Yu Kiat Lin,Hui Yi Leong,Tau Chuan Ling,Dong-Qiang Lin,Shan-Jing Yao,*

1 Key Laboratory of Biomass Chemical Engineering of Ministry of Education,College of Chemical and Biological Engineering,Zhejiang University,Hangzhou 310027,China

2 Institute of Biological Sciences,Faculty of Science,University of Malaya,Kuala Lumpur 50603,Malaysia

ABSTRACT Downstream processing or product recovery plays a vital role in the development of bioprocesses.To improve the bioprocess efficiency,some unconventional methods are much required.The continuous manufacturing in downstream processing makes the Process Analytical Technologies(PATs)as an important tool.Monitoring and controlling bioprocess are an essential factor for the principles of PAT and quality by design.Spectroscopic methods can apply to monitor multiple analytes in real-time with less sample processing with significant advancements.Raman spectroscopy is an extensively used technique as an analytical and research tool owing to its modest process form,non-destructive,non-invasive optical molecular spectroscopic imaging with computer-based analysis.Generally,its application is essential for the analysis and characterization of biological samples,and it is easy to operate with minimal sample.The innovation on various types of enhanced Raman spectroscopy was designed to enhance the Raman analytical technique.Raman spectroscopy could couple with chemometrics to provide reliable alternative analysis method of downstream process analysis.Thus,this review aims to provide useful insight on the application of Raman spectroscopy for PAT in downstream processing of biotechnology and Raman data analysis in biological fields.

Keywords:Raman spectroscopy Process analytical technology Downstream bioprocessing Biotechnology

1.Introduction

In the past decades,biotechnology has emerged as a groundbreaking industry.Bioprocess efficiency and optimization are essential to control and maintain the quality of the product besides achieving low-costs,profitable and environmentally friendly feature.Hence to improve biomanufacturing,bioprocessing requires Process Analytical Technology (PAT) to control adjustments,to increase the quality and efficiency of the process,and enhancing regulatory compliances by real-time process [1].

The PAT is a system for designing and controlling of the production operations based on knowing the scientific standard principles which involves identification of various factors influencing product.The PAT technique uses a variety of physical principles to measure a quantity and send a signal of some form to an interface where it can be understood and acted upon by a human or automated operator.PAT has been implemented for its timely measurements,designing,analyzing and controlling to ensure good quality products [2].There is an improvement in operational control and compliance owing to the potential of PAT in continuous real-time quality assurance,thus PAT is gaining higher interest in many industries [1].With the PAT,the improved product quality could be gained,which involves flexible use,non-destructive,and enabled a prognostic feature in controlling the process at minimum level[3].Moreover,PAT is considered as a valuable analytical tool for the pharmaceutical researcher from the preclinical development to manufacturing the quality product with safety processing since PAT provides the benefit of eliminating and safely monitors the health hazards such as corrosives,acute toxins,allergens,carcinogens and mutagens[4].There are a few challenges for PAT in an initial evaluation compared to the traditional approaches.The sampling condition and needs vigilant in designing are the common challenges for PAT to bioprocessing.Therefore,incorporating more advanced spectroscopic sensor devices could be broad-based effects on PAT to bioprocessing [1,2,5,6].

Spectroscopic methods are considered as the most important PAT tools for the biotechnological industries,which are more focus to enable the adoption of quality by design (QbD).Sensitivity and specificity,feasibility and cost-effectiveness are the most important criteria in selecting the spectroscopic method and measurement set-up for PAT [7].At present,the conventional sensors(UV,conductivity and pH)are non-specificity and have several limitations in the biopharmaceutical operation products.Several spectroscopic techniques have successfully been used to monitor the product variables such as Raman,nuclear magnetic resonance(NMR),Near-infrared (NIR) spectroscopy,mass spectrometry,surface plasmon resonance,and chemometric methods[6,8].Comparing among them,Raman spectroscopy contains many advantages such as little or no sample preparation requirement,acquired quickly in less time,not hindered by water,cost-effectiveness,organic and inorganic materials are suitable,and could be easily analyzable (Fig.1).

