Iron acetate solution prepared from steel wool and vinegar for ebonizing wood

Abstract

Woodworkers darken wood by reacting iron in solution with tannins in the wood to form blue–black phenolate complexes, a process called ebonizing. According to the literature, the darkness and color are controlled by the tannin concentration (higher, darker) and the counterion of iron (acetate, darker). Iron acetate is not readily available, so woodworkers prepare it in the woodshop by reacting steel wool and vinegar. This reaction was studied. Products were ferrous acetate Fe(CH₃COO)₂^.4H₂O (and H₂) and subsequently basic iron acetate [Fe₃O(C₂H₃O₂) ₆(H₂O)₃](CH₃COO) by air oxidation, giving a red-colored solution. Both inorganic phosphate and air (O₂) were required for maximal reaction rate. A volume of 85 mL of distilled white vinegar per gram of super fine (#0000) aged steel wool provided a final reaction mixture with fully dissolved steel and no precipitated products. Maximal darkening of cherry and red oak wood with minimal application of iron solution was attained by applying a 0.125 M iron acetate solution at a rate of 1 mL per 125 cm² of wood surface. A protocol for creating and using an iron acetate ebonizing solution is described and was demonstrated to be a much-improved version of the one currently promoted in the woodworking literature.

Introduction

In the craft of woodworking, ebonizing is the process of darkening wood, such as cherry (Prunus serotina) and red oak (Quercus rubra), in furniture and other wood pieces to make them appear to be ebony (Diospyros ebenum) [1, 2]. Ebony is uncommon today in woodworking due to its rarity and expense, yet its deep black color is valued. Chemical ebonizing relies on the reaction of wood tannins, polyphenols, with iron (or other metallic) salts to produce brown and blue–black amorphous complexes that remain trapped in the wood fibers. Tannins can form chelates with iron through hydroxy groups in ortho position, yielding mono-, bis-, and tris-type compounds [3]. The reactions give rise to a variety of colors depending on the wood and the iron counterion, as depicted in Fig. 1.

Fig. 1

Birch (Betula alleghaniensis), maple (Acer saccharum), cherry (Prunus serotina), and red oak (Quercus rubra) veneers treated with iron solutions. The iron–tannin reaction produces a variety of colors in the ebonized wood: light green (chloride with birch and maple); light gray (sulfate and gluconate with birch and maple); light brown (acetate with birch and maple, and sulfate and gluconate with cherry); dark gray (sulfate and gluconate with red oak); dark brown to black (chloride and acetate with cherry and red oak). Veneers were photographed 1 week after treatment

The typical amounts of tannins in the four wood species listed in Fig. 1 vary from lower concentrations in birch and maple to higher concentrations in cherry and red oak [4, 5]. The tannin concentration strongly affects the color of iron-treated wood. At low concentrations, iron simply hydrolyzes to the oxyhydroxide Fe(III)OOH which subsequently is slowly reduced to gray-colored magnetite Fe(II)Fe(III)₂O₄, a reduction coupled to oxidation of the tannin phenol group [6]. Indeed, Yamauchi [7] detected Fe(III)OOH in Japanese cedar on exposure to iron nails, but found higher amounts of Fe(II) in ancient trees exposed to iron because of the longer reaction time. At higher tannin concentrations the darker, more stable iron–phenolate complexes predominate [6]. Comparison of ferrous acetate-treated birch and maple with cherry and red oak in Fig. 1 demonstrates the point clearly.

The anion of the iron salt also plays a significant role in the color of ebonized wood. Yamauchi et al. [8] studied four trees (Japanese cedar, chestnut, oak, zelkova) buried or submerged for long periods, called umoregi, and found colors due to reaction with environmental iron that fell into two pairs. One pair (chestnut and oak) contained no chloride and displayed black colors due to Fe(III) complexes with octahedral geometry. The second group (cedar and zelkova) displayed lighter colors and lower iron content, indicating lower tannin content, with one member having an olive green color and a higher amount of chloride. The geometry of the complex was undefined, not clearly octahedral and possibly tetrahedral. Figure 1 demonstrates the same finding—a green color in woods with low tannin content and chloride as the anion of the iron salt. In addition, iron(II) polyphenol complexes oxidize in air (O₂) to the more stable Fe(III) complexes, and the rate of this autooxidation depends on the anion of the iron salt [9]. With ferrous sulfate, the autooxidation is slow, Fe(II) persists [10], fewer deeply colored Fe(III) complexes form, and the colors of ebonized woods are lighter (Fig. 1). With ferrous acetate, the opposite is true, with Fe(III) the predominant oxidation state in treated wood [10]. As a consequence, darker colors develop (Fig. 1).

