scovilles Let's talk Scoville scale

[Lucky Dog Hot Sauce]

Yep - same with "Thai spicy" - 5 stars is not their hottest rating. If a round eye orders 5 stars they think, "silly American man" and bring you 3 stars.

Order "Thai spicy" and they will bring you molten lava on a plate. I do not order Thai spicy. I'm a 3 stars kinda guy.

 

[j.t.delaney]

Having worked on some DoD-funded projects, I can assure you such a connection does not automatically imply infallibility. As you have discovered to your own chagrin.

Exactly. I don't care if it's "directly connected" with the Pope -- if you don't have a validation report showing that you know how to make the method work, and you know when it's appropriate, and where it's limitations rest, then it's "exploratory" at best.

I wouldn't necessarily say only hope; we're talking about detecting the concentrations of certain molecules in a matrix of other molecules. I can think of at least a couple of other approaches that might work, but apparently haven't been developed.

Yes, there are papers reporting GC-MS with high-temp columns, are that are folks who have reported MALDI-TOF, and I'm sure you could do something with an immunoassy like ELISA, and I'm sure there are a few others out there that could be developed, but since HPLC is by far the most widely used technique, who wants the trouble of the added method development. Seriously.

So yes, for now, HPLC is the only practical hope. edit: And apparently not all that practical either.

I don't agree with your comments about protocols and think they only reinforce the negative opinions being aired here. I'm not trying to pick a fight here, but please consider this: I think we can all agree that a certain volume or mass of pepper, powder, or sauce contains certain concentrations of a number of different capsacinoids. If a particular lab cannot measure those concentrations within some reasonable degree of accuracy and have the results confirmed by a different lab using a different protocol, then either one or both of those protocols are fatally flawed. One of the hallmarks of good science is repeatability and reproducibility.

I directed an analytical lab for seventeen years (not HPLC but an alphabet soup of other techniques), so I know what you mean about a culture of innumeracy. It disappoints me greatly that this culture has made such inroads into fields that demand the most rigorous levels of numeracy. We can see the results right here in this thread.

Long time members shouldn't have to do the work to troubleshoot. If they submit a sample and pay the fee, they are entitled to expect valid results.

I'm not sure how what you are saying is in disagreement with what I said. If multiple uncontrolled samples are sent to multiple labs, and the labs are told to "run HPLC", but not given a single procotol, then how is that not going to fail spectacularly?

Yes, it's not fair to expect small-scale hot sauce makers to be analytical test method developers, and in a more perfect world, they should be able to open the phone book (or google), look for a certified lab that does Scoville heat unit measurements, and pay them money to get data from a validated test. Aparently, our world is more flawed than that (which signals a business opportunity to me), but that means then that the up-and-coming hot sauce mogul needs to be a bit more savvy in test requests and how that data is to be consumed. If you just ask for HPLC without understanding that there are easily hundreds of thousands of protocols already out there, and an infinite number more possible, well, you just might not end up with what you expect.

One of the interesting things for me from this discussion is that most of these super hot sauces that people are talking about involve concentrations that are well above the normal ppb/ppm range of HPLC. Are they using testing protocols that are range-appropriate? Who knows? I don't, since nobody here is bothering with the details, which tells me the problem is actually a lot more fundamental.

PM me if you'd like to collaborate on a improved method. Invention's my game these days and I have the background. It sounds like you do, too.

Interesting.

I agree that for consumable products, precise quantification isn't critical, but if measurement techniques can be refined to be more reliable, why not quantify? And yes, the military does need to know precise levels, at high enough concentrations, capsaicin can be lethal. But there are numerous other applications where precise measurement is critical as well. Pain creams, animal repellents, bear spray, self-defense spray, all these applications rely on precise compositional analysis.

I want to be clear here that I am talking about quantifying the capsaicinoid concentrations only. Converting those numbers to Scoville units is a whole 'nother ball of bee's wax.

There are people who do need to quantify capsaicin content, and in consumables most of all. Like i said elsewhere, capsaicin is an API in many formulations used as topical pain relieving agents, so it falls under FDA regulation. As a consequence, having the ability to measure these things is more important than just a hot sauce marketing gimmick, and people have spent a good amount of time and money on being able to reliably make those measurements. I know of some contract labs in town that do this, and I suspect there are others elsewere, but if this many people can't get good data, then there is clearly an unserved market.

 

[Sawyer]

who wants the trouble of the added method development. Seriously.

