WEBVTT 00:00:00.000 --> 00:00:01.973 align:middle line:90% [SQUEAKING] 00:00:01.973 --> 00:00:03.446 align:middle line:90% [RUSTLING] 00:00:03.446 --> 00:00:06.883 align:middle line:90% [CLICKING] 00:00:06.883 --> 00:00:16.440 align:middle line:90% 00:00:16.440 --> 00:00:18.840 align:middle line:84% JUEJUN HU: So let's continue our discussion 00:00:18.840 --> 00:00:20.320 align:middle line:90% on glass strengthening. 00:00:20.320 --> 00:00:21.920 align:middle line:84% So in the last lecture, we highlighted 00:00:21.920 --> 00:00:23.837 align:middle line:84% there are three different techniques, or three 00:00:23.837 --> 00:00:26.160 align:middle line:84% different groups of techniques, for strengthening glass 00:00:26.160 --> 00:00:26.860 align:middle line:90% materials. 00:00:26.860 --> 00:00:28.800 align:middle line:84% You can either engineer the chemistry 00:00:28.800 --> 00:00:31.840 align:middle line:84% to enhance the intrinsic material parameter K1c. 00:00:31.840 --> 00:00:35.360 align:middle line:84% Or you can try to essentially decrease or eliminate 00:00:35.360 --> 00:00:38.480 align:middle line:90% the defects, surface cracks. 00:00:38.480 --> 00:00:41.320 align:middle line:84% And the third one is that also, even if there are surface 00:00:41.320 --> 00:00:43.620 align:middle line:84% defects, if you apply surface compressive stress, 00:00:43.620 --> 00:00:45.320 align:middle line:84% you can prevent the opening of cracks, 00:00:45.320 --> 00:00:47.040 align:middle line:84% and therefore essentially apparently make 00:00:47.040 --> 00:00:48.420 align:middle line:90% the glass material more strong. 00:00:48.420 --> 00:00:50.658 align:middle line:84% So let me elaborate a little bit and talk about all 00:00:50.658 --> 00:00:51.700 align:middle line:90% the different techniques. 00:00:51.700 --> 00:00:53.560 align:middle line:90% Let's start from the first one. 00:00:53.560 --> 00:00:56.360 align:middle line:84% How do you actually engineer the intrinsic glass chemistry 00:00:56.360 --> 00:00:58.660 align:middle line:90% to boost its K1c? 00:00:58.660 --> 00:01:00.903 align:middle line:84% So in the previous lecture, we have already 00:01:00.903 --> 00:01:02.320 align:middle line:84% given you an analytical expression 00:01:02.320 --> 00:01:04.000 align:middle line:90% of what K1c looks like. 00:01:04.000 --> 00:01:06.300 align:middle line:84% So effectively, K1c is a material constant. 00:01:06.300 --> 00:01:09.280 align:middle line:84% And it depends on two parameters in a brittle material, which 00:01:09.280 --> 00:01:15.120 align:middle line:84% is gamma the surface energy, and E is the elastic modulus. 00:01:15.120 --> 00:01:20.320 align:middle line:84% Now, gamma is directly dictated by the bond energy, 00:01:20.320 --> 00:01:22.880 align:middle line:84% because when you are trying to form new surfaces, 00:01:22.880 --> 00:01:28.120 align:middle line:84% you are essentially spending energy to break chemical bonds. 00:01:28.120 --> 00:01:30.800 align:middle line:84% And the excess energy you accumulate 00:01:30.800 --> 00:01:32.820 align:middle line:84% is basically because of the chemical bonds. 00:01:32.820 --> 00:01:36.120 align:middle line:84% They are now essentially dangling on the surface. 00:01:36.120 --> 00:01:38.680 align:middle line:84% The electrons are no longer paired with one another 00:01:38.680 --> 00:01:40.100 align:middle line:90% to lower the overall energy. 00:01:40.100 --> 00:01:43.880 align:middle line:84% So the basic material theory says 00:01:43.880 --> 00:01:47.120 align:middle line:84% that gamma is directly proportional 00:01:47.120 --> 00:01:50.120 align:middle line:90% to the chemical bond energy. 00:01:50.120 --> 00:01:53.400 align:middle line:90% 00:01:53.400 --> 00:01:56.600 align:middle line:84% So therefore, in order to increase K1c, 00:01:56.600 --> 00:02:01.590 align:middle line:84% an immediate way is to basically increase the chemical bond 00:02:01.590 --> 00:02:02.090 align:middle line:90% energy. 00:02:02.090 --> 00:02:03.930 align:middle line:84% You have stronger chemical bonds, 00:02:03.930 --> 00:02:05.790 align:middle line:90% you tend to increase gamma. 00:02:05.790 --> 00:02:08.789 align:middle line:84% It turns out stronger chemical bonds also 00:02:08.789 --> 00:02:11.030 align:middle line:84% tend to increase elastic modulus as well, which 00:02:11.030 --> 00:02:11.797 align:middle line:90% is understandable. 00:02:11.797 --> 00:02:13.630 align:middle line:84% But the correlation is actually a lot weaker 00:02:13.630 --> 00:02:15.005 align:middle line:84% because elastic modulus has to do 00:02:15.005 --> 00:02:17.190 align:middle line:84% with specific shape of essentially 00:02:17.190 --> 00:02:19.590 align:middle line:84% the interaction potential between the different atoms. 00:02:19.590 --> 00:02:21.390 align:middle line:84% But in general, rule of thumb, if you 00:02:21.390 --> 00:02:24.270 align:middle line:84% want to have material that's actually mechanically stronger, 00:02:24.270 --> 00:02:25.670 align:middle line:84% can resist crack propagation, you 00:02:25.670 --> 00:02:27.670 align:middle line:84% want something with strong chemical bonds, which also 00:02:27.670 --> 00:02:29.110 align:middle line:90% makes a lot of intuitive sense. 00:02:29.110 --> 00:02:31.910 align:middle line:84% Now, even for the same chemical bonds, 00:02:31.910 --> 00:02:35.350 align:middle line:84% I also want to mention that the network topology also 00:02:35.350 --> 00:02:38.110 align:middle line:84% play a role in dictating the mechanical properties of glass 00:02:38.110 --> 00:02:39.070 align:middle line:90% materials. 00:02:39.070 --> 00:02:41.007 align:middle line:84% So this is most clearly seen by this plot 00:02:41.007 --> 00:02:42.590 align:middle line:84% here, where we are essentially looking 00:02:42.590 --> 00:02:46.030 align:middle line:84% at the variation of the elastic modulus 00:02:46.030 --> 00:02:48.830 align:middle line:84% as a function of arsenic percentage in arsenic 00:02:48.830 --> 00:02:50.698 align:middle line:84% selenium [INAUDIBLE] [? glass. ?] 00:02:50.698 --> 00:02:51.990 align:middle line:90% So there are three curves here. 00:02:51.990 --> 00:02:54.630 align:middle line:90% E is a Young's modulus. 00:02:54.630 --> 00:02:59.270 align:middle line:84% K is essentially the volumetric modulus. 00:02:59.270 --> 00:03:01.750 align:middle line:90% G is the shear modulus. 00:03:01.750 --> 00:03:05.270 align:middle line:84% So it's quite remarkable that if you look at all three curves, 00:03:05.270 --> 00:03:09.670 align:middle line:84% they all peak where arsenic corresponds to 40% 00:03:09.670 --> 00:03:13.810 align:middle line:84% So this is actually the As2Se3 composition. 00:03:13.810 --> 00:03:16.670 align:middle line:84% Now, if you recall, this also corresponds 00:03:16.670 --> 00:03:19.670 align:middle line:84% to where you have a mean coordination number of 2.4, 00:03:19.670 --> 00:03:22.990 align:middle line:90% the magic number. 00:03:22.990 --> 00:03:24.830 align:middle line:84% This is actually quite a common phenomenon 00:03:24.830 --> 00:03:27.410 align:middle line:84% if you look at some other groups of glass materials. 00:03:27.410 --> 00:03:30.550 align:middle line:84% What I'm showing you here, if you focus on these black dots 00:03:30.550 --> 00:03:32.590 align:middle line:84% here, these are actually the binary germanium 00:03:32.590 --> 00:03:34.390 align:middle line:90% selenide glasses. 00:03:34.390 --> 00:03:37.150 align:middle line:84% And it's plotted in these two plots here. 00:03:37.150 --> 00:03:39.190 align:middle line:84% This one is basically the toughness measured 00:03:39.190 --> 00:03:40.190 align:middle line:90% by indentation method-- 00:03:40.190 --> 00:03:41.950 align:middle line:84% I talked about it in the last lecture 00:03:41.950 --> 00:03:44.048 align:middle line:84% as a function of the mean coordination number. 00:03:44.048 --> 00:03:46.590 align:middle line:84% On this one on the right side, this is basically brittleness, 00:03:46.590 --> 00:03:48.630 align:middle line:84% similarly inferred from limitation measurements 00:03:48.630 --> 00:03:51.390 align:middle line:84% as a function of mean coordination [? number. ?] 00:03:51.390 --> 00:03:54.110 align:middle line:84% What we are seeing here is that a critically connected glass 00:03:54.110 --> 00:03:56.390 align:middle line:84% network also corresponds to the point 00:03:56.390 --> 00:04:00.790 align:middle line:84% where you have the maximum toughness and the lowest 00:04:00.790 --> 00:04:03.190 align:middle line:90% brittleness. 00:04:03.190 --> 00:04:05.870 align:middle line:90% The two are correlated together. 00:04:05.870 --> 00:04:06.930 align:middle line:90% It's less brittle. 00:04:06.930 --> 00:04:09.770 align:middle line:84% You can actually accommodate more plastic deformation. 00:04:09.770 --> 00:04:11.200 align:middle line:90% You absorb more energy. 00:04:11.200 --> 00:04:12.450 align:middle line:90% Therefore, it becomes tougher. 00:04:12.450 --> 00:04:17.670 align:middle line:84% So the takeaway message is that the network topology also 00:04:17.670 --> 00:04:21.990 align:middle line:84% plays a significant role in dictating the intrinsic glass 00:04:21.990 --> 00:04:23.690 align:middle line:84% mechanical properties there as well. 00:04:23.690 --> 00:04:28.710 align:middle line:90% 00:04:28.710 --> 00:04:32.670 align:middle line:84% All right, so that's about the glass chemistry. 00:04:32.670 --> 00:04:35.590 align:middle line:84% Now we'll spend a bit more time discussing 00:04:35.590 --> 00:04:40.110 align:middle line:84% about how to introduce a compressive stress on glass 00:04:40.110 --> 00:04:43.030 align:middle line:84% surfaces to prevent crack propagation. 00:04:43.030 --> 00:04:45.310 align:middle line:84% And there are, in general, three different ways 00:04:45.310 --> 00:04:46.550 align:middle line:90% that people do it. 00:04:46.550 --> 00:04:50.110 align:middle line:84% You have thermal toughening, chemical toughening 00:04:50.110 --> 00:04:52.810 align:middle line:84% or ion exchange, as well as so-called intrinsic toughening. 00:04:52.810 --> 00:04:55.150 align:middle line:84% OK, so let me go through them one by one. 00:04:55.150 --> 00:04:57.687 align:middle line:84% So thermal toughening is probably the easier. 00:04:57.687 --> 00:04:59.270 align:middle line:84% It's basically how you actually create 00:04:59.270 --> 00:05:00.850 align:middle line:90% any kind of tempered glasses. 00:05:00.850 --> 00:05:01.930 align:middle line:90% You have a glass pane. 00:05:01.930 --> 00:05:03.870 align:middle line:84% You heat it up to a high temperature 00:05:03.870 --> 00:05:07.750 align:middle line:84% above the glass transition temperature. 00:05:07.750 --> 00:05:09.790 align:middle line:84% So then you quickly cool it down. 00:05:09.790 --> 00:05:12.630 align:middle line:84% And because of the difference in terms of cooling rate 00:05:12.630 --> 00:05:14.690 align:middle line:84% from the outside and the interior, 00:05:14.690 --> 00:05:16.710 align:middle line:84% you eventually build up a gradient 00:05:16.710 --> 00:05:19.030 align:middle line:84% of stress inside the material here. 00:05:19.030 --> 00:05:20.850 align:middle line:84% The surface will be compressive stress. 00:05:20.850 --> 00:05:23.430 align:middle line:84% The interior will be tensile stress 00:05:23.430 --> 00:05:27.670 align:middle line:84% The typical stress profile will be what I'm illustrating here. 00:05:27.670 --> 00:05:30.630 align:middle line:84% As I'll show you in the next few slides, 00:05:30.630 --> 00:05:32.830 align:middle line:84% the stress profile, theoretically, it 00:05:32.830 --> 00:05:35.590 align:middle line:84% will follow like a parabolic shape 00:05:35.590 --> 00:05:37.170 align:middle line:84% So because of the parabolic shape, 00:05:37.170 --> 00:05:39.030 align:middle line:84% you can actually easily calculate 00:05:39.030 --> 00:05:41.750 align:middle line:84% the surface layer that's under compressive strain 00:05:41.750 --> 00:05:45.730 align:middle line:84% will be about 20% of the overall thickness of the glass pane. 00:05:45.730 --> 00:05:48.630 align:middle line:84% So on each side, you have basically 20% of region that's 00:05:48.630 --> 00:05:49.880 align:middle line:90% under compressive strength. 00:05:49.880 --> 00:05:52.540 align:middle line:84% And inside, there's about 60% of the thickness 00:05:52.540 --> 00:05:53.920 align:middle line:90% that's under tensile strength. 00:05:53.920 --> 00:05:58.540 align:middle line:90% 00:05:58.540 --> 00:06:02.180 align:middle line:84% One basic rule coming from mechanics saying that because 00:06:02.180 --> 00:06:06.100 align:middle line:84% of equilibrium, the requirement for doing equilibrium, 00:06:06.100 --> 00:06:09.060 align:middle line:84% if you look at a thin pane of glass that's uniform 00:06:09.060 --> 00:06:12.140 align:middle line:84% in composition, the area under the curve 00:06:12.140 --> 00:06:16.040 align:middle line:84% must be the same for compressive and tensile stress. 00:06:16.040 --> 00:06:19.100 align:middle line:84% So effectively, we're saying that these two triangular 00:06:19.100 --> 00:06:22.380 align:middle line:84% regions here, the area should add up 00:06:22.380 --> 00:06:25.380 align:middle line:84% to the same area as the area under this curve 00:06:25.380 --> 00:06:26.540 align:middle line:90% here in the middle. 00:06:26.540 --> 00:06:29.180 align:middle line:84% That's just because of simple mechanical equilibrium 00:06:29.180 --> 00:06:31.800 align:middle line:90% requirements. 00:06:31.800 --> 00:06:34.740 align:middle line:84% Now, what that also means is that, 00:06:34.740 --> 00:06:37.260 align:middle line:84% because there's a very significant thickness 00:06:37.260 --> 00:06:39.260 align:middle line:84% of the material under large compressive strain 00:06:39.260 --> 00:06:40.660 align:middle line:90% on the surface, 20%. 00:06:40.660 --> 00:06:43.580 align:middle line:84% The magnitude of the tensile strain inside the glass 00:06:43.580 --> 00:06:48.220 align:middle line:84% is also very significant, just to counterbalance the surface 00:06:48.220 --> 00:06:49.240 align:middle line:90% compressive strain. 00:06:49.240 --> 00:06:53.020 align:middle line:84% So this is actually one of the main problems with thermally 00:06:53.020 --> 00:06:57.032 align:middle line:84% tempered glasses Because glass materials, 00:06:57.032 --> 00:06:58.740 align:middle line:84% if you have any kind of intrinsic defects 00:06:58.740 --> 00:07:01.580 align:middle line:84% under compressive strain, great, it's not going to open. 00:07:01.580 --> 00:07:05.060 align:middle line:84% Under tensile strain, however, any kind of defects 00:07:05.060 --> 00:07:08.480 align:middle line:84% can easily propagate because it's getting opened. 00:07:08.480 --> 00:07:10.860 align:middle line:84% The crack tip is getting opened by this intrinsic tensile 00:07:10.860 --> 00:07:11.860 align:middle line:90% strain. 00:07:11.860 --> 00:07:15.300 align:middle line:84% So the tensile strain region essentially 00:07:15.300 --> 00:07:19.182 align:middle line:84% becomes Achilles heel for thermally tempered glasses 00:07:19.182 --> 00:07:21.140 align:middle line:84% You already see the example when you're looking 00:07:21.140 --> 00:07:22.660 align:middle line:90% at the Prince Rupert's drop. 00:07:22.660 --> 00:07:25.180 align:middle line:84% In Prince Rupert's drop the tail essentially 00:07:25.180 --> 00:07:29.360 align:middle line:84% allows you to access the tensile strain region inside. 00:07:29.360 --> 00:07:31.000 align:middle line:84% So if you accidentally break the tail, 00:07:31.000 --> 00:07:33.780 align:middle line:84% it quickly propagates and the whole thing breaks apart. 00:07:33.780 --> 00:07:37.140 align:middle line:84% So as we'll see later, a thermally tempered glass 00:07:37.140 --> 00:07:39.942 align:middle line:84% has the issue of spontaneous breakage, 00:07:39.942 --> 00:07:41.400 align:middle line:84% even though the probability is low. 00:07:41.400 --> 00:07:44.020 align:middle line:84% But they do sometimes spontaneously break basically 00:07:44.020 --> 00:07:47.220 align:middle line:84% because of defects in the tensile strength region. 00:07:47.220 --> 00:07:50.980 align:middle line:84% So because of that, you obviously 00:07:50.980 --> 00:07:54.380 align:middle line:84% don't want a cell phone to explode suddenly in your pocket. 00:07:54.380 --> 00:07:56.860 align:middle line:84% So that's why people come up with an alternative technique 00:07:56.860 --> 00:07:59.740 align:middle line:84% for applications like smartphone screens. 00:07:59.740 --> 00:08:02.700 align:middle line:84% The technique they use is called ion exchange or chemical 00:08:02.700 --> 00:08:03.380 align:middle line:90% toughening. 00:08:03.380 --> 00:08:05.980 align:middle line:84% And in this case, what you do is immerse your glass material 00:08:05.980 --> 00:08:08.120 align:middle line:84% into essentially a molten salt bath. 00:08:08.120 --> 00:08:10.280 align:middle line:84% If it's a glass containing sodium, 00:08:10.280 --> 00:08:13.700 align:middle line:84% you can put it in, for example, a potassium salt bath. 00:08:13.700 --> 00:08:16.060 align:middle line:84% And then the potassium ion diffuses in the glass. 00:08:16.060 --> 00:08:17.240 align:middle line:90% Sodium ion also diffuses. 00:08:17.240 --> 00:08:18.615 align:middle line:84% So eventually, potassium is going 00:08:18.615 --> 00:08:22.720 align:middle line:84% to occupy the original site occupied by the sodium ions. 00:08:22.720 --> 00:08:23.720 align:middle line:90% And potassium is bigger. 00:08:23.720 --> 00:08:26.178 align:middle line:84% So then that can essentially create this compressive strain 00:08:26.178 --> 00:08:26.880 align:middle line:90% on the surface. 00:08:26.880 --> 00:08:27.780 align:middle line:90% Simple. 00:08:27.780 --> 00:08:30.860 align:middle line:84% Now, what I want to highlight here is this strain profile-- 00:08:30.860 --> 00:08:32.059 align:middle line:90% stress profile. 00:08:32.059 --> 00:08:37.059 align:middle line:84% You have very high compressive stress on the surface. 00:08:37.059 --> 00:08:38.980 align:middle line:84% But the thickness of this layer, this 00:08:38.980 --> 00:08:41.460 align:middle line:84% is called K steps, the surface layer where 00:08:41.460 --> 00:08:44.169 align:middle line:90% you have compressive stress. 