Based on the previous discussion, true ebonization—making wood black, like ebony—only occurs when wood with a high concentration of tannins is treated with an Fe(III) salt or an Fe(II) salt that can undergo rapid autooxidation. Therefore, woodworkers can blacken cherry and oak and walnut, among the hardwoods with higher tannin content, or do the same with other species that have been pretreated with aqueous solutions of tannins, such as quebracho tea [11]. The choice of iron salt is restricted, since only a few are readily available to woodworkers: ferric chloride solution is a chemical etchant for printed circuit board and photoengraving processes and can be ordered from online retailers; solid ferrous sulfate can be found at garden centers as a soil acidifier and as a nutritional supplement at drug stores; iron complexed with glycinates or gluconates is used to amend lawns and other plantings. As noted above none of these iron compounds are completely satisfactory, and ferrous acetate, which does nicely blacken wood, is not easily obtained.

As a consequence, woodworkers must prepare their own ferrous acetate, most commonly by mixing steel wool from the hardware store and vinegar from the grocery store. This concoction has been not-so-affectionately called a “witches brew” and “liquid nightmare” [12]. The terms are apt, since many recipes exist [13,14,15], many different outcomes have been reported [11], and the whole process seems to woodworkers more magic than chemistry [5]. Neither the woodworking literature nor the scientific literature offers an in-depth study and explanation of the reaction of steel wool with vinegar. To fill this void, the present study aimed to better understand the reaction and to design a foolproof, reproducible method for preparing and using the resulting solution for ebonizing wood.

Materials and methods

Materials

Steel wool

Steel wool was obtained from local hardware stores or by ordering online. Single pieces of steel wool were cut from steel wool finishing pads (Super Fine #0000 grade from Rhodes America company) or from a continuous roll of oil-free steel wool (Extra Fine #0000 grade from BriWax company). Each Rhodes steel wool pad could be unfolded to show 6 segments, with each segment weighing between 1.5 and 2.2 g. No further cutting or shredding of the steel wool in preparation for reaction was performed.

Vinegar

Heinz-brand Distilled White (DW) vinegar (5% acidity), Multi-Purpose (MP) vinegar (6% acidity), Apple Cider vinegar, and other brand and type vinegars were purchased from local grocery stores.

Reagents

Water, purified by reverse osmosis, was obtained from the laboratory faucet. Glacial acetic acid, iron wire, potassium hydrogen phosphate monobasic, various iron salts, and other chemicals were American Chemical Society (ACS) grade or better and purchased from chemical supply companies.

A stock inorganic phosphate solution (10.00 mM) was prepared by dissolving 136 mg pure KH₂PO₄ in water and diluting to exactly 100.0 mL. Standard solutions of phosphate were prepared by volumetric dilution of the stock solution with water. Aliquots of the same stock solution were added to acetic acid solutions to adjust their phosphate concentrations.

A stock iron solution (250 ppm) was prepared by dissolving 0.250 g pure iron wire in 20 mL concentrated nitric acid in a 250-mL Erlenmeyer flask. The reaction solution was heated until all of the solid was dissolved and all of the brown gaseous nitrogen oxides were removed. The solution was cooled, diluted to exactly 1.000 L, and stored in a polyethylene bottle. Standard solutions of iron were prepared by volumetric dilution of the stock solution with 1 vol% nitric acid.

Wood veneers for ebonizing

A pack of wood veneers (heartwood) with 0.60 mm thickness were purchased from a woodworking supply. Birch (Betula alleghaniensis), maple (Acer saccharum), cherry (Prunus serotina), and red oak (Quercus rubra) veneers were studied. The veneer was sanded on one face to a final sanding with 180-grit paper. Dust was blown off with a jet of air, wiped with a damp (water) cloth to raise the grain, allowed to air dry, and then sanded again with 180-grit paper. Dust was again removed with a jet of air. A 5.0 cm² area (2.0 cm × 2.5 cm) was marked on the veneer, and 50 µL (or other volume in some experiments) of iron solution was applied. The solution was spread out evenly over the entire area using a plastic pipette tip. After 24 h the area was rubbed lightly for 10 s with a damp cloth and then for 10 s with a dry cloth to remove loose solid.

Methods

Titration for total acid concentration

The acidity of the vinegars was determined by titration with standardized (against potassium hydrogen phthalate) sodium hydroxide solution (0.1090 M) to a phenolphthalein endpoint [16]. It was assumed that acetic acid was the only significant titratable acid present in the vinegars, so the titration gave the concentration of acetic acid. The acidity of the reaction mixture at the conclusion of the reaction was determined in the same way, except that iron interfered with the endpoint. Excess sodium hydroxide was added first to precipitate green ferrous hydroxide and then to change the phenolphthalein to red (easily seen against the white stir bar). The excess hydroxide volume was computed by knowing the mass of iron (steel wool) to start, and the volume correction was made to give the true acid concentration.

