- Email: [email protected]
Vacuum ultraviolet absorption spectra of oxygen in liquid and crystalline argon and nitrogen
JOllRNAL
OF MOLECULAR
Vacuum
SPECTROSCOPY
Ultraviolet
Absorption
and Crystalline
Spectra
Argon
14. BASS AKD
ARNOLD
Kational
221-230 (1964)
1%
B~ureau of Standards,
of Oxygen
in Liquid
and Nitrogen* H. P. BROIDA Washington,
D. C.
The ultraviolet absorption spectra of oxygen have been observed in liquid and solid media at low temperature. By freezing slowly from the liquid phase, samples of crystalline argon and nitrogen were prepared with added amounts of oxygen ranging between 0.01% and 10yC.The spectra were recorded photoelectrically, through fused silica windows, over the wavelength range from 1750 to 2900 b. The Schumann-Runge bands of oxygen (B Q&m+ S 2,, 1750 to 2000 d) were observed in solid and liquid mixtures of oxygen in argon or r&rogen at concentrations less than 5%, over the temperature range 60 to 90%. The bands are broad, with half-intensity width of 150 to 550 cm-1 and show no structure. The bands show a slight shading toward the violet in contrast to the bands in t.he gas phase, and are displaced toward lower frequencies by 50 to 400 cn-1 as compared with the positions in the gas phase. These shifts appear to he slightly dependent upon temperature. The changes in position or shape of the bands between liquid and solid or between argon and nitrogen media are remarkably small. At concentration of oxygen above 5y0 the Herzberg hands of oxygen (A Q,+ c S %,-) were observed in the region 2400 to 3000 A.
These hands are also broad (200 to 600 cm-l) and show some partially resolved structure. No differences were observed for changes either in temperature ot rrlatrix material. I. INTRODUCTION
The ultraviolet
absorption
of oxygen
densed phases has been studied previously. studied the spectrum a number of bands
in the region 2400 to 3000 A in conMcLennan,
Smith, atd Wilhelm
(1)
of liquid oxygen in the region 2400 to 3000 A and observed corresponding to the “high-pressure” bands in gaseous
oxygen reported by Finkelnberg and Steiner (2). Prikhotko (3) studied the absorption spectrum of oxygen in crystals of solid solutions of Or-N2 and Oe-Ar at liquid hydrogen temperature. In the region 2400 to 2800 8 she found that the spectra of all the mixtures consist of a series of triplet absorption bands. The triplet group spacing decreased toward shorter wavelengths from about 700 cm-’ to about 3.50 cm-‘. These bands coincide with the absorption spectra of solid oxygen, and have a structure identical with the “high-pressure” bands. * Work support)ed in part by the National Aeronautirs and Space Administration. 221
222
BASS AKD
BROIDA
From her data she calculated a potential energy curve for the upper state of the transition which is very similar to the A 3Z,,+ state of the Herzberg bands. Romand and Granier-Mayence (4) studied the spectrum of solid oxygen (condensed from the gas phase on a surface at 20°K) in the region from 1500 to 2000 i and observed a continuous absorption, increasing in strength toward shorter waveleng:hs. They did not observe any discrete absorption in the region 2000 to 2400 A. H&l (5) prepared thin polycrystalline layers of solid a-oxygen by deposition from the gas phase onto surfaces cooled to near 4.2”K by liquid helium with film thicknesses estimated to be a few hundredths of a millimeter. Between 2400 and 2700 i a number of broad and diffuse bands were observed with half-widths of the order of 200 to 300 cm-‘. Horl suggested that the bands in the solid are the Herzberg bands of oxygen, (A ‘S,+ +- X 3Z,-) shifted to shorter wavelengths by about 90 to 120 cm-‘. Dressler and Schnepp (6, 7) studied thin layers of oxygen deposited in mixtures with argon and nitrogen on a surface at 4.2’K. Between 1750 and 2000 A eleven bands, 50 to 150 cm-’ broad, were observed and identified as the Schumann-Runge bands of oxygen (B 32,- +X “2,). By measuring the isotope shift produced with 0:” they were able to establish that the bands in the solid are shifted toward longer wavelengths by roughly 300 to 400 cm-‘. In the present work absorption spectra of oxygen inocondensed phases have been recorded over the wavelength range 1750 to 2900 A. Both the SchumannRunge bands of O2 and the diffuse oxygen bands in the region of 2400 to 3000 A have been studied. Observations of the oxygen have been made in liquid and crystalline argon and nitrogen at temperatures between 60 and 90°K. Some small variations with change of parameters were observed. However the surprising result has been that the observed spectra show little change between the solid and liquid matrix and are quite similar to the spectra observed at 4.2”K. II. EXPERIMENTAL
Solid crystalline samples were produced in a glass Dewar vessel (Fig. l), by the method described by Bole et al. (8). Argon or nitrogen gas, with varying concentrations of oxygen, was admitted to the central chamber at a pressure of 1 atmosphere. When liquid coolant (nitrogen or oxygen) was added slowly to the surrounding compartment it was possible to produce, between the two fused silica windows (25mm path length), a transparent solid. Solid argon was grown at temperatures from 60 to 82°K. This solid was similar to that described by Bolz et al. (8)) in which they measured grain-sizes of up to 4 mm. By pumping on liquid oxygen, the temperature of the coolant bath could be adjusted as desired in the range 60 to 90°K. Thus the same sample could be observed in the liquid phase as well as in the solid. The concentration of oxygen in the sample was controlled by adjusting the ratio of oxygen to argon (or nitrogen) in the gas prior to condensation. The
ABSORPTION
GAS---W
SPECTRA
OF OSY(;EK
REFRIGERANT
VACUUM
I
EN
FIG. 1. Dewar sropic studies.
