字幕列表 影片播放
>> All right.
Well, I will leave these and get started then, I guess.
All right, so I think I want to finish
up what I'll call basic NMR spectroscopy today.
In other words, things that all sort of fall at the level
of basic interpretation of structures, and in turn things
that we'll get on the mid-term.
I think probably where I'll pick up next time is going
to be introducing 2D NMR and then our next homework set,
not Monday's, but the one after that,
we'll start using 2D NMR spectroscopy.
So what I wanted to do today was to talk more
about carbon 13 chemical shifts.
And I gave you, when we talked last time, I gave you this sort
of general information that just like proton NMR, carbon NMR,
aliphatics tend to be upfield, aromatics tend to be downfield,
things that are next to electron withdrawing groups,
particularly oxygen, tend to be mid -- we'll call it mid field,
sort of in that 50 to 70 range.
I also indicated -- and we said
that the range is a lot, lot bigger.
It's about 20 times bigger in ppm for C13 NMR.
In other words, an aldehyde CH is at roughly 10 ppm whereas
in carbon NMR an aldehyde carbonyl is at roughly 200 ppm.
So it's sort of like 20 times bigger range.
Now, C13 shifts have a bigger range.
And there's also more richness.
In other words, when we talked
about proton NMR, it was pretty easy.
And we were able to come up with some really simple, you know,
back of the envelope calculations
where you could typically peg the chemical shift to within,
you know, a few tenths of a ppm.
We talked about if you're next to an ester, you know, let's --
or if you're next to an oxygen, figure you're going to be
about 2 ppm downfield of where you'd normally be.
If you're next to a benzene ring or a double bond or a carbonyl,
figure you're about 1 ppm further downfield.
And I gave you several ways of thinking about this.
And you should all be able to pretty much estimate things.
We talked about the effects of alpha substituents and said
that alpha substituents have a really big effect being next
to a CH next to a halogen next to a nitrogen next
to an oxygen shifts you down, you know,
a couple of tenths -- a couple of ppm.
We talked about beta effects.
And we said that they're smaller.
Beta effects in proton NMR shift you down by like .2 to .5 ppm.
C13 NMR is a little bit more subtle and a little bit --
I won't say less predictable, because we're going
to see it's actually quite predictable.
But the factors, there are more factors.
For example, gamma effects as well
as beta effects tend to be big.
And there are some really interesting steric effects.
Now, along with this richness comes tremendous power.
Because this means that carbon NMR also can give you tons
of rich information about structure
and can be really useful for figuring out structures,
confirming structures, disproving structures,
and at the end I'll show you a beautiful example
of fraudulent work that was disproven
by Professor Rychnovsky's laboratory.
And also going ahead and basically having a tool
that can get you a lot more information than meets the eye.
So, I want to show you some of the factors that contribute
to these sorts of general ranges,
and particularly now, perturb them.
And let's start with something pretty simple, inductive effects
and resonance effects.
[ Writing on Board ]
And electron density, of course, plays a huge role
in chemical shift because electrons contribute
to the shielding or de-shielding
if they're absent, of various nuclei.
So, if you have substituents that increase
or decrease the electron density in a carbon, you're going
to shift that carbon upfield or downfield.
Let me show you what I mean.
We'll start with a simple benzene example.
Now, the easiest way to, in your mind,
think about chemical shifts is to start
with a base value and then perturb it.
So a great way to think about benzene is the normal position
for benzene is 128.5 ppm.
And then if you put substituents on it,
you perturb things in a rational way.
So let me show you what I mean.
If we put a methoxy on the benzene,
the oxygen is electron withdrawing
through the sigma bond.
And so the carbon attached
to the oxygen shifts substantially downfield.
So you go to 160 ppm.
In other words, you shift downfield by 30 ppm,
more than 30 ppm by putting that oxygen there.
Now, what's interesting, then, is the ortho carbons end
up having by resonance extra electron density at them.
In other words, the two ortho carbons you can push your
arrows, and you see they're electron rich.
It's the same reason why electrophilic aromatic
substitution occurs ortho and para
when you have a methoxy group on there.
So, the ortho carbons appear at 114 ppm in methoxybenzene.
You don't get a big effect at the meta carbons,
which makes sense because you have inductive effects
that are now quite removed.
So it's very small.
And resonance doesn't pump
up the electron density at the meta carbon.
You go to the para carbon,
and now you also see an upfield shifting, although it'll be
at a smaller upfield shifting of 121 ppm -- at 121 ppm.
