Calculate results from titrations and gravimetry, compare methods statistically, read chromatograms and spectra, validate methods to ICH Q2, plan sampling, and work in GMP, GLP, EPA and ISO 17025 labs.

Analytical Chemistry Methods

## Why it matters

A method that works once on the developer's bench is not yet a method. Before results are used to release a medicine, certify drinking water or support a publication, the lab must show with data that the method is **fit for its intended purpose**. That evidence is **method validation**. In the pharmaceutical world it follows the ICH guideline **Q2(R2), Validation of Analytical Procedures**, with Q14 covering method development ([ICH quality guidelines](https://www.ich.org/page/quality-guidelines), checked on 2026-10-23). Environmental and ISO 17025 labs use very similar characteristics.

## The validation characteristics

| Characteristic | Question | Typical evidence |
|---|---|---|
| **Specificity** | Does it measure only the analyte? | Blank, placebo, spiked impurities, stressed samples; peak purity |
| **Linearity** | Is the response proportional to concentration? | At least 5 levels; slope, intercept, residuals, r^2 |
| **Range** | Over which interval are accuracy, precision and linearity acceptable? | For an assay often 80 to 120 % of the target |
| **Accuracy** | Is the result close to the true value? | Recovery of spiked known amounts, or a CRM |
| **Precision** | How close are repeated results? | Repeatability (same day, analyst, instrument), intermediate precision (different days, analysts, instruments), reproducibility (between labs) |
| **Detection limit (LOD)** | What is the smallest amount that can be detected? | Signal-to-noise about 3, or 3.3 sigma / slope |
| **Quantitation limit (LOQ)** | What is the smallest amount that can be measured with acceptable accuracy and precision? | Signal-to-noise about 10, or 10 sigma / slope |
| **Robustness** | Does it survive small deliberate changes? | Vary pH, temperature, flow rate, column lot |

Not every method needs every characteristic. An identity test needs specificity; an impurity limit test needs specificity and LOD; a quantitative impurity method and an assay need most of the table.

## The protocol comes first

Validation is planned in a **protocol** with predefined **acceptance criteria** (for example "mean recovery 98.0 to 102.0 %", "repeatability RSD at most 1.0 %"). Results are compared with those criteria in a **report**. Setting criteria after seeing the data is not validation.

## Worked example: linearity, LOD and LOQ

A fictional UV method is calibrated at six levels:

| Concentration (mg/L) | 0 | 2 | 4 | 6 | 8 | 10 |
|---|---|---|---|---|---|---|
| Absorbance | 0.003 | 0.104 | 0.203 | 0.305 | 0.402 | 0.503 |

Least-squares regression gives:

- slope S = **0.04994** AU per mg/L, intercept = **0.0036** AU
- r^2 = **0.99997**
- residual standard deviation (standard error of the regression, n - 2 degrees of freedom) sigma = **0.00113** AU

Using the ICH calibration-curve approach:

- LOD = 3.3 x sigma / S = 3.3 x 0.00113 / 0.04994 = **0.075 mg/L**
- LOQ = 10 x sigma / S = 10 x 0.00113 / 0.04994 = **0.23 mg/L**

The LOQ must then be **confirmed** experimentally: analyze samples near 0.23 mg/L and show acceptable accuracy and precision there.

Two cautions about r^2. A high r^2 does not prove linearity (a curved response can still give r^2 above 0.99), so always **plot the residuals**: they should scatter randomly around zero with no pattern. And r^2 says nothing about accuracy.

## Accuracy as recovery

Spike a placebo or sample with known amounts at, for example, 80, 100 and 120 % of target, in triplicate:

```text
recovery % = (found - unspiked) / added x 100
```

If a sample containing 5.00 mg/L is spiked with 2.00 mg/L and measures 6.96 mg/L, recovery = (6.96 - 5.00) / 2.00 x 100 = **98.0 %**.

## Verification, transfer and lifecycle

- **Verification:** a compendial method (for example from the USP) is not revalidated, but the lab shows it works in its hands.
- **Transfer:** a validated method moved to another lab is shown to give equivalent results there.
- **Revalidation:** needed after significant changes (new instrument type, formulation change, method change). The ICH Q14 / Q2(R2) approach treats this as a **lifecycle** with ongoing monitoring.

