UTSKRIFT 🔗
1
00:00:00,626 --> 00:00:02,910
(upbeat music)
2
00:00:02,910 --> 00:00:06,810
- This is "Engenuity," a
Boilermaker Engineering podcast.
3
00:00:06,810 --> 00:00:10,293
This episode discusses quantum
transport and Spintronics.
4
00:00:12,330 --> 00:00:15,540
- Welcome to this episode
of our "Engenuity" podcast,
5
00:00:15,540 --> 00:00:17,550
featuring some of the latest, inspiring
6
00:00:17,550 --> 00:00:20,100
and innovative stories
from Purdue Engineering.
7
00:00:20,100 --> 00:00:22,170
I'm Arvind Raman, Boilermaker alumnus
8
00:00:22,170 --> 00:00:24,180
and Dean of Engineering.
9
00:00:24,180 --> 00:00:26,850
Now quantum transport in electrons
10
00:00:26,850 --> 00:00:29,550
might sound like something from Star Trek,
11
00:00:29,550 --> 00:00:33,060
but understanding how
electrons flow and spin
12
00:00:33,060 --> 00:00:36,180
and interact in ever shrinking spaces
13
00:00:36,180 --> 00:00:39,420
barely a few tens or
hundreds of atoms wide
14
00:00:39,420 --> 00:00:41,040
is what makes possible
15
00:00:41,040 --> 00:00:44,910
modern semiconductor technology advances
16
00:00:44,910 --> 00:00:48,090
like billion plus
transistors in the smartphone
17
00:00:48,090 --> 00:00:50,220
or future quantum computers.
18
00:00:50,220 --> 00:00:52,470
Researchers working in this space
19
00:00:52,470 --> 00:00:55,200
study how to manipulate spin
in the flow of electrons
20
00:00:55,200 --> 00:00:59,220
and electronic devices
that power our modern world
21
00:00:59,220 --> 00:01:03,450
to increase computational power
for emerging needs like AI,
22
00:01:03,450 --> 00:01:07,140
autonomy, optimization
of large complex systems,
23
00:01:07,140 --> 00:01:10,020
edge computing and the internet of things,
24
00:01:10,020 --> 00:01:12,123
and to lower power consumption.
25
00:01:13,110 --> 00:01:17,100
Today's episode of "Engenuity"
features Supriyo Datta,
26
00:01:17,100 --> 00:01:19,290
the Thomas Duncan Distinguished Professor
27
00:01:19,290 --> 00:01:21,210
in Electrical and Computer Engineering
28
00:01:21,210 --> 00:01:24,017
at the Elmore Family School of Electrical
29
00:01:24,017 --> 00:01:24,850
and Computer Engineering.
30
00:01:24,850 --> 00:01:27,030
Professor Datta is in the forefront
31
00:01:27,030 --> 00:01:29,790
of this group of researchers
and has been called
32
00:01:29,790 --> 00:01:31,890
one of the most original thinkers
33
00:01:31,890 --> 00:01:34,233
in the field of nanoscale electronics.
34
00:01:35,756 --> 00:01:38,214
Supriyo Datta's interdisciplinary work
35
00:01:38,214 --> 00:01:39,731
on quantum mechanical transport
36
00:01:39,731 --> 00:01:42,017
spans chemistry, physics,
electrical engineering,
37
00:01:42,017 --> 00:01:44,670
and has produced many firsts:
38
00:01:44,670 --> 00:01:48,812
the first rigorous quantum
mechanical simulations
39
00:01:48,812 --> 00:01:51,975
of nano and molecular
scale electronic devices,
40
00:01:51,975 --> 00:01:54,850
the first concept for a spintronic switch,
41
00:01:54,850 --> 00:01:57,390
the so-called Datta-Das Spin Transistor,
42
00:01:57,390 --> 00:01:59,640
and a new kind of spin-based logic.
43
00:01:59,640 --> 00:02:02,838
His conceptual approach
and computational methods
44
00:02:02,838 --> 00:02:04,830
are now widely used by
scientists and engineers
45
00:02:04,830 --> 00:02:05,880
throughout the world.
46
00:02:06,792 --> 00:02:07,800
And his ideas for spintronics
47
00:02:07,800 --> 00:02:10,593
focused international
attention on the field.
48
00:02:12,172 --> 00:02:15,158
It is a great pleasure not
only to welcome Professor Datta
49
00:02:15,158 --> 00:02:18,353
to "Engenuity," but to
congratulate him on his election
50
00:02:18,353 --> 00:02:20,760
to the prestigious National
Academy of Sciences.
51
00:02:20,760 --> 00:02:24,480
These select scientists that
are elected by their peers
52
00:02:24,480 --> 00:02:25,800
to membership in the Academy
53
00:02:25,800 --> 00:02:28,590
for outstanding contributions to research.
54
00:02:28,590 --> 00:02:32,250
Some 500 members have won Nobel Prizes.
55
00:02:32,250 --> 00:02:34,050
They work alongside their colleagues
56
00:02:34,050 --> 00:02:37,050
at the National Academies
of Engineering and Medicine
57
00:02:37,050 --> 00:02:40,650
to advise the nation on how
to solve complex problems
58
00:02:40,650 --> 00:02:43,473
and make informed public policy decisions.
59
00:02:44,310 --> 00:02:46,710
Professor Datta is the
only Purdue faculty member
60
00:02:46,710 --> 00:02:47,710
to have been elected
61
00:02:48,550 --> 00:02:50,550
to both the National Academy of Sciences
62
00:02:50,550 --> 00:02:52,710
and the one for engineering.
63
00:02:52,710 --> 00:02:54,710
Welcome, Professor Datta to "Engenuity."
