Slashdot Mirror

While using decoy-state in quantum key distribution is a fairly new (from 2003 by Hwang) and useful idea, the fact that the Toshiba guys have implemented it is hardly a breakthrough. They are not even the first to do it. A group in Toronto did it more than a year ago (http://www.arxiv.org/abs/quant-ph/0601168).

Second, the decoy state method doesn't improve the level of security (it can be secure without using decoy), but the distance and rate at which one can do secure key distribution improves greatly.

Third, scientists have known since the early days of the laser that laser diodes sometimes emit a variable number of photons (for those of you who know statistics, the photon number in a pulse is random with a Poissonian distribution). This has been taking into account when making security proofs.

Except from those points, the article is not too far from the truth. Using decoy states in quantum key distribution is a very nice idea. Using it requires one extra component, a variable attenuator, and a software update. And the typical secure distance is increased tenfold (from ~10 km to ~100 km) to almost the same distance one would have for a single photon source.

While using decoy-state in quantum key distribution is a fairly new (from 2003 by Hwang) and useful idea, the fact that the Toshiba guys have implemented it is hardly a breakthrough. They are not even the first to do it. A group in Toronto did it more than a year ago (http://www.arxiv.org/abs/quant-ph/0601168).

Second, the decoy state method doesn't improve the level of security (it can be secure without using decoy), but the distance and rate at which one can do secure key distribution improves greatly.

Third, scientists have known since the early days of the laser that laser diodes sometimes emit a variable number of photons (for those of you who know statistics, the photon number in a pulse is random with a Poissonian distribution). This has been taking into account when making security proofs.

Except from those points, the article is not too far from the truth. Using decoy states in quantum key distribution is a very nice idea. Using it requires one extra component, a variable attenuator, and a software update. And the typical secure distance is increased tenfold (from ~10 km to ~100 km) to almost the same distance one would have for a single photon source.

entwined pairs of course! duh!

I was always under the impression that superclusters would stay with w>-1. The paper I cited doesn't seem very confident about that outcome. Since superclusters only slow expansion, rather than resisting it completely, there could be a critical size beyond which an accelerating expansion cannot be resisted. Indeed, this paper purports to do such a calculation (in a spherical approximation), and claims that existing definitions of "superclusters" are vague and are not always fully bound.

So why not apply Ockham's razor? Things usually tend to make more sense when you do that. A great example is its application to QM interpretation by Mohrhoff: http://aps.arxiv.org/abs/quant-ph/9903051

I find it a very attractive feature of Mohrhoff's interpretation of QM that he completely does away with causality at a fundamental level. http://aps.arxiv.org/abs/quant-ph/9903051

That's a different paper; the one cited in the story is posted on arXiV here http://arxiv.org/abs/hep-th/0610213 The difference between the two papers is the causal patches you mention. The causal patch concept isn't in the former. The latest paper takes care of the entropy problem caused by a contracting universe by distributing entropy among each causal patch. Just before the Big Rip, expansion reverses and the patches contract into an infinite number of separate universes, each taking a little bit of entropy. One patch becomes/stays our universe. These patches are called causal patches because they don't interact with each other -- no light, gravity or other force can bridge the space between the patches. There's a press release about this http://www.unc.edu/news/archives/jan07/newmodel012 907.html

I don't wholly agree with your analysis. An atom is an electromagnetically bound system and resists expansion; you can't just take a straight Friedmann GR solution and make conclusions from that. That's why the Big Rip scenario was invented: it is an acceleration so severe that even bound systems become unbound by the cosmological expansion. A cosmological constant (w=-1) is right on the boundary: atoms become unbound only asymptotically. There is no actual finite time at which atoms (or galaxies) fall apart; a simple analysis on the basis of comparing the system size to the Hubble radius doesn't work. What actually happens is that space in the vicinity of the atom doesn't expand as fast as the rest of the universe; it's the same reason why your atoms aren't expanding with the universe right now.

See the Big Rip paper, page 2.

I think the work being referred to may be in this paper, in which the universes are "causal patches" which are disconnected from each other causally by the Big Rip.

Mmh, after a brief look at the blog, i hope their use of a "tunable flux transformer" ( http://arxiv.org/pdf/cond-mat/0608253 ) is only a coincidence... (=

This kind of work is not something you take on by looking it up on a general encyclopaedia like Wikipedia. If you're at a point where you can actually make an attempt on such a problem, you're probably already familiar with specialist literature and you more likely than not have heard of the problem long ago and not yet tackled it.