vessel
for
prrpttrutionof crystalline
solids of srgtm suitable
for spwtro-
concentrations used in this way ranged from about 0.01 5’ to 10 L% (and also 100 “h oxygen). However, the concentration of oxygen in the condensed mat,erial was probably different from that in the input gas. This was verified in one case where mass spectrometric analysis indicated the “input” gas composition to be 10.3 7 Or , 89.7 7:’ ss ) while the composition of the gas eraporating from the condensed sample to be 4.3 c/? O2 , 95.6 “4’. The temperature of the sample was assumed to be the same as that of t’he surrounding bath, as determined from the vapor pressure of the bath refrigerant. Boiling points and triple points of the sample obtained in this way generally agree with the known values of these fixed points. However, the assumption that, the sample and the surrounding bath are at same temperature is not always reliable. Occasionally we have found that, even with vapor pressure of the bath corresponding to a temperature well below that of the triple point of argon (83.78”K~‘i, a sample of argon in the inner chamber will remain liquid for some
224
BASS AND BROIDA
time. Thermocouple measurements of the liquid argon show that the liquid is not supercooled. We have not been able to explain this lack of thermal equilibrium, but it does cause some uncertainty as to the accuracy of our knowledge of the sample temperature. A side-arm cooled by liquid hydrogen was used to trap out any substances which might otherwise condense on the outside of the fused silica windows of the sample chamber. Prior to the use of this trap, in the region below 1900 8, a continuum absorption increased with the time that coolant was in the cell. The absorption spectra of the samples were obtained by using a vacuum grating-spectrograph of 2-meter focal length. The grating has 1200 grooves/mm, providing a reciprocal dispersion at the focal plane of about 4.2 A/mm. Spectra were obtained both photographically and photoelectrically. Photographs with relatively narrow slits (10 to 20 microns) showed that wider slits could be used with no loss of information. Thus in the photoelectric measurements, entrance and exit slits of 100 microns (6 cm-l at 2600 A, 12 cm-’ at 1800 A) were used. The detector was an EM1 62563 photomultiplier tube with a fused silica window. The background continuum was provided by a water-cooled hydrogen discharge lamp operated at about 1 kw and a pressure of 14 mm Hg. It was found that the continuum intensity was enhanced if the hydrogen gas flowed through liquid nitrogen-cooled charcoal traps. The hydrogen lamp radiation was collimated through the sample chamber by means of a lithium fluoride lens. A thin quartz plate before the lens prevented the discoloration of the lithium fluoride through formation of color centers by the far ultraviolet radiation from the lamp. III. RESULTS
AND DISCUSSION
In the wavelength range 1750 to 2000 A the Schumann-Runge bands of oxygen (B 3Z,- t X 3&-) were observed in liquid and solid media ranging in temperature from about 60 to 90°K. The bands could be observed over a range of concentrations of about 0.01% to 3 ‘%. At higher concentrations there was total absorption in this region and no band structure could be observed. A typical absorption curve is shown in Fig. 2(c) with comparison spectra of gas phase absorption at 90°K and at 295°K. These spectra were recorded with the same slit widths of 100 microns so that the resolution conditions are the same for all. The broadening of the bands and the disappearance of the rotational structure in solid argon are apparent. Furthermore, the bands in the solid appear to be degraded slightly toward shorter wavelengths in contrast to the red-shading of the gas phase bands. In Table I are collected representative measurements of the positions of absorption maxima of the Schumann-Runge bands. The vibrational numbering of the bands (u’, v”) follows that suggested by Dressler and Schnepp (6, 7) and is based upon their work using Of”. The values of relative intensity in Table I represent the fractional absorption below the baseline which results from the general continuous background. The half-intensity widths of the bands range from about 150 to about 550 cm-l,
ABSORPTION
SPECTRA
22; i
OF OXYGEIY
l/l/llllllIIiIlllIIlll~ll 2000
WAVELENGTH FIG. 2. Schumann-Runge
1800
1900
(i,
vacuum)
absorption hands of clxygen (Lc %,;cX 3X,,-) Iwrature, 760 mm Hg pressure. (h) gOoK, 760 mm Hg pressure. (c) Oxygen argon, 77°K; dashed curve is the transmission of the empty sittttple cell.
