Skip to main content

Formulating Alum Hematoxylin

H & E Staining

Probably the single most commonly used dye in histotechnology is hematoxylin or hematein, mostly used to demonstrate nuclear morphology. Such solutions usually contain hematoxylin and an alum, and are called hemalum or alum hematoxylin solutions. Many formulae have been suggested. This site, for instance, lists more than 50 hemalum solutions. They vary in the amount of hematoxylin, the amount and type of aluminum salts, solvents, oxidizing agents, and stabilizers. The variations are, at times, almost contradictory and make the underlying principles on which these solutions are based difficult to make clear.


At its minimum, there are three items needed to produce an effective nuclear staining alum hematoxylin. These are:

  • Hematoxylin or hematein, as the dye.
  • An aluminum salt, as the mordant.
  • A solvent, the simplest being water.

In addition to these three items, other ingredients may be added. These are not essential but modify the behavior in some fashion. They include:

  • Oxidizing agents to convert hematoxylin (the dye precursor), to hematein, (the dye).
  • Acids to adjust pH, primarily for extending the useful life of the solution but which can, depending on the solution, also affect the selectivity of the staining.
  • Stabilizers to inhibit further oxidation once the hematein has been formed, since this can shorten staining life and produce poorly defined nuclei.
  • Additions to the solvent are also common, often to inhibit evaporation or precipitation.

Due to its importance, the oxidation, or ripening, of hematoxylin to hematein is the subject of a separate discussion. The principles involved in the attachment of mordants to some dyes are also the subject of a separate discussion, although the specific application to hematein is mentioned briefly here. This page will focus on the relationship between the dye and the mordant, and the effect of adding acid.


Hematoxylin chemical structure


Hematein chemical structure


From the two structural formulae above it can be seen that hematoxylin and hematein differ by only one hydrogen atom. Removal of this hydrogen from hematoxylin is accomplished either naturally by atmospheric oxygen or by using mild oxidizing agents and results in a compound with a hydroxyl group adjacent to a carbonyl group. This configuration is one that facilitates the formation of coordination complexes with metals including, but not limited to, aluminum. Thus the aluminum may be perceived as a link, or bridge, between the anionic dye hematein and a negatively charged nuclear phosphate group.

Adjacent hydroxyl and carbonyl groups of hematein

Adjacent hydroxyl and carbonyl groups of hematein

Co-ordination complex between hematein and aluminum

Co-ordination complex between hematein and aluminum

Co-ordination complexes of this kind are formed with a covalent bond between the aluminum and the oxygen of the hydroxyl group, and a co-ordinate bond between the carbonyl oxygen and the same aluminum atom. In this way the aluminum is firmly attached to the molecule by a process called chelation, therefore the compound is sometimes called a chelate. However, since a dye is involved, it is more commonly called a lake. It is the lake which is the staining component of hemalum solutions. The simplest view is that it reacts as a cation and attaches to tissue anions, such as phosphate groups of DNA and carboxyl groups of proteins. A more complete view is that the attachment, at least to DNA, is also by means of covalent and co-ordinate bonding.

The amount of hematoxylin in the various solutions varies widely from about 1 gram per liter in Mayer’s solution to twenty times that in Masson’s. Obviously, the actual dye content must have an effect on the staining properties, and it can be observed that the more concentrated the dye content the more likely it is that the solution stains regressively. Most progressive solutions have about 1 gram per liter or a little more, and the regressive solutions have 4 or more grams per liter. However, the dye content alone does not determine this characteristic. Both the amount of mordant present and the pH also have an influence. In general, however, if the concentration of mordant remains constant, a solution with a lower dye content will likely be more nuclear selective than that with a higher content. This can be shown easily enough by adding increasing amounts of a ripened 10% alcoholic hematoxylin solution to 5 mL of a 10% solution of alum, then adjusting the volume to 10 mL with water.


The usual mordant for nuclear staining with hemalum is an alum, or aluminum double sulfate. Ammonium or potassium aluminum sulfates are the most common, although there is no reason at all why sodium aluminum sulfate should not be used. The usual explanation for the use of alums is that they were available in good purity in the late 1800s when these solutions were being introduced. It was therefore a simple convenience to use them in preference to other sources of aluminum. Of course, the second metal in the double sulfate should not be capable of acting as a mordant.

Other aluminum salts have also been specified. Examples are aluminum acetate in Haug’s formula, or aluminum nitrate in Rawitz’ 1909 variant. The more modern Gill’s hemalums use aluminum sulfate. From these examples, it would appear that the source of the aluminum is not that critical in practice. In addition, the question of purity is no longer a concern.