Well, if it leads to a method that avoids the problems being discussed in this thread, I do. Materials analysis is a very actively-evolving field and the best is almost certainly yet to come. Besides the techniques you mention, I'm thinking of more radical technologies such as functionalized oscillating cantilevers, direct precipitation, or microfluidics to name a few. I'm not aware of any papers relating these techniques to capsaicin quantification, but that doesn't preclude the possibility.

If multiple uncontrolled samples are sent to multiple labs, and the labs are told to "run HPLC", but not given a single procotol, then how is that not going to fail spectacularly?

By having the labs understand the implication that a request to "run HPLC" really means "quantify the concentration of capsaicinoids in this sample".

If you just ask for HPLC without understanding that there are easily hundreds of thousands of protocols already out there, and an infinite number more possible, well, you just might not end up with what you expect.

Here, I think, is where we disagree. In my opinion, it is incumbent on the testing lab to understand fully what information the client seeks, in this case, the concentrations of capsaicinoids in the sample. Looking at it from the clients perspective, "Here's my sample. Tell me the concentrations of capsaicinoids in it." is really all that should be required. Clients don't care about protocols and shouldn't have to care. It should be up to the lab to determine which protocol will give accurate results. If the lab doesn't have a protocol to fill the bill, they shouldn't accept the client's business or take their money.

which signals a business opportunity to me

agreed

There are people who do need to quantify capsaicin content, and in consumables most of all.

Perhaps I should have said "ingestible condiments" instead of "consumables", since I was in this context speaking only of sauces, powders, etc. Certainly precise measurements are required for the other applications discussed, not to mention promising potential applications to internal medicine.

I suspect the problems with HPLC being described here have mostly to do with sauces and that the problem lies in sample prep before column injection. It seems like the people in this thread who have had problems are all sauce makers, though I may be mistaken.

 

[ta2edfreak]

This video was posted on another thread but I like it. It doesn't really break down into scoville units but it's an informative video anyway. http://www.pbs.org/saf/1105/video/watchonline.htm

I learned a few years ago about how to order hot food in asian restaurants. When I order pad thai spicy in my local sushi restaurant, the new waitress said it already is. I said I know, tell them to make it spicy. It was. They don't question me now hehe. Normally though, when I hit a new place, I try their version first before I ask for spicy. I have been known to ask for a little extra sauce after my meal comes out.

 

[j.t.delaney]

Well, if it leads to a method that avoids the problems being discussed in this thread, I do. Materials analysis is a very actively-evolving field and the best is almost certainly yet to come. Besides the techniques you mention, I'm thinking of more radical technologies such as functionalized oscillating cantilevers, direct precipitation, or microfluidics to name a few. I'm not aware of any papers relating these techniques to capsaicin quantification, but that doesn't preclude the possibility.

In terms of analysis, the rich vein of HPLC hasn't been mined to exhaustion here. Moreover, there's zero evidence that (fill-in-the-blank) will address the problem better. All we have here is the contention that HPLC is "unreliable", when there's no experimental details given, and no definition of what reliability means. Hell, there's no actual data. If we're looking for a direct, rugged, robust, cheap-to-develop way to giving a reliable measurement of things like hot sauce, HPLC represents the most obvious low-hanging fruit: it already works in plenty of labs, it uses fairly ubiquitous equipment, and existing protocols can be refined to achieve some localized optima. Sure, one could develop a lab-on-a-chip or surface plasmon resonance spectroscopic tool or whatever glitzy gadget you can find in Nature these days, but there's no reason to think that the precision is going to be better, or that it will be cheaper, or that it won't be a huge investment of time and money with minimal return.

But please, by all means: go out there and prove me wrong. Dazzle me. 'Better' is always 'better'.

By having the labs understand the implication that a request to "run HPLC" really means "quantify the concentration of capsaicinoids in this sample".

Here, I think, is where we disagree. In my opinion, it is incumbent on the testing lab to understand fully what information the client seeks, in this case, the concentrations of capsaicinoids in the sample. Looking at it from the clients perspective, "Here's my sample. Tell me the concentrations of capsaicinoids in it." is really all that should be required. Clients don't care about protocols and shouldn't have to care. It should be up to the lab to determine which protocol will give accurate results. If the lab doesn't have a protocol to fill the bill, they shouldn't accept the client's business or take their money.

The cynic in me thinks the labs understood just fine: "gimme a number I can slap on a bottle; it's not really all that important." The test requestors assumed that test method development isn't necessary (from the tone of the discussion, there's no way they'd be interested in covering method development costs), and got back what sounds like embarrasing, extrordinarily cheap and flimsy data -- the kind of first-pass attempt that you get with chromatography if you're mostly after a quick-and-dirty qualitative result. They should have explained the scope and limitations, but somehow, niether party in that interaction bothered with with asking or offering those details.

