Sign in with
Sign up | Sign in
Your question

Is thin layer of silver / gold a better heat conductor?

Last response: in CPUs
Share
November 9, 2006 5:57:06 PM

If somebody uses a very thin layer of Silver / gold metal between CPU and HSF will that be better than the thermal paste as silver/gold are good conductors of heat?

What is the thermal paste made of?

Just curious

Regards.

More about : thin layer silver gold heat conductor

November 9, 2006 6:09:31 PM

Artic Silver actually contains silver. Cheaper thermal pastes can use other conductive metals. The point of using a paste is that it fills in small gaps, grooves and imperfections in the surfaces of your CPU's heat spreader and the bottom of the heatsink.

You don't want to use a layer of silver (or gold) unless it was "lapped" perfectly smooth, and your heatspreader and heatsink were also "lapped".
But, if you have a "lapped" heatsink and heat spreader, you wouldn't need the layer of silver, since the heatsink and spreader would make direct contact.
November 9, 2006 6:36:08 PM

exactly.
November 9, 2006 7:47:02 PM

Quote:
If somebody uses a very thin layer of Silver / gold metal between CPU and HSF will that be better than the thermal paste as silver/gold are good conductors of heat?

What is the thermal paste made of?

Just curious

Regards.

The silver/gold itself yes, however, the interface of the hypothetic sheet would be a pretty poo conductor because even if it was smooth @ molecular level, the heatsink and heatspreader surfaces are not and the contact surface would still be reduced to the heatspreader/heatsink grain pattern; Without filling these gaps, the surface may be reduced up to 1/8, even if your haked eye sees it smooth. That's why all thermal materials have some viscosity.
November 9, 2006 7:52:02 PM

Heh heh heh.... you said poo.
November 9, 2006 8:31:36 PM

It all depends on how much heat each individual element can transfer. Look at the Coolaboratory Liquid Pro, it's an amalagam of different metals that also transfers heat decently (we won't talk about how it eats Alumminum...)

Ask a chemist these questions, I'm sure they can tell you what would facilitate heat transfer better. I'm an EE, we don't deal with heat... (At least theoretically)
November 9, 2006 8:47:51 PM

Quote:
I'm an EE, we don't deal with heat... (At least theoretically)


hey, i'm an EE, and i deal with heat (although, that's because i build high-powered amplifiers)

but anyway, silver is the best heat/electricity conductor, then copper, gold, aluminium, etc.
(silver is best for heatsinks, but damn expensive, copper is barely far behind, and a lot cheaper, although Al is even cheaper, any oxidisation doesn't matter on heatsinks)

If anyone was wondering, the only reason gold is used for quality audio connectors, for example, is because silver and copper oxidize, gold is inert. And copper (at least, OFC) cables are the best, unless you want to waste your money on Silver cables (ie silveraudio.com)

ps, look here:
http://en.wikipedia.org/wiki/Thermal_conductivity#Some_...


and as everyone else said, the "paste"ness is to eliminate any tiny air-bubbles, and smooth out imperfections. Normal (non-AS5) heatsink paste is silicon-based, and conducts heat pretty poorly, but won't short-circuit anything when you get it everywhere...
November 9, 2006 8:55:10 PM

For the best air cooling, build your own heatsink from sterling silver. It might cost a bomb - but it will cool a lot better than aluminium will.
Anonymous
a b à CPUs
November 9, 2006 9:26:32 PM

hey PF haven't seen you up there in ages...Just reading and occasionnal posting?
November 9, 2006 9:42:59 PM

Is there some diamond-sinks?
November 9, 2006 9:54:57 PM

I highly doubt it. But they should make diamond powder/silver paste.
November 9, 2006 10:12:18 PM

deployed right now so dont get to post as much as i want to, only on liberty time.
November 9, 2006 10:50:20 PM

Quote:
For the best air cooling, build your own heatsink from sterling silver. It might cost a bomb - but it will cool a lot better than aluminium will.