In modern practice, alum solutions of about 50 grams to a liter of solvent, mostly water, is fairly typical. The saturation points of ammonium alum and potassium alum are 142 grams and 139 grams per liter respectively in water. Formulae have been published with amounts of alum ranging from 6 grams per liter up to 142 grams per liter (one specifies 300 grams, but amounts in excess of the saturation point are wasted). A comparison chart shows the dye and mordant content of most hemalum formulae.

When alum and hematein are combined a lake forms. Sometimes heat is used to speed up the process, but it eventually happens regardless of that. This lake is a distinct compound, and in strong formulations, any excess may precipitate as a dark sludge, perhaps due to the continued oxidation of hematoxylin. It should also be noted that hematein continues to oxidize to other compounds, and these also may precipitate out of solution in conjunction with alum. In practical terms, this means that hemalum solutions should be periodically filtered and that doing so is more important for stronger formulations.

The ratio of dye to mordant has a significant effect on the staining characteristics of the various formulae. Just as altering the dye content while keeping the alum concentration constant can have an influence on the nuclear selectivity of the solutions, so can increasing the alum concentration while keeping the dye concentration constant. In general, the higher the alum content, the more nuclear selective the solution is likely to be. Once again, this is easily shown by making a series of alum solutions from 0.1% to 10% and adding 0.1 mL of ripened 10% hematoxylin to 10 mL of each of them.

The usual explanation for this phenomenon is that in mass action systems there is a “competition” for the dye between the mordant attached to the tissue and the mordant still in solution. There is a tendency for the dye to equilibrate between the two, so that when there is a higher dye content, i.e. the ratio of mordant to dye is low, the equilibrium favors the tissue. When the mordant-to-dye ratio is high due to a lower dye concentration, then the equilibrium favors the solution and less dye attaches to the tissue.


Alum solutions are acidic, a 0.2 molar (5.16%) solution of potassium alum has a pH of 3.3, for instance. This is very close to the 50 grams per liter commonly used. At this pH the lake is soluble. As the solution is used and alkaline tap water is introduced into the solution, the pH rises until the lake begins to precipitate. This is shown by a color change from cherry-red to purple-red. It does not take too much contamination from tap water for this to happen, and non-acidified hemalums do not have a long useful life. Where acid is not used, it is advantageous to rinse the sections with distilled water before putting them into the hemalum so as to reduce any alkaline carryover.

It has become common practice to add small amounts of acids to hemalum solutions, initially to extend their useful life. Acetic acid at 2-5% is probably the most common, although others have been used, 0.1% citric acid in Langeron’s 1942 formula, for instance, a solution usually erroneously referred to as “Mayer’s” hemalum. Solutions containing added acid have a much longer useful life, and when they eventually do begin to change color, may be rejuvenated by the addition of a further small amount of acid. There is a limit, of course, to the number of times this can be done. It is a truism, however, that the useful life of hemalums is more closely linked to the depletion of the acid content than to depletion of the dye content. This is especially so with the stronger, regressive, formulae where the dye content is so high that it would take months of continual, intensive, use for it to be significantly lowered by the small amounts removed by attaching to DNA. Again, rinsing with water or with dilute acid of the kind used to acidify the hemalum, is advantageous in reducing alkaline carryover and extends the useful life of the solution.

An observed effect of the addition of acid is that nuclear staining is sometimes more selective. This shows up more obviously with those hemalums having a lower dye content. In fact, some of the stronger type, such as Ehrlich’s or Lillies’, do not show any significant difference. Coles’s 1943 formula, however, has a noticeably different appearance, with much greater nuclear selectivity. Unacidified, and applied in a progressive stain for about ten minutes, this solution resembles a differentiated regressive stain. If acidified and similarly applied, it is selectively nuclear with noticeably less cytoplasmic staining. This is one of the best examples of how a small amount of acid can improve nuclear selectivity.

The improved nuclear selectivity is probably due to a slight lowering of the pH when extra acid is added. This may be sufficient to eliminate some of the reactions with acidic groups of cytoplasmic proteins. As to why it does not do so with stronger formulae, it may be that the high dye content just makes it more likely that cytoplasmic groups would react. It should also be mentioned that hematein can participate in non-ionic attachment of dye to tissues, and the presence of salts can promote these dipole-dipole interactions. Hematein can, in fact, be used in a Best’s carmine type procedure to demonstrate glycogen, a method that is believed to be dependent on hydrogen bonding.