If you don't care about the method, then why bother caring about the result? You don't get the latter without the former. Why bother with testing at all? The method provides the context for making sense of the measurement. If you are interested in getting a number, you really ought to care what that number actually means, and that involves having some sort introductory level of understanding of what's being tested, and a realistic idea of how meaurements work. Sending a test request to a lab and expecting that the number you get back is going to represent the mystical platonic ideal of reality is absurd (reminder: the "Oracle of Dephi" was a myth -- not a contract lab.) If you think any quantitative measurement is giving you a truly absolute value (and I'm sure you don't), get ready for a big surprise. Real data has error bars, not all chromatography is a fungible commodity, and methods matter. That's why having a trusted, validated protocol run by a vetted lab is important, and that's why it's critical to know the approximate scope and limitations of the technique. Without those sort of details, instrumental analysis is just a fancy magic 8 ball.

If the test lab had claimed to be using a validated method, or that they were following GLP, and they gave still reported crap data like that, I would would certainly demand my money back... but I don't think that's what they asked for. This is why it pays to be a savvy consumer of test data. A report where measurements are given roughly in the form of "N ± n SHU, following protocol XYZ", along with supplementary raw data, a copy of the protocol, and any notes or observations from the analyst would be much more satisfying and would allow for meaningful comparisons.

As a consequence, if the data you get back looks fishy and you paid good money, it's not impolite to call 'shenannigans'. Test method conditions should be part-and-parcel of any decent report, and asking them to show their homework is why you pay for the service in the first place. If a lab has delivered poor service, why keep quiet about it? There are commerical food science labs that offer validated capsaicin analysis by HPLC; it's perfectly good form to ask for these details, since it's what makes the data actually mean something. If I had been given results like the ones people are talking about here, I know what I would do and how I would complain, but that's because I work in an lab. Being assertive and skeptical and asking for these things are the sorts of habits that make people better equipped to deal with situations like these.

Perhaps I should have said "ingestible condiments" instead of "consumables", since I was in this context speaking only of sauces, powders, etc. Certainly precise measurements are required for the other applications discussed, not to mention promising potential applications to internal medicine.

I suspect the problems with HPLC being described here have mostly to do with sauces and that the problem lies in sample prep before column injection. It seems like the people in this thread who have had problems are all sauce makers, though I may be mistaken.

I completely agree. There are plenty of opportunities for matrix effects-- especially involving emulsions of lipophilic alkaloids with all sorts of other bio-detritus.

 

[j.t.delaney]

Incidentally, for anybody interested, Wilbur L. Scoville's famous method just turned 100 years old this month. Somebody earlier asked for a copy, and here it is. It's written with an enjoyable, old-timey feel to it. The varieties of peppers are equally old-timey sounding: "Japan chilliess", "Zanzibar chillies" and "Mombassa chillies".

Happy pepper-squeezin'...

 

[mrz1988]

Thanks j.t.delaney. I'm glad there's someone else here who values the chemistry work going on here and understands that HPLC is a very important method of finding concentrations of all sorts of things, not just capsaicinoids. If DONE RIGHT it is extremely precise. The work still needs to be done to determine which labs will do SHU HPLC testing most reliably, and which ones are just out for a quick buck. All too often peole put too much pressure on the results and scoff at them simply because the science going on for the HPLC testing is all done in a veiled lab. Sometimes the number seems almost made up. The reality is that the limitations occur when labs are not held to a certain standard, and unfortunately the power of the scoville scale to determine heat is quite limited too. If Defcon had a super extract hit only 35K on the scale, either the extract itself purified something that the lab wasn't prepared to test for, or the lab wasn't held to very high standards of consistency/accuracy. It says nothing about the accuracy of HPLC, only about the reliability of the lab that did the testing. Just like anything else, there are people out there trying to make a quick buck off a cruddy product, and quite frankly there are a lot of folks that will pay for a cruddy number.

I would hold an enormous amount of respect to any hot sauce bottler whos label had a read out of the ppm of each individual capsaicinoid. THAT is the only way to have high scientific accuracy for heat measurements. I'd imagine the testing would be quite expensive though and would need to be performed by a trusted lab.

 

[Lucky Dog Hot Sauce]

I would hold an enormous amount of respect to any hot sauce bottler whos label had a read out of the ppm of each individual capsaicinoid. THAT is the only way to have high scientific accuracy for heat measurements. I'd imagine the testing would be quite expensive though and would need to be performed by a trusted lab.

And one would have to wonder at the value-add of such a thing.