64
00:02:55,920 --> 00:02:57,820
- Thank you very much.
65
00:02:57,820 --> 00:03:00,270
Thank you for having me here,
and this is real pleasure.
66
00:03:00,270 --> 00:03:02,034
- It is great to have you here.
67
00:03:02,034 --> 00:03:04,470
Now, Professor Datta,
you've just been elected
68
00:03:04,470 --> 00:03:07,260
to the National Academy of Sciences,
69
00:03:07,260 --> 00:03:12,260
really a fitting culmination
to a lifetime of contributions
70
00:03:12,433 --> 00:03:14,670
to innovative research and development.
71
00:03:14,670 --> 00:03:16,780
But I'd like to take us back
72
00:03:17,797 --> 00:03:19,890
to the starting point of that arc.
73
00:03:19,890 --> 00:03:22,493
What got you started on your way?
74
00:03:22,493 --> 00:03:27,030
Can you unpack, how did
you first get interested
75
00:03:27,030 --> 00:03:31,055
in science and engineering
and in electrons?
76
00:03:31,055 --> 00:03:32,332
You know, how did that all happen?
77
00:03:32,332 --> 00:03:33,890
- Frankly speaking, you see,
78
00:03:33,890 --> 00:03:36,556
growing up had no particular interest
79
00:03:36,556 --> 00:03:37,703
in science or engineering,
80
00:03:38,632 --> 00:03:41,310
and it was only during
my first or second year
81
00:03:41,310 --> 00:03:45,780
in undergraduate in college
Load full transcript
that a friend referred me
82
00:03:45,780 --> 00:03:47,373
to this Feynman's Lectures.
83
00:03:48,435 --> 00:03:49,436
- I see.
84
00:03:49,436 --> 00:03:50,269
- I mean, that was not our textbook,
85
00:03:50,269 --> 00:03:53,160
but he told me about
these amazing lectures,
86
00:03:53,160 --> 00:03:55,620
and in those days, Feynman
wasn't as well known.
87
00:03:55,620 --> 00:03:59,250
I mean, and so I went to
the library in Kharagpur.
88
00:03:59,250 --> 00:04:01,680
This is back in the 70s, early 1970s,
89
00:04:01,680 --> 00:04:04,594
and I found these lectures,
90
00:04:04,594 --> 00:04:06,344
and I really found that fascinating
91
00:04:07,414 --> 00:04:09,180
because I was really intrigued
92
00:04:09,180 --> 00:04:13,537
by his way of looking at the
world, which is so different.
93
00:04:13,537 --> 00:04:15,900
And I really wanted to
understand it deeply
94
00:04:15,900 --> 00:04:20,370
because, so I hoped I could
at least gain a fraction
95
00:04:20,370 --> 00:04:24,333
of this grand view of things
that he seemed to enjoy see.
96
00:04:25,292 --> 00:04:27,420
Anyway, so that's what
kind of got me started.
97
00:04:27,420 --> 00:04:31,516
And later, I guess when I joined Purdue,
98
00:04:31,516 --> 00:04:34,888
I was looking for an area
99
00:04:34,888 --> 00:04:37,930
where I'd be able to do quantum physics
100
00:04:38,880 --> 00:04:42,891
while in an engineering context, right?
101
00:04:42,891 --> 00:04:44,790
And at that time it wasn't clear
102
00:04:44,790 --> 00:04:47,308
how small devices would become.
103
00:04:47,308 --> 00:04:49,145
You know, you said about
atoms, in those days
104
00:04:49,145 --> 00:04:50,587
it was devices of millions of atoms,
105
00:04:50,587 --> 00:04:54,969
but I think people kind of had the idea
106
00:04:54,969 --> 00:04:55,802
that it would gradually shrink.
107
00:04:55,802 --> 00:04:57,783
No one knew it would
continue for 40 years,
108
00:04:58,846 --> 00:05:00,932
but they thought maybe
it could continue for 10.
109
00:05:00,932 --> 00:05:03,503
So I figured it may be a good
direction to get started.
110
00:05:05,412 --> 00:05:06,670
- Well, that's great.
111
00:05:06,670 --> 00:05:08,580
So you expected Moore's law to continue
112
00:05:08,580 --> 00:05:10,969
and dimensions to keep shrinking
113
00:05:10,969 --> 00:05:13,050
and quantum effects to
become more important.
114
00:05:13,050 --> 00:05:14,266
That's great.
115
00:05:14,266 --> 00:05:16,568
- Yeah, it wasn't clear
how far it would have to go
116
00:05:16,568 --> 00:05:17,788
before that happened.
117
00:05:17,788 --> 00:05:18,621
- Right, right.
118
00:05:18,621 --> 00:05:19,454
- Yeah.
119
00:05:21,108 --> 00:05:22,590
- So that brings me to, you
know, a following question.
120
00:05:22,590 --> 00:05:24,600
You know, Moore's law,
121
00:05:24,600 --> 00:05:27,406
but as you look to the future, you know,
122
00:05:27,406 --> 00:05:30,612
there are questions about the Moore's Law.
123
00:05:30,612 --> 00:05:32,526
You know, can that
sustain,
124
00:05:32,526 --> 00:05:35,430
you mentioned four decades of
shrinking device dimensions
125
00:05:35,430 --> 00:05:38,310
and sizes and the huge
amounts of innovation
126
00:05:38,310 --> 00:05:41,520
and challenges needed to overcome those.
127
00:05:41,520 --> 00:05:46,110
But when you look to
the future of computing,
128
00:05:46,110 --> 00:05:50,610
broadly speaking,
electronics could be one way,
129
00:05:50,610 --> 00:05:52,666
could be one of the ways.