Since lowering the pH can make nuclear staining more selective, lowering it enough to completely inhibit staining from carboxyl groups, leaving only nuclear phosphate to stain, would seem to make sense. Krutsay’s hemalum is such a solution. It contains hydrochloric acid and it is very highly selective for nuclei with a very clean, unstained background. It may be the most nuclear selective, progressive hemalum of them all. However, give usually involves take, and the disadvantage is that the very low pH results in the removal of calcium deposits with no indication of their location, and the presence of small calcium deposits may be a useful indicator.

The effect of adding acid to hemalum solutions can be seen by repeating the two previously described exercises with the addition of 0.2 mL glacial acetic acid or one drop of concentrated hydrochloric acid to each solution.


Tissues stained regressively usually require the removal of some of the blue material. The amount depends on the particular characteristics of the formula used, but also on the personal preference of the pathologist who will be viewing the slides. This latter can vary significantly, ranging from almost none to quite significant degrees of differentiation. There is no “correct” degree. Some pathologists prefer a darkly stained background so they can detect ground substances and mucins, which they find useful in reaching a conclusion. Others prefer a clean background with clear and sharp nuclear staining.

Mordanted hematoxylin can be extracted from tissues in several ways. The two commonest, however, are extraction by acids and by mordants. The former usually uses 0.5% or 1% hydrochloric acid in 70% ethanol (acid alcohol), while the latter usually is confined to iron hematoxylin staining of the Heidenhain type and uses the mordant at reduced strength. It is generally considered that acid alcohol produces the sharpest nuclear delineation, particularly with alum mordants. Mordant differentiation of hemalums produces indistinct definition and should be avoided.


The three most influential factors determining the type of staining to be expected from individual hemalums are:

  • The amount of dye in the solution
  • The amount of mordant, and the ratio between the dye and mordant
  • The pH

To these, we should also add, perhaps, the time for which the solution is applied.

In practical application, the formulae that are most popular seem to fall into two groups. The first is those which have 1 gram dye in conjunction with 50 grams alum, or thereabouts, and which stain progressively. The second is those which have 5 grams dye with 50 grams alum, or thereabouts, and which stain regressively. The ratios are 1:50 and 1:10 respectively and are as suitable to use as reference points for comparison as any.

Having said that, one very popular formula falls outside it. Harris’ hemalum has 5 grams of dye with 100 grams alum (1:20). Interestingly, this solution is often said to be regressive without acid, and progressive with acid. This may be overstated as it also requires a short staining time of less than one minute and gives a noticeably blue background, but it does serve to illustrate the point. Nevertheless, the staining characteristics of an untested hemalum formula can be fairly accurately estimated by looking at these amounts and ratios.

Although the majority of formulae are clearly intended to be used either progressively (1 or 2 grams of dye per liter) or regressively ( 5-6 grams per liter) a few are not clearly one or the other. These intermediate hemalum formulae have between 2 and 4 grams per liter with much the same concentration of alum as the other types. Their staining also falls between, giving darkly stained nuclear with more background staining than given by an obviously progressive formula.

Perhaps, rather than think of hemalum solutions as falling into one of two defined groups, we should consider the various formulae to represent a continuum of staining characteristics ranging from the sharp and highly selective staining obtained with Krutsay’s hemalum to the dense overstaining of a formula such as Lillie’s or Ehrlich’s, with a continuous darkening of the staining as the dye content increases, modified only by the amounts of alum and acid.

It should also be possible for a custom hemalum to be formulated based on these principles. Adjusting the amounts of dye and mordant to increase or decrease nuclear selectivity should enable a progressive hemalum to be made that gives the desired degree of nuclear and background staining that an individual worker would find most useful. Apart from tailoring the appearance, for this reason, it is also a useful exercise to do simply for itself as it teaches control of the H&E staining process. This method is of such importance that all technologists should have a thorough understanding of the principles involved and be capable of manipulating the end appearance as required.

These principles also apply with other mordants, such as metachrome iron hematoxylins, although the application is less striking. The ratio between dye and mordant is also a factor with other dyes, including at least one hemalum substitute. Mordant blue 3, in particular, responds similarly. In practical use, sharper progressive staining with mordant blue 3 is obtained as the amount of mordant decreases, whereas hematoxylin selectivity improves as the amount of mordant increases.


  1. Baker, John R., (1958)
    Principles of biological microtechnique
    Methuen, London, UK.
  2. Susan Budavari, Editor, (1996)
    The Merck Index, Ed. 12
    Merck & Co., Inc., Whitehouse Station, NJ, USA
  3. Baker, J. R., (1962),
    Experiments on the action of mordants: 2. Aluminium-haematein.
    Quarterly Journal of Microscopical Science, v. 103, pt. 4, pp. 493-517.