UV-Vis Spectroscopy: Principle, Strengths and Limitations and Applications

Ultraviolet-visible (UV-Vis) spectroscopy is a widely used technique in many areas of science ranging from bacterial culturing, drug identification and nucleic acid purity checks and quantitation, to quality control in the beverage industry and chemical research. This article will describe how UV-Vis spectroscopy works, how to analyze the output data, the technique's strengths and limitations and some of its applications.

UV-Vis spectroscopy is an analytical technique that measures the amount of discrete wavelengths of UV or visible light that are absorbed by or transmitted through a sample in comparison to a reference or blank sample. This property is influenced by the sample composition, potentially providing information on what is in the sample and at what concentration. Since this spectroscopy technique relies on the use of light, let’s first consider the properties of light.

Light has a certain amount of energy which is inversely proportional to its wavelength. Thus, shorter wavelengths of light carry more energy and longer wavelengths carry less energy. A specific amount of energy is needed to promote electrons in a substance to a higher energy state which we can detect as absorption. Electrons in different bonding environments in a substance require a different specific amount of energy to promote the electrons to a higher energy state. This is why the absorption of light occurs for different wavelengths in different substances. Humans are able to see a spectrum of visible light, from approximately 380 nm, which we see as violet, to 780 nm, which we see as red.¹ UV light has wavelengths shorter than that of visible light to approximately 100 nm. Therefore, light can be described by its wavelength, which can be useful in UV-Vis spectroscopy to analyze or identify different substances by locating the specific wavelengths corresponding to maximum absorbance (see the Applications of UV-Vis spectroscopy section).

How does a UV-Vis spectrophotometer work?

Whilst there are many variations on the UV-Vis spectrophotometer, to gain a better understanding of how an UV‑Vis spectrophotometer works, let us consider the main components, depicted in Figure 1.

Figure 1: A simplified schematic of the main components in a UV-Vis spectrophotometer. Credit: Dr. Justin Tom.

Light source

As a light-based technique, a steady source able to emit light across a wide range of wavelengths is essential. A single xenon lamp is commonly used as a high intensity light source for both UV and visible ranges. Xenon lamps are, however, associated with higher costs and are less stable in comparison to tungsten and halogen lamps.

For instruments employing two lamps, a tungsten or halogen lamp is commonly used for visible light,² whilst a deuterium lamp is the common source of UV light.² As two different light sources are needed to scan both the UV and visible wavelengths, the light source in the instrument must switch during measurement. In practice, this switchover typically occurs during the scan between 300 and 350 nm where the light emission is similar from both light sources and the transition can be made more smoothly.

Wavelength selection

In the next step, certain wavelengths of light suited to the sample type and analyte for detection must be selected for sample examination from the broad wavelengths emitted by the light source. Available methods for this include:
 

• Monochromators - A monochromator separates light into a narrow band of wavelengths. It is most often based on diffraction gratings that can be rotated to choose incoming and reflected angles to select the desired wavelength of light.^1,2 The diffraction grating's groove frequency is often measured as the number of grooves per mm. A higher groove frequency provides a better optical resolution but a narrower usable wavelength range. A lower groove frequency provides a larger usable wavelength range but a worse optical resolution. 300 to 2000 grooves per mm is usable for UV-Vis spectroscopy purposes but a minimum of 1200 grooves per mm is typical. The quality of the spectroscopic measurements is sensitive to physical imperfections in the diffraction grating and in the optical setup. As a consequence, ruled diffraction gratings tend to have more defects than blazed holographic diffraction gratings.³ Blazed holographic diffraction gratings tend to provide significantly better quality measurements.³
   
• Absorption filters - Absorption filters are commonly made of colored glass or plastic designed to absorb particular wavelengths of light.²
   
• Interference filters - Also called dichroic filters, these commonly used filters are made of many layers of dielectric material where interference occurs between the thin layers of materials. These filters can be used to eliminate undesirable wavelengths by destructive interference, thus acting as a wavelength selector.^1,2
   
• Cutoff filters - Cutoff filters allow light either below (shortpass) or above (longpass) a certain wavelength to pass through. These are commonly implemented using interference filters.
   
• Bandpass filters -Bandpass filters allow a range of wavelengths to pass through that can be implemented by combining shortpass and longpass filters together.