(a) 11oot~1 tczrtl(0.25:;)
in solid
aud seem to be substantially greater than the widths observed by Dressier (6) at 42°K. Both the half-widths and the peak intensities tend to increase aud reach a maximum value in the neighborhood of U’ = 8-10. Similar behavior is report,ed by Hethke (9) for the absolute integrated absorption coefficients of
52748.2
53302.1
53803.9
54303.6
1938.7
1916.8
1895.8
1876.1
1858.6
1841.5
8
18
38
57
67
85
102
106
79
51
630
790
990
10,o
ll,o
12,0
13,0
*
55567.9
55928.4
1788.0
55187.6
1799.6
1812.0
54770.5
51581.0
1784.3
1796.5
1808.3
1823.6
1839.5
1857.9
1876.2
1895.2
1917.6
1939.3
1964.2
5@w+.4
55663.8
55300.6
54836.6
54362.6
53824.2
53299.2
52764.9
52148.5
51565.0
50911.3
Solid Argon 760 K A vat
1936.4
1961.2
Avac
1786.8
1797.5
1810.1
1824.3
1839.2
1855.7
1874.5
1894.5
55966.0
55632.8
55245.6
54815.5
54371.4
53888.0
53347.6
52784.4
52261.1
51642.2
50989.2
cmZ.c
Liquid Nitrogen
1914.2
L. Wallace,Astrophys.J., Suppl. No. 68, 2, 165 (1962).
830
570
490
1825.8
52170.3
1962.3
4
3,O
50960.6
(1985.6)(50362.6)
1
290
v', vn
Liquid Argon 85O K crniic Avac
Relative Intensity
1875.0
1893.7
1915.3
1937.9
1961.9
Avac
53333.3
52806.7
52211.1
51602.2
50971.0
cm;;,
Solid Nitrogen
1882.4
1902.5
1924.2
1947.3
1971.9
A;;:”
1782.9
1792.5
1803.7
1816.4
1830.7
1846.4
56088.4
55788.0
55441.6
55054.0
54623.9
54159.4
53659.6
53123.7
52562.4
51969.6
51353.2
50712.5
““$i,
Gas Phase*
1863.6
TabIe I. Absorptionmaxima of the Schumann-RungeBands of 02 in liquid and solid argon and nitrogen
the Schumam-Bunge band of 02 in the gas phase, and is attributed by him to the effect, of overlapping of the bands for large u’. Over the range of conditions covered in our work the half-widths appear to he somewhat smaller in the solid than in the liquid, with the maximum difference being of the order of 30 ‘; Wavelength measurements of the absorption maxima in Table I are uncertaitl to about, 1 L$because of the width of the bands. There are small hut real diffrrencw in band positions in licluid argon, solid argon, and liquid or solid nitrogen. C:omparisons of the positions of the absorption maxima in the condensed gases with the position of the gas phase band heads show shifts toward lower frequencies (longer wa\Ttlengths) from 30 to 400 cm?. The shift appears also to increase linearly with temperature with a coefficient of roughly 2 cm-’ 1°K in the rang{’ 60 to 90°K. The wavelength shift of the absorption tnaxima in condensed phases with respect to the gas-phase band heads, is large for small values of P’ (appros. 400 cn~’ ) and decreases slowly as 1’’ increases. The decrease in the shift is somewhat more rapid for oxygen in solid argon than in liquid argon, and more rapid ill liquid argon than in licluid nitrogen. In the wa\-elength range 2100 to 3000 A a number of absorption bands were ohser\wl in pure liquid oxygen as well as in condensed mixtures of oxygen in argon 01 nitrogen. In contrast to the Schumann--Runge system, the longer wal-elcngth hands could he observed only in concentrations greater than a few percent ( in the %-mm path length used). The hands are broad (200 to 600 cm- ’ ) and show some partially resolved structure. Fig. 3 shows typical traces in this region. 