Hah... I'll get right on a custom fabrication of a 100% silver HS. It's not like it will cost more than water cooling or anything! :lol: 
November 9, 2006 11:07:28 PM

Artic Silver 5 contains silver as it is mentioned. Soldering the hs and cpu would provide better thermal transfer. :lol:  That is if you are rich and would not worry about throwing out cpu and hsf everytime you upgrade.

Anyways there's this from Frozencpu.com. This thermal material, Coollaboratory Liquid PRO Thermal Interface Material. It's like liquid metal and then sets dry, but I heard that this liquid metal stuff basically binds the cpu and the heatsink and it could be a pain in the arse to take them apart.

So as for you, the Artic Silver 5 or Artic Silver Ceramique would be good enough.
November 10, 2006 10:15:47 AM

Quote:
Heh heh heh.... you said poo.

ain't it the same concept :D 
November 10, 2006 10:33:12 AM

Silver is the only conductor better than copper.
Gold is pimp, though :lol: 
November 10, 2006 10:42:55 AM

:lol:  right,... you can always weld the HSF but then you won't be able to upgrade CPU!
November 10, 2006 11:31:53 AM

however welding is hot and ironically you would burn out the cpu by applying the heatsink
November 10, 2006 11:40:48 AM

Quote:
I highly doubt it. But they should make diamond powder/silver paste.


man you stole my idea! i had the same thought about a month ago but if you inhale powdered diamond it'll cut up for lungs and you'll basiclly drown in blood so it would be extremly dangerus, other than that it should work great! and cost insane amounts of money!
November 10, 2006 12:02:18 PM

Not to mention after you'd welded/soldered the heatsink to the cpu, you'd never get it back into the socket properly.

--Shodar
a c 140 à CPUs
November 10, 2006 12:02:47 PM

Quote:
I highly doubt it. But they should make diamond powder/silver paste.


man you stole my idea! i had the same thought about a month ago but if you inhale powdered diamond it'll cut up for lungs and you'll basiclly drown in blood so it would be extremly dangerus, other than that it should work great! and cost insane amounts of money!
Ummmmmm ouch!!!
November 10, 2006 12:31:38 PM

Artic Pimp Deluxe thermal paste: diamond powder with gold.
Yo broda! 8)
:lol:  :lol:  :lol: 
November 10, 2006 12:41:45 PM

The other advantage of gold, and the reason it is used for contacts for every connector in a PC, is that it bonds to itself. Gold bars literally stick together and have to be forced apart due to this cohesion. So, if processors came gold plated and milled smooth, a heatsink with a milled gold surface would outperform any thermal paste, because it would basically weld itself together at room temperature.
November 10, 2006 12:47:06 PM

Yea, I've been grinding up diamonds for years and using them as thermal interface. It's really easy, just swipe enough wedding rings (you'll need roughly 2 carats, color and clarity don't matter) Then you have to crush them up. I use a dremel moto-tool and the griding attachment w/ carbide tip. You'll need to have about 20 of these tips cuz they get eaten up pretty quickly. Suspend the diamond in a vice above a bowl of water and get to grinding. Once done, let the water evaporate out of the bowl and scrape the diamond residue out. To make the thermal interface compound you'll need approx. 1 teaspoon of ground diamond residue, 1 egg white, 2 Tblsp corn starch, 1 Teaspoon clear silicone (NOT RTV), a pinch of sugar, a teaspoon of vanilla extract, and lemon rind to garnish. Bring all to a steady boil over medium heat and whip. Serves up to 4 CPU's.
November 10, 2006 12:49:08 PM

The scary thing is.... someone is going to try this.
November 10, 2006 1:07:36 PM

Quote:
The scary thing is.... someone is going to try this.


Please, don't.
November 10, 2006 1:41:14 PM

Quote:
For the best air cooling, build your own heatsink from sterling silver. It might cost a bomb - but it will cool a lot better than aluminium will.


