OF MOLECULAR SPECTROSCOPY 38, 16-32 (1971) Porphyrins XVII. Vapor Absorption Spectra and Redox Reactions: Tetraphenylporphins and Porphin 1 L. EDWARDS AND D. H. DOLPHIN Department of Chemistry, Harmrd UnI:versity, Cambridge, 1\1. GOUTER1-IAN Department (if Chemistry, Unit.ersily of Washington, Seattle, Wa.shingt.on 98105 AND A. D. ADLER New England [!t.stitute, Ridge.fieltl, Conn.ecticut 06877 Vapor absorption in the range 800-200 IllIi are given for porphin, free·base tptraphenylporphin (H,TPP), and the Mg, Cr(HI), Mn(HI), Fe(HI), Co, Ni, CIl, Zn, Cd, Sn(IV), and PbTPP complexes. Electronic bands Q, B, and ]If characteristic of the ring are found in the regions 540 mIL (18 500 cm- l ), 405 mIL (24 700 em-I), and 205 IllIi (48 800 cm- l ). The region between the B and the III shows rather unstructured absorption, an order of magnitude less int.ense thalL the B region. Based on spectroscopic measurements, the Ml of vaporization and absolute vapor pressures are calculated. The pressures average t.o Torr at 390°C. Vacuum uv spectra of H.TPP, NiTPP, CuTPP, and ZnTPP down to 145 IllIi are presented which show absorption across the entire range with more intense bands at 185 mIL (54000 cm-'), 164 mj.! (61000 rm-'), and 150 mIL (67 000 cm- l ) , and no evidence of Rydberg structnre. The vapors of Cr, Fe, and 8n complexes show unusually different. spectra from their solu· tions. Spectral changes on electrolytic reduction establish that in the Mn, Fe. and Rn cases the vapors are those of t.he divalent metal. Rolution spectra of H,TPP, CIlTPP, and FeCITPP were studied lip to 300°C. Broadening com parable t.o the vapor spectra is observed. Also t.he visible bands shift to the rpd while the Soret band shows no change. 'This research was supported in part by Public Health Service Research Grant 14292 from the Institute of General Medical Hciences. and by U.s. Army Research O1flC(' (Dnrharnl Clrant DA·31·124·ARO(D)-10l (formerly held by A. A. at the of PPllllsylvania) . Hi
17
Embed
Porphyrins XVII. Vapor Absorption Spectra and Redox Reactions ...
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
JOUR~AL OF MOLECULAR SPECTROSCOPY 38, 16-32 (1971)
Porphyrins XVII. Vapor Absorption Spectra and Redox Reactions:
Tetraphenylporphins and Porphin1
L. EDWARDS AND D. H. DOLPHIN
Department of Chemistry, Harmrd UnI:versity, Cambridge, :lfa8!lachll.~eU.~ 021.~8
1\1. GOUTER1-IAN
Department (if Chemistry, Unit.ersily of Washington, Seattle, Wa.shingt.on 98105
AND
A. D. ADLER
New England [!t.stitute, Ridge.fieltl, Conn.ecticut 06877
Vapor absorption in the range 800-200 IllIi are given for porphin, free·base tptraphenylporphin (H,TPP), and the Mg, Cr(HI), Mn(HI), Fe(HI), Co, Ni, CIl, Zn, Cd, Sn(IV), and PbTPP complexes. Electronic bands Q, B, and ]If
characteristic of the ring are found in the regions 540 mIL (18 500 cm- l ), 405 mIL
(24 700 em-I), and 205 IllIi (48 800 cm- l ). The region between the B and the III shows rather unstructured absorption, an order of magnitude less int.ense thalL the B region. Based on spectroscopic measurements, the Ml of vaporization and absolute vapor pressures are calculated. The pressures average t.o ~0.05 Torr at 390°C. Vacuum uv spectra of H.TPP, NiTPP, CuTPP, and ZnTPP down to 145 IllIi are presented which show absorption across the entire range with more intense bands at 185 mIL (54000 cm-'), 164 mj.! (61000 rm-'), and 150 mIL (67 000 cm- l ) , and no evidence of Rydberg structnre. The vapors of Cr, ~fn, Fe, and 8n complexes show unusually different. spectra from their solu· tions. Spectral changes on electrolytic reduction establish that in the Mn, Fe. and Rn cases the vapors are those of t.he divalent metal. Rolution spectra of H,TPP, CIlTPP, and FeCITPP were studied lip to 300°C. Broadening com parable t.o the vapor spectra is observed. Also t.he visible bands shift to the rpd while the Soret band shows no change.
