There are often reports that during the blank run as a part of a preliminary study of gradient elution without a sample, a number of peaks are detected as in figure (A). When conducting the same test but with the column was disconnected, the ghost peak disappeared as shown in figure (B). Consequently, the cause was thought to be the column.

However, despite flushing or replacing the column, the baseline would not improve. Therefore many factors other than the column were examined; the cause was found to be, in this case, the water used to prepare eluents. Standard distilled water (which is not considered suitable for HPLC) had been mistakenly used, rather than HPLC grade distilled water which leads to an excellent baseline as in figure (B).

Water purity can have a great impact on gradient elution. Even HPLC grade distilled water will become contaminated with time, causing ghost peaks. A situation which has no significant effect with isocratic elution can cause a problem in gradient elution. In gradient elution, a column is equilibrated with an eluent with low content of organic solvent so that impurities in the eluent are adsorbed and concentrated in the column. After initiating analysis, content of organic solvent increases and impurities begin to be eluted resulting in the ghost peak. The heights of ghost peaks are dependent on the duration of equilibration. The ghost peak did not appear when no column is used because of the absence of adsorption and concentration during equilibration.