130
00:05:52,666 --> 00:05:55,070
Hopefully there's gonna
be an electronic device
131
00:05:55,070 --> 00:05:57,060
in the future, not just, you know,
132
00:05:57,060 --> 00:05:59,684
besides optical devices, you know,
133
00:05:59,684 --> 00:06:01,650
who knows what the future
of computing might be,
134
00:06:01,650 --> 00:06:03,630
computing and memory.
135
00:06:03,630 --> 00:06:05,130
But in the context of electronics,
136
00:06:05,130 --> 00:06:10,023
as you look to the future,
the next 50 years, let's say,
137
00:06:11,310 --> 00:06:14,040
what do you see are the drivers
138
00:06:14,040 --> 00:06:19,040
that are going to drive changes
in innovation in this area,
139
00:06:21,000 --> 00:06:22,863
much like Moore's Law did?
140
00:06:23,706 --> 00:06:25,650
Now you can argue Moore's
law will keep going on,
141
00:06:25,650 --> 00:06:29,310
but maybe there are
other drivers, you know,
142
00:06:29,310 --> 00:06:34,310
whether it's AI, whether it's
autonomy, internet of things,
143
00:06:34,620 --> 00:06:37,980
edge computing, optimization problems
144
00:06:37,980 --> 00:06:42,090
that are today simply too
complex for us to solve
145
00:06:42,090 --> 00:06:43,669
in the future.
146
00:06:43,669 --> 00:06:46,037
What do you think are the drivers
147
00:06:46,037 --> 00:06:49,527
and what might be, you
know, some promising,
148
00:06:51,690 --> 00:06:56,230
you know, substrates or
solutions that might help
149
00:06:57,180 --> 00:06:59,943
unlock these challenges in
the next 50 years let's say?
150
00:07:01,829 --> 00:07:02,813
Big question, sorry.
151
00:07:04,024 --> 00:07:05,029
- No, no, I get it.
152
00:07:05,029 --> 00:07:07,145
It's like the, well, you know,
153
00:07:07,145 --> 00:07:08,789
when you look back on the last 50 years,
154
00:07:08,789 --> 00:07:11,423
I guess one humbling
thought is that, you know,
155
00:07:11,423 --> 00:07:13,765
none of us foresaw all this.
156
00:07:13,765 --> 00:07:15,639
I mean, the smartphone, even 20 years ago,
157
00:07:15,639 --> 00:07:18,323
I don't think we had any idea
what it would become, right?
158
00:07:19,224 --> 00:07:21,420
And so that is a big warning sign
159
00:07:21,420 --> 00:07:23,553
about predicting 50 years, right?
160
00:07:24,485 --> 00:07:25,482
- Fair enough.
161
00:07:25,482 --> 00:07:28,480
- Right, but the one big change I'm seeing
162
00:07:29,341 --> 00:07:33,300
in the last 10, 15 years
is something like this.
163
00:07:33,300 --> 00:07:37,290
When we started back in 1980s and all,
164
00:07:37,290 --> 00:07:39,222
gradually what happened
165
00:07:39,222 --> 00:07:41,544
was our electrical engineering community
166
00:07:41,544 --> 00:07:45,360
kind of split into two
groups, rather clear.
167
00:07:45,360 --> 00:07:49,551
One group was about
making a small transistor
168
00:07:49,551 --> 00:07:53,310
so that you could put pack
more of them into chips.
169
00:07:53,310 --> 00:07:56,230
That's what you might
call the physical side.
170
00:07:56,230 --> 00:07:57,750
And then there was the other side,
171
00:07:57,750 --> 00:08:02,041
which was about using all this added power
172
00:08:02,041 --> 00:08:04,261
to do amazing things.
173
00:08:04,261 --> 00:08:05,910
You know, to implement new algorithms,
174
00:08:05,910 --> 00:08:08,626
powerful functions and so on.
175
00:08:08,626 --> 00:08:11,861
And so it is almost like
the two parts are different.
176
00:08:11,861 --> 00:08:13,264
One's making a better transistor.
177
00:08:13,264 --> 00:08:14,916
The other is using it, right?
178
00:08:14,916 --> 00:08:18,198
Physical side, lot of
overlap with physics,
179
00:08:18,198 --> 00:08:19,598
material science, chemistry.
180
00:08:20,580 --> 00:08:25,200
The other side, the virtual
side, more about mathematics,
181
00:08:25,200 --> 00:08:26,616
computer science.
182
00:08:26,616 --> 00:08:29,190
It's almost two different languages.
183
00:08:29,190 --> 00:08:33,878
But what I'm seeing more and
more in the last 10, 15 years
184
00:08:33,878 --> 00:08:35,760
is the two coming together.
185
00:08:35,760 --> 00:08:38,370
And probably because
Moore's law is slowing down.
186
00:08:38,370 --> 00:08:40,710
And so people figure you can't just say
187
00:08:40,710 --> 00:08:43,596
that we'll make a transistor,
use it anyway you want.
188
00:08:43,596 --> 00:08:46,860
We'll say, well, what algorithms
are you interested in?
189
00:08:46,860 --> 00:08:50,379
We'll try to tailor the device to that,
190
00:08:50,379 --> 00:08:53,220
so that you are hearing more and more of,
191
00:08:53,220 --> 00:08:55,410
I guess even in, you see when you walk in,
192
00:08:55,410 --> 00:08:57,240
you see this many places
193
00:08:57,240 --> 00:09:00,171
where the virtual meets
the physical, right?
194
00:09:00,171 --> 00:09:01,355
- Yep, yep.
- That's kind of what.
195
00:09:01,355 --> 00:09:03,660
- Yeah, you can see that
in the lobby of MSEE
196
00:09:03,660 --> 00:09:06,870
- But this is not something
I think you would've seen
197
00:09:06,870 --> 00:09:07,914
20 years ago.
