674
Anal. Chem. 1802, 54, 674-677
(4) Flte, W. L.; Irvlng, P. J. Vac. Sc/. Techno/. 1074, 11, 351-356. (5) Sweely, C. C.; Elliot, W. H.; Fries, I.; Ryhage, R. Ana/. Chem. 1066, 38, 1549-1553. (6) Jeltsch, E.; Graf, W. KFA Julich Report No. 1007-RX, 1973. (7) Nlshl, I.; Sugal, S.; Tanaka, K.; Tomizawa, G. Mass Specfrosc. 1076, 24, 07-105. (6) Flucklnger, R.; Schalcher, M. 7th International Mass Spectrometer Conference, Florence, Italy. 1076. (9) Klenltz, K. Z.Anal. Chem: 1058, 164, 60-00. (IO) Dobrozemskl, R.; Farber, W. Vak.-Tech. 1071, 6,231-239. (11) Raimondi, D. L.; Winters, H. F.; &ant. P. M.; Clarke, D. c. IBM J . Res. Dev. 1071, 15, 307-312.
(12) Tunnlcliff, D. D.; Wadsworth, P. A. Anal. Chem. 1065, 37, 1082-1065. (13) Margenau, H.; Murphy, 0. M. “The Mathematics of Physics and Chemlstry”, 2nd ed.; D. van Nostrand Co.: Princeton, NJ, 1956.
RECEIVED for review October 22,1981. Accepted December 21, 1981. Financial support for this work was received from the “Ehdesminister fiir Forschung und Technohie” under Contract No. 02 U 5090.
Characterization of Poly(carb0xypiperazine) by Mass Analyzed Ion Kinetic Energy Spectrometry Salvatore Fotl,’ Angelo Llguorl,* Pletro Maravigna,‘ and Glorglo Montaudo”’ Istltuto Dipartimentale di Chimica e Chimica Industriale, Universita di Catanla, 95 125 Catania, Italy, and Dipartimento di Chimica, Universlta della Calabrla. Arcavacata dl Rende (Cosenza), Italy
Thls work Is concerned wlth a study of poly(carboxyplperazlne) by dlrect pyrolysls-mass spectrometry. The Identity of a key compound In the mlxture of thermal ollgomers orlglnatlng in the polymer pyrolysis was demonstrated by mass analyzed Ion kinetlc energy spectrometry. The mass spectral data allow one to assess that the pyrolytic breakdown of thls polyurea occurs through a slngie-stage decomposltlon mechanlsm that leads to fragments wlth amlno end groups and carbon monoxlde. This appllcatlon of MIKE spectrometry Illustrates the potentlalltles of this technique In the analysis of mlxtures obtalned by direct pyrolysls of polymers in the mass spectrometer.
Although mass spectrometry (MS) is considered an essential technique to elucidate the structure of low molecular weight organic and inorganic compounds, it has been much less used in the case of polymers. This constitutes a relevant difference with respect to other widely applied spectrometric techniques, such as IR and NMR, whose respective importance in the structure elucidation of polymers is similar to that for low molecular weight compounds. It appears evident that, since MS techniques require transfer of the sample in the gas phase, the low volatility of macromolecules has constituted a serious drawback to the application of direct mass spectrometric analysis to polymer systems. In the direct pyrolysis-mass spectrometry technique (1-3), polymers are introduced via the direct insertion probe and the temperature is increased gradually up to a point at which thermal degradation reactions occur; the volatile oligomers formed are then ionized and detected. The mass spectrum of a polymer obtained in these conditions is therefore that of the mixture of oligomers formed by pyrolysis. A general advantage of this technique is that pyrolysis is accomplished under high vacuum, and therefore the thermal oligomers formed are volatilized and removed readily from the hot zone. This, together with the low probability of Universitg di Catania. Universiti della Calabria. 0003-2700/82/0354-0674$01.25/0
molecular collision and fast detection reduces to a great extent the occurrence of secondary reactions, so that almost exclusively primary fragments are detected. Consequently, the informationthus obtained is of particular importance in order to assess the primary thermal degradation mechanism of a polymer. Furthermore, since pyrolysis is achieved very close to the ion source and no problem of transport exists, fragments of high mass, which are often essential for the structural characterization of the polymer, can be detected, whereas they are often lost using other techniques. The main problem connected with this technique is, however, the