00:08:44.169 --> 00:08:48.170 align:middle line:84% K step typically is pretty small for chemically toughened glass. 00:08:48.170 --> 00:08:51.050 align:middle line:84% It's of the order of a few tens of microns. 00:08:51.050 --> 00:08:53.808 align:middle line:84% So what that means is, if you think 00:08:53.808 --> 00:08:55.850 align:middle line:84% about it from the same perspective of essentially 00:08:55.850 --> 00:08:58.090 align:middle line:84% mechanical equilibrium, Again, the area 00:08:58.090 --> 00:09:00.370 align:middle line:84% under either tensile or compressive strain 00:09:00.370 --> 00:09:01.210 align:middle line:90% has to be equal. 00:09:01.210 --> 00:09:04.490 align:middle line:84% But because the surface compressive layer is so thin, 00:09:04.490 --> 00:09:08.290 align:middle line:84% the magnitude of tensile strain in the interior of the glass 00:09:08.290 --> 00:09:12.210 align:middle line:90% pane is much smaller. 00:09:12.210 --> 00:09:13.930 align:middle line:84% So because of that, you pretty much 00:09:13.930 --> 00:09:17.850 align:middle line:84% eliminate any risks of spontaneous breakage because 00:09:17.850 --> 00:09:20.150 align:middle line:84% of the large tensile strain in the glass material. 00:09:20.150 --> 00:09:23.250 align:middle line:84% So that's why your cell phone is not going to explode suddenly 00:09:23.250 --> 00:09:24.110 align:middle line:90% in your pocket. 00:09:24.110 --> 00:09:27.930 align:middle line:90% 00:09:27.930 --> 00:09:31.290 align:middle line:84% All right, now last one-- intrinsic toughening, 00:09:31.290 --> 00:09:33.548 align:middle line:84% this usually applies to essentially selective 00:09:33.548 --> 00:09:35.090 align:middle line:84% modification of the chemical property 00:09:35.090 --> 00:09:36.570 align:middle line:90% on the surface of glass. 00:09:36.570 --> 00:09:38.650 align:middle line:84% This can be done by crystallization. 00:09:38.650 --> 00:09:41.190 align:middle line:84% And as I mentioned in the last class, 00:09:41.190 --> 00:09:43.650 align:middle line:84% glass ceramics is one way to actually strengthen 00:09:43.650 --> 00:09:44.770 align:middle line:90% glass materials. 00:09:44.770 --> 00:09:47.970 align:middle line:84% In fact, if you look at some of the high-end smartphone 00:09:47.970 --> 00:09:52.140 align:middle line:84% products, Apple has, what's called Ceramic Shield-- 00:09:52.140 --> 00:09:53.890 align:middle line:84% I don't know if you heard that brand name. 00:09:53.890 --> 00:09:56.670 align:middle line:84% In some of their more recent generation of phones, 00:09:56.670 --> 00:09:58.190 align:middle line:90% it has Ceramic Shield. 00:09:58.190 --> 00:09:59.530 align:middle line:90% It's basically a glass. 00:09:59.530 --> 00:10:01.350 align:middle line:84% The screen is made of pure glass. 00:10:01.350 --> 00:10:02.810 align:middle line:90% It's glass ceramics. 00:10:02.810 --> 00:10:04.170 align:middle line:90% Samsung is following suit. 00:10:04.170 --> 00:10:06.370 align:middle line:84% I think Corning has a new product called 00:10:06.370 --> 00:10:14.010 align:middle line:84% gorilla Armor, which is basically also glass ceramics. 00:10:14.010 --> 00:10:16.437 align:middle line:84% So you can essentially create a gradient 00:10:16.437 --> 00:10:17.770 align:middle line:90% in terms of chemical properties. 00:10:17.770 --> 00:10:20.130 align:middle line:84% You can, by essentially diffusion, 00:10:20.130 --> 00:10:22.043 align:middle line:84% create a chemical gradient, and use 00:10:22.043 --> 00:10:24.210 align:middle line:84% that to actually help to create a compressive stress 00:10:24.210 --> 00:10:24.910 align:middle line:90% on the surface. 00:10:24.910 --> 00:10:28.010 align:middle line:84% So we're mostly focusing on the first two categories 00:10:28.010 --> 00:10:30.730 align:middle line:84% because they are the most commonly used examples 00:10:30.730 --> 00:10:32.270 align:middle line:84% in strengthening glass materials. 00:10:32.270 --> 00:10:36.850 align:middle line:84% So I'll start from the first one. 00:10:36.850 --> 00:10:40.290 align:middle line:84% So thermal tempering is probably the most widely used 00:10:40.290 --> 00:10:45.916 align:middle line:84% for everyday products like architectural glasses 00:10:45.916 --> 00:10:49.410 align:middle line:90% or everyday home appliances. 00:10:49.410 --> 00:10:51.478 align:middle line:84% And the reason is that it's quite simple to make 00:10:51.478 --> 00:10:52.770 align:middle line:90% compared to chemical tempering. 00:10:52.770 --> 00:10:54.110 align:middle line:90% It's actually much lower costs. 00:10:54.110 --> 00:10:58.230 align:middle line:84% So the way it works, let's start by looking at this plot here. 00:10:58.230 --> 00:11:00.810 align:middle line:84% This is basically what I'm plotting here, two things-- 00:11:00.810 --> 00:11:04.050 align:middle line:84% temperature profile and the stress profile 00:11:04.050 --> 00:11:06.810 align:middle line:90% inside a flat pane of glass. 00:11:06.810 --> 00:11:09.570 align:middle line:84% You first start by heating up to somewhere around 600, 00:11:09.570 --> 00:11:12.970 align:middle line:84% 650 degrees Celsius for soda-lime glass. 00:11:12.970 --> 00:11:17.050 align:middle line:84% Now, for soda-lime glass, let me remind you that it's Tg. 00:11:17.050 --> 00:11:19.650 align:middle line:90% It's about 510 degrees Celsius. 00:11:19.650 --> 00:11:22.490 align:middle line:90% So what is Tg? 00:11:22.490 --> 00:11:26.290 align:middle line:84% If you recall, it is also called the straining point. 00:11:26.290 --> 00:11:29.490 align:middle line:84% It means that if you heat up the glass material above the Tg, 00:11:29.490 --> 00:11:32.330 align:middle line:84% any internal stress is going to relax. 00:11:32.330 --> 00:11:35.390 align:middle line:84% Below Tg, the stress will take too long to relax. 00:11:35.390 --> 00:11:37.640 align:middle line:84% So you can actually build up internal stress. 00:11:37.640 --> 00:11:40.832 align:middle line:84% So therefore, if you heat it up above Tg, 00:11:40.832 --> 00:11:42.540 align:middle line:84% you have a uniform temperature initially. 00:11:42.540 --> 00:11:44.440 align:middle line:84% And then the stress should also be 0 00:11:44.440 --> 00:11:46.760 align:middle line:84% because all the stress gets relaxed. 00:11:46.760 --> 00:11:48.545 align:middle line:90% Then you start to cool down. 00:11:48.545 --> 00:11:49.920 align:middle line:84% Now, when you start to cool down, 00:11:49.920 --> 00:11:51.462 align:middle line:84% you build up the temperature gradient 00:11:51.462 --> 00:11:53.080 align:middle line:90% like what I'm showing here. 00:11:53.080 --> 00:11:55.120 align:middle line:84% Initially, the temperature is still high. 00:11:55.120 --> 00:11:57.640 align:middle line:84% So there should still be no stress at all. 00:11:57.640 --> 00:12:00.640 align:middle line:84% Now, the profile of this temperature here, 00:12:00.640 --> 00:12:03.240 align:middle line:84% we argue that this should be an ideal case scenario where 00:12:03.240 --> 00:12:04.860 align:middle line:90% the glass is uniform. 00:12:04.860 --> 00:12:07.160 align:middle line:84% And this probably doesn't depend on temperature. 00:12:07.160 --> 00:12:10.340 align:middle line:84% This curve should have exactly a parabolic shape. 00:12:10.340 --> 00:12:11.980 align:middle line:90% You can prove it mathematically. 00:12:11.980 --> 00:12:14.920 align:middle line:84% But in fact, it's quite easy to explain, 00:12:14.920 --> 00:12:17.880 align:middle line:84% just intuitively here, why this is a parabolic shape. 00:12:17.880 --> 00:12:21.480 align:middle line:84% If you think about a pane of glass, 00:12:21.480 --> 00:12:26.260 align:middle line:84% and this is the temperature curve, by symmetry, 00:12:26.260 --> 00:12:28.420 align:middle line:84% you can think of the heat coming from this side, 00:12:28.420 --> 00:12:29.580 align:middle line:90% this half of the glass. 00:12:29.580 --> 00:12:31.800 align:middle line:84% They should escape from this surface. 00:12:31.800 --> 00:12:33.580 align:middle line:84% So if you look at one point here, 00:12:33.580 --> 00:12:40.280 align:middle line:84% let's say [? A, ?] then across this surface, 00:12:40.280 --> 00:12:43.118 align:middle line:84% the heat released will be proportional 00:12:43.118 --> 00:12:45.160 align:middle line:84% to essentially the slope of the temperature curve 00:12:45.160 --> 00:12:46.860 align:middle line:90% here, just by Fourier's law. 00:12:46.860 --> 00:12:52.120 align:middle line:84% Now, the amount of heat released comes from the thickness 00:12:52.120 --> 00:12:54.560 align:middle line:90% of the glass pane. 00:12:54.560 --> 00:12:56.353 align:middle line:84% So you can expect that the amount of heat 00:12:56.353 --> 00:12:58.520 align:middle line:84% that you need to release across a particular surface 00:12:58.520 --> 00:13:00.240 align:middle line:84% inside the glass pane here should 00:13:00.240 --> 00:13:03.460 align:middle line:84% be proportional to the thickness of the shaded area. 00:13:03.460 --> 00:13:04.960 align:middle line:84% The larger the shaded area, the more 00:13:04.960 --> 00:13:07.880 align:middle line:84% heat you have to escape just because you have more material 00:13:07.880 --> 00:13:09.960 align:middle line:84% if you have uniform cooling there. 00:13:09.960 --> 00:13:12.760 align:middle line:84% So what that means is that the heat dissipation 00:13:12.760 --> 00:13:14.920 align:middle line:84% rate, or the slope of this curve, 00:13:14.920 --> 00:13:19.520 align:middle line:84% should be proportional to the thickness of this shaded area. 00:13:19.520 --> 00:13:22.240 align:middle line:84% That's exactly the curve of a parabola. 00:13:22.240 --> 00:13:24.680 align:middle line:84% The slope of parabola scales linearly 00:13:24.680 --> 00:13:28.040 align:middle line:90% with the horizontal coordinate. 00:13:28.040 --> 00:13:30.400 align:middle line:90% Does that make sense? 00:13:30.400 --> 00:13:34.100 align:middle line:84% So that's why we argue that this must be a parabolic shape, 00:13:34.100 --> 00:13:35.740 align:middle line:90% [INAUDIBLE] the ideal scenario. 00:13:35.740 --> 00:13:40.080 align:middle line:84% So now in this case, as I said, at high temperature, 00:13:40.080 --> 00:13:43.140 align:middle line:84% essentially the temperature is non-uniform, but stress is zero. 00:13:43.140 --> 00:13:45.440 align:middle line:90% You start to cool down. 00:13:45.440 --> 00:13:48.580 align:middle line:84% If things happen in more or less an ideal manner, 00:13:48.580 --> 00:13:51.480 align:middle line:84% then this temperature profile is going to stay constant. 00:13:51.480 --> 00:13:56.600 align:middle line:84% It's going to decrease uniformly across the entire glass pane. 00:13:56.600 --> 00:14:00.140 align:middle line:84% Now, if you decrease-increase across the glass pane uniformly, 00:14:00.140 --> 00:14:02.560 align:middle line:84% it means that the entire piece of glass 00:14:02.560 --> 00:14:05.840 align:middle line:84% is contracting in a uniform manner, 00:14:05.840 --> 00:14:09.180 align:middle line:84% as long as the temperature profile stays the same shape. 00:14:09.180 --> 00:14:11.120 align:middle line:90% You just drop down. 00:14:11.120 --> 00:14:16.560 align:middle line:84% So in this case, there's still no stress. 00:14:16.560 --> 00:14:20.160 align:middle line:84% You only start to build up stress when the surface gets 00:14:20.160 --> 00:14:22.470 align:middle line:90% close to room temperature. 00:14:22.470 --> 00:14:24.220 align:middle line:84% When the surface gets to room temperature, 00:14:24.220 --> 00:14:25.960 align:middle line:90% it will not cool down anymore. 00:14:25.960 --> 00:14:28.640 align:middle line:84% So contraction will stop, but the inside 00:14:28.640 --> 00:14:31.640 align:middle line:84% is still at high temperature because of the parabolic shape. 00:14:31.640 --> 00:14:33.760 align:middle line:84% So the inside will continue to cool down 00:14:33.760 --> 00:14:36.680 align:middle line:84% and it will continue to contract. 00:14:36.680 --> 00:14:38.880 align:middle line:84% So in this process, this is when you actually 00:14:38.880 --> 00:14:41.180 align:middle line:90% build up this stress gradient. 00:14:41.180 --> 00:14:42.620 align:middle line:90% The surface becomes compressed. 00:14:42.620 --> 00:14:44.560 align:middle line:90% The inside has tensile stress. 00:14:44.560 --> 00:14:47.400 align:middle line:84% So the key takeaway message here is 00:14:47.400 --> 00:14:49.720 align:middle line:84% that when you're doing the thermal tempering process, 00:14:49.720 --> 00:14:52.507 align:middle line:84% you do not build up the stress under high temperature. 00:14:52.507 --> 00:14:54.840 align:middle line:84% You actually build up stress when the material is cooled 00:14:54.840 --> 00:14:59.600 align:middle line:90% down close to room temperature. 00:14:59.600 --> 00:15:03.175 align:middle line:84% So only when you start to have essentially the surface 00:15:03.175 --> 00:15:04.800 align:middle line:84% start to reach room temperature, that's 00:15:04.800 --> 00:15:06.633 align:middle line:84% when you actually build up the stress there. 00:15:06.633 --> 00:15:08.960 align:middle line:90% 00:15:08.960 --> 00:15:11.623 align:middle line:84% All right, so this plot, essentially, I 00:15:11.623 --> 00:15:12.540 align:middle line:90% give you the equation. 00:15:12.540 --> 00:15:14.498 align:middle line:84% I already tell you that the temperature profile 00:15:14.498 --> 00:15:16.040 align:middle line:90% should have a parabolic shape. 00:15:16.040 --> 00:15:20.440 align:middle line:84% So as a result, if the thermal expansion is uniform again, 00:15:20.440 --> 00:15:22.680 align:middle line:84% then the stress profile should also 00:15:22.680 --> 00:15:25.000 align:middle line:90% follow the same parabolic shape. 00:15:25.000 --> 00:15:28.020 align:middle line:84% So then, from the argument of mechanical equilibrium, 00:15:28.020 --> 00:15:30.590 align:middle line:84% the area under the tensile and compressive stress curve 00:15:30.590 --> 00:15:31.290 align:middle line:90% should be equal. 00:15:31.290 --> 00:15:34.070 align:middle line:84% You can then determine where this kind of zero stress plane 00:15:34.070 --> 00:15:34.990 align:middle line:90% should be. 00:15:34.990 --> 00:15:37.630 align:middle line:84% And that's where we come up with the conclusion that about 20% 00:15:37.630 --> 00:15:39.310 align:middle line:84% of the glass thickness on the surface 00:15:39.310 --> 00:15:41.170 align:middle line:90% is under compressive strain. 00:15:41.170 --> 00:15:44.790 align:middle line:90% 00:15:44.790 --> 00:15:48.030 align:middle line:84% All right, so now, here's a few important conclusions 00:15:48.030 --> 00:15:50.350 align:middle line:84% we can draw from this very simple equation here. 00:15:50.350 --> 00:15:53.790 align:middle line:84% The first thing is that assuming, again, 00:15:53.790 --> 00:15:57.790 align:middle line:84% we have this ideal parabolic stress distribution curve, what 00:15:57.790 --> 00:16:01.830 align:middle line:84% you will see is that the surface stress, 00:16:01.830 --> 00:16:05.530 align:middle line:84% they should scale quadratically with the thickness of the glass. 00:16:05.530 --> 00:16:08.670 align:middle line:84% So in this case, sigma, this is basically the surface stress. 00:16:08.670 --> 00:16:12.070 align:middle line:84% X is essentially the horizontal coordinate 00:16:12.070 --> 00:16:14.330 align:middle line:84% which corresponds to thickness of the glass. 00:16:14.330 --> 00:16:17.790 align:middle line:84% So the thicker the glass, the larger the stress 00:16:17.790 --> 00:16:19.510 align:middle line:84% gradient you can actually build up. 00:16:19.510 --> 00:16:24.110 align:middle line:84% And hence, the higher the surface compressive stress 00:16:24.110 --> 00:16:26.590 align:middle line:84% So this can be demonstrated by what 00:16:26.590 --> 00:16:28.550 align:middle line:84% I'm showing on the right-hand side plot here. 00:16:28.550 --> 00:16:30.675 align:middle line:84% For the right-hand side plot, what I'm showing here 00:16:30.675 --> 00:16:33.870 align:middle line:84% is essentially a surface stress measured on essentially 00:16:33.870 --> 00:16:35.630 align:middle line:90% soda-lime glass panes. 00:16:35.630 --> 00:16:38.678 align:middle line:84% That's cooled down just simply using an air quench. 00:16:38.678 --> 00:16:39.970 align:middle line:90% So it's pretty straightforward. 00:16:39.970 --> 00:16:41.637 align:middle line:84% You just get a glass pane, put it there. 00:16:41.637 --> 00:16:44.490 align:middle line:84% Just through convective cooling, it cools down from the furnace. 00:16:44.490 --> 00:16:49.110 align:middle line:84% So each of these curves correspond to a different glass 00:16:49.110 --> 00:16:51.710 align:middle line:84% thickness, ranging from about 2 millimeters 00:16:51.710 --> 00:16:55.470 align:middle line:84% all the way to about more than 1 inch in thickness. 00:16:55.470 --> 00:16:58.910 align:middle line:84% Now, you do see that the stress increases 00:16:58.910 --> 00:17:00.950 align:middle line:84% as you go to thicker and thicker glass panes. 00:17:00.950 --> 00:17:04.550 align:middle line:84% However, if you compare, for example, this top one, 00:17:04.550 --> 00:17:07.270 align:middle line:84% this one is more than 10 times thicker 00:17:07.270 --> 00:17:09.030 align:middle line:90% than the bottom curve here. 00:17:09.030 --> 00:17:11.410 align:middle line:84% So if it follows this quadratic relation, 00:17:11.410 --> 00:17:13.710 align:middle line:84% the surface stress should be 100 times higher 00:17:13.710 --> 00:17:17.910 align:middle line:84% than the thinner glass pane, which obviously is not the case. 00:17:17.910 --> 00:17:21.609 align:middle line:84% So that means that our ideal theory, 00:17:21.609 --> 00:17:24.150 align:middle line:84% this kind of parabolic stress distribution, it 00:17:24.150 --> 00:17:27.410 align:middle line:84% has not taken into account some practical considerations there. 00:17:27.410 --> 00:17:29.990 align:middle line:84% It turns out what happens there is actually 00:17:29.990 --> 00:17:34.310 align:middle line:84% because of viscoelastic relaxation 00:17:34.310 --> 00:17:37.190 align:middle line:84% because we know glass, around the glass transition regime, 00:17:37.190 --> 00:17:38.690 align:middle line:84% is actually viscoelastic material. 