SiC Technology

[Atul Agrawal]

I have recently started working on application of SiC technology in
VLSI application, but unfortunately I couldn't find much literature
avilable on internet. It would be grateful if any one can help me out
on finding them or giving me guidence.

 

[John Larkin]

On 30 Jul 2005 12:15:17 -0700, "Atul Agrawal" <atul.ee@gmail.com>
wrote:

I have recently started working on application of SiC technology in
VLSI application, but unfortunately I couldn't find much literature
avilable on internet. It would be grateful if any one can help me out
on finding them or giving me guidence.

Is that even possible? The defect density of SiC is so high it's hard
to make reliable diodes out of the stuff.

John

 

[The Phantom]

On Sat, 30 Jul 2005 14:28:23 -0700, John Larkin
<jjlarkin@highNOTlandTHIStechnologyPART.com> wrote:

On 30 Jul 2005 12:15:17 -0700, "Atul Agrawal" <atul.ee@gmail.com
wrote:

I have recently started working on application of SiC technology in
VLSI application, but unfortunately I couldn't find much literature
avilable on internet. It would be grateful if any one can help me out
on finding them or giving me guidence.


Is that even possible? The defect density of SiC is so high it's hard
to make reliable diodes out of the stuff.

John

You haven't been keeping up, John. Cree research has an ad on page 12 of the June issue
of Power Electronics for their Schottky diodes. Very nice specs:

http://www.cree.com/Products/pwr_index.asp

 

[John Larkin]

On Sat, 30 Jul 2005 15:01:23 -0700, The Phantom <phantom@aol.com>
wrote:

On Sat, 30 Jul 2005 14:28:23 -0700, John Larkin
jjlarkin@highNOTlandTHIStechnologyPART.com> wrote:

On 30 Jul 2005 12:15:17 -0700, "Atul Agrawal" <atul.ee@gmail.com
wrote:

I have recently started working on application of SiC technology in
VLSI application, but unfortunately I couldn't find much literature
avilable on internet. It would be grateful if any one can help me out
on finding them or giving me guidence.


Is that even possible? The defect density of SiC is so high it's hard
to make reliable diodes out of the stuff.

John

You haven't been keeping up, John. Cree research has an ad on page 12 of the June issue
of Power Electronics for their Schottky diodes. Very nice specs:

http://www.cree.com/Products/pwr_index.asp

I've used them, and some GaN parts, too. But it was/is very hard to
make reliable diodes from SiC, and they were actually available years
after they were announced, due to fab problems. Higher-voltage parts,
up to 1400v maybe, have been in the labs and journals for years but
are still not commercially feasible.

Infineon has a few nice parts, high voltage but low junction
capacitance.

Some people are starting to make SiC fets, too. They're more like
tubes, high-voltage, depletion mode, huge pinchoff voltages. Leaky as
hell.

John

 

[The Phantom]

On Sat, 30 Jul 2005 15:41:21 -0700, John Larkin
<jjlarkin@highNOTlandTHIStechnologyPART.com> wrote:

On Sat, 30 Jul 2005 15:01:23 -0700, The Phantom <phantom@aol.com
wrote:

On Sat, 30 Jul 2005 14:28:23 -0700, John Larkin
jjlarkin@highNOTlandTHIStechnologyPART.com> wrote:

On 30 Jul 2005 12:15:17 -0700, "Atul Agrawal" <atul.ee@gmail.com
wrote:

I have recently started working on application of SiC technology in
VLSI application, but unfortunately I couldn't find much literature
avilable on internet. It would be grateful if any one can help me out
on finding them or giving me guidence.


Is that even possible? The defect density of SiC is so high it's hard
to make reliable diodes out of the stuff.

John

You haven't been keeping up, John. Cree research has an ad on page 12 of the June issue
of Power Electronics for their Schottky diodes. Very nice specs:

http://www.cree.com/Products/pwr_index.asp


I've used them, and some GaN parts, too. But it was/is very hard to
make reliable diodes from SiC, and they were actually available years
after they were announced, due to fab problems. Higher-voltage parts,
up to 1400v maybe

Didn't you see the 1200 volt parts on that page?

, have been in the labs and journals for years but
are still not commercially feasible.

Infineon has a few nice parts, high voltage but low junction
capacitance.

Some people are starting to make SiC fets, too. They're more like
tubes, high-voltage, depletion mode, huge pinchoff voltages. Leaky as
hell.

John

 

[John Larkin]

On Sat, 30 Jul 2005 16:26:56 -0700, The Phantom <phantom@aol.com>
wrote:

On Sat, 30 Jul 2005 15:41:21 -0700, John Larkin
jjlarkin@highNOTlandTHIStechnologyPART.com> wrote:

On Sat, 30 Jul 2005 15:01:23 -0700, The Phantom <phantom@aol.com
wrote:

On Sat, 30 Jul 2005 14:28:23 -0700, John Larkin
jjlarkin@highNOTlandTHIStechnologyPART.com> wrote:

On 30 Jul 2005 12:15:17 -0700, "Atul Agrawal" <atul.ee@gmail.com
wrote:

I have recently started working on application of SiC technology in
VLSI application, but unfortunately I couldn't find much literature
avilable on internet. It would be grateful if any one can help me out
on finding them or giving me guidence.


