1、Designation: E3177 18Standard Guide onSampling for Process Analytical Technology1This standard is issued under the fixed designation E3177; the number immediately following the designation indicates the year oforiginal adoption or, in the case of revision, the year of last revision. A number in pare
2、ntheses indicates the year of last reapproval. Asuperscript epsilon () indicates an editorial change since the last revision or reapproval.1. Scope1.1 This document is to be used as a guide to ProcessAnalytical Technology (PAT) instrument sampling, and coversboth the sample from which PAT data is co
3、llected and thesample that is taken for reference assay. The ASTM definitionof a guide is a compendium of information or series of optionsthat does not recommend a specific course of action. Theintention of a guide is to increases the awareness of informa-tion and approaches in a given subject area,
4、 as such this guideshould serve as a collation of points to consider when deter-mining a sample practice for PAT instruments. It is notintended to serve as a practice to be followed. As a first step,one should define the overall goal of the PAT measurement.Once defined, this guide describes various
5、considerations asthey relate to the specific requirements that must be met toachieve the overall PAT goal, including the attributes to bemeasured, impact of the scale of the process, and interfacing ofthe measurement system to manufacturing equipment (includ-ing sampling system reliability). Additio
6、nally, it discusses theestimation and validation of the effective sample size and theoverall contribution to the measurement. Related aspects ofdata collection and data processing as well as the use of riskassessments to optimize sampling and to understand the impactof potential sampling errors are
7、also covered. Furthermore,considerations for process control and aspects pertaining tosample withdrawal and retention are also included. Lastly,continuous manufacturing processes require special consider-ations due to the time dependency associated with continuousoperations as compared to batch manu
8、facturing and specialconsiderations are needed for sampling of such processes.1.2 This guide is limited to a high level overview ofsampling considerations for PAT applied to any type ofpharmaceutical manufacturing (for example, active pharma-ceutical ingredient (API), solid oral dosage form, etc.).
9、It is notintended to provide technology- or application-specific sam-pling guidance, or both. Instead, the intent is to evoke a thoughtprocess around sampling when developing a PAT application.While the focus is mainly on sampling considerations foron/in-line applications in solids, liquids, and gas
10、es (that is, insitu PAT measurements), many of the considerations also applyto at-line and off-line applications in which a sample iswithdrawn from the process and subsequently presented foranalysis.1.3 This international standard was developed in accor-dance with internationally recognized principl
11、es on standard-ization established in the Decision on Principles for theDevelopment of International Standards, Guides and Recom-mendations issued by the World Trade Organization TechnicalBarriers to Trade (TBT) Committee.2. Referenced Documents2.1 ASTM Standards:2D4177 Practice for Automatic Sampli
12、ng of Petroleum andPetroleum ProductsE105 Practice for Probability Sampling of MaterialsE122 Practice for Calculating Sample Size to Estimate, WithSpecified Precision, the Average for a Characteristic of aLot or ProcessE456 Terminology Relating to Quality and StatisticsE1402 Guide for Sampling Desig
13、nE2363 Terminology Relating to Process Analytical Technol-ogy in the Pharmaceutical Industry2.2 ASME Standard:3ASME BPE Bioprocessing Equipment3. Terminology3.1 DefinitionsFor an extensive list of terminology relatedto pharmaceutical manufacturing, refer to Terminology E2363.4. Significance and Use4
14、.1 Application of this guidance should enable PAT methoddevelopers to design and implement reliable PAT applications1This guide is under the jurisdiction of ASTM Committee E55 on Manufactureof Pharmaceutical and Biopharmaceutical Products and is the direct responsibility ofSubcommittee E55.01 on Pro
15、cess Understanding and PAT System Management,Implementation and Practice.Current edition approved Sept. 1, 2018. Published September 2018. DOI:10.1520/E1377-18.2For referenced ASTM standards, visit the ASTM website, www.astm.org, orcontact ASTM Customer Service at serviceastm.org. For Annual Book of
16、 ASTMStandards volume information, refer to the standards Document Summary page onthe ASTM website.3Available from American Society of Mechanical Engineers (ASME), ASMEInternational Headquarters, Two Park Ave., New York, NY 10016-5990, http:/www.asme.org.Copyright ASTM International, 100 Barr Harbor
17、 Drive, PO Box C700, West Conshohocken, PA 19428-2959. United StatesThis international standard was developed in accordance with internationally recognized principles on standardization established in the Decision on Principles for theDevelopment of International Standards, Guides and Recommendation