00:17:38.690 --> 00:17:41.970 align:middle line:84% And what viscoelastic material does is that it can relax. 00:17:41.970 --> 00:17:45.510 align:middle line:84% So even if you start to build up this kind of stress and strain 00:17:45.510 --> 00:17:49.010 align:middle line:84% gradient, because of relaxation phenomena, 00:17:49.010 --> 00:17:51.190 align:middle line:84% essentially the strain or the stress you build up 00:17:51.190 --> 00:17:53.310 align:middle line:90% would gradually decay over time. 00:17:53.310 --> 00:17:56.390 align:middle line:84% So that's why when you actually get to the final steady state, 00:17:56.390 --> 00:17:58.790 align:middle line:84% the steady state stress you will measure 00:17:58.790 --> 00:18:01.350 align:middle line:84% is much less compared to what you would calculate just based 00:18:01.350 --> 00:18:04.590 align:middle line:90% on the simple theory there. 00:18:04.590 --> 00:18:09.190 align:middle line:84% So that's why if you only use a simple convective cooling, 00:18:09.190 --> 00:18:11.630 align:middle line:84% that's actually a limit on how much stress 00:18:11.630 --> 00:18:13.633 align:middle line:90% you can actually build up. 00:18:13.633 --> 00:18:15.550 align:middle line:84% This can be seen from this plot here where you 00:18:15.550 --> 00:18:16.770 align:middle line:90% increase the glass thickness. 00:18:16.770 --> 00:18:20.150 align:middle line:84% You get diminishing returns in terms of building up the surface 00:18:20.150 --> 00:18:20.750 align:middle line:90% stress. 00:18:20.750 --> 00:18:22.390 align:middle line:84% The maximum stress you can build up 00:18:22.390 --> 00:18:25.860 align:middle line:84% is of the order of about 30 megapascals, which 00:18:25.860 --> 00:18:28.300 align:middle line:84% is still quite remarkable because the atmospheric pressure 00:18:28.300 --> 00:18:28.980 align:middle line:90% is what? 00:18:28.980 --> 00:18:31.140 align:middle line:90% It's like 0.1 megapascal. 00:18:31.140 --> 00:18:34.588 align:middle line:84% So this is, like, 300 times atmospheric pressure 00:18:34.588 --> 00:18:36.380 align:middle line:84% in terms of compressive stress on the glass 00:18:36.380 --> 00:18:38.460 align:middle line:84% surface if it's a thick [INAUDIBLE]. 00:18:38.460 --> 00:18:39.860 align:middle line:90% So it is a lot. 00:18:39.860 --> 00:18:43.323 align:middle line:90% But people want to do better. 00:18:43.323 --> 00:18:45.240 align:middle line:84% Oh, by the way, before I go to the next point, 00:18:45.240 --> 00:18:46.980 align:middle line:84% I also want to mention one quick thing. 00:18:46.980 --> 00:18:48.940 align:middle line:84% Also, our simplified theory also ignores 00:18:48.940 --> 00:18:54.300 align:middle line:84% one important consideration here about fictive temperature. 00:18:54.300 --> 00:18:58.340 align:middle line:84% So I have mentioned about the assumption of this kind 00:18:58.340 --> 00:19:00.100 align:middle line:90% of parabolic stress buildup. 00:19:00.100 --> 00:19:01.850 align:middle line:90% The glass should be uniform. 00:19:01.850 --> 00:19:04.100 align:middle line:84% The inside and outside should have the same properties 00:19:04.100 --> 00:19:05.780 align:middle line:90% throughout the cooling process. 00:19:05.780 --> 00:19:07.620 align:middle line:84% It is not because the surface actually 00:19:07.620 --> 00:19:09.100 align:middle line:90% cools down faster effectively. 00:19:09.100 --> 00:19:10.253 align:middle line:90% Inside cools down slower. 00:19:10.253 --> 00:19:12.920 align:middle line:84% So that means that the structure of the glass is also different. 00:19:12.920 --> 00:19:15.140 align:middle line:84% The surface should have a fictive temperature 00:19:15.140 --> 00:19:17.920 align:middle line:84% that's actually higher compared to the inside. 00:19:17.920 --> 00:19:21.300 align:middle line:90% 00:19:21.300 --> 00:19:22.660 align:middle line:90% You can think about it. 00:19:22.660 --> 00:19:24.140 align:middle line:90% Yeah, OK. 00:19:24.140 --> 00:19:25.560 align:middle line:84% You should know it by this point. 00:19:25.560 --> 00:19:27.852 align:middle line:84% You can draw your cooling plot, and then figure it out. 00:19:27.852 --> 00:19:30.080 align:middle line:84% So now if you have a fictive temperature gradient, 00:19:30.080 --> 00:19:31.275 align:middle line:90% what does that translate to? 00:19:31.275 --> 00:19:32.900 align:middle line:84% Actually, different fictive temperature 00:19:32.900 --> 00:19:34.140 align:middle line:90% means different structure. 00:19:34.140 --> 00:19:35.840 align:middle line:84% The larger the fictive temperature, 00:19:35.840 --> 00:19:37.580 align:middle line:90% the larger the volume. 00:19:37.580 --> 00:19:40.580 align:middle line:84% So that's also going to change because of volume difference. 00:19:40.580 --> 00:19:42.980 align:middle line:84% That's also going to contribute to essentially 00:19:42.980 --> 00:19:44.910 align:middle line:84% an actual accumulation of the stress there. 00:19:44.910 --> 00:19:46.660 align:middle line:84% So that's actually additional contribution 00:19:46.660 --> 00:19:49.220 align:middle line:84% that we have not taken into account due to this essentially 00:19:49.220 --> 00:19:51.700 align:middle line:84% cooling rate gradient in our simplified 00:19:51.700 --> 00:19:55.460 align:middle line:84% theory of a parabolic stress distribution. 00:19:55.460 --> 00:19:58.260 align:middle line:84% Anyway, so coming back to my point 00:19:58.260 --> 00:20:00.140 align:middle line:84% here, if you just do a simple convective 00:20:00.140 --> 00:20:04.300 align:middle line:84% cooling, the glass you get, this is called a thermally tempered 00:20:04.300 --> 00:20:06.660 align:middle line:90% glass. 00:20:06.660 --> 00:20:11.260 align:middle line:84% And you can only build up a maximum stress. 00:20:11.260 --> 00:20:15.040 align:middle line:84% Actually, as I said, it's about 30 megapascals, give or take. 00:20:15.040 --> 00:20:19.980 align:middle line:84% So this kind of heat-strengthened glass 00:20:19.980 --> 00:20:24.960 align:middle line:84% has a limited surface stress up to between 24 to 52 megapascals. 00:20:24.960 --> 00:20:26.740 align:middle line:84% So this is actually regulated, by the way. 00:20:26.740 --> 00:20:29.020 align:middle line:84% So ASTM actually has a standard that 00:20:29.020 --> 00:20:32.660 align:middle line:84% specifically defines two classes of thermally tempered classes. 00:20:32.660 --> 00:20:35.780 align:middle line:84% One is called heat-strengthened glass. 00:20:35.780 --> 00:20:37.740 align:middle line:84% The other is called tempered glass. 00:20:37.740 --> 00:20:39.900 align:middle line:84% For heat-strengthened glass, this 00:20:39.900 --> 00:20:43.220 align:middle line:84% is done usually just using simple convective cooling. 00:20:43.220 --> 00:20:47.460 align:middle line:84% The surface stress is relatively low. 00:20:47.460 --> 00:20:50.140 align:middle line:84% I mean, it's still very high pressure there, very high 00:20:50.140 --> 00:20:52.340 align:middle line:84% compressive stress, but it's smaller 00:20:52.340 --> 00:20:53.480 align:middle line:90% compared to tempered glass. 00:20:53.480 --> 00:20:57.540 align:middle line:84% So it gives you a strength that's about twice stronger 00:20:57.540 --> 00:20:59.680 align:middle line:90% compared to annealed glass. 00:20:59.680 --> 00:21:02.420 align:middle line:84% Annealed glass essentially means that there's no internal stress. 00:21:02.420 --> 00:21:06.020 align:middle line:84% It's just a soda-lime glass without any specific treatment. 00:21:06.020 --> 00:21:08.060 align:middle line:90% So it's not strengthened. 00:21:08.060 --> 00:21:10.720 align:middle line:84% Now, you can also go to a tempered glass. 00:21:10.720 --> 00:21:13.780 align:middle line:84% A tempered glass, by this regulation, 00:21:13.780 --> 00:21:17.780 align:middle line:84% it should have a surface stress greater than 69 megapascals. 00:21:17.780 --> 00:21:20.930 align:middle line:84% And it gives you about four to five times stronger 00:21:20.930 --> 00:21:22.970 align:middle line:90% compared to annealed glass. 00:21:22.970 --> 00:21:25.350 align:middle line:84% Now, to get to the tempered glass, 00:21:25.350 --> 00:21:27.950 align:middle line:84% you cannot just do a simple convective cooling. 00:21:27.950 --> 00:21:29.820 align:middle line:84% You have to use essentially some way 00:21:29.820 --> 00:21:31.570 align:middle line:84% to actually accelerate the cooling process 00:21:31.570 --> 00:21:33.110 align:middle line:90% to build up even more stress. 00:21:33.110 --> 00:21:36.330 align:middle line:84% The typical way of doing it, you can either use compressed air. 00:21:36.330 --> 00:21:39.970 align:middle line:84% You can use essentially forced cooling, and blow compressed air 00:21:39.970 --> 00:21:41.310 align:middle line:90% to expedite the cooling. 00:21:41.310 --> 00:21:45.410 align:middle line:84% You can also do a liquid quenching. 00:21:45.410 --> 00:21:48.370 align:middle line:84% In that case, you either use a mineral oil 00:21:48.370 --> 00:21:54.530 align:middle line:84% or you actually use some salt solutions like potassium nitrite 00:21:54.530 --> 00:21:55.695 align:middle line:90% or sodium nitrite. 00:21:55.695 --> 00:21:57.570 align:middle line:84% Those are typical compositions that you would 00:21:57.570 --> 00:21:59.370 align:middle line:90% use to quench these glasses. 00:21:59.370 --> 00:22:02.730 align:middle line:84% There's more recent techniques people are also using. 00:22:02.730 --> 00:22:06.410 align:middle line:84% People also using things like fork cooling. 00:22:06.410 --> 00:22:08.250 align:middle line:84% So fork cooling, they essentially 00:22:08.250 --> 00:22:11.050 align:middle line:84% spray very tiny little droplets of liquid 00:22:11.050 --> 00:22:13.610 align:middle line:84% onto a surface of glass to actually help with cooling down. 00:22:13.610 --> 00:22:15.490 align:middle line:90% That's a more recent technique. 00:22:15.490 --> 00:22:20.730 align:middle line:84% People are also using actually solids as a cooling media. 00:22:20.730 --> 00:22:23.265 align:middle line:84% So in that case, they actually put in a fluidized bed. 00:22:23.265 --> 00:22:25.890 align:middle line:84% In the fluidized bed, you have these tiny little beads, usually 00:22:25.890 --> 00:22:26.930 align:middle line:90% made of alumina. 00:22:26.930 --> 00:22:29.470 align:middle line:84% They have a small diameter of a few hundred microns. 00:22:29.470 --> 00:22:32.610 align:middle line:84% Under appropriate conditions, the collection of particles 00:22:32.610 --> 00:22:33.715 align:middle line:90% behaves like liquid. 00:22:33.715 --> 00:22:35.590 align:middle line:84% And then you can essentially put it in there. 00:22:35.590 --> 00:22:37.007 align:middle line:84% And then these particles are going 00:22:37.007 --> 00:22:39.930 align:middle line:84% to roll across the glass surface and induce cooling there. 00:22:39.930 --> 00:22:41.570 align:middle line:84% So those are the more recent techniques 00:22:41.570 --> 00:22:45.950 align:middle line:84% compared to the liquid or forced air, compressed air cooling OK. 00:22:45.950 --> 00:22:47.090 align:middle line:90% Yes. 00:22:47.090 --> 00:22:50.530 align:middle line:84% AUDIENCE: What are the, I guess, constraints 00:22:50.530 --> 00:22:54.810 align:middle line:84% on mineral oil or salt solution [INAUDIBLE]? 00:22:54.810 --> 00:23:00.228 align:middle line:84% Is it just water is too good in [INAUDIBLE] glass [INAUDIBLE]. 00:23:00.228 --> 00:23:01.270 align:middle line:90% JUEJUN HU: Good question. 00:23:01.270 --> 00:23:02.645 align:middle line:84% I actually don't know exactly why 00:23:02.645 --> 00:23:05.790 align:middle line:84% they use that specific chemical formula for cooling there. 00:23:05.790 --> 00:23:06.670 align:middle line:90% I have to look it up. 00:23:06.670 --> 00:23:07.430 align:middle line:90% That's a good question. 00:23:07.430 --> 00:23:09.230 align:middle line:84% I know that that's what they use, the chemical formula. 00:23:09.230 --> 00:23:11.147 align:middle line:84% I don't know exactly why they actually use it. 00:23:11.147 --> 00:23:12.265 align:middle line:90% So let me look it up. 00:23:12.265 --> 00:23:13.890 align:middle line:84% I can't answer your question right now. 00:23:13.890 --> 00:23:16.570 align:middle line:90% 00:23:16.570 --> 00:23:19.630 align:middle line:84% AUDIENCE: And for the salt solutions, 00:23:19.630 --> 00:23:23.870 align:middle line:84% if you're quenching from a high enough temperature, 00:23:23.870 --> 00:23:26.130 align:middle line:84% is there any chemical strengthening 00:23:26.130 --> 00:23:32.650 align:middle line:90% that happens at the same time? 00:23:32.650 --> 00:23:35.690 align:middle line:84% JUEJUN HU: I would say no, or [? at least ?] quite minimally. 00:23:35.690 --> 00:23:39.930 align:middle line:84% In fact, what happens is that we will see, 00:23:39.930 --> 00:23:43.610 align:middle line:84% for chemical strengthening, it requires probably more extreme 00:23:43.610 --> 00:23:44.310 align:middle line:90% conditions. 00:23:44.310 --> 00:23:47.373 align:middle line:84% And also, it requires much longer duration. 00:23:47.373 --> 00:23:48.790 align:middle line:84% You can also use-- in fact, we are 00:23:48.790 --> 00:23:50.130 align:middle line:84% going to show you one example where people 00:23:50.130 --> 00:23:52.530 align:middle line:84% use supercritical fluid to do strengthening [INAUDIBLE] 00:23:52.530 --> 00:23:53.590 align:middle line:90% chemical strengthening. 00:23:53.590 --> 00:23:56.250 align:middle line:84% And in that case, people usually use [? hours ?] for diffusion 00:23:56.250 --> 00:23:57.330 align:middle line:90% to take place. 00:23:57.330 --> 00:24:00.350 align:middle line:84% During the quenching process, maybe like surface monolayer, 00:24:00.350 --> 00:24:02.010 align:middle line:90% but that's really negligible. 00:24:02.010 --> 00:24:04.250 align:middle line:90% Yeah. 00:24:04.250 --> 00:24:07.430 align:middle line:84% OK, so bottom line, tempered glass is stronger. 00:24:07.430 --> 00:24:09.690 align:middle line:90% It has more stress. 00:24:09.690 --> 00:24:12.010 align:middle line:84% And you can actually see that from actually 00:24:12.010 --> 00:24:14.610 align:middle line:84% the breaking behavior of thermally strengthened glass 00:24:14.610 --> 00:24:15.890 align:middle line:90% and tempered glass. 00:24:15.890 --> 00:24:18.390 align:middle line:84% For thermally strengthened glass, when it breaks, 00:24:18.390 --> 00:24:20.610 align:middle line:84% it's behaving in a way very similar 00:24:20.610 --> 00:24:22.930 align:middle line:90% to just an untreated glass. 00:24:22.930 --> 00:24:27.530 align:middle line:84% You end up with big shards with sharp edges and corners, 00:24:27.530 --> 00:24:30.210 align:middle line:84% which can be dangerous to users, of course. 00:24:30.210 --> 00:24:33.650 align:middle line:84% For thermally tempered glass, when they break, 00:24:33.650 --> 00:24:37.050 align:middle line:84% they tend to break into small pieces, tiny little pieces, 00:24:37.050 --> 00:24:39.330 align:middle line:84% in lots of cases, also with rounded corners. 00:24:39.330 --> 00:24:41.670 align:middle line:84% So it's actually safer to use this. 00:24:41.670 --> 00:24:43.250 align:middle line:90% There are downsides, though. 00:24:43.250 --> 00:24:45.090 align:middle line:84% For thermally tempered glass, they 00:24:45.090 --> 00:24:49.050 align:middle line:84% have the risk of spontaneous breaking. 00:24:49.050 --> 00:24:50.610 align:middle line:90% That happened to me actually. 00:24:50.610 --> 00:24:54.490 align:middle line:84% Like, the microwave oven glass door in my home? 00:24:54.490 --> 00:24:56.570 align:middle line:84% One day, we wake up and it's completely 00:24:56.570 --> 00:25:00.077 align:middle line:90% shattered for no reason. 00:25:00.077 --> 00:25:01.410 align:middle line:90% Well, there's actually a reason. 00:25:01.410 --> 00:25:02.470 align:middle line:90% We won't talk about it. 00:25:02.470 --> 00:25:05.270 align:middle line:84% But apparently, nobody tried to touch it or whatever. 00:25:05.270 --> 00:25:06.070 align:middle line:90% It just broke. 00:25:06.070 --> 00:25:07.990 align:middle line:90% OK, so it does happen. 00:25:07.990 --> 00:25:10.800 align:middle line:84% Any tempered glass products, it could break spontaneously. 00:25:10.800 --> 00:25:15.042 align:middle line:84% The reason for that is because of-- well, we're 00:25:15.042 --> 00:25:16.500 align:middle line:84% going to talk about it in a second. 00:25:16.500 --> 00:25:18.792 align:middle line:84% But before that, I want to mention that that's actually 00:25:18.792 --> 00:25:22.613 align:middle line:84% because of the very high tensile stress inside the material. 00:25:22.613 --> 00:25:24.780 align:middle line:84% If you have large compressive stress on the surface, 00:25:24.780 --> 00:25:26.820 align:middle line:84% obviously, you have high tensile stress. 00:25:26.820 --> 00:25:29.080 align:middle line:84% The tensile stress region is the Achilles heel 00:25:29.080 --> 00:25:32.060 align:middle line:84% for this kind of thermally tempered glass. 00:25:32.060 --> 00:25:33.480 align:middle line:84% And because of defects that we're 00:25:33.480 --> 00:25:34.700 align:middle line:84% going to talk about in a couple of slides, 00:25:34.700 --> 00:25:36.992 align:middle line:84% that's what's causing this kind of spontaneous breakage 00:25:36.992 --> 00:25:38.557 align:middle line:90% in the tempered glass. 00:25:38.557 --> 00:25:40.640 align:middle line:84% For heat-strengthened glass, because of the stress 00:25:40.640 --> 00:25:44.860 align:middle line:84% is very low, you don't have that risk. 00:25:44.860 --> 00:25:46.440 align:middle line:84% Now I'm actually going to show you-- 00:25:46.440 --> 00:25:48.187 align:middle line:84% I already showed the YouTube video there. 00:25:48.187 --> 00:25:50.020 align:middle line:84% The reason I want to show this YouTube video 00:25:50.020 --> 00:25:53.640 align:middle line:84% is because Prince Rupert's drop, which you see in the Glass Lab 00:25:53.640 --> 00:25:59.640 align:middle line:84% demo, represents the ultimate or extreme example 00:25:59.640 --> 00:26:02.120 align:middle line:90% of thermal tempering of glasses. 00:26:02.120 --> 00:26:05.100 align:middle line:84% So because of this extremely fast cooling rate, 00:26:05.100 --> 00:26:06.900 align:middle line:84% you're actually starting not from 600 C. 00:26:06.900 --> 00:26:08.160 align:middle line:84% You're starting from molten glass which 00:26:08.160 --> 00:26:09.285 align:middle line:90% is even higher temperature. 