3. Results and discussion
3.3. Mass spectrometric behavior of B 12 -cluster-containing compounds
Nozzle
~150 V
Skimmer
~100 V
Ion-trap MS
~1 mbar ~10-3 mbar ~10-5 mbar Metal capillary,
ID 100-200 µm
±2-5 kV
~0 V
Skimmer-CID region
MS/MS region
Sample solution
Nozzle
~150 V
Skimmer
~100 V
Ion-trap MS
~1 mbar ~10-3 mbar ~10-5 mbar Metal capillary,
ID 100-200 µm
±2-5 kV
~0 V
Skimmer-CID region
MS/MS region
Sample solution
Figure 16. Schematic diagram of an ESI mass spectrometer. Fragmentation can occur in the skimmer-CID region, or in the MS/MS region.
In sCID, fragmentation is induced by colliding the sample ions with the background gas in the intermediate-pressure region (the so-called "nozzle-skimmer region", gas pressure about 10-1 mbar) of the ESI interface. Increasing the voltage difference between the nozzle and the skimmer (VsCID) raises the kinetic energy of the ions passing through this region, hereby increasing the collisional energy of the ions with the background gas (consisting mainly of nitrogen and of some residual solvent vapor) and eventually causing fragmentation. The VsCID applied in this investigation were between 75 and 100 V and are indicated in the mass spectra shown in Figure 17.
87.1
123.0 141.1
0 2 4 6 8 x104 Intens.
60 80 100 120 140 160 180 m/z
V(sCID)= 96 V
V(sCID)= 76 V V(sCID)= 85 V
Figure 17. ESI-MS of 32, using increasing sCID voltage.
In contrast, tandem-MS (or MS/MS) takes places in the mass analyzer, i.e. in the high-vacuum region of the MS (pressure about 10-5 mbar). In MS/MS, the ion of interest is isolated within the
ion trap by ejecting all other ions out of the trap. Then this ion is accelerated (by applying a suitable high-frequency AC voltage) and collides with a collision gas (usually He) present at a pressure of 5⋅10-6 mbar. The fragment ions thus generated are then detected by a normal mass scan. The degree of fragmentation depends on the AC voltage amplitude; in this investigation, this amplitude was adjusted empirically so that a low intensity signal of the parent ion remained visible in the fragment ion spectrum. Alternatively, the residence time of the ion was increased.
Generally, ESI-MS is considered to be the softest of all known MS ionization methods;
consequently, fragmentation of sample ions is usually not observed unless it is deliberately induced by either sCID or MS/MS. Hence, the negative-ion ESI mass spectrum of 32 was expected to show only the signal of the doubly charged molecular anion at m/z 123 (2-).
However, the spectrum (Figure 18) shows two additional peaks: a doubly charged ion at m/z 87 (2-), which was attributed to [B12H11SH]2-, and a singly charged ion at m/z 141 (1-) whose mass and isotopic pattern suggest the molecular formula [B12H11]-.
87.1
123.1
141.2
0.0 0.5 1.0 1.5 2.0 x104 Intens.
60 80 100 120 140 160 180 m/z
[B12H11] -[B12H11S(CH2)2COOH] 2-[B12H11SH]
2-Figure 18. ESI-MS of 32.
Thus, the mass spectrum seemed to indicate a mixture of compounds rather than a pure compound, which, however, was not in accordance with results from HPLC, NMR (Figure 19) and IR. In order to investigate whether the additional ion signals originated from the MS measurement, a series of sCID experiments were carried out. The results are shown in Figure 17.
HPLC chromatogram of 3
(pp m)1.5 1.0 0.5 0.0 -0.5 -1.0 -1.5 2.0
2.5 3.0 3.5 4.0 4.5
5.0 4.5 4.0 3.5 3.0 2.5 2.0(pp m)1.5 1.0 0.5 0.0 -0.5 -1.0 -1.5
5.0 4.5 4.0 3.5 3.0 2.5 2.0(pp m)1.5 1.0 0.5 0.0 -0.5 -1.0 -1.5
5.0 4.5 4.0 3.5 3.0 2.5 2.0(pp m)1.5 1.0 0.5 0.0 -0.5 -1.0 -1.5
5.0
2.0-(-0.4) (11H, B12H11) 3.1 (24H, N+(CH3)4)
2.36 (2H, -CH2-) 2.07 (2H, -CH2-)
Figure 19. NMR spectra of acid 32 (solvent DMSO-d6).