Flame atomic absorption spectrophotometry (FAAS) for total iron concentration

A Perkin–Elmer Analyst 700 instrument was used to measure atomic absorbance and determine total iron in the reaction mixtures. A multielement hollow cathode lamp emitted the characteristic wavelength of 248.3 nm for iron, no background correction was employed, and a lean air–acetylene flame was used for atomization. Standard concentrations ranged from 1.00 ppm to 7.00 ppm to prepare a best-fit linear calibration line.

Visible spectrophotometry for total inorganic phosphate

The method used for determining inorganic phosphate involved the formation of phosphomolybdates and their reduction with ascorbic acid to a blue–violet species, solubilized with sodium lauryl sulfate [17]. Reagent volumes were scaled up by a factor of 40 from the published method: sample 4.00 mL; Mo/Sb reagent 0.400 mL; ascorbic acid 0.600 mL; total volume 5.00 mL. The reagents were added in the order listed in the prior sentence, the mixture briefly vortexed after each addition, and the absorbance was measured after a wait time of 15 min. The absorbance spectrum had two peaks in the 500–1000 nm range: 710 nm (lesser absorbance) and 890 nm (wavelength of measurement). Standard concentrations ranged from 2 µM to 20 µM to prepare a best-fit linear calibration line.

Redox potential measurement for the ratio of Fe(II)/Fe(III)

An Orion combination Pt and Ag/AgCl reference electrode was used to measure the potentials of solutions with varying ratios of [Fe²⁺]/[Fe³⁺]. Aqueous standards of FeCl₂^.4H₂O and FeCl₃ were prepared by dissolving pure solid in DW vinegar. Aliquots of each solution, summing to 1.00 mL, were added to 10.0 mL aqueous 1.0 M KCl to prepare each of a series of solutions with varying concentration ratio from 100:1 to 1:100. A best-fit linear calibration line was constructed by plotting potential in mV versus log [Fe²⁺]/[Fe³⁺].

Capillary electrophoresis measurement for the ratio of Fe(II)/Fe(III)

The method of Gotti et al. [18] was modified slightly to determine [Fe²⁺]/[Fe³⁺] by capillary electrophoresis (CE). Standard or sample (50 μL) was mixed with 50 μL of 1,10-phenanthroline (20 mM in water) for 3 min, then 50 μL of CDTA (20 mM in 0.10 mM NaOH) was added and mixed for 3 min, and finally 50 μL sodium tetraborate buffer (50 mM, pH 9.2), the background electrolyte (BGE), was added and mixed. All reactions were performed at room temperature. Separations were performed on a Sciex P/ACE MDQ Plus system with a 40-cm long × 50-µm i.d. fused silica capillary, integrated UV detection module at 254 nm, and 32-Karat software. Separations were accomplished using a constant voltage of 25 kV and were completed in under 5 min.

Steel wool–vinegar/acetic acid reaction

All reactions were carried out at room temperature (21–25 °C) in air or under nitrogen (as specified), in ambient light, and without agitation. An exact mass of steel wool was placed in the reaction vessel, and an exact volume of either 5 vol% acetic acid or vinegar was added. The steel wool was pushed to the bottom of the vessel several times to remove trapped air. The reaction mixtures were sampled at various reaction times by removing 1 mL, centrifuging to remove any unreacted steel wool, and then quantitatively diluting an aliquot (10–100 µL) of the supernatant in 1 vol% nitric acid to provide a solution within the standard iron concentration range for FAAS. For potentiometry, 1.00 mL of the reaction solution was taken and centrifuged, and then the supernatant was mixed with 10.0 mL of 1.0 M KCl. Prior to each sampling, the reaction mixture was stirred with a glass rod to homogenize the solution.

Reaction residue

After the reaction was judged complete, the entire reaction mixture was centrifuged and then filtered (Whatman #42 paper) to remove solid. Any unreacted steel wool was separated from other solids using a strong Nd–Fe–B magnet to hold the steel in the centrifuge tube as the rest of the mixture was poured from the tube and filtered. The supernatant was stored in a capped, glass bottle. The brown–black solid on the filter paper was rinsed with a small amount of cold vinegar before allowing it to thoroughly dry on the paper at room temperature. The mass of the solid was recorded.

X-ray powder diffractometry

Powder X-ray diffraction (PXRD) data were collected using CuK_α radiation on a Rigaku Ultima IV diffractometer in Bragg–Brentano geometry with a scan rate of 4° 2θ per minute.

