Recently, a report has surfaced that Apple (AAPL) is prototyping phones with sapphire screens, likely made by GT Advanced Technologies (GTAT). No doubt, Apple is hoping to make its products more competitive in the premium space. Sapphire, however, may not be the right tool for this job.
The most talked-about attribute of sapphire is scratch-resistance. On a Mohs scale, which measures scratch hardness, sapphire scores 9 - the second-hardest substance after diamond (Mohs 10). While Gorilla Glass, made by Corning Inc. (GLW), is not as hard as sapphire (Corning is not providing a Mohs number, only Vickers hardness: 622-701 for Gorilla, 2300 for sapphire), scratch-resistance is not the only measure of the overall durability of cover glass material. Similarly, durability alone is not the only measure of the overall suitability of a material as a smartphone cover glass. Let us have a look at some other characteristics of both Gorilla and sapphire.
Tensile strength. In this video, sapphire shatters under much lower force than Gorilla Glass. This type of stress test could be a good approximation of real-world events such as accidentally sitting or stepping on a phone. Gorilla Glass may have an advantage in these situations.
Flexibility and impact-resistance are additional areas where Gorilla Glass excels. Impact-resistance is especially important when dropping a phone on a hard surface. I have not seen a Gorilla vs. sapphire impact-resistance test, but I did see a similar test performed on a small and probably thick piece of sapphire, encased in a metal frame. Still, a comparison test of screen-sized, equally thin, not encased in anything, Gorilla Glass and sapphire plates would be very helpful.
Cost is an additional consideration. There is probably a limit of how low the cost of sapphire can be driven, due to a large amount of energy required to produce it. Still, during a recent conference call, GTAT ambitiously stated as follows:
"Our most recent analysis suggests that first generation screens could be manufactured for $10 to $15 above the cost of strengthened glass. We expect this cost will drop quickly thereafter as our partners and customers achieve higher volume and improved efficiencies... In addition, we believe that, over time, our Hyperion solution, currently in the R&D phase, could enable further cost reductions that could bring sapphire solutions to cost parity with strengthened glass."
It is interesting that they used the words "sapphire solutions." I think they are referring to sapphire-laminated glass here, otherwise they would have just said "sapphire." But let's dig a little bit deeper into these ambitious claims.
A fairly recent MIT technology report disagrees with GTAT, quoting $30 per screen, with the cost perhaps falling to $20 in a couple of years. In comparison, Gorilla Glass display currently costs less than $3. I am also not sure cost parity is possible for pure sapphire cover glass: the law of conservation of energy has proven remarkably stubborn so far. Melting aluminum oxide to grow sapphire crystals is not exactly easy: it has a melting point of 3,762°F (2,072°C). In comparison, iron has a melting point of 2,800°F (1,538°C); titanium, 3,034°F (1,668°C); and even vanadium has a melting point of 3,470°F (1,910°C)! In comparison, Gorilla Glass has a softening point of 837°C.
I realize that melting point is not the same as the amount of energy required to melt something (heat of fusion), but there is a correlation. The energy required to heat sapphire to its melting point from room temperature is also fairly high, given that specific heat of sapphire is 671 J/(Kg*K) and the same value for iron, for example, is 444 J/(Kg*K). The amount of energy required to melt sapphire, once it has been heated to the melting point, is quite high at 3200 J/cm3. Iron also has heat of fusion of 2141 J/cm3; vanadium, 2460 J/cm3; titanium, 1886 J/cm3. Per-unit-mass measurements would be different, and since I found a sapphire value per cm3, I converted the values for comparison metals to the same units, but the point remains. I could not find heat of fusion for Gorilla Glass, but based on Corning's comments and Gorilla's relatively low softening point, it should be considerably less than that of sapphire.
According to Corning's Q&A, sapphire:
- Requires 99% more energy to make that Gorilla Glass. Based on the above, this is a credible claim. More energy inevitably translates into higher cost.
