CoChat

The role of Raman spectroscopy in biopharmaceuticals from development to manufacturing

Open access
Esmonde-White, Karen A., Cuellar, Maryann, Lewis, Ian R. · 2022 · Analytical and Bioanalytical Chemistry

The DOI matches its Crossref record; 113 of the 114 references it cites were found in a registry or an index; 26 of the 115 sources its claims cite were checked for support.

Abstract

Biopharmaceuticals have revolutionized the field of medicine in the types of active ingredient molecules and treatable indications. Adoption of Quality by Design and Process Analytical Technology (PAT) frameworks has helped the biopharmaceutical field to realize consistent product quality, process intensification, and real-time control. As part of the PAT strategy, Raman spectroscopy offers many benefits and is used successfully in bioprocessing from single-cell analysis to cGMP process control. Since first introduced in 2011 for industrial bioprocessing applications, Raman has become a first-choice PAT for monitoring and controlling upstream bioprocesses because it facilitates advanced process control and enables consistent process quality. This paper will discuss new frontiers in extending these successes in upstream from scale-down to commercial manufacturing. New reports concerning the use of Raman spectroscopy in the basic science of single cells and downstream process monitoring illustrate industrial recognition of Raman’s value throughout a biopharmaceutical product’s lifecycle. Finally, we draw upon a nearly 90-year history in biological Raman spectroscopy to provide the basis for laboratory and in-line measurements of protein quality, including higher-order structure and composition modifications, to support formulation development. Graphical abstract

Overview

CoChat Summary

TL;DR

This review argues that Raman spectroscopy has become a first-choice process analytical technology for biopharmaceutical upstream bioprocessing and is expanding into downstream monitoring, formulation, and higher-order structure analysis. It synthesizes regulatory, instrumentation, chemometric, and application literature spanning single-cell studies to cGMP manufacturing.

Key findings

  • Raman is described as a first-choice PAT for upstream bioprocessing since its first industrial upstream description in 2011, with advantages in specificity, aqueous compatibility, and in-line sampling; applications now include glucose, lactate, amino acids, pH, cell attributes, and product concentration.
  • Generic Raman models can transfer across cell lines and scales: in Webster et al. 2018, glucose models gave R²p of 0.99, 0.98, and 0.99 with RMSEP of 0.47, 0.43, and 0.41 g/L across two new cell lines and a 10 L scale, while glutamate and product concentration were exceptions. Local just-in-time learning models improved RMSE from 4.4% (sodium) to 44.89% (ammonium) for PLS and from 9.2% (sodium) to 53.46% (ammonium) for Gaussian process models.
  • Downstream applications are emerging: Yilmaz et al. reported an IgG model RMSEP of 0.18 g/L across IgG subclasses; Feidl et al. reported an optimized model average RMSEP of 0.12 mg/mL within 0–2.82 mg/mL and an average limit of detection of 0.24 mg/mL; Rafferty et al. found Raman-based feed volumes were within 6% or less of capacitance-based feed volumes after day 3.
  • Raman has a nearly 90-year history in biological spectroscopy, supporting protein higher-order structure and formulation work; examples include deep-UV resonance Raman for mAb structure, possible glycation markers at ~1040 cm⁻¹ and 1121 cm⁻¹, and lysozyme fibril formation studies.

Methods

  • Design: narrative/critical review, not a primary experimental study; it updates the authors’ 2016 review and synthesizes literature on Raman in biopharmaceutical development and manufacturing.
  • Data or sample: cited studies cover CHO cells, T-cells, E. coli, mAbs, IgG, lysozyme, and other proteins, with scales from high-throughput miniature bioreactors (15 μL–250 mL) to pilot/commercial manufacturing (>20 L to thousands of L); one cited study spanned 35 cultivations, 4 CHO cell lines, 8 clones, 4 scales, and 2 cultivation modes.
  • Main analysis: thematic synthesis of regulatory frameworks, Raman variants and instrumentation, chemometric model development, and applications in upstream monitoring/control, downstream monitoring, and product formulation/higher-order structure; tables summarize model figures of merit and preprocessing approaches.