Now, I'll come in a moment
to empirical additivity relationships,
but one way to think about this is to think about it
if you have a methoxy group or an alkoxy group on a benzene,
that it shifts the ortho protons upfield
by about 14 or 14 1/2 ppm.
And if you have a methoxy group,
it shifts the para protons upfield by about 7 1/2 ppm.
And if you have a methoxy group or an alkoxy group,
it shifts the meta protons downfield
by just a fraction of a ppm.
And what we'll see in a moment is that you can add
up all these effects and then calculate
for different aromatics,
the effects of different substituents.
All right, let's take a look at some other examples
of electron -- of inductive and resonance effects.
So, let's take an alkene.
And I'll give you cyclohexene as a base value.
In cyclohexene, your alkene is at 128 -- 127.4 ppm.
Let's compare that to cyclohexanone.
And in cyclohexanone, we see a very big effect
at the beta position and just a little effect,
just a little inductive effect at the alpha position.
The beta position is 151 ppm
and the alpha position is just shifted downfield by a hair.
And that makes sense because you look at this and you say okay,
now you can think of a resonance structure
in which you're electron deficient, right?
We all know that enones are Michael acceptors.
That nucleophiles like to attack at the beta position.
Are you doing this with orbitals in Van Vranken's class now?
>> Yes.
>> Yeah, so you know
about frontier orbitals and electron density.
And so you see, the effect is actually very substantial,
right?
Both of these carbons, they're symmetrical, they're at 127.
You go more than 20 ppm downfield
by decreasing the electron density.
These effects can be absolutely humongous.
And one of the things that I've tried to emphasize
when I've talked about these ranges here are these are
general ranges.
These are not carved in stone.
And so you already see, for example,
that this one inductive oxygen here brings us even outside
of this very generic range here.
Let me show you just how huge the effects can be.
So, ketene acetal is a good example, right?
Alkenes are normally like 110 to 150 ppm.
But if hugely perturb the electron density,
you can have huge effects.
So it probably doesn't surprise you too much if I tell you
that you now, by having two methoxys
on an alkene can shift it downfield to 169.7.
But what's really huge is you look at the position here,
the beta position on the alkene, and now we're
so electronic rich, this thing is so nucleophilic
at this position, there's so much electron density
at that position, that we're at 45.5 ppm.
You look at a spectrum of that
and you wouldn't even know it's an alkene because you'd say, oh,
that's got to be aliphatic.
That's got to be somewhere over here.
And we've just pumped up the electron density hugely.
The most radical example I know off the top
of my head is this sort of push me, pull you system here,
where you have two electron-donating groups
and then two electron-withdrawing groups,
two cyano groups.
And so this alkene, you go to 39.1 ppm.
And then this carbon here is at 171.
And 171, you'd say all right, well, it's really downfield.
But you'd say it makes sense.
You've got these beta electron withdrawing groups.
But you look, 39.1, who would of thunk that that is an alkene.
[ Erasing Board ]
>> Can I ask you a question?
>> Yeah.
>> What is that letter to the -- CW, or CN?
>> CN.
>> Okay.
>> So these are two nitrile groups.
These are two cyano groups, CN groups.
>> Okay.
>> All right.
So, substituents have substantial effects whether
they're alpha, whether they're beta, whether they're gamma.
And you can really see this.
I'll give you -- we're going to walk through this
and I'll give you some examples.
So first let's talk about alpha alkyl substitution.
And if I want to give you a general principle, in general,
alpha alkyl substitution leads to more downfield shifting.
So if we, again, take our benzene example,
and remember we said that benzene was at 128.5.
If we put a methyl group on it and make it
into toluene, now we go to 137.
The point is we shift downfield by about 9 ppm by putting
on a methyl group onto benzene.
Let's take a look at alkyl systems.
So, we'll look at propane.
And we'll look at the central methylene
of the central carbon of propane.
You put on an alkyl group onto propane and you get isobutane.
And now you're down at 25 ppm.
And so you notice, it's kind of about the same.
Here we moved down by about 8 or 9 ppm.
And here we, again, moved down by about 9 ppm.
So in other words, we're talking on the order of, eh,
10 ppm or thereabouts.
All right, you put on another alkyl substituent
and the effect isn't as dramatic.
But again, you move further downfield.
Now we're at 28 ppm.
Now, what's nice about this is these ideas are generalizable.
There is really science here to it.
And this is the point, that you can take little bits
of knowledge and generalize and build
up in your mind what's going on.