13~ way of comparison, for sample (h) with 4 Y’; oxygen, the regiotl of the Sclllmlanrl-Itunge bands, 1750 t,o 2000 A, was completely absorbed, while for sample ( a’), with 1.6 “; oxygen, the Schumam---Runge hands could tw (Aset,\-ed superimposed on a very strong continuous background. Table II lists the absorption maxima observed in this region. Because the hands are quit,c broad the positions of the maxima are not as well known as in the case of t,tw Sch~lmann-Rmlge bands, and the uncertainty in the wavelength of the maxima may he as large as 2 w. Ko differences were observed for changes either in t,cmperaturc or matrix material. Alt~hough our instrumental resolving power was adequate for resolution of the triplets as observed by T’rikhotko (31, the bands observed in the 60 to !1O”Ii range were not fully resell-ed. It is possible that the lower temperature at which l’rikhotko made measurements (--’ 20%) , is necessary to produce bands narrow enough for complete separation. However, H&l’s spectra at 4.2”K show only broad unresolved bands, indicating that the temperature is not the only factor that cont,rols t’his aspect of the spectrum. The measurements of both H&l (\S) and I’rikhotko suggest that the bands in the 2400 to 2900 8 region should he correlated with the Herzherg system (A 3Z,,+ c- 9 “Z,,-) of OZ. The triplet, structure of the bands in condensed media--which is identical to that of the “high-pressure” bands-is presumed to arise from splitting of the ‘z,,- ground
228
BASS AND BROIDA
/
I
I
I
/
I
I
TI
I
I I
I 2900
I
I 2000
I
I 2700
WAVELENGTH
I
I
2600
I 2500
I
I 2400
(A. air)
FIG. 3. Absorption bands of oxygen in the near ultraviolet. (a) Oxygen (1.6%) in liquid argon, 84°K. (b) Oxygen (4%) in liquid argon, 84°K. (c) Liquid oxygen, 84°K. by the perturbing influence of the matrix molecules. The “high-pressure” bands have sometimes been attributed to 04, based in part on their occurrence in systems having a high density of oxygen. Since what appears to be the same spectrum is observed for concentrations of oxygen in liquid argon of as little as state
2683.O
36224
36486 36637 36754
37171 37386
2759.8
2740.4 2728.7 i2720.0
2689.5 I 2674.0
2642.7 2632.5
6
15 18 17
;I:
88 78
40180 40650
2522.2 2488.0 2459.3
148
75
2520.6
2553.5
2588.9
Aair
39661
39150
38615
-1 vat
Clll
O2 in Solid Nitrogen
L. Wallace, Astrophys. 3. Suppl. No. 68, 2, 165 (1962).
39636
39004 39118
2563.2 2555.6
202
38432 38576
2601.2 2591.5
37803 37983
3'7261
36414
38186
V8.C
cm-1
156
37829 37975 2644.5 2632.0
2745.4
35722 35922
2798.6 2783.0
14 11 2762.7
34980
2857.9
7
34239
2919.8
A air
2
vat
-1
33857
Cm
02 in Liquid Argon
2952.7
A air
Liquid Oxygen 56 to 90" K
1
Relative IntensiLy vv
990 10,O
890
770
690
590
40
390
290
1.0
090
V')
air
39681 40167 40585 0920
2519.3
2463.3 2443.1
39138
38546
37910
37325
36526
35780
35008
-1 cm vat
2488.8
2554.2
2593.5
2637.0
2684.9
2737.1
2794.0
2855.9
A
Herzberg Bands (gas phase)*
Table II. Absorptionmaxima of oxygen in liquid oxygen, liquid argon, and solid nitrogen in the region 2400-3000 A.
230
BASS AND
BROIDA
2 Tj, as for liquid oxygen, the assignment of these bands to O4 , rather than to 0, , is open to question. RECEIVED:
May 6, 1968 REFERENCES
1. J. G. MCLENNAN, H. D. SMITH, AND J. 0. WILHELM, Trans. Roy. Sot. Can. 84, 1 (1930). 2. W. FINKELNBERG AND W. STEINER, 2. Physik 79, 69 (1932). 3. A. PRIKHOTKO, Acta Physicochim. URSS 10, 913 (1939). 4. J. ROMANI) AND J. GRANIER-MAYENCE, J. Phys. Radium 16, 62 (1954). 5. E. HURL, J. Mol. Spectroscopy 2, 42 (1958). 6. K. DRESSLER, J. Quant. Sped. Rad. Trans. 2, 683 (1962). i’. Ii. DRESSLER AND 0. SCHNEPP (unpublished work). 8. I,. H. BOLZ, H. P. BROIDA, AND H. S. PEISER, Ac:tn Crystalloy. 9. G. W.
BETHKE, J. Chenz. Phys.