It is seen that with increasing VsCID the relative intensity of the ion signal at m/z 123 (2-) decreases while the other two ion signals increase (since all spectra are normalized to the highest intensity peak, which is m/z 87 (2-) in all cases, the relative intensity increase of this ion is not directly visible). Since generally upon increasing the fragmentation energy the intensity of a parent ion is expected to decrease while that of the fragments should increase, this finding clearly shows that the ion signals at m/z 87 (2-) and 141 (1-) must be fragments of the molecular ion at m/z 123 (2-).
We have observed the signal at m/z 141 (1-) rather frequently in ESI mass spectra of substituted B12-clusters; its origin had not yet been investigated. In the case of the molecule considered here, the mass spectrum shows two fragment peaks, which raises the question of the fragmentation pathway. With sCID this cannot be clarified because all ions present in the nozzle-skimmer region at a given time might be fragmented. Hence it cannot be distinguished whether the primary fragmentation product of the molecular ion is m/z 87 (2-) or whether m/z 141 (1-) arises independently. We therefore performed an MS/MS experiment to obtain further information on this issue. In contrast to sCID, MS/MS is selective because the ion to be investigated is isolated before fragmentation. The negative-ion ESI-MS/MS mass spectrum of m/z 123 (2-) is shown in Figure 20.
Figure 20. MS/MS of peak 123.
The spectrum shows that the main fragmentation product is m/z 87 (2-), i.e. the [B12H11SH]2- ion.
This fragmentation reaction can be easily explained by abstraction of acrylic acid (Scheme 30).
Upon protonation, the resulting [B12H11SH]2- ion can lose H2S, to give m/z 141 (1-).
S OH
O
SH
2
-m/z=123.1 z=-2
m/z=87.1 z=-2
m/z=141 z=-1
+ OH +
O
H2S H+
Scheme 30. Generation of fragment ion of 32 in ESI-MS.
This is also in accordance with the empirical rule that mass spectrometric fragmentation of even-electron ions occurs preferentially by abstraction of small neutral molecules; frequently, the driving force for the reaction is the stability of the neutral molecule. In this case, both the resulting fragment ion and the neutral acrylic acid are stable particles. For [B12H11SH]2-, H2S as the stable neutral fragment is generated by protonation of the sulfur.
The obvious conclusion from this result is that the ion at m/z 141 (1-) must be a secondary fragmentation product of m/z 123 (2-), i.e. it must originate from the [B12H11SH]2- ion (m/z 87 (2-)). To test this, we subjected the ion at m/z 87 (2-) from the MS/MS spectrum shown in Figure 20 again to MS/MS conditions, i.e., we performed an MS3 experiment. The resulting spectrum showed a weak signal at m/z 141 (1-), which was found to increase with increasing residence time of the m/z 87 (2-) ion in the ion trap. Thus, the m/z 141 (1-) signal does not only stem from
the sCID region, but can also be formed under MS/MS conditions. The same signal was observed with an MS/MS spectrum of [B12H11SH]2- (Figure 21).
86.9
141.0
0 200 400 600 800 1000 Intens.
60 80 100 120 140 160 180m/z
86.9
141.0
0 200 400 600 800 1000 Intens.
60 80 100 120 140 160 180m/z
Figure 21. MS/MS of [B12H11SH]2-.
Obviously, the formation of m/z 141 (1-) requires the presence of the background gas which is present in the nozzle-skimmer region, but whose concentration is low in the high vacuum of the ion trap. In the skimmer region, this background gas, although it consists mainly of nitrogen, also contains residual solvent vapor as source for protons; we can tentatively explain this fragmentation reaction to occur via a gas-phase proton attachment to the doubly charged BSH anion, followed by abstraction of H2S (see Scheme 30). The same explanation might hold for the high vacuum conditions in the ion trap; here, due to the low background (and hence also solvent) gas pressure, longer residence times of the precursor ions are needed to form detectable amounts of the product ion.