So let me compare us, say, to ethanol.
If we start at 58 for ethanol, and now we envision going
to isopropanol, what would you predict your carbon
to be at here?
>> 67.
>> 67 would be a very good prediction, because we say,
okay, we take 16, we add 9, we get to 25.
You add 9, you get to 67.
And that's a very good guess.
64 is the answer.
All right.
Now imagine we go further and we go to tert-butanol.
Now what do you think, for tert-butanol?
>> 66.
>> 66?
>> I don't think it's going to change.
>> You don't think it's going to change.
Okay. Other guesses on that this estimates?
>> 67.
>> 67. All right.
And the answer is 69.
No, you're doing good.
You don't have to peg it to the nearest ppm.
[ Inaudible Audience Response ]
Oh, okay. The Price is Right approach.
The Chemical Shift is Right.
Do you think we have a show here?
Maybe really late night TV.
[ Laughter ]
>> Let's call up Bob Barker [inaudible].
>> Ooh.
Maybe we can do a Webcast of this.
Try it out in a test market.
All right.
So in the case of proton NMR, we also saw that shifting,
adding an alkyl group shifts you downfield, right?
We said a methyl group is typically
at about .9 parts per million.
And methylene group we said 1.3 to 1.5.
So in other words, you're shifting down 1.3 --
1.2 to 1.4 I think I said.
So in other words, we're shifting downfield,
you know, to about .5 ppm.
A methine is down at like 1 1/2 or 2 ppm.
These are the baseline values.
So in other words, we go a few tenths of a ppm more.
And so it kind of makes sense.
If you get half a ppm for going --
adding one alkyl group and you know, another couple of tenths
of a ppm for adding another alkyl group,
it kind of makes sense here with a scale 20 times bigger
that we get about 9 ppm or 10 ppm for adding one alkyl group
and about 5 ppm for adding another alkyl group.
With the case of proton NMR,
beta alkyl groups don't make a huge effect in chemical shift.
In the case of C13 alkyl groups, though, they do.
Beta substitution ends up also leading --
also leading down to further downfield shifting.
Not quite as dramatic.
But let's take an example.
And what I'll do is I'll take butane as an example
and we'll look at this carbon here.
And then we're going to put substituents over here.
So they'll be beta to it.
In other words, instead of directly
on it, there'll be one over.
So if we add one methyl group here,
so I'll go from butane to 2-methyl butane.
Now this carbon goes 7 ppm more downfield.
We go to 32 ppm.
And then if I add another one,
we go another 5 ppm more downfield, we go to 37.
Not as big, but still pretty darn big.
Okay, so that's beta substitution.
Let's look at gamma substitution.
Now, gamma substitution is interesting
because gamma substitution leads to upfield shifting,
and it's really a steric effect.
So, with gamma substitution,
what basically you're doing is now getting interactions
that are by having steric repulsion,
pushing more electron density onto the carbon
and shifting you upfield.
So let me show you what I mean.
So we'll still keep to this same type of general structure.
But now we'll consider this carbon here.
And so if you look at the carbon here,
and you have hydrogens banging into it,
you basically will have hydrogens banging together.
This is going to lead to more electron density on the carbon.
In other words, electrons of the CH bond repelling electrons,
push electron density into the carbon.
And since you don't want to have steric repulsion,
this occurs most pronouncedly
when the molecule cannot avoid steric repulsion.
Let me show you what I mean.
In other words, if we start with butane, and we start over here,
the methyl group of butane is 13 ppm.
If I add one methyl substituent gamma, so in other words
if I attached it directly to this carbon, we'd be alpha.
If I attach the methyl substituent here we'd be beta.
If I attached the methyl substituent here we'd be gamma.
If I add one methyl substituent,
that carbon shifts from 13 to 12.
And you'd say, ah, that's not a big deal.
That's not a huge effect.
And part of it is this methyl group can adopt a rotational
isomer, a rotamer, where it keeps out of the way.
But, if I add one more methyl group, then no matter what,
we're going to be staring that other methyl into the face.
So if I go like that and go to 2, 2-dimethylbutane,
now there's no way to avoid that steric repulsion.
And we shift over to 9.
So now we've shifted upfield by about 4 ppm relative
to our initial compound.
So even this very remote substituent can make a
big difference.
You can really see this
in conformationally constrained systems like cyclohexane.
So if I take cyclohexane and I consider the carbons.
All the carbons, of course, in cyclohexane are the same.
And they're all at 27 parts per million.