31, 669 (1959).
16, 810 (19G2).
OF MOLECULAR
Vacuum
SPECTROSCOPY
Ultraviolet
Absorption
and Crystalline
Spectra
Argon
14. BASS AKD
ARNOLD
Kational
221-230 (1964)
1%
B~ureau of Standards,
of Oxygen
in Liquid
and Nitrogen* H. P. BROIDA Washington,
D. C.
The ultraviolet absorption spectra of oxygen have been observed in liquid and solid media at low temperature. By freezing slowly from the liquid phase, samples of crystalline argon and nitrogen were prepared with added amounts of oxygen ranging between 0.01% and 10yC.The spectra were recorded photoelectrically, through fused silica windows, over the wavelength range from 1750 to 2900 b. The Schumann-Runge bands of oxygen (B Q&m+ S 2,, 1750 to 2000 d) were observed in solid and liquid mixtures of oxygen in argon or r&rogen at concentrations less than 5%, over the temperature range 60 to 90%. The bands are broad, with half-intensity width of 150 to 550 cm-1 and show no structure. The bands show a slight shading toward the violet in contrast to the bands in t.he gas phase, and are displaced toward lower frequencies by 50 to 400 cn-1 as compared with the positions in the gas phase. These shifts appear to he slightly dependent upon temperature. The changes in position or shape of the bands between liquid and solid or between argon and nitrogen media are remarkably small. At concentration of oxygen above 5y0 the Herzberg hands of oxygen (A Q,+ c S %,-) were observed in the region 2400 to 3000 A.
These hands are also broad (200 to 600 cm-l) and show some partially resolved structure. No differences were observed for changes either in temperature ot rrlatrix material. I. INTRODUCTION
The ultraviolet
absorption
of oxygen
densed phases has been studied previously. studied the spectrum a number of bands
in the region 2400 to 3000 A in conMcLennan,
Smith, atd Wilhelm
(1)
of liquid oxygen in the region 2400 to 3000 A and observed corresponding to the “high-pressure” bands in gaseous
oxygen reported by Finkelnberg and Steiner (2). Prikhotko (3) studied the absorption spectrum of oxygen in crystals of solid solutions of Or-N2 and Oe-Ar at liquid hydrogen temperature. In the region 2400 to 2800 8 she found that the spectra of all the mixtures consist of a series of triplet absorption bands. The triplet group spacing decreased toward shorter wavelengths from about 700 cm-’ to about 3.50 cm-‘. These bands coincide with the absorption spectra of solid oxygen, and have a structure identical with the “high-pressure” bands. * Work support)ed in part by the National Aeronautirs and Space Administration. 221
222
BASS AKD
BROIDA
From her data she calculated a potential energy curve for the upper state of the transition which is very similar to the A 3Z,,+ state of the Herzberg bands. Romand and Granier-Mayence (4) studied the spectrum of solid oxygen (condensed from the gas phase on a surface at 20°K) in the region from 1500 to 2000 i and observed a continuous absorption, increasing in strength toward shorter waveleng:hs. They did not observe any discrete absorption in the region 2000 to 2400 A. H&l (5) prepared thin polycrystalline layers of solid a-oxygen by deposition from the gas phase onto surfaces cooled to near 4.2”K by liquid helium with film thicknesses estimated to be a few hundredths of a millimeter. Between 2400 and 2700 i a number of broad and diffuse bands were observed with half-widths of the order of 200 to 300 cm-‘. Horl suggested that the bands in the solid are the Herzberg bands of oxygen, (A ‘S,+ +- X 3Z,-) shifted to shorter wavelengths by about 90 to 120 cm-‘. Dressler and Schnepp (6, 7) studied thin layers of oxygen deposited in mixtures with argon and nitrogen on a surface at 4.2’K. Between 1750 and 2000 A eleven bands, 50 to 150 cm-’ broad, were observed and identified as the Schumann-Runge bands of oxygen (B 32,- +X “2,). By measuring the isotope shift produced with 0:” they were able to establish that the bands in the solid are shifted toward longer wavelengths by roughly 300 to 400 cm-‘. In the present work absorption spectra of oxygen inocondensed phases have been recorded over the wavelength range 1750 to 2900 A. Both the SchumannRunge bands of O2 and the diffuse oxygen bands in the region of 2400 to 3000 A have been studied. Observations of the oxygen have been made in liquid and crystalline argon and nitrogen at temperatures between 60 and 90°K. Some small variations with change of parameters were observed. However the surprising result has been that the observed spectra show little change between the solid and liquid matrix and are quite similar to the spectra observed at 4.2”K. II. EXPERIMENTAL
Solid crystalline samples were produced in a glass Dewar vessel (Fig. l), by the method described by Bole et al. (8). Argon or nitrogen gas, with varying concentrations of oxygen, was admitted to the central chamber at a pressure of 1 atmosphere. When liquid coolant (nitrogen or oxygen) was added slowly to the surrounding compartment it was possible to produce, between the two fused silica windows (25mm path length), a transparent solid. Solid argon was grown at temperatures from 60 to 82°K. This solid was similar to that described by Bolz et al. (8)) in which they measured grain-sizes of up to 4 mm. By pumping on liquid oxygen, the temperature of the coolant bath could be adjusted as desired in the range 60 to 90°K. Thus the same sample could be observed in the liquid phase as well as in the solid. The concentration of oxygen in the sample was controlled by adjusting the ratio of oxygen to argon (or nitrogen) in the gas prior to condensation. The
ABSORPTION
GAS---W
SPECTRA
OF OSY(;EK
REFRIGERANT
VACUUM
I
EN
FIG. 1. Dewar sropic studies.