This would be a better place to start:
http://arxiv.org/

If you can't even understand the papers here in the field you've chosen, you've got a lot of work to do and it may even be easier to pursue it formally as part of a postgrad degree.

The myth that you can just walk into a problem and solve it is rubbish. Einstein may have been a patent clerk when he had his breakthrough "miracle" year but he was looking at problems for many many years and got to know a lot of mathematical and scientific literature in a less than formal setting which is one reason he was able to see past all the old thinking and realise that things he was seeing (notably the Lorentz transformations/Michelson-Morley experiment) were literally true.

"When the published and unpublished trials [from a FOIA request] were pooled, the placebo duplicated about 80% of the antidepressant response; 57% of these pharmaceutical company-funded trials failed to show a statistically significant difference between antidepressant and inert placebo. [...] This modest efficacy and extremely high rate of placebo response are not seen in the treatment of well-studied imbalances such as insulin deficiency, and casts doubt on the serotonin hypothesis."

The article is also available at: http://lanl.arxiv.org/abs/cond-mat/0607706

Full article available on the arXiv since July 2006

Still do skyflats :-) but it does depends on what passband you care about for imaging - twilight sky flats work pretty well in B. These are sort of bothersome on larger telescopes because you don't want to saturate but you do want good statistics but you don't want to cut into observing time, and you have to slew between each one to reject any bright early rising stars. A lot of big telescopes use quartz lamps to illuminate a screen and image that. Dome flats are pretty common these days, especially in spectroscopy, but for photometry its nice to still get a set of sky flats. I take a bunch of flats for each instrument setup and median them before flat fielding. There are more sophisticated methods around the corner that will vastly improve calibration for projects like Pan-STARRS and later LSST - http://arxiv.org/abs/astro-ph/0609260 (Disclosure: I work with some of the people on said paper but not on this project)

Dark frames aren't actually as useful anymore for instruments on larger telescopes that use LN2 or a cryotiger for cooling.

This is not quite true - it depends heavily on the particular journal, and the particular author. I've run a fairly decent-quality online journal [jmir.org] for the past few years, and I can attest that the copyediting, typesetting and formatting aspect of publishing is the most onerous and time-consuming.

These things improve the reading experience, and I appreciate them, but I don't think they're truly necessary. The arXiv does just fine with an almost fully-automated system. Aside from the occasional obvious crackpot, the scientific quality is comparable to the second-tier physics journals (I'm thinking of the Elsevier journals like Physics Letters).

Right now, the arXiv's goal is not to replace journals, but to be as fast and as efficient as possible at hosting preprints. If they set out to replace journals outright, and added a slighly more sophisticated review system and automated validation of input, they could do what most journals do with about 90% of the quality for about 10% of the cost. The remaining 10% could probably come from advertising, donations, government grants, or some combination thereof. Alternatively, you could sell "2-week advance notice subscriptions" for cheap, and make access free to anything over two weeks old.

The fact is, there is a great range in submission quality (in both content and layout), and for a fairly advanced journal that generates XHTML and PDF versions of articles from a standardized XML format (the one Pubmed Central uses), this can take on average 2-4h of copyediting and layout work per article.

I don't know too much about XML journal article formats, and would appreciate more info. PubMed's explanation didn't really help much. Why bother with XML? Why not just require authors to submit in LaTeX? Then you've got PDFs or DVIs or whatever "camera-ready" output you want with no further human interaction. If you combine it with a package like RevTex 4, you get standardized appearance and easy harvesting of meta-data. LaTeX predates XML and SGML (if you count SGML's age from when it was formalized as a standard). Why re-invent the wheel for publishing? My best guess so far is to accomodate authors who can't be bothered to learn LaTeX and insist on submitting Word docs, but that's about all I can think of. Why not just tell the Word users to suck it up and learn LaTeX?

Subscription-only journals aren't dying: they're dead. See http://www.arxiv.org/ or http://jhep.sissa.it/ for more details. The fact that Nature etc. have left it until 2007 (!) to seriously worry about this just shows how little chance they have.

"Solar Resonant Diffusion Waves as a Driver of Terrestrial Climate Change".

Click on the link at the end of TFA and you'll see the real scientific paper on which this was based.
Lots of nice graphs at the end of it.