Limitations

  • The paper states that Raman has inherent drawbacks, including the “excessive feebleness” of the effect and competition with fluorescence; fluorescence-reduction and enhancement variants such as time-gated Raman and SERDS are described as early-stage, often limited to off-line or laboratory use, and not clearly advantageous over spontaneous Raman at higher wavelengths or software-based correction.
  • The paper states that downstream publications have been limited until recently, partly because downstream operations are fast and cycle time limits spectral acquisition with standard immersion probes, and because finding a cost-effective benefit over existing measurements such as UV–Vis can be difficult; SERS is also noted as having an acknowledged limitation in translating enhancement into reliable, quantifiable measurement.
  • Evident from the text: as a review, it does not present new experimental data or a systematic meta-analysis, and its cited model performance values come from heterogeneous studies with different cell lines, scales, preprocessing, and reference methods.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1007/s00216-021-03727-4

Cites

113 of 114 references verified0 citation errors · 1 not checkable

Statistics

Nothing to recompute.

Figures

Fig. 1 Schematic of Raman-based flow cell measurements. In one design (a), Raman signal was collected from a cross-section of a cuvette in which liquid was streamed. Within the measurement chamber, a non-contact optic was used to focus laser light into a cuvette and collect back-scattered Raman signal. Transmitted photons were focused back into the cuvette using a concave mirror on the opposite side of the cuvette from the non-contact optic. In another design (b), laser light was delivered into a measurement chamber by fiber optics. A fiber adaptor and non-contact objective were used to focus laser light into the flow path for measurements in a longitudinal direction. A reflector at the end of the flow path was used to focus light back into the flow path. In both cases, the fiber optic probe used a backscattered fiber geometry. Figure permissions: panel a was reused by permission of the publisher, John Wiley and Sons. Panel b was used under the open-access license. © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open-access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license
Fig. 1 Schematic of Raman-based flow cell measurements. In one design (a), Raman signal was collected from a cross-section of a cuvette in which liquid was streamed. Within the measurement chamber, a non-contact optic was used to focus laser light into a cuvette and collect back-scattered Raman signal. Transmitted photons were focused back into the cuvette using a concave mirror on the opposite side of the cuvette from the non-contact optic. In another design (b), laser light was delivered into a measurement chamber by fiber optics. A fiber adaptor and non-contact objective were used to focus laser light into the flow path for measurements in a longitudinal direction. A reflector at the end of the flow path was used to focus light back into the flow path. In both cases, the fiber optic probe used a backscattered fiber geometry. Figure permissions: panel a was reused by permission of the publisher, John Wiley and Sons. Panel b was used under the open-access license. © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open-access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) licenseCC BY · PubMed Central
Fig. 2 Example of experimental approach for Raman model development. Optimization of model performance is an iterative process involving data preprocessing, algorithm selection, calibration, and validation. While there are steps toward automating model development, it remains a manually intensive process. Raman data were first input into various methods for variable selection including principal components analysis (PCA), multiple linear regression (MLR), principal components regression (PCR), partial least-squares or projection to latent structures (PLS), or variable importance in projection (VIP). Data were then pre-processed using Savitzky-Golay first or second derivative, SNV, multiplicative scatter correction, and mean center or autoscale. MLR, PCR, and PLS regression were used to model data. Figure permissions: figure reused from Kozma et al. paper with permission from publisher, [51] © 2018 Elsevier B.V. All rights reserved
Fig. 2 Example of experimental approach for Raman model development. Optimization of model performance is an iterative process involving data preprocessing, algorithm selection, calibration, and validation. While there are steps toward automating model development, it remains a manually intensive process. Raman data were first input into various methods for variable selection including principal components analysis (PCA), multiple linear regression (MLR), principal components regression (PCR), partial least-squares or projection to latent structures (PLS), or variable importance in projection (VIP). Data were then pre-processed using Savitzky-Golay first or second derivative, SNV, multiplicative scatter correction, and mean center or autoscale. MLR, PCR, and PLS regression were used to model data. Figure permissions: figure reused from Kozma et al. paper with permission from publisher, [51] © 2018 Elsevier B.V. All rights reservedCC BY · PubMed Central
Fig. 3 The effect of laser wavelength on Raman spectra throughout a high-density CHO cell culture. The initial use of 785 nm excitation (panels A and B) resulted in the observation of late-stage fluorescence which prevented accurate prediction of glucose after the third day of the culture. Using higher laser wavelengths at 830 nm (panels E and F) and 993 nm (panels C and D) decreased fluorescence. Using the 993 nm wavelength, which provided the most fluorescence reduction, enabled Raman-based glucose control throughout the duration of the culture. Figure permissions: figure reused from Matthews et al. paper with permission from publisher, John Wiley and Sons (75)
Fig. 3 The effect of laser wavelength on Raman spectra throughout a high-density CHO cell culture. The initial use of 785 nm excitation (panels A and B) resulted in the observation of late-stage fluorescence which prevented accurate prediction of glucose after the third day of the culture. Using higher laser wavelengths at 830 nm (panels E and F) and 993 nm (panels C and D) decreased fluorescence. Using the 993 nm wavelength, which provided the most fluorescence reduction, enabled Raman-based glucose control throughout the duration of the culture. Figure permissions: figure reused from Matthews et al. paper with permission from publisher, John Wiley and Sons (75)CC BY · PubMed Central