vessel
for
prrpttrutionof crystalline
solids of srgtm suitable
for spwtro-
concentrations used in this way ranged from about 0.01 5’ to 10 L% (and also 100 “h oxygen). However, the concentration of oxygen in the condensed mat,erial was probably different from that in the input gas. This was verified in one case where mass spectrometric analysis indicated the “input” gas composition to be 10.3 7 Or , 89.7 7:’ ss ) while the composition of the gas eraporating from the condensed sample to be 4.3 c/? O2 , 95.6 “4’. The temperature of the sample was assumed to be the same as that of t’he surrounding bath, as determined from the vapor pressure of the bath refrigerant. Boiling points and triple points of the sample obtained in this way generally agree with the known values of these fixed points. However, the assumption that, the sample and the surrounding bath are at same temperature is not always reliable. Occasionally we have found that, even with vapor pressure of the bath corresponding to a temperature well below that of the triple point of argon (83.78”K~‘i, a sample of argon in the inner chamber will remain liquid for some
224
BASS AND BROIDA
time. Thermocouple measurements of the liquid argon show that the liquid is not supercooled. We have not been able to explain this lack of thermal equilibrium, but it does cause some uncertainty as to the accuracy of our knowledge of the sample temperature. A side-arm cooled by liquid hydrogen was used to trap out any substances which might otherwise condense on the outside of the fused silica windows of the sample chamber. Prior to the use of this trap, in the region below 1900 8, a continuum absorption increased with the time that coolant was in the cell. The absorption spectra of the samples were obtained by using a vacuum grating-spectrograph of 2-meter focal length. The grating has 1200 grooves/mm, providing a reciprocal dispersion at the focal plane of about 4.2 A/mm. Spectra were obtained both photographically and photoelectrically. Photographs with relatively narrow slits (10 to 20 microns) showed that wider slits could be used with no loss of information. Thus in the photoelectric measurements, entrance and exit slits of 100 microns (6 cm-l at 2600 A, 12 cm-’ at 1800 A) were used. The detector was an EM1 62563 photomultiplier tube with a fused silica window. The background continuum was provided by a water-cooled hydrogen discharge lamp operated at about 1 kw and a pressure of 14 mm Hg. It was found that the continuum intensity was enhanced if the hydrogen gas flowed through liquid nitrogen-cooled charcoal traps. The hydrogen lamp radiation was collimated through the sample chamber by means of a lithium fluoride lens. A thin quartz plate before the lens prevented the discoloration of the lithium fluoride through formation of color centers by the far ultraviolet radiation from the lamp. III. RESULTS
AND DISCUSSION
In the wavelength range 1750 to 2000 A the Schumann-Runge bands of oxygen (B 3Z,- t X 3&-) were observed in liquid and solid media ranging in temperature from about 60 to 90°K. The bands could be observed over a range of concentrations of about 0.01% to 3 ‘%. At higher concentrations there was total absorption in this region and no band structure could be observed. A typical absorption curve is shown in Fig. 2(c) with comparison spectra of gas phase absorption at 90°K and at 295°K. These spectra were recorded with the same slit widths of 100 microns so that the resolution conditions are the same for all. The broadening of the bands and the disappearance of the rotational structure in solid argon are apparent. Furthermore, the bands in the solid appear to be degraded slightly toward shorter wavelengths in contrast to the red-shading of the gas phase bands. In Table I are collected representative measurements of the positions of absorption maxima of the Schumann-Runge bands. The vibrational numbering of the bands (u’, v”) follows that suggested by Dressler and Schnepp (6, 7) and is based upon their work using Of”. The values of relative intensity in Table I represent the fractional absorption below the baseline which results from the general continuous background. The half-intensity widths of the bands range from about 150 to about 550 cm-l,
ABSORPTION
SPECTRA
22; i
OF OXYGEIY
l/l/llllllIIiIlllIIlll~ll 2000
WAVELENGTH FIG. 2. Schumann-Runge
1800
1900
(i,
vacuum)
absorption hands of clxygen (Lc %,;cX 3X,,-) Iwrature, 760 mm Hg pressure. (h) gOoK, 760 mm Hg pressure. (c) Oxygen argon, 77°K; dashed curve is the transmission of the empty sittttple cell.