X-ray crystallography (single crystal)

The data were collected from a shock-cooled single crystal at 150(2) K on a Bruker AXS D8 Quest three-circle diffractometer with a fine focus sealed tube X-ray source using a Triumph curved graphite crystal as monochromator and a PhotonII charge-integrating pixel array (CPAD) detector. The diffractometer was equipped with an Oxford Cryosystems low temperature device and used MoK_α radiation (λ = 0.71073 Å). All data were integrated with SAINT, and a multi-scan absorption correction using SADABS was applied [19, 20]. The structure was solved by dual methods using SHELXT and refined by full-matrix least-squares methods against F² by SHELXL-2018/3 using ShelXle [21, 22]. All non-hydrogen atoms were refined with anisotropic displacement parameters. Carbon-bound hydrogen atoms were refined isotropically on calculated positions using a riding model with C–H distances of 0.98 Å. Water H atom positions were refined and O–H distances were restrained to 0.84(2) Å. U_iso values were constrained to 1.5 times the U_eq of their pivot atoms.

Optical reflectometry

The color and brightness of ebonized wood was measured using a hand-held reflectometer (Chroma Meter, model CR-400; Konica Minolta, USA). Values were produced in the CIELAB/L*A*B color sphere system [23] and standardized against an ultra-white disk provided by the manufacturer.

Results and discussion

Characterization of reactants

Steel wool

Steel wool is usually produced from low carbon steel that contains iron, carbon (up to 0.3 wt%), manganese (up to 1.5 wt%), copper (up to 0.6 wt%), silicon (up to 0.6 wt%), and possibly other metals [24]. FAAS analysis of the two steel wools studied here showed the following contents after dissolution in nitric acid. The Rhodes steel wool (statistically identical values for unwashed and washed): 100 ± 4 wt% Fe, 0.73 ± 0.05 wt% Mn, and no Cu. The BriWax steel wool: 98 ± 4 wt% Fe, 0.56 ± 0.04 wt% Mn, and no Cu. Carbon and silicon were not determined. The two steel wools had nearly identical compositions in terms of the measured metals and were essentially pure iron.

Steel wool available in the hardware store comes in different grades, from coarse (#2) to fine (#0) to super fine (#0000). The wire diameters were measured with an optical microscope, giving values nearly identical to the nominal values: 40 µm for Rhodes fine and 25 µm for Rhodes super fine. Measurements of both newly purchased and 2-year-old Rhodes steel wool gave the same values. The newly purchased BriWax steel wool was composed of a steel ribbon, rather than wire, of measured dimensions 5 µm × 20 µm.

Vinegar

Two different Heinz brand vinegars were used in this study. Multi-Purpose (MP) vinegar is labeled as the most acidic at 6 wt% acetic acid, while the Distilled White (DW) vinegar nominally contains 5 wt% acetic acid. The measured (by titration) acetic acid concentrations (and wt% acetic acid) were: 1.031 M ± 0.002 M (6.2 wt%) for MP, and 0.872 M ± 0.002 M (5.2 wt%) for DW. The actual and nominal acidities matched very well.

The measured amount of inorganic phosphate P_i in the vinegars was 1.88 ± 0.02 mM (MP) and 1.62 ± 0.02 mM (DW). Two other brands of DW vinegar had similar P_i concentrations. It is reasonable to find phosphate in vinegar, since phosphate is added as a nutrient for the acetobacter organism that is used to catalyze the oxidation of alcohol to acetic acid in the vinegar-making process [25]. No iron was found in either of the vinegars.

Reaction products

Primary reaction product/reaction stoichiometry

The reaction of steel wool and vinegar produces ferrous acetate tetrahydrate. Within 2–7 days, depending on factors discussed below, the reaction goes to completion. In every case, except when product precipitated out of solution (solubility exceeded), the mass of steel wool to start was not significantly different from the mass computed from the total iron solution concentration determined by FAAS at the end of the reaction. In addition, after 7 days of reaction only a trace of dark material (bits of steel wool) was collected from the surface of a strong Nd–Fe–B magnet added to the reaction solution.

The acetic acid concentration in vinegar at the start of the reaction and at the conclusion of the reaction with steel wool was measured by titration to reveal that 2.11 ± 0.04 mol of hydronium ion was consumed per mol Fe. On a per gram basis, iron (steel wool) requires 0.036 mol acetic acid or 41 mL of DW vinegar for full reaction {1.0 g/55.85 g/mol) × 2 = (0.872 M)(0.041 L)}.

The volume of hydrogen gas produced by the reaction was measured by water displacement and converted to moles using the ideal gas law to reveal that 0.85 ± 0.04 mol H₂ was produced per mol Fe. This value was corrected for gas collection efficiency, as measured from the known reaction of zinc metal and hydrochloric acid to produce hydrogen gas. However, the zinc reaction was completed in a few hours, while the steel wool reaction went on for several days, and this could have led to more loss of hydrogen gas and a coefficient of less than 1. In addition, oxygen in the air may play a role as a reactant (see below), thus decreasing the total volume of gas collected. In any case, the reaction equation of iron (steel wool) and acetic acid (vinegar), as determined to the nearest whole number stoichiometry, was: Fe + 2 CH₃COOH + 4H₂O → Fe(CH₃COO)₂^. 4H₂O + H₂. This was as expected.