 

Monochromators are most commonly used for this process due to their versatility. However, filters are often used together with monochromators to narrow the wavelengths of light selected further for more precise measurements and to improve the signal-to-noise ratio.

Sample analysis

Whichever wavelength selector is used in the spectrophotometer, the light then passes through a sample. For all analyses, measuring a reference sample, often referred to as the "blank sample", such as a cuvette filled with a similar solvent used to prepare the sample, is imperative. If an aqueous buffered solution containing the sample is used for measurements, then the aqueous buffered solution without the substance of interest is used as the reference. When examining bacterial cultures, the sterile culture media would be used as the reference. The reference sample signal is then later used automatically by the instrument to help obtain the true absorbance values of the analytes.

It is important to be aware of the materials and conditions used in UV‑Vis spectroscopy experiments. For example, the majority of plastic cuvettes are inappropriate for UV absorption studies because plastic generally absorbs UV light. Glass can act as a filter, often absorbing the majority of UVC (100‑280 nm)² and UVB (280‑315 nm)² but allowing some UVA (315‑400 nm)² to pass through. Therefore, quartz sample holders are required for UV examination because quartz is transparent to the majority of UV light. Air may also be thought of as a filter because wavelengths of light shorter than about 200 nm are absorbed by molecular oxygen in the air. A special and more expensive setup is required for measurements with wavelengths shorter than 200 nm, usually involving an optical system filled with pure argon gas. Cuvette-free systems are also available that enable the analysis of very small sample volumes, for example in DNA or RNA analyses.

Detection

After the light has passed through the sample, a detector is used to convert the light into a readable electronic signal. Generally, detectors are based on photoelectric coatings or semiconductors.

A photoelectric coating ejects negatively charged electrons when exposed to light. When electrons are ejected, an electric current proportional to the light intensity is generated. A photomultiplier tube (PMT)⁴ is one of the more common detectors used in UV‑Vis spectroscopy.^2,5 A PMT is based on the photoelectric effect to initially eject electrons upon exposure to light, followed by sequential multiplication of the ejected electrons to generate a larger electric current.⁴ PMT detectors are especially useful for detecting very low levels of light.

When semiconductors are exposed to light, an electric current proportional to the light intensity can pass through. More specifically, photodiodes⁶ and charge‑coupled devices (CCDs)⁷ are two of the most common detectors based on semiconductor technology.^2,5

After the electric current is generated from whichever detector was used, the signal is then recognized and output to a computer or screen. Figures 2 and 3 show some simplified example schematic diagrams of UV-Vis spectrophotometer arrangements.

Figure 2: Schematic diagram of a cuvette-based UV-Vis spectroscopy system.

Figure 3: Schematic diagram of a cuvette-free UV-Vis spectroscopy system.⁸

UV-Vis spectroscopy analysis, absorption spectrum and absorbance units

UV-Vis spectroscopy information may be presented as a graph of absorbance, optical density or transmittance as a function of wavelength. However, the information is more often presented as a graph of absorbance on the vertical y axis and wavelength on the horizontal x axis. This graph is typically referred to as an absorption spectrum; an example is shown in Figure 4.

Figure 4: An example absorption spectrum taken from a UV-Vis spectrophotometer. The sample examined was expired hemoglobin dissolved in neutral pH phosphate buffer. Credit: Dr. Justin Tom.

Based on the UV‑Vis spectrophotometer instrumentation reviewed in the previous section of this article, the intensity of light can be reasonably expected to be quantitatively related to the amount of light absorbed by the sample.

The absorbance (A) is equal to the logarithm of a fraction involving the intensity of light before passing through the sample (I_o) divided by the intensity of light after passing through the sample (I). The fraction I divided by I_o is also called transmittance (T), which expresses how much light has passed through a sample. However, Beer–Lambert's law is often applied to obtain the concentration of the sample (c) after measuring the absorbance (A) when the molar absorptivity (ε) and the path length (L) are known. Typically, ε is expressed with units of L mol^‑1 cm^‑1, L has units of cm, and c is expressed with units of mol L^‑1. As a consequence, A has no units.

Sometimes AU is used to indicate arbitrary units or absorbance units but this has been strongly discouraged.

Beer–Lambert's law is especially useful for obtaining the concentration of a substance if a linear relationship exists using a measured set of standard solutions containing the same substance. Equation 1 shows the mathematical relationships between absorbance, Beer–Lambert's law, the light intensities measured in the instrument, and transmittance.^5,9

Equation 1: A set of equations showing the relationships between absorbance A, Beer–Lambert's law, the light intensities measured in the instrument, and transmittance.