All based on 18O/16O ratios

It's all here on the archive in glorious pdf-ed Latex.

http://arxiv.org/PS_cache/astro-ph/pdf/0701/070111 7.pdf

As I said, the paper is not specific to string theory and makes no calculations within the context of string theory. It gives tests of any generic theory that obeys unitarity, analyticity, Lorentz invariance, and crossing. That includes both string theory and quantum field theory as subsets. So no, it is not a string theory paper: it is more general than that. But it does have implications for string theory.

Try reading the paper yourself. You would save yourself less embarrassment that way.

Here is the pdf of the paper on arXiv for those interested.

http://arxiv.org/abs/hep-ph/0604255

Gravitation has been measured to travel up to 1.2 times faster than the speed of light, although these measurments are not agreed upon by the entire scientific community.

See http://arxiv.org/abs/astro-ph/0302294 for statement.

It has been measured, see http://www.arxiv.org/abs/gr-qc/0501001 . The result of this measurement - speed of gravity is between 0.8c and 1.2c, which is consistent with gravity propagating at 1c.

If you're talking on the order of a 6 month embargo, I could see that. If you're talking two years, I'd have to say that's too long.

I may be biased on the matter, as one of my duties is to distribute some public research data. The data that we generate is released immediately, except for new missions, which have had embargos until they could finish testing the instruments. The data we get from other locations may be embargoed for a few months.

For those who are new to the topic, I'd suggest you take a look at the OpenScience Project, the Science Commons, the NSF's 2003 Cyberinfrastructure Report, the NSB's Government Funding of Scientific Research and the Astronomer's Data Manifesto for a bit of background (specifically, see Ray Norris's Can Astronomy Manage its Data?).

However, in the last 3-5 years,some very respectable and highly cited open access journals have come up - check out www.plos.org or Biomedcentral http://www.biomedcentral.com/ - they are open access publishers who don't charge for access - instead, they charge the authors for the publication costs.

For Physics, Mathematics, CS and Quantitative Biology:

Arxiv

It's esily verificable because some possible experiments has been already proposed:
http://en.wikipedia.org/wiki/Heim_theory

This theory also succesfuly predicts masses of all known and some unknown particles. The predicted masses have been derived by Heim using only 4 parameters - h (Planck's Constant), G (Gravitational constant), vacuum permittivity and permeability. For example theory predicted that neutrino got mass in 1980s - long before it has been found by experiment.

Also another prediction of gravitomagnetic force (Heim-Lorentz Force) proposed here:
http://www.hpcc-space.com/publications/documents/A IAA2005-4321-a4.pdf

seems to be validated by ESA (European Space Agency) experiments:
http://www.esa.int/SPECIALS/GSP/SEM0L6OVGJE_0.html
http://www.hpcc-space.com/publications/documents/A rtificialGravity.pdf
http://arxiv.org/pdf/gr-qc/0603033

This theory also explains the "Dark Matter" phenomenon.

If Heim theory is true then it will mean that we will most likely be able to travel with superluminar speed and produce artificial gravitational fields (antigravitation):
http://www.hpcc-space.com/publications/documents/A IAA2006-4608LetterExtndVersionRevised.pdf

It's much easier to test then string theory, so why do not to direct some extra attention and money for further testing it??

The swampland is a term new to me, all the universes that almost could exist but under string theory are inconsistant and mathematically sink into the mire. Still the landscape (they term for the 100's of millions of universes that string theory can describe, is so big string theory can't make many, if any, predictions about our universe. On the other hand, DSR is starting to look pretty cool. Only recently these guys: http://arxiv.org/abs/hep-th/0701113 rewrote the Dirac equation (the equation for electrons and other fermions) in DSR. And suddenly no more off shell virtual particles. Instead of the energy momentum 4-vector not squaring to the mass (which is what off shell means), the particle stays on a 5d light cone, so instead of borrowing energy, the particle just side steps through a mass like 5th dimension.

Actually, the speed of light in a vacuum is not constant at all, according to several current theories. Professor R.T. Cahill's process physics theory(i mention this one, cause i've had some lectures on it, but there are others) states that the speed of light is actually inconstant, and depends on the flow of space around it. I don't claim to understand it, being a humble chemist, but it's interesting stuff http://www.arxiv.org/abs/gr-qc/0203015 for a cahill paper

Sorry, screwed up the link. The arXiv preprint for the paper that TFA mentions can be found here:

http://arxiv.org/PS_cache/physics/pdf/0608/0608300 .pdf
(warning: pdf)

Amazingly, Roland actually gave the link to the arXiv paper at the end of his writeup.

Well I am a chemist... and in fact my Ph.D. thesis had alot to do with these kinds of chromophores!