The Raman spectroscopy was invented following the discovery of Raman Effect by Indian scientist named C.V.Raman in 1928.Raman spectroscopy is considered as a developmental tool for its positive impact on bioprocessing and many industries with endless scientific and financial benefits [9].It is highly available for a broader audience because it measures samples relatively and consistently;most promising online,and at-line PAT tools compared to the other detection methods [10].Raman spectroscopy has involved in many applications including pharmacology,mineralogy,in-situ measurements,corrosion,semiconductor,catalysts analysis,and single-molecule detection,which continuously increase by developing advance enhancements in the technology.Fig.2 shows the current capabilities of Raman spectroscopy involved in analytical and measurement techniques.Raman spectroscopy could be used for both organic and inorganic materials in routine qualitative and quantitative measurements by solving complex analytical issues.Apart from that,gases,aerosols,vapors,solids and liquids can also be analyzed,as well asinsituidentification,optical configurations,multichannel detections schemes and quantitation of combustion products.

In the past few decades,Raman spectroscopy has been developed and proven to be a global PAT strategy for bioprocessing.Which is involved in designing,data acquisition and analysis,process analyzers in cellular performance,process controlling and knowledge management tools,and alleviate the risk from substandard drug products to both manufacturer and patient[11].In general,PAT framework must be prudently selected,including sensor,data analysis,strategies in controlling and optimization,and analytical unit operation in bioprocessing technology,in order to get the benefits of PAT [12].

Raman spectroscopy emerged as sensor technology and laser sampling in various industries,particularlyin-situRaman spectroscopy in bioprocessing [13].Raman spectroscopy has been effectively involved in monitoring the substrates,biomass,and product parameters in quantitative,sensitive,and specific forms.Moreover,Raman spectroscopy acts as a valuable analytical tool in biopharmaceutical mammalian or microbial cell culture manufacturing processes,which consists of both upstream and downstream processes.It can improve the quality of the products(such as monoclonal antibodies) by real-time monitoring [14].The applications of Raman spectroscopy in downstream bioprocess also include the measurement of product concentration,product aggregation,glycosylation,and membrane fouling.In addition,it can measure in water,inherently high molecular specificity,with minimal samples for the pre-treatment and configurations [15].Raman spectroscopy-based PAT offers fast measurement times,which an important factor in real-time process monitoring and control,present easy in-line implementation while maintaining comparable costs for downstream processing,which hugely benefit biological downstream process.

The goal of this review is to provide an updated overview of Raman spectroscopy application in biotechnological processingand the value of Raman spectroscopy utilization as PAT tool in bioseparation and purification processes,as the downstream bioprocessing is a highly significant process from the time and capital investment perspective.

Fig.1.Advantages of Raman spectroscopy.

Fig.2.Analytical and measurement techniques of modern Raman spectroscopy.

2.Theory,Types and Biological Applications of Raman Spectroscopy

Fig.3 shows the working mechanism of Raman spectroscopy.Raman spectroscopy evaluates light’s scattering through the sample,the scattered light passes through a filter then dispersed by the diffraction grating,and the dispersed light was then collected by charged coupled device detector to generate Raman spectra.The chemical and structural information are from the shifts in the wavelength of the inelastically scattered radiation,while the vibrational modes of a specimen from the Raman scattering of a molecule exposed by a monochromatic light are measured by Raman spectroscopy[16].The number of photons Raman scattered is relatively small as the Raman scattering is a weak process,however,the sensitivity can be enhanced by various variants of Raman spectroscopy.To identify the molecules and to investigate the properties of molecules,modern Raman spectroscopy involves various non-invasive reflection techniques based on the Raman Effect.As of now,there are more than 25 processes/variants of Raman spectroscopy are available with advanced analytical techniques to enhance the sensitivity of a Raman measurement.As the Raman spectroscopic technique can be performed as refraction measurement with less sample and little preparation,Vibrational Raman spectroscopy where the molecules shift in the vibrational state is well suited and widely used for analyzing the chemicals,molecules in solution and solids [17].