The term optical density (OD) is sometimes incorrectly used interchangeably with absorbance. OD and absorbance both measure the amount of light intensity lost in an optical component, but OD takes into consideration loss from light scattering whereas absorbance does not. If very little light scattering is present in a measurement, then OD may be approximated directly using absorbance and Beer–Lambert's law may be used.

Knowing the experimental conditions during measurements is important. Cuvettes designed for a 1 cm path length are standard and are most common. Sometimes, very little sample is available for examination and shorter path lengths as small as 1 mm are necessary. Where quantitation is required, absorbance values should be kept below 1, within the dynamic range of the instrument. This is because an absorbance of 1 implies that the sample absorbed 90% of the incoming light, or equivalently stated as 10% of the incoming light was transmitted through the sample. With such little light reaching the detector, some UV‑Vis spectrophotometers are not sensitive enough to quantify small amounts of light reliably. Two simple possible solutions to this problem are to either dilute the sample or decrease the path length.

As mentioned above, recording a baseline spectrum using a “blank” reference solution is essential. If the instrument was absolutely perfect in every way, the baseline would have zero absorbance for every wavelength examined. In a real situation, however, the baseline spectrum will usually have some very small positive and negative absorbance values. For best practice, these small absorbance values are often automatically subtracted from the sample absorbance values for each wavelength of light by the software to obtain the true absorbance values.¹

Depending on the purpose of the analysis, the construction of a calibration curve may be desirable. Building a calibration curve requires some data analysis and extra work but it is very useful to determine the concentration of a particular substance accurately in a sample based on absorbance measurements. There are however, numerous circumstances in which a calibration curve is not necessary including OD measurements for bacterial culturing, taking absorbance ratios at specific wavelengths for assessing the purity of nucleic acids or identifying certain pharmaceuticals.

In UV-Vis spectroscopy, the wavelength corresponding to the maximum absorbance of the target substance is chosen for analysis. This choice ensures maximum sensitivity because the largest response is obtained for a certain analyte concentration.¹ An example of a UV Vis absorption spectrum of Food Green 3 and a corresponding calibration curve using standard solutions are provided in Figure 5. Note that two maximum absorbance peaks are present in the Food Green 3 dye, a smaller maximum absorbance peak at 435 nm and a more intense maximum absorbance peak at 619 nm. To gain maximum sensitivity when calculating an unknown concentration of Food Green 3, the maximum absorbance peak at 619 nm was used for analysis. Standard solutions across a range of known concentrations were prepared by diluting a stock solution, taking absorbance measurements and then plotting these on a graph of absorbance versus concentration to build a numerical relation between concentration and absorbance. A calibration curve was created using a least squares linear regression equation. The closer the data points are to a straight line, the better the fit. The y intercept in the straight line equation was set to zero to indicate no absorbance when no dye was present. The equation shown in Figure 5 is used to calculate the concentration of Food Green 3 (variable x) in an unknown sample based on the measured absorbance (variable y). 

Figure 5: A UV-Vis spectrum of Food Green 3 extracted from a sample is shown on the left graph. A calibration curve shown on the right graph was developed from standard diluted solutions of Food Green 3 using a least squares linear regression equation. Credit: Dr. Justin Tom.

For data analysis, the graph of absorbance versus concentration can indicate how sensitive the system is when building a calibration curve. When a linear least squares regression equation is used, the slope from the line of best fit indicates sensitivity. If the slope is steeper, the sensitivity is higher. Sensitivity is the ability to differentiate between the small differences in the sample concentration. From Beer–Lambert's Law, the sensitivity can be partially indicated by the molar absorptivity ε. Knowing the ε values beforehand, if available, can help to determine the concentrations of the samples required, particularly where samples are limited or expensive.

For reliability and best practice, UV‑Vis spectroscopy experiments and readings should be repeated. When repeating the examination of a sample, in general, a minimum of three replicate trials is common, but many more replicates are required in certain fields of work. A calculated quantity, such as the concentration of an unknown sample, is usually reported as an average with a standard deviation. Reproducible results are essential to ensure precise, high quality measurements. Standard deviation, relative standard deviation, or the coefficient of variation help to determine how precise the system and measurements are. A low deviation or variation indicates a higher level of precision and reliability.

