the industry for decades. The idea that balances
need to be tested on a daily basis has been ingrained in the minds of generations of operators,
and almost everybody in the pharmaceutical industry, especially in QC, refers to “daily balance
tests”, usually for the accuracy of the balance (not the repeatability). This may date back to
the age of mechanical balances where – due to wear, tear and abrasion of critical mechanical
components – a daily check may have been justified and advisable. However, for electronic bal-
ances, and specifically those with a built-in weight adjustment, a daily test for sensitivity using
an external reference weight is a practice that appears no longer necessary.
Many Ph. Eur. texts and in particular general chapter 2.1.7, but also other references [4, 13-15]
focus on a testing approach where the quality management system of the user defines the
frequency of calibrations and of the described performance checks. It is current practice to
apply a risk-based approach for qualification activities within each quality management system.
For weighing, the impact of any inaccuracy and the required weighing process tolerance are
usually taken into consideration when using risk analysis to define the appropriate frequency.

4. SUMMARY/CONCLUSION
Accuracy and precision are critical to any quality-related weighing operation. The Ph. Eur. has
defined clear requirements in its new general chapter 2.1.7, and compliance with these should
ensure that weighing does not significantly contribute to the overall uncertainty of an analysis.
In addition to calibration of the balance, requirements for repeatability and accuracy are estab-
lished and characterised by an assessment against a specified tolerance (0.10 per cent). As
an important consequence of the repeatability test, the minimum weight can be calculated to
establish the smallest amount of net substance that can be weighed while still complying with
the tolerance requirement. The weighing requirements of general chapter 2.1.7 are generally
aligned with those given in the corresponding USP general chapters [3, 4].

5. REFERENCES
1. Balances for analytical purposes, general chapter 2.1.7. Ph. Eur. Suppl. 10.6. Strasbourg,
France: Council of Europe; 2021.

8
 Weighing according to Ph. Eur.: general considerations and new general chapter 2.1.7

2. Guidelines on the Calibration of Non-automatic Weighing Instruments. EURAMET Cali-

bration Guide No. 18, Version 4.0; 2015 Nov.
3. Balances, general chapter <41>. USP-NF. Rockville, USA: United States Pharmacopeial
Convention, Inc; 2021.
4. Weighing on an analytical balance, general chapter <1251>. USP-NF. Rockville, USA:
United States Pharmacopeial Convention, Inc; 2021.
5. Weighing According to US Pharmacopeia. White Paper 30540114A. METTLER TOLEDO;
2019.
6. General Notices, chapter 1. Ph. Eur. Suppl. 10.7. Strasbourg, France: Council of Europe;
2021.
7. Validation of Analytical Procedures: Text and Methodology [guideline]. Ref: Q2 (R1). ICH;
2005 Nov.
8. International vocabulary of metrology – Basic and general concepts and associated terms
(VIM). 3rd edition. JCGM; 2010.
9. Fritsch K. EURAMET cg-18 – state-of-the-art calibration guideline for non-automatic
weighing instruments. ACTA IMEKO 2019; 8 (3): 10-18.
10. Calibration of Weighing Instruments – Weighing in the Safe Weighing Range. White Paper
30260956. METTLER TOLEDO; 2015.
11. Nater R, Reichmuth A, Schwartz R et al. Dictionary of Weighing Terms – A Guide to the
Terminology of Weighing. Springer, Berlin, Heidelberg, 2009, ISBN 978-3-64202013-1.
12. OIML R111-1, Weights of Classes E1, E2, F1, F2, M1, M1-2, M2, M2-3 and M3, Part 1:
Metrological and Technical Requirements. International Organization of Legal Metrology;
2004.
13. Questions and Answers on Current Good Manufacturing Practices – Equipment, U.S.
Food and Drug Administration, Rockville, USA [available from: http://www.fda.gov/drugs/
guidancecomplianceregulatoryinformation/guidances/ucm124777.htm, accessed 2021
Dec 20].
14. LAB 14, Guidance on the calibration of weighing machines used in testing and calibration
laboratories. 6th edition. United Kingdom Accreditation Service; 2019.
15. ASTM E898-20, Standard Practice for Calibration of Non-automatic Weighing Instru-
ments. ASTM International; 2020.