Desalination

Conservation Specialties

Conservation Specialties Analysis and Imaging Preventive Care Education and Professional Development Workplace Practices About Conservation Lexicon

• Archaeological Artifacts and Sites
• Objects

Group

• Archaeological Heritage Network
• Objects Specialty Group

Section

• Analysis and Testing
• Condition Issues
• Deterioration Issues
• Treatment Techniques and Procedures

Please contribute to the wiki!

Introduction

The chemical definition of a “salt” is a compound that contains an anion and a cation. One such compound is sodium chloride, or table salt, but there are many others. They include those found in groundwater and seawater, such as chlorides, carbonates, sulfates, phosphates, and nitrates; compounds in manufactured materials, such as phosphate fertilizers or chlorides-containing prepared food; and compounds that form as a result of contact with airborne organic acids, including sulfates, formates and acetates. The compounds can be categorized as “soluble” – easily dissolved in water – and “insoluble” – not actually insoluble, but less easily soluble. Soluble salt compounds cycle through solid/liquid phases with cycling relative humidity. In low humidity, soluble salts crystallize into solids; in high humidity, those solids dissolve.

Soluble salts may be introduced into porous materials in a number of ways, including via object use (for example, as a storage container for salted food), via conservation treatment that utilizes materials such as acetic or hydrochloric acid, via off-gassing materials in a storage environment, or, for archaeological material, via groundwater or seawater, since as water seeps into porous archaeological materials in burial, any salts compounds dissolved in the water are introduced into those materials. In most terrestrial burial environments, relative humidity does not significantly fluctuate, and so the salts inside the object remain in a constant state. In marine deposition, because the object is immersed in water, the soluble salts remain dissolved. However, removed from the stable burial environment, a damaging cycle can begin.

In storage environments, when the relative humidity is low, salts will crystallize - expanding many times in size, and exerting considerable pressures on the porous material. The pressures are great enough to cause structural damage. When the relative humidity is high, the soluble salts again become liquid and mobile, and can travel via capillary action to new locations in the object, where new cycles of crystallization/dissolution will promote additional damage. Any porous material can be affected, including ceramic, bone, ivory, stone, certain glass, and organic materials.

If the salts species are chlorides, metals can also be affected. The presence of chloride salts can result in formation of chloride corrosion products that cause rapid structural damage to copper and iron alloys. In the case of archaeological material, the development of damaging corrosion may be inhibited during burial due to reduced oxygen, but once excavated, destructive corrosion species can develop quickly if associated chloride salts are not removed.

Soluble salts damage is irreversible. The goal of a desalination treatment is to remove soluble salts to prevent the possibility of that irreversible damage. However, desalination itself can also be destructive. Not only water-soluble salts, but any water-soluble component of the artifact can be altered or removed by desalination treatment, including water-soluble paint binders, coatings, and adhesives; organic elements of composite objects; some ceramic tempers; low-fired ceramics; and on a molecular level, residues that can clarify the object’s manufacture and use, biomarkers, and other compounds. Desalination treatment can therefore preclude certain types of analysis, organic residue analysis and DNA analysis in particular. It may be more important to preserve the artifact’s chemical integrity than its structural integrity. The advisability of treatment should always be evaluated case-by-case, and in consultation with researchers as necessary. It is rarely better to desalinate “just in case.” The likelihood of damage must be weighed against the potential destruction of information.

Predicating the likelihood of soluble salts damage is complex. In addition to artifact salinity, the degree of environmental RH fluctuations, salts species, the distribution of salts within the object, surface treatments, artifact thicknesses, artifact pore size, and, for ceramics, firing temperatures - among other factors - all affect whether damage will occur, where it will occur, and how much damage there will be . In archaeological field conservation, there is a balance between maintaining an efficient artifact processing workflow on-site during an excavation season and mitigating the danger of deterioration and loss. Whether the risk of damage is great enough to justify delaying the workflow so that artifact desalination can be performed is site- and object-specific.

An alternative to desalination is environmental control. If relative humidity is stabilized, the salts will not cycle through destructive crystallization/mobilization phases, and damage will be prevented. If relative humidity is kept low enough or if oxygen is removed, metals will not corrode. If environmental control is not possible storeroom-wide, microclimate storage containers can be constructed that will stabilize relative humidity or remove oxygen.

In general, archaeological objects from maritime sites have a higher risk of salts damage due to prolonged immersion in highly saline environments. Desalination of archaeological objects from maritime sites require special measures, discussed below.