'This research was supported in part by Public Health Service Research Grant (~M· 14292 from the Institute of General Medical Hciences. and by U.s. Army Research O1flC('
(Dnrharnl Clrant DA·31·124·ARO(D)-10l (formerly held by A. A. at the Universit~· of PPllllsylvania) .
Hi
TETRAPHENYLPORPHIN VAPOR ABSORPTION 17
INTRODUCTION
Earlier papers in this series (1,2), hereafter known as XV and XVI, reported the vapor and high temperature solution spectra of a number of phthalocyanines (Pc) and octalkylporphins (OAP) , these latter being represented by octaethylporphin (OEP) and etioporphyrin complexes. In this paper we report the continuation of these studies to the tetraphenylporphins (TPP) and porphin itself. As in XV and XVI we report the D.H's of vaporization based on the change of optical density with temperature. We also compare the effect of change of temperature with change of phase by studying the absorption of several TPP's in silicone oil between room temperat.ure and 315°C. As in XVI, we studied the absorption spectra of several reduced metal complexes.
Previous to XVI t.here had appeared only one spectroscopic investigation of a porphyrin vapor. Mullins, Adler, and Hochstrasser (3) vaporized and photographed the Soret bands of copper and free-base TPP as well as part of the visible bands of t.he latter. Aside from XVI the only systematic study of porphyrins in the uv is the work of Caughey et al. (4), who compared their findings to the calculations of Weiss et al. (5) and used these to assign bands.
In this work, we continue our spectroscopic study of the substituted porphyrins dmvn to 200 miL and lower, as well as extend the study to unsubstituted
Q PORPHIN c FREE BRSE TETRAPHENYlPOAPHIN Cl
~ 0 Cl
- RX 00 IUV) ~37 MAX 00 [VIS) =.432 MRX ODIUV)=.455 MRX 00 IVIS) =.345 -T=322·C B T=393·C T=395·C B T=ijijB"C
If) ----.. 4-- ----.. 4-- If) .r- Qy ~ Qy
r-OD .,; D 0
0 0 If) If)
D 0
c:i • If)
'" '" ON N '" .,; ci
0-1
0 MRGNESIUM TETRRPHENYLPORPHIN 0 Cl c: ~ Cl
- MRX ODIUVI=.442 MAX 11IHVIS) =.361 - MAX 00 IUVI =.177 MAX 00 [VIS) =.226 T=393·C B T=ij67"C T=393·C B T=ij69·C
'" ----.. Q 4-- If)
----.. ..-'" .r- r- r-
c;ci .,; 0
0
";0 5l 0
a:~ '" :::EO .,;
M 0
'-c:i .'" '" on
0", N N
D .,; 0-1
200 300 ~oo SAO 500 ?OO 800 200 sao 500 700 800 WRVELENGTH 1M I LLI MICR~NS) I HI LLI M I CR~NS)
FIG. 1. Vapor phase absorption spectra. Temperature and maximum OD (corrected to 1 em pat h length) for two runs are indicated.
18 EDWARDS ET AL.
free-base porphin. Spectroscopically, the latter is the most important porphyrin, and we report here the first observation of its far uv bands.