(a) 11oot~1 tczrtl(0.25:;)
in solid
aud seem to be substantially greater than the widths observed by Dressier (6) at 42°K. Both the half-widths and the peak intensities tend to increase aud reach a maximum value in the neighborhood of U’ = 8-10. Similar behavior is report,ed by Hethke (9) for the absolute integrated absorption coefficients of
52748.2
53302.1
53803.9
54303.6
1938.7
1916.8
1895.8
1876.1
1858.6
1841.5
8
18
38
57
67
85
102
106
79
51
630
790
990
10,o
ll,o
12,0
13,0
*
55567.9
55928.4
1788.0
55187.6
1799.6
1812.0
54770.5
51581.0
1784.3
1796.5
1808.3
1823.6
1839.5
1857.9
1876.2
1895.2
1917.6
1939.3
1964.2
5@w+.4
55663.8
55300.6
54836.6
54362.6
53824.2
53299.2
52764.9
52148.5
51565.0
50911.3
Solid Argon 760 K A vat
1936.4
1961.2
Avac
1786.8
1797.5
1810.1
1824.3
1839.2
1855.7
1874.5
1894.5
55966.0
55632.8
55245.6
54815.5
54371.4
53888.0
53347.6
52784.4
52261.1
51642.2
50989.2
cmZ.c
Liquid Nitrogen
1914.2
L. Wallace,Astrophys.J., Suppl. No. 68, 2, 165 (1962).
830
570
490
1825.8
52170.3
1962.3
4
3,O
50960.6
(1985.6)(50362.6)
1
290
v', vn
Liquid Argon 85O K crniic Avac
Relative Intensity
1875.0
1893.7
1915.3
1937.9
1961.9
Avac
53333.3
52806.7
52211.1
51602.2
50971.0
cm;;,
Solid Nitrogen
1882.4
1902.5
1924.2
1947.3
1971.9
A;;:”
1782.9
1792.5
1803.7
1816.4
1830.7
1846.4
56088.4
55788.0
55441.6
55054.0
54623.9
54159.4
53659.6
53123.7
52562.4
51969.6
51353.2
50712.5
““$i,
Gas Phase*
1863.6
TabIe I. Absorptionmaxima of the Schumann-RungeBands of 02 in liquid and solid argon and nitrogen
the Schumam-Bunge band of 02 in the gas phase, and is attributed by him to the effect, of overlapping of the bands for large u’. Over the range of conditions covered in our work the half-widths appear to he somewhat smaller in the solid than in the liquid, with the maximum difference being of the order of 30 ‘; Wavelength measurements of the absorption maxima in Table I are uncertaitl to about, 1 L$because of the width of the bands. There are small hut real diffrrencw in band positions in licluid argon, solid argon, and liquid or solid nitrogen. C:omparisons of the positions of the absorption maxima in the condensed gases with the position of the gas phase band heads show shifts toward lower frequencies (longer wa\Ttlengths) from 30 to 400 cm?. The shift appears also to increase linearly with temperature with a coefficient of roughly 2 cm-’ 1°K in the rang{’ 60 to 90°K. The wavelength shift of the absorption tnaxima in condensed phases with respect to the gas-phase band heads, is large for small values of P’ (appros. 400 cn~’ ) and decreases slowly as 1’’ increases. The decrease in the shift is somewhat more rapid for oxygen in solid argon than in liquid argon, and more rapid ill liquid argon than in licluid nitrogen. In the wa\-elength range 2100 to 3000 A a number of absorption bands were ohser\wl in pure liquid oxygen as well as in condensed mixtures of oxygen in argon 01 nitrogen. In contrast to the Schumann--Runge system, the longer wal-elcngth hands could he observed only in concentrations greater than a few percent ( in the %-mm path length used). The hands are broad (200 to 600 cm- ’ ) and show some partially resolved structure. Fig. 3 shows typical traces in this region. 