Ceramics desalination by immersion

Soluble salts activity in ceramics produces characteristic damage that includes flaking slips or glazes, microcracks that destabilize ceramic fabrics, and powdering, spalled, or lost surfaces. If desalination treatment is required, a standard desalination treatment involves soaking artifacts in water to remove enough of the salts so that remaining salts pose an acceptably low risk. (An alternative desalination treatment involves the use of poultices to remove soluble salts ). Though the actual kinetics are complex, the concept is simple: if immersed in a clean freshwater bath, solubilized salts will diffuse from the ceramic object into the water, at which point the water and salts can be discarded. It has been previously assumed that larger volumes of water would improve the efficiency and effectiveness of the treatment. It has been demonstrated that is not true in practice. Similarly, it has been assumed that periodic water changes will increase salt removal: whether that is true has been shown to be species specific. To date it is unverified that pure (deionized or distilled) water appreciably increases diffusion rate, another common assumption. Research is needed on this point, but there are indications that pure water may not be required to achieve artifact stability. In most situations, it appears that a single static bath of relatively clean water that is large enough to completely cover the object is all that is needed, making it possible to successfully perform the treatment even in field locations with poor water resources.

Quantifying salts extraction

During desalination, extraction of salts can be quantified by measuring the salinity of the desalination bath with a conductivity meter either equipped with a species-specific electrode, such as a chlorides electrode, or an electrode that measures all dissolved salts. The basic principle behind a conductivity meter is that water that contains dissolved salts conducts electricity more efficiently than water that does not. A conductivity meter measures the time it takes an electrical current to travel between two electrodes: the shorter the time, the more dissolved salts in the solution.

Normalizing conductivity measurements

Conductivity measurements must be normalized with respect to bath water volume, mass of objects, and time in the bath for reasons that are easy to understand intuitively. Imagine that 1 teaspoon of table salt is dissolved in 10 liters of distilled water. The resulting solution would not taste significantly salty – in fact the salt might not be detectable by taste at all. The conductivity of that solution would be very low. However, if the same 1 teaspoon is dissolved in only 10 ml of distilled water, the solution would taste horrendously salty, and a conductivity reading would be extremely high. Exactly the same quantity of salt is present in both cases: but without normalization, the conductivity measurements would indicate a low risk in the first case and a high risk in the second. In order to evaluate the actual degree of risk, the bath volume must be taken into account.

Now imagine that two objects are desalinating in two baths of equal volume, and the conductivity measurements of both desalinations are the same. However, the first object is a large ceramic jar weighing 1000 gm, and the second object is a small ceramic flask weighing 100 gm. Even though the conductivity measurements indicate the same amount of extracted salts in both baths, the salts in the first bath came from a mass of ceramic that is 10 times greater than the mass of the ceramic in the second bath. To put that another way, 10 times more salts per gram of material were extracted from the small ceramic than were from the large ceramic. Even though the conductivity measurements of the two baths are identical, the small ceramic may be at risk while the large ceramic may not be. For that reason, the mass of ceramic material must also be taken into account.

Finally, imagine that all parameters are the same: two equally salty objects of equal mass are desalinated in baths of equal volume. In the first bath, the conductivity is measured after 5 minutes of immersion, and the conductivity is negligible. In the second bath, the conductivity is measured after 24 hours of immersion, and the conductivity is significant. Despite the difference in the conductivity measurements, the second object is not at greater risk - both objects began desalination with equal salt contents. The higher measurement in the second case is simply because more time has elapsed, allowing a greater quantity of salts to diffuse into the bath. For that reason, elapsed time between measurements must be taken into account.

Unlike direct measurements - the parameters of which may vary with every measurement, and so cannot be compared to other measurements - every normalized conductivity measurement can be directly compared to every other normalized measurement, regardless of the institution in which the treatment was performed, the artifact, artifact mass, the desalination bath size, or elapsed time. Because direct comparison of desalination curves for all objects is possible after normalization, an overall picture of the salinity of all ceramics with a particular storage history, all ceramics of one functional type, or all ceramics from from one site, stratigraphic level of the site, or feature can be compiled – even if those ceramics are in different collections. That data can in turn inform the development of treatment protocols for that group of material.

Normalized conductivity (k_norm) is calculated as

k_norm= ∆kL / ∆tg

Where k=conductivity in μS; L=volume of bath water in liters; g=weight of objects in grams; and t=time elapsed since previous measurement in days (or fraction of a day). Because the math can be tedious, it can be helpful to record conductivity measurements in a spreadsheet that has been programmed to automatically calculate the normalized conductivity, such as an Excel spreadsheet or an OpenOffice spreadsheet. Doing so will also facilitate easy graphing, helpful for determining the endpoint of the treatment (see below).