Analogous to the OEP's we find that the vapor spectra of the 1\ln(III), Fe (III), and Sn(IV) complexes of TPP show large differences from their solution spectra. Similar spectral changes are reported for these compounds upon reduction. This gives strong grounds for suggesting that the reduced forms of these complexes are stable in the vapor. CrCITPP also showed similar spectral changes on vaporization, suggesting reduction, but could not be reduced in solution.
An attempt was made to sublime zinc tetraphenylchlorin (ZnTPC), but the compound decomposed with no yield.
EXPERIMENTAL
A. Compounds and Muss Spectral Analysis
All the compounds, except CrCITPP, were prepared as previously described (6, 7). CrCITPP has been synthesized only recently and the procedure will be reported elsewhere (7). Elemental and mass spectral analyses establish its identity. All compounds were recrystallized, chromatographed, and finally,
FIG. 2. Vapor phMe absorption spectra. Temperatnre and maximum OD (corrected 1 () 1 rrn pat h length) for two rnns are indicated.
TETRAPHENYLPORPHIN VAPOR ABSORPTION 19
vacuum sublimed. As in XVI, the visible bands of the Mn and Fe complexes were broader after sublimination.
An MS-9 mass spectrometer was used to determine purity of all compounds. For all except the Cr(III), ~In(III), Fe (III) , Sn(IV), and Pb complexes, the parent peak was essentially the only peak observed for low ionizing voltages (7-12 eV). (H2TPP and CdTPP show spurious peaks of MnTPP and CuTPP introduced by the MS-9.) PbTPP shows spurious MnTPP and CuTPP peaks as well as an H2 TPP peak. The H2 TPP appears as a small peak in the absorption spectrum (see figures). PbTPP is relatively labile and can easily convert to the free base on handling. By contrast, the parent peaks of CrCI, MnCI, and FeCITPP are about 750 as intense as the base peaks, which are Cr, Mn, and FeTPP, respectively. For FeOHTPP, no parent peak but only FeTPP was observed. That the starting material was FeOHTPP is inferred from its spectra (8) and synthesis. We also note that exact mass studies of sublimed 8amples of MnCI and FeCITPP at low ionizing voltage show a mass peak corresponding to Cl substituting for H. This peak is absent at such low ionizing voltages in the unsublimed materials. However, as t.he ionizing voltage is increased, the mass spectrum of unsublimed FeCITPP shows this peak. This
FIG. 3. Three vapor spectra, presented as in other Figs., and one liquid spectrum. The lower left spectra is attributed to Fe(II) complex. See text.
20
ci ci
EDWARDS ET AL.
MAX IIlIVISI:.209 T:ijSS·C ..-
IIlIUVI: .522 B T:IIDfj·C --
MAX IIlIVISI:.182 T:II62·C ..-
8~~~~~~~~~~~~~~~ HAX OO(VISJ=.", ...:
'" .r-C:ci D
200 300 ijOQ WAVELENGTH
T:ij3ij"C ..-
300 ijOO 500 sao 700 WAVELENGTH (MILLIMICR~NSJ
'" r-o
800
FIG. 4. Three vapor-phase spectra presented as in other Figs. and one liquid spe(·t rUIll
(room temperature). The upper left spectrum appears initially and contains some Mn(I111 while the upper right spectrum taken after 10 min is largely Mn(II). The Sn vapOf Ape(·t flllll is attributed to a mixtnre of SIl(Il) and Sn(IV). See text.
implies that FeTPP is reacting with the greater concentration of atomic chlorilH'. and the resultant chlorinated TPP is still thermally stable at these temperature~. }lore details of the mass spectral studies are given elsewhere (9, 10).
Analytic data shows that starting Sn(IV)TPP was Sn(Ac)2TPP. HOWl'vt'l', upon ~ublimation the principal product appears to bt' Sn(OH)~TPP. Knell Imblimed material was used for the vapor phase studies.