198
00:09:07,914 --> 00:09:10,290
So this is a new thing
which will probably carry
199
00:09:10,290 --> 00:09:12,000
for another decade at least.
200
00:09:12,000 --> 00:09:15,540
Beyond that, as I said,
I don't want to venture,
201
00:09:15,540 --> 00:09:16,885
yeah, any guesses.
202
00:09:16,885 --> 00:09:20,030
- Well, I mean, a decade into
the future is pretty good.
203
00:09:20,030 --> 00:09:21,510
I appreciate that.
204
00:09:21,510 --> 00:09:26,160
So Datta, can you tell
our listeners, you
know,
205
00:09:26,160 --> 00:09:30,120
in a simple way, you know,
about what is quantum transport
206
00:09:30,120 --> 00:09:32,358
and what is spintronics?
207
00:09:32,358 --> 00:09:35,911
You know, to many, it almost
sounds like science fiction,
208
00:09:35,911 --> 00:09:39,120
but if you could unpack some of that,
209
00:09:39,120 --> 00:09:41,169
that would be really helpful.
210
00:09:41,169 --> 00:09:42,990
- Yes, see these transistors are devices
211
00:09:42,990 --> 00:09:46,096
that control the flow of electrons.
212
00:09:46,096 --> 00:09:48,811
What I guess formally you'd
call a transport problem.
213
00:09:48,811 --> 00:09:52,425
Like heat transport,
there's electron transport.
214
00:09:52,425 --> 00:09:55,860
Now 40 years ago, devices
were large, millions of atoms,
215
00:09:55,860 --> 00:09:58,878
and it was a classical transport problem.
216
00:09:58,878 --> 00:10:00,720
So you could largely think of electrons
217
00:10:00,720 --> 00:10:04,413
as just classical particles
and describe their flow.
218
00:10:05,370 --> 00:10:06,870
Now what's happened,
219
00:10:06,870 --> 00:10:10,118
today's devices' dimensions
are down to nanometers.
220
00:10:10,118 --> 00:10:12,120
We are talking of devices
that feature sizes
221
00:10:12,120 --> 00:10:15,300
that are like hundreds of
atoms, tens of atoms even.
222
00:10:15,300 --> 00:10:19,725
And so we have to deal with
elements of quantum transport
223
00:10:19,725 --> 00:10:24,280
that kind of have to include
this view of electrons
224
00:10:24,280 --> 00:10:27,237
as the wave nature of electrons.
225
00:10:27,237 --> 00:10:32,234
And one example of that you
may have heard of is tunneling.
226
00:10:32,234 --> 00:10:36,003
So modern devices have
tunneling involved many places.
227
00:10:37,110 --> 00:10:39,780
Another interesting element here
228
00:10:39,780 --> 00:10:42,450
is the spin of the electron.
229
00:10:42,450 --> 00:10:46,150
And this one, I'm thinking
how best to explain it
230
00:10:47,014 --> 00:10:49,334
because I guess it usually helps
231
00:10:49,334 --> 00:10:51,662
if it is analogous to
something you know well.
232
00:10:51,662 --> 00:10:54,016
- How about baseball?
233
00:10:54,016 --> 00:10:55,560
- Okay.
234
00:10:55,560 --> 00:10:58,133
- You know, every ball has a spin on it.
235
00:10:58,133 --> 00:10:59,711
- A spin, yes.
236
00:10:59,711 --> 00:11:03,275
But so I would say, yeah,
237
00:11:03,275 --> 00:11:05,238
I guess you could say
the electron is spinning,
238
00:11:05,238 --> 00:11:08,315
and that's what gives
it a magnetic moment.
239
00:11:08,315 --> 00:11:11,232
Because a spin is like
as if there's a current,
240
00:11:11,232 --> 00:11:12,795
and anytime there's a current loop,
241
00:11:12,795 --> 00:11:14,155
there's a magnetic moment to it,
242
00:11:14,155 --> 00:11:18,537
and it could be spinning in, up or down.
243
00:11:18,537 --> 00:11:19,892
And this is the part
244
00:11:19,892 --> 00:11:21,210
that's kind of counterintuitive somewhat.
245
00:11:21,210 --> 00:11:24,435
And I often use the example of light
246
00:11:24,435 --> 00:11:26,910
because that's usually
more familiar to people,
247
00:11:26,910 --> 00:11:30,328
that photons have a polarization.
248
00:11:30,328 --> 00:11:31,890
That's something you hear
about in freshman physics.
249
00:11:31,890 --> 00:11:34,035
You're familiar with that.
250
00:11:34,035 --> 00:11:36,995
So you know, photons, these
are electromagnetic waves.
251
00:11:36,995 --> 00:11:38,150
There's electric field.
252
00:11:38,150 --> 00:11:39,491
So what's the polarization?
253
00:11:39,491 --> 00:11:40,523
It's the direction of the electric field.
254
00:11:41,700 --> 00:11:45,013
And usually you deal
with unpolarized light,
255
00:11:45,013 --> 00:11:48,422
meaning the electric field
is all over the place.
256
00:11:48,422 --> 00:11:52,317
It's a mixture of fields
in all directions.
257
00:11:52,317 --> 00:11:55,454
But then you could build
interesting devices
258
00:11:55,454 --> 00:11:57,002
if you used polarized light.
259
00:11:57,002 --> 00:11:59,970
And that's used in a
lot of optical devices.
260
00:11:59,970 --> 00:12:03,411
So I'd say that the electron
spin is a little bit
261
00:12:03,411 --> 00:12:05,534
like the photon polarization.