I would NOT recommend making a heat sink out of silver unless you really like polishing. Silver oxidizes a ton (if you've ever owned sterling sliverware, jewelry, or other silver stuff, you'll know that you need to polish frequently with special silver polish, and it's quite a pain) and oxides are electrical/thermal insulators, in general. (read the post about gold conductors on electronics equipment)

If you're interested in what's in Arctic Silver, go here: http://www.arcticsilver.com/msds.htm

For the most common AS, being AS5: http://www.arcticsilver.com/PDF/as5msds.pdf

Lots of oxides in there, which I found quite curious. I'm sure they're used as high heat conduction insulating compounds or something. Basically, AS is a matrix of silver, alumina, zinc oxide, and BN (white graphite) suspended in ester oil. So, the thermal dissipation you're getting is probably just that of ester oil, plus a few watts.

Anywho, I would decidedly recommend against soldering or welding a chip to anything. The heat will kill your electronics. Special manufacturing techniques are applied to seal the chip package by applying very quick, local heating to the areas that need to be sealed so that the chip doesn't overheat. Think of frying your chip by overclocking too much with not enough cooling. Processors aren't terribly thermally robust.

Now, that said, welding the housing to a heat spreader is actually a pretty good idea. Even a relatively basic heat spreader (or, crappy, by the standards applied by people that frequent this forum) that was welded to the chip housing would be far superior to even some of the best heat spreaders out there right now that use standard thermal compounds, like AS. The problems with this are that 1. Chip makers Intel and AMD (and they'd be the ones that would need to implement this) don't seem particularly interested in the problem, since they are steadfastly opposed to overclocking (think - locked processors), and it would mean adding a brand-new manufacturing process, which means less $, 2. The market isn't terribly large, since people like y'all reading this post (remember, you'd be the target audience for superior thermally conductive processors) are inclined to want to tweak your own gear, and being forced to deal with a stock heat spreader is not a very preferable thing for any DIY comp builder, and 3. As opposed to your standard CPU housing now, where there's relatively little metal, and thus not a ton of heat dissipation on its own, if you're attaching a heat spreader directly, you're effectively adding a weld chill, thus making localized heating for processor sealing (remember - gotta protect the processor) a real pain in the arse.

There was a topic a week or two back about the best thermal conductor. Interesting reading if you go back and find it (I'm too lazy to link). As it turns out, basic fluoride toothpaste is a superior thermal conductor (and by a fair margin, mind you) than even the latest and greatest AS compound. Of course, toothpaste will dry out and corrode everything in your computer, but if you're just going to try to break the world overclocking record for 10 minutes, toothpaste may be the right thermal compound for you!

Cheers.
November 10, 2006 6:53:07 PM

Quote:
[...] oxides are electrical/thermal insulators, in general. (read the post about gold conductors on electronics equipment)

I think a little clarification is necessary. The post you referred to stated "oxidisation doesn't matter on heatsinks". I do know for sure that oxides are poor electrical conductors. So what about oxidation of heat sinks? Better still, does it matter where? For example, oxidation of heat sink fins versus oxidation of the interface surface to the CPU.

This would be good food for thought considering the growing popularity of Cu and Cu/Al heat sinks versus simple Al heat sinks, since CU is far more reactive than Al. Plus, I'm hoping electrolysis won't end up being an issue over time in Cu/Al compromises.

-Brad
November 10, 2006 7:49:27 PM

Actually, speaking off the top of my head, I believe Al is way more active than Cu. Now, granted, Cu will form a green scale (think: Statue of Liberty) over time, whereas Al will be more or less stable, or worst case, start a bit of pitting. The reason for this is that all pieces of Al you encounter (aside from very, very esoteric circumstances, like a specially cleaned piece in vacuum or intert atmosphere) have a layer of aluminum oxide on the surface (gamma phase, for those who care, as opposed to the more stable alpha phase which you get when anodizing). This layer of gamma alumina (common term for aluminum oxide) is very thin - think nanometers. Therefore, it isn't a real big deal when trying to figure out heat transfer. However, this layer can grow if exposed to an oxidizing atmosphere (air, or high oxygen air) in the right conditions (heat, probably high humidity would help I'm guessing).

