Oddy Test Program Planning
Back to AIC Wiki Main Catalogs Page
Back to Research and Technical Studies page
Jump to AIC's new Oddy Test (& More) Results Tables
Jump to AIC's Oddy Test homepage
Jump to MWG's Oddy Test Program Operation page
Wiki Contributors: Julia Sybalsky, Christian Derek Aguilar, Devon Lee, Eric Breitung, Elizabeth Holford, Sarah Freshnock, David Thickett, Laura Gaylord Resch, Rachel Heyse, Elena Torok, and Catherine H. Stephens.
Planning a Program: Budgeting Time and Money[edit | edit source]
There are many issues to consider when setting up an Oddy testing program. These choices can impact the repeatability and reliability of results, initial and ongoing costs, and the time required to complete testing. Since conducting one’s own testing requires a significant investment of resources, before diving in, one should weigh that against alternative ways to evaluate materials which might be better suited to a particular institution or project.
Three basic protocols based on those developed at the Metropolitan Museum of Art (MMA), the Indianapolis Museum of Art/Winterthur Museum (IMA/W), and the British Museum (BM) have been evaluated by the Materials Working Group. They fit two general categories: the IMA/W and BM methods utilize a borosilicate test tube and silicone stopper as the vessel and coupon holder, respectively; the MMA method utilizes a borosilicate jar with screw-top lid and Viton™ o-ring as the vessel and a stainless steel and/or nylon hanger as a coupon holder. The reliability and repeatability of test results differs between the two test types. The cost to purchase supplies for the two test types is also different. Both aspects are addressed in the analysis below. In addition, procedural choices that apply to any Oddy method are outlined, including the reuse of particular test components, washing methods, and documentation practices.
Cost estimates are based on a detailed analysis conducted in 2024, found here. The spreadsheet also acts as an itemized budgeting tool that can be customized by scaling test capacity up or down; selecting test protocol, washing, and reuse procedures; and making line-item edits to lab equipment, tools, etc. It differentiates startup and recurring costs, as well as the per-test cost for consumables, and the total time needed to execute testing using the test capacity and procedural selections specified; and accounts for costs associated with retesting when needed (due to excessive water loss or inconsistent test results) and minimum purchasing quantities where they apply. It does not take shelf life or the anticipated usable lifetime of tools and equipment into consideration.
Testing Capacity[edit | edit source]
Testing capacity (the number of replicate pairs, sample or control, that can be run simultaneously) has an important influence on the rate at which testing can progress, as well as budget and storage needs. In planning, it is helpful to consider the size of one’s oven and the space available for storage of glassware, stoppers/lids, etc. when they are not in use. Capacity is reflected primarily in the cost of consumables and test components like glassware and lids/stoppers. Minimum purchasing quantities such as those for metals (10cm square sheet, equivalent to 48 coupons), jars, tubes, and replacement lids/stoppers, (all sold in cases of 10) come into play. For this reason some may find 24 tests a convenient number to use in planning, particularly for recurring costs.
Universal Costs[edit | edit source]
Many equipment and supply needs associated with Oddy testing are universal and are not impacted by the testing protocol that one uses. For example, standard equipment includes a lab oven, a balance, and a lamp with magnification for viewing coupons. Most small tools used in handling coupons and assembling the test are also universal. These include items like tongs, tweezers, scalpels, wash basins, and drying racks that are common in conservation studios and labs. In addition to these one-time startup needs, one must also budget for the ongoing cost of low-use consumables purchased occasionally (for example, weighing paper, Mylar(™), or aluminum foil used in sample preparation) and high-use consumables like metals and gloves that can be estimated on a per-test basis.
Procedural Choices[edit | edit source]
Procedural choices impact the cost and time expenditure required to execute Oddy testing as well as how likely it is that the levels of corrosion in repeat tests will match. Beyond the testing protocol, washing procedures, reuse of certain test components, and documentation methods should also be considered.
Test Protocol[edit | edit source]
The selection of a testing protocol may take into account existing supplies, equipment, and expertise that can be leveraged to streamline setup and minimize costs. However, there are key distinctions between the MMA and stopper methods (IMA/W and BM) that result in different results, and different levels of reliability and repeatability. These differences are important to consider in deciding which protocol is best for your needs. The MMA test costs more money to set up and run than the IMA/W method. However, it results in fewer tests with results that are difficult to interpret and/or require retesting. When this is taken into account, the need for more frequent retesting of the same materials can lead to higher supply and time investments that in turn make the stopper methods more costly than the MMA over the long term.