identification of the products in the spectrum of the multicomponent mixture produced by thermal degradation. In fact, in the overall end spectrum of a polymer, the molecular ions of the thermal formed oligomers will appear mixed with the fragment ions formed in the ionization step. In some instances, identification of thermal degradation products can be achieved by using soft ionization methods, by using exact mass measurements, and by matching spectra of authentic samples with those obtained from the polymer (1). A technique which appears attractive for the direct mass spectral analysis of mixtures if that of comparing mass analyzed ion kinetic energy (MIKE) spectra of selected ions in the spectrum of the mixture with MIKE spectra of reference compounds (4). Although it has been pointed out that some care must be exerted in the evaluation of such data (5),the method has proved to be valuable in several cases (4). Applications of the method to the elucidation of mass spectra of mixtures obtained by direct pyrolysis-mass spectrometry of synthetic polymers have not appeared in the literature to date. In the following, we report the utilization of MIKE spectra in the mass spectral characterization of poly(carboxypiperazine). EXPERIMENTAL SECTION Synthesis. l-4-Bis(piperazinocarbonyl)piperazine was prepared from l-4-bis(chlorocarbonyl)piperazine with an excess of piperazine (1:lO) in benzene at room temperature with stirring. After 30 min the solvent was distilled out at reduced pressure, piperazine and piperazine hydrochloride were extracted with water, and the residue was recrystallized from methylene chloride/hexane, mp 210 O C , according to the literature (6). @ 1982 Amerlcan Chemlcal Society
ANALYTICAL CHEMISTRY, VOL. 54, NO. 4, APRIL 1962
675
422
550
450
350
36O0C
L
846 I L U
.OL
850
mn
70ev
b
x5
113
I
I'
20
50
Flgure 1.
250
150
Mass spectra of poly(carboxypiperazine)at 360 OC and 16 eV (a) and at 360
OC
and 70 eV (b).
Poly(carboxypiperazine)was prepared by dispersion polycondensation from piperazine and phosgene: yield 55%, mp >350
"C. Direct Pyrolysis. Mass spectra of poly(carb0xypiperazine) were obtained on an LKB 9000 S mass spectrometer operated under the following conditions: electron energy 70 or 16 eV; accelerating voltage 3.5 kV; ion source temperature 310 "C; trap current 60 wA. A polymer sample was introduced via the direct insertion probe and heated from room temperature to 390 O C at about 10 OC/min. As soon as the thermal degradation of the polymer began, mass spectra were repetitively scanned, following the technique deI m spectra were obtained on a Varian scribed elsewhere (1-3). M Mat CH5 mass spectrometer with reverse Nier-Johnson geometry. RESULTS AND DISCUSSION Poly(carboxypiperazine) starts originating thermal oligomers a t probe temperature above 270 O C . Repetitive mass spectra recorded at increasing temperature show that the mass spectral pattern remains almost unmodified in the temperature range in which formation of thermal fragments occurs.
Flgure 2. Thermal degradation mechanism for poly(carboxypiperaz1ne).
A typical spectrum recorded at 360 "C is reported in Figure la. In order to minimize the electron impact induced fragmentation, we recorded the spectrum a t a reduced electron energy of 16 eV. Despite of this, a large number of intense peaks appear in the spectrum up to about 700 mlz, making it difficult to identify the oligomers formed in the thermal process without gathering some additional evidence. When compared with the spectrum recorded a t 70 eV (Figure lb), the low electron energy mass spectrum shows a sharp increase in the relative intensities of peaks corresponding to m / z values of 84 -t 112n and 86 + 112n, with n = 0,1, 2, 3, .... This increment of intensity is a strong indication that these series of peaks correspond to molecular ions of oligomers formed in the thermal degradation process.
676
ANALYTICAL CHEMISTRY, VOL. 54, NO. 4, APRIL 1982
120 OC
254
16 * V
60113
199
267
84
213
142 130
69 20
242 166
~
1
I
169
JI
1,
L h
tc,
50
LL.
11
I .
1
1
.
250
150
-
mir
Flgure 3. Mass spectrum (120 OC, 16 eV) of compound I .
