High Performance, Low Cost…

Chameleon - Sample Results

It is common practice in instrumental analysis to use the calibration curve method to establish the instrumental response as a function of analyte concentration. Once a calibration curve has been measured, interpolation is used to determine the concentration of the unknown. For this reason, a high quality calibration curve is a necessary prerequisite in order to give the analyst (you ?) confidence that accurate results will be obtained.

Below are some typical results using the three different flavours of Chameleon in various colorimetric, fluorometric and nephelometric applications. These results testify to the quality of the data that can be obtained using Chameleon instruments, when due care is also taken with the analytical work in making up standard solutions. These experiments (and ones similar to these) are commonly performed as undergraduate analytical chemistry laboratory exercises and many are simple enough to be useful in senior high school classes.

Source Stability Tests

To test our system we first performed a stability measurement using a green, data logging Chameleon (C-525-C), using a TSL230 sensitivity setting of 10x, a scaler setting of /2 and a counting period of 300 msec. A 120Ω resistor on the source PCB set the LED current to 10 mA. The data shown opposite was logged by Chameleon's LabVIEWTM front panel and then stored into an ASCII file for subsequent manipulation using IGOR Pro (Wavemetrics). Total recording time is 10000 sec, nearly 3 hours.

The instrument exhibits a warm-up period, with a steady downward drift by ~ 1% in the LED’s intensity over the first 10 minutes. After ~30 minutes, however, the LED output has stabilized to a standard deviation of 3 counts (in 21000) or approximately 1 part in 7000 - without any signal averaging. This shows that absorbances as small as a few tenths of a milli-absorbance unit can be easily measured using the instrument.

Colorimetric Determination of Phosphate in Coca ColaTM

Phosphorus is an essential element for life. In soils and waters, phosphorus plays a critical role because it is a limiting nutrient for plant and algal growth. In soils and in natural and waste waters, phosphates are the dominant form of phosphorus, and these may occur as orthophosphates (H2PO4-, HPO42-, PO43-), and condensed phosphates (pyro-, meta- and poly-phosphates).

Orthophosphates react with molybdate in the presence of sulphuric acid to produce phosphomolybdic acid. This acid may be reduced using reductants such as tin (II) chloride or ascorbic acid to produce the distinctly coloured phosphomolybdenum blue (PMB) complex, that can be measured spectrophotometrically.

The wavelength of maximum absorption of the PMB (and hence best analytical sensitivity) depends on the reductant used. Some condensed and organic phosphates may be hydrolyzed under the acidic conditions of the PMB reaction, so the analysis does not measure orthophosphate alone. For this reason the chemical species that are detected are referred to as "reactive" phosphates.

A C-880-L Chameleon instrument is first used to measure the absorbances of set of calibration standards, prepared by delivering volumes of a 1 mgP/l working stock solution of potassium dihydrogen phosphate (KH2PO4) into 100 ml volumetric flasks to give final concentrations spanning the range 0 - 800 ppb.

The calibration curve from a typical series of measurements is shown opposite. The performance of the system is excellent, with an R2 value of 0.9996.

Household detergents and drinks such as Coca-ColaTM are common examples of consumer products that contain phosphorus. As the phosphomolybdenum blue method is extremely sensitive, solutions of these samples typically require considerable dilution to bring the concentration into the range spanned by the above calibration curve.

A sample of Coca-ColaTM was analyzed using the colorimeter, after ultrasonication to remove bubbles and then diluting 1000-fold. The phosphate level in the original Coca-ColaTM sample was found to be 177 +-3 ppm. This result is in excellent agreement with the manufacturers stated figure of 172 ppm.

An experiment describing a pH measurement for Coca-ColaTM can be found here.

Colorimetric Determination of Nitrite in Bacon

Nitrite is an important species in the nitrogen cycle. Heterotrophic bacteria convert the nitrogen from plants and animals into ammonia; this is oxidized to nitrite (NO2-) and then nitrate (NO3-) in the process known as nitrification.