262
00:12:05,534 --> 00:12:07,112
Now there are interesting differences,
263
00:12:07,112 --> 00:12:08,910
and I could kind of go on
for a while about that one,
264
00:12:08,910 --> 00:12:10,860
but let me avoid doing that.
265
00:12:10,860 --> 00:12:14,500
But it is a little bit
like photon polarization.
266
00:12:14,500 --> 00:12:17,752
And so usually your electronic devices
267
00:12:17,752 --> 00:12:20,091
are dealing with unpolarized electrons.
268
00:12:20,091 --> 00:12:21,294
So spins in all directions,
269
00:12:21,294 --> 00:12:22,822
so you don't even worry about it.
270
00:12:22,822 --> 00:12:24,990
But you could have
these spintronic devices
271
00:12:24,990 --> 00:12:28,826
where the idea is generate
spin polarized electrons.
272
00:12:28,826 --> 00:12:31,680
So what comes in is polarized
in a particular direction.
273
00:12:31,680 --> 00:12:35,331
And then when you detect
it in somewhere else,
274
00:12:35,331 --> 00:12:38,040
you can select the angle of detection.
275
00:12:38,040 --> 00:12:40,290
And one interesting, I've
mentioned about the difference
276
00:12:40,290 --> 00:12:43,260
between photons and electrons.
277
00:12:43,260 --> 00:12:45,286
And let me just say something.
278
00:12:45,286 --> 00:12:47,875
For example, with photons,
those who are familiar with it,
279
00:12:47,875 --> 00:12:50,894
they would say, if you are
polarized say in the X direction,
280
00:12:50,894 --> 00:12:53,730
the way you'd stop it
is by having an analyzer
281
00:12:53,730 --> 00:12:55,217
that's in the Y direction
282
00:12:55,217 --> 00:12:58,419
'cause that's orthogonal, 90 degrees.
283
00:12:58,419 --> 00:12:59,580
With electrons.
284
00:12:59,580 --> 00:13:01,140
If you're polarized
up, the way you stop
it
285
00:13:01,140 --> 00:13:04,383
is by having the analyzer be
down, which is 180 degrees.
286
00:13:05,310 --> 00:13:08,310
So that's why you kind of need
a new math to describe it.
287
00:13:08,310 --> 00:13:09,330
It's not vectors.
288
00:13:09,330 --> 00:13:11,700
You have to use spinners,
something different.
289
00:13:11,700 --> 00:13:13,680
Anyway, so the interesting differences.
290
00:13:13,680 --> 00:13:15,550
But finally, I'd say yes,
291
00:13:15,550 --> 00:13:17,646
Spintronics is about trying to exploit
292
00:13:17,646 --> 00:13:20,746
this degree of freedom,
this polarization aspect,
293
00:13:20,746 --> 00:13:22,910
- Professor Datta, there are, you know,
294
00:13:22,910 --> 00:13:25,122
as you spoke about spintronics,
295
00:13:25,122 --> 00:13:29,580
spintronics has been considered
as a potential platform
296
00:13:29,580 --> 00:13:34,580
for future quantum computing
and applications like that.
297
00:13:35,220 --> 00:13:37,950
And in fact, they've been
described as, you know,
298
00:13:37,950 --> 00:13:39,843
poor man's cubit in some sense.
299
00:13:40,686 --> 00:13:43,500
So when you look to the
future of quantum computing,
300
00:13:43,500 --> 00:13:47,914
what is the role that
Spintronics could play?
301
00:13:47,914 --> 00:13:52,914
Is it a specialized, is it a broad role?
302
00:13:52,921 --> 00:13:54,287
Will it have a role?
303
00:13:54,287 --> 00:13:55,553
What do you see there?
304
00:13:57,566 --> 00:13:59,730
- Yes, I guess I've been,
you know, quantum computing,
305
00:13:59,730 --> 00:14:02,067
as you know is a vast area.
306
00:14:02,067 --> 00:14:04,189
A lot of people working on it.
307
00:14:04,189 --> 00:14:06,520
Now the part that always intrigued me
308
00:14:07,426 --> 00:14:09,630
so this is more a
philosophical thing than a,
309
00:14:09,630 --> 00:14:12,094
from an application point of view.
310
00:14:12,094 --> 00:14:14,190
And that is that what is the distinction
311
00:14:14,190 --> 00:14:17,187
between this quantum and
classical probability?
312
00:14:17,187 --> 00:14:19,390
'cause everyone realizes
quantum is very different
313
00:14:19,390 --> 00:14:21,270
from the deterministic world,
314
00:14:21,270 --> 00:14:24,391
but it always intrigues
people that, you know,
315
00:14:24,391 --> 00:14:29,391
that it involves probability
in a very fundamental way.
316
00:14:29,524 --> 00:14:33,455
And yet there is something
non-intuitive about it.
317
00:14:33,455 --> 00:14:36,690
So that is different from
classical probability.
318
00:14:36,690 --> 00:14:41,070
And again, many years ago in
the 80s when I first came here,
319
00:14:41,070 --> 00:14:43,272
I was very intrigued by a little remark
320
00:14:43,272 --> 00:14:45,432
by Feynman again in a lecture,
321
00:14:45,432 --> 00:14:47,556
where he said, you know, I've
worried a lot about this.
322
00:14:47,556 --> 00:14:49,470
Why is quantum so intriguing?
323
00:14:49,470 --> 00:14:50,970
Why does it bother people so much?
324
00:14:50,970 --> 00:14:53,640
And he said, and the best I can say is
325
00:14:53,640 --> 00:14:57,486
that it is as if you're
dealing with probability,
326
00:14:57,486 --> 00:14:59,363
but these probabilities can be negative.
327
00:15:00,869 --> 00:15:03,470
And that's about, in his mind,
328
00:15:03,470 --> 00:15:05,467
that was the key difference, you see.