- Is difficult to process. This makes sense. Cutting thin, uniform sapphire sheets is likely difficult, may require cutting with diamond saw or laser and polishing with a special polishing slurry. Hyperion ion implanter can cleave off only very thin layers of sapphire, perhaps suitable for lamination of a less-expensive substrate (Gorilla Glass?).
More on cost. Recently, Matt Margolis wrote about GTAT and the cost of sapphire coming down to near par with Gorilla Glass. Among other things, he stated:
The Key Drivers to Reduce Sapphire Screen Costs
- Next generation ASF Furnace 200kg vs 115kg → 74 % increase in boule size
- Intego's Sapphire Inspection Tool → 20-30% more usable yield per boule
- Thermal Technologies Improved Annealing (Finishing) Furnace → 10-20% more efficient
However, I did not understand what he wrote further:
Let's assume each furnace cycle (crucible) run makes enough sapphire for 1 screen and it costs $13 to produce 1 screen. However, if you could you increased the size of the size of the crucible by 74% you would be able to produce 1.74 screens per cycle if you divide $13 by 1.74 screens it would result in a per screen cost of $7.47 or a savings of $5.53.
This argument does not take into account that the energy required to produce a bigger boule will also increase by about a factor of 1.74. Also, will a larger boule be more prone to cracking? Defects? Would it take more time to grow than 2 smaller boules of identical combined volume? I would check out this recent discussion from Intech for further research into these questions.
Matt further wrote:
The next item is the inspection tool which increases usable sapphire by 25%, let's assume each furnace cycle still only makes 1 sapphire screen for $13, but now because of the inspection tool you no longer lose the edges to impurities (kind of like a loaf of bread that no child wants to eat) except this sapphire is so cloudy it wouldn't look pretty on your new iPhone. The inspection tool will allow you to make 1.2 screens per cycle run for $13, if you divide the $13 by 1.25 screens it would result in a per screen cost of $10.40 or a savings of $2.60.
First of all, I cannot find the 25% figure anywhere. But let's think about what this tool is. From the GTAT press release:
The SIRIUS sapphire inspection tool was developed by Intego in collaboration with GT's engineering team utilizing the capabilities of GT's sapphire R&D facility in Salem, Massachusetts. Using advanced sapphire inspection techniques and optimization algorithms, the SIRIUS tool can quickly provide an optimized core or material layout on a sapphire slab and transfer that pattern to the CNC machine tool.
So the inspection tool finds the defects and optimizes the pattern of cutting the sapphire boule to maximize the amount of clean sapphire harvested from it. If a boule is near-perfect, the improvement gained form the inspection tool will be near zero. So even if the 25% figure is correct, the efficiency improvement will be up to 25%, not a uniform 25% across the board. Also, complex cutting patterns may increase the consumption of diamond saw blades or energy for cutting lasers, increase processing time, not to mention that the equipment itself is not free and requires upgrades and maintenance. So I would be careful about assuming a uniform 25% increase in yield.
Finally, I could not find any confirmation for a 10-20% efficiency improvement from Thermal Technology's annealing furnace. An efficiency improvement compared to what? I would also like to know the definition of efficiency in this case - less energy use? Fewer defects? Again from Intech:
There is no magic. Each method of sapphire growth has its advantages and disadvantages.
Weight. Sapphire is 67 percent heavier than Gorilla Glass per unit volume: 3.98 g/cm3 vs. 2.45 g/cm3. So, a 5-inch diagonal cover glass, measuring 3 x 4 inches and being 1 mm thick would weigh 30.8 grams if made entirely of sapphire and about 19 grams if made entirely of Gorilla Glass. Thus, a full sapphire glass would add 11.8 grams to the weight of a 5-inch screen. The current iPhone 5s weighs 112 grams, and while the weight of the future model of an iPhone is unknown, using sapphire glass could make it about 10% heavier.