Fig.3.Simplified diagram of the working mechanism of Raman spectroscopy.

Spontaneous Raman spectroscopy,tip-enhanced Raman spectroscopy(TERS),coherent anti-Stokes Raman spectroscopy(CARS),surface-enhanced Raman spectroscopy (SERS),and Fourier transform Raman spectroscopy(FTRS)are the common techniques used with more interest among all the areas [10,16].Table 1 shows the different features and working principle of various types of Raman spectroscopy.Spontaneous Raman spectroscopy is a label-free and chemically selective hyperspectral imaging technique,it contains weaker signal compared to other Raman spectroscopy,as the technique is based on Raman scattering using normal far-field optics[24].TERS can achieve resolutions in the range of nanometers below the limit of light diffraction with the high spatial resolution of a scanning probe microscopy [25].Using TERS can provide Raman imaging of a small area within the sample and detect the spatial resolution from the nanometer scale to collect the spectral information [26].CARS is a form of Raman spectroscopy that is used to detect the vibrations of chemical bonds and molecules,it requires two pulsed laser sources compared to a single continuous wave laser for Raman spectroscopy.It uses multiple excitation laser sources to generate a signal with the frequency higher than the excitation frequency [27,28].SERS is the most useful and reliable technique compared to others as it enables analyze compounds at lower concentrations in lesser time with additional resolution[29].SERS is a method that enhances the Raman scattering up to a factor of 1010by increasing the incident electromagnetic field and using adsorbed molecules to detect single molecules on the surface or nanostructures [27,30].FTRS is commonly used with NIR laser and appropriate detectors like Germanium or Indium gallium arsenide,it reduced the fluorescence effect as working in NIR higher frequency region.FTRS can be used for discernment among edible oils and fats and yielded about 94%with more factors in determining the authenticity of edible oils and fats rapidly and simply with chemometric analysis [16].

For characterizing biological systems and determining the chemical composition,Raman spectroscopy can consider as an extremely dominant technique.However,no single spectroscopic method can provide all the desired information in a biological system as it completely depends on circumstances.The various scientific applications of combined technique or enhanced Raman spectroscopy are rapidly increasing as a valuable analytical tool(Table 2).For instance,to construct the maps and to detect the bacteria such asSalmonella enterica,Escherichia coli,Listeria monocytogenes,Lactococcus lactisin skimmed milk,SERS was used [31].By using SERS,studies shown that spatial changes in human embry-onic colonies,metabolism of antitumor drug 6-mercaptopurine in living cells,stretched DNA using micro-Raman spectroscopy were reported [32,33].Using mid-infrared images and multivariate curve resolution on Raman spectroscopy,sucrose,lactose,fat and whey are identified in white chocolate and milk[24].Using Raman images,the ageing of polymeric surfaces was demonstrated in identifying the organic molecules and unsaturated compounds[34].In fermentation monitoring,water is a weak Raman scatterer in the Raman spectrum,it is transparent at the measurement wavelengths.Compared to MIR,Raman spectroscopy has an advantage by performing through a glass window,as the Raman bands are narrow and distinct.However,the fluorescent background from excited cellular material is the major limitation to the use of Raman spectroscopy.Raman spectroscopy can focus through a window in a fermenter,thus forestalling the necessity for further additional probes into the bioreactor.Using microscopes for trace analysis,fiber optic probes for in situ identification,liquid chromatography columns for cleaning sample methods were greatly reduced with the utility of Raman spectroscopic instrumentation.In another study,for protein conformation in solid states and solution,Raman spectroscopy has been applied widely for its inferring structural information from proteins could be obtained with little secondary order structure compared to Xray crystallography and circular dichroism [35].

Table 1 The feature and working principle of numerous types of Raman spectroscopy

Table 2 Application of combined techniques of Raman spectroscopy in biotechnology fields.