329
00:15:06,404 --> 00:15:08,852
And as I said, you know,
most of my career, you know,
330
00:15:08,852 --> 00:15:09,964
I've been trying to understand
331
00:15:09,964 --> 00:15:12,513
all these little short
statements of Feynman.
332
00:15:13,731 --> 00:15:14,749
- That's right.
333
00:15:14,749 --> 00:15:15,582
- Spread all over.
334
00:15:15,582 --> 00:15:17,040
- That's right, so that reference
335
00:15:17,040 --> 00:15:22,040
to Feynman's book 50
years ago continues to,
336
00:15:24,150 --> 00:15:27,288
you know, create new
directions and new ideas.
337
00:15:27,288 --> 00:15:28,686
This is fantastic.
338
00:15:28,686 --> 00:15:30,190
- So in the modern context, I'd say,
339
00:15:30,190 --> 00:15:31,290
what I've been intrigued by is
340
00:15:31,290 --> 00:15:35,040
what can you do with quantum computing
341
00:15:35,040 --> 00:15:36,060
that is very different
342
00:15:36,060 --> 00:15:39,003
from what you could do with
a probabilistic computer?
343
00:15:40,242 --> 00:15:41,610
Again, the difference with
a deterministic computer
344
00:15:41,610 --> 00:15:44,310
is very clear, but what is it better?
345
00:15:44,310 --> 00:15:47,760
How is it different from a
simple probabilistic computer?
346
00:15:47,760 --> 00:15:51,307
And that's where I felt
that these spins and magnets
347
00:15:51,307 --> 00:15:53,383
give you a natural story.
348
00:15:53,383 --> 00:15:55,423
And that is, it says spin.
349
00:15:55,423 --> 00:15:58,342
I mean, as I said, basically
an electron, the spin,
350
00:15:58,342 --> 00:15:59,763
it is a quantum magnet.
351
00:16:01,230 --> 00:16:04,380
And then of course there
are magnets we have at home,
352
00:16:04,380 --> 00:16:07,049
which are classical deterministic magnet.
353
00:16:07,049 --> 00:16:08,783
They point in some direction,
354
00:16:08,783 --> 00:16:11,305
and there's a region in between,
355
00:16:11,305 --> 00:16:14,258
which is if you make a magnet
small, it becomes unstable.
356
00:16:14,258 --> 00:16:16,505
So it doesn't quite point up or down.
357
00:16:16,505 --> 00:16:18,463
It kind of fluctuates between them.
358
00:16:18,463 --> 00:16:20,010
But it's a very classical thing, you see?
359
00:16:20,010 --> 00:16:23,066
And those who make magnetic
devices always worry about it,
360
00:16:23,066 --> 00:16:25,590
because usually you use
magnets to store information,
361
00:16:25,590 --> 00:16:28,874
and if it fluctuates up and
down, that's not a great thing.
362
00:16:28,874 --> 00:16:31,785
So that's what we figured
363
00:16:31,785 --> 00:16:34,753
was a very good physical representation
364
00:16:34,753 --> 00:16:37,247
of something we called a pbit,
365
00:16:37,247 --> 00:16:39,941
meaning it's a probabilistic bit, right?
366
00:16:39,941 --> 00:16:41,941
So it has some features of cubits,
367
00:16:41,941 --> 00:16:43,659
which is the quantum magnet,
368
00:16:43,659 --> 00:16:45,764
but then it's a very classical thing.
369
00:16:45,764 --> 00:16:47,730
And my question was, if
you built a P computer out
370
00:16:47,730 --> 00:16:51,270
of these Pbits, you know, what things,
371
00:16:51,270 --> 00:16:54,119
what part of quantum
computing could you duplicate
372
00:16:54,119 --> 00:16:56,324
and what part you couldn't?
373
00:16:56,324 --> 00:16:57,157
And of course, the basic answer,
374
00:16:57,157 --> 00:16:58,382
I think I still would say
375
00:16:58,382 --> 00:17:01,230
it is about these negative probabilities.
376
00:17:01,230 --> 00:17:02,670
If the quantum algorithm
377
00:17:02,670 --> 00:17:07,670
takes advantage of negative
probabilities, then it would be,
378
00:17:07,753 --> 00:17:09,810
it would be equivalent.
379
00:17:09,810 --> 00:17:11,370
If it is negative probabilities,
380
00:17:11,370 --> 00:17:13,865
then you couldn't do it classically.
381
00:17:13,865 --> 00:17:14,863
- Right.
382
00:17:14,863 --> 00:17:15,865
- So you really need the quantum hardware.
383
00:17:15,865 --> 00:17:16,883
- Right, right.
384
00:17:16,883 --> 00:17:18,360
- But if you're not
taking advantage of it,
385
00:17:18,360 --> 00:17:21,415
then a classical hardware
might have done the same.
386
00:17:21,415 --> 00:17:24,000
- So it's kind of on the
border somehow of the two.
387
00:17:24,000 --> 00:17:26,373
No, that's, yes, that's fascinating.
388
00:17:27,293 --> 00:17:30,779
You already spoke about
semiconductor technology
389
00:17:30,779 --> 00:17:31,777
and computing, you know,
390
00:17:31,777 --> 00:17:35,070
the fact that what you're seeing
looking ahead is, you know,
391
00:17:35,070 --> 00:17:40,070
the two worlds of algorithms and bits,
392
00:17:40,680 --> 00:17:43,675
you know, kind of working hand in hand now
393
00:17:43,675 --> 00:17:45,825
with the design of
devices from the get go.
394
00:17:46,955 --> 00:17:51,955
What are, you know, some
potential futuristic directions
395
00:17:52,937 --> 00:17:57,057
that you see that are gonna
be exciting in this area
396
00:17:57,057 --> 00:17:58,507
for semiconductor technology?