Is that even possible? The defect density of SiC is so high it's hard
to make reliable diodes out of the stuff.

John

You haven't been keeping up, John. Cree research has an ad on page 12 of the June issue
of Power Electronics for their Schottky diodes. Very nice specs:

http://www.cree.com/Products/pwr_index.asp


I've used them, and some GaN parts, too. But it was/is very hard to
make reliable diodes from SiC, and they were actually available years
after they were announced, due to fab problems. Higher-voltage parts,
up to 1400v maybe

Didn't you see the 1200 volt parts on that page?

Yup, screaming progress. I think they're up to using 2" wafers now.

John

 

[The Phantom]

On Sat, 30 Jul 2005 16:43:47 -0700, John Larkin

I've used them, and some GaN parts, too. But it was/is very hard to
make reliable diodes from SiC, and they were actually available years
after they were announced, due to fab problems. Higher-voltage parts,
up to 1400v maybe

Didn't you see the 1200 volt parts on that page?



Yup, screaming progress. I think they're up to using 2" wafers now.

I think you're being overly pessimistic about them. From their 1999
Annual Report:

"The wafer business was a solid performer this year as worldwide R&D
and production drove sales higher. Also the materials business
targeted for gemstone applications enjoyed a significant year of
development and revenue growth. During this past fiscal year, Cree has
dramatically improved yields of the silicon carbide crystals supplied
to C3. In addition to improving the yields, Cree has been driving to
increase the size of the crystals enabling C3 to potentially increase
the number of moissanite gemstones available for distribution. As a
result of our development of three-inch size crystals, Cree will
gradually migrate to the larger three-inch size wafers for LED
production which will increase the number of LED chips available for
shipment later this fiscal year."

So they were up to 3" wafers a while ago.

From the 2004 Annual Report:

"Over the past few years, we have continued to expand our product line
of three-inch wafers, which are better suited for the manufacture of
power and microwave devices. We continue to develop SiC wafers that
are larger and of higher quality."

In the 2004 report, they say that sales of wafers were 11%, 9% and 7%
of revenue for 2002, 2003 and 2004, respectively. Sales of LEDs
increased over the same time period.

About 2% of revenue came from sales of bulk SiC crystals for gemstone
use (moissanite).

Sales of LED's represented 78% of revenue in 2004, but power
semiconductor devices only generated 1% of revenue.

So, maybe as the production of 3" wafers increases, we'll see more
power semiconductor devices. I'd like to see a SiC power FET. I
remember a description from them some years ago in one of the IEEE
journals of a power FET that could operate at 350 degrees. It
wouldn't take much of a heat sink to get rid of a lot of heat at that
temperature!

John

 

[John Larkin]

On 31 Jul 2005 04:33:06 -0500, The Phantom <phantom@aol.com> wrote:

On Sat, 30 Jul 2005 16:43:47 -0700, John Larkin
I've used them, and some GaN parts, too. But it was/is very hard to
make reliable diodes from SiC, and they were actually available years
after they were announced, due to fab problems. Higher-voltage parts,
up to 1400v maybe

Didn't you see the 1200 volt parts on that page?



Yup, screaming progress. I think they're up to using 2" wafers now.

I think you're being overly pessimistic about them. From their 1999
Annual Report:

"The wafer business was a solid performer this year as worldwide R&D
and production drove sales higher. Also the materials business
targeted for gemstone applications enjoyed a significant year of
development and revenue growth. During this past fiscal year, Cree has
dramatically improved yields of the silicon carbide crystals supplied
to C3. In addition to improving the yields, Cree has been driving to
increase the size of the crystals enabling C3 to potentially increase
the number of moissanite gemstones available for distribution. As a
result of our development of three-inch size crystals, Cree will
gradually migrate to the larger three-inch size wafers for LED
production which will increase the number of LED chips available for
shipment later this fiscal year."

So they were up to 3" wafers a while ago.

From the 2004 Annual Report:

"Over the past few years, we have continued to expand our product line
of three-inch wafers, which are better suited for the manufacture of
power and microwave devices. We continue to develop SiC wafers that
are larger and of higher quality."

In the 2004 report, they say that sales of wafers were 11%, 9% and 7%
of revenue for 2002, 2003 and 2004, respectively. Sales of LEDs
increased over the same time period.

About 2% of revenue came from sales of bulk SiC crystals for gemstone
use (moissanite).

Sales of LED's represented 78% of revenue in 2004, but power
semiconductor devices only generated 1% of revenue.