Nearly all metals get a layer of oxide on them; this is bad for some things, but it also serves as a protective boundary for the underlying metal. So, you're going to have a layer of oxide on the surface of pretty much any metal you work with that has been exposed to air.

As long as the layer remains thin, no problem. Think of a coffee mug (this is a ceramic materal, and specifically an oxide ceramic, but I don't know what all normally goes into pottery ceramics). The layers are thick, and heat transfer is poor. So your mug can contain boiling hot water, but you don't get burned. (so if your processor is the coffee, and your fingers are the heat spreader, you're not getting rid of very much heat) If you had a really, really, really thin mug, your fingers would be feeling a temperature very nearly that of the boiling hot coffee. In a metal cup, you'd get burned nearly instantly because the heat transfer is much better.

So, if you're using a copper heat spreader, just remember to keep your area relatively dry (i.e. not a swamp, or in a bay, like the Statue of Liberty). If you start to see that green scale, your thermal emissivity is going down. Polish it off, and you're fine.

The reason this is a real issue for silver is because silver, to a greater extent than many other metals, will oxidize significantly, and in thick amounts. That's why you need to polish sterling silverware every year or so, even if it sits packed away and unused. Conversely, your stainless steel silverware is pretty robust. Unless you have crappy silverware with low chromium/nickel content (stainless on the cheap) and it starts to rust (rust is an oxide of iron), you don't need to polish the stuff.

So, to wrap this up and answer your question, yes, oxidation does matter in heat sinks. And it matters everywhere. The most critical area is definitely going to be your CPU/heat sink interface. That's where the heat is coming from, and every little bit of wattage you block is energy that stays in your CPU. Getting the heat from the processor into the metal core (everything beneath the oxide) of your heat spreader is the most important step. However, fins also matter. If you were using, say, anodized aluminum fins instead of regular aluminum with that aforementioned thin layer of gamma alumina, your thermal transfer would be awful. But, as long as you're not using a computer near the ocean or in a salt mine (more highly corrosive environments), you probably don't need to worry too much about oxide on the fins.

By the way, I'm not sure if you mean a Cu/Al alloy (which I'm not familiar with) or a Cu and Al combination heat spreader, in which case you want to make sure it's brazed together properly (good metal to metal connection). But both Al and Cu should hold up OK in a normal office or indoors environment.
November 10, 2006 8:28:53 PM

Awesome reply, thanks so much for taking the time!

And no, I have never seen a Cu/Al alloy. Some heat sinks being marketed to the OC and pimp rig crowd feature a Cu core and interface surface to take advantage of Cu's better thermal coefficient, and fasten it to Al fins in order to reduce weight (as much as a couple hundred grams) and cost. I doubt any of these are brazed - more likely riveted or using screws and compression fit. Of course, the brazing material would add yet another variable, wouldn't it...

Also very interesting remark about anodizing, since a few HSFs on the market do in fact use anodized Al. I'm also thinking this is pretty interesting due to the popularity of black anodized heat sinks I've seen in so many applications I can hardly count.

Thanks,
-Brad
November 10, 2006 10:45:13 PM

You have succeeded in drifting off on a tangent and entirely avoiding reality because you have taken a simple concept and extended it beyond it's reasonable conclusion without testing along the way.

Oxidation does not matter on heatsinks to any practical extent. It is fairly easy to see that one can take an oxidized copper or aluminum 'sink in all but the very worst of (normal) operating environments and achieve no benefit whatsoever from polishing it... except of course the base, but there was never an issue about the base because it is relatively slow to react to oxygen while interfaced with the CPU and coated with grease- It merely had to be as flat as reasonable prior to installation.

