Such data could be obtained manually drawing fine vertical and horizontal auxiliary lines on spectra or on enlarged photocopies. The measured distances are linearly transformed to the respective values of the physical or chemical quantities. This is tedious, prone to error, and can only obtain a limited number of usable X–Y data points. A simple method of digitizing data using a personal computer and a pen plotter (1) requires extensive user input and generates data sets of limited size. More sophisticated methods involving light pens or other X–Y data-tracking devices require special hardware and software.

Figure 1. Processing of a retouched bitmap image originally scanned from a figure in an instrumental analysis textbook (2), containing three absorption spectra (molar absorptivity vs wavelength, in nm) of aqueous solutions of Co²⁺, Ni²⁺, and Cu²⁺. The Ni²⁺ spectrum is shown being processed, and SpecScan has been instructed to generate an .XLS file containing 501 equidistant spectral points (at 1 nm intervals).

Using SpecScan is convenient and accurate. An image file is first obtained by scanning a printed plot and saving the image in .BMP format. Some retouching may be necessary to remove black pixel smudges, and the orientation of the bitmap image may need adjustment to align the diagram axes as closely as possible to vertical and horizontal. Most image-processing programs can perform this retouching. The image is then imported into SpecScan, which generates and exports a set of data points that corresponds to the imported plot.

Alternatively, SpecScan allows the user to select the experimental points to be included by visual inspection of the graph. This manual mode must be applied when the printed signal is represented by isolated points and not by a continuous line, and for graphs depicting cyclic processes (for example, cyclic voltammograms), which cannot be automatically handled by SpecScan.

The X–Y data file generated by SpecScan can be imported to any data-analysis program. The original signal then can be reconstructed, rescaled, or subjected to any mathematical transformation or treatment to extract further numerical information, for example a least-squares regression analysis, or combining two or more X–Y data sets to generate a graph of composite signals.

Educators may use SpecScan to demonstrate the wealth of numerical information that can be extracted from figures found in texts as well as showing how one graph can be transformed to another—how an absorbance spectrum can be calculated from a transmittance spectrum or how a background-corrected absorbance can be obtained by a point-to-point subtraction of two absorbance spectra.

Literature Cited

1. Manche, E. P.; Fox, J. T. J. Chem. Educ. 1993, 70, 994.
2. Skoog, D. A.; Holler, F. J.; Nieman, T. A. Principles of Instrumental Analysis, 5th ed.; Saunders College Publishing: Philadelphia, 1998; p 338.