Oxidized reaction product

Throughout most of the reaction time the reaction mixture was colorless, reflecting the colorless/light green color of solid ferrous acetate. As the reaction reached completion, the solution color turned greenish yellow, then orange. At reaction end, all of the iron and about half of the acetic acid was consumed, so the solution pH increased. Without Fe⁰ present the disproportionation reaction Fe⁰ + Fe³⁺ → 2Fe²⁺ was no longer possible, and with a rise in pH, the air (O₂) oxidation rate likely increased [26]. Therefore, the concentration ratio Fe(II)/Fe (III) decreased, and a dark red complex, likely [Fe(II)Fe(III)₂O(OH)]⁺, formed to impart color to the solution [27]. See Fig. 2 for a visual representation of these changes. Potentiometry and CE were applied to the determination of Fe(II)/Fe(III) over the course of the reaction. Both techniques demonstrated that Fe(II) predominated during most of the reaction time, and Fe(III) grew in relative concentration as the reaction reached completion and beyond. For example, in one reaction mixture stored in a capped glass bottle, the measured ratio was 15:1 three days after completion of the reaction, 1:1 three weeks later, and 1:2 three months later.

Fig. 2

Liquid portion (supernatant) of a 4-day reaction mixture of aged Rhodes steel wool and DW vinegar: a colorless reaction solution not treated further; b with added NaOH (enough to neutralize half of the acetic acid in the vinegar), green precipitates of presumably ferrous hydroxide and green rust I form; c with added NaOH, shaken, and allowed to stand for 2 min, ferric ion forms, producing a yellow–orange color; d with added NaOH, shaken, and some Fe(III) chloride added, the dark red complex is likely formed. e Dark red–brown liquid portion (supernatant) of a 14-day reaction mixture of aged Rhodes steel wool and DW vinegar, containing red-colored basic iron acetate

Controlled evaporation of a red-colored iron acetate solution in air over a week yielded a few semi-transparent, light red–brown crystals (Additional file 1: Fig. S3). By crystallography, the solid was identified as the mixed-valence trinuclear iron acetate–aqua complex [Fe(II)Fe(III)₂O(C₂H₃O₂)₆(H₂O)₃]⁺, also known as basic iron acetate. Basic iron acetate has three iron centers each of which are octahedrally bound to six oxygens, including an oxide at the center [28]. The counterion is likely acetate in the investigated structure. The data (Additional file 1: Table S1) matched well two previously published structures (CSD ref codes ZUSJEC and ZUSJEC01) [29]. They differed by the content of the solvate molecules which are ordered water molecules for the published structures and are highly disordered unidentified molecules for the current data set. The metal complex was unchanged among the three structures. In summary, the reaction product solution of steel wool and vinegar contained both ferrous acetate and basic iron acetate, tending toward the latter as the solution aged. As further proof, rapid evaporation in air of some of the product solution left a brown–red solid that gave an X-ray diffraction pattern nearly identical to a library spectrum of basic iron acetate (Additional file 1: Fig. S2). Any ferrous acetate in the reaction solution oxidized to basic iron acetate during solution evaporation. The term “iron acetate” will be used below to indicate this mixture of ferrous acetate and basic iron acetate.

Factors affecting the reaction rate

Factors that could affect the reaction rate were investigated using 2.00 g steel wool and 170 mL of vinegar. The reaction was followed by measuring the amount of dissolved iron in solution by FAAS, and reaction rates over the first 48 h of reaction were computed. A summary of the results is given in Table 1, and the results are fully discussed in the text below.

Table 1 Mass Fe (g) reacted in 48 h relative to the value for the attribute in the first column

Vinegar

The results with Apple Cider vinegar were inconsistent with only 2 of 6 mixtures showing complete reaction by day 7, but in all cases the reaction was much slower compared to DW and MP vinegars. Perhaps proteins and other ingredients of the cider coated the steel, decreasing the contact between the reactants. Because the reaction was usually incomplete, Apple Cider vinegar was not further studied. DW and MP vinegars gave similar reaction rates. The more common DW vinegar was selected as the vinegar for further study.

Steel wool

Recall that Rhodes steel wool was made of wire of 25-µm diameter, and BriWax steel wool was made of 5 µm × 20 µm ribbon. Smaller dimensions should provide a larger surface area per gram and a higher reaction rate. Experiment confirmed this hypothesis. The BriWax #0000 reacted more than twice as fast as the Rhodes #0000, and the BriWax reaction was complete on average in 2.5–3 days compared to 6–7 days for the Rhodes.

Oils as lubricants are commonly used in steel wool manufacture, so some woodworkers try to remove the oil before reaction with vinegar in hopes that this speeds the reaction. Rhodes steel wool was washed with detergent in tap water and then rinsed thoroughly with tap water. Or it was soaked in hexanes, squeezed to remove hexanes, and residual hexanes allowed to evaporate. Washed steel wool reacted with vinegar at about the same rate as unwashed steel wool. Therefore, washing/removing oil proved to be unnecessary and was not further included in the protocol. BriWax steel wool is produced without the use of oil, according to its manufacturer.