Yes the molecules in question are "azobenzenes" (benzenes linked via N=N) and "stilbenes" (benzenes linked via C=C). These are well-established classes of molecules that have strong "nonlinear optical" properties.

The reason they are "nonlinear optical molecules" is because (in basic terms) the electron distribution is highly asymetric. You can see the chemical structures in the arXiv preprint (pdf). One end of the molecule has a group that 'attracts' electrons, and the other end has a group that 'donates' electrons, and the end result is that the electron distribution is strongly skewed. This means that when light hits the molecule, the electron cloud oscillates not like a normal sine wave (harmonic oscillator) but in a much more skewed way (think of a sawtooth wave). This means that when it re-emits light, that light can be very different from the incident light.

That's why these molecules can be used as amplifiers in lasers, and "frequency doublers" (where you input a certain frequency of laser light, and what comes out has double the frequency (i.e. half the wavelength)). They are remarkable molecules, really. This new paper is certainly noteworthy, but I'm not sure it's going to revolutionize the world of telecommunications anytime soon...

Quantum cal-culations using sum rules have been used to place an upper-bound on the molecular susceptibilities; [1, 2, 3, 4] but, the largest nonlinear susceptibilities of the best molecules fall short of the fundamental limit by a factor of 10^(3/2).[4, 5] A thorough analysis shows that there is no reason why the molecular hyperpolarizability can not exceed this apparent limit.[6] In this letter, we report on a novel set of molecules where the one with modulated conjugation[7] is found to have a hyperpolarizability that breaches the apparent limit.

(skims through comments)

What, no-one's mentioned slacktitude yet?

Basically, slacking off now on long computer/tech projects means you can take advantage of newer, faster, tech down the track to finish the job quicker, thus winding up finishing at the same time as you would have in the first place.

Interesting comment, Satanic Puppy. Can you clarify one thing for me--when you say that "having free will" doesn't mean anything, do you mean that the question has no meaning because it can never be decisively answered, or that, if we somehow magically got the answer, that it would be unimportant?

If it's the former, I'm in agreement with you for the following reason: we're much better off believing we have free will and being wrong, than believing we don't have free will and being wrong. If there is no free will, there can be no morality, and it doesn't really matter (in a moral sense) what we do. Jailing someone for something a brain tumor made them do would be a bummer for them, but of no moral consequence. On the other hand, if we have free will but we act amorally because we believe otherwise, we may do all sorts of immoral things.

If you were claiming the latter--that even an answer to the question of free will would be irrelevant--I can provide a counter-example. Recently, John Conway and Simon Kochen (of the Kochen-Specker Theorem) published a paper claiming that, for a minimal definition of free will (amounting to little more than non-determinism), humans must have free will or else the Aspect experiment and other test's of Bell's Theorem have little meaning.

In other words, if humans have "free will" (even just in the limited sense described by Conway and Kochen), the Aspect experiment and others like it show that reality is described by quantum mechanics or some other non-local physics. If we don't have free will, the experiments have no meaning, and physics could still be local.

On another track, if you were somehow presented with proof of the non-existence of free will, would you continue living your life the same way (and please avoid the glib "I'd live it however was pre-determined" type remarks)? It seems to me that any rational person would have to respond by ditching morality and living a hedonistic life. Sociopaths would be the new role models, as they would be free of all the now-irrational constraints of morality. This is why it's better we assume (wrongly, perhaps) that we do have free will...

You can read the technical papers on which this press release is based:

http://arxiv.org/abs/astro-ph/0612445

http://arxiv.org/abs/astro-ph/0612447

The basic idea is that the astronomers used an infrared
space telescope to take very deep images. They then tried
to remove all the obvious sources of light, and examined
the resulting "blank" images very carefully. They claim that
there are very faint sources of infrared radiation which
remain, and that the spatial correlation of these sources
is roughly what one would expect if they were young galaxies
in the very early universe.

There are limited opportunities for other astronomers
to examine the same regions with other telescopes and
at other wavelengths; that could provide evidence that
might support the claim, or weaken it (if, for example,
radio telescopes detect some of these sources and
show that they are ordinary galaxies in the relatively
nearby universe, that would weaken the claim in
the press release).

We can also just wait a decade or so for JWST, a more
powerful infrared space telescope, to observe the same
field.