18、s issued by the World Trade Organization Technical Barriers to Trade (TBT) Committee.1that avoid many common sources of error around sampling.Sampling is a key element of method and process validationplans.4.1.1 Many ASTM standards discuss sampling; however,almost all are very specific to a certain
19、field or application. Forexample, the “Standard Practice for Automatic Sampling ofPetroleum and Petroleum Products” (D4177) specifically cov-ers information for the design, installation, testing, and opera-tion of automated equipment for the extraction of representa-tive samples of petroleum and pet
20、roleum products from aflowing stream and storing them in a sample receiver.4.1.2 Other useful ASTM standards include: E105 (Practicefor Probability Sampling of Materials), E122 (Standard Prac-tice for Calculating Sample Size to Estimate, With a SpecifiedPrecision, the Average for a Characteristic of
21、 a Lot or Process),E1402 (Standard Guide for Sampling Design), and E456(Terminology Relating to Quality and Statistics). These stan-dards review similar considerations as those addressed in thisguidance and can be consulted for additional insight on how todeal with specific sample types or situation
22、s. However, suchstandards should be carefully reviewed for relevance to phar-maceutical applications.5. Summary of Practice5.1 Representative sampling is a key aspect of successfulPAT measurements. There are many aspects to be considered todevelop a suitable sampling approach. Scientific and statist
23、icalprinciples should be used in combination with appropriate riskassessment tools to ensure that the sampling approach issuitable for the application.5.2 This guide is organized into sections each of whichdescribes a particular aspect of sampling practices for PATapplications. Presented below is a
24、brief description of each ofthe sampling aspects as well as the key objective to beaddressed.5.2.1 Attribute to be measured (see Section 6): Attribute(s) of interest, Scale or physical characteristic of the attribute: macro-scopic versus microscopic, and Direct or indirect measurement.5.2.1.1 The ke
25、y objective is to clearly define the attributethat is being measured.5.2.2 Process scale and nature (see Section 7): Scale, and Dynamics.5.2.2.1 The key objective is to understand the impact of thescale and dynamics of the process on sample size, frequency ofsampling, and sampling locations.5.2.3 Es
26、timation of the mass of sample investigated (effec-tive sample size) (see Section 8): What is the area or volume under scrutiny? Depth of penetration? Numbers of replicate measurements to achieve therequired signal to noise ratio and capability of the measurementsystem?5.2.3.1 The key objective is t
27、o establish that the effectivesample size and the level of scrutiny (degree of examination)are appropriate.5.2.4 Interfacing of measurement systems to manufacturingequipment (see Section 9): Locations of sensors, Numbers of sensors, Rationale for position of sensors, Mechanical interfacing: Effect o
28、f sample interface/probe on the process, and Cleaning of sensors. Effect of time/temperature/other parameters on thesample interface.5.2.4.1 The key objective is to establish how representativethe sampling schedule is of the process under investigation,that is, does the sampling plan ensure that the
29、 pertinentvariability in the process is captured. Also, there is a need toestablish that the sampling interface is stable to changes in theprocess and material characteristics of the sample that may beencountered during normal operation. Furthermore, it has tobe ensured that the sample interface its
30、elf has no impact on themanufacturing process and product itself.5.2.5 Rationalization of the contribution of a sample to ameasurement (see Section 10): Sample mass contributing to a single measurement; Heterogeneity of the sample material at the microscopiclevel; Speed of analysis, data transfer ra
31、te, and relativedisplacement of the sample; Measurement reliability; and How representative the measured sample is of theprocess or product, or both.5.2.5.1 The key objective is to ensure that the validation ofa PAT sampling system is focused on the appropriate param-eters.5.2.6 Measurement cycle ti
32、me (see Section 11): Frequency of measurement, and Numbers of measurements to be averaged, etc.5.2.6.1 The key objective is to establish that the timescale ofthe measurement is appropriate relative to the timescale of theprocess.5.2.7 Risk assessment (see Section 12): Use of appropriate risk assessm
33、ent tools.5.2.7.1 The key objective is to ensure that the risks ofmaking a sampling error are assessed and mitigated.5.2.8 Process control (see Section 13): Impact of sampling on the ability to control a process.5.2.8.1 The key objective is to establish the sample size thathas the ability to reliabl