A difference in reaction rate was noted when a new package of Rhodes steel wool was purchased for use in the midst of the study. The age of the steel wool had a positive effect on the reaction rate that could not be reproduced by artificial aging (heating steel wool in a moist atmosphere; 20 min at 218 °C in a toaster oven), see the next section for a discussion of this aging effect.

Surface reactions and phosphate

Iron has a complex surface chemistry. In air and in solution, the surface of iron quickly forms a layer of iron oxides and oxyhydroxides, the main components of iron rust (Misawa 1974). More specifically at low pH, the large number of possible chemical species include Fe²⁺, Fe(OH)⁺, green rust, Fe³⁺, and forms of FeOOH [27]. Green rust I is [Fe(II)₄Fe(III)₂(OH)₁₂]CO₃, where the planar carbonate coordinates with a mixture of ferrous and ferric ions [30]. Fe(OH)²⁺, forms of Fe₂O₃, and Fe₃O₄ can be found upon further air oxidation [31]. Rhodes steel wool was exposed to DW vinegar for 3.0 min, thoroughly washed with distilled water, and air dried. The newly purchased wool did not change in measured diameter (25 µm), but the 2-year-old wool after treatment had a reduced diameter of 15–20 µm. This demonstrated that an iron oxide/oxyhydroxide layer, soluble in acetic acid, had formed over time. In addition, more iron was removed by vinegar from aged steel wool in 48 h (1.0 relative amount) than from newly purchased steel wool (0.59), due to the greater solubility of the surface layer and the resulting smaller wire diameter.

That steel wool reacted very slowly in 5 vol% glacial acetic acid in water in comparison with the same reaction in DW vinegar was puzzling, see Fig. 3. Knowing the content of typical enzymatic reactors that convert ethanol to acetic acid (Additional file 1) [25] and applying a factorial design to screen the major components for their effect on the reaction rate (Additional file 1: Fig. S1), inorganic phosphate was found to act as a catalyst. Indeed, the addition of 1.0 mM P_i brought the reaction rate in acetic acid in line with that in vinegar (Fig. 3).

Fig. 3

Reaction progress curves. Aged Rhodes steel wool (2.0 g) reactions with 170 mL DW vinegar (squares), 5 vol% acetic acid (circles), nitrogen-sparged 5 vol% acetic acid (triangles). Reactions with no added phosphate (orange and black; top line and bottom two lines), 0.10 mM phosphate added (green; 3rd line from top), and 1.0 mM phosphate added (blue; 2nd line from top and 4th line from top)

The likely reason for the catalytic effect of phosphate was revealed by careful reading of the relevant literature. Green rust I oxidizes to ferrihydrite, an amorphous hydrated form of FeOOH, that in turn can crystallize into goethite α-FeOOH [32]. Geothite is × 1000 less soluble in aqueous solution than ferrihydrite [33]. Thus, once goethite forms on the surface of steel wool the iron underneath becomes much less available for attack by acetic acid. However, phosphate can form an inner sphere complex with ferrihydrite at low pH, binding via two Fe–O–P covalent bonds [34, 35] and thereby inhibiting the conversion to goethite [30]. As a consequence, the iron oxyhydroxide layer on the surface of the steel wool remains highly soluble, the iron surface is easily exposed to acetic acid as the ferrihydrite dissolves, and the reaction proceeds at a faster rate. This process appeared to continue throughout the course of the reaction, since the accelerated rate extended to reaction end.

Air (oxygen)/reaction vessel

As described above, the oxide/oxyhydroxide layers on steel wool are slowly formed in an air/oxygen-rich environment as steel wool ages on the shelf and appear to continue to form and play a role in the steel wool–vinegar reaction in solution. Therefore, it was hypothesized that oxygen was an important factor in the dissolution of steel wool in vinegar.

Reaction vessels of different dimensions and different air–liquid interfacial areas were studied for their effect on reaction rate, see Table 2. Glass jars and plastic cups are two of the most common vessels used by woodworkers.

Table 2 Reaction vessel characteristics in study of impact of air (O₂) on reaction rate

The air–liquid interfacial areas were measured when the vessels were filled with 170 mL of liquid. The canning jar had the highest diffusional area and, thus, could likely provide the most oxygen to the reaction. Indeed, the reaction rate in the canning jar was slightly higher than in the plastic cup and significantly higher than in the Erlenmeyer flask, whose top restricted air flow above the solution. Assuming that the vessel material had no significant effect on the reactions, the data demonstrated that air (O₂) was an important component of the reaction. As further evidence, experiment showed that nitrogen-sparged 5 vol% acetic acid was slower to react with steel wool than air-saturated 5 vol% acetic acid (Fig. 3). Likewise, N₂-sparged vinegar was slower to react (data not shown).