You can read the technical papers on which this press release is based:

http://arxiv.org/abs/astro-ph/0612445

http://arxiv.org/abs/astro-ph/0612447

The basic idea is that the astronomers used an infrared
space telescope to take very deep images. They then tried
to remove all the obvious sources of light, and examined
the resulting "blank" images very carefully. They claim that
there are very faint sources of infrared radiation which
remain, and that the spatial correlation of these sources
is roughly what one would expect if they were young galaxies
in the very early universe.

There are limited opportunities for other astronomers
to examine the same regions with other telescopes and
at other wavelengths; that could provide evidence that
might support the claim, or weaken it (if, for example,
radio telescopes detect some of these sources and
show that they are ordinary galaxies in the relatively
nearby universe, that would weaken the claim in
the press release).

We can also just wait a decade or so for JWST, a more
powerful infrared space telescope, to observe the same
field.

The journal articles that go along with the story:

New Measurements of Cosmic Infrared Background Fluctuations from Early Epochs
On the Nature of the Sources of the Cosmic Infrared Background

(These were posted in the article, but only under a tiny "More info" link at the bottom that is easy to overlook.)

The journal articles that go along with the story:

New Measurements of Cosmic Infrared Background Fluctuations from Early Epochs
On the Nature of the Sources of the Cosmic Infrared Background

(These were posted in the article, but only under a tiny "More info" link at the bottom that is easy to overlook.)

Also, there's a paper available that provides more details.

http://arxiv.org/abs/physics/0607135

From back in August?

There's a 2004 analysis by Columbia University of how Skype does its sneaky business here: http://arxiv.org/ftp/cs/papers/0412/0412017.pdf. The JANET network in the UK didn't like what happened when it allowed a PC to become a supernode, as you can read here: http://www.ja.net/development/voip/skype&janet.pdf . They found in 24 hours it had talked to over 38,000 different IP addresses, maintaining 'over 660-690 open connections to different hosts'.

You keep repeating that things like photons, electrons and the like are "merely models". I have to take issue with this, as they happen to be effective models.

I would *love* to see how you would *begin* to explain how light and matter interact at a *fundamental* level, without using the concept of electrons and photons.

These guys are not cranks - the (free, as in beer) preprint seems to be a pretty typical quantum transport paper, albeit with a slightly "sexed up" angle.

Models are good, if they work.

But expansion is not the only answer to the red-shift in light.

It's the only answer so far proposed that works. The alternatives you suggest do not.

The first answer most people give is stellar motion (expansion). Which is the loss of light energy and its wavelength stretched or red-shifted from the light source moving away.

In general relativity, redshift due to expansion of space can't be completely disentangled from velocity Doppler shift. However, it is known that the cosmological redshift cannot be due to galaxies merely moving away from us in a flat, non-expanding spacetime; among other things, an expanding universe predicts different luminosity-redshift curves than a non-expanding one (guess which we observe).

There is also gravitation, in which the star's own gravity may lengthen the wavelength of the light and/or from starlight passing near a massive object, like a galaxy.

Cosmological redshift is not due to gravitational redshift of the source star; the latter has a very specific value determined by the star's mass and radius, which is not obeyed by cosmological objects such as distant galaxies. Cosmological redshift is also not due to gravitational redshift of gravitationally lensed light; the redshift of the latter is essentially zero (cancelled out by an equivalent blueshift as the light approaches the lensing body).

Second-order doppler effect (Transverse doppler shift) can also explain the red-shift. As a light source moving at right angles to an observer will always be red-shifted. And when you use this for an answer, it implies that the universe may be in circular motion instead of expansion.

Nope. We don't see any such anisotropy in the universe. (Consider that there will be two antipodal points exhibiting zero redshift, not to mention a general angular dependence of redshift that is not observed.)

Furthermore, none of these explanations can account for the cosmic microwave background radiation, even if they worked for Hubble redshifts.

As far as there being a beginning, it's just a matter of thermodynamics, as someone else said everything is falling apart which means it had to start at some point because if it was infinite all energy would be at equilibrium, meaning it would be the same everywhere, which is just not the case.

Your argument fails. See, for instance, the cyclic model of Steinhardt and Turok, specifically how local entropy density does not build up from cycle to cycle in their model. (This does not violate the second law of thermodynamics; total entropy does increase, but it gets "diluted" in such a way that local universes can continually be produced.)

hey haven't even done IN VITRO modeling yet.

It's easy to pin models down with in vitro modelling in simple systems, harder to do with cancer, except in a largely qualitative way.

It's not as obvious as "10^20 > 10^12, therefore cosmic rays would have created them by now".