From the protonated form of B12H12(2-) ion, H2 abstraction has been found to be energetically possible and leads to a rather stable B12H11(1-) anion. The energy required has been calculated from only 3 kcal/mol (Mebel 1989) to 7.5-11 kcal/mol (Mebel 1999) (Scheme 31).
B12H11
-B12H13+
+3-11 kcal/mol -H2
Scheme 31.
The MS/MS spectrum of [B12H11]- showed no fragment ions, but instead a series of peaks at higher masses with mass differences of 16 amu. To clarify the processes occurring here we chose to repeat the MS/MS experiment with monoisotopic exitation, i. e. only the highest intensity peak of the B12H11 isotopic distribution was selected and fragmented. The resulting spectrum is
shown in Figure 22. Two series of peaks can be distinguished, which can be attributed to two different processes:
• The first process (red line) starts from [B12H11]- at m/z 141 to give the peak m/z 157. The difference in 16 amu gives us the idea about attachment of H2O and abstraction of H2. Thus, the peak at m/z 157 should correspond to the structure [B12H10OH]-. Hydroxylation can be observed until 7 hydroxyl substituted ions (Figure 22) and finishes with [B12H4(OH)7]- (peak m/z 253).
H O 141.2
157.2
173.2 189.2
205.2 221.2
255.2
271.1
287.2 303.2 237.2
0 200 400 600 800 Intens.
100 120 140 160 180 200 220 240 260 280 300m/z 253.2
+H2O
141 157
[B12H11]- [B12H10OH] --H2
m=16
+H2O -H2
m=16
[B12H9(OH)2] -173
...
Figure 22. MS/MS of 141. Process 1 (hydroxylation).
In Figure 23 we present the presumable mechanism of hydroxylation with abstraction of one molecule of hydrogen.
B B
B
H H
H
O H
H
B B
B
OH H
-H2
H
Figure 23. Presumable mechanism of hydroxylation.
In collaboration with Dr. Matthias Hofmann from Anorganisch-Chemisches Institut Heidelberg, theoretical calculations were done. The calculations showed that hydroxylation of [B12H11]- unit occurs with substitution of hydrogen preferentially in position 7, 8 and 10 (Figure 24).
Figure 24. Hydroxylation of [B12H11]-.
• The second process (Figure 25, blue line) is first observed with the fourfold hydroxyl-substituted ion [B12H9(OH)4]- (peak m/z 207). The origin of this peak can be explained by attachment of H2 (probably from the previous process) to the [B12H7(OH)4]- ion.
Further hydroxylation proceeds analogously to the first process, finally leading to [B12H3(OH)10]- (peak m/z 303).
141.2
157.2
173.2 189.2
205.2
239.2 255.2
271.1
287.2 303.2 223.2
207.2
0 200 400 600 800 Intens.
100 120 140 160 180 200 220 240 260 280 300m/z
[B12H7(OH)4]
-m=2
+H2O -H2
m=16
[B12H9(OH)4]- ...
+H2
205 207
[B12H8(OH)5] -223
+H2O -H2
m=16
Figure 25. MS/MS of 141. Process 2 (attachment of H2 and subsequent hydroxylation).
We suppose that the presumable mechanism of the second process can occur with forming two endo BH bonds (Figure 26).
B B B
H H
H
O H
H
B B
B
OH H
-H2
B B
B
H H
H
H H
+H2
H
Figure 26. Presumable mechanism of attachment of H2.
Such kind of attachment of H2 is not known yet. In the literature we found only one example of a B12-containing compound with a bigger amount of hydrogen than 12. Grimes and co-workers (Brewer 1984) synthesized the neutral dodecaborane B12H16 from the K salt of B6H9.
These two processes are connected with the presence of molecular water in the skimmer-CID region or even in the ion trap of the electrospray mass spectrometer. There are several sources of water: 1. Background gas. Nitrogen, which is used as a background gas in s-CID, can contain 16 ppm of water. 2. The walls of the ion trap contain water. 3. Solvents. Before measurement, the sample must be dissolved in an appropriate solvent. There are several solvents which can possibly be used for ESI-MS, as H2O, methanol, ethanol, acetonitrile, etc. These solvents are of the purest quality obtainable, but are not dried before use.