Protocol for preparing an iron acetate ebonizing solution

Volume of vinegar per gram of steel wool

Goals for preparation of an iron acetate solution for use in the woodworker’s shop are (1) a solution that is as concentrated as possible and (2) little need at reaction end for filtering out solid—leftover steel wool and iron acetate that exceeds its solubility.

Ferrous and ferric salts precipitate in neutral to basic solutions, because their hydroxides have very low water solubilities. Thus, it is important that the hydrogen ion concentration does not drop too much as acetic acid is consumed during the reaction. For example, if too much steel wool is added to a volume of vinegar, the product might not be ferrous acetate tetrahydrate but instead ferrous and ferric hydroxides. Ferrous hydroxides are green in color (see Fig. 2), and a green color has been observed in poorly designed steel wool–vinegar reactions. To maintain an acidic pH, excess acetic acid should be used in the reaction. Since 41 mL of DW vinegar {0.872 M × 0.041 L = 0.036 mol} are required for full reaction per gram of steel wool {(1.0 g/55.85 g/mol) × 2 = 0.036 mol}, then about 20% more or 50 mL of vinegar should suffice to fully dissolved the steel, keep the pH below 3, and prevent iron precipitation.

When the ratio of vinegar to steel wool was at or above 50 mL per gram, a solid residue—some as flat crystals floating on the solution and some as particles on the sides and at the bottom of the reaction vessel—developed toward the end of the reaction. More solid was found as the volume of vinegar relative to the amount of steel wool (3.00 g) decreased: 0.02 g/250 mL; 0.80 g/200 mL; 1.1 g/150 mL. Saturation of the solution with iron acetate explained these results. In a separate experiment, solid, reagent-grade ferrous acetate was added in excess to a known volume of DW vinegar, and the mixture was shaken for 7 days. Excess solid was removed by centrifugation and filtering, and the supernatant was tested for total iron content by FAAS. Repeated trials gave a solubility of ferrous acetate tetrahydrate in vinegar of 0.25 M ± 0.02 M, and the solubility did not vary significantly upon dilution of the vinegar with water up to a dilution of 20:1. To prevent precipitation of iron acetate toward the end of the reaction, a sufficient starting volume of vinegar is necessary. The required volume is 72 mL per gram of steel wool {0.018 mol/0.072 L = 0.25 M}, but to allow for evaporation over a 3- or more-day reaction about 20% more or 85 mL vinegar per gram of steel wool should be used. Note that this volume is more than necessary to also fully dissolve the steel and keep the pH low, as described above.

Recommended protocol

The protocol for preparation of an iron acetate ebonizing solution is given in Table 3. The protocol yields an approximately 0.125 M solution from steel wool pads (e.g., Rhodes) or an approximately 0.25 M solution from wool made from ribbon (e.g., BriWax), both in 3-day time. Only about half of the steel wool pad dissolves in 3 days, while the wool made from the smaller ribbon should be fully dissolved, see Fig. 4.

Table 3 Protocol for preparing an iron acetate solution for ebonizing wood

Fig. 4

At left are two photos of pairs of reaction mixtures. Each pair has newly purchased Rhodes steel wool in DW vinegar (left) and BriWax steel wool in DW vinegar (right). The photo pair at left is the reaction after 1 day and the pair at right after 3 days. After 3 days, nearly all of the BriWax steel wool was consumed, but about half of the Rhodes steel wool remained. At right are two photos of the filter paper used to remove the small amount of solid remaining and the filtrate of the BriWax-DW vinegar reaction mixture on day 3 and again 1 day later (far right). The filtrate initially contained mostly ferrous acetate (greenish brown solution) and then began to be oxidized in air to basic iron acetate (red–brown solution)

Application of the ebonizing solution

Aliquots (50 µL) of iron acetate solution were spread evenly across 20 mm × 25 mm pieces of prepared cherry and red oak veneer. Increasing concentrations of iron solution, diluted in vinegar, gave rise to progressively darker wood, see Fig. 5. The maximum darkness, indicated by a minimum in the L value from the reflectometer, was obtained using solutions of 0.125 M and higher. A second application of iron solution did not significantly increase the darkness, though it did fill in some lighter areas.

Fig. 5

Cherry and red oak veneer treated with 50 µL of varying concentrations of iron acetate solution, the product of aged Rhodes steel wool–DW vinegar reaction. The brown to black color of the ebonized wood was measured by reflectometry using the CIELAB system. That system is diagrammed to the right of the table [23]. Values of a (green–red scale) decreased slightly and values of b (blue–yellow scale) decreased by about 12 units (bluer). The largest changes, as reported, were along the L axis

Increasing volumes of 0.125 M solution produced darker veneer until a maximum was reached at a volume between 25 µL and 50 µL, see Fig. 6. The required amount of iron solution was estimated to be about 40 µL, since excess solution on the surface of the wood was noted upon addition of 50 µL. Thus, 5 µmole or more of iron added to 5 cm² of wood (cherry and red oak) was expected to give the lowest L value, the most darkening.