Strengths and limitations of UV-Vis spectroscopy

No single technique is perfect and UV‑Vis spectroscopy is no exception. The technique does, however, have a few main strengths listed below that make it popular.

• The technique is non‑destructive, allowing the sample to be reused or proceed to further processing or analyses.
• Measurements can be made quickly, allowing easy integration into experimental protocols.
• Instruments are easy to use, requiring little user training prior to use.
• Data analysis generally requires minimal processing, again meaning little user training is required.
• The instrument is generally inexpensive to acquire and operate, making it accessible for many laboratories.

Although the strengths of this technique seem overwhelming, there are also certain weaknesses:

• Stray light - In a real instrument, wavelength selectors are not perfect and a small amount of light from a wide wavelength range may still be transmitted from the light source,¹ possibly causing serious measurement errors.⁹ Stray light may also come from the environment or a loosely fitted compartment in the instrument.¹ 
   
• Light scattering - Light scattering is often caused by suspended solids in liquid samples, which may cause serious measurement errors. The presence of bubbles in the cuvette or sample will scatter light, resulting in irreproducible results.
   
• Interference from multiple absorbing species - A sample may, for example, have multiple types of the green pigment chlorophyll. The different chlorophylls will have overlapping spectra when examined together in the same sample. For a proper quantitative analysis, each chemical species should be separated from the sample and examined individually.
   
• Geometrical considerations - Misaligned positioning of any one of the instrument's components, especially the cuvette holding the sample, may yield irreproducible and inaccurate results. Therefore, it is important that every component in the instrument is aligned in the same orientation and is placed in the same position for every measurement. Some basic user training is therefore generally recommended to avoid misuse.

Applications of UV-Vis spectroscopy

UV‑Vis has found itself applied to many uses and situations including but not limited to:

DNA and RNA analysis

Quickly verifying the purity and concentration of RNA and DNA is one particularly widespread application. A summary of the wavelengths used in their analysis and what they indicate are given in Table 1. When preparing DNA or RNA samples, for example for downstream applications such as sequencing, it is often important to verify that there is no contamination of one with the other, or with protein or chemicals carried over from the isolation process.

The 260 nm/280 nm absorbance (260/280) ratio is useful for revealing possible contamination in nucleic acid samples, summarized in Table 2. Pure DNA typically has a 260/280 ratio of 1.8, while the ratio for pure RNA is usually 2.0. Pure DNA has a lower 260/280 ratio than RNA because thymine, which is replaced by uracil in RNA, has a lower 260/280 ratio than uracil. Samples contaminated with proteins will lower the 260/280 ratio due to higher absorbance at 280 nm.

 

Wavelength used in absorbance analysis in nanometers	What does UV absorbance at this wavelength indicate the presence of?	What causes UV absorbance at this wavelength?
230	Protein	Protein shape10
260	DNA and RNA	Adenine, guanine, cytosine, thymine, uracil
280	Protein	Mostly tryptophan and tyrosine

Table 1: Summary of useful UV absorbance when determining 260/280 and 260/230 absorbance ratios.

 

Absorbance ratio	Typical values
260/280	1.8 absorbance ratio typical for pure DNA 2.0 absorbance ratio typical for pure RNA
260/230	Absorbance ratio varies; 2.15 to 2.50 typical for RNA and DNA11

 
Table 2: Summary of expected UV absorbance ratios for DNA and RNA analysis.

The 260 nm/230 nm absorbance (260/230) ratio is also useful for checking the purity of DNA and RNA samples and may reveal protein or chemical contamination. Proteins can absorb light at 230 nm, thus lowering the 260/230 ratio and indicating protein contamination in DNA and RNA samples.¹⁰ Guanidinium thiocyanate and guanidinium isothiocyanate, two of the common compounds used in purifying nucleic acids, strongly absorb at 230 nm which will lower the 260/230 absorbance ratio too.

