How Long Will Inkjet Prints Last:
Estimating Print Life Using Accelerated Test Methods

There are a number of things that affect our perception of picture quality—whether the picture is printed through photographic, thermal or inkjet: color saturation and accuracy, sharpness, contrast, depth, and durability.

Our own internal evaluations and tests have indicated a high degree of satisfaction with prints made through all three technologies from both scanned film and prints and from digital cameras. The immediate perception of image quality thus indicates parity in capture and output technologies. The question of durability—how long the pictures will last under a range of conditions—however could not be measured in these kinds of tests.

Several trade publications¹ and web sites² have recently been quoting print-life estimates for prints based solely on a single, highly accelerated light fade test. A Kodak white paper published last year (http://www.kodak.com/US/en/digital/printers/claims/printStability.jhtml) summarizes the factors involved in predicting the long-term stability of inkjet prints and the limitations of relying on one criterion for measuring print life. A key point in that paper is that darkfastness (the stability of the print to heat, humidity, and other environmental factors, such as air quality) can be just as important as lightfastness in determining the useful life of a print. As with photographic and thermal prints, any credible estimate of absolute print life must account for all the factors that can lead to unacceptable image degradation.

For example, several recent combinations of ink and paper have been flagged as being especially sensitive to ambient levels of ozone in the air. Although accelerated light fade tests had indicated a print life of about 10 years in the case of Epson, and greater than 20 years in the case of Canon, these prints lose significant cyan density in just a few days to weeks when exposed to low levels of ozone. For an ongoing discussion of this phenomenon, check out http://www.p-o-v-image.com/epson.

There are several accelerated test methods, but there are also limitations in applying them to determine the useful life of photo-quality prints, with an emphasis on prints produced by inkjet printers. Using a simple, comparative approach for benchmarking the performance of different ink-paper combinations can eliminate many of the misconceptions and misleading claims relative to print life.

Limitations of Accelerated Lightfastness Testing

The basic premise behind accelerated lightfastness testing is that by using a very high intensity light exposure—for example, 100 times the typical ambient light level—the print will fade in 1/100^th of the time it would take to fade under ambient conditions. This is known as the law of reciprocity.

Another way of looking at reciprocity is to say that equivalent cumulative exposures, expressed as the product of exposure intensity times length of exposure, should produce equivalent amounts of fade. For example, if an accelerated test were carried out at 50,000 lux (50 klux) light intensity for 100 hrs and produced a 30% level of fade, the cumulative exposure would be 5,000 klux-hr (50 klux x 100 hr). For an average ambient light level of say 100 lux, (0.1 klux) it should take 50,000 hrs (2083 days, or 5.7 years) of continuous exposure to reach the same cumulative exposure (0.1 klux x 50,000 hrs = 5,000 klux-hrs). If the law of reciprocity holds, then for a 30% fade end-point, the accelerated test would predict a print-life of 5-6 years.

In practice, though, most prints tend to fade somewhat sooner than is predicted by highly accelerated exposures. This phenomenon is called reciprocity failure, and it is probably the single biggest limitation of accelerated lightfastness testing in predicting print life for ink jet output.

In a recent study of reciprocity effects of hundreds of combinations of inkjet inks and receivers, Bugner and Suminski³ found that all combinations displayed some level of reciprocity failure. In several cases, the apparent rate of fade at the lower intensity exposure (5,400 lux, or 5.4 klux) was more than ten times faster than predicted by a much higher exposure (67,000 lux, or 67 klux). Porous receivers were found to exhibit consistently higher levels of observed reciprocity failure.

For this reason alone, basing a print-life claim on a single, highly accelerated lightfastness test is extremely risky.⁴ At Kodak, we test all of our inkjet materials under at least two light intensities: 5.4 klux and 60-80 klux. Fortunately, owing to the non-porous nature of our Premium and Ultima Picture Papers, the amount of reciprocity failure is quite low. However, the downside is that it can take a long time to reach an end-point under the lower intensity conditions. For example, it takes over one year of continuous exposure at an intensity of 5.4 klux in order to reach 30% fade for test prints made with the Kodak Personal Picture Maker on Kodak Premium Picture Paper.

Another major issue in applying high intensity fade testing in order to estimate print life relates to the assumptions about the ambient light level a print might see under typical home display conditions. A published study of light levels in typical homes found average intensities in the range of 100-200 lux.⁵ Others have proposed that an average light level of 450 lux (a level more typical of an office or commercial environment) should be used as a "standard" day for the purposes of estimating print life.² The higher the assumed level of ambient exposure, the lower the print-life estimate will be. In addition, at higher assumed ambient light levels, dark fade's effects on print life diminish. Nonetheless, the relative lightfastness of various manufacturers' inkjet prints tested under the same conditions is independent of the assumptions with respect to the ambient light level.