Determining the endpoint

Early desalination protocols advised removal of all soluble salts. We now know that it is unlikely that all salts can actually be removed. In consideration of damage that prolonged soaking may produce, updated protocols aim to desalinate for the shortest time necessary to lower the risk of soluble salts damage. At present there are several ways to calculate the endpoint of the treatment. While additional research is critically needed to establish how much salt removal is necessary to mitigate risk, experientially, it has been reported that only removing that portion of salts that initially quickly diffuse into the bath can lower the risk to an acceptable point. That being the case, a practical way to determine an endpoint is to establish the point at which the extraction rate plateaus (graphing the desalination data may aid with that determination).

A numerical endpoint of k_norm = 2 has additionally been suggested. Based on experiential data only, that numerical endpoint appears to be valid in certain cases: additional research is still needed to determine whether that numerical endpoint is valid across a range of salt species mixtures, archaeological burial conditions, and other variables.

For ceramics that are not from marine archaeological contexts, there are indications that sufficient salts extraction may happen quickly – in some cases, in a matter of hours. In the field, on archaeological sites that have a lower salt content, routine pottery washing may provide sufficient desalination without the need for additional measures that would force a time delay between excavation and processing by the ceramic specialists.

Ceramics desalination by immersion: step-by-step

The following ceramics desalination procedure is adapted from White, Pool, and Carroll 2010 pages 48–49.

1. Determine whether desalination by immersion is the best treatment for the situation. Consider the danger to any water soluble components and the impact on research value.
2. Identify a desalination container that can accommodate the ceramic to be desalinated plus enough water to completely submerge the ceramic.
3. Create a spreadsheet or table in which to record bath water volume, ceramic mass, date and time, and conductivity in µS.
4. Measure and record the conductivity of the water to be used in the desalination bath.
5. Record the dry ceramic mass in grams.
6. Record the volume of water to be used in the bath in increments of liters.
7. Completely submerge the ceramic in the bath water in the desalination container. The ceramic may be in the container as the water is added, or may be placed in the container after the full volume of water is added.
8. Record immersion time to the minute.
9. Observe the ceramic for a short period to verify that it remains stable underwater.
10. Record an initial conductivity measurement within the first five minutes and as often as desired during the period of observation. For each measurement, record the time to the minute and the conductivity.
11. For each conductivity measurement, calculate the k_norm (or utilize a spreadsheet that calculates it for you).
12. Record additional conductivity measurements, the time taken, and the k_norm at approximately one to five hour intervals. A set schedule is not necessary. Prior to measuring, gently stir the bath to distribute salts homogeneously.
13. Remove the ceramic from the bath at the point at which the k_norm values plateau, or alternatively, at k_norm = 2. That point may be reached in a matter of hours; other desalinations may take several days.
14. Remove any sediment on the surface of the object with clean water and a brush.
15. Allow the ceramic to air dry.

Metals desalination

[This section is under development]

Marine site desalination

[This section is under development]
If the artifact is coming from a maritime environment the most common way to achieve desalination is through a series of staged baths. These baths are a mixture of tap water and seawater, usually starting with a 25/75 ratio to safely remove the salts. “In this way the concentrated salt solution within the material diffuses out into the less concentrated solution” and the object begins the long process of chloride removal. After the bath has plateaued, meaning the object is no longer releasing chlorides and the bath water solution has stagnated, the conservator moves the object to the next staged bath where the ratio of tap water is increased and the seawater is decreased. This offers the conservator the ability to regulate the process, allowing the chlorides to be removed from the object at a safe rate that will not destroy the structural integrity.

Factors such as the size and material of the object play into the amount of time an artifact will have to go through the desalination process. For example, the cannons brought up from the Beaufort Inlet Wreck, thought to be Blackbeard’s Queen Anne’s Revenge, are subject to years of the desalination process and will not be put on display until they are safely desalinated. To ensure that the desalinization process is safe for the object the “conductivity (which is dependent on the salt concentration) or the chloride ion content of the washing water" is closely monitored; permitting the researcher to know the rate at which the salt is coming out.

This process is vital for any maritime artifact that cannot be left in-situ. The removal of salts, specifically chlorides, insures the structural integrity and stability of the artifact. A series of staged baths is the most common way to desalinate and prep the object for life out of the water. Conservator’s knowledge of the process of desalinization makes bringing to life shipwrecks, like the Beaufort Inlet Wreck, possible. Safely conserving maritime artifacts is not only an art, but a long process worth the wait.