Data from an early version of the MMA’s Oddy test that utilized silicone stoppers supports this judgment. The method, which used a stopper to hold the metal coupons, and a screw-top lid to keep the stopper in place, was developed to solve one of the main issues with the IMA/W test: when inserted into a test tube, the stoppers often dislodge, allowing significant loss of water and volatile chemicals. Users of the IMA/W protocols have attempted to solve this issue in myriad ways including wrapping the top of the test tube with teflon tape, and using boards held in place above and below a set of test tubes with velcro. The authors are unaware of IMA/W users that have tracked test weight to monitor this loss, likely because it is difficult to know how much weight the tape or other binding system has lost during heating, and thus to know whether the test tube was well-sealed during the test. Because of this limitation, the MMAs early version of the test, which utilized a screw-top lid, silicone o-ring, and silicone stopper, is the closest approximation to the IMA/W test setup that allows for tracking weight loss.
Analysis of data from the MMA’s stopper tests indicates that this system lost more than 30% of its weight about 55% of the time. Of these, about 1 in 5 tests produced at least one set of non-matching coupons between the A and B jars. Conceptually, this means that for jars with weight loss at this level, significant amounts of volatile chemicals (including water) were likely lost from the test. In contrast, for MMA tests where a lid and Viton(™) o-ring were used instead of a stopper, the rate of high weight loss like this was 4%. In both systems, the rate of non-matching results in tests with minimal weight loss is 8%. While this does not directly show how often repeat testing is required when using the IMA/W or BM methods, it does indicate a significant difference in the reliability of tests that utilize stoppers versus those that utilize viton-gasketed lids with either nylon or metal coupon holders; and provides a basis for estimating the potential extent of retesting required by each protocol.
| Viton o-rings, screw-top lids, and steel or nylon coupon holders | Silicone stoppers, silicone o-rings, and screw-top lids | |
|---|---|---|
| % of tests with volatiles/water loss >30% | 4% | 45% |
| % of tests with non-matching results in A and B jars, volatiles/water loss >30% | 0% | 10% |
| % of tests with non-matching results in A and B jars, volatiles/water loss <30% | 8% | 8% |
| Overall rate of irregular test results | 12% | 63% |
Specific costs associated with the MMA and IMA/W protocols reflect tools and test components (ex. jars, o-rings, tubes, glass vials, etc.) and consumables (ex. solvents and replacement parts) required in addition to the universal costs above. One’s approach to washing and the reuse of test components will also directly impact some of the financial and time costs associated with either protocol - these are discussed further in separate sections below.
Startup costs for the MMA method are greater than those for the IMA/W, with the difference being attributable to small tools and testing supplies. The recurring cost of consumables for the MMA method is also higher, particularly if one is not reusing test components. However when all equipment/supplies, recommended strategies for component reuse, and anticipated retesting are taken into account, the IMA/W test has a higher cost per test. The retest rate associated with each protocol also impacts the anticipated time investment in test setup and evaluation; washing procedures have smaller impact.