Darkfastness: The Dirty Little Secret

Most—if not all—published print-life estimates^1,2 conveniently ignore the fact that prints can also degrade in the absence of light. For certain types of print materials, the "dark keeping" is a greater issue than light stability with respect to overall print-life. This is the case for the ozone-induced dark fade on porous inkjet receivers. Stability can be further aggravated if the prints are stored at elevated temperatures and/or humidity. Clearly, print-life estimates made solely on lightfastness testing can be very misleading. A better approach would be to determine print-life estimates for both light fade and dark fade in home conditions, and to quote an overall print-life that accounts the effects of both.

Dark stability phenomena fall into two distinct categories: chemical reactions that lead to dye decomposition and physical changes in the dye environment that lead to dye migration. The former leads to loss of density (dye fade) and/or hue shift, while the latter results in density gain and/or loss of sharpness. To further complicate matters, both phenomena can be accelerated to differing degrees by temperature and humidity.

Dark fade is one of the principal contributors to reciprocity failure. At a lower light intensity, the length of time it takes to reach a given level of dye fade increases. The rate of dark fade, however, is independent of light level. Rather it depends on the rate of chemical reaction. Therefore, longer term, lower intensity light fade tests are more likely to uncover the greater contribution of any dark fade mechanisms to the observed level of fade. If the chemical rate is slower than light fade, the net result is that the observed level of dye fade is higher than what is predicted by shorter duration tests conducted at much higher levels of light intensity.

Accelerated Darkfastness Testing

So how do we accelerate dark fade? When dark fade reactions happen too slowly at room temperature to measure in a reasonable timeframe, the traditional approach⁶ has been to apply what is known as Arrhenius methodology.⁷ In this method, raising the temperature at constant 50% relative humidity accelerates the rate of dark fade. By comparing the rates of fade at several temperatures above room temperature, one can deduce the rate of fade at room temperature. From the rate of fade at room temperature, one can calculate the amount of time it would take to reach a given degree of fade, such as a 30% fade end-point.

In practice, there are several limitations to the Arrhenius method, especially when applied to inkjet prints. For example, there is a limit to how high above room temperature that measurements of dark fade can be made. Many inkjet receivers undergo what is know as a glass transition at temperatures not too far above room temperature. Whenever a phase change such as a glass transition occurs, measurements made at or above that temperature cannot be used to predict what will happen below that temperature. When the temperature is raised above the glass transition temperature, T_g, not only can the rate of dark fade increase dramatically, but dye migration, leading to density gain, can also occur in concert with dye fade. (We will discuss this effect in more detail in the context of humidity-induced dye migration.) Given this limitation, it is difficult to accelerate the rate of dark fade for inkjet materials over the available temperature range to achieve a meaningful level of fade in a reasonable period of time.

Another possible way to accelerate dark fade is to increase the concentration of the agent responsible for causing the dye to fade. For inkjet prints, ozone is known to be one such agent. Assuming that ozone is the primary cause of dark fade, one can accelerate the rate of dark fade by treating the print to elevated levels of ozone. Work is being conducted at Kodak and at other labs around the world aimed at developing such a test. Results will be presented at the upcoming NIP 17: 2001 International Conference on Digital Imaging Technologies.

Key to the application of such a test to the estimation of dark fade print-life is the knowledge of the ambient level of ozone that the typical print might be exposed to. To the best of our knowledge, studies of average ambient indoor levels of ozone have not been published.

Humidity-Induced Dark Stability Issues

In addition to chemical reactions that cause dye fade, physical changes in the print can lead to dye migration and/or loss of sharpness. As noted above, one such physical change occurs when the temperature is raised at constant humidity until it exceeds the T_g of the ink-receptive coating. Another way to induce the same effect is to increase the humidity at a given temperature. As the humidity in the air increases, the amount of water absorbed by the ink-receptive layer increases. Absorbed water can also lower the T_g of the ink-receptive layer. At high enough humidity, the T_g can actually fall below room temperature. This results in dye migration just as in the case above in which temperature was raised above T_g.

For most inkjet prints, at room temperatures in the range of 20-25oC (68-77oF), the onset of observed dye migration occurs in the range of 60-80% relative humidity. Unlike the light and dark fade phenomena, which tend to be of a cumulative nature, humidity effects tend to be more of a rate phenomenon that depends on the physical state of the media. That is, at a relative humidity near 50% at room temperature, the rate of dye migration is negligibly low. Over a relatively narrow range of humidity, as it increases to the point that the T_g of the ink-receptive layer is depressed below room temperature, the rate of dye migration becomes significantly faster.