B. A pparatu8
All vapor phase spectra were taken as in XV. The high-temperature solution :,;pectra and electrolytic spectra were taken as in XVI. The reduction of the Sn(IV)TPP waR effected chemically by successively adding small amountR of NaBH4 to a methanol solution of the compound in a Cary 14 cell until the Hpeetrum changed to the solution spectrum shown in Fig. 4. The spectrum wa:,; al~o produced by reduction with ascorbic acid in dimethylformamide. Furthermore. coulometry during controlled potential reduction showed that a reversiblP twoelectron reduction product gave the same spectrum.
TETRAPHENYLPORPHIN VAPOR ABSORPTION
V~CUUM UV ~BSCIAPT I ON OF COPPER AND FREE BASE TPf> WAVELENGTH (MILLIMICRONS)
251) 230 210 190 170 150 2
SCLlD LINE IS C!II'PEA TETftAPHENYLPIIV'HIN AT S9ti'C ~ DIITTEO LINE IS FREE BRSE TETAAf'HENYlP!IfI'HIN AT 385'C
'" z ~I!l --,0 ei
XIIl XI21
"
• I , ! , I I I • ! I I"' , ! I
VACUUM UV ABSORPTI ON OF NICKEL AND ZINC TPP ~VELENGTH (MILLIMICRONS)
o
XCS)
25Il 290 210 190 17D 151) 2
SOLID LINE IS NICKEL TETAII'I£NTlP!IfI'HIN AT ijIS'C >- roTTED lINE IS llNC TETPilPHENTlP!IfI'HIN AT ijQS'C ..... v; z ~~ --,0 a: u lt>.. o
X(2)
E 51) ~ ® ~ FREaUENCY ITHClUSANDS ClF WAVENUMBERS)
Q
XIS)
70
FIG. 5. Vacuum ltV absorption spectra: OD corrected to 1 em path lengt,h.
RESULTS
A. Vapot Spectra of Divalent Species
21
The results of these studies are given in Fig. 1-5 and Tables I-III. In the spectra, we can always identify the three bands Q, B, and M which are labeled in the graphs and whose maxima are given in Tables I and II. The first two have long been characterized (8,12). We note that MgTPP, CoTPP, ZnTPP, CdTPP, and PbTPP show different distribution of intensity in their visible bands in vapor from the solution (8, 9). Specifically, MgTPP and CdTPP show Q(G-O) absorption in solution, but in vapor the intensity is greatly reduced. ::VlgTPP, CoTPP, ZnTPP, and CdTPP show Q(o-l) hot bands in vapor which are labeled in the figures. In addition, CoTPP shows a red band (,,-,700-800 mIL) in vapor not present in solution. The greatest change from solution to vapor is manifested by PbTPP. In solution, PbTPP shows a clear vibronic progression of four bands monotonically decreasing in intensity from the Q(O-O). However, in vapor only the first two are seen and these are much broader than in solution.
The only other band which consistently appears in the spectrum is the M band, which has been previously assigned in XVI. In addition to these bands, we often see an N band, assigned previously (4). In the octalkylporphins (2), the band appears much more consistently than in the TPP's. We have also labeled certain L bands, which have been identified as characteristic of the porphyrin
22 EDWARDS ET AL.
., 0 0 .,
SOLVENT 15 CH,Cl,
"' l!' ..... 0 c::J •
• 0
0
XO 0
'" a:"! 0 :>:0 "-ci
'" .", N ON 0 0
0 E! "! 5!l.VENT IS CH,Cl,
'" II> ..... .... c::J • 0 • 0 c
xijl 0
'" a: . 0 ::0:0
" ci. III C '" d ci
8 8 SIILVENT IS CH,Cl,
.\!! \!! c·
.0 0 c
xijl 5l a:. 0 ::0:0
" ci l(l c~
0 0
350 ij50 550 650 750 WAVELENGTH (MILLIMICRONS)
FIG. 6. Lower two panels show the spectra produced by electrolytic reduction of MnClTPP, whose spectrum is given in top panel. For Mn(II)TPP with OR, E(570m/ol)!E(435) = 0.031. For Mn(II)TPP with pyridine, E(611ffi/o1)!E(436m/ol) = 0.039.
ring (4), but, as pointed out in XVI, they are more properly considered metal dependent.