Table I
a
1
o
n R
H-N
m/z ( n ) : 86 (0); 198 (1);310 ( 2 ) ; 422 (3); 534 (4); 646 (5)
m/z ( n ) : 84 (0); 196 (1);308 (2); 420 (3); 532 (4); 644 ( 5 ) I10
280
The structures that can be tentatively assigned to these series of oligomers are reported in Table I. This implies that the pyrolytic breakdown of our polyurea occurs through the single-stage decomposition mechanism shown in Figure 2. It can be noted that this scheme predicts the formation of thermal oligomers with two types of amino end groups, by hydrogen transfer from a piperazine methylene to the nitrogen. Of course, the identity of the thermal oligomers formed is to be proved unambigously, in order to substantiate the proposed mechanism. Therefore compound I corresponding to one of the expected thermal oligomers was synthetized and its E1 and MIKE spectra were obtained.
-n
H N U N
t
-"-NUN
n - C-NUN-
b
I
I
300
340 2bo
H
I
,
420
380 I
7
I
I
250
I
I
I
460 1
I
500 E
1 3 0 0
m z/
Flgure 4. Comparison of MIKE spectrum of mlz 310 from compound I (a) with MIKE spectrum of m l r 310 from poly(carb0xypiperazine)
The E1 mass spectrum of compound I is reported in Figure 3. It shows a strong molecular ion at m / z 310. The main fragmentation pathway is the loss of a neutral fragment of m/z 56 from the molecular ion, producing an intense fragment ion at m / z 254 (base peak). Actually, loss of m / z 56 is observed in the mass spectrum of poly(carboxypiperazine) (Figure la) from peaks at m/z 198, 310,422,534, and 646. The fragmentation found in compound I allows one therefore to rationalize the intense peaks at m / z 142, 254, 366, 478, and 590 in Figure 1 as originating by E1 from the molecular ions of the homologues of compound I (Table I). The MIKE spectrum of the m/z 310 ion of compound I is compared in Figure 4 with the MIKE spectrum of the m / z 310 ion of the polymer. The two spectra appear to be essentially identical, apart from an intense narrow peak at an apparent mass number of 286 observed in the polymer spectrum. The particular slope of the latter peak suggests that it is an artifact, originated by a metastable transition in
(b).
the first field region (3),which gives fortuitously a peak at apparent mass 310. The relationship Eapp/Eo = malmbcombined with mb2/m, = 310 provides the needed variables to compute the metastable transition unambigously. Calculations shows that the artifact peak corresponds to the transition 366' 337+. It can be observed that intense ions at m / z 366 and 337 are present in the mass spectrum of the polymer (Figure la). Therefore the MIKE spectra allows one to ascertain the complete identity between the structure of the model compound I and that of the thermal oligomer which gives rise to peak at m / z 310 in the mass spectrum of the polymer, and, consequently, to confirm the structure attributed to the series of ions a t m / z 86 112n. Summarizing our results the identification of compound I among those arising from the pyrolysis of poly(carboxy-
-
+
Anal. Chem. 1982, 54, 677-682
piperazine) is a key factor in proving the thermal decomposition mechanism postulated in Figure 2. This application of MIKE spectrometry illustrates the potentialities of this technique in the analysis of mixtures obtained by direct pyrolysis of polymers in the mass spectrometer.
677
(2) Lijderwald, I.; Przybylskl, M.; Rlngsdorf, H.; Foti, S.; Montaudo, G. I n “Analytical Pyrolysls”; Jones, C. E. Roland, Cramers, Carl A,, Eds.; Elsevler: Amsterdam, 1977; p 297. (3) Balllstrerl, A.; Fotl, S.; Montaudo, G.; Pappalardo, S.; Scamporrlno, E. J . Polym. Scl., Polym. Chem. Ed. 1979, 17, 2469. (4) Kondrat, R. W.; Coaks, R. G. Anal. Chem. 1978, 50, 81 A. (5) Ast, T.; Bozorgradeh, M. H.; Wlchers, J. L.; Beynon, J. H.; Brenton, A. G., Org. Mass Spectrom. 1979, 14, 313. (6) Rivett. D. E.; Wllshlre, J. F. K. Aust. J . Chem. 1966, 19, 869.
LITERATURE CITED (1) Fotl, S.; Montaudo, 0.In “Analysls of Polymer Systems”; Bark, L. S., Allen, N. S., Eds.; Applied Science: London, 1982.
for review August 7, lg8’* Accepted November 20, 1981.