Another problem with trying to estimate print-life with respect to humidity-induced dye migration is that there is no prior art documenting what constitutes an unacceptable end-of-life situation with respect to either density gain, hue shift, and/or loss of sharpness. What we can say at this point is that if humidity is kept below a certain level, the long term effect is minimized to the point that dark and light fade mechanisms will most likely dominate the overall image stability and will be the primary determinants of print life.

The Case for Comparative Testing

Given the foregoing discussion of the myriad of issues surrounding absolute print-life estimates, it seems that a focus on comparative testing might be a reasonable alternative. If we accept the proposal to isolate the effects of dye migration by keeping temperature and humidity under control, then a combination of comparative light fade and dark fade testing can be used to determine which ink-receiver combinations should last the longest.

Comparative light fade testing should be carried out at a light level that produces meaningful results in a reasonable time. But it cannot be so highly accelerated that reciprocity effects distort the results. The UV content of the light should be filtered to match as close as possible the intended display environment. Plexiglas-filtered 5.4 klux fluorescent lighting comes closest to meeting these requirements for a typical home or office display condition.⁵

Figure 1 compares the cumulative exposure required under these conditions to elicit a 30% fade in various ink-receiver combinations. The combination of Kodak Premium Picture Paper and the Kodak Personal Picture Maker clearly out performs the other manufacturers' (OMs') matched systems by a wide margin, outperforming the nearest system by a factor of almost 5:1. Kodak Premium Picture Paper printed with the OMs' inks performs comparably to, or, in the case of Epson 870 inks significantly better than, the matched OM systems. Tests for Kodak Ultima Picture Paper are still ongoing, as are tests involving the latest versions of the OMs' inks and receivers. Note that a cumulative exposure of 50,000 klux-hr takes over one year of continuous treatment with a light intensity of 5.4 klux.

As noted, ozone has been identified as a primary contributor to the dark fade of inkjet prints. As with comparative light fade testing, one should choose an elevated level of ozone that will produce sufficient acceleration of the effect, without being so high that side effects become an issue. Ambient levels of ozone indoors ranging from 0 to 100 parts per billion (ppb) are not uncommon, especially in the summer.⁸ One proposal for comparative ozone testing is to use an ozone concentration in the range of 1000 to 5000 ppb, and determine the cumulative exposure, in units of ppb-hr, required to reach a 30% fade point. Details of this type of study will be presented at the upcoming NIP17 conference in October 2001.

Figure 2. Cumulative exposure to reach 30% fade for various ink-receiver combinations using Plexiglas-filtered 5.4 klux fluorescent lighting.


Summary

Print-life estimates based on single, highly accelerated light fade tests are misleading. Any credible estimate of print life must account for all of relavant factors, including those that continue to occur regardless of the room brightness. These factors include temperature, humidity, and ozone-induced fade. In addition, one must also account for reciprocity effects that can occur when very high light levels are used to accelerate dye fade. Comparative standardized testing at multiple light levels provides for a more complete assessment of the lightfastness of different combinations of ink and paper. Standardized test methods for stability in a range of humidity and ozone conditions are also needed so that comparative testing for these factors can be included in any overall measure of image stability.

¹ A. Yegyazarian, "Fight Photo Fade-out," PC World, pp. 48-51, July, 2001.
²(http://www.wilhelm-research.com: "Also to be posted is a discussion of changes being made in the image permanence test methods at Wilhelm Imaging Research, Inc. to better take into account the potentially large reciprocity failures that may occur in high-intensity accelerated light fading tests, humidity-fastness problems, and susceptibility of inkjet prints to the effects of ozone and other atmospheric contaminants. The influence of different types of inkjet media on resistance to gas fading, susceptibility to high-intensity reciprocity failures, and humidity-fastness problems are described and recent industry developments in this area are discussed. All of this is aimed at better understanding and simulating the behavior of prints during long-term display and storage in the wide range of conditions that may be encountered in homes and offices throughout the world." -posting noted on July 25, 2001.
³ D. E. Bugner and C. Suminski, Proceedings of IS&T's NIP16: 2000 International Conference on Digital Imaging Technologies, pp. 90-94 (2000).
⁴ "More Flak on Epson's Ozone Problem," The Hardcopy Observer, 6(10), 32, (2000).
⁵ S. I. Anderson and R. J. Anderson, J. Imaging Tech., 17, 127 (1991).
⁶ ANSI Standard IT9.9 (1996).
⁷ S. A. Arrhenius, Zeit. fur Phys. Chem., 4, 226 (1889).
⁸ C. Ireland, "Report Finds High Ozone Levels Here," Rochester, N.Y. Democrat and Chronicle, May 19, 2001.