To prove the presence of water inside mass spectrometer and to prove the origin of the peak 141, we decided to synthesize [B12D12]2-(Leites 1982).
The MS/MS spectra of [B12D11]- showed a slightly complicated picture (Figure 27). D2O was used as injection solvent. Peak 152 corresponds to the [B12D11]- ion. For the MS/MS experiment of the [B12D11]- ion the peak with highest intensity of isotopic distribution was chosen. As we expected, the difference between hydroxylated ions changed from 16 to 15. The first process of hydroxylation is shown in Figure 27 by the red line. But now it is difficult to determine the amount of hydroxylated products.
149.2
152.1
227.1 273.1
0 25 50 75 100 125 Intens.
100 120 140 160 180 200 220 240 260 280 m/z
166.2 181.2
195.2 211.2
242.1 257.1
B12D11
-+H2O -HD
B12D10OH
-+H2O -HD
B12D9(OH)2- . . . 152
m= 15
167 182
m= 15
Figure 27. MS/MS of 152 B12D11- (process 1).
The second process, the blue line, starts from attachment of H2 to [B12D6(OH)5]- (peak m/z 227) to give [B12D6H2(OH)5]- (peak 229). Subsequent hydroxylation of the latter is observed (Figure 28).
149.2
152.1 166.2
181.2 195.2
211.2 227.1
242.1 257.1
273.1
0 25 50 75 100 125 Intens.
100 120 140 160 180 200 220 240 260 280 m/z
+H2O -HD
m= 15 B12D6(OH)5- +H2
B12D6H2(OH)5- B12D5H2(OH)6- +H2O -HD m= 15
B12D4H2(OH)7- . . .
227 229 244 259
Figure 28. MS/MS of 152 [B12D11]- (process 2).
The spectrum shows several signals, which are not explained yet. The peaks with mass 1 amu less than the peaks of the first process can possibly be explained by exchange of deuterium in [B12D10(OH)]- to hydrogen (Figure 29, purple line). Further hydroxylation of the product obtained by H-D exchange is also observed (the difference between the peaks is 15 amu).
149.2
152.1 166.2
181.2
195.2211.2
242.1 257.1
273.1
0 25 50 75 100 125 Intens.
100 120 140 160 180 200 220 240 260 280 m/z
227.1
B12D9H(OH) -166
+H2O
-DH B12D8H(OH)2- +H2O -DH
B12D7H(OH)3
-181 196
m=15 m=15
. . . B12D10OH- +H2
-DH 167
Figure 29. MS/MS of 152 [B12D11]- (process 3).
The last series of peaks starts from attachment of H2 to [B12D5H(OH)5]-. This ion, as the previous ones, can exchange deuterium atoms to hydroxyl group (Figure 30, turquoise line).
149.2
152.1 166.2
181.2
195.2211.2 227.1
242.1 257.1
273.1
0 25 50 75 100 125 Intens.
100 120 140 160 180 200 220 240 260 280 m/z
+H2O -DH
+H2O -DH
m=15 m=15
. . . +H2
B12D5H(OH)5- B12D5H3(OH)5- B12D4H3(OH)6- B12D3H3(OH)7
-226 228 243 258
Figure 30. MS/MS of 152 [B12D11]- (process 4).
Since we observe exchange of deuterium to hydrogen, the process of exchanging deuterium atoms to HD is also possible (Figure 31). All these signals are present in the spectra and can belong to different processes.
B12D4H3(OH)4
-+HD
B12D5H4(OH)
-209 212
B12D2H3(OH)5
-+HD
B12D3H4(OH)5 -226 223
B12D3H3(OH)5
-+HD
B12D4H4(OH)5
-224 227
B12D4H2(OH)5
-+HD
B12D5H3(OH)5
-225 228
B12D5H1(OH)5
-+HD
B12D6H2(OH)5
-226 229
Figure 31. Exchanging of D to HD.
Our explanations of the MS/MS spectra of the [B12H11]- and [B12D11]- ions are only tentative and must be verified by additional experiments. High resolution MS measurements could confirm the molecular formulas of the proposed ion structure. Unfortunately, this was not possible with the available equipment, but probably MS equipment of new generation can solve this task.