## Common mistakes

- Treating r^2 as proof of linearity without looking at residuals.
- Calculating LOQ but never confirming it.
- Writing acceptance criteria after the experiments.
- Validating precision with replicate injections of one solution (that is system precision) instead of independent sample preparations.

## Quick recap

- Validation shows a method is fit for purpose: specificity, linearity, range, accuracy, precision, LOD, LOQ, robustness.
- Predefine acceptance criteria in a protocol.
- LOD = 3.3 sigma / S; LOQ = 10 sigma / S; confirm LOQ experimentally.
- Recovery % = (found - unspiked) / added x 100.

## Try it

Open the **Do** tab. A NumPy validation script fits the calibration; you write the LOD and LOQ formula.


Define the ICH Q2 validation characteristics and calculate linearity, LOD, LOQ and recovery from data.

Method validation

Understanding ICH Q2(R2) Guidelines for Analytical Validation | Complete Overview

Chapter 5: Limit of Detection and Limit of Quantitation | CHM 214 | 040

Mix about 9 tablespoons of rice with 1 tablespoon of lentils in a jar. Do not shake it yet. This is your fictional 'lot' with a true lentil fraction of about 10 % by volume.

Take a 'grab sample': one level teaspoon from the top. Count the lentils and the rice grains (or estimate by volume) and record the lentil fraction.

Pour everything onto a clean plate or tray, mix it well, pile it into a cone, flatten it, and divide it into four quarters. Keep two opposite quarters and set the others aside.

Repeat the coning and quartering once more, then take a teaspoon from the kept portion and record the lentil fraction again.

Take three more teaspoons from different places after mixing, record each fraction, and calculate the mean and the spread.

Compare the grab sample with the quartered samples. Which is closer to 10 %? Which is more consistent?

On paper, write a sampling plan for a fictional lot of 200 drums of a powdered raw material: how many drums you would sample, how you choose them, how you take increments from each drum, how you reduce the sample, and which blanks or duplicates you would add.

Put the rice and lentils back in the kitchen (they are fine to cook after rinsing) and clean up.

## Why it matters

The most precise instrument in the world cannot rescue a sample that does not represent the material it came from. In many analyses, **sampling and sample preparation contribute more uncertainty than the measurement itself**. A truckload of ore, a batch of tablets, a lake, a field of soil: the analyst may test a gram, or a milliliter, and the result is supposed to describe all of it. Planning how to get from "the lot" to "the test portion" without bias is part of the analytical method, not an afterthought.

> **Safety and scope.** Real sampling and preparation use hazardous reagents, solvents, digestion acids and sometimes hazardous materials. They are done only in a teaching or employer lab, with your instructor or supervisor and the lab's safety training. The activity uses dry kitchen foods only (rice and lentils), nothing hazardous.

## From lot to test portion

1. **Lot (population):** everything the result must describe.
2. **Gross (bulk) sample:** increments taken from the lot according to a sampling plan.
3. **Laboratory sample:** the gross sample reduced to a size the lab can handle.
4. **Test portion:** the amount actually weighed or measured into the analysis.

Each reduction step must keep the composition. That requires **homogenizing** (grinding, mixing, blending) and **dividing** properly.

## Representative sampling

- **Random sampling:** every part of the lot has an equal chance of being sampled. Avoids the "grab the top of the drum" bias.
- **Systematic sampling:** at regular intervals (every 10th container, every 15 minutes from a pipe). Simple, but can miss periodic variation.
- **Stratified sampling:** the lot is divided into known layers or groups (top, middle, bottom of a tank; each production shift) and each is sampled. Useful when the lot is known to be heterogeneous.
- **Composite sample:** several increments combined into one for an average result. Cheaper, but hides variation between increments.

**Heterogeneity** drives everything. A well-mixed solution is homogeneous, so one sample is enough. A powder blend, a soil or an ore is particulate: the more variable the particles and the larger they are, the bigger the sample needed to keep sampling error small. That is why grinding to a smaller particle size before sub-sampling reduces error.

## Dividing a sample

- **Coning and quartering:** pile the material into a cone, flatten it, divide it into four quarters, keep two opposite quarters, and repeat.
- **Riffle splitter:** a box of alternating chutes that splits a stream of particles into two halves.
- **Rotary divider:** the most reproducible method for powders.

