Introduction

Spectrophotometry is a versatile and widely used method. One of the most fundamental uses of spectrophotometry is the use of Beer’s law to determine the concentration of an absorbing species from the measured absorbance. This quantitative use has applications in many subject areas including healthcare, pharmaceuticals, agriculture, environmental monitoring, food production and analysis, and research. One barrier to the further use of the technique is cost, with the cost of a spectrophotometer typically over €1,000. This is prohibitive to the study of the technique at most schools and in the application of the technique in many parts of the world. The typical lack of portability and requirement for mains electricity for most spectrophotometers is also a barrier to use in different areas outside the laboratory. The described model is an attempt to provide a simple, cheap, functional, open-source, basic spectrophotometer that is robust and portable, requires no expertise in electronics to construct, and can be adapted by users in ways to suit different applications.

There have been a number of attempts to create low-cost do-it-yourself (DIY) spectrophotometers, including with the use of narrow bandwidth–emitting light-emitting diodes (LEDs) as well as those based on diffraction of a polychromatic light source. Many previously published models were primarily aimed at acting as a learning tool for the principles of spectrophotometry [1–5] rather than alternatives to commercially produced spectrophotometers. The use of LEDs has been tested using other methods of detection with good adherence to Beer’s law shown, although those that have published results that were comparable to commercial spectrophotometers have been relatively complex projects requiring familiarity with electronics [5–9].

The use of a smartphone’s ambient light sensor (ALS) to detect and measure light in conjunction with smartphone software to calculate absorbance values has obvious advantages: it is a technology that for many is preexisting, it requires no need for electronic knowledge, and it has a useful range of wavelengths detectable (on models tested here, the detectable range covers at least 400–933 nm). The portability and robustness of smartphones, alongside long battery life, are also benefits to use of a smartphone in DIY scientific equipment, especially outside of a laboratory. Smartphones have been utilised previously in the construction of a low-cost microscope [10] and also in the potential acoustic detecting, identifying, and mapping of mosquito species [11], exploiting the portability and robustness of smartphones.

A potential limitation of the use of LEDs in spectrophotometry is in the relatively broad spectral bandwidth of light emitted. It is often stated that obedience to Beer's law only strictly applies when measurements are made with monochromatic source radiation. In practise, the light emitted from many spectrophotometers is not truly monochromatic but a narrow range of wavelengths [12, 13]; thus, the traditional, functional distinction between colourimetry and spectrophotometry is blurred. The narrower the spectral bandwidth of a light source, the greater the ability of the spectrophotometer to distinguish between molecules present with similar absorbance spectra, and so the resolution also increases, allowing for an accurate determination of the molar absorptivity constant [12, 14]. Narrow spectral bandwidth is obviously vital for some applications but not all. Indirect methods, in which a signal molecule is formed that has a high molar absorptivity and/or is generated in high concentration, are less likely to be affected by relatively wide spectral bandwidths than direct spectrophotometric analysis of complex solutions containing molecules with similar absorbance spectra at similar concentrations/molar absorptivities.

The model described here is offered as a contribution to the increasing amount of free open-source hardware provided by the community of science makers/DIYers [15, 16]. The plans and editable files are available on GitHub (https://github.com/VascoRibeiroPereira/phone-spectrophotometer), and we would encourage users to suggest ideas for this project by creating ‘an issue’ (instructions: https://help.github.com/en/articles/creating-an-issue) and/or adapting and sharing modified designs within the scientific community and beyond, using this or other environments. To keep the cost of the described smartphone spectrophotometer as inexpensive as possible and as open to all, commercially available LEDs with specific wavelengths stated in the descriptions (as peak emission wavelengths) were sourced from retailers on Amazon. The cost of the LEDs used in the tests shown here was <€0.10 each, although they could not be bought singularly (see S1 Table for bill of materials).

Description of smartphone spectrophotometer

Fig 1 shows the cuvette holder and the interchangeable part with an LED and CR2032 battery. Both parts are 3D printed in polylactic acid (PLA) (diagrams shown in S1B Text). The cuvette holder has a rubber piece with a hole (Fig 1A), allowing the equipment to be well positioned and fixed and preventing outside light to reach the sensor. As in the Lego version (Fig 1D and 1E), the front aperture lines up with the rear aperture and, when attached, with the smartphone’s ALS (Fig 1C). Fig 1B shows the switch mechanism used in the interchangeable part—when the battery is up, the LED is off, and when the battery is down, it touches the mounted LED legs, turning it on. The app used to calculate absorbance from the light measurements (‘Shoebox Spectrophotometry’) is available for free (and advert free) from Google Play store [17]. The real-time light measurement on the app allows the correct positioning of the cuvette holder, with the LED fully aligned with the ALS. When the LED was switched off, the real-time light measurement should be 0 to demonstrate that the sensor is sufficiently protected from external light when attached to the phone. Full assembly instructions are available on GitHub (https://github.com/VascoRibeiroPereira/phone-spectrophotometer).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1.