The amount of energy available to create new particles is governed by the total energy of the incoming particle and its target, as measured in their center of momentum frame. The LHC is a collider, so all of that 7 TeV is available to create particles. But an incoming cosmic ray of energy 10^20 eV striking an atmospheric atom has an energy in the cosmic ray-atom center of momentum frame which is proportional to the square root of its energy in the Earth's rest frame (see here); i.e., on the order of only sqrt(10^20) or 3 x 10^10 eV.

So, crudely, the LHC at energies of ~10^13 eV can be more effective at producing new particles than ultra high energy cosmic ray (UHECR) atmospheric showers at 10^20 eV.

However, I've read that cosmic rays with center-of-momentum energies of ~10^14 eV have been observed, so I think those proportionality constants I'm ignoring will turn out to matter, and that UHECR particle production energies are more significant than LHC (but only by an order of magnitude or so). For a more detailed discussion of black hole production, try Landsberg's review, or Ringwald and Tu.

Incidentally, for doomsday worriers, Hut and Rees showed back in 1984 that it is likely that two such cosmic rays have struck each other in our past (not the Earth's atmosphere) — essentially an ad-hoc collider experiment — which means a center of momentum energy ~10^20 eV. So we're safe from the universe-destroying scenarios (like tunneling out of a false vacuum), since our universe hasn't ended yet, but not necessarily from black hole production (which we would not have necessarily noticed before now).

It's not as obvious as "10^20 > 10^12, therefore cosmic rays would have created them by now".

The amount of energy available to create new particles is governed by the total energy of the incoming particle and its target, as measured in their center of momentum frame. The LHC is a collider, so all of that 7 TeV is available to create particles. But an incoming cosmic ray of energy 10^20 eV striking an atmospheric atom has an energy in the cosmic ray-atom center of momentum frame which is proportional to the square root of its energy in the Earth's rest frame (see here); i.e., on the order of only sqrt(10^20) or 3 x 10^10 eV.

So, crudely, the LHC at energies of ~10^13 eV can be more effective at producing new particles than ultra high energy cosmic ray (UHECR) atmospheric showers at 10^20 eV.

However, I've read that cosmic rays with center-of-momentum energies of ~10^14 eV have been observed, so I think those proportionality constants I'm ignoring will turn out to matter, and that UHECR particle production energies are more significant than LHC (but only by an order of magnitude or so). For a more detailed discussion of black hole production, try Landsberg's review, or Ringwald and Tu.

Incidentally, for doomsday worriers, Hut and Rees showed back in 1984 that it is likely that two such cosmic rays have struck each other in our past (not the Earth's atmosphere) — essentially an ad-hoc collider experiment — which means a center of momentum energy ~10^20 eV. So we're safe from the universe-destroying scenarios (like tunneling out of a false vacuum), since our universe hasn't ended yet, but not necessarily from black hole production (which we would not have necessarily noticed before now).

It's not as obvious as "10^20 > 10^12, therefore cosmic rays would have created them by now".

The amount of energy available to create new particles is governed by the total energy of the incoming particle and its target, as measured in their center of momentum frame. The LHC is a collider, so all of that 7 TeV is available to create particles. But an incoming cosmic ray of energy 10^20 eV striking an atmospheric atom has an energy in the cosmic ray-atom center of momentum frame which is proportional to the square root of its energy in the Earth's rest frame (see here); i.e., on the order of only sqrt(10^20) or 3 x 10^10 eV.

So, crudely, the LHC at energies of ~10^13 eV can be more effective at producing new particles than ultra high energy cosmic ray (UHECR) atmospheric showers at 10^20 eV.

However, I've read that cosmic rays with center-of-momentum energies of ~10^14 eV have been observed, so I think those proportionality constants I'm ignoring will turn out to matter, and that UHECR particle production energies are more significant than LHC (but only by an order of magnitude or so). For a more detailed discussion of black hole production, try Landsberg's review, or Ringwald and Tu.

Incidentally, for doomsday worriers, Hut and Rees showed back in 1984 that it is likely that two such cosmic rays have struck each other in our past (not the Earth's atmosphere) — essentially an ad-hoc collider experiment — which means a center of momentum energy ~10^20 eV. So we're safe from the universe-destroying scenarios (like tunneling out of a false vacuum), since our universe hasn't ended yet, but not necessarily from black hole production (which we would not have necessarily noticed before now).

at a cost of $18 unless you are a subscriber

Or get it for free at arXiv...