Scooping from the top of a jar is not dividing: fine particles settle and large ones rise when a powder is shaken.

## Preservation and holding times

Samples change after collection: volatile compounds escape, microbes consume organics, metals stick to container walls. Methods specify **containers** (glass vs plastic), **preservatives** (for example acidification for dissolved metals), **temperature** (often around 4 degrees C) and **maximum holding times**. A sample analyzed after its holding time is usually reported as compromised.

## Sample preparation

The goal is to get the analyte into a form the instrument can measure, free of interferences:

- **Dissolution and dilution** of soluble samples.
- **Acid digestion** (including microwave digestion) to destroy organic matter for metals analysis.
- **Liquid-liquid extraction** and **solid-phase extraction (SPE)** to isolate and concentrate an analyte.
- **Filtration** and **centrifugation** to remove particles that would block a column.
- **Derivatization** to make a compound volatile for GC or detectable for UV.

## Quality controls in sampling and prep

- **Field blank:** clean matrix taken to the site and handled like a sample; detects contamination from containers and transport.
- **Method blank:** carried through the whole preparation; detects lab contamination.
- **Matrix spike:** a known amount added to a real sample; shows recovery in that matrix.
- **Duplicates:** two samples or two preparations; show sampling and preparation precision.

## Worked example: a tablet assay

The lot is 500,000 tablets. The plan takes **20 tablets at random** across the batch, weighs them, grinds them to a uniform powder, and weighs out a test portion equal to about one average tablet weight, in duplicate. The result is reported per tablet using the average tablet weight. Testing a single tablet would tell you about that tablet (content uniformity is a separate test), not the batch average.

## Common mistakes

- Sampling only the most accessible part of a lot.
- Sub-sampling a powder without mixing or proper division.
- Ignoring holding times or preservation.
- Skipping blanks and spikes, so contamination and losses go unseen.

## Quick recap

- Lot, gross sample, laboratory sample, test portion: each step must keep the composition.
- Random, systematic, stratified and composite sampling suit different lots.
- Reduce particle size and divide properly (coning and quartering, riffles).
- Preserve samples, respect holding times, and use blanks, spikes and duplicates.

## Try it

Open the **Do** tab. You will see sampling error for yourself with rice and lentils.


Plan representative sampling from lot to test portion and choose preservation, preparation and QC samples.

Sampling and prep

Sampling (Steps in Chemical Analysis)—Part 1

Sampling in Analytical Chemistry Explained | Why Representative Samples Matter.

## Why it matters

The same HPLC method can run in a university research lab, a pharmaceutical QC lab, a contract environmental lab and an accredited testing lab, but the rules around it differ enormously. Who approves the method, how data are recorded, what happens when a result is out of specification, and who inspects the lab all depend on the **regulatory framework**. Analytical chemists move between these sectors, and employers expect you to know which rules apply where.

> **Requirements drift.** Regulations and standards are revised. Use the official links (checked on 2026-10-23) for current text rather than summaries.

## GMP: making medicines

**Good Manufacturing Practice** (in the US, current GMP or cGMP, [21 CFR Parts 210 and 211](https://www.ecfr.gov/current/title-21/chapter-I/subchapter-C/part-211); FDA overview: [CGMP regulations](https://www.fda.gov/drugs/pharmaceutical-quality-resources/current-good-manufacturing-practice-cgmp-regulations)) governs the manufacture and testing of medicines. The QC lab tests raw materials, in-process samples, finished products and stability samples against **specifications**.

Key features:

- Methods are **validated** (ICH Q2) or verified (compendial methods such as USP).
- Instruments are **qualified** (IQ/OQ/PQ) and calibrated.
- **Out-of-specification (OOS)** results trigger a formal investigation: first a lab investigation for an assignable cause (calculation error, instrument fault), then, if none is found, a full investigation including manufacturing. Retesting until a passing result appears ("testing into compliance") is prohibited.
- **Data integrity** follows **ALCOA+**: data must be Attributable, Legible, Contemporaneous, Original and Accurate, plus complete, consistent, enduring and available. Electronic records follow 21 CFR Part 11 (audit trails, access control, electronic signatures).
- The FDA (and other national regulators) inspect.