So, maybe as the production of 3" wafers increases, we'll see more
power semiconductor devices. I'd like to see a SiC power FET. I
remember a description from them some years ago in one of the IEEE
journals of a power FET that could operate at 350 degrees. It
wouldn't take much of a heat sink to get rid of a lot of heat at that
temperature!

Oh, don't get me wrong, I've been eagerly waiting for kilovolt SiC
schottkies and high-voltage mesfets (dare I hope for phemts?) for
years now... many years now. It's just that deliveries have lacked
promises by 10 years on average, which slows down the design cycle
some. I designed in some hv SiC schottkies about 5 years before I
actually got any that would work (we based our designs on Motorola and
Microsemi datasheets, none of which became real parts.) We're
currently using three Infineon 400 volt SiC schottkies in series to
get the high voltage and low capacitance we need.

TI sampled us some wonderful SOT-89 mesfets, Idss = 7 amps, but bailed
on production. It must be tough working with compound semiconductors.

The specs I've seen on SiC and GaN mesfets have been discouraging,
too. Huge pinchoff voltages, high capacitances, stuff like that.

Oh well, in another 10 years...

We're very lucky that silicon exists.


John

 

[The Phantom]

On Sun, 31 Jul 2005 11:20:38 -0700, John Larkin
<jjlarkin@highNOTlandTHIStechnologyPART.com> wrote:

On 31 Jul 2005 04:33:06 -0500, The Phantom <phantom@aol.com> wrote:

On Sat, 30 Jul 2005 16:43:47 -0700, John Larkin
I've used them, and some GaN parts, too. But it was/is very hard to
make reliable diodes from SiC, and they were actually available years
after they were announced, due to fab problems. Higher-voltage parts,
up to 1400v maybe

Didn't you see the 1200 volt parts on that page?



Yup, screaming progress. I think they're up to using 2" wafers now.

I think you're being overly pessimistic about them. From their 1999
Annual Report:

"The wafer business was a solid performer this year as worldwide R&D
and production drove sales higher. Also the materials business
targeted for gemstone applications enjoyed a significant year of
development and revenue growth. During this past fiscal year, Cree has
dramatically improved yields of the silicon carbide crystals supplied
to C3. In addition to improving the yields, Cree has been driving to
increase the size of the crystals enabling C3 to potentially increase
the number of moissanite gemstones available for distribution. As a
result of our development of three-inch size crystals, Cree will
gradually migrate to the larger three-inch size wafers for LED
production which will increase the number of LED chips available for
shipment later this fiscal year."

So they were up to 3" wafers a while ago.

From the 2004 Annual Report:

"Over the past few years, we have continued to expand our product line
of three-inch wafers, which are better suited for the manufacture of
power and microwave devices. We continue to develop SiC wafers that
are larger and of higher quality."

In the 2004 report, they say that sales of wafers were 11%, 9% and 7%
of revenue for 2002, 2003 and 2004, respectively. Sales of LEDs
increased over the same time period.

About 2% of revenue came from sales of bulk SiC crystals for gemstone
use (moissanite).

Sales of LED's represented 78% of revenue in 2004, but power
semiconductor devices only generated 1% of revenue.

So, maybe as the production of 3" wafers increases, we'll see more
power semiconductor devices. I'd like to see a SiC power FET. I
remember a description from them some years ago in one of the IEEE
journals of a power FET that could operate at 350 degrees. It
wouldn't take much of a heat sink to get rid of a lot of heat at that
temperature!


Oh, don't get me wrong, I've been eagerly waiting for kilovolt SiC
schottkies and high-voltage mesfets (dare I hope for phemts?) for
years now... many years now. It's just that deliveries have lacked
promises by 10 years on average, which slows down the design cycle
some. I designed in some hv SiC schottkies about 5 years before I
actually got any that would work (we based our designs on Motorola and
Microsemi datasheets, none of which became real parts.) We're
currently using three Infineon 400 volt SiC schottkies in series to
get the high voltage and low capacitance we need.

How long has it been since you got those parts? The Infineon web site
(http://www.infineon.com/cgi/ecrm.dll/ecrm/scripts/prod_cat.jsp?oid=-8681)
doesn't currently list any 400 volt parts. Maybe your parts aren't
standard parts. Anyway, the Cree parts seem pretty similar (I
compared 10 amp, 600 volt parts from both manufacturers), although I
think the Cree parts have a larger chip, since their capacitances are
larger and Rtheta j-c is substantially smaller. I notice that
Infineon also says they are using 3 inch wafers.

Given how long it has taken to get real SiC parts, how long till we
get diamond Schottkies?

TI sampled us some wonderful SOT-89 mesfets, Idss = 7 amps, but bailed
on production. It must be tough working with compound semiconductors.

The specs I've seen on SiC and GaN mesfets have been discouraging,
too. Huge pinchoff voltages, high capacitances, stuff like that.

Oh well, in another 10 years...

We're very lucky that silicon exists.


John