Optical transmission or clarity. Per Corning's Q&A:
[Sapphire] Transmits about six percentage points less light and may introduce optical distortions.
Corning provides the following diagram (presumably thickness around 1mm, although thickness is unspecified):
Valley Design Corp. provides the following diagram for a 1mm sapphire sheet:
One can see that Gorilla Glass does indeed transmit more visible light (over 90% compared with under 90% for sapphire) through a sheet of a similar thickness.
In addition, sapphire's refractive index is about 1.7. Gorilla Glass has a refractive index of about 1.5. Higher refractive index of sapphire may be the reason why Corning believes that sapphire may introduce optical distortions.
Implications for Apple, Corning and GTAT: Sapphire is not an all-around better cover glass solution than Corning Gorilla Glass. Even though it is more resistant to scratches, it shatters under lower stress and is less flexible than Gorilla Glass. It remains more costly at this time, and is heavier. Sapphire's optical properties are not as desirable as those of Gorilla Glass for the purposes of making smartphone cover glass.
In addition, the degree of increased scratch-resistance of sapphire over Gorilla Glass may not be large enough to justify a significant price premium. I own a 2-year old non-Apple phone, which probably has the original Gorilla Glass (not as good as Gorilla 3) and have been using it without any sort of additional screen cover. I dropped the phone a few times and carried it together with change and keys in the same pocket. My screen has several tiny, barely-noticeable scratches that do not impede my use of the phone in any way. For users who are more careful with their newer, Gorilla 3 phones, scratches may hardly ever happen.
Premium watchmakers have been using sapphire crystals for years, with great success. However, luxury timepieces are meant to be enjoyed for a decade or even several decades, while most phones will likely be replaced after 2-3 years. Apple's sapphire iPhones may end up being recycled with flawless screens.
In fact, if a 2-3 year old phone picks up a few barely noticeable scratches, it may be easier for a user to part with it and get a new, shiny, twice as capable device. It could even be that Apple is shooting itself in the foot by making its new product too robust, since this may delay users' decision to upgrade the phone, lengthening the upgrade cycle.
However, all is not lost for sapphire. Using a sapphire-laminated cover glass can significantly mitigate the concerns I outlined for full-thickness sapphire cover glass, while still furnishing superior scratch resistance. In fact, Corning's Gorilla Glass may be the perfect base layer to be laminated with sapphire. If this "Gorilla in a saphire jacket" is adopted by Apple, Corning will continue to make money from Gorilla Glass. GTAT will initially sell much less sapphire than the market expects, but the volumes will ramp up eventually, when manufacturers other than Apple also adopt sapphire laminate as their premium cover glass.
Also, if the back of the new iPhone is made of glass, as it is in the iPhone 4, this role will likely belong to Gorilla Glass due to cost considerations and also as a way to reduce the weight of a phone.
Mid-market and value phones from other manufacturers will probably continue to use Gorilla Glass, since it is good enough and, for at least the near future, much less expensive. This bodes well for Corning's Gorilla Glass sales.
Conclusion: Pure sapphire cover glass has drawbacks, which can be largely mitigated by the use of sapphire-laminated glass, such as Corning's Gorilla Glass.
Apple may not derive as large a quality improvement as the market hopes by using pure sapphire cover glass for its phones. Apple may also not derive as much of a differentiation as it would like by using a more sensible sapphire laminate. It would seem Apple is between a rock and a hard place on this one, or definitely between two hard surfaces. It is difficult to recommend buying Apple, even after the recent drop.
Corning will probably continue to sell plenty of Gorilla Glass, if not to Apple then to other manufacturers, since it is good enough as a screen, especially given that smartphones get replaced every 2-3 years. At the recent price levels, Corning looks like a hold.
GTAT will likely sell some sapphire but probably not as much as investors hope, due to the use of sapphire laminates and adequacy of Gorilla Glass. After a recent run-up, it may be time to take some chips off the table with GTAT.