## GLP: safety studies

**Good Laboratory Practice** ([21 CFR Part 58](https://www.ecfr.gov/current/title-21/chapter-I/subchapter-A/part-58), and the OECD GLP principles internationally) governs **nonclinical** safety studies submitted to regulators, such as toxicology studies and the bioanalysis that supports them. It is about the integrity and reconstructability of a **study**:

- A **study director** is the single point of responsibility.
- A **protocol** is approved before the study starts, and deviations are documented.
- An independent **quality assurance unit (QAU)** inspects the study and audits the final report.
- Raw data, specimens and records are archived so the study can be reconstructed years later.

GLP is not the same as "good lab habits," and it does not apply to routine product release testing (that is GMP).

## EPA methods: environmental testing

Environmental labs measure pollutants in water, soil, air and waste for regulatory compliance. In the US they use **EPA-approved methods**: for example the Clean Water Act methods in 40 CFR Part 136 ([EPA CWA methods](https://www.epa.gov/cwa-methods)), drinking-water methods, and SW-846 for solid waste. Features:

- Methods are often **prescriptive**: specified sample containers, preservation, holding times, calibration and QC (method blanks, laboratory control samples, matrix spikes, duplicates).
- Results carry **qualifiers** when QC fails.
- Labs are usually **certified or accredited by state programs** (and many through NELAP/TNI) for each method and matrix.
- Chain of custody is routine, because results may end up in enforcement cases.

## ISO/IEC 17025: competence of testing labs

**ISO/IEC 17025** ([ISO 17025](https://www.iso.org/ISO-IEC-17025-testing-and-calibration-laboratories.html)) is an international standard for the competence of testing and calibration laboratories in any sector: food, materials, forensics, calibration, environmental. Labs are **accredited** by bodies such as A2LA or ANAB in the US. Features:

- A management system (document control, corrective action, internal audits, management review).
- Validated or verified methods and trained, authorized staff.
- **Metrological traceability** of measurements to SI units through calibrated equipment and reference materials.
- **Measurement uncertainty** estimated for results.
- **Proficiency testing**: analysing blind samples from an external provider and comparing results with other labs.
- Impartiality and confidentiality requirements.

## Side-by-side

| | GMP | GLP | EPA methods | ISO/IEC 17025 |
|---|---|---|---|---|
| Purpose | Product quality (medicines) | Integrity of nonclinical safety studies | Environmental compliance | Lab competence, any sector |
| Unit of work | Batch / product | Study | Sample / permit | Test or calibration |
| Key roles | QC unit, QA | Study director, QAU | Lab and state certification | Quality manager, technical management |
| Failure process | OOS investigation | Protocol deviation | Qualified data, corrective action | Nonconforming work, corrective action |
| Overseen by | FDA and other regulators | FDA, EPA, OECD monitoring authorities | EPA and states | Accreditation bodies |

## Working in these labs

Employers typically ask for a **bachelor's degree in chemistry** (ACS-approved programs are preferred by many employers; see [ACS approval](https://www.acs.org/education/policies/acs-approval-program.html)) with substantial hands-on lab courses, and they provide sector-specific GMP, GLP or quality-system training on the job. Research leadership roles usually need an MS or PhD.

## Common mistakes

- Calling any careful lab "GLP." GLP is a specific regulation for nonclinical studies.
- Retesting an OOS result without an investigation.
- Assuming ISO 17025 accreditation covers every test a lab offers (it covers a defined scope).
- Ignoring holding times and QC requirements written into EPA methods.

## Quick recap

- GMP: product quality, OOS investigations, ALCOA+ data integrity.
- GLP: nonclinical safety studies, study director, protocol, QAU.
- EPA methods: prescriptive environmental methods with built-in QC and state certification.
- ISO/IEC 17025: competence, traceability, uncertainty, proficiency testing; accreditation with a defined scope.

## Try it

Open the **Do** tab for the quiz: match situations to the framework that applies.


Compare GMP, GLP, EPA-method and ISO/IEC 17025 laboratories and what each requires of an analyst.

Regulated labs

What is Good Manufacturing Practice GMP in Pharmaceuticals?

Good Laboratory Practice (GLP) Regulation for Pharmaceuticals

Validate a method to ICH Q2, plan representative sampling and preparation, and compare regulated lab frameworks.