B: Anti-Tarnish testing

Accelerated Life Test Results

of Angel Guard/Midas Tarnish Shield

Rob Brown, August 14, 2019

(editor's note: Midas Tarnish Shield by Rio Grande is repackaged as Angel Guard by Angel Gilding, and so are the same product.)

(editor's note number 2: at the bottom of this page is a link to download a pdf version of this article)

Summary

Midas Tarnish Shield is shown to profoundly inhibit tarnishing of a front surface silver coating in an accelerated life test (immersion in aggressive sulfide plus salt solution). The relative life of a front surface silver mirror with this protection should extend likewise, with the limiting life factors being related to accumulation and cleaning of dust, care in handling, and other factors not related to tarnish.

Unprotected silver coatings on telescope mirrors have had mixed results, mainly due to differences in environmental exposure. A 12.5” was stained with white spots in a single night of storage in high humidity in its second month, and recoated at 6 months. A 28” has gradually become similarly stained and at 11 months is still performing well (but the owner wishes to make an improvement.)

A method of protection that prevents staining and scratching has not been evaluated.

Preparation of samples

Silver coatings were deposited onto flat window glass using the Angel Gilding 2-part spray kit as per their instructions. Special care was given to cleaning the glass to get uniform coatings out to the edges. Several samples were stored in a garage wrapped in tarnish inhibiting cloth for 9 months, and one was left out in the open in a house for the same time. The samples wrapped in the tarnish inhibiting cloth were as good as new, while the unprotected sample had developed a significant gold tint.

One of the “good as new” pieces was sectioned into a size of about 30 x 125 mm using scribe-and-break. It was dipped into a small tub of Midas Tarnish Shield with a depth of 90mm so that 2/3 of the glass got the protective coating. The mirror was marked with a permanent marker at the coating line.

Silver coated slide in tub of Midas

Accelerated Life Test Concept

Since we don’t have the luxury to wait until protected silver tarnishes or otherwise reaches its end of life in a real-world situation, an accelerated life test is used. The idea is that we put the silver into an environment containing the most significant agents that cause tarnish and decay, and concentrate those agents highly so that the effects take place over a matter of hours and not years. We can then compare the behavior of both unprotected and protected silver to determine the improvement over a “standard lifetime” of unprotected silver. Such testing is not without risk or error, but is nevertheless useful for comparison purposes and rough estimates of real-world performance.

Tarnish is mainly driven by the presence of H₂S in a humid atmosphere. Salt (sodium chloride) may be an agent that causes separation of the silver from the glass, and since it is a common material in many local environments it is included in this test.

Preparation of Sulfide + Salt solution

The solution is made of 0.5M NaCl and 0.01M Na₂S

· 500mL distilled water

· 0.4g Na₂S (provided by Jerry Merz)

· 14.6g NaCl (kosher salt, no iodine)

Materials weighed in triple beam balance, mixed in room temp water, and placed outdoors in dry, warm air (~75F).

Test at T=0

Relative reflectivity was measured using the camera method. The camera was a Fujifilm X-E2 mirrorless, manual mode, RAW, 1/30s, ISO 320, F/5.6. A reference white coupon made of Teflon was placed on the mirror with the mirror on the floor, well-lit by overhead LED T-8 tubes lights (four) and some indirect sunlight. The photos were taken at 45 degree angle of incidence, as indicated by a drafting triangle.

The measurement technique is worthy of another write-up. In quick summary here: Photos show the reference white in both direct view and in reflection from the mirror. Histograms in each color channel give the mean pixel values, and the ratio of these values (reflected:direct) give the reflectivity in each color band. RAW images are used, and ImageJ is my software of choice for the analysis. No claim is made to the accuracy of the absolute reflectivity, only the relative reflectivity as compared to the initial state is of concern here.

Top: Mirror with Teflon.

Bottom: Side view showing drafting triangle to indicate 45^o. (Different mirror, sorry) The triangle is visible in the left image to the left of the mirror. It nearly is edge-on. Black in marks show the location of the Midas protection boundary, where the protected section is the longer section to the right.

Accelerated Life Test in Sulfide + Salt Procedure

A plastic cup was filled with some of the solution to a depth of 90mm such that the unprotected section of mirror was fully immersed as well as half of the protected section. This was done so that the upper half of the protected section was exposed only to strong H₂S gas produced by the solution.

Protected Silver Coated Slide in Sulfide + Salt Solution

The slide was immersed for the following durations, rinsed and measured at each step:

The bare silver tarnished very quickly in the strong solution, while the protected silver barely changed to the eye. Graphs of Reflectivity vs. Time illustrate the changes in each color band:

Spectral Average of 3 color bands

Analysis

Reflectivity decay in the unprotected zone goes quickly and seems to follow an exponential function. But the reflectivity in the protected zone goes much more slowly and is perhaps best assumed to be linear. There appears to be no change for the first hour, which could be real or an artifact of the camera technique. But in any case, they are following different curves.

Both zones, protected and unprotected, in each color band were approximated to the 80% point to yield an estimate of the tarnish inhibition in this highly accelerated test.

The results are shown below:

First, the exponential decay of the unprotected silver is estimated from a simple curve fit.

Fit = R_o*e^-t/k

where R_o is the starting reflectivity, t is time in minutes, and k is the time constant in minutes. Where t = k the reflectivity is 1/e times the starting value.