If I now add a methyl group gamma to this particular carbon.
So remember, if I attached it here, it would be alpha.
If I attach it here it would be beta.
If I attach it here we're gamma.
If I add one methyl group, that methyl group keeps
out of the way because it keeps equatorial.
And so we don't see any change.
But if I add a second methyl group gamma,
then one methyl group has to, has to, has to be axial.
And there's no way to avoid 1, 3-diaxial interactions.
And so now that carbon shifts up to 21 ppm.
And that's a pronounced effect.
And I'll show you the implications
of that in just a moment.
All right.
So these are all of the sorts of biggie effects in C13 NMR.
And now I'll show you the last biggie effect,
and that's heavy atom effect.
[ Writing on Board ]
All right.
So let's take a look at what happens
when you make various halomethanes.
So again, it's very helpful to have a baseline value.
Methane occurs at negative 2.3 parts per million.
So we're going to compare the halomethanes to methane itself.
And so we'll look at monohalomethanes, CH3X,
dihalomethanes, CH2X2, trihalomethanes, CHX3,
and tetrahalomethanes, CX4.
And we'll start with X being chlorine.
And so, no great surprises here.
You put a chlorine onto methane, you get chloromethane.
And you go downfield.
It's 25 ppm.
You add another chlorine, you go further downfield, it's 54 ppm.
Chloroform?
Where does that show up at?
>> 78.
>> 77. 77 is chloroform.
So you go further downfield.
You put yet another chlorine on there,
and you go further downfield.
And so you look and you say, hey,
what am I wasting your time for?
This doesn't seem like any sort of surprise.
Well, if we move down the periodic table,
the atoms get bigger.
And they start to have d orbitals, and they start
to have their orbitals extend outward.
And then we see something very surprising.
You put on a bromine and you'd say 10.
And you'd say, well that's not so surprising.
Bromine is less electronegative than chlorine.
It's going to shift you less further downfield.
You add two bromines, you go to dibromomethane, and you say 21.
And you say, okay, that doesn't sound so surprising.
We're moving further downfield.
Three gets a little bit weird,
because we start to move upfield.
And by the time we're at four, we're at negative 29.
And then iodine gets downright funky, so iodine --
iodomethane is at negative 21.
Diiodomethane is at negative 54.
Triiodomethane is at negative 140.
And tetraiodomethane is at negative 292.
And what's happening is the orbitals are extending so far
out with these heavy atoms,
as you move down the periodic table,
that their electron clouds are actually enveloping that carbon
and shielding it from the applied magnetic field.
And as you can see, in some cases shielding it
very dramatically.
Now, the practical implication for this becomes,
if you go ahead and do a reaction to, say,
make an alkyl iodide in synthesis,
and you don't go ahead and look upfield,
you may miss your carbon and say, oh my god,
I can't see my carbon.
There's a parameter in the NMR called sweep width
that you adjust.
That sweep width is the spectral window that you look at.
If you have a compound that you expect to be very far downfield
or very far upfield, this is true in the proton NMR as well,
if you have a compound that you expect to be very far downfield,
or very far upfield, you can just increase your sweep width,
the field width, by whatever number of hertz you like.
You type in the parameter SW, get a bigger spectral width
and then you can see, for example,
your carbon that's attached to iodine.
Same thing with protons attached to iodine.
I once took a spectrum of HI.
I was doing a de-protection reaction in an NMR tube.
And the HI, the peak was at negative 10 ppm
in the proton NMR, which was absolutely wild.
>> Why, like in the bromine one,
that at first it puts it more downfield.
And then when you get three bromines,
then all of a sudden now it pushes it back upfield?
>> I think what it is --
so the question is why are two bringing it more downfield.
Obviously, you have the inductive effect pulling away.
I think by the time you get to three you just have
that carbon completely surrounded
by the electron cloud.
So now you just have electron clouds sticking
out in all directions.
And in four, it's just all around it.
These are, in part, relativistic effects.
When you're doing MO,
when you're doing molecular orbital calculations
on heavy atoms, you actually need
to have special terms taking
into account relativistic effects of the orbitals.
All right, so one of the take home messages
from this is there are a lot of factors that go
into carbon 13 chemical shift.
And for this reason, carbon 13 chemical shifts are very rich
and very valuable.
And being able to predict them can lead you to lots
of useful information on structure that can be extracted,
sometimes can give you a problem-solving tool
that you might not otherwise have.
And so what I'd like to do now is to show you some
of these problem-solving tools and their applications.
