Journal of the Chinese Chemical Society, 2006, 53, 1067-1072 1067

A Low Cost LED Based Spectrometer

Tai-Sheng Yeh* ( ) and Shih-Shin Tseng ( )

Department of Electronic Engineering, Fortune Institute of Technology, Kaohsiung, Taiwan, R.O.C.

A low cost LED based spectrometer is described. This LED based spectrometer could be operated as a
standalone instrument or under PC control via serial link. A total of seven wavelength selections are avail-
able by the plug-and-measure LED light module. With the seven wavelength selections, the LED based
spectrometer could provide qualitative visible absorption spectra that predict the absorption maximum.
Based upon the qualitative visible spectra, quantitative photometric information could be obtained.

Keywords: LED; Spectrometer; USB 2000; Low cost; Microcontroller.

INTRODUCTION This spectrometer was not robust or inexpensive enough

for large scale deployment in the field. Cantrell et al.13 de-
There is a need in many resource limited countries signed a SLIM spectrometer with red, yellow, and blue
and laboratories for a low cost analytical instrument that LEDs. The total cost to construct the SLIM spectrometer is
could provide both qualitative and quantitative chemical about US$ 25. A TSL230 programmable light-to-fre-
information. There are also great needs for low cost auton- quency converter from Texas Advanced Optoelectronic
omous instruments that could help monitor the fragile envi- Solutions (TAOS) was used as photo-detector for the SLIM
ronment around the world on a global scale.1,2 With the ad- spectrometer. The SLIM spectrometer is low cost, low
vances in modern electronics, microcontrollers could pro- power consumption and robust, thus it is very suitable for
vide great capabilities to acquire and process sensor sig- autonomous instruments employed in the field. The only
nals. The development of compound semiconductor manu- limitation is the available wavelength in the SLIM spec-
facturing technology could also produce many LEDs with trometer.
wavelengths from IR down to the 365 nm UV range. With This work intended to build a robust, low cost, multi
its compact size, a wide range of selections of available wavelength LED based spectrometer using plug-and-
wavelengths and low power consumption, LEDs are used measure LED light source modules. A TCS230 program-
extensively in analytical instruments as light sources.3-15 mable light-to-frequency converter from Texas Advanced
Dasgupta et. al. gave a very detailed review for LED appli- Optoelectronic Solutions (TAOS) was used as the photo-
cation in flow injection analyses.3,4 LED light sources for detector in the current design. The total cost of the present
sensor applications were also extensively reviewed by design is about US$ 20. With a wide range wavelength se-
Kostov et al.5-7 Hauser et al.8-10 built a multi-wavelength lection and low cost solid-state electronic components, this
spectrometer to analyze aluminum, copper, ammonia, cal- LED based spectrometer is a very cost-effective instrument
cium, phosphate, chromium, and nitrile. In his work, LEDs for resource limited developing countries and laboratories.
with seven wavelengths and a 2 ´ 7 fiber optic coupler were
used to construct a multi-wavelength spectrometer. The 2 ´
7 fiber optic coupler would greatly increase the cost of the EXPERIMENTAL
spectrometer and would not be suitable for large scale em-
ployment. Hamilton et al.11 built a four color LED photom- The Chemical Reagent and Electronic Components
eter with a material cost of about US$ 200. The LED pho- The Universal Indicator of pH solution is purchased
tometer was designed with analog circuits and thus was not from Panreac, Inc. The pH dye Bromocresol Purple and
very flexible or user configurable. Knagge et al.12 built a Bromocresol Green were purchased from Lancaster Inc.,
spectrometer with a mini light bulb, diffraction grating and and Phenol Red and Thymol Blue were purchased from
LEGO block. The material cost was also about US$ 200. TCI, Inc. All pH buffer solutions were purchased from
1068 J. Chin. Chem. Soc., Vol. 53, No. 5, 2006 Yeh and Tseng

