Editing Ultraviolet–visible spectroscopy (section)

=== Practical considerations ===
The Beer–Lambert law has implicit assumptions that must be met experimentally for it to apply; otherwise there is a possibility of deviations from the law.<ref name=dev>{{cite web |last= Metha |first= Akul |title= Limitations and Deviations of Beer–Lambert Law |website= PharmaXChange.info |date= 14 May 2012 |url= http://pharmaxchange.info/press/2012/05/ultraviolet-visible-uv-vis-spectroscopy-%e2%80%93-limitations-and-deviations-of-beer-lambert-law/}}</ref> For instance, the chemical makeup and physical environment of the sample can alter its extinction coefficient. The chemical and physical conditions of a test sample therefore must match reference measurements for conclusions to be valid. Worldwide, pharmacopoeias such as the American (USP) and European (Ph. Eur.) pharmacopeias demand that spectrophotometers perform according to strict regulatory requirements encompassing factors such as [[#Stray light|stray light]]<ref>{{Cite web | url=https://www.mt.com/ch/en/home/library/white-papers/lab-analytical-instruments/stray-light-and-performance-verification.html | title=Stray Light and Performance Verification}}</ref> and wavelength accuracy.<ref>{{Cite web | url=https://www.chemeurope.com/en/whitepapers/126586/wavelength-accuracy-in-uv-vis-spectrophotometry.html | title=Wavelength Accuracy in UV/VIS Spectrophotometry}}</ref>

====Spectral bandwidth====
Spectral bandwidth of a spectrophotometer is the range of wavelengths that the instrument transmits through a sample at a given time.<ref>{{Cite web |title=Persee PG Scientific Inc. – New-UV FAQ: Spectral Band Width |url=http://www.perseena.com/index/newsinfo/c_id/37/n_id/7.html |archive-url=https://web.archive.org/web/20180921065923/http://www.perseena.com/index/newsinfo/c_id/37/n_id/7.html |url-status=usurped |archive-date=21 September 2018 |access-date= |website=www.perseena.com}}</ref> It is determined by the light source, the [[monochromator]], its physical slit-width and optical dispersion and the detector of the spectrophotometer. The spectral bandwidth affects the resolution and accuracy of the measurement. A narrower spectral bandwidth provides higher resolution and accuracy, but also requires more time and energy to scan the entire spectrum. A wider spectral bandwidth allows for faster and easier scanning, but may result in lower resolution and accuracy, especially for samples with overlapping absorption peaks. Therefore, choosing an appropriate spectral bandwidth is important for obtaining reliable and precise results.

It is important to have a monochromatic source of radiation for the light incident on the sample cell to enhance the linearity of the response.<ref name="dev" /> The closer the bandwidth is to be monochromatic (transmitting unit of wavelength) the more linear will be the response. The spectral bandwidth is measured as the number of wavelengths transmitted at half the maximum intensity of the light leaving the monochromator. 

The best spectral [[bandwidth (signal processing)#Photonics|bandwidth]] achievable is a specification of the UV spectrophotometer, and it characterizes how [[monochromatic]] the incident light can be. If this bandwidth is comparable to (or more than) the [[spectral linewidth|width]] of the absorption peak of the sample component, then the measured extinction coefficient will not be accurate. In reference measurements, the instrument bandwidth (bandwidth of the incident light) is kept below the width of the spectral peaks. When a test material is being measured, the bandwidth of the incident light should also be sufficiently narrow. Reducing the spectral bandwidth reduces the energy passed to the detector and will, therefore, require a longer measurement time to achieve the same signal to noise ratio.

====Wavelength error====
The extinction coefficient of an analyte in solution changes gradually with wavelength. A peak (a wavelength where the absorbance reaches a maximum) in the absorbance curve vs wavelength, i.e. the UV-VIS spectrum, is where the rate of change of absorbance with wavelength is the lowest.<ref name=dev /> Therefore, quantitative measurements of a solute are usually conducted, using a wavelength around the absorbance peak, to minimize inaccuracies produced by errors in wavelength, due to the change of extinction coefficient with wavelength.

====Stray light====
{{See also|Stray light}}
Stray light<ref>{{Cite web |date=2015-06-12 |title=What is Stray light and how it is monitored? |url=https://lab-training.com/what-is-stray-light-and-how-it-is-monitored/ |access-date= |language=en-US}}</ref> in a UV spectrophotometer is any light that reaches its detector that is not of the wavelength selected by the monochromator. This can be caused, for instance, by scattering of light within the instrument, or by reflections from optical surfaces.

Stray light can cause significant errors in absorbance measurements, especially at high absorbances, because the stray light will be added to the signal detected by the detector, even though it is not part of the actually selected wavelength. The result is that the measured and reported absorbance will be lower than the actual absorbance of the sample.