Nitrite is added to meats during curing to inhibit growth of bacteria such as Salmonella and clostridum botulinum. Nitrite tends to preserve both the flavour and aroma of meats and also leads to the desirable pink colour of cured meats as it converts to nitric oxide (NO), which then forms the complexes nitrosomyoglobin and nitrosohaemoglobin. Ingestion of large amounts of nitrite is fatal, while lower doses causes methaemoglobinemia. This condition arises when nitrite oxidizes haemoglobin to methaemoglobin, the production of the latter species resulting in impairment of oxygen transport.

At high temperatures such as those used in cooking bacon, known carcinogens such as nitrosamines can be formed through reactions of nitrates and nitrites with meat proteins.

Nitrite can be determined colorimetrically at 543 nm - see for example the series of articles in Vol. 26 of the Journal of the Science of Food and Agriculture.

In this experiment we first prepared a series of standards for analysis with nitrite concentrations in the range from 0 to 35 μM.

For nitrite determinations a C-525-L Chameleon instrument is used to obtain two replicate sets of measurements of the absorbance vs. concentration calibration curve using two different sets of solutions (here shown as red filled circles and green triangles). Least squares fits using a three term (quadratic) polynomial to each data set are also shown.

Reproducibility of the two data sets is excellent, with the data being linear up to a concentration of approximately 10 μM. Slight, but definite curvature is seen in the calibration curve at the higher concentrations. The negative deviation from Beer’s law observed at the upper end of the concentration range is due to the finite bandwidth of the green LED and this can be predicted theoretically (see later discussion).

Bacon samples are prepared for analysis by placing 5 g of finely minced pieces into a 100 ml flask. 40 ml of hot water (80C) are next added to the flask and the meat broken up as much as possible with a glass rod. The rod is rinsed with small portions of hot water so that all washings remain in the flask. The flask is then transferred to an ultrasonic bath and sonicated at 80C for 20 minutes. The mixture is then centrifuged at 4000 rpm for 15 minutes and the supernatant collected in a separate 250 ml volumetric flask. The extraction should be carried out two more times, retaining the total supernatant in the 250 ml flask for analysis. Finally, deionized water is added to make up a total volume of 250 ml before allowing the solution to cool to room temperature.

A triplicate analysis of nitrite in bacon found a mean level of 78 mg/kg, well within the level permitted by Australian standards of 125 mg/kg.

Fluorometric Determination of Quinine in Tonic Water

The fluorescence of aromatic molecules with acidic or basic substituents can be strongly pH dependent, both in terms of the fluorescence yield and the excitation and emission wavelength maxima.

Here an F-370-L Chameleon fluorometer is used to measure a fluorescence calibration curve for a series of quinine bisulphate (QBS) solutions (0-10 ppm) that have been acidified to pH 3.

When exciting at 370nm, quinine emits at approximately 450 nm (ie in the blue region of the spectrum). An optical filter installed in front of the detector in the instrument is important for reducing Rayleigh scattering below ~410nm. In this data we observe an offset due to a slight residual fluorescence from the inexpensive plastic cuvette being used for the measurements. The very slight curvature evident in this data at the higher concentrations is expected theoretically.

A typical tonic water sample (measured by diluting 10x) and acidifying to pH 3 is found to contain 60-70 ppm of quinine.

Chloride in Sea Water by Fluorescence Quenching

QBS fluorescence is quenched (diminished) in the presence of halide ions. Quenching is described mathematically by the Stern-Volmer equation :

F0/F = 1 + Ksv[Cl-]

Here F0 is the fluorescence intensity of QBS in the absence of quencher (here the NaCl), while I is the fluorescence intensity with NaCl present. Ksv is called the Stern-Volmer constant for the quenching. A plot of F0/F vs. [Cl-] should then be a straight line, with the slope yielding Ksv.

Experimental results using an F-370-L Chameleon to measure fluorescence from 50 ppm QBS solutions in 0.05M H2SO4 with additions of 0, 50, 100, 300, 1000 and 2000 mg/L chloride (Cl-) are shown opposite.

The red trace is data obtained using a Chameleon fluorometer, while the black trace was measured on a commercial instrument (Varian). The two data sets are in excellent agreement, with the slope of the Stern-Volmer plot yielding a value for Ksv of 224 M-1. Our result is in good agreement with previous studies that typically quote values in the range 70 – 250 M-1 .