Panreac, Inc. The LED was bought from RS Components control software written in Visual Basic 6.0 was used to set
and local parts vendors. The TCS230 sensor chips were the operation mode for the experiment. The spectra of the
bought from TAOS, Inc. The PIC Micro 16F877A and plug-and-measure LED light sources are shown in Fig.
electronic components were sourced from local electronic 2(a). The emission peaks for the LED light sources were
parts vendors. 389 nm, 407 nm, 462 nm, 527 nm, 572 nm, 587 nm and 620
nm. The actual device hardware is shown in Fig. 2(b), 2(c),
The Circuit and Hardware Tuning 2(d).
Fig. 1 is the circuit design for the LED based spec- The LED light intensity and the emission bandwidth
trometer. The TCS230 chip from TAOS, Inc. was used as could be adjusted by the current-limiting resistor. Because
photo-detector. Output-frequency scaling and photodiode the LED light intensity and the emission bandwidth would
color selection is set by a dip-switch connected to the S0, affect the measured absorbance by the LED based spec-
S1, S2, S3 input pin on the sensor chip. The sensor fre- trometer, calibration against a commercial spectrometer
quency output is received by a PIC Micro 16F877A micro- with standard solutions is necessary for the LED based
controller via PORTC.2. The signal after processing by a spectrometer. The best empirical value for the current-lim-
microcontroller is displayed on the LCD and transmitted to iting resistor is set by fitting the absorbance spectra ob-
a PC via a RS232 serial connection. Thus the LED based tained by a LED based spectrometer to that obtained by a
spectrometer could operate standalone or be controlled by a Ocean Optics USB 2000 spectrometer. A series of standard
PC to have a better visualized view of the obtained data. PC solutions to calibrate the LED based spectrometer were

Fig. 1. Circuit for LED based spectrometer.

A Low Cost LED Based Spectrometer J. Chin. Chem. Soc., Vol. 53, No. 5, 2006 1069

prepared by adding 5 ml Universal pH Indicator to 2 mL The resistor value for each LED light source was de-
buffer solutions from pH 1 to pH 13. The spectra for the 13 termined by comparing it with an Ocean Optics USB 2000
standard solutions obtained by Ocean Optics USB 2000 spectrometer. Herein the spectra of different buffer solu-
spectrometer are shown in Fig. 3(a). The obtained spectra tions from pH 1 to pH 13 were measured, and the resistor
serve as reference for setting the value of LED current- value was tuned to best fit the spectra obtained by the USB
limiting resistors. The best value of the current-limiting re- 2000 spectrometer. After the resistor value was fine tuned
sistor was set by comparing the spectra obtained from a for the LED based spectrometer, four indicator solutions:
USB 2000 spectrometer. Fig. 3(b) shows the best value for Bromocresol Purple, Bromocresol Green, Phenol Red and
a blue LED (462 nm) was 2 kW. Thymol Blue were measured and compared again with the
USB 2000 spectrometer. This is the testing procedure for
Experiment Procedure for Spectra Mode testing the performance of the spectra mode.
Typical LEDs have FWHM values of about 25 nm
and a typical FWHM molecular absorption band is over 50 Experiment Procedure for Photometric Mode
nm. Thus it is possible is use a discrete LED absorbance A series of the Universal Indicator solution from 1.25
measurement to approximate the absorbance spectra of mo- ml to 200 ml is dropped to 2 mL pH 7 buffer solution. The
lecular absorption.

Fig. 2a. Spectra of LED light source. Fig. 2c. LED spectrometer with cover removed.

Fig. 2b. Plug-and-measure LED light source module Fig. 2d. LED spectrometer with cover in measurement
and TCS230 detector module. mode.
1070 J. Chin. Chem. Soc., Vol. 53, No. 5, 2006 Yeh and Tseng

absorbance variation caused by concentration difference of spaced closely. For the absorption spectra of Thymol Blue,
indicator solution is compared with the USB 2000 spec- the LED based spectrometer didn’t predict the absorption
trometer for the photometric mode. maximum very well. The reason was that in the present
spectrometer design there was no 430 nm LED light source.
For the absorption spectra of Phenol Red, the LED based
RESULTS AND DISCUSSION spectrometer didn’t produce very good absorption spectra.
This is caused by the large FWHM values of the blue LED
Spectra Mode (peak 462 nm) and green LED (peak 527 nm). According to
The absorption spectra obtained by the LED based Beer’s law, a polychromatic light source would give a large
spectrometer are shown in Fig. 4. For the Bromocresol deviation compared to a monochromatic light.14 The lack
Purple and Bromocresol Green dye solutions, the LED of a good 430 nm LED light source plus the lack of a light
based spectrometer showed very good qualitative agree- source between blue LED and green LED also contributed
ment with the USB 2000 spectrometer and provided a to the ill-behaved absorption spectra.
good prediction of their absorbance spectra. In the spectra
region for the maximum absorption of Bromocresol Pur- Photometric Mode
ple and Bromocresol Green, the LED light sources are Comparison of photometric performance with an

Fig. 4a. Comparison of Bromocresol Purple spectral

Fig. 3a. Spectra of pH buffer solution from pH 1 to pH data with a USB 2000 spectrometer.
13 for Universal pH Indicator.