| Universal | MMA | IMA/W | |
|---|---|---|---|
| Start-up Costs, 24 tests (USD, 2024) | $6167 to $6747 | $4772 to $5222 | |
| Equipment | $4000–$4500 and up | ||
| Small Tools | $300 | $455 | $0 |
| Testing Supplies | $57 | $1320–$1400 | $335 |
| Recurring Costs, 24 tests (USD, 2024) | $1741 to $2371 | $1567 to $2229 | |
| Low-Use Consumable Testing Supplies | $170 | $0 | $10 |
| High-Use Consumable Testing Supplies | $930 | $525–$1035 | $245–$327 |
| Retesting - High Estimate | 12%: $175 to $236 | 63%: $740 to $792 | |
| Retesting - Low Estimate | 8%: $116 to $157 | 18%: $212 to $226 | |
| Total Expense, per test | MMA | IMA/W |
|---|---|---|
| Active Time (minutes) | 112 – 130 min. | 112 – 125 min. |
| Retesting - High Estimate | 12%: 9-11 minutes | 63%: 48-56 minutes |
| Retesting - Low Estimate | 8%: 6-8 minutes | 18%: 14-16 minutes |
| Total Active Time | 118 to 141 minutes | 126 to 181 minutes |
| Consumables | $51 | $46 |
| Retesting - High Estimate | 12%: $6 | 63%: $29 |
| Retesting - Low Estimate | 8%: $4 | 18%: $8 |
| Total Cost (USD, 2024) | $55-$57 | $54-$75 |
The MMA protocol requires selection of either steel or sintered nylon coupon holders. Steel and sintered nylon coupon holders specified in the protocol have been tested extensively by the MMA and found to provide equivalent outcomes, except with regard to the detection of sulfides. Sintered nylon holders were found to provide slightly higher sensitivity in the detection of sulfides as a silver tarnisher and may be a better choice when this is important to one’s testing application. However because sintered nylon is inherently porous, these holders are more likely to absorb contaminants in testing and should be considered single-use. Steel holders do not absorb contaminants and are indefinitely reusable. Thus if one selects nylon holders to improve the sensitivity of the test, that selection will be reflected in higher recurring costs and an increase in plastic waste. An acceptable compromise is the use of one steel and one nylon holder, which allows for maximum sensitivity with moderate recurring costs.
| Holder Type | Pros | Cons | Recommendation |
|---|---|---|---|
| Steel |
Reuseable, no recurring cost or plastic waste |
Slightly reduced reactivity of sulfides with silver coupon |
Appropriate when sulfides are known not to be present or detection is not a priority |
| Sintered Nylon | Slightly higher reactivity of sulfides with silver coupon |
Single-use, highest recurring costs, most plastic waste |
Appropriate when sulfides are suspected and detection is a priority |
| Combination (Steel in one replicate, nylon in the other) | Balances sulfide detection with cost and sustainability |
Higher recurring cost than using only steel holders, more plastic waste than using only nylon |
Good alternative balancing sulfide detection with lower recurring cost and reduced single-use plastic waste |
Washing Procedures[edit | edit source]
One’s approach to washing reusable test components like jars, tubes, lids, vials, and stoppers impacts budget, the time needed to prepare a test, and potentially the likelihood that contaminants will make their way into the test setup. In practice, one cannot optimize all of these factors at the same time. The MMA and IMA/W methods both specify washing procedures to be used. Striking a balance between cost, efficiency, and repeatability requires assessing the scale of one’s program, time and budget limitations, and one’s tolerance for the possibility of irregularities in the test results leading to retesting.
| Washing Procedure | Startup Cost ($,USD, 2024) | Recurring Cost ($,USD, 2024) | Time to Complete (minutes per test) | Pros/Cons | Recommendation |
|---|---|---|---|---|---|
| Detergent & deionized water handwash | $6755 and up | $530-$645 | 10-17 min. | Efficient; low-cost; possible elevated risk of contamination from residues | May be appropriate for a small testing program with limited space and time available |
| Acid-base handwash | $6820 and up | $725-$840 | 17-26 min. | Time-consuming; low-cost; lowest risk of contamination from residues | May be cost-efficient for a smaller testing program where minimizing irregular test results is a priority |
| Laboratory dishwasher | $24,725 and up | $1125-$1210 | 4-8 min. | Efficient; high-cost; lowest risk of contamination from residues | May be cost-efficient for a larger (>1000 tests) testing program where minimizing irregular test results is a priority |
Detergent & Deionized Water Hand Wash[edit | edit source]
This method involves soaking or washing test components in a solution of lab detergent and tap water, then rinsing thoroughly with tap water, followed by additional rinsing with deionized water. The largest associated cost is for a deionizing system. If water purification is already installed, further investment is modest. Detergent and water filters are recurring costs. Though it is possible that tenacious residues from previous use may not be fully removed, the likelihood of introducing impurities from the water is low. More work is needed to determine whether this method is associated with more irregular results than methods that use acid and base (below).
Acid-Base Hand Wash[edit | edit source]
This approach, described in detail in the MMA test protocol, supplements the Detergent and Deionized Water Handwash above with acid and base baths. After detergent washing, glassware is soaked first in base, then in acid, before a final rinse. Non-glass components are cleaned as above, with no exposure to acid or base. Additional startup and recurring costs for lidded soaking containers, chemicals and pH strips are modest, but this method requires the greatest time investment.