So we'll talk a little bit now
about C13 chemical shift prediction.
And we're going to talk
about three different ways of doing this.
One is what is called empirical additivity relationships.
That's the simplest.
[ Writing on Board ]
The idea there is we're going to go and say okay,
the effect of an ortho substituent is this
if it's an oxygen.
The effect of a meta substituent is this if it's a chlorine.
The effect of an alpha substituent is this.
The effect of a beta substituent is this.
The effect of a gamma substituent is this.
And we're going to add up all of those effects.
If you use ChemDraw and you have one
of the more advanced versions of it
that does chemical shift prediction,
that's exactly what it's doing.
Those calculations also for all sorts of systems, aliphatic
and aromatic systems, are shown in your book
in your structure determination, the orange book
that you have for the course.
You know, the supplemental book.
And you can do it for all sorts of system.
And they do it to a limited extent in Silverstein.
And I'll show you that just for aromatic systems.
Unfortunately, the ChemDoodle program,
which tries to do chemical shift prediction,
thus far they're still very much in development.
But they've thus far got it wrong.
There is also a small CD, Windows only,
that goes with the orange book in the course
with the supplemental book.
That has some limited chemical shift prediction based all
on empirical additivity relationship.
Another way of doing this is based on databases
where you come up and rather than simply saying, all right,
we're going to have alpha effects and beta effects
and gamma effects, you have a training set.
And then from that training set,
or that set of reference compounds,
you find parameters that fit them.
And then you can extrapolate.
So you first interpolate to encompass all
of the compounds in the set.
And then you can extrapolate to encompass other compounds.
And so there are generalized databases.
And then there are also specific ones.
So for example, for various stereochemical problems.
And the third way of doing this is going to be
by electronic structure calculations.
And we're not going to do this in the course,
but I'll show you an example at the end.
Basically, since carbon 13 chemical shifts come
from electron density,
if you can calculate the molecular orbitals
at an appropriate level of theory to figure
out how the electrons are distributed in your molecule,
then you can figure out the electron density
around any carbon, and hence, its shielding
from the applied magnetic field.
All right.
Let's take a look at these various methods.
And I'll start you with a simple page right out of Silverstein.
And Pretch has a -- the orange book,
Pretch, has a similar table.
Ooh, that doesn't look good.
Ah.
Helps to plug it in.
All right.
So this should be on your handout.
And as I said, this is just one of the many tables in Pretch.
But a tool is so much more useful
when you actually have used it and know how to use your tool.
All right.
So this is a table of empirical additivity relationships
for substituents on a benzene ring.
And literally what was done to make this table was,
as we said, benzene's 128.5.
So somebody took a spectrum of phenol
and said the carbon that's directly attached to the OH
of phenol is shifted downfield by 26.6 ppm --
or 20, what did I say?
Yeah, 26.6 ppm.
The carbon that's ortho is shifted upfield by 12.7.
The carbon that's meta is shifted downfield by 1.6.
And the carbon that's para is shifted downfield by 7.3 ppm.
All right.
let's take a look at the following compound,
and then I'll give you two spectra.
And I'll show you how we'll use these.
So, let us imagine for a moment that we were trying
to distinguish the following compound from other isomers.
So let us suppose for a moment that we were trying
to distinguish, say, this compound, 2, 4-diochlorophenol
from this isomer, 3, 4-dichlorophenol.
That would be a tough one to do based on coupling patterns
because -- in the proton NMR.
Because if you look at the proton NMR,
you would say we would expect to have one proton
that is split by an ortho coupling.
So we would expect it to be a doublet.
And its J value would be somewhere on the order
of about 7 or 8 hertz.
We would see another proton.
What would you expect this proton here to appear as?
As a doublet of doublets with J values of what?
>> 7-ish and it's --
>> Metacoupling?
>> 3?
>> 3?
>> 3. So we'd expect this to be a doublet of doublets
with a big J of about 7-ish and a small J of 3.
And this one we would expect to be?
>> 3 [inaudible].
>> 3 as a doublet.
And you would expect this exact same pattern of coupling here.
A doublet of doublets with a big J and a small J. A doublet
with a big J and doublet with a small J.
So we wouldn't be able to tell them apart.
But let's look at this compound here and use these principles.
If we want to calculate the chemical shift for,
let's just take -- I'll just do three of the carbons here.
So if we were to want
to calculate the chemical shift here, we would start
with the base value of 128.5.