A 3D-printed version (A–C) and Lego version (D and E) of the cuvette holder. (A) Cuvette holder and interchangeable part with narrow-bandwidth LED. (B) shows the battery mechanism of the interchangeable part (pushed up for off, pushed down for on). (C) shows the holder in place. (D) shows Lego cuvette holder with front aperture, which aligns with the interchangeable Lego block with LED inserted (E). LED, light-emitting diode.

https://doi.org/10.1371/journal.pbio.3000321.g001

Similarity of results to commercial spectrophotometers

To assess the functionality of the smartphone spectrophotometer, 3 calibration graphs were constructed for 3 molecules, which absorb maximally at wavelengths suiting the emission spectra of the 3 LEDs: crystal violet (587 nm), methyl orange (465 nm), and p-nitroaniline (400 nm) (emission spectra of the LEDs are shown in S1A Text, and the absorbance spectra of the solutions are shown in S1C Text).

Fig 2 shows similar absorbance, trend line equations, and R² values for the low-cost smartphone spectrophotometer compared with 3 other commercially available spectrophotometers in calibration graphs for crystal violet (Fig 2A), methyl orange (Fig 2B), and p-nitroaniline (Fig 2C) at 587 nm, 465 nm, and 400 nm, respectively. There is negative deviation from Beer’s law seen for all spectrophotometers at concentrations of p-nitroaniline above 200 μM, and the smartphone spectrophotometer showed similar deviation to the commercial spectrophotometers (Fig 2C).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2.

Calibration graphs for crystal violet (A), methyl orange (B), and p-nitroaniline (C) from 3 commercial spectrophotometers and the smartphone spectrophotometer. Points are from a single observation of the same solution for all spectrophotometers. Concentrations of p-nitroaniline demonstrating negative deviation (at concentrations > 200 μM) have been eliminated from the trend lines equations for all spectrophotometers. Data are available in S1 Data.

https://doi.org/10.1371/journal.pbio.3000321.g002

Functionality assessment with a protein assay

To further assess the utility of the smartphone spectrophotometer, a protein assay was used. The Bradford assay [18] is a widely used method for the determination of protein concentration. It exploits a change in absorbance spectrum when Coomassie Blue (in Bradford reagent) reacts with proteins in solution, causing the maximum absorbance to move from approximately 465 nm to approximately 590 nm. As shown in Fig 3, individual absorbance values, trend line equations, and R² values were similar for the smartphone spectrophotometer and the Pasco 2600 at 587 nm.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Calibration graph for protein using the Bradford assay for smartphone spectrophotometer and Pasco 2600 spectrophotometer.

Points are means of 4 observations; error bars represent ±1 SD. Data are available in S1 Data.

https://doi.org/10.1371/journal.pbio.3000321.g003

Overview

Here, using calibration data from 3 different molecules tested, we have shown that results at 3 different wavelengths from the smartphone spectrophotometer are similar to the results from 3 commercially available spectrophotometers. Results from the protein assay were also in agreement with a commercially available spectrophotometer.

Together, these results suggest that the model shown here shows functionality. The low cost, relatively easy construction, and portability would make this model useful for many applications. It would also be of potential use for a range of environments/situations, such as field work in which a typical, mains-powered, relatively large commercial spectrophotometer would not be of use. The relatively low cost of the components and availability of a range of LEDs that emit maximally at different wavelengths could allow many more people to be introduced to and use spectrophotometric analysis. This could include in resource-challenged countries, where other open-source, easily constructed equipment has helped provide pipettes and micromanipulators [16]. Although many potential users may not have access to equipment to allow determination of emission spectra, this information is often provided by manufacturers, and this would allow users to determine the potential use of an LED for their intended application.

Conclusions

Results suggest the smartphone spectrophotometer here could be a useful tool for many spectrophotometric applications, despite the relatively broad spectral bandwidth of light emitted from the LEDs tested.

Because of the effect of battery drainage on light intensity, absorbance measurements should be taken in close temporal proximity to calibration, and/or an appropriate resistor should be used.

Materials and methods

Supporting information