00:26:09.285 --> 00:26:12.020 align:middle line:84% So you quickly cool down through a liquid. 00:26:12.020 --> 00:26:15.200 align:middle line:84% This gives rise to extremely high compressive stress build-up 00:26:15.200 --> 00:26:17.000 align:middle line:90% in the Prince Rupert's drop. 00:26:17.000 --> 00:26:21.760 align:middle line:84% And it's so strong that it can actually withstand bullets. 00:26:21.760 --> 00:26:25.568 align:middle line:84% So I'm going to show you this video here. 00:26:25.568 --> 00:26:27.950 align:middle line:90% [VIDEO PLAYBACK] 00:26:27.950 --> 00:26:29.312 align:middle line:90% - All right, Keith. 00:26:29.312 --> 00:26:30.520 align:middle line:90% This is Prince Rupert's drop. 00:26:30.520 --> 00:26:31.853 align:middle line:90% JUEJUN HU: This is the real one. 00:26:31.853 --> 00:26:34.880 align:middle line:84% - Prince Rupert's drop, the [INAUDIBLE]. 00:26:34.880 --> 00:26:38.620 align:middle line:84% Papers submitted from 1660 to the Royal Society. 00:26:38.620 --> 00:26:40.460 align:middle line:90% So this is the very early stuff. 00:26:40.460 --> 00:26:41.620 align:middle line:90% - Hey, it's me, Destin. 00:26:41.620 --> 00:26:42.620 align:middle line:90% Welcome back to SmarterEveryDay. 00:26:42.620 --> 00:26:44.080 align:middle line:84% I am in the basement of the Royal 00:26:44.080 --> 00:26:45.660 align:middle line:90% Society in London, England. 00:26:45.660 --> 00:26:47.980 align:middle line:84% And this is Keith, the head librarian. 00:26:47.980 --> 00:26:50.020 align:middle line:84% And he has a special paper written by who? 00:26:50.020 --> 00:26:51.260 align:middle line:90% - Sir Robert Moray. 00:26:51.260 --> 00:26:53.540 align:middle line:84% - And it's about the Prince Rupert's drop. 00:26:53.540 --> 00:26:54.300 align:middle line:90% - That's right. 00:26:54.300 --> 00:26:55.300 align:middle line:90% - Can you show it to me? 00:26:55.300 --> 00:26:58.400 align:middle line:84% - Yep, so here we have an account of the glass 00:26:58.400 --> 00:27:00.688 align:middle line:90% drops by Sir Robert Moray. 00:27:00.688 --> 00:27:01.980 align:middle line:90% - Will you show me the drawing? 00:27:01.980 --> 00:27:03.020 align:middle line:90% The famous drawing? 00:27:03.020 --> 00:27:05.320 align:middle line:90% - So just turn over the pages. 00:27:05.320 --> 00:27:09.900 align:middle line:84% "So is a kingdom one and strong, but when the top is broken 00:27:09.900 --> 00:27:11.340 align:middle line:90% shivers into men." 00:27:11.340 --> 00:27:13.920 align:middle line:84% - So the strength of the Prince Rupert 00:27:13.920 --> 00:27:16.680 align:middle line:84% was recognized way back in the day, [? even in ?] the beginning 00:27:16.680 --> 00:27:17.820 align:middle line:90% of the Royal Society. 00:27:17.820 --> 00:27:18.720 align:middle line:90% [INAUDIBLE] 00:27:18.720 --> 00:27:19.250 align:middle line:90% - It is. 00:27:19.250 --> 00:27:20.500 align:middle line:90% JUEJUN HU: Don't tell me that. 00:27:20.500 --> 00:27:23.430 align:middle line:90% [LAUGHTER] 00:27:23.430 --> 00:27:25.680 align:middle line:90% - I was fascinated by them. 00:27:25.680 --> 00:27:27.100 align:middle line:84% - They were fascinated back then. 00:27:27.100 --> 00:27:28.142 align:middle line:90% I'm still fascinated now. 00:27:28.142 --> 00:27:30.517 align:middle line:84% But today, we're going to use some of the best technology 00:27:30.517 --> 00:27:31.500 align:middle line:90% that we have available. 00:27:31.500 --> 00:27:33.083 align:middle line:84% We're going to use a high-speed camera 00:27:33.083 --> 00:27:36.200 align:middle line:84% and record a bullet striking the bulb over Prince Rupert's drop 00:27:36.200 --> 00:27:39.680 align:middle line:84% at several hundred thousand frames per second. 00:27:39.680 --> 00:27:41.160 align:middle line:90% Let's go get smarter every day. 00:27:41.160 --> 00:27:42.180 align:middle line:90% All right, goggle up. 00:27:42.180 --> 00:27:43.100 align:middle line:90% Science is about to happen. 00:27:43.100 --> 00:27:44.920 align:middle line:84% The first video I made about the Prince Rupert's drop here 00:27:44.920 --> 00:27:46.720 align:middle line:84% at Lookout Mountain [INAUDIBLE] was 00:27:46.720 --> 00:27:49.200 align:middle line:84% about why the Prince Rupert's drop was so strong. 00:27:49.200 --> 00:27:51.050 align:middle line:84% Today, we're going to test the limits. 00:27:51.050 --> 00:27:52.800 align:middle line:84% If you haven't seen that video, you really 00:27:52.800 --> 00:27:54.320 align:middle line:84% need to watch it because you have 00:27:54.320 --> 00:27:57.520 align:middle line:84% to understand what we're dealing with before this video will make 00:27:57.520 --> 00:27:58.460 align:middle line:90% any sense. 00:27:58.460 --> 00:28:00.040 align:middle line:90% This is the setup. 00:28:00.040 --> 00:28:04.100 align:middle line:84% I've got a .22 rifle aimed at a Prince Rupert's drop. 00:28:04.100 --> 00:28:06.150 align:middle line:84% I've got a high-speed camera right here. 00:28:06.150 --> 00:28:10.070 align:middle line:84% This is a v2511 We're going to run it over 100,000 frames per 00:28:10.070 --> 00:28:10.610 align:middle line:90% second. 00:28:10.610 --> 00:28:13.350 align:middle line:84% So I've got the rifle aligned so that we should 00:28:13.350 --> 00:28:14.850 align:middle line:90% hit the Prince Rupert's drop. 00:28:14.850 --> 00:28:16.290 align:middle line:84% The resolution is going to be a little bit lower, 00:28:16.290 --> 00:28:18.082 align:middle line:84% because that's what you have to do in order 00:28:18.082 --> 00:28:19.288 align:middle line:90% to get so much speed. 00:28:19.288 --> 00:28:20.330 align:middle line:90% But it should be awesome. 00:28:20.330 --> 00:28:20.850 align:middle line:90% Here we go. 00:28:20.850 --> 00:28:21.810 align:middle line:90% Get your goggles on. 00:28:21.810 --> 00:28:22.510 align:middle line:90% Let's do it. 00:28:22.510 --> 00:28:25.070 align:middle line:84% My Prince Rupert's drop is wobbling, 00:28:25.070 --> 00:28:27.210 align:middle line:90% so I have to time the bullet. 00:28:27.210 --> 00:28:29.150 align:middle line:90% [MUSIC PLAYING] 00:28:29.150 --> 00:28:32.510 align:middle line:90% [GUNFIRE] 00:28:32.510 --> 00:28:35.870 align:middle line:90% 00:28:35.870 --> 00:28:36.530 align:middle line:90% All right. 00:28:36.530 --> 00:28:38.490 align:middle line:90% [INAUDIBLE] look at that. 00:28:38.490 --> 00:28:39.970 align:middle line:84% JUEJUN HU: [INAUDIBLE] slow motion. 00:28:39.970 --> 00:28:43.430 align:middle line:90% 00:28:43.430 --> 00:28:47.150 align:middle line:90% Look at the bullet [INAUDIBLE]. 00:28:47.150 --> 00:28:48.250 align:middle line:90% And this is what happened? 00:28:48.250 --> 00:28:54.430 align:middle line:90% 00:28:54.430 --> 00:28:55.635 align:middle line:90% The bullet shatters. 00:28:55.635 --> 00:28:57.510 align:middle line:84% AUDIENCE: And then [? there's ?] [INAUDIBLE]. 00:28:57.510 --> 00:29:00.926 align:middle line:90% [MUSIC PLAYING] 00:29:00.926 --> 00:29:04.342 align:middle line:90% 00:29:04.342 --> 00:29:06.190 align:middle line:84% JUEJUN HU: So the Prince Rupert's drop 00:29:06.190 --> 00:29:09.910 align:middle line:84% is intact by creating this elastic [INAUDIBLE] 00:29:09.910 --> 00:29:11.090 align:middle line:90% probably to tail. 00:29:11.090 --> 00:29:13.286 align:middle line:90% [LAUGHS] 00:29:13.286 --> 00:29:16.950 align:middle line:84% - The bullet did not break the Prince Rupert's drop. 00:29:16.950 --> 00:29:18.870 align:middle line:84% It looks like it creates a shockwave that 00:29:18.870 --> 00:29:20.790 align:middle line:90% runs down the tail. 00:29:20.790 --> 00:29:23.870 align:middle line:84% And then the tail shakes, breaks, and then 00:29:23.870 --> 00:29:25.350 align:middle line:90% breaks it back the other way. 00:29:25.350 --> 00:29:27.030 align:middle line:84% I'm going to do this several more times 00:29:27.030 --> 00:29:28.590 align:middle line:90% because that is beautiful. 00:29:28.590 --> 00:29:32.470 align:middle line:84% That gun is pointed in the general direction 00:29:32.470 --> 00:29:33.610 align:middle line:90% of the high-speed camera. 00:29:33.610 --> 00:29:36.793 align:middle line:84% So I'm a little nervous about that point. 00:29:36.793 --> 00:29:38.210 align:middle line:84% JUEJUN HU: Don't try that at home. 00:29:38.210 --> 00:29:41.020 align:middle line:90% 00:29:41.020 --> 00:29:42.870 align:middle line:90% [GUNFIRE] 00:29:42.870 --> 00:29:48.150 align:middle line:84% - Just make sure [INAUDIBLE] [? Everything ?] looks good. 00:29:48.150 --> 00:29:50.390 align:middle line:90% Look at the high-speed now. 00:29:50.390 --> 00:29:51.590 align:middle line:90% God! 00:29:51.590 --> 00:29:52.810 align:middle line:90% It's awesome! 00:29:52.810 --> 00:29:55.190 align:middle line:90% [LAUGHS] 00:29:55.190 --> 00:29:57.910 align:middle line:90% Dang! 00:29:57.910 --> 00:30:00.430 align:middle line:90% That is beautiful. 00:30:00.430 --> 00:30:03.790 align:middle line:84% It's beautiful and it's material science. 00:30:03.790 --> 00:30:07.230 align:middle line:84% And if those jokers back in the 1600s had this kind 00:30:07.230 --> 00:30:09.973 align:middle line:84% of technology, they would have written way more interesting 00:30:09.973 --> 00:30:10.890 align:middle line:90% letters to each other. 00:30:10.890 --> 00:30:13.710 align:middle line:84% Not that they didn't do a good job because they're awesome. 00:30:13.710 --> 00:30:15.210 align:middle line:90% 00:30:15.210 --> 00:30:17.530 align:middle line:84% JUEJUN HU: OK, we can skip the video. 00:30:17.530 --> 00:30:21.150 align:middle line:84% He's just going to try to shoot it multiple times. 00:30:21.150 --> 00:30:23.350 align:middle line:90% I think we can move over here. 00:30:23.350 --> 00:30:25.470 align:middle line:84% There's somewhere [? where it is ?] more clear. 00:30:25.470 --> 00:30:27.870 align:middle line:84% OK, let me try to show you something, a slightly 00:30:27.870 --> 00:30:28.970 align:middle line:90% different video here. 00:30:28.970 --> 00:30:35.390 align:middle line:84% So the same guy obviously gets fascinated 00:30:35.390 --> 00:30:37.270 align:middle line:90% with this kind of experiment. 00:30:37.270 --> 00:30:40.270 align:middle line:84% He then somehow got an AK-47 and then 00:30:40.270 --> 00:30:42.193 align:middle line:84% tries to shoot Prince Rupert's drop. 00:30:42.193 --> 00:30:43.610 align:middle line:84% - Welcome back to SmarterEveryDay. 00:30:43.610 --> 00:30:45.850 align:middle line:84% I've been waiting on a sunny day to do this. 00:30:45.850 --> 00:30:47.310 align:middle line:84% If you remember the last video, I 00:30:47.310 --> 00:30:49.790 align:middle line:84% fired a .38 special versus a Prince Rupert's 00:30:49.790 --> 00:30:51.670 align:middle line:84% drop and the Prince Rupert's drop won. 00:30:51.670 --> 00:30:53.170 align:middle line:84% Well, we're going to fix that today. 00:30:53.170 --> 00:30:54.310 align:middle line:84% We've got the high-speed camera back out. 00:30:54.310 --> 00:30:54.893 align:middle line:90% [END PLAYBACK] 00:30:54.893 --> 00:30:57.070 align:middle line:84% JUEJUN HU: Let me call that fractal luminescence. 00:30:57.070 --> 00:30:59.470 align:middle line:84% It has nothing to do with amorphous structure. 00:30:59.470 --> 00:31:02.130 align:middle line:84% It has to do with polar chemical bonds. 00:31:02.130 --> 00:31:03.878 align:middle line:84% So if I have [? a ?] polar chemical bond, 00:31:03.878 --> 00:31:06.170 align:middle line:84% it means that I have charge transfer between the atoms. 00:31:06.170 --> 00:31:07.450 align:middle line:90% So we actually break it quickly. 00:31:07.450 --> 00:31:08.790 align:middle line:84% Turns out some of the charges actually 00:31:08.790 --> 00:31:09.870 align:middle line:90% remain on the two paths. 00:31:09.870 --> 00:31:11.690 align:middle line:84% And those charges have to relax somewhere. 00:31:11.690 --> 00:31:13.510 align:middle line:84% So eventually, they relax essentially 00:31:13.510 --> 00:31:16.970 align:middle line:90% by going across the bandgap. 00:31:16.970 --> 00:31:18.830 align:middle line:84% And that goes into the form of photons. 00:31:18.830 --> 00:31:22.310 align:middle line:84% So that's basically what happens there. 00:31:22.310 --> 00:31:24.350 align:middle line:84% OK, so bottom line, what you have seen here-- 00:31:24.350 --> 00:31:27.470 align:middle line:84% Prince Rupert's drop is extremely strong. 00:31:27.470 --> 00:31:29.470 align:middle line:90% It actually shatters the bullet. 00:31:29.470 --> 00:31:36.613 align:middle line:84% So about nine years ago, a group from Purdue University 00:31:36.613 --> 00:31:38.030 align:middle line:84% figured that maybe they should try 00:31:38.030 --> 00:31:41.350 align:middle line:84% to quantify how much stress is actually 00:31:41.350 --> 00:31:43.070 align:middle line:84% on the surface of Prince Rupert's drop. 00:31:43.070 --> 00:31:46.830 align:middle line:84% And what's the amount of force it can withstand before it 00:31:46.830 --> 00:31:47.890 align:middle line:90% actually really breaks? 00:31:47.890 --> 00:31:50.430 align:middle line:84% So they actually published a paper. 00:31:50.430 --> 00:31:53.390 align:middle line:84% And this one actually got a lot of public attention. 00:31:53.390 --> 00:31:55.230 align:middle line:84% They were able to actually directly measure 00:31:55.230 --> 00:31:58.588 align:middle line:84% the stress level on Prince Rupert's drop and use 00:31:58.588 --> 00:32:00.380 align:middle line:84% that to actually compute the maximum amount 00:32:00.380 --> 00:32:03.140 align:middle line:84% of compressive load they actually withstand 00:32:03.140 --> 00:32:05.420 align:middle line:90% before it actually breaks. 00:32:05.420 --> 00:32:07.940 align:middle line:84% I'm a bit more worried about how they actually measured 00:32:07.940 --> 00:32:09.393 align:middle line:90% the compressive stress. 00:32:09.393 --> 00:32:11.060 align:middle line:84% The way they measured compressive stress 00:32:11.060 --> 00:32:15.620 align:middle line:84% is using a technique of photoelasticity. 00:32:15.620 --> 00:32:22.740 align:middle line:84% This is a very, very common way of measuring stress and strain 00:32:22.740 --> 00:32:27.380 align:middle line:84% in transparent materials like glass or plastics. 00:32:27.380 --> 00:32:31.060 align:middle line:84% The way it works is that you have of polarized light source. 00:32:31.060 --> 00:32:35.220 align:middle line:84% So when a polarized light source [INAUDIBLE] on the material, 00:32:35.220 --> 00:32:38.140 align:middle line:84% in this case, if it's no stress, then the light 00:32:38.140 --> 00:32:39.040 align:middle line:90% will just go through. 00:32:39.040 --> 00:32:40.220 align:middle line:90% It comes out as transparent. 00:32:40.220 --> 00:32:41.620 align:middle line:90% Nothing really happens. 00:32:41.620 --> 00:32:45.060 align:middle line:84% However, if you apply stress, you're 00:32:45.060 --> 00:32:49.020 align:middle line:84% essentially changing the density of electron clouds. 00:32:49.020 --> 00:32:51.240 align:middle line:84% If you're changing the density of electron clouds, 00:32:51.240 --> 00:32:53.460 align:middle line:84% you're going to change the polarizability 00:32:53.460 --> 00:32:58.270 align:middle line:84% when they respond to oscillating electromagnetic waves. 00:32:58.270 --> 00:33:00.500 align:middle line:84% So essentially, it means that the refractive index 00:33:00.500 --> 00:33:02.792 align:middle line:84% is changing when you actually apply this kind of stress 00:33:02.792 --> 00:33:03.420 align:middle line:90% [INAUDIBLE]. 00:33:03.420 --> 00:33:06.460 align:middle line:90% 00:33:06.460 --> 00:33:08.080 align:middle line:90% Makes sense? 00:33:08.080 --> 00:33:09.840 align:middle line:84% You essentially compress the electron. 00:33:09.840 --> 00:33:11.298 align:middle line:84% You think of a compressed electron. 00:33:11.298 --> 00:33:14.180 align:middle line:84% If you compress electron, the reason you have refractive index 00:33:14.180 --> 00:33:15.980 align:middle line:84% is because the electrons are responding 00:33:15.980 --> 00:33:19.020 align:middle line:84% to oscillating electromagnetic field due to the light. 00:33:19.020 --> 00:33:22.180 align:middle line:84% So if that behavior changes, you change the refractive index. 00:33:22.180 --> 00:33:24.980 align:middle line:84% Now, the key thing to take away is that this kind of change 00:33:24.980 --> 00:33:26.960 align:middle line:84% happens not in an isotropic manner. 00:33:26.960 --> 00:33:28.420 align:middle line:90% It happens anisotropically. 00:33:28.420 --> 00:33:30.000 align:middle line:84% Because you apply compressive stress, 00:33:30.000 --> 00:33:31.340 align:middle line:84% you can imagine that you're squeezing 00:33:31.340 --> 00:33:32.465 align:middle line:90% material in this direction. 00:33:32.465 --> 00:33:35.440 align:middle line:84% But the material expands in the other directions. 00:33:35.440 --> 00:33:37.940 align:middle line:84% So now you expect the material actually becomes anisotropic, 00:33:37.940 --> 00:33:40.300 align:middle line:84% even though it started out isotropically. 00:33:40.300 --> 00:33:42.260 align:middle line:84% It becomes a uniaxial material that 00:33:42.260 --> 00:33:44.700 align:middle line:84% goes along the direction of the other stress. 00:33:44.700 --> 00:33:47.400 align:middle line:84% It has a principal axis aligned with that direction. 00:33:47.400 --> 00:33:51.220 align:middle line:84% So what photoelasticity is looking at 00:33:51.220 --> 00:33:56.520 align:middle line:84% is essentially birefringent materials or birefringence 00:33:56.520 --> 00:33:57.220 align:middle line:90% in materials. 00:33:57.220 --> 00:34:00.860 align:middle line:90% 00:34:00.860 --> 00:34:03.140 align:middle line:84% Birefringence just means that the material 00:34:03.140 --> 00:34:09.820 align:middle line:84% has anisotropic refractive index distribution. 00:34:09.820 --> 00:34:11.580 align:middle line:84% So if a polarized light [INAUDIBLE] 00:34:11.580 --> 00:34:13.820 align:middle line:84% on a birefringent material, it's going 00:34:13.820 --> 00:34:19.860 align:middle line:84% to split into two radiations or two light beams. 