In food industry,Raman spectroscopic technique are used to detect and evaluate the quality of frozen foods and assess the compositional,physicochemical,and sensory characteristics of the frozen foods without any pretreatment [36].In the fermentation industry,Raman spectroscopy showed a promising role in process monitoring.Studies have shown that the cultivation process ofSaccharomyces cerevisiaewas monitored by using Raman spectroscopy integrated with chemometric methods [37].In another study,Raman spectroscopy was used along with a robust sapphire ball probe to correct the Raman signal for the attenuation to enhance the quantification of reaction components and measuring yeast cell concentrations [38].In addition,Raman spectroscopy can be applied in extra-cellular matrix imaging applications for its exclusive comprehensions into the structure and composition of tissues and cells at thein-vitro,in-vivoandex-vivoanalysis[39].Using the high sensitivity SERS method,circulating tumor cells can be detected,which is value for early diagnosis,rapid evaluation for treatment and tumor recurrence [40].Likewise,Raman spectroscopy is applied in many industries for its unique features of a non-destructive nature,a quick spectrum response and sensitivity.

3.Raman Data Analysis for Biological Systems

In biological application of Raman data analysis,correcting artefacts,regulating the spectral data and translating the signals into the highly valuable information are divided into three groups,which are data pre-treatment,pre-processing and modelling(chemometric as presented in Fig.4.The non-sample-dependent artefacts are corrected in pre-treatment and the block of procedures such as de-smoothing,baseline correction,normalization and dimension reduction is in data pre-processing,followed by chemometric methods consisting of model analysis,and prediction of the models,the performance of the models is then evaluated[41].

Fig.4.The framework of Raman spectroscopic data analysis in the biological application.

Collecting the sample’s information from Raman spectra for data analysis with accurately representative concentrations remains challenging.Chemometric methods coupled with Raman spectroscopy generally used rapidly for assessing the quality and characterization of cell-culture media.Chemometric methods handle huge data and playing a function in determining the unknown biological samples and the corrupting signals,which can be eliminated using the training data to deliver better quality by predication analysis [42].Data pre-processing is to the elimination of higher frequency components (de-spiking) or limit the disturbances such as spikes (de-smoothing),Gaussian or Poisson noise(baseline corrections),fluorescence and systematic differences(normalization) during the sampling analysis [42].As shown in Fig.4,all the contaminations in the sampling are done in preprocessing stage to avoid unnecessary spectral variations,which decline the subsequent analysis performance [43].Derivatives and vector normalization are widely used as model-based preprocessing techniques to remove unwanted interferents for quantifying and separating in the spectra [44].

Although several software implementing the steps in the data analysis,the results are vary based on the performance in the processing stages.Finding and aligning the peaks in the mean spectrum needs to be avoided at pre-processing stages to avoid errorprone issues and spoiling the data is the major concern for the research analysts.The calibration model’s performance plays an important role in determining the accuracy and reliability of analyzing data with in-line monitoring.The extended multiplicative scatters correction was developed in NIR spectroscopy in food science had become a major outline for model-based preprocessing in vibrational spectroscopy [42].In statistical modelling,Raman spectra are used after pre-processing to translate the spectral signals into high-level information by decreasing computational effort and improving the generalization performance of the model with a dimension reduction step involved principal component analysis (PCA) or partial least squares (PLS) [45].

PCA and PLS are commonly used for univariate and multivariate models bioprocessing and experimental designs[46].Guntheret al.[47]showed that the design of experiment could be rapidly screened by using PCA and PLS modeling techniques to determine outliers.The PCA is a classic multivariate analytical tool as it suits large data sets in bioprocess industries and converting the data sets into a model of smaller dimensionality by interpreting the correlation configuration[48].The PCA models were used for identifying the several process faults using Hotelling’s T square statistic for mammalian cell culture processes [47].Depending on the manufacturing process,factors,calibration errors,and coefficient of determination (R2) are the criteria that can be used to assess the model suitability.The PLS models have a definite input–output structure compared with PCA model,the PLS model defines correlations in model structure,a model for the unmeasured component may be generated from the correlation’s information.PLS modeling could allow an appropriate control strategy to be designed for specific batch components,it has been applied to estimateEscherichia coliculture quality[49];it is very useful from PAT perspective.