397
00:17:59,476 --> 00:18:00,750
You know, are there specific materials?
398
00:18:00,750 --> 00:18:02,817
Are there specific concepts
399
00:18:02,817 --> 00:18:04,315
that you're excited about personally?
400
00:18:04,315 --> 00:18:07,080
Again, this is not asking
you to predict the future,
401
00:18:07,080 --> 00:18:09,490
but rather it is simply a question about
402
00:18:10,373 --> 00:18:13,440
what are current exciting,
what you see as exciting,
403
00:18:13,440 --> 00:18:15,375
you know, emerging trends.
404
00:18:15,375 --> 00:18:19,048
- Yeah, this is the part
I probably couldn't,
405
00:18:19,048 --> 00:18:21,201
don't have much to say
406
00:18:21,201 --> 00:18:24,297
based because it is
just things I've heard.
407
00:18:24,297 --> 00:18:26,269
What I'm hearing is,
408
00:18:26,269 --> 00:18:28,537
well, transistors will continue to shrink,
409
00:18:28,537 --> 00:18:31,129
but even if they don't shrink too much,
410
00:18:32,430 --> 00:18:35,819
chips will probably get more
and more powerful and amazing.
411
00:18:35,819 --> 00:18:38,555
And a good part of that would be
412
00:18:38,555 --> 00:18:40,572
this integration of the
virtual and the physical
413
00:18:40,572 --> 00:18:43,915
that would help make things
more energy efficient,
414
00:18:43,915 --> 00:18:45,065
faster and all of that.
415
00:18:46,315 --> 00:18:48,420
And in terms of algorithms, I think
416
00:18:48,420 --> 00:18:53,310
we are used to thinking of
computers as precise calculators,
417
00:18:53,310 --> 00:18:58,297
right, you know, where 13
times 11 is 143, not 141, 142.
418
00:18:58,297 --> 00:19:00,750
And we're always told
that you need human beings
419
00:19:00,750 --> 00:19:02,670
to do this intuitive type of thing,
420
00:19:02,670 --> 00:19:06,120
that I just see a part of your
face, and I know it's you.
421
00:19:06,120 --> 00:19:08,897
Right, and a computer can do that,
422
00:19:08,897 --> 00:19:11,355
is what I was hearing 20 years ago.
423
00:19:11,355 --> 00:19:12,838
Now we all know that, you know,
424
00:19:12,838 --> 00:19:14,441
computers are being able to do all that.
425
00:19:14,441 --> 00:19:17,655
You know, ChatGPT is of
course all over the place.
426
00:19:17,655 --> 00:19:21,171
And, of course, the question then is
427
00:19:21,171 --> 00:19:23,880
if you're doing algorithms like that,
428
00:19:23,880 --> 00:19:25,920
are there other hardware?
429
00:19:25,920 --> 00:19:27,607
Can you modify the transistor
430
00:19:27,607 --> 00:19:29,401
so it'll be much more efficient
431
00:19:29,401 --> 00:19:31,290
at implementing things like that.
432
00:19:31,290 --> 00:19:35,370
So these are the paradigms
I'm hearing a lot.
433
00:19:35,370 --> 00:19:37,560
And I know I have many
colleagues here at Purdue
434
00:19:37,560 --> 00:19:39,450
who are leading all of this,
435
00:19:39,450 --> 00:19:42,822
and they probably can speak
with much more authority
436
00:19:42,822 --> 00:19:44,861
on all of this.
437
00:19:44,861 --> 00:19:46,604
- Yeah, no, this is great.
438
00:19:46,604 --> 00:19:48,054
What does your election
439
00:19:48,054 --> 00:19:50,670
to the National Academy
of Sciences mean to you?
440
00:19:50,670 --> 00:19:53,610
And, you know, what are
some new opportunities
441
00:19:53,610 --> 00:19:57,210
that it might open up to influence policy
442
00:19:57,210 --> 00:19:59,733
at a national level to you personally?
443
00:20:02,430 --> 00:20:05,190
- I guess it's of course
very nice to be recognized
444
00:20:05,190 --> 00:20:08,730
by this esteemed group, you
know, and I appreciate that.
445
00:20:08,730 --> 00:20:11,529
And much of my work, as you know,
446
00:20:11,529 -->
00:20:14,478
has been about nanoscale
electronic devices,
447
00:20:14,478 --> 00:20:18,660
which is clearly of
engineering interest, right?
448
00:20:18,660 --> 00:20:22,230
But it also requires
this new way of thinking
449
00:20:22,230 --> 00:20:25,410
about current flow, conceptual advances,
450
00:20:25,410 --> 00:20:28,231
which are of general scientific interest,
451
00:20:28,231 --> 00:20:31,110
which go beyond just transistors.
452
00:20:31,110 --> 00:20:35,670
And it is nice to see that
aspect recognized, right?
453
00:20:35,670 --> 00:20:36,503
That is very,
454
00:20:37,500 --> 00:20:40,021
but finally, in the end, of course,
455
00:20:40,021 --> 00:20:44,384
the real reward is seeing
your work being used
456
00:20:44,384 --> 00:20:47,836
by this new generation of
scientists and engineers.
457
00:20:47,836 --> 00:20:49,111
And all these awards
458
00:20:49,111 --> 00:20:51,562
are kind of a formal recognition for that.
459
00:20:51,562 --> 00:20:54,996
But what I really hope is that I can,
460
00:20:54,996 --> 00:20:57,540
yeah, continue working
with this new generation
461
00:20:57,540 --> 00:21:00,192
of amazing people and students
462
00:21:00,192 --> 00:21:04,660
and help them launch new
frontiers and some of which
463
00:21:04,660 --> 00:21:06,743
we are not even being
able to imagine today.