For one, the average consumer wouldn't understand it as well as say, a little picture of a thermometer with an arrow pointing to the red. For another, the higher that arrow gets on the bottle, the smaller your customer base is.

To me, a guy who aspires to soon compete in the "med-hot to hot" market and will only likely make a "superhot" as a novelty/product expansion (and one that I'll expect to sell minimally) this never ending competition to out-hot each other seems a bit senseless and even counter productive. The vast majority of consumers, even hot sauce consumers, sees the 1,000,000+ or other lofty number (regardless of scale used) and it might as well say "DO NOT EAT! THIS IS NOT FOOD. USING THIS WILL RUIN YOUR FOOD".

All due respect to hot heads, who I love dearly, but you people are just not "normal". That's ok though - you know you're not normal and you don't typically pretend to be normal.

But if I'm starting a company to sell something, I want that something to appeal to as many people as possible - and to be used as much as possible so people want/have to buy more. Of course I love love love hot sauce and have a passion for it. It's not all about the money - but if I'm putting money into something and trying to do business, I want to make money doing it. So narrowing my product appeal by advertising it to the least number of people seems counterproductive.

I wonder how much companies listing their Scovilles or HPLC ratings serves to confuse the general populace who's been pretty well trained to respect the age-old tried and true "x-tra hot/hot/med/mild" or thermometer formats? I'm actually not against a re-education....if it could be achieved, it would help the med and mild sauces equally because at a glance the general populace could grab a bottle and know exactly how well it matched their tastes.

Seems like the only time you hear about the numbers is with the up arrow though. I've never seen someone advertise that their mild hot sauce "has just 10,000 Scoville Units!" - though I think this would be of equal and potentially even greater value since the most common question I get from strangers is one of fear, "this isn't going to burn my mouth off, is it?"

Food for thought...

 

[Sawyer]

WARNING: Wall of text ensues.

In terms of analysis, the rich vein of HPLC hasn't been mined to exhaustion here.

No argument there.

Moreover, there's zero evidence that (fill-in-the-blank) will address the problem better.

Sometimes that's the nature of invention. All that's necessary to begin is to acknowledge that a problem exists. Also there's not exactly "zero" evidence either, resonant cantilever detection, for example, can detect single molecules in a solution. I would agree, though, that the most likely solution to the present problem is to develop (or adopt a preexisting) HPLC sample prep method that isolates the capsaicinoids from whatever confounding factors are apparently present in these sauces. I'm an inventor, though, so of course I'm going to speculate on alternative approaches.

All we have here is the contention that HPLC is "unreliable", when there's no experimental details given, and no definition of what reliability means. Hell, there's no actual data.

The "data" is the experience of sauce makers sending the same sauce to different labs and receiving wildly disparate results in return.

If we're looking for a direct, rugged, robust, cheap-to-develop way to giving a reliable measurement of things like hot sauce, HPLC represents the most obvious low-hanging fruit: it already works in plenty of labs, it uses fairly ubiquitous equipment, and existing protocols can be refined to achieve some localized optima.

Still no argument here on that point. I've already confessed to some surprise and dismay that such a mature technique as HPLC is failing to address the needs of the sauce-making sector of society.

Sure, one could develop a lab-on-a-chip or surface plasmon resonance spectroscopic tool or whatever glitzy gadget you can find in Nature these days, but there's no reason to think that the precision is going to be better, or that it will be cheaper, or that it won't be a huge investment of time and money with minimal return.

Perhaps not cheaper, at first anyway, but not to explore new possibilities leads to stagnation. Newer techniques, or refined
HPLC techniques, could easily reach ppb or ppt sensitivities, not that that would matter to the sauce makers, who don't seem to be getting even pp(whatever level is useful to them) results.

But please, by all means: go out there and prove me wrong. Dazzle me. 'Better' is always 'better'.

You seem a little tense.

The cynic in me thinks the labs understood just fine: "gimme a number I can slap on a bottle; it's not really all that important."

Then I'd say the labs really didn't understand. I think the sauce makers' own comments in this thread indicate they wanted some reasonably accurate numbers. I don't think they would have paid hundreds of dollars per test (assumed based on comments here and one local labs prices) to send samples to multiple labs for multiple tests for numbers they felt are "not really all that important." I would like to hear from the makers who've requested these tests on this point, but they seem to have abandoned this increasingly tedious thread. (Not talking about you Lucky Dog, I understand where you stand and think yours is a perfectly valid position, but by all means chime in if you wish.)