From the figures and the results in XVI we see that the TPP's show less structured absorption than the OEP's in the 200-400 mJ' range. Thus the N band and even more so the L band are sometimes washed out by the surrounding absorptions. This makes the assignment of bands more difficult and more tenuous. In contrast, the TPP Soret bands are significantly narrower than the corresponding OEP's, especially the free-base ones.
Some interesting trends are evident from Table 1. The closed-shell metal TPP's show a red shift (vapor ---lo room-temperature solution) for the Q(l-Q), while the open-shell metal TPP's show a blue shift. The closed-shell metals
8.---..----'~~*-~....,!..W~4-!.!~"*_.)!!..,...!~~=-..---r--,~· SIIlVENT IS CHI Cl2
~50 sso 7SO WAVELENGTH (HILLIHICRONS)
23
FIG. 7. Lower two panels show the spectra produced by electrolytic reduction of FeCITPP, whose spectrum is given in the top panel. For Fe (II)TPP with OH, e(522mp.)/.(420) = 0.033. For Fe(II)TPP with pyridine, e(528mp)/E(423mp.) = 0.106.
(except Pb) have a larger red shift of the Soret band than the closed shell with PbTPP having the least shift.
In the course of the absorption studies, new bands appeared with time in the visible absorption of both NiTPP and porphin. In the case of NiTPP, a band appeared at 750 mIL with intensity about equal to the 531.5 m,. band. There was considerable decomposition during the run, which is unusual for the divalent TPP complexes.
In the case of porphin, a band appeared at 642 mil- many minutes after the temperature was high enough for visible absorption. Upon reheating the cell many days after the initial run, the 642 mIL peak appeared immediately along
24
.:e o· .0
o
><&1 CI:. :0:0
" o .11>
ON ci
II> .... O· .0
o
o .Ln
o,,! o
II> .... o· .0
o
350
EDWARDS ET AL.
550 650 (HILLIMICR~NSJ
8
SOLID LINE GARPHS ARE I!? ATAIIIIH TEIf'EAATlft ci
DOTTED LINE GRAPHS ARE &1 CL 2911'C ci VIS-25S'C
I(l ci
0 0
SOLID LINE GAAI'HS RIlE II> AT fIIlIIII ... TElftAATIft ci
D!lTTED LINE GRAPHS ARE Ii! nJ- 3lJII'C 0 VlS-3Qlj'C
III N
0
8
SIl.ID LINE GAAPHS AfIE II> AT ROClH ... TEHPEAATUAE 0
mTTED LINE GAAPHS Fft: &1 M- 3QO'C 0 VlS-3Qlj'C
II> N
0
750
FIG. 8. Solution absorption spectra at high temperatures; OJ) at high temperatures has been corrected for the expansion of the solvent.
with the porphin visible bands. There was no evidence of decomposition during this or the initial experiment. On opening the cell and dissolving it.s contents in benzene, the normal porphin spectrum was observed along with a band at 63i .. 'i IllJ.£ about h the intensity of the Qu(1-o) absorption. Some residue failed to dissolve in the benzene. Chlorin, first reported by Eisner and Linstead (13), has a strong visible absorption in benzene at exactly this wavelength. l\loreover, they noted that the compound is stable in boiling dichlorobenzenf' (bp 130°C), and that it is not oxidized to the corresponding porphyrin by oxygf'n at room temperature. We conclude then that we did indeed form chlorin ill thf' vapor and that it is stable unlike ZnTPC and Co etiochlorin. (We found that the former fails to vaporize while the latter converts to the porphyrin.) It should be pointed out that since the Qx(O-O) vapor peak of porphin was largely ob-
Table 1. Q and B Band Max1ma (m~) Band Wldths (m~), and Vapor-Solutlon Sh1rts (cm-~). Compound Solvent Q(O-O) Q(l-O) B
scured by the 642 m~ peak, its quantitative determination was impossible and its intensity appears overestimated in Fig. 1.