And then we would say, okay, we have an ortho alkoxy
and we'd look in our lookup table
and we'd say -- or an ortho hydroxy.
And we'd subtract 12.7.
So minus 12.7.
Now, we would then look at the effects --
can we see from this angle?
Should I move these calculations over?
Let me move this over.
128.5 is our base value.
Then we have an ortho hydroxy, and so we'd take away 12.7.
Now we're going to look at the effects
of chlorine for this compound.
And so we have two meta chlorines.
So we'll go to our lookup table and we'll look
at the effect of a chlorine.
So I'll draw a line here.
And so, now we have two meta chlorines.
So we have a meta chlorine over here.
And that's going to have an effect of plus 1.0
and we're going to have --
I guess I should make my numbers line
up if I'm going to run a tally.
And we have another meta chlorine,
and that effect is plus 1.0.
So we tally all of those up and we predict 117.8.
And now I can do -- so that's for this carbon here.
Now I can do this next carbon.
So you help me out with this calculation for this carbon.
So what do we use as our base value?
>> 128.5?
>> 128.5. And then what do we do for the oxygen?
[ Inaudible Audience Responses ]
>> One -- so we add.
So the oxygen is meta OH, so we add plus 1.6, okay?
And then what do we do next?
James?
>> Add .2.
>> .2 for?
>> The meta chloro?
>> For?
>> Sorry, the ortho chloro.
>> Ortho chloro, so we do O chloro is -- you said plus .2?
>> Plus .2.
>> Plus .2.
And then somebody else, what do we do last?
>> Subtract 2 for the para?
>> Subtract 2 for a para chloro.
And so we tally all of that up and we predict 128.3.
All right.
I'm not going to do any more.
We could do it for all six of them.
But, let's flip to the next page in your handout.
And I grabbed from the Aldrich Library of Spectra the proton
and carbon spectra of the two compounds.
And we're just going to look at the carbon spectrum.
And the thing that jumped out at me in the carbon spectrum
for a difference that was really cool was in the carbon spectrum
of the first compound.
Remember the quats are small and the CH's are big due
to relaxation, differences in relaxation time
and nuclear Overhauser effect.
And what jumped out at me immediately was
that here we have two of the CH's close to 130.
And one of them -- so two basically downfield of 120,
and one upfield, so at about 115.
And what jumped out at me here was the dramatic difference
where here we have one at about 150, about 130.
And then two of them at the range between 110 and 120.
And of course, we could do this for the quats.
There are dramatic -- there are substantial differences as well.
You can see, here are your quats.
So all I did in preparing
for today's discussion was I just calculated
for the two compounds.
And so we did 118.5 and -- let's see, was that right?
100 -- wait.
Oh, what did I do?
I did my -- yeah.
I don't know how I managed to tally things up wrong here.
Wait, 100 and -- let's see, 128, 129, 130.
118. -- wait, 118.
-- what was it.
>> No, that's right.
It'd be 119.2.
Oh, sorry, that's minus.
I'm sorry, I got lost.
>> Wait a second.
Anyone have a calculator?
>> You were right.
[inaudible].
>> Was it 128.5?
Wait. 117.8.
Okay. So I get -- I don't know what I was writing down here.
117.8 and 128.3.
And over here -- I think -- what did end up doing?
130.7. And then over here 115.2 and 117.5
and I think here 131.7.
All right.
So the point of this is even though it might be very hard
looking at the proton NMR to assign which isomer we have,
by looking at the carbon NMR,
you can say oh wait a second, okay.
The first isomer that has the 130.7,
128.3 and 117 matches this spectrum.
And we could do this with the quats as well.
So this spectrum corresponds to the first isomer.
The second isomer that has the 115.7,
117.5 and 131 matches this spectrum.
And so imagine for a moment
that you're doing an electrophilic-aromatic
bromination of phenol, you want
to determine whether you got the product you expect.
You know you have a dichlorophenol.
You don't know regiochemistry.
You can very quickly check yourself.
And the good news is these types of relationships generally good
to let's say plus, plus or minus let's say 3
to 5 ppm on the average.
So the point is that you can,
with very little difficulty, tell things apart.
All right.
There are many, many more sophisticated ways
of doing this.
As I said, you can do this for aliphatic compounds as well.
With aliphatic compounds you've got to add up alpha effects,
beta effects, gamma effects.
There's even a little bit of delta and epsilon effects.
So you have to look at the whole molecule.
There's steric effects.
You can do this easily by pencil and paper, but it's even easier
to do it with computer software.