00:34:19.860 --> 00:34:21.320 align:middle line:84% One is called ordinary light beam, 00:34:21.320 --> 00:34:23.620 align:middle line:84% the other called extraordinary light beam. 00:34:23.620 --> 00:34:26.400 align:middle line:84% They will then propagate along slightly different directions. 00:34:26.400 --> 00:34:28.239 align:middle line:84% So then when they come off the material, 00:34:28.239 --> 00:34:30.659 align:middle line:84% if you put another polarizer at the end, 00:34:30.659 --> 00:34:33.260 align:middle line:84% the two beams are going to recombine at the polarizer 00:34:33.260 --> 00:34:34.900 align:middle line:90% and they will interfere. 00:34:34.900 --> 00:34:38.380 align:middle line:84% So now it depends on how much birefringence you have, 00:34:38.380 --> 00:34:42.616 align:middle line:84% which alters the path lengths of the two different light beams. 00:34:42.616 --> 00:34:44.199 align:middle line:84% You may get constructive interference. 00:34:44.199 --> 00:34:46.100 align:middle line:84% You may get destructive interference. 00:34:46.100 --> 00:34:48.817 align:middle line:84% So you will essentially come up with essentially a bright, dark 00:34:48.817 --> 00:34:50.900 align:middle line:84% pattern if you use essentially monochromatic laser 00:34:50.900 --> 00:34:52.190 align:middle line:90% light to illuminate material. 00:34:52.190 --> 00:34:54.690 align:middle line:84% So for that, you can infer how much stress you have. 00:34:54.690 --> 00:34:57.250 align:middle line:84% Or if you illuminate it using essentially 00:34:57.250 --> 00:34:58.930 align:middle line:84% a white light source, you come up 00:34:58.930 --> 00:35:00.610 align:middle line:84% with color patterns, color fringes 00:35:00.610 --> 00:35:03.027 align:middle line:84% like what I'm showing here, because different [? colors ?] 00:35:03.027 --> 00:35:04.690 align:middle line:84% essentially interfere differently. 00:35:04.690 --> 00:35:08.070 align:middle line:84% So I can actually show you how that works. 00:35:08.070 --> 00:35:09.310 align:middle line:90% I brought along with me-- 00:35:09.310 --> 00:35:10.310 align:middle line:90% I'm going to show you. 00:35:10.310 --> 00:35:13.770 align:middle line:84% I have of polarized light source. 00:35:13.770 --> 00:35:18.450 align:middle line:84% Any of you want to guess what's the polarized light source? 00:35:18.450 --> 00:35:20.690 align:middle line:90% My screen. 00:35:20.690 --> 00:35:22.542 align:middle line:84% Most of your laptop screens are polarized. 00:35:22.542 --> 00:35:24.250 align:middle line:84% You can come around here and take a look. 00:35:24.250 --> 00:35:26.470 align:middle line:84% We can probably take a video after later. 00:35:26.470 --> 00:35:27.310 align:middle line:90% Oh I know. 00:35:27.310 --> 00:35:29.930 align:middle line:90% 00:35:29.930 --> 00:35:35.330 align:middle line:84% So this is basically just white, nothing. 00:35:35.330 --> 00:35:39.130 align:middle line:90% 00:35:39.130 --> 00:35:40.310 align:middle line:90% This is a polarizer. 00:35:40.310 --> 00:35:44.367 align:middle line:90% 00:35:44.367 --> 00:35:46.450 align:middle line:84% OK, this is [INAUDIBLE] the polarization direction 00:35:46.450 --> 00:35:50.110 align:middle line:90% line, turn around, stuck. 00:35:50.110 --> 00:35:55.070 align:middle line:84% OK, so now we'll do it in a cross polarized orientation. 00:35:55.070 --> 00:35:56.037 align:middle line:90% We'll put it like this. 00:35:56.037 --> 00:35:58.370 align:middle line:84% You put the component that actually has intrinsic stress 00:35:58.370 --> 00:36:01.050 align:middle line:90% or strain in there. 00:36:01.050 --> 00:36:03.630 align:middle line:84% And then you will see color fringes immediately. 00:36:03.630 --> 00:36:14.570 align:middle line:90% 00:36:14.570 --> 00:36:16.310 align:middle line:84% It's very strange very, very-- strange. 00:36:16.310 --> 00:36:18.270 align:middle line:84% Now the reason this is a very heavily strained. 00:36:18.270 --> 00:36:19.990 align:middle line:90% Remember how these are made? 00:36:19.990 --> 00:36:22.810 align:middle line:84% It's the same way as blow molding. 00:36:22.810 --> 00:36:24.210 align:middle line:90% That's a technical term. 00:36:24.210 --> 00:36:26.270 align:middle line:84% You have a [INAUDIBLE], and then you blow it. 00:36:26.270 --> 00:36:28.890 align:middle line:84% So there's very large plastic deformation takes place 00:36:28.890 --> 00:36:30.473 align:middle line:84% when making this kind of containers. 00:36:30.473 --> 00:36:32.390 align:middle line:84% So that's why there's lots of residual stress. 00:36:32.390 --> 00:36:35.610 align:middle line:84% And that's why you see a very strong photoelasticity there. 00:36:35.610 --> 00:36:38.210 align:middle line:90% This is for amorphous polymer. 00:36:38.210 --> 00:36:41.110 align:middle line:84% I also bring along with me another piece I got from Amazon. 00:36:41.110 --> 00:36:43.610 align:middle line:84% This is actually a glass [? butcher ?] ball made 00:36:43.610 --> 00:36:45.390 align:middle line:90% of a thermally tempered glass. 00:36:45.390 --> 00:36:47.610 align:middle line:84% I mean, it's a lot more uniform, so you don't see 00:36:47.610 --> 00:36:49.030 align:middle line:90% a lot of stress accumulation. 00:36:49.030 --> 00:36:50.930 align:middle line:84% However, I probably need someone to help me 00:36:50.930 --> 00:36:52.350 align:middle line:90% hold in this orientation here. 00:36:52.350 --> 00:36:55.050 align:middle line:90% 00:36:55.050 --> 00:36:56.850 align:middle line:90% Orthogonal direction. 00:36:56.850 --> 00:36:59.435 align:middle line:84% So in most regions, you don't see any stress there. 00:36:59.435 --> 00:37:00.310 align:middle line:90% I mean, a little bit. 00:37:00.310 --> 00:37:05.170 align:middle line:84% You see a little bit of color change right at the corner. 00:37:05.170 --> 00:37:06.790 align:middle line:84% Now you can rotate it by 90 degrees, 00:37:06.790 --> 00:37:08.030 align:middle line:84% you actually see a different pattern there, 00:37:08.030 --> 00:37:09.072 align:middle line:90% slightly different color. 00:37:09.072 --> 00:37:10.450 align:middle line:90% Rotate it. 00:37:10.450 --> 00:37:14.593 align:middle line:84% So this is where you actually get more stress accumulation, 00:37:14.593 --> 00:37:17.010 align:middle line:84% obviously, because of geometry, because the cooling is not 00:37:17.010 --> 00:37:18.292 align:middle line:90% very uniform at corners. 00:37:18.292 --> 00:37:20.250 align:middle line:84% So that's why you accumulate more stress there. 00:37:20.250 --> 00:37:21.770 align:middle line:84% So you can use this way actually to analyze 00:37:21.770 --> 00:37:23.970 align:middle line:84% any [? residual ?] stress, any transparent component 00:37:23.970 --> 00:37:25.270 align:middle line:90% by just buying that. 00:37:25.270 --> 00:37:27.870 align:middle line:84% That's like half a dollar polarizing film. 00:37:27.870 --> 00:37:29.770 align:middle line:90% And with a laptop, you do that. 00:37:29.770 --> 00:37:31.850 align:middle line:84% You can get polarized light, actually. 00:37:31.850 --> 00:37:33.070 align:middle line:90% AUDIENCE: [INAUDIBLE]. 00:37:33.070 --> 00:37:35.592 align:middle line:90% 00:37:35.592 --> 00:37:37.550 align:middle line:84% JUEJUN HU: You can get natural polarized light. 00:37:37.550 --> 00:37:39.467 align:middle line:84% Just look at reflected light on water surfaces 00:37:39.467 --> 00:37:40.970 align:middle line:90% and that kind of stuff. 00:37:40.970 --> 00:37:43.930 align:middle line:84% The reflected light in lots of cases is polarized. 00:37:43.930 --> 00:37:44.950 align:middle line:90% I won't go into why. 00:37:44.950 --> 00:37:46.380 align:middle line:84% If you're interested in learning, 00:37:46.380 --> 00:37:47.820 align:middle line:90% why, go to my 346 class. 00:37:47.820 --> 00:37:49.440 align:middle line:90% [CHUCKLES] 00:37:49.440 --> 00:37:54.400 align:middle line:84% OK, so that's basically how they actually quantify the strain 00:37:54.400 --> 00:37:56.000 align:middle line:90% level in Prince Rupert's drop. 00:37:56.000 --> 00:37:58.457 align:middle line:84% I will directly jump to the conclusion. 00:37:58.457 --> 00:38:00.040 align:middle line:84% This is a surface stress they measure. 00:38:00.040 --> 00:38:05.280 align:middle line:84% They measure surface stress up to 700 megapascals, 00:38:05.280 --> 00:38:08.800 align:middle line:84% which is 7,000 times atmospheric pressure-- 00:38:08.800 --> 00:38:10.493 align:middle line:90% enormous. 00:38:10.493 --> 00:38:13.160 align:middle line:84% And that's why-- now if you look at the [? converse ?] tool, how 00:38:13.160 --> 00:38:15.517 align:middle line:84% much compressive load this Prince Rupert's drop 00:38:15.517 --> 00:38:16.100 align:middle line:90% can withstand? 00:38:16.100 --> 00:38:17.160 align:middle line:90% A tiny little bit. 00:38:17.160 --> 00:38:22.480 align:middle line:84% It can withstand the compressed load exceeding 10,000 newtons. 00:38:22.480 --> 00:38:25.840 align:middle line:90% That's more than a ton. 00:38:25.840 --> 00:38:28.097 align:middle line:84% So in order to break that Prince Rupert's drop 00:38:28.097 --> 00:38:30.680 align:middle line:84% from the front side, you need to put a load of more than a ton 00:38:30.680 --> 00:38:31.520 align:middle line:90% to break it. 00:38:31.520 --> 00:38:35.760 align:middle line:84% That's why next time, if you see on YouTube, people will try to, 00:38:35.760 --> 00:38:37.260 align:middle line:84% they try to put all kinds of things. 00:38:37.260 --> 00:38:41.360 align:middle line:84% They try to smash it, they do whatever, it doesn't break. 00:38:41.360 --> 00:38:42.440 align:middle line:90% That's far from breaking. 00:38:42.440 --> 00:38:45.780 align:middle line:84% You need a ton of mass to actually really break it. 00:38:45.780 --> 00:38:48.420 align:middle line:84% So don't be too surprised if you see that kind of video. 00:38:48.420 --> 00:38:51.520 align:middle line:84% OK, so bottom line-- this is essentially compressive strength 00:38:51.520 --> 00:38:52.840 align:middle line:90% at its finest. 00:38:52.840 --> 00:38:56.600 align:middle line:84% You can really get to extremely, extremely strong glasses 00:38:56.600 --> 00:38:58.262 align:middle line:84% by just doing this kind of treatment. 00:38:58.262 --> 00:38:59.720 align:middle line:84% All right, now [? it comes ?] back, 00:38:59.720 --> 00:39:02.160 align:middle line:84% I have mentioned that there's a Achilles 00:39:02.160 --> 00:39:05.520 align:middle line:84% heel for Prince Rupert's drop because they 00:39:05.520 --> 00:39:07.680 align:middle line:90% break spontaneously. 00:39:07.680 --> 00:39:09.360 align:middle line:90% Here comes the problem. 00:39:09.360 --> 00:39:11.660 align:middle line:84% So if you have any defects inside the material, 00:39:11.660 --> 00:39:16.160 align:middle line:84% if you have any cracks in the tensile strain region, 00:39:16.160 --> 00:39:18.360 align:middle line:84% then those cracks can very easily propagate, 00:39:18.360 --> 00:39:20.520 align:middle line:84% and then lead to eventually catastrophic failure. 00:39:20.520 --> 00:39:23.320 align:middle line:84% Now unfortunately, there's one kind of defect 00:39:23.320 --> 00:39:25.620 align:middle line:84% that's almost universally present in all glasses. 00:39:25.620 --> 00:39:27.380 align:middle line:84% It's very difficult to eliminate. 00:39:27.380 --> 00:39:29.600 align:middle line:90% It's a nickel sulfide. 00:39:29.600 --> 00:39:32.640 align:middle line:84% Now, nickel sulfide, where it comes from, 00:39:32.640 --> 00:39:34.945 align:middle line:90% people are not totally clear. 00:39:34.945 --> 00:39:36.320 align:middle line:84% People sometimes add a little bit 00:39:36.320 --> 00:39:38.520 align:middle line:90% of sulfur in the glass batch. 00:39:38.520 --> 00:39:40.420 align:middle line:84% Sulfur is called the fining agent. 00:39:40.420 --> 00:39:42.580 align:middle line:84% It helps to reduce viscosity of glass melt, 00:39:42.580 --> 00:39:44.220 align:middle line:84% so it makes it a bit easier to process. 00:39:44.220 --> 00:39:45.595 align:middle line:84% So sometimes people intentionally 00:39:45.595 --> 00:39:47.280 align:middle line:84% add a little bit of sulfur there. 00:39:47.280 --> 00:39:50.680 align:middle line:90% But nickel is an impurity. 00:39:50.680 --> 00:39:52.280 align:middle line:90% Where does it come from? 00:39:52.280 --> 00:39:53.420 align:middle line:90% People don't really know. 00:39:53.420 --> 00:39:55.320 align:middle line:84% It may be coming from a raw ingredient. 00:39:55.320 --> 00:39:59.760 align:middle line:84% It may be coming from the stainless steel handling 00:39:59.760 --> 00:40:02.268 align:middle line:84% instruments, some powders getting there or whatever. 00:40:02.268 --> 00:40:04.060 align:middle line:84% But there's trace amounts of nickel anyway. 00:40:04.060 --> 00:40:07.600 align:middle line:84% So in the end, it forms this nickel sulfide 00:40:07.600 --> 00:40:10.000 align:middle line:90% in the form of inclusion. 00:40:10.000 --> 00:40:12.640 align:middle line:84% The reason nickel sulfide is so problematic is 00:40:12.640 --> 00:40:14.800 align:middle line:84% because when you look at this phase diagram here, 00:40:14.800 --> 00:40:16.840 align:middle line:90% it has two phases. 00:40:16.840 --> 00:40:19.540 align:middle line:84% It has a phase-- this is the high-temperature phase. 00:40:19.540 --> 00:40:22.040 align:middle line:84% This is the low-temperature phase. 00:40:22.040 --> 00:40:23.640 align:middle line:84% So when you are processing a glass, 00:40:23.640 --> 00:40:25.182 align:middle line:84% when you're doing a thermal tempering 00:40:25.182 --> 00:40:27.060 align:middle line:84% process, the temperature, if you recall, 00:40:27.060 --> 00:40:29.960 align:middle line:84% we heat up to about 600 C. And then start to cool down right. 00:40:29.960 --> 00:40:34.520 align:middle line:84% 600 C nickel sulfide is in a high temperature phase. 00:40:34.520 --> 00:40:37.200 align:middle line:84% Now, when you start to cool down, 00:40:37.200 --> 00:40:42.200 align:middle line:84% it now has a tendency to convert to the low-temperature phase. 00:40:42.200 --> 00:40:47.120 align:middle line:84% It's a slow process because it's usually kinetically inhibited. 00:40:47.120 --> 00:40:50.480 align:middle line:84% But once it converts into the low temperature phase, 00:40:50.480 --> 00:40:52.680 align:middle line:90% the volume increases. 00:40:52.680 --> 00:40:55.800 align:middle line:84% The two phases have a significant volume difference. 00:40:55.800 --> 00:40:57.505 align:middle line:84% Now, with increased volume, it's going 00:40:57.505 --> 00:40:58.880 align:middle line:84% to essentially create microcracks 00:40:58.880 --> 00:41:02.240 align:middle line:84% by squeezing on the glass matrix around it. 00:41:02.240 --> 00:41:05.440 align:middle line:84% These microcracks is where the problem happens. 00:41:05.440 --> 00:41:08.280 align:middle line:84% They can essentially nucleate much larger cracks, eventually 00:41:08.280 --> 00:41:10.280 align:middle line:84% leading to catastrophic failure, because they're 00:41:10.280 --> 00:41:11.400 align:middle line:84% located in the inside of the glass 00:41:11.400 --> 00:41:13.260 align:middle line:84% where nothing prevents it from opening. 00:41:13.260 --> 00:41:20.640 align:middle line:90% 00:41:20.640 --> 00:41:23.640 align:middle line:84% Now, you can try to reduce the incidence 00:41:23.640 --> 00:41:27.680 align:middle line:84% of this kind of breakage by having high purity glass 00:41:27.680 --> 00:41:29.480 align:middle line:90% ingredients and all that stuff. 00:41:29.480 --> 00:41:33.338 align:middle line:84% It's quite difficult. It's quite expensive as well. 00:41:33.338 --> 00:41:34.880 align:middle line:84% The more common way people are trying 00:41:34.880 --> 00:41:37.720 align:middle line:84% to eliminate this kind of failure 00:41:37.720 --> 00:41:40.402 align:middle line:84% is using a process called heat soaking. 00:41:40.402 --> 00:41:41.610 align:middle line:90% Heat soaking is quite simple. 00:41:41.610 --> 00:41:45.370 align:middle line:84% Essentially, you preheat all the thermally tempered glass pieces, 00:41:45.370 --> 00:41:48.947 align:middle line:84% try to accelerate the phase transition of a nickel sulfide 00:41:48.947 --> 00:41:50.530 align:middle line:84% to convert into low-temperature phase. 00:41:50.530 --> 00:41:52.950 align:middle line:84% So if there are any pieces that has this kind of inclusion 00:41:52.950 --> 00:41:56.810 align:middle line:84% there, they may break during the thermal heat soaking process. 00:41:56.810 --> 00:42:00.230 align:middle line:84% And therefore, you can only pick up the good ones. 00:42:00.230 --> 00:42:01.870 align:middle line:84% Unfortunately, that process does not 00:42:01.870 --> 00:42:06.190 align:middle line:90% eliminate all the faulty parts. 00:42:06.190 --> 00:42:10.150 align:middle line:84% So that's why all the thermally tempered glass products, 00:42:10.150 --> 00:42:12.390 align:middle line:84% whether it's the door on the microwave oven 00:42:12.390 --> 00:42:15.250 align:middle line:84% or door in your shower room or whatever, 00:42:15.250 --> 00:42:19.110 align:middle line:84% architectural glasses all have the risk of spontaneously 00:42:19.110 --> 00:42:19.610 align:middle line:90% breaking. 00:42:19.610 --> 00:42:23.110 align:middle line:84% Even though the probability is very, very low, it does happen. 00:42:23.110 --> 00:42:24.710 align:middle line:84% OK, so don't be too surprised if you 00:42:24.710 --> 00:42:28.190 align:middle line:84% see some glass pieces suddenly break into small pieces. 00:42:28.190 --> 00:42:28.945 align:middle line:90% What do you do? 00:42:28.945 --> 00:42:30.570 align:middle line:84% Remember what I said in the last class. 00:42:30.570 --> 00:42:33.110 align:middle line:90% Do your fractography. 00:42:33.110 --> 00:42:35.408 align:middle line:84% So this is an example of a nickel sulfide inclusion 00:42:35.408 --> 00:42:37.950 align:middle line:84% Well, in fact, you can actually really do it for photography. 00:42:37.950 --> 00:42:39.150 align:middle line:90% So what you would do? 00:42:39.150 --> 00:42:42.310 align:middle line:84% Typically, if you have some glass broken 00:42:42.310 --> 00:42:44.390 align:middle line:84% due to this kind of nickel sulfide intrusion, 00:42:44.390 --> 00:42:48.790 align:middle line:84% the breaking pattern has this kind of butterfly geometry. 