Semimicro Assay of Crystalline Phases by X-ray Powder Diffractometry Ludo K. Frevei“ and William C. Roth The Dow Corning Corporation, Mldland, Mlchigan 48640
Semlmicro gravlmetric technlques are descrlbed for preparlng a thln unlform powder layer from 2 to 10 mg of powdered sample. Equations are glven for quantltatlve X-ray powder diffractometry pertalnlng to a thin powder layer. Both the Internal standard method utlllrlng a-Al,O, and the direct method utlllring a precisely welghed sample of the crystalline phase to be assayed are fully described. Two procedures are detailed for measurlng the mass absorption coefficient of a powder specimen. Assays of identlfled crystalllne phases can be performed in the presence of unidentified phases. The methods have been applled to the assay of a-quartz In resplrabie dusts. Accuracy of assay varles from 3 % to 50% depending on the appilcablllty of the analyte standard-chosen for obtalnlng callbratlon constants. Standards tailored to the particular appiicatlon will assure meaningful and reliable assays.
Quantitative acisays of crystalline phases were initiated in the mid-1930s by making proper allowance for the attenuation of the diffracted ‘X-raysfrom identified phases. For a comprehensive review of the literature one is referred to ref 1. The bulk of the literature pertains to the assay of a-quartz in respirable dusts or in mineral samples. Even for this seemingly simple assay one encounters serious controversies (2-4). It would appear that the X-ray diffraction method, fraught as it is with variouia experimental difficulties, would not be amenable for assays of samples available only in 2-10 mg quantities. However, the authors have developed semimicro gravimetric techniques which yield quantitative data for the weight fraction of an identified crystalline phase in a mixture which may contain unidentified phases, crystalline or amorphous. THEORY The following symbols and units obtain: J, S, X = upper case subscripts designating respectively the identified crystalline phase J, the crystalline internal standard S, and the phase(s) X other than phase J; i, n = lower case subscripts designating respectively the ith diffraction peak of phase J and the nth diffraction peak of standard S; mJ = maas of phase J in a powder specimen consisting of phase J and phase(s) X, 0003-2700/82/0354-0677$01.25/0
gram; ms = mass of standard S added to weighed specimen, gram; m = mass of mixture = mJ mx + ms, gram; p ~ ps, , p m = respective mass apsorption coefficients for phase J, standard S, and mixture, cm2 g-’; pl = linear absorption coefficient, cm-l; Og, W n ~= Bragg angles for above diffraction peaks, degree; Ig, Id := integrated intensities for above peaks, arbitrary unit (only intensity ratios involved); Io = intensity of the primary beam, arbitrary unit; t = layer thickness of mixture, cm; 2r = diameter of circular cross section of layer, cm. For the conventional Bragg-Brentano parafocusing powder diffractometer irradiating a thin flat layer of mixture m uniformly distributed over the circular area rr2,the ratio of IQto Id is given by eq 1,where ksAs/kdA,s is a calibration
+
-=1,s
m,JkiJAjJ(l- e-2Prnfi COSeC 0IJ) msknsAns(l - edpnni
co8ec
I
(1)
constant determined from known binary mixtures of phase J and standard S, m = m/rr2,and A i j and Ans are the respective irradiated areas of the flat powder specimen. As seen from expression 2 the irradiated area is not necessarily constant since it depends Ion Os, r, and h, the cross-sectional height of the rectangular collimated X-ray beam in the equatorial plane of the goniometer axis. The width of the Soller slit is
greater than 2r so that the full plateau of the collimated X-ray beam bathes the spinning sample. For 02 less than sin-’ (h/2r), Ag remains constant while kiJ decreases. In practice, this is of no consequence because kiJ is unique for each line i and can be experimentally measured. However, it is convenient to replace kiJAij by a single calibration constant CiJ. If pm is determined by X-ray absorptometry (see below), then eq 1 will yield an experimental value for mJ for each pair ( i , n). The internal-standard calibration constant, CQ/C,, can also be determined from the intensities of 100% phase J and 100% standard S, provided Io,r , and t are kept constant for the two separate measuremenk. Equations 3 and 4 yield eq 5 for the calibration constant, where m; is the weight of 100% phase 0 1982 American Chemical Society