4.Application of Raman Spectroscopy in Downstream Processing

In upstream processes,PAT usually applied to control the cell culture or fermentation processes to ensure good product quality and variability.As an example,a generic model in controlling anE.coliculture was applied to attain its optimum specific growth rate profile.The PAT control can keep an optimal environment within the bioreactor,but the models themselves are not necessarily fully known at fundamental biochemical processes’level,as the optimal environment was based on productivity information [50].

Downstream processing plays an important role in biomanufacturing.It comprises a series of steps from the biological materials to increase the purity and the quality of the product by taking advantage of the physical and chemical properties.Biochemical processes are therefore should be studied,which is causing it different in contrast with the upstream process[51].Products that produced from media components,which derived from animals usually required high downstream process costs for cleanup and virus removal [52].Herein,Raman spectroscopy is considered as a simple and accurate analytical method as it offers the remote location of the analyzer,non-contact sampling optics,and compatibility with numerous flow cells [51].The typical flow scheme for downstream processing of recombinant protein is illustrated in Fig.5.The following sections discuss some processes on Raman spectroscopy involving in downstream bioprocessing.

4.1.Extraction

Extraction involves a component separation between two immiscible liquids (generally polar and non-polar medium) and it depends on components relative solubility in the respective media used.The conditions such as temperature or pH are conventionally tracked in a two-liquid phase extraction process as a process control parameter.

Raman spectroscopy could couple with chemometrics to provide a dependable and non-destructive alternative for the determination of dissimilar types of saccharides,apart from the real-time keep track of extraction operation during Wenxin granules production.An approach that was based upon Raman spectroscopy and a competitive adaptive reweighted sampling-partial least squares(CARS-PLS) model was built for saccharides fast determination[53].Moreover,SERS spectra of (tetramethylthiuram disulfide),a dimethyl ditheocarbamate fungicide,were studied after the binding on plasmonic silver nanowires from a system of organic solvents,nanoparticles and water.Considering the requirement for trace analysis of pollutants from environmental samples,which are usually highly complex,a selective and highly sensitive analytic technique is considerable demand.The capability of detecting the low concentration of molecules and the possibility to analyze dominant herbicides quantitatively.It had shown SERS methodology’s effectiveness for compounds’ sensitive analysis with low aqueous solubility,a SERS spectrum of thiram was acquired with superb signal/noise ratio at low concentrations makes SERS an interesting and suitable approach for monitoring and environmental analysis [54].

Fig.5.Typical flow scheme of downstream processing of recombinant proteins.

Raman spectroscopy enabled pattern recognition based on rapid quantification and analysis of compounds as a PAT tool for the extraction process.Extraction process usually involves between two immiscible liquids,where the major component is water,the weak Raman scattering property of water makes Raman spectroscopy has high potential on analysis application for the extraction process.

4.2.Precipitation

The precipitation is an important technique particularly to produce micronized powders in many industries in bioprocessing.Precipitation usually occurs through the reaction of two soluble components to form less soluble product at supersaturation and results in rapid nucleation producing a large amount of micronized crystals [55].Hence,some reliable Raman spectroscopy analysis method have been applied to precipitation process.Raman spectroscopy has been used to monitor precipitation in-line:Parallel determination of the profile of the supersaturation and polymorphic forms.Inline Raman spectroscopy has been applied in monitoring the precipitation of calcium phosphate from the reaction between CaCl2and KH2PO4under ambient temperature and pressure.The findings revealed that the real changes in both the mother liquor and the solid phase were successfully applied by Raman spectroscopy[56].For the washing of TCA-precipitated proteins using ethanol/HCl,the mechanistic that activates protein conformational alteration at each recovering operation was then successfully studied by Raman spectroscopy [57].Raman analysis can be employed to the other complicated aqueous precipitation systems of electrolytes to detect and quantify the dissociated anions.It is proposed that the potential application of Raman spectroscopy as a precise and proper real-time operation monitoring method that can be employed in control processing and designing in precipitation studies.However,the case studies of precipitation coupled with Raman spectroscopy is not as common as other downstream processes,this suggests that further work still needs to be done to make Raman spectroscopy a practical alternative method for precipitation process,a lot of Raman spectroscopy combination techniques still not applied to enhance the analysis capabilities for precipitation process.