Polishing may actually REDUCE heatsink performance slightly on the fins because it reduces surface area. On the other hand, it may reduce dust accumulation so in longer term use may do as well on average. Remember we're talking about normal levels of oxidation, as seen in a (computer system?) that would otherwise be operable in same environment unless stated otherwise (and other extensive measures taken just to keep system running far beyond consideration of the surface of a heatsink if an environement were so hostile that this is significant).

The most crucial element of a typical heatsink is base conduction efficiency (all else equal besides the oxidation of surface). If the fans have a minor amount of oxidation, or anodization, they will merely rise a few degrees (if that, probably no more than fractions) higher- with similar enough transfer rate away from the hot component, and having the slightly higher temp relative to the ambient medium (air) will have similar enough transfer rate.
November 11, 2006 2:08:53 AM

A few years ago I was reading several reviews for a silver-base heatsink made by Swiftech. Back then, heatsinks were smaller, yet the Swiftech already was on the pricey side for an extra 2-3 degrees of cooling. I'm guessing people found out a good copper heatsink was just as effective and easier on the wallet.

About AS5. It is almost all thermal conductors; the 1/1000 part oil is a suspension material to keep the metals from turning solid. Pure oil would be an awful thermal compound and would cause a CPU to run warmer than with pure water.

But the thermal conductivity of AS5 isn't a big deal considering how thin the layer is. Rather, we're interested in the maximal interfacing surface it gives us for the heat to pass through. The conductivity is just a safeguard in that should we apply a thick layer of AS5 (perhaps due to rough mating surfaces), we'd still see good thermal conductivity.
November 11, 2006 3:23:43 AM

Quote:

About AS5. It is almost all thermal conductors; the 1/1000 part oil is a suspension material to keep the metals from turning solid. Pure oil would be an awful thermal compound and would cause a CPU to run warmer than with pure water.

But the thermal conductivity of AS5 isn't a big deal considering how thin the layer is. Rather, we're interested in the maximal interfacing surface it gives us for the heat to pass through. The conductivity is just a safeguard in that should we apply a thick layer of AS5 (perhaps due to rough mating surfaces), we'd still see good thermal conductivity.


Well, as long as you use enough oil...... http://www.tomshardware.com/2006/01/09/strip_out_the_fa...

Actually, I think all those metals are solid in the suspension; just fine particulates. And technically, most of them are ceramics, not actually metals (though I believe the silver is indeed in metallic form). Though on second thought, I'm thinking you meant "stick to each other" instead of "turning solid."

btw, unlike Cu and Al, silver WILL oxidize under normal conditions, and significantly. Don't buy silver heat sinks for anything you need to keep cool more than a couple months.

And the thermal conductivity of AS5 actually is somewhat of a big deal. You're right that the thinner it is, the less the actual thermal conductivity matters. But that doesn't mean the thermal conductivity doesn't matter. Your transfer goes as follows:
Chip (actual silicon) ---> walls of CPU (metal) ---> thermal compound ---> heat spreader material ---> air, normally via fan, or some other fluid (like a liquid in liquid cooled rigs)

I'm not sure of everything inside the actual CPU unit, but I imagine the thermal transfer kinda sucks between the actual silicon and the metal. Dunno how this is bonded (anybody else have insight here?). The surface of the CPU and the surface of the heat sink both play significant roles in the equation, and it isn't just surface area here. A very rough surface gives you more surface area, but usually not better thermal conduction. This is for two reasons, I believe. 1. Surface roughness makes it difficult to mate properly to another surface. Little bumps mean it's tougher for a thermal compound to get into the little crevices in between. So, if there are any gaps (bubbles) where the surface of the processor isn't in contact with the thermal compound, you've gone from conductive heat transfer (best) to convective (much worse - think how heat travels in air from one room in your house to another. Very slow) and radiative (worst of all). Whenever possible, you want 100% conductive heat transfer through only highly thermally conductive materials. 2. Probably to a lesser extent in your heat transfer problem, a rougher surface is going to have more surface oxides (or, more volume of surface oxides) between your two base metals. To visualize this, take a thick marker and draw a jagged line and a straight line. Think of the mark the marker lays as the oxide. You can see the path the heat has to take through the rough surface looks a little longer.