Pharmaceutical analysis

One of the most common uses of UV-Vis spectroscopy is in the pharmaceuticals industry.^{12,13,14,15,16,17} In particular, processing UV-Vis spectra using mathematical derivatives allows overlapping absorbance peaks in the original spectra to be resolved to identify individual pharmaceutical compounds.^12,17 For example, benzocaine, a local anesthetic, and chlortetracycline, an antibiotic, can be identified simultaneously in commercial veterinary powder formulations by applying the first mathematical derivative to the absorbance spectra.¹⁷ Simultaneous quantification of both substances was possible on a microgram per milliliter concentration range by building a calibration function for each compound.¹⁷

Bacterial culture

UV-Vis spectroscopy is often used in bacterial culturing. OD measurements are routinely and quickly taken using a wavelength of 600 nm to estimate the cell concentration and to track growth.¹⁸ 600 nm is commonly used and preferred due to the optical properties of bacterial culture media in which they are grown and to avoid damaging the cells in cases where they are required for continued experimentation.

Beverage analysis

The identification of particular compounds in drinks is another common application of UV-Vis spectroscopy. Caffeine content must be within certain legal limits,^1,19 for which UV light can facilitate quantification. Certain classes of colored substances, such as anthocyanin found in blueberries, raspberries, blackberries, and cherries, are easily identified by matching their known peak absorbance wavelengths in wine for quality control using UV-Vis absorbance.²⁰ 

Other applications

This technique may also be used in many other industries. For example, measuring a color index is useful for monitoring transformer oil as a preventative measure to ensure electric power is being delivered safely.²¹ Measuring the absorbance of hemoglobin to determine hemoglobin concentrations may be used in cancer research.²² In wastewater treatments, UV-Vis spectroscopy can be used in kinetic and monitoring studies to ensure certain dyes or dye by‑products have been removed properly by comparing their spectra over time.²³ It also finds great utility in food authenticity analysis and air quality monitoring.

UV‑Vis spectroscopy is also qualitatively useful in some more specialized research. Tracking changes in the wavelength corresponding to the peak absorbance is useful in examining specific structural protein changes^24,25,26 and in determining battery composition.²⁷ Shifts in peak absorbance wavelengths can also be useful in more modern applications such as characterization of very small nanoparticles.^28,29 The applications of this technique are varied and seemingly endless.

References

1.      Harris DC. Quantitative Chemical Analysis. 7th ed, 3rd printing. W. H. Freeman; 2007.

2.      Diffey BL. Sources and measurement of ultraviolet radiation. Methods. 2002;28(1):4-13. doi:10.1016/S1046-2023(02)00204-9

3.      Namioka T. Diffraction Gratings. In: Vacuum Ultraviolet Spectroscopy. Vol 1. Experimental Methods in Physical Sciences. Elsevier; 2000:347-377. doi:10.1016/B978-012617560-8/50018-9

4.      Mortimer Abramowitz and Michael W. Davidson. Photomultiplier Tubes. Molecular Expressions. Accessed April 25, 2021. https://micro.magnet.fsu.edu/primer/digitalimaging/concepts/photomultipliers.html

5.      Picollo M, Aceto M, Vitorino T. UV-Vis spectroscopy. Phys Sci Rev. 2019;4(4). doi:10.1515/psr-2018-0008

6.      What is a Photodiode? Working, Characteristics, Applications. Published online October 30, 2018. Accessed April 29, 2021. https://www.electronicshub.org/photodiode-working-characteristics-applications/

7.      Amelio G. Charge-Coupled Devices. Scientific American. 1974;230(2):22-31. http://www.jstor.org/stable/24950003

8.      Hackteria. DIY NanoDrop. Accessed June 15, 2021. https://hackteria.org/wiki/File:NanoDropConceptSpectrometer2.png

9.      Sharpe MR. Stray light in UV-VIS spectrophotometers. Anal Chem. 1984;56(2):339A-356A. doi:10.1021/ac00266a003

10.    Liu P-F, Avramova LV, Park C. Revisiting absorbance at 230nm as a protein unfolding probe. Anal Biochem. 2009;389(2):165-170. doi:10.1016/j.ab.2009.03.028

11.    Kalb V., Bernlohr R. A New Spectrophotometric Assay for Protein in Cell Extracts. Anal Biochem. 1977;82:362-371. doi:10.1016/0003-2697(77)90173-7

12.    Bosch Ojeda C, Sanchez Rojas F. Recent applications in derivative ultraviolet/visible absorption spectrophotometry: 2009–2011. Microchem J. 2013;106:1-16. doi:10.1016/j.microc.2012.05.012

13.    Domingo C, Saurina J. An overview of the analytical characterization of nanostructured drug delivery systems: Towards green and sustainable pharmaceuticals: A review. Anal Chim Acta. 2012;744:8-22. doi:10.1016/j.aca.2012.07.010