The test requestors assumed that test method development isn't necessary (from the tone of the discussion, there's no way they'd be interested in covering method development costs), and got back what sounds like embarrasing, extrordinarily cheap and flimsy data -- the kind of first-pass attempt that you get with chromatography if you're mostly after a quick-and-dirty qualitative result. They should have explained the scope and limitations, but somehow, niether party in that interaction bothered with asking or offering those details.

If you don't care about the method, then why bother caring about the result? You don't get the latter without the former. Why bother with testing at all? The method provides the context for making sense of the measurement. If you are interested in getting a number, you really ought to care what that number actually means, and that involves having some sort introductory level of understanding of what's being tested, and a realistic idea of how meaurements work. Sending a test request to a lab and expecting that the number you get back is going to represent the mystical platonic ideal of reality is absurd (reminder: the "Oracle of Dephi" was a myth -- not a contract lab.) If you think any quantitative measurement is giving you a truly absolute value (and I'm sure you don't), get ready for a big surprise. Real data has error bars, not all chromatography is a fungible commodity, and methods matter. That's why having a trusted, validated protocol run by a vetted lab is important, and that's why it's critical to know the approximate scope and limitations of the technique. Without those sort of details, instrumental analysis is just a fancy magic 8 ball.

As I've said before, in my opinion, it's incumbent on the lab to make sure they understand what the client is requesting and also to make sure the client understands the limitations of the results. if you work in an in-house analysis facility and have a predominately scientifically-savvy clientele, then it's reasonable to expect some familiarity with the testing protocols from that clientele. When you deal with the public, that expectation is not reasonable at all. The general contractor told by the EPA his demolition work has to stop and he has to bring in a hazmat team because the building tiles contain asbestos, doesn't care about the details of EDS when he brings me a sample for testing. He only wants to know if the EPA is right and his company is subsequently going bankrupt. (They were and it did.) The grieving widow who wants to sell her treasured "gold" statue to cover her husbands burial expenses doesn't care about details of the technique used to determine the statue is, in fact, made of yellow brass. They trust us as scientists to know what we're doing and to be honest in presenting our results. (Strange how the sad stories stick with you.) Of course in both of those cases I could have and would have provided those details if requested (and have in other instances as an expert witness.) A sauce-maker is certainly capable of caring about and understanding results that give capsaicinoids concentrations in some standard unit of mass per unit volume without needing to know the details of the technique used to determine those numbers. Especially when that technique is advertised to be as mature and as precise as HPLC is presented to be.

If the test lab had claimed to be using a validated method, or that they were following GLP, and they gave still reported crap data like that, I would would certainly demand my money back... but I don't think that's what they asked for. This is why it pays to be a savvy consumer of test data. A report where measurements are given roughly in the form of "N ± n SHU, following protocol XYZ", along with supplementary raw data, a copy of the protocol, and any notes or observations from the analyst would be much more satisfying and would allow for meaningful comparisons.

Unless some of the disillusioned sauce-makers tune back in and comment, we won't really know what they requested, expected, or received. I would like to know more details of their experiences. I would think any lab that had the equipment, though, would want to present itself as reputable and do whatever necessary to maintain a good reputation. They should make it clear the results will be in the form of capsaicinoids concentrations, with error bars. While I agree the protocol used should be included in the report, using one protocol or another is not an excuse to provide wildly inaccurate numbers. Again, when people start converting to SHU, everything gets a lot murkier. I think/hope most people here understand the subjective nature of the Scoville test and scale.

As a consequence, if the data you get back looks fishy and you paid good money, it's not impolite to call 'shenannigans'. Test method conditions should be part-and-parcel of any decent report, and asking them to show their homework is why you pay for the service in the first place. If a lab has delivered poor service, why keep quiet about it? There are commerical food science labs that offer validated capsaicin analysis by HPLC; it's perfectly good form to ask for these details, since it's what makes the data actually mean something. If I had been given results like the ones people are talking about here, I know what I would do and how I would complain, but that's because I work in an lab. Being assertive and skeptical and asking for these things are the sorts of habits that make people better equipped to deal with situations like these.

I mostly agree, although business people are generally some of the most assertive and skeptical folks I know. I don't see why there isn't a thread in the vendor's vault for capsaicin analysis labs. (I searched for HPLC in that forum and got zero hits.) It seems like an entirely appropriate thing to have. And if there were, people could quickly see which labs do provide good results and which labs don't. Instead, we have the present situation where what should be a well-accepted go-to analysis technique is being shunned by a portion of the food industry because of shoddy results.

If you thought I was criticizing HPLC, in principle, then either I misspoke or you misunderstood my words. My criticism is for labs that present shoddy results under the guise of good science and the attitude that non-specialists have to be cognoscenti in order to obtain useful numbers.