And as I said, even ChemDraw
in the more sophisticated versions incorporates this.
There are other empirical additivity relationships
that are out there.
But in addition to that, there are databases
that have training sets.
So for example, this is just one poster I pulled that's a
database approach.
And what they were looking at was how well their system
that was trained on a training set would predict the C13 NMR
of taxol.
And in general, their values were good to plus
or minus 1 ppm on the average.
And so for example, they weren't able to distinguish --
here's another one that would be hard to do by NMR.
They were able to use their database approach.
And here we have two tautomers.
And so if you synthesized this molecule, you might not know
which tautomer you have, whether you have the molecule
with the carbonyl or the molecule with the hydroxy.
And it's not so easy that you can take an IR spectrum
because by that point, you end
up with very strong stretches associated with the CN bonds
of the pyridine tautomer.
And so the IR spectrum is not clearly going
to scream carbonyl at you.
The C13 NMR, the difference
between what you would expect here
and what you would expect here
for the carbonyl is small enough.
You can't just look at say obviously I have the pyridone
or obviously I have the other compound.
But, in fact, they can go ahead and calculate
to within a couple of ppm.
So they were able to experimentally measure the value
and then say, okay, what's the deviation?
And here they're getting an average deviation
of about 2 ppm.
In other words, some of the shifts will be off by 3 ppm,
some of the shifts will be off by 1 ppm.
But on the average, they'll be good to within about 2 ppm.
All right.
And so that's an example of a database approach.
And in general, their method's good to about 3 ppm.
Many times you encounter specific problems.
You were working on a research project
that involves a specific class of molecules.
For example, the Rychnovsky group, here, and the Kishi group
at Harvard are very interested in 1, 3-diols.
1, 3-diols are important classes of natural product.
They come from biosynthetic pathways that give rise
to molecules of many different structures.
And the big issue becomes determining stereochemistry.
The Rychnovsky group noticed after the fact --
they were working with 1, 3-diol acetonides.
And the question was do they have a syn-diol
or an anti-diol, right?
So a syn-diol would be a 1,
3-diol where the two OH's are like so.
And an anti-diol would be one in which they are like so.
And the beautiful thing is if you're a clever graduate student
or a clever professor and you keep your eyes open and look
at your data, you say hey, we've got a pattern here.
And that pattern tells us something.
And that pattern means something.
And when they went back and looked at all
of their C13 NMR spectra, of the 1, 3-diol acetonides
that they had prepared, they saw something very interesting.
The methyl groups of the 1, 3-diol acetonides,
when the diols were syn, appeared split at about
into two sets of diastereotopic peaks.
One peak at about 20 ppm, the other peak at about 30 ppm.
And when they had the syn-diol acetonides,
both of the methyl groups appeared at about 25 ppm.
In other words, you could now take a diol
of unknown stereochemistry, make the acetonides and just
by looking at the C13 NMR spectrum, say oh,
that's a syn-diol or that's an anti-diol.
It's hard to do that otherwise.
That's extremely valuable.
Why does this occur?
It's that gamma effect again.
When you make the syn-diol, you end up with a chair.
And in that chair, one of your methyls is axial and one
of your methyls is equatorial.
And remember, the axial one gets steric compression.
So it is shifted upfield.
The equatorial one doesn't get that steric compression,
so it's shifted downfield.
On the other hand, when you make the anti-diol acetonide,
then you can go ahead and both your methyl groups are sort
of in the middle.
Neither of them really has -- you get a twist conformation.
Neither of them has unusual effect.
And they end up in the middle.
So just by a quick C13 NMR with this bit of knowledge in hand,
you can go ahead and assign the stereochemistry.
The Kishi group at Harvard has taken this type
of approach with their diols.
And this is a really, really cool use of synthesis.
So they have diols where you have alternating hydroxy
and methyl groups.
These are compounds that occur in a variety
of different natural products.
And so they wanted to be able
to tell the stereochemistry of these patterns.
So they went ahead and they said all right, we've got lots
of natural products with this type of unit
where you have a diol, a methyl --
an alcohol, a methyl, an alcohol, a methyl.
And so what they did was make a training set
of all the possible stereoisomers.
So they made ones in which they were all syn,
in which one is anti, in which two are anti
in different positions, and so forth.
And they found that there were characteristic, that each carbon
in their training set -- so they used this molecule
as the training set --
each molecule in the training set had an average chemical
shift of a certain position.