Laboratory Dishwasher[edit | edit source]
Using a laboratory dishwasher capable of running deionized water, lab detergent, acid and base rinses mitigates the risk of contamination, while providing much greater efficiencies. Large glassware (jars, tubes), steel and nylon hangers, o-rings, and lids can be cleaned in a dishwasher, but vials with narrow openings are not effectively cleaned this way and must be washed by hand using one of the methods above, or with solvent. Despite its high initial cost, if managing contamination risk is a priority, a dishwasher may be cost-efficient for larger testing programs in which at least 1000 or more tests will be conducted over time. Additional recurring costs include acid and base solutions. A more detailed cost-benefit assessment must take into account specific equipment and product selections, protocol selection, and the cost of staff time for testing. Detailed instructions for lab dishwasher use are provided in the MMA test protocol.
Re-Use Procedures[edit | edit source]
Reusing certain test components such as lids and o-rings (MMA protocol) or stoppers (BM and IMA/Winterthur protocols) can reduce ongoing costs and the environmental impact of testing. However, reuse, particularly of porous components, can introduce contaminants to the test since these components may absorb reactive chemicals emitted during previous tests. For this reason, both protocols contain recommendations that limit the conditions under which specific components should be reused.
| Re-use Strategy | Description | Replacement cost, per test
(USD, 2024) | |
|---|---|---|---|
| MMA – lids, rings | IMA/W – stoppers | ||
| None | Only new components (lids, rings, stoppers) are used. | $21 | $5 |
| P Only | Only components from tests with a “Permanent” result are reused. Components from “Temporary” and “Unsuitable” tests are discarded. | $19 | $5 |
| A: New & P B: T up to 3 times |
New components and those from tests with a “Permanent” result are used in one replicate (A); lids/stoppers from “Temporary” tests are reused up to three times in the other replicate (B). | $12 | Not recommended |
| P & T, up to 3 times | Components from tests with a “Permanent” result are reused freely. Lids/stoppers from “Temporary” tests are reused up to three times. Rings from ”Temporary” tests (MMA), lids/stoppers after the fourth “Temporary” test, and all reusable components from “Unsuitable” tests are discarded. | $11 | Not recommended |
Protocol Based on the MMA Method[edit | edit source]
Data from tests conducted at the American Museum of Natural History and the Metropolitan Museum of Art was analyzed to better understand how reuse of lids utilized in the MMA method affects test outcomes. When lids, rings, and sintered nylon coupon holders from “Temporary” tests were reused up to 4 times (designated T1, T2, T3, T4), increasing instances of reuse were associated with a steady upward trend in the retest rate. A “retest” is assigned as the test result when there is a mismatch in the type of corrosion on any set of coupons from a pair of replicate jars, A and B; or when a difference in the extent of corrosion on the A and B coupons results in a difference in coupon rating.
| Instances of Re-Use in B Replicate | ||||||
|---|---|---|---|---|---|---|
| 0 (P/New) | T1 | T2 | T3 | T4 | ||
| Reused Lids, Rings, Nylon holders | Total Tests | 60 | 128 | 59 | 10 | 5 |
| Retests | 5 | 16 | 18 | 4 | 0 | |
| % Retest | 8% | 13% | 31% | 40% | 0% | |
| Reused Lids | Total Tests | 73 | 229 | 99 | 46 | 13 |
| Retests | 6 | 17 | 13 | 6 | 1 | |
| % Retest | 8% | 7% | 13% | 13% | 8% | |
When reuse of “Temporary” components was limited to lids alone, this trend did not appear. In tests with new and reused lids from “Permanent” tests on both A and B replicates, the retest rate was 8%. Thus 8% is understood to represent a baseline for how often mismatching coupons occur in the test result regardless of reuse. With increasing reuse of lids (only) from “Temporary” tests, the retest rate varied 1-5%, but did not steadily increase, and is likely a reflection of the inherent variability of the system. It is not yet known whether an increase in irregular test results would emerge if “Temporary” lids were reused indefinitely. Further work is needed.