34、y separate signal from noise for thepurpose of process control.5.2.9 Sample withdrawal and retention prior to referenceanalysis4(see Section 14): Time between PAT measurement and referenceanalysis, and Procedure for sample withdrawal.5.2.9.1 The key objective is to consider the impact of samplewithd
35、rawal and time between PAT measurement and reference4In this guide, reference analysis is defined as the (chemical or physical, or both)analysis of a sample withdrawn from a process, typically after a PAT measurementhas been performed on it, to establish the reference value for the PAT measurement.T
36、his analysis is often done in a laboratory using conventional analytical techniques.E3177 182analysis. This covers the stability of the sample between thetime of removal from the process until time of analysis suchthat the sample analyzed is representative.5.2.10 Continuous processing (see Section 1
37、5): Time dependency of continuous processes.5.2.10.1 The key objective is to recognize that continuousprocesses require special considerations due to their time-dependent nature.6. Attribute to be Measured6.1 When devising a sampling interface, device, or plan forany PAT method, one of the first asp
38、ects that has to beconsidered is the attribute of interest, that is, what is itspecifically that is going to be measured.6.1.1 The physical scale of the attribute is of significantimportance as the sampling strategy may change depending onwhether the attribute to be measured is microscopic (forexamp
39、le, excipient distribution, morphology) or macroscopic(for example, crystal or granule properties such as density,size/distribution, etc.) in nature; this ties with the physicalproperties of the material(s) being examined.6.1.2 Additionally, the type of measurement has to be takeninto account becaus
40、e the sampling requirements for a directmeasurement, that is, attribute of interest is measured directly,can be different as compared to those for an indirectmeasurement, that is, attribute of interest is derived from themeasurement of another (set of) attribute(s) or process param-eters.6.2 If mult
41、iple attributes are to be determined by means ofa single measurement process or system, then the samplingplan has to cover requirements associated with all the indi-vidual attribute measurements. The goal is to implement theappropriate PAT measurement system(s) and associated sam-pling plan with the
42、 appropriate sensitivity for the attribute ofinterest and ruggedness/insensitivity with respect to interfer-ences from other factors.7. Impact of Scale of the Process7.1 The scale of the manufacturing process may have animpact on the sampling requirements.7.1.1 During development at small scale, the
43、 samplingfrequency may be higher than at full commercial scale asproduct and process knowledge and understanding are thefocus. At commercial scale a lower frequency of sampling maybe appropriate if the process is well characterized, understood,predictable, and controlled.7.2 The fact that commercial
44、 scale manufacturing may besubject to an increased likelihood of subpopulations (substrata)which would increase sample-to-sample variations should beconsidered. Sample heterogeneity would have to be taken intoaccount in this case. Further, depending on the size andphysical structure of the manufactu
45、ring equipment used, adifferent number of sensors may be needed to accomplish thesame measurement at small and large scale.Additionally, as thesize and physical structure of the manufacturing equipmentchanges not only the number of sensors may need to change,but also the location thereof. Lastly, de
46、pending on differencesin process dynamics related to scale, different sampling plansmay be required (see also Section 10). Note though that a largermanufacturing scale does not always automatically necessitatesan increase in number of sensors, a change in location, or adifferent sampling plan; this
47、will depend on the process,system, and PAT measurement.8. Estimation of the Amount of Sample Which is BeingInvestigated (Effective Sample Size)8.1 On-line and in-line PAT measurement applications typi-cally do not involve removal of samples from the system orprocess. Measurements are generally made
48、using sensors orprobes that are in direct contact with, or inserted into thesystem or process.8.1.1 However, even though there may be no physicalremoval of samples from the system or process, all such PATmeasurement techniques are effectively evaluating a sub-set ofthe material under investigation.
49、This derives from the fact thatsuch techniques have a limited field of view or operation; forexample, they will penetrate a sample matrix or process to afinite depth and can only make measurements at a finite rate.8.1.2 It is recognized that estimating or calculating aneffective sample size for analysis by the PAT system may bedifficult. For powders or solids it may be possible to approxi-mate the effective sample size for a spectroscopic techniqueusing some reasonable assumptions. In such cases, forexample, the effective sample size can be a function