Verification

Claims the paper makes

21 Valid · 17 Not valid · 70 Not checked · 81 Own results and interpretations

63 more claims · 4 with their cited sources checked

Create a free account to see all 189 claims and every source verdict.

See all claims — free account
  • It may state two separate claims as one. 67
  • Probably not a substantial claim. 85

Create a free account to see these claims.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

3 of 3

Citations

Cites

Research works printed but not linked: 6

  1. National Research Council. Putting biotechnology to work: bioprocess engineering [Internet]. Washington, D.C.: National Academies Press; 1992 [cited 2021 Jan 12]. Available from: http://www.nap.edu/catalog/2052 25144056
  2. Eesley GL. Coherent Raman spectroscopy. Elsevier; 2013.
  3. Kaschif Ahmed S, Guenard R, Romero-Torres S, Antoniou C. Hybrid model identification for monoclonal antibody production bioreactor – a digital twin [Internet]. American Pharmaceutical Review. 2019 [cited 2021 Mar 25]. Available from: http://www.americanpharmaceuticalreview.com/Featured-Articles/517739-Hybrid-Model-Identification-for-Monoclonal-Antibody-Production-Bioreactor-A-Digital-Twin/
  4. Romero-Torres S, Wolfram K, Armando J, Kaschif Ahmed S, Ren J, Hill D, et al. Biopharmaceutical process model evolution- enabling process knowledge continuum from an advanced process control perspective [Internet]. American Pharmaceutical Review. 2018 [cited 2021 Mar 25]. Available from: http://www.americanpharmaceuticalreview.com/Featured-Articles/352447-Biopharmaceutical-Process-Model-Evolution-Enabling-Process-Knowledge-Continuum-from-an-Advanced-Process-Control-Perspective/
  5. Esmonde-White KA Cuellar M Uerpmann C Lenain B Lewis IR Raman spectroscopy as a process analytical technology for pharmaceutical manufacturing and bioprocessing Anal Bioanal Chem 2016 4 409 637 649 PMC5233728 27491299
  6. Mungikar A, Kamat M. Use of in-line raman spectroscopy as a non-destructive and rapid analytical technique to monitor aggregation of a therapeutic protein. American Pharmaceutical Review [Internet]. 2010 Nov 1 [cited 2016 May 18];13(7). Available from: http://www.americanpharmaceuticalreview.com/Featured-Articles/37071-Use-of-In-line-Raman-Spectroscopy-as-a-Non-destructive-and-Rapid-Analytical-Technique-to-Monitor-Aggregation-of-a-Therapeutic-Protein/

Cited by

See all 152

Metadata

Authors
Esmonde-White, K. A., Cuellar, M., & Lewis, I. R.
Venue
Analytical and Bioanalytical Chemistry
Published
2021-10-20
Volume
414
Issue
2
Pages
969-991
ISSN
  • 1618-2642 (ISSN-L)
  • 1618-2650
OpenAlex
W3206542495
Cited by
152
References
112
Citations by year
2026: 43 · 2025: 43 · 2024: 36 · 2023: 20 · 2022: 10
FWCI
2.21
Type
journal article
Language
en
Open access
Licence
Locations
  • Analytical and Bioanalytical Chemistry (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Biophysics › Spectroscopy Techniques in Biomedical and Chemical Research
  • Physical Sciences › Chemistry › Analytical Chemistry › Spectroscopy and Chemometric Analyses
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Protein purification and stability
Keywords
Raman spectroscopy, biopharmaceuticals, process analytical technology, quality by design, upstream bioprocessing, real-time process monitoring, scale-down, bioprocess control, single-cell analysis, protein quality, higher-order structure, protein composition, formulation development
MeSH terms
Biological Products, Quality Control, Spectrum Analysis, Raman, Technology, Pharmaceutical, Spectroscopy, Near-Infrared

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-07, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-07 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.