00:42:48.790 --> 00:42:50.730 align:middle line:84% It looks like a butterfly in the center here. 00:42:50.730 --> 00:42:52.230 align:middle line:84% The reason it looks like a butterfly 00:42:52.230 --> 00:42:54.050 align:middle line:84% is because-- we can zoom in a little bit. 00:42:54.050 --> 00:42:55.750 align:middle line:90% This is actually what happens. 00:42:55.750 --> 00:43:01.710 align:middle line:84% Right at this point, right here, the intersection 00:43:01.710 --> 00:43:04.390 align:middle line:84% of several of the radial cracks, this 00:43:04.390 --> 00:43:06.250 align:middle line:90% is where the defect initiates. 00:43:06.250 --> 00:43:09.990 align:middle line:84% If you look carefully, most likely you see a nickel sulfide 00:43:09.990 --> 00:43:11.830 align:middle line:90% defect right there 00:43:11.830 --> 00:43:13.790 align:middle line:84% Now, the reason you have this butterfly pattern 00:43:13.790 --> 00:43:15.970 align:middle line:84% is because when the crack forms, initially, 00:43:15.970 --> 00:43:19.610 align:middle line:84% it propagates radially outwards from the point of defects. 00:43:19.610 --> 00:43:22.830 align:middle line:84% And then because you have a lot of compressive strain 00:43:22.830 --> 00:43:26.830 align:middle line:84% in a thermally tempered glass, this compressed strain 00:43:26.830 --> 00:43:29.030 align:middle line:84% means that there's lots of elastic energy. 00:43:29.030 --> 00:43:32.810 align:middle line:84% This elastic energy has to be released during fracture. 00:43:32.810 --> 00:43:36.450 align:middle line:84% And they get released in the form of surface energy. 00:43:36.450 --> 00:43:38.267 align:middle line:84% That's why thermally tempered glass breaks 00:43:38.267 --> 00:43:40.350 align:middle line:84% in such a different way compared to conventionally 00:43:40.350 --> 00:43:41.790 align:middle line:90% annealed glasses. 00:43:41.790 --> 00:43:44.750 align:middle line:84% It doesn't break in large shards. 00:43:44.750 --> 00:43:48.230 align:middle line:84% So as the crack propagates, the excess elastic energy 00:43:48.230 --> 00:43:50.350 align:middle line:84% is released in the form of essentially this kind 00:43:50.350 --> 00:43:54.870 align:middle line:84% of secondary cracks that go in this direction, 00:43:54.870 --> 00:43:56.473 align:middle line:90% in the orthogonal direction. 00:43:56.473 --> 00:43:58.890 align:middle line:84% So because of the interaction of these two sets of cracks, 00:43:58.890 --> 00:44:01.550 align:middle line:84% that's why you tend to form this kind of butterfly-like pattern 00:44:01.550 --> 00:44:02.050 align:middle line:90% there. 00:44:02.050 --> 00:44:04.610 align:middle line:84% But this is where the defect originates. 00:44:04.610 --> 00:44:07.990 align:middle line:90% 00:44:07.990 --> 00:44:10.590 align:middle line:90% All right, makes sense? 00:44:10.590 --> 00:44:13.042 align:middle line:84% OK, I forgot to do my fractography analysis when 00:44:13.042 --> 00:44:14.250 align:middle line:90% I saw my microwave door open. 00:44:14.250 --> 00:44:15.542 align:middle line:90% I really should have done that. 00:44:15.542 --> 00:44:16.310 align:middle line:90% [CHUCKLES] 00:44:16.310 --> 00:44:22.350 align:middle line:84% Anyway, so it's exactly because of this issue here 00:44:22.350 --> 00:44:24.790 align:middle line:84% of spontaneous breakage when they come up 00:44:24.790 --> 00:44:27.590 align:middle line:84% with some applications, like consumer electronics, 00:44:27.590 --> 00:44:29.158 align:middle line:90% cover glasses for your phones. 00:44:29.158 --> 00:44:30.950 align:middle line:84% People want to come up with a different way 00:44:30.950 --> 00:44:32.050 align:middle line:90% to do strengthening. 00:44:32.050 --> 00:44:35.000 align:middle line:84% That's where ion exchange or chemical strengthening come in. 00:44:35.000 --> 00:44:36.900 align:middle line:84% So I've already basically mentioned 00:44:36.900 --> 00:44:38.925 align:middle line:90% about this simple process here. 00:44:38.925 --> 00:44:41.300 align:middle line:84% The most common way of doing it is you immerse your glass 00:44:41.300 --> 00:44:44.660 align:middle line:90% products in a molten salt bath. 00:44:44.660 --> 00:44:46.620 align:middle line:84% So if the glass contains sodium, you 00:44:46.620 --> 00:44:48.940 align:middle line:90% put it in a potassium salt bath. 00:44:48.940 --> 00:44:50.960 align:middle line:84% Also, if you cannot contain lithium, 00:44:50.960 --> 00:44:52.900 align:middle line:84% you can put it in a sodium salt bath. 00:44:52.900 --> 00:44:54.940 align:middle line:84% People also use mixed salt baths. 00:44:54.940 --> 00:44:56.820 align:middle line:84% Lithium glass can be simultaneously 00:44:56.820 --> 00:44:59.020 align:middle line:84% [? ion ?] exchanged in a mixed salt baths of sodium 00:44:59.020 --> 00:45:00.460 align:middle line:90% and potassium, for example. 00:45:00.460 --> 00:45:03.988 align:middle line:84% So then through interdiffusion process, 00:45:03.988 --> 00:45:05.780 align:middle line:84% you build up the surface compressive strain 00:45:05.780 --> 00:45:07.863 align:middle line:84% in the region where you have [INAUDIBLE] diffusion 00:45:07.863 --> 00:45:09.380 align:middle line:90% of the ions. 00:45:09.380 --> 00:45:14.580 align:middle line:84% The kinetics is described basically by interdiffusion. 00:45:14.580 --> 00:45:20.740 align:middle line:84% So what I'm listing here is the interdiffusion coefficient. 00:45:20.740 --> 00:45:30.380 align:middle line:84% Now, for those of you who have taken 3.033, kinetics, or 3.030. 00:45:30.380 --> 00:45:35.300 align:middle line:84% What's the name of the equation for interdiffusion coefficient? 00:45:35.300 --> 00:45:36.960 align:middle line:90% Pop quiz. 00:45:36.960 --> 00:45:43.594 align:middle line:84% AUDIENCE: [INAUDIBLE] [? Einstein. ?] No. 00:45:43.594 --> 00:45:44.867 align:middle line:90% [INAUDIBLE] 00:45:44.867 --> 00:45:46.200 align:middle line:90% JUEJUN HU: [INAUDIBLE] equation. 00:45:46.200 --> 00:45:48.340 align:middle line:84% OK, this is not [INAUDIBLE] equation. 00:45:48.340 --> 00:45:50.040 align:middle line:90% [CHUCKLES] It doesn't matter. 00:45:50.040 --> 00:45:50.940 align:middle line:90% This is called the-- 00:45:50.940 --> 00:45:53.106 align:middle line:90% AUDIENCE: [INAUDIBLE] Einstein. 00:45:53.106 --> 00:45:54.520 align:middle line:90% JUEJUN HU: Nernst-Planck. 00:45:54.520 --> 00:45:55.838 align:middle line:90% Close. 00:45:55.838 --> 00:45:56.880 align:middle line:90% It's about the same time. 00:45:56.880 --> 00:45:58.080 align:middle line:84% They all lived [? within the ?] same time. 00:45:58.080 --> 00:46:00.540 align:middle line:84% But we don't credit everything to Einstein, OK? 00:46:00.540 --> 00:46:01.505 align:middle line:90% [CHUCKLES] 00:46:01.505 --> 00:46:05.020 align:middle line:90% AUDIENCE: [INAUDIBLE] 00:46:05.020 --> 00:46:07.140 align:middle line:84% JUEJUN HU: So now the reason the coefficient 00:46:07.140 --> 00:46:08.620 align:middle line:84% is a little bit different from [INAUDIBLE] equation 00:46:08.620 --> 00:46:10.940 align:middle line:84% is because [INAUDIBLE] equation describes essentially 00:46:10.940 --> 00:46:13.880 align:middle line:84% charged neutral particles, like atoms when they interdiffuse. 00:46:13.880 --> 00:46:15.220 align:middle line:90% That's what you use. 00:46:15.220 --> 00:46:17.420 align:middle line:84% For ions, because they actually carry 00:46:17.420 --> 00:46:21.180 align:middle line:84% charge, what that means is [INAUDIBLE] ions move, 00:46:21.180 --> 00:46:24.100 align:middle line:84% the charge distribution also modifies. 00:46:24.100 --> 00:46:26.380 align:middle line:84% And when you change charge distribution, 00:46:26.380 --> 00:46:28.340 align:middle line:84% you change internal electric fields. 00:46:28.340 --> 00:46:31.060 align:middle line:84% And the electric field is going to modify again 00:46:31.060 --> 00:46:32.320 align:middle line:90% the motion of the ions. 00:46:32.320 --> 00:46:35.280 align:middle line:84% So they are inter-linked together. 00:46:35.280 --> 00:46:37.900 align:middle line:84% So that's why for any interdiffusion process 00:46:37.900 --> 00:46:40.890 align:middle line:84% in a charged species, you should not use [INAUDIBLE] equation. 00:46:40.890 --> 00:46:43.140 align:middle line:84% But rather, you should use the Nernst-Planck equation. 00:46:43.140 --> 00:46:44.557 align:middle line:84% That's what I'm [INAUDIBLE] there. 00:46:44.557 --> 00:46:50.750 align:middle line:90% 00:46:50.750 --> 00:46:52.500 align:middle line:84% Now, what I want to mention is that if you 00:46:52.500 --> 00:46:54.920 align:middle line:84% look at the Nernst-Planck equation, 00:46:54.920 --> 00:46:56.580 align:middle line:90% let's see what actually happens. 00:46:56.580 --> 00:46:58.860 align:middle line:84% So suppose you look at the glass surface 00:46:58.860 --> 00:47:01.540 align:middle line:84% where you have about the N is essentially the amount. 00:47:01.540 --> 00:47:04.640 align:middle line:84% N is the molar fraction of the ions. 00:47:04.640 --> 00:47:07.920 align:middle line:84% D is the diffusion coefficient of each of the ions. 00:47:07.920 --> 00:47:11.780 align:middle line:84% So if you have about the same amount of sodium and potassium 00:47:11.780 --> 00:47:16.260 align:middle line:84% ions, then if you look at the denominator here, 00:47:16.260 --> 00:47:19.202 align:middle line:84% we know that D, the diffusion coefficient of sodium, 00:47:19.202 --> 00:47:20.660 align:middle line:84% should be larger than the diffusion 00:47:20.660 --> 00:47:21.702 align:middle line:90% coefficient of potassium. 00:47:21.702 --> 00:47:24.320 align:middle line:84% Because they're smaller, it's a lot easier to squeeze around. 00:47:24.320 --> 00:47:26.600 align:middle line:84% So that's why you expect D sodium to be larger. 00:47:26.600 --> 00:47:32.500 align:middle line:84% So therefore, the denominator, this term should be dominant. 00:47:32.500 --> 00:47:34.620 align:middle line:84% So if this term is dominant, you can cancel out 00:47:34.620 --> 00:47:37.900 align:middle line:84% this DNa, the diffusion coefficient of sodium 00:47:37.900 --> 00:47:39.940 align:middle line:84% in both numerator and denominator. 00:47:39.940 --> 00:47:42.660 align:middle line:84% What is left behind basically becomes the diffusion 00:47:42.660 --> 00:47:44.940 align:middle line:90% coefficient of potassium. 00:47:44.940 --> 00:47:46.040 align:middle line:90% This also makes sense. 00:47:46.040 --> 00:47:46.880 align:middle line:90% What is this? 00:47:46.880 --> 00:47:49.980 align:middle line:84% In the interdiffusion process, the overall kinetics 00:47:49.980 --> 00:47:53.800 align:middle line:84% is basically dictated by the slower diffusion species, 00:47:53.800 --> 00:47:56.604 align:middle line:90% which is the bigger ion. 00:47:56.604 --> 00:47:59.980 align:middle line:84% Interdiffusion process, the diffusion coefficient more 00:47:59.980 --> 00:48:02.940 align:middle line:84% or less equals to a [? decay, ?] the diffusion coefficient 00:48:02.940 --> 00:48:05.140 align:middle line:84% of potassium or the heavy ion species. 00:48:05.140 --> 00:48:09.158 align:middle line:90% 00:48:09.158 --> 00:48:11.700 align:middle line:84% All right, so now if you notice a diffusion coefficient, what 00:48:11.700 --> 00:48:14.760 align:middle line:84% you can also do again goes back to your kinetics class. 00:48:14.760 --> 00:48:20.620 align:middle line:84% You can use it to now estimate what is the diffusion length. 00:48:20.620 --> 00:48:22.740 align:middle line:84% Diffusion length is given basically a square root 00:48:22.740 --> 00:48:24.770 align:middle line:84% of diffusion coefficient and tau. 00:48:24.770 --> 00:48:28.970 align:middle line:84% Tau is basically your time of the diffusion process. 00:48:28.970 --> 00:48:32.130 align:middle line:84% So the longer you diffuse, the longer 00:48:32.130 --> 00:48:34.970 align:middle line:84% you go through the process, the thicker the surface compressive 00:48:34.970 --> 00:48:35.710 align:middle line:90% layer gets. 00:48:35.710 --> 00:48:38.870 align:middle line:84% And that diffusion length has a name. 00:48:38.870 --> 00:48:41.290 align:middle line:84% It's called K steps in the context 00:48:41.290 --> 00:48:43.330 align:middle line:84% of a chemically strengthened glass. 00:48:43.330 --> 00:48:46.810 align:middle line:84% So K steps is basically the thickness of the surface 00:48:46.810 --> 00:48:49.290 align:middle line:90% compressive layer. 00:48:49.290 --> 00:48:51.250 align:middle line:84% For soda-lime glass, it's typically, 00:48:51.250 --> 00:48:53.330 align:middle line:90% as I said, about 30 microns-ish. 00:48:53.330 --> 00:48:55.810 align:middle line:84% People try to modify it by changing glass chemistry, which 00:48:55.810 --> 00:48:57.290 align:middle line:84% I'll talk about in the next slide. 00:48:57.290 --> 00:48:59.450 align:middle line:84% You can increase the K steps to close to maybe, 00:48:59.450 --> 00:49:02.930 align:middle line:84% like, 100 microns in a lot of consumer electronic products. 00:49:02.930 --> 00:49:04.570 align:middle line:84% In extreme cases, they can even reach 00:49:04.570 --> 00:49:07.610 align:middle line:84% something like about a millimeter for things 00:49:07.610 --> 00:49:11.650 align:middle line:84% like glass windows used in military vehicles. 00:49:11.650 --> 00:49:15.233 align:middle line:84% They need extremely high compressive stress. 00:49:15.233 --> 00:49:16.650 align:middle line:84% Now, there's one thing I also want 00:49:16.650 --> 00:49:18.470 align:middle line:84% to mention about this process here, 00:49:18.470 --> 00:49:20.865 align:middle line:84% which is different from thermal tempering process. 00:49:20.865 --> 00:49:22.490 align:middle line:84% Remember for thermal tempering process, 00:49:22.490 --> 00:49:25.067 align:middle line:84% you heat it up to above the glass transition point 00:49:25.067 --> 00:49:27.650 align:middle line:84% because you want all the stress to relax at high temperatures. 00:49:27.650 --> 00:49:30.010 align:middle line:84% And you cool down to build up the stress gradient. 00:49:30.010 --> 00:49:32.070 align:middle line:84% In the chemical toughening process, 00:49:32.070 --> 00:49:35.490 align:middle line:90% you never, never exceed Tg. 00:49:35.490 --> 00:49:36.170 align:middle line:90% Why? 00:49:36.170 --> 00:49:38.250 align:middle line:84% Above Tg stress, you're going to get relaxed. 00:49:38.250 --> 00:49:41.050 align:middle line:84% So if you do your ion diffusion above Tg, 00:49:41.050 --> 00:49:42.650 align:middle line:84% then while you're actually diffusing, 00:49:42.650 --> 00:49:44.790 align:middle line:84% the stress is also actively getting relaxed. 00:49:44.790 --> 00:49:46.530 align:middle line:90% You get nothing. 00:49:46.530 --> 00:49:48.090 align:middle line:84% So that's the main difference there. 00:49:48.090 --> 00:49:51.610 align:middle line:84% Chemical strengthening has to be performed below Tg. 00:49:51.610 --> 00:49:54.530 align:middle line:84% So you retain the stress throughout the process. 00:49:54.530 --> 00:49:59.290 align:middle line:84% Thermal tempering has to go above Tg. 00:49:59.290 --> 00:50:03.210 align:middle line:84% Now, we know that diffusion is a thermally activated process. 00:50:03.210 --> 00:50:05.530 align:middle line:84% So higher the temperature, the faster it diffuses. 00:50:05.530 --> 00:50:07.530 align:middle line:84% The faster it diffuses, the larger the K steps, 00:50:07.530 --> 00:50:09.490 align:middle line:84% the stronger the glass product gets. 00:50:09.490 --> 00:50:11.690 align:middle line:84% So that's actually motivation for people 00:50:11.690 --> 00:50:14.850 align:middle line:84% to raise the temperature during the ion exchange process 00:50:14.850 --> 00:50:17.010 align:middle line:90% just to expedite diffusion. 00:50:17.010 --> 00:50:18.870 align:middle line:90% Now, Tg sets a limit. 00:50:18.870 --> 00:50:20.810 align:middle line:90% You cannot really go above Tg. 00:50:20.810 --> 00:50:23.490 align:middle line:84% So if you want to get the more efficient chemical strengthening 00:50:23.490 --> 00:50:25.370 align:middle line:84% process for ion exchange, you need 00:50:25.370 --> 00:50:28.930 align:middle line:84% to find glass materials with higher Tg, which 00:50:28.930 --> 00:50:31.610 align:middle line:84% means that you can afford doing ion exchange at relatively 00:50:31.610 --> 00:50:35.570 align:middle line:84% high temperatures to expedite the kinetics. 00:50:35.570 --> 00:50:38.050 align:middle line:84% So we're going to talk in the next slide 00:50:38.050 --> 00:50:40.850 align:middle line:84% about some of the strategies, how to essentially boost 00:50:40.850 --> 00:50:43.625 align:middle line:90% the ion diffusion kinetics. 00:50:43.625 --> 00:50:45.250 align:middle line:84% But before that, I just want to mention 00:50:45.250 --> 00:50:46.370 align:middle line:90% a couple of quick things. 00:50:46.370 --> 00:50:48.410 align:middle line:84% One is that in this kind of simple model, 00:50:48.410 --> 00:50:50.670 align:middle line:84% we assume the diffusion coefficient is the constant. 00:50:50.670 --> 00:50:51.443 align:middle line:90% In fact, it's not. 00:50:51.443 --> 00:50:53.110 align:middle line:84% It also depends on compressive strength. 00:50:53.110 --> 00:50:54.330 align:middle line:90% It's quite intuitive. 00:50:54.330 --> 00:50:57.810 align:middle line:84% If you have compressive strain, you actually impede diffusion. 00:50:57.810 --> 00:51:01.490 align:middle line:84% You can squeeze on ions, slow down the diffusion kinetics, 00:51:01.490 --> 00:51:03.338 align:middle line:90% which makes sense. 00:51:03.338 --> 00:51:05.130 align:middle line:84% The other thing I also want to mention here 00:51:05.130 --> 00:51:06.530 align:middle line:84% is about the advantages of this process, why 00:51:06.530 --> 00:51:08.238 align:middle line:84% people want to use chemical strengthening 00:51:08.238 --> 00:51:09.425 align:middle line:90% versus thermal tempering. 00:51:09.425 --> 00:51:11.550 align:middle line:84% The first one, obviously, is that they don't break. 00:51:11.550 --> 00:51:14.570 align:middle line:84% They don't explode spontaneously, number one, 00:51:14.570 --> 00:51:17.370 align:middle line:84% because of the much lower level of tensile strength 00:51:17.370 --> 00:51:19.400 align:middle line:84% inside the glass material-- first thing. 