In general, a spectrophotometer working in the solar radiation wavelength region measures either the transmittance or the reflectance of a sample. That is, the absorbance is not measured directly, but is calculated from the measured transmittance and reflectance (if both can be measured). When a spectrophotometer gives you the absorbance A’ on the logarithmic form, i.e. the optical density OD, which is a common output from many spectrophotometers (OD = A’ = log10(1/T) = alog10(e)x = a’x, "10" supposed to be as a subscript here, i.e. the 10-log, where T is the transmittance, x is the penetration length or depth x, and a and a’ denote the absorption coefficients depending what form is used, which is deduced from the Beer-Lambert law), it is in reality a calculation from the measured transmittance ignoring the reflectance. A spectrophotometer measures the radiation which is collected by the detector from either the transmitted or the reflected beam. Furthermore: T + A + R = 1 (100 %) where T = transmittance, A = absorbance and R = reflectance between 0 to 1 (or between 0 to 100 %). Note that the absorbance A in the above is not the absorbance A’ on the logarithmic form, i.e. not the optical density OD, which is a common output from many spectrophotometers. You may see further details in e.g. chapter 3, 4, 7.6.3 and 7.6.7 in the following article: B. P. Jelle, ”Solar Radiation Glazing Factors for Window Panes, Glass Structures and Electrochromic Windows in Buildings - Measurement and Calculation”, Solar Energy Materials and Solar Cells, 116, 291-323, 2013. In addition, you need to have control over if the transmittance and reflectance are specular (directly transmitted or reflected, i.e. mirror-like) or diffuse (scattered in all directions), where you in the latter case would use integrating sphere accessories to collect the scattered transmitted/reflected radiation.
Absorption coeffcient is usually measured for neat samples (such as films). If you have a solution, you could use the molar extinction coefficient, if your sample concentration in your chosen solvent is known. But if it's not a solution but just a liquid, i would use the absorption coefficient you described above, and the thickness of your sample would be a cuvette length, as Artur Braun pointed out.
In general, a spectrophotometer working in the solar radiation wavelength region measures either the transmittance or the reflectance of a sample. That is, the absorbance is not measured directly, but is calculated from the measured transmittance and reflectance (if both can be measured). When a spectrophotometer gives you the absorbance A’ on the logarithmic form, i.e. the optical density OD, which is a common output from many spectrophotometers (OD = A’ = log10(1/T) = alog10(e)x = a’x, "10" supposed to be as a subscript here, i.e. the 10-log, where T is the transmittance, x is the penetration length or depth x, and a and a’ denote the absorption coefficients depending what form is used, which is deduced from the Beer-Lambert law), it is in reality a calculation from the measured transmittance ignoring the reflectance. A spectrophotometer measures the radiation which is collected by the detector from either the transmitted or the reflected beam. Furthermore: T + A + R = 1 (100 %) where T = transmittance, A = absorbance and R = reflectance between 0 to 1 (or between 0 to 100 %). Note that the absorbance A in the above is not the absorbance A’ on the logarithmic form, i.e. not the optical density OD, which is a common output from many spectrophotometers. You may see further details in e.g. chapter 3, 4, 7.6.3 and 7.6.7 in the following article: B. P. Jelle, ”Solar Radiation Glazing Factors for Window Panes, Glass Structures and Electrochromic Windows in Buildings - Measurement and Calculation”, Solar Energy Materials and Solar Cells, 116, 291-323, 2013. In addition, you need to have control over if the transmittance and reflectance are specular (directly transmitted or reflected, i.e. mirror-like) or diffuse (scattered in all directions), where you in the latter case would use integrating sphere accessories to collect the scattered transmitted/reflected radiation.
You can use the relation " absorption = 2.303*A/d" but assumes that "d" is the internal width of the cuvette which is nearly 4mm or it may vary dependent on your spectrophotometer.