Now, before anybody goes polishing their processors and heat sinks to kingdom come to try to get perfectly smooth surfaces to mate against each other, the flatness of your parts is going to be way more important than the surface finish - if you manage to polish either the processor top or the heat sink mating face with any degree of concavity (VERY easy to do unless you're using a dead-flat polishing tool), you're making your problem bigger by making the gap between processor and heat sink bigger. Unless you're pretty good at metallurgical preparations and have the right tools, you're probably better off just wiping with some isopropanol (note: not acetone; it leaves a residue which adds another barrier for the heat) and attaching everything.

As far as Cu+Al heat sinks, if they aren't brazed, steer clear. Riveting metals together does almost nothing to help you. If they aren't brazed, all you have is a block of copper that looks fancy. Also, I'd definitely avoid anodized aluminum. Anodizing significantly increases the layer of oxide on the surface (from tens of nanometers to as much as a couple hundred microns), plus it adds paint. You're essentially insulating your heat sink.

I: with regard to oxides on heat sinks: you're right that you should never need to worry about oxides except on the base in normal conditions. But just remember: people do crazy things to computers (my favorite story is when a friend of mine accidentally dropped powdered laundry detergent into her desktop - corroded everything). The base connection with your CPU should never oxidize beyond the condition it's in when you attach it - with thermal compound in the way, no air can touch, so nothing is going to happen to it. Definitely make the effort to clean the base before using it, though. Machine oils, your finger oils from touching it, dust, and any residues will decrease the efficiency of the thermal transfer.
November 11, 2006 5:42:49 AM

Here's my take on this:

1) THG did once cool an entire computer with cooking oil and alternately with distilled water. The cooling method used was convection, not conduction, as oil is a heat insulator.

2) AS5 has solid particles measuring in the ballpark of one micrometer in diameter. However, those particles alone would constitute a very fine powder. To properly bridge mating surfaces, that powder has to be turned into a sticky paste.

3) Cu, Al, and Ag all oxidize. That does not make any of the materials unsuitable for use as a long-term heatsink base because there is no supply of oxygen to the critical surface between the CPU and the heatsink base in a regular installation. However, of the three, only silver is unsuitable fin material for the obvious reason that heavy oxidization through pollutant catalysis would damage the structural integrity of the fins. Copper has no such integrity problems past one decade in a normal computer environment.

4) Thermal conductivity of AS5 in a thin layer application is trivial. The path from the CPU to the fins typically consists of: ~1-2mm of silver-based solder/AlN, copper, and nickel plating to the IHS surface; ~0.1-0.5mm of thermal compound; ~5-10mm of heatsink conducting metals (copper, aluminum, silver, possibly a specialized heatpipe). You see that the conductivity of AS5 is dwarfed at least 10-20x by the much more important conductivity of the heatsink base material. It also explains why using peanut butter as a thin-layer interface won't fry a CPU, whereas placing a large blob will most certainly cause it to throttle or burn up.

5) Anodized aluminum makes for a much more durable heatsink fin surface, as pure aluminum is very soft. Again, the oxide layer is so thin that its six-fold drop in thermal conductivity from pure Al is trivial.

6) Under the IHS, thermal conductivity of solder material can approach 70-80% that of pure aluminum, and the IHS material is copper itself. Regardless of the solder properties, it's a thin enough layer such that there is no great benefit obtained by prying the IHS off and trying to cool the core itself.