The results of the vacuum uv studies are shown in Fig. 5. It is clear from Fig. 5 that there is absorption across the spectrum with more intense bands at 206 mJl. (48500 cm-I ), 185 mJl. (54000 cm-I ), 163 mJl. (61 000 cm-I ) , and 150 m~ (66500 cm-I ). The first band is the M band and the others are designated X(I), X(2), and X(3), respectively. The oscillator strengths for t.hese last. three bands are 3.0 ± 0.5, 1.0 ± 0.5, and 1.5 ± 0.7, respectively.
From the absorption strengt.h of either the Q bands or the B band, estimates of the vapor pressure as a function of temperature can be made, if it is assumed that the intensity is temperature- and phase-independent. Based on this assumption and using the extinction coefficients of Dorough et al. (8), the vapor pressures of the compounds at the temperature of the experiments were calculat.ed and are listed in Table III.
- of the Soret band. b Calculation impossible due to lack of temperature equilibrium. c ±3.
29
AHc (kcal/mole)
26 35 30 b
26 33 26 23 25
d For the Soret of porphin Av, = 1000 cm-1 and E = 264000 (11). For the TPP's the f'S
of Dorough. et al. (8) are used with the half-widths of Table I.
B. Vapor and Solution Spectra oj Polyvalent Species
In XVI we concluded that the stable species in the vapor experiments on MnCIOEP, l\lnAcOEP, and FeCIOEP were the divalent metal complexes. In the cases of MnCITPP, FeCITPP, and FeOHTPP, our results give even stronger reasons for drawing the same conclusions. The vapor spectra (Figs. 3 and 4) differ markedly from the solution spectra of the trivalent species but agree nicely with the spectra produced on electrolytic reduction (Figs. 6 and 7). This is particularly apparent when comparing the vapor Soret-band region with those of the di- and trivalent solution spectra. Moreover, FeCITPP and FeOHTPP give identical vapor spectra, as might be cxpected if the counter ion were cleaved. In the MnCITPP vapor spectrum, a transient peak at 474 mJL can be seen which is cleary attributable to the trivalent metal. The peak survives for some minutes in the vapor spectrum. Further evidence for the homolytic cleavage of the counter ion is given by the mass spectra (discussed above) and by the facts that the vapor peaks shift a few mJL and the relative band intensities of the iron and chromium complexes change during the two-hour course of a vapor experiment. These changes can be attributed to substitution on the ring. Comparison of the solution spectra of sublimed and unsublimed material alRO suggests that on sublimation some substitution is occurring.
The vapor spectra of Cr(nI) and Sn(IV)TPP (Figs. 3 and 4) also differ markedly from the solution spectra. In these cases the nature of the principal species is not so clear. CrCITPP behaves very similarly to MnCITPP and FeCITPP: The principal peak of the mass spectrum is not the parent CrCITPP but CrTPP. Like the iron case, the vapor spectrum changes slowly with time and
30 EDWARDS ET AL.
there is extensive decomposition. Unlike the ~lnCI and FeCITPP, the mass spectrum of sublimed CrCITPP shows no CI substitution on the ring. l\Ioreover, attempts to reduce CrCITPP in solution by chemical or electrolytic meanH failed as did all attempts to oxidize it. In the case of Sn(IV)TPP, a spectrum that corresponds to the vapor can be produced by chemical reduction as shown in Fig. 4. (The baseline of this spectrum rises towards short wavelengths.) Two small peaks in the vapor spectrum that are in correspondence with the Sn(I"\,) TPP solution spectrum (8,9) are noted in Fig. 4. The tin vapor spectrum dol'S not change Vlith time and there is no decomposition.
c. Temperature-Dependent Solution Spectra
To compare the relative importance of change of temperature with change of phase, we studied several TPP's in silicone oil, which can be taken above 300°C without decomposition. The results are presented in Table I and shown in Fig. 8. Similar spectra of ZnTPP are given in Ref. (9).