14.    Gaikwad J, Sharma S, Hatware KV. Review on Characteristics and Analytical Methods of Tazarotene: An Update. Crit Rev Anal Chem. 2020;50(1):90-96. doi:10.1080/10408347.2019.1586519

15.    Gendrin C, Roggo Y, Collet C. Pharmaceutical applications of vibrational chemical imaging and chemometrics: A review. J Pharm Biomed Anal. 2008;48(3):533-553. doi:10.1016/j.jpba.2008.08.014

16.    Lourenço ND, Lopes JA, Almeida CF, Sarraguça MC, Pinheiro HM. Bioreactor monitoring with spectroscopy and chemometrics: a review. Anal Bioanal Chem. 2012;404(4):1211-1237. doi:10.1007/s00216-012-6073-9

17.    Sánchez Rojas F, Bosch Ojeda C. Recent development in derivative ultraviolet/visible absorption spectrophotometry: 2004–2008. Anal Chim Acta. 2009;635(1):22-44. doi:10.1016/j.aca.2008.12.039

18.    Stevenson K, McVey AF, Clark IBN, Swain PS, Pilizota T. General calibration of microbial growth in microplate readers. Sci Rep. 2016;6(1):38828. doi:10.1038/srep38828

19.    Tadesse Wondimkun Z. The Determination of Caffeine Level of Wolaita Zone, Ethiopia Coffee Using UV-visible Spectrophotometer. Am J Appl Chem. 2016;4(2):59. doi:10.11648/j.ajac.20160402.14

20.    Yu J, Wang H, Zhan J, Huang W. Review of recent UV–Vis and infrared spectroscopy researches on wine detection and discrimination. Appl Spectrosc Rev. 2018;53(1):65-86. doi:10.1080/05704928.2017.1352511

21.    Leong Y, Ker P, Jamaludin M, et al. UV-Vis Spectroscopy: A New Approach for Assessing the Color Index of Transformer Insulating Oil. Sensors. 2018;18(7):2175. doi:10.3390/s18072175

22.    Brown JQ, Vishwanath K, Palmer GM, Ramanujam N. Advances in quantitative UV–visible spectroscopy for clinical and pre-clinical application in cancer. Curr Opin Biotechnol. 2009;20(1):119-131. doi:10.1016/j.copbio.2009.02.004

23.    Pinheiro HM, Touraud E, Thomas O. Aromatic amines from azo dye reduction: status review with emphasis on direct UV spectrophotometric detection in textile industry wastewaters. Dyes Pigm. 2004;61(2):121-139. doi:10.1016/j.dyepig.2003.10.009

24.    Kristo E, Hazizaj A, Corredig M. Structural Changes Imposed on Whey Proteins by UV Irradiation in a Continuous UV Light Reactor. J Agric Food Chem. 2012;60(24):6204-6209. doi:10.1021/jf300278k

25.    Lange R, Balny C. UV-visible derivative spectroscopy under high pressure. Biochim Biophys Acta BBA - Protein Struct Mol Enzymol. 2002;1595(1-2):80-93. doi:10.1016/S0167-4838(01)00336-3

26.    Tom J, Jakubec PJ, Andreas HA. Mechanisms of Enhanced Hemoglobin Electroactivity on Carbon Electrodes upon Exposure to a Water-Miscible Primary Alcohol. Anal Chem. 2018;90(9):5764-5772. doi:10.1021/acs.analchem.8b00117

27.    Patel MUM, Demir-Cakan R, Morcrette M, Tarascon J-M, Gaberscek M, Dominko R. Li-S Battery Analyzed by UV/Vis in Operando Mode. ChemSusChem. 2013;6(7):1177-1181. doi:10.1002/cssc.201300142

28.    Begum R, Farooqi ZH, Naseem K, et al. Applications of UV/Vis Spectroscopy in Characterization and Catalytic Activity of Noble Metal Nanoparticles Fabricated in Responsive Polymer Microgels: A Review. Crit Rev Anal Chem. 2018;48(6):503-516. doi:10.1080/10408347.2018.1451299

29.    Behzadi S, Ghasemi F, Ghalkhani M, et al. Determination of nanoparticles using UV-Vis spectra. Nanoscale. 2015;7(12):5134-5139. doi:10.1039/C4NR00580E