00:51:19.400 --> 00:51:23.760 align:middle line:84% The second thing is that because of the surface compressed layer 00:51:23.760 --> 00:51:26.560 align:middle line:84% of [INAUDIBLE] thing, any kind of photoelastic changes 00:51:26.560 --> 00:51:28.920 align:middle line:84% or birefringence effect is a lot weaker 00:51:28.920 --> 00:51:30.320 align:middle line:90% because it's such a thin layer. 00:51:30.320 --> 00:51:32.960 align:middle line:84% So in some kinds of high [? energy ?] applications, 00:51:32.960 --> 00:51:35.960 align:middle line:84% these actually lead to less distortion of the images 00:51:35.960 --> 00:51:38.320 align:middle line:84% or less essentially better optical quality. 00:51:38.320 --> 00:51:40.120 align:middle line:84% So that's one consideration there. 00:51:40.120 --> 00:51:44.760 align:middle line:84% The last thing I also want to mention is that for glass pieces 00:51:44.760 --> 00:51:47.780 align:middle line:84% that have strange geometries, not like a flat plane-- 00:51:47.780 --> 00:51:50.160 align:middle line:84% you may have glass containers, for example, 00:51:50.160 --> 00:51:53.440 align:middle line:84% if you want to strengthen those kinds of irregular shaped glass 00:51:53.440 --> 00:51:56.400 align:middle line:84% products, people usually don't use thermal tempering 00:51:56.400 --> 00:51:58.400 align:middle line:84% because this irregular geometry means 00:51:58.400 --> 00:51:59.740 align:middle line:90% that cooling is not uniform. 00:51:59.740 --> 00:52:01.440 align:middle line:84% You're going to build up lots of internal stress, 00:52:01.440 --> 00:52:02.400 align:middle line:90% like what I've shown there. 00:52:02.400 --> 00:52:04.260 align:middle line:84% The corner is actually where the stress [? gets ?] built up. 00:52:04.260 --> 00:52:05.820 align:middle line:84% So if you an irregular shape, you can imagine, 00:52:05.820 --> 00:52:07.237 align:middle line:84% there is a lot of residual stress, 00:52:07.237 --> 00:52:09.300 align:middle line:90% which is harmful for the glass. 00:52:09.300 --> 00:52:11.800 align:middle line:84% So for strengthening of, for example, bio bottles 00:52:11.800 --> 00:52:14.400 align:middle line:84% that people use in the pharmaceutical industry, 00:52:14.400 --> 00:52:16.040 align:middle line:84% chemical strengthening is what they do. 00:52:16.040 --> 00:52:18.603 align:middle line:84% They immerse the whole thing into the chemical salt bath. 00:52:18.603 --> 00:52:20.520 align:middle line:84% The surface is going to get uniformly diffused 00:52:20.520 --> 00:52:22.920 align:middle line:90% with the ionic species. 00:52:22.920 --> 00:52:24.543 align:middle line:84% So that's why, for these reasons, 00:52:24.543 --> 00:52:25.960 align:middle line:84% people use chemical strengthening. 00:52:25.960 --> 00:52:27.820 align:middle line:90% Yes. 00:52:27.820 --> 00:52:30.080 align:middle line:84% AUDIENCE: Are there-- is thickness 00:52:30.080 --> 00:52:33.205 align:middle line:84% of the glass a constraint on thermal tempering as well? 00:52:33.205 --> 00:52:34.080 align:middle line:90% JUEJUN HU: Of course. 00:52:34.080 --> 00:52:35.120 align:middle line:90% Yeah, I mean, the case that-- 00:52:35.120 --> 00:52:37.500 align:middle line:84% AUDIENCE: If it's too thin like for your phone or something? 00:52:37.500 --> 00:52:38.960 align:middle line:84% JUEJUN HU: Yeah, then you also build up 00:52:38.960 --> 00:52:40.060 align:middle line:90% a lot of tensile stress. 00:52:40.060 --> 00:52:42.268 align:middle line:84% Yeah, and it's quite apparent that the case depth has 00:52:42.268 --> 00:52:45.040 align:middle line:84% to be considerably smaller than half of the glass [INAUDIBLE] 00:52:45.040 --> 00:52:46.560 align:middle line:84% thickness if you want to eliminate 00:52:46.560 --> 00:52:49.100 align:middle line:84% the risk of loss of tensile strength, so same idea. 00:52:49.100 --> 00:52:51.040 align:middle line:90% Yeah 00:52:51.040 --> 00:52:53.377 align:middle line:84% And by the way, the other reason people 00:52:53.377 --> 00:52:54.960 align:middle line:84% don't use chemical strengthening also, 00:52:54.960 --> 00:52:56.460 align:middle line:90% it's a more expensive process. 00:52:56.460 --> 00:52:58.960 align:middle line:84% [INAUDIBLE] temperature is really simple [INAUDIBLE]. 00:52:58.960 --> 00:52:59.760 align:middle line:90% That's why. 00:52:59.760 --> 00:53:02.920 align:middle line:84% So in architectures-- for example, in my microwave oven, 00:53:02.920 --> 00:53:05.160 align:middle line:84% that's why they use thermal tempering, not 00:53:05.160 --> 00:53:07.280 align:middle line:90% chemical tempering. 00:53:07.280 --> 00:53:09.817 align:middle line:84% Now, one last thing I also want to mention here is-- 00:53:09.817 --> 00:53:10.900 align:middle line:90% you see this bullet point. 00:53:10.900 --> 00:53:14.920 align:middle line:84% It says that only applies to alkaline earth glasses, which 00:53:14.920 --> 00:53:16.920 align:middle line:90% makes sense. 00:53:16.920 --> 00:53:19.840 align:middle line:84% For alkaline, that's what people use most commonly 00:53:19.840 --> 00:53:21.400 align:middle line:90% across the industry. 00:53:21.400 --> 00:53:24.495 align:middle line:84% For example, with sodium, you use potassium ion exchange. 00:53:24.495 --> 00:53:25.620 align:middle line:90% That's the most common one. 00:53:25.620 --> 00:53:28.280 align:middle line:84% Lithium, you can use sodium or you can use potassium. 00:53:28.280 --> 00:53:31.800 align:middle line:84% Alkaline earth is really actually a new process. 00:53:31.800 --> 00:53:35.920 align:middle line:84% It only was developed less than a decade ago. 00:53:35.920 --> 00:53:39.000 align:middle line:84% And the motivation behind it is that actually, 00:53:39.000 --> 00:53:41.860 align:middle line:84% for some high-end applications-- especially in electronics, 00:53:41.860 --> 00:53:45.868 align:middle line:84% people want glass formulations that's free of alkaline ions. 00:53:45.868 --> 00:53:48.160 align:middle line:84% The reason people don't like alkaline ions is that they 00:53:48.160 --> 00:53:49.380 align:middle line:90% are actually very mobile. 00:53:49.380 --> 00:53:51.100 align:middle line:84% They tend to diffuse very, very fast. 00:53:51.100 --> 00:53:53.588 align:middle line:84% And if you have glass containing sodium, 00:53:53.588 --> 00:53:55.380 align:middle line:84% the sodium can easily diffuse, for example, 00:53:55.380 --> 00:53:56.922 align:middle line:84% into amorphous silicon, which is what 00:53:56.922 --> 00:54:00.240 align:middle line:84% you use to make transistors, and eventually, essentially 00:54:00.240 --> 00:54:03.660 align:middle line:84% harm the electronic properties of semiconductor materials. 00:54:03.660 --> 00:54:07.572 align:middle line:84% So for those applications, people want alkaline-free glass. 00:54:07.572 --> 00:54:09.280 align:middle line:84% But they still want to chemically toughen 00:54:09.280 --> 00:54:11.280 align:middle line:84% it to get better mechanical properties. 00:54:11.280 --> 00:54:14.710 align:middle line:84% So that's why this is where the alkaline earth based 00:54:14.710 --> 00:54:16.990 align:middle line:84% ion exchange process comes into play. 00:54:16.990 --> 00:54:20.270 align:middle line:84% Now, the process was only developed less than 10 years 00:54:20.270 --> 00:54:24.470 align:middle line:84% ago, basically, because the alkaline earth sources tend 00:54:24.470 --> 00:54:26.652 align:middle line:90% to have high melting points. 00:54:26.652 --> 00:54:27.610 align:middle line:90% They tend to decompose. 00:54:27.610 --> 00:54:29.450 align:middle line:84% They also can be highly corrosive, whatever. 00:54:29.450 --> 00:54:31.510 align:middle line:84% But the problem is this high melting point. 00:54:31.510 --> 00:54:36.990 align:middle line:84% Their melting point is higher than the Tg of most glasses. 00:54:36.990 --> 00:54:38.470 align:middle line:84% So now you see what the problem is. 00:54:38.470 --> 00:54:40.250 align:middle line:84% If you try to do a molten salt bath, 00:54:40.250 --> 00:54:42.250 align:middle line:84% the strain will be relaxed all the time. 00:54:42.250 --> 00:54:44.550 align:middle line:84% You just cannot accumulate stress. 00:54:44.550 --> 00:54:49.430 align:middle line:84% So that's why people tend to aqueous solutions 00:54:49.430 --> 00:54:52.470 align:middle line:90% of alkaline earth ions. 00:54:52.470 --> 00:54:54.470 align:middle line:84% But for aqueous solution, normally, 00:54:54.470 --> 00:54:56.710 align:middle line:84% you cannot get to the temperatures needed to expedite 00:54:56.710 --> 00:54:58.410 align:middle line:90% the diffusion process. 00:54:58.410 --> 00:55:02.510 align:middle line:84% So what they will do is put it in basically a pressure cooker. 00:55:02.510 --> 00:55:05.790 align:middle line:84% By applying a large pressure, 45 megapascals, 00:55:05.790 --> 00:55:11.170 align:middle line:84% you can essentially raise the boiling point of the liquid. 00:55:11.170 --> 00:55:14.863 align:middle line:84% In fact, when they do it, they do it in a supercritical fluid. 00:55:14.863 --> 00:55:17.030 align:middle line:84% So essentially, that's a state where you cannot tell 00:55:17.030 --> 00:55:17.715 align:middle line:90% the difference. 00:55:17.715 --> 00:55:19.590 align:middle line:84% There's no clear difference between gas phase 00:55:19.590 --> 00:55:20.630 align:middle line:90% and liquid phase. 00:55:20.630 --> 00:55:22.510 align:middle line:90% There's one simple fluid phase. 00:55:22.510 --> 00:55:25.310 align:middle line:84% So at that high temperature, still, it 00:55:25.310 --> 00:55:29.070 align:middle line:84% takes like hours in the aqueous salt bath 00:55:29.070 --> 00:55:31.150 align:middle line:84% to get the diffusion [? there. ?] 00:55:31.150 --> 00:55:31.950 align:middle line:90% So that's why. 00:55:31.950 --> 00:55:33.190 align:middle line:90% That answers your question. 00:55:33.190 --> 00:55:35.610 align:middle line:84% If you just do it in the thermal tempering process, 00:55:35.610 --> 00:55:36.670 align:middle line:84% there's very little salt diffusion 00:55:36.670 --> 00:55:38.295 align:middle line:84% because there's simply not enough time. 00:55:38.295 --> 00:55:40.512 align:middle line:84% Now, the other interesting finding also, 00:55:40.512 --> 00:55:41.970 align:middle line:84% if you look at this particular one, 00:55:41.970 --> 00:55:45.150 align:middle line:84% they are using barium ions to replace calcium ions 00:55:45.150 --> 00:55:48.270 align:middle line:84% in the glass to build the compressive stress. 00:55:48.270 --> 00:55:51.350 align:middle line:84% When the barium ion diffuses, it not only replaces the calcium 00:55:51.350 --> 00:55:54.550 align:middle line:84% ion, it also displaces some of the boron 00:55:54.550 --> 00:55:59.310 align:middle line:84% ions, which is normally not what you would expect. 00:55:59.310 --> 00:56:01.290 align:middle line:84% But this is a borosilicate glass, by the way. 00:56:01.290 --> 00:56:03.550 align:middle line:84% So anyway, there's some interesting materials studies 00:56:03.550 --> 00:56:04.883 align:middle line:90% if you want to see more details. 00:56:04.883 --> 00:56:06.390 align:middle line:84% This is actually the seminal paper 00:56:06.390 --> 00:56:08.307 align:middle line:84% that actually describes this process developed 00:56:08.307 --> 00:56:10.050 align:middle line:90% for alkaline earth ion exchange. 00:56:10.050 --> 00:56:13.190 align:middle line:90% 00:56:13.190 --> 00:56:18.350 align:middle line:84% OK, so now, next slide, I'm trying 00:56:18.350 --> 00:56:23.470 align:middle line:84% to summarize how we actually increase the K steps. 00:56:23.470 --> 00:56:25.270 align:middle line:84% Increasing K steps is actually important 00:56:25.270 --> 00:56:26.830 align:middle line:84% because increased K steps essentially 00:56:26.830 --> 00:56:29.150 align:middle line:84% means that you have a thicker surface layer that's 00:56:29.150 --> 00:56:30.950 align:middle line:90% under compressive strain right. 00:56:30.950 --> 00:56:33.550 align:middle line:84% This usually also gives rise to higher surface compressive 00:56:33.550 --> 00:56:35.550 align:middle line:84% strain, and therefore stronger glass products 00:56:35.550 --> 00:56:36.470 align:middle line:90% after the process. 00:56:36.470 --> 00:56:40.430 align:middle line:90% 00:56:40.430 --> 00:56:43.410 align:middle line:84% So increasing the K steps is really a kinetic problem. 00:56:43.410 --> 00:56:45.630 align:middle line:84% You're trying to expedite the diffusion kinetics. 00:56:45.630 --> 00:56:48.270 align:middle line:90% 00:56:48.270 --> 00:56:49.950 align:middle line:90% So some common ways-- 00:56:49.950 --> 00:56:53.090 align:middle line:84% you can essentially apply additional external stimuli. 00:56:53.090 --> 00:56:54.530 align:middle line:90% You can use electric fields. 00:56:54.530 --> 00:56:56.450 align:middle line:90% You can use ultrasonication. 00:56:56.450 --> 00:56:57.710 align:middle line:90% You can use microwaves. 00:56:57.710 --> 00:56:59.810 align:middle line:84% They can also expedite diffusion processes. 00:56:59.810 --> 00:57:01.390 align:middle line:84% That's one way people are doing it. 00:57:01.390 --> 00:57:06.458 align:middle line:90% 00:57:06.458 --> 00:57:11.088 align:middle line:84% Now, the second one, for a lot of these consumer 00:57:11.088 --> 00:57:12.630 align:middle line:84% electronic products, people typically 00:57:12.630 --> 00:57:16.550 align:middle line:90% use aluminum silicate glasses. 00:57:16.550 --> 00:57:18.250 align:middle line:84% Now, the reason they use aluminum, 00:57:18.250 --> 00:57:21.230 align:middle line:84% you may recall that in the early lectures when 00:57:21.230 --> 00:57:23.170 align:middle line:84% we talked about aluminum silicate glasses, 00:57:23.170 --> 00:57:25.730 align:middle line:84% aluminum is somewhat similar to boron. 00:57:25.730 --> 00:57:27.670 align:middle line:84% It can exist in two different kinds of states. 00:57:27.670 --> 00:57:30.310 align:middle line:84% It can either form non-bridging oxygen 00:57:30.310 --> 00:57:32.390 align:middle line:90% acting as a network modifier. 00:57:32.390 --> 00:57:35.710 align:middle line:84% Or it can get converted into these four-fold coordinated 00:57:35.710 --> 00:57:39.830 align:middle line:84% states, and it increases the connectivity of the network. 00:57:39.830 --> 00:57:42.870 align:middle line:84% So it is exactly the latter property that was actually 00:57:42.870 --> 00:57:44.710 align:middle line:90% used in this case here. 00:57:44.710 --> 00:57:47.270 align:middle line:84% So for this application, they typically 00:57:47.270 --> 00:57:51.190 align:middle line:84% use an alkali-alumina ratio of about 1. 00:57:51.190 --> 00:57:54.430 align:middle line:84% If you go back to the lecture, this ratio 00:57:54.430 --> 00:57:57.230 align:middle line:84% corresponds to where you have the maximum amount 00:57:57.230 --> 00:58:01.510 align:middle line:84% of four-fold coordinated alumina. 00:58:01.510 --> 00:58:03.430 align:middle line:84% So having more four-fold coordinated alumina 00:58:03.430 --> 00:58:05.430 align:middle line:84% means that you have better network connectivity. 00:58:05.430 --> 00:58:10.020 align:middle line:84% And therefore, you increase the Tg of the glass. 00:58:10.020 --> 00:58:11.840 align:middle line:84% By increasing the Tg of the glass, 00:58:11.840 --> 00:58:15.060 align:middle line:84% you can now do your ion exchange process at high temperatures 00:58:15.060 --> 00:58:18.660 align:middle line:84% without worrying about stress relaxation. 00:58:18.660 --> 00:58:20.780 align:middle line:84% And that's how you can expedite the kinetics, 00:58:20.780 --> 00:58:23.640 align:middle line:84% by just essentially leveraging the Arrhenius law. 00:58:23.640 --> 00:58:27.540 align:middle line:90% 00:58:27.540 --> 00:58:31.100 align:middle line:84% Now, the next one, you can use ion pairs 00:58:31.100 --> 00:58:34.180 align:middle line:84% with less size disparity, which also makes sense. 00:58:34.180 --> 00:58:35.660 align:middle line:84% If you have the two ions [? and ?] 00:58:35.660 --> 00:58:37.920 align:middle line:84% the exchange process has a huge size difference, 00:58:37.920 --> 00:58:39.840 align:middle line:84% it becomes very difficult for one ion, the bigger one, 00:58:39.840 --> 00:58:41.923 align:middle line:84% to actually squeeze through the pocket left behind 00:58:41.923 --> 00:58:42.740 align:middle line:90% by the smaller one. 00:58:42.740 --> 00:58:46.960 align:middle line:84% So if you do a lithium-sodium pair, 00:58:46.960 --> 00:58:48.660 align:middle line:84% this is kinetically faster compared 00:58:48.660 --> 00:58:50.560 align:middle line:90% to a sodium-potassium pair. 00:58:50.560 --> 00:58:53.340 align:middle line:90% 00:58:53.340 --> 00:58:55.740 align:middle line:90% Next bullet point-- yes? 00:58:55.740 --> 00:59:02.580 align:middle line:84% AUDIENCE: If you use a smaller replacement ion, [INAUDIBLE] 00:59:02.580 --> 00:59:04.392 align:middle line:84% trade-off of, it's less effective 00:59:04.392 --> 00:59:06.600 align:middle line:84% JUEJUN HU: [INAUDIBLE] less effective in that regard. 00:59:06.600 --> 00:59:08.240 align:middle line:90% AUDIENCE: But it's much easier. 00:59:08.240 --> 00:59:10.220 align:middle line:84% JUEJUN HU: But you get much larger K steps, 00:59:10.220 --> 00:59:13.420 align:middle line:84% less amount of a compressive strength in general. 00:59:13.420 --> 00:59:14.760 align:middle line:90% So that's the trade-off there. 00:59:14.760 --> 00:59:15.860 align:middle line:84% Or in fact, what people will do-- let 00:59:15.860 --> 00:59:17.040 align:middle line:90% me get to the last point. 00:59:17.040 --> 00:59:19.700 align:middle line:84% People actually do mixed salt baths 00:59:19.700 --> 00:59:22.140 align:middle line:84% So for example, in a glass containing lithium, 00:59:22.140 --> 00:59:24.180 align:middle line:84% you put it in salt baths containing both sodium 00:59:24.180 --> 00:59:25.860 align:middle line:90% and potassium. 00:59:25.860 --> 00:59:27.660 align:middle line:90% So the sodium will go in deeper. 00:59:27.660 --> 00:59:29.200 align:middle line:90% It'll give you a large K steps. 00:59:29.200 --> 00:59:31.640 align:middle line:84% But potassium will go more in the surface region. 00:59:31.640 --> 00:59:34.180 align:middle line:84% But that gives you a very high surface compressive stress. 