7) Only a disreputable heatsink manufacturer would join different metals by riveting. Is that why some of those heatpipe designs perform so poorly compared to others?
November 11, 2006 6:00:25 AM

WORD 8)

Your avatr reminds me of the movie IT
November 11, 2006 6:30:29 AM

on a slightly different note.. which is better for OC, arctic silver or ceramanique?
Anonymous
a b à CPUs
May 20, 2010 2:55:58 PM

Calls for calm in South Korea ship row

By Mark Willacy and Stephen McDonell

Updated 1 hour 34 minutes ago
The stern of the South Korean warship PCC-772 Cheonan

The stern of the South Korean warship PCC-772 Cheonan is placed on a barge after it was sunk two months ago. (AFP: Byun Yeong-Wook)

* Video: North Korea blamed for warship sinking (The Midday Report)
* Video: Nth Korea threatens war over torpedo claims (7pm TV News NSW)
* Audio: Report finds North Korea sunk South Korean ship (The World Today)
* Audio: Korea Investigation sparks threats of war (PM)
* Audio: Beijing accepts validity of investigation into South Korean ship sinking (PM)
* Related Story: 'Act of aggression' sank Korean ship

Australia is urging South Korea to show restraint after an investigation found the communist North was responsible for sinking one of Seoul's warships.

A report by a multi-national team of investigators found a torpedo fired by a North Korean submarine was responsible for the sinking of the 1,200-tonne corvette Cheonan.

Australia, Britain and the United States are among the nations which have already condemned North Korea for what has been described as a deliberate naval attack.

Forty-six sailors died when the ship was torn in half two months ago.

South Korea's president Lee Myung-bak has vowed resolute counter-measures, which will most likely include seeking tough new sanctions against Pyongyang.

North Korea has responded with heated denials and threats of all out war.

Speaking in Tokyo, Foreign Minister Stephen Smith counselled Seoul against a military retaliation.

"That of itself would not be in our view in South Korea's interest, or in North Asia's interest, or in the interest of our region the Asia Pacific," he said.

Mr Smith says the international community still needs to send North Korea a message.

"It's a flagrant breach of their obligations to international law, a flagrant breach of their obligations to the international community and in breach of their armistice agreement," he said.

"So we condemn it absolutely and stand shoulder to shoulder with South Korea."

Seoul is almost certain to ask the United Nations Security Council for tough new sanctions against North Korea.

South Korea's vice-minister for foreign affairs and trade, Chun Yung-Woo, says his country will not necessarily react with military might.

"I think there are different ways of making them pay a price for their misbehaviour, even without direct military action," he said.

"So there is a whole range of options we can work with other members of the international community, our allies and friends and international organisations."

The Chinese Government says it is studying the report into the incident.

China's foreign ministry spokesman Ma Zhaoxu is urging all sides to "stay calm" and "exercise restraint" so as to avoid an escalation of the developing crisis on the Korean Peninsula.

He also appears to be giving the South Korean report some credibility.

"China welcomes a scientific and objective approach [to this incident]," he said.

Mr Ma added that China was currently assessing the report to form an official view.

As the only country in the region with significant leverage in North Korea, China is coming under considerable international pressure to help find a way to deal with their North Korean neighbours.

International investigation

The team investigating the incident included five investigators from the Australian Navy.

It concluded that part of the torpedo found on the sea floor carried lettering that matched a North Korean design.

The communist regime in the North has rejected the report as a fabrication.

One of the heads of the Korean investigation team, Yoon Duk-yong, says there is no other possible explanation for the sinking of the ship.

"Based on all such relevant facts and classified analysis, we have reached the clear conclusion that the South Korean ship was sunk as a result of an external underwater explosion caused by a torpedo made in North Korea," he said.

"The evidence points overwhelmingly to the conclusion that the torpedo was fired by a North Korean submarine. There is no other plausible explanation."

North Korean radio said its National Defence Commission would conduct its own investigation.

"As the group of traitors declared that the sinking of the warship is linked with us, the National Defence Commission of the Democratic People's Republic of Korea will dispatch an inspection group to the area in South Korea to verify material evidence concerning the linkage," the radio report said.

Tags: government-and-politics, foreign-affairs, world-politics, unrest-conflict-and-war, north-korea, south-korea

First posted 1 hour 42 minutes ago
!