The combined AMNH and MMA reuse data informs the following recommendations when using the MMA method: lids and rings from “Permanent” tests, and lids (only) from “Temporary” tests. Both components should be discarded after an “Unsuitable” result. It is also recommended to use at least one new lid/o-ring setup per test. Until further work is completed to understand outcomes associated with greater reuse of “Temporary” lids, a conservative approach is to discard reused lids after the fourth “Temporary” test. Finally, though further work is needed to confirm the presence of contaminants in reused nylon coupon holders, their reuse is not recommended based on the above data.
When following a reuse strategy that discards components from “Unsuitable” tests, the reduction in cost and waste for the reused components is driven by the average distribution of “Permanent”, “Temporary”, and “Unsuitable” test results, and the number of reuses allowed. Among tests conducted at the MMA and AMNH, 9% of valid test results were “Permanent”; 56% were “Temporary”, and 35% had an “Unsuitable” result. Consequently, in each test cycle, roughly two-thirds of reusable components at most were actually reused, putting a natural check on how many times any particular lid or ring will remain in use. If reuse is allowed to continue indefinitely, lids will generally encounter an “Unsuitable” material before the ninth test, so in practice reuse is limited to about eight cycles.
Protocol Based on the IMA/W Method[edit | edit source]
This protocol recommends limiting reuse of stoppers to those from “Permanent” tests in order to reduce the risk of contamination. An analysis of data from stopper tests conducted at the Metropolitan Museum of Art supports the premise that reuse may be associated with a greater need for retesting due to non-matching results in tests with minimal water loss. The table below shows reuse statistics from an early version of the MMA’s Oddy test that utilized silicone stoppers to hold the coupons inside the same lidded jars used in the current MMA test. When combining new and used stoppers in the same test, 26% produced different types or significantly different levels of corrosion on at least one set of coupons in replicate jars. When new stoppers were used in both jars, differences in corrosion between replicates were only observed 8% of the time. Further work is needed to distinguish outcomes reusing “Permanent” and “Temporary” stoppers, and accounting of instances of retesting when reusing stoppers for the IMA/W method is underway.
| Tests utilizing new stoppers in both replicates | Tests utilizing a reused stopper in one replicate | |
|---|---|---|
| Total Tests | 223 | 43 |
| Retests | 17 | 11 |
| % Retest | 8% | 26% |
Coupon Photography[edit | edit source]
To accurately document test results for reporting and sharing, it is important to photograph coupons under tightly controlled lighting conditions. The Metropolitan Museum of Art (MMA) and Smithsonian National Museum of the American Indian (NMAI) have each developed imaging workflows for coupon documentation. They are comparable in the time needed to implement them (approximately 19 minutes per test), but one utilizes a rigid frame to which camera and lighting are secured, while the other utilizes a modified studio setup. Both workflows rely on tethered capture, a color reference, as well as diffuse and glancing-angle light.
Rigid Frame Setup (MMA)[edit | edit source]
The MMA’s rigid frame setup is constructed from T-slotted extrusion secured to an optical breadboard base, with adjustable camera and light attachments. Two different light sources (a 12-inch square LED light panel and a 27.4mm high-flux LED integrated array) are specified for diffuse and glancing-angle lighting. While most parts are pre-cut and can be assembled without specialized tools, some simple wiring and soldering are required. The workflow also requires color profiling and image processing software. For accurate color, the setup should be re-profiled any time it is moved to a different lighting environment. Being a rigid structure with dedicated equipment, images are of a consistent and repeatable quality. The total associated costs, including a dedicated DSLR camera, a 100mm macro lens, and other items specified, are estimated at $5300. A basic description is provided by the MMA , but detailed instructions for constructing and using this system are still to come.
Modified Studio Setup (National Museum of the American Indian (NMAI)[edit | edit source]
The imaging workflow developed at the NMAI modifies a traditional studio setup. It utilizes a set of monolights (Broncolor Minipuls C80 or equivalent) on light stands with softboxes and frosted bulb covers to create diffuse and glancing-angle lighting conditions. The workflow also requires a flash sync unit and image processing software (Adobe Camera RAW). Costs vary with the camera, lens, and lighting units selected, and will depend on whether existing studio equipment is available for use. Including a DSLR camera, lens, monolights, software, and other items specified, they are estimated at $4125.