4.3.Filtration

The filtration process is the most frequently used process in biotechnology,as a normal flow filtration or tangential flow filtration applications to achieve a countless function.A pre-coating filtration process combined with SERS has been studied.SERS was used to determine the distance of solid particles across the precoating in filter slurry.The effects of thicknesses and nonidentical sizes of the pre-coating on the filtration of a specific filter slurry were investigated.The utilization of SERS greatly solved the rapid termination of filtration,which originated by the rapid rise of the specific resistance of filter cake in the classic filtration operation and helped to discover the trajectory of the filter particles moving with the filtrate in dissimilar pre-coating[53].Besides that,the application of SERS-online sensing approach could detect biofouling in drinking water membrane filtration.A customdesigned flow-cell was created to determine alternation of surface-foulants (Brevundimonas dimiuta)bacteria and adenine[58].The Raman spectrometer with a laboratory Raman probe coupled with gold nanoparticle SERS-sensing area on filtermembranes was used for the measurement.Under the pressuredriven filtration conditions,the low concentrations of surfacefoulants were detected using real-time detection.The polymer coagulant aid could be effectively enhancing the coagulationultrafiltration process,which can effectively remove particulate matter and bacteria from water to ensure the stability and biosafety of effluent water quality.However,when coagulant aid entered the filtration,it may also cause serious membrane fouling as polymer.Moreover,some SERS combined characterization techniques may have limitations in their applications considering the sample complexity.Combined Raman spectroscopy techniques In-situ Raman spectroscopy and electrochemical impedance spectroscopy(EIS) are applied to monitor the effects of coagulant aids on the membrane.The equivalent circuit fitting is performed on the EIS data and the Raman spectral data are statistically analyzed after peak fitting.EIS and the cluster analysis of Raman spectroscopy provided earlier feedback on membrane fouling layers compared to flux.The cause of membrane fouling was explainedviavariation of characteristic functional groups obtained by Raman spectroscopy.The results of Raman spectroscopy and EIS are consistent for the analysis of membrane fouling,it proved that Raman spectroscopy can provide early feedback on the occurrence of fouling by online monitoring [59].

Filtration process usually applied as a last step in biological production.According to the previous studies,Raman spectroscopy has high potential to monitor and estimate the quality of production.Variety of approaches have been proposed in the studies shown that implementation of Raman spectroscopy on filtration process is likely to result in process robustness and provides assurance that the product meets the quality criteria.

4.4.Chromatography

Raman spectroscopic technique has been used to study liquid chromatographic to examine the spectroscopic differences for bonded ligands used in liquid chromatographic separations.Intact function,high sensitivity,and scalability are the advantages of chromatography and lasts as the foundation of downstream processing.

SERS could be used as detectors of thin-layer chromatography and liquid chromatography because of the reproducibility,dynamic range and application potential evaluation as it could provide structural information of components [60].In short,Raman spectroscopy is a promising and rapid monitoring technique for analysis of biopharmaceuticals,complex bioprocesses in biotechnology.The performance of Raman spectroscopy for the analytical quality control (AQC) of an anticancer drug,5-fluorouracil was compared by using UV/Vis HPLC as a reference.The statistical analysis confirms a strong correlation in the results between Raman spectroscopy and the UV/Vis HPLC.The potential of non-intrusive AQC performed by Raman spectroscopy was indicated,the technique could apply for medication safety [61].On the other hand,Raman spectroscopy has been utilized as a detection system for high-performance liquid chromatography eluents.A novel DNA gene probe has been built based on SERS label detection [24].The surface-enhanced Raman gene probes are based on SERSactive labels attachment,such as erythrosin,aminoacridine or cresyl fast violet,to oligonucleotide probes to be utilized for hybridization to target DNA sequences.These probes’ spectral specificity possesses a major advantage comparing other alternatives to radioactive probes,such as luminescence techniques or absorption.The SERS probes possess applications in genes identification,bacterial and viral components detection,and utilization in conjunction with the polymerase chain reaction for diagnostic applications and biotechnology.The use of HPLC together with SERS has been explored for the illicitly sold drugs’ingredient identification method [62].An individual component of a sample was favorably attained by using an acetonitrile free eluent.The targeted fractions were cumulated in microliter volumes in microtiter plate’s well,which comprised of a matrix-stabilized silver halide dispersion.Here,the latter that play a role as the precursor for the SERS-active surface was produced by using the probing laser beam.The detection limit of 1 μg of analyte per well of the microtiter plate can be achieved.If analyte binding is not restricted,because of strong amplification of the Raman signal,SERS spectra of high quality can be read even of tiny quantities of an analyte.