464
00:21:08,040 --> 00:21:10,576
- Speaking of that, Professor Datta,
465
00:21:10,576 --> 00:21:14,281
you often speak about
your research and teaching
466
00:21:14,281 --> 00:21:19,143
being completely intertwined,
entangled if you will,
467
00:21:20,370 --> 00:21:21,203
for you.
468
00:21:22,176 --> 00:21:25,172
So what does that mean
on a day-to-day basis?
469
00:21:25,172 --> 00:21:28,710
I mean, all this incredible
discovery that you do,
470
00:21:28,710 --> 00:21:30,990
how do you, you know, bring it
471
00:21:30,990 --> 00:21:32,665
and speak to students about it?
472
00:21:32,665 --> 00:21:34,040
How do you engage students?
473
00:21:34,040 --> 00:21:35,840
How is that one and the same for you?
474
00:21:36,970 --> 00:21:41,970
- Yes, again, because I think new students
475
00:21:42,229 --> 00:21:44,730
who are seeing it for
the first time, you know,
476
00:21:44,730 --> 00:21:45,870
ask you questions
477
00:21:45,870 --> 00:21:49,160
that you kind of stopped
asking on the way somewhere
478
00:21:49,160 --> 00:21:51,759
as you became an expert, right?
479
00:21:51,759 --> 00:21:56,070
And so, and then the trick
is not to be annoyed by it,
480
00:21:56,070 --> 00:21:58,195
but think seriously about it.
481
00:21:58,195 --> 00:21:59,484
- Right.
482
00:21:59,484 --> 00:22:01,023
- What exactly is bothering them?
483
00:22:01,023 --> 00:22:02,850
And I'd say, you know, we
talked a little bit about spin.
484
00:22:02,850 --> 00:22:04,148
- Right.
485
00:22:04,148 --> 00:22:06,682
- And that, I think back in the 80s
486
00:22:06,682 --> 00:22:08,400
came from students asking me,
487
00:22:08,400 --> 00:22:10,444
what exactly is this spin
about?
488
00:22:10,444 --> 00:22:12,450
And then I started thinking of analogies,
489
00:22:12,450 --> 00:22:16,440
and that led to these ideas
for devices, applications,
490
00:22:16,440 --> 00:22:18,741
all of that kind of came from there.
491
00:22:18,741 --> 00:22:22,957
So usually I've felt
that it really helps to,
492
00:22:22,957 --> 00:22:26,609
and even now when I have
a complex topic in mind
493
00:22:26,609 --> 00:22:30,600
and I want to make sure
I really understand it,
494
00:22:30,600 --> 00:22:32,761
I can, even if I don't
have a real student,
495
00:22:32,761 --> 00:22:35,887
I imagine I'm explaining it to someone.
496
00:22:35,887 --> 00:22:39,201
And that always helps me, you
know, clear up my own mind.
497
00:22:39,201 --> 00:22:41,428
So I've often said that, you know,
498
00:22:41,428 --> 00:22:43,989
I really teach to learn really,
499
00:22:43,989 --> 00:22:46,830
and it's nice if other
people are also learning,
500
00:22:46,830 --> 00:22:49,980
but I'm really doing it only in a way
501
00:22:49,980 --> 00:22:51,483
that really helps me a lot.
502
00:22:52,320 --> 00:22:54,527
- Well, that's really incredible.
503
00:22:54,527 --> 00:22:57,381
So interacting with
our students, you know,
504
00:22:57,381 --> 00:22:59,290
the spark that they have,
505
00:22:59,290 --> 00:23:04,243
and the questions they ask is what, like,
506
00:23:04,243 --> 00:23:08,160
in answering those or in
trying to resolve them in ways
507
00:23:08,160 --> 00:23:12,090
is what allow you to be
inspired in some sense
508
00:23:12,090 --> 00:23:14,868
for new directions.
509
00:23:14,868 --> 00:23:17,973
I mean, that's a wonderful,
wonderful explanation, right?
510
00:23:19,082 --> 00:23:21,632
- And with students, I
think a cardinal rule is to,
511
00:23:22,560 --> 00:23:25,410
that nothing is too silly.
512
00:23:25,410 --> 00:23:26,708
- Right.
513
00:23:26,708 --> 00:23:28,353
- Or all questions are
very welcome always.
514
00:23:28,353 --> 00:23:29,461
- Right.
515
00:23:29,461 --> 00:23:31,530
- And if I don't know the
answer, I'll tell you as much.
516
00:23:31,530 --> 00:23:33,660
- Oh, this is incredible.
517
00:23:33,660 --> 00:23:36,540
To learn more about the
National Academy of Sciences
518
00:23:36,540 --> 00:23:39,600
and its role in our nation's
education and research,
519
00:23:40,850 --> 00:23:45,850
visit its website at
https://www.nasonline.org.
520
00:23:47,341 --> 00:23:49,800
This is Arvind Raman signing off
521
00:23:49,800 --> 00:23:52,074
this episode of "Engenuity."
522
00:23:52,074 --> 00:23:54,741
(upbeat music)
523
00:23:55,717 --> 00:23:56,550
- This has been "Engenuity,"
524
00:23:56,550 --> 00:23:58,743
a Boilermaker Engineering podcast.
525
00:23:58,743 --> 00:24:00,690
For more information on some of the issues
526
00:24:00,690 --> 00:24:03,720
discussed in this podcast, go
to the engineering homepage
527
00:24:03,720 --> 00:24:06,975
and click on the podcast
icon on the upper right.
528
00:24:06,975 --> 00:24:09,642
(upbeat music)