R₀ = 0.99, k = 82

Then, a multiplier is found for k in which the resulting reflectivity crosses the 80% mark where the protected silver is 80%. The equation is

Extrapolation = R_o*e^-t/(k*m)

m = 12

So, the protected silver has decayed 12 times slower than the unprotected silver in this experiment. The process is repeated for the green and blue channels.

R₀= 98.1, k = 48.7

m = 22

R₀= 95.2, k = 43.2

m = 30

The summary table above helps reveal that the silver decays faster in the blue than it does in the red (smaller time constant in blue.) The multipliers, m, increase from red to blue. Interestingly, the time constant for blue is about half that of red, while the multiplier for blue is about twice that of red. Therefore, it may be said that the life extension induced by the Midas is roughly the average of the multipliers, 21, which is very close to the green channel multiplier.

A 20X improvement over bare silver is very substantial! Current experiments with spray silver coatings indicate that a mirror can be expected to last anywhere from 6 months to 1 year depending primarily on how it is stored and cared for. The accelerated test results suggest that Midas treatment could extend that to 10 years or more, which is competitive with protected aluminum.

However, silver with Midas is not the same as protected aluminum. It remains a soft metal and is easily scratched. Great care must be taken to keep the mirror clean rather than let it collect dust and dirt, because cleaning the mirror will inevitably damage it. An airtight mirror cover is highly recommended. Dust removal should be non-contact, perhaps nothing more than a jet of dry air from a bulb. Distilled water rinses should be done sparingly, with non-contact drying preferred.

Inevitably, the life of the mirror will be limited by the care it gets. But when it does need a fresh coating we now have an easy and inexpensive alternative to aluminum.

Inspection of Mirrors

Evidence for mirror failure other than tarnishing also come out of the tests performed. This section is simply an album of photos of the coating at each time interval. In each photo, the unprotected silver zone is at left, the protected and immersed silver is in the middle, the white Teflon block straddles the boundary, and at right is the section of protected silver that did not get immersed. Note how the middle section becomes damaged after about an hour, rather than changing color as would be the case for tarnish.

T = 0

T = 1 minute (the bare silver almost instantly tarnishes visibly, losing about 10% reflectivity.)

T = 10 minutes

T = 20 minutes

T = 30 minutes. A faint, dark, horizontal streak has started to form in the protected zone.

T = 60 minutes. We can now see strong evidence of damage in the protected zone. Note also the dark vertical streak at the immersion boundary, off to the right.

T = 120 minutes

T = 120 minutes, taken to show damage more clearly. Right hand zone is nearly pristine, some coloration in corners of glass due to handling.

T = 240 minutes (4 hours).

T = 240 minutes (4 hours). The protected silver looks badly damaged but not exactly tarnished. The right section looks very good.

The right-hand section that did not get immersed is interesting, as it is not damaged or tarnished. Being in very close proximity to the Na₂S solution it was exposed to strong H₂S vapors which will tarnish bare silver very quickly. So here we have evidence of the substantial protection that the Midas treatment gives in a more realistic environment (albeit still harsh.)

Conclusions

A highly accelerated life test was performed on a front-surface silver mirror to evaluate changes in reflectivity over time. The test results indicate a substantial improvement over bare silver when treated with Midas Tarnish Shield for 5 minutes. The amount of improvement seems to be on the order of 21X, by extrapolation of the exponential decay time constants, indicating that a mirror could last anywhere from 10 to 20 years before becoming objectionably tarnished. However, silver coatings are still fragile and require careful storage, use, handling, and maintenance. Nevertheless, it is expected that a silver mirror treated with Midas Tarnish Shield will provide years of serviceable life.

The measurement of the silver reflectivity is a relative one, and is not intended to indicate absolute reflectivity. The use of a consumer digital camera (Fujifilm X-E2) enabled a fast, inexpensive, and convenient method of comparing the reflectivity as the tarnishing proceeded. A detailed report of this measurement is needed, in which test results are compared to NIST traceable methodologies and equipment, including spectrometers and lasers. Careful study and process might someday yield a reliable camera-based method of measuring absolute reflectivity.

Appendix: Interferogram of silver mirror, half treated with Midas

Although unrelated to life testing, there are many questions as to the thickness of Midas Tarnish Shield on silver. A 2” diameter optical flat was silvered, then dipped halfway into Midas for 5 minutes. Black tic marks delineate the boundary, which runs vertically in the views shown here. The treated side is on the left. The interferograms do not show convincing evidence of a coating layer.

There are some sub-fringe hints however, that are only visible by maximum processing of the data (left side of report window, below. Two thin vertical shadows are barely visible, one of them coincidentally perfectly aligned with the tic marks. If only the latter feature were present there would be an argument that this is caused by the phase shift at the border of the treated area, however the presence of a second nearly identical feature means it is far more likely to be something else. Perhaps it’s a processing artifact or a pair of fine scratches. In any case, the feature is so small as to be totally insignificant. Midas causes no ill effect on the surface figure of a mirror.

Zygo report window. Left: 3D Processed wavefront, all terms removed. Upper right: 2D processed wavefront. Lower right: Unprocessed fringe image.

3D Wavefront, processed. Note faint vertical lines. Whatever they are, they are in the noise.

2D wavefront, processed. No evidence of boundary.

Fringe image. Mirror was marked M for Midas, B for Bare, tic marks at the boundary. The saddle shape of the interferogram indicates that the mirror was stressed in its mount, very likely caused by hard-epoxying it in an aluminum cell. The mirror and cell were removed from the product and set aside for non-critical lab use, such as redirecting laser beams. And so, the mirror now enjoys a new life as a silvering substrate.