13~ way of comparison, for sample (h) with 4 Y’; oxygen, the regiotl of the Sclllmlanrl-Itunge bands, 1750 t,o 2000 A, was completely absorbed, while for sample ( a’), with 1.6 “; oxygen, the Schumam---Runge hands could tw (Aset,\-ed superimposed on a very strong continuous background. Table II lists the absorption maxima observed in this region. Because the hands are quit,c broad the positions of the maxima are not as well known as in the case of t,tw Sch~lmann-Rmlge bands, and the uncertainty in the wavelength of the maxima may he as large as 2 w. Ko differences were observed for changes either in t,cmperaturc or matrix material. Alt~hough our instrumental resolving power was adequate for resolution of the triplets as observed by T’rikhotko (31, the bands observed in the 60 to !1O”Ii range were not fully resell-ed. It is possible that the lower temperature at which l’rikhotko made measurements (--’ 20%) , is necessary to produce bands narrow enough for complete separation. However, H&l’s spectra at 4.2”K show only broad unresolved bands, indicating that the temperature is not the only factor that cont,rols t’his aspect of the spectrum. The measurements of both H&l (\S) and I’rikhotko suggest that the bands in the 2400 to 2900 8 region should he correlated with the Herzherg system (A 3Z,,+ c- 9 “Z,,-) of OZ. The triplet, structure of the bands in condensed media--which is identical to that of the “high-pressure” bands-is presumed to arise from splitting of the ‘z,,- ground
228
BASS AND BROIDA
/
I
I
I
/
I
I
TI
I
I I
I 2900
I
I 2000
I
I 2700
WAVELENGTH
I
I
2600
I 2500
I
I 2400
(A. air)
FIG. 3. Absorption bands of oxygen in the near ultraviolet. (a) Oxygen (1.6%) in liquid argon, 84°K. (b) Oxygen (4%) in liquid argon, 84°K. (c) Liquid oxygen, 84°K. by the perturbing influence of the matrix molecules. The “high-pressure” bands have sometimes been attributed to 04, based in part on their occurrence in systems having a high density of oxygen. Since what appears to be the same spectrum is observed for concentrations of oxygen in liquid argon of as little as state
2683.O
36224
36486 36637 36754
37171 37386
2759.8
2740.4 2728.7 i2720.0
2689.5 I 2674.0
2642.7 2632.5
6
15 18 17
;I:
88 78
40180 40650
2522.2 2488.0 2459.3
148
75
2520.6
2553.5
2588.9
Aair
39661
39150
38615
-1 vat
Clll
O2 in Solid Nitrogen
L. Wallace, Astrophys. 3. Suppl. No. 68, 2, 165 (1962).
39636
39004 39118
2563.2 2555.6
202
38432 38576
2601.2 2591.5
37803 37983
3'7261
36414
38186
V8.C
cm-1
156
37829 37975 2644.5 2632.0
2745.4
35722 35922
2798.6 2783.0
14 11 2762.7
34980
2857.9
7
34239
2919.8
A air
2
vat
-1
33857
Cm
02 in Liquid Argon
2952.7
A air
Liquid Oxygen 56 to 90" K
1
Relative IntensiLy vv
990 10,O
890
770
690
590
40
390
290
1.0
090
V')
air
39681 40167 40585 0920
2519.3
2463.3 2443.1
39138
38546
37910
37325
36526
35780
35008
-1 cm vat
2488.8
2554.2
2593.5
2637.0
2684.9
2737.1
2794.0
2855.9
A
Herzberg Bands (gas phase)*
Table II. Absorptionmaxima of oxygen in liquid oxygen, liquid argon, and solid nitrogen in the region 2400-3000 A.
230
BASS AND
BROIDA
2 Tj, as for liquid oxygen, the assignment of these bands to O4 , rather than to 0, , is open to question. RECEIVED:
May 6, 1968 REFERENCES
1. J. G. MCLENNAN, H. D. SMITH, AND J. 0. WILHELM, Trans. Roy. Sot. Can. 84, 1 (1930). 2. W. FINKELNBERG AND W. STEINER, 2. Physik 79, 69 (1932). 3. A. PRIKHOTKO, Acta Physicochim. URSS 10, 913 (1939). 4. J. ROMANI) AND J. GRANIER-MAYENCE, J. Phys. Radium 16, 62 (1954). 5. E. HURL, J. Mol. Spectroscopy 2, 42 (1958). 6. K. DRESSLER, J. Quant. Sped. Rad. Trans. 2, 683 (1962). i’. Ii. DRESSLER AND 0. SCHNEPP (unpublished work). 8. I,. H. BOLZ, H. P. BROIDA, AND H. S. PEISER, Ac:tn Crystalloy. 9. G. W.
BETHKE, J. Chenz. Phys.
31, 669 (1959).
16, 810 (19G2).