And so then they took the average value and they said
if you compare it to the average value, you have a fingerprint
where certain ones are shifted upfield,
certain ones are shifted downfield.
And each of the possible stereoisomers has their
own pattern.
So that means you can then take an unknown compound based
on this training set and the database
and take an unknown compound and say, ah!
The pattern matches this stereochemistry.
All right.
Last example I want to talk about.
And this was really cool.
So this is -- this is all calculating based
on known information.
We started with the simplest thing.
You just took phenol and you said okay, what's the effect
of an ortho substituent?
What's the effect of a meta substituent?
What's the effect of a para substituent?
Took chlorine, took chlorobenzene, did the same.
And then say you could predict another compound.
These diols are also examples where you have a training set.
But for really funky and unusual compounds
that have very unusual rings and strained rings,
there may not be a good model out there.
And there was a really cool compound that was synthesized --
end quote -- by a single researcher in San Diego
who published this on his own in "Angewandte" with no co-authors
and no university affiliation.
His name was James J. La Clair.
Still is James J. La Clair.
And he published this synthesis that looked really cool
because the molecule he said he synthesized was this,
hexacyclinol.
One that Phil Baron, I think, would appreciate
as a challenging target.
And like a student who maybe didn't quite get an A in Syn.
1, his synthesis had some stuff that was funny about it
in addition to him being the sole author.
There were steps that just didn't seem to make sense.
And so the buzz in the community was something wasn't right here.
Now, nobody wants to tell their entire lab, hey guys!
Let's reproduce James J. La Clair's experiments
and see if it really works.
That's several graduate students
and dissertation's worth of work.
That's not exactly a great job to be doing.
So, Professor Rychnovsky used electronic structure
calculations to see how well he could first predict the chemical
shifts of a known molecule.
So he took a known compound, elisapterosin,
and did these electronic structure calculations
and plotted a graph of the actual chemical shift
versus the observed chemical shifts.
And as you could see, most of the chemical shifts calculated,
the calculated ones and the observed ones matched
within a couple of tenths of a ppm.
So in other words, here was have a match of plus
or minus 2 ppm on the average.
All right.
So confident, confident of his methodology,
he then took the structure of hexacyclinol --
and I'll put this in quotes here as you'll see in a moment --
and found a terrible match.
The vertical scale on this graph is from 0
to 5 parts per million.
You look at the scale on this graph,
it's from 0 to 25 parts per million.
The deviations on the average are within about 7 ppm.
It's many of the shifts are way off.
So that's the average.
Now, he then went ahead and thought,
this guy got the structure wrong.
And he had a hypothesis of what --
so not only did he think that the synthesis was a fraud,
the thought that the structure wasn't even right
because it didn't match.
So he reinterpreted the published structure
of the molecule, the one from the Journal of Natural Products
where they had isolated this molecule,
and thought it was this structure,
reinterpreted the data,
and considering biosynthetic factors,
came up with what he believed the correct structure to be was.
And there the structure matched.
And he did two different conformers of it.
And there, for this structure, matches within plus
or minus 2 ppm average.
By 2 ppm average, I mean RMS deviation,
root-mean-square deviation.
So on the average within 2 ppm.
So, the answer became, not only did James J. La Clair --
what would is say?
Cheat, forge a synthesis of the molecule,
but he forged a synthesis of the wrong molecule.
And it was an impossible synthesis.
Anyway, it was pretty cool
and the Rychnovsky group was really happy for a while.
>> Who is this guy?
And what is he, exactly?
Is he known?
>> A very unusual character.
So the other thing that was interesting, and it's all
out there, is if you look at his spectrum of hexacyclinol,
hexacyclinol is a terpene, the has methyl singlets in it,
which are big, tall peaks.
If you look at his spectrum of his synthetic hexacyclinol,
the methyl groups are missing C13 satellites,
those little side peaks spaced 125 ppm, 125 hertz around there
at .5 percent of the height,
which for a 20 centimeter height peak you can actually see
in a spectrum.
So he fabricated this synthesis
and his spectrum out of whole cloth.
[ Multiple Speakers ]
>> I assume this person had his PhD [inaudible].
>> It wasn't part of his PhD -- if you go and you earn your PhD,
you're a legitimate, and then later
on as an independent scientist,
working at the Xenobe Research Institute,
your own research institute, publish absolute nonsense
in the literature, yeah, I don't know.
>> How did he get published?
>> Yeah, how did he published?
>> Ah, interesting question.
A very controversial story. ------------------------------dc9c2cfd829b--