00:59:34.180 --> 00:59:37.460 align:middle line:90% So that also helps. 00:59:37.460 --> 00:59:40.140 align:middle line:84% Now, the last bullet point I want to mention here 00:59:40.140 --> 00:59:43.820 align:middle line:84% is to add some minor quantity of magnesium oxide or calcium 00:59:43.820 --> 00:59:46.863 align:middle line:84% oxide, about 2% usually, give or take. 00:59:46.863 --> 00:59:47.780 align:middle line:90% That's what people do. 00:59:47.780 --> 00:59:49.088 align:middle line:90% Why? 00:59:49.088 --> 00:59:51.380 align:middle line:84% There have been lots of studies about why that actually 00:59:51.380 --> 00:59:53.620 align:middle line:84% helps increase the diffusion kinetics. 00:59:53.620 --> 00:59:55.120 align:middle line:84% I'm not going to go through details. 00:59:55.120 --> 00:59:56.662 align:middle line:84% I don't think it's fully settled yet. 00:59:56.662 --> 00:59:58.420 align:middle line:84% But in practice, that's what people do. 00:59:58.420 --> 00:59:59.920 align:middle line:84% Empirically, that's what they found. 00:59:59.920 --> 01:00:07.040 align:middle line:90% 01:00:07.040 --> 01:00:08.940 align:middle line:84% So if you're happy with that, now let's 01:00:08.940 --> 01:00:12.700 align:middle line:84% actually look at one real-life example of a [? chemically ?] 01:00:12.700 --> 01:00:14.300 align:middle line:90% toughened glass. 01:00:14.300 --> 01:00:18.300 align:middle line:84% So the example I've given here is Gorilla Glass from Corning. 01:00:18.300 --> 01:00:20.140 align:middle line:84% Gorilla Glass was the one that was actually 01:00:20.140 --> 01:00:23.125 align:middle line:84% developed when Steve Jobs tried to approach Corning. 01:00:23.125 --> 01:00:25.500 align:middle line:84% Essentially, when he was trying to make the first iPhone, 01:00:25.500 --> 01:00:27.800 align:middle line:84% and he was complaining about plastics being terrible, 01:00:27.800 --> 01:00:29.747 align:middle line:84% it creates all the scratches, he wants glass. 01:00:29.747 --> 01:00:31.080 align:middle line:90% So that's why he called Corning. 01:00:31.080 --> 01:00:33.460 align:middle line:90% And Corning made it for them. 01:00:33.460 --> 01:00:35.020 align:middle line:84% And if you also remember, when we 01:00:35.020 --> 01:00:37.240 align:middle line:84% talk about the manufacturing of glass process, 01:00:37.240 --> 01:00:39.300 align:middle line:84% we talk about this fusion process-- 01:00:39.300 --> 01:00:40.700 align:middle line:90% [INAUDIBLE] fusion. 01:00:40.700 --> 01:00:42.220 align:middle line:84% Fusion process was really developed 01:00:42.220 --> 01:00:44.860 align:middle line:90% by Corning decades ago. 01:00:44.860 --> 01:00:45.880 align:middle line:90% It was shelved. 01:00:45.880 --> 01:00:47.460 align:middle line:90% Nobody found any use for that. 01:00:47.460 --> 01:00:50.240 align:middle line:84% Now, when they come to make Gorilla Glasses 01:00:50.240 --> 01:00:53.140 align:middle line:84% that's what they thought of when they reinvigorated or revived 01:00:53.140 --> 01:00:55.780 align:middle line:84% their process specifically for making high-quality cover 01:00:55.780 --> 01:00:57.500 align:middle line:90% glasses. 01:00:57.500 --> 01:01:00.610 align:middle line:84% Over the years, Gorilla Glasses have developed 01:01:00.610 --> 01:01:02.770 align:middle line:90% many, many generations. 01:01:02.770 --> 01:01:05.070 align:middle line:90% They are all slightly different. 01:01:05.070 --> 01:01:07.730 align:middle line:84% They A little bit better one generation than the other. 01:01:07.730 --> 01:01:09.958 align:middle line:84% I picked this one, Gorilla Glass 6. 01:01:09.958 --> 01:01:11.750 align:middle line:84% This is actually not the latest generation. 01:01:11.750 --> 01:01:13.330 align:middle line:90% The latest ones are the-- 01:01:13.330 --> 01:01:15.730 align:middle line:90% I think they have Gorilla 7i. 01:01:15.730 --> 01:01:17.670 align:middle line:90% They also have Gorilla Armor 2. 01:01:17.670 --> 01:01:19.330 align:middle line:90% They have Gorilla Invictus 2. 01:01:19.330 --> 01:01:21.027 align:middle line:90% Their 3 is the latest ones. 01:01:21.027 --> 01:01:23.110 align:middle line:84% I'm not going to go through the differences there. 01:01:23.110 --> 01:01:26.590 align:middle line:84% But the reason I picked Gorilla Glass 6 01:01:26.590 --> 01:01:28.090 align:middle line:84% is because this is the one they have 01:01:28.090 --> 01:01:30.690 align:middle line:90% the most complete data online. 01:01:30.690 --> 01:01:33.490 align:middle line:84% The latest generations, they hide some of the data. 01:01:33.490 --> 01:01:35.702 align:middle line:84% So that's why I decided to pick this one here. 01:01:35.702 --> 01:01:37.410 align:middle line:84% But again, they are more or less the same 01:01:37.410 --> 01:01:39.930 align:middle line:84% if you look at the data that was published. 01:01:39.930 --> 01:01:41.190 align:middle line:90% So now let's go through them. 01:01:41.190 --> 01:01:43.410 align:middle line:90% So first one-- viscosity. 01:01:43.410 --> 01:01:46.312 align:middle line:84% Essentially, they are looking at different reference points. 01:01:46.312 --> 01:01:47.770 align:middle line:84% These are the temperature reference 01:01:47.770 --> 01:01:49.730 align:middle line:84% points we talked about before-- softening point, annealing 01:01:49.730 --> 01:01:50.830 align:middle line:90% point, straining point. 01:01:50.830 --> 01:01:52.330 align:middle line:84% The key point I want to mention here 01:01:52.330 --> 01:01:58.130 align:middle line:84% is this guy, essentially annealing point. 01:01:58.130 --> 01:02:01.050 align:middle line:84% Sorry, I should mention that this is basically the glass 01:02:01.050 --> 01:02:03.070 align:middle line:90% transition temperature. 01:02:03.070 --> 01:02:04.570 align:middle line:84% So for glass transition temperature, 01:02:04.570 --> 01:02:05.370 align:middle line:90% this is annealing point. 01:02:05.370 --> 01:02:06.310 align:middle line:90% It's not straining point. 01:02:06.310 --> 01:02:07.710 align:middle line:84% I made a mistake the previous lecture. 01:02:07.710 --> 01:02:08.918 align:middle line:90% It should be annealing point. 01:02:08.918 --> 01:02:11.310 align:middle line:84% Straining point is beyond which you don't have any-- 01:02:11.310 --> 01:02:15.330 align:middle line:90% 01:02:15.330 --> 01:02:20.090 align:middle line:84% you don't have any further strain relaxation. 01:02:20.090 --> 01:02:22.090 align:middle line:84% Annealing point is when you have an expedited 01:02:22.090 --> 01:02:23.850 align:middle line:90% relaxation of a strain. 01:02:23.850 --> 01:02:26.410 align:middle line:84% But bottom line, this is close to the [? essential ?] glass 01:02:26.410 --> 01:02:27.770 align:middle line:90% transition temperature. 01:02:27.770 --> 01:02:31.210 align:middle line:84% So for aluminum silicate glasses, 01:02:31.210 --> 01:02:34.430 align:middle line:84% the annealing point in this case is about 620 degrees Celsius, 01:02:34.430 --> 01:02:37.130 align:middle line:84% which is much higher compared to soda-lime glass, which 01:02:37.130 --> 01:02:39.490 align:middle line:90% is about 510 degrees Celsius. 01:02:39.490 --> 01:02:42.050 align:middle line:84% So that's why they can afford using high temperatures 01:02:42.050 --> 01:02:45.010 align:middle line:84% in the ion exchange process to get the large K steps. 01:02:45.010 --> 01:02:47.902 align:middle line:84% In terms of the mechanical properties, 01:02:47.902 --> 01:02:49.610 align:middle line:84% you can look at density, Young's modulus, 01:02:49.610 --> 01:02:51.110 align:middle line:90% Poisson ratio, shear modulus. 01:02:51.110 --> 01:02:54.870 align:middle line:84% One little practice we can also do is in the previous lecture, 01:02:54.870 --> 01:02:57.970 align:middle line:84% I mentioned about this criteria of the ratio between the bulk 01:02:57.970 --> 01:02:59.810 align:middle line:90% modulus and shear modulus. 01:02:59.810 --> 01:03:02.170 align:middle line:84% The ratio turns out to be greater than 0.42, 01:03:02.170 --> 01:03:03.190 align:middle line:90% the critical number. 01:03:03.190 --> 01:03:05.110 align:middle line:84% So it should be brittle, which makes sense. 01:03:05.110 --> 01:03:06.890 align:middle line:84% The glass is [INAUDIBLE] brittle, 01:03:06.890 --> 01:03:09.370 align:middle line:84% although you also have a little bit of intrinsic ductility 01:03:09.370 --> 01:03:12.090 align:middle line:84% coming from the volume change during compressive stress, which 01:03:12.090 --> 01:03:13.230 align:middle line:90% we're going to cover. 01:03:13.230 --> 01:03:17.010 align:middle line:84% Now, if you look at the hardness, 01:03:17.010 --> 01:03:19.290 align:middle line:84% hardness is basically the ability of the material 01:03:19.290 --> 01:03:20.770 align:middle line:90% to resist deformation. 01:03:20.770 --> 01:03:24.090 align:middle line:84% So you see that after the ion exchange process, 01:03:24.090 --> 01:03:27.290 align:middle line:84% the hardness actually increases, because when you're actually 01:03:27.290 --> 01:03:29.870 align:middle line:84% measuring hardness in this case, they are using 01:03:29.870 --> 01:03:31.610 align:middle line:90% viscous indentation method. 01:03:31.610 --> 01:03:33.250 align:middle line:84% Essentially, it's the pyramidal shape 01:03:33.250 --> 01:03:35.950 align:middle line:84% of the stylus to compress on the surface. 01:03:35.950 --> 01:03:39.290 align:middle line:84% So you have a compressive stress on the surface. 01:03:39.290 --> 01:03:41.130 align:middle line:84% This compressive stress is going to resist 01:03:41.130 --> 01:03:43.350 align:middle line:90% deformation due to the stylus. 01:03:43.350 --> 01:03:45.250 align:middle line:84% And therefore, apparently, it will show up 01:03:45.250 --> 01:03:47.250 align:middle line:90% with a higher hardness. 01:03:47.250 --> 01:03:48.030 align:middle line:90% Makes sense? 01:03:48.030 --> 01:03:51.130 align:middle line:84% OK, now next one, this is the more tricky one-- fracture 01:03:51.130 --> 01:03:52.690 align:middle line:90% toughness. 01:03:52.690 --> 01:03:55.800 align:middle line:84% The fracture toughness of all the Gorilla Glasses, [INAUDIBLE] 01:03:55.800 --> 01:04:00.240 align:middle line:84% you find, it's all around like 0.7 megapascal square root 01:04:00.240 --> 01:04:01.200 align:middle line:90% meter. 01:04:01.200 --> 01:04:05.020 align:middle line:84% It's basically the same as untreated soda-lime glass. 01:04:05.020 --> 01:04:08.520 align:middle line:90% 01:04:08.520 --> 01:04:09.720 align:middle line:90% Why? 01:04:09.720 --> 01:04:12.720 align:middle line:84% The simple argument is that the fracture toughness 01:04:12.720 --> 01:04:15.600 align:middle line:84% is an intrinsic material parameter. 01:04:15.600 --> 01:04:19.400 align:middle line:84% So it should not depend on how you actually process it. 01:04:19.400 --> 01:04:22.200 align:middle line:84% In fact, I specifically asked my colleagues at Corning 01:04:22.200 --> 01:04:24.800 align:middle line:84% how they actually measure this fracture toughness number. 01:04:24.800 --> 01:04:27.600 align:middle line:84% It was done using a chevron-notch test, which 01:04:27.600 --> 01:04:29.220 align:middle line:90% is one of ASTM's standards. 01:04:29.220 --> 01:04:33.560 align:middle line:84% And it was done on a unstrengthened Gorilla Glass 01:04:33.560 --> 01:04:35.120 align:middle line:90% sample. 01:04:35.120 --> 01:04:39.840 align:middle line:84% So when you look at this number, it really 01:04:39.840 --> 01:04:42.260 align:middle line:84% only measures the intrinsic glass material property. 01:04:42.260 --> 01:04:45.240 align:middle line:84% It hasn't taken into account the compressive stress yet. 01:04:45.240 --> 01:04:47.560 align:middle line:84% Realistically, if you look at the material 01:04:47.560 --> 01:04:51.200 align:middle line:84% under actual use conditions, you have 01:04:51.200 --> 01:04:53.680 align:middle line:84% lots of compressive stress essentially forcing 01:04:53.680 --> 01:04:55.020 align:middle line:90% on the tip of the crack. 01:04:55.020 --> 01:04:58.800 align:middle line:84% So if you apply your external load, which 01:04:58.800 --> 01:05:04.920 align:middle line:84% is our sigma infinity, in order to induce opening of the crack, 01:05:04.920 --> 01:05:10.320 align:middle line:84% your sigma infinity has to overcome the compressive stress. 01:05:10.320 --> 01:05:12.480 align:middle line:84% And then on top of that, the remaining amount 01:05:12.480 --> 01:05:16.640 align:middle line:84% has to exceed essentially the fracture stress of the glass. 01:05:16.640 --> 01:05:20.000 align:middle line:84% So the compressive stress would actually 01:05:20.000 --> 01:05:23.040 align:middle line:84% increase the apparent ability for the material 01:05:23.040 --> 01:05:25.000 align:middle line:90% to resist breaking. 01:05:25.000 --> 01:05:28.680 align:middle line:84% But it does not get reflected in this measure 01:05:28.680 --> 01:05:30.920 align:middle line:90% number of fracture toughness. 01:05:30.920 --> 01:05:32.720 align:middle line:84% In fact, there are some people who 01:05:32.720 --> 01:05:34.600 align:middle line:84% are arguing, saying that you should not 01:05:34.600 --> 01:05:37.300 align:middle line:84% use unstrengthened glass to measure fracture toughness. 01:05:37.300 --> 01:05:39.400 align:middle line:84% You should essentially do some calculations 01:05:39.400 --> 01:05:42.100 align:middle line:84% to look at correct geometry, account for compressive stress, 01:05:42.100 --> 01:05:45.040 align:middle line:84% and use that modified fracture stress to actually calculate 01:05:45.040 --> 01:05:46.680 align:middle line:90% the fracture toughness. 01:05:46.680 --> 01:05:48.060 align:middle line:90% Most people don't accept that. 01:05:48.060 --> 01:05:48.640 align:middle line:90% That's why. 01:05:48.640 --> 01:05:52.560 align:middle line:84% They still do the untoughened glass. 01:05:52.560 --> 01:05:54.160 align:middle line:84% Now this is something to keep in mind. 01:05:54.160 --> 01:05:57.320 align:middle line:84% Because if you look at some news reports, 01:05:57.320 --> 01:06:00.680 align:middle line:84% they will just be citing this number as figure of merit 01:06:00.680 --> 01:06:02.240 align:middle line:84% to gauge a glass material's ability 01:06:02.240 --> 01:06:05.420 align:middle line:84% to resist fracturing or resist damage. 01:06:05.420 --> 01:06:08.840 align:middle line:84% This is not quite accurate because this is measured 01:06:08.840 --> 01:06:12.440 align:middle line:90% on the unstrengthened sample. 01:06:12.440 --> 01:06:15.360 align:middle line:84% All right, so now about chemical strengthening, finally here, 01:06:15.360 --> 01:06:17.640 align:middle line:84% you can see it has a compressive stress 01:06:17.640 --> 01:06:22.840 align:middle line:84% on the surface of maximally exceeding 900 megapascals. 01:06:22.840 --> 01:06:24.680 align:middle line:84% This is even higher than what you 01:06:24.680 --> 01:06:28.578 align:middle line:84% can get using in the Prince Rupert's drop. 01:06:28.578 --> 01:06:31.120 align:middle line:84% So you can really get to a very, very high compressive stress 01:06:31.120 --> 01:06:33.160 align:middle line:84% with chemical toughening process. 01:06:33.160 --> 01:06:37.292 align:middle line:84% The K steps, it's also larger compared to soda-lime glass. 01:06:37.292 --> 01:06:39.000 align:middle line:84% We mentioned soda-lime glass, the K steps 01:06:39.000 --> 01:06:40.940 align:middle line:90% is around 30 microns. 01:06:40.940 --> 01:06:43.320 align:middle line:84% You can actually get close to 100 microns with this kind 01:06:43.320 --> 01:06:45.000 align:middle line:90% of a modified formula there. 01:06:45.000 --> 01:06:45.720 align:middle line:90% Yes. 01:06:45.720 --> 01:06:48.960 align:middle line:84% AUDIENCE: Can you [INAUDIBLE] do [INAUDIBLE] 01:06:48.960 --> 01:06:50.180 align:middle line:90% in [? after you ?] do it. 01:06:50.180 --> 01:06:51.023 align:middle line:90% [INAUDIBLE] 01:06:51.023 --> 01:06:51.940 align:middle line:90% JUEJUN HU: To do what? 01:06:51.940 --> 01:06:55.820 align:middle line:90% AUDIENCE: [INAUDIBLE] 01:06:55.820 --> 01:06:59.200 align:middle line:84% JUEJUN HU: Yes, nobody really does that. 01:06:59.200 --> 01:07:02.820 align:middle line:84% That's the point about why would you actually want to do that? 01:07:02.820 --> 01:07:03.760 align:middle line:90% [CHUCKLES] 01:07:03.760 --> 01:07:05.957 align:middle line:84% Because if you do a thermal tempering, 01:07:05.957 --> 01:07:07.540 align:middle line:84% and then you do a chemical toughening, 01:07:07.540 --> 01:07:10.800 align:middle line:84% you combine the worst aspects of both. 01:07:10.800 --> 01:07:13.860 align:middle line:84% You get a large tensile stress region inside, 01:07:13.860 --> 01:07:16.800 align:middle line:90% which may be prone to damage. 01:07:16.800 --> 01:07:19.920 align:middle line:84% And then the amount of stress you accumulate on the surface, 01:07:19.920 --> 01:07:21.340 align:middle line:90% this is pretty much kept. 01:07:21.340 --> 01:07:22.350 align:middle line:90% That's very high stress. 01:07:22.350 --> 01:07:24.600 align:middle line:84% You get lots of relaxation already at that high stress 01:07:24.600 --> 01:07:25.300 align:middle line:90% level. 01:07:25.300 --> 01:07:26.760 align:middle line:84% So by combining both, I don't think 01:07:26.760 --> 01:07:29.860 align:middle line:84% you can get too much additional out of it. 01:07:29.860 --> 01:07:30.508 align:middle line:90% So that's why. 01:07:30.508 --> 01:07:32.300 align:middle line:84% And also, it's just simply, more expensive. 01:07:32.300 --> 01:07:34.320 align:middle line:84% Why not just do a simple chemical toughening? 01:07:34.320 --> 01:07:38.770 align:middle line:84% So that's why people usually do one but not the other 01:07:38.770 --> 01:07:40.020 align:middle line:90% We actually [? ran on ?] time. 01:07:40.020 --> 01:07:42.640 align:middle line:90% Any more questions? 01:07:42.640 --> 01:07:43.560 align:middle line:90% OK, I will stop here. 01:07:43.560 --> 01:07:44.680 align:middle line:84% In the next lecture, we are going 01:07:44.680 --> 01:07:46.590 align:middle line:84% to talk a little bit about the comparison 01:07:46.590 --> 01:07:51.260 align:middle line:84% of the sum of the different carbon materials beyond glasses. 01:07:51.260 --> 01:08:45.000 align:middle line:90%