The results in Fig. 8 are typical and show that upon increasing the temperature all bands broaden and the visible bands shift 4-9 mil to the red. The Soret bands do not shift within experimental error. From Table I we see that, the bandwidths at the highest oil temperatures are approaching the vapor bandwidths implying that the vapor bandwidth is due mainly to temperature. This was also observed for the OAP's (2) but not for the phthalocyanines (1). Noteworthy is the behavior of H2TPP. The Qy(O-Q)-Q.,(O-Q) splitting is 2764 cm-l
in CH2CI2, 2864 cm-l in DMOE and benzene, 2836 cm-l in oil at 30°C, 2910 cm-l in oil at 204°C, and 3255 cm- l in vapor at 448°C. A plot of the splitting of the Q(O-Q) bands vs. temperature shows no possibility of extrapolating to the vapor phase data (9). This is similar to the case of H2Pc (1) and in contrast to the case of H20EP (2,9).
DISCUSSION
A. Electronic Transitions
In XVI we compared our results to the latest calculations on metal porphin (14). As expected, the TPP spectrum is predicted with about the same accuracy as the OAP's. An important difference is the absorption at 184 mil (54 000 cm- l ) , which is much more intense than its OAP counterpart (2). It is likely that this added intensity is due to a strong transition localized on the phenyl groups, i.e., the benzene transition observed at 178 mil (56300 cm-I ) (15). Hammond and Price (16) photoelectrically measured the extinction coefficient of benzene in this region and determined that} = 0.88 with f max (56 300 em-I) =
52 500 liters/mole cm. It might be valuable to compare our absorption spectra of free-base porphin
with recent calculations (14), since these latter raised several questions. Calculations were done for several values of .1a, the difference between the Cou-
TETRAPHENYLPORPHIN VAPOR ABSORPTION 31
lomb energy of aza- and pyrrole nitrogens. These calculations for Aa = 5.5 eV, half the normal value, are listed in Table IV and compared to our experiments on porphin. The energies of Qx and Qy are nicely predicted as are their weak intensities. However, the Soret-band intensity is predicted to spill out into three states, labeled Bx , By , and Nx spread over 4200 cm-I . This does not seem to happen experimentally. However, the data does show that the free-base Soret bands are substantially broader (half-width 1900 cm-I ) than the narrower metal Soret bands: MgTPP, CuTPP, and ZnTPP Soret-band half-widths are between 950 and 1350 cm-I • If the extra bandwidth is due to a Soret-band splitting and Gaussian shapes are assumed, it can be shown that a Soret-band splitting "-'1500 cm-I is needed to account for the extra broadening. The calculations would then correspond to our experiments if some of the calculated N x intensity is transferred to the Bx bands. Our spectrum of H20EP in XVI shows similar Soret-band broadening.
The fact that H2TPP and H2Pc (1) show a phase dependent Qy-Qx splitting suggests that upon vaporization a geometry change occurs. Such a change is consistent with the rearrangement of intensity in the visible bands of some of the metal TPP's. That such a geometry change is much larger in H2Pc than H2TPP is probably because the H2Pc protons are bridged among three nitrogens (1) while the H2TPP protons are bonded to only one (17).