Formazin Nephelometry - Light Scattering

Turbidity in water is caused by suspended and colloidal matter such as clay, silt, finely divided inorganic matter and microscopic organisms such as plankton. The presence of this material causes light to be scattered rather than being transmitted without a change in direction or flux.

A turbidimeter is an instrument that directly measures the intensity of the scattered light due to the presence of particles in solution. The correlation of the measured turbidity with the weight or concentration of suspended matter is often difficult because the size, shape and refractive index of the particles affect the light-scattering properties of the suspension.

The usual approach taken is to compare the turbidity of a sample with that of a calibration standard. This is commonly done using a “nephelometer”, a light scattering instrument that employs a detector positioned at 900 with respect to the excitation direction.

The usual standard for turbidity measurements is a suspension of the polymer “formazin”. A series of calibration standards having turbidities of 3, 5, 7, 10, 20, 50, 100, 200, 400 and 600 NTU were first prepared by diluting a 4000 NTU* formazin stock solution.

A Chameleon fitted with a high brightness white LED and a TSL230 detector mounted in the 900 position (designator = N-W-L) was used to record the calibration curve opposite. Although the scattering intensity shows some curvature at the highest concentrations, an excellent fit can be obtained using a quadratic equation (R2=1). The detection limit of the system in this configuration has been determined to be 0.5 NTU, more than adequate to study real samples taken from streams and rivers.

* NTU stands for "nephelometric turdbity units"

Beer's Law Deviations due to Finite Source Spectral Bandwidth

Light emitting diodes make extremely convenient sources for colorimetric measurements. It is important to realize that the spectral bandwidths of LED's are typically much greater than are found in instruments that employ a broadband light source and a monochromator. Here we explore the implications of finite spectral bandwidths on Beer's law plots and show how the response of an instrument can be predicted theoretically if the emission spectrum of the LED and the absorption spectrum of the analyte are both measured. To this end we describe results obtained using a simple Mathematica notebook that calculates absorbance versus concentration to be performed and apply it to the case of nitrite absorption, as studied previously .

In their textbook, “Spectrochemical Analysis”, Ingle and Crouch discuss various reasons why non-linear Beer’s law calibration plots may be obtained. In the case of an instrument that makes use of polychromatic radiation, negative deviations occur, resulting in the calibration curve bending towards the concentration axis. These authors give plots of the ratio of the observed absorbance to the absorbance at the maximum of the absorption band as a function of the bandwidth ratio s/dλ, where s is the source bandwidth and dλ is the half-width of the absorption band of the analyte. They conclude that s should be less than 10% of dλ when the wavelength is set to the peak maximum to ensure good linearity of Beer’s law calibration plots.

It is straightforward to predict the response of a system with finite bandwidths for both the source emission and the absorber by introducing into the Beer’s law expression the wavelength dependence of both the light source and the sample’s extinction coefficient


In the treatment which follows we will assume that the detector has a flat spectral response over the absorption band of interest and compute the measured absorbance by numerical integration of the functions I0(λ) and I(λ).

Here, a Mathematica notebook is used to perform the calculations to illustrate deviations from Beer's law. For convenience, the sample’s absorption profile and the emission profile of the LED have both been parameterized as Gaussian functions.

The parameters used in the simulation are values taken from spectral measurements for the green LED used in the nitrite determination (Imax at 535 nm, with a FWHM of 60 nm, and from spectrophotometric measurements on a solution of the pink azo dye (Amax occurs at 535 nm, with a FWHM of 85 nm and an εmax of 25739 M-1cm-1.

This plot shows the absorbance versus concentration behaviour predicted for this LED/sample combination, assuming that the detector response is uniform over the spectral region under consideration.

Curvature first begins to appear in this plot at nitrite concentrations exceeding 20 μM; this closely follows what one observes for the experimental data shown earlier.

The final plot presented here shows the ratio of the measured absorbance to the true absorbance as a function of the sample concentration.

The LED-based instrument measures absorbance values that are between 20 and 25% lower than would be measured on a spectrophotometer with a narrow bandpass.

This does not mean that concentrations measured with an LED-based colorimeter will be inaccurate, since the calibration curve method takes account of these reduced absorbances across the concentration range under study.