The stray light is an important factor, as it determines the ''purity'' of the light used for the analysis. The most important factor affecting it is the [[Monochromator#Stray light|''stray light'' level of the monochromator]].<ref name="dev" />

Typically a detector used in a UV-VIS spectrophotometer is broadband; it responds to all the light that reaches it. If a significant amount of the light passed through the sample contains wavelengths that have much lower extinction coefficients than the nominal one, the instrument will report an incorrectly low absorbance. Any instrument will reach a point where an increase in sample concentration will not result in an increase in the reported absorbance, because the detector is simply responding to the stray light. In practice the concentration of the sample or the optical path length must be adjusted to place the unknown absorbance within a range that is valid for the instrument. Sometimes an empirical calibration function is developed, using known concentrations of the sample, to allow measurements into the region where the instrument is becoming non-linear.

As a rough guide, an instrument with a single monochromator would typically have a stray light level corresponding to about 3 Absorbance Units (AU), which would make measurements above about 2 AU problematic. A more complex instrument with a [[Monochromator#Double monochromators|double monochromator]] would have a stray light level corresponding to about 6 AU, which would therefore allow measuring a much wider absorbance range.

====Deviations from the Beer–Lambert law====
At sufficiently high concentrations, the absorption bands will saturate and show absorption flattening. The absorption peak appears to flatten because close to 100% of the light is already being absorbed. The concentration at which this occurs depends on the particular compound being measured. One test that can be used to test for this effect is to vary the path length of the measurement. In the Beer–Lambert law, varying concentration and path length has an equivalent effect—diluting a solution by a factor of 10 has the same effect as shortening the path length by a factor of 10. If cells of different path lengths are available, testing if this relationship holds true is one way to judge if absorption flattening is occurring.

Solutions that are not homogeneous can show deviations from the Beer–Lambert law because of the phenomenon of absorption flattening. This can happen, for instance, where the absorbing substance is located within suspended particles.<ref>{{cite journal |last1=Berberan-Santos |first1=M. N. |title=Beer's law revisited |journal=Journal of Chemical Education |date=September 1990 |volume=67 |issue=9 |pages=757 |doi=10.1021/ed067p757 |bibcode=1990JChEd..67..757B }}</ref><ref>{{cite journal |last1=Wittung |first1=Pernilla |last2=Kajanus |first2=Johan |last3=Kubista |first3=Mikael |last4=Malmström |first4=Bo G. |title=Absorption flattening in the optical spectra of liposome-entrapped substances |journal=FEBS Letters |date=19 September 1994 |volume=352 |issue=1 |pages=37–40 |doi=10.1016/0014-5793(94)00912-0 |pmid=7925937 |s2cid=11419856 |doi-access= }}</ref> The deviations will be most noticeable under conditions of low concentration and high absorbance. The last reference describes a way to correct for this deviation.

Some solutions, like copper(II) chloride in water, change visually at a certain concentration because of changed conditions around the coloured ion (the divalent copper ion). For copper(II) chloride it means a shift from blue to green,<ref>{{cite journal |last1=Ansell |first1=S |last2=Tromp |first2=R H |last3=Neilson |first3=G W |title=The solute and aquaion structure in a concentrated aqueous solution of copper(II) chloride |journal=Journal of Physics: Condensed Matter |date=20 February 1995 |volume=7 |issue=8 |pages=1513–1524 |doi=10.1088/0953-8984/7/8/002 |bibcode=1995JPCM....7.1513A |s2cid=250898349 }}</ref> which would mean that monochromatic measurements would deviate from the Beer–Lambert law.

====Measurement uncertainty sources====
The above factors contribute to the [[measurement uncertainty]] of the results obtained with UV-Vis [[spectrophotometry]]. If UV-Vis spectrophotometry is used in quantitative chemical analysis then the results are additionally affected by uncertainty sources arising from the nature of the compounds and/or solutions that are measured. These include spectral interferences caused by absorption band overlap, fading of the color of the absorbing species (caused by decomposition or reaction) and possible composition mismatch between the sample and the calibration solution.<ref>{{cite journal |last1= Sooväli |first1= L. |last2= Rõõm |first2= E.-I. |last3= Kütt |first3= A. |last4= Kaljurand |first4= I. |last5= Leito |first5= I. |year= 2006 |title= Uncertainty sources in UV–Vis spectrophotometric measurement |journal= Accreditation and Quality Assurance |volume= 11 |issue= 5 |pages= 246–255 |doi= 10.1007/s00769-006-0124-x |s2cid= 94520012 |display-authors= 3}}</ref>