TABLE IV
TRANSITIONS OF FREE-BASE PORPHIN
Label SCMO-PPP-CI" Experimental E(cm-I )
Q. 15 000 Qy 17 000 E. 27 500 By 28 800 N. 31 700 Ny 35000 Ly 38400 Lx 40 200 Mx 47 000 llfy 47 600
47 900 49 100
• From Ref. (14) using .ill! = 5.5. b Based on data given in Ref. (11). c Half-width = 2300 em-I.
d In solution, N. and Ny are resolved and 1400 em-I apart. e Broad.
f
.01
.06b
1. 15b ,c
<.1
~.1
~.3
32 EDW ARDR ET AL.
B. Fapor Phase Reduction
As discussed above, our results make clear that the ::\In(III), l'e(III) , alld Sn(IV)TPP complexes that we studied convert to the divalent species on vaporization. We would also like to postulate that the difference between the solution spectra of Cr(III)TPP and its vapor spectra arises because the principal vapor species is the divalent metal. However, this interpretation is still open to quet:tion in the case of CrClTPP because we ,vere unable to oxidize or reduce the compound in solution.
ACKNOWLEDGMENT
Dr. S. Dym and Dr. J. Mullins and Messrs. F. Kampas and R. Holland assisted with f he preparation and purification of some of the materials. One of us (L.E.) would like to thank t.he National Institutes of Health for a predoctoral fellowship. We would like to thank Mr. J. Larabee, Dr. R. Hoffman, and Dr. Y. Tanaka of the Air Force Cambridge Research LlIhoratories, Bedford, Massachusetts, for valuable assistance in the vacuum ltV 8tlldie~
Received: October 30, 1969
REFERENCE:-3
1. Paper XV: L. EDW.\ROS and M. GOUTERMAN, J. Mol. Speclrosc. 33, 292 (1970). 2. Paper XVI: L. EDWARDS, D. H. DOLPHIN, and ~I. (}OUT~~RMAN, J. Mol. Spectl"o.,c. 36,
90 (1970). 3. J. A. MULLIN:';, A. D. AnLt:u, and R. M. HOCHSTRASSJ<:R, J. Chem. Phys. 43, 2548 (1965). 4. W. ~. CA"{;"GHm', R. M. DluL, C. WEISR, and M. (~OLTERM.\::-I, J. Mol. Speclroxc. 16,
451 (1965). 5. Paper III: C. W~;ISS, H. KOBAYASHI, and M. GOUT~:HM.\K, J. Mol. Spec/rosf'. 16. 41.')
(1965). 6. A. D. ADL~;R, L. SKL.\R, F. R. LONGO, J. D. FINARI,LLr, and M. n. FI::-I.\R.:r,LI,.l. Ht·l
erocycl. Chem. 6, 669 (1968), and references therein. 7. A. D. ADU~R, F. R. LONGO, F. KAMPAs, andJ. KIM. J. Inorg.Nuc. Chem. 32,2443 (197/j"1. 8. G. D. DOROUGH, J. H. MILI,ER, and F. M. HrExxEKEN8, J. A 1111'''. Chem. So('. 73, 4:n.')
1969. 10. A. D. ADLt;R, J. H. OUEEN, and M. MAUTNICR, Org. Ma.~., Spectrom. 3, 955 (1970). 11. C. HIMINGTON, S. F. MASON, and O. KENN.-I.RD, SpectTochim. Acla 12, 65 (19581 if). Paper I: M. GOUTERMAN, J. Mol. Spectrose. 6, 138 (1961). 13. U. EIRNER and R. P. LINSTI~AD, J. Chem. Soc. 1966, 3742, 3749 14. A. J. McHUGH, M. GOUTERMAN, and C. WEISS, ullpublished. 15. P. G. WILKINSON, Can. J. Ph1ls. 34, 596 (1956). 16. V. J. HAMMOND and W. C. PRICE, Trans. Faradall Soc. 61, 605 (1955). 17. r;. RrLVl-JRR and A. Tt:I,YNSKY,.I. Amer. Chelll. Soc. 86, 927 (H)64).