These studies proved that analysis of biological active compounds is possible by combining chromatography with Raman spectroscopy.However,direct analysis of biological fluids by Raman spectroscopy might still challenging because of possible matrix interferences from fluids.The sample clean-up of the biological fluid before Raman spectroscopy measurements may be necessary.But it is fair to consider Raman spectroscopy is a reliable tool for rapid quality control and providein-situstructural information under chromatography downstream processing.

5.Limitations of Raman Spectroscopy

To overcome the limitations in industries,many efforts are needed to improve the Raman spectroscopy.Several limitations need to be resolved for conventional Raman spectroscopy,for example,its low sensitivity preventing it from ultrasensitive detection in biomedical and pharmaceutical analysis.However,some techniques like SERS and Ultraviolet resonance Raman spectroscopy can improve the sensitivity,which can somewhat minimize these effects.In addition,the high level of fluorescence overlaying the Raman band is one of the major limitations in Raman spectroscopy.Nevertheless,this issue could be avoided by shifting the laser wavelength to the NIR spectral region.Apart from that,in Raman spectroscopy,the high excitation intensities will lead to sample decomposition.Raman spectroscopy,therefore,needs a good subsampling as it could interrelate with the light results in a limited depth of light penetration.Next,Raman line is sensitive to temperature,if the temperature of the scattering sample increased,certain Raman lines would become more diffuse,the sample should therefore be prevented from high temperature [63].Finally,yet importantly,the expensive cost of the advanced Raman spectroscopy equipment comparing with conventional spectrometers is also one of the most obstacles in many routine analysis laboratories.But with the development of science and technology as time goes by,the cost of Raman spectroscopy is expected to be reduced and Raman spectroscopy will become more widely adopted in biological industry.To reduce computing costs of Raman spectroscopy operation,software engineer should develop fully automated Raman based analytical model for downstream processes monitoring.The trend of PAT in Raman spectroscopy is toward fully automation by development of new chemometric techniques to operate Raman spectroscopy with minimum manpower[64].

In summary,Raman spectroscopy has both advantages(i.e.,simple and effective,high sensitivity and specificity,highly reproducibility and advance technology) and limitations(i.e.,substrates,hardware are expensive,for biogenic dynamics,analyte-substrate dynamics could be complicated;light,degrade the samples) in the application.

6.Conclusions

PATs are important tool for monitoring and controlling the bioprocesses for the advanced bio manufacturing.Although Raman spectroscopy has some limitation for bioproduction applications due to its sensitivity and sub-sampling,it features many advantages like fast measurement times,easy in-line implementation and low maintainable costs compared to the other spectroscopic techniques.Raman spectroscopy has been proved a valuable tool for PAT in downstream bioprocess application for various fields,which is an important device for downstream bioprocessing because of its speed,high sensitivity,high analytic efficiency and non-destructiveness.This review shows the applications of Raman spectroscopy as PAT tool in the operations of biological downstream processes.It can be concluded that Raman spectroscopy has potential on real-time decisions and product quality measurement,thereby increasing the consistent of product quality and efficiency of operation.It is also expected that more Raman spectroscopy applications will be see in food safety,medicine,materials,environmental protection,archaeology,gem identification,and many other fields.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (No.21878263,22078286).Dr.Yu Kiat Lin would like to thank Zhejiang University and Talent-Introduction Program of China for Postdoctoral Researcher for the financial support.