Fig. 6

Cherry and red oak veneer treated with varying volumes of 0.125 M iron acetate solution, the product of aged Rhodes steel wool–DW white vinegar reaction. The brown to black color of the ebonized wood was measured by reflectometry using the CIELAB system. The largest changes, as reported, were along the L axis

The iron solution seeps from the wood surface and into the pores and lower layers. Optical microscopy gave an estimate of the depth of penetration of the iron solution by measurement of the dark layer of the wood viewed edge on and along the grain (Fig. 7). The dark layer was not of uniform color or thickness and difficult to judge in some samples. The best estimate of the average ebonized layer thickness was 100–120 µm for cherry and 60–80 µm for red oak wood in the 600-µm thick veneer. Since cherry wood 630 kg/m³ is less dense than red oak 740 kg/m³ [36], the measured deeper penetration of the iron solution in cherry is reasonable. Canevari [37] reported treatment of popular wood with gall dyestuff and found penetration of high iron concentrations to 30 µm. They found SEM–EDX data for iron to be “rather noisy” and inconsistent. That finding was confirmed.

Fig. 7

Cherry (left) and red oak (center) veneer treated with iron acetate solution, the product of aged Rhodes steel wool–DW vinegar reaction, and viewed edge on along the grain (vertical axis) by optical microscopy. The scale of the photos and the average thickness of the veneer was 600 µm. The ebonized wood layer is the slightly darker band on the left side, and its thickness was estimated by ruler measurement of the photo. The third photo (right) is of the holder used to secure the veneer edge perpendicular to the optical path of the microscope. The edge of the wood (and top of the holder) was sanded to 4000 grit prior to microscopy

With a surface area of 5 cm² and an average depth of 0.0070 cm, the volume of penetration of iron solution in red oak was 0.035 cm³, per 40 µL of applied 0.125 M iron acetate, giving an iron concentration in the wood of 143 µmol/cm³ {(0.125 M)(40 µL)/(5 cm²)(0.0070 cm)}. Dagher [38] reported that the concentration of di-substituted phenolics in red oak (500 µmol per g wood) was well-correlated with wood color after ebonization. Assuming the same tannin content of the red oak sample in this study, a rough estimate was that the mole ratio of phenolics to iron in the dark layer of the treated red oak was 2.6 {(500 µmol/g)(0.74 g/cm³)/(143 µmol/cm³)}. This indicated the likelihood that mostly bis-phenolate and tris-phenolate complexes formed, an observation that is quite reasonable and supported by the report of Perron [9].

Conclusion

The findings and recommendations presented here differ considerably from the information that is currently available to woodworkers. In a wikiHow article on ebonizing wood [39], representative of the content of woodworkers’ sites on the internet, the following mistaken notions are espoused. 1 “Steel wool usually has an oil coating that you’ll need to remove before making the solution.” The reaction rate is not significantly changed upon washing with detergent and water or rinsing in nonpolar solvent. 2 “New steel wool is best.” Aged steel wool has an oxyhydroxide coating that is easily dissolved in vinegar leaving a smaller dimensioned wire that reacts at a faster rate. 3 “Apple cider vinegar works better.” Apple cider vinegar reacts inconsistently and at a slower rate. DW vinegar is best. 4 “Fill a glass jar with steel wool and vinegar.” No metrics are given. The relative amounts of vinegar and steel wool are quite important—85 mL per g is best. In addition, the shape of the reaction vessel has an effect. 5 “Let the steel wool and vinegar sit for a week.” An effective iron ebonizing solution can be created in just 3 days.

In the WoodWeb forum [12], ferrous sulfate was said to produce the same results as iron acetate. Instead, this study found the colors produced were markedly different (Fig. 1), and the literature suggests that sulfate anion slows the rate of conversion of Fe(II) to Fe(III), minimizing the production of dark Fe(III) phenolate complexes. Iron acetate is best for ebonizing. Application of the iron acetate solution at a rate of 1 mL per 125 cm² or about ¼ cup of liquid per 7 ft² is sufficient for oak wood. This seems to be a much lesser application than is typically shown in woodworking videos on the internet [5]. Using excess iron solution does nothing but leave surface residue to clean up after the solution evaporates. By following the protocol in Table 3, 340 mL of ebonizing solution are prepared, providing coverage for about 40 ft² of finished lumber.

By way of this work, the theory and practice of ebonizing wood by chemical means is improved, and important knowledge and protocols have been added to the woodworker’s toolkit.