Fig. 4b. Comparison of Bromocresol Green spectral

Fig. 3b. Tuning resistor to fit USB 2000 Absorbance. data with a USB 2000 spectrometer.
A Low Cost LED Based Spectrometer J. Chin. Chem. Soc., Vol. 53, No. 5, 2006 1071

Ocean Optics USB 2000 spectrometer is shown in Fig. 5. the Ocean Optics USB 2000 spectrometer for the full mea-
The LED based spectrometer showed a similar trend with surement range. For the concentration range from 10 ml to
80 ml, the agreement with the USB 2000 spectrometer was
very good for the LED based spectrometer. The LED based
spectrometer could yield very good quantitative results for
absorbance values under 0.8. In daily practice, absorbance
measurements seldom exceed 1.0. Many commercial in-
struments also fail at high absorbance values. The devia-
tion of the LED based spectrometer in the high absorbance
region could be due to the polychromatic nature of the LED
light sources.14

CONCLUSIONS
Fig. 4c. Comparison of Thymol Blue spectral data
with a USB 2000 spectrometer
Both spectra mode and photometric mode for the
LED based spectrometer had shown very comparable re-
sults compared to the Ocean Optics USB 2000 spectrome-
ter. With low cost electronic components, it is possible to
build a cost effective instrument for resource limited labs.
The electronic components for a LED based spectrometer
had a total cost of about US$ 20. Commercial instruments
usually would cost more than US$ 1,000. In this respect,
the LED based spectrometer was a very cost-effective in-
strument. Inclusion of a kinetic mode for a LED based
spectrometer is possible by rewriting the firmware in a PIC
microcontroller. This kind of LED based spectrometer will
be very useful for both autonomous environment monitor-
ing and educational purposes. The capability of the present
Fig. 4d. Comparison of Phenol Red spectral data with a LED based spectrometer can be further extended by incor-
USB 2000 spectrometer. porating more LEDs with different wavelengths.

ACKNOWLEDGEMENTS

The authors sincerely thank the National Science

Council of Taiwan (No. NSC 93-2113-M-268-001) for the
financial support of this work.

Received December 15, 2005.

REFERENCES
Fig. 5. Comparison of photometric data with a USB
2000 spectrometer.
1. Diamond, D. Anal. Chem. 2004, 76, 279A-286A.
1072 J. Chin. Chem. Soc., Vol. 53, No. 5, 2006 Yeh and Tseng

2. Sequeira, M.; Bowden, M.; Minogue, E.; Diamond, D. 9. Hauser, P. C.; Rupasinghe, T. W. T.; Tan, R. Chimia 1995,
Talanta 2002, 56, 355-363. 49, 492-494.
3. Dasgupta, P. K.; Bellamy, H. S.; Liu, H. G.; Lopez, J. L.; 10. Hauser, P. C.; Rupasinghe, T. W. T. Fresensius J. Anal.
Loree, E. L.; Morris, K.; Petersen, K.; Mir, K. A. Talanta Chem. 1997, 357, 1056.
1993, 40, 53-74. 11. Hamilton, J. R.; White, J. S.; Nakhleh, M. B. J. Chem. Educ.
4. Dasgupta, P. K.; Eom, I.-Y.; Morris, K. J.; Li, J. Anal. Chim. 1996, 73, 1052-1054.
Acta 2003, 500, 337-364. 12. Knagge, K.; Raftery, D. Chem. Educator 2002, 7, 371-375.
5. Kostov, Y.; Rao, G. Rev. Sci. Instrum. 1999, 70, 4466. 13. Cantrell, K. M.; Ingle, Jr. J. D. Anal. Chem. 2003, 75, 27.
6. Kostov, Y.; Rao, G. Rev. Sci. Instrum. 2000, 71, 4361. 14. Skoog, D. A.; Holler, F. J.; Nieman, T. A. Principles of In-
7. Kostov, Y.; Rao, G. Rev. Sci. Instrum. 2003, 74, 4129. strumental Analysis; 5th ed.; Harcourt Brace: Orlando, FL,
8. Hauser, P. C.; Rupasinghe, T. W. T.; Cates, N. E. Talanta 1998.
1995, 42, 605-612.